The strain described here (Pond1) was first isolated in Pondicherry, India, in November 2002 from the blood of a 23-year-old man admitted with fever and with no history of having received antimicrobial chemotherapy (6). The isolate was resistant to ciprofloxacin and nalidixic acid, and the MIC of ciprofloxacin was 8 mg/L. It was sensitive to all other antimicrobial drugs tested by disk diffusion: ampicillin, chloramphenicol, cotrimoxazole, gentamicin, and ceftriaxone. Repeat testing showed that zones of inhibition indicating susceptibility were seen around both ofloxacin (17 mm) and ciprofloxacin (21 mm) 5-μg disks; however, in light of the ciprofloxacin MIC and resistance to nalidixic acid, Pond1 was considered resistant to fluoroquinolones. No similar isolates have been seen in this area since the initial report, although the total number of paratyphoid fever cases has increased.

Polymerase chain reaction (PCR) amplification (Table 1) and direct DNA sequencing of both strands of the full length of gyrase (gyrA and gyrB) and topoisomerase IV (parC and parE) subunit genes was performed with an ABI Prism dye terminator cycle sequencing kit (Perkin Elmer, Foster City, CA, USA) on an ABI 3730 automated sequencer. Results showed 3 mutations: 2 in gyrA and 1 in parC. A comparison of these mutations with those previously described in fluoroquinolone-resistant S. Paratyphi A is shown in Table 2. A rise of fluoroquinolone resistance over time is apparent, and although the point mutations do not fully explain the MIC data, we noted general associations: a single mutation in gyrA is always associated with resistance to nalidixic acid and reduced susceptibility to ciprofloxacin and ofloxacin, and a double mutation in gyrA is always accompanied by mutations in parC and is associated with high-level (>4 μg/mL) resistance. Although this sample is limited, heterogeneity in mutations found at all sites suggests that this finding is not the result of a single successful strain's sequentially acquiring mutations but rather that resistance is arising in several different strains.

The QRDR within topoisomerases contains hotspots for mutations around the active site, which are associated with raised MIC values for fluoroquinolones (10). For gyrA from nalidixic acid–resistant Salmonella isolates, 2 mutations are most frequently observed in clinical isolates: Ser83→Phe and Ser83→Tyr (8,9,11). The association with resistance of mutations seen in parC, however, is less clear (12). Mutations in parC of gram-negative bacteria are usually within the QRDR at amino acids 80 and 84 (Ser 80→Ile, Glu 84→Gly, Lys). The first reported mutation was Thr57→Ser, and other mutations have been described: Asp 79→Ala, Gly 78→Asp (9). Each mutation, in gyrA or parC, can give rise to different MICs in different isolates, which means that other factors must also influence the resistance phenotype in S. Paratyphi A. The most likely cause is changes in expression levels of proteins involved with permeability barriers and efflux pumps. These changes could be the result of either point mutations in transcription promoters and regulators or downstream effects of mutations in topoisomerases. Known mechanisms of fluoroquinolone resistance that we were able to screen for include transmissible plasmidborne resistance and efflux pumps. The qnr-containing plasmid was not detected with PCR using primers 5´ GGG TAT GGA TAT TAT TGA TAA 3´ and 5´ CTA ATC CGG CAG CAC TAT TA 3´ (13) in Pond1, and sensitivity testing gave the same zone size around both tetracycline and chloramphenicol discs when compared with a nalidixic acid–susceptible isolate. This finding, combined with the ciprofloxacin MIC of 8 μg/mL, argues against the presence of multiple antimicrobial drug resistance efflux pumps. Thus the high-level resistance seen in Pond1 appears to be associated directly with 3 point mutations in the topoisomerase genes.

To confirm the identity of the isolate as S. Paratyphi A, we used MLST. The primer sequences and MLST data are available from the MLST database at the Max-Planck-Institut für Infektionsbiologie (http://web.mpiib-berlin.mpg.de/mlst/). Six of 7 sequenced loci matched exactly the previously described sequences, and 1 was a unique allele. This finding means that the isolate described here is within the clonal MLST group described as S. Paratyphi A but is a recognizable variant. This finding supports previous typing data that show very little variation (14); typing S. Paratyphi A is problematic because genomic restriction analyses (pulsed-field gel electrophoresis) of isolates from an outbreak are not always identical, and susceptible and resistant strains cannot be differentiated. For molecular epidemiologic studies to be carried out, several methods need to be used (15). A broader study of single base-pair differences between strains of S. Paratyphi A could provide a usable typing scheme.

Resistance in S. Paratyphi A populations must be monitored because the acquisition of resistance to fluoroquinolones, coupled with the reduction in S. Typhi by the use of typhoid-specific vaccination, may cause S. Paratyphi A to become the main cause of enteric fever. Disc susceptibility testing does not always detect resistance, and screening with nalidixic acid and MIC testing remains the method of choice. The isolate described here, Pond1, contains a unique combination of mutations that provides a way to track the spread of this strain of S. Paratyphi A.