Genome plasticity in Yersinia pestis

Lyndsay Radnedge1, Peter G. Agron1, Patricia L. Worsham2 and Gary L. Andersen1

Biology and Biotechnology Research Program, Lawrence Livermore National Laboratory, L-441, 7000 East Avenue, Livermore, CA 94550, USA1
United States Army Research Institute of Infectious Diseases, Fort Detrick, MD 21702, USA2

Author for correspondence: Gary L. Andersen. Tel: +1 925 423 2525. Fax: +1 925 422 2282. e-mail: Andersen2{at}LLNL.GOV


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Yersinia pestis, the causative agent of bubonic plague, emerged recently (<20000 years ago) as a clone of Yersinia pseudotuberculosis. There is scant evidence of genome diversity in Y. pestis, although it is possible to differentiate three biovars (antiqua, mediaevalis or orientalis) based on two biochemical tests. There are a few examples of restriction fragment length polymorphisms (RFLPs) within Y. pestis; however, their genetic basis is poorly understood. In this study, six difference regions (DFRs) were identified in Y. pestis, by using subtractive hybridization, which ranged from 4·6 to 19 kb in size. Four of the DFRs are flanked by insertion sequences, and their sequences show similarity to bacterial genes encoding proteins for flagellar synthesis, ABC transport, insect toxicity and bacteriophage functions. The presence or absence of these DFRs (termed the DFR profile) was demonstrated in 78 geographically diverse strains of Y. pestis. Significant genome plasticity was observed among these strains and suggests the acquisition and deletion of these DNA regions during the recent evolution of Y. pestis. Y. pestis biovar orientalis possesses DFR profiles that are different from antiqua and mediaevalis biovars, reflecting the recent origins of this biovar. Whereas some DFR profiles are specific for antiqua and mediaevalis, some DFR profiles are shared by both biovars. Furthermore, the progenitor of Y. pestis, Y. pseudotuberculosis (an enteric pathogen), possesses its own DFR profile. The DFR profiles detailed here demonstrate genome plasticity within Y. pestis, and they imply evolutionary relationships among the three biovars of Y. pestis, as well as between Y. pestis and Y. pseudotuberculosis.


View this table:
[in this window]
[in a new window]
 
Table 1. Characteristics of DFRs on the Y. pestis CO92 genome

 
Keywords: bacterial genome, comparative genomics, subtractive hybridization, Yersinia pseudotuberculosis

Abbreviations: DFR, difference region; RFLP, restriction fragment length polymorphism; SSH, suppression subtractive hybridization

The GenBank accession numbers for the sequences reported in this paper can be found in Table 1; the GenBank accession number for DFR4 is AF426171.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Yersinia pestis, the causative agent of bubonic plague, possesses a genome that is highly conserved among different isolates. Y. pestis is predominantly a pathogen of rodents, but it is transmitted into the bloodstream of humans via flea bites. Y. pestis has been classified into three biovars (antiqua, mediaevalis and orientalis), based on the ability of strains to ferment glycerol and to reduce nitrate. It has been proposed that antiqua strains are derived from those responsible for the plague pandemic of the sixth century and that mediaevalis strains are responsible for the Black Death of the 14th century; the current pandemic has been associated with the most-recently evolved biovar, orientalis (Devignat, 1951 ).

Several studies have demonstrated a lack of diversity between different strains of Y. pestis. Different strains cannot be classified by differences in serotypes or in phagetypes and, at the genetic level, there are only two reported sequence polymorphisms in seven loci examined in 58 strains (Achtman et al., 1999 ; Adair et al., 2000 ). Moreover, the genome of Y. pestis is highly related to that of Yersinia pseudotuberculosis (an enteric pathogen transmitted by the faecal–oral route), based on hybridization studies (Bercovier et al., 1980 ) and 16S rRNA sequences (Trebesius et al., 1998 ). These observations led to the proposal that Y. pestis is a recent, highly uniform clone of Y. pseudotuberculosis that arose between 1500 and 20000 years ago (Achtman et al., 1999 ). Some genome plasticity was demonstrated by restriction fragment length polymorphism (RFLP) analysis of IS100 locations, enabling the construction of a phylogenetic tree in which the different biovars of Y. pestis are clustered together (Achtman et al., 1999 ; McDonough & Hare, 1997 ). Similarly, PFGE analysis of SpeI fragments of Y. pestis genomic DNA shows no variability within biovars, but some variation is seen between biovars (Lucier & Brubaker, 1992 ). Another study of variable number tandem repeats (VNTRs), which are regions with potentially high variability, has shown these regions to be a useful tool for strain discrimination, but, in contrast to RFLP analysis, the VNTR alleles were not biovar specific (Adair et al., 2000 ). Small genomic differences between different isolates of Y. pestis can also be demonstrated by ribotyping (Guiyoule et al., 1994 , 1997 ).

