THURJ Vol. 2 Issue 1

$POHSBUVMBUJƲOTȁƺ
/PSNBO:BP
GPǚXJOƢJOł
#Ĝ˞.BOVTDSJQU
Volume 2 Issue 1 | Spring 2009 LETTERS
March 2009
Dear Harvard Community,
It is a true delight for me to introduce this issue of The Harvard Undergraduate Research Journal.
At a time when the college re-commits itself to the importance of hands-on learning and research
through the new program in General Education, it is wonderful to see undergraduates already well
advanced in their efforts as researchers. As important as books and journals are to the educational
enterprise, there is no substitute for hands-on experience, especially in the world of science. One
successful—or even unsuccessful—original experiment can bring students to a much deeper
understanding of scientific principles than can reading about or listening to descriptions of the
experiments of others. This is a position that has been repeatedly reaffirmed by many curricular
committees, and is the central conviction energizing the wonderful Program for Research in Science
and Engineering, in which the College has invested so deeply.
The Harvard Undergraduate Research Journal demonstrates that our efforts in the sciences have
combined with the remarkable skills and passion of our undergraduates to produce impressive results
that will contribute not only to the formation of the next generation of investigators and innovators,
but also to a more scientifically literate society.
Sincerely,
Jay M. Harris
www.thurj.org i
LETTERS Volume 2 Issue 1 | Spring 2009
The Harvard Undergraduate Research Journal

March, 2009
Dear Harvard Community,
We are honored to share with you Volume 2, Issue 1 of The Harvard Undergraduate Research
Journal (THURJ), a biannual publication that features the research and writing endeavors of Harvard
undergraduates. This issue has a number of unique features, and we hope you willfind reading it both
enjoyable and intellectually stimulating. The work contained in these pages showcases the talented
intellectual inquiry taking place daily on Harvard’s campus. It also emphasizes the importance of
research in the undergraduate experience. Our research articles feature subjects ranging from
neurofilament networks to priming effects as related to bingo. They represent undertakings in
biomechanics, materials science, psychology, health and medicine, and evolutionary and molecular
biology. Our prize winning manuscript, “Elasticity in ionically cross-linked neurofilament networks,”
represents in its discussion of the mechanical properties of neurofilaments the type of cross-cutting
query that both THURJ and Harvard are so proud of. We also continue in this issue our commitment
to science writing. Our writers have delved into microfluidic chips, the origins of life, and even robotic
flies. Their pieces are essential to our mission of sharing science with the larger Harvard community.
Our journal has grown tremendously in the past year. With an expanded staff of over fifty, we have
broadened our research focus into the social sciences and solidified our ties with many of the depart-
ments on Harvard’s campus. Our members have been trained by professionals from Cell, the New
England Journal of Medicine, and the National Association of Science Writers. Yet, we are constantly
looking for new partners, new directions, and new ideas.
We hope this issue provides a taste of the Harvard undergraduate community’s intellectual vigor.
We would like to thank the people who made publication of this issue possible. Firstly, thanks to
our Peer Review Board and the Harvard faculty, graduate students, and associates who reviewed our
manuscripts and ensured their scientific quality. Next to our Content, Design, Business, Social and
Public Relations, and Strategic Planning and Operations Boards for their tireless work and creativity.
Perhaps most importantly, this issue would not have been possible without the generous support
of HMS Dean Jeffrey Flier, Professor Steven Freedman, Provost Steven Hyman, Faculty of Arts and
Sciences (FAS) Dean Michael Smith, FAS Dean for the Physical Sciences Jeremy Bloxham, Professor
Xiao-li Meng, Harvard University, and Harvard College. And of course, a special thank you to the
authors whose work is showcased in these pages for sharing their inquiries with THURJ and the
Harvard Community. Enjoy.
Sincerely,
John Zhou Lisa Rotenstein

Co-Editor-in-Chief Co-Editor-in-Chief
ii The Harvard Undergraduate Research Journal

Volume 2 Issue 1 | Spring 2009 CONTENTS
Summaries
Preview the research articles with
summaries before reading them.
1
Research Features
Biomechanics 32
Elasticity in ionically cross-linked
neurofilament networks
Evolutionary Biology 38 Searching for life’s origins, on Earth
7
and beyond
Evolution of marine cyanobacteria in
the Red Sea
Health and Medicine 44
Biomechanical response of the in
situ primate lens in its natural versus
empty state as assessed with an ex
Quantum dots, vacuum energy, and
microfluidic chips 10
vivo accommodation simulator
Materials Science 49
Properties of silk III fibroin at the
air-water interface
A window into humanity
13
Molecular Biology 56
Tandem repeats in promoter regions
of S. cerevisiae generate variability
in gene expression with phenotypic
consequences
It’s a bird! It’s a plane! It’s a robotic fly!
17
www.thurj.org iii
CONTENTS Volume 2 Issue 1 | Spring 2009
Research Features
Psychology 66
Young people play bingo, too:
Reducing priming effects through
mindfulness Statistics: Your chance for happiness
(or misery) 21
Cover Image
This issue’s cover image comes from work done by the
Lieber group, mentioned in the article “Quantum dots,
vacuum energy, and microfluidic chips.” It features a
SEM image of hybrid measurement chip where neurons
are interfaced with multiple independently-addressable
nanowire devices.
A Marxist utopia in your backyard?
28
Visit http://www.thurj.org for news, details about the organization,

guidelines for submission, and other information on research at Harvard
iv The Harvard Undergraduate Research Journal

Volume 2 Issue 1 | Spring 2009 CONTENTS
Boards
About Us
Business
Hua Szu Yang ‘12 The Harvard Undergraduate Research Journal (THURJ) showcases
Design peer-reviewed undergraduate student research from all science and
Lisa Chen ‘12 - Associate Design Chair quantitative social science disciplines. As a biannual publication,
Francis Deng ‘12 - Associate Design Chair THURJ familiarizes students with the process of manuscript
Ritchell van Dams ‘11 submission and evaluation. Moreover, it provides a comprehensive
Evelyn Park ‘11 forum for scientific discourse on the cutting-edge research that
Yan Yan Mao ‘10
impacts our world today.
Lauren Kaye ‘10
Katie Goldin ‘11 At its core, THURJ allows students to gain insight into the peer
Ingrid Pierre ‘12 review process, which is central to modern scientific inquiry.
Kathleen Tang ‘12 All THURJ manuscripts are rigorously reviewed by the Peer
Joyce Yang ‘12 Review Board (consisting of Harvard undergraduates), and the
Content top manuscripts that they select are further reviewed by Harvard
Alissa D’Gama ‘11 - Associate Managing Editor graduate students, post-doctoral fellows, and professors. This
Sophie Wharton ‘11 - Associate Managing Editor process not only stimulates faculty-student collaboration and
Jen Jian Gong ‘12 provides students with valuable feedback on their research, but
Isha Jain ‘12 also promotes collaboration between the College and Harvard’s
Social and Public Relations many graduate and professional schools. In addition to publishing
Alyssa Blaize ‘12 original student research papers, THURJ is also an important
Roshane Campbell ‘12 medium for keeping the Harvard community updated on science
Angela Primbas ‘12 research-related news and developments.
Shoshana Tell ‘10
Peer Review and Submissions
Meng Xiao He ‘11 - Associate Manager Contact Executive Board
Monica Liu ‘12 -Associate Manager
Charlotte Seid ‘10 - Associate Manager General Co-Editors-in-Chief
Jessica Zeng ‘12 - Associate Manager Email: contact@thurj.org John Zhou ‘10
Helen Yang ‘11 - Head Copy Editor Lisa Rotenstein ‘11
Advertising
Lisa Chen ‘12 - Copy Editor
Email: advertising@thurj.org
Darius Li ‘12 - Copy Editor Business Manager
Jacob Cedarbaum ‘12 Subscriptions Alexander Piñero ‘11
Eric Chen ‘12 Email: subscriptions@thurj.org
Sway Chen ‘12 Managing Editor of Content
Submissions
Hyunje (Grace) Cho ‘12 Fernando Racimo ‘11
Email: submissions@thurj.org
Francis Deng ‘12
Ben Dobkin ‘12 Website Managing Editor of Peer
Kelly Fitzgerald ‘10 http://www.thurj.org Review and Submissions
Eva Gillis-Buck ‘12 John Liu ‘11
Jen Jian Gong ‘12 Copyright 2009 The Harvard
Johnny Hu ‘11 Undergraduate Research Journal.
Design Chair
Edward Kogan ‘12 John Mei ‘12
No material appearing in this
David Levary ‘12 publication may be reproduced
Shravani Mikkilineni ‘12 without written permission of the Co-Managers for Social and
Briana Prager ‘12 publisher, with the exception of the Public Relations
Abby Schiff ‘11 rights of photographs which may only Hyunje (Grace) Cho ‘12
Nicholas Tan ‘12 be granted by the photographer. The Gordon Bae ‘12
Jacob Weatherly ‘12 opinions expressed in this magazine
Vanisha Yarbrough ‘10 are those of the contributors and are
Manager for Strategic Planning
Chi Zhang ‘12 not necessarily shared by the editors.
and Operations
All editorial rights are reserved.
Strategic Planning and Operations Tengbo Li ‘12
Eric Lin ‘12
Anne Polyakov ‘12
www.thurj.org v
CONTENTS Volume 2 Issue 1 | Spring 2009
Faculty Advisory Board David Jeruzalmi, Ph.D Special Thanks to Our

Associate Professor of Molecular and Cellular
Biology Reviewers
Alán Aspuru-Guzik, Ph.D
Assistant Professor of Chemistry and Chemical Efthimios Kaxiras, Ph.D
Biology
Daniel Lieberman, Ph.D
Gordon McKay Professor of Applied Physics and Professor of Anthropology
Professor of Physics
Paul Bamberg, Ph.D Andrew Murray, Ph.D
Senior Lecturer on Mathematics George Lauder, Ph.D Herchel Smith Professor of Molecular Genetics
Professor of Biology and Alexander Agassiz
Michael Brenner, Ph.D Professor of Zoology David Mooney, Ph.D
Glover Professor of Applied Mathematics and Gordon McKay Professor of Bioengineering
Applied Physics Richard Losick, Ph.D
Maria Moors Cabot Professor of Biology L. Mahadevan, Ph.D
Myron Essex, D.V.M., Ph.D Lola England de Valpine Professor of Applied
Mary Woodard Lasker Professor of Health Mathematics
L. Mahadevan, Ph.D
Sciences in the Faculty of Public Health
Lola England de Valpine Professor of Applied
Mathematics
James McCarthy, Ph.D
Brian Farrell, Ph.D Alexander Agassiz Professor of Biological
Professor of Biology Oceanography
David Mooney, Ph.D
Jeffrey Flier, M.D. Associate Dean for Applied Chemical/Biological Yundan Pi
Dean, Harvard Medical School, and George C. Sciences and Engineering and Gordon McKay Ph.D Candidate in Earth and Planetary
Reisman Professor of Medicine Professor of Bioengineering Sciences
Hongkun Park, Ph.D Robin Greenwood, Ph.D

Nicole Francis, Ph.D Professor of Chemistry and Chemical Biology
Associate Professor of Molecular and Cellular Associate Professor of Business Administration
Biology Steven Pinker, Ph.D Maxine Isaacs, Ph.D
Johnstone Family Professor of Psychology Adjunct Lecturer in Public Policy, Harvard
Steven Freedman, M.D., Ph.D Kennedy School
Associate Dean for Clinical and Translational Tobias Ritter, Ph.D
Research and Associate Professor of Medicine Assistant Professor of Chemistry and Chemical Matt Chingos
Biology Ph.D Candidate in Government
Guido Guidotti, Ph.D
Higgins Professor of Biochemistry Eugene Shakhnovich, Ph.D Ellen Langer, Ph.D
Professor of Chemistry and Chemical Biology Professor of Psychology
David Haig, Ph.D
George Putnam Professor of Organismic and Irwin Shapiro, Ph.D Phyllis Kanki, Ph.D
Evolutionary Biology Timken University Professor Professor of Immunology and Infectious
Diseases, Harvard School of Public Health
Marc Hauser, Ph.D
Professor of Psychology
Zhigang Suo, Ph.D Marty Hirsch, Ph.D
Allen E. and Marilyn M. Puckett Professor of
Professor of Immunology and Infectious
Mechanics and Materials
Dudley Herschbach, Ph.D Diseases, Harvard School of Public Health
Frank B. Baird Jr. Professor of Science
David Weitz, Ph.D
Mallinckrodt Professor of Physics and of Katharine Jensen
John Hutchinson, Ph.D Ph.D Candidate in Physics
Applied Physics
Abbott and James Lawrence Professor of
Engineering and Gordon McKay Professor of
Applied Mechanics
vi The Harvard Undergraduate Research Journal

Volume 2 Issue 1 | Spring 2009 SUMMARIES
Biomechanics: Elasticity in ionically cross-linked neurofilament networks
The mechanical response of all eukaryotic cells, intermediate cross-linkers for these networks.
cells depends in large part on the filaments are highly specialized, They are then able to propose a
structure and elasticity of their e.g., with specific neurofilaments model which accounts for the
cytoskeleton. The cytoskeleton is in neurons, and desmin filaments physical origins of neurofilament
composed of a variety of biopoly- in muscle. Nevertheless, these fila- elasticity. This model, based on
mers, ranging from microtubules ments generally play a key role in thermal stretching, quantitatively
and filamentous actin to inter- the stable organization of the cy- explains the network’s remark-
mediate filaments and binding tosol and in giving cells enhanced able resilience and also allows for
proteins. compliance to external stress. the extraction of microstructural
While both individual micro- Thus, understanding intermediate parameters. These microstruc-
tubules and actin, as well as their filament assemblies and their me- tural network parameters, which
networks, have been extensively chanics represents an important include the polymer’s effective
studied in vitro and in vivo, much biophysical challenge. length and the average distance
less is known about the properties By examining both the me- between cross-links are extremely
of intermediate filament networks. chanical properties of in vitro difficult to measure directly.
This stems, in part, from their sheer neurofilament networks, the
diversity: while microtubules and authors are able to determine that
actin are essentially the same in multivalent ions act as effective p. 32
Evolutionary Biology: Evolution of marine cyanobacteria in the Red Sea
Each and every day, ecosystems By taking carbon compounds out to determine how each strain
around the world are being trans- of the atmosphere and convert- has adapted to unique environ-
formed by global climate change. ing them into nutrients usable by mental conditions. Synechococcus
Soon, the planet’s environments other aquatic species, cyanobacte- sequences showed 26 genes related
may no longer be inhabitable by ria allow bodies of water to act as to nitrogen processing, suggesting
their current residents. Among ‘carbon sinks’ and slow the effects that they have adapted genetically
the first to feel these effects will of global warming. to their environment. Prochloro-
be the most ancient organisms This research studied the genetic coccus, however, showed only
adapted specifically to their compositions of Prochlorococcus one nitrogen-related gene. It is
habitats, including cyanobacteria. and Synechococcus cyanobacteria possible that, while Synechococcus
The Prochlorococcus and Synecho- to understand how they have evolved genetically to survive in
coccus strains of these microscopic adapted to the specific nitrogen- a nitrogen-stressed environment,
organisms are the smallest photo- stressed environment of the Red Prochlorococcus made physical or
synthetic organisms on Earth, yet Sea. By sequencing, analyzing, behavioral changes instead. These
they account for two-thirds of the and comparing their DNA se- two different approaches dem-
oceans’ photosynthetic reactions. quences, the investigator tried onstrate that cyanobacteria have
www.thurj.org 1
SUMMARIES Volume 2 Issue 1 | Spring 2009
Photo by Lamiot/Wikipedia
multiple ways to adapt to changing marine environments, but their specific evolutionary mechanisms should
be further investigated.
p. 38
Health and Medicine: Biomechanical response of the in situ primate lens in its natural
versus empty state as assessed with an ex vivo accommodation simulator
Near vision, or presbyopia, is the substance and decreased elastic- rhesus monkey, and human eyes
most common refractive disorder ity of the lens capsular bag with of varying ages were stretched
of the elderly. It is characterized age. One novel technique in the in their natural and empty states
by the loss of accommodation, restoration of accommodation is in an ex vivo Accommodation
the ability to focus distant to near Phaco-Ersatz, or lens refilling. The Simulator in eight, 0.25mm steps,
objects. Lens and capsule-based promise of Phaco-Ersatz can be mimicking the changes that occur
theories of presbyopia assert that assessed by characterizing the bio- in vivo when focusing objects. The
the decrease in accommoda- mechanical properties of the lens diameter-force relationship of
tive ability can be attributed to in its natural versus empty state. the natural and empty lens were
increased hardening of the lens Postmortem cynomolgus monkey, characterized and compared.
2 The Harvard Undergraduate Research Journal

Illustration by Evelyn Park/THURJ Staff
There was no relationship between load-diameter slope decreased sig- conducted as long as the lens
the empty-bag diameter slope and nificantly with age, showing that contents have proper viscoelastic
age, indicating that the lens cap- it is the lens material, and not the properties. Thus, Phaco-Ersatz
sule’s mechanical properties do elasticity of the capsular bag, that is a viable future treatment for
not change the setting of accom- contributes to presbyopia. The re- presbyopia.
modation. Moreover, the ratio of sults confirm the postulation that
the empty capsule to natural lens accommodation can be efficiently p. 44
Materials Science: Properties of silk III fibroin at the air-water interface
The macromolecule fibroin,

found in the silk of silkworms, is
found in the silk I and II confor-
mations, which have well charac-
Illustration by Charlotte Seid/THURJ Staff

terized structures and properties.
Since ancient times, these materi-
als have lent themselves to a variety
of applications, and now that their
sturctures are fully apprehended,
they can be used for increasingly
complex biomedical applications,
including the rapid repair of
torn anterior cruciate ligaments
(ACL). It was hypothesized that
the fibroin crystalline structure
silk III could also have applicable
properties, such as high tensile stood due to its relatively recent properties of silk III fibroin at
strength, biocompatibility, and discovery and the lack of an es- the air-water interface in order to
liquid crystallinity. tablished manufacturing process facilitate future manufacturing
Silk III is not yet fully under- for it. This research assessed key and biomedical applications. A
www.thurj.org 3
SUMMARIES Volume 2 Issue 1 | Spring 2009
procedure was first determined for at such an interface. A successful the physical integration of silk III
the consistent production of silk preliminary biocompatibility in hybrid materials. Overall, the
III at the interface. Mechanical assessment affirmed possible fu- investigator evaluated the novel
strength and tensile properties ture use of silk III in humans for material of silk III and its capacity
were measured and analyzed to fit biomedical applications. Imaging for use in biological and physical
a mathematical model. The data tests also revealed important struc- applications of interest.
indicated which concentrations are tural features of the silk III film,
optimal if silk III is manufactured important for applications such as p. 49
Molecular Biology: Tandem repeats in promoter regions ofS. cerevisiae generate variability
in gene expression with phenotypic consequences
Most genomes are made

up of substantial portions of
repetitive DNA. In humans, for
example, as much as 49% of the
genome consists of such repeats.
Tandem repeats, sequences
which are repeated head-to-tail
at one specific locus within the
genome, are especially interest-
ing because of their high level
of variability. The focus of this

study is on tandem repeats which
occur in the promoter regions
of genes in Saccharomyces cer-
evisiae (brewer’s yeast). Here, the
investigator shows that repeats
in promoters are indeed hyper-
variable and often differ between
evolutionarily closely related
sub-populations of yeast (natural
yeast strains). Studies conducted
concerning the effect of repeat
size on the transcriptional activ- variation in transcriptional activ- phase that corresponded with the
ity of the candidate genes SDT1 ity. Growth assays were conducted gene expression of the respective
and YKL071w found that changes with SDT1 strains in the presence strain, suggesting that changes
in the number of repeat units of 6-azauracil, which yielded dif- in transcription levels result in a
in the genes corresponded to ferences in the length of the lag phenotypic change. In YKL071w,

binding sites for the stress-response that just as variable repeats located quick evolution of gene expression
transcription factor Yap1 overlap within coding regions allow swift levels to changing environments
with the variable tandem repeats. evolution of protein function, and selective pressures.
Together, these results indicate repeats in promoters may allow p. 56
Psychology: Young people play bingo, too: Reducing priming effects through mindfulness
In psychology, the concept of film displayed more aggressive Bargh and colleagues and extend-
mindfulness refers to making behaviors in a subsequent hockey ed it by introducing the concept
a conscious effort to be “in the game than did those who instead of mindfulness, to see whether
moment” and aware of one’s watched a nonviolent film about by making participants mindful,
environment, which allows for bike racing. they could reduce the priming
more thoughtful decision-making. In one well-known study on effects. Although non-significant,
In this study, the researchers priming, researchers investigated the study’s results showed an
examined how the concept of how being exposed to words asso- interaction in the expected direc-
mindfulness might inform our ciated with the elderly stereotype tion with decreased walk-time for
understanding of another concept affected how quickly participants participants in the elderly-mindful
in psychology: priming. Priming walked down a hallway at the condition, lending tentative sup-
is the idea that exposing people conclusion of an experiment. They port to the researchers’ hypothesis.
to certain stimuli can affect their found that participants who were This concept has important impli-
subsequent behavior by increasing exposed to words related to the cations for social issues such as
the accessibility of other informa- elderly stereotype walked more stereotyping, as the results imply
tion even without them being slowly than did those who were that mindfulness may be an effec-
aware of it. For example, research- exposed to neutral words. tive means of reducing stereotype
ers in one study found that chil- In the current study, the re- activation.
dren who watched a violent police searchers replicated the study by p. 66
www.thurj.org 5
Features
Volume 2 Issue 1 | Spring 2009 FEATURES
!!"#$%&'()*!+,%!-(+#./
!!!!,%(*()/0!,)!1$%2'
!!!!!!!!!!$)3!4#5,)3
By Sophie Wharton, THURJ Staff
Illustration by Katherine Goldin/THURJ Staff
!
ince the beginning of recorded history, the recognized the need to bring together researchers
questions “How did life begin?” and “Are we working on separate pieces of the same puzzle, to
alone in the universe?” have mystified humans, form a bridge between the physical and life sciences.
dominating religious and political debate. Today, As Dimitar Sasselov, director of the initiative, puts it,
Harvard scientists from across four major disciplines “The big issues faced by humanity today and in the
have united to grapple with the same age-old ques- near future can only be solved by combining exper-
tions, using innovative new technologies. tise from the life sciences and the physical sciences.”
Biologists, chemists, earth and planetary scientists, Sasselov is an astronomer whose research has
and astronomers together comprise the Harvard focused on the diversity of planetary environments.
Origins of Life Initiative, a grassroots project officially The collaborative and interdisciplinary nature of
formed in 2005 after recommendations from the Task the initiative are helping him gain new perspectives
Force on Science and Technology. The Task Force had on his own work: “This project is allowing me to
www.thurj.org 7
FEATURES Volume 2 Issue 1 | Spring 2009
follow a path in the discovery and study of distant primitive cell membranes and is studying their
planets which is much richer in what I look to find dynamic nature and the conditions necessary for them
on them.” Just as discoveries on a micro-level are to spontaneously enlarge and divide.
guiding the theoretical and observational work of the The chemists in George Whitesides’ Research Group
astrophysicists and planetary scientists, so too are the – also involved in the Origins Team – are trying to
laboratory experiments conducted by chemists and find out how energy was available to drive the earliest
chemical reactions on pre-biotic Earth. They have
“Across all levels of inquiry made progress in explaining how cells became able to
harness the potassium and sodium potentials created
– from the cosmic to the by the different concentrations of these compounds
outside and inside the cell, and use them as a source of
cellular – enormous and free energy. Another group in the Initiative – Scot T.
Martin’s lab – managed to demonstrate that in prebi-
often unanticipated strides otic conditions, a reverse version of the Krebs cycle (a
have been made since the key biochemical process in cellular respiration) might
have produced the first biomolecules. The reaction
project was launched.” may have been catalyzed by sunlight combining with
a particular mineral that is thought to have existed in
Earth’s early waters.
molecular biologists informed, via their setup and On a different front, astronomers on the Origins
initial conditions, by the macro-level work of the other Team developed a new laser that dramatically
team members. improves our ability to measure the size of a star’s
Thus, across all levels of inquiry – from the cosmic “wobble”: a type of oscillation caused by the gravita-
to the cellular – enormous and often unanticipated tional pull of orbiting bodies around it, and a proxy
strides have been made since the project was launched. for the existence of extrasolar planets that might be
The research conducted by Jack Szostak, a professor of
genetics at Harvard Medical School and a researcher at “As Sasselov notes, ‘We
Massachusetts General Hospital, has helped advance
the scientific understanding of how the very first cell are uniquely positioned to
may have formed and evolved. His lab developed
offer Harvard students an
opportunity to combine
projects in the intersection
of life sciences and physical
sciences.’”
capable of harboring life. Until now, only the wobble of
giant, gaseous planets with very noticeable effects on
stars could be detected. However, with this new laser
technology, the sensitivity of detection has increased
by about a hundred times, and it is now possible to
detect smaller, rockier planets, which are better candi-
dates for sustaining life.
Such exciting findings are evidence that the Initia-
tive is successfully tackling key questions about life’s
origins. But research is only one part of its mission:
education and outreach are two of the most important taking one of the eight undergraduate courses given by
goals. Carol Knell, Program Coordinator for the Origins team members at Harvard, including the core
Initiative, has been on board since the beginning and course Science A-54 “Life as a Planetary Phenomenon,”
remarks on how publicity and outreach events have which is taught by Sasselov.
“This interdisciplinary enterprise… boldly confronts some

of the most fundamental questions about our existence
and our universe.”
progressed since the project’s infancy: “so many more After the $8 billion hit that Harvard’s endowment
people, nonscientists and scientists, have become took a few months ago, there is fear that funding to the
aware and excited about the initiative.” Knell helps to Origins of Life Initiative may suffer too. The progress
organize the Origins Forums, monthly events aimed at of the Initiative is hindered by the lack of a dedicated
giving the University community a taste of the current space for labs, which are currently scattered through-
research undertaken by the Initiative. Large audiences out the University. Nevertheless, the development of a
regularly come to hear presentations from researchers new science campus in Allston – on the other side of
from Harvard, other universities, or international the Charles River – offers a promising solution.
speakers. During the more informal monthly “Chalk Certainly, this interdisciplinary enterprise is one that
Talks,” graduate students and post docs discuss excit- merits Harvard’s support, as it boldly confronts some
ing advances made in their research. of the most fundamental questions about our existence
Undergraduates are encouraged to attend all such and our universe. And although this team of scientists
events, but those interested in a more hands-on experi- does not embark upon scientific inquiries with the
ence are welcome to apply for research fellowships to intention of contributing to philosophical or religious
work on a particular project in a team member’s lab. debate, Sasselov holds that one cannot deny that such
Fellowships are available for Harvard undergraduates “new fundamental knowledge is always bound to shape
during the academic year (wage stipends are provided) humans’ total worldview.”
and summer fellowships are open to undergraduates And that fundamental knowledge cannot come
at other colleges as well. Knell reports that feedback about without scientific collaboration across many
from students involved in research has been fantastic fields. The questions of whether we are alone in the
across the board. As Sasselov notes, “We are uniquely universe or how we came to be here will most likely
positioned to offer Harvard students an opportunity not be answered by a team composed solely of biolo-
to combine projects in the intersection of life sciences gists or solely of astronomers. It will be up to teams
and physical sciences.” Students can also learn about of scientists from a variety of disciplines – like the
the astrophysics and biochemistry of life’s origins by Origins of Life Initiative – to take on that challenge.
www.thurj.org 9
Quantum dots,
vacuum energy,
and
microfluidic
chips: Bridging the
sciences at
By Isha Jain, the
THURJ Staff
nanoscale
level

I
n the past, the sciences were often regarded as questions through a nanoscale lens. Rather than
a hierarchy. Professor Federico Capasso – from simply considering single units at the nanoscale level,
Harvard’s School of Engineering and Applied the Center strives to understand systems built from
Sciences – describes society’s original outlooks: nanoscale components. In addition, CNS serves as a
“The fields that are more theoretical, more far from technological melting pot. Supported by grants from
reality, particularly in science and technology, by the NSF, the
some ill-conceived argument were viewed as superior offers the Harvard research community access
intellectually.” But for Capasso, “this mentality is a total to a myriad of technological resources for imaging,
disaster: we create the differences between these fields. nanofabrication and materials synthesis.
“The advantage of CNS is that you have a

lot of interdisciplinary activity, because
the lines between biology, chemistry and
physics are being blurred at the nanoscale
level.”
Nature does not distinguish whether a phenomenon

is chemistry, physics or biology…it is a completely Why here? Why now? What next?
W
artificial thing.” Recently, we have come to appreciate hen it comes to the life sciences, Harvard’s research
science as more of a coplanar playing field. Every is in many ways unparalleled. But when it comes
subfield exists to serve its own purpose as well as to to engineering, more technically focused institutions
augment progress in other subfields. One of the newest like MIT tend to be seen as at the forefront of scientific
embodiments of this mantra is Harvard’s very own efforts. With such a distinctive reputation, the Center
Center for Nanoscale Systems (CNS). faced the challenges of establishing an engineering
The term “nano” tends to conjure up an image and physics based institution in the midst of a largely
of the unknown, the mysterious that resides in a biology-oriented community. As Professor Capasso
parallel world, see-sawing reality and science fiction. explains, “The advantage of CNS is that you have
But the groundbreaking research being conducted a lot of interdisciplinary activity, because the lines
in the CNS seeks to conquer such ambiguities in the between biology, chemistry and physics are being
form of applicative innovations. It is often difficult to blurred at the nanoscale level. A number of professors
conceptualize orders of magnitude that deviate from had the wisdom to propose a service center, a facility.
our day-to-day encounters. A nanometer is 10-9 of a These fields have developed so much that you need a
meter. A human hair is approximately 100 microns or centralized facility. It cannot be done in the old style
100,000 times a nanometer in width. A typical covalent garden variety manner.” CNS is now at the forefront of
bond is approximately one-tenth of a nanometer. So, competitive nanotech research due to the collaborative
what truly lies behind all the hype of this esoteric efforts of its faculty and to the top-notch technology it
nano-world? The primary motivation driving this field houses.
of study is the increased importance of minute forces
we tend to disregard on larger scales. These forces Projects through a Peephole
D
drastically alter the classical laws of physics. We enter r. Federico Capasso exudes a love of science. With
a realm of phenomena such as quantum-tunneling, a distinctive Italian accent, he remarks, “I like to
carbon nanotubes and quantum dots. think of myself as an engineer. I work in a number of
The Center for Nanoscale Systems is an endeavor interlocked, interdisciplinary areas. I like to bridge the
by the Harvard faculty to approach multidisciplinary gap between so called fundamental stuff, basics and
www.thurj.org 11
applications.” In one of its projects, Capasso’s group circuits.

