Extracellular signals perceived by the cells invoke an appropriate intracellular response triggering cascades of events that alter the activities of various signalling molecules. The variations in functional states of the proteins are brought about by covalent and non-covalent molecular interactions. Intracellular signalling in prokaryotes and eukaryotes is often achieved by phosphorylation of target proteins. The phosphorylated proteins undergo changes in their three-dimensional (3-D) shapes resulting in altered functional states. Phosphorylation events also lead to changes in the electrostatic interactions and ionisation states of residues at the catalytic site as exemplified by regulation of isocitrate dehydrogenases and HMG-CoA reductase 12. Covalent attachment of the phosphoryl groups to various proteins is catalysed by protein kinases. These enzymes transfer a phosphate group from adenosine tri phosphate (ATP) onto an acceptor amino acid in a substrate protein. The acceptor groups on amino acid residues could be alcoholic as in Ser and Thr, phenolic group as in tyrosine or basic amino acid such as histidine (N^π or N^τ), arginine (guanidino group), and the ε-NH2 group of lysine, carboxyl groups of aspartate and glutamate or thiol group of cysteine 34. In addition to nucleoside triphosphates as donors of the phosphate groups, certain class of phosphorylating enzymes makes use of either phospho-enzyme intermediates or low molecular weight metabolites such as phosphoenol pyruvate, acetyl phosphate or carbamoyl phosphates and polyphosphates as donors 567.

Despite the occurrences of various phosphorylating enzymes, cellular signalling mechanisms are dependent on a subset of the list of enzymes mentioned above, for intra-cellular communications. In eubacteria and archaea, two component signal transduction systems 8, also referred to as His-Asp phospho-relay systems, mediate adaptive responses to changes in environmental conditions. The two-component system is composed of a histidine kinase (HK) and a cognate regulatory protein referred to as a response regulator (RR). HKs consist of an ATP-binding kinase domain and the H-box domain carrying the histidine which is the site of phosphorylation. HKs have also been identified in eukaryotes, such as yeast, Neurospora and Arabidopsis thaliana, and their functional roles have been studied 910.

However, in eukaryotes, Ser/Thr and Tyr kinases play a key role in cellular signal transduction. The Ser/Thr and Tyr kinases differ completely from the HKs both in sequence and structure. The Ser/Thr and the Tyr kinase share a common catalytic core of about 270 residues made up of N-terminal lobe consisting of mainly β-strands and a C-terminal lobe composed predominantly of α-helices 11. The eukaryotes also possess two other distinct classes of protein kinases, the myosin heavy chain kinase/EF-2 kinase 12 and the mitochondrial protein kinases that include the pyruvate dehydrogenase kinase and branched-chain alpha-ketoacid dehydrogenase kinase 13 that differ from the Ser/Thr kinases in sequence and structure. Protein phosphorylation is hence considered to have originated independently in prokaryotes and eukaryotes in order to meet their diverse cellular and environmental requirements.

However studies on the eukaryotic protein-Ser/Thr/Tyr kinase like sequences (STYKs) in the prokaryotes emerged with the identification of proteins phosphorylated on Ser/Thr and Tyr residues in bacteria 14. Phosphorylation of Ser/Thr residues in archaeal proteins has also been reported for Sulfolobus acidocaldarius by earlier studies 15. Subsequent identification and characterisation of genes encoding STYKs in Myxococcus xanthus 16 and Streptomyces species 17 confirmed the presence of homologues of eukaryotic protein kinases in bacteria. Following these reports of STYKs, and the release of large number of bacterial genome sequences, a number of groups have surveyed for the occurrences of STYKs in them 1819202122. Their surveys suggested the occurrences of significant number of STYKs in both archaea and eubacterial genomes and they have been further classified based on their sequence similarity to various subfamilies 19. A previous study 20 proposed that the Ser/Thr/Tyr protein kinases in bacteria have bacterial origin. Kinase activity in the STYKs has been demonstrated in a few species. These studies have suggested the involvement of STYKs in secondary metabolism 17, morphology of the aerial hyphae in Streptomyces granticolor 23, regulation of fruiting body formation in M.xanthus 24. STYKs in Pseudomonas aeruginosa and Yersinia pseudotuberculosis 2526 are shown to be critical for progression of infection and expression of virulence factors during certain stages of infection. The role of STYKs in the repression of proline utilisation operon and degradation of proline in S.typhimurium has been suggested 27.