As the nucleotide sequences of more bacterial genomes are uncovered, it is becoming increasingly apparent that they are composed of DNA sequence mosaics, some of which are acquired by lateral (or horizontal) gene transfer (Jain et al., 1999 ; Lawrence & Ochman, 1998 ; Ochman et al., 2000 ; Perna et al., 2001 ). Large genomic differences that have arisen from lateral gene transfer events commonly originate from mobile genetic elements, such as transposons, insertion elements or bacteriophage integration. The excision and acquisition of large genomic fragments quickly generate new strain variants, with the balance between the two processes maintaining the genome size (Lawrence & Ochman, 1998 ). There is now evidence for genome plasticity in Y. pestis CO92, since three large genomic rearrangements are seen – one translocation and two inversions – which apparently arise during growth of the organism (Parkhill et al., 2001 ). Furthermore, a large unstable fragment has been characterized – the 102 kb pgm locus – within which lies the high-pathogenicity island (Buchrieser et al., 1999 ). Instability of the pgm locus is mediated by a recombination event between two copies of IS100 that flank the locus (Fetherston et al., 1992 ). Insertion elements are also found to be unusually abundant in the genome of Y. pestis CO92 and are found at the boundaries of the three variable regions in this strain (Parkhill et al., 2001 ).

In this study, suppression subtractive hybridization (SSH) (Diatchenko et al., 1996 ; Akopyants et al., 1998 ) was used to identify genomic differences between different strains of Y. pestis. Here we report the identification of six difference regions (DFRs) that are variable between different isolates of Y. pestis, four of which contain insertion elements. Five of the DFRs are present in the genome of Y. pestis CO92 (biovar orientalis) (Parkhill et al., 2001 ), but are absent in some isolates. One DFR is absent in Y. pestis CO92, but is present in other strains, providing evidence that both the acquisition and excision of large genomic segments in Y. pestis have contributed to the evolution of its genome.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains.
The bacterial strains examined in the comparisons using SSH were Y. pestis D14 Salazar (orientalis), Y. pestis D15 Yokohama (antiqua), Y. pestis D27 KIM (mediaevalis), Y. pestis D46 KIM10 (mediaevalis), Y. pestis Antiqua (antiqua) and Y. pestis CO92 (orientalis) (Table 1). The first four strains were generous gifts from Dr Robert Brubaker (Michigan State University) and the last two strains were from the USAMRIID culture collection. DNA from the additional strains tested for difference products is listed in Table 2, and this DNA was generously provided by the following colleagues: strains 1, 2, 11, 14, 30, 31, 39, 45, 46, 56, 59 and 72 were from the CDC collection, courtesy of Dr May Chu; strains 5–10 were from the Dept of Health Services, Berkeley, CA, courtesy of Dr Will Probert; strains 18, 19, 20, 21, 41, 42, 51, 60, 61, 68, 70 and 79 were courtesy of Dr Robert Brubaker. Strains 80–82 were purchased from the American Type Culture Collection (ATCC), Manassas, VA. DNA from the remaining strains was supplied by the USAMRIID collection.


View this table:
[in this window]
[in a new window]
 
Table 2. The appearance of each DFR in 78 strains of Y. pestis, and in four isolates of Y. pseudotuberculosis

 
SSH.
The method of SSH used here is essentially as described by Akopyants et al. (1998) . SSH identifies DNA sequences (termed difference products) that are specific to one genome (designated the tester) and which are absent in another genome (designated the driver). Briefly, genomic DNA from the tester and driver are cut with a frequently cutting restriction enzyme (RsaI). After ligation of the tester DNA to oligonucleotide adaptors and hybridization with restricted driver DNA, SSH uses PCR amplification to enrich for unique segments of restricted tester DNA and simultaneously limits non-target amplification by suppression PCR. Difference products were cloned into the pGEMT-Easy vector (Promega) and transformed into 50 µl of competent DH5{alpha} cells (Life Technologies). DNA from clones containing putative tester-specific difference products was purified using magnetic beads (Skowronski et al., 2000 ) and sequenced on an ABI 3700 automated sequencer (Applied Biosystems). The resulting data were analysed using the ABI Sequencing Analysis software (version 3.2), and assembled and edited using Phred, Phrap (Ewing & Green, 1998 ; Ewing et al., 1998 ) and Consed (version 7.0; Gordon et al., 1998 ).

Analysis of tester-specific sequences.
Oligonucleotide primers were designed using the putative tester-specific sequences and were supplied by Genosys. The primers were designed either manually or by using Consed (Gordon et al., 1998 ); they had a melting temperature of >60 °C. The primers were initially screened against genomic DNAs prepared from both the tester and the driver. To determine whether a primer pair was tester-specific, 75 pg of the tester and the driver DNA was used as template in PCR reactions using the following parameters: 94 °C (15 s), 65 °C (15 s) and 72 °C (30 s) for 27 cycles. The products were visualized on a 1·5% agarose gel run in 0·5xTBE. If a PCR product was present when tester DNA was used as template and absent when the driver DNA was used a template then the sequence was designated tester-specific. The confirmed tester-specific oligonucleotides were then used to amplify PCR products from genomic DNA prepared from the strain collection listed in Table 2. A positive control, to test the integrity of the genomic DNA template, was performed using primers (23S Forward, 5'-ctaccttaggaccgttatagttac-3'; 23S Reverse, 5'-gaaggaactaggcaaaatggt-3') specific for a region of the 23S gene conserved within the Enterobacteriaceae. The nucleotide sequences of all of the primers were searched against the Y. pestis CO92 sequence, to ensure that each sequence occurred only once within the genome.

