Correspondence: Roy Curtiss III, Center for Infectious Diseases and Vaccinology, The Biodesign Institute, School of Life Sciences, Arizona State University, PO Box 875401, 1001 S. McAllister Avenue, Tempe, AZ 85287-5401, USA. Tel.: +1 480 727 0445; fax: +1 480 727 0466; e-mail: email@example.com
Yersinia pestis PsaA is an adhesin important for the establishment of bacterial infection. PsaA synthesis requires the products of the psaEFABC genes. Here, by prediction analysis, we identified a PsaA signal sequence with two signal peptidase (SPase) cleavage sites, type-I and type-II (SPase-I and SPase-II). By Edman degradation and site-directed mutagenesis, the precise site for one of these Spase-I PsaA cleavage sites was located between alanine and serine at positions 31 and 32, respectively. Yersinia pestis psaA expression and the role of the PsaB and PsaC proteins were evaluated in recombinant attenuated Salmonella Typhimurium vaccine strains. PsaA was detected in total extracts as a major 15-kDa (mature) and 18-kDa (unprocessed) protein bands. PsaA synthesis was not altered by a ΔA31–ΔS32 double-deletion mutation. In contrast, the synthesis of PsaA (ΔA31–ΔS32) in Y. pestis and delivery to the supernatant was decreased. Otherwise, substitution of the amino acid cysteine at position 26 by valine involved in the SPase-II cleavage site did not show any effect on the secretion of PsaA in Salmonella and Yersinia. These results help clarify the secretion pathway of PsaA for the possible development of vaccines against Y. pestis.
The Yersinia are Gram-negative bacteria with 11 species including the gastrointestinal pathogens Yersinia pseudotuberculosis and Yersinia enterocolitica, and the systemic pathogen Yersinia pestis, which is typically fatal without treatment. Genetic and whole-genome studies indicate that Y. pestis is closely related to Y. pseudotuberculosis. In contrast, Y. enterocolitica is only distantly related to Y. pestis and Y. pseudotuberculosis, displaying a more variable genomic arrangement (Achtman et al., 1999). Yersinia pestis is the etiological agent of plague in humans (Perry & Fetherston, 1997) and a recently recognized re-emerging disease. The widespread aerosol dissemination combined with high mortality rates make Y. pestis a deadly pathogen (Inglesby et al., 2000).
The signal peptide of Y. enterocolitica MyfA was identified (Iriarte et al., 1993) and needs to be determined for PsaA in both Y. pseudotuberculosis and Y. pestis. In bacteria, a signal peptide present on proteins that are destined to be secreted or to be membrane components, it is usually present at the amino terminal and absent from the mature protein. The signal peptide is removed by signal peptidases (SPases) as an SPase-I or SPase-II (processing of prolipoproteins) (Yamaguchi et al., 1988; Tuteja, 2005).
Recently, a new generation of improved recombinant attenuated Salmonella Typhimurium vaccine (RASV) strains, such as Salmonella enterica serovar Typhimurium χ9558, have been developed and tested using heterologous antigens (Li et al., 2009). These RASV strains will facilitate investigations into the role of selected amino acids in the biogenesis of Y. pestis PsaA. The focus of this present study is a better understanding of the PsaA translocation process and improvement of its secretion, with the eventual goal of developing a subunit vaccine against Y. pestis.
Materials and methods
Bacterial strains and media
Escherichia coli, Salmonella, Y. pestis strains and plasmids used in this study are listed in Table 1. Escherichia coli and Salmonella strains were grown in Luria–Bertani (LB) medium (1% Bacto tryptone, 1% NaCl, 0.5% yeast extract), 1.5% LB agar or on McConkey (Difco); when required, the medium was supplemented with 50 μg mL−1 ampicillin, 10 μg mL−1 nalidixic acid, 0.2% mannose or 50 μg mL−1 diaminopimelic acid for growing the strain with ΔasdA mutation.
DNA manipulations were carried out as described by Sambrook & Russell (2001). All primers (Integrated DNA Technology) were flanked with restriction enzymes (uppercase in the primer sequences), as shown in Supporting Information, Table S1. The psaEFABC genes were amplified by PCR from Y. pestis KIM6+ strain chromosome, and constructions were verified by DNA sequencing (Arizona State University Facilities). Fifteen codons from Y. pestis psaA were substituted with the most frequently used codons found in Salmonella genes for optimization of Y. pestis psaA expression in RASV strains. All amino acid substitutions and deletions in Y. pestis psaA were performed using a Quick-Change site-directed mutagenesis kit (Stratagene). The presence of a desired mutation was verified by DNA sequencing (Fig. 1a, Table 1).
