Escolar Documentos
Profissional Documentos
Cultura Documentos
The appearance of multicellular forms of life has been The emergence of multicellularity required the simultaneous
tightly coupled to the ability of an organism to retain its development of cell adhesion and recognition properties. In
own anatomical integrity and to distinguish self from addressing the unresolved question of what may have been the
a N2 stream. Pronase (type E from Merck) was dissolved in 0.1 M Tris, content by Dionex high pH anion-exchange chromatography, both after
pH 8, 1 mM CaCl2 at a concentration of 1 mg/ml and preincubated for 20 6 M HCl hydrolysis at 100 °C for 12 h.
min at 60 °C. Subsequently 1 ml of preincubated pronase (1 mg/ml) was Preparation of Cells and Glyconectin-coated Beads—Sponge cells
added to 20 mg of dried glyconectin pellet and incubated for 12 h at were dissociated and washed in ACMFSW at 0 °C. In some experiments
60 °C. This proteolysis digestion procedure was repeated twice by add- cells were fixed with 1% glutaraldehyde (14, 15).
ing fresh aliquots of preincibated pronase. Released glycans were sep- A volume of 200 l of 2% latex-amidine bead suspension, 0.6 m in
arated from free amino acids, salts, and small peptides by gel filtration diameter (Molecular Probes, Inc., Eugene, OR), was washed three times
and subsequent ion exchange chromatography. with 1 ml of SWT by centrifugation at 3000 ⫻ g for 10 min. Beads were
Preparation of Monospecific Polyclonal Antibodies against Glyconec- resuspended in 200 l of SWT and bath-sonicated for 5 min. To allow
tin Carbohydrates—Purified glyconectin carbohydrates were used for adsorption to beads on the basis of charge, 100 g of either glyconectins
preparation of polyclonal antibodies according to the procedure de- or glycans isolated from these glyconectins was added to a 100 l of
scribed previously (14). Monospecific polyclonal antibodies were ob- bead suspension that was then incubated at 22 °C for 15 min followed
tained by affinity purification on GN glycans cross-linked to a solid by three washes as described above. The glyconectin-coated beads were
resuspended in 100 l of SWT. Because the diameter of the beads is 600
nm and the size of glyconectins are in the same range, each bead bound
1
The abbreviations used are: ACMFSW, artificial Ca2⫹- and Mg2⫹- at least 1 molecule of glyconectin and about 20 –50 large glycan
free seawater; SWT, seawater-Tris; CSW, Ca2⫹- and Mg2⫹-free sea- molecules.
water buffered with Tris; GC/MS, gas chromatography-coupled mass Recognition, Adhesion, and Overlay Assays with Cells and Beads—
spectrometry; GN, glyconectin; EI-MS, electronic impact mass spec- Cell-cell adhesion assays were performed with 106 cells bearing surface
trometry; MALDI-TOF MS, matrix-assisted laser desorption ionization glyconectins in 100 l of ACMFSW, without or with 10 mM Ca2⫹, in the
time-of-flight mass spectrometry; Py(4,6)Gal, 4,6-O-(1-carboxyethyli- presence or absence of anti-GN-glycan polyclonal antibodies. For glut-
dene)-␣-D-Galp; ACTH, adrenocorticotropic hormone; HPLC, high pres- araldehyde-fixed cells assays were done in SWT with or without Ca2⫹.
sure liquid chromatography. For the cell-bead assay, 5–10 l of 2% bead suspension in SWT was
Molecular Recognition between Glyconectins 15581
TABLE I
Physicochemical properties of sponge glyconectins
Molecular weight (Mr) and sedimentation coefficient (s20,w) of sponge
GNs were determined by analytical centrifugation. The carbohydrate
and amino acid composition (mol %) of glyconectins were obtained from
gas chromatography and HPLC, respectively. Carbohydrate/protein ra-
tio, content of uronic acids, and SO2⫺
4 were calculated from amino acid,
monosaccharide, and ion composition. The standard deviation from
three analyses was maximally 1% of each value (see “Experimental
Procedures”).
