Characterization of Erythromycin-Resistant Porcine Isolates of Campylobacter coli

Abstract

Erythromycin-resistant Campylobacter organisms were isolated from swine, and their resistance to the antibiotic was characterized. One hundred fourteen Campylobacter organisms were isolated from 572 swine intestinal samples. All isolates were identified as Campylobacter coli by sequence analysis of 16S rRNA gene and polymerase chain reactions with primers specific to hippuricase gene in Campylobacter jejuni and aspartokinase gene in C. coli. Minimal inhibitory concentrations (MICs) of erythromycin were determined by using the agar dilution method, and 80 isolates were found to be resistant to erythromycin (MIC ≥4 μg/ml). Of these, 31 isolates had low-level resistance (MIC =4–16 μg/ml), and 49 isolates had high-level resistance (HLR, MIC ≥32 μg/ml). The HLR isolates carried a point mutation at position A2075 → G in domain V of the 23S rRNA gene, whereas the low-level resistance isolates carried no mutation. These 49 HLR isolates were characterized by pulsed-field gel electrophoresis and multilocus sequence typing to study their genetic diversity. Pulsed-field gel electrophoresis identified 16 distinct types with 50% genetic similarity as the cutoff. On the other hand, 28 different sequence types (STs), including 10 new STs, were identified with multilocus sequence typing. Forty-six of 49 erythromycin HLR isolates showed crossresistance to 6 macrolide derivatives. The correlation between the inhibitory activity of carbonyl cyanide m-chlorophenylhydrazone and the existence of cmeB, which is responsible for efflux in HLR isolates, was found to be low. Erythromycin resistance was transferred from 38 of the 43 HLR isolates to susceptible C. coli by natural transformation, with a frequency of 1.217 × 10⁻⁸–4.618 × 10⁻⁵ per recipient cell. All transformants were erythromycin resistant and had A2075 → G mutation in at least one of three copies of the 23S rRNA gene. Results indicate that variable genotypes of HLR C. coli coexist in swine and high-level erythromycin resistance can be transferred to other strains.

Introduction

C ampylobacter jejuni and Campylobacter coli are the most common Campylobacter species associated with diarrhea in humans; these two species are clinically indistinguishable.⁴ Animal food products are considered to be the most important sources of Campylobacter-induced infections in humans.⁴⁰ Campylobacter infections tend to be self-limiting and do not normally require treatment; however, therapy may be proposed for patients with severe, prolonged, or relapsing illness.

Macrolides, the drug of choice for treating Campylobacter infections, have been widely used in the veterinary field for prophylactic and therapeutic purposes.² Massive consumption of these drugs has been suspected to render the emergence of macrolide-resistant isolates, resulting in problems in the treatment of resistant Campylobacter infections in animals.¹⁵ Antimicrobial agents in the macrolide class that are used for animals in Korea are erythromycin, tylosin, lincomycin, and josamycin.²⁵ Since resistant isolates can be potentially transmitted to humans, the presence of macrolide-resistant isolates in the food chain has raised concerns that treatment of human infections will be compromised.⁴ Macrolide resistance is more prevalent in Campylobacter isolates of animal origin than in those of human origin, especially C. coli from swine and C. jejuni and C. coli from poultry.^{1,8,15,38,43,44} Macrolide antibiotics bind to 23S rRNA in the 50S subunit of bacterial ribosome and produce an antibacterial effect by inhibiting protein synthesis.⁶ Two mechanisms, target modification and efflux, are involved in macrolide resistance in Campylobacter, but there is no evidence that Campylobacter produces macrolide-modifying enzymes.³³ Campylobacter has developed resistance to erythromycin by target mutations; for example, mutations in 23S rDNA.^21,28,32,41 Recent studies have shown that macrolide resistance in C. jejuni and C. coli isolates may also involve an efflux by the CmeABC pump, which contributes to the multidrug-resistant phenotype of clinical bacteria.¹⁵ This research was performed with focus on the 49 isolates that were classified into the high-level resistance (HLR) group; isolates were divided into groups according to three types of erythromycin susceptibility to determine the characteristics of the mechanism of resistance, molecular similarity, and the transfer of resistance from the isolates.

Materials and Methods

Samples

During a 4-month period from July to October 2004, 572 porcine intestinal samples were immediately obtained after slaughter at an abattoir where animals were transported from 24 pig farms located in 10 different regions in the Kyung-gi and Chungchung provinces, Korea. All samples were kept in an insulated container and immediately transported to the laboratory that is within 1 hr of travel time. Samples were washed twice in a sterile phosphate-buffered saline (pH 7.4) to remove the intestinal contents. The washed intestinal membranes were then smeared and streaked onto Brucella agar (BBL; Becton Dickinson) supplemented with 10% horse serum, amphotericin B (2.5 μg/ml, Fungizone; Sigma), and Skirrow's supplement (polymyxin B, 2.5 IU/ml; vancomycin, 10 μg/ml; trimethoprim, 5 μg/ml; Sigma).³⁹ After 48 hr incubation at 42°C under microaerobic conditions, one presumptive Campylobacter colony per sample was subcultured on Brucella agar supplemented with 10% horse serum, amphotericin B, and Skirrow's supplement in a microaerobic atmosphere for 48 hr at 37°C. One hundred fourteen isolates were determined as presumptive Campylobacter based on typical characteristics of this genus; for example, an S-shaped rod form under microscopy, Gram-negative stain, and catalase- and oxidase-positive tests. Isolates were stored at −70°C for further study.

Identification by polymerase chain reaction with species-specific primers and sequencing of 16S rRNA gene

All genomic DNA templates for polymerase chain reaction (PCR) amplification were prepared by the cetyltrimethylammonium bromide method.¹⁹ Oligonucleotide primers were used to amplify a 1,004-bp fragment within the coding region of the 16S rRNA gene for the genera Campylobacter, Arcobacter, and Helicobacter.³¹ For sequencing reactions, PCR products were purified with a QIAquick Gel Extraction Kit (Qiagen). Sequence reactions were performed with an ABI 3100 automated sequencer (Applied Biosystems) distributed by Bionix (Seoul, Korea). DNA sequences were compared with a database by using the online BLAST algorithm at the National Center for Biotechnology Information Web server (www.ncbi.nlm.nih.gov).

