Coryneform Bacteria in Throat Cultures of Healthy Individuals

Abstract

Throat swabs from 113 healthy individuals from Hamburg, Germany, and Zurich, Switzerland, were investigated for coryneform bacteria with nonselective and selective media. Ninety specimens contained 123 strains. Surprisingly, 76% of them were strains of Corynebacterium durum (47%) and Rothia dentocariosa (29%). Only two were strains of Corynebacterium pseudodiphtheriticum, and none were strains of C. striatum, C. amycolatum, or C. diphtheriae.

Some coryneform bacteria inhabit the skin and/or mucous membranes of humans and animals (7). Their speciation in throat cultures has, with the exception of Corynebacterium diphtheriae, Arcanobacterium haemolyticum (7), and Rothia dentocariosa (1a), rarely been tried; Corynebacterium xerosis (probably Corynebacterium amycolatum [7]), Corynebacterium pseudodiphtheriticum, and Corynebacterium striatum are said to occur in the nasopharynx (7). Earlier studies on these bacteria in throat cultures (9, 10) had employed neither selective media nor modern identification techniques and taxonomy and were, therefore, deficient in sensitivity and speciation.

In the framework of our studies on the ecology of coryneform bacteria (7) we have examined 113 throat cultures with selective and nonselective media for the presence of coryneform bacteria.

The probands were 94 healthy individuals aged 20 to 55 years (average age, 36 years) belonging to the staff of the Endo Clinic in Hamburg, Germany, a hospital specializing in orthopedic surgery, and 19 healthy individuals from the Department of Medical Microbiology at the University of Zurich (average age, 23 years). While the two groups had no contact, there were contacts within each group that were impossible to specify. No proband had been on antibiotics. For sampling, the swab technique of Johnston and Bodey (8), which employs a cotton swab on an applicator, was used. Following vortexing for 3 min in 0.5 ml of 0.85% NaCl, 0.1 ml of each sample was plated on blood agar (Columbia agar [bioMérieux, Marcy-l’Etoile, France] with 5% sheep blood), blood agar with 1% Tween 80 (E. Merck, Darmstadt, Germany), blood agar with 100 mg of fosfomycin-sodium succinate (kindly supplied by Boehringer Mannheim Schweiz AG, Rotkreuz, Switzerland) per liter plus 12.5 mg of glucose-6-phosphate per liter, and blood agar with the above concentrations of Tween 80 and fosfomycin. Tween 80 was used to enlarge colonies of lipophilic corynebacteria, and fosfomycin was used for selecting coryneforms (3, 12, 16). Incubation was for 48 h at 37°C. Plates were picked for colonies of different morphologies, which were subcultured into Columbia Broth (bioMérieux) and examined microscopically. Bacteria with a coryneform morphology (7) were then identified according to a scheme developed by two of us (14) and enlarged to take into consideration more recently recognized species (4–6, 11). For Corynebacterium durum, Rothia, Aureobacterium, Microbacterium, and Arthrobacter strains, cellular fatty acids, cell wall diamino acids, and the presence or absence of mycolic acids were also determined (7).

Because fosfomycin-blood agar had never been systematically investigated for its inhibitory effects on coryneforms, we checked three strains each of the following species from our Zurich reference strain collection for growth on fosfomycin-blood agar: Corynebacterium jeikeium, C. urealyticum, C. striatum, C. minutissimum, C. amycolatum, C. pseudodiphtheriticum, C. propinquum, C. afermentans subsp. afermentans and subsp. lipophilum, C. diphtheriae, C. coyleae (6), C. glucuronolyticum, C. auris, C. argentoratense, C. xerosis, C. accolens, C. macginleyi, Corynebacterium group G, Dermabacter hominis, Turicella otitidis, R. dentocariosa (catalase positive), Actinomyces neuii, Actinomyces radingae, Actinomyces turicensis, Actinomyces europaeus (4), Brevibacterium casei, Propionibacterium acnes, Propionibacterium avidum, Propionibacterium granulosum, Arcanobacterium bernardiae, Listeria monocytogenes, a Lactobacillus sp., and a Bacillus sp., as well as Enterococcus faecalis ATCC 29212, Staphylococcus aureus ATCC 29213, and Escherichia coli ATCC 25922. One loopful from a 24-h Columbia Broth culture was streaked on a plate. At >50 mg of fosfomycin per liter, only the following failed to grow: all strains of Actinomyces spp., D. hominis, the Bacillus sp., two strains of catalase-positive R. dentocariosa, and one strain each of three Propionibacterium spp., as well as the S. aureus, E. coli, and E. faecalis American Type Culture Collection strains. Furthermore, one strain of C. xerosis and one strain of C. striatum were cultured for 48 h at 37°C in Columbia Broth. Then, 1-μl samples of two 1/10 dilutions were plated on blood agar and on fosfomycin-blood agar. Equal quantities were recovered from both media.

