INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Leptospirosis is a widespread zoonosis caused by Leptospira interrogans. This bacterial pathogen can infect most mammalian species through either direct or indirect contact with contaminated body fluids from an infected animal (7). Leptospirosis can be fatal in humans. In livestock, Leptospira infection may result in death or a chronic infection may ensue, leading to abortion, stillbirth, infertility, or decreased milk production. Leptospira-infected humans are often treated with -lactam antibiotics. It has recently been suggested that leptospirosis in livestock can be treated with -lactam antibiotics (19). In vitro, pathogenic leptospires are very sensitive to -lactam antibiotics (16). The MIC of ampicillin is between 0.025 and 0.78 µg/ml, and that of penicillin G is between 0.39 and 3.13 µg/ml. The minimal bactericidal concentrations observed for penicillin G are up to 100 µg/ml or more. In contrast, ampicillin exhibits high bactericidal activity, as evidenced by low minimal bactericidal concentrations (<25 µg/ml).

-Lactams exert their effects by acting as substrate analogs of the peptidoglycan biosynthetic enzymes transpeptidase and D-alanine carboxypeptidase (21). These enzymes are located within the cytoplasmic membrane and play an integral role in the synthesis of peptidoglycan. These proteins are commonly called penicillin-binding proteins (PBPs) because of their ability to covalently bind radiolabeled penicillin (20). There are two distinguishable groups of PBPs: low-molecular-weight PBPs and high-molecular-weight (HMW) PBPs. The low-molecular-weight PBPs are monofunctional enzymes acting as DD-carboxypeptidases involved in the remodeling of peptidoglycan during cell growth. The HMW PBPs have a multidomain structure. These proteins are anchored to the cytoplasmic membrane by an N-terminal pseudo-signal peptide and are essentially composed of two modules localized on the outer face of the cytoplasmic membrane. The N-terminal domain, which is several hundred amino acids long, is fused to the C-terminal penicillin-binding domain. This domain displays the transpeptidase activity that catalyzes cross-linking of the peptidoglycan peptides. Pairwise comparison and multiple alignments of amino acid sequences lead to the conclusion that HMW PBPs fall into two classes, A and B, which differ in their N-terminal domain (8, 13). In Escherichia coli, PBPs 1a and 1b of class A behave as bifunctional proteins exhibiting both transglycosylase (N-terminal module) and transpeptidase (C-terminal module) activities. They catalyze polymerization of the peptidoglycan from undecaprenyl diphosphate-linked disaccharide peptides, probably by producing primers for PBP 2 and PBP 3 to act upon during cell elongation and cell division. PBP 2 and PBP 3 of class B are likewise considered bifunctional proteins, though the role of the N-terminal module is not clearly established. PBP 3 is specifically involved in polymerization of the septal peptidoglycan during cell division (14). Little is known about the PBPs of Leptospira. During analysis of subcellular fractions, five PBPs were identified in Leptospira kirschneri (10). However, neither the proteins nor the genes that encode them have been characterized.

To establish a framework by which leptospiral peptidoglycan structure can be analyzed, we isolated and characterized the L. interrogans ponA and pbpB genes, encoding PBP 1 and PBP 3, respectively, which play an important role in peptidoglycan synthesis. Comparison of these sequences from two strains (serovar icterohaemorrhagiae strain Verdun and serovar pomona strain RZ11) also provides information on genetic drift between distinct serovars of the same species.

(This work represents a portion of a thesis submitted by Audrey Brenot to the University of Paris VII, Paris, France, for the Ph.D. degree.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains, plasmids, and growth conditions. Bacterial strains and plasmids used are detailed in Table 1. E. coli strains were grown at 37°C in Luria-Bertani broth (18)_. Antibiotics and substrates were used in selective media at the indicated concentrations: isopropyl--D-thiogalactopyranoside (IPTG) at 500 µM, 5-bromo-4-chloro-3-indolyl--D-galactoside at 80 µg/ml, ampicillin (AMP) at 50 µg/ml, and kanamycin at 30 or 50 µg/ml. L. interrogans serovar icterohaemorrhagiae strain Verdun (National Reference Center for Leptospira, Institut Pasteur, Paris, France) and serovar pomona strain RZ11 (24) were grown in EMJH medium at 28°C (6, 12).

