Environmental and population influences on mummichog (Fundulus heteroclitus) gut microbiomes

ABSTRACT

The mummichog, Fundulus heteroclitus, an abundant estuarine fish broadly distributed along the eastern coast of North America, has repeatedly evolved tolerance to otherwise lethal levels of aromatic hydrocarbon exposure. This tolerance is linked to reduced activation of the aryl hydrocarbon receptor (AHR) signaling pathway. In other animals, the AHR has been shown to influence the gastrointestinal-associated microbial community, particularly when activated by the model toxic pollutant 3,3′,4,4′,5-pentachlorobiphenyl (PCB-126) and other dioxin-like compounds. To understand host population and PCB-126 exposure effects on mummichog gut microbiota, we sampled two populations of wild fish, one from a PCB-contaminated environment (New Bedford Harbor, MA, USA) and the other from a much less polluted location (Scorton Creek, MA, USA), as well as laboratory-reared F2 generation fish originating from each of these populations. We examined the microbes associated with the gut of these fish using amplicon sequencing of bacterial and archaeal small subunit ribosomal RNA genes. Fish living in the PCB-polluted site had high microbial alpha and beta diversity compared to fish from the low PCB site. These differences between wild fish were not present in laboratory-reared F2 fish that originated from the same populations. Microbial compositional differences existed between wild and lab-reared fish, with the wild fish dominated by Vibrionaceae and the lab-reared fish by Enterococceae. These results suggest that mummichog habitat and/or environmental conditions have a stronger influence on the mummichog gut microbiome compared to population or hereditary-based influences. Mummichog are important eco-evolutionary model organisms; this work reveals their importance for exploring host-environmental-microbiome dynamics.

IMPORTANCE

The mummichog fish, a common resident of North America's east coast estuaries, has evolved the ability to survive in waters contaminated with toxic chemicals that would typically be deadly. Our study investigates how living in and adapting to these toxic environments may affect their gut microbiomes. We compared mummichogs from a polluted area in Massachusetts with those from a non-polluted site and found significant differences in their gut microbes. Interestingly, when we raised the next generation of these fish in a lab, these differences disappeared, suggesting that the environment plays a more crucial role in shaping the gut microbiome than genetics. Understanding these changes helps shed light on how animals and their associated microbiomes adapt to pollution, which can inform conservation efforts and our broader understanding of environmental impacts on host-microbe dynamics.

INTRODUCTION

Microbial communities are central to animal biology, and the animal host plus its symbiotic microbial community is increasingly considered a single complex assemblage, termed the “holobiont” (1, 2). The microbiome can influence critical host processes including physiology, development, and behavior, which in turn contributes to higher order phenomena including host adaptation, co-evolution of host and microbiome, and genetic divergence (3). As animals adapt to human-altered environments, it is important to incorporate host-microbe interactions as a factor affecting host ecology and evolution.

Fish are valuable field and laboratory models for investigating vertebrate host-microbiome interactions (4). In fish, gut microbiomes influence a variety of host processes, including the digestion of plant material (5), immune regulation (6), and neurological development (7). Unlike terrestrial mammals, fish live in a dense aquatic microbial environment that may interact with and influence the host microbiome. Changing environmental conditions such as pollution can cause rapid and profound shifts in environmental microorganisms, as well as aquatic host genetics and microbiome structure and function. Recent research has shown that external variables such as season, diet, and population density as well as host sex and immune system variation contribute to changes in fish gut microbiome richness and taxonomic composition (reviewed in reference [8]). A key research question is whether the composition of the microbiome is affected primarily by external environmental factors or by the host’s physiology. Answering such questions in fish requires attention to the host, its microbiome, and the environment (9).

The mummichog (Fundulus heteroclitus; also known as Atlantic killifish) is a unique natural model system for the interdisciplinary study of the effect of host genetics (and the adaptation of the host to a particular environment) on gut microbiome assembly and function in relation to different environments. Mummichogs are nonmigratory and have small home ranges, typically less than 1,000 m, leading to restricted gene flow and the development of genetic structure even at small spatial scales (10–12). For decades, this estuarine fish has served as a mature model system for studies in ecophysiology, evolution, and immunology (13). These fish possess a high-standing genetic diversity, which allows them to adapt to changing environmental conditions, including pollution, salinity, and eutrophication. Because of their broad distribution across a wide range of environmental conditions, mummichog has been proposed as a teleost model for environmental effects monitoring (13, 14). A better understanding of mummichog microbiomes will improve their utility as an ecosystem indicator and as a model organism for investigating host-microbe interactions in changing environments.

