The Presence of Antibiotic Resistance and Integrons in Escherichia coli Isolated from Compost

Abstract

The compost microcosm is a prime site for bacterial interaction that carries the inherent potential for disseminating antibiotic resistance through genetic exchange and subsequent land application. One hundred and thirty-six Escherichia coli isolates from compost heaps in South Carolina and California from a total of 277 compost samples were classified to phylogenetic groups, tested for resistance to 10 antibiotics, and screened for shiga toxins and integrons. Isolates that had identical antibiotic resistance patterns, grouped phylogenetically, and came from the same sample source were tested with pulsed-field gel electrophoresis. All isolates were negative for shiga toxins 1 and 2 as determined by polymerase chain reaction (PCR) assay. Resistance phenotypes comprised various combinations of seven antibiotics with a prevalence of ampicillin resistance in 63% of isolates (n = 56) from California, and tetracycline resistance in 37% of isolates (n = 62) from South Carolina. This disparity may be attributed to the differences in regional application of antibiotics as well as the origin of the waste materials in the compost itself. Phylogenetic PCR revealed that the majority of E. coli isolates (64%) belonged to groups A or B1. However, among these isolates, only 7% were resistant to two or more antibiotics as compared with 35% of isolates belonging to groups B2 and D. Integron detection by PCR revealed that nine (7.6%) E. coli isolates were positive for class I integrons, with six of the nine detected in isolates from groups B2 or D. Sequenced integrons (n = 5) were found to carry genes aadA, conferring reduced susceptibility to streptomycin, and dfrA, conferring resistance to trimethoprim. Our results suggest that E. coli isolates in compost belonging to phylogenetic groups B2 and D are more likely to contain integrons and higher levels of antibiotic resistance. The presence of multiple antibiotic resistances as well as integrons in E. coli implicates the potential for antibiotic resistance spread in the compost environment.

Introduction

T he widespread dissemination of antibiotic resistance among bacterial communities has become a topic of great concern, and is often attributed to the massive overuse as well as misuse of antimicrobials in agriculture (McEwen and Fedorka-Cray, 2002; Hoyle et al., 2006; Nunnery et al., 2006). The prophylactic use of antimicrobials for therapy and growth promotion in animal production systems selects for resistant enteric microorganisms that can then be introduced to the environment through animal feces (Smith et al., 2007). One mechanism for the spread of resistance is lateral gene transfer, which has been seen to occur in the gut of livestock, in manure, as well as in composted wastes (Blake et al., 2003; Guan et al., 2007). Zhao et al. (2001) demonstrated that antibiotic resistance genes carried on plasmids, transposons, and integrons are capable of being transferred from commensal to pathogenic bacterial strains. Integrons are genetic elements consisting of two conserved segments bordering a variable region. An integrase gene and recombination site allow for acquisition of linear pieces of DNA, gene cassettes, often coding for antibiotic resistance (Cocchi et al., 2007).

Escherichia coli is an organism that is naturally found in the gut of most humans and animals. Although the majority of strains are harmless, some are capable of causing a wide variety of diseases of the intestinal and urinary tracts of humans, septicemia, and neonatal meningitis (Walk et al., 2007; Talan et al., 2008; Varaiya et al., 2008; Zou et al., 2008). They fall into four main phylogenetic groups (A, B1, B2, and D) with extraintestinal virulent strains most commonly associated with groups B2 and D (Clermont et al., 2000). Pathogens commonly associated with outbreaks of foodborne illness such as enterohemorrhagic E. coli have been shown to belong to phylogenetic group B1 or B2 (Bando et al., 2007; Tramuta et al., 2008). Several studies have investigated the relationship between antibiotic resistance and virulence and have come up with contrasting conclusions (Johnson et al., 1991; Pallecchi et al., 2007; Clermont et al., 2008; Piatti et al., 2008). Regardless of whether or not commensal or pathogenic strains are more commonly associated with antibiotic resistance, the potential for bacteria to transfer genes has been demonstrated (Blake et al., 2003). Therefore, the presence of either pathogenic or nonpathogenic strains of E. coli with antibiotic resistance in feces is potentially hazardous if dispersed into the environment.

The most common form of animal waste treatment is on-farm composting, which is a process designed to eliminate harmful microorganisms through elevated temperature, desiccation, and microbial competition. While effective under specific conditions, studies have shown that some commensal and pathogenic organisms alike may survive for prolonged periods in compost and soils after land application (Soares et al., 1995; Shepherd et al., 2007). Further, the potential exists that pathogen presence in soils may result in contamination of surfaces or internal structures of fresh produce (Solomon et al., 2002; Islam et al., 2004; Johannessen et al., 2005). Although some research has been performed on the survival of pathogens in compost, the population characteristics and antibiotic resistance of E. coli isolated from compost have not been studied. The objective of this study was to isolate E. coli from compost samples and characterize them for antibiotic resistance.