Specific primers (hip-1, 5′-ATG ATG GCT TCT TCG GAT AG-3′; hip-2, 5′-GCT CCT ATG CTT ACA ACT GC-3′) designed by Hani and Chan¹⁷ were used to amplify the hippuricase gene in C. jejuni, whereas primers (CC18F, 5′-GGT ATG ATT TCT ACA AAG CGA G-3′; CC519R, 5′-ATA AAA GAC TAT CGT CGC GTG-3′) designed by Linton et al.²⁷ were used to amplify the aspartokinase gene in C. coli. These gene amplification procedures were performed to differentiate the two species. C. jejuni ATCC 33560 and C. coli ATCC 33559 were included as controls.

Minimal inhibitory concentration test

Minimal inhibitory concentrations (MICs) were measured on Mueller–Hinton agar (MHA) plates containing 5% sheep blood in accordance with the recommendations of the Clinical Laboratory Standards Institute (CLSI; formerly the National Committee for Clinical Laboratory Standards).⁹ Isolates were microaerobically grown for 48 hr on Brucella agar containing Skirrow's supplement or Trypticase soy agar containing 5% sheep at 42°C. MICs were measured by using erythromycin, ciprofloxacin (Korea Research Institute of Chemical Technology), enrofloxacin (Dr Ehrenstofer GmbH), ampicillin, chloramphenicol (Fluka), gentamicin, and tetracycline at concentrations ranging from 0.5 to 128 μg/ml. All chemicals used in this study were purchased from Sigma unless otherwise stated. C. jejuni ATCC 33560 and C. coli ATCC 33559 were included in each batch of the agar dilution tests, and CLSI-approved MIC quality control limits for these strains were used for the control of agar dilution performance. Wild-type cutoff values defined by EUCAST were employed as breakpoints for C. coli resistance to antimicrobials except enrofloxacin.¹⁴ MIC interpretive standards of veterinary pathogens was employed as breakpoints for Campylobacter resistance to enrofloxacin.⁹

Effect of efflux pump inhibitor on macrolide susceptibility

The effect of an efflux pump inhibitor (EPI) on various macrolides was determined with carbonyl cyanide m-chlorophenylhydrazone (CCCP; Sigma), which inhibits the RND family efflux pumps.¹⁰ A 10-μL aliquot containing 5 μg of a 14-membered ring macrolide (e.g., erythromycin and roxithromycin), 5 μg of a 15-membered ring macrolide (e.g., azithromycin), 5 μg of a 16-membered ring macrolide (e.g., spiramycin and josamycin), or a lincosamide (clindamycin) was loaded on a pair of AA-size filter disks (diameter, 6 mm; Whatman); then, 1 μg CCCP was added to one of the two disks. After the disks were air dried, they were placed on the surface of MHA supplemented with 5% sheep blood that was inoculated with a Campylobacter suspension (0.5 McFarland). Susceptibility to various macrolides was determined by the disk diffusion method in accordance with the recommendations of CLSI.⁹ Diameters of growth-inhibition zones were measured with a ruler after 48-h incubation at 37°C under microaerobic conditions. C. jejuni ATCC 33560 and C. coli ATCC 33559 were used to control disk diffusion test performance.

Detection of a mutation in 23S rRNA gene

A new set of primers (cam1, 5′-TGA TCG AAG CCC GAG TAA AC-3′; cam2, 5′-CCA GAC ATT GTC CCA CTT GA-3′) corresponding to nucleotides 1889 to 1908 and 2251 to 2233 of the 23S rRNA gene of C. jejuni NCTC11168 were designed. A 353-bp internal fragment of the 23S rRNA gene that contains position 2075 was amplified to detect macrolide-related mutation. PCR was carried out in a final volume of 50 μL containing 1 × buffer, 1.5 mM MgCl₂, 200 μM dNTPs, 1 μM of each primer, 1.25 U of Taq polymerase (Takarabio), and 2 μL of genomic DNA prepared as described above. The PCR thermal regimen consisted of 1 cycle at 94°C for 5 min; 30 cycles at 94°C for 30 sec, 53°C for 30 sec, and 72°C for 30 sec; and 1 cycle at 72°C for 7 min in a Perkin-Elmer GeneAmp 9700 thermocycler (Applied Biosystems). C. coli ATCC 33559 was used as a control. PCR products were purified with a QIAquick Gel Extraction Kit, and DNA sequences were analyzed as described above.

Detection of cmeB

Primers (cmeB1, 5′-TGT TTT TGT ACC AGT TTC TT-3′; cmeB2, 5′-TTT TCT TTG AGT GAA CTT GT-3′) corresponding to nucleotides 1362 to 1381 and 1862 to 1843 of the complete efflux pump gene in C. coli CIT 382 (GenBank No. AY598796) were designed to amplify 501 bp of cmeB.¹⁰ The amplification reaction was carried out in a final volume of 25 μL containing 1 × buffer, 2.5 mM MgCl₂, 200 μM deoxynucleoside triphosphates, 1 μM of each primer, 1 U of Taq polymerase, and 1 μL of genomic DNA prepared as described above. PCR was performed with 1 cycle at 94°C for 5 min; 30 cycles at 94°C for 30 sec, 51°C for 30 sec, and 72°C for 30 sec; and 1 cycle at 72°C for 7 min. C. coli ATCC 33559 was used as a negative control. PCR products were confirmed by electrophoresis.

