If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Shelf-stable milk is consumed worldwide, and this market is expected to continue growing. One quality challenge for UHT milk is age gelation during shelf life, which is in part caused by bacterial heat-stable proteases (HSP) synthesized during the raw milk storage period before heat processing. Some Pseudomonas spp. are HSP producers, and their ability to grow well at refrigeration temperature make them important spoilage organisms for UHT processors to control. Previous studies have shown that lactose oxidase (LO), a natural and commercially available enzyme that produces hydrogen peroxide and lactobionic acid from lactose, can control bacterial growth in raw milk. In this research, we investigated the ability of LO to control HSP producer outgrowth, and thus delay age gelation in UHT milk. Six strains of Pseudomonas spp. were selected based on their ability to synthesize HSP and used as a cocktail to inoculate both raw and sterile (UHT) milk at a level of 1 × 105 cfu/mL. Groups were treated with and without LO, stored for 4 d at 6°C, and monitored for cell count and pH. Additionally, a sample from each was tested for HSP activity via particle size analysis (average effective diameter at 90° angle and 658 nm wavelength) and visual inspection on each day of the storage period. The HSP activity results were contrasted using Tukey's HSD test, which showed that in UHT milk, a LO treatment (0.12 g/L) effectively prevented gelation as compared with the control. In raw milk, however, a concentration of 0.24 g/L of LO was needed to obtain a similar effect. This test was scaled up to 19-L pilot plant batches of raw milk where they were challenged with Pseudomonas cocktail, treated with LO for 3 d, and then UHT processed. Resulting UHT milk bottles were monitored for gelation. Significant differences in particle size between the LO-treated samples and the control were observed as early as 1 mo after processing, and gelation was not detected in the LO-treated samples through 6 mo of storage. These results demonstrated that LO can be used to delay age gelation in UHT milk induced by HSP-producing Pseudomonas spp., representing an opportunity to improve quality and reduce postproduction losses in the shelf-stable milk market sector.
Globally, the UHT milk market is growing steadily, as consumers continue to demand safe, nutritious, and convenient products with extended shelf life (
). From an economic point of view, the manufacture of UHT milk provides some inherent advantages because its distribution does not require refrigeration, which consequently lowers energy costs (
) and facilitates commercialization in markets where refrigerated transportation and storage are challenging. A market analysis and forecast performed by Knowledge Sourcing Intelligence LLP (
) reported that the global UHT milk market is expected to reach almost 125 million tons by 2024, increasing at a compound annual growth rate of 4.57% from 2019 to 2024. However, despite this product's microbiological stability, its shelf life can be threatened by an irreversible phenomenon called “age gelation,” which can cause financial losses for dairy producers.
first described age gelation as the increase in viscosity of a liquid UHT-processed product that sometimes occurs during storage; this process leads to the loss of fluidity and, ultimately, the formation of a gel. It is believed that the triggering factor responsible for this phenomenon is the proteolytic activity of endogenous milk plasmin, or the activity of extracellular proteases produced by psychrotrophic bacteria, of which some such proteases remain active after UHT processing (
concluded that the rate at which age gelation occurs mainly depends on the extent of bacterial growth before heat treatment. Furthermore, various studies have reported that synthesis of bacterial HSP starts during the late exponential or early stationary phases of growth when counts are approximately between 7 and 8 log cfu/mL (
). These findings emphasize the importance of low microbial counts in raw milk destined for UHT milk production, as this will minimize the effects of HSP during subsequent shelf life.
Current guidelines for the handling, transportation, and storage of grade “A” raw milk bound for ultra-pasteurization dictate that its temperature is to be maintained below 7°C until the milk is processed (
in: Section 7. Standards for grade “A” milk and/or milk products. U.S. Department of Health and Human Services, Public Health Service,
Washington, DC2017: 34
). This attribute makes species of Pseudomonas important spoilage organisms for UHT processors to control. During preprocessing operations, not only do Pseudomonas increase their concentration, but also their potential for protease production. Based on the work of
, it can be concluded that by the end of milk collection and transportation operations, any attendant Pseudomonas will have already completed the lag phase of growth and will be in a position to enter the exponential phase, with potential for concomitant production of HSP during the plant's standard holding time before UHT treatment. Thus, holding time in dairy facilities is a critical factor that will have a major influence on the quality of the final product; if no additional preventative measures are added to the conventional regimen, there is high potential that the resulting UHT milk will be subject to age gelation during shelf life. Nevertheless, raw milk exhibits a natural mechanism that can be used to control the development of HSP-producing Pseudomonas during this stage.
The lactoperoxidase system (LPOS) in milk is an endogenous antimicrobial system that contributes to raw milk preservation. It has 3 main components: lactoperoxidase, thiocyanate, and hydrogen peroxide (H2O2). In this system, lactoperoxidase catalyzes the oxidation of thiocyanate by H2O2, generating compounds such as hypothiocyanite ions, which ultimately exhibit an antimicrobial effect (
). The efficacy of the LPOS depends on the concentration of thiocyanate and H2O2. Although endogenous thiocyanate can be present in close-to-optimal concentrations, depending on animal breed, udder health, and type of feed (
Lund, M., C. L. Nikolajsen, and J. M. van den Brink, inventors. 2019. Use of cellobiose oxidase for reduction of Maillard reaction. World Intellectual Property Organization. Pat. No. WO 2019/110497 A1.
Lund, M., C. L. Nikolajsen, and J. M. van den Brink, inventors. 2019. Use of cellobiose oxidase for reduction of Maillard reaction. World Intellectual Property Organization. Pat. No. WO 2019/110497 A1.
explored the effects of LO as an activator of the LPOS when used against a strain of P. fragi in different scenarios; their findings served as a reference for the present study.
