Skip to main content
Applied and Environmental Microbiology logoLink to Applied and Environmental Microbiology
. 2011 Oct;77(20):7171–7175. doi: 10.1128/AEM.05069-11

Effective Antibiotic Resistance Mitigation during Cheese Fermentation

Xinhui Li 1, Yingli Li 1,, Valente Alvarez 1, Willis James Harper 1, Hua H Wang 1,2,*
PMCID: PMC3194865  PMID: 21784910

Abstract

Controlling antibiotic-resistant (ART) bacteria in cheese fermentation is important for food safety and public health. A plant-maintained culture was found to be a potential source for ART bacterial contamination in cheese fermentation. Antibiotics had a detectable effect on the ART population from contamination in the finished product. The decrease in the prevalence of antibiotic resistance (AR) in retail cheese samples from 2010 compared to data from 2006 suggested the effectiveness of targeted AR mitigation in related products.

TEXT

The rapid emergence of antibiotic-resistant (ART) pathogens has been a major public health concern (11). The recent findings on the prevalence of antibiotic resistance (AR) gene-containing ART commensal bacteria in a broad spectrum of ready-to-consume items (3, 12, 18, 20) suggested that the food chain likely has served as an important avenue for dissemination of ART bacteria to the general public. From 2005 to 2006, small-scale surveillance studies using retail samples purchased from national chain grocery stores in the Columbus, OH, area revealed the presence of high levels of ART bacteria and a representative AR gene pool in most cheese samples (multiple types and brands) examined (14, 18). Various commensal bacteria, such as Streptococcus thermophilus, Pseudomonas sp., Staphylococcus sp., Lactococcus lactis, and Lactococcus sp., were identified to be the isolates carrying AR genes (18). The AR genes from certain food-borne isolates were further transmitted to human-pathogenic and residential bacteria in laboratory settings by horizontal gene transfer (HGT) mechanisms, leading to acquired resistance in the recipient cells, which suggests the functionality and mobility of the food-borne AR genes (5, 12, 18). The high prevalence of ART bacteria and the AR gene pool in cheese products and the identification of certain isolates carrying AR genes as lactic acid bacteria raised concerns about the potential impact of the dairy fermentation process on the emergence and amplification of ART bacteria. Thus, the objective of this study was to reveal critical control points (CCP) in reducing the number of ART bacteria in dairy fermentation for targeted mitigation.

CCP assessment in pilot-plant-scale cheese making.

Cheddar-type cheeses were made at the Ohio State University (OSU) pilot plant to determine the impact of starter cultures, the fermentation process, and antibiotics on the development of ART bacteria. The mesophilic culture F-DVS R-604 (culture A, containing a mix of L. lactis subsp. lactis and L. lactis subsp. cremoris) and the thermophilic culture F-DVS RSF-621 (culture B, containing L. lactis, Lactobacillus helveticus, and S. thermophilus), both from the Chr. Hansen Company (Milwaukee, WI), were used in the study. Cultures were activated by inoculating 2 to 3 g of a frozen pellet into 940 ml shelf-stable ultrahigh-temperature (UHT) milk (Parmalat 2% reduced-fat milk; Farmland Dairies, Wallington, NJ) and incubated at 33°C (mesophilic) or 38°C (thermophilic) for 18 h. Raw milk was obtained from the OSU dairy farm (Columbus, OH) and pasteurized at 71.7°C for 30 s using a plate heat exchanger (APV Corporation, Buffalo, NY) at the OSU dairy pilot plant. Standard procedures for Cheddar cheese making were followed (10) using chymosin (Chy-Max Extra from Chr. Hansen). Pasteurized milk was inoculated with 2% (wt/wt) of the mesophilic culture or 1% (wt/wt) of the rapidly growing thermophilic starters in cheese making. Proper sanitation practices were applied during the cheese-making process. Vacuum-packed cheese samples were stored at 7°C for ripening for up to 6 months.

To investigate the impact of the fermentation process on the development of ART bacteria, three vats of cheese (batch 1, in 5-gal vats) were made with either mesophilic culture A, thermophilic culture B, or thermophilic culture B with a subinhibitory concentration of tetracycline (0.25 μg/ml). Samples of raw milk, pasteurized milk, starter cultures, whey, curds, and cheese were collected during the cheese making and every month throughout the ripening process. A replicate (batch 2) of the pilot-plant cheese making was conducted for repeatability. All samples were subjected to assessments for total bacteria as well as tetracycline-resistant (Tetr) bacteria as described previously (12, 18) but with modifications. Brain heart infusion (BHI) and MRS plates supplemented with 16 μg/ml tetracycline (Sigma-Aldrich, St. Louis, MO) were used to recover the corresponding ART bacteria, and all plates were incubated aerobically at 30 or 37°C for 48 h. The cell numbers reported were the mean values from at least two replicates.

