Author response: Key features of the genetic architecture and evolution of host-microbe interactions revealed by high-resolution genetic mapping of the mucosa-associated gut microbiome in hybrid mice Article Swipe

MAKE IMPACT

Shauni Doms , Hanna Fokt , Malte Rühlemann , Cecilia J. Chung , Axel Kuenstner , Saleh Ibrahim , André Franke , Leslie M. Turner , John F. Baines · YOU? · · 2022 · Open Access · · DOI: https://doi.org/10.7554/elife.75419.sa2

Article Figures and data Abstract Editor's evaluation eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Determining the forces that shape diversity in host-associated bacterial communities is critical to understanding the evolution and maintenance of metaorganisms. To gain deeper understanding of the role of host genetics in shaping gut microbial traits, we employed a powerful genetic mapping approach using inbred lines derived from the hybrid zone of two incipient house mouse species. Furthermore, we uniquely performed our analysis on microbial traits measured at the gut mucosal interface, which is in more direct contact with host cells and the immune system. Several mucosa-associated bacterial taxa have high heritability estimates, and interestingly, 16S rRNA transcript-based heritability estimates are positively correlated with cospeciation rate estimates. Genome-wide association mapping identifies 428 loci influencing 120 taxa, with narrow genomic intervals pinpointing promising candidate genes and pathways. Importantly, we identified an enrichment of candidate genes associated with several human diseases, including inflammatory bowel disease, and functional categories including innate immunity and G-protein-coupled receptors. These results highlight key features of the genetic architecture of mammalian host-microbe interactions and how they diverge as new species form. Editor's evaluation This paper uses inbred hybrid mouse lines to estimate the heritability of the mucosa-associated microbiome and map variants in the mouse genome that are associated with the composition of the microbiome. The findings are of broad interest to microbiome researchers and improve on knowledge in the field, as the mapping design facilitates the identification of narrow association intervals and points to a novel correlation between heritability and cospeciation rates. The manuscript provides useful information about the approach to heritability estimation, allowing the results to be more readily placed in context. Congratulations on this important contribution to the literature. https://doi.org/10.7554/eLife.75419.sa0 Decision letter Reviews on Sciety eLife's review process eLife digest The digestive system, particularly the large intestine, hosts many types of bacteria which together form the gut microbiome. The exact makeup of different bacterial species is specific to an individual, but microbiomes are often more similar between related individuals, and more generally, across related species. Whether this is because individuals share similar environments or similar genetic backgrounds remains unclear. These two factors can be disentangled by breeding different animal lineages – which have different genetic backgrounds while belonging to the same species – and then raising the progeny in the same environment. To investigate this question, Doms et al. studied the genes and microbiomes of mice resulting from breeding strains from multiple locations in a natural hybrid zone between different subspecies. The experiments showed that 428 genetic regions affected the makeup of the microbiome, many of which were known to be associated with human diseases. Further analysis revealed 79 genes that were particularly interesting, as they were involved in recognition and communication with bacteria. These results show how the influence of the host genome on microbiome composition becomes more specialized as animals evolve. Overall, the work by Doms et al. helps to pinpoint the genes that impact the microbiome; this knowledge could be helpful to examine how these interactions contribute to the emergence of conditions such as diabetes or inflammatory bowel disease, which are linked to perturbations in gut bacteria. Introduction The recent widespread recognition of the gut microbiome's importance to host health and fitness represents a critical advancement of biomedicine. Host phenotypes affected by the gut microbiome are documented in humans (Ley et al., 2006; Turnbaugh et al., 2009; Lynch and Pedersen, 2016), laboratory animals (Bäckhed et al., 2004; Turnbaugh et al., 2008; Rolig et al., 2015; Rosshart et al., 2017; Gould et al., 2018), and wild populations (Suzuki, 2017; Roth et al., 2019; Suzuki et al., 2020a; Hua et al., 2020), and include critical traits such as aiding digestion and energy uptake (Rowland et al., 2018), and the development and regulation of the immune system (Davenport, 2020). Despite the importance of the gut microbiome, community composition varies significantly among host species, populations, and individuals (Benson et al., 2010; Yatsunenko et al., 2012; Brooks et al., 2016; Rehman et al., 2016; Amato et al., 2019). While a portion of this variation is expected to be selectively neutral, alterations of the gut microbiome are on the one hand linked to numerous human diseases (Carding et al., 2015; Lynch and Pedersen, 2016), including diabetes (Qin, 2012), inflammatory bowel disease (IBD) (Ott et al., 2004; Gevers et al., 2014), and mental disorders (Clapp et al., 2017). On the other hand, there is evidence that the gut microbiome can play an important role in adaptation on both recent (Hehemann et al., 2010; Suzuki and Ley, 2020b) and ancient evolutionary timescales (Rausch et al., 2019). Collectively, these phenomena suggest that it would be evolutionarily advantageous for hosts to influence