Efficacy of Various Pasteurization Time-Temperature Conditions in Combination with Homogenization on Inactivation of Mycobacterium avium subsp. paratuberculosis in Milk

doi:10.1128/AEM.71.6.2853-2861.2005

Appl Environ Microbiol. 2005 June; 71(6): 2853–2861.

doi: 10.1128/AEM.71.6.2853-2861.2005

PMCID: PMC1151836

Efficacy of Various Pasteurization Time-Temperature Conditions in Combination with Homogenization on Inactivation of Mycobacterium avium subsp. paratuberculosis in Milk

Irene R. Grant,¹^* Alan G. Williams,³ Michael T. Rowe,^1,² and D. Donald Muir³

Department of Food Microbiology, Queen's University of Belfast,¹ Food Microbiology Branch, Department of Agriculture and Rural Development for Northern Ireland, Agriculture and Food Science Centre, Newforge Lane, Belfast BT9 5PX, Northern Ireland, United Kingdom,² Hannah Research Institute, Ayr KA6 5HL, Scotland, United Kingdom³

^*Corresponding author. Mailing address: Department of Food Microbiology, Queen's University of Belfast, Agriculture and Food Science Centre, Newforge Lane, Belfast BT9 5PX, United Kingdom. Phone: 44 28 9025 5299. Fax: 44 28 9025 5009. E-mail: i.grant/at/qub.ac.uk.

Received September 24, 2004; Accepted December 16, 2004.

This article has been cited by other articles in PMC.

Abstract

The effect of various pasteurization time-temperature conditions with and without homogenization on the viability of Mycobacterium avium subsp. paratuberculosis was investigated using a pilot-scale commercial high-temperature, short-time (HTST) pasteurizer and raw milk spiked with 10¹ to 10⁵ M. avium subsp. paratuberculosis cells/ml. Viable M. avium subsp. paratuberculosis was cultured from 27 (3.3%) of 816 pasteurized milk samples overall, 5 on Herrold's egg yolk medium and 22 by BACTEC culture. Therefore, in 96.7% of samples, M. avium subsp. paratuberculosis had been completely inactivated by HTST pasteurization, alone or in combination with homogenization. Heat treatments incorporating homogenization at 2,500 lb/in², applied upstream (as a separate process) or in hold (at the start of a holding section), resulted in significantly fewer culture-positive samples than pasteurization treatments without homogenization (P < 0.001 for those in hold and P < 0.05 for those upstream). Where colony counts were obtained, the number of surviving M. avium subsp. paratuberculosis cells was estimated to be 10 to 20 CFU/150 ml, and the reduction in numbers achieved by HTST pasteurization with or without homogenization was estimated to be 4.0 to 5.2 log₁₀. The impact of homogenization on clump size distribution in M. avium subsp. paratuberculosis broth suspensions was subsequently assessed using a Mastersizer X spectrometer. These experiments demonstrated that large clumps of M. avium subsp. paratuberculosis cells were reduced to single-cell or “miniclump” status by homogenization at 2,500 lb/in². Consequently, when HTST pasteurization was being applied to homogenized milk, the M. avium subsp. paratuberculosis cells would have been present as predominantly declumped cells, which may possibly explain the greater inactivation achieved by the combination of pasteurization and homogenization.

The causative agent of Johne's disease in dairy cattle, Mycobacterium avium subsp. paratuberculosis, may also be associated with Crohn's disease, an inflammatory bowel disease in humans, although the potential link remains the matter of much debate. Recent reviews of the available evidence concluded that there was insufficient evidence to prove or disprove that M. avium subsp. paratuberculosis is a causative agent of some or all cases of Crohn's disease, but a causal link remains a plausible hypothesis (4, 21, 22, 24). If there is a link between Johne's disease and Crohn's disease, then M. avium subsp. paratuberculosis may be transmitted to humans via cows' milk. M. avium subsp. paratuberculosis may be naturally present in milk produced by an infected cow (26, 28) or it may gain access to the milk as the result of fecal contamination during the milking process. Currently, the number of M. avium subsp. paratuberculosis cells present in naturally infected raw milk cannot be accurately determined, as the culture of M. avium subsp. paratuberculosis from milk necessitates a chemical decontamination treatment which will have an adverse effect on the viability of a proportion of the M. avium subsp. paratuberculosis cells present (3, 7). Consequently, any colony count reported for milk when chemical decontamination has been applied is an underestimate of the real number of M. avium subsp. paratuberculosis cells present.

The concern to the milk and milk products industry about M. avium subsp. paratuberculosis centers on its apparent heat resistance. Time-temperature conditions used to pasteurize milk were originally designed to kill Mycobacterium bovis, the cause of tuberculosis in cattle and potentially humans and a relative of M. avium subsp. paratuberculosis. The expectation was that M. avium subsp. paratuberculosis present in milk would also be effectively killed by pasteurization of milk at 72°C for 15 s (so-called high-temperature, short-time [HTST] pasteurization). However, there is a growing body of evidence, from both laboratory pasteurization studies involving milk spiked with laboratory-grown M. avium subsp. paratuberculosis (2, 6, 10, 13, 15, 17, 19, 20, 27) and studies of pasteurized naturally infected milk (8, 12), showing that low levels of viable M. avium subsp. paratuberculosis may survive HTST pasteurization of milk on occasion. Thermal death information suggests that the heat resistance of M. avium subsp. paratuberculosis may be influenced by its tendency to occur as clumps of cells (9, 15). Bacteria in the center of large clumps may be protected and may require additional heating to be inactivated, or alternatively, because the clumps contain up to 10,000 cells, a proportion of the cells will survive a nonsterilizing heat treatment. Despite the uncertainty surrounding the mechanism of apparent heat tolerance, much of the United Kingdom dairy industry responded to the potential M. avium subsp. paratuberculosis survival problem in 1998 by extending the holding time for heat treatment of liquid milk to 25 s at 72°C (1). This action was prompted by the findings of laboratory pasteurization studies at Queen's University Belfast (QUB) which appeared to indicate that an extended holding time would be more effective in killing M. avium subsp. paratuberculosis than an elevated pasteurization temperature (11). However, the findings of subsequent studies by the same researchers employing naturally infected milk did not provide any clear evidence that this was the case (8, 12), since surviving M. avium subsp. paratuberculosis cells were isolated from commercially pasteurized milk samples heat treated for both holding times.

