ABSTRACT

   The gut microbiota and its metabolic activities are crucial for
   maintaining host homoeostasis and health, of which the role of
   probiotics has indeed been emphasized. The current study delves into
   the performance of probiotics as a beneficial managemental strategy,
   which further highlights their impact on growth performance, serologic
   investigation, gut microbiota, and metabolic profiling in yaks’ calves.
   A field experiment was employed consisting of 2 by 3 factorial
   controls, including two development stages, namely, 21 and 42 days
   (about one and a half month), with three different feeding treatments.
   Results showed a positive impact of probiotic supplements on growth
   performance by approximately 3.16 kg (P < 0.01) compared with the blank
   control. Moreover, they had the potential to improve serum antioxidants
   and biochemical properties. We found that microorganisms that threaten
   health were enriched in the gut of the blank control with the depletion
   of beneficial bacteria, although all yaks were healthy. Additionally,
   the gut was colonized by a microbial succession that assembled into a
   more mature microbiome, driven by the probiotics strategy. The gut
   metabolic profiling was also changed significantly after the probiotic
   strategy, i.e., the concentrations of metabolites and the metabolic
   pattern, including enrichments in protein digestion and absorption,
   vitamin digestion and absorption, and biosynthesis of secondary
   metabolites. In summary, probiotics promoted gut
   microbiota/metabolites, developing precise interventions and achieving
   physiological benefits based on intestinal microecology. Hence, it is
   important to understand probiotic dietary changes to the gut
   microbiome, metabolome, and the host phenotype.

   IMPORTANCE The host microbiome is a composite of the trillion
   microorganisms colonizing host bodies. It can be impacted by various
   factors, including diet, environmental conditions, and physical
   activities. The yaks’ calves have a pre-existing imbalance in the
   intestinal microbiota with an inadequate feeding strategy, resulting in
   poor growth performance, diarrhea, and other intestinal diseases.
   Hence, targeting gut microbiota might provide a new effective feeding
   strategy for enhancing performance and maintaining a healthy intestinal
   environment. Based on the current findings, milk replacer-based
   Lactobacillus feeding may improve growth performance and health in
   yaks’ calves.

   KEYWORDS: microbiome, metabolomics, Lactobacillus, gut development, yak
   calf, gut development

INTRODUCTION

   The yak (Bos grunniens) is an ancient ruminant with a mystique of the
   plateau (>3,000 m), thriving under extreme living conditions, such as
   low temperature and hypoxia ([48]1). It has been connected intimately
   with the local human civilization and agriculture, providing basic
   survival resources, e.g., warm hides, dung for fuel, meat, and
   transportation ([49]2). The traditional grazing method relies mostly on
   natural pasture with limited supplementary feeding, so the yaks face
   deficient foraging resources on the plateau, especially during the
   calving season ([50]3). In this context, neonatal calves are extremely
   susceptible to intestinal diseases caused by enteric bacterial
   imbalance. It results in inefficient digestion diarrhea, and poor
   growth, that further cause high mortality and morbidity ([51]4).
   According to a number of recent studies, the mortality rate during
   early calves rearing remains remarkably high in most countries, e.g.,
   17% in Germany ([52]5) and 5% in United States ([53]6). Studies on milk
   or milk replacer (MR) feeding suggested that early feeding management
   is necessary to boost the immune system maturation, body development,
   and health of the calves ([54]7). Additionally, when mother’s milk and
   early feeding programs are implemented concurrently, the goal of better
   welfare during early period of rearing can be achieved easily ([55]8).
   Hence, we hypothesized that early nutrition management might be
   important for the improvement of growth performance and well-being in
   yaks’ calves.