Figure 1 shows that Tetr bacteria were detected only in the raw milk, indicating that pasteurization effectively eliminated Tetr bacteria. Tetr bacteria were not detected in either mesophilic or thermophilic commercial starter cultures. No Tetr bacteria were detected in samples collected throughout the cheese-making process (Fig. 1) or the 6-month ripening period (data not shown). The results indicated that with proper sanitation controls, production of ART-bacterium-free cheese is achievable. In a contained system, the presence of a subinhibitory level of an antibiotic did not lead to the development of corresponding ART bacteria.

Fig. 1.

Fig. 1.

Dynamics of counts of total bacteria and Tetr bacteria during Cheddar-type-cheese making. Results represent samples collected from cheese made with a mesophilic starter culture. Similar results were found with a thermophilic culture (data not shown). ND, not detected. The arrow indicates the inoculation point of the starter culture. ▪, batch 1 total plate counts; ▴, batch 2 total plate counts; □, batch 1 Tetr counts; ▵, batch 2 Tetr counts.

To further examine the impact of antibiotics on the ART-bacterium population and the potential of HGT events in a contaminated system during cheese making, a tet(M)+ tet(L)+ Enterococcus faecium isolate, M7M2, from a cheese sample previously made at the OSU pilot plant was used to spike pasteurized milk during cheese making and the dynamic changes of ART bacteria in the fermentation process were examined. Three vats of cheese were made with either culture B (control), culture B with a low dose (approximately 1 CFU/ml) of M7M2, or culture B with a low dose of M7M2 and 0.25 μg/ml of tetracycline. The experiment was repeated twice (batches 3 and 4). Figure 2 illustrates that the overall change in the numbers of Tetr bacteria remained within 1 log value during the 6-month ripening period, indicating that ripening did not significantly affect the dynamics of Tetr bacteria in cheese, even when it was made from milk with a low concentration of the antibiotic. However, cheese samples made from milk with Tetr E. faecium and tetracycline had slightly higher Tetr bacterial counts than those spiked only with Tetr E. faecium. A lab test was further conducted to reveal the potential cause of this phenomenon. Tubes containing 36 ml of milk with 0.25, 0.5, and 1 μg/ml of tetracycline were coinoculated with 4 g of activated starter culture B and the same dose of Tetr E. faecium (approximately 4 to 10 CFU/ml, with 3 replicates) and incubated at 37°C for 6 h. Then a 30% (wt/vol) NaCl solution was added to each tube to a final concentration of 2% (wt/wt), and all tubes were further incubated at room temperature for 18 h. Results showed that a decreased acid production by the starter culture and an increased amplification of the ART bacteria in the microbial population were associated with the increased tetracycline concentration (data not shown). It was further confirmed by PCR using E. faecium-specific primers (Table 1) targeting the ddl gene that all 182 ART isolates recovered from the end products were E. faecium. Similar results were found after screening 144 Tetr isolates from ripened cheese samples of batches 3 and 4. These data indicated that the Tetr bacteria were the originally spiked ART M7M2 strain, and there was no indication of an HGT event(s) between M7M2 and the starter culture under the experimental conditions.

Fig. 2.

Fig. 2.

Dynamics of the counts of Tetr bacteria during ripening of the cheese made with the thermophilic starter spiked with Tetr E. faecium and tetracycline. ▪, batch 3 spiked with Tetr E. faecium; ▴, batch 4 spiked with Tetr E. faecium; □, batch 3 spiked with Tetr E. faecium and tetracycline; ▵, batch 4 spiked with Tetr E. faecium and tetracycline.

Table 1.

Primers and probes used in this study

Gene Sequence (5′ to 3′) (primer or probe type)a Amplicon size (bp) Reference(s)
ddl CCAAGGCTTCTTAGAGA (F) 535 4b,c
CATCGTGTAAGCTAACTTC (R)
tet(S) GAACGCCAGAGAGGTATTAC (F) 1,050 12b
TACCTCCATTTGGACCTCAC (R)
tet(M) GAACTCGAACAAGAGGAAA (F) 979 12b
CCAATACAATAGGAGCAAGC (R)
tet(L) TTGGATCGATAGTAGCCATG (F) 908 12
GTAACCAGCCAACTAATGAC (R)
tet(K) AGGATAGCCATGGCTACAAG (F) 981 12
ACAAGGAGTAGGATCTGCTG (R)
16S rRNA AGAGTTTGATCCTGGCTCAG (F) 1,498 12, 19
TACCTTGTTACGACTT (R)
tet(S) GTATGTTCATCTTTCTAAG (F) 190 14d
GCAATAACATCTTTTCAAC (R)
tet(M) GAACATCGTAGACACTCAATTG (F) 168 9d
CAAACAGGTTCACCGG (R)
tet(S) FAM-CCATGTGTCCAGGAGTATCTAC-BHQ (P) 14
tet(M) FAM-CGGTGTATTCAAGAATATCGTAGTG-BHQ (P) 9
a

