HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 43, 40055-40064, October 26, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/43/40055    most recent M105557200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Seigneuret, M. Articles by Conjeaud, H. Search for Related Content PubMed PubMed Citation Articles by Seigneuret, M. Articles by Conjeaud, H. Social Bookmarking                      What's this?

Structure of the Tetraspanin Main Extracellular Domain

A PARTIALLY CONSERVED FOLD WITH A STRUCTURALLY VARIABLE DOMAIN INSERTION^*

Michel Seigneuret, Alix Delaguillaumie^§, Cécile Lagaudrière-Gesbert^§¶, and Hélène Conjeaud^§

From the  Unité Mixte de Recherche, 7033 CNRS, Laboratoire de Physicochimie Biomoléculare et Cellulaire, Université Paris 6, 4 Place Jussieu, 75005 Paris, France and the ^§ U332 INSERM, Institut Cochin de Genetique Moleculaire, 22 Rue Méchain, 75014 Paris, France

Received for publication, June 15, 2001, and in revised form, July 31, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The tetraspanin family of membrane glycoproteins is involved in the regulation of cellular development, proliferation, activation, and mobility. We have attempted to predict the structural features of the large extracellular domain of tetraspanins (EC2), which is very important in determining their functional specificity. The tetraspanin EC2 is composed of two subdomains: a conserved three-helix subdomain and a variable secondary structure subdomain inserted within the conserved subdomain. The occurrence of key disulphide bridges and other invariant residues leads to a conserved relative topology of both subdomains and also suggests a structural classification of tetraspanins. Using the CD81 EC2 structure as a template, the structures of two other EC2s were predicted by homology modeling and indicate a conserved shape, in which the variable subdomain is located at one side of the structure. The conserved and variable subdomains might contain sites that correspond, respectively, to common and specific interactions of tetraspanins. The tetraspanin EC2 seems to correspond to a new scheme of fold conservation/variability among proteins, namely the insertion of a structurally variable subdomain within an otherwise conserved fold.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Tetraspanins are a family of membrane glycoproteins that seem to be involved in various aspects of the regulation of cellular development, proliferation, activation, and mobility. Several recent studies suggest that their role is at least in part mediated by their ability to interact with other proteins such as integrins, coreceptor molecules, and major histocompatibility complex antigens as well as other tetraspanins (1-11). Several tetraspanins have also been shown to be involved in binding of specific viruses (12-16) or toxins (17-19). Current hypotheses view tetraspanins as "molecular facilitators," the function of which would be to bring into close proximity other proteins involved in cellular processes (for a review see Ref. 19). Such properties might lead to the formation of large membrane complexes of the involved proteins that may be associated with lipid rafts (20). Tetraspanins are type II proteins characterized by four transmembrane segments (TM1-4)¹ with both intracytoplasmic N- and C-terminal extremities linked by one short extracellular, one short intracellular, and one large extracellular (EC2) stretch. In addition to these general features, tetraspanins possess a number of conserved residues in both the intra- and extramembrane regions. Among these is the so-called tetraspanin signature, which consists of a specific stretch of residues located in the TM2/intracellular loop/TM3 region. Distant members of the tetraspanin family also exist and share the same transmembrane pattern and many conserved residues but lack the tetraspanin signature. As for "true" tetraspanins, the functions of many tetraspanin-like proteins such as RDS/ROM proteins or uroplakins appear to involve self-association and/or association with other proteins (21, 22).

Some data indicate that part of the specific activity of tetraspanins is determined by the large extracellular EC2 region (23-25). Unlike the TM regions that are significantly conserved, this region is highly variable in size and sequence composition. There are nonetheless a few invariant residues, which include an ubiquitous CCG motif. Recently Kitadokoro et al. (26) reported that the crystallographic structure of a soluble form of the tetraspanin CD81 EC2 domain is a five-helix bundle stabilized by two disulfide bridges and suggest that key structural features and the protein fold are conserved among tetraspanins. Such a hypothesis must be tested with other tetraspanins, especially in view of the very low conservation and variable size of the EC2 region. Indeed, the central part of the EC2 region (located between the CCG motif and the last canonical cysteine), which in CD81 corresponds to a 32-residue stretch that contains three helices, varies widely in composition and length for other tetraspanins ranging from 24 to 87 residues.

