Typing of Human Campylobacter jejuni Isolates in Finland by Pulsed-Field Gel Electrophoresis

Abstract

A total of 69 pulsed-field gel electrophoresis (PFGE) types were identified among 176 Campylobacter jejuni isolates from Finnish patients. In two geographic areas studied, five predominant PFGE types comprised over 40% of the isolates. One-third of the isolates had unique PFGE types. In small outbreaks, identical PFGE patterns were demonstrated, indicating a common source of infection.

Campylobacter jejuni and Campylobacter coli are the most common bacterial enteropathogens in developed countries. Contaminated drinking water, unpasteurized milk, and poultry consumption have been shown to be risk factors both in epidemics and in sporadic cases (5, 7, 16). Serotyping has been the most common typing method (8, 12), although biotyping and phage typing have also been employed (11) to show the epidemiologic association of strains isolated from patients or to trace the possible routes of transmission from animals to humans. Molecular methods, such as pulsed-field gel electrophoresis (PFGE) (4, 15), are more distinguishing than serotyping because several genotypes are found within a serotype. PFGE has been shown to be a highly discriminatory method if a combination of two restriction enzymes, SmaI and SacII, is used (2, 4, 15). In the present study, PFGE typing was applied for a longitudinal survey of human C. jejuni infections in two different geographic areas, and the PFGE patterns of human isolates were compared to those of C. jejuni isolated at the same time from Finnish chickens.

Bacterial isolates.

Two geographic areas were chosen, one to represent urban living (Helsinki, area 1) and the other to represent a more rural lifestyle (Satakunta, area 2). C. jejuni isolates from fecal samples of enteritis patients with no foreign travel within 2 weeks prior to the illness were collected during a 1-year period (area 1) or during 14 months (area 2). The samples were cultured on Campylobacter blood-free selective medium (charcoal cefoperazone deoxycholate agar; Oxoid Ltd., Basingstoke, Hampshire, England). Chicken fecal samples (48 isolates) and meat samples (25 isolates, representing 10% of all meat samples tested) from area 1 were cultured on Campylobacter charcoal differential agar medium directly (fecal samples) or after enrichment (meat samples). The chickens originated from three major Finnish chicken producers. All isolates were gram negative and oxidase, catalase, and hippurate positive, and they grew at 37 and 43°C in an atmosphere of 5% O₂–10% CO₂–85% N₂. After the original isolation, they were stored at −70°C. All isolates from area 1 were tested for their susceptibility to ciprofloxacin (5-μg disk; Oxoid) and erythromycin (15 μg; Oxoid). The susceptibility of chicken isolates to enrofloxacin (5-μg disk [6]) was tested.

Typing of isolates by PFGE.

For PFGE analysis, the isolates were grown on brucella blood agar for 2 days at 37°C in a microaerobic atmosphere. The bacterial cells were harvested and treated with formaldehyde to inactivate endogenous nucleases (3). Otherwise, DNA was prepared by the method of Maslow et al. (9). The bacteria were embedded in 1% low-melting-point agarose plugs (SeaPlaque GTG; FMC Bioproducts, Rockland, Maine). After DNA purification, 2-mm slices of the agar plugs were digested with SmaI or SacII restriction enzyme (New England Biolabs, Hertfordshire, United Kingdom) as described by the manufacturer. The DNA fragments were separated with Gene Navigator (Pharmacia LKB Biotechnology AB, Uppsala, Sweden) in a 1% agarose gel in 0.5× TBE buffer (45 mmol of Tris, 45 mmol of boric acid, 1 mmol of EDTA) at 200 V. SmaI fragments were separated with a ramped pulse from 0.5 to 25 s for 20 h, and SacII fragments were separated with a ramped pulse from 0.3 to 18 s for 20 h. For the interpretation of PFGE results, the guidelines suggested by Tenover et al. were used (17). SmaI patterns were designated by roman numerals, and if isolates with a certain pattern could be further subdivided by SacII, the additional patterns were designated by capital letters (such as I/B).

With the two restriction enzymes, 48 combined PFGE types were found among the area 1 isolates (107) and 30 such types were found among the area 2 isolates (69). Nine PFGE types were identical in both areas. Similar results were reported in a study showing that, in spite of a high degree of diversity in SmaI-digested DNA profiles of C. jejuni, some identical PFGE patterns were demonstrated in three geographically separated sampling locations (10). Although most cases seemed to be sporadic in our study, five predominant PFGE types covered 42 and 44% of the isolates in area 1 and area 2, respectively (Table 1). The most common SmaI pattern, I (Fig. 1, lanes 1, 2, and 11), was further subdivided by SacII into nine different patterns. A total of 68% of the SmaI pattern I isolates were of SacII type B (I/B [Fig. 2]). SacII pattern E (Fig. 2, lanes 10 and 11) was the most common pattern among chicken isolates. Similarly, SmaI pattern XI (Fig. 1, lane 5) was subdivided into five patterns by SacII, of which T, Q, and U (Fig. 3, lanes 1 to 6) were seen only among chicken isolates, whereas patterns R and S (Fig. 3, lanes 7 to 13) were associated with human infections. SmaI patterns I and IX formed two lineages with genetically related isolates. PFGE type VII was identified only among human isolates. The interpretation of SmaI patterns I and II was complicated, as only five fragments were produced and small changes in the molecular sizes of two fragments of about 300 to 350 kb (Fig. 1, lanes 1, 2, 6, 8, and 11) were seen. SacII pattern analysis increased the information on the relatedness of these isolates. For example, in Fig. 1, isolates in lanes 6 and 8 differ in one of five fragments produced by SmaI but in only 1 of 11 fragments produced by SacII (Fig. 2, lanes 1 to 4).

