481621

research-article2013

JBXXXX10.1177/1087057113481621Journal of Biomolecular ScreeningBusch et al.

Technical Note

Does Your Lab Coat Fit to Your Assay?

Journal of Biomolecular Screening

18(6) 744747
2013 Society for Laboratory
Automation and Screening
DOI: 10.1177/1087057113481621
jbx.sagepub.com

Michael Busch1, Heinz Bjoern Thoma1, and Ingo Kober1

Abstract
An explanation for randomly occurring spikes on microplates in fluorescence-based assays employing shorter-wavelength
readouts is presented. It is demonstrated that lint originating from standard (white cotton) lab coats is most likely to be
responsible for such artifacts in assays applying wavelengths at 380 nm excitation and 450 nm emission. The fluorescence
properties of this lint are discussed and compared with those of optical brighteners. An alternative to the use of cottonbased lab coats is presented, which led to a reduction of spikes in a high-throughput screening campaign by 90%.
Keywords
HTS, fluorescence, spikes, lab coat, dust, lint

There are many laboratories developing assays involving

different ranges of wavelengths depending on the target
under study.
The projects they serve vary and hence so do the type and
format of assays. Very often, the readouts of these assays are
optical ones (e.g., fluorescence-based measurements).
False-positives due to compound interference
especially at shorter wavelengthsare a well-known phenomenon in high-throughput screening (HTS).1 The wavelengths used in most cases can be chosen in a way that
compound interference is minimized. Longer wavelengths
(red region >500 nm) are useful here.2
However, in some cases, the use of shorter wavelengths
cannot be avoided, for example, when coumarin derivatives
are used in protease assays. This needs a certain amount of
attention, not only with regard to compound interference.
Here we show examples of how lint shed from standard
lab coats interferes with fluorescence-based assays run at
about 380 nm excitation and emission at about 450 nm.
We describe our findings during assay development of
fluorescence-based assays at these wavelengths with respect to
spikes, which were found to be abundant under standard lab
conditions. We could track most of the spikes back to lint originating from standard white cotton lab coats. This kind of lab
coat is used worldwide, and the coats are likely to be cleaned
according to common industry cleaning procedures using
detergents and brighteners. The latter seem to interfere with the
assay signals in the wavelength region mentioned above.

Materials and Methods

Measurements were performed on an Envision 2104 from
Perkin Elmer (Waltham, MA). The fluorescence measurements

were performed at ex 380 nm, em 450 nm, five flashes per well,
1 s per well. For the three-dimensional (3D) spectrum, the
monochromator function of this instrument was used.
For the tests, empty microplates were used, which had
been exposed to different types of lab coats in order to collect lint from them (see the text below). Microplates were
from Greiner (384 wells, black, small volume).
Standard lab coats were rented from Bardusch and
cleaned according to standard industry-grade washing procedures. Disposable plastic lab coats were from different
vendors (see the text below).

Results and Discussion

During the assay development of two different assays formats, both involving excitation wavelengths at about 380
nm and emission wavelengths at about 450 nm, spikes (i.e.,
strong and unexpected individual fluorescence signals)
were found regularly when running the assays in microplates. One of the assays was based on fluorescence intensity and the other on fluorescence lifetime. In both cases,
about 5 to 10 randomly distributed spikes (signals up to 8
times higher than a normal assay signal) were found on
1

Global Research & Early Development, Small Molecule Platform,

Molecular Pharmacology, Merck Serono, Darmstadt, Germany.
Received Oct 30, 2012, and in revised form Jan 7, 2013; Feb 6, 2013.
Accepted for publication Feb 8, 2013.
Corresponding Author:
Michael Busch, Global Research & Early Development, Small Molecule
Platform, Molecular Pharmacology, Merck Serono, Frankfurter Str. 250,
D 64293 Darmstadt, Germany.
Email: michael.busch@merckgroup.com

