|Accueil | Aperçu | Revues | Soumettre | Nous Contacter | English|
The LcrV protein is a multifunctional virulence factor and protective antigen of the plague bacterium and is generally conserved between the epidemic strains of Yersinia pestis. We investigated the diversity in the LcrV sequences among non-epidemic Y. pestis strains which have a limited virulence in selected animal models and for humans. Sequencing of lcrV genes from 19 Y. pestis strains belonging to different phylogenetic groups (“subspecies”) showed that the LcrV proteins possess four major variable hotspots at positions 18, 72, 273, and 324–326. These major variations, together with other minor substitutions in amino acid sequences, allowed us to classify the LcrV alleles into five sequence types (A-E). We observed that the strains of different Y. pestis “subspecies” can have the same type of LcrV, including that conserved in epidemic strains, and different types of LcrV can exist within the same natural plague focus. Therefore, the phenomenon of “selective virulence” characteristic of the strains of the microtus biovar is unlikely to be the result of polymorphism of the V antigen. The LcrV polymorphisms were structurally analyzed by comparing the modeled structures of LcrV from all available strains. All changes except one occurred either in flexible regions or on the surface of the protein, but local chemical properties (i.e. those of a hydrophobic, hydrophilic, amphipathic, or charged nature) were conserved across all of the strains. Polymorphisms in flexible and surface regions are likely subject to less selective pressure, and have a limited impact on the structure. In contrast, the substitution of tryptophan at position 113 with either glutamic acid or glycine likely has a serious influence on the regional structure of the protein, and these mutations might have an effect on the function of LcrV. The polymorphisms at positions 18, 72 and 273 were accountable for differences in the oligomerization of LcrV.
Plague is an acute and highly lethal disease caused by Yersinia pestis. Based on the ability to ferment glycerol and to reduce nitrate, Y. pestis strains can be assigned to the biovars antiqua, medievalis and orientalis, which are thought to be responsible for three major plague pandemics (Devignat, 1951; Perry & Fetherston, 1997). In addition, there is a group of atypical isolates of Y. pestis, circulating in populations of different species of voles (Microtus spp.), which have not been associated with epidemics of human plague thus far. These non-epidemic strains which currently include eight “subspecies” (Li et al., 2009) that were originally named Pestoides (Martinevsky, 1969), and in contrast to the most of epidemic-causing isolates, are capable of fermenting rhamnose and melibiose (Anisimov et al., 2004); however, five of “subspecies” have lost the ability to ferment arabinose (Li et al., 2009; Zhou et al., 2004a; Zhou et al., 2004b). In addition, some of these strains are sensitive to pesticin 1 of Y. pestis, and, generally have significantly reduced virulence in guinea pigs as well as in rabbits, rhesus monkeys, sheep, and humans (Anisimov et al., 2004; Zhou et al., 2004b). Soviet Union researches recommended that strains of this group, specifically altaica, caucasica, hissarica, ulegeica, and talassica, were to be classified into a “subspecies” of Y. pestis and named according to the areas in which they were isolated (Anisimov et al., 2004). Recently, Chinese researchers described similar strains isolated in China and proposed to assign rhamnose-positive and arabinose-negative isolates into the biovar microtus, because the main rodent host for these strains in the natural plague reservoirs is a vole (Microtus genus) (Zhou et al., 2004b; Zhou et al., 2004c). More recently it has been proposed to include all non-epidemic, rhamnose-positive variants into biovar microtus as the Y. pestis subspecies altaica, caucasica, hissarica, ulegeica, talassica, xilingolensis, qinghaiensis and angola (Li et al., 2009). In this paper we use the term “microtus” as a synonym for “Pestoides.” The microtus strains form a separate phylogenetic group when compared with the epidemic strains of other biovars by different methods of genotyping (Achtman et al., 2004; Cui et al., 2008; Li et al., 2009; Motin et al., 2002; Vergnaud et al., 2007). Two strains of this biovar have been sequenced and shown to have a unique genomic profile of gene loss and pseudogene distribution, which most likely accounts for the human attenuation of this new biovar (Garcia et al., 2007; Song et al., 2004; Zhou et al., 2004c).
