Skip Navigation


Glycobiology Advance Access originally published online on July 11, 2006
Glycobiology 2006 16(10):959-968; doi:10.1093/glycob/cwl021
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
16/10/959    most recent
cwl021v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Patra, M.
Right arrow Articles by Mandal, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Patra, M.
Right arrow Articles by Mandal, C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org

Search for glucose/galactose-binding proteins in newly discovered protein sequences using molecular modeling techniques and structural analysis

Madhumita Patra and Chhabinath Mandal1

Indian Institute of Chemical Biology, 4, Raja S. C. Mullick Road, Jadavpur, Kolkata 700032, West Bengal, India

1 To whom correspondence should be addressed; e-mail: cnmandal{at}iicb.res.in

Received on March 8, 2006; revised on June 21, 2006; accepted on July 9, 2006


   Abstract
Top
 Abstract
Introduction
 Results and discussion
 Conclusions
 Materials and methods
 Supplementary data
 Conflict of interest statement
 Acknowledgments
 References
 
Sugar moieties serve as specificity markers in a wide variety of biochemical functions, and periplasmic glucose/galactose-binding proteins (GGBPs) serve as the primary receptors for transport and chemotaxis. Recently, complete genome sequencing projects have revealed many open reading frames for such receptors. On the basis of the homology search with the known x-ray structures (PDB ID: 3GBP/1GCA) of a periplasmic receptor protein from Salmonella typhimurium, we selected four putative proteins with amino acid identities between 30 and 48% for the prediction of three-dimensional (3D) structures of the proteins as well as their complexes with glucose and galactose. We could successfully identify the key residues involved in coordination with calcium ion spanning over two loop structures. We calculated the ligand-binding affinities and hydrogen bonding patterns of the modeled structures and compared with those of the x-ray structures. The calculation of free energies of binding of the modeled structures to glucose and galactose in the presence of water suggested that two of four putative proteins can form complexes with dissociation constants in the micromolar range (1–10 µM). Electrostatic potentials on the surfaces near the sugar and calcium-binding sites of the modeled structures were predominately negative as found in case of the x-ray structure. Taken together, our results suggest that the products of two newly discovered genes would serve as receptors for the transport of glucose and galactose.

Key words: galactose / glucose / modeling / receptor / structure


   Introduction
Top
 Abstract
 Introduction
Results and discussion
 Conclusions
 Materials and methods
 Supplementary data
 Conflict of interest statement
 Acknowledgments
 References
 
Sugar moieties associated with glycoproteins and polysaccharides are involved in the determination of virulence and pathogenicity of many infectious agents (Casadevall and Liise-anne, 1999Go; Weissman et al., 2006Go). In mammalian cells, some kinases are sensitive to different oligosaccharides leading to dramatic perturbation of signaling pathways (Clodfelder-Miller et al., 2005Go; Nagaoka et al., 2005Go). The periplasmic GGBPs of the gram-negative bacteria serve as the primary receptors for transport and in some cases for chemotaxis in response to compounds such as sugars, ions, and small peptides. The binding of a ligand to a free periplasmic receptor activates that protein, allowing it to be recognized by distinct membrane components specific to transport and in some cases to chemotaxis (Mowbray et al., 1990Go). X-ray structures of a periplasmic glucose/galactose-binding receptor protein from Salmonella typhimurium bound to glucose, galactose, or unliganded form have been solved (PDB ID: 3GBP/1GCA/1GCG) at 2.4 Å/1.7 Å/1.9 Å resolutions (Mowbray et al., 1990Go; Zou et al., 1993Go; Flocco and Mowbray, 1994Go). The protein consists of two highly similar {alpha}/ß-domains; each domain is made up of a core of parallel ß-sheets sandwiched between two layers of {alpha}-helices. The domains are linked by a hinge made up of three polypeptide strands. The ligand is almost completely buried in a binding site located in the cleft between the two domains.

In our present study, we have searched for the newly discovered protein sequences that have significant sequence homology with the bacterial protein, but their three-dimensional (3D) structural information is not yet available. Using BLAST search, taking the periplasmic receptor protein sequence as query in the "nr" database, we have identified four homologous sequences in the complete genome sequences of the bacterial species, that is, Clostridium tetani (Brüggemann et al., 2003Go), Pseudomonas syringae (Feil et al., 2005Go), Actinobacillus pleuropneumoniae (Accession number: ZP_00134897; Gillaspy, A.F. et al., unpublished data), and Treponema pallidum (Fraser et al., 1998Go), to examine their mode of binding with the ligand. The homologous protein genes were predicted by the conceptual translation of the identified open reading frames.