The operational conditions of LO draw attention to its suitability as a processing aid, especially where medium to high temperatures are applied. With regard to milk that will ultimately be UHT-processed, LO can exert its function in the early stages of processing, and will subsequently be inactivated during preheating operations, avoiding the need for ingredient declaration. Treating enzymes that are inactivated as processing aids in foods is common practice in different processes, such as those of the baking industry (
In this study, we investigated the suitability of LO as a processing aid in high-quality UHT milk production. The main objective was to assess the capability of LO as an LPOS activator in raw milk during cold storage in dairy facilities to minimize HSP synthesis and potentially delay age gelation in UHT milk. Our approach consisted of first selecting Pseudomonas strains based on their ability to synthesize HSP, then exploring the effect of LO on these strains in the absence of background microflora (LO + UHT milk matrix) as well as in raw milk (LO + raw milk matrix), concluding with scaling-up and shelf stability assessment.
MATERIALS AND METHODS
Selection of Pseudomonas spp
Twenty-eight strains of Pseudomonas spp., all isolated from dairy facilities and products and obtained from the Cornell University Food Safety Laboratory, were screened for their ability to synthesize HSP. The screening process consisted of 2 phases: 1 for proteolytic activity, 1 for protease heat stability. First, cultures stored at −80°C were streaked onto Brain Heart Infusion (BHI) agar (BD Diagnostics, Franklin Lakes, NJ) and incubated for 24 h at the appropriate temperature for each species. Then, single colonies were streaked onto Milk Agar medium consisting of 100 g of skim milk powder (Difco, BD, Sparks, MD) and 15 g of agar (Difco, BD) per liter and were incubated between 6°C and 21°C for up to 14 d (
). Details on incubation settings can be found in the Supplemental Table S1 (https://doi.org/10.7298/rakk-ry78). The presence or absence of a clear zone around the colonies was a positive or negative indication of protease production, respectively (Supplemental Figure S1a).
Only strains that showed positive results at this stage were tested for protease heat stability. Testing was carried out using the method proposed by
with slight modifications. Single colonies from Milk Agar plates were inoculated into 50 mL of UHT skim milk (Parmalat USA Corp., Grand Rapids, MI) and incubated in an orbital shaker at 200 rpm (refer to Supplemental Table S1). After the incubation period, 20 mL was collected from each microbial culture and analyzed for their thermoresistant protease activity (TRPA), using the following steps.
Analysis of TRPA
Samples were centrifuged at 12,000 × g for 10 min to obtain cell-free supernatants (
). To test the heat stability of potential bacterial proteases without the interference of proteolysis induced by native plasmin, the supernatants were treated at 95°C for 8.45 min (
), and then cooled quickly using an ice bath. After that, each one was mixed with a 1% (wt/vol) solution of sodium azide (Sigma-Aldrich, St. Louis, MO) for a final concentration of 0.1% (
). Positive controls were also prepared; each consisted of a 1:1 solution of unheated supernatant with 1% sodium azide. For each supernatant, 2 samples and 1 positive control were individually mixed with UHT whole milk (Parmalat USA Corp.) in a 1:2 proportion (1 mL of heated or unheated supernatant with sodium azide and 2 mL of UHT milk in a 15-mL test tube); UHT whole milk was used as a negative control.
Finally, coagulation assessment was carried out via visual inspection at the end of 14 d at the appropriate temperature (Supplemental Table S1, https://doi.org/10.7298/rakk-ry78). Refer to Supplemental Figure S1b for examples of positive proteolytic activity versus no activity. Experiments were done in triplicate unless stated otherwise.
Application of LO in UHT Milk
We tested the ability of LO to control the growth of HSP producers, as well as the time needed for protease synthesis during cold storage. For that purpose, a cocktail of HSP producers was prepared as follows: strains were streaked onto BHI agar and incubated for 24 h at 25°C (P. fluorescens: 30°C); then, single colonies were inoculated into 5 mL of BHI broth (BD Diagnostics) to prepare overnight cultures (18 h at 25°C with shaking at 200 rpm); finally, optical density at 600 nm was measured for all overnight cultures to determine the appropriate volume needed from each to obtain a cocktail with approximately equal amounts of colony-forming units per strain.
After cocktail preparation, two 250-mL polypropylene conical tubes (Corning Inc., Corning, NY), each containing 150 mL of UHT skim milk, were inoculated to reach an initial concentration of 1 × 105 cfu/mL. One conical was used as a control; we added an enzymatic solution of LO (LactoYield, Chr. Hansen, Milwaukee, WI) to the other for a final concentration of 0.12 g/L (
), equivalent to 40 U of LO/kg of lactose. Both were kept at 6°C for 4 d to simulate cold storage conditions. Cell concentration and pH were monitored every 12 h; cell concentration was obtained via total plate count analysis doing spread plating on Standard Plate Count agar (Difco, BD Diagnostics, Franklin Lakes, NJ), followed by a 24-h incubation at 25°C. Every 24 h, 20 mL was collected for analysis of TRPA, as described in the previous section. Coagulation assessment was performed over a period of 4 wk during which the samples were held at 30°C. To complement visual inspection, particle size analysis via dynamic light scattering was also conducted as described below.
Particle Size Analysis
Particle size measurements were conducted using a 90Plus Nanoparticle Size Analyzer with a Peltier temperature control system (Brookhaven Instruments Corporation, Holtsville, NY). Analyses were performed at a fixed 90° angle with a laser wavelength of 658 nm at a constant temperature of 20°C; the sample temperature was allowed to equilibrate before each measurement. Milk samples were diluted to 3:1,000 using reverse osmosis water as a solvent to keep the signal intensity of the instrument above 400 kilocounts per second for all runs. The presented data correspond to the average effective diameters in nm. Finally, on d 4 of the storage period, both the control and LO treatments were exposed to heat (95°C for 8.45 min) and stored at 25°C for 2 additional days to monitor LO activity by means of pH change.
