Identification and characterization of psychrotolerant sporeformers associated with fluid milk production and processing

Abstract

Psychrotolerant spore-forming bacteria represent a major challenge to the goal of extending the shelf life of pasteurized dairy products. The objective of this study was to identify prominent phylogenetic groups of dairy-associated aerobic sporeformers and to characterize representative isolates for phenotypes relevant to growth in milk. Analysis of sequence data for a 632-nucleotide fragment of rpoB showed that 1,288 dairy-associated isolates (obtained from raw and pasteurized milk and from dairy farm environments) clustered into two major divisions representing (i) the genus Paenibacillus (737 isolates, including the species Paenibacillus odorifer, Paenibacillus graminis, and Paenibacillus amylolyticus sensu lato) and (ii) Bacillus (n = 467) (e.g., Bacillus licheniformis sensu lato, Bacillus pumilus, Bacillus weihenstephanensis) and genera formerly classified as Bacillus (n = 84) (e.g., Viridibacillus spp.). When isolates representing the most common rpoB allelic types (ATs) were tested for growth in skim milk broth at 6°C, 6/9 Paenibacillus isolates, but only 2/8 isolates representing Bacillus subtypes, grew >5 log CFU/ml over 21 days. In addition, 38/40 Paenibacillus isolates but only 3/47 Bacillus isolates tested were positive for β-galactosidase activity (including some isolates representing Bacillus licheniformis sensu lato, a common dairy-associated clade). Our study confirms that Paenibacillus spp. are the predominant psychrotolerant sporeformers in fluid milk and provides 16S rRNA gene and rpoB subtype data and phenotypic characteristics facilitating the identification of aerobic spore-forming spoilage organisms of concern. These data will be critical for the development of detection methods and control strategies that will reduce the introduction of psychrotolerant sporeformers and extend the shelf life of dairy products.

Fig 1

Midpoint-rooted maximum-likelihood (ML) phylogenetic tree of partial rpoB sequences from Bacillus spp. and related species isolated from pasteurized milk (red), raw milk (blue), and dairy farm environments (green). The scale represents the estimated number of nucleotide substitutions per site. Source information is shown for clades that contain 7 or more isolates. Numerical values represent the percentage of bootstrap replications that support the respective node. Only bootstrap values greater than 60 are shown. Bootstrap values for the Bacillus aerophilus sensu lato (s.l.), Bacillus pumilus, and Bacillus safensis clades are based on a separate ML analysis that included only rpoB ATs within these clades. AT158 was considered a plant-specific contaminant (since all 157 isolates were obtained from the same plant) and is therefore included once in the count shown. Group designations refer to both well-supported (i.e., groups I, II, and IV; BS, >70) and artificial (i.e., group III; BS, <70) groups. Species identification of clades and ATs was based on 16S rRNA gene sequence analyses as detailed in Materials and Methods. Clades and ATs that could not be identified to the species level were assigned a genus but no species (e.g., Bacillus sp. clade 2). B. cereus sensu lato also includes Bacillus anthracis and Bacillus pseudomycoides.

Fig 2

Midpoint-rooted maximum-likelihood phylogenetic tree of partial rpoB sequences from Paenibacillus isolated from pasteurized milk (red), raw milk (blue), and dairy farm environments (green). The scale represents the estimated number of nucleotide substitutions per site. Source information is shown for clades that contain 7 or more isolates. Numerical values represent the percentage of bootstrap replications that support the respective node. Only bootstrap values greater than 60 are shown. Group designations (i.e., groups V to XI) refer to both well-supported (i.e., groups V to VII and X; BS, >70) and artificial (i.e., groups VIII and XI; BS, <70) groups. Species identification of clades and ATs was based on 16S rRNA gene sequence analyses as detailed in Materials and Methods. Clades and ATs that could not be identified to the species level were assigned a genus but no species (i.e., Paenibacillus sp. clade 1 to Paenibacillus sp. clade 11).

Fig 3

Growth, in skim milk broth at 6°C, of isolates representing the most common rpoB allelic types found among Bacillus and related spp. (551 isolates) (A) and among Paenibacillus spp. (737 isolates) (B). Each data point represents the average for 3 independent biological replicates; error bars indicate standard deviations. Bacillus isolates tested represented AT001 (Bacillus licheniformis sensu lato clade 1; 2 isolates), AT003 (B. weihenstephanensis), AT017 (Viridibacillus sp.), AT020 (B. pumilus clade 1), AT135 (Bacillus aerophilus sensu lato), AT141 (B. safensis), and AT158 (Bacillus cereus sensu lato clade 1). Paenibacillus isolates tested represented AT015 (P. odorifer clade 1), AT023 and AT111 (Paenibacillus amylolyticus sensu lato), AT039 (P. graminis clade 2), AT045 (P. graminis clade 1), AT100 (Paenibacillus cf. xylanilyticus), AT157 (Paenibacillus cf. peoriae), AT159 (P. lautus), and AT260 (P. odorifer clade 3).