Cloning and sequencing of the ponA and pbpB genes. The ponA gene of L. interrogans serovar pomona strain RZ11 was cloned from a previously described plasmid-based BamHI library (24) and identified during sequence analysis of randomly picked clones (4). The 3' end of this gene was amplified using a PCR-based genome walking technique (Universal Genome Walker Kit; Clontech, Palo Alto, Calif.) using the conditions described previously (25). Specific amplification of ponA was initiated with primer 173 (Table 2) on BstUI-digested DNA to which adapters had been ligated. The pbpB gene was amplified from strain RZ11 using primers 921 and 922 (Table 2), which were derived from the strain Verdun sequence. Amplicons were ligated with pCR2.1 vector (Invitrogen Corp., Carlsbad, Calif.) and used to transform E. coli INV F'. The resulting plasmid, p921-1, contained the pbpB gene downstream of the T7 promoter.

RT-PCR. The methods used to extract total genomic L. interrogans strain RZ11 RNA and perform reverse transcriptase PCR (RT-PCR) were described previously (25). Primers used to detect the strain RZ11 ponA transcript were 173 and 121 (Table 2). Total RNA from L. interrogans strain Verdun was prepared from 400 ml of culture at 10⁹ bacteria/ml (3) using Tri reagent (Sigma Chemical Co., St Louis, Mo.). Residual RNA was removed by treatment for 1 h at 37°C with RNase-free DNase (Pharmacia Biotech, 1 U/µg) and extracted using the RNeasy kit (Qiagen, Hilden, Germany). The RT-PCRs were done using the Access RT-PCR System (Promega Corp., Madison, Wis.) according to the manufacturer's recommendations with primers listed in Table 2.

LD-PCR. PCRs were used to determine the distance and orientation between the ponA and pbpB genes from strains Verdun and RZ11. Long-distance PCR (LD-PCR) products were amplified from strain RZ11 genomic DNA using Tth polymerase (Clontech) using the amplification parameters described previously (25) and primer 184, located downstream of ponA, paired with either primer 185 or 186, oriented in opposite directions within pbpB (Table 2). LD-PCR products were also amplified from a cosmid containing the strain Verdun ponA and pbpB genes with the Advantage 2 PCR kit (Clontech). For strain Verdun, the primers were designed to hybridize at the beginning and the end of both genes and were directed outward of these genes (Table 2). Two additional primers, oligo5.4UP and oligo5.4RP, which anneal to opposite strands of a 5.4-kb XbaI fragment located between ponA and pbpB genes, were also used in LD-PCR analysis of the strain Verdun locus. For amplification of the strain Verdun locus, the cycling parameters were as recommended by the supplier to amplify 10- to 20-kb templates.

Protein expression of pbpB and ponA in E. coli. The pbpB gene from strain Verdun was amplified by PCR using the PfuTurbo DNA polymerase (Stratagene, La Jolla, Calif.) using primers PBP3M and PBP3L (Table 2). Primers PBP3M and PBP3L (1 µM each) were added to 10 ng of pG-PBP3 plasmid DNA. This allowed amplification of the pbpB sequence from 95 nucleotides after the start codon to the stop codon (which corresponded to the periplasmic predicted part of the protein). The 1,737-bp PCR product was cloned into pCRII-TOPO, and E. coli TOP10F' cells were screened for lacZ inactivation as described by the supplier (Invitrogen Corp.). A 1,723-bp BamHI-XhoI insert containing the pbpB coding region was inserted into pET26b(+) (Novagen, Inc., Madison, Wis.), resulting in plasmid pET-PBP3. The resulting plasmid created a translational fusion of the pelB leader sequence with the predicted periplasmic part of PBP 3.