Of 113 throat cultures, 90 (80%) grew coryneforms. A total of 123 strains were isolated (Table 1); 49 of them were found together with other coryneforms, but each component generally grew on a different medium. Whenever strains of the same species grew on different media, they were counted as one. The number of colonies of coryneforms per plate was between 5 and 50, while colonies of viridans group streptococci and neisseriae on nonselective blood agar plates were generally 10 to 100 times more numerous than those of coryneforms, confirming the necessity of employing selective media (16). On fosfomycin plates, the growth of viridans group streptococci and neisseriae was markedly inhibited (3).

TABLE 1.

Coryneforms isolated from 113 throat cultures

Species	No. of strains		Medium or mediab (no. of strains)
	Total	Found with other coryneformsa
C. durum	58	17	FoBAd
R. dentocariosa
Catalase positive	26	11	BAe
Catalase negative	10	6	BAf
Corynebacterium group G	7	5	FoTwBA
C. jeikeium	4	3	FoTwBA
Aureobacterium sp.	4	1	BA
Corynebacterium spp.c	4	1	BA (1), FoBA (2), TwBA (1)
C. pseudodiphtheriticum	2	0	FoBA
Microbacterium sp.	2	1	FoBA
Brevibacterium sp.	1	1	BA
Arthrobacter sp.	1	0	BA
C. argentoratense	1	0	FoBA
C. minutissimum	1	1	FoBA
Corynebacterium group F-1	1	1	TwBA
Actinomyces sp.	1	1	FoBA

^aSee text. 

^bBA, blood agar; FoBA, fosfomycin-blood agar; FoTwBA, fosfomycin–Tween 80–blood agar; TwBA, Tween 80–blood agar. 

^cUnidentified. 

^dThree on blood agar only. 

^eTwo on fosfomycin-blood agar also. 

^fThree on fosfomycin-blood agar also. 

To our surprise, the largest contingent (47%) of strains belonged to the recently described C. durum, an adherent, fermentative, nonlipophilic species (11). The original work on this species was done on five strains (11). We also found β-galactosidase-positive and/or mannitol-negative strains as well as weakly ribose-positive and weakly mannitol-positive or even mannitol-negative strains which were morphologically typical of C. durum. One strain that was β-galactosidase positive, one that was mannitol negative, and one that was weakly ribose positive were subjected to DNA-DNA hybridization with the type strain, CCUG 37331 (11), to which they proved to be ≥75% related. In the API (RAPID) Coryne system (API bioMérieux, La Balme-les-Grottes, France), 24-h readings were thus not only 3000135, 3040135, or 3001135 (11) but also 3400115, 3400135, 3400305, 3400325, or 3400335. Within 4 days of incubation, some negative mannitol, ribose, and maltose reactions had turned positive.

The next most frequent species was R. dentocariosa (29%), an organism found frequently in throat cultures of healthy persons by others as well (1a). As expected from our controls, most strains of this species grew on blood agar only. Again to our surprise, 10 strains proved to be catalase negative. Such strains have also been observed by others and may prove to be a new species (1). The third surprise was that we did not isolate C. amycolatum and C. striatum, i.e., species isolated by others from sputa (2, 15), albeit of predisposed and/or hospitalized patients. Among the more frequently reported corynebacteria, C. jeikeium was isolated from four individuals with no known predisposing condition. The only other lipophilic taxa isolated were Centers for Disease Control and Prevention (CDC) group G and group F-1 corynebacteria. Other nonlipophilic coryneforms were isolated only rarely. There were no qualitative or quantitative differences or differences related to sex of the probands between the bacteria recovered from the samples from Hamburg and Zurich (data not shown). It is also noteworthy that the spectrum of coryneforms isolated from prosthetic joint and open fracture infections at the Endo Clinic (13) was entirely different from the spectrum found in the present study.

Although the large amount of normal throat flora on nonselective blood agar plates may have reduced the chance to find coryneforms inhibited by fosfomycin, we believe that our findings are representative at least of corynebacteria and Rothia. They point to a reservoir for these species, which have been considered rare. On the other hand, the widespread opinion that hitherto well-known corynebacteria are part of the normal throat flora should be reassessed.