Journal of Antimicrobial Chemotherapy (2004) 53, 371–374
DOI: 10.1093/jac/dkh063
Advance Access publication 16 January 2004
BRO β-lactamase alleles, antibiotic resistance and a test of the BRO-1
selective replacement hypothesis in Moraxella catarrhalis
F. Levy1* and E. S. Walker2,3
Departments of 1Biological Sciences, Box 70703 and 2Internal Medicine, Box 70622,
East Tennessee State University, Johnson City, TN 37614; 3James H. Quillen Veterans Affairs Medical
Center (11C), Mountain Home, TN 37684, USA
Received 18 July 2003; returned 9 October 2003; revised 7 November 2003; accepted 10 November 2003
Objectives: The hypothesis that BRO-1 selectively replaced the BRO-2 isoform of the Moraxella catarrhalis
BRO β-lactamase was tested by examining the temporal distribution, antibiotic resistance and epidemiological
characteristics of isolates from a long-term collection at a single locale.
Methods: A rapid, one-step PCR assay conducted on 354 isolates spanning 1984–1994 distinguished bro
alleles in over 97% of the β-lactamase-producing isolates. Probes of dot blots were used to distinguish PCR
failure from non-β-lactamase-mediated penicillin resistance.
Results: BRO-2 isolates comprised 0–10% of the population per year with no evidence of a decline over time.
All β-lactamase producers exceeded the clinical threshold for penicillin resistance. Bimodality of penicillin
MICs for β-lactamase producers was caused by variation within BRO-1 rather than differences between
BRO-1 and BRO-2. Non-β-lactamase factors also confer resistance to penicillin and may contribute to the
BRO-1 bimodality. The 13 BRO-2 isolates were associated with diverse genotypes within which there was
evidence of epidemiologically linked clusters. The exclusive association of BRO-2 with four unrelated genotypes suggested maintenance of BRO-2 by recurrent mutation or horizontal exchange.
Conclusions: The relative rarity of BRO-2 throughout the study, the absence of a declining temporal trend,
and genetic diversity within BRO-2 all failed to support the hypothesis that BRO-2 was more common in the
past and has been selectively replaced by BRO-1.
Keywords: disease transmission, molecular epidemiology, selection
Introduction
Soon after the first reports of a β-lactamase in Moraxella catarrhalis,
phenotypic assays revealed differences in enzyme isoelectric points
that accompanied a difference in β-lactamase enzymic activity.1 The
high and low activity variants, referred to as BRO-1 and BRO-2,
respectively, differed primarily in amounts of enzyme produced and
a slower in vitro rate of substrate metabolism by BRO-2 relative to
BRO-1.1,2
The nucleotide sequence of the bro-2 allele differs from bro-1 by
five nucleotides within the coding region, one of which causes an
amino acid replacement of unknown significance. The upstream
region of bro-1 includes 21 bp encompassing a second copy of a 16 bp
repeat motif that is absent in bro-2 and is presumed to function as a
promoter enhancer.2
BRO-2 is relatively rare, invariably occurring in less than 15% of
isolates.1,3 On the basis of sequence similarity between bro-2 and
bro-negative isolates, Bootsma et al.4 hypothesized that a bro-2-like
allele was originally transferred into M. catarrhalis and that bro-1
was generated by a duplication in the promoter with subsequent
spread enhanced by selection for more active enzymic activity via
greater enzyme production.
This study examined evidence for the bro-1 replacement hypothesis,
investigated the relationship between allelic types and antibiotic
resistance phenotypes, documented the role of non-β-lactamase
factors in penicillin resistance and tested the efficacy of a one-step,
length-based PCR assay to distinguish bro alleles.
Materials and methods
Study population
Assays were conducted on 354 isolates composed of 25–50 randomly
selected isolates from each of the 10 years (July 1984 to June 1994) repre-
..................................................................................................................................................................................................................................................................
*Corresponding author. Tel: +1-423-439-6926; Fax: +1-423-439-5958; E-mail: [email protected]
...................................................................................................................................................................................................................................................................
371
JAC vol.53 no.2 © The British Society for Antimicrobial Chemotherapy 2004; all rights reserved.