Further research was therefore warranted in order to identify more effective pasteurization conditions or additional milk treatments that could be used to break up or remove clumps of M. avium subsp. paratuberculosis before heat treatment. However, there were important constraints on the way in which this information could be collected if the results were to be relevant to commercial practice. A major technical difficulty lay in extrapolating the effects of heat treatment using laboratory techniques (e.g., capillary tubes) to mimic the conditions in commercial heat exchangers. Laboratory pasteurization studies employing a variety of heating apparatus have generally been criticized for this reason (16, 25). Commercial pasteurizing plants are designed to accommodate a degree of fouling (deposition of material from the milk on the heat exchange surface) and to allow efficient use of heat. The surface area of the heat exchanger is increased beyond that required in theory to reduce the impact of fouling. Thermal efficiency is improved by incorporating a regeneration section in a heat exchanger. This is, in effect, a countercurrent heat exchanger where outgoing heated milk transfers its heat to incoming cold, raw milk. The consequence of these practical arrangements is that the cumulative heat load that milk experiences in a commercial plant is substantially different from that in laboratory equipment both in terms of total energy input and in the profile of the heating and cooling regimen (16). Thus, it was imperative that studies to provide information for commercial application were based on the type of heat exchange equipment used in industry. A similar argument applies to other processing steps such as homogenization, a process used to uniformly disperse solid particles throughout a liquid. In the case of milk, homogenization is the application of pressure to disperse milk fat globules evenly throughout the milk.

A further complication lay in the practical constraints associated with treatment of milk, which limited the range of processing options available for study. For example, clumps of bacteria are theoretically susceptible to disruption by different types of homogenization: conventional pump homogenizer, microfluidizer, or ultrasonic treatment. In reality, only conventional pump homogenizers are widely available within the United Kingdom dairy industry. Heat exchange can be carried out in a variety of ways, but in the United Kingdom, most companies operate heat exchangers with indirect heating and cooling of the milk. It was considered important that any solution to the M. avium subsp. paratuberculosis problem should be sought using processes already available within the United Kingdom dairy industry. An analysis of current United Kingdom practice revealed that the following processes were of potential practical benefit in solving the M. avium subsp. paratuberculosis survival problem: modified pasteurization time-temperature conditions, homogenization, clarification/bactofugation (use of centrifugal force to remove heavier particles such as spores and potentially bacterial clumps), or microfiltration (use of 1.4-μm filters to remove bacteria and spores from skim milk). The focus of this study was on the application of heat treatment and homogenization to reduce the numbers of viable M. avium subsp. paratuberculosis cells in milk. The objectives of this study were to identify processing strategies that would achieve a minimum 4-log₁₀ reduction in numbers of M. avium subsp. paratuberculosis cells in raw milk, within the constraints imposed by commercial considerations, and to attempt to elucidate the mechanism that confers heat resistance in M. avium subsp. paratuberculosis. In this study, great importance was placed upon ensuring that the pilot-scale pasteurizing and homogenization equipment used was of a design similar to that of a commercial plant and that it was fully calibrated and accredited to United Kingdom dairy industry standards before pasteurization trials commenced.

MATERIALS AND METHODS

Bacterial strains and culture media.

The majority of the heat trials (experiments 1 to 4 and 6 to 8 [Table [Table1])1]) were carried out using raw bovine milk spiked with a type strain of M. avium subsp. paratuberculosis, NCTC 8578. The particular NCTC 8578 culture used during this study had been purchased from the National Collection of Type Cultures and preserved on Cryobeads at −80°C at QUB until required for research purposes. For experiment 5, three other M. avium subsp. paratuberculosis strains were employed: 806R, an M. avium subsp. paratuberculosis isolate from raw milk; B4, an M. avium subsp. paratuberculosis isolate from a bovine in Northern Ireland; and 796PSS, an M. avium subsp. paratuberculosis isolate from pasteurized milk. Inoculum for the pilot plant runs was prepared at QUB by inoculating 50 ml Middlebrook 7H9 broth (Difco Laboratories, Detroit, Mich.) containing 10% OADC supplement (Difco), 2 μg/ml mycobactin J (Synbiotics Europe SAS, Lyon, France), and 0.05% Tween 80 (Sigma Chemical Co. Ltd., Poole, England) with a colony of the appropriate strain of M. avium subsp. paratuberculosis from a Herrold's egg yolk medium (HEYM) slope culture and incubating for 8 weeks at 37°C. At the time of spiking, the milk M. avium subsp. paratuberculosis cultures would have been in the stationary phase of growth. M. avium subsp. paratuberculosis inoculum was sent, as and when required, to the Hannah Research Institute (HRI) by overnight courier service.

TABLE 1.

Overview of heat trials carried out using pilot-scale pasteurizing plant

Preparation of spiked raw milk.

Raw milk was obtained from the bulk tank at either the HRI or Scottish Agricultural College, Auchincruive, Scotland, dairy farms. There had never been any sign of Johne's disease, or any recorded veterinary intervention for Johne's disease, in either herd. In all experiments (except those in which the inoculum effect was under investigation), raw milk was inoculated with an 8-week-old stationary-phase M. avium subsp. paratuberculosis culture (20 ml containing approximately 10⁸ cells/ml) and allowed to equilibrate at 4 to 6°C in a refrigerator overnight (18 h). This spiked milk (2 liters) was made up to 20 liters with raw milk immediately prior to heat treatment, producing an initial M. avium subsp. paratuberculosis concentration of approximately 10⁵ cells/ml. In trials in which the initial level of M. avium subsp. paratuberculosis was varied (experiments 4 and 5), the broth culture was diluted 10-, 100-, and 1,000-fold before the addition of 20 ml to 20 liters of raw milk to achieve inoculum levels of approximately 10⁴, 10³, and 10² M. avium subsp. paratuberculosis cells/ml.

Description of pilot pasteurization and homogenization plant.

A pilot-scale pasteurizer comparable in characteristics to certain full-scale commercial equipment was constructed in a microbiological hazard level 2 containment facility at HRI. Dairy industry experts were fully consulted concerning the design and operation of this equipment. The heat exchanger plate pack, holding sections, temperature, pressure and flow monitoring probes, and logging equipment were externally accredited by Honeywell Control Systems Limited (Glasgow, United Kingdom) and Pressure Technical Services Limited (Manchester, United Kingdom). Plate pack integrity was confirmed upon completion of the plant trials to establish that sample contamination by raw milk in the plate section had not occurred. The pasteurizing plant could be configured in three ways: homogenization applied upstream of heat treatment (as in certain conventional industrial plants), homogenization applied within the holding section of the pasteurizer, or homogenizer not operational.