Application of LO in Raw Milk
To observe the effects of LO on HSP producers in raw milk, 100 mL of raw milk (Cornell Dairy, Ithaca, NY) was inoculated with the Pseudomonas cocktail to reach an initial concentration of 1 × 105 cfu/mL; LO was then added at 0.12 g/L, 0.24 g/L, or 0.60 g/L. Additionally, sodium thiocyanate (NaSCN; VWR International, Solon, OH) was tested in combination with all treatments of LO at a final concentration of 14 mg/L. The NaSCN was prepared according to
as a 32.7% (wt/vol) solution, using a solvent water filtered through a Milli-Q Advantage A10 system (MilliporeSigma, Burlington, MA); the solution was filter-sterilized through a 0.20-µm pore, surfactant-free cellulose acetate filter (Corning Inc.). All treatments were then kept at 6°C for 4 d, and cell concentration and pH were analyzed every 24 h. Raw milk, inoculated with the Pseudomonas cocktail but without the addition of either LO or NaSCN, was used as a control for cell concentration, and the raw milk alone was used as a control for pH. Furthermore, 20-mL samples were collected from these raw milk cultures on d 3 and 4 for TRPA, executed according to the procedure described previously and culminating in coagulation assessment via visual inspection over a period of 5 wk, during which the samples were held at 30°C.
Scaling up and Shelf Stability
Based on the results seen in raw milk, the most effective LO concentration was then scaled up to 18.9-L (5-gallon) batches held in 20-L carboys (Nalgene Nunc International, Rochester, NY). For this instance, LO [0.24 g/L] was tested against raw milk with the above-mentioned Pseudomonas cocktail. Raw milk containing the Pseudomonas cocktail but no LO was used as a positive control, and raw milk alone was used as a negative control. The inoculation was set to 1 × 105 cfu/mL, and carboys were stored at 6°C in a Thermo Scientific Forma Environmental Chamber (Thermo Fisher Scientific, Waltham, MA). Cell concentration and pH samples were collected every 24 h for 3 d. After the cold storage period, UHT treatment took place at the Cornell Food Venture Center Pilot Plant (Cornell AgriTech, Geneva, NY) using a MicroThermics Development-S series unit (MicroThermics Inc., Raleigh, NC). The run order for each group was randomized in all replicates, and steam sterilization was performed between runs at 121°C for 20 min. Parameters for UHT processing were set following the
method with some modifications: preheating was done at 90°C for 30 s, followed by homogenization at 20.7 MPa (3,000 psi), and finally, UHT treatment at 140°C for 4 s. After that, milk treatments were aseptically aliquoted into sterile 500-mL clear polyethylene terephthalate bottles (Corning Inc.) and stored at 25°C (for pH, particle size, and visual assessment) and at 30°C (for particle size and visual assessment) for 6 mo. Microbial concentration was checked immediately after the UHT treatment to verify absence of microbial contamination via total plate count analysis.
Statistical Analysis
Proteolytic activity and visible coagulation data are presented for descriptive purposes. Experiments were carried out in triplicate, unless stated otherwise, and percentage of response was computed based on the number of samples that exhibited the reported result as a fraction of the total that were run for that analysis. Inferential statistical tests were performed with JMP software (SAS Institute, Cary, NC) with significance level set to 0.05. Student's t-test was used for the comparison between 2 means on the same time point, and “Each Pair, Student's t” function was used for nonindependent datapoints in longitudinal data analysis. Analysis of variance and Tukey's honest significant difference test were used to compare differences among means of several treatments at the same time point. All experiments were done in triplicate and error bars were generated representing the standard deviation of the mean. Cell concentration and particle size data were log-transformed and analyzed using a linear scale.
RESULTS AND DISCUSSION
Production of HSP Identified in Strains of P. fluorescens, P. lundensis, and P. fragi
Table 1 presents the results of the process of screening various Pseudomonas spp. in the search for HSP producers. Eleven strains belonging to P. fluorescens were tested; all of them presented both proteolytic activity and protease heat stability. The relevance of P. fluorescens as the main contributor to milk spoilage has been studied widely over the years (
showed that various strains of P. fluorescens demonstrated extracellular HSP production, and that their proteases kept up to 50% of their activity at 25°C. This is of particular importance because not only is P. fluorescens a frequent source of HSP in milk, but it is also the case that UHT milk is stored at room temperature.
Table 1Proteolytic activity of Pseudomonas spp.; genus and species identities were obtained from the Food Microbe Tracker database (www.foodmicrobetracker.com;
Proteolytic activity, hydrolysis of milk proteins (skim milk agar, n = 3): “− − −” = no activity detected in any replicate, no halo around colonies; “+ + +” = activity detected in all replicates, halo around colonies.
Protease heat stability (at 95°C for 8.5 min), coagulation of milk proteins (UHT whole milk, n = 3): “− − −” = no activity detected in any replicate, no coagulation; “+ + +” = activity detected in all replicates, evident coagulation.
Isolate does not display a significant 16S rDNA sequence similarity (i.e., <99.9%) to any strain.
+ + +
+ + +
Heat-treated 2% fat milk
1 Cornell University Food Safety Laboratory (FSL) isolate designation.
2 Proteolytic activity, hydrolysis of milk proteins (skim milk agar, n = 3): “− − −” = no activity detected in any replicate, no halo around colonies; “+ + +” = activity detected in all replicates, halo around colonies.
3 Protease heat stability (at 95°C for 8.5 min), coagulation of milk proteins (UHT whole milk, n = 3): “− − −” = no activity detected in any replicate, no coagulation; “+ + +” = activity detected in all replicates, evident coagulation.
4 Genus and species identification based on 16S rDNA, Food Microbe Tracker database.
5 Isolates with 16S rDNA sequence matching (with 99.1% similarity) that of a strain of P. fragi.
6 Isolates with 16S rDNA sequence matching (with 98.7% similarity) that of a strain of P. lundensis.
7 Isolate does not display a significant 16S rDNA sequence similarity (i.e., <99.9%) to any strain.
8 Results from 5 replicates.
9 NA = nonapplicable; strain not tested for protease heat stability.
10 Strain was tested for protease heat stability in 2 out of 5 replicates, showing no coagulation.
Seasonal influence on heat-resistant proteolytic capacity of Pseudomonas lundensis and Pseudomonas fragi, predominant milk spoilers isolated from Belgian raw milk samples.
studied the microbiotas in Belgian raw milk samples and identified P. fragi and P. lundensis as the main contributors to Belgian milk spoilage. In our study, most P. fragi strains exhibited low proteolytic activity, which was the case for strains E2–0614, E2–615, and E2–616; these strains showed very little proteolytic activity in the first stage of the screening process; therefore, they were not tested for protease heat stability. Likewise, the proteolytic activity of W7–0098 and M7–0328 was minimal, to the point that only 2 out of 5 replicates showed a barely noticeable halo around the colonies; the other 3 replicates showed no halo at all. Based on our methodology, this implied that their proteolytic activity was negligible. On the other hand, P. fragi strains P4–0823 and P4–0824 showed effective proteolytic activity and heat stability, such as those demonstrated by strains of P. fluorescens. Finally, regarding strains identified as P. lundensis, all engaged in efficient HSP production.