Pulsed-field gel electrophoresis

Pulsed-field gel electrophoresis (PFGE) was performed according to the protocol of Pulsenet (www.Cdc.gov/pulsenet/protocols.htm) for Campylobacter. Briefly, fresh Campylobacter cells were suspended in Tris EDTA buffer (pH 8.0), and the absorbance at 600 nm was adjusted to 1. The cell suspension (100 μL) was mixed with an equal volume of molten 1% InCert agarose (FMC Products), 5 μl of 20 mg/ml proteinase K, and 10 μl of 10 mg/ml lysozyme. Genomic DNA in agarose-embedded plugs was digested with 40 U SmaI. PFGE was performed by using a contour clamped homogeneous electric field (CHEF-DRIII; Bio-Rad) under the following conditions: 0.5 × Tris borate EDTA, 1% SeaKem gold agarose (FMC Products), 14°C, 6 V/cm for 20 hr, with switch times ranging from 6.8 to 35.4 sec, with an including angle of 120°. A lambda ladder PFGE marker (New England Biolabs) was used as a reference marker. After the gel was stained with 2 μg/ml ethidium bromide and destained with distilled water, DNA fragments were visualized with GelDOC EQ (Bio-Rad). Analysis of PFGE was performed by Bionumerix software (Applied Maths) according to the manufacturer's recommendations.

Multilocus sequence typing

Multilocus sequence typing (MLST), PCR, and sequencing reactions were performed according to the methods of Dingle et al.¹³ Fragments of seven housekeeping genes (aspA for aspartase A, glnA for glutamine synthetase, gltA for citrate synthase, glyA for serine hydroxymethyl transferase, pgm for phosphoglucomutase, tkt for transketolase, and uncA for ATP synthase subunit) were amplified by PCR with primers used by others.¹² The amplified fragments were purified by DNA purification kit (Genall), and sequencing was performed by Bionics. Allele profiles and sequence types (STs) were designated using the public Campylobacter MLST profile database (www.pubmlst.org). Each sequence is assigned with an allele number, and the combination of alleles yields an ST. Related STs can be clustered into an appropriate clonal complex (CC). The concatenated MLST gene sequence data for each isolate were used to draw a neighbor-joining tree by using MEGA software (version 4.0.2, available at http://megasoftware.net).¹²

Transformation with naked DNA

Natural transformation was performed according to the previously published protocol²⁴ with some modifications. DNAs from 43 erythromycin HLR isolates were used as donors. Swine-derived, erythromycin-susceptible C. coli ATCC 33559 and CCARM 13264 were used as recipients. Negative controls were identically processed, except that no genomic DNA was added. Genomic DNAs from erythromycin-resistant donors were extracted by the cetyltrimethylammonium bromide method.¹⁹ One microliter of the erythromycin-susceptible recipient from 48-h culture on MHA grown microaerobically at 42°C was spotted onto MHA plates in triplicate; then, 2 μL genomic DNA (∼0.1 μg) from the donor was added to each spot. The plates were incubated overnight (18–19 hr) at 42°C under microaerobic conditions. An agar block for each spot was punched out with a sterile Pasteur pipette, collected in a microcentrifuge tube containing 100 μl of saline, and vigorously mixed. The whole supernatant was spread on an MHA plate (9-cm diameter) containing 10 μg/ml of erythromycin (EMHA). EMHA plates were incubated at 42°C under microaerobic conditions for 48 hr, and growth of Campylobacter in the presence of erythromycin was observed. Cell density of negative controls (without donor DNA) was determined by serial dilution and spreading onto MHA in triplicate. Colonies were counted after microaerobic incubation at 42°C for 48 hr. The transformation frequency was determined as the ratio of colony-forming units of transformants per milliliter to colony-forming units per milliliter of the negative control. Mutation in 23S rRNA in transformants was verified by PCR and sequence analysis as described above. Antimicrobial susceptibility of transformants to erythromycin, tetracycline, ciprofloxacin, and gentamicin was measured by the disk diffusion method according to CLSI recommendations.⁹

Nucleotide sequence accession numbers

The partial 23S rRNA gene sequences obtained from 49 HLR strains in this study were deposited in GenBank and assigned with accession numbers EU677797, EU677798, and GU434169–GU434215.

Results

Isolation and identification of Campylobacter

One hundred fourteen presumptive Campylobacter organisms were isolated from 572 porcine intestinal samples. Catalase- and oxidase-positive colonies with the characteristic translucent morphology of Campylobacter were isolated and examined under a microscope after Gram staining. All these isolates were identified as C. coli by sequence analysis of 16S rRNA gene and PCRs with primers specific to the hippuricase gene in C. jejuni and the aspartokinase gene in C. coli.

Erythromycin MIC of C. coli

Erythromycin resistance rates were 46.5%, and MIC₅₀ and MIC₉₀ were 4 and ≥128 μg/ml, respectively (Table 1). Isolates were classified into three groups according to the level of erythromycin resistance: group 1, susceptible group containing 34 isolates (29.8%); group 2, low-level resistance (LLR) group containing 31 isolates (27.2%) (MIC = 4–16 μg/ml); group 3, HLR group containing 49 isolates (43.0%) (MIC ≥64 μg/ml).

Table 1.

Minimal Inhibitory Concentrations to Erythromycin of 114 Campylobacter coli Isolates from Swine

MIC (μg/ml)	≤0.5	1	2	4	8	16	32	64	≥128	MIC₅₀	MIC₉₀
Number of isolates	8	10	16	27	3	1	0	1	48	4	≥128
Total	34			31			49
Resistance type	S			LLR			HLR
Number of isolates with A2075 → G mutation in 23S rDNA	0 (0%)			0 (0%)			46 (93.9%)

HLR, high-level resistance; LLR, low-level resistance; MIC, minimal inhibitory concentration; S, susceptible.

EPI effect on erythromycin resistance

In accordance with macrolide resistance and the influence of EPI-CCCP, 49 HLR isolates were grouped into three types (types I, II, and III) (Table 2). Type 1 consists of 27 isolates that were resistant to all 6 macrolides and showed no changes in susceptibility with the addition of CCCP. This result suggests that the efflux pump is not involved in macrolide resistance. Type II consists of 19 isolates that were also resistant to 6 macrolides; their inhibition zone diameter changed from 1 to 24 mm upon the addition of CCCP, indicating the involvement of a CCCP-susceptible efflux in the resistance. Type III consists of three isolates that were resistant to some of the macrolides and showed various effects from EPI.