The homogenizer used was a Crepaco Multi-flo model 3DDL fitted with two-stage, aseptic SPB valves. It was operated at a pressure of 2,500 lb/in² in all experiments except experiment 7, when homogenization at both 2,500 lb/in² and 4,000 lb/in² was carried out. The pumping speed, and hence turbulent flow rate through the heat exchanger, was controlled at approximately 1,650 ml/min using a Mitsubishi transistorized inverter model Freqrol A140E (EC) to achieve the accredited hold time.

Pasteurization trials.

Prior to each run, the entire pilot plant was steam scalded at 85°C for a minimum period of 30 min. At the completion of a run, the plant was thoroughly cleaned and sanitized at an elevated temperature with a strongly alkaline 2.5% (vol/vol) solution of LQ32 (Diversey Lever). The plant was a closed system, and waste product from the pasteurizer was fed into a reservoir of LQ32 sterilant. All waste in the reservoir was heated to and maintained at temperatures in excess of 80°C for 2 to 3 h, cooled, and held for at least 12 h before disposal to ensure that any residual viable M. avium subsp. paratuberculosis was destroyed.

Over the course of this study, a series of replicated experiments was carried out in which the pasteurization temperature (72.5 to 84.5°C), hold time (15, 25, or 60s), homogenizer setting (none, upstream, or in hold), inoculum level (10¹ to 10⁵ M. avium subsp. paratuberculosis cells/ml), and mycobacterial strain were varied to evaluate their effects on M. avium subsp. paratuberculosis survival in milk after pasteurization. The experimental conditions that were studied are summarized in Table Table1.1. Up to four different time-temperature-homogenization combinations were carried out using a single 20-liter batch of spiked milk on each processing day, with the most stringent processing conditions applied first. Samples of heat-treated milk were collected aseptically from the pasteurizer and transported by overnight courier, on ice, to QUB for M. avium subsp. paratuberculosis testing.

M. avium subsp. paratuberculosis testing of milk samples.

Milk samples (200 ml) were received for testing at QUB within 24 h of processing at HRI. Upon receipt, each sample was aseptically divided into three 50-ml portions in sterile centrifuge tubes and then placed at 4°C overnight. Testing commenced the next morning, so samples were 48 h postpasteurization at the time of testing. All heat-treated milk samples were then centrifuged at 4,000 rpm for 15 min. A chemical decontamination treatment was not applied to any of the milk samples. The cream and pellet fractions were simply combined and resuspended in 4.5 ml phosphate-buffered saline containing 0.05% Tween 20, and the skim fraction was discarded. Five hundred microliters of this 10× concentrated sample was inoculated directly into a vial of BACTEC 12B medium (Becton Dickinson UK Ltd., Cowley, Oxford, United Kingdom) supplemented with mycobactin J (Synbiotics Europe SAS, Lyon, France) and PANTA PLUS antibiotic supplement (Becton Dickinson), and 250 μl was inoculated onto two slopes of HEYM containing mycobactin J and VAN antibiotics (vancomycin, amphotericin B, and nalidixic acid). A total of 816 heat-treated milk samples were tested over the course of this study.

A sample of uninoculated raw milk was also submitted for testing after each production run to confirm that the raw milk used in the pasteurization trials was M. avium subsp. paratuberculosis free. This sample was diluted in Maximum Recovery diluent (MRD; Oxoid), and 10⁻² and 10⁻³ dilutions were inoculated onto two HEYM slopes and into one vial of BACTEC 12B medium.

Milk cultures were incubated for up to 18 weeks at 37°C. BACTEC cultures were read weekly on the BACTEC 460 machine, and when a positive growth index reading was obtained, a 1-ml portion of the BACTEC culture was withdrawn and subjected to Ziehl-Neelsen acid-fast staining. If acid-fast cells were observed, the remainder of the BACTEC culture was centrifuged and the pellet was seeded onto HEYM slopes to obtain discrete colonies. When typical colonies were observed on HEYM slopes, a colony was subcultured into 2 ml Middlebrook 7H9 broth. When growth occurred (i.e., turbidity was observed), a portion of the broth culture was subjected to DNA extraction and IS900 PCR to confirm that the isolate was M. avium subsp. paratuberculosis.

Estimation of initial M. avium subsp. paratuberculosis inoculum levels in spiked raw milk.

It did not prove possible to obtain a count from the inoculated raw milk because other milk microorganisms grew more rapidly than M. avium subsp. paratuberculosis and contaminated the HEYM and BACTEC cultures. Instead, a subsample of the Middlebrook broth culture of M. avium subsp. paratuberculosis added to the raw milk at HRI was tested. Appropriate dilutions were prepared in MRD and inoculated onto HEYM slopes (250 μl per slope) to obtain a colony count (CFU/ml), and 500 μl of the 10⁻² dilution was inoculated into one vial of BACTEC medium to obtain a BACTEC count (cells/ml) using a previously published formula (18). The number of M. avium subsp. paratuberculosis cells present in the inoculated raw milk was assumed to be 1,000-fold less than the M. avium subsp. paratuberculosis count obtained, as 20 ml of culture had originally been diluted to 20 liters with raw milk. The target inoculum level for experiments 1, 2, 3, 6, and 7 was 10⁵ cells/ml. In experiments 4 and 5, 10⁻¹, 10⁻², and 10⁻³ dilutions of the M. avium subsp. paratuberculosis broth culture were added to the raw milk, so the initial counts in these inoculated raw milks were derived accordingly.

Other microbiological and biochemical analyses.

Microbiological and biochemical analyses to confirm effective pasteurization were carried out on the raw milk and all heat-treated samples on the same day as the heat treatment had taken place. Total viable counts were determined on milk agar after incubation aerobically at 30°C for 48 h, and the presence of coliforms was monitored on MacConkey agar after incubation at 37°C for 24 h. Potential postpasteurization contamination by gram-negative bacteria was assessed by a preincubated plate count whereby the sample was incubated at 21°C for 25 h followed by subsequent enumeration on milk agar at 21°C for 25 h. The incubation and enumeration stages were performed in the presence of crystal violet (2 mg/liter), penicillin (20,000 U/liter), and nisin (40,000 U/liter) to inhibit gram-positive bacteria (14). Confirmation of effective pasteurization by the plant was also established by milk biochemistry measurements. Alkaline phosphatase levels, in all raw milk and heat-treated samples, were determined by the CHARM method (CHARM Scientific Inc., Lawrence, Kan.). Confirmation that the treated milk had not been overheated was established by monitoring residual lactoperoxidase activity with 2,2′-azino-di-(3-ethylbenzthiazoline-6-sulfonic acid) as the chromogen.

