Diet Is a Stronger Covariate than Exercise in Determining Gut Microbial Richness and Diversity

Obesity is a common metabolic disorder caused by a sedentary lifestyle, and a high-fat and a high-glucose diet in the form of fast foods. High-fat diet-induced obesity is a major cause of diabetes and cardiovascular diseases, whereas exercise and physical activity can ameliorate these disorders. Moreover, exercise and the gut microbiota are known to be interconnected, since exercise can increase the gut microbial diversity and contribute to the beneficial health effects. In this context, we analyzed the effect of diet and exercise on the gut microbiota of mice, by next-generation sequencing of the bacterial V4 region of 16S rRNA. Briefly, mice were divided into four groups: chow-diet (CD), high-fat diet (HFD), high-fat diet + exercise (HFX), and exercise-only (EX). The mice underwent treadmill exercise and diet intervention for 8 weeks, followed by the collection of their feces and DNA extraction for sequencing. The data were analyzed using the QIIME 2 bioinformatics platform and R software to assess their gut microbial composition, richness, and diversity. The Bacteroidetes to Firmicutes ratio was found to be decreased manifold in the HFD and HFX groups compared to the CD and EX groups. The gut microbial richness was comparatively lower in the HFD and HFX groups and higher in the CD and EX groups (ACE, Chao1, and observed OTUs). However, the Shannon alpha diversity index was higher in the HFD and HFX groups than in the CD and EX groups. The beta diversity based on Jaccard, Bray–Curtis, and weighted UniFrac distance metrics was significant among the groups, as measured by PERMANOVA. Paraprevotella, Desulfovibrio, and Lactococcus were the differentially abundant/present genera based on the intervention groups and in addition to these three bacteria, Butyricimonas and Desulfovibrio C21c20 were differentially abundant/present based on diet. Hence, diet significantly contributed to the majority of the changes in the gut microbiota.


Introduction
The obesity pandemic has been the largest contributor to ill health according to the World Health Organization (WHO) [1]. According to the latest WHO fact sheets, more than 1.9 billion adults and 600 million individuals have been classified as overweight or obese worldwide, respectively, and these rates are projected to increase at a faster pace in the coming decades [2]. Obesity is caused by a combination of genetic and environmental factors including genetic predisposition, the disruption of energy metabolism, and environmental and social factors [3]. A high-fat diet (HFD) and inadequate physical activity are major environmental factors causing obesity, while genetics also plays an important role in terms of body weight homeostasis [4]. Obesity is, therefore, expected to place a serious microbiota, using diet (involving chow and HFD) and treadmill exercise interventions, the context of diet-induced obesity.

Animal and Fecal Sample Collection
Four-week-old C57BL/6 female mice (n = 5) were randomly assigned to four group namely chow diet (CD), high-fat diet (HFD), high-fat diet + exercise (HFX), and exercis only (EX) groups. Normal mice were fed a standard rodent chow diet (protein 18%, 5%, fiber 5%, ash 5% of total weight) and HFD mice were fed a high-fat diet (protein 20.5 fat 34.9%, fiber, ash 8.0%, calcium, phosphorus 0.5% of total weight). The mice we housed in sterile cages, at a temperature of 22 ± 2 °C and relative humidity of 50 ± 10 under 12 h light/dark cycle. After 8 weeks of diet and treadmill exercise intervention, t mice were fasted for 12 h and sacrificed by isoflurane overdose in a random order. T body weight of the mice was recorded every week until the last week. Fecal samples fro individual mice were collected by restraint, in sterile microcentrifuge tubes, snap froze and stored at −80 °C, until further processing. The design of the study is presented Figure 1.

Exercise Protocol
The mice were housed in a cage for a week and adapted to treadmill exercise for fi days prior to the start of the exercise experiment. Exercise was performed for 8 weeks a treadmill chamber. Mice from the EX and HFX groups underwent treadmill exercise f a duration of 60 min/day on a 5% slope, with a gradually increasing intensity from m/min to 15 m/min for the initial 30 min, then increasing to 20 m/min for another 30 m followed by 5 min of cooling at 5 m/min.

