Isolated from Human Feces - Journal of Clinical Microbiology

4 downloads 0 Views 2MB Size Report
World Health Organization Collaborating Centre for Enteric Campylobacter, St. Pieters ... organisms from blood cultures of pediatric patients in South. Africa; Walmsley and Karmali (39) isolated "C. upsaliensis" ..... A total of 60 strains grew in the presence of 1% glycine, and .... taking a representative strain from each class.
JOURNAL OF CLINICAL MICROBIOLOGY, May 1990, p. 1039-1046 0095-1137/90/051039-08$02.00/0

Vol. 28, No. 5

Copyright ©) 1990, American Society for Microbiology

Characterization and Description of "Campylobacter upsaliensis" Isolated from Human Feces HERMAN GOOSSENS,l* BRUNO POT,2 LINDA VLAES,' CHANTAL VAN DEN BORRE,' RUDY VAN DEN ABBEELE,' CHRISTINE VAN NAELTEN,' JACK LEVY,' HUGO COGNIAU,' POL MARBEHANT,' JAN VERHOEF,3 KAREL KERSTERS,2 JEAN-PAUL BUTZLER,' AND PETER VANDAMME2 World Health Organization Collaborating Centre for Enteric Campylobacter, St. Pieters University Hospital, B-1000 Brussels,' and Laboratorium voor Microbiologie en Microbiële Genetica, Rijksuniversiteit, B-9000 Gent,2 Belgium, and Department of Microbiology, University of Utrecht, Utrecht, The Netherlands3 Received 1 November 1989/Accepted 18 January 1990

During a 3-year period, "Campylobacter upsaliensis" was isolated from 99 patients. Phenotypic characterization and numerical analysis of protein electrophoregrams showed evidence that "C. upsaliensis" is a distinct Campylobacter species with unique characteristics. The MBCs of 13 antibiotics were determined. In general, these organisms were highly susceptible to drugs that were present in the selective isolation media, making none of the available selective media suitable for the isolation of "C. upsaliensis." Ten strains were found to be resistant to erythromycin (MBCs, .12.50 mg/liter). Plasmid DNA was detectable in 89 of the 99 strains; 16 plasmid profiles could be identified. Plasmid pattern 16, containing four plasmids of 52, 32, 5.5, and 2.6 megadaltons, represented 60.7% of the plasmid-containing strains. None of the "C. upsaliensis" strains could be agglutinated with antisera against heat-labile antigens from C. jejuni, C. coli, or C. laridis. "C. upsaliensis" was found to be susceptible to serum killing and was readily phagocytized by human polymorphonuclear cells.

Catalase-negative or weakly positive (CNW) strains of Campylobacter were first isolated from dogs by Sandstedt et al. (31). In the same report, those investigators were able to demonstrate, by DNA hybridization, that these organisms belonged to a new species; the name "Campylobacter upsaliensis" was proposed in 1986 (K. Sandstedt and J. Ursing, Abstr. XIV Int. Congr. Microbiol. P.B8-17, p. 61, 1986). The species description was minimal, although a type strain (NCTC 11541 = LMG 8850) was designated. Neither the species name nor the type strain has standing in the literature, because the species description was not published in a peer-reviewed journal. CNW strains were isolated from five children with diarrhea by Mégraud and Bonnet in France (F. Mégraud and F. Bonnet, Letter, J. Infect. 12:275-276, 1986); Steele and McDermott (32) recovered eight CNW isolates from diarrheic stools of Australian children and one CNW isolate from an adult; Patton et al. (27) showed evidence that "C. upsaliensis" is responsible for enteritis in both adults and children in the United States; Taylor et al. (36) isolated "C. upsaliensis" from seven patients, both adults and children, with diarrhea in Canada; Lastovica et al. (16) isolated these organisms from blood cultures of pediatric patients in South Africa; Walmsley and Karmali (39) isolated "C. upsaliensis" from six children. Most recently, Fox et al. (8) isolated "C. upsaliensis" from three asymptomatic cats. In this report, we characterize and describe the "C. upsaliensis" group, using strains isolated in our hospital. We use the species name in quotation marks to indicate its proposed name and refer to the "type strain" as a reference strain.

*

MATERIALS AND METHODS

Isolation procedure. Stool specimens were cultured for the presence of "C. upsaliensis" from 1 July 1986 to 30 June 1989. A total of 15,185 fecal samples were cultured for Campylobacter species by the filter method (32), and two solid selective isolation media were used: (i) our previously described modified solid medium (9), which is offered commercially by Virion AG (Zurich, Switzerland), and (ii) the blood-free medium (4), which is offered commercially by Oxoid Ltd. (CM739 and SR125; Oxoid Ltd., Basingstoke, United Kingdom). A newly described semisolid selective medium (11) was also used for the last 4,878 stool specimens. Plates were incubated in an atmosphere of 84% N2-10% C02-6% 02. Suspected Campylobacter colonies were stained with crystal violet. The oxidase and catalase tests were carried out, and the biotypes of all the campylobacters that were isolated were determined by the biotyping scheme of Lior (17). Campylobacter isolates which were CNW or which could not be identified on the basis of the biotyping scheme of Lior (17) were further characterized phenotypically and were finally identified on the basis of their protein electrophoregrams.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of whole-cell proteins. We prepared whole-cell protein extracts of 130 Campylobacter strains, including (i) 105 campylobacters isolated in our hospital; (ii) 6 reference strains of "C. upsaliensis" isolated in Sweden, South Australia, the United Kingdom, and the United States; and (iii) 19 reference strains of other human enteropathogenic campylobacters including C. jejuni subsp. jejuni, C. jejuni subsp. doylei, C. coli, C. laridis, and C. hyointestinalis (Table 1). All strains were grown at 37°C on Mueller-Hinton agar (CM337; Oxoid Ltd.) supplemented with 5% (vol/vol) horse blood and incubated in a microaerophilic atmosphere consisting of approximately 4% C02-5% 02-4% H2-87% N2. The protein

