Moellerella wisconsensis: identification, natural antibiotic susceptibility and its dependency on the medium applied

1. Introduction 

Within the Enterobacteriaceae, strains of the genus Moellerella represent a monophyletic species, named M. wisconsensis, which was first recognized as Enteric Group 46 in 1980 from cultures that had been sent to the Centers for Disease Control (CDC, Atlanta, GA) in the United States. In 1984, the name M. wisconsensis was proposed for Enteric Group 46 (Hickman-Brenner et al., 1984); the specific epithet ′wisconsensis′ was coined because the majority of the strains examined had been isolated from clinical specimens in the Wisconsin area (Hickman-Brenner et al., 1984).

M. wisconsensis strains resemble in all respects well-established Enterobacteriaceae, representing Gram-negative, facultative anaerobic, nitrate-reducing and Oxidase-negative rods producing enterobacterial common antigen Hickman-Brenner et al 1984, Ramia et al 1982. Taxonomi-cally, M. wisconsensis is 23–32% related to Providencia spp. and 18–26% related to Proteus spp.; to other Enterobacteriaceae, its relatedness is lower than 19% (Hickman-Brenner et al., 1984).

After its initial isolation in the United States, strains of M. wisconsensis have been isolated in several European countries, i.e., the Czech Republic and Slovakia (Cabadajova & Kudrna, 1988), France Ohanessian et al 1987, Richard 1989, Wallet et al 1994, Germany (Wittke et al., 1988), and the UK (Marshall et al., 1986). M. wisconsensis seems to be distributed widely in nature and several strains have been isolated from animals (Bangert et al., 1988), water Aldová 1992, Hickman-Brenner et al 1984 and human food (Aldová, 1992). However, the majority of moellerellae has been isolated from human clinical specimens, in particular feces Hickman-Brenner et al 1984, Marshall et al 1986, and there is evidence that M. wisconsensis is associated with human diarrhea (Brenner, 1992). Unfortunately, there is only one study on the prevalence of Moellerella in stool specimens (Marshall et al., 1986) and research in virulence factors of Moellerella was never performed. Apart from human stools, M. wiscoensensis has also been isolated from non-fecal specimens, i.e., gall bladders Ohanessian et al 1987, Wittke et al 1985, bronchial aspirates Aldová 1992, Wallet et al 1994 and one peritoneum exsudate (Aldová, 1992), indicating clinical significance.

There is evidence that in the last decades strains of Moellerella have been perceived or misidentified. M. wisconsensis is indistinguishable from Escherichia coli on MacConkey and eosin-methylene blue agars (Hickman-Brenner et al., 1984) and may therefore be missed in clinical specimens. In addition, several strains of Moellerella resemble phenotypically some klebsiellae, in particular certain strains of Klebsiella pneumoniae subsp. ozaenae (see below).

The main scope of this research was to create a data compilation of the natural susceptibility of M. wisconsensis strains originating from different sources to a wide range of antibiotics. Antibiotic susceptibility data on Moellerella include sole strains and/or few antimicrobial agents. Natural antibiotic susceptibility patterns of Moellerella have not been published. In addition, we were interested in probable medium-dependencies in antibiotic susceptibility testing. Finally, this study was performed to evaluate two commercially available identification systems that contain M. wisconsensis in their data bank.

2. Materials and methods 

2.1. Bacterial strains 

A total of 17 moellerellae isolated from humans (n = 11), food (n = 5) and water (n = 1) was examined. The majority of the strains tested derived from the Culture Collection of the University of Göteborg in Sweden. M. wisconsensis 66-8 (Sr 1 21969/81), isolated from a clinical sample in 1997, was obtained from Alexander von Graevenitz (University of Zürich, Switzerland). M. wisconsensis M-1, isolated from a specimen of a outpatient in South Germany in 1998, was sent as ‘Klebsiella pneumoniae subsp. ozaenae’ by G. Stempfel (Gärtner Laboratories, Weingarten, Germany). M. wisconsensis ATCC 35017T and Escherichia coli ATCC 25922 derived from the German culture collection of microrganisms and cell cultures in Braunschweig (DSMZ). These strains and M. wisconsensis ATCC 35618 served as controls for antibiotic susceptibility testing. An overview on the origin of the strains examined is shown in Table 1.

