|Home | About | Journals | Submit | Contact Us | Français|
An examination of the two Leptospira interrogans genomes sequenced so far reveals few genetic differences, including an extra DNA region, 54 kb in length, in L. interrogans serovar Lai. This locus contains 103 predicted coding sequences that are absent from the genome of L. interrogans serovar Copenhageni, of which only 20% had significant BLASTP hits in GenBank. By analyzing the L. interrogans serovar Lai genome by pulsed-field gel electrophoresis, we also found that this 54-kb DNA fragment exists as a circular plasmid. This was confirmed by amplification of a DNA fragment corresponding to that of the predicted fragment if this region excised from the chromosome and its left and right ends joined together. In addition, cloning of the putative rep gene of this DNA region was responsible for autonomous replication in Leptospira spp., therefore generating a new Escherichia coli-Leptospira sp. shuttle vector. Taken together, our results show that this genomic island can excise from the chromosome and form a replicative plasmid. Analysis of the distribution of this genomic island revealed that highly related sequences exist in other L. interrogans virulent strains. This genomic island, containing a high proportion of novel genes, may have an important role in spreading genes, including virulence factors, among bacterial populations.
The phylum of spirochetes has a deep branching lineage, and numerous studies have shown that they differ in a number of morphological, structural, biochemical, and genetic respects from well-studied gram-negative and gram-positive bacteria. Spirochetes are the causative agents of several important human diseases such as syphilis, Lyme disease, and leptospirosis.
Leptospirosis is a worldwide-distributed zoonosis, which is endemic in tropical areas. With the occurrence of large outbreaks in the last decade in areas including Nicaragua (23), Brazil (12), and India (26), leptospirosis is now recognized as an important emerging infectious disease. In industrialized countries, the chance of contracting leptospirosis can increase depending on recreational or occupational exposure. It is estimated that more than 500,000 cases of severe leptospirosis occur annually in the world, with a mortality rate of up to 23% (25). Transmission to humans occurs through direct or indirect contact with urine of infected animals, such as small rodents (15).
Leptospira interrogans is the most frequently reported agent of leptospirosis, with L. interrogans serogroup Icterohaemorrhagiae representing more than half of the leptospires encountered in human infections. Recently, the completion of the genome sequences of L. interrogans serovar Lai and L. interrogans serovar Copenhageni, both belonging to serogroup Icterohaemorrhagiae, was achieved (16, 20). The Lai and Copenhageni sequenced strains were isolated from patients with severe leptospirosis in China (20) and Brazil (16), respectively. It should be noted that while the regular reservoir host for serovar Lai is the striped field mouse (Apodemus agrarius), the host for serovar Copenhageni is the domestic rat (Rattus norvegicus) (15). The L. interrogans genome consists of a 4.33-Mb large circular chromosome and a 350-kb small chromosome (16, 20). The majority of predicted coding sequences (CDSs) of the L. interrogans genome fail to exhibit similarities to proteins of known function in other organisms. The genome is highly conserved between the serovars Lai and Copenhageni, exhibiting 95% identity at the nucleotide level (16). Comparative genomics of these two L. interrogans serovars reveals few genetic differences, including a 54-kb insertion specific to serovar Lai, not found in serovar Copenhageni (Fig. (Fig.1).1). In this study, we show that this 54-kb DNA fragment can both excise from the Lai chromosome and form a circular plasmid.
We obtained reference strains belonging to L. interrogans, L. kirschneri, L. noguchi, and L. borgpetersenii from the collection maintained by the National Reference Laboratory for Leptospira at the Institut Pasteur (Paris, France). All strains were cultured at 30°C in EMJH liquid medium (7, 11).
Genomic DNA of Leptospira was isolated using the phenol-chloroform method as previously described (17). For pulsed-field gel electrophoresis (PFGE), cells were embedded in agarose plugs as previously described (5). For digestion, DNA plugs were washed in Tris-EDTA buffer, followed by equilibration in 1× restriction enzyme buffer, and then incubated overnight at 37°C in fresh 1× restriction enzyme buffer containing 30 U NotI restriction enzyme. PFGE was performed in a contour-clamped homogeneous electric field DRII apparatus (Bio-Rad Laboratories, Richmond, CA). Programs with a ramping from 5 to 100 s for 40 h at 170 V or from 10 to 100 s for 40 h at 150 V were used for the resolution of DNA fragments. The DNA was amplified using Taq polymerase (Amersham Biosciences, Little Chalfont, England) under the following conditions: 2 min at 95°C, followed by 35 cycles of 10 s at 95°C, 15 s at 55°C, and 5 min at 72°C, and then one cycle of 10 min at 72°C. For Southern blot analysis, genomic DNA was digested, subjected to electrophoresis in a 1% agarose gel, and transferred onto nylon membranes as previously described (17). Probes were generated by PCR and radiolabeled with [α-32P]dATP using a commercial kit (Megaprime; Amersham Biosciences). Membranes were hybridized overnight at 55 or 60°C in rapid hybridization buffer (Amersham Biosciences) and then washed as previously described (17).