Corresponding author. 1039

1040

J. CLIN. MICROBIOL.

GOOSSENS ET AL.

TABLE 1. Reference strains included in the sodium dodecyl Other strain designation(s)a b Bacteria and straina.b

sulfate-polyacrylamide gel electrophoresis study Sourceb

Source of isolation

"C. upsaliensis" LMG 8850C LMG 8851 LMG 8852 LMG 8853 LMG 7533 LMG 7915

NCTC 11541 NCTC 11840, 926/H184 NCTC 11540 NCTC 11926, IMVS 1089 CCUG 14913 CCUG 20818, CDC D533

NCTC NCTC NCTC NCTC E. Falsen E. Falsen

Canine feces (Sweden) Human feces (United Kingdom) Canine feces (Sweden) Human feces (South Australia) Canine feces (Sweden) Human blood (United States)

C. jejuni subsp. jejuni LMG 6444T LMG 7534 LMG 9217 LMG 9218 LMG 9219

CCUG 11284T CCUG 14914 A452 E277 A458

E. Falsen E. Falsen Our isolate Our isolate Our isolate

Bovine feces

NCTC 11951T

NCTC

Human feces

C. coli LMG 6440T LMG 7535 LMG 8530

CCUG 11283T CCUG 10369 CCUG 8320

E. Falsen E. Falsen E. Falsen

Porcine feces Porcine placenta Porcine origin

C. laridis LMG 8846T LMG 8844 LMG 8845 LMG 7607 LMG 7929

NCTC 11352T NCTC 11844 NCTC 11458 CCUG 12774 CCUG 12773

NCTC NCTC NCTC E. Falsen E. Falsen

Cloacal swab of a herring gull Seawater Human feces Human feces Canine feces

Campylobacter sp. strain

CCUG 18267

E. Falsen

River water

CCUG 14169T CCUG 20823, CDC D2189 CCUG 20822, CDC D2170 CCUG 24265

E. E. E. E.

C. jejuni subsp. doylei LMG 8843T

Human feces Human feces Human feces

LMG 7791d

C. hyointestinalis LMG 7817T LMG 8216 LMG 8753 LMG 9151

Falsen Falsen Falsen Falsen

Porcine intestine Human feces Human feces Porcine origin

a Type strains are indicated by a superscript T. b Source abbreviations: CDC, Centers for Disease Control, Atlanta, Ga.; IMVS, Institute of Medical Veterinary Science, Adelaide, South Australia, Australia; LMG, Culture Collection Laboratorium voor Microbiologie, Gent, Belgium; NCTC, National Collection of Type Cultures, Central Public Health Laboratory, London, United Kingdom; E. Falsen, Culture Collection of the University of Goteborg (CCUG), Department of Clinical Bacteriology, University of Goteborg, Goteborg, Sweden. C Reference strain of Sandstedt et al. (31). d

Urease-positive thermophilic Campylobacter strain of Bolton et al. (3).

were prepared, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed by using small modifications of the procedure of Laemmli (14) as described previously (13). Numerical analysis of the protein electrophoregrams. The densitometric analysis and normalization and interpolation of the protein electrophoregrams were performed as described by Pot et al. (29). The dense protein bands from the 39,000- to 60,000-molecular-weight region were excluded from the numerical analysis. The numerical analysis was performed as described by Kersters and De Ley (12) on points 6 to 130 and points 180 to 350 of each interpolated trace. The similarities between all pairs of traces were expressed by the Pearson product moment correlation coefficient (r), and clustering was performed by the unweighted pair group method by using average linkage. Phenotypic characterization. Phenotypic characterization was performed on all 99 "C. upsaliensis" strains isolated in our hospital, together with the type strains of C. jejuni subsp. jejuni, C. jejuni subsp. doylei, C. coli, C. laridis, C. hyointestinalis, C. fetus subsp. fetus, and the proposed type

extracts

strain of "C. upsaliensis," LMG 8850 (Table 1). All cultures were incubated under microaerophilic conditions for 48 to 72 h at 37°C, unless specified otherwise. The morphology ofthe cells was evaluated by Gram staining (by using 0.3% carbol fuchsin as a counter stain) and electron microscopy (by negative staining of cells with 2% uranyl acetate and observation by electron microscopy [magnification, x150,000; EM 201; Philips]). Motility was observed in young cultures by examining wet mounts in broth by phase-contrast microscopy. After growth on Mueller-Hinton agar (CM337; Oxoid Ltd.) with 5% sheep blood for 48 h, coccoid transformation was recorded daily, and the plates were left on the laboratory bench. Temperature tolerance was tested on MuellerHinton agar, with and without 5% sheep blood, at 25, 37, and 42°C. Aerobic growth was carried out on Mueller-Hinton agar, with and without 5% sheep blood, at 37°C. Anaerobic growth was tested at 37°C on Mueller-Hinton agar, with and without 5% sheep blood, under atmospheres of 90% N2-10% C02 and 75% N2-15% H2-10% C02. Anaerobic growth in the presence of trimethylamine-N-oxide hydrochloride (TMAO) was carried out as described previously (2). For the