F. Levy and E. S. Walker
Table 1. Genotypica, phenotypic and epidemiologicalb characteristics of BRO-2-producing
Moraxella catarrhalis isolates at the Veterans Affairs Medical Center
MIC (mg/L)
Genotype
206BC
206BC
206BC
206BC
349_ _c
113CC
113CC
441CCc
441CC
441CC
151DC
585CCc
757CCc
Epidemiology
ID
penicillin
cefamandole
location
date (month-day-year)
206
221
223
226
349
467
499
441
488
536
523
585
757
4
4
4
4
0.25
1
2
1
2
2
0.25
0.50
0.25
2
2
2
2
1
1
2
2
1
2
1
2
2
01
01
01
NA
05
01
17
01
01
17
28
01
NA
08-25-86
12-03-86
12-04-86
12-18-86
06-29-87
03-20-88
05-10-88
02-24-88
04-14-88
08-01-88
06-12-88
01-09-89
12-12-90
NA = data not available.
aThree-locus genotype; strain identifier number.
bCoded hospital location of patient; date strain was isolated.
cGenotype with exclusive association with BRO-2.
sented in a collection of M. catarrhalis from the Veterans Affairs Medical Center at Mountain Home, TN (VAMC). Details of the collection and
methods for isolate selection, antibiotic testing, DNA extraction, and
genotyping are given in Walker & Levy5 and Walker et al.6
β-Lactamase gene assays
PCR primers amplified nucleotides –216 to +19 of the bla locus and
encompassed the 21 bp difference that distinguishes the two bro alleles
(GenBank U49269 and Z54180). Primer sequences were left: 5′-CACCCYGTRGGACAAGC-3′ and right: 5′-AATGACGGCGTTGCATC3′. PCR products of 235 bp (bro-1) or 214 bp (bro-2) were distinguished
on 2.0% agarose gels. Isolates that failed to generate a bro allele PCR
product were tested for the presence of the bla gene using PCR primers
encompassing 862 bp of the 945 bp of the coding region. Primer
sequences were left: 5′-TTTGGATTGGGGTGAATGAT-3′ and right:
5′-TGGGGCTGGGTGATAAATAG-3′. PCR protocols consisted of
94°C for 10 min followed by 25 cycles of: 94°C for 30 s, 55°C for 60 s,
72°C for 40 s and ending with 72°C for 7 min. ATCC 43627, 43628 and
25238 were used as BRO-1, BRO-2 and BRO-negative controls, respectively. All negative PCRs were tested twice and DNAs were demonstrated
to be suitable PCR substrates in control assays. Digoxigenin-labelled
probes for dot blots were constructed by PCR using DNA from ATCC
43627 and BRO-1 clinical isolate 327.
Statistical analysis
Penicillin MICs for BRO-1 and BRO-2 samples were compared using
ANOVA.
Results
The bro-2 allele, found in 13 of 354 isolates, comprised 2–10% of the
samples from 1986 to 1991 but was absent during earlier and later
sample years. BRO-2 was associated with seven different genotypes
and showed evidence of spatial and temporal clustering of isolates
with identical and related genotypes (Table 1). For example, three
206BC BRO-2 isolates were recovered within a 15-day period and
two were recovered from patients in the same ward. All three occurrences of genotype 441CC were BRO-2 and they were recovered
within a 6-month span concurrently with the two BRO-2 isolates of
the closely related (differing by one mutational step) genotype
113CC. Further, all BRO-2 isolates of genotypes 113CC and 441CC
were isolated from patients in two wards. The pairwise number of
differences between the remaining BRO-2 genotypes ranged from
three to 12 mutational steps.
BRO-1 isolates showed significantly higher penicillin MICs than
BRO-2 isolates (F = 6.51; df = 1, 266; P = 0.01; geometric means:
BRO-1 = 3.29, BRO-2 = 1.24) with a minimum MIC of 0.25 mg/L for
a β-lactamase-producing isolate (Table 2). BRO-1 isolates also
showed significantly higher rates of nitrocefin substrate hydrolysis
compared with BRO-2 isolates, but there was no relationship
between rate of hydrolysis and penicillin MIC (data not shown).