BLAST searches using tester-specific DNA sequences were performed via the National Center for Biotechnology Information (NCBI) website (http://www.ncbi.nlm.nih.gov/BLAST/). ORFs were identified and compared to non-redundant protein sequence databases using ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). Sequences that were present in the tester were searched against the Sanger Centre Y. pestis CO92 database (http://www.sanger.ac.uk/Projects/Y_pestis/blast_server.shtml) (Altschul et al., 1997 ; Parkhill et al., 2001 ) and the preliminary sequence of Y. pseudotuberculosis IP32953 (held at http://bbrp.llnl.gov/bbrp/bin/y.pseudotuberculosis_blast).

The sequence context of the difference products that did not map to the Y. pestis CO92 genome was determined by the identification of clones in a library of Y. pestis D15 Yokohama DNA fragments with an average size of 5 kb. The library was prepared by generating a partial digest of genomic DNA with Sau3A I, gel-purifying the fragments that fell within the 4–6 kb size range and then cloning them into the BamHI site of vector pUC9. The library was probed with oligonucleotides designed from the difference product identified as tester-specific.

Nomenclature.
The term ‘difference product’ refers to the tester-specific restriction fragment derived from SSH. The term ‘difference region’ (DFR) is used to describe the region of the genome to which several closely located difference products map. The term ‘DFR profile’ refers to the pattern of alleles for each of the six DFRs identified.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Mapping of difference products to DFRs
Subtractions with various strains of Y. pestis identified a collection of clones whose nucleotide sequences could be aligned with that of Y. pestis CO92. Sixteen such clones were found to map to five DFRs (DFR1, DFR2, DFR3, DFR5 and DFR6) that were distributed throughout the Y. pestis CO92 genome (Fig. 1A). The sixth DFR (DFR4) was absent from Y. pestis CO92, and was mapped immediately downstream of DFR3 (Fig. 1B; see below). Difference products within the pgm locus of Y. pestis were identified by BLAST analyses and were discarded (data not shown).



View larger version (22K):
[in this window]
[in a new window]
 
Fig. 1. Distribution of the difference products within each DFR. (A) An illustration of the five DFRs that are present in the Y. pestis CO92 genome. The numbered solid boxes show the location of each difference product (clone); short vertical lines delineate the boundaries of contiguous difference products. The locations of insertion elements are shown as expanded hatched boxes. Arrows show the approximate locations of direct repeats (DR). The rightmost DR of DFR2 is interrupted by IS1541 (dotted line). The boundary of each DFR is shown as its maximum size (dashed line) and minimum size (solid line), as determined by PCR (see Methods for explanation). (B) A graphical representation of DFR4, which does not map to Y. pestis CO92. The expanded hatched boxes show the locations of IS1541, the rightmost of which, IS1541', is truncated.

 
The boundaries of the DFRs that mapped to Y. pestis CO92 were determined by designing oligonucleotides based on the Y. pestis CO92 sequence upstream and downstream of the locations of the difference products. These oligonucleotides were used to amplify PCR products from templates prepared from the tester, the driver and Y. pestis CO92. This approach assumed that the tester sequence was identical to that of Y. pestis CO92. Indeed, the tester DNA was always reported as positive in these amplifications. No products were seen when the nucleotide sequence of the driver DNA differed from that of Y. pestis CO92. It is highly unlikely that false-negative results would be seen, given the clonal nature of the Y. pestis genome (Achtman et al., 1999 ; Adair et al., 2000 ). Amplification products were seen when the nucleotide sequences of the driver and Y. pestis CO92 resumed collinearity. The borders can thus be crudely, but rapidly, mapped to within approximately 1 kb, based on their location between the last negative result and the first positive PCR result, when DNA from the driver is used as the template. Although this approach does not provide a precise location for sequence divergence, it can rapidly establish the context of multiple loci within a DFR. The GenBank accession numbers for each difference product are listed in Table 1.

One clone was identified from a subtraction where DNA from Y. pestis D15 Yokohama (antiqua) was used as the tester and Y. pestis KIMD27 (mediaevalis) was used as the driver which did not map to the Y. pestis CO92 genome (difference product no. 2 of DFR4, Fig. 1B). A library of partial Sau3A I digests of Y. pestis D15 Yokohama DNA was probed with oligonucleotides specific to this difference product. One clone was identified that contained a 5645 bp insert, and this was termed DFR4. The insert sequence was determined, and it showed a 5' 1180 bp region that overlapped the 3' region of DFR3, indicating that DFR4 is deleted from Y. pestis CO92 immediately downstream of the rightmost copy of IS1541 in DFR3. Overlapping clones from the Y. pestis D15 Yokohama library were not identified using the DFR4 difference product no. 2 specific primer pair, but two additional difference products (nos 1 and 3) were located within the same library clone. Furthermore, three additional difference products (DFR4 nos 4, 5 and 6) were identified that had identical properties, namely they were present in Y. pestis D15 Yokohama and were absent from Y. pestis KIM D27. These three difference products were also found to be similar to three non-overlapping contigs in the Y. pseudotuberculosis database. Primers were designed at the ends of these contigs, to map their relative position in Y. pestis D15 Yokohama. Successfully amplified products were cloned, sequenced and assembled into a single contig to complete the sequence of DFR4 (accession no. AF426171). The genome of Y. pestis D15 Yokohama resumes collinearity with Y. pestis CO92 15·6 kb downstream of IS1541 in DFR4.