their microbiome. An intriguing observation made in comparative microbiome research in the last decade is that the pattern of diversification among gut microbiomes appears to mirror host phylogeny (Ochman et al., 2010). This phenomenon, coined 'phylosymbiosis' (Brucker and Bordenstein, 2012a; Brucker and Bordenstein, 2012b; Lim and Bordenstein, 2020), is documented in a number of diverse host taxa (Brooks et al., 2016) and also extends to the level of the phageome (Gogarten et al., 2021). Several non-mutually exclusive hypotheses are proposed to explain phylosymbiosis (Moran and Sloan, 2015). However, it is likely that vertical inheritance is important for at least a subset of taxa, as signatures of cospeciation/codiversification are present among numerous mammalian associated gut microbes (Moeller et al., 2016; Groussin et al., 2017; Moeller et al., 2019), which could also set the stage for potential coevolutionary processes. Importantly, experiments involving interspecific fecal microbiota transplants indeed provide evidence of host adaptation to their conspecific microbial communities (Brooks et al., 2016; Moeller et al., 2019). Furthermore, cospeciating taxa were observed to be significantly enriched among the bacterial species depleted in early onset IBD, an immune-related disorder, suggesting a greater evolved dependency on such taxa (Papa et al., 2012; Groussin et al., 2017). However, the nature of genetic changes involving host-microbe interactions that take place as new host species diverge remains underexplored. House mice are an excellent model system for evolutionary microbiome research, as studies of both natural populations and laboratory experiments are possible (Suzuki, 2017; Suzuki et al., 2019). In particular, the house mouse species complex is composed of subspecies that hybridize in nature, enabling the potential early stages of codiversification to be studied. We previously analysed the gut microbiome across the Central European hybrid zone of Mus musculus musculus and Mus musculus domesticus (Wang et al., 2015), which share a common ancestor ~0.5 million years ago (Geraldes et al., 2008). Importantly, transgressive phenotypes (i.e. exceeding or falling short of parental values) among gut microbial traits as well as increased intestinal histopathology scores were common in hybrids, suggesting that the genetic basis of host control over microbes has diverged (Wang et al., 2015). The same study performed an F2 cross between wild-derived inbred strains of M. m. domesticus and M. m. musculus and identified 14 quantitative trait loci (QTL) influencing 29 microbial traits. However, like classical laboratory mice, these strains had a history of rederivation and reconstitution of their gut microbiome, thus leading to deviations from the native microbial populations found in nature (Rosshart et al., 2017; Org and Lusis, 2018), and the genomic intervals were too large to identify individual genes. In this study, we employed a powerful genetic mapping approach using inbred lines directly derived from the M. m. musculus–M. m. domesticus hybrid zone, and further focus on the mucosa-associated microbiota due to its more direct interaction with host cells (Fukata and Arditi, 2013; Chu and Mazmanian, 2013), distinct functions compared to the luminal microbiota (Wang et al., 2010; Vaga et al., 2020), and greater dependence on host genetics (Spor et al., 2011; Linnenbrink et al., 2013). Previous mapping studies using hybrids raised in a laboratory environment showed that high mapping resolution is possible due to the hundreds of generations of natural admixture between parental genomes in the hybrid zone (Turner and Harr, 2014; Pallares et al., 2014; Škrabar et al., 2018). Accordingly, we here identify 428 loci contributing to variation in 120 taxa, whose narrow genomic intervals (median 1.53 × 10–6), a total of 1030 genome-wide significant associations were identified for individual taxa (p<1.53 × 10–6, Supplementary file 2), of which 428 achieved study-wide significance (p<1.29 × 10–8). Apart from the X chromosome, all autosomal chromosomes contained study-wide significant associations (Figure 2). Out of the 153 mapped taxa, 120 had at least one significant association (Table 1). For the remainder of our analyses, we focus on the results using the more stringent study-wide threshold, and combined significant SNPs within 10 Mb into significant regions (Supplementary file 3). The median size of significant regions is 1.91 Mb, which harbour a median of 14 protein-coding genes. On average, we observe five significant mouse genomic regions per bacterial taxon. Figure 2 with 2 supplements see all Download asset Open asset Heatmap of significant host loci from association mapping of bacterial abundances. Karotype plot showing the number of significant loci found using a study-wide threshold, where (A) plots the significance intervals and (B) the significant single nucleotide polymorphisms (SNP) markers on the chromosomes. The position of the SNPs on panel (B) has been amplified by 0.5 Mb to visualise it. The position of the genes closest to SNPs with the lowest p-values (Tlr4, and Irak4 and Adamsts20) are indicated. Table 1 Overview of mapping statistics. Loci with a P-value below the study-wide threshold (P<1.29 × 10–8) are considered significant. 