Currently, little is known about the natural variation in mummichog microbiota and how it has changed following host adaptation to environmental shifts, in particular to increased pollution. Multiple waterways on the Atlantic coast are impacted by high levels of pollution in the form of polycyclic aromatic hydrocarbons, chlorinated dibenzo-p-dioxins, and polychlorinated biphenyls (PCBs). Mummichog in these locations has experienced rapid and parallel adaptation to tolerate otherwise lethal concentrations of these chemicals (15). The mechanism of aromatic hydrocarbon tolerance in mummichog involves selection for downregulation of the aryl hydrocarbon receptor (AHR) pathway (15–17). In addition to affecting host genetics, environmental chemical contaminants including aromatic hydrocarbons may also shift the mummichog gut microbiome directly by altering microbial growth or indirectly by affecting host physiology (for a review, see reference [18]).

Adaptation to polluted environments through altered function of AHR pathway genes may contribute to changes in the gut microbiota of mummichog. The AHR is prominent in regulating intestinal immunity (19) and has been shown in a variety of animals to influence the gut microbiome. In mice, experiments have shown that AHR contributes to the community structure of the cecal microbiota (20–22). Furthermore, exposure to the potent AHR agonist PCB-126 increased inflammation and disrupted the gut microbiota, although the role of AHR was not directly investigated (23). In zebrafish, AHR activation by PCB-126 is associated with altered growth of gut microbes and inflammation (24). Examining the microbiome of PCB-exposed and PCB-adapted mummichog can help us further understand the consequences of pollution for the holobiont.

Here, we compared the microbiomes of mummichog originating from Scorton Creek (SC) and New Bedford Harbor (NBH), Massachusetts, USA. Mummichog from NBH, an EPA Superfund site (25), has adapted to tolerate high levels of PCB pollution in the sediment as a result of decades of industrial waste dumping, while the SC mummichog population is sensitive to PCB exposure (15, 26). The SC and NBH mummichog populations have previously been shown to have strong population genetic structure within AHR-related loci (27). We examined gut microbiomes of these two mummichog populations taken directly from the wild, as well as after being reared for two generations (3–5 years) in a common garden lab environment. We hypothesized that the tolerant wild mummichog microbiomes would be distinct from the sensitive wild mummichog microbiomes and that this effect would persist to some degree in the lab environment. In particular, we expected the tolerant mummichog gut microbiomes to show signatures of dysbiosis, such as pathobiont expansion or increased beta diversity.

MATERIALS AND METHODS

Fish and water sampling

Wild mummichog was collected from Scorton Creek, MA, USA (GPS: 41.746178, –70.426606) and New Bedford Harbor, MA, USA (GPS: 41.657413, –70.918928) on 9 September 2020 and 22 September 2020, respectively. Forty adult males were collected from each site using minnow traps placed at shallow depths (0.5–1.0 m) and transported live in a cooler with aerated water to Woods Hole, MA, USA. Replicate 2 L surface water samples was collected from each site and filtered through 0.22 µm pore size Supor filters (25 mm; Pall Corporation), and the filters were frozen in cryovials at −80°C. Sediment PCB concentration in SC and NBH was previously measured at 1 and 22,666 ng/g, respectively (15). Fish were sedated within the same day in 1.5% MS-222 and sodium bicarbonate and sacrificed by severing the spinal cord. The progenitors of the F2 fish were collected from SC and NBH in the years 2016–2018 and were maintained for two generations separately in otherwise similar tanks with continuous flow through seawater at the Atlantic Coastal Environmental Science Division as described previously (28). F2 fish experienced the same seawater and feeding conditions and can be considered common-gardened. A mixture of male and female F2 fish from each originating population (n = 15 SC F2; n = 44 NBH F2) were collected on 16 June 2021, flash-frozen and shipped to Woods Hole, MA, USA, where they were thawed and dissected. Aquaria water samples were not available for comparison.

Second-generation F2 fish were used rather than F1 fish because F1 fish had been exposed to contaminants passed on in the yolk. It has been shown previously that PCB levels do not differ between F2 fish from these two populations (29).