In these four sequences, the residues that are playing important roles in glucose/galactose binding as well as in calcium binding in 3GBP/1GCA are found to be highly conserved. Only sequences with identities in the range from 30 to 50% were selected because higher identities would always lead to very similar structure. 3D structures of these proteins have been predicted by homology modeling, and the glucose or galactose molecule has been docked to arrive at the structures of the complexes using computer-aided modeling techniques. The analysis of these structures allowed us to examine the feasibility of the formation of calcium-binding loops (CBLs) and the nature of interaction of glucose/galactose with these proteins. Another striking observation was the highly negatively charged environment of the galactose/glucose-binding pocket (Mowbray et al., 1990Go; Flocco and Mowbray, 1994Go). To verify whether this is a necessary condition for glucose/galactose binding, we calculated the electrostatic potentials of all the modeled proteins and examined their nature in the vicinity of the binding pocket.


   Results and discussion
Top
 Abstract
 Introduction
 Results and discussion
Conclusions
 Materials and methods
 Supplementary data
 Conflict of interest statement
 Acknowledgments
 References
 
In this study, we are in search of bacterial GGBPs in the newly sequenced genomes. The tertiary structures of a bacterial glucose/galactose-binding protein (GGBP) from the pathogenic bacteria S. typhimurium have been determined in complex with glucose at 2.4 Å (Mowbray et al., 1990Go) and galactose at 1.7 Å (Zou et al., 1993Go) resolutions as well as in the unliganded form at 1.9 Å resolution (Flocco and Mowbray, 1994Go). Earlier the structure of L-arabinose-binding protein (ABP) from Escherichia coli was reported at 1.7 Å resolution (Quiocho and Vyas, 1984Go) having a novel binding site geometry that accommodates both its {alpha}- and ß-anomers. It has a sequence very different from GGBP of S. typhimurium but shows very high similarity with its 3D structure where 70% of the structures superpose with a root mean square deviation (RMSD) of 2.0 Å (Vyas et al., 1991Go).

To find out the protein sequences homologous to GGBP from S. typhimurium, BLAST search was performed with its amino acid sequence as the query. Among the BLAST hits, we selected four sequences on the basis of their high homology in a particular region of the sequence with the query sequence, which formed the CBL in the structure of the query sequence but <50% amino acid identity in the entire sequence. All these selected protein sequences belong to bacteria. The conserved hypothetical protein from A. pleuropneumoniae showed 48% sequence identity and ~72% sequence similarity with the query sequence. The hypothetical protein from the bacteria C. tetani showed 47% identity and ~64% similarity with the query sequence, and the hypothetical protein from T. pallidum showed 36% identity and 56% similarity with GGBP sequence from S. typhimurium. The uncharacterized protein from P. syringae showed only 30% identity and ~48% similarity. Amino acid sequences of these four proteins were used as target sequences for 3D structure prediction and analysis of the calcium as well as glucose/galactose-binding sites. After the initial structures were predicted by the combination of ANALYN and MODELYN (Mandal, 1998Go), they were refined by the regularization of segments with deletions or insertions using DISCOVER module of InsightII of Accelrys (San Diego, CA).

Predicted structures were checked for main chain conformations using PROCHECK (Laskowski et al., 1993Go) which showed that initially >90% of the {phi} –{psi} plots were in the core regions and <3% together in the generously allowed and disallowed regions of Ramachandran’s plot. Most of the residues that fell in the disallowed regions were from the loop regions in which major insertions and deletions were made. On further refinement of the main chain conformations selectively in these regions, {phi} –{psi} plots of all the nonglycine residues of all the modeled structures were outside the disallowed regions. Side chain planarity of the planar groups in phenylalanine, tyrosine, tryptophan, histidine, arginine, glutamine, asparagines, glutamic acid, and aspartic acid was checked using PROCHECK where deviations from planarity were identified by measuring RMS distances of planar atoms from the best-fitted plane; residues having RMS distances >0.03 Å for rings and 0.02 Å for other groups were marked as outliers (Laskowski et al., 1993Go). Modeled proteins from A. pleuropneumoniae, C. tetani, T. pallidum, and P. syringae showed 1.8, 8.0, 5.7, and 2.0% outliers, respectively, of the total planar group-containing residues compared with 2.8% in the x-ray structure (3GBP) of S. typhimurium. RMSDs of bond lengths and bond angles of all the modeled structures were within 0.02 Å and 3.5°, respectively, from the standard values indicating reasonably good structural parameters.

We checked protein geometry of the modeled structures by calculating clashscores and rotamer outliers using MOLPROBITY (Davis et al., 2004). For modeled structures of proteins from A. pleuropneumoniae, C. tetani, T. pallidum, and P. syringae, all atom clashscores were 1.68, 1.2, 0.4, and 2.1, respectively, compared with a clashscore of 1.7 for the x-ray structure (3GBP) of S. typhimurium; rotamer outliers for the respective structures were 11, 13, 13, and 19% as compared with 7% for the x-ray structure as calculated using MOLPROBITY. The multiple sequence alignment of the query and the modeled proteins is shown in Figure 1 that shows the highly conserved amino acid residues; the residues involved in coordination with calcium ion are shown as boldfaced letters. All the modeled structures were structurally superposed with respect to the calcium atoms of the common core of the structures and shown in Figure 2.


