|Home | About | Journals | Submit | Contact Us | Français|
Fluoroquinolone (FQ) resistance of Bacillus anthracis is a serious concern in the fields of biodefense and bioterrorism since FQs are very effective antibiotics and are recommended as first-line treatment against this lethal bacterium. In this study, we obtained 2 strains of B. anthracis showing resistance or intermediate resistance to ciprofloxacin (CIP) by a stepwise selection procedure with increasing CIP concentrations. Fifteen genetic variations were identified between the parental and CIP-resistant strains by next-generation sequencing. Nonsynonymous mutations in the quinolone resistance-determining region (QRDR) of type II DNA topoisomerase were identified in the resistant strain but not in the intermediate-resistant strain. The GBAA0834 (TetR-type transcriptional regulator) locus was also revealed to be a novel “mutation hot spot” that leads to the increased expression of multidrug efflux systems for CIP resistance. As an initial step of CIP resistance in B. anthracis, such disruptive mutations of GBAA0834 appear to be more easily acquired than those in an essential gene, such as that encoding type II DNA topoisomerase. Such an intermediate-resistant phenotype could increase a cell population under CIP-selective pressure and might promote the emergence of highly resistant isolates. Our findings reveal, in addition to QRDR, crucial genetic targets for the investigation of intermediate resistance of B. anthracis to FQs.
Anthrax caused by the spore-forming bacterium Bacillus anthracis is one of the most severe zoonoses and poses a serious threat to both public and animal health (9, 15). B. anthracis belongs to the Bacillus cereus group of bacteria, which comprises closely related Gram-positive organisms with highly divergent virulent properties (15, 20). Infection with this bacterium can occur through the skin, gastrointestinal tract, or respiratory apparatus following contact, ingestion, or inhalation of spores, respectively (9, 15). The recent “postal anthrax attack” in the United States that aimed to intentionally release B. anthracis spores underlines the growing importance of the identification of B. anthracis at the strain level in forensic and epidemiological investigations (3, 21).
Fluoroquinolone (FQ) resistance is a major concern in medical treatment following anthrax bioterrorism because FQs are first-line antibiotics for the treatment of B. anthracis infection (9, 27). FQs act as broad-spectrum bactericidal antibiotics by inhibiting type II DNA topoisomerases, DNA gyrases (GyrA and GyrB), and type IV DNA topoisomerases (ParC and ParE). The mechanism responsible for FQ resistance has been well documented with bacteria, in which frequent mutations of topoisomerase genes have been identified and designated the quinolone resistance-determining region (QRDR) (22, 23), the most fundamental region for a primary investigation of FQ resistance.
Recent studies have reported several mutations in QRDR in in vitro-selected FQ-resistant strains of B. anthracis (2, 8, 18). Aside from QRDR, multidrug resistance proteins are reportedly involved in the development of FQ resistance by scavenging intracellular FQ (13, 22, 23). In fact, recent reports have also suggested a possible contribution of multidrug efflux pumps to FQ resistance in B. anthracis (2, 18). However, the responsible efflux pump has not yet been identified, and thus, mutations responsible for additional resistance mechanisms remain to be elucidated. Recently, next-generation sequencing has facilitated the identification of genomewide single nucleotide polymorphisms (SNPs) and insertions-deletions (indels) (4, 11, 28). Using such a method, the present study reveals novel genetic variations responsible for FQ resistance in B. anthracis.
Japanese isolates of B. anthracis BA103 or BA104 were used as the parental strains. The BA103 strain was isolated from beef cattle in 1991, while BA104 was isolated from pigs in a sporadic incident in 1982. CIP-resistant mutants were isolated by serial passage on Muller-Hinton (MH) agar supplemented with 2-fold stepwise increasing concentrations of ciprofloxacin (CIP) (Wako, Osaka, Japan), from 0.0625 to 16 mg/liter CIP at 37°C. The MIC to antibiotics was determined using Etest (AB Biodisk, Solna, Sweden) according to the manufacturer's instructions.