VOL. 28, 1990

oxidase test, a piece of filter paper was moistened with a few drops of a 1% (wt/vol) solution of tetramethyl-p-phenylene-

diamine dihydrochloride (Sigma Chemical Co., St. Louis, Mo.). A few colonies were smeared on the filter paper. For the catalase test, bacterial growth from a blood agar plate was suspended in a drop of 3 or 10% hydrogen peroxide on a glass slide; the drop was covered with a cover slip and observed for 2 min under the microscope. An immediate reaction (within 10 s) was considered positive; a reaction after that was considered weakly positive. Nitrate and nitrite reduction was tested in a semisolid bruceila medium composed of brucella broth (152-0860; GIBCO Ltd., Paisley, Scotland) and 1.5 g of Bacto-Agar (Difco Laboratories, Detroit, Mich.) per liter with 1% KNO3, which was inoculated with a few drops of a 48-h-old culture in thioglycolate medium (152-04780; GIBCO Ltd.). The test was done as described previously (21). Hydrogen sulfide production in triple sugar iron (TSI) agar was carried out as described previously (21). The rapid H2S test in FBP medium was done as described by Lior (17). Oxidation or fermentation of carbohydrates was carried out by inoculating a few drops of bacterial suspension in a semisolid brucella medium (GIBCO Ltd.) supplemented with 1% glucose and 0.002% phenol red. The hippurate hydrolysis test was done in tubes (17) and by gas-liquid chromatography (1). Urea hydrolysis was tested in urea broth base (CM71; Oxoid Ltd.) inoculated with three loopfuls of bacteria. Growth in the presence of 1% glycine, 1% oxgall, and 3.5% NaCl was done in a semisolid brucella medium (GIBCO Ltd.). A pellicle of growth similar to that which occurred in the unsupplemented brucella medium (GIBCO Ltd.) was considered a positive reaction. Growth on agar plates containing 1.5% NaCl was carried out as described previously (2). The capacity to grow in the presence of2,3,5-triphenyltetrazolium chloride (TTC) was demonstrated on brucella agar (GIBCO Ltd.), with and without 5% sheep blood, containing 0.4 and 1.0 g of TTC per liter. The presence of red colonies over more than one-third of the plate after 5 days was considered positive. Tolerance to nalidixic acid and cephalothin was determined as described previously (21). The DNA hydrolysis test was done as described by Lior (17), and for the indoxyl acetate test the method of Mills and Gherna (20) was followed. All tests were done by two investigators working separately, and each investigator performed the tests in duplicate. Antibiotic susceptibility testing. MICs and MBCs were determined as described previously (10). The following drugs were tested: ampicillin (Beecham), gentamicin (Schering-Essex), cefoperazone, streptomycin, and tetracycline (Pfizer), colistin (Bellon), vancomycin (Eli-Lilly), rifampin (Dow-Lepetit), chloramphenicol (Lepetit), amphotericin B (Squibb), erythromycin (Abbott), and trimethoprim and cotrimoxazole, the 20:1 combination of sulfamethoxazole-trimethoprim (Roche). Escherichia coli ATCC 25922 and Staphylococcus aureus ATCC 25923 were tested simultaneously to ensure the potency of each drug. Plasmid analysis. Plasmid DNAs of all "C. upsaliensis" strains were extracted by the method of Portnoy et al. (28), except that the pH of the lysis buffer was 12.02. Each "C. upsaliensis" strain was tested at least five times for the presence of plasmids. Serotyping. All "C. upsaliensis" strains were serotyped with heat-labile antigens by the method of Lior et al. (18). Bactericidal assay and phagocytosis. A suspension of Campylobacter cells with an optical density of 0.600 at 660 nm

CHARACTERIZATION OF "C. UPSALIENSIS"