Figure 1
View larger version (33K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Multiple alignments of the partial sequences showing the calcium- and sugar-binding regions of the modeled proteins and the starting scaffold (3GBP). The amino acid residues involved in coordination with the calcium ion are boldfaced. The conserved set of amino acid in all the sequences are identified by an "*" and semi-conserved amino acids marked with "colon (:)" or " dot (.)."

 

Figure 2
View larger version (67K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. Ribbon representation of the superposed structures of the modeled proteins with the crystal structure of 3GBP. The ligand molecule is shown in space filling representation and is colored by atoms. Ribbon colors are as follows: cyan for GGBP from Salmonella typhimurium; light pink for model of protein from Actinobacillus pleuropneumoniae; light green for model of protein from Clostridium tetani; light blue for model of protein from Treponema pallidum; and light yellow for model of protein from Pseudomonas syringae. Color figure is given as supplementary material online.

 

Calcium atoms were added to our modeled structures by superposing them with the x-ray structure with respect to a set of calcium atoms, which showed RMSD <0.3 Å, followed by the transfer of the calcium ion. Potential coordinating atoms in a modeled structure were identified measuring distances of the residues within 4 Å around the calcium atom, and the equivalent amino acids were identified. Energy minimization and molecular dynamics were performed to get a regularized structure of the calcium-binding environment similar to that of the x-ray structure; distance constraints were applied to obtain a coordination shell of similar geometry. In periplasmic GGBP of S. typhimurium, the calcium ion is surrounded by seven ligands (Vyas et al., 1987Go, 1991Go) in two CBLs (CBL-1 and CBL-2). The amino acid sequence from 134 to 142 (DLNKDGKIQ) forms the major CBL (CBL-1), and this sequence is found to be highly conserved in the three target sequences selected for modeling. The loop CBL-2 only participates in calcium binding via Glu-205 side chain, whereas the side chains of Asp-134, Asn-136, Asp-138, and Gln-142 and the main chain of Lys-140 of loop CBL-1 are involved in calcium binding (Figure 3). We have identified all the conserved equivalent residues and calcium-coordinated atoms in the modeled structures as discussed below and presented in Table I and Figure 4. In the modeled proteins from A. pleuropneumoniae and C. tetani, the CBLs also consist of the residues from 134 to 142, the residues involved in calcium binding are Lys-140 main chain oxygen and the side chains of Asp-134, Asn-136, Asp-138, Gln-142, and Glu-205 (Table I). All the residues involved in coordination are identical, and only two amino acids in the entire loop are different from those of GGBP. The situation is very similar in case of modeled protein from T. pallidum although it has less overall sequence identity with 3GBP. In this protein also, the CBLs consist of the residues from 134 to 142, and the residues participating in calcium coordination are Ile-140 main chain oxygen and side chains of the residues Asp-134, Asn-136, Asp-138, Gln-142, and Glu-205 differing only in one amino acid, the coordinating backbone oxygen of Lys-140 is replaced with that of Ile-140 (Table I). Although the modeled protein from P. syringae has only two residues, Asp-134 and Asp-138, identical with those involved in calcium coordination in GGBP of S. typhimurium, the calcium coordination shell is partially satisfied by other residues. The main chain oxygen of Gly-140 and the side chains of Asp-134, Asp-138, Asp-141, and Asn-205 are involved in calcium coordination in this case, and two other coordination positions are occupied by water molecules to satisfy all the seven positions (Table I). Similar seven-ligand calcium coordination was also observed in case of a fucose-binding lectin (Mitchell et al., 2002Go).


Figure 3
View larger version (74K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. Mode of calcium binding in 3GBP. The calcium ion is shown in pink. The residues of protein involved in binding are colored by atom and shown in sticks. The distances of calcium ion from their coordinating atoms are also shown. The calcium ion has a classical seven-ligand coordination. Color figure is given as supplementary material online.

 

View this table:
[in this window]
[in a new window]

  		Table I. Residues involved in coordination with calcium ion and also hydrogen bonding in the complexes of proteins with glucose and galactose

 

Figure 4
View larger version (83K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. Mode of interaction of calcium ion with modeled proteins. The residues of protein involved in coordination with the calcium ion are colored by atom and shown in stick representation. Six residues are involved in calcium binding in case of modeled protein from Actinobacillus pleuropneumoniae (A), Clostridium tetani (B), and Treponema pallidum (C) similar to that of 3GBP, whereas five residues and two water molecules are involved in case of modeled protein from Pseudomonas syringae (D). The calcium ion is shown as pink sphere. Color figure is given as supplementary material online.