Our observations indicated that the assayed strains of P. fluorescens and P. lundensis showed the most efficient proteolytic activity, a result that was consistent in all replicates. We also noted that the HSP of some Pseudomonas spp. were more powerful than others; that is, the coagulation effect was noticeable sooner and to a greater extent. For this reason and considering the natural diversity of raw milk's microbiotas, we considered it appropriate to assess the effects of LO on the level of HSP production exhibited by a cocktail comprising strains of multiple Pseudomonas species isolated from multiple sources. Test tubes were constantly monitored during the visible coagulation test, which facilitated the selection of strains based on the speed and degree of coagulation (data not shown). As a result, strains P4–0824, P4–0825, W5–0203, W5–0325, M7–0606, and M7–0636 were chosen for analysis in subsequent experiments.
This study provides a methodology for screening TRPA in Pseudomonas spp., minimizing potential interference from plasmin activity (
); however, the conditions of this methodology are amenable to further optimization with regard to standardization of incubation times and temperatures, as well as standards by which results may be assessed qualitatively.
Lactose Oxidase Inhibition of Pseudomonas-Induced Gelation
Experiments in UHT milk showed that in the absence of background microbiota, LO [0.12 g/L] was able to control the growth of the Pseudomonas cocktail inoculum, as presented in Figure 1. Cell counts of the sample treated with LO proved to be significantly lower than those of the control as early as 12 h after inoculation (P = 0.016). Moreover, even though the control reached cell counts greater than 9 log cfu/mL upon 4 d of storage, the LO-treated sample consistently stayed around 3 log cfu/mL.
Figure 1Cell counts of Pseudomonas spp. in UHT milk treated with lactose oxidase (LO) during storage at 6°C. Error bars represent the SD of 3 biological replicates.
, bacteriostatic activity is defined as a reduction of 3 log or less in the number of viable cells following exposure to an antimicrobial agent; based on this definition, our results suggested that LO was able to exert a bacteriostatic effect on the Pseudomonas cocktail. These results are consistent with
, who reported the effects of LO [0.12 g/L] on a strain of P. fragi in skim milk at 6°C; their findings showed that microbial counts were significantly different starting on d 1. Throughout the ensuing 7-d storage period, these counts consistently remained below the ones taken immediately following inoculation. This microbial inhibition was assumed to be due to H2O2 production. Despite the ability of Pseudomonas spp. to produce catalase and protect itself from H2O2, this self-defense mechanism appeared to be insufficient to neutralize it entirely. The results of
showed that a strain of Pseudomonas aeruginosa was unable to survive a 50 mM H2O2 dose, resulting in a 3.5-log reduction over a period of 1 h in spite of a 50% increase in catalase activity.
Figure 2 presents the effect of LO on UHT milk pH. At the end of the 4-d storage period, LO treatment resulted in a significantly lower pH than that of the control (P = 0.015). In the presence of LO, pH decreased at an average rate of 0.08 units/d, resulting in an overall decrease of 0.24 and 0.32 units on d 3 and 4, respectively. The observed drop in pH positively indicated that under the conditions of initial pH and storage temperature of the treatment, LO is actively engaged in the enzymatic production of LBA. Even though the control also underwent a decrease in pH over time (∼0.07 units/d during exponential growth), this was a consequence of Pseudomonas spp. metabolism (
Figure 2pH of UHT milk inoculated with Pseudomonas spp. and treated with lactose oxidase (LO) during storage at 6°C. Datapoints shown after d 4 represent pH following the application of heat (dotted vertical line). Error bars represent the SD of 3 biological replicates.
Following the application of heat, pH remained relatively constant during subsequent readings (Figure 2); there were no significant differences between the time point occurring before the heat treatment (d 4) and any of the following measurements, presumably due to the inactivation of LO. According to
Lund, M., C. L. Nikolajsen, and J. M. van den Brink, inventors. 2019. Use of cellobiose oxidase for reduction of Maillard reaction. World Intellectual Property Organization. Pat. No. WO 2019/110497 A1.
, the residual activity of LO becomes negligible at temperatures above 80°C. This indicates that LO can effectively exert its inhibitory action in milk during the cold storage period, with no subsequent deleterious effects on the quality of the milk following the preheating stage of UHT processing, as the LO would be inactivated during this step.
We used particle size as a proxy for the detection of initial stages of gelation before it could be seen macroscopically; these results are shown in Figure 3.
studied the influence of storage temperature in UHT milk age gelation and concluded that milk's resistance to gelation was the lowest at 30°C; therefore, experiments that assessed HSP's activity were performed at such temperature to accelerate the phenomenon.
Figure 3Particle size as an indication of heat-stable protease activity in UHT milk inoculated with Pseudomonas spp. and treated with lactose oxidase (LO). Storage time represents the day on which samples were collected, processed, and incubated at 30°C for analysis of thermoresistant protease activity and coagulation assessment. The UHT whole milk was used as a negative control. Error bars represent the SD of 3 biological replicates.
Based on samples collected on different days of the storage period, only the ones that came from the control treatment on d 3 and 4 showed signs of TRPA, whereas LO-treated samples did not show evidence of an increase in particle size. The range of particle size observed in this study was consistent with those reported by
in UHT milk during storage at 20°C after inoculation with a strain of P. fluorescens.