   The healthy nutritional management through probiotics is considered to
   be a vital feeding strategy for the prevention of neonatal
   gastrointestinal disturbance ([56]9). Several studies have proved that
   probiotics effectively enhance immunity and feed efficiency and reduce
   diarrhea in neonate Holstein calves ([57]10, [58]11). The proposed
   mechanism for direct-fed probiotics includes organic acids, digestive
   enzyme activities, immune system stimulation, and the production of
   various antagonistic factors, such as hydrogen peroxide, bacteriocins,
   and diacetyl, of which all have an obvious inhibitory effect against a
   wide range of pathogenic bacteria ([59]12). The short-term therapy of
   Lactobacillus reuteri modulates intestinal microflora and improves
   proliferation, differentiation, immune defense, and barrier function
   formation in intestinal epithelial cells ([60]13, [61]14). Yan et al.
   demonstrated that neonatal colonization of Lactobacillus rhamnosus GG
   improves the functional maturation of intestines and conferred lifelong
   health outcomes by enhancing the effectiveness against intestinal
   damage and inflammation ([62]15). The potential role of certain
   Lactobacillus rhamnosus GG in regulating gut microbiota modulation and
   promoting intestinal metabolic profile is accomplished by the increase
   in beneficial microorganisms. So, the reduction of such microorganisms
   associated with health threats ([63]16). Additionally, commensal
   bacteria in the gut systemically effect the molecules of microbial
   metabolism, such as short-chain fatty acids ([64]17). These findings
   emphasized the need for identifying the role of certain Lactobacillus
   species that allow animals to develop a potential for intestinal
   maturation and growth. So, a comparable effect of Lactobacillus
   supplementation on stimulating neonatal intestine development is
   conceivable but requires further exploration in yaks’ calves.

   The development of the early intestinal microbiome acts as a bridge in
   the host-microbial relationship, which contributes to the maintenance
   of host health and normal physiological functions throughout the life.
   Ruminants have a massive and complex intestinal microbiome. Such an
   extraordinary symbiosis between a host and gut microorganism is linked
   to their mother’s vaginal microbiome, which influences the development
   of their immune system and nutrition, as well as early colonization to
   form the microbiota’s first colonizer community ([65]18). The enormous
   potential for interactions between intestinal microorganisms indicates
   that early colonizers could affect the establishment of microbiota in
   long term ([66]19). A recent study emphasizes the significance of
   intestinal microorganisms on the development and maturation of the
   neonatal immune system in the gut ([67]20). Among those different
   microorganisms that colonize the host intestines, some entangle with
   the body’s metabolism. For example, the potential metabolic capacities
   in the intestinal microbiome assist in absorbing the body’s own
   inaccessible energy from diet and help in biotransformation to various
   xenobiotics ([68]21, [69]22). The profiles of intestinal microbes may
   also alter the environment in the gut because of the metabolites. These
   microbes may be involved in immune system regulation and host-generated
   signaling ([70]23, [71]24). Similarly, abundant metabolites that
   execute extensive metabolic activities in the gut are used efficiently
   by microbes for their own proliferation ([72]25). As an example,
   tryptophan metabolites obtained from microbial sources might alter host
   physiology and behavior by decreasing the amount of tryptophan and
   activating the aryl hydrocarbon receptor by producing indole
   derivatives ([73]26). In addition, short-chain fatty acids (SCFAs),
   produced by gut microorganisms during the digestion of dietary fiber,
   may also influence intestinal epithelial cells by modulating cell
   proliferation, gene expression, and immune response ([74]27). Since the
   gut microbiota and metabolism have been connected to a wide range of
   disorders, it is not surprising that the intestinal microbiome and
   metabolism have been linked to a series of diseases ([75]28, [76]29).
   The discovery of such relation between the gut microbiome and
   metabolome has promoted extensive research to characterize its
   taxonomic diversity ([77]30). However, it is unclear how the
   composition and dynamics of microbiome and microbially mediated
   metabolic processes interact to maintain physiological benefits in
   calves. This limitation might be attributed to the variation in
   microbiome pathways and physiology adaptability between yak and other
   cattle species on plateau.