 

Glucose has been docked into the binding site of the modeled structures by superposing them with the x-ray structure, bound to glucose (3GBP), with respect to the structurally conserved regions followed by the transfer of the glucose molecule to the binding site. The optimization of the structures of the complexes was performed by repeated molecular dynamics and energy minimization in the presence of water as described in Materials and methods, and the free energies of complex formation were calculated both in the presence and in the absence of calcium. The binding energies are summarized in Table II, and their hydrogen bonding (H-bonding) patterns are given in Table I. In all the cases including the crystal structure (3GBP) of GGBP from S. typhimurium bound to glucose, there are small differences in binding energies in the presence and absence of calcium. In a recent study, it has been shown that calcium depletion has a small effect on the secondary structures of GGBP of E. coli resulting in the reduction of its thermal stability, and glucose binding eliminated the effect induced by calcium depletion by restoring secondary structures similar to that of the native protein (D’Auria et al., 2006Go). Hence, this small difference in binding energies in the presence and absence of calcium is quite reasonable.


View this table:
[in this window]
[in a new window]

  		Table II. Empirical free energies of the complex formation between proteins (crystal structure and the modeled structures) with glucose and galactose in aqueous solution in the presence and absence of calcium ion

 

Total free energies of the complexes have two components: van der Waals (vdW) and electrical. The contribution of the vdW interactions is lower, varying between –16 and –26 kcal/mol, compared with the electrical contributions with variations in the range from –58 to –92 kcal/mol (Table II). Values of {Delta}Gbind were calculated by the linear interaction energy approximation as described in Materials and methods (Aqvist and Mowbray, 1995Go) for the complexes of glucose with all the models as well as the x-ray structure and presented in Table II. It may be noted that the calculated {Delta}Gbind values for the complexes of glucose with the proteins from S. typhimurium (3GBP), A. pleuropneumoniae, and C. tetani are negative. Therefore, these proteins will form stable complexes with glucose, as these complexes have lower free energy in the protein–water environment compared with water. The values of {Delta}Gbind for proteins from S. typhimurium (3GBP), A. pleuropneumoniae, and C. tetani (Table II) correspond respectively to dissociation constants (Kd) of 0.018, 8.1, and 10.0 µM, respectively. Our calculated {Delta}Gbind value of –10.71 kcal/mol for the GGBP–glucose complex is very close to the range of experimental values from –9.1 to –10.1 kcal/mol (Aqvist and Mowbray, 1995Go). On the contrary, the calculated {Delta}Gbind values for T. pallidum and P. syringae are positive (Table II), indicating that the complex formation of glucose with these proteins in the aqueous medium is thermodynamically unfavorable.

A substantial portion of the electrical energy arises from H-bonding interaction, which plays a major role in the binding affinity. The side chains of the residues Asp-14, Asn-91, and Asn-256 from domain 1 and His-152, Asp-154, Arg-158, Asn-211, and Asp-236 from domain 2 interact with the hydroxyl groups and ring oxygen of the sugar molecule in the crystal structure (Table I). The sugar molecule is also involved in H-bonding with the water molecule through hydroxyl group oxygen atom (O3) at position C-3 of sugar. The residues involved in H-bonding with glucose are found to be almost conserved in the modeled proteins from A. pleuropneumoniae, C. tetani, and T. pallidum. In the modeled protein from A. pleuropneumoniae, it is fully conserved, and in the modeled protein from C. tetani, Asn-256 of domain 1 is not involved in H-bonding with glucose and also His-152 of domain 2 is replaced by Arg-92 of domain 1. Modeled protein from T. pallidum lacks Asn-256 from domain 1 as well as Arg-158 from domain 2. Here, Asp-14 and Asp-154 are replaced by Asp-183 and Ser-154 of domain 2. In case of modeled protein from P. syringae, a different set of residues are involved in H-bonding with glucose. In this case, side chains of the residues Asp-10 from domain 1 and Asn-152, Try-183, and Glu-211 from domain 2 are involved in H-bonding with glucose showing significantly negative free energy of complex formation with the sugar, although it has less sequence homology with the crystal structure. The H-bonding interaction of sugar hydroxyl group oxygen atom (O3) with the water molecule is found to be conserved in case of modeled proteins from A. pleuropneumoniae and C. tetani. In the modeled protein from T. pallidum, there is no such type of H-bond, but in the modeled protein from P. syringae, a different hydroxyl group oxygen atom (O6) is involved in H-bonding with a water molecule.