DNA Extraction
The mouse fecal samples (frozen) were thawed at 4 °C and weighed (200 mg) f DNA extraction by SPINeasy DNA Kit for Feces (MP Biomedicals, Irvine, CA, USA), a cording to the manufacturer's instructions. Briefly, the stool samples were homogeniz

Exercise Protocol
The mice were housed in a cage for a week and adapted to treadmill exercise for five days prior to the start of the exercise experiment. Exercise was performed for 8 weeks in a treadmill chamber. Mice from the EX and HFX groups underwent treadmill exercise for a duration of 60 min/day on a 5% slope, with a gradually increasing intensity from 4 m/min to 15 m/min for the initial 30 min, then increasing to 20 m/min for another 30 min, followed by 5 min of cooling at 5 m/min.

DNA Extraction
The mouse fecal samples (frozen) were thawed at 4 • C and weighed (200 mg) for DNA extraction by SPINeasy DNA Kit for Feces (MP Biomedicals, Irvine, CA, USA), according to the manufacturer's instructions. Briefly, the stool samples were homogenized using a Fast-Prep 24 (MP Biomedicals, Irvine, CA, USA) bead homogenizer in tubes prefilled with high-quality beads, which ensured the disruption of the cell walls and the release of DNA into the solution. The extracted DNA was examined for integrity using electrophoresis on 1% agarose gel and visualized using ChemiDoc (Bio-Rad, Hercules, CA, USA).

DNA Quantification
DNA quantification and quality measurements were performed using two spectrophotometric devices, Spectra Max iD3 (Molecular Devices, San Jose, CA, USA) and Qubit 4 (Thermo Fisher Scientific, Waltham, MA, USA), based on the absorbance and fluorescence readouts, respectively. The A260/A280 absorbance ratio determined using Spectra Max iD3 was used to estimate the quality and purity of the extracted DNA, while the concentration was measured using Qubit 4.

16S rRNA-Based Amplicon Sequencing
The extracted DNA was amplified using barcoded primers that targeted the bacterial V4 region of the 16S rRNA gene. We used V4 region-specific primers along with locusspecific overhang sequences for PCR. The sequences used were 515F-5 -TCGTCGGCAGCGT CAGATGTGTATAAGAGACAG-GTGCCAGCMGCCGCGGTAA-3 and 806R -5 -GTCTCGT GGGCTCGGAGATGTGTATAAGAGACAGGGACTACHVGGGTWTCT-AAT-3 . The PCR thermal cycling conditions were 95 • C for 3 min, 72 • C for 5 min, 25 cycles of 95 • C for 30 s, 55 • C for 30 s, and 72 • C for 30 s. We attached the indices and Illumina sequencing adapters through index-PCR using the Nextera XT index kit v2 (Illumina, USA). The PCR thermal cycling conditions for index-PCR were 95 • C for 3 min, 72 • C for 5 min, eight cycles of 95 • C for 30 s, 55 • C for 30 s, and 72 • C for 30 s. PCR clean-up was performed to purify the 16S V4 amplicons using AMPure XP beads (Beckman Coulter, Brea, CA, USA), after every PCR amplification. The samples were normalized, pooled, mixed with the PhiX control v3 (Illumina), and sequenced using the Illumina iSeq 100 platform.

16S rRNA-Sequencing Analysis
Paired-end FASTQ sequences were imported into QIIME2 v2021.04 [37] and inspected with FastQC for quality control parameters, such as PhiX contamination, read trimming, and adapter sequences. The analysis was carried out using QIIME 2, unless otherwise stated. The PCR primers were removed using Cutadapt v3.4 [38] and the sequences were merged using VSEARCH v2.7.0 [39], followed by quality filtration (Q20) and dereplication. De-novo clustering was performed to select OTUs, followed by chimera removal, and representative sequences and abundances were extracted. A naïve Bayes classifier [40] was trained using 16S rRNA V4 region sequences extracted from the Greengenes 13.8 database [41], and the representative sequences were classified by taxon using the trained classifier. Alpha diversity indices, such as observed OTUs, Chao1, ACE, and Inverse Simpson, were estimated using the phyloseq package v1.38 of R, by importing the data from QIIME 2 in biom format. However, Shannon and Pileou's evenness were estimated using QIIME 2. Alpha diversity significance was calculated using the Kruskal-Wallis test, followed by the improved Benjamini-Hochberg procedure for false discovery rate (FDR) correction. Principal coordinate analysis (PCoA) was performed based on Jaccard, Bray-Curtis, and weighted UniFrac distance metrics to stratify the samples and determine the microbial community structure. Permutational multivariate analysis of variance (PERMANOVA) was employed to determine the statistical significance of the PCoA results, which were later confirmed by PERMDISP. Longitudinal ANOVA and ADONIS were used to determine the significant covariates of diversity. Analysis of the composition of microbiomes (ANCOM) was employed to determine differentially abundant taxa. The random forest model [42] was used to generate a heatmap of important taxa. Statistical significance was calculated using two-way ANOVA, one-way ANOVA, and Tukey's post hoc test. The results are expressed as the mean ± SEM. The significance level was set at p < 0.05.