Significant differences were detected between the particle sizes of the examined groups in samples taken on d 3, as seen during wk 4 of the coagulation assessment (P = 0.020), and in samples taken on d 4, during wk 3 of the assessment (P = 0.041). This showed that the concentration of HSP, as produced by the Pseudomonas cocktail in milk at 6°C, increased with storage time and caused coagulation problems starting on d 3. Samples taken on d 4 of the storage period also showed increased particle size by wk 4 of the assessment; however, 2 out of 3 replicates exceeded the 10-µm detection limit of the particle size analyzer (
) and were not included in the statistical analysis.
Visible coagulation assessment results (Table 2) were consistent with those observed during particle size analysis. In the control, gel formation was present on d 3 and 4, with evidence of more rapid gelation on the latter. Previous research has shown that the rate and extent of coagulation in milk is heavily dependent on the microbial concentration in the sample.
reported that raw milk that had been inoculated with a strain of P. fluorescens and that reached 7.7 log cfu/mL before the UHT process gelled after 14 d; although, when the cell counts before UHT processing were lower (6.9 log cfu/mL), gelation took up to 10 wk at 20°C. Our results indicated that raw milk with initial counts of 1 × 105 cfu/mL—compliant with the Pasteurized Milk Ordinance limits (
in: Section 7. Standards for grade “A” milk and/or milk products. U.S. Department of Health and Human Services, Public Health Service,
Washington, DC2017: 34
)—stored at 6°C can present gelation problems after UHT treatment if the raw milk is not processed within 2 d of reception.
Table 2Visible coagulation (assessed in wk 2, 3, and 4) as an indication of heat-stable protease activity in UHT milk inoculated with Pseudomonas spp. and treated with lactose oxidase (LO)
Storage time (0, 1, 2, 3, and 4) represents the day on which samples were collected, processed, and incubated at 30°C for analysis of thermoresistant protease activity and coagulation assessment. UHT whole milk was used as a negative control. Unless stated otherwise, signs represent 100% response (n = 3).
1 Storage time (0, 1, 2, 3, and 4) represents the day on which samples were collected, processed, and incubated at 30°C for analysis of thermoresistant protease activity and coagulation assessment. UHT whole milk was used as a negative control. Unless stated otherwise, signs represent 100% response (n = 3).
Results from these analyses revealed the possibility of assessing UHT milk for TRPA in a short period of time by means of visual assessment of coagulation, as it correlated well with particle size analysis. Refer to Supplemental Figure S1b (https://doi.org/10.7298/rakk-ry78) for representative examples of positive coagulation versus no coagulation.
Effects of Concentration of LO and Supplementation of NaSCN on Raw Milk Quality and Subsequent Pseudomonas-Induced Gelation
After observing the effects of LO on Pseudomonas spp. in UHT milk, we also assessed the effect of this enzyme on microbial growth in raw milk. Figure 4 presents cell concentrations resulting from the Pseudomonas cocktail combined with the microbiota intrinsic to raw milk following exposure to different concentrations of LO, both alone and in combination with NaSCN. All samples were compared with a control consisting of raw milk inoculated with the bacterial cocktail and kept under the same storage conditions (dashed lines).
Figure 4Total plate count in raw milk (RM) treated with different concentrations of lactose oxidase (LO), with or without sodium thiocyanate (NaSCN) [14 mg/L] during storage at 6°C. Treatments presented in graphs c–h correspond to cell concentrations of raw milk previously inoculated with Pseudomonas spp. (the cocktail). Solid lines represent sample cell counts, and dashed lines represent the cell counts of raw milk inoculated with the cocktail but with neither LO nor NaSCN (control). Error bars represent the SD of 3 biological replicates.
Figure 4a shows the progression of microbial counts in raw milk, which went from an initial concentration of 3.4 log cfu/mL to a final cell count of 6.9 log cfu/mL over the course of 4 d of storage at 6°C. Previous researchers have determined the common raw milk microflora to be mostly composed of gram-negative genera including Pseudomonas, Achromobacter, Aeromonas, and Serratia, among others (
; our results are consistent with those observations.
When LO was incorporated into raw milk at a concentration of 0.12 g/L (Figure 4b) the cell count remained consistent around 3 log cfu/mL. The mechanism behind this bacteriostatic effect is different from the one observed in UHT milk. In raw milk, LO's production of H2O2 activates the LPOS (
explored the effects of LO at 0.12 g/L as a potential preservative of raw milk when exposed to suboptimal storage conditions (21°C). Their results showed that microbial counts of samples treated with LO were not significantly different from those of raw milk alone. According to
, hypothiocyanite ions are not stable at neutral pH, and at such stressed storage temperatures, they can spontaneously decompose back to thiocyanate, which has no demonstrated antimicrobial effect in milk. Low temperatures slow down this reduction of hypothiocyanite (
), providing an explanation the increased microbial inhibition of LO at refrigeration temperatures.
Bacterial growth in the control treatment reached an average of 8.3 log cfu/mL on d 4. This number differs from what we observed in UHT milk, where the final microbial count was on average 9.5 log cfu/mL. Various reasons could be responsible for this discrepancy, the main one being the action of the LPOS. The naturally occurring concentrations of H2O2 and thiocyanate in raw milk, along with the action of lactoperoxidase, seem to have a slight bacteriostatic effect (
). Additional H2O2-producing lactic acid bacteria are generally present in raw milk and have also been shown to work synergistically with the LPOS to inhibit psychrotrophic bacteria (
). In contrast, we would not expect this system to be active in UHT milk due to the great sensitivity that lactoperoxidase exhibits to high temperatures (
Figure 4c shows the effect of LO [0.12 g/L] on raw milk that had been inoculated with the Pseudomonas cocktail. Although, at the end of the 4-d storage period, the LO-treated sample was still significantly different from the control (P = 0.007); microbial levels reached 6.8 and 7.7 log cfu/mL on d 3 and 4, respectively. Based on previous observations from the present study, such microbial counts are within the range of potential HSP synthesis. It is known that the effectiveness of LPOS depends on the initial amount and type of microbiological contamination (
). Given that LO promotes the LPOS in this context, its effectiveness could also be dependent on the same factors.