We docked galactose into the binding site of the modeled proteins on the basis of the x-ray structure of GGBP–galactose complex (1GCA [PDB] ) and optimized for getting the stable structure of the complex in the same way as that of glucose complex. The analysis of their free energy of complex formation shows that the modeled proteins have similar values with galactose as glucose (Table II). The contributions of the vdW interactions for all the complexes range from –18 to –29 kcal/mol, whereas the electrical contributions show higher variations ranging from –58 to –85 kcal/mol. The values of {Delta}Gbind for proteins from S. typhimurium (3GBP), A. pleuropneumoniae, and C. tetani as summarized in Table II correspond respectively to Kd of 0.039, 1.45, and 4.90 µM. Kd value for the GGBP–galactose complex (1GCA) from S. typhimurium is more than two times that of the GGBP–glucose complex (3GBP) suggesting a weaker complex with galactose. The experimentally determined Kd value is about 3.5 times less than the lowest value in the range from 0.14 to 4.0 µM (Aqvist and Mowbray, 1995Go). Calculated {Delta}Gbind values for the complexes of galactose with the proteins from T. pallidum and P. syringae are positive; hence, no complex formation is expected in the aqueous medium.

The residues involved in the H-bonding with galactose in all the modeled proteins are found to be identical in case of the complexes with glucose (Table I). Here, the sugar hydroxyl group oxygen atom (O3) is found to be involved in H-bonding with the water molecule, and this interaction is found to be conserved in case of modeled proteins from A. pleuropneumoniae and C. tetani. In the modeled protein from T. pallidum, there is no such type of H-bond, but in the modeled protein from P. syringae, a different set of hydroxyl group oxygen atoms (O6 and O4) as well as hydrogen atom of the hydroxyl group (HO6) at position C-6 is involved in H-bonding with water molecules. These results confirm the dual specificity of these receptors for glucose and galactose.

ABP from E. coli has only 14% sequence identity with the GGBP from S. typhimurium, but a high degree of structural similarity is present (Quiocho and Vyas, 1984Go; Vyas et al., 1991Go). The ABP is also composed of two globular domains connected by three polypeptide segments, and the ligand-binding site is located in the cleft between the two domains. In the crystal structure (3GBP) as well as the modeled structures, the residues involved in H-bonding with the sugar have similarity with ABP from E. coli. In ABP, the side chain of the residues Lys-10, Glu-14, Asp-90, Arg-151, Asn-205, and Asn-232 and also two water molecules are involved in H-bonding with the sugar (Vyas et al., 1991Go). In the crystal structure of S. typhimurium and also in the modeled proteins from A. pleuropneumoniae and C. tetani, the side chains of the residues Asp-14, Asn-91, Arg-158, Asn-211, and Asp-236 as well as one water molecule are involved in H-bonding with the hydroxyl groups and ring oxygen of the sugar molecule (Table I) which are identical or very similar to the residues involved in ABP. Modeled protein from T. pallidum lacks Arg-158 as well as water molecule involved in H-bonding with the sugar, and also Asp-14 is replaced by Asn-14. In case of modeled protein from P. syringae, a different set of residues, that is, side chains of the residues Asp-10, Asn-152, Try-183, and Glu-211, and one water molecule are involved in H-bonding with the glucose.

The calculation of electrostatic potential on the surfaces around the binding site shows that the calcium-binding site is extensively negative due to the presence of acidic residues required for calcium binding except for the hypothetical protein of P. syringae. Electrostatic potential surfaces of the crystal structure of S. typhimurium and a few modeled proteins are shown in Figure 5. The glucose/galactose-binding site is mostly negatively charged with a small patch of a positively charged region in case of crystal structure and modeled proteins from A. pleuropneumoniae and C. tetani. The ligand-binding site is mostly negatively charged with a small patch of a hydrophobic region in case of modeled protein from T. pallidum and P. syringae. The highly negative potential surface around the positively charged calcium ion-binding site is necessary for the neutralization of the positive charge as is observed in other calcium-binding lectins.


Figure 5
View larger version (62K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5. Electrostatic potential surfaces of the modeled proteins. Blue region represents positively charged environment, red for negatively charged, and white for hydrophobic surroundings. The calcium-binding site is shown by yellow bands and glucose-binding site by black bands. (A) 3GBP; (B) a hypothetical protein of Actinobacillus pleuropneumoniae with 48% amino acid identity; (C) a hypothetical protein of Treponema pallidum with 36% amino acid identity; (D) a hypothetical protein of Pseudomonas syringae with 30% amino acid identity. Color figure is given as supplementary material online.

 


   Conclusions
Top
 Abstract
 Introduction
 Results and discussion
 Conclusions
Materials and methods
 Supplementary data
 Conflict of interest statement
 Acknowledgments
 References
 
We have modeled the structures of four bacterial proteins taking the crystal structure of a periplasmic glucose/galactose-binding receptor protein from S. typhimurium (PDB ID: 3GBP/1GCA) as template. Glucose and galactose were separately docked into the binding sites of the structures to study the physiochemical forces involved in such binding. The acidic residues involved in the coordination with calcium ion form a major CBL in 3GBP. It is also found to be conserved in modeled protein sequences except in the periplasmic protein from P. syringae due to less overall sequence homology. Seven calcium-coordinating atoms could be identified for all the structures, which also include one or two water molecules.