   In this study, a controlled-field experiment was performed to
   investigate the growth, serum antioxidants, serum biochemical
   indicators, and gut development (i.e., bacterial communities and
   metabolites) under a milk replacer-based Lactobacillus (Lactobacillus
   reuteri). The study focused on 16S rRNA gene sequencing and untargeted
   liquid chromatography-mass spectrometry (LC-MS) metabolomics profiling
   to identify the dynamic distribution and interaction of the gut
   bacterial communities and metabolites in calves. The findings allowed
   us to develop a possible way for sustaining the health and well-being
   of yaks’ calves.

RESULTS

Growth performance.

   [78]Fig. 1B presented values for initial and final body weight (BW). In
   our findings, the initial BW was not significantly different among
   treatment groups. Therefore, the difference in growth performance
   caused by the initial values was excluded. As expected, the final BW
   for calves receiving MR and milk replacer-based Lactobacillus (LAB-MR)
   tended to be higher than control calves (P < 0.05 or P < 0.01). Results
   showed a positive impact of LAB-MR on growth performance by
   approximately 3.16 kg (P < 0.01) and 0.8 kg (P > 0.05) compared with
   blank control and control groups, respectively. Additionally, a
   management strategy for milk replacer-based Lactobacillus was found to
   be a potential tool for improving serum antioxidant properties (i.e.,
   total antioxidant capacity [T-AOC], superoxide dismutase [SOD], and
   malondialdehyde [MDA]) ([79]Fig. 1C and [80]D) and serum biochemical
   parameters (i.e., aspartate aminotransferase [AST], total cholesterol
   [TC], and urea) (P < 0.05 or P < 0.01) ([81]Fig. 2C and [82]D).

FIG 1.

   [83]FIG 1
   [84]Open in a new tab

   Body weight and serum antioxidant properties of yaks’ calves,
   supplemented with LAB in three different groups. (A) yaks’ calves. (B)
   Initial and final body weight of calves. (C) Serum antioxidant
   properties of 21-day-old calves. (D) Serum antioxidant properties of
   42-day-old calves. *, P < 0.05; **, P < 0.01.

FIG 2.

   [85]FIG 2
   [86]Open in a new tab

   Effects of LAB-MR supplementation on serum biochemical indexes and
   alpha diversity of gut microbiome in three different groups. (A) Alpha
   diversity of 21-day-old calves. (B) Alpha diversity of 42-day-old
   calves. (C) Serum biochemical indexes of 21-day-old calves. (D) Serum
   biochemical indexes of 42-day-old calves. *, P < 0.05; **, P < 0.01.

Feeding strategy increased the diversity of the gut microbiome.

   We used amplicon sequencing to evaluate how the functional potential of
   the intestinal microbiome diversifies over the course of feeding
   strategy. Fecal samples from the calves at two different intervals (21
   and 42 days) were collected, which passed strict quality control for
   raw sequence data (2,882,125) to obtain effective sequences
   (2,749,133). The average number per sample was 76,365 from bacterial
   populations. Multisample evidence for rarefaction curve, Shannon index,
   species accumulate curves, and coverage values demonstrated that almost
   all the bacterial populations were identified. Moreover, the current
   sequencing depth sufficiently covered the diversity of microbial
   communities (see Fig. S1A to C in the supplemental material). Based on
   97% similarity, all of the V3/V4 regions of sequences were clustered
   into 2,166 operational taxonomic units (OTUs).

   The multiple alpha diversity indices were measured to investigate the
   overall differences of microbial community richness and diversity. The
   intestinal diversity in the blank control was significantly lower than
   that of calves based on feeding management ([87]Fig. 2A and [88]B). The
   same condition was noticed for other diversity measurements (for Chao1,
   WTC, WTM, and WTL yielded 373, 435, and 455; and WFC, WFM, and WFL
   yielded 397, 445, 490; for ACE, WTC, WTM, and WTL yielded 362, 437, and
   447; and WFC, WFM, and WFL yielded 387, 450, and 469; P < 0.05 or
   P < 0.01).