The oligonucleotides listed in Table Table11 were used to amplify DNA fragments from L. interrogans serovar Lai.
A 1.5-kb fragment encompassing the LA1839 gene of L. interrogans serovar Lai was amplified with primers 1839a and 1839c (Table (Table1)1) and inserted into pCR2.1-TOPO by using a TOPO TA cloning kit (Invitrogen Life Technologies, Carlsbad, CA). After PvuII digestion, the DNA fragment containing LA1839 was inserted into the dephosphorylated SmaI site of the spectinomycin-resistant plasmid pGSpc, which is derived from pGEM-7Zf(+) (Promega, Madison, WI). Plasmids from Escherichia coli were recovered using a QIAprep spin miniprep kit (QIAGEN GmbH, Hilden, Germany). Leptospira cells were electrotransformed as previously described (22). Spectinomycin-resistant clones were verified by isolation of plasmid DNA and subsequent restriction fragment length polymorphism analysis.
The nucleotide sequences of the genomes of L. interrogans serovar Lai strain Lai 56601 (20) and L. interrogans serovar Copenhageni strain Fiocruz L1-130 (16) were analyzed using MaGe software (http://www.genoscope.cns.fr/agc/mage/) (24). PCR products were sequenced at Genome Express (Meylan, France).
The accession numbers of the nucleotide sequences of the serovar Lai genomic island are EF088430 (5′ flanking sequence of the chromosomal locus), DQ890383 (chromosomal empty locus), and DQ890384 and EF100899 (circular intermediates).
In bacteria, genomic islands (including those related to virulence, pathogenicity islands) represent large chromosomal regions that are present in a subset of strains of the same species. Their G+C content usually differs from that of the host chromosome. They are flanked by direct repeats and are often inserted into tRNA-encoding genes. They also encode mobility enzymes, such as integrase, allowing for excision from the host chromosome (6). Analysis of the 54-kb DNA region of L. interrogans serovar Lai (from LA1768 to LA1847) not found in serovar Copenhageni (Fig. (Fig.1)1) suggests that this locus is a genomic island, further referred to as Lai genomic island I (LaiGI I).
This island is not associated with tRNA genes but is flanked by a putative rhs (rearrangement hot spot)-like gene and an insertion sequence (LA1848 and LA1849) belonging to the IS4 family (Fig. (Fig.11 and and2).2). Rhs elements belong to a set of composite elements found in the chromosome of Escherichia coli. These genes, the function of which remains unknown in E. coli, are typically long CDSs with repetitions and can be found in association with insertion sequences (10). The reason for the association between LaiGI I and the rhs-like gene is not clear.
The presence of LaiGI I in the large chromosome of L. interrogans serovar Lai was verified by PCR amplification across the putative DNA junctions (Fig. (Fig.22 and and3),3), followed by sequencing of the PCR fragments. Sequencing of the flanking sequences of LaiGI I indicated the absence of an 812-bp direct repeat (Fig. (Fig.2),2), corresponding to the 5′ end of LA1766, which was previously believed to be present in the L. interrogans serovar Lai genome (20). This discrepancy may be due to an error in the assembly of genomic DNA sequences or to DNA rearrangements of the sequenced strain after in vitro passages. The 170-bp direct repeat found in L. interrogans serovar Copenhageni was found in the flanking regions of LaiGI I (Fig. (Fig.2).2). Sequence analysis of LaiGI I revealed an average G+C content of 33%, in comparison to 36% for its chromosomal host.
By using AMIGene (Annotation of MIcrobial Genes) software (1), which allows the calculation of coding prediction curves, the pattern of codon usage of genes belonging to LaiGI I was not distinguishable from that of codon usage in the rest of the genome (data not shown). LaiGI I contains 103 predicted CDSs, 23 of which had significant BLASTP hits in GenBank (see the supplemental material). The first gene of LaiGI I, LA1768, encodes a putative protein which shows features related to members of the RNase H-like superfamily (Pfam 3e-11), which includes the human immunodeficiency virus type 1 integrase and the Mu transposase (21). The LA1768 protein contains the three amino acids of the DDE motif (Asp-86, Asp-152, and Glu-185) found in the catalytic domains of transposases and retroviral integrases (14). This putative protein may be required for the integration and/or excision of the 54-kb DNA fragment.