To investigate these features in more detail, we have attempted to predict the structure of the tetraspanin EC2 of all tetraspanin and tetraspanin-like proteins using multiple sequence alignments, secondary structure, and accessibility prediction methods as well as homology modeling. The predictions obtained for CD81 are in agreement with the experimental structure (26), indicating the reliability of this approach. On the other hand, our data indicate that the structural features of the CD81 EC2 are conserved only partially among tetraspanins. It seems that the EC2 domain corresponds to a structural conservation and variability scheme that is unique among proteins. The EC2 seems to be organized in two subdomains. The first subdomain, despite significant sequence divergence, appears to have a structurally conserved fold. A second subdomain, with an even higher heterogeneity, is extremely variable in size, secondary structure, and fold. The variable subdomain is inserted within the conserved subdomain and their relative topology is governed by the occurrence of key disulfide bridges, the number of which corresponds to distinct subtypes of tetraspanins.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Multiple Sequence Alignment-- All amino acid sequences were obtained from the SWISS-PROT/TrEMBL data base at the Expasy web site (www.expasy.ch). Homologous sequences were searched for using the BLAST server of Expasy. To gather tetraspanin and tetraspanin-like sequences from the data base, BLAST searches were performed using a number of sequences from well established members of the tetraspanin superfamily (i.e. CD81, CD82, CD9, CD53, CD63, UPK, RDS, and ROM). A multiple sequence alignment was initially achieved with the ClustalX software (27). The alignment was then improved manually using the Genedoc software.

Secondary Structure and Solvent Accessibility Prediction-- To predict the secondary structure of the EC2 domain, two methods (available on the World Wide Web) based on a consensus assignment were used. The first method, Jpred (jura.ebi.ac.uk:8888/), takes a multiple sequence alignment as input and performs a consensus average of nine different alignment-based secondary structure prediction methods. Alignment-based prediction methods have been demonstrated to have a significantly better accuracy than those using single sequences, and consensus averaging by Jpred has been shown to increase the accuracy to 72.9% (28). The second method, NPS (npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_seccons.html), performs a consensus average of secondary structure prediction methods based on single sequences. The use of alignment-based secondary structure prediction methods requires the sequences to have a degree of homology of at least ~25%. Because the initial alignment indicated that most EC2 sequences had a lower overall homology (except of course for sequences of identical proteins in different organisms), the following procedure was adopted. EC2 sequences for which homologues were available were submitted to the Jpred prediction. Those EC2 sequences that had no homologue at a 25% level were submitted to the NPS method (which actually was used to evaluate all sequences one-by-one as a routine control). Both methods unambiguously indicated the presence of several constant secondary structural elements in the overall EC2 family (see "Results"). It was also found that for each of these secondary structural elements higher homologies could be found between larger groups of sequences. Alignments corresponding to such groups were therefore reprocessed through the Jpred method to predict the secondary structure of the corresponding elements with a better accuracy. Solvent accessibility was predicted using the RCNPRED server (prion.biocomp.unibo.it/rcnpred.html) (29). Accessibility was also measured on the CD81 EC2 structure (Protein Data Bank entry 1G8Q, chain A) (26) using the Naccess 2.1 program (30) with a cutoff of 40%.

Homology Modeling-- Homology modeling of the CD53 and Q9V3R4 EC2 was performed with the Modeller 4 program (31). The CD81 EC2 structure (Protein Data Bank entry 1G8Q, chain A) (26) was used as a template. To obtain the best possible structure, four or six slightly different pairwise alignments of the target protein sequence to the CD81 EC2 were used, and five structures were calculated in each case. For Q9V3R4, the additional disulfide bridge was introduced as a constraint. To model the loop regions corresponding to nonaligned sequence stretches, the Whatif protein fragment data base (32) was searched first for homologous regions. Because this approach proved unsuccessful, loops were predicted using the "Loop" simulated annealing routine (33) of Modeller 4. After addition of hydrogen atoms using Whatif (34), the resulting 20 or 30 structures were submitted to energy minimization using the CHARMM module of Quanta 95. The geometric, steric, and compactness quality of the structures was then examined using Whatif, and the best structure was selected in each case. Graphical representations of molecular structures were done with the MOLMOL software (35).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Sequence Conservation and Variability of the Tetraspanin EC2-- The tetraspanin or tetraspanin-like sequences that were retrieved from the SWISS-PROT/TrEMBL data base are indicated in Table I. All sequences had in common 1) the occurrence of four hydrophobic spans of ~20-25 residues corresponding to the TM segments, two extracellular loops of unequal size separating TM1-TM2 and TM3-TM4, respectively, and a short intracellular domain between TM2 and TM3 and 2) a few very highly conserved residues in the second extracellular loop including the CCG motif. The list contains true tetraspanins that have in common the tetraspanin signature as well as distant members. Apart from the well identified tetraspanin proteins (i.e. the CDs) and well known distant members (RDS, ROM, and UPK), most sequences arise from genomic sequence cDNA translations.