In the context of raw milk, our methodology faced a limitation not seen in UHT milk, which involves the inability to differentiate between HSP-producing bacteria and non-HSP producers. This limitation was largely overcome through inclusion of additional analyses, however. The main purpose of our research was to determine whether a delay is seen in UHT milk gelation through the prevention of HSP synthesis via application of LO. As such, regardless of the milieu of raw milk versus UHT milk, the analysis of TRPA provided us with a much more direct means of measuring progress toward this goal than did the determination of concentration of pure cultures or complex microbial communities alone.
Considering that LO [0.12 g/L] did not exhibit the expected effectiveness in raw milk that was seen in UHT milk, we tested its effect at greater concentrations: 0.24 and 0.60 g/L (Figure 4, Figure 4). At a concentration of 0.24 g/L, LO proved to have a more pronounced inhibitory effect and reduced the microbial concentration by 0.61 log cfu/mL during the first 24 h of storage as compared with a decrease of 0.25 log cfu/mL with the addition of 0.12 g/L. Moreover, counts of these 2 LO treatments differed by 1.2 log cfu/mL on the last day of storage, and the 0.24g/L treatment did not exceed 6.5 log cfu/mL.
The application of LO at 0.60 g/L resulted in almost a 5-log difference as compared with the control at the end of the trial. These results showed that microbial count reduction was directly proportional to LO concentration. Determinations regarding what proportion of H2O2 entered the LPOS to produce antimicrobial compounds, versus what proportion remained uninvolved in the LPOS and instead acted as an antimicrobial itself, are outside the scope of this research. After examining these results, we decided to also investigate supplementation with NaSCN to determine if its availability was a limiting factor that may have influenced LO's promotion of the LPOS.
showed that in pasteurized milk, NaSCN at a concentration of 14 mg/L in combination with LO [0.12 g/L] demonstrated a bacteriostatic effect after a 7-d storage period at 6°C. On a similar note, we assessed the effects of NaSCN [14 mg/L] in combination with all LO concentrations previously tested; these results can be found in Figure 4, panels d, f, and h. After the 4-d storage period, LO at a concentration of 0.12 g/L, supplemented with NaSCN, produced an effect that was not significantly different from that of the highest LO concentration alone (P = 0.71). In addition, we observed no significant differences between NaSCN treatments throughout the entire storage period. This suggests that when added alone at a concentration of 0.12 g/L, LO can produce a concentration of H2O2 theoretically sufficient to boost the LPOS, but the raw milk fails to provide enough thiocyanate to complete the LPOS reaction. Our observations may also support the idea that greater concentrations of added LO result in a 2-pronged system of microbial inhibition, with a proportion of the resultant H2O2 feeding into the production of ionic hypothiocyanite and the rest of the H2O2 exerting its antimicrobial influence in an unmodified form.
The effects on pH of different concentrations of LO, as well as LO in combination with NaSCN [14 mg/L], are presented in Figure 5. All samples were compared with the pH of raw milk held at the same temperature (dashed lines). Results revealed that the addition of LO at a rate of 0.12 g/L caused a reduction of almost 0.2 pH units by the end of the simulated storage period. Similarly, when a LO concentration of 0.24 g/L was applied, the final pH was almost 0.3 units lower than that of the raw milk at the beginning of the experiment; however, this minor difference was not large enough to be statistically significant. On the other hand, at 5 times the original concentration of LO (i.e., 0.60 g/L), the pH drop was significantly different compared with those caused by the 2 lower concentrations, presumably due to a greater production of LBA.
Figure 5pH of raw milk (RM) treated with different concentrations of lactose oxidase (LO), with or without sodium thiocyanate (NaSCN) [14 mg/L] during storage at 6°C. Treatments presented in graphs c–h correspond to pH of raw milk previously inoculated with Pseudomonas spp. Solid lines represent sample pH, and dashed lines represent the pH measurements of uninoculated and untreated raw milk (control). Error bars represent the SD of 3 biological replicates.
Figure 5, panels d, f, and h show the effects of LO in combination with NaSCN on raw milk pH. Based on these results, supplementation with NaSCN had almost no effect on pH when compared with pH readings of the respective unsupplemented counterparts. Only samples that were treated with 0.60 g/L of LO exhibited lower pH values that were statistically different.
Table 3 presents the qualitative assessment of TRPA for all treatments (a to i) and includes a negative control of commercial UHT whole milk, as it constitutes the coagulation milieu into which the supernatants were mixed. Based on results observed in experiments with UHT milk, this assessment was carried out on the last 2 d of the storage period (d 3 and 4). Degree of coagulation was evaluated visually once a week for 5 wk. As seen previously, the supernatant obtained from the milk sample that had been inoculated with the cocktail but that contained no LO (treatment c in Table 3) triggered rapid gelation of the UHT whole milk.
Table 3Visible coagulation (assessed in wk 1, 2, 3, 4, and 5) as an indication of thermoresistant protease activity in raw milk (RM) inoculated with Pseudomonas spp. cocktail and treated with lactose oxidase (LO) and sodium thiocyanate (NaSCN) on d 3 and 4 of the storage period at 6°C
Coagulation assessment was conducted at 30°C wherein UHT whole milk was used as a negative control. Signs represent 100% response (3 out of 3 replicates).
3
4
Wk 1
Wk 2
Wk 3
Wk 4
Wk 5
Wk 1
Wk 2
Wk 3
Wk 4
Wk 5
UHT whole milk
−
−
−
−
−
−
−
−
−
−
(a) RM
−
−
−
−
−
−
−
−
−
−
(b) RM + LO [0.12 g/L]
−
−
−
−
−
−
−
−
−
−
(c) RM + cocktail
+
+
+
+
+
+
+
+
+
+
(d) RM + cocktail + LO [0.12 g/L]
−
−
−
−
−
−
−
−
−
+
(e) RM + cocktail + LO [0.24 g/L]
−
−
−
−
−
−
−
−
−
−
(f) RM + cocktail + LO [0.60 g/L]
−
−
−
−
−
−
−
−
−
−
(g) RM + cocktail + LO [0.12 g/L] + NaSCN
−
−
−
−
−
−
−
−
−
−
(h) RM + cocktail + LO [0.24 g/L] + NaSCN
−
−
−
−
−
−
−
−
−
−
(i) RM + cocktail + LO [0.60 g/L] + NaSCN
−
−
−
−
−
−
−
−
−
−
1 Coagulation assessment was conducted at 30°C wherein UHT whole milk was used as a negative control. Signs represent 100% response (3 out of 3 replicates).