Genomic islands can excise from the chromosome and form circular intermediates (6). Interestingly, genes encoding proteins that could be involved in plasmid segregation were identified in LaiGI I. The segregation of low-copy-number plasmids in bacterial cells is an efficient process that ensures that every daughter cell receives a copy of plasmid DNA, thereby enhancing their inheritance and stability (9). LaiGI I contains putative addiction (LA1780-LA1781) and partition (LA1837-LA1838) modules (see the supplemental material). Addiction modules contain two small genes encoding a potent cell toxin and an antidote protein. In the presence of the addiction module-containing plasmid, the toxicity is neutralized by the antidote protein, which is labile. In contrast, upon loss of the plasmid, the residual stable toxin will likely kill plasmid-free cells. Other toxin-antitoxin loci were identified in the L. interrogans genome (18, 27). Partition modules usually consist of the parA and parB genes, whose products act in conjunction with a centromere-like locus to facilitate faithful plasmid segregation (9). The identification of homologs of partitioning and postsegregational killing genes indicates that LaiGI I could be highly stable in L. interrogans serovar Lai.
The origins of replication of low-copy-number plasmids are often found in the vicinity of parAB. For instance, the rep gene of the leptophage LE1 was found immediately downstream of parAB (2). In LaiGI I, the gene found downstream of the parAB genes, LA1839, encodes a putative protein (381 amino acids in length) unrelated to any other eukaryotic or prokaryotic protein. However, the C-terminal region of the LA1839 protein (from amino acids 153 to 358) shows 50% similarity to the LB376 and LIC20276 proteins, which are encoded by genes located in the putative replication origin of the small chromosome (CII) of L. interrogans serovar Lai and L. interrogans serovar Copenhageni, respectively (16, 20). A 1.5-kb fragment encompassing the LA1839 protein was amplified and then cloned into an E. coli plasmid to generate pORLS. After electroporation with pORLS, we observed plasmid autonomous replication in our laboratory strain Leptospira biflexa and in the pathogen L. interrogans serovar Canicola strain Hond Utrech but not in L. interrogans serovar Lai and L. interrogans serovar Copenhageni (data not shown). However, transformation efficiency was very low (10 transformants per μg of DNA). The absence of transformants in serovars Lai and Copenhageni can therefore be due to poor transformation of the strains and not the absence of plasmid replication. The maintenance of pORLS in leptospiral cell strains suggests that LA1839 is the rep gene of LaiGI I. Consistent with the replication initiation protein (Rep protein), LA1839 encodes a protein containing a predicted helix-turn-helix motif at its N-terminal end that could be involved in DNA binding activity. Finally, the origin of replication of LaiGI I coincides with the polarity switch of the GC skew of the circular plasmid (see the supplemental material). The L. biflexa-E. coli shuttle vector derived from LE1 has already proved to be useful in studies of gene function in saprophytes (2). Our results enable the construction of a new shuttle vector for the genetics of Leptospira, including pathogenic species.
PFGE analysis of chromosomal DNA after digestion with the rare-cutting restriction enzyme NotI revealed the presence of an additional fragment, larger than 50 kb in size, in comparison to predictions made from the whole-genome sequence of L. interrogans serovar Lai (Fig. (Fig.4).4). Southern hybridization of the blot with specific probes of LaiGI I gave a strong signal with this ≈50-kb fragment. While DNA probes B and C are physically separated by a NotI restriction site when LaiGI I is inserted into the chromosome, our data indicate that these DNA probes are linked in the ≈50-kb NotI-digested DNA fragment (Fig. (Fig.4).4). Since LaiGI I contains a single NotI restriction site, this restriction profile is consistent with the linearized form of a circular plasmid corresponding to LaiGI I. No linear replicon was detected by PFGE of undigested DNA and Southern blot hybridization (Fig. (Fig.4).4). Comparison of the in silico restriction map of L. interrogans serovar Lai (20) and PFGE-separated, NotI-digested chromosomal DNA fragments also reveals that a 22-kb fragment predicted by in silico analysis was missing in PFGE. This 22-kb restriction fragment is associated with the chromosomal LaiGI locus (Fig. (Fig.4).4). Consistent with the absence of this segment, the 720-kb NotI-digested DNA deduced from the published sequence should be ≈688 kb in size (Fig. (Fig.4).4). However, it remains uncertain by PFGE whether the restriction fragment is 688 kb rather than 720 kb (the restriction fragment is located between the 679- and 727-kb fragments of the molecular size standard) (Fig. (Fig.4).4). Evaluation of the ethidium bromide staining intensities of the chromosomal restriction fragment less than 100 kb in length (Fig. (Fig.4)4) also suggests that the majority of LaiGI I DNA remains as a low-copy-number circular plasmid (similar to the chromosome copy number) rather than an integrated form in L. interrogans serovar Lai.