F, forward primer; R, reverse primer; P, probe for real-time PCR; FAM, 6-carboxyfluorescein; BHQ, black hole quencher.

b

With modification.

c

E. faecium-specific primers.

d

Primers for real-time PCR.

CCP assessment in commercial Swiss cheese-manufacturing plants.

To assess CCP for ART bacteria under commercial conditions, samples, including raw milk, pasteurized milk, starter culture, cheese curds, whey, and cheese samples during ripening, were collected throughout the Swiss-cheese-making process from two commercial manufacturing facilities in Ohio, designated plant I and plant II. A conventional PCR was used to detect the presence of representative tetracycline resistance genes [tet(M), tet(S), tet(L), and tet(K)] of Tetr isolates as described previously (12). Templates for PCR were prepared by the bead-beating method (18) or an alternative Triton X-100 boiling method (7). The presence of AR genes and of identified isolates carrying those AR genes is summarized in Table 2.

Table 2.

Identified AR genes and isolates carrying the AR genes in cheese plant samples

Cheese plant and type of sample Resistance gene identified (no. of positive isolates/total no. of isolates screened) Isolate(s) carrying the AR gene (no. of isolates carrying the AR gene/total no. of isolates identified)
Plant I
    Adjunct culture maintained by cheese plant tet(S) (40/40) Streptococcus thermophilus (4/4)
    Cheese curds after pressing tet(S) (20/20) Streptococcus thermophilus (6/6)
    Cheese before brining tet(S) (24/24) Streptococcus sp. (3/4), Lactococcus sp. (1/4)
    Cheese after brining tet(S) (2/3) Leuconostoc sp. (2/3)
tet(L) (1/3) Staphylococcus sp. (1/3)
Plant II
    Whey tet(S) (4/4) Streptococcus thermophilus (1/1)
    Cheese curds before pressing tet(S) (3/3) Streptococcus thermophilus (2/2)
    Cheese curds after pressing tet(S) (53/53) Streptococcus thermophilus (14/14)
    Cheese before brining tet(S) (33/36) Streptococcus thermophilus (2/3)
tet(M) (3/36) Streptococcus sp. (1/3)
    Cheese after brining tet(S) (19/19) Streptococcus thermophilus (6/6)
    Cheese after warm-room ripening tet(S) (6/35) Lactobacillus sp. (12/13)
tet(M) (29/35) Streptococcus thermophilus (1/13)
    Cheese after cold-room ripening tet(M) (25/25) Lactobacillus sp. (2/2)

As in the case of the pilot-plant Cheddar-type-cheese study, no Tetr bacteria were detected in pasteurized milk. The Tetr bacterial counts were 102 CFU/ml (g) or less in most samples collected during the Swiss-cheese-making process. The commercial starter cultures were free of Tetr bacteria. However, Tetr bacteria (9.2 × 103 CFU/ml) were found in the adjunct starter culture maintained by plant I, and selected isolates were identified to be S. thermophilus [tet(S)+], suggesting that the in-house-maintained culture is a CCP for AR mitigation in cheese fermentation. Plant II did not carry in-house-maintained starters in the plant.

Prevalence of AR genes in retail cheese samples.

Twelve retail cheese samples purchased in 2010, involving 8 brands (most brands and cheese types were the same as in the 2005 and 2006 studies) from 2 national chain stores in the Columbus, OH, area, were analyzed again for both Tetr bacteria and the tet(S) and tet(M) gene pools by following procedures described previously but with slight modifications (14). The real-time PCR thermo profiles were 95°C for 3 min, followed by 40 cycles of 95°C for 30 s, 55°C for 30 s, and 68°C for 20 s, with a final extension of 68°C for 5 min. Tetr bacteria were detected in only 3 out of 12 samples (Fig. 3), including one with a total of 20 colonies (detected on two plates from the least-diluted sample) and two others containing 103 CFU/g and 104 CFU/g of cheese, while 6 out of 11 samples in the 2006 study contained 104 to 106 CFU/g of cheese. Accordingly, the tet(S) and tet(M) gene pools in most of the 2010 cheese samples [except the tet(S) gene pool in one sample] were present at around 104 to 105 copies per g of sample, which is around or slightly above the detection limit of the real-time PCR method (104 copies per g), while 7 out of 11 cheese samples in the 2006 study contained 107 or more copies of the tet(S) gene per g of food (14). Because most brands and cheese types examined were the same as those tested in the 2006 study, the data from this small-scale study illustrated a trend toward a reduction in the prevalence of both ART bacteria and the AR gene pool in retail cheese samples in the past 4 years.