Table 1. Moellerella strains of this study

	Strain (Additional designations)	Origin	Country
	CCUG 180421	Human gallbladder, 71-year-old male with acute cholecystitis	Germany
	CCUG 18768 (ATCC 35618, CDC 1826-79)2	Human feces, patient with diarrhoea	Virginia, USA
	CCUG 18773 (ATCC 35619, CDC 2552-77)2	Human feces, 62-year-old male with gastroenteritis	Wisconsin, USA
	CCUG 18774 (ATCC 35620, CDC 2897-78)2	Human feces, 38-year-old female	Wisconsin, USA
	CCUG 18775 (ATCC 35621, CDC 3065-75)2	Human feces, 5-year-old female with diarrhea	Wisconsin, USA
	CCUG 29901 (Aldová 27223)	Foam of ham	Czech Republic
	CCUG 29902 (Aldová 27224)	Sausage filling	Czech Republic
	CCUG 29903 (Aldová 27472)	Sausage, fowl	Czech Republic
	CCUG 29904 (Aldová 27625)	Smear	Czech Republic
	CCUG 29905 (Aldová 28037)	Human peritoneum, vesicular exsudate	Czech Republic
	CCUG 29906 (Aldová 28518)	Water	Czech Republic
	CCUG 29907 (Aldová 28536)	Sausage	Czech Republic
	CCUG 29908 (Aldová 28663)	Human sputum, chronic bronchitis	Czech Republic
	CCUG 29958 (Aleksic H1297/84)	Human wound	Germany
	ATCC 35017T	Human feces, 16-year-old female	Wisconsin, USA
	66-8 (Sr 1 21969/81)	Human clinical specimen	Switzerland
	M-1	Human clinical specimen	Germany

1 Strain published by Wittke et al. (Wittke et al., 1985) 2 Strain published by Hickmann et al. (Hickmann et al., 1984)

3. Results 

3.1. Identification 

All strains examined were unambiguously identified as M. wisconsensis by the Api20E system. Five strains (including M. wisconsensis M-1, originally labelled as K. pneumoniae subsp. ozaenae) showed unusual biochemical properties in the MCN-E system and were rhamnose- and xylose-positive (these strains gave rhamnose-negative reactions with Api20E). They were identified as K. pneumoniae subsp. ozaenae with sufficient or questionable suitability by the MCN-E data base. Applying conventional plate tests, the respective strains were rhamnose- and xylose-negative after an 24 h- and 48 h-incubation, but weakly positive after 7 days (Table 2).