Fig. 1 shows the multiple sequence alignment of all EC2 domains. Only sequences corresponding to distinct subcategories of a single tetraspanin type or to identical tetraspanins from distinct organisms can be aligned straightforwardly along their entire lengths. Otherwise, sequences have been aligned to match several completely or partially conserved residues as well as chemically similar residues. The only canonical residues are the CCG motif and two other cysteine residues. A motif formed of an aspartic acid or asparagine followed by an aromatic residue occurs in ~85% of the sequences. A few residues, mainly cysteines, serines, and prolines are conserved between large groups of sequences. The overall homology between different tetraspanin EC2s ranges from 5 to 40%, and the average pairwise homology is only 15%. EC2 sequences also vary greatly in length, a feature that introduces large gaps in the alignment. It is mainly the central region of the sequences (i.e. between the conserved CCG and the last conserved cysteine) that varies in length, and therefore the corresponding region of the alignment carries the larger gaps. This difference between the side regions and the central region of the alignment also manifests itself in the degree of homology. Although a certain extent of homology is apparent in the side regions between particular subgroups of sequences, there is virtually no homology between EC2s from distinct tetraspanin types in the central regions apart from the aforementioned residues (in the Fig. 1 sequences, stretches between conserved residues in this region are aligned for different tetraspanin types only to save space). The number of conserved cysteine residues in the EC2 sequences is variable, being either four, six, or eight, a feature that allows the classification of sequences into three groups. Various data indicate that the cysteines form disulfide bridges as indicated in Fig. 1. The existence of two disulfide bridges involving the four cysteines conserved in all sequences is documented from the known structure of the CD81 EC2 (26). Sequences displaying only these four cysteines and therefore two disulfides correspond to the simplest EC2 type (group 1). From Fig. 1 it is clear, however, that the majority of EC2 sequences (group 2) possesses an additional pair of cysteines. Disulfide formation by these two latter cysteines seems very likely because it has already been suggested for two members of this group, ROM and RDS, for which their mutation (as well as that of the four canonical cysteines) precludes a proper protein insertion (36). There are actually two subtypes within group 2 depending on the exact position of one cysteine relative to the other. The first additional cysteine of group 2 sequences is found in close proximity to the third canonical cysteine. For the less common group 2a sequences, the first additional cysteine is present in a conserved PCSC motif, whereas for the more common group 2b sequences it occurs in a conserved PXSCC motif (although there are several alignment possibilities for these motifs, the presence of conserved serine and proline residues in both groups 1 and 2 led us to the proposal in Fig. 1 in agreement with Ref. 26). A few tetraspanin EC2 sequences (group 3) contain two more cysteine residues that are relatively close to each other. Because no sequence exists that contains only one of these cysteines, it seems also likely that these form a fourth disulfide bridge, although currently no experimental data exist in this regard. In the alignment, gaps occur in all regions separating the conserved cysteine residues.

View larger version (84K): [in this window] [in a new window]   Fig. 1.   Multiple sequence alignment of the EC2 domains of tetraspanin and tetraspanin-like proteins. The background colors of sequence names correspond to the different tetraspanin groups (see "Results" for details): green, group 1; light blue, group 2a; darker blue, group 2b; red, group 3. The lines connecting conserved cysteine residues correspond to disulfide bridges. The positions of the experimental helical regions of the CD81 EC2 structure are indicated as magenta tubes at the top of the figure.

The experimentally determined secondary structure of the CD81 EC2 (26) is also indicated in Fig. 1. Interestingly, three of the five helices (A, B, and E) correspond to the three nongapped regions at the sides of the alignment, a feature that suggests the presence of regions of at least conserved size. These three helices also correspond to the most conserved regions among sequences, specially for helix B. Although the extent of homology per se (i.e. the percentage of identical residues in pairwise alignments) is not significantly higher, there seems to be more conserved residue types. Such conserved residues types in helices A, B, and E were already noted in Ref. 26, and our data confirm these observations for a larger number of sequences. It appears that 14 specific positions located in all three helices are almost invariably occupied by hydrophobic residues. On the other hand, there are only two conserved polar positions, both located in helix B, that are occupied by an aspartic acid and a glutamine in most sequences, although not in CD81.