LcrV (V antigen) is an essential virulence factor and protective antigen of Yersinia spp., which was discovered more than 50 years ago (Burrows, 1956). This protein is attributed to have a variety of immunomodulatory effects on the host (Abramov et al., 2007; Brubaker, 2003; Depaolo et al., 2008; Foligne et al., 2007; Heesemann et al., 2006; Motin et al., 1997; Nakajima et al., 1995; Nedialkov et al., 1997; Overheim et al., 2005; Reithmeier-Rost et al., 2004; Reithmeier-Rost et al., 2007; Sharma et al., 2005; Sing et al., 2002a; Sing et al., 2002b; Sing et al., 2003; Sing et al., 2005; Welkos et al., 1998), and indirectly regulates expression and secretion of type 3 secretion system (T3SS) virulence factors Yops (Bergman et al., 1991; Matson & Nilles, 2001; Nilles et al., 1998; Nilles et al., 1997; Price et al., 1991; Skrzypek & Straley, 1995), as well as being involved in their translocation to eukaryotic cells (Broms et al., 2007; Broms et al., 2003; Cowan et al., 2005; DeBord et al., 2001; Goure et al., 2005; Holmstrom et al., 2001; Lee et al., 2000; Marenne et al., 2003; Mota, 2006; Mueller et al., 2005; Pettersson et al., 1999; Philipovskiy et al., 2005; Roggenkamp et al., 1999; Sarker et al., 1998; Weeks et al., 2002). Recently, the crystal structure of this multifunctional protein has been resolved, and it was revealed that LcrV is a dumbbell-like molecule with two globular domains on either end separated by a coiled-coil motif that is uncommon in bacterial proteins (Derewenda et al., 2004).
The sequence of the lcrV gene is generally conserved between the epidemic strains of Y. pestis; however, sequencing of a few representatives of the microtus biovar revealed a surprising polymorphism in this virulence factor (Adair et al., 2000; Motin et al., 1992) (Garcia et al., 2007; Song et al., 2004). In this study we sequenced lcrV genes from 19 strains of Y. pestis, 17 of each belonged to the microtus strains of Y. pestis isolated in the territory of the Former Soviet Union (FSU) and Mongolia. The polymorphism in amino acid sequences of LcrV was used to evaluate the possible influence of these variations on the three-dimensional structure of the protein and functional activity of this major virulence factor of Y. pestis.
The strains of Y. pestis used in this study as a source of lcrV genes represented five "subspecies" circulating in the Eurasian natural plague foci and differed in their epidemiological significance (Anisimov et al., 2004). Two isolates (I-1996 and I-2638) were the epidemic-type strains of the biovar antiqua, while other isolates belonged to the group of atypical strains (Table 1).
The nucleotide sequence of each lcrV gene was determined by the direct sequencing of the PCR fragment obtained after amplification of the part of the lcrGVH operon of the corresponding strain. The primers LcrVF (5'-CAGCCTCAACATCCCTACGA-3') and LcrVR (5'-TGTCTGTCGTCTCTTGTTGC-3'), both flanking the lcrV gene, were located within the lcrG and lcrH genes, respectively. The additional primer LcrVFI (5'-GCAAAATGGCATCAAGCGAG-3') resided inside the lcrV gene. Determined sequences of the lcrV genes were deposited to the GenBank (see accession numbers in the Table 1) and compared with the available sequences of this gene from other Y. pestis strains (strain KIM, medievalis biovar, accession no. M26405; strain CO92, orientalis biovar, accession no. AL117189, strain Pestoides F, likely subspecies caucasica (Motin et al., 2002), accession no. AF167309; atypical strain Angola (subspecies angola) (Li et al., 2009; Motin et al., 2002), accession no. AF167310; strain 91001, microtus biovar (subspecies xilingolensis) (Li et al., 2009), accession no. AE017043; strain Antiqua, antiqua biovar, accession no. CP000311). Y. pestis strains and DNA isolates are from the State Research Center for Applied Microbiology and Biotechnology, Obolensk (Moscow Region, Russia) or were kindly provided by Prof. S. V. Balakhonov (Antiplague Research Institute of Siberia and Far East, Irkutsk, Russia).