Interestingly, supernatant taken on d 4 from the sample treated with LO at 0.12 g/L (treatment d in Table 3) showed coagulation by wk 5. This result reveals that LO at this concentration was able to inhibit HSP producers and to reduce to a certain extent the amount of protease synthesized during cold storage, but not to the levels seen in UHT milk. As stated before, the microbial concentration in this sample reached levels on d 3 and 4 that are considered potentially risky for HSP synthesis. Therefore, although not observed in the 3 replicates of the experiment discussed here, there seems to be little preventing the same visible coagulation of this sample on d 3.
On the other hand, LO at 0.24 g/L (treatment e in Table 3) successfully delayed milk gelation throughout the entire incubation period. These results correlate to the lower microbial concentrations observed in this sample (Figure 4e). Similarly, treatments with LO [0.60 g/L] and those with NaSCN supplementation were also effective in delaying gelation.
Altogether, in raw milk, LO at a concentration of 0.24 g/L was more successful in inhibiting the growth of HSP-producing Pseudomonas than was the previously proposed treatment (0.12 g/L) and prevented gelation with a minimal drop in pH. Furthermore, its effect was evident even in the absence of any additional LPOS activator (e.g., NaSCN); thus, this concentration was selected to move forward to the next stage of the study.
Lactose Oxidase Improvement of UHT Milk Shelf Life by Delaying Age Gelation
Up to this point our results have underlined the importance of preventing HSP-producing Pseudomonas spp. from reaching high cell concentrations approximately above 6.5 log cfu/mL as a means of delaying UHT milk gelation; in this regard, LO proved effective in benchtop scale experiments. The next phase of our research aimed to evaluate similar treatment protocols that would be applicable at a larger scale. The
in: Section 7. Standards for grade “A” milk and/or milk products. U.S. Department of Health and Human Services, Public Health Service,
Washington, DC2017: 34
Pasteurized Milk Ordinance standards for grade “A” raw milk destined for ultra-pasteurization mandate that milk holding tanks be emptied at least every 72 h for cleaning and sanitization procedures. Therefore, to mimic conditions that could be seen in dairy facilities, the storage time in our pilot plant–scale experiments was set to 3 d. During this stage, 3 treatments were tested: raw milk (hereafter referred to as RM), raw milk inoculated with the Pseudomonas cocktail (hereafter referred to as “control”), and raw milk plus the bacterial cocktail plus LO at a final concentration of 0.24 g/L (hereafter referred to as LO).
Figure 6a shows microbial concentrations during the simulation of the cold storage period. Results revealed that the RM and LO samples reached cell counts at the same log level as were seen in our small-scale experiments; however, the control showed lower microbial counts than those observed previously in RM. While final counts were not significantly different from one another, the control exhibited almost a 1-log difference from the LO-treated sample, reaching a final count of 6.6 log cfu/mL.
Figure 6Observed changes in pilot plant–scale treatments of raw milk (RM) treated with lactose oxidase (LO) during storage at 6°C and before UHT treatment: (a) cell counts and (b) pH. Error bars represent the SD of 3 biological replicates.
The pH changes during cold storage are shown in Figure 6b, where only the LO-treated sample proved to be statistically different from the other treatments (P = 0.021) after the 3-d storage period. Even though low pH can promote the solubilization of calcium phosphate and can potentially cause protein aggregation, the pH values observed with the proposed concentration of LO did not drop below 6.66 and would not be sufficient to cause significant changes in the composition of milk.
reported that casein micelle aggregation was only evident at pH below 5.5, at which point colloidal calcium phosphate also reaches its maximum solubilization rate. Potential fouling that could be triggered by low pH has been associated with aggregation phenomena (
); thus, we would not expect to observe such fouling at the reported pH.
After storage, carboys were carefully transported to the pilot plant facility in Geneva, NY, where temperatures were kept within 1°C of the setpoint. In this stage of the study, minimizing plasmin-induced proteolysis that could contribute to UHT milk gelation, as well as other factors that could cause other physical changes in milk (e.g., sedimentation and creaming) were considered important. Even though most of these other characteristics could be visibly discerned (
), we intended to prevent them from confounding the effects of heat-stable bacterial proteases in our samples. For that reason, the UHT process was designed to reduce these risks: homogenization was incorporated to curtail creaming, indirect heat was used to reduce sedimentation (
concluded that at temperatures ≥90°C, there is an important decrease in the activation energy required for the inactivation of plasmin and plasminogen (plasmin's precursor). Additionally,
determined the minimum preheating treatment required to delay plasmin-induced proteolysis, sedimentation, and gelation in UHT-processed milk that had been reconstituted from low-heat skim milk powder. Their study showed that preheating at 90°C for 30 s successfully prevented gelation and sedimentation for up to 6 mo following the heat treatment. In the present study we preheated all milk treatments at said conditions and subjected them to homogenization at 20.7 MPa (3,000 psi), followed by UHT treatment at 140°C for 4 s (
Ensuing UHT treatment, samples were aseptically aliquoted and packaged in 500 mL PET bottles, which were then stored at room temperature (∼25°C) for 6 mo. During this period, individual representative aliquot bottles were monitored for pH, particle size, and visible coagulation. Particle size analysis and visual inspection were also done for samples kept at 30°C, as this temperature promotes the gelation phenomenon (
Particle size data indicating level of HSP activity are presented in Figure 7. At 25°C (Figure 7a), there was no significant difference between LO and RM for up to 6 mo following the UHT treatment; however, the control particle size showed early signs of coagulation starting 1 mo after heat treatment, at which point it proved to be significantly different from those of the other groups (P = 0.0002).