Secondary Structure and Solvent Accessibility of the Tetraspanin EC2-- The results of secondary structure prediction of the EC2 domains of tetraspanins and tetraspanin-like proteins are shown in Fig. 2A. For the sake of clarity, in the case of sequences from different organisms and sequences that present more than 85% homology in EC2, only one member has been reproduced. For homologous sequences or sequence regions, the secondary structure shown in Fig. 2A was predicted using a multiple sequence alignment as input for the Jpred consensus method and is therefore valid for all sequences. Such sequences were also examined one-by-one by the single sequence-oriented NPS method as a control. Consistent results were obtained in the great majority of cases. The NPS method was the sole usable approach for secondary structure determination of those EC2 sequences or sequence regions of Fig. 2A that had no homologues in Fig. 1. The recent determination of the three-dimensional structure of the EC2 domain of the tetraspanin CD81 (26) provides us with an opportunity to check the reliability of the prediction methods. The experimental secondary structure found for this domain is also indicated in Fig. 2A. It can been seen that the secondary structure prediction for the CD81 EC2, which corresponds to the Jpred result, matches quite accurately the five actual helices of a 1-2 residue error level. A similar accuracy is found when the CD82 EC2 sequence is analyzed with the NPS method. This suggests that both methods yield satisfactory results with regard to the secondary structure of tetraspanin EC2s.

View larger version (97K): [in this window] [in a new window]   Fig. 2.   A, secondary structure prediction of tetraspanin EC2. Sequences positions highlighted in magenta and yellow correspond, respectively, to helices and strands. Sequences positions highlighted in blue are potential glycosylation sites. Squared positions correspond to conserved residues. The background colors of sequence names correspond to the different tetraspanin groups (see the legend to Fig. 1). The positions of the experimental helical regions of the CD81 EC2 structure are indicated as magenta tubes at the top of the figure. The lines connecting conserved cysteine residues correspond to disulfide bridges. Sequences having homologues in Fig. 1 were analyzed as multiple sequence alignments using the Jpred method. Other sequences were singly analyzed using the NPS method. Several sequences that after a first prediction run were found to have more than 25% homology in one of the three conserved helical regions were reprocessed together as a multiple sequence alignment using Jpred to refine the prediction of that particular region. B, solvent accessibility of the conserved regions of tetraspanin EC2. Sequence positions highlighted in blue correspond to buried residues. The positions of the experimental helical regions of the CD81 EC2 structure are indicated as magenta tubes at the top of the figure. Buried residues of the CD81 EC2 structure are also indicated as sequence positions highlighted in green.

It is obvious from Fig. 2A that a very different situation occurs among the aligned EC2 sequences with regard to the predicted secondary structure for the N- and C-terminal parts on one hand and for the central part on the other hand. On the sides of each sequence, the prediction invariably indicates that the secondary structure elements are the same three helices, two on the N-terminal side and one on the C-terminal side. These appear to be conserved in all EC2 domains and also correspond to the first two (A and B) and the last experimental helices (E) of the CD81 EC2 domain. It is noteworthy that a similar secondary structure is predicted despite the relatively low sequence homology in particular for the first helix.

Oppositely, for the central region of the EC2 alignment, virtually no conservation of secondary structure occurs among the sequences. As already pointed out, the central region of all EC2 sequences appears to be formed of a limited number of sequence stretches separated by cysteine residues. As shown in Fig. 2A, not only are these sequence stretches very different in length and without homology, but they bear no conserved secondary structure. The secondary structure type present in these sequence stretches (within the current ~73% accuracy of prediction) is mostly coil (indicating that loops often occur between the disulfide-forming cysteines) but also occasionally strands or helices. The number of such disulfide-separated variable secondary structure sequence stretches is respectively two, three, and five for EC2 sequences from groups 1, 2, and 3. For the EC2 of CD81 (group 1), the two sequence stretches forming the variable central region appear to correspond to two helices (termed C and D in Ref. 26), but this specific motif does not appear to be conserved among tetraspanins. Even for tetraspanins that have a similarly sized EC2 central region and the same number of conserved cysteines than CD81, the predicted secondary structural elements located between these cysteines are not helices.

The results of residue solvent accessibility prediction of the EC2 domain of tetraspanins and tetraspanin-like proteins is shown in Fig. 2B for the conserved right and left sides of the sequences. Again the structure of the CD81 EC2 (26) serves as a validation test, because the external accessibility, measured for the monomer, is also shown. The prediction seems to be relatively accurate. Fig. 2B also indicates that there is an obvious periodicity of the predicted accessibility, which is conserved on the alignment. Moreover, the periodicity is roughly 3-4 residues, a feature that is typical of surface helices. This confirms the structurally conserved helical character of the N- and C-terminal parts of the EC2.

