SAVES: Structural Analysis and Verification Server; MEGA: Molecular Evolutionary Genetic Analysis; LBP: Local Bootstrap values; ACC: Auto Cross Covariance; PI: Protrusion Index; NEFF: Number of Effective sequence homologs

Leptospirosis is one of the most widespread zoonotic diseases in the world and is caused by the pathogen Leptospira [1]. Susceptible animals, including humans are infected by direct contact with urine from rodents, or indirectly through contaminated water. Transmission occurs via dermal abrasions or inoculation of the mucous or conjunctival membranes [2]. It is also known as Weil’s syndrome and the clinical manifestations of leptospirosis include high fever, bleeding, and renal failure. The major target of Leptospira is the renal proximal tubular cells of kidney. Mortality ranges from 10-15% in cases of the traditional Weil’s disease and can be more than 70% in cases of severe pulmonary haemorrhage syndrome (SPHS) [3-5]. Leptospirosis is endemic in most of the southern states of India like Kerala, Tamil Nadu and certain parts of Andhra Pradesh [6-8].

The Leptospira genus is sub-classified into 18 genomospecies that included both saprophytic and pathogeneic species [1,9]. Based on serologic methods, approximately 300 serovars have been identified; of which more than 200 are pathogenic [1,2,10]. The availability of genome sequence data for different Leptospira strains drives the discovery of new diagnostic tools and vaccines for Leptospirosis [11]. The major problem associated with Leptospirosis is the diagnosis of the disease, as it shows multiple symptoms; very often it has been confused with other diseases. The diagnosis of the leptospiral infection is very much complicated when compared with other ailments.

A number of leptospiral outer membrane proteins (OMPs) have been characterized including OmpL1 [12], LipL41 [13], LipL36 [14], LipL32 [14], LipL21 [15], LipL46 [16], LenA [17], Loa22 [16] and Omp52 [12]. However their performance in diagnostic assays for acute leptospirosis or as vaccine candidates has been problematic [12,18]. Among these OMPs, LipL41 appears to be a great interest since it is one of the key virulence determinants involved in host-pathogen interactions and it is being formed only in the host during the infection [13,15,18]. LipL41 is essential for virulence of L. interrogans in the animals [19]. Recent studies have revealed that LipL41 may play an important role in the infection and produces immunological responses in the host during the infection of Leptospira [12].

Considering the large number of pathogenic leptospiral serovars and broad distribution of leptospiral host reservoirs, the potential effect of selective pressure on the evolutionary mapping of the LipL41 proteins was not studied so far. The availability of genomic sequences of various serovars and strains opened up opportunities to identify evolutionary relationships among different pathogenic strains of L. interrogans and others representing various kinds of serotypes (serogroups and serovars). Given the potential of the LipL41 proteins as diagnostic antigens and vaccine candidates, we examined the evolutionary relationship of LipL41 with 87 sequences from various serovars and strains followed by protein threading and mapping of B-cell epitopes from the conserved region of the alignment to develop vaccine for Leptospirosis.

Sequence retrieval, sequence alignment and dendrogram construction

Amino acid sequences used in this study were retrieved from protein knowledgebase (UniProt KB) (http://www.uniprot.org/uniprot/). A total of 87 sequences of LipL41 from different serovars and strains (Table 1) were collected from UniProtKB. Sequences with significant identity were aligned with ClustalW algorithm implemented in Molecular Evolutionary Genetic Analysis (MEGA 5.2.2) (http://www.megasoftware.net) by using distance matrix and then it was trimmed to consensus. Neighbour Joining (NJ) trees were constructed with 1000 bootstraps at uniform divergence rates with distance ‘p’ as the evolutionary model and with a data subset to use with gaps/ missing data treatment as complete deletion [20]. Posterior probability and conserved regions among the closely related sequences were done with MEGA 5.2.2.

Table 1: List of LipL41 sequences of Leptospira species used for phylogenetic analysis in figure 1.

Modelling, energy minimization and validation of the model

The amino acid sequence of LipL41 of Leptospira interrogans serogroup Icterohaemorrhagiae serovar Lai (strain 56601) was retrieved from UniProt KB (Uniprot Id: Q8F8E1). The sequence was submitted into the RaptorX server [21] (http://raptorx.uchicago.edu/) to derive the 3-dimensional structure. The modelled protein structures were viewed in Swiss-PdbViewer (http://www.expasy.org/spdbv/) and the individual residues were collected from 100 cycles of steepest descent algorithm carried out in GROMOS96 [22] until the side chain interactions in the vicinity is readjusted and brings up lower potential energy and becomes more stable. Energy minimized models were assessed by PROCHECK [23] to analyse the stereo chemical quality and residual geometry of the model by submitting the co-ordinate file in Structural Analysis and Verification Server (SAVES) (http://nihserver.mbi.ucla.edu/SAVES/). The value of the predicted LipL41 model was analyzed by using PYMOL [24,25].