1041

was labeled with 100 ,uCi of 51Cr for 1 h at 37°C. After washing, a final concentration of labeled bacteria of 1.3 x 1010 CFU/ml (optical density, 0.200 at 660 nm) was used for further studies. Isolates were assayed for their susceptibilities to complement-mediated bactericidal activity and phagocytosis by using pooled normal human serum. The bactericidal activity of the serum was tested by mixing the 51Cr-labeled Campylobacter cells with 30% pooled normal human serum for 1 h at 37°C. The percent lysis was determined as the percentage of bacterial radioactivity that was released into the supernatant. Phagocytosis with human leukocytes was determined by incubating human polymorphonuclear cells for 12 min with Campylobacter cells, which were preopsonized for 30 min with 30% normal human serum. The percent phagocytosis was calculated as the percentage of radioactivity that was associated with leukocytes. RESULTS Isolation. Organisms resembling Campylobacter strains were isolated from 802 patients during the 3-year study; 753 strains could be identified further. C. jejuni subsp. jejuni, C. jejuni subsp. doylei, C. coli, C. laridis, C. hyointestinalis, and "C. upsaliensis" were isolated from 556, 4, 91, 1, 2, and 99 patients, respectively. The following isolation rates for C. jejuni, C. coli, and C. laridis were found with the following different isolation media: our solid blood-based Virion medium (9), 91.6%; the solid blood-free Oxoid medium (4), 93.8%; and our new semisolid blood-free medium (12), which was tested on the last 4,878 stool specimens only, 93.2%. The isolation rate was 81.6% by the filter method (32). All C. jejuni subsp. doylei strains were isolated by the filter method only. One C. hyointestinalis strain was isolated with both solid selective media only, and the other C. hyointestinalis strain was isolated by the filter method and with the solid blood-free Oxoid medium (the semisolid medium had not yet been tested at the time that both C. hyointestinalis strains were isolated). All 99 "C. upsaliensis" strains were isolated by the filter method; however, in only 4 cases, these organisms were recovered simultaneously from the selective media. Further characterization and description of these 99 "C. upsaliensis" strains is reported below. Polyacrylamide gel electrophoresis of whole-cell proteins. Duplicate protein extracts of several strains were prepared to check the reproducibility of extract preparations, which was at least 94%. We found an excellent agreement between the electrophoretic clusters and the taxonomic groups that were studied. The numerical analysis of the protein electrophoregrams revealed four major clusters at a correlation level of r = 0.86; two of these clusters could be subdivided into two subclusters (Fig. 1). Cluster I contained all except two CNW strains isolated in our hospital and the six reference strains of "C. upsaliensis" that were grouped above r = 0.88. Cluster II consisted of two subclusters; subclusters IIa and IIb contained the reference strains of C. laridis and C. coli, which grouped at r > 0.94 and r > 0.98, respectively. Cluster III contained all C. jejuni strains: subcluster IIIa contained the type strain of C. jejuni subsp. doylei, the two remaining CNW strains (LMG 9130 and LMG 9143), as well as two unidentified catalase-positive strains (LMG 9243 and LMG 9255) isolated at our hospital; all of these strains grouped at r > 0.90; subcluster IIIb contained all C. jejuni subsp. jejuni reference strains, which grouped at r > 0.88.

l~ V

GOOSSENS ET AL.

1042

J. CLIN. MICROBIOL.

r

0.8

'

À

I

Ila

p

I

0.8

.

(105)

il1

9230

.i'ttl.i ^"

1I

I

I

"C. upsaliensis" (6) + CNW strains (99)

Il

1~

UIt

fl

.UlUIl

9124 9226 7915 8851 8843T 9130 9143 9255 9243 _ 7607 7929 8846T. 7535 644 0T. 9217

I (6) C. laridis

(6) C. hyointestinalis

1.0

FIG. 1. Simplified dendrogram derived from the unweighted pair linkage of correlation coefficients (r) of the protein patterns of all strains studied. Roman numerals indicate cluster numbers; numbers in parentheses refer to the number of strains belonging to a given cluster. CNW, Catalase-negative or weakly positive strains; CP, catalase-positive strains.

group average

Cluster IV consisted of the C. hyointestinalis reference strains and two unidentified catalase-positive strains (LMG 9260 and LMG 9276), which grouped at r > 0.86. Figure 2 shows the protein profiles of representative strains (Table 1) of each electrophoretic cluster. Phenotypic characterization. Table 2 summarizes the main phenotypic characteristics of "C. upsaliensis" in comparison with those of the type strains of the other Campylobacter species that we studied. All "C. upsaliensis" strains were gram negative, actively motile, curved or spiral rods with a single flagellum at one or both ends. All cells were 0.2to 0.5-,tm wide and 0.5- to 8->xm long. Swarming occurred in 37 strains. All strains were nonhemolytic. Coccoid transformation was observed for all strains after 1 to 14 days (mean, 5.3 days). None of the strains grew in an aerobic atmosphere, nor did any of the strains grow at 25°C in a microaerophilic atmosphere. A total of 19 strains did not grow at 42°C under microaerophilic conditions on MuellerHinton agar; however, 11 strains grew well when the agar base was supplemented with 5% sheep blood. Growth was enhanced by the presence of H2. None of the strains showed growth in an anaerobic atmosphere without H2; however, with 15% H2 under anaerobic conditions, 77 strains showed

I

I

Il I I

"C. upsaliensis'

C. jejuni subsp. doylei C. laridis C. Col

6444T_411J

C. jeluni subsp. jejuni

8216

C. hyointestinalis

9276 9151

I

r

1

I

té e t

A-4Ib (3) $. coli IeC.juni subsp. doylei ) +CNW strains (2) L CP strains (2) llb (5) C. jejuni subsp. jeluni

I

9222 91 33 8853 8850 8852

f4 * |*

1.0

99260 781 7T_.. MWM

FIG. 2. Protein profiles of a representative selection of strains of each electrophoretic cluster. The positions of the following molecular weight markers (lane labeled MWM) are indicated from left to right, respectively: lysozyme, 14,500; trypsin inhibitor, 20,100; carbonic anhydrase, 29,000; glyceraldehyde-3-phosphate dehydrogenase, 36,000; egg albumin, 45,000; bovine albumin, 66,000; and P-galactosidase, 116,000.