   To assess the difference between intestinal microbial communities in
   the three groups, we calculated β-diversity. The unweighted pair-group
   method with arithmetic mean (UPGMA) matrix distance and principal
   coordinates analysis (PCoA) clearly showed the variations in
   evolutionary information per sample (Fig. S1D and E). The similarity of
   each group showed a tendency of separation and partial overlap at the
   two different sampling points. The members in each group clustered in
   branches separately, suggesting that the early colonization of the gut
   microorganisms in juvenile animals is a nonfixed dynamically evolving
   structure. Additionally, the UPGMA matrix distance and PCoA clearly
   showed that the microbiota cluster in the blank control group was
   separated from the other groups ([89]Fig. 3). The samples in control
   and LAB-MR groups clustered separately with the gradually expanding
   distance of the microbiome.

FIG 3.

   [90]FIG 3
   [91]Open in a new tab

   Effects of LAB-MR supplementation on gut bacterial communities’
   structures. (A) Rarefaction curve; (B) rank abundance curve; (C)
   species accumulate curve; (D, E) represent the microbial similarity
   between groups by using PCoA scatter plot and UPGMA, respectively; (F,
   G) represent the analysis of microbial communities’ structures at the
   phylum and genus levels, respectively.

Feeding strategy optimized the structure and composition of the gut
microbiome in calves.

   We evaluated the relative proportion of dominant bacteria in all
   samples at different taxonomical levels (Fig. S1F and G). Following the
   phylum level, Firmicutes was the most dominant phyla regardless of
   sampling time point and feeding management, which consisted of over
   50.2% total sequences. Followed by Bacteroidetes in all groups (WTC,
   WTL, WTM yielded 18.02%, 34.75%, and 27.79%; WFC, WFL, and WFM yielded
   17.04%, 35.25%, and 21.51%), except for the WTC group, the high
   abundance was replaced by Proteobacteria (20.83%). While at the genus
   level, Faecalibacterium and Escherichia-Shigella were notably enriched
   in WTC and Bacteroides in WTM, WTL, WFM, and WFL.

   We next executed differential relative abundance analysis using
   Metastat to accurately explore the microorganisms among different
   groups ([92]Fig. 4). The relative abundance of several intestinal
   communities between control groups showed a significant shift at phylum
   and genus levels, e.g., phyla Fusobacteria, Tenericutes, Acidobacteria,
   and Deferribacteres and genera Bacillus, Barnesiella, Oscillospira,
   Clostridium_sensu_stricto_1, Anaerostipes, and Neisseria. On the first
   sampling, the relative abundances of phylum Patescibacteria and genera
   Bacteroides, Parabacteroides, Phascolarctobacterium, Blautia, and
   Anaerostipes were significantly higher in the WTL group. The gut
   microbiome of the WTM group had a higher proportion of Veillonella and
   Neisseria than that of the WTL group. Moreover, the abundances of
   phylum Bacteroidetes and genera Christensenellaceae_R-7_group,
   Lachnoclostridium_10, Clostridium_sensu_stricto_1, Bacteroides,
   Parabacteroides, Ruminococcaceae_UCG-004, Phascolarctobacterium,
   Blautia, and Anaerostipes were significantly increased in WTL compared
   with those of WTC, while Actinobacteria, Tyzzerella_4, Veillonella, and
   Neisseria were significantly reduced. On the second sampling, the
   beneficial functional communities gradually enriched and microorganisms
   that threaten health reduced in the intestines of the LAB-MR group. In
   a brief view, the phyla Bacteroidetes, Patescibacteria, Tenericutes,
   and Deferribacteres and genera Candidatus_Saccharimonas, Bacillus,
   Senegalimassilia, Odoribacter, Parabacteroides, and Anaerostipes were
   more abundant in the LAB-MR group than those in control who were taking
   milk replacer. The phyla Bacteroidetes and Patescibacteria and genera
   Candidatus_Saccharimonas, Alistipes, Lachnoclostridium_10,
   Defluviitaleaceae_UCG-011, Bacteroides, Ruminococcaceae_UCG-004,
   Saccharofermentans, Barnesiella, Faecalibacterium, Senegalimassilia,
   and Parabacteroides were enriched in LAB-MR compared with those in the
   blank control group. Furthermore, the [Clostridium]_innocuum_group and
   Porphyromonas were more common in the blank control group. These
   findings conveyed a message that the principal component of microbiota
   remained separate and exhibited partial overlap between control and
   blank control groups. At the same time, milk replacer-based probiotic
   feeding motivated a different intestinal microbiome profile, which was
   dominated by a variety of functional beneficial bacteria.