Diversity of DFRs in different strains of Y. pestis
The presence or absence of each DFR was determined in 78 strains representing all three biovars of Y. pestis and many worldwide origins, to examine the possibility of biovar specificity. Four strains of Y. pseudotuberculosis, the progenitor of Y. pestis, representing two serogroups were also included in the study. PCR using oligonucleotides specific to the multiple difference products (clones) that map within each DFR demonstrated the presence or absence of biovar specificity. In all cases, the PCR results were identical from each difference product within each DFR, with each plus or minus in Table 2 representing identical data for at least nine PCRs.

With the exception of DFR5, there seems to be little correlation between the presence and absence of each individual DFR with the biovars of Y. pestis. DFR5 is present in all orientalis biovars, whereas it is absent in all mediaevalis biovars (Table 2). DFR5 is found in only one strain of the antiqua biovar: Y. pestis Nicholisk 51 (strain 65, Table 2). Interestingly, this strain represents a variant of the antiqua biovar with restored ability to ferment glycerol (Motin et al., 2002 ). The PCR data from the Y. pestis collection were correlated as profiles designated A–M (Table 2). Profiles A, B and C were found almost exclusively in orientalis strains, and all contained DFR5. The only exception was Nicholisk 51 (antiqua biovar), which fell into profile A. Profiles G, H and I were seen for the mediaevalis biovars; profile I was composed of Central Asian isolates, including five KIM strains (strain 61; Table 2). Profiles J, K and M were seen for strains within the antiqua biovar, whereas profiles D, E and F were found in both the antiqua and mediaevalis biovars. Two strains whose biovar could not be determined, since they could not utilize glycerol or reduce nitrate, are listed in Table 2 (strains 77 and 78). Strain 77 had the same profile (A) as the orientalis biovars, whereas strain 78 had profile I. All four strains of Y. pseudotuberculosis fell into the same profile (L), indicating that the presence or absence of the six DFRs is independent of its serovar. Profile L was not found for any strain of Y. pestis tested. DFR2 and DFR5 were both absent in Y. pseudotuberculosis, and may represent regions acquired since Y. pestis emerged as a derivative of Y. pseudotuberculosis. DFR2 was seen in all but two strains that originated in South America (biovar orientalis, profile C, strains 50 and 51; Table 2). DFR2 was isolated from subtractions using driver DNA from Y. pestis D14 Salazar (strain 51; Table 2).

BLAST analysis of the DFRs
The results of BLAST analyses for each DFR are shown in Table 3, and show good concordance with the genome of Y. pestis CO92 (Parkhill et al., 2001 ). DFR1 contains 15 ORFs, eight of which show similarity to genes involved in lateral flagellar synthesis (McCarter & Wright, 1993 ). Two ORFs in DFR1 encode proteins that are 39 and 59% identical to ORFA and ORFB of IS1397, respectively (Bachellier et al., 1997 ). Seven ORFs were identified within DFR2 which show identity to bacteriophage sequences, including a putative replicon, primase and integrase. There are two directly repeated regions of 287 bp, the rightmost of which is interrupted by a copy of IS1541, and which appears to be located at the right boundary of DFR2 (Fig. 1A). A copy of IS100 is seen at the approximate location of the left boundary of DFR2, all of which are features commonly found in pathogenicity islands (Hacker & Kaper, 2000 ). DFR3 contains 10 ORFs, some of which show identity to enzymes involved in purine salvage pathways, which are encoded by xapRAB. There are two directly repeated copies of IS1541 located at the boundaries of DFR3; the rightmost IS1541 interrupts the xapB ORF (YPO1172; Parkhill et al., 2001 ). Downstream there are two ORFs with no similarity to the non-redundant database. These are followed by three ORFs with similarity to yohI, penicillin-binding protein 7 and D-lactate dehydrogenase of Escherichia coli. Interestingly, ORF analysis of DFR4 reveals sequence similarity to the C-terminal portion of XapB immediately downstream of IS1541. Thus, it would appear that IS1541 has been inserted into the xapB ORF in Y. pestis D15 Yokohama and lies immediately downstream of DFR3. DFR4 has been deleted from Y. pestis CO92 and is absent from all orientalis biovars tested. The sequence downstream of this copy of IS1541 in DFR4 encodes 17 additional ORFs, the first of which shows similarity to the C terminus of XapB. The rest of DFR4 encodes homologues of hydH/hydG from Salmonella typhimurium, and hypothetical ORFs from Pseudomonas aeruginosa and E. coli. The ORF of Y. pestis hydH contains an insertion when compared to hydH of S. typhimurium, and the hydG ORF is incomplete. The right boundary of DFR4 is delimited by another (truncated) copy of the insertion sequence element IS1541. DFR4 is present in some antiqua and mediaevalis strains, and is present in all four strains of Y. pseudotuberculosis tested (Table 2). DFR5 has 14 ORFs and is located near to the replication terminus of Y. pestis (accession no. AAF68950). The smallest DFR, DFR6, encodes three ORFs, one of which shows 73% amino acid sequence identity to SepC from the insect pathogen Serratia entomophila. The sepABC operon is a member of a family of toxin complex (tc) genes that encode insecticidal proteins (Waterfield et al., 2001 ). The G+C content of the entire Y. pestis CO92 genome is 47·6 mol%, whereas the G+C contents of the DFRs range from 43·1 to 52·5 mol%.