Results
The body weight and food intake of each mouse were recorded once a week for 8 weeks. Table 1 shows the average initial and final body weights of the mice at the end of the experiment. After the 8-week intervention, the HFD group showed a significant increase in the final body weight compared to the CD group (p < 0.0001). As expected, the HFD group mice showed a higher body weight than the mice in the EX group (p = 0.0003). Similarly, the HFX group mice gained significantly more weight than the CD group mice (p = 0.0046). However, the CD and EX groups showed similar weight gain after 8 weeks. HFD can induce weight gain and increase body fat content in mice, as reported previously [43]. The mice in the HFX group weighed less on average compared to HFD-fed mice; however, no significant difference was recorded in weight gain between the HFD and HFX groups. The average food intake of the groups is shown in Table 1, with no significant differences observed among the groups.

Gut Microbial Richness and Diversity
The gut microbial richness was determined using different alpha diversity indices. The HFX group had low richness as measured by ACE, Chao1 and observed OTUs, where no significant difference was observed compared to the HFD group. In contrast, the CD group showed the highest richness, with no significant difference compared to the EX group ( Figure 2). Diversity estimation with the inverse Simpson metric showed patterns and results like those of the richness estimation analysis, with significantly lower diversity in the HFD and HFX groups than in the CD and EX groups ( Figure 2). The Shannon diversity did not differ between the HFD and HFX groups, and between the CD and EX groups. Moreover, the Shannon index, which considers richness and evenness, was higher in the HFD and HFX groups than in the CD and EX groups. A similar pattern was observed for Pileou's evenness ( Figure 2).
The structure of the gut microbial community of the different groups was examined using PCoA, and statistical significance was determined using PERMANOVA and confirmed using PERMDISP ( Figure 3). There was a partial separation between the HFD and HFX groups along axis 1 and axis 2, in Bray-Curtis and weighted UniFrac distance-based PCoA. Jaccard, which is a qualitative non-phylogenetic beta diversity metric, showed significant differences among all groups (p ≤ 0.02), except for the HFD and HFX groups ( Figure 3). Similarly, the Bray-Curtis metric was found to be significant among the groups (p ≤ 0.02, CD/EX, p = 0.024), except for the HFD and HFX groups. In contrast, the weighted UniFrac distance metric-based PCoA revealed significant differences among the groups (p < 0.02), except for CD/EX and HFD/HFX. Overall, the HFX and HFD groups showed higher variability, while the CD group had the lowest variability in Jaccard and Bray-Curtis based PCoA ( Figure 3). Overall, the dissimilarities between the CD and EX groups were evident without overlap in Jaccard and Bray-Curtis-based PCoAs, and with some overlap in the weighted UniFrac-based PCoA.

Microbial Taxonomic Profiling of Mouse Gut
Bacteroidetes and Firmicutes are the two largest phyla in the gut, which together account for approximately 90% of the total gut microbiota. Bacteroidetes accounted for approximately 89% and about 83% of the OTUs in the CD and EX groups (p = 0.14, ns), respectively ( Figure 4a). Likewise, the profile of Bacteroidetes in the HFD and HFX groups were similar but decreased to approximately half (p < 0.0001) of those estimated for the CD and EX groups. Moreover, Firmicutes accounted for about 24% and 16% of the OTUs in the HFD and HFX groups, respectively, with no significant difference between the two groups ( Figure 4a). In addition to the increased proportion of Firmicutes in the HFD and HFX groups, Proteobacteria also showed substantial expansion close to 40% in these groups (p < 0.0001) compared to the CD and EX groups (Figure 4a). Conversely, the relative abundance of Proteobacteria in the CD and EX groups was below 10%, and the profile was similar between the two groups. The relative abundance of the phylum Cyanobacteria significantly increased by approximately 4.7-fold, 7.8-fold, and~30-fold in the EX group only, compared to the CD group (p = 0.04), HFX group (p = 0.02), and HFD group (p = 0.009), respectively. Nonetheless, the relative abundance of Verrucomicrobia was rare in the intervention groups (Figure 4a). The structure of the gut microbial community of the different groups was examined using PCoA, and statistical significance was determined using PERMANOVA and confirmed using PERMDISP (Figure 3). There was a partial separation between the HFD and