FIG 4.

   [93]FIG 4
   [94]Open in a new tab

   Significant differences in the relative abundance of gut microbiota
   among three different groups at phylum and genus levels. *, P < 0.05;
   **, P < 0.01.

Effects of Lactobacillus-feeding strategy on the gut metabolic profiling.

   Gut metabolite evaluation was performed in the positive and negative
   ion modes. A total of 668 and 927 different features were examined. All
   data from 18 samples, including quality-control (QC) samples were
   examined by PCA and orthogonal projections to latent structures
   discriminant analysis (OPLS-DA), following positive and negative mode
   ionization in order to characterize the overall metabolomic changes of
   different groups. As shown in PCA scatterplots, yak calves in the blank
   control and control were separated clearly from yaks in the LAB-MR
   group, regardless of ion modes or sampling ([95]Fig. 5I to [96]L).
   Additionally, a further OPLS-DA analysis showed that the three groups
   could be obviously separated according to their metabolic differences.
   Also, the data within the Hotelling T2 ellipse (with the class
   separation; R2Y) accounted for more than 0.991 ([97]Fig. 5A to [98]D).
   Meanwhile, the permutation test for OPLS-DA demonstrated that original
   points on the right were higher than R2 and Q2 values on the left.
   Also, the Q2 regression curve had a negative intercept, which shows the
   reliability and validity of the OPLS-DA model ([99]Fig. 5E to [100]H).

FIG 5.

   [101]FIG 5
   [102]Open in a new tab

   Different metabolic patterns among three different groups. OPLS-DA
   score plot of gut metabolome in the ESI+ (A, B) and ES− modes (C, D).
   The permutation test for OPLS-DA in the ESI+ (E, F) and ES− modes (G,
   H) demonstrated the reliability of current OPLS-DA model. PCA score
   plot in the ESI+ (I, J) and ES− modes (K, L) assessed the metabolites’
   similarity between groups.

   The gut metabolome was changed significantly after the feeding of
   LAB-MR. The differential metabolites between FC (i.e., WFC) and FW
   (i.e., WFL) (150 positively ionized metabolites and 224 negatively
   ionized metabolites) and differential metabolites between FM (i.e.,
   WFM) and FW (185 positively ionized metabolites and 81 negatively
   ionized metabolites) were identified (see Table S1, Table S2, and Fig.
   S2A to D in the supplemental material). The 150 different metabolites
   between FC and FW could be classified mainly into 71 downregulated and
   79 upregulated metabolites. Also, the 185 different metabolites between
   FM and FW were classified into 16 downregulated metabolites and 169
   upregulated metabolites for the positive ionization analysis (Fig. S2A
   and B). The pathway enrichment analysis showed a total of 7 metabolic
   pathways that had a significant change (P < 0.05 or P < 0.01),
   including estrogen signaling pathway, GnRH secretion, steroid
   biosynthesis, cAMP signaling pathway, axon regeneration, thermogenesis,
   and shigellosis in the comparison between FC and FW. Vitamin digestion
   and absorption, glycerophospholipid metabolism, choline metabolism in
   cancer, novobiocin biosynthesis, alpha-linolenic acid metabolism,
   tropane metabolism, piperidine, and pyridine alkaloid biosynthesis were
   significantly changed between FM and FW ([103]Fig. 6A and [104]B). As
   shown in the volcano map in negative ionization, the 224 different
   metabolites between FC and FW could be classified mainly into 28
   downregulated and 196 upregulated metabolites. A total of 77
   upregulated and 4 downregulated metabolites were identified between FM
   and FW (Fig. S2C and D). Among different metabolic pathways (P < 0.05
   or P < 0.01), five pathways with a significant impact value were
   pyrimidine metabolism, lipopolysaccharide biosynthesis, carbon fixation
   in photosynthetic organisms, ascorbate and aldarate metabolism,
   chlorocyclohexane, and chlorobenzene degradation between FC and FW. The
   other three pathways with significant impact were secondary bile acid
   biosynthesis, phenylpropanoid biosynthesis, and 2-oxocarboxylic acid
   metabolism between FM and FW ([105]Fig. 6C and [106]D).