We amplified the lcrV gene of the strains Angola, KIM and Pestoides F by using a forward primer, GCGGGATCCATTAGAGCCTACGAACAAAACCCAC, and reverse primer, CGGAATTCTCATTTACCAGACGTGTCATCTAGCA, containing BamHI and EcoRI sites, respectively. The PCR fragment was cloned into expression vector pRSET A (Invitrogen, Carlsbad, CA) digested by BamHI and EcoRI that resulted in a construct containing a fusion of the lcrV with the vector-encoded N-terminal His-Tag and the leader sequence of the T7 gene 10 under control of the T7 promoter. This N-terminal tail increased the molecular weight of the native LcrV protein by 3.9 kDa. After the sequence of lcrV was verified, the constructs were transformed to an E. coli BL21 (DE3) host. Expression and purification of recombinant protein was performed essentially as described by us previously (Motin et al., 1996). We analyzed purified proteins by using SDS-PAGE followed by silver stain or immunoblot with the monoclonal antibody to LcrV (Brubaker et al., 1987; Motin et al., 1994). Protein sample buffer contained a reducing agent dithiothreitol (DTT) at a concentration of 100 mM. Protein concentration was determined by using a BCA kit (Pierce, Rockford, IL) and prepared to equalize the amounts of protein loaded to SDS-PAGE.
The crystal structure of LcrV deposited in the Protein Data Bank (PDB) under the code 1R6F (Derewenda et al., 2004) was used as a structural template for modeling the LcrV sequence from Y. pestis Angola strain. This crystal structure is of the truncated (Δ1–22) triple mutant variant (K40A/D41A/K42A) from the KIM strain LcrV protein. In the deposited triple mutant structure of LcrV, PDB entry 1R6F, there are several disordered regions, these included the N- and C-termini (M1-H27 and T323-K326, respectively), Y90 and two large loops (Y50-A60 and N263-S273). Several atoms are also missing in the following amino acids: K49, N61, R62, D91, S274, and K276. The structural elements missing from the crystal structure were modeled by “grafting” suitable fragments from structures in the PDB using a Local-Global Alignment method (LGA) (Zemla, 2003). The secondary structure prediction from the PSIPRED method (Jones, 1999) was used to aid in the modeling of two long loops (50–60, and 263–273). The N-terminus was not modeled for lack of sequence identity to any known structural elements from the PDB. Finally, the 3D model of LcrV from the Angola strain was used as a structural template to generate models for all of the Y. pestis LcrV protein sequences analyzed in this study using the amino acid sequence to tertiary structure system (AS2TS) (Zemla et al., 2005). In all models, coordinates of side chain atoms were left unchanged for residues that are identical between the sequence of the modeled structures and the sequence of the protein in the crystal structure (1R6F). Side chain atom positions for the remaining residues were calculated using SCWRL (Canutescu et al., 2003).