The handling of live B. anthracis was performed at biosafety level 3 (BSL-3) according to the Regulations on the Safety Control of Laboratories Handling Pathogenic Agents (SCLHPA) authorized by the biorisk management committee of the National Institute of Infectious Diseases (NIID), Japan. The management of group 2 agents, such as B. anthracis, was also regulated and required obligatory special permission, special biosafety facilities, and other measures for its possession and use as stated in the Law Concerning the Prevention of Infectious Diseases and Medical Care for Patients of Infections (the Infectious Diseases Control Law) by the Ministry of Health, Labor, and Welfare.
Library preparation was performed using a genomic DNA sample prep kit (Illumina, San Diego, CA), and DNA clusters were generated on a slide using the cluster generation kit (version 2) on an Illumina cluster station (Illumina), according to the manufacturer's instructions. To obtain ~10 million clusters for one lane, the general procedure as described in the standard recipe (Illumina) was performed as follows: template hybridization, isothermal amplification, linearization, blocking, denaturation, and hybridization of the sequencing primer (Illumina). All sequencing runs were performed with the Illumina genome analyzer II (GA II) using the Illumina sequencing kit (version 3). Fluorescent images were analyzed with the Illumina base-calling pipeline 1.3.2 to obtain FASTQ formatted sequence data.
A schematic flow chart of the data processing procedure is shown in Fig. S2 in the supplemental material. To identify specific SNVs/indels compared with the reference sequence of B. anthracis Ames 0581 (GenBank accession number NC_007530), Maq software (version 0.7.1) (12), a mapping assembler for short reads generated by the next-generation sequencer, was used with the easyrun Perl command as a default parameter. Read alignment for the validation of SNVs/indels was performed using the MapView graphical alignment viewer (1). Strain-specific SNVs/indels were extracted from the “cns.final.snp” or “cns.indelse” file, respectively, by comparison of the parental or CIP-resistant strains to Ames 0581. SNVs located in repetitive sequence regions (e.g., variable-number tandem repeat [VNTR], rRNA, and insertion sequence [IS]) were excluded from the analysis. Regarding the reliable detection of indels, the cns.indelse file, which includes potential indels showing abnormal alignment patterns, was processed by the recommended filtering technique to reduce the number of false positives, as described in the manual. Furthermore, we analyzed the 100 nucleotides surrounding the indels with a BlastN search against all sequenced short reads (parameter −F F −e 1.0E−10 −m 3). Putative SNVs/indels were finally verified by Sanger sequencing using the BigDye terminator version 3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA) (see Fig. S6 in the supplemental material). PCR and sequence primers are listed in Table S2 in the supplemental material.
For the preparation of total RNA from B. anthracis, bacterial cells were cultured in brain heart infusion broth at 37°C to the mid-log phase, and then total RNA was extracted with Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Total RNA was treated with the turbo DNA-free kit (Ambion, TX). Quantitative reverse transcription-PCR (qRT-PCR) was performed using 100 ng of total RNA, gene-specific primers designed by PrimerQuest (see Table S3 in the supplemental material), and the SuperScript III Platinum SYBR green one-step qRT-PCR kit with ROX (Invitrogen) and analyzed using the ABI Prism 7900HT real-time PCR system (Applied Biosystems). We used the following qRT-PCR program: RT reaction, 50°C for 3 min; initial denaturation, 95°C for 5 min; and 2 steps of amplification (40 cycles), 95°C for 15 s and 60°C for 30 s. The rplB gene encoding the 50S ribosomal protein L2 was used as the internal control.
The MIC was determined using a CIP Etest (AB Biodisk) on MH agar containing 10 mg/liter reserpine (Sigma-Aldrich, St. Louis, MO) (14). Dimethyl sulfoxide was used as the control solvent.
The short-read archives have been deposited in the DNA Data Bank of Japan (DDBJ; accession numbers DRA000067, DRA000068, DRA000069 and DRA000070 for BA103, BA104, BA103-CIPr, and BA104-CIPr, respectively).