Table 2.

Characterization of High-Level Macrolide-Resistant Campylobacter coli Isolates from Swine

								Natural transformation frequency/recipient cell
Isolate no.	Mutation in 23S rRNA	Resistant to	Resistance type	Influence of CCCP	cmeB	PFGE group	ST	Recipient 1	Recipient 2
13267	A2075 → G	AzCmEmJmRmSm	I	−	+	7	829	2.107 × 10⁻⁶	0
13269	A2075 → G	AzCmEmJmRmSm	I	−	+	16	2715	8.705 × 10⁻⁶	0
13178	A2075 → G	AzCmEmJmRmSm	I	−	−	4	1055	0	2.384 × 10⁻⁷
13183	A2075 → G	AzCmEmJmRmSm	I	−	+	4	829	0	2.068 × 10⁻⁷
13184	A2075 → G	AzCmEmJmRmSm	I	−	+	4	829	0	2.839 × 10⁻⁶
13194	A2075 → G	AzCmEmJmRmSm	I	−	−	1	854	0	2.993 × 10⁻⁷
13198	A2075 → G	AzCmEmJmRmSm	I	−	+	2	4173	0	2.628 × 10⁻⁷
13207	A2075 → G	AzCmEmJmRmSm	I	−	−	16	829	0	1.324 × 10⁻⁶
13213	A2075 → G	AzCmEmJmRmSm	I	−	−	2	4169	0	1.307 × 10⁻⁶
13227	A2075 → G, C2097 → T	AzCmEmJmRmSm	I	−	+	3	890	0	2.117 × 10⁻⁷
13239	A2075 → G	AzCmEmJmRmSm	I	−	−	6	1056	NT	NT
13240	A2075 → G	AzCmEmJmRmSm	I	−	−	6	2716	0	1.217 × 10⁻⁸
13242	A2075 → G	AzCmEmJmRmSm	I	−	−	8	2716	0	1.630 × 10⁻⁷
13243	A2075 → G	AzCmEmJmRmSm	I	−	+	6	829	0	2.370 × 10⁻⁶
13247	A2075 → G	AzCmEmJmRmSm	I	−	+	5	829	NT	NT
13249	A2075 → G	AzCmEmJmRmSm	I	−	+	15	854	0	1.131 × 10⁻⁶
13251	A2075 → G	AzCmEmJmRmSm	I	−	+	6	4171	4.618 × 10⁻⁵	0
13255	A2075 → G	AzCmEmJmRmSm	I	−	+	8	4176	3.627 × 10⁻⁶	0
13259	A2075 → G	AzCmEmJmRmSm	I	−	+	4	1203	6.805 × 10⁻⁶	0
13260	A2075 → G, C2097 → T	AzCmEmJmRmSm	I	−	−	4	2716	NT	NT
13263	A2075 → G	AzCmEmJmRmSm	I	−	−	4	4174	3.282 × 10⁻⁶	0
13266	A2075 → G	AzCmEmJmRmSm	I	−	+	7	829	NT	NT
13268	A2075 → G	AzCmEmJmRmSm	I	−	−	4	1142	1.062 × 10⁻⁵	0
13271	A2075 → G	AzCmEmJmRmSm	I	−	+	4	854	2.297 × 10⁻⁶	0
13245	A2075 → G	AzCmEmJmRmSm	I	−	−	14	1103	0	9.270 × 10⁻⁷
13195	—^a	AzCmEmJmRmSm	I	−	+	2	4168	0	1.825 × 10⁻⁶
13214	A2075 → G	AzCmEmJmRmSm	I	−	−	NT	2220	0	1.509 × 10⁻⁷
13204	A2075 → G	AzCmEmJmRmSm	II	+	+	9	829	0	6.131 × 10⁻⁷
13206	A2075 → G	AzCmEmJmRmSm	II	+	−	3	4172	0	0
13161	A2075 → G	AzCmEmJmRmSm	II	+	−	10	4170	0	1.504 × 10⁻⁶
13163	A2075 → G	AzCmEmJmRmSm	II	+	−	7	4166	0	4.136 × 10⁻⁶
13166	A2075 → G	AzCmEmJmRmSm	II	+	−	12	1143	0	7.105 × 10⁻⁷
13171	A2075 → G	AzCmEmJmRmSm	II	+	−	8	1143	0	2.336 × 10⁻⁷
13176	A2075 → G	AzCmEmJmRmSm	II	+	+	4	828	0	2.603 × 10⁻⁷
13186	A2075 → G	AzCmEmJmRmSm	II	+	−	1	867	0	1.046 × 10⁻⁷
13188	A2075 → G	AzCmEmJmRmSm	II	+	+	7	4167	0	6.496 × 10⁻⁷
13221	—^a	AzCmEmJmRmSm	II	+	−	11	1055	0	1.849 × 10⁻⁷
13203	A2075 → G	AzCmEmJmRmSm	II	+	−	2	828	0	1.971 × 10⁻⁷
13233	A2075 → G	AzCmEmJmRmSm	II	+	+	7	3299	0	5.912 × 10⁻⁷
13236	A2075 → G	AzCmEmJmRmSm	II	+	−	8	3299	0	1.217 × 10⁻⁷
13258	A2075 → G	AzCmEmJmRmSm	II	+	−	6	2716	2.739 × 10⁻⁵	0
13261	A2075 → G	AzCmEmJmRmSm	II	+	+	7	829	NT	NT
13262	A2075 → G	AzCmEmJmRmSm	II	+	+	4	1016	1.675 × 10⁻⁶	0
13252	A2075 → G	AzCmEmJmRmSm	II	+	−	6	825	4.542 × 10⁻⁶	0
13265	A2075 → G	AzCmEmJmRmSm	II	+	−	10	4172	NT	NT
13272	A2075 → G	AzCmEmJmRmSm	II	+	−	4	1173	0	0
13159	A2075 → G	AzEmRmSm	III	+	−	12	1143	0	0
13230	A2075 → G	CmEmJmRmSm	III	−	−	5	829	0	0
13256	—^a	RmSm	III	+	−	6	2716	0	0

Az, azithromycin; CCCP, carbonyl cyanide m-chlorophenylhydrazone; Cm, clindamycin; Em, erythromycin; Jm, josamycin; NT, not tested; PFGE, pulsed-field gel electrophoresis; Rm, roxithromycin; Sm, spiramycin. 0, below the detection limit (5.2 × 10⁻⁸ for recipient 1 and 7.2 × 10⁻⁹ for recipient 2), which was based on a minimum of 1 colony per plate, applied to the volumes used in the procedure.

