The activities of different microbes in the cow rumen have been shown to modulate the host's ability to utilize plant biomass, while the host-rumen interface has received little attention. As datasets collected worldwide have pointed to Campylobacteraceae as particularly abundant members of the rumen epithelial microbiome, we targeted this group in a subset of seven cows with meta- and isolate genome analysis. We show that the dominant Campylobacteraceae lineage has recently speciated into two populations that were structured by genome-wide selective sweeps followed by population-specific gene import and recombination. These processes led to differences in gene expression and enzyme domain composition that correspond to the ability to utilize acetate, the main carbon source for the host, at the cost of inhibition by propionate. This trade-off in competitive ability further manifests itself in differential dynamics of the two populations in vivo. By exploring population-level adaptations that otherwise remain cryptic in culture-independent analyses, our results highlight how recent evolutionary dynamics can shape key functional roles in the rumen microbiome.

Simultaneous identification and metabolic analysis of microbes with single-cell resolution and high throughput are necessary to answer the question of "who eats what, when, and where" in complex microbial communities. Here, we present a mid-infrared photothermal-fluorescence in situ hybridization (MIP-FISH) platform that enables direct bridging of genotype and phenotype. Through multiple improvements of MIP imaging, the sensitive detection of isotopically labeled compounds incorporated into proteins of individual bacterial cells became possible, while simultaneous detection of FISH labeling with rRNA-targeted probes enabled the identification of the analyzed cells. In proof-of-concept experiments, we showed that the clear spectral red shift in the protein amide I region due to incorporation of C atoms originating from C-labeled glucose can be exploited by MIP-FISH to discriminate and identify C-labeled bacterial cells within a complex human gut microbiome sample. The presented methods open new opportunities for single-cell structure-function analyses for microbiology.

Infectious disease is widely considered to be a major driver of evolution. A preponderance of signatures of balancing selection at blood group-related genes is thought to be driven by inherent trade-offs in susceptibility to disease. B4galnt2 is subject to long-term balancing selection in house mice, where two divergent allele classes direct alternative tissue-specific expression of a glycosyltransferase in the intestine versus blood vessels. The blood vessel allele class leads to prolonged bleeding times similar to von Willebrand disease in humans, yet has been maintained for millions of years. Based on in vivo functional studies in inbred lab strains, it is hypothesized that the cost of prolonged bleeding times may be offset by an evolutionary trade-off involving susceptibility to a yet unknown pathogen(s). To identify candidate pathogens for which resistance could be mediated by B4galnt2 genotype, we here employed a novel "pathometagenomic" approach in a wild mouse population, which combines bacterial 16S rRNA gene-based community profiling with histopathology of gut tissue. Through subsequent isolation, genome sequencing and controlled experiments in lab mice, we show that the presence of the blood vessel allele is associated with resistance to a newly identified subspecies of Morganella morganii, a clinically important opportunistic pathogen. Given the increasing importance of zoonotic events, the approach outlined here may find useful application in the detection of emerging diseases in wild animal populations.