Skip to main content

Destabilization and off-flavors generated by Pseudomonas proteases during or after UHT-processing of milk



Pseudomonads play a major role in the spoilage of UHT processed dairy products, due to their growth-related protease production in raw milk.


To assess the off-flavor generating capacity of these AprX proteases in milk after UHT-processing, six major milk spoiling Pseudomonas groups were investigated. Sensory evaluation of the different processed milk samples showed large differences in the degree of proteolysis related to onset of off-flavors. Nevertheless, it was illustrated that P. fragi has the greatest spoilage potential within the tested Pseudomonas groups, when it comes to generating off-flavors.


No clear correlation could be obtained between protein hydrolysis and the presence of off-flavors in UHT milk.


Refrigerated storage of raw milk is universally accepted for prolonging shelf life and preventing spoilage by mesophilic bacteria. Due to evolutions in the dairy market in which dairies have become more and more centralized, milk is now stored longer at refrigerated temperatures (Gaafar and Ali 1995). To ensure good dairy products, the Belgian legislation foresees, that milk on farm should be collected within 72 h post-production (Anonymous 2007). Indeed, quality problems may arise if milk is stored too long at these refrigerated temperatures. This is mainly due to an outgrowth of pyschrotrophic microorganisms in the raw milk. Psychrotrophic Pseudomonas (especially P. fragi, P. lundensis and members of the P. fluorescens-like group) are the dominant microbiota of raw milk and are known to compromise heat-treated milk (e.g., UHT) due to the production of heat-stable enzymes during their growth in raw milk (Marchand et al. 2009a). While pseudomonads are readily eliminated by UHT heating conditions (minimal 135 °C for 1 s), their heat-stable proteases may remain active in the heat-treated products (Chen et al. 2003; Griffiths et al. 1981). The presence of heat-stable Pseudomonas protease, encoded by the AprX gene, may result in spoilage and destabilization of UHT milk during extended storage (Dufour et al. 2008). Although, this protease gene is widespread over numerous Pseudomonas spp. (Chessa et al. 2000), the production process of this protease is still not completely understood and appears to be very complex. Quorum sensing (Juhas et al. 2005), temperature (Nicodème et al. 2005), iron content (Woods et al. 2001) and phase variation (van den Broeck et al. 2005) regulate and influence the production process of proteases at different levels. Typically, within Pseudomonas spp. only one protease, AprX, an alkaline zinc metalloprotease with a pH optimum of 6.5–8, is produced (Woods et al. 2001). This AprX protease is solely responsible for casein hydrolysis as evidenced by casein zymography (Marchand et al. 2009b). The family of serralysin proteases, to which the AprX Pseudomonas protease belong, appears to be highly conserved in some domains. Typical similarities in amino acid sequence are observed: a zinc-binding motif (xxxQTLTHEIGHxxGLxxGLxHPx), a calcium binding domain characterized by the presence of four glycine rich repeats (GGxGxD), a high content of hydrophobic amino acids and no cysteine residues (Rawlings and Klostermeyer 1995; Kumeta et al. 1999). However, despite of this general interspecies conservation, genetic differences in the AprX sequence might be responsible for observed inter individual differences in proteolytic capacity and/or the specific activity of these sequence divergent AprX proteases.

Proteolysis of UHT milk causes the development of bitter off-flavors, through the generation of hydrophobic peptides by hydrolysis of casein (Chen et al. 2003; Datta and Deeth 2003). Next to Pseudomonas proteases, proteolysis in UHT milk may also be attributed to the native milk enzyme plasmin. Milk plasmin is associated with the casein micelle and the milk fat globular membrane and is also quite heat-resistant (Saint-Denis et al. 2001; Fox and Kelly 2006). It may even partially survive mild UHT-processing conditions. Plasmin exists in milk in both its active form as well as its inactive precursor plasminogen. Its activity in milk is controlled by a complex network of enzyme activators and inhibitors (Fox and Kelly 2006). In addition, Pseudomonas proteases may contribute to overall plasmin activity by acting as plasminogen activators to convert plasminogen to plasmin (Fajardo-Lira et al. 2000).

Former research of the authors of this paper identified six major Pseudomonas protease groups with great milk spoilage behavior (Marchand et al. 2009a, b). However, no data was gathered yet on the off-flavor generation in milk with relation to that proteolytic capacity. Therefore, this paper addresses the differences in off-flavor generation by six representatives of the major milk Pseudomonas protease groups and assesses the correlation between protein hydrolysis and of off-flavor perception in processed UHT milk.