Expression of psaABC in E. coli and Y. pestis
The recombinant PsaA-AU1-6XHis protein was overexpressed in E. coli strain LMG194, transformed with the pYA3883 (Table 1) and grown in 1 × minimal salts media (Curtiss, 1965), supplemented with 0.2% (v/v) glycerol, 2% casamino acids, 50 μg mL−1 thiamine and 100 μg mL−1 ampicillin at 37 °C to an OD600 nm of 0.5, and then induced with arabinose at 0.05% for 5 h. Yersinia pestis harboring a different psaA expression pYA4787 plasmid derivative was grown in 3 mL of heart infusion broth at 28 °C until an OD600 nm of 0.5, and then centrifuged. The pellet was then resuspended in 100 μL of brain heart infusion broth and inoculated into 3 mL of brain heart infusion broth with 0.5% yeast extract pH 6 and grown at 37 °C for 8 h.
The recombinant PsaA-AU1-6XHis protein was purified by nickel-nitrilotriacetic acid agarose chromatography under denaturing conditions. Protein concentration was determined using the Bradford assay with bovine serum albumin as the standard. The PsaA-AU1-6XHis purified protein was used to immunize rabbits for production of PsaA polyclonal antibody.
Cell fractionation and immunoblots
Cell fractions were prepared from 1 mL of culture using the PeriPreps periplasting Kit (Epicentre) following the manufacturer's instructions. To isolate proteins released into the culture supernatants, 1 mL of each sample was filtered (0.22-mm pore size, Corning), and precipitated with 10% trichloroacetic acid, then pelleted by centrifugation and resuspended in 100 μL of LDS sample buffer. Each 10 μL fraction was separated on a 4–12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (NuPAGE Bis-Tris, Invitrogen) and transferred to nitrocellulose sheets (Bio-Rad). The recombinant protein was immunolocalized using rabbit anti-PsaA serum (1 : 15 000) followed by alkaline phosphatase-conjugated anti-rabbit immunoglobulin G (Sigma). The rabbit anti-AU1 epitope tag (1 : 5000) (Bethyl) was used to monitor the eluted fractions during the purification procedure of PsaA-AU1-6XHis (Jenson et al., 1997) (data not shown). All experiments were performed in triplicate.
Amino acid sequence of the N-terminal region of Y. pestis PsaA protein
The PsaA protein from Y. pestis P325 transformed with pYA4788 (Table 1) was isolated from the periplasmic fraction using the PeriPreps periplasting Kit (Epicentre) by cutting the identified band from a polyvinylidene fluoride membrane (Invitrogen) after separation and transfer from an SDS-PAGE gel. The Edman (1960) degradation method was used to determine the amino-terminus sequence of the mature PsaA in two independent experiments (Arizona State University Facilities).
Determination of the cleavage site of Y. pestis PsaA
PsaA is predicted to be a 163 amino acid protein with an estimated molecular mass of 17.93 kDa. Sequence analysis of PsaA with the computer algorithm signalp 3.0, lipop 1.0 and dolop (Bendtsen et al., 2004; Babu et al., 2006) predicted a prokaryotic signal sequence at its amino-terminal region with a potential SPase-I cleavage site between alanine at position 31 and serine at position 32 (ANA▾S+1T+2) with an expected mature form of 14.6 kDa. In addition, a putative SPase-II cleavage site was identified between the alanine at position 25 and the cysteine at position 26 (IAA▾C+1G+2) with an expected PsaA mature form of 15.1 kDa. To identify the cleavage site in PsaA, the first 10 amino acids of the mature PsaA form were determined by Edman degradation and matched exactly from serine at position 32 to serine at position 41 of Y. pestis PsaA, indicating that the cleavage site is located between alanine 31 and serine 32 and it is recognized by the SPase-I (Fig. 1a).
Validation of the amino acid residues involved in the SPase sites of Y. pestis PsaA
The substitution of amino acid residues involved in the SPase-I and SPase-II sites of Y. pestis PsaA was evaluated in the Y. pestis P325 strain, which does not express the PsaA unless the strain is complemented with the pYA4788 harboring the psaEFABC locus (Fig. 2a). When the PsaA amino acid alanine 31 (pYA4792) or serine 32 (pYA4793) involved in SPase-I cleavage site was deleted, the PsaA was observed in all subcellular fractions (data not shown). However, secretion of the PsaA ΔA31–ΔS32 double deletion (pYA4794) was drastically decreased in the supernatant fraction (Fig. 2b, lane 4). In contrast, when the cysteine 26 involved in the SPase-II recognition site was changed by valine, PsaA synthesis and secretion were not affected (data not shown); similarly, substitution of cysteine at position 10 by valine (pYA4789) or C10V–C26V double substitution (pYA4791) (Fig. 2b, lane 3) did not affect PsaA synthesis and its secretion. These data confirm that the predicted SPase-I cleavage site is located between alanine 31 and serine 32.