View this table:
[in this window]
[in a new window]
 
Table 3. ORFs found in the DFRs

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
It is becoming increasingly apparent that the acquisition of genes via lateral transfer is the basis of bacterial genome diversity (Ochman et al., 2000 ; Perna et al., 2001 ). Lateral gene transfer enables strains to rapidly acquire large DNA segments that encode functions facilitating their survival. The very close genetic similarity of Y. pestis and Y. pseudotuberculosis has long been established (Bercovier et al., 1980 ). The few sequence polymorphisms that have been reported for Y. pestis represent small changes in nucleotide sequence, and these would not be detected by SSH. It is likely that lateral gene transfer has contributed to the evolution of Y. pestis, perhaps even since it emerged as a clone derived from Y. pseudotuberculosis. Indeed, RFLP experiments indicate genome plasticity due to the mobility of larger regions of DNA (Achtman et al., 1999 ; McDonough & Hare, 1997 ; Lucier & Brubaker, 1992 ), as does the instability of the 102 kb pgm locus (Fetherston et al., 1992 ). The genome sequence of Y. pestis CO92 demonstrates the genetic flexibility of this strain (Parkhill et al., 2001 ). The three genomic anomalies reported in Y. pestis CO92 were not detected by SSH since they represent rearrangements; SSH depends on the presence of a sequence in one genome (tester) and its absence in another (driver).

Data presented here indicate the presence of six large genomic DFRs ranging in size from 4·6 to 19 kb (Fig. 1), which demonstrates for the first time the genetic basis of genomic plasticity among the three biovars of Y. pestis. Interestingly, the DFR profile (L) of Y. pseudotuberculosis is not seen in Y. pestis. One of the Y. pestis DFRs, DFR2, is found in all but two of the 78 Y. pestis strains tested, but it is absent from all four isolates of Y. pseudotuberculosis tested (Table 2). Therefore, DFR2 appears to be a region that was acquired early in the evolution of Y. pestis. It possesses many hallmarks of lateral gene transfer: the 15 kb region encodes ORFs with similarities to many bacteriophage genes (including replication, primase and integration functions), it possesses an insertion element (IS1397) and it is delineated by two direct repeats. However, there is no evidence that DFR2 is inserted into a tRNA gene and its 44·5 mol% G+C content is not significantly different from that of the rest of the Y. pestis CO92 genome (47·6%), as is common with pathogenicity islands (Hacker & Kaper, 2000 ). Since Y. pestis emerged less than 20000 years ago (Achtman et al., 1999 ), it is more likely that new sequences were acquired from genomes with a similar G+C content, rather than amelioration whereby an incoming sequence is adjusted to the base composition of the resident genome (Lawrence & Ochman, 1997 ).

In contrast, the presence or absence of some of the DFRs (DFR1, DFR3 and DFR6) shows no correlation with the biovar of the isolates. These DFRs are also present in Y. pseudotuberculosis and have been subsequently deleted in some strains of Y. pestis. Eight of the 15 ORFs in DFR1 have similarity to genes involved in lateral flagellar synthesis responsible for surface swarming in Vibrio parahaemolyticus (McCarter & Wright, 1993 ). Since Y. pestis is non-motile, the presence of DFR1 does not apparently confer motility on those strains in which it is present. ORFs identified in DFR3 encode putative proteins involved in the purine salvage pathway (Seeger et al., 1995 ). A single ORF was found in DFR6 with significant similarity to SepC, an insect toxin encoded on a plasmid in Serratia entomophila (Hurst et al., 2000 ). Given the insect vector-borne aetiology of bubonic plague, this is an interesting observation to bear in mind as the mode of action of the SepABC insecticidal proteins is being uncovered.

Another DFR, DFR5, is present in only the orientalis biovar of Y. pestis, which is considered to have emerged most recently (Buchrieser et al., 1999 ). DFR5 is absent from all mediaevalis and all but one of the antiqua biovars (Nicholisk 51), as well as from Y. pseudotuberculosis (Table 2). Therefore, DFR5 appears to be more recently acquired than DFR2 (which is seen in all three biovars). Two of the 14 ORFs in DFR5 are similar to bacteriophage genes, again implicating a lateral gene transfer event. Interestingly, DFR5 is located adjacent to the replication terminus of the genome, a site where increased recombination has been observed (Perkins et al., 1993 ), and which may explain genome rearrangement involving DFR5 in the absence of insertion sequences.

DFR4 is absent from all orientalis biovars tested, indicating that the evolution of the biovars results from both the acquisition and deletion of genetic material. DFR4 is also deleted from Nicholisk 51, an antiqua strain, which has recently been shown to have an IS100-based fingerprint more typical of the orientalis biovar (Motin et al., 2002 ). This observation led to the hypothesis that this antiqua strain has restored ability to ferment glycerol. The data presented here support this hypothesis, as Nicholisk 51 possesses the same six DFR alleles (profile A) as the majority of orientalis strains.