Both BA103 and BA104 were susceptible to CIP at 0.064 mg/liter, the MIC as determined using Etest (Fig. (Fig.11 D). Resistant mutants were identified by repeating the selection procedure until a final concentration of 16.0 μg/ml CIP. Two CIP-resistant mutants, BA103-CIPr (MIC, 1.5 mg/liter) and BA104-CIPr (MIC, 6 mg/liter), were isolated from BA103 and BA104, respectively. Their CIP resistance levels increased by 23- and 94-fold, respectively (Fig. (Fig.1D).1D). The two selected mutants also showed reduced susceptibility to another quinolone antibiotic, levofloxacin (see Fig. S1 in the supplemental material): for BA103 and BA104, the MIC was 0.047 mg/liter; for BA103-CIPr, 0.125 mg/liter; and for BA104-CIPr, 0.50 mg/liter. Cross-resistance to other antibiotics was not observed with the selected mutants: for penicillin, the MIC was 0.023 mg/liter; for tetracycline, <0.016 mg/liter; and for amikacin, 0.25 mg/liter (see Fig. S1 in the supplemental material).
The Illumina GA II sequencer produced 6.85 to 10.31 million 50-base-long reads per strain after applying the quality filter of the Illumina base-calling pipeline 1.32 (see Table S1 in the supplemental material). More than 94.4% of the filter-passed reads were aligned to the reference sequences of the Ames 0581 strain using Maq software, resulting in more than a 56-fold coverage depth on average and a 92.7 to 99.34% coverage of the sequence, excluding ambiguous repetitive regions (Table S1 in the supplemental material). Since genomic analysis of laboratory strains of B. subtilis using Illumina GA II was performed at a maximum coverage of 51.7-fold (26), the coverage in the present study would be sufficient for the identification of SNVs and indels.
At the first screening step, with reads mapping to Ames 0581, 93 possible SNVs were found in BA103 and 85 in BA103-CIPr (see Fig. S3A in the supplemental material). The Venn diagram in Fig. S3A indicates that 16, 6, and 75 positions could be SNV candidates for BA103-specific, BA103-CIPr-specific, and common SNVs, respectively. Incidentally, these SNV candidates appeared to be wrongly identified due to a misalignment with Maq, because 5 of the 6 BA103-CIPr-specific SNV candidates were found to be incorrect by Sanger sequencing methods. In most cases of such incorrect extractions, mapped reads showed a lower coverage depth at a 50% coefficient of variation (CV) at SNV candidates than those around the SNV position. Thus, the original filtering method was performed by calculating the CV with a coverage depth value around 100 bp at each SNV candidate. This filtering enabled us to refine the potential SNVs and led to the identification of one BA103-CIPr-specific SNV (Fig. S3A).
This BA103-CIPr-specific SNV was at genomic position 842518 and was a C-to-T base substitution in the GBAA0834 gene, encoding the TetR family transcriptional regulator. This substitution resulted in an amino acid substitution of threonine to isoleucine at the 38th amino acid (Thr38Ile) (Table (Table1;1; Fig. Fig.1A).1A). SNV was not detected in the pXO1 and pXO2 plasmids between BA103 derivatives.
Using the same filtering methods described above, we identified 3 BA104-CIPr-specific SNVs, which were verified by Sanger sequencing (see Fig. S3A in the supplemental material). Among these 3 BA104-CIPr-specific SNVs, an SNV at genomic position 6850 was a G-to-C nucleotide substitution resulting in an amino acid substitution (Ala86Pro) in the QRDR of the gyrA gene (GBAA0006) (Table (Table1).1). There are no previous reports of this SNV associated with CIP resistance in B. anthracis; however, the substitution has been found as Ala84Pro in Escherichia coli (25) and Ser85Pro in Staphylococcus aureus (10). The second SNV was located at genomic position 2138798 and was an A-to-G nucleotide substitution resulting in an amino acid substitution (Lys80Glu) in GBAA2291 encoding the major sporulation sensor histidine kinase. The third SNV was located at genomic position 4595343 and was a substitution of T to C in the intergenic region between GBAA5072 and GBAA5074. SNV was not detected in the pXO1 and pXO2 plasmids between BA104 derivatives.