Microbial Taxonomic Profiling of Mouse Gut
Bacteroidetes and Firmicutes are the two largest phyla in the gut, which together account for approximately 90% of the total gut microbiota. Bacteroidetes accounted for approximately 89% and about 83% of the OTUs in the CD and EX groups (p = 0.14, ns), respectively ( Figure 4a). Likewise, the profile of Bacteroidetes in the HFD and HFX groups were similar but decreased to approximately half (p < 0.0001) of those estimated for the CD and EX groups. Moreover, Firmicutes accounted for about 24% and 16% of the OTUs in the HFD and HFX groups, respectively, with no significant difference between the two groups ( Figure 4a). In addition to the increased proportion of Firmicutes in the HFD and HFX groups, Proteobacteria also showed substantial expansion close to 40% in these groups (p < 0.0001) compared to the CD and EX groups (Figure 4a). Conversely, the relative abundance of Proteobacteria in the CD and EX groups was below 10%, and the profile was similar between the two groups. The relative abundance of the phylum Cyanobacteria significantly increased by approximately 4.7-fold, 7.8-fold, and ~30-fold in the EX group only, compared to the CD group (p = 0.04), HFX group (p = 0.02), and HFD group (p = 0.009), respectively. Nonetheless, the relative abundance of Verrucomicrobia was rare in the intervention groups (Figure 4a).  Paraprevotellaceae was the most abundant bacterial family of the phylum Bacteroidetes in the CD group and was significantly prevalent compared to its negligible proportions in the HFD and HFX groups (p < 0.0001) (Figure 4b). The EX group showed a significant  Paraprevotellaceae was the most abundant bacterial family of the phylum Bacteroidetes in the CD group and was significantly prevalent compared to its negligible proportions in the HFD and HFX groups (p < 0.0001) (Figure 4b). The EX group showed a significant decrease in the Paraprevotellaceae family compared to the CD group (p = 0.01). Another family of Bacteroidetes, S24-7, was significantly higher in relative abundance in the highfat diet-fed groups (HFD and HFX) in comparison to the chow-fed groups (CD and EX) (p < 0.0001), regardless of the effect of exercise activity (Figure 4b). Likewise, Desulfovibrionaceae of the phylum Proteobacteria showed a similar pattern of relative abundance, with an increased proportion of about 20-30% in the HFD and HFX groups compared to about 2-4% in the CD and EX groups (p < 0.0001). Streptococcaceae, which belongs to the phylum Firmicutes, was rarely identified in the CD group, but was significantly abundant in the HFD group (~16%) (p < 0.0001). Interestingly, the abundance of the Streptococcaceae family was significantly reduced in the HFX group~1.6-fold compared to the HFD group (p = 0.01) (Figure 4b). The bacterial taxonomic families responsible for the increased abundance of the phyla Proteobacteria and Firmicutes in the high-fat diet-fed groups (HFD, HFX) were Desulfovibrionaceae and Streptococcaceae, respectively ( Figure 4). In contrast, the reduction of Paraprevotellaceae in the high-fat diet-fed groups (Figure 4b) might be the major contributing taxonomic family for the decreased proportion of phylum Bacteroidetes observed in these groups (Figure 4a).