These data indicate that the EC2 domain of tetraspanins possesses two very different subdomains. A first subdomain encompassing the N- and C-terminal sides of the sequence has a similar size and a conserved secondary structure among all tetraspanins despite a limited but significant sequence conservation. This secondary structure appears to be composed of three helices. A second central subdomain is extremely variable in size and secondary structure and bears almost no sequence conservation apart from disulfide-forming cysteines. The occurrence of such conserved disulfides suggests, however, that even the variable domain corresponds to an at least partially conserved topology.

Topology of the Tetraspanin EC2-- From our data and the known structure of the CD81 EC2, it is possible to build general models for the topological organization of the EC2 polypeptide chain. The organization of the three conserved helices in all tetraspanins appears similar to that found in the CD81 EC2 structure. This is first suggested by the fact that these three helices have similar lengths. Moreover, interactions occurring between these helices are significantly conserved. Indeed, for CD81, interactions that occur among the A, B, and E helices occur at hydrophobic and charged side chains located at specific sequence positions (as already pointed out in Ref. 26 for the A-E pair). Careful examination of the alignment of Fig. 1 indicates that although residues at these positions are variable, the potential for interaction is generally conserved (i.e. by having appropriately paired hydrophobic, hydrogen-bonding, or charged residues). It is therefore likely that the tetraspanin EC2 general topology is determined by a common core, formed by the three conserved helices and a variable domain, the global topology of which with respect to the common core is stabilized by two, three, or even four disulfide bonds as shown in Fig. 3. The minimal topology for EC2 is that of group 1 tetraspanins in which two disulfide bridges determine the topology of a variable domain composed of two peptide stretches of variable secondary structure relative to the conserved domain made of three helices. In group 2 tetraspanins, the second variable peptide stretch is replaced by a pattern formed by two variable peptide stretches separated by a cysteine that forms a third disulfide bridge that can either crossover, or not, one of the canonical disulfide depending on the subtype (2a or 2b). In the group 3 EC2, it is the first variable peptide stretch that is replaced by a disulfide-separated two-stretch pattern. Therefore, the chain topology of more complex tetraspanin is built from a basic topology by the duplication of specific peptide stretches and disulfide insertion.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Chain topography of several tetraspanins EC2 belonging to group 1 (top), groups 2a and 2b (middle), and group 3 (bottom). Rectangles, arrows, and thinner lines correspond to helices, strands, and coil, respectively. Conserved and variable subdomains are indicated in blue and red, respectively. Cysteines and disulfide bridges are indicated in yellow. Other important residues are in green.

Three-dimensional Structure of the Tetraspanin EC2-- It thus seems that tetraspanins share a conserved structural core made of three helices (helices A, B, and E of Ref. 26), upon which is built a variable subdomain with an overall structure determined by a conserved disulfide linking and a globally conserved chain topology albeit an unconserved secondary structure. It appears also likely that not only the chain topology but also the protein topology in a more general sense (i.e. the relative position of secondary structural elements in space) are conserved. This is suggested strongly by the observation that the structural motifs that determine the overall orientation of the unconserved structural elements of the variable subdomain with respect to each other and to the common core in CD81 EC2 are also conserved among tetraspanins. These motifs, which impose specific orientation constraints to the main chain, include not only the two disulfide bridges (the relative orientation of which is constrained by the consecutive character of two of the involved cysteines that have a unique backbone conformation) but also two specific residues, a glycine and a proline, which in each case are neighbor to a disulfide-linked cysteine and promote specific bends of the main chain as emphasized in Ref. 26 for CD81. The invariant character of these residues indicates that the local backbone geometries are conserved among tetraspanins. Furthermore, the CD81 EC2 structure contains two other bends strategically located at the start and at the end of the C helix, respectively. It is noteworthy that in the alignment of Fig. 1, the corresponding sequence loci often contain glycine and asparagine or proline residues, respectively, that are well known to occur often at turns or bents. All these features suggest that the overall orientation of the variable structural elements relative to the conserved subdomain is conserved among tetraspanins. On this basis we have attempted to perform homology modeling of two relatively simple tetraspanin EC2s, namely CD53 and Q9V3R4, using the framework of the CD81 EC2 structure. For CD53, which is also a group 1 tetraspanin, the C and D helices of CD81 are replaced by two loops (Fig. 4A, center). In addition, the group 2 tetraspanin Q9V3R4 bears an additional disulfide bridge. This latter protein currently bears no particular functional interest because its role is unknown, its sequence being a cDNA translation from the Drosophila melanogaster genome. However, this sequence happens to have a striking homology with CD81 in critical parts of the variable subdomain of EC2 (see Fig. 1), although both sequences correspond to different tetraspanin EC2 groups. In particular, six residues neighboring the two extra cysteine residues of Q9V3R4 are also found in CD81 at similar sequence positions. Furthermore, the two CD81 residues that are replaced by the disulfide-forming extra cysteines in Q9V3R4, according to the corresponding pairwise alignment (see Fig. 1), have spatially close and facing side chains in the x-ray structure (26). This suggests that these can be used as templates for homology modeling of the extra disulfide bridge. The resulting modeled structure is shown on Fig. 4A (right). It is likely that all group 2 tetraspanins have a similar disulfide topology. It should be stressed that the modeling of peptide loops of more than 5-6 residues is currently only approximate (33), so the modeled structures of Fig. 4A may certainly be subject to future improvement. Homology modeling of more complex tetraspanins requires further prediction or experimental work to be performed because it depends on the tertiary interactions of the extra secondary structural elements.