seeks to understand a newly discovered phenomenon. In the biotech realm, the Westervelt group has con-
Contrary to what one may believe, a vacuum is a structed a “hybrid integrated circuit and microfluidic
surprisingly dynamic condition. Specific materials pos- chip.” This fusion of technologies enables researchers
sess intrinsic properties at the nanoscale level. In the to trap and move thousands of small droplets and
mid 1900s, a theory known as the Casimir effect first living biological cells in specific configurations. This
highlighted the existence of attractive forces between biological etch-a-sketch has been used to move yeast
macroscale objects. Now, the Capasso group has found and mammalian cells at speeds in the tens of microns
the first evidence of repulsive forces caused by a so per second.
called “vacuum energy.” Two plates made of special
materials can now be engineered to have a levitating
effect. The weight of the upper weight can be countered “The future lies in
by the repulsive force between the two plates.
The group is also pioneering development in laser
lowering the barrier
technology. “Last year our group developed a tiny
laser spectrometer on a chip, something like a finger
between science,
nail. The Holy Grail is to go to the longest possible engineering and
wavelength. At this wavelength, light can penetrate
non metallic enclosures, so it could be a substitute technology.”
to x-rays. It could be used in security check points to
detect weapons.”
Robert Westervelt – a Professor of Physics and At the forefront of integrative bio-nanotechnology
Applied Physics at the CNS – works to engineer at the CNS is the work of Dr. Charles Lieber. There are
circuitry using quantum dots. It was originally believed two approaches generally used by nanotechnologists:
that the number of microprocessors able to fit on top-down and bottom-up. The Lieber lab focuses on
a chip would double every three years. However, the bottom-up approach to design nanoscale building
physical laws place an upper limit on this prediction. blocks and then eventually build complex and elabo-
Then came the revolutionizing idea of quantum rate systems. For example, the Lieber group devises
computers, based on quantum dots. Quantum dots sensory technology to detect specific protein-protein
contain a specific number of electrons (each with an up interactions or the presence of bio-hazardous materials
or down spin). This is analogous to the binary system in trace amounts. The basis of this technology is the
currently utilized in common computers. Westervelt manner in which the presence of a particular molecule
has led the pioneering research of connecting quantum can be translated into an electrical signal. In another
dots, which in turn could be used to build nanoscale project, Lieber attempts to model engineering at the
nanoscale level off the architecture of the brain. Today’s
computers have to be restricted to the concept of the
2-dimensional flat chip. Dr. Lieber has been crafting

the foundation for branching nanowires, to overcome
this restriction.
The Future of the CNS
C apasso, Westervelt, Lieber and many other faculty

at the CNS are paving the path for integrative
research in the study of nanoscale systems. As Capasso
asserts, “the future lies in lowering the barrier between
science, engineering and technology.” The Center
for Nanoscale Systems combines all three: studying,
modifying, and creating new connections between the
basic particles that make up our world.
Illustration by Kathleen Tang/THURJ Staff
Online collaboration and its implications

by Jen Jian Gong, THURJ Staff
R ecent years have seen the massive rise of
web-based social networking , blogs, and
free data sharing through websites like
Project Gutenberg and Wikipedia. Not surprisingly,
the way scientists look at social interactions is now
changing dramatically. Professor Yochai Benkler, the
Jack N. and Lillian R. Berkman Professor of Entrepre-
neurial Legal Studies of the Harvard Law School, has
made human behavior his primary research focus and
is now helping to lead the effort in introducing social
implications to the study of the Internet.
Professor Benkler started his career at Harvard
as a Teaching Fellow, worked as an associate at the
Ropes & Gray law firm in Boston and then as a law
clerk to Justice Stephen Breyer of the United States
Supreme Court. He later became a professor of law
at the New York University School of Law, continu-
ing his teaching career at Yale, and finally coming
back to Harvard in 2007, where he became faculty
co-director of the Berkman Center for Internet and
Society. His book, The Wealth of Networks: How Social
Production Transforms Markets and Freedom, has won
many awards, including an award from the American
www.thurj.org 13
Sociological Association in 2008
Illustration by Kathleen Tang/THURJ Staff

and the 2007 Don K. Price award
from the American Political Sci-
ence Association for the best book
on science, technology, and politics
“How does the

internet affect
democracy?
And what does
collaboration
online mean to
society and human
relationships in
general?”
in the last 3 years.
Benkler’s research aims to answer
two fundamental questions: how
does the Internet affect democracy?
And what does collaboration online
mean to society and human rela-
tionships in general? In the 1990s,
the Internet was regarded as “a new
frontier,” where “everyone could
say what they wanted.” However,
there was not enough real data to
back up this idealism. But now, “the
physical capital means in our hands
– computers, mobile phones – are
as necessary as the physical capital
in the core economic activities of
the most advanced economies” and
in recent years have started to affect
“behaviors we’ve always had as
social beings.”
Benkler is currently studying
the same process, but in reverse.
Originally the puzzle was, “‘why
do we see so much cooperation
online?’ and the answer had to do
with things we’ve always been doing
becoming economically important.” he proposes that “one of the big text from a variety of sources in
But now that we’ve been exposed challenges now is to try and zoom order to use machine analysis, in
to so much online cooperation, the in on what is happening within the collaboration with other groups in
real question is: “what can we learn blogs, within the text itself.” Insteadthe University, such as the Institute
about human motivation or inter- of integrating all the blogs and for Quantitative Social Science at
action from that very exposure?” then recording the data, Benkler Harvard, to find out “where ideas
Professor come from,
Benkler
is also
“Benkler is also questioning the ‘basic, how they move
through the
questioning
the “basic,
dominant assumption’ that ‘everybody is network public
sphere so that
dominant more or less the same, we’re all rational we can actu-
assump- ally begin to get
tion,” that and we’re all self-interested.’” a map of ideas
“everybody and statements.”
is more or less the same, we’re hopes to maintain the integrity Instead of staying away from the
all rational and we’re all self- of the individual nature of each content of each blog, MediaCloud’s
interested.” Instead, he posits that blog and then search for similar goal is “zooming in, both by human
perhaps all the networking and characteristics between them. coding of what the actual practices
social interactions occurring online For example, political affiliations are, and by a large-scale text analy-
can show us “deviations from define interesting differences in the sis to look at what it is that people
selfishness.” blogosphere. “We’ve been looking are doing.”
An important aspect of Benkler’s at the organization of technologies The research spearheaded by
research deals with the perspective adopted by blogs, and we’re finding Professor Benkler taking place at
we take when looking at the data very interesting differences between the Berkman Center investigates
about different sites and practices the left and right wings of the what all the collaborative efforts on
on the Internet. While the current blogosphere, with the left adopting the Internet mean to society’s un-
tendency is to “look at every single technologies that allow for much derstanding of human nature and
blog as though it is the same thing,” more participation: people who human motivation: “How do we
are not the primary construct a new model for human
authors can quite behavior that takes into account
easily find a place the reality that we’re observing in
on the front page lots of different disciplines, that
of a site, whereas people don’t behave in practice
the right tends to according to the predictions of
Photo by Jen Jian Gong/THURJ Staff
be much more solo selfishness?” The Center’s research

and hierarchical in on human cooperation focuses on
the sense that users free software and Web 2.0 websites
can’t contribute that like Wikipedia and Facebook,
much.” investigating the evolution of social
Following this behaviors evidenced and enabled
line of thought, by the Internet.
Benkler is currently In questioning our current
associated with economic assumptions about
MediaCloud, a human nature (that we are fully
scientific project rational and selfish), Benkler also
that attempts to joins with others in “a great group
take in massive that meets weekly with people from
Professor Benkler quantities of online evolutionary dynamics and com-
www.thurj.org 15
puter science and political science communities and tries to see what could understand how cooperation
and the business school – all sorts typifies them: “What are the kinds works online, not only could you
of people who are interested in this of things they’re doing? What are play a part in creating institutions
question of how we understand the the kinds of motivations? How that support working in cool things
micro-foundations of cooperation.” much do people rely on money, online, you could also take part in
David Rand, a member of the how much do people rely on a broader trend we’ve been seeing
Program of Evolutionary Dynamics moral implications, how much do in the past few years: internet
at Harvard, coordinates these meet- people rely on just creating person culture manifesting itself in the
ings. He works with Benkler on the to person interactions and com- mainstream. If you understood it
cooperation
group project, “If human nature is indeed different, as
investigating
how the
cooperation over the Internet seems to suggest,
collaborative
efforts that
then ‘solidarity is a force that can potentially be
have appeared harnessed to help groups work together’”
online question
our current beliefs about human munications?” These observations, on this level, you could generalize it
nature. Rand believes that “in the Benkler notes, provide the perfect to the larger societal space.”
face of changing conflicts and contrast with experimental work, Professor Benkler’s work and
shifting alliances, it is essential “[combining] both greater preci- the related research happening at
for group identities to be flexible,” sion and tractability and greater the Berkman Center and all across
rather than centered on individual realism in terms of looking at the Harvard demonstrate the impor-
selfish interests. If human nature way humans are actually behaving.” tance of interdisciplinary study.
is indeed different, as cooperation Tim Hwang, a recent Harvard “[T]he net provides a breathtaking
over the Internet seems to suggest, College graduate, worked part-time expanse where you can actually go
then “solidarity is a force that can at the Berkman center during his and look at lots of different things
potentially be harnessed to help undergraduate years. He is now a without physically traveling to a
groups work together if institutions research associate there, and works thousand different places.” This
are designed in a way that prevents with Benkler on many projects, research links a variety of depart-
the negative and discriminatory pursuing his interest in online ments and subjects, pooling knowl-
aspects of such behavior.” communities. Hwang’s enthusiasm edge and using technology and
Another study at the Berkman for the cooperation project and networking as windows into human
Center seeks to analyze a couple these studies on online communi- nature and social interactions.
of thousand online collaboration ties arises from the idea that “if you

It’s a bird!
It’s a plane!
Photo by John Mei/THURJ Staff
It’s a robotic fly!

!"#$%%"&'($)*"+'(,-..('"/(#"0((*1."
!"#$%$%&%'#()*+&
!"#$%&''(#)*+(,(-# most undergaduates’ travels—you suggestions and sharing in their

find yourself amid the hustle and enthusiasm.
!"#$%&'()* bustle of undergraduates, graduate What’s so special about his
students, and postdocs peering product? It is the first at-scale
P
rofessor Robert Wood’s cre-
through microscopes at actuators, robotic insect with sufficient thrust
ations have been featured in
using lasers to create air frames, to take off—in other words, fly—
an exhibit at the Museum
and tweaking mechanical wings with external power.
of Modern Art, on two Discovery
and transistors.
Channel series, and in Time
Yet the most excited of all is Getting ready for takeoff
Magazine. When you walk into his
Professor Wood himself, walking
Harvard lab at 60 Oxford Street—a
building beyond the boundary of
around and peering over the
shoulders of his students, making
P rofessor Wood’s journey to
flight began when he entered
graduate school at the University
www.thurj.org 17
of California, Berkeley. With a research projects was material All insect wings have some degree
background in electrical engineer- design—saying “I want to make this of compliance due to aerodynamic
ing, Wood hoped to find a research shape, so how do I do that?” That and inertial loads experienced in
project in his interest as well as thinking eventually led him down flight. Wood hopes to understand
“inherently why insects
interesting of
itself.”
!"""#$%&'$($)*#+&*',-)./&($01& have this com-
pliance and how
At that time, -20$.$3$2#&'$($0#+&415*1&-',&*-6-(7,&$8& it benefits them
one of the first during flight.
efforts to create 7$*$3$)$.+&-.9&*',-)./&.,4&3-0,'5-7#& Soft robots
a robotic insect aren’t like
was starting up, 4501&5.1,',.0&*$3675-.*,&75:,&-');*5-7& the tin-can
and the project
caught Wood’s
32#*7,#&-.9&,3(,99,9&#,.#$'#"! creations from
comic books—
attention—yet rather, they are
what he initially thought he would two primary paths. robots that can easily change their
work on was not what he ended “One is microrobotics, creating shape and size and are made from
up researching. Wood was geared particular robotic insects, things flexible materials.
towards working on controlling the that fly and crawl that are inspired Unlike those working on creating
fly, but the fly didn’t exist yet! by various aspects of arthropod human robots—sensationalized
“My adviser said, ‘You have to morphology and control materials,” on the big screen in movies like I,
build the fly first,’” said Wood. “I said Wood. “The second is soft Robot, Wood has had to focus on
had to switch gears and become a robotics, creating both autonomous many aspects larger-scale roboti-
different sort of engineer.” robots, which are capable of cists take for granted and that can’t
After receiving his PhD, Wood locomotion, and also creating new be applied directly on a smaller
carried over his microrobotics work materials with inherent compliance scale to robotic insects.
to Harvard where he was appointed like artificial muscles and embed- “Instead of jumping right to
as an Assistant Professor of Electri- ded sensors.” questions of control or behavior
cal Engineering and Computer Inherent compliance in wings re- we have to start from nothing and
Science. The catalyst behind his fers to their “degree of floppyness.” do everything from scratch,” said
Wood.

Fly, Robot, Fly
I
n 2007, Wood’s
fly took flight for
the first time, fixed
to guide wires that
allowed only vertical
motion. The robotic
fly was integrated from
four components:
an actuator (the flight
muscles), transmission
(thorax), an air frame
(exoskeleton), and wings.
Weighing in at only 60 mg
with a 3 cm wingspan, the
tiny mechanical creature
could move freely in the
vertical direction with a trajec- pieces” of previous models. wartime or as a means to locate
tory nearly identical to biological As Wood noted, once his lab survivors trapped in hazardous
counterparts. demonstrated the first successful
This robotic fly can be put flight, “that opened up a dozen
together in under a week – a user- different paths—we have an endless !"#$%&'()&*+,&
friendly advancement.
“In the past the fabrication
supply of things we need to study.”
He hopes to develop robotic
-%."#)/0+/%-&/'%&
methods we used took a great deal insects that can fly in multiple 10)/&)2$$%))32*&
of skill and time,” said Wood. “That dimensions (rather than solely
was fine when it was really just me vertically) without an external 4(5'/6&7/'+/&
making them, but when you get power source and with the ability
new students involved nobody has to work together in groups. !"#$#%&'"&(&%!)#$&
this very esoteric skill set that they
wouldn’t use for anything else.”
His project is funded by the
U.S. Defense Advanced Research
-(8%0%#/&9+/'):
Now the emphasis is on the Projects Agency, who hopes that ;%&'+<%&+#&%#-*%))&
empirical portion of the research— at-scale robotic insects can be
building structures and trying to used as spy technology during )299*=&"3&/'(#5)&;%&
characterize their
performance to #%%-&/"&)/2-=*+
“fill in the
empty environments.
But Wood said he encouraged
people to look beyond the face
value.
“Any place you wouldn’t want to
put a human or an animal, there’s
a big class of scenarios for applica-
tion of a mobile robot,” said Wood.

“And in reality, we’re developing
new techniques for fabrication and
actuation at an interesting scale
which individually could be ap-
plied to understand many different
phenomena, for example, the fluid
mechanics of different
wings.”
www.thurj.org 19
hours at the lab each day interact- out one of his students, he works science, bioengineering…every-
ing with the students, as well as to raise money and make sure thing,” said Wood. “So when you’re
scheduling weekly lab meetings for everyone is happy. stuck, you have a lot of different
voices from a lot of different
areas.”
!"#$%&'()%&*$('+&,-./*&,&0.-.1)2&3,-&(2&*$,*& And by putting all those
(*42&5%06&7/318(2)(93(',06:;%&$,5%&2*/8%'*2& voices together, Wood and
his students have created a
;$.&,0%&2*/86('+&7%)$,'(),3&%'+('%%0('+<& robotic fly that is, indeed,
ready for take-off and may
%3%)*0(),3&%'+('%%0('+<&).79/*%0&2)(%')%<& one day search out soldiers
injured in the battlefield or
-(.%'+('%%0('+=%5%06*$('+<!"#$%&"'((&) miners trapped underground.
As Prof. Wood’s website
relates: “Remarked one
the lab as a whole and for different “The nice thing about a robotics unimpressed Yale researcher, ‘Leave
subgroups. When he’s not tinkering lab is that it’s very multidisci- it to the Harvard fellows to invent
with the robotic flies themselves, plinary—we have students who are new and exciting ways to be irritat-
he says he will often embark on studying mechanical engineering, ing.’”
his own little project that will help electrical engineering, computer
As an introduction to undergraduates seeking out lab research, Prof Wood sat down with THURJ and described
the projects different members of his group are currently working on.
Postdoctoral student: Dr. Sangbae Kim Undergraduates: Elliot Hawkes, Brandon Eum, and
Geneva Trotter
Dr. Kim is working on the integration of soft actuators
into compliant sheets. To put it simply, he is creating Most undergraduates work closely with a graduate student
“programmable matter.” “There are sheets which or post-doctoral student. Case in point: Hawkes is working
fold themselves to create whatever useful devices with Kim to integrate novel actuation within self-folding
you might want to have around,” said Wood. sheets, and Eum creates experiments to quantify the motion
of the wings developed by graduate students.
Graduate student: Ben Finio Trotter is working on creating an experimental device

that would allow Wood and his group to apply arbitrary
Finio is working on the thoracic mechanisms, or pre-loads on a certain class of actuators. She previously
transmission, of the robotic flies. He is trying to worked on characterizing conductive polymers for use in a
come up with ways to separately include actuators compliance circuit.
for power and low-power actuators for control. “We
want to create a compact mechanical device that When you put the work of all these students together, what
can not only flap its wings and crate motion but also results is a nanorobotic fly that we would commonly swat
crate body torque,” added Wood. at—but in this case, has widespread applications.

Volume 2 Issue 1 | Spring 2009 Op-Ed
Statistics:
Your chance for happiness
(or misery) By Professor Xiao-Li Meng
Whipple V. N. Jones Professor of
Statistics and Department Chair
1. “I keep saying the sexy
job in the next ten years
Photo by Rose Lincoln/Harvard News Office

will be statisticians.”
Hal Varian, Google’s Chief

Economist, recently was inter-
viewed by The McKinsey Quar-
terly, and was quoted (see www.
mckinseyquarterly.com/Strategy/
Innovation/):
“I keep saying the sexy job in the

next ten years will be statisticians.
People think I’m joking, but who
would’ve guessed that computer Professor Meng and his “Happy Team” on the opening day of Stat 105
engineers would’ve been the sexy Cassandra Wolos, Kari Lock, Xiao-Li Meng, Yves Chretien, and Paul Edlefsen
job of the 1990s? The ability to take
data—to be able to understand it, to to remind you that a professor of caused the housing bubble, and how
process it, to extract value from it, to any subject can find quotes – tons to properly forecast those prices based
visualize it, to communicate it—that’s of them – to demonstrate the on all the variables involved (e.g.,
going to be a hugely important skill importance of his or her beloved interest rates, inventories, short sales,
in the next decades, not only at the foreclosures, delinquencies, etc.). I am
subject.
professional level but even at the actively seeking those individuals
Wait! Does the “reminder”
educational level for elementary who have the background and desire
school kids, for high school kids, for have anything to do with being a to apply their Stat/analytical skills
college kids. Because now we really do statistician? Well, let’s label this specifically in the Real Estate me-
have essentially free and ubiquitous question as Puzzle One, and read dium. … The trend in these unique
data. So the complimentary scarce on. And while we are at it, let me economic times is that companies
factor is the ability to understand that throw in another quote, this time want the more scientific/mathemati-
data and extract value from it.” from a recruiter representing Wall cal/engineering backgrounds to help
Street – yes, they are still hiring – them back solve [sic] these very new
As a professor of statistics, you but read this carefully: and volatile markets. My clients these
guessed it, I of course cannot days are actually shying away from
disagree less (just to check if you “Now more than ever, they are MBA-types because today’s equity
looking for the best and brightest to markets have much more to do with
have had enough coffee!). But
help get an understanding as to what randomness and psychology than
as a statistician, I am obligated
business fundamentals.”
www.thurj.org 21
w Op-Ed Volume 2 Issue 1 | Spring 2009
Here, the word randomness is the other hand, the same variation who look like a “Chinese female
what brings statistics and statisti- causes uncertainty. If all I tell you student.” You then need to do
cians into the picture. Statistics, is to pick up a Chinese female something creative on your own in
in a nutshell, is a discipline that student by the name Xiao-Li order to pick up the right one, such
studies the best ways of dealing (meaning “Little Beauty” (!") as making a name sign.
with randomness, or more precisely in Chinese, not “Plough at Dawn”
and broadly, variation. As human (#$)!as in my
beings, we tend to love information Chinese name - an “Statistics, in a nutshell,
but hate uncertainty, especially example of uncer-
when we need to make decisions. tainty in translation, is a discipline that studies
But information and uncertainty
actually are two sides of the same
or lost in transla-
tion!), then my
the best ways of dealing
coin. If I ask you to go to the description is not with randomness, or more
airport to pick up a new student informative enough
you have never met, my description precisely because it precisely and broadly,
of her is information only because still allows too many
there are variations - if everyone “variations” - there variation”
at the airport looks identical, then may be a substantial
my description has no value. On number of individuals at the airport Then again, the name sign is use-
ful for her to identify that you are
the one who is picking her up, only
because there is variation among
names. Indeed, if it happens that
there are two “Xiao-Li” name signs
outside the terminal, she will need
to do something creative on her
own in order to find the right one.
This is of course a trivial fact, and
any of us would recognize and
deal with the situation when we

encounter it. But we may or may
not recognize the deeper principle
behind it, that is, information is
there for the same reason that
uncertainty is there.
While we are at the airport,
let me throw in this almost well-
known joke. Mr. Skerry needs to
take a flight, but he is terrified by
the possibility, however small, that
someone could bring a bomb onto
his plane. So he decides to pack a
bomb himself, as he reasons that
the chance that two individuals
bringing bombs onto the same
plane is much smaller than that of
one individual bringing a bomb.
You, of course, are chuckling at
this. However, which probabilistic/
statistical principle is he trying (i.e., undergraduates) I have spoken anity, and Jesus himself is teaching.”
to use, or rather violating? Can with, the number one reason that (If you can come up with more
you easily explain to your fellow you did not even consider majoring impressive praise than this, email
students why Mr. Skerry’s argument (or concentrating, to be true to me at chair@stat.harvard.edu!).
is ridiculous? If you cannot, then the Harvard spirit!) in statistics Another colleague, Joe Blitzstein,
let’s label this as Puzzle Two. is because the AP Statistics you has single-handedly doubled the
Regardless of whether you can or took convinced you that statistics enrollment of Stat 110, Introduction
cannot, I hope the discussion above is the most boring subject. We to Probability, from 90 students
has helped you to see more clearly, statisticians, of course, are to when he took over in 2005-2006,
and fundamentally, why Google be blamed for this unfortunate to 188 students this past fall. He
and Wall Street, among many situation. Statistics is an urgently is now an international sensation,
others, are increasingly interested demanded but vastly underap- so to speak – a student was telling
in hiring statisticians. We are now preciated field; urgently demanded her friend in Germany that she was
squarely in the information age, for reasons discussed above, and taking this cool stat course with Joe,
with almost everything digitized. vastly underappreciated because and her friend responded “Oh, you
Each of us is trying to see what all too few statisticians, relatively mean that YouTube stat professor?”
the data (which don’t have to be speaking, have effectively conveyed (You can satisfy your curiosity by
numerical) out there are telling us, the excitement of statistics, as a way googling “Stat 110 at Harvard.”)
on issues from personal health to of scientific thinking for whatever Last year, we also launched Stat
the global economic crisis. There you do, instead of a collection of 105, Real Life Statistics: Your Chance
is so much variation in almost tools you may or may not need one for Happiness (or Misery), and I
everything we want to know or day. Tremendous efforts have been am teaching it again this semester.
study, so what is real information made, for example, by the Con- This course was designed by my
and what is just noise? Mr. Skerry’s sortium for the Advancement of Happy Team, which consisted of 8
reasoning surely is ridiculous, but Undergraduate Statistics Education Master’s and Ph.D students from
how many of us have the statistics depart-
realized that the
many “small prob-
“Each of us is trying to see what ment, over a period
of two years and
abilities” reported in all the data out there are telling many happy dinners
the media and even (not happy meals!) at
in scientific publica- us, on issues from personal health the best restaurants
tions, such as prob- Boston can offer.
abilities of crime to the global economic crisis” The course aims at
evidence, were based introducing students
on exactly the same ridiculous (CAUSE, http://www.causeweb. to the wonderland of statistics, by
reasoning, that is, multiplying org/). But clearly more is needed, showcasing how it is used (and mis-
probabilities inappropriately? as surely any successful educational used) in real-life situations every
program requires on-going effort. student should be able to relate to,
2. “AP Statistics was the At Harvard Statistics, we are either happily or miserably!
most boring course I took fortunate to have several first-class Unlike many traditional
statistical educators who are at the statistical courses, which arrange
in high school!”
forefront of teaching introductory the material by statistical topics
statistical courses. For example, my in the approximate order of their
As a professor of statistics, I hear
colleague, Ken Stanley, who teaches complexity, Stat 105 arranges the
this almost every time I tell some-
Stat 104, Introduction to Quantita- material by what we call “Real-Life
one that I teach statistics: “Oh, that
tive Methods for Economics, has Modules.” For last year’s offering,
was really a hard course for me!” or,
been so effective that one student the five modules were (1) Finance
“I really didn’t like my stat course!”
wrote in his/her CUE evaluation, (e.g., stock market), (2) Romance
And for nearly every one of you
“It is like taking a course in Christi- (e.g., on-line dating models; not
www.thurj.org 23
dating on-line models!),

(3) Medical Sciences (e.g.,
Viagra trial; not trying
Viagra!), (4) Law (e.g.,
OJ Simpson trial), and
(5) Wine and Chocolate
Tasting (depending on your
age!). This semester, we are

replacing the Law module
by an Election module,
given the historic election
we all just witnessed (and
now that OJ is behind bars).
More information about
the first offering can be
found in the Valentine’s Day
edition of Harvard Gazette
exactly one year ago (http://
www.news.harvard.edu/
gazette/2008/02.14/11-stats.
html). For the current
offering, check the Stat
105 course website (open
to anyone with a Harvard
ID) and view the video for
the first-day introductory
lecture to enjoy a virtual
chocolate tasting, with or research, regardless of the subject, phone because of those convincing
without wine! but also in your life. Our happiness “infomercial statistics,” chances are
All these efforts are aimed to or misery often literally depends that you would have been much
make “statistics not just palatable, on (but of course is not necessarily more satisfied by trying out the
but delicious” (the title of the determined by) our understanding chocolates or wine offered by our
aforementioned Gazette article) to of statistics, whether we realize it Stat 105 class! (And of course if
all of you, who, I am 98% sure (that or not. Statistics or, more generally, you have a relative who had been
quantitative evidence is being convinced by the dazzling “return
“Our happiness or used everywhere in the media, statistics” of Mr. “Made-Off,” then
scientific publications, etc., to no amount of chocolates or wine
misery often literally persuade us to buy a product, an could compensate!)
argument, a theory, etc. Some
depends on our of the claims are statistically 3. “Honey, I know you are
understanding of and scientifically sound, and
many others are not. A good
in excruciating pain, but
which treatment do you
statistics, whether percentage of them are even
want?”
deliberate lies, intended to
we realize it or not” deceive the public in order to
make a profit. If you have been Here is another real-life scenario
is the highest assurance any profes- one of those flipping channels that literally makes your happiness
sional statistician would give!), will in the wee hours and have given or misery depending on your
need statistics not only in your own your credit card number over the understanding of statistics, if you,

like me, unfortunately suffer from I reported above are from an actual are its general implications? Did it
kidney stones. Two treatments for study (Charig et. al., British Medical actually happen with some studies
kidney stones were evaluated in a Journal (Clinical Research Ed), you have done or read? Let’s label
medical study. Treatment A has a March 1986, 292 (6524): 879–882), this Puzzle Three and read on,
success rate of 78% and treatment and you can verify them yourself: unless you really suffer from kidney
B, 83%. Which one should you for treatment A, there were 350 stones, in which case let me distract
choose? Surely treatment B, right? patients, 87 carrying small stones, you by telling you how I was treated
Well, what if I tell you that when among which
treatment A and treatment B are
applied to those who suffer small
treatment
A was suc-
“Simpson’s Paradox ... is
stones, the success rates become, cessful for 81 responsible for a vast quantity
respectively, 93% and 87%, and patients; for
when they are applied to those the remaining of misinformation in the
who carry large stones, the success 263 patients
rate for treatment A is 73% and with large literature and in the public”
for treatment B it is 69%? That is, stones, treat-
regardless of the sizes of the stones, ment A was successful for 192 of by Dr. Coe from The University
treatment A has a higher success them. For treatment B, there were of Chicago (where I taught from
rate. Surely you then should choose also 350 patients, with 270 suffering 1991-2001), a world renowned
treatment A, right? small stones, and among them 234 nephrologist, who treats his
Confused? You should be, if were successfully treated by treat- patients with statistical principles!
you don’t understand Simpson’s ment B; for the remaining 80 with Once Dr. Coe learned that I was
Paradox (no relationship with OJ, large stones, treatment B was found a statistician, he said, as I recall,
though there could be a paradox successful for 55 of them. “Well, you then should understand
with him too, if he is still looking Now you do the math! And this well. The kidney stones are ac-
for himself), one of the most then think statistically – how could tually formed by a Poisson process,
fundamental statistical phenomena, this happen? That is, how could with those crystals bumping into
which is responsible for a vast treatment B have a better success each other. So what you need to
quantity of misinformation in rate overall than treatment A, and do is to drink a lot of liquid, any
literature and in the public. There yet a worse rate in each subgroup kind of liquid, water, juice, coffee,
is actually no paradox at all in the defined by the stone size? What even beers and wine, anything that
mathematical sense. The numbers caused such a “paradox”? What helps to reduce the Poisson rate for
www.thurj.org 25
crystals to bond with each other.” close to the boundaries, either 1 unnecessarily worried and gone
He was obviously pleased to finally or 2, suggests something to watch through additional tests, simply
find a patient who understood for. Indeed, my physician asked because of an elementary statistical
“Poisson process,” and surely the me to schedule an appointment mistake in setting the “normal
feeling was mutual as I was pleased with an endocrinologist for further limits.” Incidentally, I was told by a
that I was treated by a doctor who studies. This of course is a rational medical student that when a pa-
understood statistics! Of course suggestion, given the “normal” tient’s list of test results come back
I have followed his advice closely, interpretation of my test result and from the lab, abbreviation WNL
and have not had any episodes of the fact that I was having various after the name of a test indicates
that the result was “within
“I was sad to think how many other normal limits.” The inside joke is
that it really stands for “we never
people had unnecessarily worried and looked.” Having incorrectly set
normal limits could be even
gone through additional tests, simply worse than “we never looked”!
I hope by now I have distracted
because of an elementary statistical you enough from your kidney-
mistake in setting the ‘normal limits’” stone suffering, and that you
understand what Dr. Coe was
laughing and sighing about. If
kidney stones for the past 15 years symptoms, which could have been not, let’s label this as Puzzle Four,
or so. And I have never had any due to a thyroid disorder of some and read on again.
surgery for kidney stones, nor am I sort.
on any other treatment now other Since the quality of doctors mat- 4. “The best thing about
than a lot of drinking – so next ters (obviously!), and it happened being a statistician is that
time you see me pouring myself a that I had a regular follow-up visit
you get to play in every-
glass of wine, I may be just trying to with Dr. Coe shortly after that test,
reduce my Poisson rate for crystals I asked him if he could recommend one’s backyard.”
bonding! a colleague who is an endocrinolo-
Here is another example where gist. He naturally asked me why, This quote is attributed to John
Dr. Coe saved me much trouble and I showed him the test results. Tukey, a statistical giant who also
and worry because of his – and my He laughed and sighed at the same coined the terms “software” and
– understanding of statistics. While time: “Well, these doctors really “bit” (see http://www.princeton.
I was at The University of Chicago, don’t know anything,” (I assume it’s edu/pr/news/00/q3/0727-tukey.
I suffered for a long period from OK for a well-known doctor to say htm or The New York Times, July
fatigues and various pains of that!). He continued, “For years I 28, 2000). This is literally true,
unknown cause. So my primary- have told them that they shouldn’t as many statisticians, myself
care physician did all sorts of tests provide “normal limits” as such included, can personally testify.
on me. One of them was checking when the distribution is highly Other than teaching the delicious
my thyroid function. One result skewed! You actually have the most Stat 105 class and other courses
came back on the “borderline” – I typical value in the population! (e.g., I also co-teach, with Joe
don’t recall which test and what You of course understand that Blitzstein, Stat 303, The Art and
were the exact values, but for the they should have taken a log or Practice of Teaching Statistics,
sake of the story, let’s say my value something.” As a statistician, I was aimed at training more and better
was 1.1 and the normal range listed both happy and sad. I was happy future statistical educators), I am
was (1.0 – 2.0). Most people would of course that I had no reason to currently conducting – together
consider this interval (1.0 – 2.0) worry about my thyroid (and I still with researchers from the Harvard-
to imply that values close to 1.5 don’t to this date). I was sad to Smithsonian Observatory – a
to be “normal” and a test result think how many other people had workshop on AstroStat for dealing