Downloaded from jbx.sagepub.com at NANYANG TECH UNIV LIBRARY on April 25, 2015

745

Busch et al.
53
53
57
68
62
69
52
58
78
68
62
58
54
55
42
56

62
40
56
47
55
64
60
72
75
70
63
65
48
60
67
54

63
60
68
60
55
72
59
66
52
62
64
62
53
56
74
57

70
56
74
74
67
64
62
77
58
57
76
57
47
56
65
71

77
70
66
75
66
63
71
86
87
50
57
54
59
61
53
60

62
57
77
65
65
62
86
77
65
80
75
86
63
81
59
64

61
64
53
76
51
55
59
58
62
70
74
71
55
68
62
80

58
64
44
60
51
53
55
64
66
68
55
61
55
60
71
76

61
57
69
62
48
66
60
63
71
44
65
64
55
62
67
49

69
59
67
54
64
64
61
61
56
51
69
69
68
64
67
72

70
73
60
65
82
89
76
73
66
65
62
63
53
59
63
62

53
70
66
67
61
71
73
57
55
59
53
64
51
66
65
66

63
63
73
66
86
59
73
58
46
71
51
63
50
48
58
62

82
59
67
58
78
65
68
57
57
77
72
50
73
75
48
60

60
65
63
75
62
67
62
61
73
69
50
66
65
64
57
87
RFU:

82
60
51
58
68
52
63
59
54
71
69
52
56
42
60
64

63
67
62
54
59
67
71
55
80
68
65
44
80
68
53
72

50
57
78
71
68
58
87
60
84
77
59
65
65
62
50
63

56
65
73
61
56
59
74
80
80
71
55
60
80
63
69
59

10

100

500

1000

61
46
44
61
55
60
66
52
54
64
65
60
63
55
67
67

59
75
56
79
74
70
56
47
63
61
74
59
79
80
43
61

90
82
59
60
80
49
73
51
76
63
70
62
82
76
40
56

66
83
62
74
61
74
82
68
58
44
69
64
57
60
46
58

63
57
48
62
60
63
79
60
65
92
71
68
51
50
35
55

5000 10000 20000 30000 40000

Figure 1. Empty microplate, freshly taken out of a closed bag. Fluorescence readout at 380/450 nm ex/em. The medium signals were
less than 70 RFU.

53
63
69
66
59
61
57
61
59
56
47
46
67
79
57
65

81
60
89
68
65
56
64
86
56
75
67
64
78
76
1195
61

182
71
50
42
83
59
54
71
43
53
40
90
61
56
68
73

58
63
67
91
72
74
59
59
146
69
57
63
77
67
61
50

61
60
73
71
66
68
85
58
68
59
89
200
69
60
59
51

113
67
59
82
59
61
81
63
53
67
56
81
49
75
61
59

69
69
56
63
47
71
50
69
68
240
44
73
88
72
61
71

56
48
51
76
54
60
52
52
62
60
3448
64
65
57
59
80

54
52
64
58
57
66
54
74
56
50
80
54
76
65
63
49

63
38
51
50
76
49
66
66
57
66
58
66
66
65
56
74

58
704
62
81
72
58
63
68
61
76
56
59
81
61
76
53
67
61
64
46
56 10122
55
50
68
74
52
59
53
64
228
49

66
54
269
60
84
56
62
89
58
61
53
66
70
61
57
66

53
63
81
65
76
53
58
50
67
81
62
61
53
64
48
52

51
56
50
47
63
56
63
59
71
66
65
48
55
53
57
66
RFU:

60
57
59
60
47
50
59
52
52
43
85
59
63
59
57
62

87
78
64
82
59
50
68
55
66
73
59
63
44
53
49
78

79
108
59
52
62
59
62
47
79
68
70
47
55
52
64
82

87
64
54
75
74
63
49
48
60
57
54
59
86
64
48
106

10

100

500

1000

83
61
70
67
79
46
73
52
62
77
69
69
48
137
2171
56

57
66
94
68
58
61
70
70
70
636
66
67
58
58
58
57

56
68
55
61
88
70
71
68
63
60
68
54
56
58
58
64

87
63
49
73
65
61
75
88
64
143
66
71
76
53
60
69

108
58
67
69
58
69
48
44
76
75
62
65
58
73
60
54

5000 10000 20000 30000 40000

Figure 2. Measurement of the same plate after moving it for 2 min in free laboratory air. Spikes up to 10000 RFU were observed
here.