TABLE 1.

Predominant PFGE types among 176 human C. jejuni isolates and their presence in urban area 1, rural area 2, and chickens studied

SmaI/SacII patterna	Total no. of isolates	Presence in:
		Urban area 1	Rural area 2	Chickens
I/B	22	+b	+c	+
IV	21	+b	+c	+
V	10	+b	−	+
VI	5	+b	+	+
VII	13	+b	+c	−
XIV	5	+	+c	−
XV	8	+	+c	−

^aIf further subdivided by SacII restriction enzyme. 

^bOne of the five predominant PFGE types which comprised 42% of all isolates in the urban area. 

^cOne of the five predominant PFGE types which comprised 44% of all isolates in the rural area. 

FIG. 1.

Examples of PFGE patterns of SmaI digests of C. jejuni. Lanes 1, 2, and 11, pattern I; lane 3, pattern V; lane 4, pattern VII; lane 5, pattern XI; lanes 6 and 8, pattern II; lanes 7 and 10, pattern VI; lane 9, pattern IV; lane 12, pattern VIII; lane 13, pattern XX. MW, lambda concatamer molecular size markers (48.5 kb). Numbers on the left indicate molecular sizes of DNA marker bands.

FIG. 2.

Examples of PFGE patterns of SmaI type I strains after digestion with SacII. Included also are SacII patterns of related SmaI pattern II (two strains [lanes 1 and 4]) and SmaI pattern V (one strain [lane 12]). Lanes 2, 3, 5, 6, and 7 represent pattern I/B (SmaI/SacII), and lanes 8, 9, 10, and 11 represent patterns I/C, I/D, I/E, and I/E, respectively. MW, lambda concatamer molecular size markers (48.5 kb). Arrows at the left indicate molecular sizes 48.5, 97, 145, 194, 291, and 436.5 kb (bottom to top, respectively).

FIG. 3.

PFGE SacII patterns of SmaI type XI isolates from chickens (lanes 1 to 6) and from humans (lanes 7 to 13). Lanes 1 to 3, pattern XI/T (chickens from retail shops); lanes 4 and 5, pattern XI/Q; lane 6, pattern XI/U (chicken fecal samples from a production plant); lanes 7, 8, 9, and 13, pattern XI/R (human isolates, infection acquired during a cruise in area 1, and a human isolate from area 2); lanes 10 to 12, pattern XI/S (human isolates, infection acquired after eating at a certain restaurant). Numbers on the left indicate molecular sizes of DNA marker bands (lanes MW).

According to the isolation season, a peak was seen in July-August in both sampling areas, although this was more pronounced in area 1. Certain types like I/B, V, and VII predominated in June to August, and type IV was common in area 2 during all seasons, whereas type V was identified only in area 1 (Table 1). During late autumn, winter, and early spring, PFGE types were unique.

Handling and eating poultry have been shown to be clear risk factors in sporadic campylobacteriosis (7). In the present study, chicken was suspected to be the source of infection in 14 cases with 11 different PFGE types. Eight of these 11 PFGE types, including the predominant pattern, I/B, were identical with those seen in poultry. This further confirmed the actual role of poultry products in the infection of these particular patients. Four of the 10 most common human PFGE types were not found among chicken isolates.

Several small outbreaks were noted. In one outbreak, three patients who had been on the same cruise had identical PFGE types XI/R (Fig. 3, lanes 7 to 9). In addition, three patients with identical and unique PFGE types XI/S (Fig. 3, lanes 10 to 12) had eaten at the same restaurant. Furthermore, at least five small family outbreaks with identical PFGE patterns were identified.

Fluoroquinolone resistance has been reported to be an increasing problem in many countries (1, 14). In 1990, 9% of campylobacters isolated from Finnish patients were resistant to ciprofloxacin (13), and during the study period, the overall ciprofloxacin resistance in campylobacters was 32% at the Department of Bacteriology and Immunology. In the present study, all human isolates tested were susceptible to erythromycin, only a few (3 of 107) isolates were resistant to ciprofloxacin, and all chicken isolates were susceptible to enrofloxacin. This could be because quinolones are not used in Finnish poultry production.

In conclusion, the application of PFGE typing for a longitudinal survey of Campylobacter infections in two restricted geographic areas provided new information on the epidemiology of apparently sporadic infections. Although PFGE analysis distinguished 69 types among 176 isolates, five predominant types covered over 40% of the isolates in both areas studied. In particular, type I/B and other types related to it covered 17% of the isolates studied. This suggests the existence of certain common infection sources, although one-third of the patients were infected with unique genotypes. In addition, geographic differences were obvious, as only nine PFGE types were identical in the two areas studied. PFGE typing not only was useful in discriminating different Campylobacter isolates but also could suggest epidemiologic associations in several small outbreaks.