FIG 6.

   [107]FIG 6
   [108]Open in a new tab

   Differential gut metabolites and metabolic pathways driven by LAB-MR
   supplementation. The metabolic pathways that were significantly changed
   in the ESI+ (A, B) and ES− modes (C, D). Volcano map showed the
   metabolites that were significantly up- or down- abundant driven by
   LAB-MR supplement in the ESI+ (E, F) and ES− modes (G, H).

Correlation analysis between differential intestinal microbes and
metabolites.

   The multiomic nature of the data set identified the microbial
   characteristics and metabolites that were significantly different in
   their abundance after LAB-MR feeding. There might be a mechanism in
   place that connects metabolite concentrations to the abundance of
   microorganisms; however, this mechanism may be ameliorated by LAB-MR
   feeding strategies. For example, an increase in microorganism abundance
   may be accompanied by an increase in the concentrations of some
   by-products. To identify such a covariant relationship, we performed
   Spearman correlation analysis between altered fecal metabolites and
   perturbed intestinal microbes. The results showed that the increases in
   the genera Candidatus_Saccharimonas, Barnesiella, Odoribacter, and
   Parabacteroides were positively correlated with the upregulated
   metabolites of l-isoleucine, l-proline, and l-tyrosine, whereas it was
   negatively correlated with the downregulated metabolites, e.g.,
   phosphorylcholine and choline ([109]Fig. 7). Another metabolite was
   included in Tyzzerella_4 genera, which had an unfavorable association
   with the aforementioned metabolites. Therefore, the changes in
   intestinal metabolites might associate with the alteration of gut
   microbiota to LAB-MR feeding.

FIG 7.

   [110]FIG 7
   [111]Open in a new tab

   Potential correlations between LAB-MR strategy-linked microbes and
   metabolites in the ESI1 and ES2 modes. (A, B) Represented covariation
   associations between FC and FW groups. (C, D) represented covariation
   associations between FM and FW groups.

DISCUSSION

   The gut microbiome is home to trillions of microbes that play a vital
   role in maintaining host homoeostasis ([112]31). Although microbiota
   colonization is important, its composition and structure induce
   metabolic variations that further may cause alterations in phenotypes
   ([113]32). Such colonization of the intestinal microbiome is highly
   malleable and influenced by different factors, e.g., growth,
   environmental conditions, diet, and physical activities ([114]1). The
   conjoint analysis of microbiome and metabolome has been recognized as
   the most promising method for assessing host-microbiome interactions
   ([115]33). However, few studies have used this method in yak calves. In
   the current study, we used a feeding strategy to examine the
   physiological effectiveness of Lactobacillus on the bacterial community
   and metabolic profile in yaks’ calves. First, we confirmed the role of
   growth performance, serum antioxidant, and serum biochemical
   indicators. Then we provided a detailed evaluation of the microbiome
   and metabolome mediated by Lactobacillus, which contributed to the
   identification of a possible feeding strategy for maintaining host
   health. The current study provided the first evidence that
   Lactobacillus-based milk replacer might be a potential feeding strategy
   for modulating the composition and relative abundance of intestinal
   microorganisms and metabolites. Also, we found multiple links between
   the gut microbiome (structure and taxonomic composition) and intestinal
   metabolites.