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Homology modeling of the tetraspanin EC2. A, ribbon representation of the experimental structure (26) of the CD81 EC2 (left) and the predicted structures of the CD53 (center) and Q9V3R4 (right) EC2. Conserved and variable subdomains are colored in blue and red, respectively. Cysteine side chains and disulfide bridges are in yellow. B, molecular surface representation of the same structures. The top view corresponds to that of A, and the bottom view has each molecule rotated vertically by 180°. Conserved and variable subdomains are colored in blue and red, respectively. Grayish blue and light-blue colored regions correspond, respectively, to conserved hydrophobic side chains and to asparagine side chains involved in glycosylation sites.

Fig. 4A emphasizes that although each EC2 domain has a distinct secondary structure of the variable peptide subdomain, the overall orientation of each variable peptide stretch remains relatively similar. This has consequences on the shape of the domain as shown in molecular surface representations of the experimental structure of the CD81 EC2 and the modeled structure of the CD53 and Q9V3R4 EC2 (Fig. 4B). Despite the secondary structure differences, the three domains have a similar shape in which the variable subdomain has a specific location. Interestingly, for CD53 the variable region is located relatively close to two glycosylation sites. One of these sites is conserved among a large number of tetraspanins (CD63, CD82, CD151, UPKB, TSP3, TSP6, net6, Q9U3V4, and O96961) and is very exposed at the extremity of the first helix (note that this locus in the structure also corresponds in RDS and ROM to a cysteine residue responsible for the functionally important covalent dimer formation in rod outer segment disc membranes) (36). The other potential glycosylation site also represented for CD53 in Fig. 4B is located downstream of the conserved CCG motif and is also relatively conserved (CD53, CD63, CD82, CD151, TSP3, A15, RDS, NET1, and Q9QZA6). A third potential glycosylation site not present in CD53 is often present 6-15 residues upstream from helix E (see Fig. 1). In all three structures of Fig. 4B, the variable domain is located on one side of the EC2 domain, which is opposite to a hydrophobic patch conserved in many sequences and was suggested in Ref. 26 to be involved in interactions between tetraspanins. On the other hand, another hydrophobic patch found in the structure of the CD81 EC2 (26) and located in the variable subdomain is not conserved in CD53 or Q9V3R4, in which the residues of this region are relatively hydrophilic. Therefore, this latter hydrophobic patch is specific to CD81. In the corresponding region, both CD53 and Q9V3R4 display an aspartic acid residue (not shown) conserved in almost all tetraspanins that may be involved in specific recognition processes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Our data indicate that all EC2 domains of the tetraspanin superfamily are formed of two subdomains, one conserved and one variable. The conserved subdomain is made of a three-helix bundle corresponding, respectively, to the two first helices (A and B) and the last helix (E) of the CD81 crystallographic data (26). The variable subdomain corresponds to a sequence insertion within the conserved three-helix bundle subdomain. It differs from one tetraspanin to the other in length and secondary structure and contains only a few conserved residues. The relative topology of the conserved and variable domains is likely to be conserved and corresponds to a bipolarity of the EC2 domain. These data imply that the protein fold found by Kitadokoro et al. (26) for the CD81 EC2 is not conserved among tetraspanins but that only a "subfold," corresponding to the three-helix bundle, is conserved.

