Phase variation in Mannheimia haemolytica challenges the static genome paradigm

Abstract

Mannheimia haemolytica is a key agent in bovine respiratory disease (BRD), driving antibiotic use in feedyard cattle. As a facultative anaerobe commonly found in the upper respiratory tracts of cattle, its role under anaerobic conditions in BRD has not been extensively studied despite the known importance of anaerobiosis in human respiratory infections. Utilizing a combined omics approach, we refuted the null hypothesis that the M. haemolytica genome sequence is independent of the environment. This finding provides the rationale for research exploring how anaerobiosis-driven genome plasticity might contribute to a transition from a commensal to a virulent state. Genome sequencing of colony morphology variants from aerobic and anaerobic cultures revealed phase variation via homologous recombination between ribosomal RNA (rRNA) operons and slipped strand mispairing at simple sequence repeats (SSRs), frameshifting genes "on" and "off." Homologous recombination was exclusive to anaerobic conditions. SSR length variation in a DNA methyltransferase gene, targeting 5'-GACAT, correlated with methylation status. We observed statistically significant differences in the fraction of methylated motifs in isolates derived from anaerobic and aerobic cultures, as well as in the transcript abundances of genes in a nitrate reduction pathway and in a ribosomal protein class among anaerobic culture-derived colonial variants. We propose that instead of representing a M. haemolytica isolate's genome as a single static sequence, dynamic genome models should be developed. These models should account for the stochastic changes in genome sequence induced by phase variation, represented as a probability-weighted mixture of homologous recombinants and SSR switches.IMPORTANCEThe anaerobic growth of Mannheimia haemolytica generates phase variants through homologous recombination and slipped strand mispairing at simple sequence repeats. Homologous recombination between ribosomal RNA operons occurred exclusively in anaerobic conditions, likely similar to those present in infected lung tissue. Phase variation may yield variants better adapted for persistence in pulmonary tissues, potentially impacting antimicrobial persistence, as observed in persisters of other species. The genomic diversity resulting from phase variation complicates analyses based on static genome sequences, highlighting the need for dynamic genome models that reflect the stochastic nature of phase variation to understand M. haemolytica's role in host disease biology.

Keywords: DNA methylation; aerobiosis; anaerobiosis; bacterial infection; cattle; homologous recombination; hypoxia; nitrate reduction; persistence; phase variation; respiratory disease; ruminants; simple sequence repeat; slipped strand mispairing; tandem repeat.

Fig 1

Phase variants derived from anaerobic cultures. Spread plates of M. haemolytica USDA-ARS-USMARC-184 phase variants on BABH stained according to Wessman (18) sampled from four different vessels containing 120 mL of chemically defined minimal medium (CDMM, see Anaerobic culture at 37°C pre-purged with N₂ so that no dissolved O₂ was detected. Each vessel was inoculated with 4 mL of the same USDA-ARS-USMARC-184-derived CDMM inoculum SampleID 100158. The appearance of smooth and nonsmooth variants after staining is shown on the right. Without staining, smooth colonies appeared more opaque, creamy, larger, and convex, while nonsmooth colonies tended to be smaller and concave. Individual smooth and nonsmooth colonies were collected from nonstained BABH plates on day 3 and grown in 2 mL CDMM overnight aerobically, from which glycerol stocks were prepared. Spread plates in vessels 3 and 4 on day 8 were diluted 10-fold to achieve countable colony densities.

Fig 2

Mauve (20) alignment of closed genomes of smooth M. haemolytica USDA-ARS-USMARC-184 subculture phase variants resulting from growth under anaerobic conditions in a CDMM from day 3 colonies. Arrows denote homologous regions between genomes, identified as LCBs (19) in Mauve, and their direction relative to the parent reference genome CP006957.2. Manual curation of the start and stop of homologous regions was necessary because the automated algorithm missed regions of homology within the rRNA operons. Changes in the direction or location of homologous regions identify recombined genomic regions. The lengths of homologous regions are not drawn to scale but rather to accommodate the rRNA operon genes. Anchors denote rRNA operons not involved in recombination. There are six rRNA operons in the genome.

Fig 3

Mauve (20) alignment of closed genomes of nonsmooth M. haemolytica USDA-ARS-USMARC-184 subculture phase variants resulting from growth under anaerobic conditions in a CDMM from day 3 colonies. Arrows denote homologous regions between genomes, identified as LCBs (19) in Mauve, and their direction relative to the parent reference genome CP006957.2. Manual curation of the start and stop of homologous regions was necessary because the automated algorithm missed regions of homology within the rRNA operons. Changes in the direction or location of homologous regions identify recombined genomic regions. Lengths of homologous regions are not drawn to scale but rather to accommodate the rRNA operon genes. Anchors denote rRNA operons not involved in recombination. There are six rRNA operons in the genome.

Fig 4

Differences in the fraction of methylated motifs from phase variants generated under anaerobic and aerobic conditions from day 3 colonies for M.Mha184IV, M.Mha184II, M.Mha184I, M.Mha184III, and M.Mha184VI MTases. A indicates m⁶A modifications. Motifs methylated by Dam and Type III MTases exhibit significantly more variation in methylation fraction between anaerobic and aerobic sources compared to motifs methylated by Type I MTases. Boxes represent the interquartile range (25th to 75th percentile), red lines indicate the median, and whiskers extend to the most extreme data points not considered outliers; observations outside whiskers are outliers.

Fig 5

Direct RNA transcriptional analysis of nonsmooth versus smooth phase variants derived from day 3 anaerobic cultures. (a) Position and directions of genes in the M. haemolytica USDA-ARS-USMARC-184 (version 2) genome coding for transcripts found using the Pathway Tools Omics Dashboard (22) to be significantly differentially abundant in a pathway or class. These genes were in the nitrate reductase IV pathway or annotated with the GO: translation term (see Supplemental Figures A and B). The differential abundance P-value was the significance statistic, with a multiple test correction applied with a q = 0.01 level of significance using the Fisher exact statistics with the Benjamini-Hochberg (23) correction. The nitrate reduction pathway and GO: translation class were the only subsystems that fell below the multiple test threshold. (b) Eleven categories for 158 genes were indentifed in the transcriptome analysis: (i) GO: translation, (ii) translation regulation, (iii) nitrate reduction, (iv) anaerobiosis, (v) DNA metabolism, (vi) transporter, (vii) OMP (outer membrane protein), (viii) tRNA (intergenic), (ix) tRNA ribo16S23S (between 16S and 23S genes), (x) tRNA riboOpEdge (adjacent to 5S rRNA on the edge of the rRNA operon), and (xi) tRNA processing genes. See Annotation and categorization of differentially abundant genes for details.

Fig 6

(a) Histogram of the number of genes in the eleven categories from phase variants derived from day 3 anaerobic cultures. (b) Boxplot of the same group of 158 genes in Fig. 5b in each of the eleven categories in the histogram above. The variation in differential abundances from the 25th and 75th percentiles of each category is represented by the boxes; the red line indicates the median value, whiskers extend to the most extreme data points not considered outliers, and red crosses represent outliers. The eleven categories analyzed were (i) GO: translation, (ii) translation regulation, (iii) nitrate reduction, (iv) anaerobiosis, (v) DNA metabolism, (vi) transporter, (vii) OMP (outer membrane protein), (viii) tRNA (intergenic), (ix) tRNA ribo16S23S (between 16S and 23S genes), (x) tRNA riboOpEdge (immediately adjacent to, and upstream of, the 5S rRNA gene on the edge of the rRNA operon), and (xi) tRNA processing genes. See Annotation and categorization of differentially abundant genes for details.