Selection of milk Pseudomonas strains

The selection of the different Pseudomonas strains was based on the findings of Marchand et al. (2009a, b). They defined six major Pseudomonas AprX protease groups. From each of these groups (A-B-C1-C2-D and the P. lundensis group) a representative was chosen to use in this study: respectively, Pseudomonas sp. Z34a, Pseudomonas sp. W12b, Pseudomonas sp. Z34b, Pseudomonas sp. W2a, P. fragi W41b and P. lundensis W52b.

Growth media

Cryopreserved Pseudomonas strains were first recovered in Brain Heart Infusion Broth (BHI) (Oxoid,

Basingstoke, Hampshire, England) before inoculation in UHT milk. The isolates were incubated in BHI at room temperature until growth was visually present. Next, 100 μL of incubated BHI broth was inoculated in 10 mL of UHT milk and incubated overnight at ambient temperature. The strains were checked for purity and bacterial counts showed that 24 h incubation at room temperature in UHT milk resulted in approximately 108 cfu mL−1. The six cultures were diluted in ringer solution (Oxoid) until 107 cfu mL−1.

Raw milk collection, pasteurization, inoculation with Pseudomonas strains

575 L of raw milk was collected from a farm in East-Flanders, Belgium. This full fat milk was pasteurized (72 °C,15 s) and aseptically divided in seven batches of 60 L. One batch was used as the control milk for further follow up of the experiment. The other six pasteurized milk batches were each inoculated with 6 mL of the ringer solution containing approximately 7 log mL−1 pseudomonads, in order to reach a final concentration of approximately 3 log pseudomonads in the 60 L batch.

Total colony counts and Pseudomonas counts

Total colony counts of the raw and pasteurized milk were determined by pour plating serial dilutions on Nutrient agar (Oxoid) with incubation at 30 °C for 3 days. Immediately after inoculation, Pseudomonas counts of the six 60 L milk batches were determined on a selective medium for Pseudomonas that contains cetrimide (10 mgL−1), fucidin (10 mgL−1) and cephalosporin (50 mgL−1) (CFC agar) (Oxoid) with incubation at 22 °C for 3 days.

Cold milk storage

The inoculated milk batches were further stored for 3 (t3), 4 (t4) and 5 (t5) days at 6,5 °C until skimming and further UHT-processing.

Skimming and UHT-processing

After cold storage, the seven different milk batches were further processed. Before UHT-processing the milk was skimmed using a Elecrem decreamer (Type 315 L/H,7800 tpm; Tomega, Marche-en-Famenne, Belgium). Indirect UHT-processing was performed on a Junior N326L apparatus, Process Pilot Plant, 200 Lh-1 (APV, Aartselaar, Belgium) under the following conditions: 2 steps homogenization: 200 bar, 65 °C; indirect UHT-processing: 5 s,140 °C; cooling to 20 °C. Milk was aseptically filled in high density poly ethylene (HDPE) bottles of 0.5 L and stored at 37 °C to accelerate possible proteolysis events. To ensure safe sensory evaluation, all produced milk samples were tested for sterility. Therefore 2 (HDPE) bottles of each milk batch were chosen randomly and incubated at 30 °C for 3 days. Total plate counts of the milk samples were determined by undiluted pour plating and incubation at 30 °C for 3 days.