Mastersizer experiments with broth cultures of M. avium subsp. paratuberculosis.

The size distribution of the cell clumps in Middlebrook broth cultures of M. avium subsp. paratuberculosis was measured by small-angle static light scattering using a Mastersizer X spectrometer (Malvern Instruments Limited, Malvern, Worcestershire, England). Comparative measurements on different strains were made using irradiation-inactivated cultures (treated with a dose of 10 kGy using a Gammabeam 650 irradiator). The effects of homogenization on the particle size distribution profile of the culture following high-pressure treatment with a Crepaco Multi-flo model 3DDL homogenizer were determined using viable broth cultures of M. avium subsp. paratuberculosis NCTC 8578. Particle-free filter-sterilized (0.22-μm porosity filter) MRD was used as the diluent.

Statistical analysis of results.

Results from all the experiments involving milk spiked with NCTC 8578 at a level of 10⁵ cells/ml were analyzed together. Given that the inoculum for each experiment was prepared in exactly the same manner, was used after 8 weeks incubation, and was added to 20 liters of milk in the same proportion, this was a statistically valid approach. The proportions of culture-positive milk samples recovered after each different processing combination (time, temperature, and homogenization setting) were subjected to statistical analysis by logistic regression analysis followed by pairwise comparisons (t test). Fewer processing runs were carried out with milk spiked at the lower inoculum levels and with the other M. avium subsp. paratuberculosis isolates, so this meant that there were insufficient data available to perform statistical analysis on the results of these experiments.

RESULTS

Verification of correct pilot plant operation.

The efficacy of pilot plant operation was confirmed by monitoring various microbiological and biochemical indicators of pasteurization immediately after each heat treatment (Table (Table2).2). The microbiological quality of the heat-treated milk produced by the plant was high, with total viable counts in the range of 10¹ to 10³ CFU/ml. Coliforms were not detected in any sample, and there was no evidence of gram-negative bacteria being present, which would have indicated postpasteurization contamination. Alkaline phosphatase levels in all heat-treated milk samples were <25 mU/liter, a value well below the threshold value for effective pasteurization of 350 mU/liter (CHARM method). Lactoperoxidase determinations confirmed that the milk had been correctly processed.

TABLE 2.

Microbiological and biochemical data for raw milk used in pasteurization trials and heat-treated milk produced by pilot-scale pasteurizer

Extent of M. avium subsp. paratuberculosis inactivation after various pasteurization treatments.

Of the 816 heat-treated 50-ml milk samples tested during this study, 27 (3.3%) were confirmed to contain viable M. avium subsp. paratuberculosis cells (i.e., subcultures of typical colonies from HEYM [n = 5 of 27] or from BACTEC [n = 22 of 27] tested IS900 PCR positive). Generally, only one or two of the triplicate 50-ml samples per treatment tested culture positive, which probably reflects the fact that survivors were present in low numbers and may not have been uniformly distributed throughout the milk due to clumping. Many of the M. avium subsp. paratuberculosis culture positives were isolated from milk samples that had been pasteurized at 72.5°C (11 of 27 positives), which is close to the legal minimum pasteurization temperature, or from pasteurized milk that had not been homogenized (13 of 27 positives). If we consider the five pasteurized milk samples which were culture positive on HEYM, in all five cases, only one of the triplicate 50-ml aliquots yielded any colonies, and in four of the five cases, only one of the duplicate HEYM slopes inoculated per 50 ml had a colony present. The number of M. avium subsp. paratuberculosis cells present in these five pasteurized milk samples was calculated to be 10 to 20 CFU/150 ml. As no chemical decontamination treatment was applied to the milk samples before culture, the number of M. avium subsp. paratuberculosis cells surviving heat treatment will not have been underestimated. By subtracting the colony counts after pasteurization from the number of M. avium subsp. paratuberculosis cells spiked into these milk samples before pasteurization, an estimate of the minimum log₁₀ kill achieved by HTST pasteurization was obtained. A 4.0- to 5.2-log₁₀ reduction in numbers of viable M. avium subsp. paratuberculosis cells was achieved in the case of these five milk samples, so it would follow that in instances where no viable M. avium subsp. paratuberculosis cells were isolated after pasteurization, which was the situation for 96.7% of pasteurized milk samples tested during this study, a >4- to 5-log₁₀ inactivation of M. avium subsp. paratuberculosis was being achieved by pasteurization.

All of the confirmed M. avium subsp. paratuberculosis isolates arose from milk originally spiked with NCTC 8578. Figure Figure11 provides an overview of the temperature, time, and homogenization conditions under which M. avium subsp. paratuberculosis NCTC 8578 was not completely inactivated. The vast majority (25 of 27) of the confirmed isolates were from milk originally spiked with 10⁵ cells/ml, and the other two were recovered from milk originally spiked with 10³ and 10¹ to 10² cells/ml. The other three M. avium subsp. paratuberculosis strains (806R, 796PSS, and B4) spiked into milk at levels of 10¹ to 10⁴ cells/ml were all completely inactivated by heating at 72.5 and 78.5°C for 25 s with upstream homogenization. However, it is important that these three strains were not tested over as wide a range of processing conditions as strain NCTC 8578 and attained a maximum inoculum level of only 10⁴ cells/ml (10-fold less than NCTC 8578 in most experiments). In particular, no processing runs without homogenization were carried out using milk inoculated with 806R, 796PSS, or B4. When the processing and inoculum level conditions after which the culture positives occurred during the present study are considered, it is clear that the majority of the culture positives arose after pasteurization of milk spiked with 10⁵ NCTC 8578/ml, at 72.5°C, and when the milk had not been homogenized. Two of these three parameters were not studied in the case of 806R, 796PSS, and B4, so there is no basis to conclude that NCTC 8578 is any more resistant than the other three M. avium subsp. paratuberculosis strains studied.

FIG. 1.