   Supplementing Lactobacillus has been reported to improve feeding
   efficiency and body weight by modulating intestinal microflora,
   limiting pathogen invasion, and increasing villus height ([116]34,
   [117]35). Similarly, Cox et al. demonstrated that probiotic promotes
   steady gut microbiota, stimulating digestibility and mucosal immunity
   ([118]12). Timmerman et al. emphasized that Lactobacillus
   supplementation works best during early stressful periods. This
   hypothesis was confirmed by feeding five Lactobacillus and one
   Enterococcus faecium ([119]36). In another study, probiotics were given
   to newborns and young animals, which were proven to induce some
   therapeutic effects in reducing the duration of intestinal disease
   ([120]37). However, the effect of supplementing probiotics on the
   growth performance of the host seems uncertain. Szabo et al. reported
   that pigs supplemented with Enterococcus faecium had no difference in
   weight gain compared with negative-control pigs after challenge with
   Salmonella enterica serovar Typhimurium ([121]38). These inconsistent
   responses might be caused by many potential factors, including pathogen
   loads, animal species, animal stress, and environmental factors
   ([122]39). In our current findings, among three treatments groups,
   calves fed LAB-MR had more body weight than the blank control group
   (P < 0.01), whereas calves fed MR were intermediate among the three
   groups. It might be due to the crude protein (CP) content of the milk
   replacer that drove the young animals’ frame growth ([123]40). It could
   also be correlated with improving gastrointestinal tract development
   ([124]41). This study provided the first evidence of improvement with
   Lactobacillus-based milk replacer in yak calves. Also, it further
   confirmed the possibility of LAB-MR as a feasible feeding strategy
   through serum antioxidant and serum biochemical indicators.

   Animals develop and maintain a complex host-specific gut microbial
   community that includes three domains of life, i.e., Archaea, Bacteria,
   and Eukarya ([125]19). The gastrointestinal tract of a newborn is
   colonized immediately by various microorganisms after birth. The
   process of such colonization has been identified as a coevolution,
   caused by the interaction between host and microbes ([126]42, [127]43).
   This remarkable symbiosis, formed by a variety of environmental
   factors, began during direct contact with the maternal vaginal
   microbiome during birth ([128]18). The gut microbiome is a community of
   microorganisms playing a vital role in most physiological processes
   ([129]44), while complex microbiotas also have a role in providing
   fertile ground for noncommunicable diseases and infections. For
   example, the gut microbiome has been linked to a variety of ailments,
   such as inflammatory bowel disease and irritable bowel syndrome
   ([130]45, [131]46), whereas others are scarcely thinkable, e.g., autism
   spectrum disorder and Parkinson’s disease ([132]47, [133]48). A pattern
   of abrupt dynamic changes in microbiota is correlated with
   interference, which demonstrates a deterministic mechanism throughout
   the development ([134]49). These shifts might be explained by the
   developing immune system, diet, and the initial effects of microbial
   colonies. Furthermore, the nondigested nutrient contents are
   catabolized by gut microorganisms. As a result, metabolites are
   transported, absorbed, or excreted by highly dynamic metabolic pathways
   ([135]50). Previous research has revealed that the characteristics and
   concentration of gut metabolites contribute to our understanding of the
   effect of metabolism on their host ([136]51). Importantly, intestinal
   metabolites nourish the gut epithelial cells as well as adjusting
   downstream signaling pathways ([137]52), which further acts as a bond
   between the gastrointestinal tract and host health. It implies that
   intestinal microflora dysbiosis probably is not the sole inducer for
   changing host physiology, as the gastrointestinal tract-derived
   metabolites also have significant systemic impact, e.g., immune system
   modulation. Therefore, it is important to investigate host gut
   microbiome-metabolite interactions within specific animal species and
   management practices. Previous research reported that the
   administration of Lactobacillus to newborn calves increased weight gain
   and decreased diarrhea incidence. These efficiencies were greater in
   early-weaned calves than those in adult calves, indicating more
   effectiveness for intestinal communities ([138]53).