Sensory evaluation

Preliminary sensory evaluation and proteolysis measurements started after 3 days storage at 37 °C. The six Pseudomonas protease milk samples were compared with the reference control milk (by a taste panel of 5 persons) on a daily basis. If off-flavors were experienced, a larger sensory analysis took place. The panel consisted of 35 people (8 men and 27 women) who were staff of the Institute of Agriculture, Fisheries and Food Research. Their mean age was 37 years (range 24–55 years). All panelists had earlier experience in sensory evaluation of milk. Evaluations were conducted in a sensorial cabinet that was equipped with individually partitioned booths. Milk samples (30 mL) were served at 14 °C. The set-up was as follows: Milk with off-flavor was diluted in the following way: A) Pseudomonas protease milk undiluted, B) 2/3 Pseudomonas milk + 1/3 control milk, C) 1/3 Pseudomonas milk + 2/3 control milk, D) Control milk undiluted. Next, the taste panel was asked to rank the milk samples (A-B-C-D) according to preference. Statistical evaluation of the results was based on the Rank Test to Kramer (Kramer 1960) for α = 0.05 but also compared with a newly developed sensory evaluation test. In short: The four milk samples (A-B-C-D) were presented at random to the tasting panel. The tasting panel was asked to arrange the milk samples on a line scale of 10 cm length according to preference. 10 cm was considered as an excellent taste, 5 cm satisfactory and 0 cm was considered as having a very bad taste. Two correction parameters were added to this test; (an example is given in Fig. 1). First, a correction within the tasting panel: the sensory analyst was enabled to choose between five statements to indicate the difference degree between the most extreme marks on the line scale. Each of these statements was correlated with a corrected difference in cm going from 0 cm (no noticeable difference) till 4 cm (very obvious difference) (Fig. 1). And second, a correction for the sensory analyst: the reference milk (the control undiluted milk, which is unknown to the tasting panel) and the other milk samples under evaluation can be placed by the sensorial analyst anywhere on the line scale. However, for the evaluation of the test, the control undiluted milk (D), is considered as satisfactory and is thus arbitrary associated with 5 cm on the line scale. Therefore, regardless of the place where the sensorial analyst has placed the reference, this milk gets de facto scaling 5. To obtain the corrected position for the other milk samples under evaluation, the distances between them and the reference mark need to be measured accurately. If the taste of the other milk samples was considered worse than the reference, the analyst would have placed its mark left from the reference on the line scale. If the taste was better, on the other hand, it would have been placed on the right hand side from the reference. Dependent if the mark is left or right from the reference, the measured distance between the two of them will be subtracted or added, respectively, from the arbitrary 5 cm scaling, resulting in the corrected values. All values derived from the different sensorial analysts were grouped and mean values and standard deviations (sd) for each milk sample were determined. Milk samples with mean values (±1 x standard deviation) were considered significant according to the Rank test to Kramer (Kramer 1960). Simultaneously, proteolysis was determined in each milk dilution (A-B-C-D).

Fig. 1
figure 1

Example of the newly developed sensorial evaluation test

Measurement of proteolysis

Hydrolysis of proteins was measured by the determination of the release of a-amino groups directly in milk by the trinitrobenzenesulfonic acid (TNBS) method (Polychroniadou 1988), in which free amino groups react with the TNBS reagent (Sigma-Aldrich, Bornem, Belgium) at pH 9.2 in the absence of light. A yellow-orange color develops and its intensity is determined in duplicate by absorption measurements at 420 nm. The amount of proteolysis in the Pseudomonas milk samples and dilutions is calculated from the increase in absorption and expressed as μmol glycine equivalents mL−1 milk using glycine (Sigma-Aldrich) as a standard curve.

Results and discussion

Milk processing and bacterial counts

575 L raw (full fat) milk was pasteurized and aseptically divided in seven batches of 60 L. Total bacterial count of the raw milk was 33.000 cfu/mL. After pasteurization total bacterial counts were reduced to 1900 cfu/mL in the pasteurized milk. Six 60 L batches were subsequently inoculated with Pseudomonas strains. Pseudomonas counts and total bacterial counts were determined at the moment of inoculation (t0) and before UHT-processing (t3, t4) for every milk batch. Bacterial counts can be retrieved in Table 1. All produced milk samples were sterile and were used for further sensory evaluation. In addition, the experiments showed that milk, which had been stored for 5 days or longer cannot be processed anymore under UHT conditions, because of destabilization of the milk, resulting in clogging of the heating exchanger.

Table 1 Bacterial counts in milk before the different processing steps (In bold: Pseudomonas counts sufficient to induce off-flavors in the processed milk samples)

Sensory evaluation of the processed milk samples

Sensory evaluation of the different processed milk samples showed large differences in the onset of off-flavors. The majority of the Pseudomonas inoculated milk samples (strains Z34b, Z34a, W2a and W52b) were stored 4 days at 6,5 °C before sufficient proteases were produced. In Pseudomonas sp. Z34b milk, Pseudomonas sp. Z34a milk and Pseudomonas lundensis W52b milk, off-flavors occurred after 13 days of storage at 37 °C post UHT-processing, while in Pseudomonas sp. W2a, off-flavors were already present after 10 days. However, it was illustrated that not all pseudomonads contain equal spoilage threats for the dairy industry: Pseudomonas fragi Z41b and Pseudomonas sp. W12b produced already sufficient proteases after 3 days of storage at 6,5 °C. Off-flavors were detected in those milk samples, after 15 days and 10 days storage at 37 °C, post UHT production, respectively (Table 2). From these results, it can be deduced that refrigerated storage of milk should be limited in order to prevent or reduce Pseudomonas protease production. In addition, the sensorial analysis evaluation by the method of Kramer (Kramer 1960) and the method described in this paper showed identical results (Table 2).