Table 2. Biochemical features of the M. wisconsensis strains testeda

	Biochemical Test	Method	M. wisconsensis				K. pneumoniae subsp. ozaenae
			This study		Hickmann-Brenner et al. (1984) (n=9)	Farmer (1995)	Farmer (1995)
			ATCC 35017T	All strains (n=17)
	Amino acid deaminase	Api20E, MCN-E	−	0	0	0	0
	Arginine dihydrolase	Api20E, MCN-E	−	0	0	0	0
	Cytochromoxidase	Test strip	−	0	0	0	0
	β-Galactosidase1	Api20E, MCN-E	+	100	78 (24h) 100 (48h)	90	80
	Gelatinase	Api20E	−	0	0	0	0
	β-Glucuronidase2	MCN-E	−	0	NT	NT	NT
	β-Glucosidase3	MCN-E	−	0	0 (48h) 33 (7d)	0	80
	H2S production	Api20E, MCN-E	−	0	0	0	0
	Lysine decarboxylase	Api20E, MCN-E	−	0	0	0	40
	Motility	Tube test (48h)	−	0	0	0	0
	NO3−→NO2−	MCN-E	+	100	89 (24h)	90	80
	Ornithine decarboxylase	Api20E, MCN-E	−	0	0	0	3
	Tryptophanase4	Api20E, MCN-E	−	12	0	0	0
	Urease	Api20E, MCN-E	−	0	0	0	10
	Voges Proskauer Test	Api20E, MCN-E	−	0	0	0	0
	β-Xylosidase5	MCN-E	−	0	NT	NT	NT
	Assimilation of
	•citrate	Api20E, MCN-E	+	88	67 (24h)7 100 (48h)	807	307
		Plate Test (48h)	+	100
	•malonate	MCN-E	−	0	NT	0	3
	Fermentation of
	•adonitol	MCN-E	+	100	67 (24h) 100 (48h)	100	97
	•amygdalin	Api20E	−	0	NT	NT	NT
	•L-arabinose	Api20E	−	0	0	0	98
	•cellobiose	Plate test (48h)	−	0	0	0	92
	•glucose	Api20E, MCN-E	+	100	NT	100	100
	•myo-inositol	Api20E, MCN-E	−	0	0	0	55
	•D-mannitol	Api20E	−	0	0 (24h) 56 (48h)	60	100
					100 (7d)
	•melibiose	Api20E	+	100	100 (24h)	100	97
	•raffinose	Api20E	+	100	100 (24h)	100	90
	•L-rhamnose	MCN-E	−	29
		Api20E	−	0
		Plate test (48h)	−	0	0	0	55
		Plate test (7d)	−	(29)6
	•salicin	Tube test (48h)	−	0	0	0	97
	•D-sorbitol	Api20E, MCN-E	−	0	0	0	65
	•sucrose	Api20E, MCN-E	+	100	100 (24h)	100	20
	•trehalose	Plate test (48h)	−	0	0	0	98
	•D-xylose	MCN-E	−	35
		Plate test (48h)	−	0	0	0	95
		Plate test (7d)	−	(29)6

Cleavage of 1ortho-nitrophenyl-β-galactopyranoside (ONPG); 2para-nitrophenyl-β-glucuronide (PGUR); 5ortho-nitrophenyl-β-D-xylopyranoside (ONPX); 3hydrolysis of esculin; 4indole production; 6weakly positive; 7Simmons citrate; NT, not tested.

a Percentages of positive reactions are given. The results for Moellerella are contrasted to the respective data of Hickmann-Brenner et al. (Hickmann-Brenner et al. 1984) and Farmer (Farmer, 1995) and to the data for K. pneumoniae subsp. ozaenae (Farmer, 1995). Key discriminating reactions between M. wisconsensis and K. pneumoniae subsp. ozaenae are shown in bold print. It should be noted that the results obtained by Farmer were read after an 48h incubation.

In contrast to the MCN-E system, moellerellae able to ferment rhamnose and xylose can be separated from K. pneumoniae subsp. ozaenae with the Api20E system by arabinose and mannitol fermentation tests (Table 2). Additional discriminating reactions between M. wisconsensis and K. pneumoniae subsp. ozaenae were fermentation tests of cellobiose, salicin and trehalose (Table 2). The latter tests are not included in the pannels of Api20E and MCN-E.

3.2. Natural antibiotic susceptibility 

The antibiotic susceptibility patterns of M. wisconsensis are shown in Table 3; ;its natural sensitivities and resistances are summarized in Table 4.