'Significant loci total P' are the loci having a significant P-value from the total model (additive and dominance effect), 'Significant loci additive P' have a significant additive effect, and 'Significant loci dominance P' have a significant dominance effect. DNARNATotalMapped taxa101142153Taxa with significant loci6593120Median interval size (Mb)1.322.51.91Total significant SNPs316596782Total significant loci4437461184Unique significant loci172305428Significant loci total P83144204Significant loci additive P144244351Significant loci dominance P88171230Median significant loci per trait558Median unique significant loci per trait334Median unique significant SNPs per locus222Median number of genes per locus325443Median protein coding genes per locus111714 SNPs: single nucleotide polymorphisms. Of the significant loci with estimated interval sizes, we find 69 intervals (16.1%) that are smaller than 1 Mb (Supplementary file 4). The smallest interval is only 231 bases and associated with the RNA-based abundance of an unclassified genus belonging to Deltaproteobacteria. It is situated in an intron of the C3 gene, a complement component playing a central role in the activation of the complement system, which modulates inflammation and to activity et al., The significant genomic regions and SNPs are in Figure 2A and respectively. SNPs were associated with to 14 taxa, and significant intervals with to 30 taxa. The SNPs with the lowest p-values were associated with the genus and two ASVs belonging to and Figure supplement 1). the RNA level this two Mb, where the is of the closest gene × additive × dominance × Figure Figure supplement 1), and Mb, where the is found within the gene × additive × dominance × Figure Figure supplement 1). the Irak4 gene, whose protein is is also of The five taxa the most associations were ASV36 and (unclassified (Figure supplement 2). and effect A total of significant SNPs were between M. m. musculus and M. m. domesticus (i.e. between To gain further into the genetic architecture of microbial trait we estimated the of dominance at significant using the where with and effects have values of and respectively. of the SNPs were not (Figure it was not possible to have a associated with one we such that it can be with to bacterial abundance. For the of loci the associated with higher abundance is or Figure On the basis of the we used to only a proportion of are However, for of the significant the transgressive i.e., abundances that are either significantly of or higher of than those of both the domesticus was associated with higher bacterial abundance in of this subset musculus Figure Figure Download asset Open asset Genetic architecture of significant (A) of the with the highest phenotypic (B) of dominance values of significant loci a of loci or showing the percentage of variance explained by the single nucleotide polymorphisms (SNP) for DNA and RNA Next, we estimated phenotypic effect by the variance explained by the of significant SNPs explain between and of the variance in bacterial abundance, with a median effect size of (Figure The combined PVE by the additive effects of all significant markers for taxon from to with an of (Figure the combined additive effects of significant loci are than the which is the as it represents the total additive genetic effect. there are several taxa for which the PVE by additive and dominance effects of all significant SNPs genus and for RNA-based there are dominance For example, has two significant regions which show and has significant regions which show of candidate genes mapping for associations between a microbial trait and a single However, the genetic architecture is likely more For example, multiple genes in a common host could influence a taxonomic the of which could also be among taxa (i.e. functional Thus, in order to reveal potential biological phenomena among the identified we performed analysis to identify interactions and functional categories enriched among the genes in significant We used et al., to a interaction of protein-coding genes to significant SNPs A total of genes were in the and the is significant enrichment by and an of After only the with the highest confidence this results in one large with and smaller (Figure 4). Figure with supplements see all Download asset Open asset confidence interaction of genes closest to single nucleotide polymorphisms (SNPs) significantly associated with bacterial abundances. are using functional enrichment et al., 2019). and their interactions. Only with an interaction higher than are The of the line the interaction calculated by The of the the of the protein in the where is not and is Next, we using functional enrichment The genes of the are of the G-protein-coupled are the and also the class of and 2018). We then calculated the top from the based on with to important that can as (Figure supplement 1). The top contributing to the were and which are all involved in the Furthermore, we performed enrichment analysis on the genes to significant SNPs using the We found 14 pathways to be and inflammatory regulation of and the in and the (Supplementary file Figure supplement 2 and Figure supplement 3). Finally, genes involved in five human diseases are among mental disorders (Figure supplement 4). Finally, due to the observation of a significant enrichment of cospeciating taxa among the bacterial species depleted in early onset et al., and the evidence that is associated with a in mucosa-associated communities et al., 2020a; et al., 2021), we possible of genes involved in et al., among the genes significant SNPs. We found 14 of the genes, which was significantly more than expected by exact Supplementary file SNPs in of the 14 genes are associated with ASVs belonging to the genus a cospeciating taxon known to the of et al., 2020). of significant loci to gut microbiome mapping studies Host loci that in multiple independent studies are more likely to positive associations on We a of unique confidence intervals of significant associations with gut bacterial taxa from previous mouse studies (Benson et al., 2010; et al., 2012; et al., 2014; et al., 2015; Org et al., 2015; et al., 2016; et al., and compared this to our significance intervals for bacterial taxa at both the DNA and RNA level than 10 Mb were from all We found which is significant