Sample processing and sequencing

Prior to dissection, measurements of the total length and body weight of each fish were taken. Then, the entire digestive tract from the esophagus to the anus was removed using sterile tweezers and frozen at −80°C. DNA was extracted from the whole guts and filters using the DNeasy PowerBiofilm kit (Qiagen) according to manufacturer’s instructions. Eight DNA extraction controls, consisting of no sample, were processed along with the biological samples. Barcoded primers 515FY (30) and 806RB (31) were used to amplify the V4 region of the small subunit rRNA gene in microbes (the primers target both bacteria and archaea). About 1 µL of DNA template was included in a 25 µL GoTaq Flexi PCR reaction. Two PCR controls consisting of 1 µL of PCR-grade water as template were also included, as well as microbial genomic DNA from a Human Microbiome Project mock community (BEI Resources, NIAID, NIH as part of the Human Microbiome Project: Genomic DNA from Microbial Mock Community B (Even, Low Concentration), v5.1L, for 16S RNA Gene Sequencing, HM-782D). PCR conditions were as follows: 37 cycles (95°C 20 s, 55°C 15 s, and 72°C 30 s) with a 2-min 95°C hot start and 10-min 72°C final elongation. Each reaction was run in duplicate and then pooled for gel purification (MinElute PCR Purification Kit, Qiagen). Purified products were quantified using the Qubit 2.0 fluorometer HS dsDNA assay (ThermoFisher Scientific), diluted to equal concentrations, and pooled for sequencing. Sequencing was performed at the University of Georgia Genomics and Bioinformatics core on a paired-end 2 × 250 bp Illumina MiSeq platform.

Data analysis

All code used for generating the data and figures for this project can be found on GitHub: https://github.cosm/microlei/killifish_manuscript. All raw sequence data have been uploaded to the Sequence Read Archive under BioProject ID PRJNA858104. Raw reads of the V4 region of the 16S rRNA gene were processed using the DADA2 package in R (v.3.6.2) (32). Reads were filtered using the default parameters of the function filterAndTrim except both forward and reverse reads were truncated at 240 bp. Read merging, chimera removal, and amplicon sequence variant (ASV) generation were also performed using DADA2. Taxonomy was assigned at the 100% identity level based on the Silva SSU rRNA gene database (v.138) (33). Contaminant reads were identified and removed, using the negative controls as a guide, by the prevalence method of the R package decontam (34). Sequences belonging to the Kingdom Eukaryota or Order Chloroplast were manually removed. Data analysis was performed in RStudio using the packages vegan and phyloseq (35, 36). Alpha diversity metrics were estimated using the function estimate_richness in vegan on the raw read counts. Differences in alpha diversity were tested using the pairwise t tests with holm correction for multiple comparisons (37). Counts were transformed to the center log ratio (CLR) after removing zeros using the package zCompositions (38). CLR transformed counts were used for the principal component analysis (PCA), and Aitchison distances (the Euclidean distance between the CLR transformed counts) were used for distance measures between samples. Aitchison distances were used because it is robust to subsetting the data and account for the compositionality of relative abundance data (39). Differentially abundant taxa were identified using the corncob package (40).

RESULTS

Sequencing statistics

After quality control and filtering, a total of 12,129,065 reads were retained from the 139 fish gut microbiome and the four water microbiome samples. Some fish gut microbiome samples had a high proportion of chloroplast reads but the median number of reads per sample was 88,125 with a range of 6,412 to 214,684 reads per sample. A total of 11,132 ASVs were identified, with a median of 83 unique ASVs per fish.

Alpha and beta diversity

Gut microbiome alpha diversity, including richness, Shannon index, and Simpson index, differed between the two wild populations but not between the F2 fish populations (Fig. 1A, pairwise t test: P < 0.0001 for all significant comparisons). The number of unique taxa (Richness) in NBH wild fish (the tolerant population) was on average 412 with standard deviation (SD) of 200, which was similar to that of the water samples (552 taxa, average of two samples). In contrast, the alpha diversity observed in SC wild fish (the sensitive population) (96, SD = 112) and all F2 fish populations (82, SD = 47) was lower, while that of the water from SC was similar to NBH water (478, average of two samples) (Table 1). Similarly, the NBH wild fish had a more even distribution of microbial taxa (Shannon index) than the other three fish populations. Overall, pairwise t tests show that NBH wild fish have significantly different diversity values than all other fish types, while SC wild, SC F2, and NBH F2 fish have similar measurements (Table S1). The Simpson index, a measure of the uniformity of the species abundances, showed similar patterns. One exception is that SC wild fish had a higher Simpson index (0.439) than NBH F2 fish (0.282). Richness of water samples collected at the wild sites was higher compared to SC but not NBH wild fish gut microbiota (Fig. 1A).