Table 3. Antibiotic susceptibility of M. wisconsensisa

	Antibiotic (Standard-Reference)	Concentrations examined [mg/L]	Medium	Number of Strains with MIC [mg/L] of
				0.01	0.03	0.06	0.13	0.25	0.5	1	2	4	8	16	32	64	128	256	512	1024
	TETRACYCLINES
	Tetracycline	0.03–64	CAMHB						7	4						6
	(DIN, 1998)		ISOB								8	3					6
	Doxycycline	0.03–64	CAMHB					3	6	2				4	2
	(DIN, 1998)		ISOB							10	1				2	4
	Minocycline	0.03–64	CAMHB				1	2	6	2	4	1	1
	(DIN, 1998)		ISOB						4	6	1	2	4
	AMINOGLYCOSIDES
	Amikacin	0.13–256	CAMHB				1	14	2
	(DIN, 1998)		ISOB					4	9	4
	Gentamicin	0.06–128	CAMHB			14	3
	(DIN, 1998)		ISOB			1	12	3	1
	Netilmicin	0.06–128	CAMHB			12	5
	(DIN, 1998)		ISOB				10	6	1
	Tobramycin	0.06–128	CAMHB			11	6
	(DIN, 1998)		ISOB			1	9	7
	Streptomycin	0.13–256	CAMHB						10		4	1	1	1
	(SFM, 1998)		ISOB							9	1	5	2
	Kanamycin	0.13–256	CAMHB				1	12	4
	(DIN, 1998)		ISOB					3	10	4
	Neomicin	0.13–256	CAMHB				17
	(SFM, 1998)		ISOB				4	11	2
	Spectinomycin	0.13–256	CAMHB										1	13			2		1
	(NCCLS, 2000d)		ISOB											8	6	1		2
	Apramycin	0.06–128	CAMHB					2	13	2
	(Troxler et al., 2000)		ISOB						4	10	3
	Ribostamycin	0.06–128	CAMHB					7	10
	(Troxler et al., 2000)		ISOB					1	8	6	2
	Lividomycin A	0.06–128	CAMHB					1	13	3
	(Troxler et al., 2000)		ISOB						1	10	6
	β-LACTAMS: PENICILLINS
	Benzylpenicillin	0.01–32	CAMHB							3	11	1	1	1
	(DIN, 1998)		ISOB						1	4	9	1	1	1
	Oxacillin	0.03–64	CAMHB												1	16
	(DIN, 1998)		ISOB												4	13
	Amoxicillin	0.06–128	CAMHB						5	10					1	1
	(DIN, 1998)		ISOB						5	10					2
	Amoxicillin/clavulanate	0.06–128	CAMHB					2	15
	(DIN, 1998)		ISOB						17
	Ampicillin/sulbactam	0.06–128	CAMHB					1	15	1
	(DIN, 1998)		ISOB					1	15	1
	Sulbactam	0.25–512	CAMHB												5	12
			ISOB												8	9
	Piperacillin	0.13–256	CAMHB				11	4	1	1
	(DIN, 1998)		ISOB				10	5	1	1
	Piperacillin/tazobactum	0.13–256	CAMHB				7	10
	(DIN, 1998)		ISOB				10	7
	Ticarcillin	0.13–256	CAMHB						5	10	2
	(DIN, 1998)		ISOB						9	5	3
	Mezlocillin	0.13–256	CAMHB						13	3	1
	(DIN, 1998)		ISOB					3	10	3	1
	Azlocillin	0.25–512	CAMHB						2	12	1	1	1
	(DIN, 1998)		ISOB					1		13	1	2
	β-LACTAMS: CEPHALOSPORINS
	Cefaclor	0.13–256	CAMHB					9	8
	(DIN, 1998)		ISOB					5	12
	Loracarbef	0.13–256	CAMHB					1	14	2
	(DIN, 1998)		ISOB						14	3
	Cefazoline	0.13–256	CAMHB							14	3
	(DIN, 1998)		ISOB							15	2
	Cefuroxime	0.03–64	CAMHB					1	3	13
	(DIN, 1998)		ISOB						4	13
	Cefotiam	0.03–64	CAMHB			1	14	2
	(DIN, 1998)		ISOB				14	3
	Cefoxitin	0.03–64	CAMHB							10	7
	(DIN, 1998)		ISOB							10	7
	Cefixim	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	Cefpodoxime	0.03–64	CAMHB			14	3
	(DIN, 1998)		ISOB			15	2
	Cefdinir	0.03–64	CAMHB		1	16
	(NCCLS, 2000a)		ISOB		1	15	1
	Cefoperazone	0.03–64	CAMHB			1		3	10	3
	(DIN, 1998)		ISOB				1	3	10	3
	Cefetamet	0.03–64	CAMHB				15	2
	(DIN, 1998)		ISOB				16	1