growth. No strains grew anaerobically in the presence of TMAO. On primary isolation the strains were slow growing, with pinpoint colonies appearing after 2 days of incubation. All strains were oxidase positive and showed weak catalase reactions with 3% as well as 10% H202; they all reduced nitrate, but not nitrite, and did not ferment or oxidize D-glucose. Hippurate was not hydrolyzed. No H2S production occurred in TSI or FBP agar. Urea was not hydrolyzed. A total of 60 strains grew in the presence of 1% glycine, and 79 strains grew in the presence of 1% oxgall. None of the strains grew in 1.5% NaCl. Two strains grew on plates with 3.5% NaCl. All strains were inhibited by 1.0 g of TTC per liter; tolerance to 0.4 g of TTC per liter occurred for 72 strains when they were grown on Mueller-Hinton agar with 5% sheep blood, but there was no tolerance when they were inoculated on Mueller-Hinton agar only. No DNase activity could be demonstrated for any of the strains. All strains were able to hydrolyze indoxyl acetate. All strains were susceptible to cephalothin (diameter range, 37 to 78 mm; mean diameter, 58.8 mm) and nalidixic acid (diameter range, 19 to 45 mm; mean diameter, 31.4 mm). Antimicrobial susceptibility. Table 3 shows the MBCs of 13 antimicrobial agents for 91 of 99 "C. upsaliensis" strains. We report the MBCs and not the MICs because interpretation of growth of these campylobacters in broth is difficult and unreliable. In general, these organisms were highly susceptible to the drugs that were present in selective isolation media, such as cefoperazone, colistin, vancomycin, rifampin, and trimethoprim. Gentamicin and chloramphenicol were highly active. The MBCs of erythromycin were -12.5 ,ug/ml for 10 strains. None of the strains were resistant to tetracycline; two strains were resistant to ampicillin (MBCs, .25 ,tg/ml). Plasmid analysis. Plasmids were detected in 89 of 99 "C.

VOL. 28, 1990

CHARACTERIZATION

OF "C. UPSALIENSIS"

1043

TABLE 2. Phenotypic characteristics of 99 "C. upsaliensis" strains isolated from human feces and Campylobacter reference and type strains % of the 99 tested "CC

Characteristic

upsaliensis" strains that were

Oxidase Catalase Urease Nitrate reduced to nitrite H2S production TSI slant FBP agar Growth at: 250C 420C Growth in: 1% Glycine 1% Oxgall 3.5% NaCl 1.5% NaCl Growth on 0.04% TTC Anaerobic growth in 0.1% TMAO Hippurate hydrolysis Susceptibilityd to: Nalidixic acid

Cephalothin DNase Indoxyl acetate hydrolysis

.

.

upsaliensis" LMG 8850b

ssubsp. * eubp Jejuni LMG

Cus. subsp. doylei LMG

C. Coli

LMG 644T

LMG 8846T

tinalis LMG

+ + -

+ + +

+ + +

+ +

+ +

+

+

-

+ -

+

+ +

jjn

C. laridis

C.

hyoint'es-

C. fetus

subsp. fetus

7817T LMG 6442T LMGT6843T

positive

100 100c 0 100

+

+ + +

0 0

-

-

O

-

-

-

-

+C

+

81

+

+

+

+

+

+

+

+

+ +

+ +

+ +

+ +

+ +

+ +

-

-

+

+

+ +C

61 80 2 0 0

-

O

-

+ -

0

-

+

+

100 100 0 100

S S

S R

+

+

-

-

-

+

+

S S

S R

R R

R S

+

+

R S

Type strains are indicated by a superscript T. Reference strain of Sandstedt et al. (31). C Weak reaction. d S, Susceptible; R, resistant.

a

b

upsaliensis" strains. Sixteen different patterns could be distinguished (Table 4). Plasmid pattern 16, consisting of plasmids with molecular sizes of 52, 32, 5.5, and 2.6 megadaltons (MDa), was the most frequently encountered pattern and was present in 60.7% of all strains that contained plasmids. Overall, the common plasmids were those with sizes of 52 (82.0%), 32 (70.8%), 5.5 (61.8%), and 2.6 (61.8%) MDa. Figure 3 illustrates the plasmid profiles of all 16 types by taking a representative strain from each class. Serotyping. None of the "C. upsaliensis" strains agglutinated in antisera against heat-labile antigens by the method of Lior et al. (18).