The presence of DNA segregation and replication mechanisms typical of circular plasmid and visualization of an extrachromosomal band by PFGE suggests that LaiGI I can replicate as a plasmid. A PCR strategy was therefore designed that would amplify a PCR product only if LaiGI I excised from the chromosome and formed a circular intermediate. PCR analysis with primers P2 and P3 successfully amplified a PCR product (Fig. (Fig.3).3). DNA sequencing of the cloned PCR product showed that the sequences were identical to that of the predicted fragment if LaiGI I excised from the chromosome and its left and right ends joined together. The occurrence of DNA junctions between distinct left and right ends of LaiGI I was also detected with primers P7 and P8 (Fig. (Fig.2),2), suggesting that imprecise excision of LaiGI I from the chromosome may result in distinct plasmid derivatives. By using a similar methodology, circular excision intermediates of genomic islands have been demonstrated in other organisms (8, 13, 19). We also hypothesized that there should be an empty chromosomal site, and we demonstrated that to be the case by PCR (Fig. (Fig.22 and and3).3). However, sequencing of the PCR product suggested that imprecise excision occurs, leaving behind a remnant of LaiGI I (Fig. (Fig.2).2). As implied by imprecise excision, the acquisition of DNA of LaiGI I to the right end of LaiGI I is accompanied by the loss of host DNA from the left end of LaiGI I. Taken together, our data suggest that LaiGI I is preferentially in an excised state and that both normal and abnormal excisions of LaiGI I occur in L. interrogans serovar Lai.
We were interested in determining whether this genomic island is restricted to L. interrogans serovar Lai or is also found in other pathogenic members of the genus. Hybridization analysis demonstrated that sequences highly related or identical to those of this genomic island exist in other L. interrogans strains, i.e., serovars Australis, Bataviae, Canicola, Hebdomadis, and Pyrogenes (Table (Table2).2). Transduction, or bacteriophage-mediated gene transfer, is thought to play an important role in the dissemination of genomic islands in other organisms. However, none of the LaiGI I CDSs were found to be homologs of phage-like structural proteins. In addition, genes of LaiGI I are not clearly clustered in large transcription units of related or interacting proteins as in most phage genomes (see the supplemental material). It is possible that the phage genes that are essential for the viral life cycle are not fully functional, so the phage cannot be propagated. This element may also be transferred, if it can really be transferred, to a new host by another mechanism such as conjugation.
In conclusion, the 54-kb DNA region described here fulfills most of the criteria proposed for genomic islands (4): the overall G+C content is lower than that found in the rest of the L. interrogans chromosome, its size is >30 kb, genes encoding mobility enzymes are present, and the LaiGI I flanking sequences contain direct repeats and insertion sequences. In addition, while the genome of L. interrogans serovar Lai was originally described as containing two circular chromosomes (20), we show that L. interrogans serovar Lai also possesses a circular plasmid. It is well established that integrated phage genomes or genomic islands are responsible for conferring new traits that can allow an adaptation of the bacteria to new environment/host conditions (4). For example, LaiGI I contains genes encoding putative regulators and membrane proteins (see the supplemental material) that may play an important role in the physiology of the bacterium. Interestingly, this island contains a high proportion of novel genes. The determination of the function of these novel genes will likely provide insights into the role of this genomic island in L. interrogans serovar Lai.
Only recently, the first evidence of gene transfer was demonstrated in L. interrogans by transposition of Himar1, a transposon of eukaryotic origin (3). The putative integrase (encoded by LA1768) and the replication protein (encoded by LA1839) from LaiGI I can be developed as novel tools for the genetics of Leptospira. A better understanding of the molecular biology of this genomic island not only will allow the development of genetic tools but also may lead to improvements in our understanding of virulence and gene dissemination in Leptospira spp.
This work was supported by the French Ministry of Research “ANR Jeunes Chercheurs” (no. 05-JCJC-0105-01).
We thank L. Frangeul for his help in the drawing of the circular map of LaiGI I and I. Saint Girons for her encouragement.
Editor: J. B. Bliska
Published ahead of print on 21 November 2006.
†Supplemental material for this article may be found at http://iai.asm.org/.