Comparison of the effects of no homogenization (A), upstream homogenization (B), and in-hold homogenization (C) in combination with pasteurization at temperatures between 72.5 and 84.5°C for hold times of 15, 25, and 60 s on inactivation of M. (more ...)

Influence of homogenization in combination with pasteurization on M. avium subsp. paratuberculosis inactivation.

Almost half of the M. avium subsp. paratuberculosis culture positives (12 of 27) were isolated from milk samples that had been pasteurized at 72.5°C, which is close to the legal minimum pasteurization temperature (72.5°C). As pasteurization temperature increased, the number of culture-positive samples tended to decline, although two of the M. avium subsp. paratuberculosis-positive milk samples were recovered after heat treatment at 82.5°C with upstream homogenization and the longer hold times (Fig. (Fig.1B).1B). Heat treatments without homogenization and those in combination with upstream homogenization yielded more of the culture-positive milk samples than heat treatments incorporating in-hold homogenization (13, 11, and 3 isolates, respectively). It was also noticeable that M. avium subsp. paratuberculosis-positive milk samples occurred across almost the entire range of temperatures studied when no homogenization (72.5 to 78.5°C) and upstream homogenization (72.5 to 78.5 and 82.5°C) were applied, whereas positive M. avium subsp. paratuberculosis cultures occurred only at 72.5 and 76.5°C after in-hold homogenization (Fig. (Fig.1C1C).

The beneficial effect of homogenization on M. avium subsp. paratuberculosis inactivation was more significant when a 25-s hold time was employed compared to the standard 15-s hold time (Fig. (Fig.2).2). Both in-hold homogenization and upstream homogenization in conjunction with a 25-s hold time produced significantly fewer culture-positive samples than when no homogenization in combination with a 25-s hold was applied (P < 0.001 and P < 0.05, respectively). In contrast, there was no significant difference between the different homogenization settings when a 15-s hold time was employed (P > 0.05). There was also no statistically significant difference between the number of culture-positive samples obtained after no homogenization/15-s and no homogenization/25-s heat treatments (P > 0.05).

FIG. 2.

Effect of holding time and homogenization treatment on the inactivation of M. avium subsp. paratuberculosis NCTC 8578 (10⁵ cells/ml) in milk processed at temperatures of 72.5 to 78.5°C. Bars indicate standard errors of the means.

Effect of extended hold time on M. avium subsp. paratuberculosis inactivation during pasteurization.

The effect of extended hold times of 25 s and 60 s on M. avium subsp. paratuberculosis inactivation was studied in conjunction with upstream homogenization and temperatures ranging from 72.5 to 84.5°C. Overall, no significant difference (P = 0.933) was found between the percentage of heat-treated milk samples containing viable M. avium subsp. paratuberculosis obtained after holding times of 15, 25, and 60 s (4.2%, 6.0%, and 6.3%, respectively) under the same temperature and homogenization conditions. Overall, no significant difference was found between 15-s and 25-s hold times in conjunction with any of the homogenization settings (Fig. (Fig.2,2, P > 0.05 in all cases).

Evaluation of clump size distribution in M. avium subsp. paratuberculosis suspensions and assessment of the impact of homogenization on clump size.

Small-angle static light scattering was employed to compare the clump size distributions in Middlebrook broth suspensions of three of the M. avium subsp. paratuberculosis inoculum strains. The profiles obtained revealed some differences in the clump size distributions of broth suspensions of irradiated cultures of NCTC 8578, 806R, and 796PSS (Fig. (Fig.3).3). Strain NCTC 8578 (which yielded some survivors after pasteurization during this study) appeared to have a greater proportion of large clumps than the other two strains (which were both completely inactivated under the pasteurization conditions studied). The proportion of particles under 25 μm in size in the broth cultures of NCTC 8578, 806R, and 796 PSS analyzed was 24.8%, 47.5%, and 40.6%, respectively.

FIG. 3.

Clump size distributions in Middlebrook 7H9 broth suspensions of three M. avium subsp. paratuberculosis strains determined using Mastersizer.

The effect of homogenization (2,500 lb/in²) on clump size distribution of NCTC 8578 was to reduce clump size to single-cell or “miniclump” (i.e., aggregates of a few cells) dimensions (size range, 1 to 10 μm). Virtually 90% of the cells were in this size range after homogenization (compare Fig. 4A and B). When the homogenized cell suspension was passed through the Mastersizer again after 10 min had elapsed, the clump size was observed to have increased again (range, 20 to 150 μm [Fig. [Fig.4C]).4C]). Similar shifts in clump size profiles were observed when a further broth suspension of NCTC 8578 was homogenized at 4,000 lb/in² and analyzed again using the Mastersizer after 5 min (data not shown). These findings suggest that M. avium subsp. paratuberculosis cells have a tendency to reaggregate into clumps within reasonably short periods of time after homogenization.

FIG. 4.

Clump size distribution in (A) an untreated Middlebrook 7H9 broth suspension of M. avium subsp. paratuberculosis NCTC 8578, (B) the same broth culture homogenized at 2,500 lb/in² and analyzed immediately, and (C) the same broth culture analyzed 10 min (more ...)

DISCUSSION

Over recent years, we have consistently observed, and reported, survival of low numbers of viable M. avium subsp. paratuberculosis cells in HTST pasteurized milk irrespective of whether we had used laboratory apparatus (13), Franklin heat exchange plates (10, 11, 13), or a commercial pasteurizer (8) to assess the impact of HTST pasteurization conditions on the viability of this potential human pathogen. The present study, an extension of our previous research, was carried out in an effort to identify a processing solution to the M. avium subsp. paratuberculosis survival issue by assessing the efficacy of higher pasteurization temperatures and longer holding times in combination with homogenization by using a pilot-scale pasteurizer comparable in characteristics to a commercial-scale plant. A further objective of the study was to try and elucidate the mechanism of survival of this organism by investigating the clumping characteristics of M. avium subsp. paratuberculosis.