Fig. 3.

Fig. 3.

Assessment of antibiotic resistance in commercial cheese samples by conventional plate counting and real-time PCR. MRS agar plates containing 16 μg/ml tetracycline were incubated aerobically at 30°C for 48 h to obtain Tetr bacterial counts. Cheese samples: A, sharp Cheddar; B, mild Cheddar 1; C, mild Cheddar 2; D, mild Cheddar 3; E, mild Cheddar 4; F, sharp white Cheddar; G, mild Cheddar 5; H, Swiss 1; I, aged Swiss; J, baby Swiss 1; K, Swiss 2; L, baby Swiss 2. The data are the means of the results from at least two replicates, and the coefficients of variation of the log copy numbers of resistance genes were less than 0.1. White bars, log numbers of Tetr CFU/g; hatched bars, log tet(S) copy numbers per g; black bars, log tet(M) copy numbers per g.

Discussion and conclusion.

The frequency of HGT is correlated to the size of the AR gene pool and to genetic features, including the compatibility of the donor and recipient strains. A successful dairy fermentation process involves the effective amplification of starter cultures and the inhibition of the growth of spoilage and pathogenic bacteria. The prevalence of ART bacteria in the final products could be due to the amplification of AR gene-containing starter cultures or ART contaminants throughout the fermentation process or could result from HGT events during fermentation and ripening. Even though no specific HGT events were detected in this study, because lactic acid bacteria are prone to HGT events in vitro as well as in vivo (6, 8, 13, 15), the second scenario still may be a concern.

Although the AR-gene-encoding plasmids from Lactococcus sp. and Enterococcus sp. were transmitted to Streptococcus mutans or Enterococcus faecalis by natural transformation and electroporation (12), we were unsuccessful in illustrating transmission of the AR gene from Tetr S. thermophilus to Tets S. thermophilus under the same laboratory conditions (data not shown). It is expected that the involvement of HGT events likely varies among isolates carrying AR genes, and thus, proper risk assessment should be based on not only the size of the AR gene pool but also the genetic characteristics of the isolates carrying AR genes. Cultures with potential fermentation or probiotic applications should be characterized at the strain (isolate) level for not only the presence of AR genes but also the potential for both acquisition and dissemination of such genes via HGT mechanisms.

Results from the study also revealed that with thorough safety screening, commercial starter cultures from major companies are free of AR genes. Further, there is no indication of detectable HGT events during fermentation and ripening, even in the presence of the corresponding antibiotic compound, under the specified experimental conditions. However, our result does not exclude the potential involvement of the dairy ART isolates in disseminating AR genes under other environmental conditions, including an HGT event in vivo (6).

The reduction in the prevalence of AR in retail cheese products indicated the effectiveness of targeted mitigation strategies by the dairy and starter culture industries in the past several years. However, the lack of safety screening in locally maintained starters and adjunct cultures, as well as the occasional ART bacterial contamination from the environment and facility during cheese making, likely contributed to the sporadic cases of contaminated cheeses. This observation is also in agreement with the recent reports on dairy ART bacteria, but those bacteria were isolated mostly from specialty cheeses and from areas with probably less access to the cutting-edge literature and knowledge (2, 16, 17). However, there is no geographic boundary for the rapid dissemination of problematic organisms nowadays. Lactic acid bacteria are also commonly used in large quantities as fermentation starters and probiotic supplements for human, food animal, and aquaculture animal consumption in developing countries. Thus, it is particularly important to properly communicate the scientific knowledge, including the risk factors involved and the mitigation strategies, to a broad audience. Uncovering the tendency, mechanisms, and conditions of HGT events and enhancing the safety screening of cultures for beneficial applications are critical for targeted AR mitigation.

ACKNOWLEDGMENTS

The study was supported by OARDC seed grant OHOA1084 and Dairy Management Inc. grant OSURF (project 60010225).

We sincerely thank industry participants for their collaboration.

Footnotes

Published ahead of print on 22 July 2011.