with astronomical amounts of course also play in statistics’ own have provided a snapshot on how
data from astrophysics; working backyard, or perhaps I should say practically useful and intellectually
with a group of geophysicists front yard, investigating statistical fulfilling it is to be a statistician, or
from the University of Illinois and foundational issues, such as to what at least to be able to reason with
the National Weather Service on extent size matters – do more data good statistical insights. I am
climate change; writing papers automatically imply more accurate certainly having great fun, both
with a team of psychiatrists from results? (This one will take more professionally and personally, as a
the Harvard Medical School thinking, so let’s consider it the last statistics professor, and I hope you
and Columbia University on Puzzle of the Day.) will be able to share some of the
estimating disparities in mental If you find the range of my fun by taking at least one statistics
health services; course, no matter
collaborating
with researchers
“I hope you will be able to share how much you
hated that idea be-
from Harvard’s some of the fun by taking at least fore. You will then,
engineering among many other
school on signal one statistics course, no matter how benefits, easily find
processing,
particularly for
much you hated that idea before” out the answers
to all five puzzles
digital cameras, listed above. If you
via wavelets methods; publishing “backyard” activities impressive, want to think hard about them now
articles with statistical geneticists check out our webpage (stat. to challenge yourself, of course that
at The University of Chicago harvard.edu), and prepare to be is part of the fun! But if you start
and deCode Genetics in Iceland dazzled by a wide range of “front to lose sleep over any of them and
on how to measure information yard” research my colleagues feel miserable, email me (chair@
in genetic studies; preparing are conducting, such as Sam stat.harvard.edu) – remember, I
reports with my ex-postdoc at Kou’s absolutely pioneering work promised you both happiness and
The University of Chicago on on statistical models for neon- misery!
AIDS reporting delay to the CDC biochemical experiments.
(Center of Disease Control). I of I hope the quotes and stories
www.thurj.org 27
BOOK REVIEW Volume 2 Issue 1 | Spring 2009
A Marxist utopia in
your backyard?
Book review: The Superorganism
By Fernando Racimo, THURJ Staff
E
dward O. Wilson once proclaimed: “Karl Marx members of the group (workers and soldiers) forfeit
was right, socialism works, it is just that he had their individual right to reproduce in order to dedi-
the wrong species.” The renowned American cate more amounts of time, effort, and resources to
biologist has dedicated his life to studying the organ- taking care of the brood and of those individuals who
isms that would have made the German philosopher do get to reproduce – either one queen or a group of
proud: ants, bees and termites – the stalwarts of social them, depending on the species under study.
living. Wilson’s new book, The Superorganism– co-au- But unlike what Marx had in mind, eusociality has
thored with the biologist no room for centralized
Bert Hölldobler – seeks planning. Wilson and
to unravel the mysteries Hölldobler aim to show
of insect eusociality: that there is no central
a social system that in administrator inside a
many ways surpasses bee hive or a termite
human societies in its Photo by Yewenyi/Wikipedia
mound: the queen is too
complexity and cohe- busy producing eggs and
sion. the workers are too busy,
The book begins by well, working. Therefore,
giving an evolutionary planning comes about as
survey of the origins of an emergent property of
eusociality before transi- each individual’s percep-
tioning into an extensive tions and actions. But
ecological account of how do thousands of
the inner workings of a insects with brains
variety of insect colonies. smaller than the tip
The last chapters are of a needle manage to
devoted to analyzing coordinate and exchange
the most extraordinary A complex social system exists within this termite so much information?
displays of eusociality, mound The book
like fungi agriculure dedicates an entire
in leaf-cutter ants and the giant structures built by chapter to the many forms of eusocial communica-
African termites to house their brood. tion, another to the development of simple decision
What exactly is eusociality? As the authors make rules associated with different inputs of information,
clear, eusocial colonies generally possess three defin- and another to the interplay of communication and
ing characteristics: cooperative care of the young, information in determining castes and allowing for
overlaps between generations of individuals, and cooperative teamwork to gather resources or attack
reproductive castes. The last feature implies that some neighboring colonies. Wilson and Hölldobler have
Volume 2 Issue 1 | Spring 2009 BOOK REVIEW
an ability to make complex biology seem like a piece Yet this is not exactly the case in the eusocial
of cake, and that ability is at their best in these three kingdom: parents, siblings and offspring are not exact
chapters. From long-lasting chemical trails to alarm copies of one another and the levels of relatedness
pheromones, the processes that underlie the exchange between each other – though uncommonly high – are
of information between workers, soldiers and queens
are detailed with true mastery. They emphasize how
little science has revealed about eusocial communica- “But unlike what Marx had
tion and how much more remains to be discovered.
Their one flaw in this analysis is perhaps their overly in mind, eusociality has
narrow focus on ant communication, while saying little
to nothing about communication in bees, termites
no room for centralized
or other eusocial insects, like aphids and thrips. This planning.”
is, nevertheless, justifiable given how much is known
about ants (especially because of the authors’ own
scientific contributions) and how little is known about not exactly as high as those between the cells in
other colonial organisms. our body. The authors show many examples of the
Though efficient mechanisms for communication chaos that sometimes ensues when a queen dies
allow the colony to maximize its energy input and and the foundations of colonial hierarchy begin to
reproductive output, eusociality also entails a measure shake. Usually, communication and agonistic rituals
of internal conflict. Chances are that Marx would have prevent widespread violence. Grabbing a subordinate’s
been slightly disappointed, even if he had applied antennae and shaking them, for example, constitutes
his sociopolitical theories to the world of ants and a display of superiority in contests for power and
termites. In fact, some female worke`r ants mutilate prevents subordinates from continuing their attempts
their brother’s wings to prevent them from mating at climbing the social ladder (biology shows that bul-
with other queens; workers may also kill queens who lying is by no means limited to high school corridors).
But conflict may go much farther than bullying. In
“Wilson and Hölldobler ponerine ants, as the queen gets old and weak, some
workers (called gamergates) acquire the ability to
have an ability to make reproduce. Once they are recognized by other workers
as reproductives, they are also able to mark other
complex biology seem like a potential gamergates with a chemical. The signal ef-
fectively activates policing behavior in workers, which
piece of cake.” proceed to spread-eagle and often kill the potential
gamergates who dare defy the reproductive right of the
recognizably do not possess the gene that causes the already established gamergates.
killing in the first place. Conflict indeed permeates Given such high degrees of cooperation, conflict
what seems like a cohesive system from an outsider’s and communication, there is one central issue that the
viewpoint. book addresses poorly: how did it all come to be that
In fact, there is so much conflict inside certain colo- way? Wilson and Hölldobler make use of multilevel
nies that the word “superorganism” may not exactly be selection theory to explain the origins of eusociality.
the right way to describe a eusocial colony. All of the Multi-level selection theory is the idea that altruistic
cells in an organism’s body carry the same versions of traits are naturally selected in the course of evolution
each gene, so it is extremely unlikely that a group of because they benefit a group of individuals as opposed
cells will try to harm any of the other cells. They are all to the individual members of the group. So long as the
in for the ride, and they all privilege the survival and group has low intra-group genetic variance (members
reproduction of the organism above anything (cancer are more related to themselves than to outsiders), the
and autoimmune diseases are rare exceptions). Genes trait will be selected and become widespread. Thus,
need to cooperate in order to be passed on to future certain behaviors and communication mechanisms in
generations. ants must – according to Wilson and Hölldobler – have
www.thurj.org 29
BOOK REVIEW Volume 2 Issue 1 | Spring 2009
evolved because they were advantageous to the group. separate processes, creating confusion as to what
Their argument seems to run opposed to inclusive fit- exactly they mean when they refer to the evolution of
ness theory: the idea that traits can be selected not just eusociality.
because they benefit an individual’s offspring but also Despite the sloppiness of the book when describing
because they benefit close genetic relatives who have evolutionary phenomena, its elegance and clarity when
a high probability of carrying the same genes as the it comes to physiological, organizational and ecological
individual with the traits. For Wilson and Hölldobler processes are unmatched. Not only do the authors
inclusive fitness theory is not a sufficient explanation succeed at making the reader understand how dynastic
for eusocial behavior. It is thus necessary to use succession affects the overarching dynamics of the
“[The Superorganism] is a telling account of a social

world that is highly distinct from human societies and
yet has managed to thrive during millions of years of
evolutionary struggle.”
multilevel selection theory to explain some aspects eusocial kingdom or how recruitment to new nest sites
of eusociality. And they stress this point in repeated occurs over time, they also make sure the reader ‘sees’
instances throughout the book from the division of these and many other behaviors in vivid photographs
labor to self-imposed restrictions to reproduction. and illustrations. An example is the image of a group
The problem with this approach is that, as many of weaver ants forming a “living bridge” with their own
scientists are now realizing, multilevel selection is just bodies to pull leaves together, which truly enlivens
another side of the coin of inclusive fitness theory. Wilson and Hölldobler’s narrative of nest construction.
Both theories are accurate because, in reality, both are Overall, The Superorganism constitutes a compre-
just two equivalent methods of explaining the same hensive analysis of a complex biological system by two
mechanism. Stating that altruistic group traits are masters in the field. But more than that, it is also a
selected when group members are closely related to telling account of a social world that is highly distinct
each other is just another way of saying that altruistic from human societies and yet has managed to thrive
individual traits will be selected when they are directed during millions of years of evolutionary struggle. Both
towards close kin. Even Wilson and Hölldobler admit of its authors have studied these creatures for decades
to this point at the very beginning of their book: “It is and, through this book, they make a compelling case
important to keep in mind that mathematical gene- for the value of this knowledge and the many questions
selectionist (inclusive fitness models) can be translated that still remain unanswered. Ants, bees and wasps
into multilevel selection models and vice-versa.” And do not live in communist utopias. Nevertheless, their
yet they don’t seem to apply their own realization to extraordinary ability to cooperate with one another
the rest of the book: the authors repeatedly make the and carry out coordinated enterprises by the millions
distinction between the two theories as if they were is something Marx would have definitely envied.

Research
RESEARCH Volume 2 Issue 1 | Spring 2009
Elasticity in ionically cross-linked

Biomechanics
neurofilament networks
Norman Y. Yao§*, Yi-Chia Lin*, Chase P. Broedersz†, Karen E. Kasza‡,
Frederick C. MacKintosh†, and David A. Weitz*‡¶
§
Harvard College 2008; *Department of Physics, Harvard University, Cambridge, MA 02138, USA;
†
Department of Physics and Astronomy, Vrije Universiteit, 1081HV Amsterdam, The Netherlands;
‡
School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA.
Neurofilaments are found in abundance in the cytoskeleton of neurons, where they act as an
intracellular framework protecting the neuron from external stresses. To elucidate the nature of the
mechanical properties that provide this protection, we measure the linear and nonlinear viscoelastic
properties of networks of neurofilaments. These networks are soft solids that exhibit dramatic strain
stiffening above critical strains of 30–70%. Surprisingly, divalent ions, such as Mg2+, Ca2+, and Zn2+ act
as effective cross-linkers for neurofilament networks, controlling their solid-like elastic response. This
behavior is comparable to that of actin-binding proteins in reconstituted filamentous actin. We show
that the elasticity of neurofilament networks is entropic in origin and is consistent with a model for
cross-linked semiflexible networks, which we use to quantify the cross-linking by divalent ions.
Introduction from the filament backbone [9]. These highly charged “side-arms”
lead to significant interactions among individual filaments as well as
The mechanical and functional properties of cells depend between filaments and ions [10]. Although the interaction of divalent
largely on their cytoskeleton, which is comprised of networks of ions and rigid polymers has been previously examined, little is known
biopolymers; these include microtubules, actin, and intermediate about the electrostatic cross-linking mechanism [11]. Networks
filaments. A complex interplay of the mechanics of these networks of neurofilaments are weakly elastic; however, these networks are
provides cytoskeletal structure with the relative importance of able to withstand large strains and exhibit pronounced stiffening
the individual networks depending strongly on the type of cell [1]. with increasing strain [12, 13]. An understanding of the underlying
The complexity of the intermingled structure and the mechanical origin of this elastic behavior remains elusive; in particular, even
behavior of these networks in vivo has led to extensive in vitro studies the nature of the cross-linkers, which must be present in such a
of networks of individual biopolymers. Many of these studies have network, is not known. Further, recent findings have shown that NF
focused on reconstituted networks of filamentous actin (F-actin) aggregation and increased network stiffness are common in patients
which dominates the mechanics of the cytoskeleton of many cells with amyotrophic lateral sclerosis (ALS) and Parkinson’s. Thus, an
[2-7]. However, intermediate filaments also form an important understanding of the fundamental mechanical properties of these
network in the cytoskeleton of many cells; moreover, in some cells networks of neurofilaments is an essential first step in elucidating the
they form the most important network. For example, in mature role of neurofilaments in a multitude of diseases [14]. However, the
axons, neurofilaments, a type IV intermediate filament, are the most elastic behavior of these networks has not as yet been systematically
abundant cytoskeletal element overwhelming the amount of actin studied.
and outnumbering microtubules by more than an order of magnitude Here, we report the linear and nonlinear viscoelastic properties
[8]. Neurofilaments (NF) are assembled from three polypeptide sub- of networks of neurofilaments. We show that these networks form
units NF-Light (NF-L), NF-Medium (NF-M), and NF-Heavy (NF-H), cross-linked gels; the cross-linking is governed by divalent ions such
with molecular masses of 68 kDa, 150 kDa and 200 kDa, respectively as Mg2+ at millimolar concentrations. To explain the origins of the
[8]. They have a diameter d ~ 10 nm, a persistence length believed network’s elasticity, we apply a semiflexible polymer model, which
to be of order lp ~ 0.2 µm and an in vitro contour length L ~ 5 µm. ascribes the network elasticity to the stretching of thermal fluctua-
They share a conserved sequence with all other intermediate tions; this quantitatively accounts for the linear and nonlinear elas-
filaments, which is responsible for the formation of coiled dimers ticity of neurofilament networks, and ultimately, even allows us to
that eventually assemble into tetramers and finally into filaments. extract microstructural network parameters such as the persistence
Unlike other intermediate filaments such as vimentin and desmin, length and the average distance between cross-links directly from
neurofilaments have long carboxy terminal extensions that protrude bulk rheology.
¶
Author to whom correspondence should be addressed: Department of Physics and School of Engineering and Applied Sciences, Harvard
University, Pierce 231, 29 Oxford Street, Cambridge, MA 02138, USA. E-mail: weitz@seas.harvard.edu.

Volume 2 Issue 1 | Spring 2009 RESEARCH
Materials and Methods σ, resulting in a total stress of the form σ(t) = σ + |δσ| sin(ωt). The
resultant strain is γ(t) = γ + |δγ| sin(ωt + φ), yielding a differential
Materials elastic modulus
and a differential viscous modulus
Biomechanics
Neurofilaments are purified from bovine spinal cords using a
[2].
standard procedure [9, 15, 16]. The fresh tissue is homogenized in
the presence of buffer A (Mes 0.1 M, MgCl2 1 mM, EGTA 1 mM, pH Scaling Parameters
6.8) and then centrifuged at a K-factor of 298.8 (Beckman 70 Ti). The To compare the experiments with theory, we collapse the differen-
crude neurofilament pellet is purified overnight on a discontinuous tial measurements onto a single master curve by scaling the stiffness
sucrose g radient with 0.8 M sucrose (5.9 ml), 1.5 M sucrose (1.3 ml) K' and stress σ by two free parameters for each data set. According to
and 2.0 M sucrose (1.0 ml). After overnight sedimentation, the theory, the stiffness versus stress should have a single, universal form
concentration of the purified neurofilament is determined with a apart from these two scale factors. We determine the scale factors
B radford Assay using bovine serum albumen (BSA) as a standard. by cubic-spline fitting the data sets to piecewise polynomials; these
The purified neurofilament is dialyzed against buffer A containing polynomials are then scaled onto the predicted stiffening curve using
0.8 M sucrose for 76 hours and then 120 μl aliquots are flash frozen in a least squares regression.
liquid nitrogen and stored at -80 °C.
Results and Discussion
Bulk Rheology
The mechanical response of the cross-linked neurofilament To quantify the mechanical properties of neurofilaments, we
networks is measured with a stress-controlled rheometer (HR Nano, probe the linear viscoelastic moduli of the network during gelation,
Bohlin Instruments) using a 20 mm diameter 2 degree stainless which takes approximately one hour; we characterize this by continu-
steel cone plate geometry and a gap size of 50 μm. Before rheological ously measuring the linear viscoelastic moduli at a single frequency,
testing, the neurofilament samples are thawed on ice, after which they ω = 0.6 rad/s. Gelation of these networks is initiated by the addition of
are quickly pipetted onto the stainless steel bottom plate of the rhe- millimolar amounts of Mg2+ and during this process we find that the
ometer in the presence of varying concentrations of Mg2+. We utilize linear viscoelastic moduli increase rapidly before reaching a plateau
a solvent trap to prevent our networks from drying. To measure the value. We measure the frequency dependence of the linear viscoelas-
linear viscoelastic moduli, we apply an oscillatory stress of the form tic moduli over a range of neurofilament and Mg2+ concentrations.
σ(t) = A sin(ωt), where A is the amplitude of the stress and ω is the To ensure that we are probing the linear response, we maintain a
frequency. The resulting strain is of the form γ(t) = B sin(ωt + φ) maximum applied stress amplitude below 0.01 Pa, corresponding to
and yields the storage modulus and the loss modulus strains less than approximately 5%; we find that the linear moduli

. To determine the frequency dependence of the are frequency independent for all tested frequencies, 0.006–25 rad/s.
linear moduli, G'(ω) and G''(ω) are sampled over a range of frequen- Additionally, neurofilament networks behave as a viscoelastic solid
cies from 0.006–25 rad/s. In addition, we probe the stress dependence for all ranges of Mg2+ concentrations tested and the linear storage
of the network response by measuring G'(ω) and G''(ω) at a single modulus is always at least an order of magnitude greater than the
frequency varying the amplitude of the oscillatory stress. To probe linear loss modulus, as shown in Fig. 1. This is indicative of a cross-
nonlinear behavior, we utilize a differential measurement, an effec- linked gel and allows us to define a plateau elastic modulus G0 [20].
tive probe of the tangent elastic modulus, which for a viscoelastic The elasticity of neurofilament networks is highly nonlinear;
solid such as neurofilaments provides consistent nonlinear measure-
ments of elasticity in comparison to other nonlinear methods [17-19].
A small oscillatory stress is superimposed on a steady pre-stress,
102
, 1mg/ml (5mM)
, 2mg/ml (5mM)
, 2mg/ml (3mM)
A B
101
G',G''(Pa)
101
G',G''(Pa)
100 100
0.5 mg/ml 2mM Mg2+

1.5 mg/ml 5mM Mg2+
10-1
2.0 mg/ml 8mM Mg2+
10-1 -3
10-3 10-2 10-1 100 10-3 10-2 10-1 100 10 10-2 10-1 100
f (Hz) f (Hz)
Fig. 1. γ
Figure 1. The frequency dependence of the linear viscoelastic
Fig. 2.
moduli of cross-linked networks for a variety of neurofilament Figure 2. The strain-stiffening behavior of neurofilament net-
and Mg2+ concentrations. a. Variations of the moduli at constant Mg2+ works at various Mg2+ and neurofilament concentrations. Close
concentration (5 mM) and changing filament concentration; b. Variations squares represent G', the elastic modulus and open squares represent G'',
of the moduli at constant neurofilament concentration (1.5 mg/ml) and the viscous modulus. Dramatic nonlinearities are seen at critical strains
changing Mg2+ concentration. ranging from 30–70%.
www.thurj.org 33
therefore hypothesize that the network elasticity is due to the stretch-

ing out of thermal fluctuations. These thermally driven transverse
fluctuations reduce neurofilament extension resulting in an entropic
Biomechanics
14 spring. To consider the entropic effects we can model the Mg2+-cross-

linked network as a collection of thermally fluctuating semiflexible
G0 (Pa)
segments of length lc, where lc is the average distance between Mg2+

cross-links. A convincing test of the hypothesis of entropic elasticity
10 is the nonlinear behavior of the network. When the thermal fluctua-
tions are pulled out by increasing strain, the elastic modulus of the
network exhibits a pronounced increase.
6 To probe this nonlinear elasticity of neurofilament networks, we
measure the differential or tangent elastic modulus K'(σ) at a con-
2 8 stant frequency ω = 0.6 rad/s for a variety of neurofilament and Mg2+
6 concentrations. If the network elasticity is indeed entropic in origin,
0.4 0.8 4 this can provide a natural explanation for the nonlinear behavior in
M)
terms of the nonlinear elastic force-extension response of individual
1.2 2
(m
cN (m 1.6 2.0 0 filaments that deform affinely. Here, the force required to extend a
F g/ml c
Mg
single filament diverges as the length approaches the full extension
) lc, since [24-26]. Provided the network deformation is
affine, its macroscopic shear stress is primarily due to the stretch-
Fig. 3. ing and compression of the individual elements of the network. The
Figure 3. The linear elastic modulus can be finely tuned by varying the
concentration of the cross-linker Mg2+ and the neurofilament concentra- expected divergence of the single-filament tension leads to a scaling
tion. of

; we therefore expect a scaling of network stiffness
with stress of the form K'(σ) ~ σ3/2 in the highly nonlinear regime [2].
Indeed, ionically cross-linked neurofilament networks show remark-
above critical strains γc of 30–70%, the networks show stiffening up to
able consistency with this affine thermal model for a wide range of
strains of 300% [21], as shown in Fig. 2. This marked strain-stiffening
neurofilament and cross-link concentrations, as shown in Fig. 4. This
occurs for a wide variety of Mg2+ and neurofilament concentrations.
consistency provides convincing evidence for the entropic nature of
In addition, by varying the neurofilament concentration cNF and
the network’s nonlinear elasticity [2, 25].
the Mg2+ concentration cMg, we can finely tune the linear storage
The affine thermal model also suggests that the functional form of
modulus G 0 over a wide range of values, as seen in Fig. 3. The strong
the data should be identical for all values of cMg and cNF. To test this,
dependence of G 0 on Mg2+ concentration is reminiscent of actin
we scale all the data sets for K'(σ) onto a single master curve. This is
networks cross-linked with the incompliant cross-linkers such as
accomplished by scaling the modulus by a factor G' and the stress by a
scruin [2, 22, 23]; this suggests that in the case of neurofilaments,
factor σ'. Consistent with the theoretical prediction, all the data from
Mg2+ is effectively acting as a cross-linker leading to the formation
various neurofilament and Mg2+ concentrations can indeed be scaled
of a viscoelastic network. Thus, the neurofilaments are cross-linked
onto a universal curve, as shown in Fig. 5. The scale factor for the
ionically on length scales comparable to their persistence length;
modulus is the linear shear modulus G' = G 0, while the scale factor
hence, they should behave as semiflexible biopolymer networks. We
104 2+
103
2mM Mg & 1.5 mg/ml
8mM Mg2+ & 1.5 mg/ml
103 5mM Mg2+ & 0.5 mg/ml
102
K' (Pa)
5mM Mg2+ & 2.0 mg/ml 3/2

K' / G0
102 3/2
1
101
10
100 100
10-2 10-1 100 101 102 10-3 10-2 10-1 100 101
Fig. 4.
σ (Pa) σ// σCc
Fig. 5.
Figure 4. The dependence of K'(σ) on σ for a variety of neuro- Figure 5. Collapse of all data sets of the σ dependence of K' onto
filament and Mg2+ concentrations. All data show an exponent of ap- a single universal curve. The solid line represents the theoretical pre-
proximately 3/2 in agreement with the affine thermal model. diction of [2]. The scaling parameters are G0, the linear elastic modulus
and σc, the critical stress. These parameters are calculated using a least
squares regression.