Materials and Methods

Sampling compost

Compost samples (n = 209) were collected aseptically from five farms in the upstate of South Carolina between October 2004 and February 2007. Among these compost samples, a total of 24 samples of vegetable waste/dairy manure compost from Clemson University's Calhoun Student Organic Farm as well as 71 samples of dairy manure compost from a field study in our lab were collected for analysis between July and November 2006. Thirty samples of poultry compost from a university research farm (farm M) composed of chicken litter and poultry carcasses were collected between October 2004 and August 2005. Forty-two samples of poultry compost from a private farm (Farm S) composed of poultry litter and pine shavings were collected between December 2006 and February 2007. Eighteen samples of chicken litter with pine fines and 24 samples of chicken litter, carcasses, and fresh wood chips from two private farms, Farm A and Farm G, respectively, were collected between November 2004 and August 2005. In addition, a total of 68 samples of California compost were collected from four sites during the fall of 2007: cattle farm (Facility 1), two mushroom facilities (Facilities 2 and 4), and a municipal garbage composting facility (Facility 3).

Isolation and characterization of E. coli isolates

Twenty-five grams of compost samples were mixed with 225 mL of universal preenrichment broth (Neogen, Lansing, MI), stomached for 60 s in a Seward stomacher (Stomacher^®400, West Sussex, United Kingdom), serially 10-fold diluted in saline, and plated on E. coli/coliform Petrifilm^® (3M, St. Paul, MN). After incubation at 35°C overnight, approximately three typical E. coli colonies (blue with gas bubble formation) per sample were picked from all positive plates and purified further.

The presumptive E. coli isolates were plated on levine eosin methylene blue agar (Difco, Detroit, MI) and verified by real-time polymerase chain reaction (RT-PCR) of the glutamate decarboxylase gene (gad) (Table 1). Gad detection has been commonly used as a highly specific method of E. coli identification and was performed as previously described by Chen et al. (2006). The SYBR green–based RT-PCR method was chosen using the Bio-Rad iCycler™ system (BioRad, Inc., Hercules, CA).

Table 1.

Polymerase Chain Reaction Primers and Control Strains

Gene	Primer name	Primer sequence (5′–3′)	Positive control strains	Reference
intI	INTI-F	CCTCCCGCACGATGATC	Salmonella Typhimurium CVM 786	Bass et al. (1999)
	INTI-R	TCCACGCATCGTCAGGC
CS	INTEGRON5-CS	GGCATCCAAGCAGCAAG	Salmonella Typhimurium CVM 786	Bass et al. (1999)
	INTEGRON3-CS	AAGCAGACTTGACCTGA
stx1	VT1A	GACTGCAAAGACGTATGTAGATTCG	Escherichia coli O157:H7 C7927	Lang et al. (1994)
	VT1B	ATCTATCCCTCTGACATCAACTGC
sxt2	VT2A	ATTAACCACACCCCACCG	E. coli O157:H7 C7927	Lang et al. (1994)
	VT2B	GTCATGGAAACCGTTGTCAC
gad	GADRT-1	GCGTTGCTGAAATATGGTTGCCGA	E. coli O157:H7 C7927	Chen et al. (2006)
	GADRT-2	CGTCACAGGCTTCAATCATGCGTT
chuA	ChuA 1	GACGAACCAACGGTCAGGAT	E. coli ATCC 25922	Clermont et al. (2000)
	ChuA 2	TGCCGCCAGTACCAAAGACA
yjaA	YjaA 1	TGAAGTGTCAGGAGACGCTG	E. coli ATCC 25922	Clermont et al. (2000)
	YjaA 2	ATGGAGAATGCGTTCCTCAAC
TSPE4.C2	TspE4C2 1	GAGTAATGTCGGGGCATTCA	E. coli ATCC 25922	Clermont et al. (2000)
	TspE4C2 2	CGCGCCAACAAAGTATTACG

Quality control strains used for PCRs are listed in Table 1. E. coli O157:H7 strain C7927 and Salmonella enterica Typhimurium CVM 786 were kindly provided by Dr. Mike Doyle at the UGA Center for Food Safety and Dr. Shaohua Zhao at the FDA Center for Veterinary Medicine, respectively. All E. coli isolates used for the study were kept at −80°C in tryptic soy broth containing 20% glycerol until they were used.