The recently reported 3-D crystal structures of Mycobacterial PKN-B 2829 has revealed an overall similarity of their structure to the eukaryotic Ser/Thr protein kinases with variations in the nucleotide binding regions. The prokaryotic homologues also include aminoglycoside kinases which phosphoryate antibiotics, and share remote sequence similarity with the STYKs. The 3-D fold of their catalytic domain is highly similar to the structures of eukaryotic STYKs 30. Studies on the evolution of STYKs in Synechocystis sp. PCC 6803 have proposed the origin of STYKs before the divergence between prokaryotes and eukaryotes during evolution 20. The conservation of these sequences in significant number of bacterial genomes despite limitation on their genomic size during evolution have therefore suggested their functions are essential in regulation of cellular activities in prokaryotes, as in their eukaryotic counterparts.

The present analysis is focussed on understanding the biological roles of the prokaryotic STYKs which have been shown to be encoded by significant number of genes in a large number of bacterial genomes by various groups. The present study is based on representative domain architectures of STYKs identified from the completely sequenced genomes of 125 prokaryotes which includes 35 genomes compiled in our early version of KinG database 31. The domain architectures of STYKs and their homologues in 90 additional genomes included in this paper are recently added to the KinG database http://hodgkin.mbu.iisc.ernet.in/~king.

About 600 STYKs and their homologues have been identified in the current study in 125 bacterial genomes, of complete genomic data, using various sequence and profile search methods. Manual analysis of each of these STYKs and homologues has been made in order to check for the presence or absence of residues known to be critical for the activity of eukaryotic protein kinases. Using profile based search method (See Materials and Methods) the extent of similarity of bacterial homologues with different subfamilies of eukaryotic protein kinases has been analysed. The analysis suggests that the bacterial homologues form a distinct class of protein kinases as indicated by the highest similarity (20–30% sequence identity) of their catalytic kinase domain with the generic kinase profiles rather than any individual subfamily profile of eukaryotic protein kinases. i.e., the percent sequence identities of bacterial kinases with specific sub-families of eukaryotic STYKs is lower.

During the course of current analysis we encountered lipopolysaccharide kinases (LPSKs) which are well known for their remote relationship with STYKs 32. RIO1 class of STYKs have been previously identified in eukaryotes 3334 and in archaea 19 and our earlier work reported the occurrence of RIO1 proteins in all the three major taxons namely, Archaea, Bacteria and Eukaryota 32.

The domain organisation of STYKs identified from various genomes has been studied in an attempt to understand their probable biological roles. Groups of STYKs conserved only among specific classes of bacteria have been identified. Diverse domain arrangements observed in bacterial STYKs are radically different from those observed in eukaryotic Ser/Thr or Tyr kinases and prokaryotic HKs. These domain arrangements unique to bacterial STYKs are therefore suggestive of their functional roles distinct from that of other well-known families of protein kinases. Consideration of known functions of domains tethered to catalytic kinase domain of bacterial STYKs enabled obtaining clues about gross functional roles of various bacterial STYKs. Involvement of STYKs in stress response, protein folding, extensive protein-protein interactions and sugar transport has been suggested by covalent tethering of catalytic kinase domains of STYKs with other enzyme and non-enzyme modules. Homologues of bacterial PKs that have been previously well characterised in terms of biochemical activity and function have also been identified in the study. Putative substrates for a small set of bacterial STYKs have also been identified based on their similarity to bacterial kinases with well characterised substrates.

The STYK-like sequences have been identified in various completely sequenced bacterial genomes and have been analysed for the occurrence of other functional domains. The domains have been identified using different profile search methods with stringent search parameters as described in the section on Materials and Methods and have also been manually checked for the conservation of key motifs. It should be noted that we have used stringent e-value cut-off of 10^-8 and 0.01 respectively in the PSSM (Position Specific Scoring Matrix) matching and HMM (Hidden Markov Model) matching procedures. All the functional domain assignments discussed below are detected within the e-value limit of 0.01 by HMMER and are ensured, manually, that there is no obvious erroneous domain assignment.

The genome-wide analyses reveal the occurrences of a large number of STYKs in bacteria. However the available knowledge about their biological functions is limited.