At the first screening step, with reads mapping to B. anthracis Ames 0581, 7 possible short indels were identified in BA103 and 10 in BA103-CIPr (see Fig. S3B in the supplemental material). The Venn diagram in Fig. S3B shows that 4, 7, and 3 positions could be candidates for BA103-specific indels, BA103-CIPr-specific indels, and common indels, respectively. Further verification with a BlastN search around the positions of these indels refined 2 potential BA103-CIPr-specific indels (Table (Table2).2). These were validated by Sanger sequencing (see Fig. S6 in the supplemental material). The first short indel, located at genomic positions 842291 to 842295, was a deletion of 5 nucleotides (TAACA) in BA103-CIPr. This indel was located in the 132-bp intergenic region between GBAA0833 and GBAA0834 (Fig. (Fig.1A;1A; see also Fig. S4A in the supplemental material). The second short indel was located at genomic positions 5215555 to 5215556 and was an insertion of 2 nucleotides (CA) within GBAA5724. This indel caused a frameshift in the coding sequence, resulting in the truncation of the amino acid sequence from amino acids (aa) 366 to 197 (see Fig. S5 in the supplemental material). Indel was not detected in the pXO1 and pXO2 plasmids between BA103 derivatives.
Using the filtering methods described above, we identified 9 BA104-CIPr-specific indels (Table (Table2).2). The first short indel was located at genomic positions 336869 to 336873 and was a deletion of 5 nucleotides (ACTTA) in GBAA0328, encoding a small multidrug resistance (SMR) family multidrug efflux pump. This indel caused a frameshift mutation leading to an increase in the coding sequence from aa 107 to aa 114 (see Fig. S5 in the supplemental material). The second short indel was located at genomic position 842291 and was a single nucleotide (T) deletion in the 132-bp intergenic region between GBAA0833 and GBAA0834 (Fig. (Fig.1A).1A). Intriguingly, this deletion overlapped the first BA103-CIPr indel (Table (Table2;2; see also Fig. S4A in the supplemental material). Moreover, the third indel was located at genomic position 842715 and was a single nucleotide (C) insertion within the GBAA0834 TetR family regulator, the same as that mentioned above, leading to a truncation of the amino acid sequence from aa 191 to aa 117 (Fig. (Fig.1A;1A; see also Fig. S4C in the supplemental material). The other 6 indels appeared to be generated by tandem duplications of 2 to 6 nucleotide bases, as summarized in Table Table2.2. These indels resulted in a frameshift causing the truncation of the coding sequences of GBAA1077, GBAA2814, and GBAA4143 (see Fig. S5 in the supplemental material). On the other hand, the tandem duplications of 6 bases in GBAA3656 (parC) and GBAA5329 resulted in the incorporation of an additional 2 amino acids without frameshift mutation (Fig. S5).
Three genes of the putative multidrug efflux system, GBAA0832, GBAA0833, and blt (GBAA0835), are located adjacent to GBAA0834 (Fig. (Fig.1A).1A). We performed qRT-PCR to determine whether the TetR-type regulator (GBAA0834) is involved in the regulation of the multidrug efflux systems. All 4 genes were significantly upregulated in both BA103-CIPr and BA104-CIPr compared with their respective parental strains (Fig. 1B and C).
To evaluate whether multidrug efflux systems are involved in CIP resistance, the MIC was determined with reserpine, which acts by blocking efflux ability (14). Reserpine reduced the MIC by 4-fold in BA103-CIPr and 8-fold in BA104-CIPr (Fig. (Fig.1D),1D), whereas it reduced the MIC by 1.3-fold in the parental strains.
Genetic variations in the QRDR of type II topoisomerases, the most common factors involved in CIP resistance, have been well characterized. It has been suggested, however, that a certain efflux system should contribute to the intermediate resistance to FQs. We performed a genomewide search using next-generation sequencing technology and identified novel genetic elements for resistance, in addition to the action of the QRDR.
BA103-CIPr and BA104-CIPr were classified as “nonsusceptible” isolates for CIP resistance according to the MIC interpretive standards (6); however, the CIP MICs for these strains are notable values for resistance in clinical treatment. A nonsynonymous mutation (Ala86Pro) of the gyrA gene in BA104-CIPr appeared to contribute to resistance, because the substitution was also reported as Ala84Pro in E. coli (25) and Ser85Pro in S. aureus (10). No variation in the QRDR was found in BA103-CIPr, suggesting that other novel genetic variations were involved in its resistance.