No mutation found.

Sequence analysis of the erythromycin target site: 23S rRNA gene

A2075 → G mutation in domain V of the 23S rRNA gene is the best-known mechanism of macrolide resistance in Campylobacter.⁶ Forty-six of 49 HLR isolates exhibited A2075 → G mutation. Two of these isolates (CCARM 13227 and CCARM 13260) carried C2097 → T mutation. Mutation was not detected in three of the HLR isolates or in erythromycin-susceptible and LLR isolates.

Detection of cmeB

When PCR was performed to amplify a 501-bp internal fragment of the cmeB gene, 21 of the 49 HLR isolates (42.8%) produced DNA fragments. EPI influence on susceptibility and the presence of an efflux gene cmeB are as shown in Table 2. Amplified DNA fragments for cmeB were obtained from 15 strains (55%) of type I isolates (27 strains) and 6 (31%) of 19 type II isolates.

Pulsed-field gel electrophoresis

The genetic relationship among the HLR isolates based on their SmaI-digested DNA fragments is shown in Fig. 1. SmaI-digested DNA fragments of HLR isolates showed various patterns; a 50% genetic similarity cutoff resulted in 16 clusters (Table 2). Cluster 4 (11 isolates), cluster 6 (7 isolates), cluster 7 (6 isolates), and cluster 8 (4 isolates) are the predominant clusters, containing 57% of the HLR isolates. CCARM 13183, 13184, and 13176 were classified into different types (Table 2) according to EPI effect; however, these three types had a high genetic similarity (>90%) in PFGE, had the same antimicrobial-resistant pattern (resistant to ciprofloxacin, enrofloxacin, erythromycin, and tetracycline), and carry cmeB. In contrast, CCARM 13269 and 13178 were classified into the same EPI type, although they have different PFGE (clusters 4 and 13) patterns, antimicrobial-resistant patterns (resistant to CmCipEfEmTc and CipEfEmGmTc), and existence of cmeB.

FIG. 1.

PFGE of SmaI-digested DNA of porcine HLR Campylobacter coli isolates. PFGE was performed with a linear ramped pulse time of 6.8–35.4 sec with a run time of 20 hr in CHEF-DRIII, as described in Materials and Methods section. The dendrogram was produced with Dice and UPGMA analysis. Am, ampicillin; Cm, chloramphenicol; Cip, ciprofloxacin; Em, erythromycin; Gm, gentamicin; HLR, high-level resistant; MLST, multilocus sequence typing; PFGE, pulsed-field gel electrophoresis; Tc, tetracycline; UPGMA, unweighted pair-group method with averages.

Multilocus sequence typing

BLAST results using the MLST database indicated that the allele sequences for all seven loci had exact matches to sequences previously deposited in the database. Sequencing yielded 28 STs, 10 of which were novel; these STs were submitted to the MLST database (Table 3). Nine of the new STs resulted from new combinations of previously described alleles; for example, ST-4171 resulted from a new combination of previously described alleles with the new allele sequence in tkt gene. Certain STs appeared to be common in the majority of isolates: ST-829 (11 isolates, 22.4%) was the most common type, whereas ST-2716 (5 isolates, 10. 2%), ST-854 (3 isolates), and ST-1143 (3 isolates) were also commonly detected (Table 2). Most (9 of 10 STs) of the new STs were represented by a single isolate. Twenty-eight STs, including 10 novel ones, belonged to the ST-828 CC. Three isolates assigned to ST-1143 could not be grouped into a previously defined clonal group. The genetic distance of the 28 STs was depicted by creating a neighbor-joining tree (Fig. 2). Among the 28 STs, 26 were clustered together (the 2 divergent STs are ST-1143 and ST-4174).

FIG. 2.

Neighbor-joining tree created with the nucleotide sequences of 28 HLR C. coli to indicate the relationship within HLR C. coli isolates.

Table 3.

New Sequence Types Reported from This Study

		Allele
New ST	No. of isolates	aspA	glnA	gltA	glyA	pgm	tkt	uncA	CC
4166	1	33	153	44	82	113	43	17	ST-828 complex
4167	1	33	153	44	82	104	43	17	ST-828 complex
4168	1	33	39	30	82	104	225	211	ST-828 complex
4169	1	33	38	44	82	113	43	17	ST-828 complex
4170	1	33	38	30	81	104	394	17	ST-828 complex
4171	1	33	262	30	82	104	44	36	ST-828 complex
4172	2	33	38	30	81	104	43	17	ST-828 complex
4173	1	33	39	44	82	113	47	17	ST-828 complex
4174	1	32	39	30	82	118	43	36	ST-828 complex
4176	1	33	153	30	82	113	43	17	ST-828 complex

The new allele identified in this study is in a boldface.

CC, clonal complex; ST, sequence type.

Transformation with naked DNA

The 43 HLR isolates were tested for their ability to transfer erythromycin resistance. Transformation results showed that 38 isolates transferred erythromycin resistance to the recipient cells, with a frequency of 1.217 × 10⁻⁸–4.618 × 10⁻⁵ per recipient cell (Table 2). The resulting transformants exhibited resistance to erythromycin and tetracycline but not to ciprofloxacin and gentamicin. Fifteen of the 38 transformants had the same A2075 → G mutation observed in donor HLR isolates. Twenty-three transformants were found to have both A and G at 2075 by visual examination of the sequence chromatograms.