The STYKs show a lower representation compared to HKs in most of the genomes analysed (Data not shown). HKs are encoded by higher number of genes in comparison with STYK encoding genes in the prokaryotes. Nearly two percent of the genomes of eukaryotic organisms encodes STYKs as suggested from the genome-wide analysis by various groups 3484858687. Difference in the relative abundance of kinases belonging to the two classes, namely HKs and STYKs in the eukaryotes and the prokaryotes probably arises from the selection of specific class of PKs during the course of evolution, as an adaptation to distinct cellular environment. Low occurrence of STYKs compared to HKs in most bacterial species with the exception of Pirellula (Rhodopirellula baltica SH1) suggests, in general, the predominant role of HKs in prokaryotes.

HisK in archaeal genomes have been suggested to have arisen from lateral gene transfer events explaining their limited representation 8889. With the exception of Buchnera, Onion phytoplasma and Thermococcus kodakaraensis, all the above mentioned genomes have at least one of the two kinds of protein kinases, namely HisK and STYK. STYKs have not been identified in genomes of the two spirochaetes, Treponema pallidum and Borrelia burgdorferi, Chlorobium tepidum, Helicobacter hepaticus, Thermotoga maritima and Xanthomonas species. The absence of either classes of protein kinases in Buchnera species, an obligate symbiont of the Aphids and Onion phytoplasma (intra-cellular plant pathogen) suggests the loss of such enzymes in the above species is likely to be an adaptation to the host-intracellular environment. Our analysis also identifies an anaerobic obligate heterotrophic thermophile, T.kodakaraensis as the only species with a non-symbiotic or non-parasitic habitat that lacks HisK and STYKs. Further studies on influence of phosphorylation on response to external stimuli will therefore be able to provide insight into the nature of enzymes and significance of phosphorylation in these organsims.

The number of STYKs in phylogenetic domains of archaea and bacteria increase with the growing size of bacterial genomic data. The assignment of probable biochemical functions and biological roles to large number of sequences are often dependent on the occurrence of homologous gene products in other phylogenetic domains despite the vast divergence between the species. Earlier studies have also indicated that the gross biological function and interaction network of multi-domain proteins are conserved by the preservation of domain composition and arrangement 909192. Therefore domain organisation of the bacterial STYKs have been studied here to get an insight into their putative biological roles in comparison with other proteins of similar domain organisation.

Enzymatic and non-enzymatic domains described in the earlier sections have been found covalently tethered to the kinase catalytic domain. Among the archaeal multi-domain STYKs, the OSGP-containing STYKs are widely represented. Knowledge of substrates of the OSGP-containing STYKs and their sites of phosphorylation is required to understand their role in regulation of functions of glycoproteins. The Ser/Thr residues serve as sites for both glycosylation and the phosphorylation 9394. In many instances the phosphorylation and glycosylation occurs at common sites in a competitive manner. The phosphorylation would therefore prevent the glycosylation of the proteins. OSGP domains are involved in the cleavage of those proteins at the sites containing glycosylated Ser/Thr residues. The phosphorylation of the target proteins by the associated kinase domains of the OSGP containing STYKs might therefore serve as a regulatory mechanism to inhibit the proteolysis of glycoproteins.

Eubacterial STYKs are associated with a large variety of signalling domains. Chase 2 domains containing STYKs identified by recent studies 646573 are also found in other classes of trans membrane receptors including histidine kinases, diguanylate cyclases and methyl accepting proteins. Presence of widely occurring sensory domain in Ser/Thr kinases of Nostoc further emphasises on its role in the perception of some critical signals that are likely to trigger cellular pathways synergistically.

STYK of P.aeruginosa with phosphatase like domain has significant similarity with spoIIE protein family in the PFAM database. Recent studies 9596 have shown the involvement of Ser/Thr kinase-Phosphatase pair, PrkC and PrpC in the sporulation of B.subtilis. Psuedomonas aeruginosa however encodes a composite kinase-phosphatase in the same gene. The phosphatase-2C (PP2C) domains are known to selectively dephosphorylate protein phosphorylated on Ser/Thr residues and their homologues have been previously identified and characterised in various bacteria 979899100101. The PP2C domains have been predicted to be functionally active in SpoIIE 102103 of Bacillus subtilis, a Gram positive bacterial species. SpoIIE controls the activation of sporulation transcription factors and critical for the development of the bacteria.