Antimicrobial susceptibility testing

All E. coli isolates were tested for susceptibility to a series of antimicrobial agents on Mueller Hinton agar (Difco, Sparks, MD) according to the procedures described by the Clinical and Laboratory Standards Institute (CLSI, 2006). E. coli isolates were grown overnight in tryptic soy broth to an OD₆₀₀ of between 0.4 and 0.5. Each culture was diluted 1:50 in 250 μL of saline in a 96-well micro plate (Costar, Corning, NY). A sterile replica plater (Sigma, St. Louis, MO) was used to inoculate plates supplemented with four different concentrations of each antibiotic. Each experiment was performed in triplicate. The antimicrobial agents in the tests included tetracycline (4–32 μg/mL), chloramphenicol (8–64 μg/mL), ampicillin (8–64 μg/mL), ceftriaxone (16–128 μg/mL), gentamicin (4–32 μg/mL), kanamycin (16–128 μg/mL), streptomycin (16–128 μg/mL), trimethoprim/sulfamethoxazole (1/19–8/152 μg/mL), nalidixic acid (8–64 μg/mL), and ciprofloxacin (0.5–4 μg/mL) (Sigma). The minimum inhibitory concentration (MIC) results were interpreted by use of the CLSI breakpoints for Enterobacteriaceae. The resistance breakpoints for the tested antibiotics were 16 μg/mL for tetracycline and gentamicin; 32 μg/mL for chloramphenicol, ampicillin, and nalidixic acid; 64 μg/mL for ceftriaxone, kanamycin, and streptomycin; 4 μg/mL for ciprofloxacin; and 4/76 μg/mL for trimethoprim/sulfamethoxazole. For streptomycin, no CLSI breakpoints are currently available, so the concentration of >32 μg/mL was used, which is recognized by many monitoring programs as a valid breakpoint (Guerra et al., 2003; Sunde and Norstrom 2005). The control strain used for antimicrobial susceptibility testing was E. coli 25922 (ATCC, Manassas, VA).

Detection of virulence genes, integrons, and phylogenetic groups by PCR

Bacterial DNA used for PCR was prepared by a boiling method. The primers used in this study are listed in Table 1. All primers were synthesized by Invitrogen Co. (Carlsbad, CA). A multiplex PCR for stx1 and stx2 was carried out as described previously using 0.4 μM of each primer (Lang et al., 1994). The PCR program contained a 5-min initial denaturation at 94°C, followed by 30 cycles each of 94°C for 1 min, 58.5°C for 1 min, and 72°C for 1.5 min. The RT detection of class I integrons, intI, was performed as previously described (Jiang et al., 2006). A triplex PCR was performed for phylogenetic analysis as described by Clermont et al. (2000). The specificity of amplification of all RT-PCR products was confirmed by melting-curve analysis, and the amplicons size was checked by gel electrophoresis in a 1.5% agarose gel.

Sequencing of integron gene cassettes

The PCR for amplification of the integron variable region was carried out under the same conditions as for integrase gene detection. The PCR products were run on a 1.5% agarose gel in 1× Tris borate EDTA (TBE) for 1 h at 70 V. Bands were removed and the product purified using an UltraClean™ Gel Spin DNA purification kit (MoBio, Carlsbad, CA). The DNA sequencing was performed using the Dye-terminator cycle sequencing kit v3.1 on a 3130 ABI sequencer (Perkin Elmer Applied Biosystems, Boston, MA) at the Clemson University Genomics Center, using the forward and reverse primers described above. DNA sequences were analyzed by searching the GenBank database of the National Center for Biotechnology Information via the Basic Local Alignment Search Tool.

Statistical analysis

The regional prevalence of antibiotic resistances between South Carolina and California was compared. In addition, phylogenetic groups A and B1 were compared against groups B2 and D for multiple antibiotic resistances (MARs) and integrons. All calculations were performed using the Fisher's exact test of the statistical analysis system (SAS 2001, Cary, NC). A difference was considered significant if the p-value was < 0.05.

Results

In this study, we analyzed a total of 277 compost samples. All sampling sites from farms in South Carolina were positive for the presence of E. coli in compost samples at different stages of the composting process (Table 2). Although the interior samples were not always positive for E. coli, the surface samples of all South Carolina compost heaps were putatively positive for E. coli. This differed from California composts that had about an equal number of positive samples for both the interior and surfaces of heaps (Table 3). California compost samples from finished cattle manure-based compost resulted in no E. coli detection, whereas the remaining sampling sites were positive for the indicator microorganism. Notably, most of the California samples that were positive were from composts that were >1 month of age. One hundred sixty-seven putative E. coli isolates were collected with 136 confirmed as E. coli by Gad PCR. Additionally, isolates from the same sample site with identical antibiotic resistance patterns and phylogenetic groupings were further screened by pulsed-field gel electrophoresis (PFGE), yielding 118 unique E. coli strains.

Table 2.

Compost Sampling Locations and Origin of Escherichia coli Isolates from Farms in South Carolina

Major waste constituent	Farm	Heap (n) ^a	Compost age	Heap location	No. of samples E. coli positive	No. of E. coli isolates picked (no of strains) ^b	MAR ^c	Integrons ^d
Poultry	A	Second (18)	2 months	Surface (n = 6)	2	1	–	–
				Interior (n = 12)	0	0	–	–
	G	First (24)	14 days	Surface (n = 6)	1	1	1	–
				Interior (n = 18)	1	0	–	–
	M	First (21)	5–6 months	Surface (n = 7)	2	7	–	–
				Interior (n = 14)	1	2	–	–
		Second (9)	7 months	Surface (n = 3)	1	1	–	1
				Interior (n = 6)	0	0	–	–
	S	First (21)	0–30 days	Surface (n = 7)	7	14 (10)	5	3
				Interior (n = 14)	2	10 (8)	3	2
		Second (21)	0–21 days	Surface (n = 7)	7	13 (8)	2	1
				Interior (n = 14)	3	5 (4)	0	0^e
Dairy	Field trial	First (71)	0–30 days	Surface (n = 31)	13	18 (16)	–	–
				Interior (n = 40)	4	2	–	–
	Organic farm	First (24)	21 days	Surface (n = 8)	5	3 (2)	–	–
				Interior (n = 16)	5	0	–	–

The number of samples taken at location.