PsaA synthesis and secretion to the extracellular medium in RASV strain
The RASV χ9558 strain was described in detail by Li et al. (2009). This Salmonella strain contains the deletion–insertion mutation ΔrelA198∷araC PBADlacI TT. This mutation encodes for arabinose-regulated LacI synthesis, which regulates the expression from Ptrc. The χ9558 strain was transformed with the plasmid pYA3705 and the expression of psaA under Ptrc was analyzed in LB 0.2% or without arabinose at 37 °C until an OD600 nm of 0.8. The PsaA protein had an unprocessed form of 18 kDa and a mature form of 15 kDa in total cell extracts of both growth conditions, with an expected reduction in synthesis with arabinose due to the lacI repressor gene expression. In the periplasmic fraction and in the culture supernatant PsaA was detected as a mature 15-kDa protein (Fig. 3a). The detection of PsaA in the supernatant was a result of its secretion, as the detection of the cytoplasmic protein σ70 was only observed in the total extract and not in the supernatant (Fig. 3b). The PsaA synthesis profile with psaA-optimized (pYA3705) was similar to the psaA wild type (pYA3704) (Fig. 4a, lanes 7 and 8). To analyze the synthesis and secretion of Y. pestis PsaA in χ9558 strain, growth was compared using 0.2%, 0.02% and 0.0% arabinose in the culture medium. Synthesis was found to be proportional at these different concentrations and we report the results without arabinose.
Function of Y. pestis PsaB and PsaC in RASV strain
The synthesis of the PsaB chaperone protein was required for export of PsaA from the cytoplasm to the periplasmic space (pYA4798), but the secretion of PsaA to the supernatant was almost undetectable (Fig. 4a, lane 5). The detection of PsaA coexpressed with PsaC (pYA4799) (Fig. 4a, lane 6) was sparsely localized to the membrane fraction. Conversely, when a set of 16 amino acid residues was added in frame at the carboxy terminal of PsaA (pYA4797), the product of psaA minus psaBC was detected as a ∼19-kDa protein band in total extracts (Fig. 4a, lane 4), barely delivered in the supernatant and was in an insoluble form in the membrane fractions. When a set of eight amino acid residues was added in frame at the carboxy-terminal end of PsaA without PsaBC (pYA4796), PsaA was not detected (data not shown). These results indicate that PsaB and PsaC are essential for the processing and translocation of PsaA from the cytoplasm to the cell surface in Salmonella.
Role of the amino- and carboxy-terminal regions of Y. pestis PsaA in the RASV strain
Deletion of the first 26 amino acids of PsaA amino-terminal region (pYA3711) prevented the translocation of PsaA from the cytoplasm to the cell surface. The unprocessed PsaA form was not observed and the mature 15-kDa protein was decreased in the total extract, cytoplasm and membrane fractions (Fig. 4b, lane 1). Deletion of the last nine amino acids of PsaA at the carboxy-terminal region (pYA4800), from threonine at position 155 to phenylalanine at position 163, drastically decreased its expression and was barely detectable as a ∼13.5-kDa product in supernatant and membrane fraction (Fig. 4a, lane 9). These results indicate that the amino-terminal region is necessary to secrete PsaA and that the carboxy-terminal region is required for its stability.
Substitution and deletion on PsaA of the amino acids involved in the SPase cleavage sites in the RASV strain
Deletion of the PsaA A31 (pYA4374) or S32 (pYA4375), which forms part of the SPase-I cleavage site, did not affect the synthesis or secretion of PsaA in any subcellular fraction (Fig. 4b, lanes 7 and 8), but with the ΔA31–ΔS32 double deletion (pYA4376), the unprocessed 18-kDa product was not detected in the total extract and barely observed in the membrane fraction (Fig. 4b, lane 9). In contrast, when the amino acids involved in the PsaA predicted SPase-II cleavage site, cysteine at position 26 changed to valine (pYA3708) and the glycine at position 27 replaced by serine (pYA3709), the PsaA synthesis was not affected (Fig. 4b, lanes 2 and 5). To determine whether the cysteine residues at positions 10 and 26 play a role in the PsaA biogenesis and stability, the cysteine10 (pYA3707) was replaced with valine and either cysteine was changed to valine (pYA3706). None of these mutations affected PsaA synthesis or secretion (Fig. 4b, lanes 4 and 6). We observed the same expression profile when the RASV strain containing each of the previously described plasmids was grown with either 0.2% or 0.02% arabinose in the culture medium (data not shown). The amino acid substitution of the putative glycosylation site, asparagine 30 to leucine (pYA3710) produced a shorter unprocessed ∼17-kDa PsaA (Fig. 4b, lane 3). These results indicate that in the absence of either A31or S32, other amino acids flanking this SPase-I cleavage site can generate alternative cleavage sites, but deletion of both A31 and S32 produces a new cleavage site, which is processed more efficiently than the original. However, the C10V and C26V or double substitution (C10V–C26V) and G27S had no effect on the translocation of PsaA to the cell surface.