each 384-well plate. This would have led to 1% to 2% of

false-positives in HTS applying these assays.
The occurrence of such spikes was not dependent on the
brand or type of the microplates. Moreover, it was not
dependent on the laboratory: Even a lab at a different company (working on the same assay) reported such spikes.
Because there were no alternatives regarding the choice
of wavelengths, the question arose as to the origin of these
artifacts, and options for their effective prevention were
considered. To spot the source of the spikes, a series of
experiments was conducted using empty microplates. The
use of empty microplates clearly demonstrates and overemphasizes the effects.
Figure 1 shows the situation found in which a microplate was taken out of a closed bag and measured immediately at ex 380/em 450 nm. Here, no spikes were found: All
of the signals were in the range of less than 70 RFU.
Figure 2 shows measurements (ex 380 nm/em 450 nm)
of the same plate after simply moving it about for 2 min
freely within the lab. We obtained 13 spikes (framed in the

picture) that had a signal larger than twofold of the base

signal. Some spikes were found up to 100-fold of the base
signal (colored spots in the picture). Repetitions of this type
of test on different days showed similar results. Microplates
taken out of bags, which had been open for a while, showed
similar results. Some kind of dustfree floating within the
labmust have been the reason for these spikes.
Apart from dust particles originating, for example, from
the air conditioning system, a different reason could be that
those particles might be the result of lint arising from abrasion and wear and tear of lab coats during their use in
daily work. To test this hypothesis and to find out the origin
of the dust particles, several types of lab coats were tested
for lint arising from abrasion.
A microplate was sealed partly, leaving only four columns
open for access. The plate then was vigorously rubbed against
a lab coat for 3 to 5 s. The open area was sealed, a new area was
opened, and it was rubbed against a different coat. Figure 3
compares the rate of spikes obtained from a standard freshly
washed cotton lab coat to different brands of disposable plastic

Downloaded from jbx.sagepub.com at NANYANG TECH UNIV LIBRARY on April 25, 2015

746

Journal of Biomolecular Screening 18(6)

60
46
48
52
41
48
80
58
56
60
45
47
63
66
49
57

62
57
53
53
58
48
59
47
49
57
59
51
65
65
63
41

56
60
62
57
54
62
65
73
59
71
73
49
47
61
43
43

59 1660 5319 2565 5427

55 1981 4690 4416 1473
62 3554 3404 6036 6581
51 5424 3640 643016072
56 4509 3177 17771 4161
65 4921 3393 10841 2838
54 9127 6042 3658 5257
66 32008 8995 5203 4573
52 4362 6533 7348 1889
61 22742 3310 2409 5202
53 1537 2562 2359 6285
67 3644 574 9305 3381
51 3545 3111 5270 1450
52 3189 3450 2315 3186
61 6107 6705 7962 6642
49 7295 3348 7684 7490

73
51
53
44
51
50
51
50
208
47
51
48
351
53
36
53

66
48
49
60
51
79
63
68
52
68
54
53
36
60
47
59

61
55
73
61
63
40
71
74
80
55
65
47
69
61
59
59

60
46
60
56
45
55
51
56
48
60
74
61
51
40
42
53

48
58
62
60
38
66
45
58
68
60
62
69
46
50
81
55

65
45
50
42
55
55
68
45
53
64
65
62
37
70
44
53

45
81
46
46
61
45
47
38
533
46
57
59
55
44
90
70
RFU:

49
39
64
64
59
72
54
63
62
55
66
49
36
59
60
49
10

59
55
47
47
51
54
64
64
59
61
46
45
70
60
58
53
100

49
55
70
54
99
61
67
56
54
58
50
58
55
68
65
60

57
49
61
62
51
83
58
74
70
65
64
62
50
156
90
55

71
41
63
50
52
63
66
50
80
56
66
62
58
49
67
72

52
53
176
50
38
57
69
56
99
67
50
72
48
63
36
39

64
62
56
55
54
64
50
49
145
55
53
50
64
62
39
131

47
77
221
200
52
72
76
36
120
68
70
63
279
59
71
54

68
61
85
73
72
54
62
79
66
72
41
53
55
52
61
37

500 1000 5000 10000 20000 30000 40000

Figure 3. An empty microplate partially exposed to different types of lab coats by extensively rubbing the exposed wells in the
material of the coats for 3 to 5 s. (A) No exposure. (B) Standard cotton white coats. (C) BioCleanD (Basan, Kelsterbach, Germany).
(D) Tyvek ClassicPlus (Rala, Ludwigshafen, Germany). (E) TyvekProTech (Rala). (F) PP coat (Rala). The colored wells showed signals
up to 30000 RFU.