GN1 GN2 GN3
Mr ⫻ 10 6
19 10 8
s20,w (S) 58 42 46
Carbohydrate/protein ratio (w/w) 63/37 21/79 36/64
Monosaccharides (mol %)
Ara 0 0 11
Fuc 28 7 31
Man 9 16 20
Gal 29 43 12
GalNAc 1 1 7
GlcNAc 17 14 13
Py(4,6)Gal 4 16 0
GlcA 11 3 6
Amino acids (mol %)
Asx 9.1 7.8 11.1 FIG. 1. Electrophoretic separation of sponge glyconectins. A,
0.75% agarose gel stained with 0.02% toluidine blue followed by 0.1%
FIG. 3. Peptide mapping of glyconectins by mass spectrometry. Tryptic peptide maps of GN1 (top), GN2 (middle), and GN3 (bottom) are
shown.
15584 Molecular Recognition between Glyconectins
TABLE II
Monosaccharide composition of GN glycans
Carbohydrate and amino acid composition (mol %) of glyconectin glycans were obtained from gas chromatography after methanolysis and HPLC
after HCl hydrolysis, respectively, as described under “Experimental Procedures.” Carbohydrate/protein ratios and content of uronic acids and
SO2⫺
4 were calculated from amino acid and monosaccharide composition. The standard deviation from three analyses was maximally 1% of each
value (see “Experimental Procedures”). The total amount of SO2⫺
4 recovered in GN glycans after pronase treatment of each GN was more than 90%.
data together with compositional and electrophoretic analyses that the large glycans of GN2 and -3 are N-linked to the protein
specify that three GN carbohydrates have a novel type of spe- core as in the case of both GN1 glycans (16, 18, 29). To confirm
cies-specific and repetitive structure residing in the large acidic the nature of the glycan-protein linkage, a more quantitative
polysaccharides. More detailed chemical sequence analyses of analysis of the isolated linkage region must be made. This will
GN glycans combined with NMR and mass spectrometry are require a new methodology to take into account the spacing of
described in the accompanying article (40). glycan chains, their internal sequences and charge distribu-
The linkage type of GN glycans to the protein core was tion, as well as the peptide sequences of the core protein (which
studied using a chemical and enzymological approach. We an- greatly influence enzymatic and chemical cleavage). Our struc-
alyzed the total glyconectin glycans using the method of Maes tural data of GN polysaccharides presented in the accompany-
et al. (30), which shows that classical methanolysis conditions ing article (40) promote the development of sequencing tech-
quantitatively cleave the N-glycosidic bond (96%), liberating niques applicable to novel classes of glycans. Such potential
glucosamine (and not its O-methylglycosides). Because other methods, when used on the purified GN fragments close to the
N-acetyl-D-glucosamine (GlcNAc) residues are quantitatively core protein together with cloning of the GN genes, will defin-
liberated as the O-methylglycosides of glucosamine, the Glc- itively confirm the exact nature of GN glycan core and linkage
NAc residue involved in the N-glycosidic bond is separated to protein.
from the others using gas chromatography of heptafluorobu- Taking in account the amount of carbohydrate present per
tyrate derivatives. GC/MS indicated that all three GNs have intact glyconectin, the quantity of each individual glycan spe-
N-linked glycans. Because of the large size of the glycans, the cies recovered from the respective glyconectin, the size of the
great amounts of internal GlcNAc, and the comparatively small glycans, and the size of the intact glyconectins, a rough esti-
amount of possible linkage GlcNAc, gas chromatography and mation was made that: 1) GN1 would have 20 chains of 200
mass spectroscopy data of GN-type molecules cannot be abso- kDa and 950 chains of 6 kDa glycans as discussed previously
lutely quantitative in terms of the exact number of moles of (14, 15, 18); 2) GN2 would have 10 repeats of 180 kDa glycan
linkage GlcNAc but rather are of a qualitative nature. Classical and 300 glycans with an average size of 6 kDa; and 3) GN3
alkali treatment of intact glyconectins was able to only par- would have 13 copies of 110 kDa glycans. The rest of the
tially release glycans (about 30 –50%), as in the classical exam- size-heterogeneous population of GN3 polysaccharides showing
ples of N-linked glycans. Peptide-N-glycanase was not able to tailing from the 110 kDa glycan band on the gel is most likely
release any of the large glycans but was able to release the either the result of degradation or the unfinished synthesis
small 6-kDa glycan in the case of GN1 as reported previously product of this major glycan species (see accompanying article
(29). Most likely peptide-N-glycanase did not cleave because of (40)). We cannot completely exclude the possibility that this
the well known steric inaccessibility of the linkage due to the product is also mixed with some less abundant yet undefined
high charge and large size of the glycans. Hydrazinolysis of polysaccharide species.