Table 2 Oversight of the sensory evaluation and proteolysis results (aMilk samples with off-flavors have values within the range indicated by Kramer and are thus significantly different from the milk samples with values outside that range)

Correlation between off-flavors and protein hydrolysis in UHT milk by the six different Pseudomonas protease groups

In each sensory evaluated milk sample proteolysis was determined. TNBS-values of each milk sample can be retrieved in Table 2. Based on these grouped results the correlation between protein hydrolysis and off-flavors can be determined. First of all, control milk was checked for the occurrence of proteolysis events. During a period of 30 days, milk was monitored for possible proteolytic activity. The TNBS value remained constant over time and had a mean value of 1,01 ± 0,04 μmol glycine equivalents mL−1. It can be concluded that no plasmin activity was present and milk (raw or processed) of good quality should thus have a TNBS-value in that range. Next, all data concerning sensorial and proteolysis analyses were compiled in Fig. 2. This graph shows that no clear correlation can be obtained between the onset of off-flavors and the rate of protein hydrolysis in milk by the different Pseudomonas protease groups. The TNBS-values of the milk samples in which off-flavors were significantly tasted were different for each Pseudomonas protease under evaluation. For example, with Pseudomonas sp. W2a proteases, the TNBS-value was allowed to rise with 1,03 μmol glycin equivalents mL−1 before any off-flavors were tasted. P. fragi proteases, on the other hand were capable in generating off-flavors after very limited proteolysis (a raise in TNBS-value of 0,15 glycin equivalents mL−1). Therefore it can be speculated that not all Pseudomonas proteases have the same specificity for their casein substrates. The amino acid recognition sites within Pseudomonas proteases might thus be fundamentally different, resulting in peptide generation with a variable hydrophobic amino acid content. Further research, however is necessary to confirm this. Nevertheless, it is now clear that presence (of high numbers) of P. fragi strains prior to UHT-processing will severely compromise the shelf life of derived dairy products. To ensure good quality dairy products, milk should therefore be processed as quickly as possible or held refrigerated (≤2 °C) (Griffiths 1989; Haryani et al. 2005) awaiting further processing.

Fig. 2
figure 2

Correlation between off-flavors and protein hydrolysis in UHT milk by 6 different Pseudomonas protease groups. * No significant proteolysis off-flavors indicated by black bars, the uncertainty range by light gray bars (the panel did not reject the milk samples; the lower limit of the bar is determined by the TNBS-value of the most diluted sample that was not rejected by sensory analysis) and the significant proteolysis off-flavors by dark gray bars (panel rejected the milk samples and tasted off-flavors; the lower limit of the bar is determined by the TNBS-value of the least diluted sample)


High Pseudomonas counts and extended cold storage severely limits UHT-processing. Therefore, to ensure good quality dairy products, raw milk should be processed as quickly as possible or kept well refrigerated (≤2 °C) during the entire dairy chain (from farm to dairy). No clear correlation can be obtained between the degree of protein hydrolysis by the different Pseudomonas AprX proteases and the generation of off-flavors in UHT-milk. Nevertheless, P. fragi has the greatest spoilage potential within the tested Pseudomonas protease groups, when it comes to generating off-flavors.


  • Anonymous. Ministrieel besluit houdende goedkeuring van het document opgesteld door de erkende inteprofessionele organismen betreffende de modaliteiten van de controle van de kwaliteit van de rauwe koemelk. Belgisch Staatsblad. 2007;7679–83.

  • Chen L, Daniel RM, Coolbear T. Detection and impact of protease and lipase activities in milk and milkpowders. Int Dairy. 2003;7:255–75.

    Article  Google Scholar 

  • Chessa JP, Petrescu I, Bentahir M, van Beeumen J, Gerday C. Purification, physico-chemical characterization and sequence of a heat-labile alkaline metalloprotease isolated from a psychrophilic Pseudomonas species. Biochim Biophys Acta. 2000;1479:265–74.

    Article  CAS  Google Scholar 

  • Datta N, Deeth HC. Diagnosing the cause in proteolysis in UHT milk. LWT Food Sci Technol. 2003;36:173–82.

    Article  CAS  Google Scholar 

  • Dufour D, Nicodème M, Periin C, Driou A, Brusseaux E, Humbert G, et al. Molecular typing of industrial strains of Pseudomonas spp. isolated from milk and genetical and biochemical characterization of an extracellular protease produced by one of them. Int J Food Microbiol. 2008;125:188–96.

    Article  CAS  Google Scholar 

  • Fajardo-Lira C, Oria M, Hayes KD, Nielsen SS. Effect of psychrotrophic bacteria and of an isolated protease from Pseudomonas fluorescens M3/6 on the plasmin system of fresh milk. J Dairy Sci. 2000;83:2190–9.