101
2.5 3
Biomechanics
G0 (Pa)
0.41
G0 /σc * cMg-0.21
101
0
10
cNF1/2*G0
-1
10
0
10
cNF (mg/ml)
1.54
G0 /σcc* cNF-2/5
c
3
0.21
100
5
cMg (mM)
1
100 1 2
σc (Pa) cNF (mg/ml)
C
Figure 6. The dependence of cNF1/2G0 on σc. The solid line is the result Figure 7. The dependence of G0/σc×cNF-1/5 on cNF. The solid line is
of Fig. 6. fit to the data and depicts an exponent of 1.54. This is in
a regression Fig. 7.
the result of a regression fit and exhibits an exponent of 0.41. This is in
agreement with the affine thermal model which predicts an exponent of agreement with the affine thermal model which predicts an exponent of
3/2. Closed squares are data obtained with Mg2+, open squares are data 2/5. The inset shows the dependence of G0/σc×cNF-2/5 on cMg and depicts an
obtained with Ca2+, and crossed squares are data obtained with Zn2+. The exponent of 0.21 obtained by a regression fit; this empirical power law
inset shows the dependence of G0 on cNF and depicts an exponent of 2.5 was used to collapse the data and to obtain the 0.41 exponent for the cNF
obtained from regression. This is also consistent with the affine thermal dependence.
model which predicts an exponent of 2.2.
for the stress is a measure of the critical stress σc at which the network
(2)

begins to stiffen. This provides additional evidence that the nonlinear
elasticity of the Mg2+-cross-linked neurofilament networks is due to [2, 22, 25]. Thus, if the network’s linear elasticity is dominated by
the entropy associated with single filament stretching. entropy, we expect the scaling cNF1/2G 0 ~ σc3/2 , where the pre-factor
To explore the generality of this ionic cross-linking behavior, we should depend only on kBT and lp; although the pre-factor will differ
use other divalent ions including Ca2+ and Zn2+. We find that the for different types of filaments it should be the same for different
effects of both of these ions are nearly identical to those of Mg2+; they networks composed of the same filament type and at the same
also cross-link neurofilament networks into weak elastic gels. This temperature, such as ours. Thus, plotting cNF1/2G 0 as a function of σc
lack of dependence on the specific ionic cross-link lends evidence that for different neurofilament networks at the same temperature should
the interaction between filaments and ions is electrostatic in nature. result in collapse of the data onto a single curve characterized a 3/2
This electrostatic interaction would imply that the various ions are power law; this even includes systems with different divalent ions
acting as salt-bridges, thereby cross-linking filaments into low energy or different ionic concentrations. For a variety of divalent ions, we
conformations. find that cNF1/2G 0 ~ σcz, where z = 1.54 ± 0.14 in excellent agreement
The ability to scale all data sets of K'(σ) onto a single universal with this model, as shown in Fig. 6. It is essential to note that the 3/2
curve also provides a means to convincingly confirm that the linear exponent found here is not a direct consequence of the 3/2 exponent
elasticity is entropic in origin. To accomplish this, we derive an obtained in Fig. 5, which characterizes the highly nonlinear regime.
expression that relates the two scaling parameters to each other. For Instead, the plot of cNF1/2G 0 as a function of σc probes the underlying
small extensions δl of the entropic spring, the force required can mechanism and extent of the linear elastic regime.
be derived from the wormlike chain model giving . For a fixed ratio of cross-links R = cMg/cNF, we expect cross-linking

Assuming an affine deformation, whereby the macroscopic sample to occur on the scale of the entanglement length, yielding lc ~ cNF-2/5
strain can be translated into local microscopic deformations, and ac- [25, 27, 30]. Thus, we expect the linear storage modulus to scale with
counting for an isotropic distribution of filaments, the full expression neurofilament concentration as G 0 ~ cNF11/5 [25]. For R = 1000, we find
for the linear elastic modulus of the network is given by an approximate scaling of G 0 ~ cNF25, consistent with the predicted
power law, as shown in the inset of Fig. 6. Interestingly, the stronger
(1) concentration dependence of G 0 may be a consequence of the dense

cross-linking that we observe. Specifically, for densely cross-linked
where κ = kBTlp is the bending rigidity of neurofilaments, kBT is the networks, corresponding to a minimum lc on the order of the typical
thermal energy, and ρ is the filament-length density [2, 25, 27, 28]. The spacing between filaments as we observe here, the model in Eq. (1)
density ρ is also proportional to the mass density cNF, and is related predicts G 0 ~ cNF25 [25]. The agreement with the affine thermal model
to the mesh size ζ of the network by [29]. Furthermore, the in both the linear and nonlinear regimes confirms the existence of an
model predicts a characteristic filament tension proportional to , ionically cross-linked neurofilament gel whose elasticity is due to the
and a characteristic stress pulling out of thermal fluctuations.
The ability of the affine thermal model to explain the elasticity of
www.thurj.org 35
Conclusion
the neurofilament network also suggests that we should be able to
quantitatively extract network parameters from the bulk rheology. We measure the linear and nonlinear viscoelastic properties
Biomechanics
The model predicts that of cross-linked neurofilament solutions over a wide range Mg2+
and neurofilament concentrations. Neurofilaments are interesting
(3)
intermediate filament networks whose nonlinear elasticity has not
and been studied systematically. We show that the neurofilament net-
works form densely cross-linked gels, whose elasticity can be well
(4) understood within an affine entropic framework. We provide direct

where ρ ≈ 2.1!1013 m-2 for neurofilament networks at a concentra- quantitative calculations of lp and lc from bulk rheology using this
tion of 1 mg/mL. This yields a persistence length lp ≈ 0.2 µm which is model. Furthermore, our data provides evidence that Mg2+ acts as
in excellent agreement with previous measurements [31]. In addition, the effective ionic cross-linker in the neurofilament networks. The
we find that lc ≈ 0.3 µm which is close to the theoretical mesh size weaker than expected dependence that we observe suggests that
≈ 0.26 μm; surprisingly, this is far below the mesh size of 4 µm
Mg2+ may be near saturation in our networks. Future experimental
inferred from tracer particle motion [1]. Such particle tracking only work with other multivalent ions is required to better understand
provides an indirect measure: in weakly cross-linked networks, for the electrostatic interaction between filaments and cross-links; this
instance, even particles that are larger than the average inter-filament would lead to a better microscopic understanding of the effects of
spacing will tend to diffuse slowly. electrostatic interactions in the cross-linking of neurofilament net-
To further elucidate the cross-linking behavior of Mg2+, we explore works. Moreover, the effect of divalent ions on the cross-linking of
the dependence of lc on both cMg and cNF, based on Eq. (3-4). Based on networks of other intermediate filaments would also be very interest-
the form of G 0 and σc, we expect that . Assuming that Mg2+ ing to explore.

is acting as the cross-linker and that lc is also the typical distance
between binary collisions of filament chains we would expect that Acknowledgments
, where le is the entanglement length. Thus, for a given

concentration of neurofilaments This work was supported in part by the NSF (DMR-0602684 and
CTS-0505929), the Harvard MRSEC (DMR-0213805), and the Sticht-
ing voor Fundamenteel Onderzoek der Materie (FOM/NWO).
(5)

References
[25]. This yields
1. Rammensee S, Janmey PA, Bausch AR. Mechanical and structural proper-
ties of in vitro neurofilament hydrogels. European Biophysics Journal with
(6) Biophysics Letters 2007;36:661-8.
2. Gardel ML, Shin JH, MacKintosh FC, Mahadevan L, Matsudaira P, Weitz
DA. Elastic Behavior of cross-linked and bundled actin networks. Science
where X is the exponent of the Mg2+ concentration. Naively, 2004;304:1301-5.
we would expect that X ≈ 1 which would imply that doubling the 3. Hinner B, Tempel M, Sackmann E, Kroy K, Frey E. Entanglement, elas-
concentration of Mg2+ would halve the average distance between ticity, and viscous relaxation of actin solutions. Physical Review Letters
1998;81:2614-7.
cross-links. Empirically we find a much weaker dependence on 4. Tharmann R, Claessens M, Bausch AR. Micro- and macrorheological prop-

cMg, where . This weaker dependence suggests that erties of actin networks effectively cross-linked by depletion forces. Bio-
mM concentrations of Mg2+ actually saturate our networks. This is physical Journal 2006;90:2622-7.
5. Storm C, Pastore JJ, MacKintosh FC, Lubensky TC, Janmey PA. Nonlinear
consistent with a calculation of the percentage of Mg2+ ions, which elasticity in biological gels. Nature 2005;435:191-4.
actually act as cross-links. The number of cross-linking ions per 6. Xu JY, Schwarz WH, Kas JA, Stossel TP, Janmey PA, Pollard TD. Mechani-

cubic meter is ; the number density of ions, N in

cal properties of actin filament networks depend on preparation, polym-
a standard 5 mM Mg2+ concentration is N ≈ 30 × 1023. Thus, there erization conditions, and storage of actin monomers. Biophysical Journal
1998;74:2731-40.
is an excess of Mg2+ ions available to act as cross-linkers; this may 7. Kas J, Strey H, Tang JX, et al. F-actin, a model polymer for semiflexible
account for the weak cross-link dependence. A similarly weak depen- chains in dilute, semidilute, and liquid crystalline solutions. Biophysical
dence has been seen previously with actin networks in the presence of Journal 1996;70:609-25.
the molecular motor heavy meromysin where X was found to be 0.4 8. Wong PC, Marszalek J, Crawford TO, et al. Increasing neurofilament sub-
unit NF-M expression reduces axonal NF-H, inhibits radial growth, and

results in neurofilaments accumulation in motor-neurons. Journal of Cell
and thus, , where cA is the actin concentration
Biology 1995;130:1413-22.
and cHMM is the heavy meromysin concentration [32]. Utilizing our 9. Leterrier JF, Kas J, Hartwig J, Vegners R, Janmey PA. Mechanical effects of
empirical power law for cMg, we are able to collapse the curves such neurofilament cross-bridges - Modulation by phosphorylation, lipids, and

interactions with F-actin. Journal of Biological Chemistry 1996;271:15687-
that which is in excellent agreement with the
94.
predicted exponent 2/5, as shown in Fig. 7. The fact that the cross- 10. Kumar S, Hoh JH. Modulation of repulsive forces between neurofilaments
linking distance lc scales directly with cMg further confirms the by sidearm phosphorylation. Biochemical and Biophysical Research Com-
role of Mg2+ as the effective ionic cross-linker of the neurofilament munications 2004;324:489-96.
networks. Thus, our findings demonstrate both the entropic origin 11. Wong GCL. Electrostatics of rigid polyelectrolytes. Curr Opin Colloid In-
of neurofilament network’s elasticity as well as the role of Mg2+ as an terface Sci 2006;11:310-5.
12. Wagner OI, Rammensee S, Korde N, Wen Q, Leterrier JF, Janmey PA. Soft-
effective ionic cross-linker. ness, strength and self-repair in intermediate filament networks. Experi-
mental Cell Research 2007;313:2228-35.

13. Kreplak L, Bar H, Leterrier JF, Herrmann H, Aebi U. Exploring the me- DA. Scaling of F-actin network rheology to probe single filament elasticity
chanical behavior of single intermediate filaments. Journal of Molecular and dynamics. Physical Review Letters 2004;93.
Biology 2005;354:569-77. 23. Bausch AR, Kroy K. A bottom-up approach to cell mechanics. Nature Phys-
14. Kumar S, Yin XH, Trapp BD, Hoh JH, Paulaitis ME. Relating interactions ics 2006;2:231-8.
Biomechanics
between neurofilaments to the structure of axonal neurofilament distribu- 24. Bustamante C, Marko JF, Siggia ED, Smith S. Entropic elasticity of Lambda-
tions through polymer brush models. Biophysical Journal 2002;82:2360- Phage DNA. Science 1994;265:1599-600.
72. 25. Mackintosh FC, Kas J, Janmey PA. Elasticity of semiflexible biopolymer net-
15. Delacourte A, Filliatreau G, Boutteau F, Biserte G, Schrevel J. Study of the works. Physical Review Letters 1995;75:4425-8.
10-nm-filament fraction isolated during the standard microtubule prepara- 26. Fixman M, Kovac J. Polymer conformational statistics 3. Modified Gaussian
tion. Biochemical Journal 1980;191:543-6. models of stiff chains. Journal of Chemical Physics 1973;58:1564-8.
16. Leterrier JF, Eyer J. Properties of highly viscous gels formed by neurofila- 27. Semenov AN. Dynamics of concentrated solutions of rigid-chain polymers
ments in vitro - a possible consequence of a specific inter-filament cross- .1. Brownian-motion of persistent macromolecules in isotropic solution.
bridging. Biochemical Journal 1987;245:93-101. Journal of the Chemical Society-Faraday Transactions Ii 1986;82:317-29.
17. Yao NY, Larsen R, Weitz DA. Probing nonlinear rheology with inertio-elas- 28. Gittes F, MacKintosh FC. Dynamic shear modulus of a semiflexible polymer
tic oscillations. Journal of Rheology 2008;52:13. network. Physical Review E 1998;58:R1241.
18. Baravian C, Benbelkacem G, Caton F. Unsteady rheometry: Can we char- 29. Schmidt CF, Barmann M, Isenberg G, Sackmann E. Chain dynamics,
acterize weak gels with a controlled stress rheometer? Rheologica Acta mesh size, and diffusive transport in networks of polymerized actin - a
2007;46:577-81. quasielastic light-scattering and microfluorescence study. Macromolecules
19. Baravian C, Quemada D. Using instrumental inertia in controlled stress 1989;22:3638-49.
rheometry. Rheologica Acta 1998;37:223-33. 30. Odijk T. On the statistics and dynamics of confined or entangled stiff poly-
20. Rubenstein M, Colby R. Polymer Physics. Oxford: Oxford University Press; mers. Macromolecules 1983;16:1340-4.
2004. 31. Dogic Z, Zhang J, Lau AWC, et al. Elongation and fluctuations of semiflex-
21. Weitz DA, Janmey PA. The soft framework of the cellular machine. Proceed- ible polymers in a nematic solvent. Physical Review Letters 2004;92.
ings of the National Academy of Sciences of the United States of America 32. Tharmann R, Claessens M, Bausch AR. Viscoelasticity of isotropically
2008;105:1105-6. cross-linked actin networks. Physical Review Letters 2007;98.
22. Gardel ML, Shin JH, MacKintosh FC, Mahadevan L, Matsudaira PA, Weitz
www.thurj.org 37
Evolution of marine cyanobacteria in

the Red Sea
Kevin Huang Lin
Harvard College 2012, kevin.lin12@college.harvard.edu
The cyanobacteria Prochlorococcus and Synechococcus are the smallest photosynthetic organisms on
Evolutionary
Earth yet account for two-thirds of the oceans’ photosynthetic reactions. Their abilities to survive in
Biology
diverse environments make them ideal for studies of genetic diversity. This research investigated the
genetic compositions of multiple Prochlorococcus and Synechococcus species isolated from the Red
Sea to determine how the organisms have evolved to nitrogen stress conditions and therefore differ
from cyanobacteria found in other bodies of water. PCR-based sequencing of cyanobacterial DNA was
followed by computer-based analyses to determine the functions of sequenced genes. A phylogenetic
analysis compared the sequences of Prochlorococcus marinus strains 8390-C9 and 13A3 to Synechoc-
occus species 4320-C2 and 4320-C3. The alignments were used to observe genotypic adaptations to
nitrogen stress, causing sequences to differ from P. marinus strain HOT0M-8F9, isolated from the
Pacific Ocean. One hundred ninety-two open reading frames, corresponding to possible genes, were
identified through sequencing and assembly. Of these, 95–100% had highest-significance matches to
other cyanobacterial genes, indicating that the geographic boundaries separating species do not cause
extreme evolutionary divergence. Twenty-six nitrogen assimilating and metabolizing genes (e.g., ami-
notransferases, methyltransferases) were identified in the two Synechococcus sequences. Only one
such gene was found in Prochlorococcus DNA. This lack of nitrogen related genes was surprising due
to high levels of nitrogen stress found in the Red Sea. The ability of Prochlorococcus to survive in
nitrogen-stressed environments despite an apparent lack of necessary genes for nitrogen metabolism
should be further investigated.
Introduction rine microbe, Synechococcus (Figure 1), the species account for two-
thirds of the oceans’ photosynthetic processes.
Since the discovery of the photosynthetic cyanobacteria Prochlo- Studies of these prokaryotes have recently focused on the genetic
rococcus (Figure 1) in the late 1980s, this organism has become a variability between different species and unique strains. Because the
crucial part of research in a wide range of scientific fields. Due to organisms’ habitats are so widespread, researchers have investigated
its dense occupation of habitats from 40°S to 40°N and survival in genetic and genomic divergences between species living in different
waters up to 200 m deep, Prochlorococcus is known to be the most areas around the globe as well as those living at different water depths
abundant photosynthetic organism in the ocean and, presumably, on (Figure 2). Since Prochlorococcus is a photosynthetic organism and
Earth (Partensky, Hess, and Vaulot, 1999). Along with a second ma- thus relies on sunlight to function, studies have shown that strains
Figure 1. Images of magnified individual Prochlorococcus (right) and Synechococcus (left) organisms (Post, 2006a).

Evolutionary
Biology
Figure 2. Map showing widespread distribution of Prochlorococcus strains (Partensky, Hess, & Vaulot, 1999).
living at high-light and low-light environments are differentiated by echococcus, we conducted a sequence of DNA preparation, sequenc-
vast genomic variations (Rocap, et al., 2003). ing, and analysis steps. The main procedures included growing and
Yet the existence of Prochlorococcus and Synechococcus in bodies lysing fosmid-containing E. coli cells, purifying fosmid DNA, shear-
of water with distinct environmental conditions makes it necessary ing and inserting DNA into a plasmid vector, growing and sequenc-
for differentiation between species for reasons other than levels of ing individual clones, and assembling and analyzing DNA sequences
light. In the Red Sea, located 28°N of the equator, annual cycles of for functional and phylogenetic studies. Detailed procedures are
mixing and stratification control the biochemical processes of the shown below.
ecosystem. Periods of nitrogen stress that occur from April to Octo-
ber see nutrient sources to fall toward the bottom of the Sea as a result Bacterial culture growth
of stratification. During these times, the photosynthetic organisms Fosmid-containing E. coli cells were transferred to sterile tubes
inhabiting the photic zone at the surface of the sea must find alternate containing SBCAP medium and grown overnight, continuously shak-
ways of acquiring nitrogen, whether it is through migration or rapid ing. The overnight cultures were then transferred into Erlenmeyer
adaptations. Since the cyanobacteria are not capable of large-scale flasks containing SBCAP medium and an induction solution then
movements due to their minute size of 0.5-2.0 μm in diameter, phe- grown again at the previous conditions. The cells were harvested im-
notypic adaptations allow some species to cope with nitrogen stress mediately after growth by centrifugation, and the supernatant was
through the assimilation of ammonium rather than nitrate (Post, decanted. Cells were stored at -20° C.
2006).
In April 2003 and August 2005, Dr. Anton Post of the Interuni- Isolation and purification of fosmid DNA
versity for Marine Sciences in Eilat, Israel, collected Prochlorococcus The cell pellet from previous centrifugation was homogeneously
and Synechococcus cyanobacteria at varying depths (Synechococcus resuspended and lysed using the QIAGEN Plasmid Midi Kit. The cell
at the seas surface, Prochlorococcus at >80 meters deep) from the Red lysis solution was incubated, centrifuged, and decanted. The superna-
Sea’s Gulf of Aqaba. Using these samples, this research investigated tant was re-centrifuged.
the evolution of these cyanobacteria species in the nitrogen-stressed Next, the QIAGEN-tip 100 device was equilibrated. The tip was
environment of the Red Sea. By sequencing the DNA samples using washed with buffer solution and eluted. The DNA was precipitated by
PCR-based methods, assembling and analyzing the sequences using adding room-temperature isopropanol to the eluted DNA and cen-
computer-based software, and comparing sequenced genes to other trifuging. The supernatant was decanted, and the pellet was washed
strains of cyanobacteria, we seek to understand the genetic evolution with 70% ethanol before recentrifuging. Once again, the supernatant
and variability between species of this major photosynthetic organ- was decanted, and the pellet was subsequently air-dried and dis-
ism along with its ability to survive in unique conditions of nitrogen solved in buffer.
stress. By investigating both the genetic makeups of Prochlorococcus
and Synechococcus and their abilities to adapt to various environ- Shearing and subcloning of high quality DNA into plasmid vector
mental conditions, we aim to better understand the role that the or- A sample of highly purified supercoiled DNA in diH2O was incu-
ganisms play in nutrient cycles and their photosynthetic impacts on bated and centrifuged. Then, the DNA was sheared using a HydroS-
the current issue of global climate change. hear machine. The concentration of sheared DNA was checked using
a spectrophotometer. The sheared DNA sample was repaired using
Methods and Materials DNA Polymerase through incubation and heat inactivation proce-
dures. EtOH was used to precipitate the DNA.
To obtain genetic information from Prochlorococcus and Syn- The resulting pellet was resuspended, and shrimp alkaline phos-
www.thurj.org 39
Coords BlastP excluding environmental Organism
1...564 methylmalonyl-CoA mutase unidentified eubacterium SCB49

580...1782 Malate/L-lactate dehydrogenase Maricaulis maris MCS10
2398...3012 Glutamine amidotransferase class-I: CTP synthase Prochlorococcus marinus str. AS9601
uncultured Prochlorococcus marinus clone
3031...3708 putative organic radical activating enzyme
HOT0M-3ES
3715...4389 hydrothetical ATPase Prochlorococcus marinus str. AS9601
4386...5699 possible p-aminobenzoate synthetase Prochlorococcus marinus str. AS9601
5696...6523 Aminotransferases class-IV Prochlorococcus marinus str. AS9601
6496...7311 putative uroporphyrin-III C-methyltransferase
HOT0M-8F9
Evolutionary
7342...8625 multidrug efflux transporter, MFS family Prochlorococcus marinus str. MIT 9312
Biology
8625...10703 Polyphosphate kinase Prochlorococcus marinus str. MIT 9312

10911...11930 type II alternative RNA polymerase sigma factor, sigma-70 family Prochlorococcus marinus str. MIT 9312
12134...12577 hypothetical protein PMT9312_1791 Prochlorococcus marinus str. MIT 9312
12570...13637 phospho-2-dehydro-3-deoxyheptonate aldolase Prochlorococcus marinus str. MIT 9312
13735...16308 aconitate hydratase B
HOT0M-3E5
16318...17676 putative chloride channe
HOT0M-8F9
17687...18196 formyltetrahydrofolate deformylase
HOT0M-8F9
18590...19672 NAD binding site:D-amino acid oxidase
HOT0M-8F9
19669...21576 molecular chaperone DnaK2, heat shock protein hsp70-2
HOT0M-8F9
20096...20302 hypothetical protein CaO19.9450 Candida albicans SC5314
21694...22554 shikimate / quinate 5-dehydrogenase
HOT0M-8F9
22658...23119 30S ribosomal protein S6 Prochlorococcus marinus str. AS9601
23123...24337 argininosuccinate synthase
HOT0M-8F9
24888...25964 putative phospho-N-acetylmuramoyl-pentapeptide-transferase
HOT0M-8F9
25975...26220 conserved hypothetical protein
HOT0M-8F9
26154...27563 putative sucrose-phosphate synthase
HOT0M-8F9
Excinuclease ABC, A subunit, ATP/GTP-binding site motif A (P-
27766...30669 Prochlorococcus marinus str. AS9601
loop): ABC transporter
30724...32403 DNA repair protein RecN, ABC transporter Prochlorococcus marinus str. MIT 9301
32472...34328 possible protein kinase:ABC1 family Prochlorococcus marinus str. MIT 9301
34329...34901 hypothetical protein P9301_19051 Prochlorococcus marinus str. MIT 9301
Threonine synthase: Pyridoxial-5’-phosphate-dependent enzymes,
34908...36011 Prochlorococcus marinus str. AS9601
beta family
36350...37507 DNA polymerase III, beta chain Prochlorococcus marinus str. AS9601
37569...38216 hypothetical protein A9601_00011 Prochlorococcus marinus str. AS9601
38220...39164 phosphoribosylformylglycinamidine synthetase II Prochlorococcus marinus str. MIT 9301
956...1771 phosphoribosylformylglycinamidine synthetase II Prochlorococcus marinus str. MIT 9301

123...425 hypothetical protein A9601_00011 Prochlorococcus marinus str. AS9601
487...957 DNA polymerase III, beta chain Prochlorococcus marinus str. MIT 9301

1...411 molecular chaperone DnaK2, heat shock protein hsp70-2
HOT0M-8F9
84...911 DNA repair protein RecN, ABC transporter Prochlorococcus marinus str. MIT 9301
Figure 3. Abbreviated annotation of Prochlorococcus strain 8380-C9 orfs, displaying coordinates of gene on the complete contig (left), functional
assignment (middle), and organism from which the functional assignment was derived (right).

phatase was added, incubated, and heat inactivated. The DNA was days, they were run on a 3730 capillary sequencer, and the resulting
precipitated as before and the sample was resuspended in H2O plus DNA sequences were retrieved through e-mail.
Qiagen 10× PCR buffer solution before being transferred into a PCR
tube. dNTPs were added along with Qiagen Taq solution and incubat- Sequence assembly and closure
ed at 72° C for 30 min. A Phenol-chloroform extract was performed The computer software Phred was used to read DNA from trace
to purify the DNA sample using phenol: chloroform: isoamyl solu- files that resulted from DNA sequencing. By reading DNA, calling
tion, followed by an EtOH precipitation and resuspension in H2O. bases, and assigning a quality value for each identified base, Phred
LB/agar gel with ampicillin antibiotic was prepared and poured obtained accurate DNA sequences in the form of nucleotides. Ad-
into Petri dishes. TOPO-TA cloning was performed for the DNA, us- ditionally, Phred trimmed sequences based on quality value to en-
ing pCR 2.1-TOPO Vector, salt solution, and DNA. The solution was sure that only high-quality sequence data were retained. The Phrap
then incubated before adding it to high-quality chemically compe- software was used to assemble contigs from shotgun DNA sequences.
tent cells and mixing with the pipette tip. Then, the cells were incu- The program found overlaps between the fragmented sequences and
bated on ice, heat shocked, and quickly moved back to ice. SOC was assembled the pieces into longer, more complete strands of cyanobac-
Evolutionary
added, and the cells were incubated for 1 hour. Finally, the cells were terial DNA. The GLIMMER program was finally used to identify
Biology
spread on the LBAmp plates, and the plates were incubated overnight coding regions in the sequences by searching for known start and
at 37° C. stop sequences within the DNA bases, thereby finding hypothetical
genes.
Individual clone growth
Kanamycin was added as an antibiotic to Super Broth (SB), a me- Functional analysis and annotation
dium for cell growth. Ninety-six well growth blocks filled partway The NCBI database (http://www.ncbi.nlm.nih.gov) BLAST search
with SB/kanamycin medium were added to each well, and the blocks was conducted to identify hits between cyanobacterial sequences and
were inoculated with cells colonies using toothpicks. The blocks were genes from other organisms in the database. The top matches were
covered with porous membrane seals and incubated for ~20 hours. examined for e-value, percent identity, function, and organism of
The blocks were centrifuged, and the supernatant was dumped origin. These data were recorded on a Microsoft Excel spreadsheet
into a waste container. The inverted blocks were patted gently on pa- (Figure 3). High-scoring comparisons to other genes in the database
per towel to absorb as much media as possible, covered with foil, and were used to determine functional assignment for cyanobacterial
frozen until sequencing preparation. genes based on nearly identical DNA sequences.
Sequencing of individual plasmid subclones Phylogenetic analysis

The polymerase chain reaction was conducted for plasmid sub- Based on functional assignments in Part 7, organisms from which
clones. A master mix of BigDye Terminator, DNA primer, DMSO, many Prochlorococcus and Synechococcus genes had close matches
reaction buffer, and diH2O was aliquoted to each well of the reaction were identified. Genes from Red Sea cyanobacteria were aligned with
plates along with an equal volume of DNA template. The plates were sequences from Prochlorococcus marinus strain HOT0M-8F9 discov-
spun quickly to settle the plate contents and then thermocycled for ered in the Pacific Ocean, which had 10 high-significance matches
60 cycles at 96° C for 10 sec, 50° C for 5 sec, and 60° C for 4 min to in Part 7, to determine how the organisms have evolved differently
increase the quantity of DNA available for sequencing. (Figure 4).
The plates were spun briefly, and the samples were washed with
isopropanol. After air-drying, pellets were resuspended in HiDi
formamide, sealed with foil, and stored in the freezer. Within three Results
6SPQPSQIZSJO***$NFUIZMUSBOTGFSBTF DPC"
.BKPSGBDJMJUBUPSTVQFSGBNJMZUSBOTQPSUFS
1PMZQIPTQIBUFLJOBTF /JUSPHFONFUBCPMJTNSFMBUFEHFOF
Prochlorococcus marinus str.

)05.' SFWFSTFE
$$POUJH
"$POUJH
Synechococcus sp.
$
$
Figure 4. Alignment of cyanobacterial DNA sequences.
www.thurj.org 41
We have obtained results through three main approaches of se- 3. The 95–100% similarity of top hits of orfs to other cyanobacterial
quencing and assembling fosmid DNA, analyzing the functions of species demonstrates the significant genetic similarities between
Prochlorococcus and Synechococcus genes, and phylogenetically ana- cyanobacteria in the Red Sea and related organisms living in vast-
lyzing the studied organism. The main results are described below. ly different environments.
4. The lack of nitrogen assimilation and metabolism genes in
Sequence and assembly Prochlorococcus marinus strains 8380-C9 and 13A3 suggests that
Four cyanobacterial genome sections were successfully se- the organisms lack nitrogen utilization genes and must rely on
quenced. Two Prochlorococcus marinus strains of 8380-C9 and 13A3 ammonium as the sole source of nitrogen.
were analyzed, as well as two Synechococcus clones of 4320-C2 and 5. The presence of multiple nitrogen genes in Synechococcus species
4320-C3. During assembly, the full-length sequences of 4320-C2 4320-C2 and 4320-C3, in contrast, shows that these cyanobacteria
and 4320-C3 were assembled into single contigs (pieces), indicating are well adapted to the nitrogen stress conditions in the Red Sea.
that all available genetic information was processed during labora- When ammonium is unavailable due to stratification in the photic
tory experiments. The sequences of 8380-C9 and 13A3 were grouped zone, they are able to utilize nitrate and nitrite compounds.
Evolutionary
in multiple contigs because connecting section between the contigs

Biology
were not obtained through software analyses. In total, 207,730 base Analysis and Future Research
pairs (bp) of DNA were sequenced and assembled, with an average of
approximately 51,900 bp per organism. Despite the high environmental stress levels found in the Red Sea,
the organisms living in this area do not necessarily demonstrate ex-
Functional analysis treme genetic variations from related species living in other, less ex-
BLAST searches against the NCBI database resulted in the identifi- treme environments. During sequence annotations, the vast majority
cation of 192 total open reading frames (orfs), corresponding to prob- of genes were identified as most similar to other cyanobacterial genes,
able genes. The average number of gene sequences per organism was and sequence alignments to Prochlorococcus marinus str. HOT0M-
48. The most significant matches for all orfs in Synechococcus 4320- 8F9, isolated from the Pacific Ocean, showed few differences between
C2, Synechococcus 4320-C3, and Prochlorococcus 13A3 originated gene layouts.
from other cyanobacterial species. In the Prochlorococcus 8380-C9 The Synechococcus species studied in this project show obvious
sequences, 95% of the orfs (35/37) had top hits to other Prochlorococ- genetic adaptations to nitrogen stress levels in the Red Sea. The abun-
cus strains. Numerous orfs were given functional assignments relat- dance of nitrogen assimilation and metabolism genes within the ge-
ed to nitrogen assimilation or metabolism, a result of high nitrogen nomes of both Synechococcus 4320-C2 and Synechococcus 4320-C3
stress levels found in the Red Sea. Synechococcus 4320-C2 displayed indicate an ability to utilize nitrate and nitrite compounds as sources
ten such genes, and Synechococcus 4320-C3 displayed 16 genes. Only of nitrogen when ammonium is not available.
one nitrogen gene, however, was identified in Prochlorococcus 13A3, However, not all cyanobacteria in the Red Sea demonstrate these
and no genes related to nitrogen assimilation or metabolism were adaptive abilities. Both sequenced Prochlorococcus strains lacked the
found in Prochlorococcus 8380-C9 sequences. nitrogen genes found in Synechococcus species, even though related
sections of each genome were compared. According to these results,
Phylogenetic analysis in relation to uncultured Prochlorococcus Prochlorococcus marinus species 8380-C9 and 13A3 lack the ability to
marinus clone HOT0M-8F9 process nitrate and nitrite compounds. Therefore, they must rely on
The four sequenced cyanobacteria samples, as well as an annotat- ammonium as the main, if not sole, source of nitrogen.
ed sequence of Prochlorococcus marinus str. HOT0M-8F9 obtained During periods of nitrogen stress, these Prochlorococcus organ-
from the GenBank database, were successfully aligned. Prochlorococ- isms must have a way of obtaining ammonium not available in their
cus marinus str. 8380-C9 and Prochlorococcus marinus str. 13A3 both immediate environment. It is possible that the cyanobacteria obtain
contained multiple contigs after sequencing, so contigs 39 and 28, sufficient ammonium through a symbiotic relationship with het-
respectively, were used for alignments. The gene uroporphyrin-III C- erotrophic bacteria. Yet, it is also possible that any nitrogen genes
methyltransferase (cobA) was identified in all five aligned sequences. have been translocated to another section of the Prochlorococcus
This gene was used as the basis of sequence alignments. The gene genome. Thus, the genotypic, as well as phenotypic, adaptations of
polyphosphate kintase (ppk) was identified in four of five aligned se- these organisms to nitrogen stress conditions in the Red Sea should
quences, but was not present in contig 39 of Prochlorococcus 13A3. be further investigated.
The vast majority of nitrogen-related genes were identified in the two
Synechococcus sequences, with the exception of one nitrogen gene in References
the sequence of Prochlorococcus 13A3.
Berman-Frank, I., Lundgren, P., Chen, Y., Küpper, H., Kolber, Z., Bergman,
Conclusions B., and Falkowski, P. (2001). Segregation of nitrogen fixation and oxygenic
photosynthesis in the marine cyanobacterium Trichodesmium. Science,
294, 1534-1537.
The sequence, assembly, functional analysis, and phylogenetic
Coleman, M.L., Sullivan, M.B., Martiny, A.C., Steglich, C., Barry, K., DeLong,
analysis of marine cyanobacteria genomes yielded five main conclu- E.F., and Chisholm, S.W. (2006). Genomic islands and the ecology and
sions: evolution of Prochlorococcus. Science, 311, 1768-1770.
1. The average sequence length of each cyanobacterial genome sam- Capone, D.G. and Carpenter, E.J. (1982). Nitrogen fixation in the marine envi-
ple (51,900 bp) indicates that the analyzed genes represented ap- ronment. Science, 217, 1140-1142.
proximately 3–5% of the entire genome, which can range from 1–2 Dufresne, A., et al. (2003). Genome sequence of the cyanobacterium Prochlo-
rococcus marinus SS120, a nearly minimal oxyphototrophic genome. Proc.
million bp. Natl. Acad. Sci., 100, 10020-10025.
2. The identification of the cobA gene in all sequences proved that Ewing, B., Hillier, L., Wendl, M.C., and Green, P. (1998). Base-calling of auto-
samples were taken from the same relative areas of the genomes. mated sequencer traces using Phred. I. accuracy assessment. Genome Res.,
Thus, alignments and comparisons were possible. 8, 175-185.