The bacterial STYKs with TPR repeats identified here are suggested to be influenced by the chaperone like protein during their trafficking, folding or interactions with substrate proteins. The involvement of HSP90 in hormonal and growth factor signal transduction mediated by protein kinases through proteins containing TPR repeats is well characterised in eukaryotic cells 42. Further phosphatase pp5 of Plasmodium falciparum with TPR domain is known to act as co-chaperone for hsp90 and it has been therefore suggested that protein dephosphorylation might play a regulatory role in protein folding 104105. Despite their high sequence divergence the TPR repeats and their constituting domain form a general interface of interactions with HSP90 like chaperones 104.

STYKs with stress response domains such as USPA have been identified. The USPA proteins are suggested to be autophosphorylated to switch on to active state. Subsequent to the induction of USPA proteins, many phosphorylated proteins have been detected 106. The two-component system involved in nitrate assimilation is known to induce bacterial universal stress proteins of the nucleotide binding class 44. Majority of these proteins are tightly regulated and expressed under various conditions of stress in the bacteria.

The USPA containing kinases of plants have been implicated in disease resistance signalling pathway based on their similarity to other protein kinases like Pto and Pto interacting proteins of tomatoes that are involved in disease resistance 43107. Based on the role of USPA containing proteins in stress response and their regulation by the some two component system, we suggest the bacterial STYKs with USPA domains are likely to be induced during stress response. In the eukaryotes, MAP kinases are also known to trigger stress induced signalling pathways. This group of bacterial kinases could therefore be functionally analogous to their eukaryotic counterparts that are induced during stress.

Protein kinase containing gene products with C-terminal pentapeptide repeats have been identified initially by an early study 73 in Nostoc sp. PCC 7120 (gi|17132363) and Synechocystis sp. PCC 7803 (gi|1653478). Nitrate assimilation and nitrogen fixation are well characterised in various species of Cyanobacteria. The key enzyme involved in this pathway, the nitrogenase, is highly sensitive to oxygen. Some multi-cellular cyanobacterial species like Nostoc sp. form specialised cells called heterocysts wherein the glycolipid layer prevents the diffusion of oxygen 77. The accumulation of glycolipids into the heterocysts is brought about by a number of proteins. Some of these proteins have the pentapeptide repeats, which are known to be critical for the localisation of glycolipids 108.

PASTA domains conserved in a group of STYKs encoded by Gram positive bacteria are found in proteins comprising the divisome complex. The representation of this group of STYKs in all gram positive bacteria, including the Lactococcus lactis (gi|12724921), and Streptococcus pyogenes (gi|13622698), which contain only one STYK suggest their role to be critical in their cell division. The role of these protein kinases in cell division is reiterated with their genes clustering with rodA involved in the regulation of cell morphology during cell division. The phosphatase 2C encoding genes also lie in the same cluster supporting the role of reversible phophorylation during cell division.

Lipopolysaccharide kinases (LPSK) engaged in the phopshorylation of the lipopolysaccharides in the outer membrane of Gram-negative bacteria share remote similarity with eukaryotic STYKs 32. They could be grouped into two classes based on their sites of phosphorylation into waaP gene products and Kdo kinases (KdoK) 109110. A remote similarity between STYKs and the two family of LPSKs has been recently reported by our group 32. The similarity in the potential 3-D structure and catalytic residues of LPSKs with the eukaryotic protein kinases suggests analogous catalytic functions to the residues that are well conserved across the two distinct families of enzymes. The occurrence of Kdo kinase is so far restricted to pathogenic Gram-negative species and is also known to be influencing the virulence in H.influenzae 109. Inferences drawn on the structure and the catalytic mechanism of LPSK based on their similarity to eukaryotic protein kinases can therefore serve as a guide to the design of inhibitors to the LPSKs in virulent organisms.

The genomic organisation of all STYKs encoded in M.tuberculosis H37Rv genome has been analysed in an earlier study 111. The gene encoding the protein kinase PKN-G of M.tuberculosis is clustered with the genes, glnH for periplasmic Gln binding proteins. Therefore PKN-G is likely to be the functional homologue of the other Gln binding periplasmic Ser/Thr protein kinase of Streptomyces coelicolor (gi|21223285). These protein kinases therefore suggests the intake of amino acids in to the cell is also likely to be influenced by phosphorylation, in addition to the sugar transport in M.tuberculosis as described previously. Recent studies have established the involvement of PKN-G in the regulation of cellular levels of Gln/Glu 112113 and in evading phagocytosis by the host cells.