	Cefotaxime	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	Ceftibutene	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	Ceftriaxone	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	Ceftazidime	0.03–64	CAMHB		1	13	3
	(DIN, 1998)		ISOB		2	11	4
	Cefepime	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	β-LACTAMS: CARBAPENEMS
	Imipenem	0.03–64	CAMHB			3	12	2
	(DIN, 1998)		ISOB			1	15	1
	Meropenem	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	β-LACTAMS: MONOBACTAMS
	Aztreonam	0.03–64	CAMHB		17
	(DIN, 1998)		ISOB		17
	QUINOLONES
	Ciprofloxacin	0.01–32	CAMHB	14	3
	(DIN, 1998)		ISOB	7	9	1
	Sparfloxacin	0.01–32	CAMHB	2	13	2
	(DIN, 1998)		ISOB	1	14	2
	Norfloxacin	0.03–64	CAMHB		2	14	1
	(DIN, 1998)		ISOB			11	6
	Ofloxacin	0.01–32	CAMHB		2	10	5
	(DIN, 1998)		ISOB			8	9
	Enoxacin	0.01–32	CAMHB			2	9	6
	(DIN, 1998)		ISOB				6	9	2
	Fleroxacin	0.01–32	CAMHB		2	15
	(DIN, 1998)		ISOB		1	14	2
	Pefloxacin	0.01–32	CAMHB		1	13	3
	(SFM, 1998		ISOB			7	10
	Pipemidic acid	0.06–128	CAMHB								17
	(DIN, 1998)		ISOB								3	12	2
	MACROLIDES
	Erythromycin	0.03–64	CAMHB											1	2	13	1
	(DIN, 1998)		ISOB													4	13
	Roxithromycin	0.03–64	CAMHB											1		6	10
	(O.-L. et al., 1997)		ISOB														17
	Clarithromycin	0.03–64	CAMHB								1				14	2
	(O.-L. et al., 1997)		ISOB												1	7	9
	Azithromycin	0.03–64	CAMHB						1				15	1
	(DIN, 1998)		ISOB										1		11	4	1
	LINCOSAMIDES
	Lincomycin	0.01–32	CAMHB												2	15
	(SFM, 1998)		ISOB													17
	Clindamycin	0.01–32	CAMHB									7	8	2
	(DIN, 1998)		ISOB											9	7	1
	STREPTOGRAMINS
	Dalfopristin	0.03–64	CAMHB										1				16
	(NCCLS, 2000c)		ISOB													1	16
	Quinupristin	0.03–64	CAMHB													1	16
	(NCCLS, 2000c)		ISOB														17
	Synercid	0.03–64	CAMHB										1			1	15
	(NCCLS, 2000c)		ISOB													1	16
	FOLATE PATHWAY INHIBITORS
	Trimethoprim	0.03–64	CAMHB		6	5	4										2
	(O.-L. et al., 1997)		ISOB		7	4	4										2
	Sulfamethoxazole	0.25–512	CAMHB									2	4	7		2	1			1
	(NCCLS, 2000a)		ISOB									2	4	7	1	2				1
	Co-trimoxazole	0.13–256	CAMHB				1	9	5					1	1
	(DIN, 1998)		ISOB					9	3	3				1	1
	GLYCOPEPTIDES
	Teicoplanin	0.06–128	CAMHB														1	16
	(NCCLS, 2000c)		ISOB															17
	Vancomycin	0.03–64	CAMHB											1			16
	(DIN, 1998)		ISOB														17
	KETOLIDES
	Telithromycin	0.03–64	CAMHB							1					16
	(Soussy et al., 2000)		ISOB											1	1	14	1
	ABT-773	0.03–64	CAMHB							1				16
	(Stone et al., 2000)		ISOB										1	13	3
	OTHER ANTIBIOTICS
	Chloramphenicol	0.06–128	CAMHB							2	14	1
	(DIN, 1998)		ISOB								16	1
	Nitrofurantoin	0.13–256	CAMHB												17
	(DIN, 1998)		ISOB												17
	Rifampicin	0.01–32	CAMHB					1			1	11	4
	(DIN, 1998)		ISOB								1	5	8	3
	Fosfomycin	0.13–256	CAMHB							1	1		7	4	4
	(SFM, 1998)		ISOB												4	8	3	2
	Fusidic acid	0.01–32	CAMHB												1	16
	(O.-L. et al., 1997)		ISOB												1	16
	Linezolid	0.03–64	CAMHB											5	12
	(EUCAST, 2001)		ISOB											2	14	1