This study has shown that low numbers of M. avium subsp. paratuberculosis cells (estimated to be around 10 to 20 CFU/150 ml) are capable of surviving pilot-scale HTST pasteurization of milk on occasion when stationary-phase organisms are present in raw milk in high numbers (10⁵ cells/ml) and/or as large clumps of cells. Surviving M. avium subsp. paratuberculosis cells were cultured from 27 (3.3%) of 816 pasteurized milk samples overall, the majority of which were isolated from liquid BACTEC cultures (22 of 27) rather than on HEYM slopes (5 of 27). This is consistent with previous findings, our own (10) and those of others (15, 19), that liquid culture media appear to be more conducive to the recovery of heat-treated M. avium subsp. paratuberculosis than solid-agar slopes where chemical decontamination has not been applied. All the evidence accumulated during the study suggests that the M. avium subsp. paratuberculosis isolates recovered from pasteurized milk were genuine isolates that had survived heat treatment, viz., that (i) gram-negative preincubated plate counts of the pasteurized milks were consistently negative throughout the study, indicating that there was no evidence of postprocess contamination; (ii) unspiked pilot plant flushes, included in several runs, all tested M. avium subsp. paratuberculosis negative, indicating that cross-contamination of the milk with viable M. avium subsp. paratuberculosis from the plant during processing had not occurred; and (iii) the large number (n = 72) of negative-control milk samples tested during experiment 8 all reported culture negative, confirming that culture-positive samples arose only from heat-treated spiked milk and not as a result of laboratory cross-contamination during sample processing. Where surviving M. avium subsp. paratuberculosis cells were encountered, the pasteurization time-temperature and homogenization combinations studied had achieved a 4.0- to 5.2-log₁₀ reduction in numbers of viable M. avium subsp. paratuberculosis cells. As there was no evidence of surviving M. avium subsp. paratuberculosis in the vast majority (96.7%) of pasteurized milk samples tested, HTST pasteurization achieved a greater than 5-log₁₀ reduction in numbers of this potential pathogen in most processing runs.

Heat treatments incorporating in-hold, but not upstream, homogenization were found to result in significantly fewer milk samples containing surviving M. avium subsp. paratuberculosis (P = 0.006 for in-hold homogenization; P = 0.089 for upstream homogenization) than heat treatments without homogenization of any kind. Only in conjunction with a 25-s hold time did the effect of upstream homogenization on M. avium subsp. paratuberculosis survival become statistically significant (P = 0.022) compared to a 25-s hold without homogenization. One previous study of the effect of commercial-scale HTST pasteurization on naturally infected milk assessed the impact of pasteurization at 73°C for 15 and 25 s with and without prior (upstream) homogenization (2,500 lb/in²) on M. avium subsp. paratuberculosis inactivation (8). The 25-s hold time was found to be no more effective in inactivating M. avium subsp. paratuberculosis than the conventional 15-s treatment, which is in agreement with the findings of the present study. Although a smaller number of M. avium subsp. paratuberculosis culture-positive samples were isolated from (upstream) homogenized pasteurized milk samples than from unhomogenized pasteurized milk samples, the effect of upstream homogenization on the survival of M. avium subsp. paratuberculosis was not found to be statistically significant overall (P = 0.19), which is also in agreement with the findings of the present study. In-hold homogenization has not been studied previously.

The pasteurization studies described here are directly comparable to three published studies involving pilot- or commercial-scale pasteurization plant and laboratory-grown M. avium subsp. paratuberculosis which were carried out in Germany (15), Australia (19), and New Zealand (23). The findings of the present study confirm those of the German and Australian studies, which both reported survival of low numbers of M. avium subsp. paratuberculosis cells after a variety of time-temperature conditions when milk containing 10¹ to 10⁵ CFU/ml of M. avium subsp. paratuberculosis was pasteurized. The findings of the New Zealand study (23) are in sharp contrast to those of the other three studies, since Pearce et al. (23) did not recover any survivors from pasteurized milk spiked with 10³ to 10⁴ CFU/ml before heating. In our opinion, differences in relation to the volume of milk tested and application of chemical decontamination directly after heating between the New Zealand study and the other studies mentioned above and our study may explain nonrecovery of any survivors by Pearce et al. (23). McDonald et al. (19) and Pearce et al. (23) applied chemical decontamination to the recently pasteurized milk samples prior to culture, whereas Hammer et al. (15) and we did not. Pearce et al. (23) tested only a single 50-ml milk sample per heat treatment, whereas McDonald et al. (19) tested 30 50-ml samples (i.e., 1.5 liters) per heat treatment. In our opinion, the testing of 1.5 liters of milk by McDonald et al. (19) possibly compensated for the detrimental effect of chemical decontamination on the viability of M. avium subsp. paratuberculosis (7) and enabled these researchers to detect low levels of this organism (1 CFU/500 ml to 1 CFU/250 ml) in batches of pasteurized milk processed at 72.5°C for 15 s, 75°C for 25 s, and 78°C for 15 s.

Hammer and coworkers (15) suggested three possible explanations for the ability of M. avium subsp. paratuberculosis to survive HTST pasteurization: (i) the influence of clumping, as survivors were observed to exist only within large clumps by means of an acid-fast viability stain (originally developed by Grant et al. [9]); (ii) the potential existence of dormant/metabolically inactive cells which were revived by resuscitation in liquid culture for several months; and (iii) possible heat activation of M. avium subsp. paratuberculosis cells at pasteurization temperatures >74°C for holding times longer than 33 s. During the present study, survival of M. avium subsp. paratuberculosis after an extended holding time of 60 s in conjunction with upstream homogenization was observed, and interestingly, all of the culture positives under these conditions were from milk samples pasteurized at temperatures of 76.5°C or above. This finding may lend some support to the suggestion of Hammer et al. (15) of possible heat activation of M. avium subsp. paratuberculosis at high temperatures and longer hold times.