Identification of PBPs by labeling with DIG-AMP. Pellets of L. interrogans and E. coli harboring plasmids containing the pbpB and ponA genes and their vector controls were suspended in phosphate-buffered saline and sonicated. Aliquots of the sonicated cells (100 µg of protein) were incubated at 37°C for 10 min with 2.5 µg of AMP per ml conjugated to digoxigenin (DIG) as described by Weigel et al. (22). Of each sample, 12.5 µg was resolved by SDS-PAGE; PBPs were identified by immunoblotting with an anti-DIG-alkaline phosphatase conjugate (Boehringer Mannheim Corp., Indianapolis, Ind.) followed by chemiluminescence from CDP Star (Boehringer Mannheim Corp.). In all competition experiments, samples were incubated for 30 min with a 400-fold excess of free AMP (Sigma Chemical Co.).

Nucleotide sequence accession number. The pbpB and ponA sequences from strain Verdun have been assigned the EMBL accession no. AJ243720 and AJ278610, respectively. The ponA and pbpB sequences from strain RZ11 have been assigned the EMBL accession no. AF282906 and AF282907, respectively.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Characterization of the L. interrogans pbpB gene. Partial sequence analysis of a 1.5-kb fragment of L. interrogans serovar icterohaemorrhagiae strain Verdun showed that it contained part of a gene similar to those encoding bacterial HMW PBPs (4). This fragment was used to screen a cosmid library of strain Verdun genomic DNA by colony hybridization to isolate a complete copy of the gene and surrounding DNA. The region surrounding this putative PBP gene was subcloned and sequenced. One open reading frame (ORF) with the potential to encode a 602-amino-acid protein, having an estimated molecular mass of 67.3 kDa according to the compute pI/M_w tool (23), was identified. The protein sequence deduced from this ORF was used to search the GenBank and EMBL databases for homologs using BLASTP. This protein was most similar to several HMW PBPs, including the Bacillus subtilis stage V sporulation protein D; PBP 1 and PBP 3 of Borrelia burgdorferi and Treponema pallidum, respectively; cell division protein FtsI of Streptomyces coelicolor; PBP A2 of Rickettsia prowazekii; and PBP 3 of E. coli. Pairwise comparison revealed that L. interrogans protein shares about 30 and 26% sequence identity with the PBP 3 proteins from T. pallidum and E. coli, respectively. Because of the strong similarity to the gene encoding PBP 3, this gene was designated pbpB.

Characterization of the L. interrogans ponA gene. A plasmid clone, pKB1, containing a portion of the strain RZ11 ponA gene, encoding PBP 1, was identified during a study using sequence analysis of randomly selected clones to improve resolution of the combined physical and genetic map of L. interrogans (4). Plasmid pKB1 contains about two-thirds of the gene, including the 5' end. A genomic walking technique was used to amplify the 3' end of the gene using primer 173. The resulting 1,300-bp amplicon was cloned, generating plasmid pK127, and sequenced. The overlapping sequences of pKB1 and pK127 revealed the presence of a 2,409-bp ORF, with the potential to encode an 802-amino-acid protein with a predicted mass of 89.8 kDa according to the compute pI/M_w tool (23). The deduced protein was used to search the GenBank database using BLASTP. This sequence was most similar to those of HMW PBPs, with 25 and 26% of its amino acids identical to those of Neisseria gonorrhoeae and E. coli PBPs 1 and 1a, respectively. The L. interrogans gene was designated ponA because of its similarity to the E. coli ponA gene. A cosmid containing the strain Verdun ponA gene was identified by colony hybridization using a 1-kb ClaI fragment derived from pKB1 as a probe.

The pbpB and ponA genes are 8 to 10 kb apart and comprise individual transcription units. The cloned ponA and pbpB genes were previously localized on the L. interrogans strain RZ11 and Verdun physical maps by hybridization (4). These data showed that the two genes were located in the same region of the genome. However, the methodology used for mapping lacks detailed resolution, and thus, it could not be determined if these two genes were closely linked in the genome. The approximate distance between pbpB and ponA was determined using LD-PCR. The primers used for this analysis were located at the beginning and end of both genes and were directed outward toward flanking sequences (Table 2). Initial LD-PCR results showed that the two genes were about 8 (strain Verdun) and 10 (strain RZ11) kb apart and were in a convergent orientation for both strains. The distances between ponA and pbpB were confirmed for both strains. For strain Verdun, using two additional primers that anneal to a 5.4-kb XbaI fragment found between ponA and pbpB genes, the 8.5-kb distance between these two genes was confirmed. For strain RZ11, the 10.4-kb distance was confirmed, indicating that there may be a small insertion between ponA and pbpB in strain RZ11 compared to strain Verdun.