The alignment of all LcrV sequences from Y. pestis known to date with those determined in this study resulted in the selection of the Angola strain of Y. pestis as a consensus sequence (Fig. 1). Four major “hot points” of the amino acid polymorphism were found at positions 18, 72, 273, and 324–326, which allowed us to classify the LcrV alleles into several types (Table 1). The reason that these positions were chosen as the major variable hotspots is that at least one of such substitutions will be repeated in two or more LcrV types. The type A (N18, R72, S273, and S324-G325-K326) group consisted of the Angola strain, as well as two strains of subspecies altaica (I-3455 and I-2359). All three strains of this type differed at a single position 113. In addition strain I-2359 contains two changes at the N-terminus (positions 2 and 3). The type B (N18, R72, C273, and S324-G325-K326) was restricted to a single strain of the subspecies ulegeica (I-2422). In comparison with the LcrV of the type A, the LcrV of this strain contained the substitution C273 in the “hot point” area, as well as three additional changes at positions 3, 7 and 84. The type C group (N18, R72, S273 and R324) included the strains A-1728 and C-582 of the subspecies hissarica and caucasica, respectively, which had a deletion of two amino acid residues at the C-terminus. The LcrV sequences of these two strains differed from each other at position 103. The type D (K18, K72, C273 and S324-G325-K326) group contained the LcrV from the epidemic strains of all three known biovars, such as antiqua (I-1996, I-2638, Antiqua), medievalis (KIM) and orientalis (CO92), as well as the strains of the subspecies ulegeica (I-2836, I-2487, I-2457), caucasica (C-715, C-540) and altaica (I-3519, I-2131). Therefore, both epidemic and non-epidemic strains (except subspecies hissarica) contained this type of LcrV. The LcrV sequences of the type D strains were 100% homologous at both nucleotide and amino acid levels. Finally, type E (K18, K72, C273 and R324) was represented by strains of the subspecies caucasica (1146, C-585, Pestoides F), xilingolensis strain 91001 isolated in China (Li et al., 2009; Song et al., 2004), altaica (I-3132, I-3088) and hissarica (A-1249). This LcrV type was not found in epidemic strains and in the isolates of the subspecies ulegeica. The LcrV sequence of type E had a deletion at the C-terminus identical to that of the strains of type C. Otherwise, the LcrV sequence of type E was the same as that of type D.
Since LcrV of the Angola strain of Y. pestis corresponded to the consensus sequence (Fig. 1), the 3D model of LcrV from this strain was constructed by using the crystal structure of this protein from Y. pestis KIM (Derewenda et al., 2004). According to the Structural Classification of Proteins (SCOP; release 1.71) (Murzin et al., 1995) the fold of the LcrV protein consists of an "all-alpha" domain, made mostly from the N-terminal region, and an "alpha+beta" domain. The domains are connected by an antiparallel coiled coil. Most of the LcrV amino acid polymorphisms, including those located in the regions that were not modeled, are found within the N-terminal region, residues 1–146. Inspection of 3D molecular models of LcrV showed that, with the exception of residue 273, all polymorphisms (72, 84, 103, 113, and 324) are located within the “all-alpha” domain (Fig. 2).
Sequence polymorphisms at positions 273 and 324 in the LcrV protein sequence occur in regions that are disordered in the crystal structure. Alignment of the LcrV of Y. pestis with its homologs from other bacterial species revealed that there are no conserved residues in the C-terminus region, 321–326, suggesting a limited structural importance of this area (Fig. 1). Although the residue at position 273 is not highly conserved, the internal loop (261–281) contained ten identical residues between all species, indicating a possible structural and functional significance of this region of LcrV. The sequence variations at positions 72, 84 and 103 probably have minimal structural consequences and lead to minor local perturbations of the structure.
The substitution of lysine for arginine at position 72 occurs on the surface of the protein and preserves the local inter-residue interaction. The substitution of leucine for isoleucine at position 84 does not occur at the protein surface; moreover, the side chain is mostly buried, but does preserve the local chemical properties and, likely, the inter-residue interactions. The substitution of glutamic acid with glycine at position 103 in the Y. pestis C-582 strain does not preserve the local chemical properties, though the mutation does occur at the surface of the protein in the middle of a helix. This mutation might destabilize the helix, but would not likely significantly disrupt the structure. Of the sequence substitutions we were able to model, the one most likely to have a serious impact on the local structure of the LcrV protein is the tryptophan to glutamic acid (strain I-2359) or glycine (strain I-3455) at position 113 (Fig. 3). Tryptophan is bulky and mostly hydrophobic, though it can contribute one hydrogen bond from the ε nitrogen. Tryptophan 113 of LcrV is mostly buried in a large hydrophobic region near the core of the N-terminal domain. The ε nitrogen stabilizes the local loop structure by hydrogen bonding to the backbone carbonyl of asparagine 110 (Fig. 3, panel A). When tryptophan is replaced by glutamic acid, the favorable hydrophobic interactions are mostly lost, though glutamate is somewhat ampipathic, and E113 can no longer hydrogen bond to the backbone carbonyl of N110. E113 might hydrogen bond with N144, which might, in turn hydrogen bond with the carbonyl of N110 to stabilize the local loop structure. When W113 is replaced with glycine, G113, all of the hydrophobic and hydrogen bonding interactions are lost, likely destabilizing the 108–112 loop and leading to a disordering of this loop, and possibly to some lesser stability of the whole protein (Fig. 3, panel C).