Figure 7Particle size of pilot plant–scale UHT milk as measured over 6 mo of storage at (a) 25°C and (b) 30°C for raw milk (RM) and lactose oxidase (LO) groups. Error bars represent the SD of 3 biological replicates. For the control group, only data collected during the first 3 mo and 1.5 mo of storage were included at 25°C and 30°C, respectively, as gelation was readily visually observed after those time points (Supplemental Figure S2, https://doi.org/10.7298/rakk-ry78).
Particle size analysis of samples stored at 30°C is shown in Figure 7b, wherein size differences between groups are seen after 1 mo of storage (P = 0.0001); these differences were more pronounced than those seen at 25°C, with the highest values (control) being more than 1 log higher than those of both RM and LO. According to
, the size of particles does not change via consistent, progressive increments, but rather is subject to sudden increases, as casein micelles aggregate over a relatively brief period of time shortly before onset of visible gelation (around 1 mo at 20°C). For us, this macroscopic gelation marked the end of particle size monitoring for the control samples stored at both temperatures, for which data are not shown after this time point.
stated that age gelation is first detectable as an increase in viscosity due to the aggregation of casein micelles; in the case of bacterial proteases, this leads to the formation of a curd or a gel with custard-like consistency throughout the whole sample (
). Consequently, we looked for signs of gelation, first by tilting the bottles to reveal any changes in the consistency of the milk, such as curd formation, and later by looking for a serum layer on the surface of the liquid. Once the serum layer appeared and gelation was evident without tilting the bottles, we interpreted that as “evident” visible coagulation. For control samples, coagulation was consistently evident by the third month for all aliquot bottles in all 3 replicates stored at 25°C, and by 1.5 mo for those kept at 30°C, examples of which are presented in Supplemental Figure S2a and b (https://doi.org/10.7298/rakk-ry78), respectively. Additionally, after 6 mo of storage at both temperatures, there were no signs of visible gelation in the LO or RM samples, whereas the control underwent further physical separation (Supplemental Figure S3).
Our research has exposed some of the shortcomings inherent in current UHT manufacturing processes. Issues in UHT milk quality are not solely the result of the most egregious hygienic and processing practices. Although producers can process their raw milk in ways that are technically in conformity with the Pasteurized Milk Ordinance's limits regarding microbial concentrations and the duration and temperature of storage conditions, such milk is still susceptible to age gelation, provided that prior to UHT processing its microbiota is largely composed of HSP producers. Our studies demonstrate these results and the superior results achieved through the application of LO.
Effect of Storage on pH
Figure 8 presents changes in pH that occurred during storage at 25°C. Immediately after UHT treatment, we noticed a slight decrease in pH in all groups, a phenomenon that is believed to occur due to heat triggering effects on Maillard reactions, resulting in the production of formic acid (
). During the storage period, the pH of all treatment samples continued dropping over the course of 6 mo, with the control group experiencing the sharpest decrease, especially during the first 15 d of room temperature storage. Other studies have also reported pH declines during storage of UHT milk (
Figure 8pH of pilot plant–scale UHT milk over 6 mo of storage at room temperature (∼25°C) for the raw milk (RM), lactose oxidase (LO), and control groups. Error bars represent the SD of 3 biological replicates.
All in all, the results from this study support the hypothesis that the application of LO at a concentration of 0.24 g/L can successfully delay age gelation in UHT milk by providing the LPOS with the H2O2 required to generate antimicrobial compounds that prevent Pseudomonas spp. from producing HSP. To the best of our knowledge, H2O2 that could remain in suspension in the product is negligible, providing an opportunity for UHT milk producers to extend product shelf life while preserving overall quality and safety.
CONCLUSIONS
This study aimed to explore the ability of added LO to control growth of various strains of Pseudomonas spp. that can potentially contribute to age gelation in UHT milk. Our results showed that 0.24 g/L inhibited bacterial growth and prevented gelation, with a decrease of <0.2 pH units over 3 d at 6°C. Subsequently, LO-treated samples avoided gelation for up to 6 mo of storage at room temperature, whereas samples inoculated only with Pseudomonas underwent gelation after 3 mo. These findings provide a strategy for UHT milk supply chains, especially those lacking uninterrupted cold storage, to improve shelf stability; producers can also maintain a clean-ingredient label, while reducing postproduction losses and waste. Further studies should investigate use of LO in the production of other dairy products that are susceptible to HSP-driven textural changes, such as cheese and yogurt, and explore the potential of LO as an antimicrobial in other stages of dairy processing.
ACKNOWLEDGMENTS
We thank the New York State Dairy Promotion Advisory Board appointed by the New York State Department of Agriculture and Markets (Albany, NY) for funding this study. We acknowledge Cornell University personnel who supported this project in different ways: Martin Wiedmann and staff of the Food Safety Laboratory for facilitating access to the microbial isolates used in this study; Kyle Kriner and staff of the Cornell Food Venture Center Pilot Plant for their valuable help with the scaling-up section of this research; and Carmen Moraru, head of the Dairy Processing Laboratory, for allowing access to the particle size analysis equipment. The authors have not stated any conflicts of interest.
REFERENCES
Adams D.M.
Barach J.T.
Speck M.L.
Effect of psychrotrophic bacteria from raw milk on milk proteins and stability of milk proteins to ultrahigh temperature treatment.
in: Section 7. Standards for grade “A” milk and/or milk products. U.S. Department of Health and Human Services, Public Health Service,
Washington, DC2017: 34
Lund, M., C. L. Nikolajsen, and J. M. van den Brink, inventors. 2019. Use of cellobiose oxidase for reduction of Maillard reaction. World Intellectual Property Organization. Pat. No. WO 2019/110497 A1.
Seasonal influence on heat-resistant proteolytic capacity of Pseudomonas lundensis and Pseudomonas fragi, predominant milk spoilers isolated from Belgian raw milk samples.
31098-5&id=gr1.jpg){kind=link}
31098-5&id=gr2.jpg){kind=link}
31098-5&id=gr3.jpg){kind=link}
31098-5&id=gr4.jpg){kind=link}
31098-5&id=gr5.jpg){kind=link}
31098-5&id=gr6.jpg){kind=link}
31098-5&id=gr7.jpg){kind=link}
31098-5&id=gr8.jpg){kind=link}