Type

peer-review

Language

en

Landing Page

https://doi.org/10.7554/elife.75419.sa2

OA Status

gold

Related Works

10

OpenAlex ID

https://openalex.org/W4286727913

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4286727913

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.7554/elife.75419.sa2

Digital Object Identifier

Title

Author response: Key features of the genetic architecture and evolution of host-microbe interactions revealed by high-resolution genetic mapping of the mucosa-associated gut microbiome in hybrid mice

Work title

Type

peer-review

OpenAlex work type

Language

en

Primary language

Publication year

2022

Year of publication

Publication date

2022-04-13

Full publication date if available

Authors

Shauni Doms, Hanna Fokt, Malte Rühlemann, Cecilia J. Chung, Axel Kuenstner, Saleh Ibrahim, André Franke, Leslie M. Turner, John F. Baines

List of authors in order

Landing page

https://doi.org/10.7554/elife.75419.sa2

Publisher landing page

Open access

Yes

Whether a free full text is available

OA status

gold

Open access status per OpenAlex

OA URL

https://doi.org/10.7554/elife.75419.sa2

Direct OA link when available

Concepts

Key (lock), Host (biology), Microbiome, Biology, Genetic architecture, Architecture, Computational biology, Evolutionary biology, Genetics, Geography, Gene, Ecology, Phenotype, Archaeology

Top concepts (fields/topics) attached by OpenAlex

Cited by

0

Total citation count in OpenAlex

Related works (count)

10

Other works algorithmically related by OpenAlex

Full payload