Three genetic variations, including 1 SNV and 2 indels, were found in the intermediate-resistant strain BA103-CIPr; in fact, 2 of the 3 mutations were located around the TetR-type transcriptional regulator (GBAA0834). Multidrug efflux pumps are highly expressed mainly as a result of mutations of either their regulatory cis element or their transcriptional regulator genes (7). It appears that these 2 mutations could be associated with the CIP resistance of BA103-CIPr. A nonsynonymous mutation (Thr38Ile) of GBA0834 was located in the conserved region of the “HTH_TETR_2” domain involved in DNA binding, in which the corresponding amino acid residue Thr40 of TetR in E. coli comes directly into contact with the cis element (17, 19). This threonine residue is well conserved in 34% of TetR family proteins, and thus, the substitution could cause a reduced DNA-binding affinity to a regulatory element.
The TetR family regulator is likely to modulate genes adjacent to the regulator gene (19). GBAA0832 and GBAA0833 are possibly involved with the small multidrug resistance protein (SMR family), and Blt (GBAA0835) shows a high amino acid sequence similarity with NorA of S. aureus, which is well characterized as a major efflux pump involved in FQ resistance (16). Indeed, our qRT-PCR data suggest the increased expression of Blt, GBAA0832, and GBAA0833 (Fig. 1B and C); in addition, the reserpine inhibition test indicates the contribution of the multidrug efflux system in CIP resistance (Fig. (Fig.1D).1D). Thus, the functional disruption of GBAA0834 increased the transcription of these multidrug resistance genes, leading to a reduced susceptibility to CIP.
Twelve genetic variations, including 3 SNVs and 9 indels, were identified in the potentially resistant strain BA104-CIPr; 1 SNV was located in the gyrA gene and was discussed above. Two of the 9 indels were located around GBAA0834; one was an insertion of a C nucleotide in the “tetracycline repressor-like, C-terminal” domain involved in dimerization and ligand binding, and another was a deletion located upstream of GBAA0834. We propose that this insertion resulted in the disruption of GBAA0834, leading to the increased transcription of the 3 adjacent putative multidrug resistance genes, as was observed for BA103-CIPr.
Regarding the synonymous or nonsynonymous SNVs, all 3 SNVs in the coding sequence of the genes caused nonsynonymous mutations (Table (Table1),1), while surprisingly, synonymous mutations were not found throughout the whole genome. A transversional substitution, which is supposed to occur at a lower frequency than transitional substitutions, was found in the gyrA gene of BA104-CIPr, implying that the detection of such a rare mutation might be the result of stringent selection pressure, such as CIP selection for adaptive mutations (5, 24).
Regarding the detection of short indels, it is recommended to use paired-end sequencing reads rather than single-end reads. Although single-end sequencing reads were used in this study, 11 novel short indels could be identified with a combination of Maq mapping, BlastN search, and Sanger sequencing. Our strategy of using short single-end reads is applicable to the identification of SNVs and short indels as strain-specific genetic markers. Variable-number tandem repeats (VNTRs) were too long to determine indels with short-read mapping. Thus, 25 VNTR loci were analyzed by multiple-locus VNTR analysis (MLVA), and a single locus (pXO2) of BA104-CIPr was found to be extended by 2 repeat units (4 bp) compared with that of BA104 (see Table S4 in the supplemental material). FQs affect DNA integrity by blocking DNA topoisomerases, suggesting that they might cause an increase in indel variation. Such unstable VNTR loci may be an unsuitable genetic marker for the traceability of B. anthracis strains.
Taken together, our findings with next-generation DNA sequencing technology enabled us to identify 15 novel genetic variations associated with CIP resistance. At this point, we can conclude that the GBAA0834 locus is a novel “mutation hot spot” related to resistance, in addition to the QRDR. Type II topoisomerase is essential; thus, genetic variation around GBAA0834 could be easily acquired as a primary step toward the development of resistance in B. anthracis. Moreover, such an intermediate resistance could increase the probability of the emergence of highly resistant isolates. Although the QRDR is the most reliable target for investigating FQ resistance, our findings will provide crucial genetic information for the detection of nonsusceptible isolates between isolates intermediate and fully resistant to CIP.
We thank Tadahito Kanda and Akio Yamada for valuable suggestions.
This work was supported by a Research on Emerging and Re-emerging Infectious Diseases grant (H20 and 21 Shinko-Ippan-6) from the Ministry of Health, Labor, and Welfare, Japan.
Published ahead of print on 12 April 2010.
†Supplemental material for this article may be found at http://aac.asm.org/.