intact GNs under standard conditions released glycans; how- Glyconectin Glycoconjugates are Cell Adhesion Molecules—
ever, these conditions are now known to cleave both N-linked The cell adhesive function of two new sponge glyconectins
and O-linked glycans. Gas chromatography coupled to mass purified from H. panicea (GN2) and C. celata (GN3) was tested
spectroscopy analyses and alkali treatment of GNs indicate and compared with that of M. prolifera (GN1) in a rotary
Molecular Recognition between Glyconectins 15585
force microscopy and glyconectin-coated bead adhesion that
Ca2⫹-dependent self-interactions between glyconectin 1 from
M. prolifera sponge provides the major driving force for cell
adhesion (1, 14, 15). To quantitatively test whether glyconec-
tins 2 and 3 from sponges H. panicea and C. celata also utilize
the same adhesive molecular mechanism via Ca2⫹-dependent
glyconectin to glyconectin interactions, we used two assay sys-
tems, Ca2⫹-induced aggregation of glyconectin-coated beads
(Fig. 6B) and glyconectins self-interaction monitored by gela-
tion (Fig. 6C). All three glyconectins revealed 100% self-asso-
ciation at 10 mM CaCl2 and no interactions in the absence of
calcium ions in either assay system (Fig. 6, B and C; see also
bead recognition assay in Figs. 8 and 9). CaCl2 could not be
substituted with MgCl2 indicating a specific role of Ca ions
(Fig. 6, E and F). Titration experiments of Ca2⫹ concentration
dependence of sponge glyconectin self-interactions revealed a
transition at 5 mM and 100% interactions at physiological 10
mM CaCl2 (Fig. 6, B and C), identical to that of Ca2⫹-dependent
glyconectin-promoted cell adhesion (Fig. 6A). These experi-
ments indicated that Ca2⫹-dependent glyconectin to glyconec-
tin interactions play a pivotal role in the cell adhesion of the
FIG. 5. NMR analyses of glyconectin glycans. A, one-dimensional spectra of GN1 glycans (top), GN2 glycans (middle), and GN3 glycans
(bottom). B, two-dimensional COSY90 spectra of GN1 (top), GN2 (middle), and GN3 (bottom).
CaCl2 (Fig. 8B). Both processes occurred at apparently similar white beads, as expected, mixed-color aggregates were formed
rates for each of the three glyconectins. In control experiments upon the addition of 10 mM CaCl2 (not shown). In the absence
with glyconectin 1 separately attached to pink, yellow, and of 10 mM CaCl2, bead aggregation did not occur either in the
Molecular Recognition between Glyconectins 15587
mixture of the three glyconectins or in one glyconectin coated to 9, B, C, D, F, G, and H). Cell-bead recognition and adhesion
three colored beads (same as Fig. 8A). occurred at apparently equal rates and ended in similar aggre-
A similar type of experiment was performed by overlaying gate sizes for all three species. Such recognition took place only
agarose gel containing three electrophoretically separated gly- when cells had glyconectin on their surfaces and in the pres-
conectins with color-coated glyconectin beads. After overnight ence of 10 mM CaCl2. This process did not require live cells
incubation at room temperature in the presence of 10 mM CaCl2 because the same results were obtained with glutaraldehyde-
under gentle agitation, species-specific staining of gel glyconec- fixed and/or metabolically attenuated cells at 0 °C (data not
tin bands identical to the ones stained with toluidine blue and shown). Self-assembly experiments with live cells, metaboli-
Amido Black showed that glyconectin to glyconectin interac- cally attenuated cells, fixed cells, and non-natural latex beads
tions are highly species-specific (Fig. 1). showed that, independently of the carrier surface, glyconectins
The combinations of the above described experiments dem- are able to recognize self and distinguish self from non-self via
onstrate species-specific molecular self-recognition of glyconec- cohesive glyconectin to glyconectin interactions.