Computational mapping of epitopes

Linear and discontinuous B-Cell epitopes of LipL41 were mapped from the generated three dimensional structure of LipL41. Linear B-cell epitopes were chosen with two different algorithms: ABCPred and BepiPred. ABCPred uses a recurrent neural network to predict B-cell epitopes (http://www.imtech.res.in/raghava/abcpred/) [26,27]. The amino acid length of 16 and the scoring threshold of 0.8 were set to predict B-cell epitopes in ABCPred. The epitope prediction in BepiPred is based on hidden Markov model (http://www.cbs.dtu.dk/services/BepiPred/) and propensity scale method [28,29]. The value, 0.35 was set as the threshold value, because at this value, the sensitivity/ specificity of predictions are maximized in BepiPred. BepiPred analyzes each amino acid independently and does not have a minimum or maximum number of amino acids to predict an epitope.

Discontinuous epitopes were predicted using Ellipro which is an Antibody Epitope Prediction server (http://tools.immuneepitope.org/tools/ElliPro/iedbinput) [30,31]. Ellipro, with the best algorithm to predict discontinuous epitopes from 3-D structures when compared to six other software programs that predict discontinuous epitopes [30]. The default threshold value was set at 0.8. The predicted epitopes were additionally verified in VaxiJen server to predict the probability of an antigen (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) [32], with a threshold of 0.4. VaxiJen uses an alignment free approach for antigen prediction and works on an auto cross covariance (ACC) transformation of protein sequences into uniform vectors of principal amino acid properties. Sixteen sequences of LipL41 were taken randomly in T-coffee programme to identify the conserved amino acids residues.

Phylogenetic analysis for genetic relatedness

The phylogenetic tree was performed by using 87 sequences of LipL41 from various serovars and strains which were retrieved from UniProt Knowledgebase (Uniprot KB) (Table 1). The phylogenetic tree of LipL41 is evidenced that the isolates are clustered with different serovars and strains (Figure 1) and diverged to form different branches in the phylogenetic tree. The following mutations are observed for Borgpetersenii strains: 33S->T, 39M->I, 40F->Y, 125A->I, 126I->L, 130S->T, 139N->S, 191D->E, 336T->V; Weilii strains: 33S->A, 247I- >V; Fiocruz strains: 80A->P, 177L->I, 183A->V, 186M->A, 197E->D; Santorasai strains: 125A->I, 126I->L, 130S->T, 139N->S, 336T->I; Krishneri strains: 176I->V, 336T->A; Noguchi strains: 269I->M, 274R- >K, 336T->A (data not shown). The closest neighbouring clusters include strains of L. weilli, L. kirschneri and L. interrogans with 100% bootstrap confidence values. Based on the phylogenetic analysis, the cluster of LipL41 of L. borgpetersenii together forms a clade showing the evolutionary relationship of same serovars and strains with the highest bootstrap value of 98 which indicates that it has uniform support. The reliability of a branch length in MEGA 5 is based on confidence probability (CP). The branch length is high when the CP is high, thus the branch length is considered to be statistically significant. MEGA 5 inferred the evolutionary tree by a Neighbour-Joining (NJ) algorithm by using a matrix of pairwise distances. In order to resolve the relationships of the sequences within each group of the constructed phylogenetic tree (Figure 1), a separate phylogenetic tree was constructed with 20 sequences from different serovars (Table 2) (Figure 2). It shows the highest bootstrap value of 99, indicate that the clade is close to 100%, which reveals that all the characters in a group believed to comprise all the evolutionary descendants of a common ancestor which is rooted with different serovars and strains as the ancestral group. In order to resolve the polytomies and to make the evolutionary relationship into dichotomies, 8 sequences of different serovars and strains (Table 3) (Figure 3) was used to make a separate branch of tree reflected Polytomies with Local bootstrap probability (LBP) values below 50% in L. interrogans Icterohaemorrhagiae serovar Lai and Copenhageni serovar (Figure 3). The phylogenetic tree evidenced that L. interrogans Icterohaemorrhagiae serovar Lai and Copenhageni serovar were closely clustered with other different serovars. The neighbouring groups also include serovars of L. autumnalis and L. hebdomadis which clustered with different bootstrap confidence values. Even though LipL41 is highly analogous protein present in all pathogenic Leptospira but the phylogenetic pattern of the present study exhibited the clonality of the sequences of the serovar Lai and Copenhageni used to analyze species separation. The evolutionary relationship was confirmed among the 87 sequences and on further confirmation with serovar Lai can be used for serodiagnosis of pathogenic leptospiral species.

Table 2: List of LipL41 sequences of Leptospira species used for phylogenetic analysis in figure 2.

Table 3: List of LipL41 sequences of Leptospira interrogans used for phylogenetic analysis in figure 3.