Clones were identified using PFGE and eliminated to account for unique strain types.

Multiple antibiotic resistances defined as resistance to two or more antibiotics; numbers do not reflect clones.

Number of integrons found in unique strains of E. coli; clones are not included.

Integron-containing isolate from this heap was already counted; however, it had an identical PFGE pattern to isolates in Farm S heap first.

PFGE, pulsed-field gel electrophoresis.

Table 3.

Compost Sampling Locations and Origin of Escherichia coli Isolates from Farms in California

Major waste constituent	Farm	Heap (n) ^a	Compost age	Heap location	No. of samples E. coli positive	No. of E. coli isolates picked (no. of strains) ^b	MAR ^c	Integrons ^d
Finished cattle manure/wheat straw	Facility 1	First (12)	1–2 months	Surface (n = 6)	0	0	–	–
				Interior (n = 6)	0	0	–	–
		Second (4)	7–14 days	Surface (n = 2)	0	0	–	–
				Interior (n = 2)	0	0	–	–
Chicken litter/horse track	Facility 2	First (8)	7 days	Surface (n = 4)	4	2	1	1
				Interior (n = 4)	2	6	1	–
		Second (8)	19 days	Surface (n = 4)	4	0	–	–
				Interior (n = 4)	4	6 (5)	2	1
Green waste/food waste	Facility 3		0 day	FGW (n = 2)	2	4 (3)	–	–
		First (8)	2–3 days	FGar (n = 2)	2	4	–	–
				Surface (n = 4)	4	9	–	–
		Second (4)	5 months	Interior (n = 4)	4	7	–	–
				Surface (n = 4)	0	0	–	–
		Third (4)	7–8 months	Surface (n = 4)	4	3	–	–
Chicken litter/horse track	Facility 4	First (8)	3 days	Surface (n = 4)	4	3	–	–
				Interior (n = 4)	4	6 (5)	–	–
		Second (8)	26 days	Surface (n = 4)	4	5	–	–
				Interior (n = 4)	3	4	2	–

The number of samples taken at location.

Clones were identified using PFGE and eliminated to account for unique strain types.

Multiple antibiotic resistances defined as resistance to two or more antibiotics; numbers do not reflect clones.

Number of integrons found in unique strains of E. coli; clones are not included.

FGW, fresh green waste; FGar, fresh garbage.

Screening by RT-PCR for shiga toxins 1 and 2 revealed that none of the E. coli isolates carried either virulence gene. Class 1 integrons were detected in seven isolates from chicken litter compost and two isolates from chicken litter/horse track compost. PFGE results revealed the different banding patterns among strains of integron containing E. coli isolates (data not shown).

The MICs for 10 antibiotics were examined for all 118 E. coli strains (Table 4). Isolates from dairy manure and poultry composts on South Carolina farms were found to be more resistant (p < 0.05) to tetracycline, 10% and 50%, respectively, as compared with E. coli isolates in compost from California, which had no resistance to this antibiotic. Conversely, E. coli isolates from California chicken litter/horse track and municipal garbage had higher levels (p < 0.05) of resistance to ampicillin, 63% and 62%, respectively, as compared with dairy manure and poultry compost in South Carolina, 5% and 0%, respectively. Multiple resistance phenotypes were observed in E. coli isolates from all composting facilities except for municipal garbage, which only had isolates that were resistant to ampicillin (Table 4). Overall, resistance phenotypes comprised various combinations of seven antibiotics with a prevalence of ampicillin resistance in 63% of isolates (n = 56) from California, and tetracycline resistance in 37% of isolates (n = 62) from South Carolina.

Table 4.

Antibiotic Resistance and Integron Phenotypes in Escherichia coli Isolates (n = 118) from Different Compost Types

			Integron sequencing
Source and E. coli strains (n)	Phylogenetic group (%) ^a	Resistance phenotypes (%)	Isolate name and group	Resistance genes	Resistance profile
South Carolina: vegetable waste/dairy manure (20)	A (5), B1 (65), B2 (10), D (10)	AMP (5), NAL (5), TET (10)	VW54 (−)	aadA1	GEN
South Carolina: poultry compost (42)	A (21), B1 (38), B2 (14), D (12)	STR (10), STR/TET (7), STR/TET/GEN (5), STR/NAL (2), TET/GEN (2), TET (26), SXT/TET (10), NAL (2), GEN (2)	Cm C (D) Cm N (D)	aadA1, dfrA1 aadA1	TET, SXT TET, GEN, STR
California: chicken litter/horse track (30)	A (27), B1 (40), B2 (3), D (23)	AMP (47), AMP/SXT (10), AMP/STR/SXT (3), AMP/KAN (3), SXT (7)	F₂H₁RS-2 (A) F₂H₂LI-1 (B2)	aadA1, dfrA15 aadA2, dfrA12	SXT AMP, SXT
California: municipal garbage (26)	A (19), B1 (42), B2 (8), D (19)	AMP (62)	–	–	–

Percentage of isolates that could not be assigned not included.