tins in an elementary reconstituted bead adhesion system, Glyconectin Carbohydrates Mediate Cell Recognition and Ad-
which fully resembles glyconectin-mediated cell-cell recogni- hesion—The final question was to examine whether glyconec-
tion and adhesion. Thus, glyconectins mediate self- and non- tin carbohydrates are involved in cell adhesion and recognition
self-discrimination in the initial step of sponge cell adhesion processes during reaggregation of Porifera cells. The facts that
and xenogeneic recognition. The degree of GN selectivity is GNs are carbohydrate-rich molecules and that GN1 glycans
analogous to that of antibody-antigen interactions in similar specifically mediate cell adhesion in M. prolifera led us to test
assay systems. for the cell adhesive function of GN2 and GN3 carbohydrates
Self-assembly of a Biological and Artificial Cell-Bead System using two different approaches. In the first immunological
via Species-specific Glyconectins Recognition—To evaluate the method, we prepared affinity-purified monospecific polyclonal
exclusive role of glyconectins in self-recognition we tested spec- antibodies against isolated GN1, -2, and -3 total glycans, es-
ificity of homophilic glyconectin-glyconectin interactions in a sentially free of proteins (see Table II), as described under
color-coded cell-bead adhesion assay. Each of the three gly- “Experimental Procedures.” Then we examined whether such
conectins was attached to color-coded beads, as in the above anti-GN-glycan antibodies can specifically inhibit cell aggre-
described experiments, and beads were mixed in nine combi- gation of the three respective sponge species. The results
nations with glyconectin-bearing cells of the three sponge spe- presented in Fig. 10A showed that each of the three anti-GN-
cies in the presence of 10 mM Ca2⫹. To simultaneously follow glycan polyclonal antibodies, anti-GN1-glycan, anti-GN2-gly-
cells and beads, visible light microscopy was used. In visible can, and anti-GN3-glycan, species-selectively inhibited cellular
light, fluorescent pink beads with glyconectin 1 appear as pink, adhesion in the presence of 10 mM CaCl2 of the respective
fluorescent green beads with glyconectin 2 are yellow, and GN-bearing cells. In the second direct approach, we coated
fluorescent blue beads with glyconectin 3 are white. Bead-cell preferentially large GN glycans, which were shown to be es-
coaggregation was observed for homotypic combinations, as sentially free of proteins (see Table II), to color-coded latex-
observed microscopically by the color and shape of uniformly amidine beads using the same adsorption procedure as de-
mixed beads and cells (Fig. 9, A, E, and I). The heterotypic scribed for the intact GNs. The adsorption is based on charge
combinations always resulted in bead and cell segregation (Fig. interactions, which in high salt concentrations are much
15588 Molecular Recognition between Glyconectins
studies on the synthetically prepared sulfated epitope of GN1 ated glyconectin to glyconectin interactions in Porifera. How-
directly showed that this oligosaccharide is capable of self- ever, the structural differences between these two systems
association (31). In the initial study using three sponge species imply conceptually distinct molecular mechanisms. First, gly-
(M. prolifera, H. panicea, and C. celata), it was shown that the conectins are 100 times larger and extend 10 times farther
two new glyconectins also are involved in species-specific ag- from the cell surface than immunoglobulin molecules. Second,
gregation (13). Here we have isolated and performed detailed the glyconectin 1 specificity and tight binding of ⬎109 M⫺1
physicochemical and functional characterizations of glyconec- reside in polyvalent carbohydrate-carbohydrate interactions of
tin 2 from H. panicea and glyconectin 3 from C. celata and ⬎1000 sites, with a low affinity for the single site (⬍103 M⫺1)
extended the analyses of glyconectin 1 from M. prolifera. In this (14, 15, 29, 32–34), whereas immunoglobulins recognize anti-
comparative study of three glyconectins we have found that gens via higher affinity ranging from 104 to 109 M⫺1, with low
they have species-specific and complex repetitive glycan struc- valency binding (34 –38). The low affinity of the monovalent
tures with some common properties for isolated glycans such as glyconectin oligosaccharide units needs reconstitution of the
a large size, a high charge content (sulfate and/or pyruvate), naturally occurring valency in GNs to recover the adhesion
and N-acetylated hexose, pentose, or deoxyhexose, which clas- function. Therefore, the use of classical inhibition experiments
sifies them as a new type of polysaccharide molecule distinct with monovalent glycan units, commonly utilized as evidence of
from classical glycosaminoglycans. A more detailed description glycan function in the conceptually different lectin-carbohy-
of this new type of glycan sequences and their possible struc- drate type of interactions, is not applicable for GN glycans.