    Article  CAS  Google Scholar 

  • Fox PF, Kelly AL. Indigenous enzymes in milk: overview and historical aspects - Part 1. Int Dairy J. 2006;16:500–16.

    Article  CAS  Google Scholar 

  • Gaafar AM, Ali AA. The role of psychrotrophic bacteria in raw milk on stability of milk proteins to UHT-treatment. Egypt J Food Sci. 1995;23:147–54.

    Google Scholar 

  • Griffiths MW. Effect of temperature and milk fat on extracellular enzyme synthesis by psychrotrophic bacteria during growth in milk. Milchwissenschaft. 1989;44:537–43.

    Google Scholar 

  • Griffiths MW, Phillips JD, Muir DD. Thermostability of proteases and lipases from a number of species of psychrotrophic bacteria of dairy origin. J Appl Bacteriol. 1981;50:289–303.

    Article  CAS  Google Scholar 

  • Haryani S, Datta N, Elliot AJ, Deeth HC. Production of proteinases by psychrotrophic bacteria in raw milk stored at low temperature. Aust J Dairy Technol. 2005;58:15–20.

    Google Scholar 

  • Juhas M, Eberl L, Tümmler B. Quroum sensing: the power of cooperation in the world of Pseudomonas. Environ Microbiol. 2005;7:459–71.

    Article  CAS  Google Scholar 

  • Kramer A. A rapid method for determining significance of differences from rank sums. Food Technol. 1960;14:576–81.

    Google Scholar 

  • Kumeta H, Hoshino T, Goda T, Okayama T, Shimada T, Ohgiya S, et al. Identification of the serralysin family isolated from the psychrotrophic bacterium, Pseudomonas fluorescens 114. Biosci Biotechnol Biochem. 1999;63:1165–70.

    Article  CAS  Google Scholar 

  • Marchand S, Heylen K, Messens W, Coudijzer K, De Vos P, Dewetinck K, et al. Seasonal influence on heat-resistant proteolytic capacity of P. lundensis and P. fragi, predominant milk spoilers isolated from Belgian raw milk samples. Environ Microbiol. 2009a;11:467–82.

    Article  CAS  Google Scholar 

  • Marchand S, Vandriesche G, Coorevits A, Coudijzer K, De Jonghe V, Dewettinck K, et al. Heterogeneity of heat-resistant proteases from milk Pseudomonas species. Int J Food Microbiol. 2009b;133:68–77.

    Article  CAS  Google Scholar 

  • Nicodème JP, Grill G, Gaillard JL. Extracellular protease activity of different Pseudomonas strains: dependence of proteolytic activity on culture conditions. J App Microbiol. 2005;99:641–8.

    Article  Google Scholar 

  • Polychroniadou A. A simple procedure using trinitrobenzenesulphonic acid for monitoring proteolysis in cheese. J Dairy Sci. 1988;55:585–96.

    CAS  Google Scholar 

  • Rawlings ND, Klostermeyer H. Evolutionary families of metallopeptidases. Methods Enzymol. 1995;248:183–228.

    Article  CAS  Google Scholar 

  • Saint-Denis BT, Humbert G, Gaillard JL. Heat inactivation of native plasmin, plasminogen and plasminogen activators in bovine milk: a revisited study. Lait. 2001;81:715–29.

    Article  CAS  Google Scholar 

  • van den Broeck D, Bloemberg VG, Lugtenberg B. The role of phenotypic variation in rhizosphere Pseudomonas bacteria. Environ Microbiol. 2005;7:1686–97.

    Article  Google Scholar 

  • Woods RG, Burger M, Beven CA, Beacham IR. The aprX-lipA operon of Pseudomonas fluorescens B52: a molecular analysis of metalloprotease and lipase production. Microbiology. 2001;147:345–54.

    Article  CAS  Google Scholar 

Download references


The authors wish to thank Hans Bultinck for performing all pasteurization and UHT operations, Hans Steurbaut, Claudine Roels and Geert Goeteyn for the help with the flask filling operations. This research was supported by an OZM grant from the Agency for Innovation by Science and Technology (IWT).

Authors’ contributions

SM carried out the samplings, the sample analyses and drafted the manuscript. BD and JDB developed the new sensory evaluation test. MH, KC, JDB and BD participated in the design of the study. All authors read and approved the final manuscript.

Competing interests

The authors declare that they have no competing interests.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Barbara Duquenne.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Marchand, S., Duquenne, B., Heyndrickx, M. et al. Destabilization and off-flavors generated by Pseudomonas proteases during or after UHT-processing of milk. FoodContamination 4, 2 (2017).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI:


  • Pseudomonas
  • Protease
  • Sensory analysis
  • Spoilage
  • Milk