Discussion

C. coli is the most common Campylobacter species in swine.⁴⁰ A wide range of macrolide antibiotics are currently available for the treatment of infected swine, including erythromycin, tylosin, lincomycin, and josamycin.²⁵ For erythromycin, the resistance rate of C. coli from swine differs among countries and seems to be related to the amount of the antibiotic used in swine in each country.¹ For instance, in neighboring Japan, resistance rates were 61.1% and 48.4%,^18,20 which are higher than the resistance rate (46.5%) in this study. In Canada, Spain, France, and Italy,^{5,11,16,34,37} erythromycin resistance rate was around 30%, which is lower than the result of this study. Ethiopia²² and Switzerland⁴⁴ showed the lowest resistance rates; that is, <10%. One interesting report showed that in the United States, C. coli in swine that were grown without the use of antibiotics showed a very high resistance rate (39%).⁴⁶ These results suggest that erythromycin resistance is mainly due to continuous consumption of the antibiotic rather than consumption of a large amount in a short period.³

Mutations in the peptidyl-transferase region of domain V of the 23S rRNA have been reported as mainly associated with erythromycin resistance,^21,29,32,41 and erythromycin cannot bind to these modified targets. The most frequently reported mutations are A2075 → G and A2074 → C.^15,40,42 Among the HLR isolates, 46 isolates (93.9%) had A2075 → G, whereas other mutations, A2075 → C, A2075 → T or A2074 → C, A2074 → G, or A2074 → T reported by others,²⁹ were not observed in this study. CCARM 13227 and CCARM 13260 have an additional mutation, C2097 → T mutation, as already described by Vacher et al.,⁴² although this mutation has not been identified to confer erythromycin resistance. As observed by others,⁶ erythromycin HLR isolates show a crossresistance to clindamycin and other macrolide analogs, including erythromycin; the 14-, 15-, and 16-membered ring macrolides; and lincosamide.

We checked the other macrolide resistance mechanism for C. coli: the efflux pump in HLR isolates. CmeABC is a multidrug-resistant pump encoded by a three-gene operon (cmeA, cmeB, and cmeC) located on a Campylobacter chromosome.⁶ The CmeABC efflux pump is composed of CmeA (the periplasmic fusion protein), CmeB (the inner membrane drug transporter), and CmeC (outer membrane protein).²⁶ These genes show homology with the multidrug efflux pump in the RND superfamily in Gram-negative bacteria.²⁶ These three proteins form the membrane channel that is known to excrete toxic materials and antimicrobial agents out of the cell.⁶ cmeR is located upstream of the cmeA gene and has a high similarity with the transcriptional repressor TetR family, which is known to control the manifestation of cmeABC gene.^7,26 If an efflux pump is involved in erythromycin HLR isolates, EPI exerts an effect, resulting in a decrease in erythromycin resistance. Although phenylalanine-arginine β-naphthylamide (PAβN) has been widely used as a broad-spectrum EPI of Gram-negative bacteria,³⁶ it is known to mainly function on LLR isolates.²⁸ Thus, CCCP, instead of PAβN, was used as the EPI in this study.

Results showed that the relationship between the change in susceptibility on addition of CCCP and the existence of cmeB gene is low. This may be due to the sequence variation of cmeB, thus yielding false-negative results in PCR.⁶ Otherwise, an additional efflux mechanism other than CmeABC may be active in Campylobacter.⁶

Many studies have reported the weak clonal population structure and the diverse genome of Campylobacter.^{23,30,35,40,45} Molecular typing methods—PFGE, restriction fragment length polymorphism of the amplified flagellin gene, and MLST—have been used to investigate the genetic diversity of C. coli. Initially, C. coli swine isolates were characterized by antimicrobial susceptibility, which yielded 19 antimicrobial-resistant types for 6 classes of antibiotics. The various characteristics shown in this study suggested that the isolates are genetically very diverse. They represented multiple PFGE patterns despite using 50% genetic similarity as the cutoff. All HLR isolates were resistant to erythromycin with at least one additional antimicrobial resistance comprising nine different antibiograms. Each individual antibiogram was not related to any specific PFGE pattern, suggesting that the resistance was due to the selective pressure of different antimicrobial usage. MLST was performed to further confirm the genetic diversity of HLR group isolates. PFGE identified 16 distinct clusters compared with 28 different STs identified with MLST. No simple linear or statistical relationship was evident between PFGE and MLST. A single ST could comprise isolates with different PFGE types (e.g., PFGE clusters 4, 5, 6, 7, 9, and 16 in ST-829). Ten STs had not been previously described, and all except 1 ST were assigned to the ST-828 complex. This group may have evolved from the main ST-828 CC due to recombination or mutational events that induced population divergence. The major CC observed in our study was ST-828, which has been previously reported to be the predominant C. coli complex in isolates from multiple sources (e.g., pig, chicken, sheep, and human diarrheal stool from sporadic cases)^13,40 Although the resulting natural transformation frequency was relatively low compared with other reports,²³ this experiment shows that natural transformation plays a significant role in the transfer of erythromycin resistance in Campylobacter. Campylobacter that acquire antibiotic resistance via natural transformation will have competitive advantages in the dynamic environment of swine intestine.

We conclude the following: (1) genetic diversity among the erythromycin HLR swine isolates is very high within the ST-828 CC category, (2) the most important mechanism of macrolide resistance is a point mutation, A2075 → G, in the 23S rRNA gene, (3) these isolates have crossresistance to other macrolide analogues, and (4) cmeABC is not the only gene responsible for the macrolide efflux mechanism. Further studies to characterize C. coli and C. jejuni from various sources, including swine, chicken, and human, are under way.

Footnotes

Acknowledgment

The authors would like to thank Dr. Won Jun Whang for the tissue samples. This study was supported by a special grant from Seoul Women's University (2010).

Disclosure Statement

No competing financial interests exist.

References

Aarestrup

F.M.

, Nielsen

E.M.