REFERENCES

  • 1.Reference deleted.
  • 2. Comunian R., et al. 2010. Susceptibility to tetracycline and erythromycin of Lactobacillus paracasei strains isolated from traditional Italian fermented foods. Int. J. Food Microbiol. 138:151–156 [DOI] [PubMed] [Google Scholar]
  • 3. Duran G. M., Marshall D. L. 2005. Ready-to-eat shrimp as an international vehicle of antibiotic-resistant bacteria. J. Food Prot. 68:2395–2401 [DOI] [PubMed] [Google Scholar]
  • 4. Dutka-Malen S., Evers S., Courvalin P. 1995. Detection of glycopeptide resistance genotypes and identification to the species level of clinically relevant enterococci by PCR. J. Clin. Microbiol. 33:24–27 (Erratum, 33:1434.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Feld L., Bielak E., Hammer K., Wilcks A. 2009. Characterization of a small erythromycin resistance plasmid pLFE1 from the food-isolate Lactobacillus plantarum M345. Plasmid 61:159–170 [DOI] [PubMed] [Google Scholar]
  • 6. Feld L., et al. 2008. Selective pressure affects transfer and establishment of a Lactobacillus plantarum resistance plasmid in the gastrointestinal environment. J. Antimicrob. Chemother. 61:845–852 [DOI] [PubMed] [Google Scholar]
  • 7. Iwamoto T., Sonobe T., Hayashi K. 2003. Loop-mediated isothermal amplification for direct detection of Mycobacterium tuberculosis complex, M. avium, and M. intracellulare in sputum samples. J. Clin. Microbiol. 41:2616–2622 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Jacobsen L., et al. 2007. Horizontal transfer of tet(M) and erm(B) resistance plasmids from food strains of Lactobacillus plantarum to Enterococcus faecalis JH2-2 in the gastrointestinal tract of gnotobiotic rats. FEMS Microbiol. Ecol. 59:158–166 [DOI] [PubMed] [Google Scholar]
  • 9. Kinkelaar D. F. 2008. M.S. thesis. The Ohio State University, Columbus, OH [Google Scholar]
  • 10. Kosikowski F. V. 1982. Cheese and fermented milk foods, 2nd ed, 3rd printing with revision F. V. Kosikowski and Associates, Brookdontale, NY [Google Scholar]
  • 11. Levy S. B., Marshall B. 2004. Antibacterial resistance worldwide: causes, challenges and responses. Nat. Med. 10:S122–S129 [DOI] [PubMed] [Google Scholar]
  • 12. Li X., Wang H. H. 2010. Tetracycline resistance associated with commensal bacteria from representative ready-to-consume deli and restaurant foods. J. Food Prot. 73:1841–1848 [DOI] [PubMed] [Google Scholar]
  • 13. Luo H., Wan K., Wang H. H. 2005. High-frequency conjugation system facilitates biofilm formation and pAMβ1 transmission by Lactococcus lactis. Appl. Environ. Microbiol. 71:2970–2978 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Manuzon M. Y., et al. 2007. Quantitative assessment of the tetracycline resistance gene pool in cheese samples by real-time TaqMan PCR. Appl. Environ. Microbiol. 73:1676–1677 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Morelli L., Sarra P. G., Bottazzi V. 1988. In vivo transfer of pAM beta 1 from Lactobacillus reuteri to Enterococcus faecalis. J. Appl. Bacteriol. 65:371–375 [DOI] [PubMed] [Google Scholar]
  • 16. Togay S. O., Keskin A. C., Acik L., Temiz A. 2010. Virulence genes, antibiotic resistance and plasmid profiles of Enterococcus faecalis and Enterococcus faecium from naturally fermented Turkish foods. J. Appl. Microbiol. 109:1084–1092 [DOI] [PubMed] [Google Scholar]
  • 17. Rizzotti L., La Gioia F., Dellaglio F., Torriani S. 2009. Characterization of tetracycline-resistant Streptococcus thermophilus isolates from Italian soft cheeses. Appl. Environ. Microbiol. 75:4224–4229 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Wang H. H., et al. 2006. Food commensal microbes as a potentially important avenue in transmitting antibiotic resistance genes. FEMS Microbiol. Lett. 254:226–231 [DOI] [PubMed] [Google Scholar]
  • 19. Weisburg W. G., Barns S. M., Pelletier D. A., Lane D. J. 1991. 16S ribosomal DNA amplification for phylogenetic study. J. Bacteriol. 173:697–703 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Zhang X. X., Zhang T., Fang H. H. 2009. Antibiotic resistance genes in water environment. Appl. Microbiol. Biotechnol. 82:397–414 [DOI] [PubMed] [Google Scholar]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)

RESOURCES