Genus/Species Level Classification
Among the 30 taxa identified at the genus or species level in the mouse gut (Figure 5a), Paraprevotella, Desulfovibrio, and Lactococcus were significantly altered in the experimental groups, as illustrated in Figure 5b. Paraprevotella of the phylum Bacteroidetes was significantly decreased in relative abundance in the EX group (p = 0.0006) compared to the CD group. However, its abundance was scarce in the HFD and HFX groups, with no difference between the groups. The decreased percentage of Bacteroidetes in the HFD and HFX groups compared to that in the CD and EX groups can be attributed to the decreased abundance of Paraprevotella (Figures 4a and 5a,b). The genus Desulfovibrio from the Proteobacteria phylum increased significantly in the HFD and HFX groups compared to its proportion in the CD and EX groups (p < 0.0001). Interestingly, the proportion of the Desulfovibrio genus was significantly higher in the HFX group than in the HFD group (p = 0.017). The relative increase in the proportion of Firmicutes in the HFD and HFX groups (Figure 4a) was reflected at the genus level and is attributed to Lactococcus. Lactococcus was significantly reduced in the HFX group~1.6-fold compared to its proportion in the HFD group (p < 0.0009), as shown in Figure 5a, b. The presence of Lactococcus was not identified in the CD and EX groups. Moreover, Bacteroides ovatus, Sutterella, Atherosperma moschatum, Anaeroplasma, Lactobacillus reuteri, and Coprobacillus were identified with negligible relative abundance in all the experimental groups (data not shown). Unidentified species fell in the range of~20-50% (Figure 5a).

Differentially Abundant and Important Taxa of Mouse Gut Microbiota
ANCOM based on diet revealed the differentially abundant or present taxa in the high-fat diet-fed groups and chow-fed groups. Among these, Paraprevotella (Paraprevotellaceae, Bacteroidetes) and Desulfovibrio C21c20 (Desulfovibrionaceae, Proteobacteria) were differentially abundant/present in the chow-fed groups (CD, EX), whereas Lactococcus (Streptococcaceae, Firmicutes), Desulfovibrio (Desulfovibrionaceae, Proteobacteria) and Butyricimonas (Odoribacteraceae, Bacteroidetes) were differentially abundant/present in the high-fat diet-fed groups (HFD, HFX) (Figure 5c). Almost half of the microbiota that were significantly altered with respect to the diet intervention belonged to S24-7 bacterial taxonomic family of the order Bacteroidales, within the phylum Bacteroidetes. Moreover, 80% of S24-7 family members identified by ANCOM based on diet were relatively significantly abundant in the high-fat diet-fed groups (Figure 5c). icantly reduced in the HFX group ~1.6-fold compared to its proportion in the HFD group (p < 0.0009), as shown in Figure 5a, b. The presence of Lactococcus was not identified in the CD and EX groups. Moreover, Bacteroides ovatus, Sutterella, Atherosperma moschatum, Anaeroplasma, Lactobacillus reuteri, and Coprobacillus were identified with negligible relative abundance in all the experimental groups (data not shown). Unidentified species fell in the range of ~20-50% (Figure 5a).

Differentially Abundant and Important Taxa of Mouse Gut Microbiota
ANCOM based on diet revealed the differentially abundant or present taxa in the high-fat diet-fed groups and chow-fed groups. Among these, Paraprevotella (Paraprevotellaceae, Bacteroidetes) and Desulfovibrio C21c20 (Desulfovibrionaceae, Proteobacteria) were differentially abundant/present in the chow-fed groups (CD, EX), whereas Lactococcus The most important taxa within the OTU table were generated using the random forest machine-learning algorithm ( Figure 6). The majority of the members of the order YS2 from the phylum Cyanobacteria were relatively abundant in the EX group compared to the other groups. On the other hand, Allobaculum (Firmicutes) was relatively abundant in the CD group, as shown in the heatmap (Figure 6). Among the four Adlercreutzia (Actinobacteria) species identified as the important taxa between the intervention groups, three were relatively abundant in the high-fat diet-fed groups (HFD, HFX) ( Figure 6). Oscillospira (Firmicutes) was insignificantly enhanced in the HFX group compared to the other groups. Two species of Odoribacter (Bacteroidetes) were identified with contrasting abundance profiles, where one taxon showed moderately higher abundance in HFD than the other groups, while the other taxon showed strikingly higher abundance in all groups except HFD ( Figure 6). Coprococcus (Firmicutes) was found to be relatively mildly higher in abundance in the high-fat diet-fed groups. Desulfovibrio and Lactococcus were abundant in the HFD and HFX groups, while Prevotella was relatively frequently observed in the CD and EX groups ( Figure 6).