The nitrocefin disc assay for β-lactamase activity and the bro
allele PCR assay showed correspondence in 98% (346/353) of the
isolates. Seven penicillin-resistant isolates lacked PCR products but
nitrocefin hydrolysis and bla probe hybridization showed the
absence of a PCR product was not caused by the absence of a functional gene, suggesting that PCR failure probably arose from mismatches between primer and target sequences. For the purpose of
detecting a β-lactamase gene, the bro allele PCR assay had a success
rate of 97.4% (262/269) and it does not require restriction digestion of
PCR products. Reduced susceptibility to penicillin also occurred in
the absence of a β-lactamase as evidenced by 12 isolates with reduced
susceptibility (MIC 0.06–0.25 mg/L) that tested negative for a
β-lactamase in enzymic and molecular assays (Table 2). One isolate
generated no PCR products and failed to hydrolyse nitrocefin (907,
MIC = 0.50 mg/L), but its DNA hybridized to the probe derived from
the clinical isolate and not to the ATCC probe. Thus, the β-lactamase
372
BRO-1 selective replacement hypothesis in M. catarrhalis
Table 2. Number of isolates within penicillin MIC (mg/L) classes by BRO type
Penicillin MIC (mg/L)
BRO type
BRO–
BRO-1
BRO-2
<0.03
0.06
0.125
0.25
0.5
1
2
4
8
>16
Total
65
0
0
1
0
0
5
0
0
6
2
3
0
16
1
0
81
2
0
34
3
0
12
4
0
42
0
0
69
0
77
256
13
gene, if present in 907, is non-functional or the DNA sequence and
specificity are much altered.
Discussion
BRO-1 replacement hypothesis
Two observations from this study fail to support the hypothesis that
BRO-2 was more common in the past and its frequency declined as a
result of replacement by BRO-1. First, BRO-2 was absent in isolates
from the early collection years, a period when β-lactamase producers
were less common,6 and second, when present, BRO-2 showed a sporadic, rather than declining temporal distribution. Incomplete collections render our analysis and data from other studies inadequate for
detecting processes that preceded the early 1980s. Nevertheless, the
presence of BRO-2 in diverse, but often rare phylogenetic lineages,
combined with prior evidence that the transition to penicillin resistance was not mediated by population bottlenecks or a selective
sweep,5 suggests generation of bro-2 alleles by either recurrent mutation or spread via horizontal transmission.
Phylogenetic and epidemiological associations among BRO-2
isolates suggest minor fluctuations in frequency reflect brief bouts of
clonal expansion rather than long-term trends caused by selection.
For example, temporal clusters of genetically identical BRO-2 isolates accounted for over half (8/13) of the BRO-2 isolates and these
appeared to have arisen via patient-to-patient transmission. The
frequency of BRO-2 isolates is probably an equilibrium balancing
increases by mutation and clonal expansion with decreases caused by
either genetic drift or selection. If BRO-1 is under positive selection
then the sporadic and continued occurrence of BRO-2 may be
explained by recurrent mutation and horizontal exchange, both of
which are consistent with observations of BRO-2 in unrelated genotypes as shown in our study and in Bootsma et al.7
Trimodality in the penicillin-susceptibility spectrum
Trimodal patterns of penicillin MICs uncovered in earlier studies led
to the hypothesis that the three peaks corresponded to BRO-negative,
BRO-2 and BRO-1 strains.6,8,9 The current survey has shown that that
hypothesis was not correct but rather, the trimodality is caused by a
unimodal pattern in BRO-negative isolates with a peak in the susceptible range (<0.03 mg/L) and bimodality within BRO-1 isolates with
peaks at 1 and >16 mg/L (Table 2).
Factors other than the characterized differences between BRO-1
and BRO-2 must underlie the BRO-1 bimodal pattern. Similar
hypotheses of additional factors have been offered to explain the
range of ampicillin MICs among β-lactamase producers10 and the
overlap in ampicillin MICs among BRO types.3
Non-β-lactamase-mediated resistance factors as MIC
modifiers
Twelve β-lactamase-negative isolates, representing five different
genotypes, displayed MICs with either reduced susceptibility or clinically-relevant penicillin resistance (≥0.25 mg/L). These strains
tested negative for the presence of the bla gene by two independent
PCR assays and a genomic dot blot assay, and they failed to hydrolyse
nitrocefin. Penicillin MICs for these isolates ranged from 0.06 to
0.25 mg/L, conclusively demonstrating that non-β-lactamase resistance factors can confer clinical resistance in M. catarrhalis.
