Paenibacillus odorifer, the Predominant Paenibacillus Species Isolated from Milk in the United States, Demonstrates Genetic and Phenotypic Conservation of Psychrotolerance but Clade-Associated Differences in Nitrogen Metabolic Pathways

Abstract

Paenibacillus is a spore-forming bacterial genus that is frequently isolated from fluid milk and is proposed to play a role in spoilage. To characterize the genetic and phenotypic diversity of Paenibacillus spp., we first used rpoB allelic typing data for a preexisting collection of 1,228 Paenibacillus species isolates collected from raw and processed milk, milk products, and dairy environmental sources. Whole-genome sequencing (WGS) and average nucleotide identity by BLAST (ANIb) analyses performed for a subset of 58 isolates representing unique and overrepresented rpoB allelic types in the collection revealed that these isolates represent 21 different Paenibacillus spp., with P. odorifer being the predominant species. Further genomic characterization of P. odorifer isolates identified two distinct phylogenetic clades, clades A and B, which showed significant overrepresentation of 172 and 164 ortholog clusters and 94 and 52 gene ontology (GO) terms, respectively. While nitrogen fixation genes were found in both clades, multiple genes associated with nitrate and nitrite reduction were overrepresented in clade A isolates; additional phenotypic testing demonstrated that nitrate reduction is specific to isolates in clade A. Hidden Markov models detected 9 to 10 different classes of cold shock-associated genetic elements in all P. odorifer isolates. Phenotypic testing revealed that all isolates tested here can grow in skim milk broth at 6°C, suggesting that psychrotolerance is conserved in P. odorifer Overall, our data suggest that Paenibacillus spp. isolated from milk in the United States represent broad genetic diversity, which may provide challenges for targeted-control strategies aimed at reducing fluid milk spoilage.IMPORTANCE Although Paenibacillus species isolates are frequently isolated from pasteurized fluid milk, the link between the genetic diversity and phenotypic characteristics of these isolates was not well understood, especially as some Bacillales isolated from milk are unable to grow at refrigeration temperatures. Our data demonstrate that Paenibacillus spp. isolated from fluid milk represent tremendous interspecies diversity, with P. odorifer being the predominant Paenibacillus sp. isolated. Furthermore, genetic and phenotypic data support that P. odorifer is well suited to transition from a soil-dwelling environment, where nitrogen fixation (and other nitrate/nitrite reduction pathways present only in clade A) may facilitate growth, to fluid milk, where its multiple cold shock-associated adaptations enable it to grow at refrigeration temperatures throughout the storage of milk. Therefore, efforts to reduce bacterial contamination of milk will require a systematic approach to reduce P. odorifer contamination of raw milk.

Keywords: Paenibacillus odorifer; nitrogen fixation; nitrogen metabolism; psychrotolerance; whole-genome sequencing.

FIG 1

Paenibacillus spp. isolated from fluid milk and dairy-associated environmental sources in the United States represent tremendous genetic diversity. (A) Bar chart displaying the sources of different rpoB allelic types of Paenibacillus species isolates represented five or more times among a collection of 1,228 Paenibacillus isolates. Colors correspond to the isolation source, and the length of each colored section of the bar reflects the percentage of isolates from a given source. (B) Distribution of Paenibacillus spp. identified using a 632-nt internal sequence of rpoB. Isolates in the “other” category represent a total of 24 isolates (representing the species P. glucanolyticus, P. lautus, P. pabuli, P. macerans, P. castanae, P. jilunlii, P. cineris, and P. illinoisensis). Isolates in the “Paenibacillus sp.” category represent isolates with rpoB alleles that could not be differentiated to the species level. (C) ANIb analysis of a subset of 58 Paenibacillus species isolates. Isolates shown in boldface type represent type strains. The vertical line represents the 95% cutoff for species identification.

FIG 2

Whole-genome sequence comparison of 25 P. odorifer isolates. (A) Phylogenetic tree constructed from 19,990 core SNPs identified in the genomes of 25 P. odorifer isolates and 3 isolates representing closely related species (Paenibacillus sp. 5 [clade D] and sp. 6 [clade C]). The maximum likelihood tree was constructed using a general time-reversible model with gamma-distributed sites and 1,000 bootstrap repetitions. Only bootstrap values of ≥70 are shown. Isolates are colored by clade. (B and C) Bar charts showing the number of isolates that contain a given gene cluster/product that was not detected in any isolates in the other clade for isolates in clade A (B) and clade B (C). Ortholog clusters encoding hypothetical proteins that were uniquely present in each clade (n = 68 and 51 for clade A and B isolates, respectively) are not shown. Complete lists of all ortholog clusters and GO terms significantly enriched in clade A and B isolates are shown in Data Sets S1 through S4 in the supplemental material. MFS, major facilitator superfamily.

FIG 3

Clade A and B isolates encode different nitrogen metabolism-associated pathways. (A) Illustration of the gene clusters in P. odorifer strain DSM 15391^T associated with nitrate and nitrite reduction and nitrogen fixation; these genes were used for BLAST searches in panel B. (B) Presence and absence of the nif (nitrogen fixation), nar (nitrate reduction), and nir (nitrite reduction) gene clusters for all 25 P. odorifer isolates. (C) Alignment of 5′ and 3′ regions flanking the nif gene cluster in isolate FSL F4-0134, compared to closed genomes for DSM 15391^T (clade A; NCBI accession number NZ_CP009428.1) and CBA7130 (clade B; NCBI accession number GCA_003255855.1). Alignments were performed using Muscle in Geneious software for 10 iterations. Identity scores shown in green (top column) have 100% nucleotide sequence identity; regions shown in yellow and red have between 30 and 99% and ≤29% nucleotide sequence identities, respectively. Gene names represent those annotated in the closed genome for DSM 15391^T.

FIG 4

P. odorifer isolates encode a variety of cold shock-associated genetic elements and have a conserved ability to grow at 6°C in skim milk. (A) Heat map displaying the number of hits identified for cold shock-associated genetic elements (e.g., proteins, the cold shock domain, and DEAD box sequences) for each isolate. Values are shaded to reflect the relative number of sequences detected (orange indicates high, and white indicates low). (B) Log difference in CFU per milliliter for isolates grown in SMB at 6°C for 14 or 21 days, compared to the CFU per milliliter in the inoculum at day 0. Results are averaged for each clade and are shown for clade A (n = 13 isolates) and clade B (n = 12 isolates). Error bars represent the standard deviations of data from three independent replicates plated in technical duplicates for all isolates in the clade.