, Madsen

, Engberg

1997. Antimicrobial susceptibility patterns of thermophilic Campylobacter spp. from humans, pigs, cattle, and broilers in Denmark. Antimicrob. Agents Chemother., 41:2244–2250.

Blaser

M.J.

1995. Campylobacter and related species. Mandell

G.L.

, Bennet

J.E.

, Dolin

Principles and Practice of Infectious Diseases. Churchill Livingstone: New York, 1948–1956.

Baquero

, Artigao

G.B.

, Cantón

, Rey

C.G.

2002. Antibiotic consumption and resistance selection in Streptococcus pneumoniae. J. Antimicrob. Chemother., 50,S2:27–37.

Bodhidatta

, Vithayasai

, Eimpokalarp

, Pitarangsi

, Serichantalergs

, Isenbarger

D.W.

2002. Bacterial enteric pathogens in children with acute dysentery in Thailand: increasing importance of quinolone-resistant Campylobacter. Southeast Asian J. Trop. Med. Public Health, 33:752–757.

Bywater

, Deluyker

, Deroover

, de Jong

, Marion

, McConville

, Rowan

, Shryock

, Shuster

, Thomas

, Vallé

, Walters

2004. A European survey of antimicrobial susceptibility among zoonotic and commensal bacteria isolated from food-producing animals. J. Antimicrob. Chemother., 54:744–754.

Cagliero

, Mouline

, Payot

, Cloeckaert

2005. Involvement of the CmeABC efflux pump in the macrolide resistance of Campylobacter coli. J. Antimicrob. Chemother., 56:948–950.

Cagliero

, Cloix

, Cloeckaert

, Payot

2006. High genetic variation in the multidrug transporter cmeB gene in Campylobacter jejuni and Campylobacter coli. J. Antimicrob. Chemother., 58:168–172.

Chuma

, Ikeda

, Maeda

, Niwa

, Okamodo

2001. Antimicrobial susceptibilities of Campylobacter strains isolated from broilers in the southern part of Japan from 1995 to 1999. J. Vet. Med. Sci., 63:1027–1029.

[CLSI] Clinical, Laboratory Standards Institute. 2005. Performance standards for antimicrobial susceptibility testing. Sixteenth informational supplement M100-S15. CLSI: Wayne, PA.

10.

Corcoran

, Quinn

, Cotter

, O'Halloran

, Fanning

2005. Characterization of a cmeABC operon in a quinolone-resistant Campylobacter coli isolate of Irish origin. Microb. Drug Resist., 11:303–308.

11.

Desmonts

M.H.

, Dufour-Gesbert

, Avrain

, Kempf

2004. Antimicrobial resistance in Campylobacter strains isolated from French broilers before and after antimicrobial growth promoter bans. J. Antimicrob. Chemother., 54:1025–1030.

12.

Dingle

K.E.

, Colles

F.M.

, Falush

, Maiden

M.C.J.

2005. Sequence typing and comparison of population biology of Campylobacter coli and Campylobacter jejuni. J. Clin. Microbiol., 43:340–347.

13.

Dingle

K.E.

, Colles

F.M.

, Wareing

D.R.A.

, Ure

, Fox

A.J.

, Bolton

F.E.

, Bootsma

H.J.

, Willems

R.J.L.

, Urwin

, Maiden

M.C.J.

2001. Multilocus sequence typing system for Campylobacter jejuni. J. Clin. Microbiol., 39:14–23.

14.

[EUCAST] European Committee on Antimicrobial Susceptibility Testing. Antimicrobial wild type distributions of microorganisms, EUCAST version 5.12. www.eucast.org. 2010 May 24.

15.

Gibreel

, Kos

V.N.

, Keelan

, Trieber

C.A.

, Levesque

, Michaud

, Taylor

D.E.

2005. Macrolide resistance in Campylobacter jejuni and Campylobacter coli: molecular mechanism and stability of the resistance phenotype. Antimicrob. Agents Chemother., 49:2753–2759.

16.

Guévremont

, Nadeau

, Sirois

, Quessy

2006. Antimicrobial susceptibilities of thermophilic Campylobacter from humans, swine, and chicken broilers. J. Vet. Res., 70:81–86.

17.

Hani

E.K.

, Chan

V.L.

1995. Expression and characterization of Campylobacter jejuni benzoylglycine amidohydrolase (Hippuricase) gene in Escherichia coli. J. Bacteriol., 177:2396–2402.

18.

Harada

, ASAI

, Kojima

, Sameshima

, Takahashi

2006. Characterization of macrolide-resistant Campylobacter coli isolates from food-producing animals on farms across Japan during 2004. J. Vet. Med. Sci., 68:1109–1111.

19.

Honoré-Bouakline

, Vincensini

J.P.

, Giacuzzo

, Lagrange

P.H.

, Herrmann

J.L.

2003. Rapid diagnosis of extrapulmonary tuberculosis by PCR: impact of sample preparation and DNA extraction. J. Clin. Microbiol., 41:2323–2329.

20.

Ishihara

, Kira

, Ogikubo

, Morioka

, Kojima

, Kijima-Tanaka

, Takahashi

, Tamura

2004. Antimicrobial susceptibilities of Campylobacter isolated from food-producing animals on farms (1999–2001): results from the Japanese Veterinary Antimicrobial Resistance Monitoring Program. Int. J. Antimicrob. Agents, 24:261–277.

21.

Jensen

L.B.

, Aarestrup

F.M.

2001. Macrolide resistance in Campylobacter coli of animal origin in Denmark. Antimicrob. Agents Chemother., 45:371–372.

22.

Kassa

, Gebre-Selassie

, Asrat

2007. Antimicrobial susceptibility patterns of thermotolerant Campylobacter strains isolated from food animals in Ethiopia. Vet. Microbiol., 119:82–87.

23.

Kemp

, Leatherbarrow

A.J.

, Williams

N.J.

, Hart

C.A.

, Clough

H.E.

, Turner

, Wright

, French

N.P.