In Neisseria gonorrhoeae, non-β-lactamase-mediated penicillin
resistance developed through the gradual accumulation of mutations,
each leading to a minor reduction in susceptibility. These mutations
gave rise to an altered penicillin-binding protein (PBP), enhanced
efflux pumps, and decreased membrane permeability.11–13 Altered
PBPs confer resistance specifically to β-lactam agents, but altered
efflux and permeability confer reduced susceptibility to non-β-lactam
agents including macrolides.12,13 Similar non-β-lactamase resistance
mechanisms may underlie prior observations of antibiotic-specific
changes in susceptibility in M. catarrhalis at the VAMC. For example,
cefamandole resistance declined after the drug was no longer used at
the medical centre, although resistance to penicillin and the proportion of β-lactamase producers continued to rise in the population.6
Additionally, a significant temporal population trend toward increasing
clarithromycin MICs was restricted to the β-lactamase producers,6 a
phenotype consistent with determinants conferring altered efflux or
permeability.
Acknowledgements
This work was supported by a grant from the ETSU-RDC to E. S.
Walker and a non-instructional assignment to F. Levy hosted by the
Division of Infectious Diseases, Department of Internal Medicine.
References
1. Wallace, R. J., Jr, Steingrube, V. A., Nash, D. R. et al. (1989).
BRO β-lactamases of Branhamella catarrhalis and Moraxella subgenus
Moraxella, including evidence for chromosomal β-lactamase transfer by
conjugation in B. catarrhalis, M. nonliquefaciens, and M. lacunata. Antimicrobial Agents and Chemotherapy 33, 1845–54.
2. Bootsma, H. J., van Dijk, H., Vorhoef, J. et al. (1996). Molecular
characterization of the BRO β-lactamase of Moraxella (Branhamella)
catarrhalis. Antimicrobial Agents and Chemotherapy 40, 966–72.
3. Fung, C. P., Yeo, S. F. & Livermore, D. M. (1994). Susceptibility of
Moraxella catarrhalis isolates to β-lactam antibiotics in relation to
β-lactamase pattern. Journal of Antimicrobial Chemotherapy 33, 215–22.
373
F. Levy and E. S. Walker
4. Bootsma, H. J., van Dijk, H., Vauterin, P. et al. (2000). Genesis of
BRO β-lactamase-producing Moraxella catarrhalis: evidence for transformation-mediated horizontal transfer. Molecular Microbiology 36, 93–104.
5. Walker, E. S. & Levy, F. (2001). Genetic trends in a population
evolving antibiotic resistance. Evolution 55, 1110–22.
6. Walker, E. S., Neal, C. L., Laffan, E. et al. (2000). Long-term
trends in susceptibility of Moraxella catarrhalis: a population analysis.
Journal of Antimicrobial Chemotherapy 45, 175–82.
7. Bootsma, H. J., van der Heide, H. G., van de Pas, S. et al. (2000).
Analysis of Moraxella catarrhalis by DNA typing: evidence for a distinct
subpopulation associated with virulence traits. Journal of Infectious
Diseases 181, 1376–87.
8. Berk, S., Kalbfleisch, J. H. & The Alexander Project Collaborative
Group. (1996). Antibiotic susceptibility patterns of community-acquired
respiratory isolates of Moraxella in Western Europe and the USA.
Journal of Antimicrobial Chemotherapy 38, Suppl. A, 85–96.
9. Baquero, F. (1996). Trends in antibiotic resistance in respiratory
pathogens: an analysis and commentary on a collaborative surveillance
study. Journal of Antimicrobial Chemotherapy 38, Suppl. A, 117–32.
10. Richter, S. S., Winokur, P. L., Brueggmann, A. B. et al. (2000).
Molecular characterization of the β-lactamases from clinical isolates of
Moraxella (Branhamella) catarrhalis obtained from 24 U.S. medical
centers during 1994–1995 and 1997–1998. Antimicrobial Agents and
Chemotherapy 44, 444–6.
11. Spratt, B. G. (1989). Hybrid penicillin-binding proteins in penicillinresistant strains of Neisseria gonorrhoeae. Nature 332, 173–6.
12. Gill, M. J., Simjee, S., Al-Hattawi, K. et al. (1998). Gonococcal
resistance to β-lactams and tetracycline involves mutation in loop 3 of the
porin encoded at the penB locus. Antimicrobial Agents and Chemotherapy 42, 2799–803.
13. Hagman, K. E., Pan, W., Spratt, B. G. et al. (1995). Resistance of
Neisseria gonorrhoeae to antimicrobial hydrophobic agents is modulated
by the mtrRCDE efflux system. Microbiology 141, 611–22.
374