a The number of strains for the corresponding MIC value is cited. A number in the lowest concentration of the antibiotic represents the maximal MIC value at this concentration (MIC = cmin → MIC ≤ cmin). An MIC value higher than the highest concentration tested is cited in the subsequent higher concentration step. MIC values in shaded areas indicate the clinically intermediate area according to the standard applied. A black thick line indicates the breakpoint between the clinically sensitive and clinically resistant strains, if the interpretation ‘intermediate’ does not exist. Abbreviations: DIN, Deutsches Institut für Normung e.V.; EUCAST, European Committee on Antimicrobial Susceptibility Testing; NCCLS, National Committee for Clinical Laboratory Standards; SFM, Comité de l’Antibiogramme de la Société Française de Microbiologie Communiqué; O.-L., Olsson-Liljequist et al.; CAMHB, cation-adjusted Mueller Hinton broth; ISOB, IsoSensitest broth.

Table 4. The natural antibiotic susceptibility of M. wisconsensis according to the standards in and NCCLS criteriaa

	Antibiotic		Medium	Naturally sensitive	Naturally intermediate	Naturally resistant	Assessment according to NCCLS Breakpoints1
	Tetracyclines	Tetracycline	CAMHB				Naturally sensitive in both media
			ISOB
		Doxycycline, Minocycline	CAMHB, ISOB				Identical assessment
	Aminoglycosides	All tested Aminoglycosides	CAMHB, ISOB				Identical assessment2
	β-Lactams	Oxacillin	CAMHB, ISOB				Identical assessment
		Penicillin G	CAMHB, ISOB				Naturally resistant
		All further tested β-Lactams	CAMHB, ISOB				Identical assessment3
	Quinolones	All tested Fluoroquinolones	CAMHB, ISOB				Identical assessment4
		Pipemidic acid	CAMHB				No breakpoints
			ISOB
	Macrolides	All tested Macrolides	CAMHB, ISOB				Identical assessment5
	Lincosamides	Lincomycin	CAMHB, ISOB				No breakpoints
		Clindamycin	CAMHB				Naturally resistant in both media
			ISOB
	Streptogramins	All tested Streptogramins	CAMHB, ISOB				Identical assessment
	Ketolides	Telithromycin, ABT-773	CAMHB, ISOB				NCCLS breakpoints were used in Table 3
	Glycopeptides	Teicoplanin, Vancomycin	CAMHB, ISOB				Identical assessment
	Antifolates	All tested Antifolates	CAMHB, ISOB				Identical assessment
	Other antibiotics	Fusidic acid	CAMHB, ISOB				No breakpoints
		Linezolid	CAMHB, ISOB				No breakpoints
		Rifampicin	CAMHB, ISOB				Identical assessment
		Fosfomycin	CAMHB				Identical assessment
			ISOB				Identical assessment
		Chloramphenicol	CAMHB, ISOB				Identical assessment
		Nitrofurantoin	CAMHB, ISOB				Identical assessment

1NCCLS criteria valid for Enterobacteriaceae (NCCLS 2000a) Neisseria gonorrhoe (NCCLS 2000b), staphylococci (NCCLS 2000c) and Pseudomonas aeruginosa and Non-Enterobacteriaceae (NCCLS 2000d) were applied. There are no NCCLS breakpoints for 2streptomycin, neomicin, apramycin, ribostamycin and lividomycin A; 3cefotiam; 4sparfloxacin and pefloxacin; 5roxithromycin. Abbreviations: CAMHB, cation-adjusted Mueller Hinton broth; ISOB, IsoSensitest broth.

a Please note that IsoSensitest broth is not recommended for antimicrobial susceptibility tests according to NCCLS criteria.

Strains of Moellerella were naturally sensitive to doxycycline, minocycline, all tested aminoglycosides, all β-lactams except oxacillin and benzylpenicillin, all fluoroquinolones, folate-pathway inhibitors, chloramphenicol and nitrofurantoin. Natural resistance was found with oxacillin, benzylpenicillin (to the latter some strains were of intermediate susceptibility), all tested macrolides including azithromycin, lincomycin, streptogramins, ketolides, glycopeptides, fusidic acid, linezolid and rifampicin.