2005. Prevalence and genetic diversity of Campylobacter spp. in environmental water samples from a 100-square-kilometer predominantly dairy farming area. Appl. Environ. Microbiol., 71:1876–1882.

24.

Kim

J.S.

, Carver

D.K.

, Kathariou

2006. Natural transformation-mediated transfer of erythromycin resistance in Campylobacter coli strains from turkeys and swine. Appl. Environ. Microbiol., 72:1316–1321.

25.

[KAHPA] Korea Animal Health Products Association. 2001–2004. Monthly market trend of Korea animal health products. http://kahpa.or.kr. (Online.)

26.

Lin

, Akiba

, Sahin

, Zhang

2005. CmeR functions as a transcriptional repressor for the multidrug efflux pump CmeABC in Campylobacter jejuni. Antimicrob. Agents Chemother., 49:1067–1075.

27.

Linton

, Lawson

A.J.

, Owen

R.J.

, Stanley

1997. PCR detection, identification to species level, and fingerprinting of Campylobacter jejuni and Campylobacter coli direct from diarrheic samples. J. Clin. Microbiol., 35:2568–2572.

28.

Mamelli

, Amoros

J.P.

, Pages

J.M.

, Bolla

J.M.

2003. A phenylalanine-arginine beta-naphthylamide sensitive multidrug efflux pump involved in intrinsic and acquired resistance of Campylobacter to macrolides. Int. J. Antimicrob. Agents., 22:237–241.

29.

Mamelli

, Prouzet-Mauleon

, Pages

J.M.

, Megraud

, Bolla

J.M.

2005. Molecular basis of macrolide resistance in Campylobacter: role of efflux pumps and target mutations. J. Antimicrob. Chemother., 56:491–497.

30.

Mannering

S.A.

, West

D.M.

, Fenwick

S.G.

, Marchant

R.M.

, O'Connell

2006. Pulsed-field gel electrophoresis of Campylobacter jejuni sheep abortion isolates. Vet. Microbiol., 115:237–242.

31.

Marshall

, Melito

P.L.

, Woodward

D.L.

, Johnson

W.M.

, Rodgers

F.G.

, Mulvey

M.R.

1999. Rapid identification of Campylobacter, Acrobacter and Helicobacter isolates by PCR-restriction fragment length polymorphism analysis of 16S rRNA Gene. J. Clin. Microbiol., 37:4158–4160.

32.

Niwa

, Chuma

, Okamoto

, Itoh

2003. Simultaneous detection of mutations associated with resistance to macrolides and quinolones in Campylobacter jejuni and C. coli using a PCR-line probe assay. Int. J. Antimicrob. Agents, 22:374–379.

33.

Payot

, Bolla

J.M.

, Corcoran

, Fanning

, Megraud

, Zhang

2006. Mechanisms of fluoroquinolone and macrolide resistance in Campylobacter spp. Microbes Infect., 8:1967–1971.

34.

Pezzotti

, Serafin

, Luzzi

, Mioni

, Milan

, Perin

2003. Occurrence and resistance to antibiotics of Campylobacter jejuni and Campylobacter coli in animals and meat in northeastern Italy. Int. J. Food Microbiol., 82:281–287.

35.

Praakle-Amin

, Roasto

, Korkeala

, Hänninen

M.L.

2007. PFGE genotyping and antimicrobial susceptibility of Campylobacter in retail poultry meat in Estonia. Int. J. Food Microbiol., 114:105–112.

36.

Pumbwe

, Piddock

L.J.

2002. Identification and molecular characterization of CmeB, a Campylobacter jejuni multidrug efflux pump. FEMS Microbiol. Lett., 206:185–189.

37.

Saenz

, Zarazaga

, Lantero

, Gastanares

M.J.

, Baquero

, Torres

2000. Antibiotic resistance in Campylobacter strains isolated from animals, foods, and humans in Spain in 1997–1998. Antimicrob. Agents Chemother., 44:267–271.

38.

Sánchez

, Fernández-Baca

, Díaz

M.D.

, Muñoz

, Rodríguez-Créixems

, Bouza

1994. Evolution of susceptibilities of Campylobacter spp. to quinolones and macrolides. Antimicrob. Agents Chemother., 38:1879–1882.

39.

Skirrow

M.B.

1977. Campylobacter enteritis: a “new” disease. Br. Med. J., 2:9–11.

40.

Thaker

, Gebreyes

W.A.

2005. Campylobacter coli in swine production: antimicrobial resistance mechanisms and molecular epidemiology. J. Clin. Microbiol., 43:5705–5714.

41.

Vacher

, Menard

, Bernard

, Santos

, Megraud

2005. Detection of mutations associated with macrolide resistance in thermophilic Campylobacter spp. by real-time PCR. Microb. Drug Resist., 11:40–47.

42.

Vacher

, Ménard

, Bernard

, Mégraud

2003. PCR-restriction fragment length polymorphism analysis for detection of point mutations associated with macrolide resistance in Campylobacter spp. Antimicrob. Agents Chemother., 47:1125–1128.

43.

Van Looveren

, Doube

, De Zutter

, Dumont

J.M.

, Lammens

, Wijdoogle

, Vandamme

, Jouret

, Cornells

, Goossens

2001. Antimicrobial susceptibilities of Campylobacter strains isolated from food animals in Belgium. Antimicrob. Agents Chemother., 48:234–240.

44.

Weisblum

1995. Erythromycin resistance by ribosome modification. Antimicrob. Agents Chemother., 39:577–585.

45.

Wittwer

, Keller

, Wassenaar

T.M.

, Stephan

, Howald

, Regula

, Bissig-Choisat

2005. Genetic diversity and antibiotic resistance patterns in a Campylobacter population isolated from poultry farms in Switzerland. Appl. Environ. Microbiol., 71:2840–2847.

46.

Wondwossen

A.G.

, Thakur

, Morgan Morrow

W.E.

2006. Campylobacter coli: prevalence and antimicrobial resistance in antimicrobial-free (ABF) swine production systems. J. Antimirob. Chemother., 56:465–768.