Fig 1.

Alpha and beta diversity measurements of mummichog microbiomes. (A) Alpha diversity indices. Asterisk indicates significant differences (P < 0.05) by pairwise t test with Holm correction between NBH wild versus all other fish and between SC wild and NBH F2. (B) Beta diversity calculated as distance to spatial median in Aitchison distance. Asterisk indicates significant differences (P < 0.05) by Tukey HSD pairwise testing. (C–E) PCA plots based on Aitchison distances of wild fish (C), F2 fish (D), and all samples (E). Ellipses around water samples, wild fish microbiomes, and F2 fish microbiomes are drawn to visually highlight the groups.

TABLE 1.

Mean alpha diversity metrics for all fish gut and water microbiomes^a

Sample type	Richness	Shannon	Simpson
New Bedford Harbor F2 (n = 44)	70.5 (32.1)	0.865 (0.378)	0.282 (0.117)
New Bedford Harbor water (n = 2)	552 (74.2)	4.25 (0.0716)	0.952 (0.0102)
New Bedford Harbor wild (n = 40)	412 (200)	3.2 (1.02)	0.826 (0.149)
Scorton Creek F2 (n = 15)	115 (64.9)	1.15 (0.602)	0.356 (0.149)
Scorton Creek water (n = 2)	478 (25.5)	4.42 (0.0528)	0.971 (0.00161)
Scorton Creek wild (n = 40)	96.7 (113)	1.12 (0.535)	0.439 (0.205)

^a Sample number is indicated after each sample type. Numbers in parentheses after metrics indicate the standard deviation of the measurement.

For examining beta diversity, a PCA of the Aitchison distance comparing microbiomes showed that water, as well as gut microbiomes of wild fish and F2 fish were distinct (Fig. 1E). The microbiomes of the two wild fish populations were separate from each other, while the two F2 fish populations were not as distinct. Separate PCAs of just wild or F2 fish show more separation between the NBH and SC populations (Fig. 1C and D). PERMANOVA comparisons of the Aitchison distances between the four fish types (SC wild, NBH wild, SC F2, and NBH F2) demonstrate that each fish type is significantly different from the others, as are the wild fish and F2 fish. However, the percent of variance explained (R²) varies between comparisons (Table 2).

TABLE 2.

PERMANOVA results comparing microbiome composition between fish types^a

Comparison	Df	Sum Sq	R 2	Pseudo-F	P value
All fish types	3	92,207	0.302	19.5	0.001
Wild versus F2	1	62,184	0.204	35.0	0.001
SC wild versus NBH wild	1	27,072	0.134	12.1	0.001
SC F2 versus NBH F2	1	2,878	0.073	4.5	0.001

^a Each row is an independent PERMANOVA test comparing the Aitchison distances of the mummichog microbiomes. All results were significant, although the R² value was variable. Df, degrees of freedom; NBH, New Bedford Harbor; Pseudo-F values derived from 999 permutations; R², R-squared value; SC, Scorton Creek.

Comparison of beta dispersion showed that the NBH wild population had a significantly higher beta diversity compared to all other fish types (Tukey HSD P < 0.001) and SC wild fish had a higher beta diversity than NBH F2 (Tukey HSD P < 0.05) (Fig. 1B). The F2 fish microbiomes did not differ in beta diversity (Tukey HSD P > 0.05). After accounting for the effect of fish type, fish length, weight, and (in the case of F2 fish) sex were not a significant factor in explaining the variance of Aitchison distances across samples (Table S2).

Core and differentially abundant taxa

To identify the core gut microbiome of the mummichog, we looked at the shared taxa between populations. The two wild populations shared 563 ASVs (6.3% of the total wild microbiome), while the two F2 populations shared 283 ASVs (19.5% of the total F2 microbiome). A total of 53 taxa were shared between all fish types. In general, there was a low percentage of overlap in ASVs between fish types, even between common-gardened F2 fish (Fig. 2A). The taxa shared between all fish types included members of the Phyla Firmicutes, Proteobacteria, and Planctomycetota. Detailed taxonomic classification of these shared ASVs can be found in Table S3.

Fig 2.