We cloned, expressed in E. coli, and purified LcrV of types A (Angola), E (Pestoides F) and D (CO92). When non-boiled samples were analyzed by SDS-PAGE, silver staining revealed a major band of around 40 kDa that corresponded to a monomeric form of LcrV with a deduced molecular weight of 41.1 kDa (Fig. 4A); however, we noticed that, in contrast to the LcrV of the Angola strain, the LcrV proteins from CO92 and Pestoides F contained additional minor bands of higher molecular weight. Immunoblot analysis performed by using an anti-LcrV monoclonal antibody confirmed that these bands belong to LcrV and likely represent multimeric forms of this protein (Fig. 4B). The diffused band of about 70 kDa was a possible product of interaction of LcrV monomer with the degraded forms of this protein that arose during LcrV purification. Since the immunoblot assay is generally more sensitive than silver staining, we could detect multimers of LcrV from Angola as well, although the intensity of these bands was significantly lower than that of the LcrV from the other two strains. Nevertheless, only LcrV of CO92 and Pestoides F could form a detectable band of approximately 160 kDa, likely corresponding to a tetrameric form of this protein. The LcrV proteins from CO92 and Pestoides F capable of forming a large amount of multimers are identical except for three terminal amino acids, while this part of the low multimer-forming LcrV from Angola is identical to that of CO92 (Fig. 1). Thus, these three last amino acids of the C-terminus are not involved in multimerization of LcrV, but other variables located at positions 18, 72 and 273 are likely responsible for the observed phenomenon.
Y. pestis, a highly lethal pathogen, is a recently evolved clone of enteropathogenic Y. pseudotuberculosis which has spread throughout the world (Achtman et al., 1999; Achtman et al., 2004; Li et al., 2009). The LcrV protein is a multifunctional virulence factor and protective antigen of the plague bacterium (Brubaker, 2003) which is generally conserved among the epidemic strains of Y. pestis (Adair et al., 2000; Motin et al., 1992). In contrast, LcrV sequences of enteropathogenic Yersinia spp. displayed a greater diversity (Motin et al., 1992; Roggenkamp et al., 1997), which can affect the functional activity of this protein (Sing et al., 2005). In addition to the epidemic strains of Y. pestis of the biovars antiqua, medievalis, orientalis (Devignat, 1951) and intermedium (Li et al., 2009), there are non-epidemic Y. pestis strains which have a limited virulence in selected animal models and for humans. These atypical isolates are referred to as the strains of microtus biovar (Li et al., 2009). Prior to this study, nucleotide sequences of only three lcrV genes that originated from this type of strains were available, such as the strains Angola and Pestoides F (Adair et al., 2000) as well as the strain 91001 (Song et al., 2004). The lcrV sequence of all of these isolates differed from the conserved lcrV characteristic of the epidemic strains. Therefore, we investigated the polymorphism of the lcrV among the strains of the different Y. pestis “subspecies” by sequencing this gene from the representative strains isolated in natural plague reservoirs located in the territory of the FSU and Mongolia.