In this study, we elucidated the role in secretion and biogenesis of the Y. pestis PsaA amino- and carboxy-terminal regions.
Using different computer analyses we identified two putative SPase cleavage sites in the PsaA signal sequence, with their tripartite consensus regions: n-, a positively charged amino terminus; h-, a hydrophobic core; and c-, terminal cleavage site. In Gram-negative bacteria the lipoproteins are anchored to either the inner or the outer membrane and an aspartic acid residue at position +2 (D+2) is proposed to determine the final destination of the lipoproteins (Yamaguchi et al., 1988). The D+2 substitution to amino acid residues such as phenylalanine, tryptophan, tyrosine, glycine and proline maintains the retention of the lipoprotein to the periplasmic face of the cytoplasmic membrane (Seydel et al., 1999). The glycine at position 27 is the amino acid +2 in the Y. pestis PsaA putative SPase-II cleavage site, and substitution of the amino acids from this cleavage site, such as C26V (pYA3708) and G27S (pYA3709), did not show any effect on the translocation process of PsaA, nor did the substitution C10V (pYA3707) or double-substitution C10V–C26V (pYA3706). Further studies using electron microscopy will be required to determine whether the PsaA structure and its assembly into multisubunit protein polymers are affected by the mutations on PsaA cysteine residues. Surprisingly, the substitution of the hydrophilic asparagine at position 30 to the hydrophobic leucine generated a shorter unprocessed PsaA form, but the mature PsaA form did not change. The asparagine at position 30 forms part of the putative glycosylation consensus sequence, N-X-S/T, where X can be any amino acid except proline (Fig. 1a) (Gavel & von Heijne, 1990). However, to date no N-glycosylation system has been reported in Salmonella or Yersinia (Upreti et al., 2003). In our analysis, the mechanism by which the substitution of N30L that generates the shorter unprocessed form of PsaA remains to be clarified. With the deletion of either A31 or S32 or both, alternative cleavage sites could be generated among the flanking amino acid residues such as asparagine, serine and threonine with similar properties (polar, hydrophilic and neutral). Surprisingly, the PsaA with the SPase-I cleavage site derived by the ΔA31–ΔS32 double-deletion mutations was more efficiently secreted in Salmonella, but in Yersinia it impaired the secretion of PsaA to the supernatant, indicating a different affinity for the SPase-I cleavage site between Salmonella and Yersinia.
Two highly conserved regions were observed between the amino acid sequence of PsaA and its counterpart MyfA in Y. enterocolitica, one at the amino-terminal region and the second at the carboxy-terminal region (Fig. 1b). Similar regions play a role in the interaction with the chaperone and are therefore essential for the biogenesis of the protein (Kuehn et al., 1993; Soto et al., 1998). Here we show that when nine alternating hydrophobic/hydrophilic residues are removed from the carboxy-terminal end of PsaA, the PsaA synthesis is drastically affected even with the coexpression of the chaperone and usher protein PsaB/PsaC. Although additional experiments are required to validate this result, this suggests that this PsaA region is essential for its biogenesis.
The coexpression of PsaA with the PsaB chaperone protein allowed the detection of PsaA in the cytoplasm, periplasm and membrane fraction. The role of PsaB in the cytoplasm possibly to avoid the degradation of PsaA and the cytoplasmic interactions between PsaA/PsaB still needs to be established. In contrast, the coexpression of PsaA with only PsaC resulted in a lack of detection of PsaA, confirming that the interaction of PsaABC proteins is essential in the secretion process of PsaA.
These results will help to provide new design strategies for delivery of PsaA in RASV strains using their own secretion pathway. Combined with a new more efficient SPase-I cleavage site, these strategies should aid in improving RASV for the effective delivery Y. pestis antigens.
We are thankful to Dr J.D. Fetherston (University of Kentucky, Lexington) who generously provided the Y. pestis P325 strain. We also thank Dr Clara Espitia (UNAM, Mexico) for her critical reading of the manuscript and Dr David S. Reiner (Burnham Institute for Medical Research) and Isabel Perez Montfort (UNAM, Mexico) for correction of the English version of this manuscript. This research was supported by the National Institutes of Health, grant AI057885.