Grula, J.W. (2005). Evolution of photosynthesis and biospheric oxygenation Palenik, B., et al. (2003). The genome of a motile marine Synechococcus. Nature,
contingent upon nitrogen fixation? Intl. J. Astrobiol., 4, 251-257. 424, 1037-1042.
Invitrogen (2007). Comments for pCR2.1 3929 nucleotides. URL www.invitro- Partensky, F., Hess, W.R., and Vaulot, D. (1999). Prochlorococcus, a marine
gen.com/contents/sfs/vectors/pcr2_1_map.pdf photosynthetic prokaryote of global significance. Microbiol. and Molec.
Invitrogen (2006). TOPO TA cloning® five-minute cloning of Taq polymerase- Biol. Rev., 63, 106-127.
amplified PCR products. Carlsbad, CA: Invitrogen. Post, A. (2006a). All stressed out?!: the molecular ecology of marine cyanobac-
Josephine Bay Paul Center for Comparative Molecular Biology and Evolution teria. Marine Biological Laboratory, Interuniversity Institute for Marine
(2004). JPBC protocol for 1/16 X sequencing reactions. URL http://jbpc.mbl. Sciences, Eilat, Israel.
edu/SeqFacility/Pages/Protocol1_16/protocol1_16.html Post, A. (2006b). Guest Lecture. MBL Summer Fellows Symposium. Meigs
Josephine Bay Paul Center for Comparative Molecular Biology and Evolution Room, Swope Center, Marine Biological Laboratory, Woods Hole, MA.
(2004). Post-reaction clean up & resuspension. URL http://jbpc.mbl.edu/ Post, A. (2005). Short protocol for high yield preparations of fosmid DNA.
SeqFacility/Pages/Precipitation/precipprotocol.html Marine Biological Laboratory, Woods Hole.
Josephine Bay Paul Center for Comparative Molecular Biology and Evolu- Robinson, D.H., and Lafleche, G.J. (2000). Essential Molecular Biology: A
tion (2006). JBPC Protocol for automated template preparations using practical approach, Volume I. Oxford University Press, New York.
BiomekFX.URL http://jbpc.mbl.edu/SeqFacility/Pages/Templateprep/ Rocap, G., et al. (2003). Genome divergence in two Prochlorococcus ecotypes
Evolutionary
templateprepprotocol.html reflects oceanic niche differentiation. Nature, 424, 1042-1047.
Lesser, M.P., Mazel, C.H., Gorbunov, M.Y., and Falkowski, P.G. (2004). Dis- Salzberg, S.L., Delcher, A.L., Kasif, S. and White, O. (1998). Microbial gene
Biology
covery of symbiotic nitrogen-fixing cyanobacteria in corals. Science, 305, identification using interpolated Markov models. Nucleic Acids Res., 26,
997-1000. 544-548.
Marcarelli, A.M., and Wurtsbaugh, W.A. (2006). Temperature and nutrient Towe, K.M. (2002). Evolution of nitrogen fixation. Science, 295, 798-799.
supply interact to control nitrogen fixation in oligotrophic streams: an Wilkinson, C.R., and Fay, P. (1979). Nitrogen fixation in coral reef sponges with
experimental examination. Limnol. Oceanogr., 51, 2278-2289. symbiotic cyanobacteria. Nature, 279, 527-529.
Mark-Welch, J., and Dhillon, A. (2006). Fosmid/cosmid subcloning. Woods Xiong, J., et al. (2000). Molecular evidence for the early evolution of photosyn-
Hole, MA: Marine Biological Laboratory. thesis. Science, 289, 724-1730.
National Center for Biotechnology Information (2006). URL http://www.ncbi. Zehr, J.P., et al. (1995). Diversity of heterotrophic nitrogen fixation genes in a
nlm.nih.gov marine cyanobacterial mat. Appl. Environ. Microbiol., 61, 2527-2532.
www.thurj.org 43
Biomechanical response of the in

situ primate lens in its natural versus
empty state as assessed with an ex vivo
accommodation simulator
Isabelle Dortonne
Harvard College 2012, idortonne12@college.harvard.edu
In ophthalmology, accommodation is the act of focusing from distant to near objects. According to
the Helmholtz theory of primate accommodation, the ciliary muscle contracts and moves anteriorly
inward as the zonules relax. Zonules are fibrous strands that connect the ciliary body with the
crystalline lens. Presbyopia, or near vision, is the age-related loss of accommodative ability and is the
most common refractive disorder of the elderly. Lens and capsule-based theories of presbyopia assert
Health and
Medicine
that the decrease in accommodative amplitude can be attributed to increased hardening of the lens
substance and decreased elasticity of the lens capsule with age. One novel technique in the restoration of
accommodation is Phaco-Ersatz, or lens refilling. The promise of Phaco-Ersatz as a treatment for near
vision can be assessed by characterizing the biomechanical properties of the lens in its natural versus
its empty state. Postmortem cynomolgus monkey, rhesus monkey, and human eyes of varying ages were
stretched in their natural and empty states in an ex vivo accommodation simulator in eight, 0.25mm
steps, which mimics the changes in zonular tension that occur in vivo. Diameter-force relationships
were graphed. There was no relationship between the empty-bag diameter slope and age, indicating
that the lens capsule’s mechanical properties do not change the setting of accommodation. Moreover,
the ratio of the load-diameter slope for empty capsule to natural lens decreased significantly with
age, showing that it is the lens material and not the capsule that contributes to presbyopia. The results
confirm the postulation that accommodation can be efficiently conducted as long as the lens contents
have proper viscoelastic properties. Thus, Phaco-Ersatz is a viable future treatment for presbyopia.
Introduction Among the many contributors to accommodation in the primate lens

are the lens material, lens capsule, ciliary muscles, ciliary body, and
Accommodation is the act of focusing from distant to near objects vitreous (Parel et al., 2006). Particularly, the lens capsule and zonular
(Bron et al., 1997). According to the Helmholtz theory of primate fibers play a significant role in transmitting the force of ciliary muscle
accommodation, the ciliary muscle contracts and moves anteri- contraction to the lens fibers during accommodation (Krag & An-
orly inward as the zonules relax. This action results in an increase dreassen, 2003).
in the anterior and posterior curvatures of the lens. These lens shape Presbyopia refers to the age-related loss of accommodative ability
changes during accommodation increase lens refractive power, the and is the most common refractive disorder of the elderly. It usually
ability of the lens to focus incoming light rays (Helmholtz, 1855). manifests itself when the patient is about 40 years old. Symptoms
Figure 2. In a non-presbyopic eye, the image falls directly on the retina

Figure 1. Accomodation in the normal eye. Parel, et al. (back of the eye). In presbyopia, the image projects behind the retina.
become progressively worse, and soon magnifying glasses or bifocals The lens capsule’s elastic properties have been measured by several
are required to see small objects (Werner et al., 2002). Most likely, researchers. Krag and Andreassen used a uniaxial procedure to
presbyopia is initiated by a medley of physical and anatomical changes produce load-deformation and stress-strain curves for the human
in the lens and other structures involved in accommodation (Parel lens capsule in relation to age. They found that the lens capsule
et al., 2006). There are three major theories of presbyopia: lens and becomes less elastic and more rigid with age (Krag & Andreassen,
capsule based theories; extralenticular theories, which focus on the 2003). Bowman, a pioneer in the study of lens capsule elastic prop-
ciliary muscle and choroid; and geometric theories, which examine erties, demonstrated the ability of the lens capsule to return to its
the changes in the geometry of zonular attachments (Werner et al., original shape after being swollen and punctured (Bowman, 1849).
2002). One of the most accepted theories of presbyopia is based on Various studies have reported the elastic modulus under diverse
changes that occur in the lens and lens capsule. The lens substance strain levels. It has been found that the lens capsule modulus in the
also becomes harder with age. In addition, the lens capsule loses young primate is more than one hundred times that of the lens sub-
elasticity, which compounds its inability to mold the more rigid lens stance, which contributes to its role as a force transmitter. Although
substance (Fincham, 1937). the reported exact values of lens elasticity in the literature vary, it is
The lens capsule is the membrane that surrounds the crystalline known that lens capsule extensibility decreases with age. This may be
lens and serves as an anchorage point for the zonular fibers. The caused by the increasing volume of lens, which stretches the collagen
zonules, in conjunction with the lens capsule, mold the crystalline network structure and limits further deformation. Cross-linking of
lens contents during the process of accommodation (Courtois, 1987). the molecular network structure may also limit further deformation
Fincham stressed the significance of the elastic lens capsule in shaping (Krag & Andreassen, 2003).
the lens into its accommodated form. In fact, lesser mammals with a Currently, there are over 2 billion presbyopes worldwide suffer-
uniform lens capsule thickness have limited accommodative ability ing from the age-related loss of accommodative ability. One novel
(Fincham, 1937). technique in the restoration of accommodation is Phaco-Ersatz, or
As humans age, the lens capsule continues to grow anteriorly in lens refilling. The nucleus and cortex are removed from the crystal-
thickness and increases in surface area as the lens contents increase line lens through a small opening in the capsule known as a mini-
Health and
Medicine
in volume. The anterior lens capsule is three to five times thicker capsulorhexis; the capsule, zonules, and ciliary body remain intact.
than the posterior lens capsule. Although the anterior lens capsule is A polymerized gel is then injected into the lens capsule through the
produced by the lens epithelium and continues to grow throughout mini-capsulorhexis (Parel et al., 2006).
life, the posterior lens capsule loses its epithelial cells during the fetal The success of lens refilling relies on the hypothesis that the
stage. Therefore, the thickness of the posterior lens capsule remains elasticity of the lens capsule has a negligible effect on its ability to
constant throughout life (Krag & Andreassen, 2003). mold the lens material. The postulation is that accommodation can
Over time, the structure of the lens capsule becomes more be efficiently conducted as long as the lens contents have proper
homogenous (Courtois, 1987). The density of the lens capsule also viscoelastic properties. This is because it is the lens contents and not
increases with age, partially due to the increase of non-collagenous the lens capsular bag that more greatly maintain accommodation.
amino acids and the decrease of collagenous amino acids (Krag & This hypothesis can be tested by characterizing the biomechanical
Andreassen, 2003). For example, in the adult capsule, collagen IV properties of the lens in its natural state versus its empty state, with
and laminine are abundant, but the glycoprotein fibronectin is scarce the simulation of accommodation.
(Courtois, 1987). Collagen was recently found to be responsible for An ex vivo accommodation simulator can mimic on postmortem
the mechanical strength of various soft connective tissues (Krag & lenses the changes in zonular tension that occur during natural (in
Andreassen, 2003). More specifically, collagen IV is a chief structural vivo) accommodation. Thus, power and diameter-force characteristics
component of the lens capsule (Krag & Andreassen, 2003). can be analyzed. If it is found valid that the biomechanical proper-
Figure 3. Ex vivo Accommodation Simulator (Manns et al., 2007)
www.thurj.org 45
ties of the lens capsular bag do not drastically change with age and 0.25 millimeter steps at a speed of 0.1 mm/s, by the joystick controlled
have a negligible effect on accommodation, then Phaco-Ersatz, the movement of the translation stage. The maximum diameter increase
lens refilling technique, may be an effective treatment of presbyopia of the outer sclera was 4 millimeters.
(Manns et al., 2007). A preconditioning stretch cycle was performed on each lens to
ensure that there were no problems with the tissues, the suture align-
Methods ment, or shoe attachment. At the end of each step, the translation
stage was stopped for 10 seconds to allow a digital picture record-
Preparation of Eyes ing of the load (force) and lens diameter changes. Each stretching
Eyes of the cynomolgus monkey (n = 26, 4–10 years, <4 days post- experiment was conducted and recorded a minimum of three times.
mortem), rhesus monkey (n = 10, 1–11 years, <5 days postmortem) A trigger pulse sent to the system signaled the start of a new cycle.
and human (n = 29, 8–76 years, <5 days postmortem) were obtained Stretching was performed on the natural lens and on the empty lens
from the University of Miami Division of Veterinary Resources and capsule after phacoemulsification or hydrodissection.
Florida Lions Eye Bank. All animal experiments abided by the ARVO
Statement for the Use of Animals in Ophthalmic and Vision Research; Diameter Measurements
all human eye experiments abided by the Declaration of Helsinki for Sharp, high-contrast images of the lens were provided by diffuse
research involving the use of human tissue. Immediately following retroillumination between steps. A digital picture of the top view of
euthanasia, the eyes were enucleated, wrapped in gauze, placed in a the lens was taken at each step during the stretching cycles. The lens
closed jar, and stored at 4 °C until experimentation. and ciliary body diameters were measured along two perpendicular
axes, one horizontal and one vertical, in units of pixels, using Canvas 9
Dissection imaging software. A calibration factor was determined by measuring
The clean scleral surface of the eye was exposed after removal of a picture of a ruler with 1mm divisions, taken during the stretching
the conjuctiva, episcleral tissue, and extraocular muscles. Fluid on experiment, in Canvas 9 software. After recording the length of the
the scleral surface was blotted with a surgical sponge and the scleral lines in Canvas 9, the measurements were divided by the calibration
Health and
Medicine
tissue was bonded, using cyanoacrylate adhesive, to eight custom- factor and inserted into a Microsoft Excel chart. The average of the
made scleral shoes made of plastic (polymethylmetacrylate, PMMA). horizontal and vertical diameters was used as the measure of the
The resulting arrangement was a ring which covered the circumfer- diameter.
ence of the (eye) globe along the equator. During bonding, the shoes The continuous load values were recorded in an Excel chart during
were held in place with fixation pins and a PMMA alignment ring. stretching and later graphed. The corrected (calibrated) load measures
After the adhesive dried, the posterior portion of the eye was cir-
cumferentially dissected and excess vitreous was removed. Extreme
caution was taken to ensure that the anterior vitreous, hyaloid mem-
brane, and ciliary body remained intact. Surgical sponges were used
to carefully remove vitreous residue on the lens (Figure 4).
The eye was removed from the alignment ring and fixation pins,
and transferred to the tissue chamber of the EVAS system, anterior
face up. The EVAS contains eight hooks that fit through a hole in each
shoe (Figure 4). During transfer, the eye was immersed in DMEM
(Dulbeco’s Modified Eagle Medium), a preservation medium, in order
to prevent dehydration. Next, the cornea was removed at the limbus
with scissors and the iris gently pulled off by its root. A diamond
knife was used to make incisions in the sclera between adjacent shoes,
producing eight independent segments.
Removal of Lens Contents

By means of hydrodissection, and phacoemulsification in cases of
stiff lens contents, the surgeon removed the lens contents through a
minicapsulorhexis, a small circular opening in the periphery of the
lens capsule (Tahi et al, 1999). Hydrodissection involves the injection
of saline (Balanced Salt Solution, BSS), via a blunt cannula, through
the minicapsulorhexis between the lens capsule and the cortex. The
liquid separates the lens cortex from the surface of the capsular bag
so that the lens contents can be removed while leaving the capsule Figure 4. Dissection and
intact. Phacoemulsification uses a piezoelectric motor connected to a Preparation. a. Eye mounted
titanium tip to ultrasonically break the hard lens nucleus into small on alignment ring and fixating
fragments, allowing for easier aspiration (Buratto et al., 2003). pins for dissection; b. Removal
of posterior portion of the eye;
EVAS Preparation and Stretching c. Bonding to scleral PMMA
Each shoe was connected by 6-0 nylon monofilament sutures to a shoes; d. Removal of cornea,
T-shaped bar, which was mounted on a translation stage powered by iris, and vitreous; incisions made
a stepper motor. Each string also ran through the two sets of pulleys between shoes; e. Mounting
to produce the radial stretching forces, equally distributed among of tissue onto translation stage
the eight shoes. EVAS was programmed to stretch the lens in eight, and shoe attachment to EVAS.

Health and
Medicine
Figure 5. Typical Load-Diameter Figure 6. Mechanical properties of all specimens as functions
Response of age.
CY 124-1267 OD Empty Bag
Age: 64 months PMT: 3hrs of force required to elicit a diameter change in the empty lens capsule
is not dependent on age. Although previous research has demon-
strated a decrease in lens capsule elasticity in age, such changes have
shown to be insignificant in the setting of accommodation. Thus,
were obtained from the plateaus of the graph, each signifying another loss of elasticity of the lens capsule is not a significant contributing
EVAS step. This same analysis was used for natural and empty lens factor to a decrease in accommodative ability. Moreover, it has been
capsule experiments. shown that the lens capsule of presbyopic humans (>45) is sufficiently
Lens capsule stretch was defined by the equation: elastic to easily deform the novel polymers designed for lens capsule

refilling, allowing for restoration of a minimum of 5 diopters of ac-
commodation in patients undergoing the Phaco-Ersatz procedure.

The ratio of the empty capsule load-diameter slope to the natural
Various load-diameter relationships were graphed using Origin lens load-diameter slope decreased significantly with age (p < 0.0001).
software. The slopes of the linear regressions of the resulting scatter It is more difficult to change the shape of the older natural lens than
plots and corresponding ratios were calculated and then analyzed as the younger natural lens, whereas this is not the case with the empty
a function of age. lens capsule. More force is required to get the same results in older
primates as when less force is applied to younger primates.
Results and Discussion The results support the hypothesis that accommodation can be
efficiently conducted as long as the lens contents have proper vis-
In all species studied, the unstretched empty bag diam- coelastic properties because it is the contents, and not the lens cap-
eter remained constant as a function of age (d = 8.1 ± 0.3 mm cyno, sular bag, that more greatly maintain accommodation. The findings
8.8 ± 0.5 mm rhesus, and 9.8 ± 0.3 mm human). Although the anterior are contrary to the capsule based theories of presbyopia, which state
lens capsule thickness increases with age, the surface area of the empty that changes in the lens capsule’s mechanical properties significantly
lens capsule of all three primate groups remained constant through- decrease accommodative amplitude. The findings promote the feasi-
out life (SA = 103.5 ± 6.9 mm2 for cynomolgus, 122.3 ± 13.7 mm2 bility of regularly using Phaco-Ersatz, a lens refilling technique, as a
for rhesus, and 149.6 ± 9.7 mm2 for human). The ratio of empty bag treatment of the millions of presbyopes worldwide.
surface area to natural lens surface area decreased with age, although
less significantly in the cynomolgus group, which had a smaller age Acknowledgements
range. These results are plausible because the empty bag surface area
stays constant, whereas the lens surface area increases with age. The The full version of this work appears in Investigative Ophthalmol-
ratio is constantly less than one, indicating in situ tension even in ogy & Visual Science 49:4490-6 (2008). It is authored by Noël M. Zie-
young animals. barth, David Borja, Esdras Arrieta, Mohamed Aly, Fabrice Manns,
There was no relationship between empty bag load-diameter re- Isabelle Dortonne, Derek Nankivil, Rakhi Jain, and Jean-Marie
sponse and age. Load values ranged from about 5-35 g/mm in all eyes. Parel. The work was conducted at the Ophthalmic Biophysics Center,
The absence of an observable relationship indicates that the amount Bascom Palmer Eye Institute of the University of Miami Miller School
www.thurj.org 47
of Medicine in Miami, Florida. (Eds.), Presbyopia/Recent Research and Reviews from the 3rd International
Symposium. New York: Fairchild Publications. 45-52.
Fincham, E. F. (1937). The mechanism of accomodation. Br. J. Ophthalmol., 8,
References 5-80.
Helmholtz, H. (1855). Über die accommodationes des auges. Arch. für Opthal-
mol., 1, 1-74.
Bowman W. (1849). Lectures on the parts concerned in the operations on
Krag, S., & Andreassen, T. T. (2003). Mechanical properties of the human lens
the eye, and on the structure of the retina (pp. 61-72). London: Longman,
capsule. Prog. Retin. Eye Res., 22, 749-767.
Brown, Green, and Longmans.
Manns et al. (2007). Optomechanical Response of Human and Monkey Lenses
Bron, A. J., Tripathi, R. C., & Tripathi, B. J. (1997). The posterior chamber and
in a Lens Stretcher. Invest. Ophth. Vis. Sci., 48(7), 3260-3268.
ciliary body. Obrecht, & Gerard, Eds. Wolff’s anatomy of the eye and orbit
Parel, J., Manns, F., Ho, A., & Holden, B. (2006). Refractive surgical procedures
(8th ed., pp. 367-370). London: Chapman and Hall Medical.
to restore accommodation. D. Azar, D. Gatinel, & T. Hoang-Xuan, (Eds.),
Buratto, L., Barboni, P., & Firrincieli, R. (2003). Developments in Cataract Sur-
Refractive Surgery (pp.501-510). 2nd ed. Philadelphia: Mosby-Elsevier.
gery. L. Buratto, L. Werner, M. Zanini, & D. Apple (Eds.), Phacoemulsifi-
Tahi H, Fantes F, Hamaoui M, Parel J-M. (1999) Small peripheral anterior con-
cation: Principles and techniques (pp. 1-2, 93-96). Thorofare, NJ : Slack Inc.
tinuous curvilinear capsulorhexis. J. Cataract Refr. Surg., 25, 744-747.
Burrato, L., Zanini, M., & Savini, G. (2003). Hydrodissection. L. Buratto, L.
Werner, L. P., Werner, L., Pandey, S. K., & Apple, D. (2002). Physiology of Acco-
Werner, M. Zanini, & D. Apple (Eds.), Phacoemulsification: Principles and
modation and Presbyopia. Agarwal, A. (Ed.), Presbyopia: A Surgical Review
techniques (pp. 1-2, 93-96). Thorofare, NJ : Slack Inc.
(pp. 29-60). Thorofare: Slack Inc.
Courtois, Y. (1987). The capsule of the crystalline lens. L. Stark, & G. Obrecht,
Health and
Medicine

Properties of silk III fibroin at the air-

water interface
Senan Ebrahim
Harvard College 2012, sebrahim@fas.harvard.edu
The silk macromolecule fibroin is used extensively in biomedicine in the silk I and II conformations, which have
well characterized structures and properties. It was hypothesized that the fibroin crystalline structure silk III,
which forms exclusively as a nanofilm at certain interfaces, could have similarly applicable properties, such as high
tensile strength, biocompatibility, and liquid crystallinity. This research qualitatively and quantitatively assessed key
properties of silk III fibroin at the air-water interface to facilitate future manufacturing and biomedical applications.
Mechanical strength and viscoelastic properties were studied. Silk III films of varied concentrations were subject to
deformation experiments in which the time-varying moduli were analyzed. Both the time at which the crystallization
began and the time at which the tan-delta value of the film reached unity were discovered to approximately follow the
diffusion model, dependent on surface concentration. Additional data was collected regarding conditions requisite
to silk III crystallization. A preliminary biocompatibility assessment, in which human fibroblast cell cultures were
grown to confluence on silk III nanofilms, affirmed possible future use of silk III in human tissue.
Introduction Background
The unique properties of silkworm (Bombyx mori) silk give Fibroin is a fibrous protein from the silk of silkworms (Bombyx
the material tremendous promise in modern applications such as mori) that is studied intensively for its desirable mechanical properties
regenerative medicine. Its biocompatibility, mechanical strength, such as tensile strength and biocompatibility. The silk emitted from
and natural occurrence are all increasingly valuable features for a the glands of Bombyx mori in pure form consists of fibroin strands
host of applications. In July 2007, scientists at Serica Technologies coated in serecin, a globular protein which gives natural silk its sheen
Materials
were recognized for the repair of anterior cruciate ligaments (ACL) and texture. Fibroin is largely responsible for the unusual strength of
Science
in animals using silk fibroin grafts. Their tests successfully demon- natural silk (silk I) due to its secondary structure, which is made of
strated that the fibroin graft met the “demanding mechanical and several crystalline antiparallel β-pleated sheets (Figure 1). The alanine
biological requirements of a functional ACL” (Serica Technologies and glycine rich polymers with amino acid sequences similar to those
Inc., 2007). Silk fibroin has been the subject of many scientific of collagen pack closely because of their β-pleated sheet conforma-
inquiries from biochemistry to material science, but one particular tion, greatly increasing the material's tensile strength.
fibroin structure, silk III, has several properties that have not yet
been examined (Valluzzi et al., 1999a). The viscoelastic properties of
silk III during formation have not yet been the subject of conclusive
study, and its hypothesized crystal structure remains unconfirmed.
Applications of silk III in the biomedical and textile industries
involving the full capabilities of the molecule would be contingent
upon an understanding of the fundamental science behind its crystal
structure, biocompatibility, and tensile strength.
At the air-water interface of aqueous fibroin solutions, a thin
film crystal structure forms with unique conformational properties
(Valluzzi et al., 1999a). However, the precise nature of the conditions
required to form this crystal structure (silk III) remains unknown.
Modulating concentrations and conditions so as to alter the sur-
factant behavior of the fibroin effects a change in the formation of
silk III. In this project, the uncharacterized properties of the silk
III film were examined using a variety of interfacial techniques.
The physical properties of silk III were studied using an Interfacial
Stress Rheometer and a Langmuir Trough at various concentrations
of fibroin. Human fibroblast cells were cultured on a thin film of
fibroin to test the biocompatibility of the surface. X-ray diffraction Figure 1. Secondary structure of fibroin in silk I. This image shows
experiments were used to accurately study the crystal structure of the secondary structure of fibroin, with interdigitation of the alanine and
fibroin molecules assembled at an air-water interface. These data glycine side chains, a property which contributes to tensile strength. The
were used to characterize the crystalline silk III film and assess its β-pleated sheets of fibroin, such as the one above, are integral parts of the
optimal conditions of formation. protein, vital to its structure and strength.
www.thurj.org 49
fibroin configurations both in vitro and in some cases such as the silk
in suture threads, in vivo. However, because medical complications
could arise from variances in crystallinity, the biocompatibility of
fibroin silk III specifically warrants further study. As the femoral
fortification proposal of Meinel et al. (2006) also relies on high tensile
strength of the selected biomaterial, an assessment of the physical
characteristics of silk III would also be helpful in ensuring its reliable
use in human bone, cartilage, ligaments, and any other load-bearing
applications.
In addition to tissue engineering applications, the work of
Hofmann et al. (2006) points to the possibility of using silk fibroin
in drug delivery. Fibroin crystalline casings were prepared from thin
films to contain drugs intended to enter the human bloodstream. It
was concluded that silk structures were desirable in pharmaceuti-
cal applications not only for their typical biocompatibility, but also
because of their crystalline variability. Studying the crystalline
properties of silk III could be helpful in determining its possible use
in drug delivery capsules.
This investigation into the material properties of silk III at an
air-water interface relied on established rheological properties. The
viscous modulus and the elastic modulus can both be calculated by
simple deformation experiments. Elasticity is an index of how much
energy of deformation is stored, and viscosity measures how much
Figure 2. Theorized structural arrangment of fi broin in silk energy is lost. The complex modulus is a collective measure of the
III. The imaging studies conducted by Professor Regina Valluzzi (Tufts dynamic mechanical properties of a material. The shear viscous
University) have led to this hypothesized structure in crystalline silk III. modulus (G'') represents the resistance to irrecoverable deformation
The fibroin bands in this image are formed from parallel helices of fibroin of a material in the complex modulus (G). The shear elastic modulus
strands, leading to a regular structure of silk III, detectable by techniques
(G') is the portion of the complex modulus resisting recoverable
such as X-ray diffraction.
deformation. They are related by the equation in the complex plane:
The molecular mass of fibroin is essential to determining the G = G’ + iG’’
mean molecular area (MMA) of fibroin when spread at the interface. Changes in the two moduli were used as a mode of quantifying
Recent SDS page testing has indicated that it is approximately 300,000 the rheometric properties of the viscoelastic crystalline film.
Materials
Daltons, the mass used throughout this project (Terry et al., 2004).
Science
Fibroin polymers arrange themselves in various solid-state con- Materials and Methods
figurations. Crystalline structures of silk I and II (silk I being the
natural unprocessed silk fiber, and silk II being spun silk fiber) have Preliminary Procedures
been studied in great detail since these relatively simple structures Silk fibroin was obtained in aqueous solution, about 10% by weight.
were identified (Jin et al., 2003). A new silk structure, now called The heterogeneity of the solution initially presented difficulties when
silk III, was first observed at the air-water interface and later at assessing the film rheology, so the fibroin solution was subsequently
organic-water interfaces (Valluzzi et al., 1999a). The theoretical centrifuged at 13000 rpm for 30 minutes. Varying amounts of fibroin
model proposed for the structure of silk III based on preliminary and de-ionized water (Millipore), or fibroin and 0.1 M acetic acid,
X-ray diffraction studies lacks the traditional pleated sheet structure were then pipetted into vials in solutions of definite concentration
between polymers, as is typically seen in other crystalline silk struc- and refrigerated in storage. Concentrations of fibroin in water were
tures. Valluzzi hypothesized that an asymmetric three-fold helical prepared daily ranging from 1.0 mg/mL to 20 mg/mL. The millipore
conformation with axially aligned helices, as shown in Fig. 2, was water used had a resistivity of 18.2 Ωcm, and was generally devoid of
forming in the film (Valluzzi et al., 1999b). confounding impurities and ions.
Fibroin structures silk I and silk II have numerous applications
in the textile, cosmetics, and biomaterials industries. Silk fibroin Isothermal and Imaging Experimentation
in its natural form (silk I) has been used as a thread for suturing Numerous isothermal compressions of the fibroin film were con-
wounds and incisions for years. The greatest potential of silk III lies ducted as a preliminary investigation of the crystallization process
in its use as a biomaterial in skin, internal tissue, bone, or cartilage, of fibroin. A basic compression experiment was set up with a Teflon
if it retains the tensile strength of silk I and II. It might also present interfacial Langmuir-Blodgett trough (172.0 cm2 ) and two equally
valuable bulk fluid properties such as liquid crystallinity with many sized Delrin barriers. Delrin surface-compressing barriers were used
applications. The work of Altman et al. (2007) with silk II in liga- instead of Teflon ones, to optimize the compression of the surface
ment construction is one of many promising applications of silk in film since the relatively hydrophilic Delrin would allow less leakage
medicine. This work by Altman et al. (2007) regarding fibroin-based than a Teflon contact. A sterilized platinum Wilhelmy plate was sus-
ligament grafts and the work by Meinel et al. (2006) that involved pended from a force detection device (KSV Instruments, Helsinki,
fibroin patches for human femur defects indicate the possibility of FI) to measure the surface pressure at the center of the film on the
using silk-based biomaterials in medicine. Biocompatibility is a trough during the compression (see Fig. 3a).
critical factor that determines whether silk III can ever be used in a The trough was filled until a large convex meniscus formed, using
similar manner. Human cells are known to be compatible with many about 120 mL of Phosphate Buffer Saline solution (PBS). PBS solution
at ten times standard saline concentration (10× PBS) was prepared
a growth on a silk III film was conducted. A silk III film was transferred
from the air-water interface onto a thin glass slide using 5 mg/mL
aqueous fibroin and the Langmuir Blodgett deposition technique.
After drying in a desiccator for 24 hours, the films were seeded with
primary adult Homo sapiens dermal fibroblast cells. The cells were
grown to confluence in media before being transferred to the plates
with the silk III films.
The following chemical components from Invitrogen (Gibco)
were mixed in solution to make the cell culture media: 5 mL 100×
penicillin/streptomycin, 5 mL 100× glutamine, 5 mL MEM (meth-
oxyethoxymethyl) Non-Essential Amino Acids, 5 mL 100× sodium
b pyruvate, 500 mL D-MEM (high glucose) without dye, 50 mL FBS
(fetal bovine serum).
The fibroblast cell concentration was estimated via a random
count procedure, and the seeding density of the cultures on the two
plates was 46 cells/mm2. The fibroblast cells were grown to confluence
on both plates in a humidified 95% air, 5% CO2 incubator at 37° C.
Interfacial Stress Rheometer