Bactericidal assay and phagocytosis. We included eight representative strains of "C. upsaliensis." All eight strains tested were susceptible to the bactericidal activity present in normal human serum and were adequately phagocytized by human polymorphonuclear cells (Table 5). The percent serum lysis and phagocytosis of 13 C. jejuni and C. coli isolates from stools (27.5 and 45.5%, respectively) was statistically not significantly different from the percent serum lysis and phagocytosis of the "C. upsaliensis" isolates obtained from stool specimens. However, a statistically significant difference was found between the percent serum lysis and phagocytosis of 20 blood isolates of C. fetus (8.7 and 33.4%, respectively) and those of the "C. upsaliensis"

TABLE 3. Susceptibilities to antimicrobial agents of 91 "C. upsaliensis" isolates isolated from human feces Antimicrobial agent

Ampicillin

Cefoperazone Gentamicin Streptomycin Colistin

Vancomycin Rifampin Chloramphenicol Amphotericin B Tetracycline Trimethoprim Co-trimoxazole Erythromycin

No. of isolates susceptible at the following MBCs (iig/ml):

-0.097

0.195

0.39

0.78

1.56

3.12

6.25

35 5 28 7 70 2 6 2 2 30 2 2 3

5 3 27 3 2 2

10 5 22 2 3 3 3 9 1 13

12 6 9 5

18 2 1 6

5 10 1 5 3 1 9 23 6 1 3 3 12

3 17 1 5 1 2 22 10 5 1 3 3 3

4

39 1 2 3

4 13

4 5 1 14 2 5 1

22

3 9 24 3 2 4 10 25

12.5

25

50

1

2 16

5

5 1 18 5 1 12

4 1 20 4 1 17

13 1 22 1 13

19

6 14 8

8 19 1

12 14

45 14

23 1 4 5 9 31 3 il

6 6 1

100

>100

1 32

4

J. CLIN. MICROBIOL.

GOOSSENS ET AL.

1044

TABLE 5. Percent susceptibilities to normal human serum and phagocytosis of eight representative "C. upsaliensis" strains

TABLE 4. Plasmid profile types and sizes of the 89 tested "C. upsaliensis" strains Plasmid profile type

1 2 3 4 5 6 7 8 9 10 il 12 13 14 15 16

Size (MDa) of

of plasmids.No. strains

Size (MDa) of plasmids

8.6, 7.9, 5.0 23, 1.3, 0.90; Fig. 1 and 2) and are therefore considered to belong to this taxon. Finally, another two unidentified catalase-positive strains (LMG 9260 and LMG 9276) were identified as C. hyointestinalis (Fig. 1 and 2). From the phenotypic analysis it was clear that "C. upsaliensis" can be easily identified and distinguished from other Campylobacter species (Table 2). Sandstedt et al. (31) reported the isolation of CNW campylobacters subsequently referred to as "C. upsaliensis" (Sandstedt and Ursing, Abstr. XIV Int. Congr. Microbiol. 1986, P.B8-17, p. 61, 1986). However, as weak catalase reactions may also occur for C. jejuni subsp. doylei strains (two of our four C. jejuni subsp. doylei strains that were isolated; 33), we suggest, to avoid further confusion, that investigators no longer refer to "C. upsaliensis" as CNW campylobacters. The very high proportion (89.9%) of "C. upsaliensis" strains containing plasmids is another feature that distinguishes "C. upsaliensis." We were not able to allocate antibiotic resistance or other phenotypic markers to these plasmids. Plasmid profiling has not been used extensively as an epidemiological typing tool for Campylobacter species (19), mainly because plasmids are recovered from clinical isolates of C. jejuni at low rates (5). If it is confirmed by other investigators that "C. upsaliensis" is quite different from C. jejuni, in that a high proportion of strains contain multiple plasmids, then plasmid profiling might become a useful marker for distinguishing "C. upsaliensis." Moreover, since serotyping on the basis of heat-stable (27) and heat-labile antigens from C. jejuni, C. coli, and C. laridis is not useful, a serotyping system specific for "C. upsaliensis" would have to be developed. Lastovica and Ambrosio (15) found that 7 of the 24 isolates of Campylobacter with no or weak catalase activity contained plasmids. However, from subsequent reports (33, 34) it appeared that most of their CNW campylobacters belonged to C. jejuni subsp. doylei. This again highlights the confusion in the literature between CNW campylobacters and "C. upsaliensis." In vitro antibiotic susceptibility testing (Table 3) of the "C. upsaliensis" strains showed that they were generally susceptible to ampicillin, gentamicin, chloramphenicol, and