It is tempting to suggest that the two subdomains correspond to distinct interaction potentialities of tetraspanins. The conserved subdomain might contain sites that correspond to interactions that are common to all tetraspanins, e.g. interactions with themselves and other tetraspanins, as already suggested for CD81 (26) as well as with integrins (37). Oppositely, the variable subdomain might promote interactions with proteins that are specific to each tetraspanin, similar to specific interactions with viruses or toxins. Indeed, Higginbottom et al. (23) have mapped the binding site of hepatitis C virus on CD81 at the end of this variable domain. This does not preclude the possibility that a binding site for a particular protein might involve both subdomains. In this regard, recent data (25) indicate that the interaction between CD151 and ₃₁ integrins involve an epitope of the CD151 EC2 (residues 186-217), which overlaps both the end of the variable region (i.e. immediately downstream the third cysteine of group 2 tetraspanins) and the terminal helix of the conserved subdomain. Because among tetraspanins CD151 is considered to display the most robust interaction with integrins, this could indicate that the interaction of tetraspanins with integrins involves mainly the conserved hydrophobic surface and that in the case of CD151 part of the variable subdomain participates in the stabilization of the interaction. In this framework, the interactions between distinct tetraspanins would be able to bring the interacting heterologous proteins in close proximity.

Crystallographic or NMR structural determination of soluble forms of various tetraspanin EC2 domains, which will be available in the near future, would certainly constitute the best method to confirm our conclusions of the EC2 domain structure. However, circular dichroism quantification of the secondary structure of several soluble EC2s would constitute a simpler structural approach. Another functionally relevant approach would be the construction of chimeras in which the variable subdomain of EC2 is swapped between distinct tetraspanin types. One might expect that in such chimeras, the functional aspects that involve interactions common to all tetraspanins will be retained. By contrast, those that are related to type-specific interactions would be swapped accordingly.

In a recent tetraspanin conference,² the need for a classification of tetraspanins was expressed. We suggest that the distinction of the structural groups made here may form one of the bases of this classification.

The occurrence of distinct domains within a protein or a protein family is a rather common feature. In particular, "insertion domains" have been evidenced during the last few years (38). The most popular of these are the so-called protein modules (39) that confer specific interaction potentialities to a number of proteins (as may the EC2 to tetraspanins). These correspond to domains with relatively variable sequences but with a conserved protein fold. The occurrence of variable sequence segments within a protein family has also been widely documented for immunoglobulins (the "hypervariable loops" (40) or olfactory receptors (41)). Again, these correspond to variable sequence stretches within a conserved structural fold. To our knowledge, the tetraspanin EC2 corresponds to a new domain insertion and variability scheme, namely the occurrence of a domain with variable structure within an otherwise conserved fold for a protein family.

	ACKNOWLEDGEMENTS

We are indebted to Dr. Miglena Angelova and Elizabeth Cogan for critically reading the manuscript.

	FOOTNOTES

^* This work was supported by grants from the Center National de la Recherche Scientifique, The Institut National de la Santé et de la Recherche Médicale, and the Association pour la Recherche sur le Cancer.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ Present address: Dept. of Pathology, Harvard Medical School, D2-143, 200 Longwood Ave., Boston, MA 02115.

To whom correspondence should be addressed. Tel.: 33-1-40-51-6487; Fax: 33-1-40-51-6454; E-mail: conjeaud@cochin.inserm.fr.

Published, JBC Papers in Press, August 1, 2001, DOI 10.1074/jbc.M105557200

² First International Meeting on Tetraspanin Proteins, Federation of American Societies for Experimental Biology Conference, Aspen, July 1-6, 2000.

	ABBREVIATIONS

The abbreviations used are: TM, transmembrane segment; EC2, second extracellular domain.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. Krishnamoorthy, J. Heimburg-Molinaro, A. M. Bargo, R. J. Nash, and R. J. Nash Heparin Binding Epidermal Growth Factor-Like Growth Factor and PD169316 Prevent Apoptosis in Mouse Embryonic Stem Cells J. Biochem., February 1, 2009; 145(2): 177 - 184. [Abstract] [Full Text] [PDF]

				N. Lineberry, L. Su, L. Soares, and C. G. Fathman The Single Subunit Transmembrane E3 Ligase Gene Related to Anergy in Lymphocytes (GRAIL) Captures and Then Ubiquitinates Transmembrane Proteins across the Cell Membrane J. Biol. Chem., October 17, 2008; 283(42): 28497 - 28505. [Abstract] [Full Text] [PDF]

				K. Lambou, F. Malagnac, C. Barbisan, D. Tharreau, M.-H. Lebrun, and P. Silar The Crucial Role of the Pls1 Tetraspanin during Ascospore Germination in Podospora anserina Provides an Example of the Convergent Evolution of Morphogenetic Processes in Fungal Plant Pathogens and Saprobes Eukaryot. Cell, October 1, 2008; 7(10): 1809 - 1818. [Abstract] [Full Text] [PDF]