Pasteurization time-temperature data collected during the recent United Kingdom milk survey provide an indication of the temperatures and hold times typically used for commercial pasteurization of milk in the United Kingdom (5). The vast majority (85%) of the pasteurized milk samples tested for the presence of M. avium subsp. paratuberculosis during the United Kingdom milk survey had been pasteurized at temperatures between 72.0°C and 74.9°C. The highest pasteurization temperature recorded for any sample was 77.9°C. There was more of a spread in terms of hold times used commercially: 49% of samples were pasteurized for 15 s, 18% were pasteurized for 16 to 24 s, 25% were pasteurized for 25 s, and 8% were pasteurized for longer than 25 s. Interestingly, some of the time-temperature and homogenization combinations that yielded viable M. avium subsp. paratuberculosis during the present study with spiked milk also yielded survivors when pasteurized naturally infected milk was tested during the United Kingdom milk survey (12). However, direct comparisons between these two studies are probably inadvisable, since stationary-phase cultures were used in the pilot plant trials, while the number and growth status of M. avium subsp. paratuberculosis cells present in the naturally infected raw milk before commercial pasteurization are unknown. The second objective of the present study was to investigate whether clumping may contribute to the heat resistance of M. avium subsp. paratuberculosis. Results obtained using small-angle static light scattering with the Mastersizer X spectrometer provided a valuable insight into the cell size distribution in suspensions of several M. avium subsp. paratuberculosis strains and the effect of homogenization on clumps of this organism. There is no doubt that laboratory-grown M. avium subsp. paratuberculosis exists in clumps, some of which are extremely large (despite the fact that Tween 80 is routinely added to Middlebrook broth to reduce clumping). When M. avium subsp. paratuberculosis cultures were tested using the Mastersizer following homogenization (2,500 or 4,000 lb/in²), clear evidence was obtained showing that clumps were effectively disrupted and that the vast majority of cells were 1 to 5 μm in size (i.e., single-cell or “miniclump” status). Given that pasteurization treatments incorporating homogenization (either upstream or in-hold configuration) resulted in greater inactivation of M. avium subsp. paratuberculosis during the present study, this finding lends support to our long-held hypothesis that clumping assists or aids survival and contributes to the apparent heat resistance of this organism. The data reported previously by Hammer et al. (15), obtained using the acid-fast viability staining technique, would also support the clumping hypothesis.

Mastersizer experiments also demonstrated that reaggregation of M. avium subsp. paratuberculosis cells occurred reasonably quickly (within 5 to 10 min) after disruption, particularly if the cells were mixed. Earlier laboratory experiments during this study indicated that reaggregation could occur in less than 5 min. When M. avium subsp. paratuberculosis suspensions were declumped by vortexing with glass beads and then left for 2 min with and without agitation, M. avium subsp. paratuberculosis colony counts were shown to increase upon declumping but reverted to original levels when subsequently mixed for 2 min but not when left static for 2 min. These observations provide a possible explanation for the greater effectiveness of treatments incorporating in-hold as opposed to upstream homogenization. In treatments involving in-hold pasteurization, the milk is homogenized at pasteurization temperature at the start of the holding section, and so as soon as M. avium subsp. paratuberculosis clumps are disrupted, the heat is able to act on individual cells. In contrast, upstream homogenization is carried out as a separate unit operation, and milk must flow through pipework to reach the pasteurizer, being mixed as it goes, potentially providing the opportunity for some M. avium subsp. paratuberculosis cells to reaggregate into clumps between homogenization and heating and hence reducing the beneficial effects of homogenization.

In conclusion, the present study was carried out to identify a solution to the problem of M. avium subsp. paratuberculosis surviving HTST pasteurization of milk by using technology existing and readily available within the United Kingdom dairy industry. This study has shown that adjustment of pasteurization temperature or holding time alone will not necessarily ensure complete inactivation of M. avium subsp. paratuberculosis. Rather, results indicate that pasteurization needs to be combined with a process such as homogenization in order to maximize the log₁₀ kill of M. avium subsp. paratuberculosis achieved. Homogenization was shown to break up clumps of M. avium subsp. paratuberculosis cells, and thus, cells would potentially be more exposed to the lethal effects of subsequent heat treatment. Where and how homogenization is applied (as a separate process or within the holding section of the pasteurizer) may be important determinants of inactivation. Upstream homogenization is currently used in liquid milk processing within the United Kingdom dairy industry, although most high-throughput commercial plants do not homogenize whole milk but separate the milk into cream and skim milk streams, homogenize the cream fraction, and then recombine it with the skim milk. In-hold homogenization has been employed in the past, but problems of protein precipitation (visible as deposits on the inside surface of glass bottles) led to its use being discontinued some years ago. The results of this study strongly suggest that where homogenization of whole milk is carried out, homogenization at pasteurization temperature (i.e., in-hold homogenization) should be revisited by the United Kingdom dairy industry in order to maximize M. avium subsp. paratuberculosis inactivation, unless some other physical process (such as centrifugation or microfiltration) is employed to reduce M. avium subsp. paratuberculosis levels in raw milk prior to pasteurization.

Acknowledgments

This study formed part of a Department of Environment, Food, and Rural Affairs (DEFRA) LINK project (FQS14) jointly funded by DEFRA, the Dairy Industry Association Limited, the Milk Development Council, and the Scottish Executive Environment and Rural Affairs Department (SEERAD) under the Food Quality and Safety LINK program.

Thanks are due to Susan Withers and Ian West for expert technical contribution, to members of the analytical group at HRI for their skilled assistance, and to members of the United Kingdom dairy industry for technical advice in relation to setting up the pilot pasteurization plant.

REFERENCES

1. Amor, D. 1998. Why milk needs reassessing. Food Ind. News 1998:10.

2. Chiodini, R. J., and J. Hermon-Taylor. 1993. The thermal resistance of Mycobacterium paratuberculosis in raw milk under conditions simulating pasteurisation. J. Vet. Diagn. Investig. 5:629-631. [PubMed]

3. Dundee, L., I. R. Grant, H. J. Ball, and M. T. Rowe. 2001. Comparative evaluation of four decontamination protocols for the isolation of Mycobacterium avium subsp. paratuberculosis from milk. Lett. Appl. Microbiol. 33:173-177. [PubMed]

4. European Commission. 2000. Possible links between Crohn's disease and paratuberculosis. Report of the Scientific Committee on Animal Health and Animal Welfare. European Commission, Brussels, Belgium. [Online.] http://www.europa.eu.int/comm/food/fs/sc/scah/out38_en.pdf.

5. Food Standards Agency. 2003. Report on the national study on the microbiological quality and heat processing of cows' milk. Food Standards Agency, London, England. [Online.] http://www.foodstandards.gov.uk/multimedia/pdfs/milksurvey.pdf.