Discussion
The B/F ratio is known to have an important influence on the homeostasis of the gut microbiome. An altered B/F ratio is associated with dysbiosis and several other pathological conditions [44]; for example, an increase in the proportion of Firmicutes has been observed in obesity, while an increase in the proportion of Bacteroidetes has been observed in inflammatory bowel disease (IBD) [45]. The B/F ratio varies depending on the food source, where a higher relative proportion of Firmicutes is associated with HFD. This can be explained by the efficient extraction of energy from food by the Firmicutes than Bacteroidetes, thereby promoting more efficient absorption of calories by the host and the consequent weight gain [46]. Our results revealed a similar B/F ratio in the CD and EX groups; moreover, no significant difference was observed in the B/F ratio between the HFD and HFX groups. However, the B/F ratio drastically decreased by about 30-fold in the HFD group in comparison to the CD group (Figure 4a). Our results corroborate with the findings of the previous studies, which showed an increased abundance of Firmicutes in obese and HFD-fed mice compared to that in the lean mice [47][48][49]. In agreement with our results, another study showed an increase in the populations of Firmicutes and Proteobacteria in response to the consumption of unsaturated fat (74%) [50]. In the present study, a higher abundance of Proteobacteria was observed in the HFD group (p < 0.0001) compared to the CD group; this increased abundance has been reported to be a potential diagnostic marker for dysbiosis and other metabolic diseases [51]. In addition, we observed a significantly increased proportion of Cyanobacteria in the EX group only, compared to all other groups (Figure 4a). Cyanobacteria are primarily recognized as photosynthetic and have presumably shaped the earth's geochemical environment over some billions of years by generating oxygen as a by-product [52]. The bacterial taxa YS2, which was previously recognized as a member of Cyanobacteria, is now part of another taxa called 'Melainabacteria', based on recent genomic evidence that showed the absence of photosynthetic abilities in YS2. The Cyanobacteria-related YS2 demonstrates many functions, such as obligate anaerobic fermentation, nitrogen fixation, and the synthesis of vitamin B and K, as illustrated by the metabolic analysis. YS2 are common in the mammalian gut [53], and in the current study, it was found in higher proportions in the gut of mice that underwent exercise (EX) (Figure 6).
Paraprevotallaceae was significantly more abundant in the EX group compared to the CD group. An earlier report showed that this taxon has decreased relative abundance in female obese swine (Western diet-fed) than in lean swine (low-fat diet-fed) after 36 weeks of diet intervention [54]. The result of the present study and the earlier one can be explained by the difference in the diet intervention periods and the choice of experimental animals. The specific majority of the members of the S24-7 microbiota family that were identified by ANCOM and based on diet showed differential abundance in the highfat diet-fed groups (HFD, HFX) compared to the chow-fed groups (CD, EX) (Figure 5c). These species might be contributing to the significant expansion of the S24-7 family in the high-fat-fed groups, as illustrated in Figure 4b. The members of Desulfovibrionaceae and Streptococcaceae were significantly enhanced in the HFD group, and exercise effectively reversed the proportions of Streptococcaceae, as shown by the HFX group in Figure 4b. Desulfovibrionaceae include sulphate-reducing and endotoxin-producing bacteria, which are enhanced in HFD [22]. A recent report illustrated the increase of Desulfovibrionaceae and Peptostreptococcaceae in the HFD-fed mice, which was reversed by exercise [55]. Moreover, Streptococcaceae was associated with high-BMI individuals and a high-fat diet intake [56]. These reports support our data in demonstrating the beneficial impact of exercise on the gut microbiota of HFD mice.
Although Ruminococcaceae (Firmicutes), Lachnospiraceae (Firmicutes), Odoribacteraceae (Bacteroidetes), and Coriobacteriaceae (Actinobacteria) did not reach statistical significance, these bacterial taxonomic families were found to be bloomed in the high-fat diet-fed groups compared to the chow-fed groups (Figure 4b). An earlier report showed that the long-term intake of a high-fat diet increases Lachnospiraceae and Streptococcaceae bacteria in C57BL/6 mice, and inclines towards an increased inflammatory profile [57]. Similarly, at the genera level, Coprococcus (Lachnospiraceae) and Aldercreutzia spp. (Coriobacteriaceae) did not attain statistical significance but were found to be intensified in the high-fat diet-fed groups in comparison to the chow-fed groups (Figures 5a and 6). However, Desulfovibrio expansion in HFD was enhanced~1.2-fold in the HFX group. A study associated Butyricimonas, Desulfovibrio, and Oscillospira genera with a BMI < 25, that is, normal-weight people [56]. The possible discordance in the results can be attributed to the variability in the exercise methodology, selection of model organism, and choice of sequencing platform.
The random forest regression model has been reported to be one of the most effective supervised machine-learning algorithms for analyzing microbiome data [58]. The relative abundance of Adlercreutzia (Coriobacteriaceae, Actinobacteria) identified using random forest classification corroborates with the moderate expansion of Coriobacteriaceae at the family level in the high-fat diet-fed groups (Figure 4b). Moreover, exercise might have a role to play in the mild enhancement of Oscillospira (Ruminococcaceae) (Figure 5a, Figure 6), which might contribute to the mild amplification of its family Ruminococcaceae (Firmicutes), as implied by the results of the HFX group (Figure 4b). Consistent with our results, Oscillospira was found to be positively associated with the blood lactate levels in the exercised rats [59].
Faith's phylogenetic diversity metric of alpha diversity and the unweighted UniFrac (qualitative, phylogenetic) metric of beta diversity were found to be insignificant among the groups, indicating that the experimental interventions did not significantly affect the phylogenetic diversity of the gut microbiome (data not shown). The HFD and HFX groups showed higher diversity (Shannon) and evenness than the CD and EX groups ( Figure 2). This finding is supported by a recent study, which reported an increased gut microbial diversity as well as energy metabolism in mice fed an HFD for 14 weeks, compared to mice fed a low-fat diet for the same period. This was explained by studying the changes in the metabolic pathways. Specifically, the expression of genes involved in non-absorbed carbohydrate metabolism pathways were found to be increased in HFD mice, leading to the colonization of intestinal pathogens and inflammation [60].
Diet appeared to be a significant contributor to the alpha and beta diversities of the mouse gut microbiota, as examined by longitudinal ANOVA and ADONIS, when compared to exercise. Overall, the results showed a higher relative abundance of Lactococcus spp. in the guts of HFD mice than in those of CD mice (p < 0.0001). Interestingly, Lactococcus spp. abundance was significantly decreased in the HFX group compared to the HFD group (p = 0.0001) ( Figure 5). The results of the present study are supported by a recent report showing the role of lactic acid bacteria in metabolic syndrome. In a previous study, yeast β-glucan was administered to HFD-induced metabolic syndrome mice for 10 weeks, which resulted in the reduction of the relative abundances of Lactobacillus and Lactococcus, explaining the possible attenuation of the metabolic syndrome in mice [61]. Since Lactobacillus and Lactococcus play a critical role in HFD-induced metabolic syndrome, the exercise-mediated decrease in the abundance of Lactococcus spp. shown by our data was encouraging. Further studies investigating the functional readout of the observed changes in the gut microbiota are required to better understand the physiological phenomena associated with different dietary or exercise interventions in the context of pathology and metabolic syndromes.