Optimum 3D structures of the complexes of glucose and galactose were predicted by energy minimization and molecular dynamics simulations of the complexes in the presence of water molecules both in the presence and in the absence of calcium ions. There are only minor differences in the calculated free energy of binding between the models in calcium bound and unbound forms; this small change is expected as a recent study (D’Auria et al., 2006Go) on calcium binding to GGBP from E. coli demonstrated small changes in the secondary structures of the protein by calcium depletion reducing its thermal stability, which was restored by glucose binding.

Total free energy of complex formation of glucose and galactose with the models, including the x-ray structures, of the proteins was found to increase in the order S. typhimurium < A. pleuropneumoniae < C. tetani < T. pallidum {approx} P. syringae indicating the order of decreasing strength of binding, which is also the order of decreasing amino acid identity/similarity with the starting scaffold used for homology modeling. The calculation of {Delta}Gbind values by the linear interaction energy approximation (Aqvist and Mowbray, 1995Go) suggests that both glucose and galactose can form thermodynamically stable complexes with the proteins from S. typhimurium, A. pleuropneumoniae, and C. tetani (Table II), whereas the proteins from T. pallidum and P. syringae gave positive values indicating no complex formation. It should be pointed out that the proteins from T. pallidum and P. syringae are less homologous to the starting scaffold (GGBP from S. typhimurium) for homology modeling, amino acid identities being only 36 and 30%, respectively. Hence, it remains uncertain whether the positive value of {Delta}Gbind is due to inaccuracy in structure prediction or intrinsic inability of these proteins in binding to glucose/galactose; however, this can only be answered by the experimental determination of their structures.

The most reliable values of Kd determined by various experimental techniques for binding of GGBP from S. typhimurium to glucose and galactose fall in the ranges 0.04–2 µM and 0.14–4 µM, respectively (Aqvist and Mowbray, 1995Go). Our calculations give the Kd values for binding of GGBP from S. typhimurium to glucose and galactose as 0.018 and 0.039 µM, respectively. Both these values are close to the lower side of the range, implying that this method predicts a little higher values in terms of association constant Ka. Considering the empirical nature of the prediction, this may be considered to be reasonably successful in the cases where experimental values are available for verification. Calculated Kd values, predicted for the complexes of the putative periplasmic proteins from A. pleuropneumoniae and C. tetani with glucose and galactose, lie between 1.45 and 10.0 µM which are in good agreement with the experimental range of values; hence, these proteins are expected to serve as glucose and galactose receptors having dual specificity of these sugars in these bacteria.


   Materials and methods
Top
 Abstract
 Introduction
 Results and discussion
 Conclusions
 Materials and methods
Supplementary data
 Conflict of interest statement
 Acknowledgments
 References
 
The initial structures of the glucose/galactose-binding periplasmic proteins were obtained by knowledge-based homology modeling using our in-house software package of ANALYN and MODELYN (Mandal, 1998Go). The starting scaffold for homology modeling was the x-ray crystallographically determined structure of a bacterial glucose/galactose-binding periplasmic receptor from pathogenic bacteria S. typhimurium (PDB ID: 3GBP/1GCA) (Mowbray et al., 1990Go; Zou et al., 1993Go; Flocco and Mowbray, 1994Go). The modeled structures were refined using the InsightII (2005) of Accelrys equipped with DISCOVER as the energy minimization and molecular dynamics module. Structural optimization involved energy minimization (100 steps each of steepest descent and conjugate gradient methods) using cff91 force field followed by dynamics simulations. A typical dynamics run for structure regularization consisted of 1000 steps of 1 fs after 100 steps of equilibration with a conformation sampling of 1 in 10 steps at 300K. At the end of the dynamics simulation, the conformation with lowest potential energy was picked for the next cycle of refinement using the ANALYSIS module of InsightII. This combination of minimization and dynamics was repeated until satisfactory conformational parameters were obtained. Special care was taken in the structural zones where major insertion or deletions were made.

The structures of the complexes were obtained by the superposition of the modeled protein structures with the experimental structure of S. typhimurium in complex with glucose/galactose (3GBP/1GCA), followed by optimization with repeated energy minimization and dynamics simulation. In this case, 1000 steps of equilibration were allowed in every span of 10 ps dynamics at 300K. To investigate the influence of water molecules on sugar and calcium ion binding, water molecules were added and associated with the complex using the Assembly/Soak option of InsightII to cover the desired area during molecular dynamics simulations and energy minimization. From the values of the free energies of complex formation of glucose/galactose in water and water–protein environments, we have calculated the absolute binding energies following the linear interaction energy approximation method of Aqvist and Mowbray (1995)Go using the relation {Delta}Gbind = 1/2 {Delta}<Vell-s>+ {alpha}{Delta}<VvdWl-s>,where {Delta}Gbind is the absolute binding energy, {Delta} stands for differences in the electrical (Vell-s) and vdW (VvdWl-s)components of the free energies of the ligand solvent (l-s) systems, that is, in pure water and protein-containing water environments. The weight factors of the electrical and vdW contributions were taken, respectively, as 0.5 (1/2) and 0.16 ({alpha}) used by them for a very similar system. Association constant (Ka) was calculated using the thermodynamic relation {Delta}Gbind = –RTlnKa, where R is the ideal gas constant and T the absolute temperature; Kd was calculated by taking the inverse of Ka.