Currently, biovar designation is based on only two phenotypes: the ability to utilize glycerol (antiqua and mediaevalis biovars) and the ability to reduce nitrate (antiqua and orientalis biovars). Thus, the only discernable difference between Nicholisk 51 (antiqua) and the profile A orientalis biovars is the ability of the former to utilize glycerol. It is possible that the loss of glycerol utilization seen in orientalis strains is due to a deletion of 93 bp in the glycerol-3-phosphate dehydrogenase gene glpD (Motin et al., 2002 ). Similarly, the single mediaevalis strain with profile F, Harbin 35 (strain 56, Table 2), only differs from antiqua strains with the same profile by its inability to reduce nitrate, which may also be a simple loss of gene function. Two strains in this collection, 316 (strain 77, Table 2) and Pestoides J (strain 78, Table 2), could not be characterized with respect to a biovar, since they are negative for both glycerol utilization and nitrate reduction. Since 316 has profile A it seems appropriate to align it with the orientalis strains, whereas Pestoides J seems to conform best with the Central Asian mediaevalis strains of profile I. Clearly, the comprehensive and accurate determination of the evolutionary development of the three biovars of Y. pestis requires analysis of more than two genetic markers.

It is possible to use the DFR profiles presented here to propose a model of the evolutionary relationships between natural isolates. The number of profiles possible for six DFRs is 64 (26), of which 12 are represented in this strain collection (though the possibility exists that other strains possess profiles not found here). The model originates with Y. pseudotuberculosis as the progenitor, which possesses DFR profile L. Assuming that only one DFR allele is changed at a time, it is possible to suggest the simple evolutionary model presented in Fig. 2 using the 12 profiles found in this study. The only Y. pestis derivative with a single DFR change from profile L is profile D, which results from the acquisition of DFR2. This would imply that the earliest Y. pestis strains would possess profile D (Antiqua, Pestoides A and Pestoides B). The profiles that appear early in the evolution of Y. pestis (profiles D, E and F) are seen in the older antiqua and mediaevalis biovars. A lineage leading to the Central Asian mediaevalis strains, including KIM10 (profile I), can be proposed via profiles K and H (Fig. 2). Profile K can arise from profile D via F or M. Since profiles K and M are seen only in antiqua biovars, this is the more likely evolutionary route. The orientalis lineage progresses from profile D to profile A via profile M by the loss of DFR4 followed by the acquisition of DFR5. Twelve of the 51 orientalis strains have profiles that contain only one DFR allele different from profile A. Specifically, the loss of DFR3 leads to profile B, and the loss of DFR2 leads to profile C. Y. pestis strains with profile B are geographically diverse, whereas the two strains with profile C are both South American isolates. Significantly, these observations conform to the results of PFGE analyses of strains EV76 (Profile B) and Salazar (Profile C), which showed different restriction patterns to the other orientalis strains studied (Lucier & Brubaker, 1992 ). There is more homogeneity within the DFR profiles of the orientalis biovar, which would be expected since the orientalis biovar has evolved over a shorter time period (since 1894) than the antiqua or mediaevalis biovars (1500–20000 years). Even though only three profiles are seen within the orientalis biovar, this is evidence that significant changes in the Y. pestis genome have occurred in as short a time frame as 100 years.



View larger version (9K):
[in this window]
[in a new window]
 
Fig. 2. Model of the evolutionary relationships of strains of Y. pestis based on different DFR profiles. Profile L is seen in the four strains of Y. pseudotuberculosis tested, whereas the remaining profiles represent those found in Y. pestis.

 
Insertion elements have played important roles in the pathogenicity of many bacteria (Hacker & Kaper, 2000 ), and are highly represented in the Y. pestis CO92 genome (Parkhill et al., 2001 ). They have been previously implicated in the strain variability of Y. pestis, namely by RFLP analysis of the IS100 location (Achtman et al., 1999 ; McDonough & Hare, 1997 ), and their presence contributes to the instability of the pgm locus (Buchrieser et al., 1998 ; Fetherston & Perry, 1994 ). In E. coli, the majority of insertion sequences are associated with lateral gene transfer, and they are frequently located at junctions of native and transferred DNA (Lawrence & Ochman, 1998 ). In this study insertion elements are seen close to the borders of DFR1, DFR2, DFR3 and DFR4. DFR2 is flanked by two different insertion elements, IS100 and IS1541, and appears to be very stable as it is present in all but two of the strains tested. There is more variability in the presence or absence of DFR3 and DFR1, which are also flanked by insertion elements. One possible mechanism for the loss of DFR3 is recombination between the two directly repeated copies of IS1541, in a manner analogous to the instability of the pgm locus. The six DFRs of Y. pestis were stable during laboratory manipulations such as serial subculture or passage through laboratory animals. The original Y. pestis CO92 isolate was compared to 16 derivatives (pgm-negative derivatives, plasmid-less derivatives, strains in which plasmid pMT1 appears to have integrated into the chromosome and strains that had been subjected to long-term serial subculture on various media at room temperature and 37 °C). All Y. pestis CO92 derivatives gave the same result: profile A (no. 17; Table 2). Further evidence for stability under laboratory conditions is seen in the Yreka derivatives A1122 and A12 (strains 38, 11 and 12; Table 2), which all possess profile A. Presumably the selective pressure driving genome evolution in the natural environment of Y. pestis is far greater than that found under laboratory conditions.