Eukaryotic signalling domains have been observed in a large number of bacterial proteins including in the STYKs 676869. The cyanobacterial STYK with a SH3b domain has a transmembrane segment between the kinase domain and the SH3 domain. The nature of signals perceived by this trans-membrane STYK is not clear. These SH3 domains are distantly related to their eukaryotic counterparts. Eukaryotic Ser/Thr kinases with SH3 domains have been well studied which is exemplified by mitogen activated protein kinase kinase 114115.

WD-repeats, a versatile protein-protein interaction module has also been observed bacterial protein kinases 6473116117 and in human protein kinases 46. Pkwa of Themospora curvata, a member of WD-repeat kinase family has been shown to be phosphorylated by exogenous protein kinase form S. granticolor 100.

The fork head associated (FHA) domain was initially found in transcription factors and has been recently identified as protein – protein interaction modules known to bind to phospho-threonine residues on target proteins. Hence the interactions of the FHA domains with their cognate proteins are largely influenced by phosphorylation. Many bacterial proteins have been recently identified with FHA domains 117. Previously it has been observed that the genes encoding FHA containing proteins cluster along with Ser/Thr protein kinases and phosphatases in the M. tuberculosis genome 111. Thus the identification of the PKs with FHA domains in cyanobacteria suggests synergistic actions of FHA domains containing proteins and enzymes involved in reversible phosphorylation as observed in eukaryotes.

Very few studies have shed light on the nature of substrates of bacterial kinase 17. The AfsR, the substrate of AfsK has an N-terminal transcription regulator domain (trans_reg domain), trans-activation domain (bacterial trans-activation domain, BTAD) followed by a signalling domain NB-ARC and a C-terminal tetratrico peptide repeat region which is known to be involved in protein-protein interactions. The BTAD domain is shared by all the members DNRI/REDD/AFSR family of transcription regulators involved in secondary metabolism. Other proteins in Streptomyces coelicolor and related species that control actinorhodin production share the trans_reg domain and the BTAD domain. A similar domain organisation among the homologues of AFSR is intriguing and is likely to contain sites of phosphorylation.

The phosphorylation site of HUα has been identified as a threonine residue which is well conserved across HUα of various species. RXXRTGR is the sequence pattern conserved at the phosphorylation site (P-site). The occurrence of R at the P-1 site suggests PKN2 to be a 'basic directed' protein kinase like Protein Kinase A (PKA) that specifically phosphorylates substrates with consensus motif RRXS/TZ' (X is any amino acid and Z is any hydrophobic amino acid). Based on the residue preferences of PKN2, at known phosphporylation sites on substrates, bacterial STYKs sharing significant similarity with PKN2 have been examined, that are likely to seek similar residue determinants on their substrates. The catalytic domain of PKN2 shares highest similarity (40% identity) with PKNB of M.tuberculosis and its homologue in M.leprae. The high sequence similarity between PKN2 and PKNB is likely to reflect in similar substrate binding modes such as preferences for N-terminal basic amino acid residues as also observed for PKA.

PKN2 of Myxococcus xanthus is known to autophosphorylate in the region following the kinase domain. The GC domain that lies C-terminal to the kinase domain contains 10 threonines. Further experimental studies are therefore necessary to ascertain if these sites could be the natural phosphorylation sites targeted by their respective enzymes.

A significant number of bacterial STYKs have been biochemically characterised previously to elucidate their biological roles. STYKs of the PKN subfamily have been identified and well characterised in Myxococcus xanthus 798081Streptomycetes coelicolor 17, Strepyomycetes granticolor 86, Synechocystis sp 118119120, Yersinia tuberculosis 33, Mycobacterium tuberculosis 111122 and Psuedomonas aeruginosa 55. The PKN representatives of Myxococcus xanthus and Streptomyces species are involved in the specialised developmental cycles characteristic to these species. They are also shown to be capable of autophosphorylation although it has been suggested to be unimportant for their kinase activity in vitro 1779808182.