Position constraints were applied to the atoms, which were >10 Å away during energy minimization and molecular dynamics of the complexes. During the regularization of the structures of the complexes with calcium and glucose/galactose, the distances between the calcium ion and the atoms involved in coordination with the calcium ions were kept constant by applying "generic distance constraints" from the DISCOVER module of InsightII.

ANALYN was used for the analysis of prealigned sequences of the target and scaffold proteins and was run on IBM-compatible PC. MODELYN was used for automated prediction of the initial target structure and for its structural analysis after refinement; it was run both on IBM-compatible PC in the Windows environment and in FUEL workstation of Silicon Graphics, Inc. in the IRIX environment. InsightII was run on FUEL workstation and Altrix 350 server of Silicon Graphics, Inc. in the IRIX environment. CLUSTALW (Thompson et al., 1994Go) was run through the Internet for multiple alignment of the amino acid sequences. The electrostatic potential surfaces of the proteins were determined by MOLMOL (Koradi et al., 1996Go). PROCHECK (Laskowski et al., 1993Go) was used for checking the structural parameters. Both MOLMOL and PROCHECK were run on FUEL in the UNIX operating system. MOLPROBITY (Davis et al., 2004) was used for all-atom contact analysis in terms of clashscores (number of atoms having atom pair overlaps ≥ 0.4 Å of 1000 atoms) and for the calculation of rotamer outliers. MOLPROBITY, being a general-purpose web service offering quality validation for 3D structures of proteins, nucleic acids, and complexes, was used through the Internet. H-Bonding patterns of the modeled and x-ray structures were obtained by adding hydrogen (x-ray structures lack hydrogen atoms, and DISCOVER needs these atoms for minimization and dynamics) followed by optimization of the complex by energy minimization and molecular dynamics. The free energies of binding of the complexes were calculated using the DOCKING module of InsightII. Protein BLAST (Altschul et al., 1997Go) was used through the Internet for finding homologous sequences.


   Supplementary data
Top
 Abstract
 Introduction
 Results and discussion
 Conclusions
 Materials and methods
 Supplementary data
Conflict of interest statement
 Acknowledgments
 References
 
Supplementary data are available at Glycobiology online (http://glycob.oxfordjournals.org/)


   Conflict of interest statement
Top
 Abstract
 Introduction
 Results and discussion
 Conclusions
 Materials and methods
 Supplementary data
 Conflict of interest statement
Acknowledgments
 References
 
None declared.


   Acknowledgments
Top
 Abstract
 Introduction
 Results and discussion
 Conclusions
 Materials and methods
 Supplementary data
 Conflict of interest statement
 Acknowledgments
References
 
This work was financially supported by the CSIR MMP Project No. CMM0017. M.P. is an SRF of UGC, India.


   Abbreviations
 
3D, three dimensional; ABP, l-arabinose-binding protein; CBLs, calcium-binding loops; GGBP, glucose/galactose-binding protein; RMSD, root mean square deviation; vdW, van der Waals


   References
Top
 Abstract
 Introduction
 Results and discussion
 Conclusions
 Materials and methods
 Supplementary data
 Conflict of interest statement
 Acknowledgments
 References
 
Altschul, S.F., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., 25, 3389–3402.[Abstract/Free Full Text]

Aqvist, J. and Mowbray, S.L. (1995) Sugar recognition by a glucose/galactose receptor. J. Biol. Chem., 270 (17), 9978–9981.[Abstract/Free Full Text]

Brüggemann, H., Bäumer, S., Fricke, W.F., Wiezer, A., Liesegang, H., Decker, I., Herzberg, C., Martínez-Arias, R., Merkl, R., Henne, A., and Gottschalk, G. (2003) The genome sequence of Clostridium tetani, the causative agent of tetanus disease. Proc. Natl. Acad. Sci. U. S. A., 100, 1316–1321.[Abstract/Free Full Text]

Casadevall, A. and Liise-anne, P. (1999) Host–pathogen interactions: redefining the basic concepts of virulence and pathogenicity. Infect. Immun., 67, 3703–3713.[Free Full Text]

Clodfelder-Miller, B., De Sarno, P., Zmijewska, A.A., Song, L., and Jope, R.S. (2005) Physiological and pathological changes in glucose regulate brain Akt and glycogen synthase kinase-3. J. Biol. Chem., 280 (48), 9723–9731.