3D spectrum of abraised lint

300000
250000

150000

RFU

200000

100000
393
402
411
420
429
438
447
456
465
474
483
492
501
510
300

coats. All disposable coats were made from plastic materials,

and none of these showed significant spikes as the signals were
in the range of 70 RFU. There are a few mini-spikes visible,
but they were presumably caused by contamination via lab air
during the de-sealing and sealing procedures. However, rubbing the plate against a white cotton coat resulted in drastic
spikes in all 64 wells. All wells showed signals at least 70-fold
above the base level of unexposed wells. Some of them were
up to 400-fold of the base level of the untreated wells. Here, the
rubbing test had captured abrasions from the surface of this
type of lab coat. When such plates were measured at ex 485/
em 520, no such high spikes were observed. At these wavelengths, the rubbing test revealed a maximum of 10-fold of
base signal in all wells exposed to cotton lab coats (data not
shown). When a reusable blue cleanroom coat was tested with
the rubbing method, the body part of this coat revealed the
same results as the disposable plastic coats showed: no significant spikes. Interestingly, the white wristband of this coat
caused spikes in the same range shown by the cotton coats.
To analyze the spectral properties of these rubbingderived lint samples, spectra were taken from wells, which
showed high signals in the tests mentioned above. All spectra were nearly identical. An example of a 3D spectrum is
shown in Figure 4. At an excitation of about 390 nm, the
fluorescence emission peak around 435 nm can be seen as
well as the low shoulder up to the 560 nm range. These
emission and excitation characteristics are in the same
range, where optical brighteners normally operate.3
Based on this good fit of the regions of wavelengths, we
concluded that the abrasions were likely to consist of optical brighteners, which are very often involved as laundry
additives in the cleaning process of reusable cotton white
working clothing.
As a consequence of these findings, employees and visitors in the labs were required to wear plastic disposable

emission / nm

50000
0

excitaon / nm

Figure 4. Three-dimensional spectrum of abrasions from

a white cotton lab coat captured in a well of a microplate.
Excitation from 300 to 475 nm, emission from 393 to 513 nm.

coats during the assay development period and, more

importantly, throughout the HTS campaign. This dress code
was started 3 wk prior to the start of any work on the assays
and the HTS campaign. The change of coats could not prevent all spikes but did reduce the amount of spikes down to
one to two per two microplates in the HTS campaign.
Compared with the situation without such preventative
measures (510 spikes per plate if cotton lab coats would
have been worn), this was a significant improvement: More
than 90% reduction of spikes could be achieved.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.

Downloaded from jbx.sagepub.com at NANYANG TECH UNIV LIBRARY on April 25, 2015

747

Busch et al.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.

References
1. Simeonov, A.; Jadhav, A.; Thomas, C. J.; Wang, Y.; Huang,
R.; Southall, N. T.; Shinn, P.; Smith, J.; Austin, C. P.; Auld,
D. S.; Inglese, J. Fluorescence Spectroscopic Profiling of
Compound Libraries. J. Med. Chem. 2008, 51, 23632371.

2. Shi, Z. D.; Motabar, O.; Goldin, E.; Liu, K.; Southall,

N.; Sidransky, E.; Austin, C. P.; Griffiths, G.L.; Zheng,
W. Synthesis and Characterization of a New Fluorogenic
Substrate for Alpha-Galactosidase. Anal. Bioanal. Chem.
2009, 394, 19031909.
3. Liu, M. O.; Lin, H. F.; Yang, M. C.; Lai, M. J.; Chang, C.
C.; Liu, H. C.; Shiao, P. L.; Chen, I. M.; Chen, J. Y. Thermal
and Fluorescent Properties of Optical Brighteners and Their
Whitening Effect for Pelletization of Cycloolefin Copolymers.
Materials Letters 2006, 60, 21322137.

Downloaded from jbx.sagepub.com at NANYANG TECH UNIV LIBRARY on April 25, 2015