We found that there are four major variable hotspots at positions 18, 72, 273, and 324–326, a finding which allowed us to classify the LcrV alleles into five types (A-E). This classification of the LcrV types accounted for the polymorphism from the consensus sequence and is now different from that assigned by us previously (Anisimov et al., 2007). We found that two of the altaica strains (I-3455 and I-2359) possessed the same type A allele as did the previously sequenced Angola strain, but this subspecies also possessed LcrV of two other types, D (I-2131, I-3519) and E (I-3088, I-3132). Similarly, the caucasica strains fell into different types of LcrV, one of which (C-582) was grouped with the hissarica strain (A-1728) of type C, while the other two (C-585 and 1146) belonged to the type E, and strains C-540 and C-715 had the D type of LcrV. The ulegeica strains also could be distinguished between type B (strain I-2422) and type D (strains I-2836, I-2487, I-2457). Thus, the strains of different Y. pestis “subspecies” can have the same type of LcrV and different types of LcrV can exist within the same natural plague reservoir.
The major result of the observed V antigen sequence heterogeneity was a finding that type D sequences characteristic of the epidemic strains of Y. pestis were widely distributed among microtus biovar isolates. This conclusion was confirmed by the sequencing of lcrV from additional isolates of subspecies ulegeica (2 strains), caucasica (10 strains) and altaica (12 strains) (data not shown). The strains of subspecies hissarica did not reveal this type of LcrV; however, only two isolates of this subspecies were available in our collection. We expect that the population of Y. pestis ssp. hissarica also contains the LcrV of type D, because the isolate A-1249 of this subspecies possessed LcrV of type E which is a truncated version of the type D sequence (see below). The existence of the LcrV of the type D in non-epidemic strains of Y. pestis was somewhat unexpected, since both previously sequenced strains of Y. pestis microtus biovar 91001 (Song et al., 2004) and Pestoides F (Adair et al., 2000; Garcia et al., 2007) had the type E sequence. The latter type of LcrV is a derivative of the type D sequence as a result of a single deletion involving two direct repeats ATGACACG at the C terminus of the protein (Adair et al., 2000; Song et al., 2004). The Y. pestis subspecies strain Angola, which does not have this C-terminal deletion, possessed a significantly distant sequence of LcrV different from the type D variant in three out of four major variable hotspots (positions 18, 72, 273). Thus, all lcrV genes sequenced to the present from epidemic strains of Y. pestis, including those from ancient Siberian foci in Russia and Mongolia, such as I-1996, I-2636 and eight more isolates (data not shown), were restricted to type D. This finding is in good agreement with the clonal nature of the origin of the plague bacterium and its spread over the globe during major pandemics (Achtman et al., 1999; Achtman et al., 2004; Devignat, 1951; Li et al., 2009). On the other hand, the LcrV from Y. pestis subspecies exhibited all of the varieties of the sequences (types A-E), which indicated these strains were the earliest in origin of Y. pestis pathogen and had more time for still continuing microevolution or such mutations were neutral in the course of adaptation to the specific host, which is a vole. Therefore, the phenomenon of “selective virulence” characteristic of the strains of the microtus biovar is unlikely to be the result of polymorphism of the V antigen.
We tried to predict whether the sequence polymorphism of LcrV detected in the strains of Y. pestis of different origin resulted in a significant disturbance of the 3D structure. The crystal structure of LcrV from Y. pestis KIM has several missing regions not visible in the electron density maps (Derewenda et al., 2004), and thus a more complete model was generated using LGA (Zemla, 2003). Unfortunately, it was not possible to model the missing N-terminus (residues 1–27) due to the lack of similar structural elements in the PDB, so the sequence polymorphisms occurring in this region were not analyzed. This region is essential for LcrV secretion and virulence (Broms et al., 2007; Skrzypek & Straley, 1995) and contains protective epitope(s) that are sufficient for partial protection (Pullen et al., 1998). The significance of the variation of the LcrV observed in this region for strains I-2359 and I-2422 remains to be determined experimentally. The alignment of all V antigens of Y. pestis produced the consensus sequence that was identical to that of the Angola strain, which belongs to the group of atypical, evolutionary ancient strains of the plague bacteria (Achtman et al., 2004; Anisimov et al., 2004; Li et al., 2009; Motin et al., 2002; Radnedge et al., 2002). Therefore, a 3D model of LcrV was made for the primary sequence of this strain using homology-based modeling system AS2TS (Zemla et al., 2005), and a possible structural difference due to each individual amino acid residue substitution was evaluated.