To make a detailed assessment of the formation process of the
silk III crystalline structure at the air-water interface, an Interfacial
Stress Rheometer (ISR) was employed in multiple experiments. The
Figure 3. a. Schematic diagram of Langmuir trough for isotherm experi- ISR measures the fluid properties of the film, specifically the viscous
ments. This diagram displays the elements of the Langmuir trough setup and elastic moduli (G's and G"s respectively). Th e kinetics of film
used to procure surface pressure curves for isothermal compression pro- formation was monitored by varying the initial concentration. Both
cesses. KSV software sampled the surface pressure through the Wilhelmy centrifuged and uncentrifuged fibroin samples were used to test the
plate approximately ever y twenty seconds. The delrin barriers were used effects of varying concentration on silk III film formation.
to compress the film during crystallization; b. At left: Interfacial Stress The ISR consists of two magnets in a Helmholtz condition creating
Rheometer (ISR) setup (LOT-Oriel Group, Europe <http://www.lot-oriel. a uniform magnetic field with a gradient of zero in the center set up
com>). At right: Schematic diagram of ISR with labeled components. The to analyze the rheology of an interface. Th e objective is to measure
ISR was used to sample the moduli of a film at the air-water interface over viscosity and elasticity by moving a needle resting on the film with a
time. The magnetic field generated by the Helmholtz coils moved the mag- controlled magnetic field and observing the response of the film to the
netic rod (a needle) which deformed the film. The film’s response to the deformation.
Materials
Science
deformation was measured and quantitatively reported by KSV software
A cleaned Langmuir-Blodgett trough fi lled with 10× PBS was
as the viscous and elastic moduli of the film.
situated centrally between the Helmholtz coils (see Fig. 3b at left). A
as a substitute for water as the supporting subphase in all interfacial glass channel (8 mm internal width) was inserted centrally between
experiments to amplify the amphiphilic nature of fibroin. After 45 the coils, resting on the trough, allowing the PBS to flow through it
minutes, the PBS surface was vacuumed to clean the interface for the freely. A high-intensity focused white light shone on the channel (see
fibroin film. Thirty microliters of fibroin solution was then spread Fig. 3b at right). A platinum Wilhelmy plate was also used to measure
evenly on top of the subphase with a syringe. Thirty microliters was the surface pressure of the film simultaneously.
calculated to be the approximate amount needed to form a nanofilm The air-water interface was swept with a vacuum and prepared for
of single molecule thickness on the surface of the trough. After a the fi broin fi lm. A magnetized Teflon-coated ferromagnetic needle
further 45 minutes, the barriers were closed inwards to a displace- was placed on the surface of the PBS in the channel, held in place
ment of 135 mm, each at a rate of 1.5 mm/min, uniformly diminishing by the constant magnetic field of the Helmholtz coils. A magnifying
the trough area and increasing the surface pressure at a variable rate. camera (A601I, Edmund Industrial Optics) situated above the trough
Several isothermal compressions were performed, varying PBS and was focused on the needle edge and connected to a PCI card frame-
fibroin concentrations, depositing centrifuged and uncentrifuged grabber. Th e entire setup was enclosed in a Plexiglas casing (Thor
fibroin, depositing dilute fibroin dissolved in acetic acid instead of Labs Inc., Newton, NJ) to prevent air currents from disrupting the
water. interface.
Multiple imaging techniques were attempted in order to visu- A sterilized syringe was used to procure 55 μL of prepared fibroin
ally characterize the films, but the only one that rendered useful from a vial, which was spread on the vacuumed interface. The elastic
information was Atomic Force Microscopy (AFM). Brewster’s Angle and viscous moduli of the film were sampled every 20 seconds by the
Microscopy (BAM) was unsuccessful because of the large nature of software. Th e experiment was stopped when the elastic and viscous
the domains of orientation, and a dichroism setup failed because the moduli had clearly stabilized. Varying the concentration of fibroin
dye (Sirius red) exhibited poor adhesion to the fibroin molecules. spread was hypothesized to affect the time needed for silk III crystal-
However, AFM conducted on glass slides of vertically deposited silk lization to proceed. Several concentrations from 2.5 mg/mL to 20 mg/
III films produced accurate visual representations of the structure mL were tested multiple times to ensure reproducibility.
under study.
Results
Biocompatibility Assessment
In order to determine whether or not use of silk III is viable in Tests with fibroin prepared via centrifugation revealed greater re-
humans as a biomaterial, a preliminary in vitro study of human cell producibility than those using uncentrifuged fibroin. The homogene-
www.thurj.org 51
Figure 5. At left: Diffraction results of silk III with corresponding iso-

therm (Kirkwood, J.) The isotherm for the diffraction image displayed at
right was compared to the isotherms of all procedures attempted, in order
to optimize the procedure for silk III formation. Grazing Incidence X-ray
Diffraction results are shown at left (original) and center (integrated),
properly indicated characteristic peaks of the silk III structure (using the
peaks defined by Valluzzi et al.). The integrated diffraction peaks shown at
Figure 4. Isotherms of silk III fibroin under various conditions. center – obtained experimentally – were compared to the definitive peaks,
Above are surface pressure curves for fibroin films made using fibroin at confirming that the structure formed using one particular procedure was
the air-water interface. Scores of films were constructed and subjected to in fact silk III, in turn identified by comparison of the isotherms for all
isothermal compression, under various conditions and using procedures procedures.
varying parameters, in order to assess which procedure best effects silk III
crystallization. Three such isotherm curves are shown above. the peaks on the micron scale displayed (Valluzzi et al., 1999b)
The biocompatibility assessment indicated that silk III is compat-
ity of the centrifuged fibroin solution revealed surface pressure plots ible with human fibroblasts, and does not inherently inhibit cell
indicative of greater crystallization during isothermal compression. growth and function. The fibroblasts on both plates with silk III films
In comparison to the aqueous fibroin isotherm curves generated in successfully reproduced multiple times and grew to confluence after
conjunction with positive X-ray diffraction results, surface pressure five days (see Fig. 7).
curves of fibroin dissolved in acetic acid seemed characteristic of The ISR tests indicated a clear correlation between time of crystal-
gelation instead of crystallization (Terry et al., 2004). The curves as lization and varying concentration. The software's measurements of
displayed in Fig. 4 allow a comparison to determine the effects of the elastic and viscous moduli as time-dependent functions revealed
centrifugation and acetic acid dissolution. curves of typical conformation (see Fig. 8).
The smooth surface pressure curves of uncentrifuged fibroin and The takeoff time, indicated by the black circle in Fig. 8, is an esti-
fibroin in acetic acid, contrast with those of centrifuged fibroin in mation of the starting point for the crystallization of the film at the
water. The procedure most conducive to crystallization was deduced interface. This takeoff time can be approximated at the initial rise of
via Grazing Incidence X-ray Diffraction (GIXD). the viscous modulus from a constant value. The change in viscos-
Materials
Science
The GIXD images displayed in Fig. 5 indicated the optimal ity is a clear indication of a change in the film’s material properties.
procedure to produce crystallization at an air-water interface (Kirk- Fig. 8 demonstrates a clear enough curve (with very little noise) that
wood, personal communication). Structural peaks in the crystalline the takeoff time could be extracted by determining when the first
film as shown in Fig. 5 taken from one of the experimental silk III significant (greater than 1%) deviation from the constant modulus
production procedures matched results of Valluzzi et al. (1999a) (not value occurred.
pictured in the figure). This diffraction pattern resulted exclusively The approach taken here, intended to model the bulk properties
from the procedure of using water as the solvent and centrifuging of an interface, does not have the sensitivity required to create a
the fibroin solution. The centrifuged fibroin isotherm also matched detailed model of the kinetics. Th e crossover time was the second
the isotherms corresponding to the positive diffraction images most characteristic studied in silk III crystallization. The crossover time is
closely (see Fig. 4, Fig. 5). This information helped create a new pro- a mathematically definitive point, analogous to the rheometric quan-
cedure for fibroin experimentation: centrifuged fibroin in water was tity of crossover frequency, at which point, the two moduli intersect
subsequently used in all experiments. (see white circle of Fig. 8).
The AFM conducted on the slides with vertically deposited films of The crossover and takeoff times were recorded for each experi-
fibroin yielded positive results as well. The images displayed (see Fig. ment with different concentrations of fibroin. The takeoff times of
6) are of a bundle of fibroin fibers in a silk III film. The coaxial helix viscous moduli with runs of different concentrations of fibroin were
theory proposed for the structure of silk III supports the aligment of used to determine the effect of initial concentration on the time
Figure 6. AFM images of silk III. At left: top-down view. At right: 3-D Figure 7. Images of fibroblast growth on two plates with silk III
representation. These AFM images provide a detailed view of the structure films. These two magnified images of human fibroblast cells growing on
of silk III as it forms at the air-water interface. The ridges in the 3-D repre- silk III films imply an intrinsic biocompatibility of silk III. Cells reached
sentation support Valluzzi’s hypothesized structure. confluence in five days.
moduli (mN/m) vs. Time (s) indicated by the gradual upward trend in Fig. 12.
10
Discussion
1
The crystallization of fibroin at the air-water interface, and the
modulus (mN/m)
0.1 formation of silk III were confirmed by the GIXD results. It was also
determined in the preliminary experimentation with isotherms that
0.01
acetic acid is not an ideal solvent when working with fibroin at the
air-water interface. Methanol as used by Meinel et al. (2006) could
help the fibroin spread and reduce clumping, an improvement over
0.001
Time (s) acetic acid as a solvent. The AFM indicated general similarities of silk
III to the coaxial model, but more detailed imaging is necessary to
Figure 8. Logarithmic vs. linear view of viscous and elastic validate the theory.
moduli versus time. These moduli are from a film of 5 mg/mL fibroin. The fibroblast cells on both plates grew to confluence in exactly
The viscous modulus (G") is shown in red and elastic modulus (G') in blue. five days—a time scale consistent with many biocompatible materi-
The white circle indicates crossover time, black circle takeoff time. als. As observed by Karamichos et al. (2009), fibroblast cells grow
to confluence within three days in Fetal Bovine Serum (FBS) media
until crystallization begins. The values of both the takeoff time and cultures and within five to seven days in structures of natural human
crossover time were characterized by an apparent decrease with an materials like collagen matrices. However, it is vital to note that while
increase in concentration. The takeoff time variation closely fit an the fibroblast cells grew successfully on the silk III films, their growth
inverse relationship as shown in Fig. 9. only proves a rudimentary biocompatibility of silk III. Fibroblasts
The trend of a generally inverse correlation is corroborated by are relatively durable cells in the human body, since they function as
the plot in Fig. 10 as the values of the takeoff time multiplied by the sturdy connective tissues, and can handle a broader range of condi-
concentration remain roughly constant. tions than more advanced cells and tissues. Therefore, the histotoxic-
A similar analysis conducted for the crossover time indicates a ity of silk III to human tissue is not categorically disproved by these
roughly inverse relationship between concentration and crossover experiments. However, the fibroblast biocompatibility was still an
time as well (see Fig. 11). However, the correlation appears weaker as important determination, as fibroblasts are critical to wound healing,
Takeoff Time (s, Takeoff Time (s,

Fibroin Concentration (mgomL,
Takeoff Time vs. Fibroin Concentration Takeoff Time
Fibroin Concentration vs. Fibroin Concentration
5000 40 000
4000
30 000
Materials
3000
Science
20 000
2000
10 000
1000
Fibroin Concentration (mgomL, Fibroin Concentration (mgomL,

5 10 15 20 5 10 15 20
Figure 9. Scatter plot of takeoff time as a function of concentra- Figure 10. Scatter plot of takeoff time multiplied by concentra-
tion. On the scatter plot above, the inverse function (k/x) fit included, tion as a function of concentration. This plot maintains a visibly
with the optimal coefficient determined by Mathematica. The plot shows near-constant y-value for any x-value. This feature demonstrates the con-
the approximately inverse relationship between the takeoff time and fibroin stant nature of the takeoff time multiplied by the concentration of fibroin
concentration for silk III at the air-water interface. Error bars represent used. Error bars represent one standard deviation with four runs at one
one standard deviation with four runs at one concentration. concentration.
Crossover Time (s,

Fibroin Concentration (mgomL,
Crossover Time
Fibroin Concentration vs. Fibroin Concentration
Crossover Time (s,
Crossover Time vs. Fibroin Concentration
40 000
8000
30 000
6000
20 000
4000
2000 10 000
Fibroin Concentration (mgomL, Fibroin Concentration (mgomL,

5 10 15 20 5 10 15 20
Figure 11. Scatter plot of crossover time as a function of con- Figure 12. Scatter plot of crossover time multiplied by con-
centration. On the scatterplot above, the inverse function (k/x) fit is in- centration as a function of concentration. Th is plot maintains a
cluded with the optimal coefficient determined by Mathematica. The plot visibly near-constant y-value for any x-value. Th is feature demonstrates
shows the approximately inverse relationship between the crossover time the constant nature of the crossover time multiplied by the concentration
and fibroin concentration for silk III at the air-water interface. Error bars of fibroin used. Error bars represent one standard deviation with four runs
represent one standard deviation with four runs at one concentration. at one concentration.
www.thurj.org 53
and silk III could one day be used in various wound treatments. a Plexiglas container, whereas the Langmuir trough and trough used
From the interfacial rheology experiments at varying concentra- for vertical deposition were not. Future experimentation should be
tions, the takeoff times and crossover times are defined by an inverse conducted in a space where environmental factors can be controlled
function. The time for the crystallization of silk III to begin (the with higher degree of sensitivity.
takeoff time) can be characterized by the diffusive model, where the
dependent condition is how long it takes the fibroin molecules to Conclusions
arrange (“diffuse”) themselves into an alignment conducive to silk
III formation. This two-dimensional diffusion relationship between While experimentation served as a preliminary exploration of
the time and length scale of the problem is written: the behavior of fibroin at the air-water interface, much remains to
The variable L is the length scale of the problem, squared to be studied about silk III. The biocompatibility of silk III was only
represent the surface area available to the molecules. The diffusion assessed using fibroblasts, which are resistant to various pathological
constant D is a system dependent value, and tt is the takeoff time. conditions to which other human cells would succumb. Further tests
To simplify the relationship, a new constant A can be defined as on silk III films need to be conducted to determine whether various
k/D. Substituting k/C for L2, as concentration and surface area are kinds of human cells, particularly more delicate tissues and cells with
inversely proportional, and rearranging variables algebraically, a specialized functions are compatible with silk III films.
function can be constructed describing the takeoff time as depending The study of the effect of concentration on the formation of silk III
on concentration: revealed that the crystallization process follows the diffusive model.
The diffusive law is reproduced in the above simplified form so The crystallization of silk III is primarily dependent on the initial
that the relationship between time and concentration becomes clear, concentration of fibroin, with time having an inverse relationship
so that time scale effects of concentration on silk III production may to concentration. The strong correlation, particularly of the takeoff
be analyzed. The above derived inverse proportionality between the times, indicates that the time is dependent on few other factors, and
takeoff time and fibroin concentration matches the inverse function that the chief determinant is concentration.
fit in Fig. 9. This relationship also accounts for the fact that the quan- Further research on silk III crystallization should be carried out
tity C × tt equals a constant At approximately equal to 12,500 mg-s/ using different concentrations of fibroin in water and organic solvents
mL (see Fig. 10. A similar relation is evident for the crossover time, of different strengths. Crystallization might proceed more rapidly,
as deduced from an identical derivation, simply substituting tc (cross- and perhaps with more identifiable domains of orientation, because
over time) for tt (takeoff time). The graph in Fig. 12 indicates that the of newly introduced intermolecular forces between fibroin and ad-
constant Ac equals about 24,000 mg-s/mL. ditional molecules in the organic solvent, keeping the fibroin spread
The application of the diffusive model to the formation of silk III is at the air-water interface and preventing the fibroin from clumping,
significant for any future manufacturing process or other procedure gelling, or forming fibers. Research involving organic solvents such
requiring time-critical formation of silk III. It is quite possible that as hexane or decane as a replacement for air at the interface may also
higher than a certain concentration the diffusive model breaks down, effect faster crystallization of silk III. Further imaging studies of silk
due to formation of a thicker, denser multilayer at the surface and III must also be conducted to view the exact structure of silk III, in
Materials
Science
more complex interactions between fibroin molecules. However, order to validate the coaxial helix theory or propose an alternative
concentrations higher than 20 mg/mL could not be accurately tested model. Additionally, experimental measurements of domain size
because of the physical limitations of the instruments. The film would over time would be important to assess the early stages of silk III
be unusually thick at higher concentrations, and the elasticity was structural formation (before the takeoff time). Optical techniques
beyond the range of the instrument. such as imaging of fluorescently tagged fibroin proteins would have
While the data closely fit the diffusive model, there are other to be used to achieve this end in further study.
possible factors in the formation of silk III at the air-water interface. Silk III is a relevantly recent discovery by comparison to silk I and
During late stage aggregation of fibroin, Lifshitz-Slyzov coarsening II but it holds promise as a material with industrial and biomedical
may be responsible for the behavior of the film as silk III forms in application. The alternative structure of silk III could solve biocom-
domains. Since according to Lifshitz-Slyzov coarsening, diffusion patibility issues that have plagued studies of drug delivery where
and coalescence of droplets follows the mathematical relation that interference of the casing material with bodily function is a serious
the domain size is directly proportional to the cube-root of the time, concern (Hofmann et al., 2006). Using silk III-based scaffolds in tissue
it would have been necessary to measure the film rheology at later reconstruction is also a possibility, given its biocompatibility and the
times for the coarsening to be a significant factor. Silk III exhibited incipient success of silk II-based scaffolds in animals (Altman et al.,
such high moduli at these later times that they could not be prop- 2007). Multilayers of silk III could also eventually be used as textiles,
erly measured by lab equipment in the late stage of film formation. or components of textiles, due to the high elasticity of the crystalline
Therefore, while the diffusive model aptly describes the formation of film observed during rheometric experimentation.
silk III films, the possibility of coarsening and other effects cannot Silk fibroin already has many unique and vital qualities, and silk III
be precluded. may make fibroin even more valuable. Fibroin polymers are large and
Errors in the rheological experimentation may have been caused uniform enough to generate varied stable intermolecular structures,
by changes in the fibroin solution over weeks of experimentation, often with very high tensile strength. Fibroin is also usually biocom-
despite storage at low temperature. Additionally, the Teflon coating patible at the chemical level, as it originates as a natural biomolecule
on the magnetic needle began to wear off at the edge, which slightly and is very similar to the protein collagen which is the most abundant
affected the needle’s interaction with the film as well as the ability in human body. Since a better understanding of silk III is necessary
of the software to detect the needle edge. However, the same needle before applying it to practical uses such as those in which silk I and
remained usable for all measurements, so this was not a confounding II are involved, the goal of this research was to better describe the
error. Errors applicable to fibroin films other than those made in the nature of silk III through experimentation. Its basic properties in
ISR trough include contamination from dust particles in the air and a nanofilm and its manner of formation were analyzed in order to
disturbance from convective air currents, as the ISR was encased in characterize it for future study and application to industry.
The determinations of this study serve as a starting point for References

future experimentation on silk III. Since procedures are not yet
established for working with silk III the same way they are for silk 1. Altman, G. H., Horan, R. L., Bramono, D. S., Simmons, Q., Chen, J., Mor-
I and II, a procedure was developed to optimize crystallization. The tarino, E., Boepple, H. E., Toponarski, I., Collette, A. L., and J. S. Prudom.
biocompatibility assessment ensured that there is a distinct possibil- (2007). Biological and biomechanical assessment of a long-term bioresorb-
ity of using silk III in biomedicine. The rheological analyses provide able silk-derived surgical mesh in an abdominal body wall defect model.
Journal of the American College of Surgeons, 205, S53–S54.
quantitative information about time needed for the crystalline film
2. Hofmann, S., Foo, C. T., Rossetti., F., Textor, M., Vunjak-Novakovic, G.,
to form depending on the initial amount of fibroin added, which Kaplan, D. L., Merkle, H. P. and L. Meinel. (2006). Silk fibroin as an organic
can be used both as a basis for testing other parameters of silk III polymer for controlled drug delivery. Journal of Controlled Release, 111,
crystallization in future research and in manufacture. Before silk III 219–227.
can enter the realm of applications, it is imperative that knowledge 3. Jin, H. J. and D. L. Kaplan. (2003). Mechanism of silk processing in insects
and spiders. Nature, 424, 1057–1061.
about it as a material be advanced. This research successfully assessed
4. Karamichos, D., Lakshman, N., and W. M. Petroll. (2009). An Experimental
various quantitative and qualitative characteristics of silk III fibroin Model for Assessing Fibroblast Migration in 3-D Collagen Matrices. Cell
that remain essential to future study and use of this new material in Motility and the Cytoskeleton 66, 1–9.
potentially revolutionary biomedical applications. 5. Meinel, L., Betz, O., Fajardo, R. Hofmann, S. Nazarian, A. Cory, E., Hilbe,
M., McCool, J., Langer, R.,Vunjak-Novakovic, G., Merkle, H. P., Rechen-
Acknowledgements berg, B., Kaplan, D. L. and C. Kirker-Head. (2006). Silk based biomaterials
to heal critical sized femur defects. Bone, 39, 922–931.
6. Serica Technologies, Inc. (2007). AOSSM Award Public Release [WWW
Many thanks to Dr. Gerald Fuller of Stanford University for Document] URL http://www.sericainc.com/pdf/AOSSM_RELEASE_7_07.
giving me the opportunity to conduct research in his laboratory, as pdf (visited 2007, September 9).
well as to An Goffin and John Kirkwood of Stanford University, and 7. Terry, A. E., Knight, D. P., Porter, D. and F. Vollrath. (2004). pH Induced
Changes in the Rheology of Silk Fibroin Solution from the Middle Division
Dr. Kate Schafer of The Harker School for their invaluable guidance
of Bombyx mori Silkworm. Biomacromolecules, 5, 768–772.
throughout the duration of the study. 8. Valluzzi, R., Gido, S. P., Muller, W. and D. L. Kaplan. (1999). Orientation of
silk III at the air-water interface. International Journal of Biological Macro-
molecules, 24, 237–242.
9. Valluzzi, R., He, S. J., Gido, S. P. and D. Kaplan. (1999). Bombyx mori silk
fibroin liquid crystallinity and crystallization at aqueous fibroin–organic
solvent interfaces. International Journal of Biological Macromolecules, 24,
227–236.
Materials
Science
www.thurj.org 55
Tandem repeats in promoter regions

of S. cerevisiae generate variability
in gene expression with phenotypic
consequences
Bianca Calderon
Harvard College 2009, calderon@fas.harvard.edu
Most genomes are made up of substantial portions of repetitive DNA. In humans, for example, as
much as 49% of the genome consists of such repeats. Tandem repeats, sequences which are repeated
head-to-tail at one specific locus within the genome, are especially interesting because of their high
level of variability. The focus of this study is on tandem repeats which occur in the promoter regions
of genes in Saccharomyces cerevisiae (brewer’s yeast). Here, I show that repeats in promoters are
indeed hyper-variable and often differ between evolutionarily closely related sub-populations of yeast
(natural yeast strains). To test the effect of repeats in promoter regions, I conducted experiments with
two candidate genes: SDT1 and YKL071w. I created constructs with varying number of repeats in the
promoter regions, studied the effect of repeat size on the transcriptional activity of the respective
genes, and found that changes in the number of repeat units in the genes correspond to variation in
transcriptional activity. To test whether these changes in transcription levels resulted in phenotypic
changes, growth assays were conducted with SDT1 strains in the presence of 6-azauracil. These
assays yielded differences in the length of the lag phase that corresponded with the gene expression
of the respective strain. In YKL071w, binding sites for the stress-response transcription factor Yap1
overlap with the variable tandem repeats. Studies of the YKL071w mutant series comparing the gene
expression of strains with a deletion of the YAP1 gene and strains with an intact YAP1 open reading
frame show a significant reduction in expression in the strains with the deletion, suggesting that
transcription factor binding regulates transcription in YKL071w. Together, these results indicate
that just as variable repeats located within coding regions allow swift evolution of protein function,
repeats in promoters may allow quick evolution of gene expression levels to changing environments
and selective pressures.
Molecular
Biology
Introduction Tandem Repeats

Tandem repeats, also known as satellite DNA, are a type of repeti-
Most genomes contain significant portions of so-called “repetitive DNA sequence. This study will focus on these repeats. Tandem
tive DNA” or “repeats”. As much as 45% of the human genome, for repeats are made up of sequences which are repeated head-to-tail
example, consists of such repeats.1 Despite their abundance, repeats at one specific locus within the genome. For example, the DNA
have been historically regarded as nonfunctional “junk” DNA.2 While sequence CGACGACGACGA is a tandem repeat made up of four
many repeats are found in “gene deserts” (i.e. parts of the genome units of three nucleotides (Figure 1). These sequences were named
without any obvious function), some repeats are located within pro- “satellites” because during separation of genomic DNA in a buoyant
moter regions of genes and even within open reading frames.3 The density gradient, repetitive DNA sequences form a secondary, or
prevalence of repeats in genomes suggests that these sequences may “satellite” band on account of the different density of repetitive DNA
have a biological function.4 This study will attempt to characterize compared to the rest of the genome.5
the variability and function of repeats, specifically tandem repeats, Tandem repeats can be divided into two classes: microsatellites
in the promoter regions of the yeast genome. and minisatellites. The first human minisatellites were discovered by
5’….TAGCGTCATGCGAGCTGACTTCGACGACGACGACGATCGTAGAGTCGATCGTACCT.....3’
Figure 1. Types of repeats. In this example, the repeat unit CGA is considered a tandem repeat because it is repeated head-to-tail at one specific
locus.

