Use of Cefoperazone MacConkey Agar for Selective Isolation of Laribacter hongkongensis

Abstract

A new selective medium, cefoperazone MacConkey agar (CMA), was developed for primary isolation of Laribacter hongkongensis from stool. Its performance in quantitative recovery and in a clinical evaluation of 4,741 human diarrheal stool specimens was superior to that of charcoal cefoperazone deoxycholate agar. In addition, with CMA, Arcobacter butzleri was unexpectedly isolated from the stools of six patients.

Primary isolation of bacterial pathogens that cause diarrhea largely relies on the availability and refinement of selective media (1, 2-4, 8, 10, 13, 17). Laribacter hongkongensis is a newly discovered bacterium that has recently been isolated from the stools of three Asian and three European patients with community-acquired diarrhea (16; P. C. Y. Woo, P. Kuhnert, A. P. Burnens, J. L. L. Teng, S. K. P. Lau, T. L. Que, H. H. Yau, and K. Y. Yuen, submitted for publication). To define its causative role in infectious diarrhea, the development of selective isolation media is of paramount importance. Although the six L. hongkongensis strains isolated from stools were initially recovered on charcoal cefoperazone deoxycholate agar (CCDA), which contains 32 μg of cefoperazone/ml, their colonies on CCDA under microaerophilic incubation were small and difficult to pick up. Moreover, the type strain, HKU1, was unable to grow on CCDA.

Based on the cefoperazone susceptibilities of L. hongkongensis, in this study we developed a new agar medium, cefoperazone MacConkey agar (CMA), which was modified from MacConkey agar by adding 32 μg of cefoperazone/ml. We compared five different isolation media—MacConkey agar, CMA, CMA-8 (MacConkey agar with 8 μg of cefoperazone/ml), CCDA, and CCDA-8 (CCDA with 8 μg of cefoperazone/ml)—on their ability to recover L. hongkongensis and suppress the enteric bacteria. An evaluation that compared the efficacies of CMA and CCDA on clinical specimens was also carried out.

Seven clinical isolates of L. hongkongensis were used for cefoperazone susceptibility tests and quantitative assessment of various media (16; Woo et al., submitted). Four standard strains of aerobic enteric bacteria were also used for evaluating the suppressive abilities of the media (Table 1). The MICs of cefoperazone were determined by the macrodilution broth method according to NCCLS guidelines (12). All agar media were prepared according to the manufacturer's (Oxoid Ltd., Basingstoke, United Kingdom) instructions, with modifications of cefoperazone (Sigma, St. Louis, Mo.) concentrations.

TABLE 1.

Mean colony counts for seven L. hongkongensis strains and four standard strains of enteric bacteria on various media

Bacterial strain	Mean log10 colony count at 48 h
	Blood agar control	MacConkey agar	CMA	CMA-8	CCDA	CCDA-8	CCDA (microaerophilic incubation)
L. hongkongensis strain 1 (HKU1)	7.97	7.94	7.94	7.95	No colony	No colony	No colony
L. hongkongensis strain 2 (HLHK2)	7.91	7.90	7.91	7.89	7.88	7.90	5.85
L. hongkongensis strain 3 (HLHK3)	7.99	7.98	8.00	7.97	7.97	7.97	5.77
L. hongkongensis strain 4 (HLHK4)	8.02	8.03	8.00	8.00	7.99	7.96	5.84
L. hongkongensis strain 5 (HLHK5)	7.98	8.00	7.97	7.99	7.96	7.97	5.86
L. hongkongensis strain 6 (HLHK6)	8.05	8.04	8.03	8.03	8.00	8.01	5.90
L. hongkongensis strain 7 (HLHK7)	7.95	7.96	7.97	7.95	7.98	7.96	5.81
Enterococcus faecalis ATCC 29212	8.03	8.01	No colony	No colony	No colony	No colony	No colony
Escherichia coli ATCC 25922	7.96	7.99	No colony	No colony	No colony	No colony	No colony
Klebsiella pneumoniae ATCC 13883	7.94	7.97	No colony	No colony	No colony	No colony	No colony
Proteus mirabilis ATCC 7002	7.99	7.97	No colony	No colony	No colony	No colony	No colony

MacConkey agar, CMA, CMA-8, CCDA, and CCDA-8 were quantitatively tested for their ability to support the growth of L. hongkongensis and inhibit the growth of four standard enteric bacteria. Working suspensions were prepared by emulsifying cultures of each isolate in 0.01 M phosphate-buffered saline and adjusted visually to a 0.5 McFarland standard. Six 10-fold dilutions were made, and 100 μl of the last four dilutions was inoculated evenly onto each medium in triplicate. All plates were incubated at 37°C in ambient air, with an additional set of CCDA at 42°C under a microaerophilic environment. The log₁₀ mean colony counts were determined at both 24 and 48 h.