A total of six LcrV polymorphisms were structurally analyzed by comparing the modeled structures of LcrV from all available Y. pestis strains. All changes except one occurred either in flexible regions or on the surface of the protein and conserved the local chemical properties (i.e. those of a hydrophobic, hydrophilic, amphipathic, charged nature) across all the strains. These variations would likely have a limited impact on the structure. Polymorphisms in flexible and surface regions are likely subject to less selective pressure. In contrast, the substitution of tryptophan at position 113 with either glutamic acid or glycine likely has a serious impact on the regional structure of the protein, and the mutations found in Y. pestis I-2359 and I-3455 might have an effect on the function of LcrV. In fact, the recombinant LcrV from I-3455 produced in Escherichia coli cells demonstrated a dramatically induced ability to agglomeration when compared with other LcrV variants which have a tryptophan residue at position 113 (Kopylov, P.Kh., Kiseleva, N.V., personal communication). Moreover, LcrV contains major murine H-2d-, H-2k- and H-2b- restricted T-cell epitopes in the area of 102–121 (Parent et al., 2005; Shim et al., 2006; von Delwig et al., 2005), suggesting that the mutations at positions 104 and 113 may affect the immunogenicity of this antigen.
In addition to the structural comparisons, we analyzed whether the variations in the LcrV sequence might be located in the areas known to be important for established properties of this antigen. A major protective epitope(s) of LcrV resides internally between amino acids 168 and 275 (Motin et al., 1994), the major antigenic region is located between the residues 135 and 245 (Hill et al., 1997) and the smallest protective fragment is comprised of amino acids 135–262 (Vernazza et al., 2009). The combined region (residues 135–275) is conserved between the LcrV proteins of Y. pestis, except for the variable position 273. Since the epitope mapping described in these studies was not precise, the significance of the variation of cysteine/serine at this position in terms of its effects on the immunogenic properties of V antigen could not be addressed at that time. However, a detailed peptide mapping of LcrV conducted later (Khan et al., 2008; Parent et al., 2005; Pullen et al., 1998; Shim et al., 2006) has revealed that the cysteine-containing internal loop (261–281) does not possess immunodominant linear T- and B-cell epitopes. Nevertheless, the region 271–300, which includes the polymorphic position 273 is crucial for the immunomodulatory property of LcrV (DeBord et al., 2006) which depends on the suppression of proinflammatory cytokines by inducing Interleukin 10 (IL-10) (Brubaker, 2003; Nedialkov et al., 1997; Reithmeier-Rost et al., 2004; Sing et al., 2002a). The N-terminal region of LcrV (residues 31–57), shown to be important for toll-like receptor 2 (TLR2)- and IL-10-stimulating activities in Yersinia enterocolitica O:8 ( Sing et al., 2002b; Sing et al., 2005) was conserved in LcrV of Y. pestis as determined by us. Also conserved were two sites of interaction of LcrV with human TLR2 (residues 32–35 and 203–205) and putative sites of binding with CD14 receptor (residues 41–43 and 95–97) (Abramov et al., 2007). Interestingly, there were no polymorphic changes in a hypervariable region (residues 225–232) found between the LcrV sequences of Yersinia spp. (Motin et al., 1992; Roggenkamp et al., 1997). Finally, none of the variations occurred at positions that are conserved between LcrV protein sequences from Y. pestis and LcrV homologs found in other bacterial species (Fig. 1).
