The demand for high quality dairy products has stimulated research regarding the diversity and spoilage potential of psychrotrophic bacteria. P. fluorescens and P. putida are important psychrotrophic microorganisms that produce thermostable hydrolytic enzymes that mainly degrade milk proteins and lipids, thus posing a major problem for the dairy industry (Decimo et al., Reference Decimo, Morandi, Silvetti and Brasca2014). According to Aguiar et al. (Reference Aguiar, Longhi, Poli-Frederico, Fagnani and Santana2019), the association of adequate cow milking and genetic studies evaluating Pseudomonas spp. strains present in the milking environment is essential to improve the quality and shelf life of dairy products.
In this work, the genetic variability of P. fluorescens and P. putida isolated from refrigerated raw milk from bulk tank trucks sent to the processing plant was determined and the relationship between the genetic profile of the species and their lipoproteolytic potential was evaluated. In addition, the lipoproteolytic potential was correlated with the number of milk producers composing a sample and the distance from the dairy property to the processing plant.
Materials and methods
Collection and source of isolates
After 48 h of refrigeration in bulk tanks in the dairy farms, milk was transported to the processing plant in trucks with isothermal tanks. Raw milk (10 samples) was collected from the tanks of the trucks containing milk from 8 to 80 dairy properties, whose distances from the processing plant ranged between 21 and 370 km.
Pseudomonas spp. was isolated from raw milk using CFC-supplemented (cefaloridine, fusidic acid, cetrimide) Pseudomonas agar base (Himedia, Mumbai, India) at 30°C for 48 h (Almeida et al., Reference Almeida, Bruzaroski, Zanol, Melo, Santos, Aragon- Alegro, Botaro and Santana2017). Milk temperature was measured using a digital thermometer and ranged between 7.5 and 9.7°C, with an average of 8.5°C.
Lipoproteolytic activity
Agar plates containing 25 to 250 colonies of Pseudomonas spp. were selected from each sample. The isolates (111) had their proteolytic (10% milk agar) and lipolytic (tributyrin agar) potential evaluated (21°C/72 h) (Frank and Yousef, Reference Frank and Yousef2004) by measuring the diameter (cm) of the characteristic halos formed after 72 h. Thereafter, the isolates were grouped in categories according to their low or high lipoproteolytic potential. Isolates with clear halos measuring ≤2 cm and <1.5 cm in diameter were classified as presenting low proteolytic (PP) and lipolytic (LP) potentials, respectively. Those with halos >2 cm and ≥1.5 cm in diameter were considered of high PP and LP potentials, respectively (Aguiar et al., Reference Aguiar, Longhi, Poli-Frederico, Fagnani and Santana2019). The cutoff values for the lipoproteolytic activity (high or low) were based on the mean and standard deviation of our data, always avoiding less than 40 observations for each category.
According to the diameter of the halos (cm), the lipoproteolytic potential was correlated with the number of milk producers composing a sample and the distance from the dairy property to the processing plant using nonparametric Spearman's test with 5% of significance.
P. fluorescens and P. putida molecular identification
The extracted genetic material (Wizard Genomic DNA Purification Kit, Promega Corporation, Madison, USA) was subjected to PCR amplification for the identification of Pseudomonas spp. (GenBank number AY486387.1) (Spilker et al., Reference Spilker, Coenye, Vandame and Lipuma2004). After, the species P. fluorescens (GenBank number CP015639.1) (Scarpellini et al., Reference Scarpellini, Franzetti and Galli2004) and P. putida (GenBank number CP015876.1) (Yamamoto and Harayama, Reference Yamamoto and Harayama1995) were identified and submitted to REP-PCR protocol. Ultrapure water was used as a negative control, while DNA from P. fluorescens (ATCC 13525) and P. putida (ATCC 31483) strains were used as positive controls.
Genetic similarity between the strains was assessed based on the presence or absence of a specific REP-PCR product (Louws et al., Reference Louws, Fulbright, Stephens and Debruijn1994). Similarity matrices were constructed using the Dice coefficient. The matrices were then grouped using the UPGMA method and graphically represented as OPTICS dendrograms using the software NTSYSpc2.10 (Numerical Taxonomy and Multivariate Analysis System; Stone Brook, 1998) and dendrograms were created for P. fluorescens and P. putida. The isolates were clustered into groups (Clusters I and II), where Cluster I had isolates with ‘low genetic diversity’ and Cluster II ‘high genetic diversity’.
The proportions between the groups (Clusters I and II) for both species were compared by Chi-square test. Fisher's exact test were used when the absolute frequency was less than 5.
Results and discussion
Sixty-seven (60.36%) Pseudomonas spp. isolates out of a total of 111 were P. fluorescens. This species was confirmed in all analyzed milk samples, with the frequency ranging from 3.57% to 100% (average 72.19%). A total of 83.9% (n = 26) of P. fluorescens isolates from Cluster I (n = 31), with low genetic diversity, presented at least 70% similarity, and the remainder exhibited 100% similarity. Cluster II (n = 36), with high genetic diversity, on the other hand, showed 44.4% (n = 16) of isolates with 70% similarity and none of them with 100% similarity. Interestingly, no isolates from samples G and H were observed in Clusters I and II, respectively (Fig. 1a). Thus, these results suggest the presence of a unique genetic profile of P. fluorescens in these two samples, which present as common characteristics related to the greater proximity between the farms and the dairy plant and samples containing the milk of a smaller number of dairy properties per tank truck.