AMP, ampicillin; GEN, gentamicin; KAN, kanamycin; NAL, nalidixic acid; STR, streptomycin; SXT, sulfamethoxazole/trimethoprim; TET, tetracycline.

The use of PCR to determine phylogenetic groups of the E. coli isolates resulted in 90% positive classification (Table 4). The majority of E. coli isolates (47%) from South Carolina compost (n = 62) were identified as group B1. Similarly, 41% of the isolates from California compost (n = 56) were classified as group B1. Isolates belonging to phylogenetic groups B2 and D were 13% (n = 8) each for South Carolina and 5% (n = 3) and 21% (n = 12) for California compost, respectively. Groups B2 and D fostered isolates containing six of the nine integrons detected, whereas two were found in group A strains and one in an unclassified strain. Seven and four isolates belonging to phylogenetic groups B2 and D also carried resistances to two or more families of antibiotics, being considered as MAR strains, in South Carolina and California compost, respectively. Groups A and B1 isolates consisted of four and one MAR strains for South Carolina and California, respectively. Among the MAR isolates, two isolates from poultry compost (South Carolina) and one isolate from chicken litter/horse track (California) were resistant to three antibiotics, and all belonged to group D. Overall, group B2 and D isolates had more integrons (p < 0.05) and MARs (p < 0.05) as compared with groups A and B1.

Among nine integron-containing isolates, five were selected for sequencing based on antibiotic resistance phenotypes and integron variable region size. Sequencing analysis of five integron variable regions revealed resistance gene cassettes belonging to the aadA and dfrA1 families (Table 4). Although all integrons carried one aadA gene cassette, only strain Cm N carried resistance to streptomycin at >32 μg/mL. Another integron containing isolate was intermediate to streptomycin with an MIC of 32 μg/mL (data not shown).

Discussion

In this study, 136 E. coli strains were isolated from seven composting sites located in South Carolina and California. PFGE results confirmed that 118 of those strains were not from the same clones. Antibiotic resistance phenotypes revealed regional prevalence of tetracycline resistance in South Carolina isolates and ampicillin resistance in California isolates. Van den Bogaard and Stobberingh (2000) found that high levels of tetracycline resistance were prevalent in fecal E. coli isolates from turkeys. A review of antimicrobial use in food animals indicates that both tetracycline and chlortetracycline are common in the poultry industry (McEwen and Fedorka-Cray, 2002). Since two of the three composting sites in South Carolina utilized poultry litter, this may account for the high prevalence of tetracycline resistance. Recent studies on fresh feces and stored cow manures have also reported high levels of tetracycline resistance in E. coli isolates (Hoyle et al., 2006; Cocchi et al., 2007; Khachatryan et al., 2008). However, in our study, the isolates obtained from dairy manure and vegetable waste composts were highly susceptible to the majority of antibiotics tested, with 75% of isolates susceptible to all antibiotics tested. This could be due to the use of other antibiotics on the farm that were not tested in the current study. Another potential reason is that plasmids carrying resistance may be lost due to the reduced fitness of the organism in compost and the absence of selective pressures.

Ampicillin is commonly administered to horses for the treatment of bacterial infections; however, many microorganisms such as Enterobacter, Klebsiella, and E. coli often carry resistance to this antimicrobial (Wilson, 2001). Mushroom facilities in California used hay containing horse manure that may account for high levels of resistance to ampicillin. Hoyle et al. (2006) detected ampicillin resistance in both horse and cattle feces from farms in Scotland. Similarly, White et al. (2002) found ampicillin resistance in horse manure (33%), with lower levels of tetracycline resistance (22%). Although similarities exist between aforementioned studies, most other studies show isolates containing mid to high levels of resistance to both tetracycline and ampicillin, whereas our study shows a clear division in resistance from South Carolina to California. Over 60% of isolates from California were resistant to ampicillin with none resistant to tetracycline. Only one isolate from South Carolina was resistant to ampicillin, whereas 37% were resistant to tetracycline. This difference could be attributed to the variation in regional application of antibiotics as well as the origin of the waste materials in the compost itself, or the small sample size, which may reflect variations in antibiotic regiments used during different times of the year. Interestingly, 62% of the E. coli isolated from municipal garbage in California were also resistant to ampicillin, suggesting that ampicillin resistance may be widespread in that region of the country, not being strictly limited to agriculture. Since information about antibiotic practices is sparse due to prophylactic use with often minimal book-keeping, it is only possible to make assumptions regarding antibiotic use based on resistance patterns, which may not accurately reflect actual usage.