The LcrV is a key component of the translocatory machine providing injection of Yop effectors into the host-cell target which forms a pentamer at the tip of the T3SS needle (Broz et al., 2007; Mueller et al., 2005). Therefore, the ability of LcrV to form multimers is an essential property of this protein. It is a well-established fact that recombinant LcrV may exist in solution as a dimer and higher order oligomers (Derewenda et al., 2004; Hamad & Nilles, 2007; Lawton et al., 2002). Moreover, a controlled refolding of this protein led to the assembly of LcrV into oligomeric doughnut-like complexes, and the C-terminal helix α12 (a.a. 279–317) was crucial for this in vitro oligomerization (Caroline et al., 2008). Previously, it was established that the helix α7 (a.a. 148–182) is involved in oligomerization as well (Lawton et al., 2002). We observed that recombinant LcrV from CO92 and Pestoides F had a significantly better capacity for oligomerization than did that from the Angola strain. However, the α7 and α12 helixes of all three strains are identical. The polymorphism of the three C-terminal residues also cannot explain the differences in the level of multimer formation, since high-oligomeric LcrV of CO92 and low-oligomeric LcrV of Angola have the identical C-terminus, which is different from that of the high-oligomeric LcrV of Pestoides F. Therefore, the rest of polymorphic residues located at positions 18, 72 and 273 should account for the differences in oligomerization. Further studies using site-directed mutagenesis will be necessary to determine which of these residues (or combination of thereof) is essential for the observed phenomena. The oligomers of LcrV were stable at denaturing conditions of SDS-PAGE and could withstand 10 min thermal denaturation at 95°C in sample buffer in the presence of the reducing agent DTT at a concentration of 100 mM (data not shown). Similar properties of LcrV dimers and tetramers were observed for LcrV expressed in DNA vaccine constructs. Interestingly, the ability of LcrV to form oligomers correlated with the efficiency of protection of mice from Y. pestis challenge followed by the DNA vaccination (Wang et al., 2004). Therefore, an understanding of the mechanism underlying a formation of LcrV multimers might be important for plague subunit vaccine optimization, especially, when truncated protective fragments of this antigen are used (Vernazza et al., 2009). It was shown recently, that aggregates and high-molecular-weight multimers (larger than the dimer and tetramer forms) of LcrV possessed TLR2-stimulating activity (Pouliot et al., 2007), which perhaps is involved in the virulence-associated, immunomodulating property of LcrV via induction of IL-10 signaling through the TLR2/TLR6/CD14 complex (Brubaker, 2003; Depaolo et al., 2008). Although the formation of large LcrV multimers has to be demonstrated in vivo, one can speculate that the observed polymorphism of LcrV provided a basis for the natural selection of more virulent variants of Y. pestis during the evolution of this pathogen.
In summary, this study demonstrated that the LcrV proteins from the strains representing a different “subspecies” of Y. pestis displayed the size, sequence and 3D structure polymorphism. These variations in the LcrV apparently did not alter the lethality of these strains in mice and their natural hosts, since these atypical Y. pestis isolates were reported to be highly virulent for these animal species (Anisimov et al., 2004; Zhou et al., 2004b). Nevertheless, we showed for the first time that LcrV derived from different isolates of Y. pestis can vary in their ability to form multimers in vitro, and the polymorphic changes responsible for this property are located outside of the α7 and α12 helixes previously assigned as essential for oligomerization (Caroline et al., 2008). The impact of the variations of LcrV antigen on multimerization and their influence on virulence and protective properties will need to be further investigated.
This work was supported by the International Science and Technology Center (project #2426), Russian Foundation for Basic Research (project 08-04-00405-a), Sealy Center for Vaccine Development of the University of Texas Medical Branch at Galveston, and the NIH/NIAID grant 1R43 AI71634-01A2. This work was implemented under the auspices of the U.S. Department of Energy under contract no. W-7405-Eng-48.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
(A part of this work was presented at the 9th International Symposium on Yersinia, October 10–14, 2006, in Lexington, Kentucky, USA) (Anisimov et al., 2007).