Core and differentially abundant amplicon sequence variants (ASVs) across fish populations. (A) A Venn diagram showing the overlap in ASVs between fish types, with 53 ASVs (0.526% of ASVs) shared among all four fish types. NBH, New Bedford Harbor; SC, Scorton Creek. (B) Pie chart of ASVs from the family Vibrionaceae shared between the two wild fish populations. Chart area is proportional to relative abundance. Colored ASVs have significantly different abundances between NBH wild and SC wild. (C) Bar graph of the relative abundances of selected bacterial families from all mummichog gut microbiomes. Colored bars are the families belonging to ASVs identified as significantly different between F2 and wild fish microbiomes.

Differential abundance (DA) analysis was performed on subsets of the data set to investigate which taxa may be driving the differences in community composition. In a comparison between all wild fish and all F2 fish gut microbiomes, DA analysis found 37 significant taxa. Overall, the dominant taxa in the wild fish were from the families Vibrionaceae or Mycoplasmataceae, while the dominant families of the F2 fish were comprised of Enterococcacea, Streptococcacae, and Clostridiacea (Fig. 2C). A bar graph showing the top 10 most abundant Families and Genera can be found in Fig. S1.

In the wild fish gut microbiomes, 40 taxa were significantly differentially abundant between SC and NBH fish—that is, enriched in either SC or NBH guts. We examined the relative abundance of these taxa in the seawater samples and found that they were at very low abundance (<1%) or not present at all. Of those that were present, only 12 out of 28 taxa had a higher abundance in the water sample corresponding to the enrichment in the gut samples (Table S4). Within wild fish, SC guts were characterized by high levels of ASV2, an unclassified Vibrionaceae, while NBH guts had a more even distribution of several Vibrionaceae ASV relative abundances (Fig. 2B).

Only three taxa were identified as significantly differentially abundant between the two F2 populations: the unclassified Vibrionaceae (ASV2), a member of the Paracoccus genus, and a member of the Colwellia genus. Of note, ASV2 was significantly enriched in both wild and F2 SC fish compared to their NBH counterparts.

DISCUSSION

In this study, we examined and compared the gut microbiomes of wild and lab-reared mummichog from two populations distinguished by their differential sensitivity to aromatic hydrocarbon pollution. We found that alpha diversity, beta diversity, and overall community composition of mummichog gut microbiomes are highly plastic and are strongly influenced by the environment, although host population (fish type) also has some effect. Mummichog size, weight, and sex were not correlated with the microbiome composition. The microbiomes from wild, PCB-resistant NBH fish were characterized by high alpha and beta diversity.

Mummichog microbiomes are dynamic and highly influenced by their environment

The composition of the gut microbiomes of all four fish types was distinguishable from each other (Fig. 1C through E), suggesting that habitat/environment as well as population influence the gut microbiome. However, high microbial diversity in the surrounding aquatic habitat does not necessarily lead to high gut microbiome diversity, as evidenced by the difference in alpha diversity metrics between the SC wild fish and its water source. This suggests that the mummichog gut microbiome is a selective environment, likely due to a combination of host control and environmental filtering. This result is in line with previous work describing fish gut microbiomes as specialized environments with potentially co-evolved relationships between host and gut bacteria (41–43). Variation in microbial composition between populations of the same species of fish have been explained by a combination of environmental and host effects. For example, in the three-spined stickleback, host genetic distance correlated with the microbiome’s UniFrac distance, while the microbiota also resembled that of local invertebrate prey (44). In a study comparing inter- and intra-species microbiome differences of Norwegian cod fishes, niche separation, and to a lesser extent evolutionary separation, were major contributors to microbiome differences (45). Mummichog gut microbiomes have also been found to differ both between wild populations and between wild and captive-raised fish (46, 47). Even short-term captivity has a large effect on microbiome composition (48).

The wild mummichog in the present study experienced two different environments (NBH and SC), and the previous work on these populations has shown them to be genetically distinct (27), suggesting that environmental and possibly host immune differences may contribute to differences in gut microbiomes. All fish were sampled in the summer to reduce the effect of seasonality, but dietary input, water condition, and other features likely vary between the two wild sites. Water condition measurements were not taken at time of sampling. Future investigations should incorporate additional measurements of the aquatic environment as well as gut content which may affect fish microbiome composition.