4. Discussion 

Within the variety of Enterobacteriaceae, there are numerous species found predominantly or exclusively in human clinical specimens, without or with only poor knowledge of their incidence and clinical significance. Such taxa include Cedecea spp., Leminorella spp., M. wisconsensis, Tatumella ptyseos and Yokenella regensburgei(Brenner et al., 1992). In 2000, it was shown that Leminorella spp. that had been formerly recovered from stool and urine only, with no clinical correlates, should be considered as ′emerging nosocomial pathogens′, capable of causing numerous clinical syndromes, e.g., urinary tract and surgical site infections (Blekher et al., 2000). Although there is no evidence that such a classification might be justified with regard to M. wisconsensis, it is likely that it is involved in the pathogenesis of human disease. To obtain more profound information on the incidence of obviously unusual pathogens, it is important that an easy and reliable identification of respective strains is applicable. Commercially available identification systems are mainly adressed to the identification of well-known organisms with high medical importance and consist of a relative limited number of key tests useful to identify significant representatives within a particular group. Incomplete or sparse databases will tend to give wrong identification results. In regard to Moellerella, the data bases of both commercial identification systems applied in the present study contained M. wisconsensis, but 5 of 17 strains were misidentified with the Micronaut-E system (see results and Table 2). Interestingly, misidentification of weakly rhamnose- and xylose-positive moellerellae was obviously due to the high sensitivity of the Micronaut-E system, combined with a sparse data base (Table 2). In contrast to Micronaut-E, the api20E system was able to separate M. wisconsensis from K. pneumoniae subsp. ozaenae, mainly based on fermentation tests of arabinose and mannitol. However, because several mollerellae are known to produce acid from mannitol at prolonged incubation times Hickman-Brenner et al 1984, Farmer 1995, it would be advisable to perform additional sugar fermentation tests to secure a reliable identification (Table 2).

According to the data, it is possible that in the past some moellerellae may have been identified as K. pneumonaie subsp. ozaenae. Unfortunately, the exact origin and the former identification procedure of M. wisconsensis M-1, sent to our laboratory as K. pneumoniae subsp. ozaenae, is unknown. K. pneumonaie subsp. ozaenae is known to be a colonizer of the oral and nasopharyngeal mucosa and is the classic agent of an atrophic rhinitis called ozena, but has been implicated in several other respiratory and non-respiratory diseases Goldstein et al 1978, Tang and Chen 1994, Tang et al 1997. Interestingly, M. wisconsensis has also been isolated from bronchial aspirates Aldová 1992, Wallet et al 1994.

Apart from the characterization of metabolic features leading to reliable identifications, the main scope of the present study was to create a data compilation on the natural antibiotic susceptibility of M. wisconsensis. Antibiotic susceptibility patterns showed that moellerellae share several natural phenotypes with other Enterobacteriaceae and there were no phenotypic features characteristic for the species. The natural resistance of M. wisconsensis to benzylpenicillin, oxacillin, macrolides, lincosamides, streptogramins, glycopeptides, rifampicin and fusidic acid can be found in several Enterobacteriaceae and is predominantly due to drug exclusion by the cell envelope (for an overview see Nikaido, 1996). In contrast to some Enterobacteriaceae, M. wisconsensis was also naturally resistant to azithromycin, indicating specific ionic interactions on the cell surface, preventing the entry of this extended-spectrum azalide. Uniform natural azithromycin resistance has been found in all close neighbors of Moellerella, i.e., Proteus(Stock & Wiedemann, 1997) and Providencia species (Stock & Wiedemann, 1998), but also in Morganella morganii (Stock & Wiedemann, 1998) and other enterobacteria (for an overview see Stock, 1999). The natural rifampicin resistance was in contrast to the results of one previous study (Richard, 1989), in which all strains examined were sensitive to rifampicin, implying either other methods (which were not stated) and/or other clinical assessment criteria (not stated). It was shown that M. wisconsensis was naturally sensitive to all β-lactams except benzylpenicillin and oxacillin. Natural β-lactam susceptibilty patterns were similar to those of natural populations of E. coli and Shigella spp. (Stock & Wiedemann, 1999b), Edwardsiella tarda(Stock & Wiedemann, 2001) and Proteus mirabilis, which are known to produce small amounts of naturally occurring (chromosomally encoded) β-lactamases Clark et al 1991, Livermore 1995, Normark et al 1980, Reger et al 1993, Stock and Wiedemann 2001. Although there are no reports on the β-lactamase of Moellerella, it is likely that at least some strains produce small amounts of enzyme (unpublished data). The natural resistance of Moellerella to benzylpenicillin and oxacillin is regarded to be connected to the limited permeability of the outer membrane for these penicillins Curtis et al 1979, Livermore 1996. Because the interior channel size of the porins of several enterobacteria is broader than the molecular size of these β-lactams, it seems likely that their hydrophobicity is responsible for the failure to cross the outer membrane. Unfortunately, studies on the cell envelope of Moellerella have not been performed.