These repeats, although they are found in non-coding regions, still

Prevalence (Percentage of
Type of Sequence serve a biological function. In humans, short alleles of the tandem
Human Genome)
repeat found in the 1f exon of the gene NOS1 are said to increase
Coding Regions 2 the risk of Alzheimer’s disease.13 In Saccharomyces cerevisiae (S288c),
Repeats 45 tandem repeats occur between the UAS region and the TATA box
in four MAL promoters.14 The tandem repeats reduce expression of
Tandem Repeats 3 maltose permease relative to maltase, thereby preventing possible
toxic effects that can occur through over-expression of the permease
Table 1. Repeats are more prevalent than coding regions in the gene. Repeats in promoter regions have a biological purpose and
human genome. The significant portion of the genome that contains deem further study.
repeats suggests that these sequences serve a biological role.
Central Hypothesis and Goals of This Study
We hypothesize that, just as repeats located within coding regions
A.R. Wyman and R. White.6 Minisatellites have repeat units that are allow swift evolution of protein function, repeats in promoter regions
10-150 bp in length and are typically more stable than microsatellites may allow quick adaptation and evolution of gene expression levels
or short tandem repeats (STRs), which are so named because these to changing environments and selective pressures. We focus on
sequences have a repeat unit length of 1-9 bp, and repeat length of the model eukaryote Saccharomyces cerevisiae because the relative
less than 150 bp.7 Minisatellites are often found near centromeres compactness of the genome, the available genetic toolbox, and the
and telomeres, while microsatellites are distributed throughout the extraordinary knowledge about this organism make it a good model
entire genome.8, 9 for this study. Initial studies completed by the Verstrepen group
Three percent of the human genome is comprised of tandem indicate that more than 1000 S. cerevisiae promoters contain tandem
repeats, which is a larger percentage than that composed of coding repeats (Table 2).
regions (Table 1). Tandem repeats, however, have been tradition- First, we explored the differences between tandem repeats within
ally regarded as junk DNA for a number of reasons. The majority promoter regions found in naturally occurring yeast strains to estab-
of repetitive sequences are located in intergenic regions rather than lish whether repeats in promoters are hyper-variable in nature. We
in coding regions.4 For example, large satellite regions (greater than hypothesized that there would be significant variation in the repeat
100 kb) are normally found near the centromere.8 In humans, alphoid regions between these evolutionarily closely related sub-populations
DNA, a satellite region, occurs at the centromere of each chromo- of yeast.
some.10 Alphoid DNA has a repeat unit of 170 bp, and makes up 3% To investigate the role of repeats in transcriptional regulation, I
of the DNA sequence in each chromosome. The modest number of created constructs with varying number of repeats in the promoter
repeats in coding regions led to the belief that repeats were not useful regions and studied the effect on the transcriptional activity of the
sequences because they do not code for proteins in most cases.11 This respective genes. For this experimental study, I focused on a few
idea was strengthened by the unsuccessful attempts of many scien- specific genes that carry repeats in their promoter regions. Interesting
tists to find a function for these sequences.12 The lack of complexity candidate genes with intra-promoter repeats identified in the in silico
in these repeat regions also led to questions about their biological screen previously performed by the Verstrepen group include:
usefulness or function. It is easy to assume that repeats are too simple t YKL071w, which is induced by stress and contains binding sites
to be of any value. In addition, repeats are extremely unstable and for the stress-response transcription factor Yap1 in the promoter
variable. Some argued that if repetitive DNA was indeed functional, region that overlap with the variable tandem repeats (Table 3).15
then it would presumably be more conserved. This study, however, t SDT1, a stress-induced pyrimidine nucleotidase whose overex-
will attempt to counter these arguments. pression suppresses the 6-Azauracil (6-AU) sensitivity of the
transcription elongation factor S-II and confers resistance to other
Tandem Repeats in Promoter Regions. pyrimidine derivatives (Table 3).16
Molecular
Regulatory (promoter) regions also contain tandem repeats. In order to examine the mechanisms by which tandem repeats
Biology
Repeats located within promoters and other regulatory sequences are affect gene expression, I studied the dependence of gene expression
less well characterized than repeats located within coding regions. levels for YKL071w on Yap1. I hypothesized that the expression levels
Category Number of Genes Prevalence (Percentage)

Genes analyzed 5695
Genes with tandem repeats 1456 25
Genes with variable tandem repeats 307 21
Genes with non-variable tandem repeats 1006 69
Genes with tandem repeats that overlap
146 10
transcription factor binding sites
Table 2. Differences between natural yeast strains in genes containing tandem repeats in the promoter region. 5695 genes were
analyzed in the yeast strains S288c and RM11-1a. 21% of the genes with tandem repeats in the promoter region showed differences in the number of
repeat units between the two strains, demonstrating that tandem repeats are naturally variable. In 10% of the genes with tandem repeats in the promoter
region, the tandem repeats overlap with at least one transcription factor binding site. One hundred forty three of the genes with tandem repeats in the
promoter region did not have enough sequence information available to determine whether the tandem repeats were variable or conserved.
www.thurj.org 57
Gene Repeat Unit Number of Repeats Function

CDC14 TG 31 Protein phosphatase required for mitotic exit.
PRE8 TTA 39.3 Endopeptidase activity.
Pyrimidine nucleotidase; overexpression suppresses the 6-AU sensitivity of transcrip-
SDT1 AT 29
tion elongation factor S-II, as well as resistance to other pyrimidine derivatives.
Late endosomal protein involved in late endosome to vacuole trafficking; functional
VPS55 GT 20.5 homolog of human obesity receptor gene-related protein
(OB-RGRP).
Repressor of G1 transcription that binds to the SCB binding factor (SBF) at SCB target
promoters in early G1; phosphorylation of WHI5p by the CDK, Cln3p/Cdc28p relieves
WHI5 AACAA 11
repression and promoter binding by WHI5;
periodically expressed in G1.
ATTAG- Induced by stress and contains binding sites for the stress-response transcription factor
YKL071w 6.8
TAATGAG Yap1 in the promoter region which overlap with the variable tandem repeats.
Table 3. Genes utilized in this study and their functions. A description of the repeat unit, repeat size, and function of the wild type for the genes
used in this study (Legendre et al., 2007; Saccharomyces Genome Database).
of YKL071w are dependent on Yap1 and that tandem repeats control number of tandem repeats than the wild type. This PCR product was
gene expression through transcription factor binding in this gene. transformed into a wild-type strain of S. cerevisiae and selected for
Together, these three specific aims will reveal how repeats confer on hygromycin B plates. Repeat length in these new strains was de-
variability and “evolvability” of transcriptional regulation in eukary- termined through PCR and gel electrophoresis. This method allowed
otes. us to control the number of repeat units added or deleted.
Materials and Methods “Loopout” Transformation Strategy

The gene URA3 was inserted within the tandem repeats of the
Strains and Growth conditions promoter region of the chosen gene in S. cerevisae by transformation,
The strains used are listed in Supplementary Table 1. YPD medium so that the gene was flanked on both sides by the repeats. The correct
contained 2% glucose (Sigma-Aldrich), 2% peptone (Difco), and 1% transformants were selected for on SC Ura— plates. URA3 was then
yeast extract (Difco); YPD plates contained 2% glucose, 2% peptone, selected against on 5-fluoroorotic acid plates. 5FOA forms a toxic me-
1% yeast extract, and 2% agarose (Invitrogen); hygromycin B plates tabolite in strains containing URA3.28 URA3 was looped out due to
contained 2%glucose, 2% peptone, 1% yeast extract, 2% agarose, and natural recombination events between the flanking tandem repeats.
200 μg ml-1 hygromycin B (Sigma Aldrich); Synthetic Complete (SC) Loss of the URA3 marker was confirmed by PCR.
medium contained 0.67% yeast nitrogen base without amino acids
and with ammonium sulphate (VWR (BD)), 2% glucose, and 0.08% RNA Isolation
CSM (Dropout mix; Sunrise Science); SC Ura— plates contained 0.67% RNA was extracted from cells by first spheroplasting yeast cells
yeast nitrogen base without amino acids and with ammonium sul- for 1 hour at 37 °C using Solution A (Zymolyase, 1mg/mL (MP
phate, 2% glucose, 0.08% CSM (Ura— Dropout mix; Sunrise Science), Biomedicals); sorbitol, 0.9 M; EDTA pH 7.5, 0.1 M, mercaptoethanol,
and 2% bacto agar (VWR (BD)); 5FOA plates contained 0.67% yeast 14 mM) and then using an ABI 6100 Nucleic Acid Prep Station and
nitrogen base without amino acids and with ammonium sulphate, reagents (Applied Biosystems).
Molecular
Biology
0.08% CSM (Dropout mix), 0.1% 5-fluoroorotic acid (Toronto Re-

search Chemicals Inc.), 0.05% uracil (Sigma Chemical), 2% glucose, cDNA Synthesis and Gene Expression
and 2% bacto agar. Yeast cultures were grown in 3 mL of YPD for cDNA was prepared using the AffinityScript QPCR cDNA Syn-
16-20 hours at 30 °C in a rotating wheel unless otherwise noted. thesis Kit (STRATAGENE). Random primers supplied in the kit were
Plated cultures were incubated at 30 °C for 3 days. used. Real-time PCR using the ABI 7500 system (Applied Biosystems)
was carried out with the appropriate enzymes and chemicals from
Tranformations Applied Biosystems as recommended by the supplier. All oligonucle-
Standard procedures for S. cerevisiae transformations with otides (Sigma-Genosys) used are listed in Supplementary Table 1.
DNA were used.27 The PCR enzyme utilized was TAKARA ExTaq
(TAKARA), and producer guidelines were followed. All oligonucle- Growth Assays
otides (Sigma-Genosys) used are listed in Supplementary Table 2. Overnight cultures were grown of the SDT1 mutant series in
YPD. These cultures were transferred to SC medium with 0.03 mg/
PCR-Based Transformation Strategy mL 6-azauracil (Sigma-Aldrich). Using the Bioscreen C MBR system
Hph, a gene conferring resistance to the antibiotic hygromycin (Oy Growth Curves Ab Ltd.), these cultures were incubated for 30
B, was inserted upstream or downstream of the gene containing hours at 30 °C with constant shaking. Growth curves were created
variable repeats in S. cerevisiae through transformation and directed for each strain using the optical density recorded by the Bioscreen C
integration. Cells containing the hph insertion were then selected for MBR system using a 600-nm filter.
on hygromycin B plates. PCR was carried out on genomic DNA from
the strains with the hph insertion to create a product with a different

a
Figure 2. Tandem repeats
are hyper-variable in natural
strains. a. Partial alignments
of the YKL071w, VPS55, and
PRE8 promoter sequences of
five naturally occurring S. cer-
evisiae strains in addition to the
respective promoter sequence
published on the Saccharomyces
Genome Database (SGD; added
to confirm that our sequencing of
the same strain (S288c) matched
the already published sequence)
show that the number of repeat
units varies between the strains
(Goffeau et al., 1996). Each
tandem repeat unit is marked
by a red box. Sequence that is
conserved between strains is
highlighted black, and sequence
that differs between strains is
white. In YKL071w, the strain
Y55 has an extra copy of the
repeat unit. The strains S288c
and W303 in VPS55 have 6.5
additional repeat units compared
to the other strains. Compared
to the strain Y55 in PRE8, the
strains D273, Sigma, W303, and
b SGD have 9.66, 11.66, 12.66, and
14.66 additional repeat units,
respectively (not all shown here).
These differences demonstrate
that these tandem repeats vary
naturally between strains; b. An
alignment of the YKL071w pro-
moter sequences. Sequences that
are conserved between strains are
highlighted pink, and sequences
that differ between strains are
white. Each tandem repeat unit
is boxed and numbered. Outside
of the tandem repeat region, the
Molecular
Biology
sequences between strains are
very well conserved.
Results in six naturally occurring yeast strains (S288c, RM11-1A, W303,

Sigma, D273, and Y55). Prior to my arrival, the entire S. cerevisiae
Tandem Repeats within Promoter Regions are Hyper-Variable in genome was scanned for tandem repeats using sequence from the
Natural Strains Saccharomyces Genome Database (SGD) and the TRF algorithm.17, 18, 19
Tandem repeats within coding regions are known to be variable Of the 5695 genes analyzed in S288c, 25% (1456 genes) contained a
in nature.4 We hypothesized that tandem repeats found in pro- tandem repeat in the promoter region (Table 2). Promoter regions
moter regions would be naturally variable as well. In order to test were defined as the 1000 nucleotides found upstream (fewer if there
this hypothesis, we located and compared the promoter sequences is another gene in this region) of the start codon in the open reading
frame of the gene. We then investigated the differences between the
www.thurj.org 59
e
Figure 3. Strains created by the PCR-based and “Loopout” strate-
gies. a. In the “Loopout” strategy, the gene URA3 is inserted within the
tandem repeats of the promoter region of the chosen gene in S. cerevisae.
URA3 is then selected against in 5FOA medium. URA3 is looped out due
to natural recombination events within the tandem repeats. When URA3
is looped out, it may also loop out some of the repeated region; b. PCR
products of the tandem repeat regions used to size repeat changes in strains generated using the “Loopout” strategy. 1-6 are mutants of the gene CDC14.
Number 1 is wild type of CDC14. 7-11 are mutants of the gene PRE8. Number 11 is the wild type of PRE8; c. In the PCR-based strategy, the Hygromycin
Molecular
Biology
resistance gene (hph) is inserted upstream or downstream of the desired gene in S. cerevisiae. The hph insertion is then selected for in Hygromycin
medium. PCR is carried out on genomic DNA from the strains with the hph insertion to create a product with a number of tandem repeats that differs
from wild type. This PCR product is then transformed into a wild-type strain of S. cerevisiae; d. PCR products from strains generated using the PCR-
based strategy. 1-2 are the wild type and complete tandem repeat deletion mutant of the gene SDT1, respectively. 3-4 are the wild type and complete
tandem repeat deletion mutant of the gene VPS55, respectively. 5-6 are the wild type and complete tandem repeat deletion mutant of the gene WHI5,
respectively. 7-8 are the wild type and complete tandem repeat deletion mutant of the gene YKL071w, respectively; e. PCR products from the series
of mutant strains created for the gene SDT1. These strains were generated using the PCR-based strategy. 1) Wild type. 2) Deletion of ⅓ of the tandem
repeat. 3) Wild type. 4) Deletion of 1/2 of the tandem repeat. 5) Wild type. 6) Deletion of ⅔ of the tandem repeat. 7) Wild type. 8) Deletion of the entire
tandem repeat. The only difference between these mutants is the number of tandem repeats found in the promoter region. Varying band lengths signify
a difference in the number of tandem repeat units.
promoter regions of these genes in the strains S288c and RM11-1a genes) of the genes with repeats were found to be conserved.
(version 1; Saccharomyces cerevisiae RM11-1a Sequencing Project. To find the most variable repeats and to see how conserved pro-
Broad Institute of Harvard and MIT). Repeats were classified as vari- moter regions are, the promoter regions of thirty-three of the variable
able if the number of complete units differed by at least one between genes found in the in silico screen, chosen based on their interesting
the two strains. 21% (307 genes) of the genes with repeats were found functions, were sequenced in the strains S288c, W303, Sigma, D273,
to have variable tandem repeats (Table 2). It is possible that this and Y55. The sequences of these five strains, along with the published
percentage is actually higher as 143 of the genes containing repeats S. cerevisiae sequence from SGD, were aligned for each gene using
did not have enough sequence information available to determine ChromasPro (Technelysium Pty Ltd). Through these alignments, I
whether the tandem repeats were variable or conserved. 69% (1006 found that the tandem repeat regions were highly variable between

a Generation of Tandem Repeat Variance in an Isogenic Strain

Background
To study the role of repeats in transcriptional regulation, I created
constructs with varying number of repeats in the promoter regions
from the progenitor strain BY4742 using two different transformation
strategies (Supplementary Table 1; Figure 3). Promoters were chosen
for this experiment based on their variability between the natural
strains. I was able to create constructs with repeat length variations
from total deletion to addition of repeat units (Figure 3b,d,e).
The “Loopout” strategy is based on natural recombination (Figure
3a). The gene URA3 is inserted through transformation within the
tandem repeats of the promoter region of the chosen gene in S.
cerevisae. URA3 insertions were selected for by growing strains in
SC minus uracil medium. Stably integrated, URA3 is then selected
b against in a medium (5-fluoroorotic acid) which forms a toxic me-
tabolite in strains containing the URA3 gene. In the strains that were
able to grow in the 5-fluoroorotic acid (5FOA) medium, URA3 was
looped out due to natural recombination events within the tandem
repeats. When URA3 was looped out, it would at times also loop out
some of the repeated region. I used gel electrophoresis to determine
if the size of the repeat region in these new strains differed from
the wild type. In the genes CDC14 and PRE8, I was able to create a
number of mutants with a repeat region that varied from the respec-
tive wild type (Figure 3b). This method, however, did not allow me to
control the size of the repeat alteration, and often yielded strains with
a tandem repeat identical to that of the parental strain.
We then developed a PCR-based strategy that allowed us to
control the number of repeat units added or deleted (Figure 3c). In
this method, hph, a gene conferring resistance to the antibiotic hygro-
mycin B, is inserted upstream or downstream of the gene containing
variable repeats in S. cerevisiae. Cells containing the hph insertion
are then selected for in hygromycin-containing medium. PCR is
carried out on genomic DNA from the strains with the hph insertion
to create a product with a different number of tandem repeats than
the wild type. This PCR product, which also contains the inserted
hph gene, is then transformed into a wild-type strain of S. cerevisiae.
Using this strategy, strains were created with repeats ranging from
a full deletion of the original repeat to a length double that of the
Figure 4. Expression levels of the YKL071w mutant strains. a. wild-type repeat (Figure 3d,e). The number of repeat units in each
The strains created for YKL071w are: KV830-wild type; KV833-deletion strain was determined through gel electrophoresis and sequencing
of ⅓ of the tandem repeat; KV835-deletion of 1/2 of the tandem repeat;
of the tandem repeat region. With this method, I created mutant
Molecular
KV834-deletion of the entire tandem repeat; KV1028-deletion of the
series (multiple strains with different numbers of repeat units in the
Biology
YKL071w open reading frame; b. Gene expression levels of the YKL071w
promoter region) in YKL071w and SDT1 (Figure 3e). These genes
mutant series were found through quantitative real-time PCR. Expres-
were chosen because of their interesting functions and successful
sion was induced through exposure to hydrogen peroxide. Because these
strains only differ in the number of tandem repeats, these differences in repeat alterations using this strategy (Table 3).
expression suggest that varying the number of tandem repeats results in
expression level differences. The expression levels were normalized with YKL071w: Transcriptional Regulation through Transcription
the expression levels of the gene ACT1. Error bars denote standard error. Factor Binding
*p ≤ 0.05 compared with wild type. Variation in Tandem Repeat Units Affects Gene Expression
To investigate whether altering the number of repeat units affects
the different strains (Figure 2). The gene YKL071w, whose twelve gene expression, we performed Quantitative real-time PCR (QPCR)
nucleotide repeat unit is ATTAGTAATGAG, contains an extra repeat on the mutant series of YKL071w, which ranges from wild type to a
unit in the strain Y55 (Table 3; Figure 2a). In VPS55, with repeat unit total deletion of the repeat unit (Figure 4a). Expression was induced
GT, the strains S288c and W303 contain 6.5 additional units (Table through exposure to hydrogen peroxide, which stimulates the
3; Figure 2a). The gene PRE8, whose repeat unit is TTA, is extremely oxidative stress response. Gene expression levels were determined
variable (Table 3). In comparison to the strain Y55, the strains D273, by averaging the data from eight experiments for each strain and
Sigma, W303, and SGD have 9.66, 11.66, 12.66, and 14.66 additional calculating the standard error. All of the gene expression levels are
repeat units, respectively (Figure 2a). These alignments were also used relative values, normalized to the constitutively expressed gene
to see if variability was confined to the tandem repeat region. We ACT1. KV1028, where the open reading frame of YKL071w has been
found that outside of the repeat, the promoter regions were indeed completely deleted, shows an expression of 4.44×10-11 (± 8.75×10-12)
highly conserved between strains (Figure 2b). (Figure 4b). KV834, with a complete tandem repeat deletion, has an
www.thurj.org 61
Figure 5. YKL071w tandem repeats overlap with Yap1 binding sites. An alignment of the DNA sequence found in the promoter region of
YKL071w in seven natural yeast strains. Each tandem repeat unit is marked in red. The repeat units containing the binding motif for Yap1 (TTAC/
GTAA) are marked (Fernandes et al., 1997). The strains YJM789 and Y55 contain an extra repeat unit, and therefore an extra binding site for Yap1.
Natural variation between these strains alters the number of Yap1 binding sites in the promoter region.
expression level of 4.14×10-8 (± 4.25×10-9). KV835 has a deletion of to decreasing repeat units), gene expression decreases in YKL071w
½ of its tandem repeat and an expression of 1.18×10-7 (± 2.90×10-8). (Figure 6). In order to determine if Yap1 actually has an effect on
KV833, where ⅓ of the tandem repeat has been deleted, has expres- transcription, we deleted the YAP1 gene in each strain of the
sion 2.49×10-7 (± 2.46×10-8). The wild type (KV830), with its tandem YKL071w mutant series and performed QPCR on these new strains.
repeat intact, shows an expression of 3.34×10-7 (± 2.32×10-8). These Gene expression was induced through exposure to hydrogen perox-
results suggest that the expression level rises as the number of tandem ide. The data was normalized to ACT1 and then normalized again
repeat units increases in the promoter region of YKL071w. according to the appropriate wild type so that the expression levels
Changing the Number of Yap1 Binding Sites Alters Gene Expression from the strains with an intact YAP1 and from the strains with the
In 10% of the genes containing intra-promoter repeats, the YAP1 deleted could be compared. Expression is delineated in units
tandem repeats overlap with at least one transcription factor binding of percentage (in decimal form) of the respective wild type. KV834,
site (Table 2). The repeats in YKL071w overlap with the TTAC/GTAA with a complete tandem repeat deletion has expression 0.12 (± 0.01)
binding motif for Yap1, a transcription factor involved in the oxida- while the strain containing the YAP1 deletion has expression 0.05
tive stress response (Figure 5).20 Between natural yeast strains, there is (± 0.003). KV835, containing a ½ deletion of the repeat, shows
variation in the number of Yap1 binding sites in the promoter region. expression 0.35 (± 0.09) while its YAP1 mutant shows expression
In YJM789 and Y55, an extra repeat unit results in six Yap1 binding 0.18 (± 0.11). KV833 has a deletion of ⅓ of its tandem repeat and an
sites in the promoter region as opposed to the five Yap1 binding sites expression of 0.75 (± 0.07), and its YAP1 mutant has an expression
found in the strains RM11-1a, S288c, Sher (D273), Sigma, and W303. level of 0.04 (± 0.006). The wild-type strain, KV 830, has expression 1,
We hypothesized that Yap1 is involved in the transcriptional regula- while its YAP1 mutant shows expression 0.08 (± 0.03). In the strains
tion of YKL071w. without YAP1, expression is significantly decreased in comparison to
As the number of Yap1 binding sites decreases (corresponding the strains containing the gene, regardless of the size of the tandem
Molecular
Biology
Figure 6. Variation in the number of Yap1 binding sites and deletion of YAP1 alter gene expression levels in YKL071w. Gene expression
levels of the YKL071w mutant series were found through quantitative real-time PCR. Expression was induced through exposure to hydrogen peroxide.
The expression levels were normalized according to ACT1 and the appropriate wild type. Decreasing the size of the tandem repeat in the promoter region
of YKL071w also reduces the number of Yap1 binding sites in that region. In the strains where only the tandem repeat has been altered (blue), expression
decreases as the number of Yap1 binding sites decreases. The gene YAP1 was deleted in four YKL071w strains (red): KV830-wild type; KV833-deletion of
⅓ of the tandem repeat; KV835-deletion of 1/2 of the tandem repeat; KV834-deletion of the entire tandem repeat. Expression is significantly decreased
in the strains without YAP1 in comparison to the strains containing the gene. Transcription factor binding, therefore, may regulate transcription in
YKL071w. Error bars denote standard error normalized by the respective wild type. *p ≤ 0.05 compared with its equivalent YAP1+ strain.

Figure 7. Expression levels and growth

assays for the SDT1 mutant series. a. The
strains created for SDT1 are: KV534-Wild type;
KV979-deletion of ⅓ of the tandem repeat; KV975-
deletion of ½ of the tandem repeat; KV981-deletion
of ⅔ of the tandem repeat; KV782-deletion of the
entire tandem repeat; KV1380-addition of the
entire tandem repeat; b. Gene expression levels of
the SDT1 mutant series found through quantitative
real-time PCR. Expression was induced through
exposure to the alkylating agent methyl methane-
sulfonate (MMS). Because these strains only differ
in the number of tandem repeats, these differences
in expression suggest that varying the number of
tandem repeats results in expression level differ-
ences. The expression levels were normalized to
the expression levels of the gene ACT1. Error bars
denote standard error normalized by the respective
wild type. *p ≤ 0.05 compared with wild type; c.
Growth curves for the SDT1 mutant series in the
presence of 0.03 mg/mL 6-azauracil (6-AU). The
length of the lag phase exhibited in each strain
compared to the others corresponds to the gene
expression levels of the strains relative to each
other. Th is suggests that varying the number of
tandem repeats results in phenotypic differences in
Molecular
addition to expression level differences.
Biology
repeats. These two phenomena suggest that tandem repeats alter gene repeat deletion, has an expression level of 0.11 (± 0.02). KV975 has a
expression in YKL071w through transcription factor (specifically, deletion of ½ of its tandem repeat and an expression of 0.156 (± 0.02).
Yap1) binding. KV979, where ⅓ of the tandem repeat has been deleted, has expres-
sion 0.162 (± 0.005). The wild type (KV534), with its tandem repeat
SDT1: Tandem Repeats Affect Phenotype intact, shows an expression of 0.13 (± 0.02). KV1380, which contains
Variation in Tandem Repeat Units Affects Gene Expression two copies of the entire tandem repeat, has expression 0.07 (± 0.006).
To determine whether altering the number of repeat units affects These results show that, in SDT1, gene expression increases as the
gene expression in SDT1, we performed QPCR on the mutant series, number of tandem repeat units in the promoter region increases until
which ranges from a tandem repeat double the length of wild type to the repeat reaches an optimal size (slightly smaller than that of wild
a total deletion of the repeat unit (Figure 7a). Expression was induced type, exact size unknown). Once the tandem repeat becomes larger
through exposure to the alkylating agent methyl methanesulfonate than this optimal size, gene expression seems to decrease.
(MMS). Gene expression levels were determined by averaging the
data from three experiments for each strain and calculating the Length of Lag Phases Corresponds to Levels of Gene Expression
standard error. All of the gene expression levels are relative values, To examine the effect of tandem repeats on phenotype, growth
normalized to the constitutively expressed gene ACT1. KV782, where assays in the presence of the stress reagent 6-azauracil (resistance
the tandem repeat of SDT1 has been completely deleted, shows an to which requires the gene SDT1) were completed with the mutant
expression of 0.08 (± 0.01) (Figure 7b). KV981, with a ⅔ tandem series of SDT1 (Figure 7c). The length of the lag phase exhibited in
www.thurj.org 63
each strain was then calculated by determining the time at which is the number of repeat units, suggesting that the alteration of the
exponential growth was initiated. KV534, the wild type, had a lag repeat region is the cause of the deviation in gene expression from
phase of 16 hours. The strain with two full repeats, KV1380, yielded that of the wild type. Similarly, the SDT1 mutant series exhibits gene
a 21.25-hour lag phase. KV979, the strain with a ⅓ deletion of the expression levels that increase as the number of tandem repeat units
tandem repeat, exhibited a lag time of 13.5 hours. The lag phase of increase until the repeat reaches an optimal size, and then decrease
the strain with ½ of its tandem repeat deleted, KV975, was 15.75 with enlargement of the repeat beyond this size. These strains also
hours. KV981, containing a ⅔ repeat deletion, had a lag time of 23.5 only differ in the number of tandem repeats in the promoter region,
hours. The mutant with a complete deletion of the repeat, KV782, indicating that the variation in the repeat region drives the change
yielded a 23.25-hour lag phase. From the expression data, we see that in transcriptional activity. Growth assays in 6-AU with the SDT1
the strains have increased expression as the number of tandem repeat strains showed that the length of the lag phase in each strain cor-
units in the promoter region increases until the repeat reaches a size responded to the pattern seen in the gene expression levels. This
slightly smaller than that of wild type, beyond which, expression correlation insinuates that the difference in repeat size also affects
decreases (Figure 7b). The lag times follow this same trend, includ- the phenotypic fitness of the respective strain. A shorter lag phase
ing similar expression in strains KV981 and KV782 resulting in corresponded to a higher level of expression. This finding suggests
comparable lag times, suggesting that an alteration in the number of that a shorter lag phase may indicate that activation of transcription
tandem repeats confers a phenotypic change (Figure 7c). occurs earlier than in strains exhibiting a longer lag phase and lower
gene expression. Therefore, a different number of repeat units may
Discussion allow certain strains to adapt more quickly to environmental stress
through rapid transcriptional activation.
The results presented in this thesis suggest that tandem repeats in
promoter regions are hyper-variable and affect transcriptional activity Mechanisms by which Tandem Repeats Regulate Transcription
through multiple mechanisms. Natural strains of S. cerevisiae show Transcription factor binding motifs overlap with tandem repeats
significant differences in the size of their tandem repeats, whereas in 10% of the genes containing intra-promoter repeats (Table 2).
the DNA sequences immediately surrounding the repeats are largely YKL071w, whose repeat region overlaps with binding sites for the
conserved, indicating that these repeats are variable. Expression transcription factor Yap1, is one of these genes. YAP1 is a vital gene
levels change as the number of repeat units is altered, which suggests in the oxidative stress response of cells and is activated by oxidizing
that tandem repeats are involved in the regulation of transcriptional agents, such as hydrogen peroxide.20, 24 While the exact function of
activity. In addition, these variations in expression correspond to YKL071w is unknown, it is thought to also be involved in the oxidative
phenotypic differences. One possible mechanism through which stress response and shows high induction in the presence of hydrogen
tandem repeats may regulate gene expression is transcription factor peroxide.15 As the number of repeat units in the YKL071w promoter
binding, shown through the reduced expression that is seen in region increase, and therefore, the number of Yap1 binding sites
YKL071w strains where the gene coding for the transcription factor increases, gene expression levels rise. This suggests that Yap1 binds
Yap1 has been deleted. The variable nature of tandem repeats allows the motifs found in the repeat region, and that as more transcrip-
for fluctuation of gene expression levels and quick evolution to novel tion factors are bound, transcriptional activity is enhanced. Strains
environments. containing a deletion of YAP1 had significantly lower expression
compared to the respective strains with an intact YAP1 open reading
Variability in S. cerevisiae Promoter Regions frame, strengthening the claim that Yap1 binding is necessary for
For millions of years, the majority of genes in different S. cer- high levels of gene expression in YKL071w. These results suggest
evisiae strains have been conserved, making any difference in DNA that transcription factor binding is a potential mechanism by which
sequence between these strains noteworthy.21 Our finding that many tandem repeats regulate transcription.
of the genes with tandem repeats in the promoter region varied in the Because only 10% of these genes overlap with transcription factor
number of repeat units between strains was, therefore, unexpected binding sites, there must be other mechanisms by which transcrip-
Molecular
Biology
and interesting when considering how conserved the strains are tional activity is controlled. The variability of tandem repeats may
otherwise (Table 2). This discovery suggests that variability is actu- affect transcription spatially as a sort of on/off switch. Changing the
ally desired in these promoters, and the unstable nature of tandem number of repeat units could alter the distance between regulatory
repeats provides an excellent vehicle by which the promoter region elements in the promoter region and thereby turn transcription on or
can expand and contract.4 A possible compensating benefit to the off in the respective gene. Another likely mechanism by which tandem
instability of these repeat regions could be swift adaptation of gene repeats could regulate gene expression is nucleosome positioning.25
expression levels to changing environments and selective pressures. Variation in the size of the repeat region may affect histone binding,
In this fashion, tandem repeats in promoter regions may provide a allowing a gene to be turned on or off depending on how tightly
functional diversity in S. cerevisiae similar to that brought about the region was bound. If the DNA was bound tightly, transcription
by tandem repeats in coding regions in multiple species, including would be hindered, and a loose interaction would allow uninhibited
bacteria, dogs, yeast, and humans.4, 22, 23 transcription. Studies are currently under way to investigate these
mechanisms.
Transcriptional Regulation through Tandem Repeat Variability
Tandem repeats in coding regions are known to confer variability Conclusion
in proteins.8 This leads one to hypothesize that tandem repeats in Little is known about the role of tandem repeats in promoter
promoter regions might also serve a biological function. Indeed, this regions. The results of this study begin to characterize the biologi-
study shows that changes in repeat size correlate with differences cal function of these repeats. This thesis demonstrates that repeats
in transcriptional activity. In YKL071w, gene expression levels rise confer variability and “evolvability” of transcriptional regulation
as the number of repeat units increase. Because these strains were in S. cerevisiae. I show that intra-promoter tandem repeats are
created in an isogenic background, the only difference between them hyper-variable in natural strains, and that this variability affects the

