Divergence of gut bacteria through the selection of genetic variations by milk exosomes Janos Zempleni,1 Fang Zhou,1 Haluk Dogan,2 Juan Cui2 University of Nebraska-Lincoln, Departments of 1Nutrition and Health Sciences and 2Computer Science and Engineering Background: Exosomes play an important role in cell-to-cell communication and are present in most body fluids including milk. Both Gram-positive and Gram-negative bacteria communicate with their environment through exosome-like outer membrane vesicles. We demonstrated that mouse pups absorb exosomes from maternal milk and a fraction of milk exosomes (MEs) escapes absorption and reaches the large intestine. MEs altered the composition of bacterial communities in murine ceca. Hypothesis: MEs select genetic variations in gut bacteria, thereby contributing to the divergence of bacterial populations. Methods: Gut content was collected from the ceca of three mice, age 7 weeks, suspended in minimal salts media and divided into two aliquots. One aliquot was cultured in media containing a nutritionally relevant concentration of MEs (1.7x1010/mL; denoted ME-supplemented, MES) under anaerobic conditions for 7 days; the other aliquot was cultured in ME-free media (MEF). DNA was sequenced using a 75-bp single end protocol (Illumina NextSeq 500; estimated coverage 150x). Genetic variations were assessed by using the MIDAS and StrainPhlAn pipelines. Results: Bioinformatics analyses were performed by using 127,935,309±30,104,915 and 138,253,606±25,740,862 reads per sample in MES and MEF cultures, respectively (N=3). MIDAS: More than 200 and 190 million sequencing reads were mapped to 11 and 19 bacterial species in MES and MEF cultures, respectively. In MES cultures, 278 and 28,594 strain-level genetic variations were detected by high stringency (detected in all 3 cultures) and low stringency (detected in 2 out of 3 cultures) analyses, respectively. Ninety-five genes in 11 bacterial species carried non-synonymous SNPs in MES cultures in the high stringency dataset. In MEF cultures, 92 and 26,382 strain-level genetic variations were detected by high and low stringency analyses, respectively. Forty-two genes in 19 bacterial species carried non-synonymous SNPs in MEF cultures in the high stringency dataset. Genetic variations were detected in enzymes catalyzing essential steps in the metabolism of tryptophan, glutamate and purines, i.e., pathways implicated in neurotransmitter synthesis in the host. StrainPhlAn: We detected 6,715 genetic variations across all loci in E. faecalis, C. sporogenes and L. johnsonii for both MES and MEF combined, including 6,694 variations in protein coding regions: 5,182 non-synonymous (77%) and 1,512 synonymous (23%) variations. We detected 62 insertions and 75 deletions among the non-synonymous variations. Conclusions: MEs contribute to the divergence of gut bacteria through the selection of genetic variations, which might affect neuronal signaling in the host. Funding: NIFA/USDA 2016-67001-25301 and 2020-67017-30834, NIH P20GM104320, USDA Hatch and W-40022 and Gates Foundation OPP1200494. J.Z. is a consultant for PureTech Health, Inc.