tures is reported and discussed in the accompanying article Here reported carbohydrate-mediated glyconectin-glyconectin
(40). interactions may provide a new paradigm for initial molecular
The degree of selectivity observed in the evolutionarily ad- self-recognition.
vanced, heterophilic, self and non-self molecular recognition The close evolutionary relationship between the Porifera of
within the immunoglobulin superfamily is closely approached today and the ancestors of multicellular organisms (33–35)
by the degree of specificity of homophilic carbohydrate-medi- indicates that xenogeneic discrimination of self-cohesive gly-
15590 Molecular Recognition between Glyconectins
conectins may have been essential for the appearance of mul- 10. Humphreys, T. (1963) Dev. Biol. 8, 27– 47
11. Jumblatt, J. E., Schlup, V., and Burger, M. M. (1980) Biochemistry 19,
ticellularity. Clearly the evolution from primordial Metazoans 1038 –1042
to complex organisms required the development of additional 12. Muller, W. E., and Zahn, R. K. (1973) Exp. Cell Res. 80, 95–104
13. Popescu, O., and Misevic, G. N. (1997) Nature 386, 231–232
cell recognition and adhesion mechanisms mediated via immu- 14. Misevic, G. N., Finne, J., and Burger, M. M. (1987) J. Biol. Chem. 262,
noglobulin, lectin, integrin, and cadherin families of molecules. 5870 –5877
Such an increase in molecular diversity enabled a correspond- 15. Misevic, G. N., and Burger, M. M. (1993) J. Biol. Chem. 268, 4922– 4929
16. Fernandez-Busquets, X., Kammerer, R. A., and Burger, M. M. (1996) J. Biol.
ing increase in functions based on a more demanding self- and Chem. 271, 23558 –23565
non-self-discrimination such as embryonal morphogenesis, re- 17. Spillmann, D., Thomas-Oates, J. E., van Kuik, J. A., Vliegenthart, J. F.,
Misevic, G., Burger, M. M., and Finne, J. (1995) J. Biol. Chem. 270,
newal of adult tissues, and defense from foreign pathogens. 5089 –5097
Despite this evolving complexity of the components participat- 18. Jarchow, J., Fritz, J., Anselmetti, D., Calabro, A., Hascall, V. C., Gerosa, D.,
ing in cellular interactions, glyconectin-like structures have Burger, M. M., and Fernandez-Busquets, X. (2000) J. Struct. Biol. 132,
95–105
been preserved in mammalian systems, suggesting their ver- 19. Brune, D. C. (1992) Anal. Biochem. 207, 285–290
satility in recognition and adhesion (39). A model system for 20. Misevic, G. N. (1989) Methods Enzymol. 179, 95–104
21. Kamerling, J. P., Gerwig, G. J., Vliegenthart, J. F., and Clamp, J. R. (1975)
primordial xenogeneic self-recognition and adhesion based on Biochem. J. 151, 491– 495
the simplest Metazoans of the phylum Porifera gives results 22. Zanetta, J. P., Timmerman, P., and Leroy, Y. (1999) Glycobiology 9, 255–266
consistent with the hypothesis that the emergence of the first 23. Ciucanu, I., and Kerek, F. (1984) Carbohydr. Res. 131, 209 –217
24. Hassell, J. R., Kimura, J. H., and Hascall, V. C. (1986) Annu. Rev. Biochem. 55,
multicellular organisms, as well as the divergence of species 539 –567
and the appearance of more complex multicellular forms of life, 25. Hascall, V. C., and Kimura, J. H. (1982) Methods Enzymol. 82, 769 – 800
may have been achieved by the evolution of glyconectin-like 26. Hascall, V. C., and Heinegard, D. (1974) J. Biol. Chem. 249, 4242– 4249
27. Dudman, W. F., and Lacey, M. J. (1986) Carbohydr. Res. 145, 175–191
proteoglycan molecules with the capacity for self-recognition 28. Spillmann, D., Hard, K., Thomas-Oates, J., Vliegenthart, J. F., Misevic, G.,
and adhesion via carbohydrates. Burger, M. M., and Finne, J. (1993) J. Biol. Chem. 268, 13378 –13387
29. Misevic, G. N., and Burger, M. M. (1990) J. Biol. Chem. 265, 20577–20584