In summary, here we report the first successful identification and characterization of large regions of DNA responsible for genomic differences between different strains of Y. pestis. This approach is particularly powerful when a reference genome is available, and in this study the genome sequences of Y. pestis CO92 and Y. pseudotuberculosis IP32953 were exploited. When no reference genome is available, a library of tester fragments can be interrogated which, although more laborious, will still provide useful data. Six DFRs were identified among Y. pestis strains, which result from both the acquisition and deletion of large regions of DNA, and these carry many hallmarks of lateral gene transfer and provide clues to the evolutionary relationships between different biovars and geographical isolates of Y. pestis. Future work will determine the exact boundaries of these DFRs and their benefits in an evolutionary context. Specifically, their role in the survival or pathogenicity of Y. pestis will be ascertained.


   ACKNOWLEDGEMENTS
 
This work was performed under the auspices of the US Dept of Energy by the University of California, Lawrence Livermore National Laboratory, under contract no. W-7405-Eng-48, and was funded by the Department of Energy, NN-20, Chemical and Biological Non-Proliferation Program. We appreciate the generosity of May Chu (CDC, Fort Collins), Robert Brubaker (Michigan State University) and Will Probert (Dept of Health Services, Berkeley, CA) who supplied genomic DNA for some strains included in this study. We appreciate the technical assistance of Sylvia Gamez-Chin, Aubree Hubbell, Madison Macht and Jessica Wollard. We greatly appreciate the sharing of unpublished data by Vladimir Motin, and gratefully acknowledge his constructive suggestions for this manuscript.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Achtman, M., Zurth, K., Morelli, G., Torrea, G., Guiyoule, A. & Carniel, E. (1999). Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis. Proc Natl Acad Sci USA 96, 14043-14048.[Abstract/Free Full Text]

Adair, D. M., Worsham, P. L., Hill, K. K., Klevytska, A. M., Jackson, P. J., Friedlander, A. M. & Keim, P. (2000). Diversity in a variable-number tandem repeat from Yersinia pestis. J Clin Microbiol 38, 1516-1519.[Abstract/Free Full Text]

Akopyants, N. S., Fradkov, A., Diatchenko, L., Hill, J. E., Siebert, P. D., Lukyanov, S. A., Sverdlov, E. D. & Berg, D. E. (1998). PCR-based subtractive hybridization and differences in gene content among strains of Helicobacter pylori. Proc Natl Acad Sci USA 95, 13108-13113.[Abstract/Free Full Text]

Altschul, S. F., Madden, T. L., Schäffer, A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25, 3389-3402.[Abstract/Free Full Text]

Bachellier, S., Clément, J. M., Hofnung, M. & Gilson, E. (1997). Bacterial interspersed mosaic elements (BIMEs) are a major source of sequence polymorphism in Escherichia coli intergenic regions including specific associations with a new insertion sequence. Genetics 145, 551-562.[Abstract/Free Full Text]

Bercovier, H., Mollaret, H. H., Alonso, J. M., Brault, J., Fanning, G. R., Steigerwalt, A. G. & Brenner, D. J. (1980). Intra- and interspecies relatedness of Yersinia pestis by DNA hybridization and its relationship to Y. pseudotuberculosis. Curr Microbiol 4, 225-229.

Buchrieser, C., Prentice, M. & Carniel, E. (1998). The 102-kilobase unstable region of Yersinia pestis comprises a high-pathogenicity island linked to a pigmentation segment which undergoes internal rearrangement. J Bacteriol 180, 2321-2329.[Abstract/Free Full Text]

Buchrieser, C., Rusniok, C., Frangeul, L., Couve, E., Billault, A., Kunst, F., Carniel, E. & Glaser, P. (1999). The 102-kilobase pgm locus of Yersinia pestis: sequence analysis and comparison of selected regions among different Yersinia pestis and Yersinia pseudotuberculosis strains. Infect Immun 67, 4851-4861.[Abstract/Free Full Text]

Devignat, R. (1951). Variétés de l’espèce Pasteurella pestis. Nouvelle hypothèse. Bull W H O 4, 247-253.