VOL. 28, 1990

tetracycline. Ten strains were found to be resistant to erythromycin (MBCs, .12.5 mg/liter). For primary isolation and subculturing, we recommend that at least 4% H2 be included in the atmosphere. Initially, subculturing on solid blood agar is recommended. We believe that none of the available selective isolation media are satisfactory for the isolation of "C. upsaliensis" (H. Goossens and J.-P. Butzler, Letter, J. Clin. Microbiol. 27:2143-2144, 1989). Only four "C. upsaliensis" strains were recovered from the selective media. Because of the wide range of susceptibilities to the drugs that were present in the selective isolation media (Table 3), we failed to develop a new selective medium for the isolation of "C. upsaliensis" from clinical specimens. Therefore, microbiologists interested in the isolation of "C. upsaliensis" should include the filter method (32) in their isolation procedure. The bactericidal assays and the study of phagocytosis of "C. upsaliensis"demonstrated a pattern similar to that of C. jejuni. Stool isolates were susceptible to serum killing and phagocytosis, whereas it was shown by Patton et al. (27) that blood isolates are usually resistant to complement-mediated bactericidal activity. Results of our study provide enough evidence that "C. upsaliensis" is a distinct Campylobacter species with unique phenotypic characteristics. We hope that the results of this study wil draw more attention to the importance of "C. upsaliensis" and encourage investigators to perform additional studies to better define their clinical, epidemiological, and pathogenic roles. ACKNOWLEDGMENTS K.K. is indebted to the Fonds voor Geneeskundig Wetenschappelijk Onderzoek, Belgium, for research and personnel grants. P.V. is indebted to the Instituut tot Aanmoediging van het Wetenschappelijk Onderzoek in Nijverheid en Landbouw, Belgium, for a scholarship. Part of this research was carried out in the framework of contract BAP-0138-B from the Biotechnological Action Program of the Commission of the European Community. We are indebted to Dirk Dewettinck for excellent technical assistance and thank E. Falsen, Culture Collection, University of Goteborg, Goteborg, Sweden, for the gift of numerous strains. LITERATURE CITED 1. Bar, W., and G. Fricke. 1987. Rapid and improved gas liquid chromatography technique for detection of hippurate hydrolysis by Campylobacter jejuni and Campylobacter coli. J. Clin. Microbiol. 25:1776-1778. 2. Benjamin, J. L., S. Leaper, R. J. Owen, and M. B. Skirrow. 1983. Description of Campylobacter laridis, a new species comprising the nalidixic acid resistant thermophilic Campylobacter (NARTC) group. Curr. Microbiol. 8:231-238. 3. Bolton, F. J., A. V. Holt, and D. N. Hutchinson. 1985. Ureasepositive thermophilic Campylobacter. Lancet i:1217-1218. 4. Bolton, F. J., D. N. Hutchinson, and D. Coates. 1984. Blood-free selective medium for isolation of Campylobacter jejuni from feces. J. Clin. Microbiol. 19:169-171. 5. Bradbury, W. C., A. D. Pearson, M. A. Marko, R. V. Congi, and J. L. Penner. 1984. Investigation of Campylobacter jejuni outbreak by serotyping and chromosomal restriction endonuclease analysis. J. Clin. Microbiol. 19:342-346. 6. Costas, M., R. J. Owen, and P. J. H. Jackman. 1987. Classification of Campylobacter sputorum and allied campylobacters based on numerical analysis of electrophoretic protein patterns. Syst. Appl. Microbiol. 9:125-131. 7. Flores, B. M., C. L. Fennell, and W. E. Stamm. 1989. Characterization of Campylobacter cinaedi and C. fennelliae antigens and analysis of the human immune response. J. Infect. Dis. 159:635-640. 8. Fox, J. G., K. O. Maxwell, N. S. Taylor, C. D. Runsick, P.

CHARACTERIZATION OF "C. UPSALIENSIS"

1045

Edmonds, and D. J. Brenner. 1989. "Campylobacter upsaliensis" isolated from cats as identified by DNA relatedness and biochemical features. J. Clin. Microbiol. 27:2376-2378. 9. Goossens, H., M. De Boeck, H. Coignau, L. Vlaes, C. Van den Borre, and J. P. Butzler. 1986. Modified selective medium for the isolation of Campylobacter spp. from feces: comparison with Preston medium, a blood-free medium, and a filtration system. J. Clin. Microbiol. 24:840-843. 10. Goossens, H., P. De Mol, H. Coignau, J. Levy, O. Grados, G. Ghysels, H. Innocent, and J. P. Butzler. 1985. Comparative in vitro activities of azthreonam, ciprofloxacin, norfloxacin, ofloxacin, HR 810 (a new cephalosporin), RU 28965 (a new macrolide), and other agents against enteropathogens. Antimicrob. Agents Chemother. 27:388-392. 11. Goossens, H., L. Vlaes, I. Galand, C. Van den Borre, and J. P. Butzler. 1989. Semisolid blood-free selective-motility medium for the isolation of Campylobacter from stool specimens. J. Clin. Microbiol. 27:1077-1080. 12. Kersters, K., and J. De Ley. 1975. Identification and grouping of bacteria by numerical analysis of their electrophoretic protein patterns. J. Gen. Microbiol. 87:333-342. 13. Kiredijan, M., B. Holmes, K. Kersters, I. Guilvout, and J. De Ley. 1986. Alcaligenes piechaudii, a new species from human clinical specimens and the environment. Int. J. Syst. Bacteriol. 36:282-287. 14. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 15. Lastovica, A. J., and R. E. Ambrosio. 1986. Plasmid profiles of clinical isolates of Campylobacter with no or weak catalase activity. S. Afr. Med. J. 70:337-338. 16. Lastovica, A. J., E. Le Roux, and J. L. Penner. 1989. "Campylobacter upsaliensis" isolated from blood cultures of pediatric patients. J. Clin. Microbiol. 27:657-659. 17. Lior, H. 1984. New extended biotyping scheme for Campylobacter jejuni, Campylobacter coli, and "Campylobacter laridis." J. Clin. Microbiol. 20:636-640. 18. Lior, H., D. L. Woodward, J. A. Edgar, L. J. Laroche, and P. Gill. 1982. Serotyping of Campylobacterjejuni by slide agglutination based on heat-labile antigenic factors. J. Clin. Microbiol. 15:761-768. 19. Mayer, L. W. 1988. Use of plasmid profiles in epidemiologic surveillance of disease outbreaks and tracing the transmission of antibiotic resistance. Clin. Microbiol. Rev. 1:228-243. 20. Mills, C. K., and R. L. Gherna. 1987. Hydrolysis of indoxyl acetate by Campylobacter species. J. Clin. Microbiol. 25: 1560-1561. 21. Morris, G. K., and C. M. Patton. 1985. Campylobacter, p. 302-308. In E. H. Lennette, A. Balows, W. J. Hausler, Jr., and H. J. Shadomy (ed.), Manual of clinical microbiology, 4th ed. American Society for Microbiology, Washington, D.C. 22. Owen, R. J., M. Costas, D. D. Morgan, S. L. W. On, L. R. Hill, A. D. Pearson, and D. R. Morgan. 1989. Strain variation in Campylobacter pylori detected by numerical analysis of onedimensional electrophoretic protein patterns. Antonie Van Leeuwenhoek J. Microbiol. 55:253-267. 23. Owen, R. J., M. Costas, and L. L. Sloss. 1988. Electrophoretic protein typing of Campylobacterjejuni subsp. "doylei" (nitratenegative Campylobacter-like organisms) from human faeces and gastric mucosa. Eur. J. Epidemiol. 4:277-283. 24. Owen, R. J., M. Costas, L. L. Sloss, and F. J. Bolton. 1988. Numerical analysis of electrophoretic protein patterns of Campylobacter laridis and allied thermophilic campylobacters from the natural environment. J. Apple. Bacteriol. 65:69-78. 25. Owen, R. J., and C. Dawson. 1986. DNA base composition and base sequence relatedness of atypical Campylobacter jejuni strains from clinical material. FEMS Microbiol. Lett. 35:283287. 26. Owen, R. J., D. D. Morgan, M. Costas, and A. Lastovica. 1989. Identification of "Campylobacter upsaliensis" and other catalase-negative campylobacters from paediatric blood cultures by numerical analysis of electrophoretic protein patterns. FEMS Microbiol. Lett. 58:145-150.