D’Auria, S., Ausili, A., Marabotti, A., Varriale, A., Scognamiglio, V., Staiano, M., Bertoli, E., Rossi, M., and Tanfani, F. (2006) Binding of glucose to the D-galactose/D-glucose-binding protein from Escherichia coli restores the native protein secondary structure and thermostability that are lost upon calcium depletion. J. Biochem., 139 (2), 213–221.[Abstract/Free Full Text]

Davis, I.W., Murray, L.W., Richardson, J.S., Richardson, D.C. (2004) MOLPROBITY: structure validation and all-atom contact analysis for nucleic acids and their complexes. Nucleic Acids Research, 32, W615–W619.

Feil, H., Feil, W.S., Chain, P., Larimer, F., DiBartolo, G., Copeland, A., Lykidis, A., Trong, S., Nolan, M., Goltsman, E., and others (2005) Comparison of two complete genome sequences of Pseudomonas syringae pv. syringae B728a and pv. tomato DC3000. Proc. Natl. Acad. Sci. U. S. A., 102 (31), 11064–11069.[Abstract/Free Full Text]

Flocco, M.M. and Mowbray, S.L. (1994) The 1.9 Å x-ray structure of a closed unliganded form of the periplasmic glucose/galactose receptor from Salmonella typhimurium. J. Biol. Chem., 269, 8931–8936.[Abstract/Free Full Text]

Fraser, C.M., Norris, S.J., Weinstock, G.M., White, O., Sutton, G.G., Dodson, R., Gwinn, M., Hickey, E.K., Clayton, R., Ketchum, K.A., and others (1998) Complete genome sequence of Treponema pallidum, the syphilis spirochete. Science, 281 (5375), 375–388.[Abstract/Free Full Text]

Koradi, R., Billeter, M., and Wuthrich, K. (1996) Related articles, MOLMOL: a program for display and analysis of macromolecular structures. J. Mol. Graph., 14, 51–55.[CrossRef][ISI][Medline]

Laskowski, R.A., MacArthur, M.W., Moss, D.S., and Thornton, J.M. (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr., 6, 283–291.

Mandal, C. (1998) MODELYN – a molecular modelling program version PC-1.0. Indian Copyright No. 9/98.

Mitchell, E., Houles, C., Sudakevitz, D., Wimmerova, M., Gautier, C., Perez, S., Wu, A.M., Gilboa-Garber, N., and Imberty, A. (2002) Structural basis for oligosaccharide-mediated adhesion of Pseudomonas aeruginosa in the lungs of cystic fibrosis patients. Nat. Struct. Biol., 9, 918–921.[CrossRef][ISI][Medline]

Mowbray, S.L., Smith, R.D., and Cole, L.B. (1990) Structure of the periplasmic glucose/galactose receptor of Salmonella typhimurium. Receptor, 1, 41–53.

Nagaoka, K., Kazuhiko, T., Tanaka, K., Yoshida, H., Steinman, R.M., Saitoh, S., Akashi-Takamura, S., Miyake, K., Sun Kang, Y., Gyu Park, C., and Inaba, K. (2005) Association of SIGNR1 with TLR4-MD-2 enhances signal transduction by recognition of LPS in gram-negative bacteria. Int. Immunol., 17 (7), 827–836.[Abstract/Free Full Text]

Quiocho, F.A. and Vyas, N.K. (1984) Nobel stereospecificity of the L-arabinose-binding protein. Nature, 310, 381–386.[CrossRef][Medline]

Thompson, J.D., Higgins, D.G., and Gibson, T.J. (1994) CLUSTALW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res., 22, 4673–4680.[Abstract/Free Full Text]

Vyas, N.K., Vyas, M.N., and Quiocho, F.A. (1987) A novel calcium binding site in the galactose-binding protein of bacterial transport and chemotaxis. Nature, 327, 635–638.[CrossRef][Medline]

Vyas, N.K., Vyas, M.N., and Quiocho, F.A. (1991) Comparison of the periplasmic receptors for L-arabinose, D-glucose/D-galactose, and D-ribose. J. Biol. Chem., 266 (8), 5226–5237.[Abstract/Free Full Text]

Weissman, S.J., Chattopadhyay, S., Aprikian, P., Obata-Yasuoka, M., Yarova-Yarovaya, Y., Stapleton, A., Ba-Thein, W., Dykhuizen, D., Johnson, J.R., and Sokurenko, E.V. (2006) Clonal analysis reveals high rate of structural mutations in fimbrial adhesins of extraintestinal pathogenic Escherichia coli. Mol. Microbiol., 59 (3), 975–988.[CrossRef][ISI][Medline]

Zou, J.Y., Flocco, M.M., and Mowbray, S.L. (1993) The 1.7 Å refined x-ray structure of the periplasmic glucose/galactose receptor from Salmonella typhimurium. J. Mol. Biol., 233, 739–752.[CrossRef][ISI][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
16/10/959    most recent
cwl021v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Patra, M.
Right arrow Articles by Mandal, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Patra, M.
Right arrow Articles by Mandal, C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?