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Analysis of transcription of pbpB and ponA from L. interrogans strain Verdun by RT-PCR. A sample of each reaction mixture was analyzed on a 1.75% agarose gel with TBE (90 mM Tris, 90 mM borate, 2 mM EDTA [pH 8]) buffer. Lanes with even numbers correspond to negative controls without RT. The size of the amplified products is indicated. (A) RT-PCR for pbpB and adjacent regions. Four reactions were performed with primers internal to the pbpB ORF: primers 4 and 8 (lanes 1 and 2) and primers 12 and 29 (lanes 7 and 8). The other reactions were performed with primers outside the ORF in combination with primers internal to the ORF: primers 4 and 11 (lanes 3 and 4), primers 4 and 36 (lanes 5 and 6), primers 12 and 38 (lanes 9 and 10), and primers 12 and 39 (lanes 11 and 12). (B) RT-PCR for ponA and adjacent regions. Four reactions were performed with primers internal to the ponA ORF: primers 3 and Z157 (lanes 1 and 2) and primers 7 and Z195 (lanes 5 and 6). The other reactions were performed with primers outside the ORF in combination with primers internal to the ORF: primers 3 and 15 (lanes 3 and 4), primers 7 and 12 (lanes 7 and 8), and primers 7 and 17 (lanes 9 and 10). (C) Diagram showing the location of the RT-PCR primers and products of transcription of pbpB and ponA from L. interrogans strain Verdun. The primers (Table 2) are indicated by solid arrows. A thin line shows the presence of a transcript with its length in base pairs. A large boldface X interrupting a broken line indicates the absence of transcript.

PBP 1 and PBP 3 bind AMP. The L. interrogans PBP 1 and PBP 3 proteins resemble other HMW PBPs, and both proteins retain signature motifs associated with penicillin binding sites (14). Based on these similarities, we predicted that both proteins would covalently bind penicillin. As a first step, we wished to compare the sizes of PBPs in L. interrogans to those of the five PBPs identified for L. kirschneri (10). L. interrogans strain Verdun PBPs were detected by incubation of cell sonicates with DIG-AMP, separated by electrophoresis, and visualized. Several PBPs were detected with estimated masses of 89 (doublet), 64, 41, 32, and 20 kDa (data not shown). Several of the L. interrogans PBPs have estimated masses similar to those previously reported for L. kirschneri serovar grippotyphosa strain RM52 (i.e., 82-kDa doublet and 64, 59, and 33 kDa) (10). Differences were detected with the PBP 1 and 2 proteins (89 kDa for strain Verdun and 82 kDa for strain RM52), PBP 4 proteins (41 kDa for strain Verdun and 59 kDa for strain RM52), and strain Verdun PBP 6 (18 kDa) not reported for L. kirschneri.

View larger version (31K): [in this window] [in a new window]   FIG. 2.   Binding of DIG-AMP to L. interrogans PBP 1 and PBP 3. Lysates of E. coli cells harboring recombinant or vector plasmids were incubated with DIG-AMP, separated by SDS-PAGE, and transferred to a membrane, and the modified proteins were detected as described in Materials and Methods. (A) Production of strain Verdun PBP 3 from pET-PBP3, a recombinant plasmid in E. coli. pET-26b(+) is the expression vector (lane 1). pET-PBP3 is the recombinant plasmid carrying pbpB from strain Verdun (lanes 2 and 3). Binding of DIG-AMP was also assayed in the presence of free AMP (lane 3). The positions of E. coli PBP 5, PBP 6, and PBP 7 and L. interrogans strain Verdun PBP 3 are indicated on the left. The migration of size standards is indicated on the right (in kilodaltons). (B) Production of strain RZ11 PBP 1 from p513-3, a recombinant plasmid in E. coli. pCR2.1 is the expression vector (lane 1). p513-3 is the recombinant plasmid carrying ponA from strain RZ11 (lane 2). The migration of size standards is indicated on the right (in kilodaltons). L. interrogans strain RZ11 PBP 1 is labeled.