The presence of class I integrons in E. coli isolates collected from both states implicates the ubiquitous nature of antibiotic resistance in agricultural environments. All sequenced integrons carried either gene aadA1 or aadA2. The gene aadA confers resistance to streptomycin and spectinomycin (Sandvang, 1999). Sunde and Norstrom (2005) discovered that E. coli isolates harboring an aadA gene may still be classified as susceptible due to low levels of resistance conferred by that gene. The fact that only one isolate (Cm N) carrying the aadA gene cassette was resistant to high levels of streptomycin as well as gentamicin, another aminoglycoside, suggests that an additional resistance gene or other mechanism may have been present in that organism external to the integron. Isolate F₂H₂LI-1 was the only strain that carried the aadA2 gene cassette, which may account for the fact that it was intermediate to streptomycin with an MIC of 32 μg/mL. Three integrons carried genes dfrA1, dfrA12, and dfrA15, encoding dihydrofolate reductases, which are known to confer resistance to trimethoprim, and are commonly detected in class I integrons (Grape et al., 2005). Zhao et al. (2001) screened shiga-toxin-producing E. coli strains for the presence of integrons and found that several carried aadA and dfrA gene cassettes.

Although none of the isolates that we screened carried the virulence genes for shiga toxins 1 and 2, ∼26% belonged to phylogenetic groups B2 and D, which are associated with extraintestinal virulent strains of E. coli as well as enterohemorrhagic E. coli and enteropathogenic E. coli (Bando et al., 2007; Orsi et al., 2008). Thirty-five percent of our isolates belonging to phylogenetic groups B2 and D were resistant to two or more antibiotics, compared with 7% of strains in groups A and B1. This implicates not only resistance acquisition but also the potential for genetic exchange between closely related commensal E. coli and pathogenic strains. Interestingly, PFGE results identified several strains of E. coli that had identical banding patterns but different antibiotic resistance phenotypes, suggesting that interactions may have occurred within the compost itself. Further study will be carried out to determine the transferability of resistance.

Our results were in agreement with Boerlin et al. (2005), who found that antibiotic resistance was higher in potentially pathogenic strains of E. coli. Not only did isolates belonging to groups B2 and D exhibit higher levels of antibiotic resistance in that study, but also 85% of isolates contained integrons. Clermont et al. (2008) also discovered high levels of antibiotic resistance in B2 strains, alluding to the selective pressures of antibiotic use as a main reason. A study by Pallecchi et al. (2007) found that E. coli strains isolated from human fecal material contained nearly identical levels of resistance among the four different phylogenetic groups. A higher percentage of integrons, however, were detected in groups B2 (30%) and D (33%), as compared with groups A (14%) and B1 (25%). In contrast, Cocchi et al. (2007) observed that integrons were most prevalent in E. coli isolates belonging to the commensal group A. Similarly, a study analyzing E. coli isolated from lactating dairy cows found higher levels of antibiotic resistance in groups A and B1, which interestingly also carried shiga toxins (Houser et al., 2008).

The Clermont triplex PCR method for differentiating E. coli based on their phylogenetic group is simple, quick, and fairly accurate. Gordon et al. (2008) found that about 85%–90% of E. coli strains were typable using the triplex method, which is in agreement with our result of 90%. In this study, 12 out of 118 E. coli isolates could not be typed by this method. Further analysis using multilocus sequence typing or ribotyping would be helpful in determining the phylogenetic groups of strains that were unclassifiable with this triplex PCR method.

This study reports the presence of class 1 integrons in commensal E. coli, which to our knowledge is the first report on the presence of integrons in nonpathogenic E. coli isolated from compost. The fact that integrons containing the same antibiotic resistance gene cassettes were isolated from compost products from both South Carolina and California further verifies the ubiquitous nature of these transmissible elements. Antibiotic resistance phenotypes of isolates from both states appear to be strongly affected by regional practice of antibiotic use. It seems evident that E. coli belonging to phylogenetic groups B2 and D has a genetic predisposition to antibiotic resistance and integrons, in addition to being generally associated with virulence. This can be especially problematic in the composting environment, where genetic exchange can occur, as has been previously reported in manure and soils (Gotz and Smalla, 1997). If these organisms find their way into the food production chain through fresh produce, they may introduce transferrable antibiotic resistances to the gut microflora, perhaps allowing for the survival or selection of pathogenic strains of E. coli or other related species. The need for good agricultural practices, especially in the area of composting, is therefore essential to limiting the survival of enteric organisms and the continued dissemination of resistance genes into the environment.

Conclusions

The results of this study indicate that the presence of integrons and antibiotic resistance is widespread in commensal E. coli isolates from compost. Furthermore, higher levels of antibiotic resistance are found in the phylogenetic groups most typically associated with pathogenic strains of E. coli. Further study is needed to determine the frequency of genetic exchange among bacteria in the composting environment.

Footnotes

Acknowledgments

The authors thank farms in South Carolina and composting facilities in California for allowing us to collect compost samples, and Dr. James Rieck at Clemson University for help on statistical analysis. This study was supported partially by the grants from USDA-NIFSI and Fresh Express Produce Safety Initiative.

Disclosure Statement

No competing financial interests exist.