The wild NBH fish were exposed to high levels of PCB and other industrial pollutants. Although PCBs are the predominant contaminant in NBH and the reason the harbor was designated a Superfund site, other contaminants have also been found. These include metals (especially copper and chromium), petroleum hydrocarbons, and chlorinated dibenzofurans (49, 50). It is possible that exposure to these other contaminants has contributed to the differences in microbiomes that we observed. Exposure to environmental chemical contaminants has been shown in multiple fish species to be disruptive to gut microbiome composition (51–53). Upon comparing the variation in microbiome composition between the two wild fish populations and the common-gardened F2 fish, we suggest that the environment is likely the dominant force driving differences in mummichog gut microbiome composition in these populations.

In addition to environmental influences, our comparison between F2 fish populations reveals the potential for host effects on the gut microbiome. The microbial community composition of SC F2 and NBH F2 fish were significantly different as assessed by PERMANOVA. However, the effect size was small, and the result may also reflect the differences in sample size (NBH F2 consisted of 44 fish while SC F2 consisted of 15 fish). DA analysis revealed another indicator of host effect: ASV2, an unclassified member of the Vibrionaceae family, is significantly enriched in both the wild and F2 SC population compared to their NBH counterparts. Although Vibrionaceae is not dominant in F2 fish generally, its abundance in both SC fish types is suggestive of host selection. Within fish populations, genetic factors such as MHC genotype in stickleback and Bacterial Cold Water Disease resistance in rainbow trout have been shown to affect microbiome composition (54, 55). The results of this study indicate that additional sampling of different populations is needed to move beyond environmental effects. Closer examination of the genetic and microbial differences of these common-gardened captive mummichog populations could reveal more concrete links between host genotype and microbiome composition.

Captive fish microbiomes are compositionally distinct from their wild counterparts

While the wild fish gut microbiomes were dominated by Proteobacteria, such as Vibrionaceae and Mycoplasmataceae, the F2 fish microbiomes were dominated by Firmicutes, including Enterococcacea, Streptococcacae, and Clostridiacea. Diet is likely a major factor in this shift. The F2 fish were fed a typical aquaria flake diet (Aquanix Cool Mix 5-in-1) of fish protein, krill, and soy, yeast, and wheat gluten. In contrast, a typical wild mummichog fish diet is unlikely to contain a similar range of carbohydrates and is primarily composed of detritus, algae, and arthropods (56). Two other studies have examined the diversity of captive versus wild mummichog gut flora, though with different approaches. As part of a broader survey of fish microbiomes (47), researchers performed 454 pyrosequencing on six mummichogs and found a reduction in alpha diversity as measured by observed species, Chao1, and Shannon diversity (47). A more recent study examined the impact of short-term (28 days) captivity on mummichog gut flora using culture-based techniques found a significant loss of diversity, less evenness, and greater dominance of fewer species after the captivity period (48). Although the present study observed a similarly dramatic turnover in dominant taxa, we did not observe consistent differences in alpha or beta diversity between wild fish and their captive counterparts. This indicates that some other factor may be influencing the number of microbial species in mummichog guts. Additional studies following wild, captive acclimated (F1 generation), and captive-reared (F2 generation) fish may generate valuable insight into how the mummichog gut microbiome is colonized.

Shifts in microbiome composition from wild to captive-reared fish other than mummichog have been documented in aquaculture. Captive rearing of fish alters many factors that have been shown to influence gut microbiota, including stocking density, salinity, diet, and antibiotic treatment (42, 43). For example, the gut microbiota of rainbow darter (Etheostoma caeruleum) decreased in alpha diversity upon acclimation to a laboratory environment, and the dominant taxa shifted from Proteobacteria and Firmicutes to mainly Firmicutes (57), similar to the changes seen in this present study. The beta diversity of the acclimated fish microbiomes was also affected. In contrast, the microbiomes of hatchery-born Atlantic salmon juveniles displayed a higher alpha diversity than wild-born salmon while wild salmon hosted a more specialized microbial community (58). Additionally, the commercial diet fed to these salmon contributed to the dominance of the Lactobacillaceae family, which was absent from the abundant taxa of the wild microbiomes.

A complete turnover of the microbial environment can have consequences for the physiology of the host. Exposure to different microbes throughout the life of a fish activates a multitude of immune signaling pathways and commensal microbes that can modulate systemic immune responses in the host (59). Thus, we may expect host regulation of gut microbiota in the lab-reared mummichog to differ from that of wild mummichog. Such a difference would restrict the application of inferred host-microbiome interactions in a controlled lab setting to wild fish. However, without further sequencing of the microbial functional genes, it is difficult to come to any conclusions about the overall metabolic potential represented by the taxa present in each component. Additionally, information on host transcriptional activity in the gut environment is needed to confirm any similarity in intestinal regulation.