Fig. 1. Dendrogram showing genetic similarity of P. fluorescens (a) and P. putida (b). Dendrogram was constructed based on numerical analysis of REP–PCR profiles of 67 P. fluorescens (a) and 44 P. putida (b) isolates from refrigerated raw milk. (a) Cluster 1: samples A (2), B (3, 18, 46), C(4, 27, 36, 37, 65), D (9, 12, 13, 17, 19, 20, 21, 24–26), E (7, 8, 28, 53), F (5, 11, 33, 47, 60), H (40), I (14) and J (50), Cluster 2: samples A (45), B (42, 49), C (23, 35, 62), D (6, 32, 41, 63, 64, 66, 67, 68), E (10, 22, 51, 52, 54–59, 61), F(15, 16), G (29), I (34, 38) and J (30, 31, 39, 43, 44, 48). (b) Cluster 1: samples E (28), G (3, 2, 5- 8, 10, 11, 13, 15–17, 34, 36, 38, 39–43, 45), H (31) and I (18, 19, 30). Cluster 2: samples E (20–25, 27,35), F (29), G (4, 9, 12, 14, 33, 37), H (32) and J (44).
Among P. fluorescens isolates showing 100% similarity, four groups were identified: the first (samples C and F) and the second (samples B, D and E) groups were composed of 3 isolates each, the third group (sample D) of 2 isolates, and the fourth one contained 9 isolates from 6 different origins (samples B, C, D, F, H, and J) (Fig. 1a). So, in this research, P. fluorescens isolates showing a 100% similarity compose the raw milk microbiota of different dairy farms. Probably, this fact is associated with sources of contamination common to these microorganisms, such as water and milking environment (Decimo et al., Reference Decimo, Morandi, Silvetti and Brasca2014), in addition to similar orientation of good cow milking, since the properties provide milk to the same dairy cooperative. Milkers hands, the surface of cows teats, teat cups and cooling tanks were associated with the presence of Pseudomonas spp. in raw milk from farms using manual or mechanical milking system, showing that, regardless of the type of system, good milking practices (GMP) are essential to avoid the contamination of raw milk (Vidal et al., Reference Vidal, Netto, Vaz, Capodifóglio, Gonçalves, Rossi, Figueiredo and Ruiz2017).
Forty-four (39.64%) strains of Pseudomonas spp. out of a total of 111 were identified as P. putida and isolated from 60% (6) of milk samples, with a frequency ranging from 12.5 to 96.43%. Most isolates (n = 20, 74%) belonging to Cluster I (n = 27) (Fig. 1b) presented at least 50% similarity. The presence of isolates with 100% similarity (Fig. 1b) shows low genetic diversity among them. On the other hand, only 35% (n = 6) of the 17 strains belonging to Cluster II, presented 50% similarity, which indicates greater genetic diversity than Cluster I. In addition, no strains from sample I was represented in Cluster II. Two groups of P. putida showing 100% similarity were identified (Fig. 1b): the first, composed of 9 isolates (samples G and I) and the second, by two (samples E and I). These results demonstrate the circulation of P. putida 100% similar in different milk producing regions.
In both clusters (I and II) observed in the dendrogram of P. fluorescens, the occurrences of high proteolytic (Cluster I 80.6%; Cluster II 75%) and lipolytic (Cluster I 22.6%; Cluster II 27.8%) potentials were equivalent (χ 2 = 0.30; P = 0.80) (Table 1). Thus, in our study, the deteriorating potential of the isolates was not linked with the level of genetic diversity of the species.
Table 1. Relative frequencies of high proteolytic and lipolytic potentials of P. fluorescens and P. putida isolated from raw milk in bulk tank truck*, clustered according to Rep-PCR analyses
* The samples (10) were collected from tank trucks containing milk from 8 to 80 dairy properties.
The lipoproteolytic intensity of P. fluorescens was associated only with the number of producers per milk sample. The greater the number of producers per sample, the greater the proteolytic capacity (P < 0.05; R = 0.28) and lipolytic (P < 0.01; R = 0.44) (Table 1) potentials of the P. fluorescens isolates. Enzymatic synthesis at refrigeration temperatures occurs mainly at the end of the log phase and during the stationary phase of the bacterial growth curve (Mahieu, Reference Mahieu and Luquet1991). Kumaresan et al. (Reference Kumaresan, Annalvilli and Sivakumar2007) observed that milk stored at 2°C resulted in lower bacterial growth and lower proteolytic and lipolytic activities, when compared to storage at 4 and 7°C for 14 d. Since a greater number of properties per tank is usually linked to a small volume of milk produced, more studies are needed to evaluate whether variables such as temperature and storage time differ in dairy farms and how the enzymatic activity of P. fluorescens is affected.
The occurrence of high proteolysis by P. putida was proportional between Clusters I (69.2%) and II (66.7%), but the occurrence of high lipolysis was greater (Exact Fisher's test; P = 0.10) in Cluster II (27.8%) (Table 1). Differences in extracellular enzymatic activity of individual strains are probably associated with a particular genetic group to which they belong (Ercolini et al., Reference Ercolini, Russo, Ferrocino and Villani2009). Although the lipolytic degradation of milk is not as intense as the proteolytic, defects resulting from the action of lipases are the first noticeable sensory changes, such as soap flavor and odor, metallic or oxidized flavor and fruity odor (Chen et al., Reference Chen, Daniel and Coolbear2003). According to the data obtained for P. putida, no significant association was detected between the number of properties per milk sample and the distance between the dairy farm and the processing plant.
In conclusion, the results indicate the circulation of P. putida and P. fluorescens with 100% similarity in different milk producing regions. The level of genetic diversity was related only to the lipolytic capacity of P. putida isolates and the lipoproteolytic intensity of P. fluorescens was positively associated with the number of farms per bulk milk tank truck.
Acknowledgements
This work was financial supported by CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior).