Conclusions
The present study revealed that variable diets in the context of fat percentage induces significant alterations in the gut microbiota composition of mice. In general, the gut microbiota were altered under different dietary interventions, while the microbial community richness was reduced as expected, but the Shannon diversity index increased in response to HFD. The B/F ratio significantly decreased in the high-fat diet-fed groups compared to the chow-fed controls. The relative abundance of bacteria belonging to Proteobacteria and Firmicutes was higher in the HFD group. Lactococcus, which was rare in the CD group and relatively abundant in the HFD group, was significantly reversed in abundance in the HFX group. A similar pattern was shown by the bacterial taxonomic family Streptococcaceae.
Regardless of the comparably less significant effect of exercise, this study enabled us to resolve the taxonomic players at the genus level that contribute significantly to the changes observed at the phylum level in response to different diets. Paraprevotella, Desulfovibrio, Lactococcus, Butyricimonas, and Desulfovibrio C21c20 were the differentially abundant/present taxa based on diet. From the results of this study, diet has emerged as a major contributor to the gut microbial richness and diversity in mice. These differential responses require detailed mechanistic studies to explore the relationship between the gut microbes and the dietary interventions (with or without exercise), by exploring the underlying players involved in bacterial metabolism and signaling.  Informed Consent Statement: Not applicable.

Data Availability Statement:
The data generated during this study are available from NCBI under the accession number PRJNA846968.

Conflicts of Interest:
The authors declare no conflict of interest.