				A. F. X. Goldberg, L. M. Ritter, N. Khattree, N. S. Peachey, R. N. Fariss, L. Dang, M. Yu, and A. R. Bottrell An Intramembrane Glutamic Acid Governs Peripherin/rds Function for Photoreceptor Disk Morphogenesis Invest. Ophthalmol. Vis. Sci., July 1, 2007; 48(7): 2975 - 2986. [Abstract] [Full Text] [PDF]

				L. Tu, X.-P. Kong, T.-T. Sun, and G. Kreibich Integrity of all four transmembrane domains of the tetraspanin uroplakin Ib is required for its exit from the ER J. Cell Sci., December 15, 2006; 119(24): 5077 - 5086. [Abstract] [Full Text] [PDF]

				S.-H. Ho, F. Martin, A. Higginbottom, L. J. Partridge, V. Parthasarathy, G. W. Moseley, P. Lopez, C. Cheng-Mayer, and P. N. Monk Recombinant extracellular domains of tetraspanin proteins are potent inhibitors of the infection of macrophages by human immunodeficiency virus type 1. J. Virol., July 1, 2006; 80(13): 6487 - 6496. [Abstract] [Full Text] [PDF]

				G. Min, H. Wang, T.-T. Sun, and X.-P. Kong Structural basis for tetraspanin functions as revealed by the cryo-EM structure of uroplakin complexes at 6-A resolution J. Cell Biol., June 19, 2006; 173(6): 975 - 983. [Abstract] [Full Text] [PDF]

				L. Cocquerel, C. Voisset, and J. Dubuisson Hepatitis C virus entry: potential receptors and their biological functions. J. Gen. Virol., May 1, 2006; 87(Pt 5): 1075 - 1084. [Abstract] [Full Text] [PDF]

				F. Martin, D. M. Roth, D. A. Jans, C. W. Pouton, L. J. Partridge, P. N. Monk, and G. W. Moseley Tetraspanins in Viral Infections: a Fundamental Role in Viral Biology? J. Virol., September 1, 2005; 79(17): 10839 - 10851. [Full Text] [PDF]

				S. Levy and T. Shoham Protein-Protein Interactions in the Tetraspanin Web Physiology, August 1, 2005; 20(4): 218 - 224. [Abstract] [Full Text] [PDF]

				A. Delaguillaumie, J. Harriague, S. Kohanna, G. Bismuth, E. Rubinstein, M. Seigneuret, and H. Conjeaud Tetraspanin CD82 controls the association of cholesterol-dependent microdomains with the actin cytoskeleton in T lymphocytes: relevance to co-stimulation J. Cell Sci., October 15, 2004; 117(22): 5269 - 5282. [Abstract] [Full Text] [PDF]

				K. Kaji and A. Kudo The mechanism of sperm-oocyte fusion in mammals Reproduction, April 1, 2004; 127(4): 423 - 429. [Abstract] [Full Text] [PDF]

				A. Cherukuri, T. Shoham, H. W. Sohn, S. Levy, S. Brooks, R. Carter, and S. K. Pierce The Tetraspanin CD81 Is Necessary for Partitioning of Coligated CD19/CD21-B Cell Antigen Receptor Complexes into Signaling-Active Lipid Rafts J. Immunol., January 1, 2004; 172(1): 370 - 380. [Abstract] [Full Text] [PDF]

				T. Shoham, R. Rajapaksa, C. Boucheix, E. Rubinstein, J. C. Poe, T. F. Tedder, and S. Levy The Tetraspanin CD81 Regulates the Expression of CD19 During B Cell Development in a Postendoplasmic Reticulum Compartment J. Immunol., October 15, 2003; 171(8): 4062 - 4072. [Abstract] [Full Text] [PDF]

				C. J.R. Loewen, O. L. Moritz, B. M. Tam, D. S. Papermaster, and R. S. Molday The Role of Subunit Assembly in Peripherin-2 Targeting to Rod Photoreceptor Disk Membranes and Retinitis Pigmentosa Mol. Biol. Cell, August 1, 2003; 14(8): 3400 - 3413. [Abstract] [Full Text] [PDF]

				D. Smyth, D. P. McManus, M. J. Smout, T. Laha, W. Zhang, and A. Loukas Isolation of cDNAs Encoding Secreted and Transmembrane Proteins from Schistosoma mansoni by a Signal Sequence Trap Method Infect. Immun., May 1, 2003; 71(5): 2548 - 2554. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 276/43/40055    most recent M105557200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Seigneuret, M. Articles by Conjeaud, H. Search for Related Content PubMed PubMed Citation Articles by Seigneuret, M. Articles by Conjeaud, H. Social Bookmarking                      What's this?