References

Bando

, Trabulsi

, Moreira-Filho

. Genetic relationship of diarrheagenic Escherichia coli pathotypes among the enteropathogenic Escherichia coli O serogroup. Mem Inst Oswaldo Cruz, 2007; 102:169–174.

Bass

, Liebert

, Lee

, Summers

, White

, Thayer

, Maurer

. Incidence and characterization of integrons, genetic elements mediating multiple-drug resistance, in avian Escherichia coli. Antimicrob Agents Chemother, 1999; 43:2925–2929.

Blake

, Hillman

, Fenlon

, Low

. Transfer of antibiotic resistance between commensal and pathogenic members of the Enterobacteriaceae under ileal conditions. J Appl Microbiol, 2003; 95:428–436.

Boerlin

, Travis

, Gyles

, Reid-Smith

, Heather Lim

, Nicholson

, McEwen

, Friendship

, Archambault

. Antimicrobial resistance and virulence genes of Escherichia coli isolates from swine in Ontario. Appl Environ Microbiol, 2005; 71:6753–6761.

Chen

, Higgins

, Maas

, Murthy

. DNA extraction and Escherichia coli quantification of anaerobically digested biosolids using the competitive touchdown PCR method. Water Res, 2006; 40:3037–3044.

Clermont

, Bonacorsi

, Bingen

. Rapid and simple determination of the Escherichia coli phylogenetic group. Appl Environ Microbiol, 2000; 66:4555–4558.

Clermont

, Lavollay

, Vimont

, Deschamps

, Forestier

, Branger

, Denamur

, Arlet

. The CTX-M-15-producing Escherichia coli diffusing clone belongs to a highly virulent B2 phylogenetic subgroup. J Antimicrob Chemother, 2008; 61:1024–1028.

[CLSI] Clinical and Laboratory Standards Institute. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically, 2006. Approved Standard, CLSI document M7–A7: Wayne, PA.

Cocchi

, Grasselli

, Gutacker

, Benagli

, Convert

, Piffaretti

. Distribution and characterization of integrons in Escherichia coli strains of animal and human origin. FEMS Immunol Med Microbiol, 2007; 50:126–132.

10.

Gordon

, Clermont

, Tolley

, Denamur

. Assigning Escherichia coli strains to phylogenetic groups: multi-locus sequence typing versus the PCR triplex method. Environ Microbiol, 2008; 10:2484–2496.

11.

Gotz

, Smalla

. Manure enhances plasmid mobilization and survival of Pseudomonas putida introduced into field soil. Appl Environ Microbiol, 1997; 63:1980–1986.

12.

Grape

, Farra

, Kronvall

, Sundström

. Integrons and gene cassettes in clinical isolates of co-trimoxazole-resistant gram-negative bacteria. Clin Microbiol Infect, 2005; 11:185–192.

13.

Guan

, Wasty

, Grenier

, Chan

. Influence of temperature on survival and conjugative transfer of multiple antibiotic-resistant plasmids in chicken manure and compost microcosms. Poultry Sci, 2007; 86:610–613.

14.

Guerra

, Junker

, Schroeter

, Malorny

, Lehmann

, Helmuth

. Phenotypic and genotypic characterization of antimicrobial resistance in German Escherichia coli isolates from cattle, swine and poultry. J Antimicrob Chemother, 2003; 52:489–492.

15.

Houser

, Donaldson

, Padte

, Sawant

, DebRoy

, Jayarao

. Assessment of phenotypic and genotypic diversity of Escherichia coli shed by healthy lactating dairy cattle. Foodborne Pathog Dis, 2008; 5:41–51.

16.

Hoyle

, Davison

, Knight

, Yates

, Dobay

, Gunn

, Amyes

SGB

, Woolhouse

MEJ

. Molecular characterization of bovine faecal Escherichia coli shows persistence of defined ampicillin resistant strains and the presence of class 1 integrons on an organic beef farm. Vet Microbiol, 2006; 115:250–257.

17.

Islam

, Morgan

, Doyle

, Jiang

. Fate of Escherichia coli O157:H7 in manure compost-amended soil and on carrots and onions grown in an environmentally controlled growth chamber. J Food Prot, 2004; 67:574–578.

18.

Jiang

, Yang

, Dettman

, Doyle

. Analysis of fecal microbial flora for antibiotic resistance in ceftiofur-treated calves. Foodborne Pathog Dis, 2006; 3:355–365.

19.

Johannessen

, Bengtsson

, Heier

, Bredholt

, Wasteson

, Rorvik

. Potential uptake of Escherichia coli O157:H7 from organic manure into crisphead lettuce. Appl Environ Microbiol, 2005; 71:2221–2225.

20.

Johnson

, Goullet

, Picard

, Moseley

, Roberts

, Stamm

. Association of carboxylesterase B electrophoretic pattern with presence and expression of urovirulence factor determinants and antimicrobial resistance among strains of Escherichia coli that cause urosepsis. Infect Immun, 1991; 59:2311–2315.

21.