Figure 1: Phylogenetic tree with 87 sequences of LipL 41. Bootstrap confidence levels are shown as percentages on nodes and confidence values are shown in branches. The phylogenetic tree showed the divergence of Leptospiral strains and serovar with LipL41.

Figure 2: A phylogram of 20 selected sequences of LipL41. The bootstrap values calculate the frequency for each taxon bipartition during replication and boot strapping denotes measures how consistently the data support given taxon bipartitions. The scale bar also represents branch length (number of amino acid substitutions/100 residues). The high bootstrap value 99 shows the uniform support and bootstrap values close to 100% which indicate that the clade is a group which means that all the characters in a group believed to comprise all the evolutionary descendants of a common ancestor which is rooted with different serovars and strains as the ancestral group.

Figure 3: A phylogram of 8 selected sequences of LipL 41. Polytomies reflect nodes with LBP values below 50% for Lai and Copenhageni serovar of L. interrogans Icterohaemorrhagiae. This tree resolves the relationship among different LipL 41 groups and the sequences within each group and serovars.

Structure prediction of LipL41 by threading method

Protein threading method is a fold recognition method of protein modelling which is based on the predicted structure properties, such as predicted secondary structures and predicted residue burial status [21]. Threading based prediction for LipL41 was done in RaptorX server which uses a non-linear scoring function to combine homologous information with structural information for the given templatesequence alignment [21] . The amino acid sequence of LipL41 of L. interrogans Icterohaemorrhagiae serovar Lai (Q8F8E1) was taken for modelling by protein threading. Given an input sequence, RaptorX predicted its tertiary structure as well as solvent accessibility and disordered regions (Figure 4A). The RaptorX assigned the confidence score which is based on P-value and uGDT (unnormalized Global Distance Test). P-value measures the relative quality and uGDT measures the absolute quality of protein model. uGDT has greater value of 50 is a indicator for a good model. The input of 355 amino acid residues of LipL41 were completely modelled with 100% and showed 2 domains (Figure 4A). In the model, 7 positions among 355 residues were predicted as disordered regions which are 1%. The model shows P-value with 3.52 e-03 and uGDT (GDT) with 137(38). The modelled LipL41 has two domains (Figure 4A); Domain 1 (254 to 355) showed p value of 5.32 e-3 and Domain 2 (1 to 253) showed p value 3.52 e-3.

Figure 4:A: Three dimensional structure of LipL41 of Leptospira interrogans serogroup Icterohaemorrhagiae serovar Lai (strain 56601 (Uniprot Id: Q8F8E1). B: LipL41 model showing haem binding regions (CS and CP) marked in red.

LipL41 is a haem binding protein with Cys-Ser (CS) and Cys-Pro (CP) domains [9] (Figure 4B). The CS and CP are conserved domains of pathogenic Leptospira which are responsible for immunoprotection [9]. Mutation of these domains fails to cause immunoprotection in mice [12]. The motifs Cys-Pro or Cys-Ser has been determined in diverse proteins binding to heme (Fe2+)/hemin (Fe3+) [33,34]. It has been reported that the cysteine containing dipeptide: CS or CP is necessary for heme binding in HRM [35-37]. These conserved residues are found in LipL41 at 140 Cys-Ser and 220 Cys-Pro are located on the surface of the predicted structure (Figure 4B), and the thiol of cysteine may be a ligand for iron on heme [36,37].

Validation and evaluation of LipL41 model

The NEFF (Number of Effective sequence homologs) score which was ranging from 1 to 20 for the predicted structure was estimated by PROCHECK. The results obtained from PROCHECK [23,38] was evaluated for protein backbone conformations by Ramachandran Plot [39,40]. The phi-psi torsion angle for 92.7% of residues of LipL41 are in the most favourable region (A, B and L); 6.6%, 0.3% and 0.2% in additionally allowed, generously allowed and disallowed regions, respectively (a, b, l, p), indicate that LipL41 model is stereo chemically good (Figure 5) and the model derived from RaptorX was of higher quality in terms of protein folding.

Figure 5: Ramachandran plot of the LipL41 model. The most favoured regions are colored red, additional allowed, generously allowed and disallowed regions are indicated as yellow, light yellow and white fields, respectively.

