Prodromal pertussis symptoms developed on March 4, 2013, in an 11-month-old white, non-Hispanic infant from New York State while the family was traveling outside the state. Cough reportedly began on March 14, 2013, and 12 days later (March 26) he was brought to his healthcare provider (HCP) with symptoms consistent with pertussis. Since the child’s birth, the diagnosing HCP had seen the child only once; no visits to other HCPs were known. Per parental report, the case-patient was experiencing paroxysmal cough, apnea, and posttussive vomiting. No thoracic radiograph was obtained. A 5-day course of oral azithromycin was prescribed; the parent reported that the infant received treatment for 3 consecutive days, beginning March 26, 2013. The infant was not reported to have any pertussis-associated complications (seizures, pneumonia, or encephalopathy) and had only light coughing as of April 11, 2013.

The infant was unvaccinated because the parents refused administration of all vaccines. Three siblings, ages 12, 10, and 8 years, lived with the infant and were undervaccinated; they had received 2, 1, and 3 doses, respectively, of pertussis-containing vaccines. No coughing illness was reported among the siblings. The mother reported that she received Tdap vaccine during her pregnancy with the case-patient, but receipt of vaccine could not be verified.

A nasopharyngeal swab specimen was collected from the infant on March 26, 2013, for testing at a commercial laboratory. The isolate was also forwarded to New York State’s public health laboratory, the Wadsworth Center, where it was found to be positive for B. pertussis by PCR targeting IS481 and BP283 (5). Both laboratories yielded positive culture results for B. pertussis. No other testing was performed.

The Wadsworth Center forwarded the isolate, designated I979, to the Centers for Disease Control and Prevention (CDC; Atlanta, Georgia, USA) for confirmatory identification and molecular typing as part of the Enhanced Pertussis Surveillance program (6). PCR amplification of the gene encoding the first subunit of Pt (ptxA) was unsuccessful while the CDC multitarget real-time PCR diagnostic assay was performed (7). Amplification of the promoter region (ptxP) and ptxA was also unsuccessful during multilocus sequence typing targeting acellular vaccine component genes ptxA, ptxP, prn, and fim3 (8).

Further characterization of I979 and French strain FR3749 (4) was undertaken by multilocus sequence typing, multilocus variable-number tandem-repeat analysis, pulsed-field gel electrophoresis (9), and whole-genome sequencing. Long sequencing reads were obtained with the Pacific Biosciences RS II (Menlo Park, CA, USA) at >120× coverage and assembled de novo into a single contig by using HGAP v3 and Quiver v1 (Pacific Biosciences). Assembly structure was confirmed with a genome optical map after restriction digestion with KpnI (OpGen, Gaithersburg, MD, USA). The final sequence was polished with short reads obtained with Illumina MiSeq and CLC Genomics Workbench v7.5.1 (QIAGEN, Valencia, CA, USA) with >90× coverage. Completed genomes were submitted to the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) with GenBank accession nos. CP010965 (FR3749) and CP010966 (I979). Basic genome metrics are listed in Table 1.

Figure 1

Figure 1. Western blot of pertussis toxin (Pt) expression in Bordetella pertussis Tohama I, I979, FR3749, and 3 additional recent isolates. All isolate lanes were loaded with 10-μg of protein, extracted after growth...

Prn production was determined by ELISA (2,11). Pt production was examined through Western blot analysis of cultures grown in cyclodextrin-modified Stainer-Scholte liquid medium to optical density (OD) 600 nm = 0.1 (12). Proteins precipitated with trichloroacetic acid were washed, reduced, and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Pt detection is described in Figure 1.

I979 and FR3749 share the same multilocus variable-number tandem-repeat analysis type 27 and prn-2 genotype, the most common recent types (3,13). I979 and FR3749 are fim3–1 and fim3–2, respectively. The fim3 locus has fluctuated between these 2 alleles recently (3,8). Pulsed-field gel electrophoresis indicated that I979 displays profile CDC306, and FR3749 displays CDC046. I979 and FR3749 both lacked Pt production, as shown by Western blot (Figure 1). I979 also failed to produce Prn, whereas FR3749 was positive for Prn production by ELISA, within the range of negative (OD 0.3–0.6) and positive (OD 1.2–1.6) controls.

Figure 2

Figure 2. Map of the 28-kb region deleted in Bordetella pertussis strains I979 and FR3749, compared with vaccine strain Tohama I. Vertical lines indicate the deletion boundaries, at a nucleotide G of the...

Comparison of assembled I979 and FR3749 genomes with that of Tohama I (GenBank accession no. NC_002929.2) (10) indicated that the entire ptx/ptl operon is missing as the result of a putative deletion spanning 28,040 bp (Figure 2). Both genomes contain a conserved, truncated IS481 immediately upstream of the deletion and a single IS481 (FR3749) or 2 tandem IS481 sequences (I979) immediately downstream (Figure 2). Within Tohama I, the region absent from I979 and FR3749 encodes 30 predicted genes bound by 2 NCATGN motifs, the target sequence for IS481 insertion (Table 2). The insertion element IS1002 is located within the 3′ end of this region, and shares a GCATGG motif with IS481 immediately downstream. The 3′ deletion boundary is between IS1002 and IS481 (Figure 2). Whole-genome alignment, using progressiveMauve (14), of I979 and FR3749 with Tohama I revealed structural variation through genomic rearrangements and inversions. In particular, I979 and FR3749 genomes differ by a single, large inversion, the coordinates of which correspond to 2 conserved insertions of IS481 in opposing orientations (Technical Appendix Figure). I979 and FR3749 differ by 31 single nucleotide polymorphisms, each differing from Tohama I by 204 and 173 single nucleotide polymorphisms, respectively. FR3749 contains wild-type prn at position 1613, whereas I979 prn contains an IS481 insertion, the most common cause of Prn-deficiency, at position 1613 (2,11).

B. pertussis strain I979, identified in our study, is both Prn- and Pt-deficient. Loss of Pt in B. pertussis is a rare occurrence; only 2 isolates have been documented in 8 years. Both I979 and FR3749 were isolated from unvaccinated infants (11 months and 3 months old, respectively), who exhibited typical pertussis symptoms, although FR3749 had difficulty colonizing and multiplying in respiratory tracts of adult mice (4). B. pertussis isolates with deletions at other sites across the genome, including part or all of prn, were reported previously (4,15). During the past 5 years, US B. pertussis isolates have become nearly 100% Prn-deficient (2,3) (unpub. data), and Prn-deficient isolates have been obtained from vaccinated persons (11). The loss of Pt may represent a higher fitness cost to B. pertussis than the loss of Prn. In addition, the possibility that only the Pt-deficient isolates were recovered from patients who were co-infected with wild-type and mutant B. pertussis cannot be discarded. Further testing in models to understand the clinical relevance of Prn- and Pt-deficient strains in vaccinated and unvaccinated persons is warranted.

Although incidence of combined Pt- and Prn-deficiency in B. pertussis is rare, any increased mutation in these or other acellular vaccine immunogens may have serious implications for the efficacy of current vaccines. Global epidemiologic, culture-based, and molecular-based monitoring of B. pertussis is critical for understanding current trends of the disease it causes.