Based on the results of quantitative assessment, CMA was compared to CCDA on 4,741 consecutive freshly collected stool specimens from patients with community-acquired diarrhea in four regional hospitals in Hong Kong during a 6-month period (July-December 2002). All specimens were directly cultured onto both agars. The CMA plates were incubated at 37°C in ambient air, and the CCDA plates were incubated at 42°C under microaerophilic incubation, a standard procedure for Campylobacter (6). All plates were examined for the presence of L. hongkongensis at 24 and 48 h, with all suspected isolates identified by phenotypic tests and 16S rRNA gene sequencing (11, 15, 16; Woo et al., submitted).

The MICs of cefoperazone on all seven isolates of L. hongkongensis were higher than 256 μg/ml. With the exception of CCDA under microaerophilic incubation, where colonies at 24 h were too small to be recognized, colony counts for each isolate on each medium at 24 and 48 h were the same. The log₁₀ mean colony counts of each isolate on each medium at 48 h are shown in Table 1. Apart from that of HKU1, which did not grow on CCDA or CCDA-8, the colony counts of the L. hongkongensis strains on MacConkey agar, CMA, CMA-8, CCDA, and CCDA-8 incubated in ambient air were comparable to those obtained with the blood agar control. However, counts on CCDA under microaerophilic incubation were significantly reduced (P < 0.01). Moreover, colonies on CCDA under microaerophilic incubation at 48 h were still of pinpoint size and were difficult to pick up. The four standard aerobic enteric bacterial strains failed to grow on CMA, CMA-8, CCDA, or CCDA-8.

The isolation rate of L. hongkongensis from human diarrheal stool specimens was higher on CMA incubated in ambient air than on CCDA under microaerophilic incubation. From the total of 4,741 specimens, L. hongkongensis was isolated in pure culture on CMA from 22 specimens from 22 patients but on CCDA it was isolated from only 7 specimens from 7 of the 22 patients (P < 0.05). However, the colonies on CCDA were very small and difficult to recognize, in contrast to the large, lactose-negative colonies on CMA. In addition, six strains of L. hongkongensis-like organisms subsequently identified as Arcobacter butzleri by 16S rRNA gene sequencing were unexpectedly present in specimens from six patients and were isolated in pure culture on CMA but not on CCDA. These strains shared similar biochemical characteristics with L. hongkongensis, except that they were negative for urease and arginine dihydrolase. All 22 strains of L. hongkongensis and 6 strains of A. butzleri were isolated sporadically from the specimens of 28 unrelated patients from among the 4,741 clinical specimens.

The historical failure to recognize L. hongkongensis from human stools is likely due to a combination of misidentification and the lack of an optimal selective medium. All six initial L. hongkongensis strains from stools were recovered on CCDA under microaerophilic incubation. They were at first mistaken for Campylobacter species but were all found to grow in an aerobic environment after aerotolerance testing. In most clinical laboratories, these strains would have been discarded as nonpathogens or wrongly reported as Campylobacter.

CMA was developed because L. hongkongensis strains were highly resistant to cefoperazone and grew well on MacConkey agar as large lactose-negative colonies. CMA and CMA-8 were superior in supporting the growth of all seven L. hongkongensis strains. To suppress more cefoperazone-resistant enteric flora that may be encountered in clinical stool specimens, CMA would probably be more advantageous than CMA-8. A clinical evaluation also showed that CMA incubated in ambient air was superior to CCDA under microaerophilic incubation in terms of both isolation rate and colony size. All 22 isolates of L. hongkongensis were recovered on pure culture on CMA, suggesting that CMA also effectively inhibits normal gut flora.

The unexpected isolation of A. butzleri on CMA suggests that the medium may also be potentially useful in isolating Arcobacter, which has also been associated with diarrhea and invasive disease by entry through the gut (9, 11, 14, 15). Owing to the lack of a specific selective medium for isolation of Arcobacter in fecal specimens, its role as a diarrheal pathogen remains undetermined (5, 7). CMA may serve as a selective medium for both L. hongkongensis and Arcobacter and assist in defining their causative roles in infectious diarrhea.