Apart from the antibiotics mentioned above, natural resistance of M. wisconsensis was found to fosfomycin in Isensensitest broth. Since there was no naturally occurring high-level resistance, it is likely that a reduced permeability of the cell envelope to this antibiotic rather than a specific fosfomycin:glutathione-S-transferase contributed to the low-level resistance of several strains. Alternatively, a transferase with low affinity to fosfomycin or a low-level enzyme expression could have been responsible for the phenotype observed (O’Hara, 1993). Results in fosfomycin susceptibility testing can be highly dependent on the constituents of media and several further factors Barry and Fuchs 1991, Patel et al 1997. In contrast to one previous study dealing with natural antibiotic susceptibilities of Enterobacter spp. (Stock et al., 2001), it was shown that different media have major influence on the fosfomycin MICs for M. wisconsensis, resulting in the absence of strains with natural resistance, applying CAMHB (Table 3). This is in agreement with respective MIC data for Staphylococcus species: Kresken et al. showed that MIC values of fosfomycin for staphylococci were significant higher in IsoSensitest broth than in Mueller Hinton media (Kresken et al., 2000). Thus, whether or not (and to which extent) a medium-dependency for fosfomycin susceptibility testing exists, seems to be highly related to the species examined. The biochemical basis of the described phenomenon is poorly understood. In contrast to Mueller Hinton media, IsoSensitest broth contains glucose that inhibits the fosfomycin uptake into the bacterial cell (resulting in higher MICs) (Kresken et al., 2000). To enhance its uptake, the addition of glucose-6-phosphate (G-6-P) is recommended for fosfomycin susceptibility testing. However, in the present study and in the study of Kresken et al. (Kresken et al., 2000) fosfomycin susceptibility testing was performed in the presence of G-6-P, indicating that a large amount of constituents acting inhibitory on the uptake of fosfomycin, not compensated by G-6-P, remain, when Isonsensitest broth in fosfomycin susceptibility testing for moellerellae and staphylococci is applied. Medium-dependencies in susceptibility testing affecting clinical assessment criteria were also found with tetracycline, pipemidic acid and clindamycin, although the medium-associated differences of the respective MICs were smaller than with fosfomycin (Table 3). It is likely that at least some of these differences can be attributed to variations in cation concentrations. In studies with tetracyclines it was shown that in tryptic soy solution, tetracycline MICs for S. aureus, E. coli and Klebsiella species were three or more doubling dilution steps higher than in Mueller Hinton broth, attributed to medium-dependent differences in cation concentrations (Amsterdam, 1996).

In conclusion, the natural resistance of M. wisconsensis to azithromycin and its decreased susceptibility to fosfomycin should give rise to attention for antibiotic therapy since both drugs are in use for treatment of diarrheagenic diseases. Although the present study is not suitable to draw conclusions on the incidence of acquired resistance, it is conspicuous that some Moellerella strains were resistant to tetracyclines, trimethoprim and sulfamethoxazole (Table 3). The natural susceptibility patterns described, the presence of strains with acquired resistance, and the probable occurrence of moellerellae misidentified as Klebsiella should direct the interest in bacterial pathogens also to M. wisconsensis.