Residing in polluted water may impact microbiome composition

Our most striking finding was the increased diversity of the NBH wild fish microbiomes compared to the three other fish types. NBH wild was the only fish population that was directly exposed to high levels of PCBs. Although the fish are apparently healthy due to their tolerant phenotype, their microbiomes suggest a difference in host-microbiome interactions.

The increased beta diversity of NBH wild microbiomes is in line with the Anna Karenina Principle of dysbiosis, in which external environmental stress alters the microbiome in unpredictable ways, increasing variation among individuals (60). The resulting high beta diversity may be due to direct perturbation of the microbiome, such as by displacing mutualists, or due to a compromised host immunity. Laboratory-raised F2 generation NBH fish were previously found to have no difference in susceptibility to Vibrio harveyi infection compared to lab-reared PCB-sensitive fish from a reference site, but no comparison was made between lab-reared and wild-caught fish (28). Without additional experimental evidence, it is not possible to determine whether NBH wild fish display increased tolerance to pathogens than their captive counterparts or whether the increase in gut diversity is related to disease at all. The increased alpha and beta diversity in NBH wild fish is consistent with the hypothesis that NBH wild fish may have a reduced ability to regulate its gut microbiome. On the other hand, low gut alpha diversity and stability in fish are commonly associated with disease (61). While a third wild site for comparison may provide further evidence for the effect of PCB exposure on microbiome composition, diversity alone is not sufficient to infer dysbiosis as both increases and decreases in diversity can be implicated in gut microbiome disorders (62).

A healthy gut microbial community provides necessary services to the host, including the degradation of fiber and production of metabolic products, contributing to host nutrient absorption and energy metabolism (63). These metabolic functions are distributed among microbial consortia in the gut and therefore the redundancy and stability of microbe-microbe interactions is a topic of interest for those studying gut microbiomes (64).

Both theoretical and experimental studies have highlighted the need to understand the interaction network of microbial communities as it relates to microbiome functioning (51, 65, 66).

Caveats and future directions

Because only one tolerant and one sensitive population was sampled, the generalizability of these results is limited. We do not know what specific environmental conditions have contributed to the differences in these microbiomes. A broader survey of distinct mummichog populations across an environmental gradient of PCB contamination and fish tolerance phenotypes would be needed to uncover any relationship between environmental pollutants and gut microbiome composition. Additionally, direct host-microbiome interactions could not be assessed without experimental manipulations. Future work on the microbiomes of these and other populations of mummichog should ideally incorporate an experimental component, such as cross-transplantation, as well as examine the functional profile of both microbiome and host.

In conclusion, this study establishes a baseline understanding of the microbiome composition of wild and captive mummichog and provides evidence for a link between acquiring a divergent microbiome and living in a highly PCB-contaminated environment. We were able to describe the differences in microbiome composition and structure in the PCB-exposed mummichog population compared to the other fish populations that were not exposed to PCBs. We discovered that the dominant phyla of the mummichog microbiome can switch after two generations of captivity, although measures of diversity may be differently affected depending on the origin population. Finally, the plasticity of the mummichog microbiome highlights the strong connection that fish microbiomes have to their environment as well as the need for further research on the consequences of environmental change on the holobiont.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/spectrum.00947-24.

Fig. S1. spectrum.00947-24-s0001.tiff.

Bar graph of the top 10 most abundant A) Families and B) Genera across all samples between the four fish types. All other Families/Genera are represented in grey.

Supplemental material. spectrum.00947-24-s0002.docx.

Captions to Tables S1 to S4 with links to the CSVs.

Table S1. spectrum.00947-24-s0003.txt.

Pairwise t-tests of microbial community alpha diversity metrics between groups of fish, with holm correction for multiple comparisons.

Table S2. spectrum.00947-24-s0004.txt.

Marginal PERMANOVA results on fish body condition.

Table S3. spectrum.00947-24-s0005.txt.

Taxonomic identity and nucleotide sequences of shared ASVs between all fish types.

Table S4. spectrum.00947-24-s0006.txt.

Abundance of ASVs in seawater calculated as significantly enriched between New Bedford Harbor wild and Scorton Creek wild fish microbiomes.

OPEN PEER REVIEW. reviewer-comments.pdf.

An accounting of the reviewer comments and feedback.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.