Diatchenko, L., Lau, Y. F., Campbell, A. P. & 8 other authors (1996). Suppression subtractive hybridization: a method for generating differentially regulated or tissue-specific cDNA probes and libraries. Proc Natl Acad Sci USA 93, 6025–6030.[Abstract/Free Full Text]

Ewing, B. & Green, P. (1998). Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res 8, 186-194.[Abstract/Free Full Text]

Ewing, B., Hillier, L., Wendl, M. C. & Green, P. (1998). Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res 8, 175-185.[Abstract/Free Full Text]

Fetherston, J. D. & Perry, R. D. (1994). The pigmentation locus of Yersinia pestis KIM6+ is flanked by an insertion sequence and includes the structural genes for pesticin sensitivity and HMWP2. Mol Microbiol 13, 697-708.[Medline]

Fetherston, J. D., Schuetze, P. & Perry, R. D. (1992). Loss of the pigmentation phenotype in Yersinia pestis is due to the spontaneous deletion of 102 kb of chromosomal DNA which is flanked by a repetitive element. Mol Microbiol 6, 2693-2704.[Medline]

Gordon, D., Abajian, C. & Green, P. (1998). Consed: a graphical tool for sequence finishing. Genome Res 8, 195-202.[Abstract/Free Full Text]

Guiyoule, A., Grimont, F., Iteman, I., Grimont, P. A., Lefèvre, M. & Carniel, E. (1994). Plague pandemics investigated by ribotyping of Yersinia pestis strains. J Clin Microbiol 32, 634-641.[Abstract]

Guiyoule, A., Rasoamanana, B., Buchrieser, C., Michel, P., Chanteau, S. & Carniel, E. (1997). Recent emergence of new variants of Yersinia pestis in Madagascar. J Clin Microbiol 35, 2826-2833.[Abstract]

Hacker, J. & Kaper, J. B. (2000). Pathogenicity islands and the evolution of microbes. Annu Rev Microbiol 54, 641-679.[Medline]

Hurst, M. R., Glare, T. R., Jackson, T. A. & Ronson, C. W. (2000). Plasmid-located pathogenicity determinants of Serratia entomophila, the causal agent of amber disease of grass grub, show similarity to the insecticidal toxins of Photorhabdus luminescens. J Bacteriol 182, 5127-5138.[Abstract/Free Full Text]

Jain, R., Rivera, M. C. & Lake, J. A. (1999). Horizontal gene transfer among genomes: the complexity hypothesis. Proc Natl Acad Sci USA 96, 3801-3806.[Abstract/Free Full Text]

Lawrence, J. G. & Ochman, H. (1997). Amelioration of bacterial genomes: rates of change and exchange. J Mol Evol 44, 383-397.[Medline]

Lawrence, J. G. & Ochman, H. (1998). Molecular archaeology of the Escherichia coli genome. Proc Natl Acad Sci USA 95, 9413-9417.[Abstract/Free Full Text]

Lucier, T. S. & Brubaker, R. R. (1992). Determination of genome size, macrorestriction pattern polymorphism, and nonpigmentation-specific deletion in Yersinia pestis by pulsed-field gel electrophoresis. J Bacteriol 174, 2078-2086.[Abstract]

McCarter, L. L. & Wright, M. E. (1993). Identification of genes encoding components of the swarmer cell flagellar motor and propeller and a sigma factor controlling differentiation of Vibrio parahaemolyticus. J Bacteriol 175, 3361-3371.[Abstract]

McDonough, K. A. & Hare, J. M. (1997). Homology with a repeated Yersinia pestis DNA sequence IS100 correlates with pesticin sensitivity in Yersinia pseudotuberculosis. J Bacteriol 179, 2081-2085.[Abstract]

Motin, V. L., Georgescu, A. M, Elliott, J. M. & 8 other authors (2002). Genetic variability of Yersinia pestis isolates as predicted by PCR-based IS100 genotyping and analysis of structural genes encoding glycerol-3-phosphate dehydrogenase (glpD). J Bacteriol 184, 1019–1027.[Abstract/Free Full Text]

Ochman, H., Lawrence, J. G. & Groisman, E. A. (2000). Lateral gene transfer and the nature of bacterial innovation. Nature 405, 299-304.[Medline]

Parkhill, J., Wren, B. W., Thomson, N. R. & 32 other authors (2001). Genome sequence of Yersinia pestis, the causative agent of plague. Nature 413, 523–527.[Medline]

Perkins, J. D., Heath, J. D., Sharma, B. R. & Weinstock, G. M. (1993). XbaI and BlnI genomic cleavage maps of Escherichia coli K-12 strain MG1655 and comparative analysis of other strains. J Mol Biol 232, 419-445.[Medline]

Perna, N. T., Plunkett, G., III, Burland, V. & 25 other authors (2001). Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 409, 529–533.[Medline]

Seeger, C., Poulsen, C. & Dandanell, G. (1995). Identification and characterization of genes (xapA, xapB, and xapR) involved in xanthosine catabolism in Escherichia coli. J Bacteriol 177, 5506-5516.[Abstract]

Skowronski, E. W., Armstrong, N., Andersen, G., Macht, M. & McCready, P. M. (2000). Magnetic microplate-format plasmid isolation protocol for high-yield, sequencing-grade DNA. Biotechniques 29, 786-790.[Medline]

Trebesius, K., Harmsen, D., Rakin, A., Schmelz, J. & Heesemann, J. (1998). Development of rRNA-targeted PCR and in situ hybridization with fluorescently labelled oligonucleotides for detection of Yersinia species. J Clin Microbiol 36, 2557-2564.[Abstract/Free Full Text]

Waterfield, N. R., Bowen, D. J., Fetherston, J. D., Perry, R. D. & ffrench-Constant, R. H. (2001). The tc genes of Photorhabdus: a growing family. Trends Microbiol 9, 185-191.[Medline]

Received 4 September 2001; revised 13 January 2002; accepted 15 February 2002.