Khachatryan

, Besser

, Call

. The streptomycin-sulfadiazine-tetracycline antimicrobial resistance element of calf-adapted Escherichia coli is widely distributed among isolates from Washington State cattle. Appl Environ Microbiol, 2008; 74:391–395.

22.

Lang

, Tsai

, Mayer

, Patton

, Palmer

. Multiplex PCR for detection of the heat-labile toxin gene and shiga-like toxin I and II genes in Escherichia coli isolated from natural waters. Appl Environ Microbiol, 1994; 60:3145–3149.

23.

McEwen

, Fedorka-Cray

. Antimicrobial use and resistance in animals. Clin Infect Dis, 2002; 34:S93–S106.

24.

Nunnery

, Angulo

, Tollefson

. Public health and policy. Prev Vet Med, 2006; 73:191–195.

25.

Orsi

, Stoppe

, Sato

, Prado

, Ottoboni

. Phylogenetic group distribution among Escherichia coli isolated from rivers in São Paulo State, Brazil. World J Microbiol Biotechnol, 2008; 24:1573–1577.

26.

Pallecchi

, Lucchetti

, Bartoloni

, Bartalesi

, Mantella

, Gamboa

, Carattoli

, Paradisi

, Rossolini

. Population structure and resistance genes in antibiotic-resistant bacteria from a remote community with minimal antibiotic exposure. Antimicrob Agents Chemother, 2007; 51:1179–1184.

27.

Piatti

, Mannini

, Balistreri

, Schito

. Virulence factors in urinary Escherichia coli strains: phylogenetic background and quinolone and fluoroquinolone resistance. J Clin Microbiol, 2008; 46:480–487.

28.

Sandvang

. Novel streptomycin and spectinomycin resistance gene as a gene cassette within a class 1 integron isolated from Escherichia coli. Antimicrob Agents Chemother, 1999; 43:3036–3038.

29.

Shepherd

, Liang

, Jiang

, Doyle

, Erickson

. Fate of Escherichia coli O157:H7 during on-farm dairy manure–based composting. J Food Prot, 2007; 70:2708–2716.

30.

Smith

, Drum

DJV

, Dai

, Kim

, Sanchez

, Maurer

, Hofacre

, Lee

. Impact of antimicrobial usage on antimicrobial resistance in commensal Escherichia coli strains colonizing broiler chickens. Appl Environ Microbiol, 2007; 73:1404–1414.

31.

Soares

, Cardenas

, Weir

, Switzenbaum

. Evaluating pathogen regrowth in biosolids compost. Biocycle, 1995; 36:70–72.

32.

Solomon

, Yaron

, Matthews

. Transmission of Escherichia coli O157:H7 from contaminated manure and irrigation water to lettuce plant tissue and its subsequent internalization. Appl Environ Microbiol, 2002; 68:397–400.

33.

Sunde

, Norstrom

. The genetic background for streptomycin resistance in Escherichia coli influences the distribution of MICs. J Antimicrob Chemother, 2005; 56:87–90.

34.

Talan

, Krishnadasan

, Abrahamian

, Stamm

, Moran

. Prevalence and risk factor analysis of trimethoprim-sulfamethoxazole-and fluoroquinolone-resistant Escherichia coli infection among emergency department patients with pyelonephritis. Clin Infect Dis, 2008; 47:1150–1158.

35.

Tramuta

, Robino

, Nebbia

. Phylogenetic background of attaching and effacing Escherichia coli isolates from animals. Vet Res Commun, 2008; 32:433–437.

36.

van den Bogaard

, Stobberingh

. Epidemiology of resistance to antibiotics: links between animals and humans. Int J Antimicrob Agents, 2000; 14:327–335.

37.

Varaiya

, Dogra

, Kulkarni

, Bhalekar

. Extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae in diabetic foot infections. Indian J Pathol Microbiol, 2008; 51:370–372.

38.

Walk

, Mladonicky

, Middleton

, Heidt

, Cunningham

, Bartlett

, Sato

, Whittam

. Influence of antibiotic selection on genetic composition of Escherichia coli populations from conventional and organic dairy farms. Appl Environ Microbiol, 2007; 73:5982–5989.

39.

White

, Zhao

, McDermott

, Ayers

, Gaines

, Friedman

, Wagner

, Meng

, Needle

, Davis

, DebRoy

. Characterization of antimicrobial resistance among Escherichia coli O111 isolates of animal and human origin. Microb Drug Resist, 2002; 8:139–146.

40.

Wilson

. Rational selection of antimicrobials for use in horses. Proceedings of the 47th Annual Convention American Association Equine Practitioners, 2001; 75–93.

41.

Zhao

, White

, Ayers

, Friedman

, English

, Wagner

, Gaines

, Meng

. Identification and characterization of integron-mediated antibiotic resistance among shiga toxin-producing Escherichia coli isolates. Appl Environ Microbiol, 2001; 67:1558–1564.

42.

Zou

, He

, Chi

, Jong

, Huang

. Involvement of Escherichia coli K1 ibeT in bacterial adhesion that is associated with the entry into human brain microvascular endothelial cells. Med Microbiol Immunol, 2008; 197:337–344.