The YpkA of Yersinia pseudotuberculosis is a secreted STYK that has been previously shown to interact with small GTPases like RhoA and Rac1 of the host 26. We have identified the homologue of YpkA, the only STYK in the completely sequenced genome of Yersinia pestis CO92 sharing 99% sequence identity with YpkA. However the genome of Yersinia pestis KIM does not encode any YpkA homologues. In addition to ABC1 and RIO1 class of STYKs, a single gene product containing only the kinase domain is identified.

PpkA of Pseudomonas aeruginosa have been implicated in their virulence and are suggested to be induced during the process of infection 55. An unusual domain arrangement of PpkA identified in the current analysis comprises of an N-terminal kinase domain followed by a stretch of 300 residues of unknown function and a C-terminal von Willebrand factor Type A (vWA) domain. The vWA domain, known to bind to glycoproteins and growth factor receptors, in PpkA is therefore suggested to aid in interactions at the surface of the host cell.

Mbk, is a Ser/Thr kinase from Mycobacterium tuberculosis encoded by the gene mbk 121 that lies in pst operon that controls phosphate transport. Mbk corresponds to STYK gi|2078052 in M.tuberculosis genome, which has an N-terminal kinase domain, a transmembrane region and a C-terminal NHL-repeats. The occurrence of membrane spanning region and extra-cellular NHL-repeats that forms an interaction surface suggests the Mbk to serve as receptors for unknown ligands regulating the phosphate transport.

The STYK, referred to as SpkA in Synechocystis is previously shown to be required for the motility of the cyanobacteria. SpkA corresponds to STYK (gi|1652312), in the genome of Synechocystis sp that has an N-terminal kinase domain and a C-terminal extension of 200 residues to which no known functional domain could be associated. The role of the C-terminal extension in the regulation or interaction with other proteins controlling the motility of the organism remains to be elucidated

The attempt made in the current analysis to study the bacterial STYKs by In silico analysis is hoped to aid in the further understanding of STYKs encoded in various genomes. The diversity of the functional domains occurring in these protein kinase like gene products provide clues to their biological functions that are yet to be explored experimentally. The involvement of bacterial STYKs in various cellular processes described in the previous sections suggests a very little overlap with the Histidine kinases and the protein kinases of the eukaryotes.

Large-scale genomic analysis are fraught with uncertainties related to the inheritance of function among remotely homologous proteins. Nevertheless search procedures employed in the current study for the study of domain organisation of gene products has been extensively bench marked 122 An accuracy level of 96% is achieved by use of stringent search parameters described in 'Materials and Methods' section. The functional roles of STYKs and related systems encoded in various prokaryotic organisms deduced using computational approaches in the present analysis and in a recent independent analysis 123 is hoped to provide useful leads for further studies on their regulation.

Experimental studies [eg., 124] elucidating the mechanisms of action STYKs and their homologues in various cellular processes as suggested from the analysis would therefore have implications in understanding their influence on the growth and development of the bacteria.

The complete set of predicted protein sequences from the ORFs of the bacterial genomes has been obtained from NCBI 125 Using sensitive sequence profile matching algorithms, STYK like sequences have been identified in the genomes.

We have employed multiple sensitive sequence search and analysis methods PSI-BLAST 126, IMPALA 127 and HMMer which matches Hidden Markov Models (HMMs) 128. These programs have been previously benchmarked 122129130 and we have used a stringent cutoff for e-values (0.0005 in PSI_BLAST, 10^-5 in IMPALA and 0.01 for HMMer) for identifying homologues of kinases. The list of predicted STYKs used for further analysis has been arrived at after careful cross-referencing between the results of these methods as well as manual scrutiny for a variety of factors such as length of the kinase domains and presence of critical functional residues. Kinase catalytic domains of the PKLA were aligned using CLUSTALW. Manual decisions are taken considering the presence/absence of functional motifs and the lengths of putative kinase domains. Those kinase-like sequences without functional residues, such as an aspartate in the catalytic base position, are segregated separately and are not considered for the detailed analysis. Number of such sequences in various genomes is provided in Table 2.

Domain assignment to the non-catalytic regions of the kinase containing genes has been carried out using the HMM search methods by querying each of the kinase containing sequences against the 6190 protein family HMMs available in the Pfam database 131. Trans-membrane segments were detected using TMHMM 132.

The 3-dimensional structures were superimposed using STAMP 133.

AK performed the computational sequence analysis and modelling. AK and NS conceived the study and participated in its design and coordination. Both the authors read and approved the final manuscript.