Prediction and immunoinformatic analysis of antigenic peptides

Two different epitope prediction software programs (ABCPred and BepiPred) were utilized to predict the most immunogenic linear B-cell epitopes on the surface of the leptospiral OMP LipL41. ABCPred and BepiPred predicted 9 different overlapping and potentially immunogenic regions within LipL41, respectively (data not shown). ABCPred is able to predict epitopes with approximately 66% accuracy using the recurrent neural network [26]. ABCPred assigns scores between 0 and 1 for each epitope it predicts. If prediction shows score closer to 1, the particular prediction can be taken as epitope and prediction closer to 0 is not suitable for an epitope. Eight B- cell discontinuous epitopes of LipL41 were mapped from the predicted 3-D structure of LipL41 by using Ellipro (Figure 6). These epitopes spans (EP1: 22-42), (EP2: 68-97), (EP3: 132-150), (EP4: 167-182), (EP5: 185- 189), (EP6: 195-213), (EP7: 267-284) and (EP8: 303-355) positions of LipL41 (Table 4). Ellipro predicts epitope with Protrusion Index (PI) value which is percentage of the protein atoms enclosed in the ellipsoid, at which the residue first becomes lying outside the ellipsoid; whereas all the residues which were lying 90% outside the ellipsoid had the PI value 9, i.e., 0.9 in Ellipro. This gives information of amino acids lying outside the ellipsoid.

Table 4: List of epitopes predicted based on amino acid sequence and structure.

Figure 6: Structures of B-cell epitopes of LipL41.

The prediction of peptides is vital not only for diagnostics but also for vaccines. It became clear that the small segments of protein called the antigenic determinants or the epitopes are sufficient for eliciting the desired immune response. Based on the threshold value, all the predicted epitopes are antigenic nature. All these epitopes predicted can be used for the development of Monoclonal antibody or epitope based diagnostic kit for the Leptospirosis.

Conservancy of LipL41 epitopes

Universal epitope vaccine development requires conserved amino acids of a protein among the various pathogenic strains [41]. Thus 16 different strains and serovars of Leptospira were taken randomly. Multiple sequence alignment of LipL41 for 16 different strains and serovars of Leptospira interrogans: Leptospira interrogans serovar Wolffi (Q6GXB5), Leptospira weilii serovar Manhao II (Q6GXB6), Leptospira interrogans serovar Canicola (Q6GXC6), Leptospira interrogans serovar Australis (Q6GXC2), Leptospira interrogans serovar Autumnalis (Q6GXC3), Leptospira kirschneri serovar Bulgarica str. Nikolaevo (M6F5M3), Leptospira interrogans str.L1207(M6M715), Leptospira interrogans serogroup Icterohaemorrhagiae serovar Lai (strain_56601) (Q8F8E1), Leptospira interrogans serogroup Icterohaemorrhagiae serovar copenhageni (strain_Fiocruz_L1-130) (Q72N71), Leptospira interrogans serovar Autumnalis (Q33BN1), Leptospira interrogans serovar Hebdomadis (Q33BN0), Leptospira interrogans serovar Manilae (Q33BM7), Leptospira interrogans serovar Australis (Q33BM9), Leptospira interrogans serovar Icterohaemorrhagiae (Q33BM8), Leptospira kirschneri str. (K6FCL0), Leptospira interrogans serovar Icterohaemorrhagiae (C9EH90) were analyzed by using T-coffee program (http://www.tcoffee.org/) [42] (Figure 7) and found that all these epitopes are conserved among all these strains and serovars.

Figure 7: Multiple amino acid sequences alignment of LipL41. Red regions in the above alignment show perfect sequence alignment among the sequences. Conserved epitopes are marked.

Prediction of transmembrane domain, signal sequence and topology of LipL41

In order to ensure that the epitopic region should not overlap with signal peptide or transmembrane domain of LipL41, amino acid sequence of LipL41 of Leptospira interrogans serogroup Icterohaemorrhagiae serovar Lai (strain_56601) was analyzed by TMHMM 2.0 server (http://www.cbs.dtu.dk/services/TMHMM/) [43]. It was found that amino acids from 1-6, 7-29, and 30-355 are located inside the plasma membrane, inside the transmembrane (TM) helix, and outside the plasma membrane of the cell, respectively. A combined trans-membrane topology and signal peptides were predicted using Phobius online server (http://phobius.sbc.su.se/) [44]. Based on the probability of occurrences of specific amino acids, the sequence from 1 to 22 was predicted to be a signal peptide and between 23 and 355 is non-cytoplasmic region of LipL41. This analysis confirms that all the eight epitopes are topologically surface exposed and do not have any signal sequences in them.

In this present study, we have characterized LipL41 for its genetic diversity among the Leptospira species. The phylogenetic relationship of Leptospira with LipL41 from 87 sequences of different serovars and serogroup have shown that the comprising lineages with largely varying rates of evolution. The alignment also has shown the presence of haem binding motifs are conserved in all the LipL41. The three dimensional structure of LipL41 was predicted by using RAPTOR X and was validated by Ramachandran plot. Eight B-cell epitopes were predicted from LipL41. Antibody developed against these conserved epitopic regions could be used to develop a detection kit or as a vaccine candidate for Leptospirosis.

The authors thank DBT-IPLS and UGC-Networking Resource Centre in Biological Sciences (NRCBS) for providing necessary facilities to carry out the work.