6. Gao, A., L. Mutharia, S. Chen, K. Rahn, and J. Odumeru. 2002. Effect of pasteurisation on survival of Mycobacterium paratuberculosis in milk. J. Dairy Sci. 85:3198-3205. [PubMed]

7. Grant, I. R., and M. T. Rowe. 2004. Effect of chemical decontamination and refrigerated storage on the recovery of Mycobacterium avium subsp. paratuberculosis from heat-treated milk. Lett. Appl. Microbiol. 38:283-288. [PubMed]

8. Grant, I. R., E. Hitchings, A. McCartney, F. Ferguson, and M. T. Rowe. 2002. Effect of commercial-scale high-temperature, short-time pasteurization on the viability of Mycobacterium paratuberculosis in naturally infected cows' milk. Appl. Environ. Microbiol. 68:602-607. [PMC free article] [PubMed]

9. Grant, I. R., H. J. Ball, and M. T. Rowe. 1997. A novel staining technique for assessing clumping and viability of Mycobacterium paratuberculosis cells during pasteurisation, p. 374-377. In R. J. Chiodini, M. E. Hines, and M. T. Collins (ed.), Proceedings of the Fifth International Colloquium on Paratuberculosis. International Association for Paratuberculosis, Rehoboth, Mass.

10. Grant, I. R., H. J. Ball, and M. T. Rowe. 1998. Effect of high temperature, short time (HTST) pasteurisation on milk containing low numbers of Mycobacterium paratuberculosis. Lett. Appl. Microbiol. 26:166-170. [PubMed]

11. Grant, I. R., H. J. Ball, and M. T. Rowe. 1999. Effect of higher pasteurisation temperatures, and longer holding times at 72°C, on the inactivation of Mycobacterium paratuberculosis in milk. Lett. Appl. Microbiol. 28:461-465. [PubMed]

12. Grant, I. R., H. J. Ball, and M. T. Rowe. 2002. Incidence of Mycobacterium paratuberculosis in bulk raw and commercially pasteurized cows' milk from approved dairy processing establishments in the United Kingdom. Appl. Environ. Microbiol. 68:2428-2435. [PMC free article] [PubMed]

13. Grant, I. R., H. J. Ball, S. D. Neill, and M. T. Rowe. 1996. Inactivation of Mycobacterium paratuberculosis in cow's milk at pasteurization temperatures. Appl. Environ. Microbiol. 62:631-636. [PMC free article] [PubMed]

14. Griffiths, M. W., J. D. Philips, and D. D. Muir. 1984. Rapid detection of post-pasteurization contamination. Hannah Res. Inst. Bull. 10:14-27.

15. Hammer, P., C. Kiesner, H.-G. Walte, K. Knappstein, and P. Teufel. 2002. Heat resistance of Mycobacterium avium ssp. paratuberculosis in raw milk tested in a pilot plant pasteurizer. Kieler Milchwirtschaftliche Forschungsberichte 54:275-303.

16. Hasting, A. P. M., C. D. W. Blackburn, and J. S. Crowther. 2001. Mycobacterium paratuberculosis and the commercial pasteurisation of milk. TransIChemE 79(C):83-88.

17. Hope, A. F., P. A. Tulk, and R. J. Condron. 1996. Pasteurisation of Mycobacterium paratuberculosis in whole milk, p. 377-382. In R. J. Chiodini, M. E. Hines and M. T. Collins (ed.), Proceedings of the Fifth International Colloquium on Paratuberculosis. International Association for Paratuberculosis, Rehoboth, Mass.

18. Lambrecht, R. S., J. F. Carriere, and M. T. Collins. 1988. A model for analyzing the growth kinetics of a slowly growing Mycobacterium sp. Appl. Environ. Microbiol. 54:910-916. [PMC free article] [PubMed]

19. McDonald, W. L., K. J. O'Riley, C. J. Schroen, and R. J. Condron. 2003. Heat inactivation of Mycobacterium avium subsp. paratuberculosis in milk, p. 312-316. In R. A. Juste, M. V. Geijo, and J. M. Garrido (ed.), Proceedings of the Seventh International Colloquium on Paratuberculosis. International Association for Paratuberculosis, Madison, Wis.

20. Meylan, M., D. M. Rings, W. P. Shulaw, J. J. Kowalski, S. Bech-Nielsen, and G. F. Hoffsis. 1996. Survival of Mycobacterium paratuberculosis and preservation of immunoglobulin G in bovine colostrum under experimental conditions simulating pasteurisation. Am. J. Vet. Res. 57:1580-1585. [PubMed]

21. National Institute for Allergy and Infectious Diseases. 1998. Crohn's disease—is there a microbial etiology? Recommendations for a research agenda. Report of a meeting held at National Institutes of Health, Washington, D.C., 14 December 1998. National Institute for Allergy and Infectious Diseases, National Institutes of Health, Washington, D.C. [Online.] http://www.niaid.nih.gov/dmid/meetings/crohns.htm.

22. National Research Council. 2003. Johne's disease and Crohn's disease, p. 104-120. In Diagnosis and control of Johne's disease. Report of the Committee on Diagnosis and Control of Johne's Disease for the National Academy of Science. The National Academies Press, Washington, D.C.

23. Pearce, L. E., H. T. Truong, R. A. Crawford, G. F. Yates, S. Cavaignac, and G. W. De Lisle. 2001. Effect of turbulent-flow pasteurization on survival of Mycobacterium avium subsp. paratuberculosis added to raw milk. Appl. Environ. Microbiol. 67:3964-3969. [PMC free article] [PubMed]

24. Rubery, E. 2001. A review of the evidence for a link between exposure to Mycobacterium paratuberculosis (MAP) and Crohn's disease (CD) in humans. A report for the United Kingdom Food Standards Agency, June 2001. Food Standards Agency, London, England. [Online.] http://www.foodstandards.gov.uk/multimedia/pdfs/MAPcrohnreport.pdf.

25. Stabel, J., L. Pearce, R. Chandler, P. Hammer, N. Klijn, O. Cerf, M. T. Collins, C. Heggum, and P. Murphy. 2001. Destruction by heat of Mycobacterium paratuberculosis in milk and milk products. Bull. Int. Dairy Fed. 362:53-60.

26. Streeter, R. N., G. F. Hoffsis, S. Bech-Nielsen, W. P. Shulaw, and M. Rings. 1995. Isolation of Mycobacterium paratuberculosis from colostrum and milk of subclinically infected cows. Am. J. Vet. Res. 56:1322-1324. [PubMed]

27. Sung, N., and M. T. Collins. 1998. Thermal tolerance of Mycobacterium paratuberculosis. Appl. Environ. Microbiol. 64:999-1005. [PMC free article] [PubMed]

28. Sweeney, R. W., R. H. Whitlock, and A. E. Rosenberger. 1992. Mycobacterium paratuberculosis cultured from milk and supramammary lymph nodes of infected asymptomatic cows. J. Clin. Microbiol. 30:166-171. [PMC free article] [PubMed]

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