transcriptional activity of the respective gene, potentially through Junk!” Curr Opin Biotechnol 17.2 (2006): 168-73.
transcription factor binding along with other mechanisms. The 12. Epplen, J.T., W. Mäueler, and E.J. Santos. “On GATAGATA and Other
“Junk” in the Barren Stretch of Genomic Desert.” Cytogenet Cell Genet
findings outlined above suggest that repeat regions are desirable in
80.1-4 (1998): 75-82.
certain genes because they allow quick evolution and adaptation to 13. Galimberti, Daniela, et al. “Association of a NOS1 Promoter Repeat with
external conditions. Alzheimer’s Disease.” Neurobiology of Aging In Press, Corrected Proof
Continued study is needed to determine the precise role of (2007).
repeats in the regulation of transcription. These results could provide 14. Bell, P.J., et al. “Tandemly Repeated 147 bp Elements Cause Structural
and Functional Variation in Divergent MAL Promoters of Saccharomyces
insights into the function of intra-promoter repeats in other eukary-
cerevisiae.” Yeast 13.12 (1997): 1135 - 44.
otes, including humans. In humans, fragile-X syndrome exhibits 15. Gasch, A.P., et al. “Genomic Expression Programs in the Response of Yeast
upregulation of FMR1 due to an expansion of the tandem repeat in Cells to Environmental Changes.” Mol Biol Cell 11.12 (2000): 4241-57.
the promoter region until the repeat reaches an optimal size, where 16. Nakanishi, Toshiyuki, and Kazuhisa Sekimizu. “SDT1/SSM1, a Multicopy
expression is then silenced.26 This same pattern was seen in SDT1. Suppressor of S-II Null Mutant, Encodes a Novel Pyrimidine 5’-Nucleoti-
Increased knowledge of the mechanism behind this phenomenon dase.” J. Biol. Chem. 277.24 (2002): 22103-06.
17. Legendre, M., et al. “Sequence-Based Estimation of Minisatellite and Mic-
could be crucial in understanding and perhaps even preventing such rosatellite Repeat Variability.” Genome Res. 17.12 (2007): 1787-96.
diseases. 18. Goffeau, A., et al. “Life with 6000 Genes.” Science 274.5287 (1996): 546,
63-7.
References 19. Benson, G., et al. “Tandem Repeats Finder: A Program to Analyze DNA
Sequences.” Nucleic Acids Res. 27 (1999): 573–80.
20. Fernandes, L., C. Rodrigues-Pousada, and K. Struhl. “Yap, a Novel Family
1. “Human Genome Project”. (March 26, 2008). 1 May 2008. <http://www.
of Eight bZIP Proteins in Saccharomyces cerevisiae with Distinct Biological
ornl.gov/sci/techresources/Human_Genome/project/info.shtml>.
Functions.” Mol Cell Biol. 17.12 (1997): 6982-93.
2. Orgel, L.E., and F.H. Crick. “Selfish DNA: The Ultimate Parasite.” Nature
21. Kellis, Manolis, et al. “Sequencing and Comparison of Yeast Species to
284 (1980): 604-07.
Identify Genes and Regulatory Elements.” Nature 423.6937 (2003): 241-54.
3. Verstrepen, Kevin J., Todd B. Reynolds, and Gerald R. Fink. “Origins of
22. Martin, P., et al. “Microsatellite Instability Regulates Transcription Factor
Variation in the Fungal Cell Surface.” Nat Rev Micro 2.7 (2004): 533-40.
Binding and Gene Expression.” Proc Natl Acad Sci U S A 102.10 (2005):
4. Verstrepen, Kevin, et al. “Intragenic Tandem Repeats Generate Functional
3800-4.
Variability.” Nat Genet 37.9 (2005): 986-90.
23. Fondon, J.W., and H.R. Garner. “Molecular Origins of Rapid and Continu-
5. John, B. Heterochromatin Molecular and Structural Aspects. Ed. R.S.
ous Morphological Evolution.” Proc Natl Acad Sci USA 101.52 (2004):
Verma: Cambridge University Press, 1988. 1–147.
18058-63.
6. Wyman, A.R., and R. White. “A Highly Polymorphic Locus in Human
24. Kuge, S., N. Jones, and A. Nomoto. “Regulation of YAP-1 Nuclear Local-
DNA.” Proc Natl Acad Sci USA 77.11 (1980): 6754-8.
ization in Response to Oxidative Stress.” EMBO J. 16.7 (1997): 1710-20.
7. Thomas, Elizabeth E. “Short, Local Duplications in Eukaryotic Genomes.”
25. Wang, Y.H., et al. “Long CCG Triplet Repeat Blocks Exclude Nucleosomes:
Current Opinion in Genetics & Development 15 (2005): 640-44.
A Possible Mechanism for the Nature of Fragile Sites in Chromosomes.” J
8. Csink, A.K., and S. Henikoff. “Something from Nothing: The Evolution
Mol Biol. 263.4 (1996): 511-6.
and Utility of Satellite Repeats.” Trends Genet 14.5 (1998): 200-4.
26. Usdin, K. “The Biological Effects of Simple Tandem Repeats: Lessons from
9. Blackburn, E.H., and J.G. Gall. “A Tandemly Repeated Sequence at the Ter-
the Repeat Expansion Diseases.” Genome Res. 18.7 (2008): 1011-9.
mini of the Extrachromosomal Ribosomal RNA Genes in Tetrahymena.” J
27. Gietz, R.D., and R.H. Schiestl. “Transforming Yeast with DNA.” Methods
Mol Biol. 120.1 (1978): 33-53.
in Mol. Cell. Biol. 5 (1995): 255-69.
10. Tyler-Smith, C., and H.F. Willard. “Mammalian Chromosome Structure.”
28. Boeke, J.D., et al. “5-Fluoroorotic Acid as a Selective Agent in Yeast Mo-
Curr. Opin. Genet. Dev. 3 (1993): 390-7.
lecular Genetics.” Methods Enzymol. 154 (1987): 164-75.
11. Morgante, M. “Plant Genome Organisation and Diversity: The Year of the
Molecular
Biology
www.thurj.org 65
Young people play bingo, too:

Reducing priming effects through
mindfulness
Stephanie M. Kaplan* and Rory Michelle Sullivan**
*
Harvard College 2010, skaplan@fas.harvard.edu; **Harvard College 2009, sulliv3@fas.harvard.edu
This study examined whether mindfulness, or active awareness, could reduce priming effects, the
subconscious effects of stimuli on one’s behavior. With a sample of 39 undergraduates, we replicated
an earlier study demonstrating the effect of priming the elderly stereotype through a word task on
participants’ walk speed, and extended it by introducing a mindfulness manipulation through the
task instructions given. We hypothesized that participants in the elderly condition who received the
mindfulness instructions would not walk more slowly. Although non-significant, our results showed
an interaction in the expected direction with decreased walk-time for participants in the elderly-
mindful condition, lending tentative support for our hypothesis. Future studies that address the
present study’s limitations can extend these results and continue to inform our understanding of the
effects of mindfulness on priming. This concept has important implications for social issues such as
stereotyping, as our results imply that mindfulness may be an effective means of reducing stereotype
activation.
Introduction related to the elderly stereotype walked more slowly than did those
who were exposed to neutral words. When debriefed, participants
Researchers have defined mindfulness as an effortless, simple expressed no knowledge of the connection between the words they
process that consists of drawing novel distinctions and noticing new had seen and the elderly stereotype, nor did they believe the words
things, allowing us to become aware of how things change depending they saw had affected them in any way. This study demonstrates
on the context and perspective from which they are viewed (Langer, the effects of priming, as participants’ behavior was affected by the
1989). Mindfulness is not about just paying attention; it is about stimuli even though they were not aware of it.
making a conscious effort to be “in the moment” and not to ignore However, perhaps if people are made mindful, these priming
the environment (Langer, 1989). By being mindful, a person is more effects will not occur. There is some evidence that priming effects
aware of the events that are happening to, and the stimuli that are might be the result of participants’ mindlessness during the studies
acting upon, him or her, and of how those might impact his or her (Djikic, Langer, & Stapleton, 2008). In one study, participants who
subsequent actions. This awareness of one’s environment allows one were made mindful during a priming word task were better able to
to make more thoughtful, calculated decisions when approaching a judge a speaker’s characteristics than were those who received the
given situation, rather than acting without a full understanding or same priming but were not made mindful (Langer & Newman, 1979).
appreciation of it (Langer & Moldoveanu, 2000). In the current study, In another study, researchers found that consciousness of a priming
we sought to examine how this concept of mindfulness might inform event allowed participants to interpret subsequent stimuli more flex-
our understanding of another concept in psychology: Priming. ibly than participants who were not conscious of the priming event
Researchers in past studies have suggested that exposing people to (Lombardi, Higgins, & Bargh, 1987).
certain stimuli can affect their subsequent behavior by increasing the In another study, men who were likely to sexually aggress were
accessibility of a schema, trait, or concept even without them being primed with “power”-related words, and subsequently found women
Psychology
aware of it (Josephson, 1987). For example, researchers in one study more attractive than did control participants, but, unaware of the
found that children who watched a violent police film displayed more manipulation, attributed their attraction to the women’s physical
aggressive behaviors in a subsequent hockey game than did those characteristics (Bargh, Raymond, Pryor, & Strack, 1995). When the
who instead watched a nonviolent film about bike racing (Josephson, men’s attentions were drawn to the source of their attraction – the fact
1987). This phenomenon is called priming. that men are sexually attracted to women whom they have authority
In one well-known study on priming, researchers in- over – the implicit effect was dampened (Bargh et al., 1995). As dem-
vestigated how being exposed to words associated with the elderly onstrated by the power-primed men, implicit cognitive effects such
stereotype affected how quickly participants walked down a hallway as priming may be reduced by focusing an individual’s attention the
at the conclusion of an experiment (Bargh, Chen, & Burrows, 1996). task (Greenwald & Banaji, 1995). Schwarz and Clore (1983; as cited
The researchers found that participants who were exposed to words in Greenwald & Banaji, 1995) conducted a telephone survey inquir-
ing about quality of life. On sunny days, reported quality of life was

higher than on rainy days. However, when people were asked a ques- 2 at the conclusion of the word task. Bargh and colleagues (1996)
tion about the weather first, quality of life reports evened out. This also used walk-time as the measure of priming effects in their study.
indicates that attending, or active thinking, disrupts the influence of The length of time it took the participant to walk this distance was
weak perceptual cues. Priming, then, can be interrupted by focusing measured by Experimenter 2, who started a stopwatch as soon as the
on the source of the prime. These findings suggest that mindfulness participant crossed a fixed line on the floor, and stopped it as soon
may be able to reduce or even eliminate priming effects. This would as the participant crossed a second fixed line on the floor 23.625 feet
mean that by being mindful, we could recognize the effects certain away.
stimuli might have on us unconsciously, and take steps to work
against this and to instead make conscious decisions. Procedure
This idea has important real-world implications for domains The study is a 2 × 2 between-subjects design. There were four
such as stereotyping and prejudice. Perhaps by being mindful of the conditions, to which participants were assigned through a block ran-
existence of a stereotype, one can recognize that the stereotype is domization of four blocks with 10 random numbers. As in the study
there and take steps not to be influenced by it. Findings by Shih and by Bargh and colleagues (1996), half of the participants were primed
colleagues (2002) support this notion. When the Asian stereotype with the elderly stereotype. Half of that group, as well as half of the
was blatantly activated for Asian participants before a math test, the control group, were also given instructions designed to make them
stereotype did not affect their performance. That is to say, when the mindful. These “mindful” instructions read as follows: “For each set
cue in question is cognized clearly – when it is attended to from the of words below, make a grammatical four word sentence and write
outset – implicit cognitive effects such as priming can be reduced or it down in the space provided. Note the similarities and differences
even eliminated, something which is certainly desirable with regard among the words as you unscramble the sentences. You do not need
to stereotyping. to write anything beyond the four word sentence down, but actively
In our experiment, we sought to investigate whether mindfulness think about how the words in each sentence might be related to each
can reduce priming effects by examining this idea in the context of the other and about how they might be different.” The other half of each
elderly stereotype-activation study (Bargh et al., 1996). We replicated group received neutral instructions not intended to induce mindful-
the study by Bargh and colleagues and extended it by introducing ness; these “not-mindful” instructions read as follows: “For each set
the concept of mindfulness to see whether by making participants of words below, make a grammatical four word sentence and write it
mindful, we could reduce the priming effects the researchers found. down in the space provided.” This design allowed us to examine the
We wanted to discover whether those participants who viewed words main effect of the priming condition in an attempt to replicate the
related to the elderly stereotype but were also made mindful would results of the study by Bargh and colleagues (1996), while simultane-
still exhibit priming effects. We hypothesized that by making these ously introducing our mindfulness manipulation. All groups were
participants mindful (i.e., leading them to pay active attention to the treated the same way by the experimenters in the study; the only dif-
words in the study), we could negate the priming effects discovered by ferences between the groups were whether the participants completed
Bargh and colleagues (1996) (i.e., these participants would not walk a word task that included words related to the elderly stereotype, and
more slowly than normal). Presumably, this would be because these whether the participants received mindfulness instructions.
participants were made aware of the elderly stereotype’s presence, The study was conducted using two tables in the back of Quincy
thus disrupting the prime. This would provide continued support for Dining Hall, on opposite sides of the dining hall. It was conducted
the idea that mindfulness (i.e., making active distinctions) can indeed between the hours of 10:30am and 11:30am, and between 3pm and
reduce priming effects, which would allow for a greater awareness of 5pm, as the dining hall is mostly empty during these times.
our actions as well as the ability to make more informed decisions. To begin, Experimenter 1 was seated in a specific chair at the back-
right table in Quincy Dining Hall, and Experimenter 2 was seated at
Method a specific chair at the back-left table in Quincy Dining Hall. When
the participant arrived, Experimenter 1 greeted him or her, told the
Participants participant to sit down across from her, and had the participant sign
The study included 40 male (n = 8) and female (n = 32) participants the consent form. This experimenter then handed the participant
who were undergraduates at Harvard College over the age of 18. the Language Ability Task and explained the instructions.1 This
Given that the study we sought to replicate and extend was conducted task was identical to that used in the original study by Bargh, Chen,
using college undergraduates, we did the same in the current study. and Burrows (1996), and was obtained directly from Professor John
The Harvard College student body was also the most easily accessible A. Bargh. Depending on the condition, the task included either
pool of potential participants. Participants were recruited through mindful instructions or not-mindful instructions, and either elderly-
emails sent out over Harvard email lists stereotype-related words or neutral words. The participant was then
Informed consent was obtained from all participants. Before com- told to reread the instructions silently, and also that he or she was
pleting the study, participants signed a written consent form, which allowed unconstrained time to complete the task. The participant’s
Psychology
informed them that there would be minimal risk associated with the task-time was timed by Experimenter 1 on her laptop, without the
study and that they could choose to discontinue participation at any participant’s knowledge.
time, and requested their permission to use their data anonymously When the participant informed the experimenter that the task
for future research. All procedures were approved by the Committee was complete, the experimenter thanked the participant, gave a fake,
on the Use of Human Subjects at Harvard University. partial debriefing explaining that the study was meant to examine
how the brain processes language, and instructed the participant
Apparatus to walk over to Experimenter 2, who would finish debriefing and
The construct of interest to our study was the display of priming compensate him or her. Experimenter 2 started a stopwatch as soon
effects. We measured this as the amount of time it took each par- as the participant crossed a fixed line on the floor, without the par-
ticipant to walk 23.625 feet from Experimenter 1 to Experimenter 1
Copies of the tasks are available from the authors upon request.
www.thurj.org 67
Walk%Time%by%Word%Recall prime. Participants who recalled the most words (more than the
5.7
median) were recoded as primed. This ANOVA yielded a significant
interaction, F(1, 39) = 5.039, p = .031, in the expected direction, with
5.6
a medium to large effect size of .129 (Figure 1). This demonstrates
5.5
that participants who received elderly words and then recalled more
words were more affected by the elderly stereotype and walked slower
Walk%Time%(s)
5.4
5.3 as a result, affirming the results of the previous research by Bargh et

5.2 al. for the existence of this priming effect.
5.1 To determine whether our mindfulness instructions had an effect
5
on participant behavior, we also assessed differences in word recall
and task-time between the groups that received mindfulness instruc-
4.9
Neutral Elderly Neutral Elderly tions and the groups that did not. Analyses revealed that participants
who received the mindful instructions remembered more words, χ2(1,
Lower7Word7Recall Higher7Word7Recall
N = 39) = 4.356, p = .037, than did those who received the not-mindful
Word%Recall
instructions, and took slightly shorter time to complete the task, χ2(1,
Figure 1. Interaction between elderly/neutral conditions and N = 39) = 1.600, p = .206, although this value did not reach the level of
higher word recall/lower word recall participants. Participants with significance.
higher word recall recalled more words than the median, and participants
with lower word recall recalled fewer words than the median. Roughly half Walk-Time
of the participants in each group had received elderly-stereotype-related We ran a 2 (elderly/neutral) × 2 (mindful/not mindful) analysis of
words. Walk-times within each of these four groups were then averaged. variance (ANOVA) on walk-time, which is the dependent measure.
A 2 (elderly/neutral) × 2 (higher word recall/lower word recall) analysis This revealed an interaction that was in the expected direction and
of variance (ANOVA) on walk-time revealed a significant interaction, F(1, approached significance, F(1, 39) = 2.529, p = .121, and had a medium
39) = 5.039, p = .031, in the expected direction, with a medium to large effect size (η2 = .067; Figure 2). Participants in the elderly-not mindful
effect size of .129, demonstrating the existence of the elderly stereotype condition walked more slowly than participants in the neutral-not
prime. mindful condition. Participants in the elderly-mindful condition
walked more quickly than those in the elderly-not mindful condition.
ticipant’s knowledge. This experimenter stopped the stopwatch on Participants in the neutral-not mindful condition walked slightly
the laptop as soon as the participant crossed a second fixed line that faster than participants in the neutral-mindful condition.
was 23.625 feet away from the first line. Experimenter 2 was blind
to the instructions received by the participant so as not to bias the
timing. This experimenter then conducted a funneled debriefing and
Discussion
gave the participant his or her compensation. During this debriefing, Walk%Time%by%Condition
the participants were asked what they believed the purpose of the
experiment was, and all answered that they thought it had to do with 5.7
language ability, demonstrating that the deception employed was 5.6

effective. 5.5
Data Analyses
Walk%time%(s)
5.4
We first ran t-tests between male and female participants’ walk- 5.3
time, task-time, and word recall data. Next, we ran a 2 × 2 ANOVA
5.2
using between-subjects factors of mindfulness (mindful instructions,
not-mindful instructions) and priming (elderly stereotype-related 5.1
words, neutral words). We then recoded word recall into a measure 5

of mindfulness by splitting the data into two groups at the median
and running a 2 × 2 ANOVA using the higher recall as the between- 4.9
Neutral Elderly Neutral Elderly
subjects mindfulness factor. We then repeated this procedure using
task-time as the between-subjects mindfulness factor. To assess how Less7Mindful More7Mindful
the mindfulness instructions affected word recall, we ran a Pearson’s Condition
Chi-Square on the mindful/not-mindful instructions conditions and

word recall, which we made into a categorical variable by splitting Figure 2. Graphical depiction of interaction between elderly/
the groups at the median. We also ran a Pearson’s Chi-Square on neutral conditions and mindful/not mindful conditions. Partici-
Psychology
mindfulness and walk-time and on mindfulness and task-time. pants in the “More Mindful” condition received mindfulness instructions
and those in the “Less Mindful” condition received neutral instructions.
Results Half of the participants in each condition received elderly-stereotype-
related words and half received neutral words. Walk-times within each of
these four groups were then averaged. A 2 (elderly/neutral) × 2 (mindful/
Preliminary Analyses
not mindful) analysis of variance (ANOVA) on walk-time revealed an
Preliminary analyses revealed no significant differences between
interaction that was in the expected direction and approached signifi-
male and female participants on walk-time, task-time, or word recall,
cance, F(1, 39) = 2.529, p = .121, and had a medium effect size (η2 = .067).
so we collapsed our results across gender. Participants who received mindfulness instructions and the elderly prime
We ran a 2 (elderly/neutral) × 2 (higher word recall/lower word walked more quickly than participants who received neutral instructions
recall) ANOVA to test for evidence of the elderly stereotype as a and the elderly prime.

Our results provide tentative support for our hypothesis that address the above limitations.
mindfulness is capable of reducing priming effects. We were able Future research should be conducted with more participants to
to replicate the results of the study by Bargh and colleagues (1996), bolster the theory that mindfulness can reduce priming effects. It
finding that participants in the elderly-not mindful condition walked should also be extended to assess mindfulness’s effect on other ver-
more slowly than did participants in the neutral-not mindful condi- sions of stereotype activation, including stereotype threat. It would
tion. We were also able to determine that our mindfulness instruc- be desirable to determine how subtle mindfulness instructions can
tions did have an effect, as the participants who received them were be while still being effective, as the more general the instructions
able to recall more words than were those who did not, suggesting can be made, the more contexts to which they can be applied, and
that these participants were made mindful. It is somewhat surprising the more pervasive the mindfulness mindset can become. It is when
that the mindful participants completed the word task more quickly, mindfulness becomes a pervasive mindset that we can be protected
but it is possible that the mindfulness instructions made these par- from subtle stereotype activation, which will particularly benefit
ticipants more focused on the task, which would explain their lower those subject to stereotype threat. By practicing mindfulness, we may
task-times. find ourselves able to make more conscious decisions in our daily
In support of our hypothesis, our mindfulness manipulation had lives.
the expected effect on walk-time. The participants in the elderly-
mindful condition walked faster than the participants in the elderly- Acknowledgments
not mindful condition, supporting our hypothesis that mindfulness
can reduce the effects of priming. Participants in the neutral-not This research was made possible by a stipend from the Harvard
mindful condition walked slightly faster than participants in the University Psychology Department for students enrolled in Psychol-
neutral-mindful condition, but this result is also non-significant ogy 1901: Methods of Behavioral Research. We also thank Berryline,
and we expect would remain so upon further testing. Although the which supported our research by granting forty $2 gift certificates
interaction was not significant, there was twice as large a difference within our $50 budget limitation to be used as compensation for
in walk-times between mindful and not-mindful participants when participants. We would like to thank Professor John A. Bargh for
primed than when neutral, suggesting that the mindful instructions providing us with easy access to his scrambled sentence task. We also
accomplished something more in the elderly condition, as they were thank Professor Ellen Langer and the students of her mindfulness lab
intended to. It is promising that with a sample size of only 39, we for their helpful comments on experimental design and task instruc-
acquired results that were in the expected direction; this suggests tions, Dr. Mariann Weierich and Michelle Wedig for all of their help,
that had there been more participants, the interaction might have and especially Michelle Wedig for her assistance and suggestions in
reached significance. analyzing and interpreting the data.
Though not entirely conclusive, our study does lend support to
the theory that mindfulness can indeed reduce priming effects, and References
prompts future research on a larger scale that incorporates more
participants. Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social behavior:
Nevertheless, the study suffered from certain limitations. For Direct effects of trait construct and stereotype activation on action. Journal
one, we were unable to replicate the conditions of the study by Bargh of Personality and Social Psychology, 71(2), 230-244.
and colleagues (1996) exactly. In the 1996 study, walking distance Bargh, J., Raymond, P., Pryor, J. & Strack, F. (1995). Attractiveness of the under-
ling: An automatic power—sex association and its consequences for sexual
was 9.75 meters (31.9882 feet) and located in a hallway between the harassment in addition, aggression. Journal of Personality and Social Psy-
experimenter and an elevator. In the present study, walking distance chology, 68, 768-781.
was 7.2009 meters (23.625 feet) and located in a mostly empty dining DeMarree, K. G., Wheeler, S. C., Petty, R. E. (2005). Priming a new identity:
hall between two experimenters. Though the closest conditions avail- Self-monitoring moderates the effects of nonself primes on self-judgments
able were used, these discrepancies could have interfered with our and behavior. Journal of Personality and Social Psychology, 89(5), 657-671.
Djikic, M., Langer, E. J., & Stapleton, S. F. (2008). Reducing stereotyping
results. First, a longer stretch of hallway may allow for a greater range
through mindfulness: Effects on automatic stereotype-activated behaviors.
of measured walking times and therefore greater variability between Unpublished manuscript, University of Toronto and Harvard University,
participants. Second, the present study required participants to walk Cambridge.
toward an experimenter, which may have created an environment Greenwald, A & Banaji, M (1995) Implicit social cognition. Psychological Re-
differing significantly from one in which the participant believed he view, 102(1),1024-1027.
Josephson, W.D. (1987). Television violence and children’s aggression: Testing
or she is merely leaving the room. Future studies that mirror the con-
the priming, social script, and disinhibition prediction. Journal of Person-
ditions of the study by Bargh and colleagues (1996) more exactly may ality and Social Psychology, 53, 882-890.
provide more conclusive results than ours did. However, researchers Langer, E.J. (1989). Mindfulness. Cambridge, MA: Da Cappo Press.
Djikic, Langer, and Stapleton (2008) found, using a similar design Langer, E. J. & Moldoveanu, M. (2000). The construct of mindfulness. Journal
to the present study in which after a partial debriefing participants of Social Issues, (56), 1-9.
Psychology
walked between two experimenters, that when participants were Langer, E. J. & Newman, H. (1979). The role of mindlessness in a typical social
psychology experiment. Personality and Social Psychology Bulletin, (5),
primed with photos of elderly people, they walked significantly faster 295-298.
when encouraged to be mindful. Lombardi, W.J., Higgins, E.T., & Bargh, J.A. (1987). The role of consciousness
It is also possible that the present study’s mindfulness instructions in priming effects on categorization: Assimilation versus contrast as a func-
were not powerful enough, since some participants mentioned upon tion of awareness of the priming task. Personality and Social Psychology
debriefing that they only remembered to “notice similarities and dif- Bulletin, 13(3), 411-429.
Shih, M., Ambady, N., Richeson, J. A., Fujita, K., & Gray, H. M. (2002). Stereo-
ferences” after the first few scrambled sentences. Nevertheless, the
type performance boosts: the impact of self-relevance and the manner of
fact that such a subtle manipulation, coupled with small variability in stereotype activation. Journal of Personality and Social Psychology, 83 (3),
walking time because of the short distance, yielded a medium effect 638-647.
size in the expected direction is promising for future studies that can
www.thurj.org 69
8JUIQBSUJDVMBSUIBOLTUP%FBO'MJFS %S'SFFENBO 1SPWPTU)ZNBO
'BDVMUZPG"SUTBOE4DJFODFT '"4
%FBO4NJUI BOE'"4%FBOGPSUIF
1IZTJDBM4DJFODFT+FSFNZ#MPYIBNGPSUIFJS
JOWBMVBCMFTVQQPSUJOQVCMJTIJOHUIJTJTTVF
You’ve done the research.
You’ve written the report.
Why not submit to THURJ?
We are Harvard’s undergraduate research journal,

dedicated to showcasing the research* of Harvard
undergraduate students.
Every research manuscript that is submitted is peer

reviewed by us and by members of our Faculty Advisory
Board, composed of Harvard faculty including Dean of
Harvard Medical School Jeffrey Flier, Nobel Laureate
Dudley Herschbach, and psychologist Steven Pinker.
We are always accepting submissions, so check

http://www.THURJ.org for more concrete dates and
details.
Did we mention there is a $500 prize for best research

manuscript?
*We publish research in all science and many social

science fields including biology, economics, statistics,
physics, psychology, chemistry, and many more.

THURJ Vol. 2 Issue 1

Enviado por

Dados do documento

Descrição original:

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

THURJ Vol. 2 Issue 1

Enviado por

Direitos autorais:

Formatos disponíveis

$POHSBUVMBUJƲOTȁƺ

Dear Harvard Community,

The Harvard Undergraduate Research Journal

Dear Harvard Community,

John Zhou Lisa Rotenstein

ii The Harvard Undergraduate Research Journal

Visit http://www.thurj.org for news, details about the organization,

iv The Harvard Undergraduate Research Journal

Faculty Advisory Board David Jeruzalmi, Ph.D Special Thanks to Our

Hongkun Park, Ph.D Robin Greenwood, Ph.D

vi The Harvard Undergraduate Research Journal

Biomechanics: Elasticity in ionically cross-linked neurofilament networks

Evolutionary Biology: Evolution of marine cyanobacteria in the Red Sea

2 The Harvard Undergraduate Research Journal

Materials Science: Properties of silk III fibroin at the air-water interface

The macromolecule fibroin,

Illustration by Charlotte Seid/THURJ Staff

Most genomes are made

Illustration by Evelyn Park/THURJ Staff

4 The Harvard Undergraduate Research Journal

Illustration by Katherine Goldin/THURJ Staff

“This interdisciplinary enterprise… boldly confronts some

Illustration by Katherine Goldin/THURJ Staff

10 The Harvard Undergraduate Research Journal

“The advantage of CNS is that you have a

Nature does not distinguish whether a phenomenon

applications.” In one of its projects, Capasso’s group circuits.

2-dimensional flat chip. Dr. Lieber has been crafting

The Future of the CNS

C apasso, Westervelt, Lieber and many other faculty

Online collaboration and its implications

Sociological Association in 2008

Illustration by Kathleen Tang/THURJ Staff

“How does the

be much more solo selfishness?” The Center’s research

16 The Harvard Undergraduate Research Journal

It’s a robotic fly!

!"#$%&''(#)*+(,(-# most undergaduates’ travels—you suggestions and sharing in their

Photo by John Mei/THURJ Staff

tion of a mobile robot,” said Wood.

Graduate student: Ben Finio Trotter is working on creating an experimental device

20 The Harvard Undergraduate Research Journal

Photo by Rose Lincoln/Harvard News Office

Hal Varian, Google’s Chief

“I keep saying the sexy job in the

deal with the situation when we

dating on-line models!),

Illustration by Evelyn Park/THURJ Staff

24 The Harvard Undergraduate Research Journal

26 The Harvard Undergraduate Research Journal

“[The Superorganism] is a telling account of a social

30 The Harvard Undergraduate Research Journal

Elasticity in ionically cross-linked

32 The Harvard Undergraduate Research Journal

0.5 mg/ml 2mM Mg2+

therefore hypothesize that the network elasticity is due to the stretch-

14 spring. To consider the entropic effects we can model the Mg2+-cross-

segments of length lc, where lc is the average distance between Mg2+

5mM Mg2+ & 2.0 mg/ml 3/2

34 The Harvard Undergraduate Research Journal

36 The Harvard Undergraduate Research Journal

Evolution of marine cyanobacteria in

38 The Harvard Undergraduate Research Journal