1046

J. CLIN. MICROBIOL.

GOOSSENS ET AL.

27. Patton, C. M., N. Shaffer, P. Edmonds, T. J. Barrett, M. A. Lambert, C. Baker, D. M. Perlman, and D. J. Brenner. 1988. Human disease associated with "Campylobacter upsaliensis" (catalase-negative or weakly positive Campylobacter species) in the United States. J. Clin. Microbiol. 27:66-73. 28. Portnoy, D. A., S. L. Moseley, and S. Falkow. 1981. Characterization of plasmids and plasmid-associated determinants of Yersinia enterocolitica pathogenesis. Infect. Immun. 31:775782. 29. Pot, B., M. Gillis, B. Hoste, A. Van de Velde, F. Bekaert, K. Kersters, and J. De Ley. 1989. Intra- and intergeneric relationships of the genus Oceanospirillum. Int. J. Syst. Bacteriol. 39:23-34. 30. Roop, R. M., Il, R. M. Smibert, J. L. Johnson, and N. R. Krieg. 1985. DNA homology studies of catalase-negative campylobacters and "Campylobacter fecalis," an emended description of Campylobacter sputorum and proposal of the neotype strain of Campylobacter sputorum. Can. J. Microbiol. 31:823-831. 31. Sandstedt, K., J. Ursing, and M. Walder. 1983. Thermotolerant Campylobacter with no or weak catalase activity isolated from dogs. Curr. Microbiol. 8:209-213. 32. Steele, T. W., and S. N. McDermott. 1984. The use of membrane filters applied directly to the surface of agar plates for the isolation of Campylobacter jejuni from feces. Pathology 16: 263-265. 33. Steele, T. W., and R. J. Owen. 1988. Campylobacter jejuni subsp. doylei subsp. nov., a subspecies of nitrate-negative

34.

35.

36.

37.

38.

39.

Campylobacter isolated from human clinical specimens. Int. J. Syst. Bacteriol. 38:316-318. Steele, T. W., N. Sangster, and J. A. Lanser. 1986. DNA relatedness and biochemical features of Campylobacter spp. isolated in Central and South Australia. J. Clin. Microbiol. 22:71-74. Tanner, A. C. R. 1986. Characterization of Wolinella spp., Campylobacter concisus, Bacteroides gracilis, and Eikenella corrodens by polyacrylamide gel electrophoresis. J. Clin. Microbiol. 24:562-565. Taylor, D. E., K. Hiratsuka, and L. Mueller. 1989. Isolation and characterization of catalase-negative and catalase-weak strains of Campylobacter species, including "Campylobacter upsaliensis" from humans with gastroenteritis. J. Clin. Microbiol. 27:2042-2045. Ursing, J., M. Walder, and K. Sandstedt. 1983. Base composition and sequence homology of deoxyribonucleic acid of thermotolerant Campylobacter from human and animal sources. Curr. Microbiol. 8:307-310. Vandamme, P., E. Falsen, B. Pot, B. Hoste, K. Kersters, and J. De Ley. 1989. Identification of EF group 22 campylobacters from gastroenteritis cases as Campylobacter concisus. J. Clin. Microbiol. 27:1775-1781. Walmsley, S. L., and M. A. Karmali. 1989. Direct isolation of atypical thermophilic Campylobacter species from human feces on selective agar medium. J. Clin. Microbiol. 27:668-670.