   Research on the gut microbiome has provided resources that connect
   intestinal microbiome interactions to health-related results. Indeed,
   in this cross-sectional cohort, we found that calves fed with LAB-MR
   had the highest alpha indices among three groups (P < 0.05 or
   P < 0.01), reflecting the richness and diversity of bacterial
   communities. Previous research found that the microbiome
   interindividual variation reduced with age except microbial diversity
   ([139]54). Newborn or young animals are more susceptible to infections
   or intestinal diseases, resulting in high morbidity and mortality. But
   this condition eases gradually with age ([140]55). So, we hypothesized
   an association between gut microbiota diversity and susceptibility
   after using Lactobacillus supplementation in order to overcome
   intestinal diseases. Moreover, six branches in the evolutionary tree
   were associated with feeding strategy, revealing an adaptability in
   calf intestinal microbiota that was linked closely to dietary shift
   resources. However, host demographics, individual temperament factors,
   and infectious diseases all had a certain degree of influence on
   microbiome structure ([141]56). As a result, more clarification of the
   gut microbiome among the yak calves was needed to adequately
   investigate the role of Lactobacillus supplementation in intestinal
   microbiota acquisition.

   The differences of specific microorganisms intuitively reflected the
   intrinsic connection between LAB-MR supplement and gut microbiota
   composition. Our study found that calves in the LAB-MR group had the
   most abundant phyla (Bacteroidetes) and genera (Lachnoclostridium_10,
   Bacteroides, Parabacteroides, and Anaerostipes), regardless of sampling
   time. For young ruminants, Bacteroidetes play an important role in
   degrading carbohydrates and proteins for facilitating gastrointestinal
   immune system ([142]57). Bacteroides were identified as potential
   microorganisms to regulate the intestinal environment for
   immunomodulation and healthy homeostasis. At the same time,
   Lachnoclostridium_10 has been reported to work in response to changes
   in gut luminal proteins ([143]58, [144]59). The members of genus
   Anaerostipes ferment xylitol to produce butyrate, a nondigestible
   carbohydrate which plays an important role in improving barrier
   function, while Parabacteroides is a producer of short-chain fatty
   acids ([145]60, [146]61). As mentioned previously, a series of
   beneficial microorganisms are positively involved in regulating
   intestinal function and the immune system and reducing susceptibility
   to intestinal diseases ([147]62). It conveys a message that the
   intestinal environment, which is more vulnerable to disease, drives the
   reduction of beneficial microbiota or the reduction in beneficial
   microorganisms. These findings might be the consequences of the host’s
   gut microbiome toward a better structure, benefited by Lactobacillus
   supplementation. Furthermore, the largest percentage of Actinobacteria
   was detected in the blank control group, which may have been
   transformed readily into pathogenic bacteria when synergy with one
   partner or host was seen ([148]63). In addition,
   [Clostridium]_innocuum_group, Tyzzerella_4, Porphyromonas, Veillonella,
   and Neisseria were enriched in the blank control of calves. These
   bacteria participate in gut bacterial dysbiosis and diseases
   propagation through opportunistic pathogens ([149]64[150]–[151]68).
   Meanwhile, the functional microbes that maintain intestinal health or
   produce short-chain fatty acids were less abundant in the WFC group,
   i.e., Candidatus_Saccharimonas, Alistipes, Lachnoclostridium_10,
   Defluviitaleaceae_UCG-011, Bacteroides, Ruminococcaceae_UCG-004,
   Bacillus, Saccharofermentans, Barnesiella, Faecalibacterium,
   Senegalimassilia, and Parabacteroides, than those in the WFL group
   ([152]69[153]–[154]71). In animal husbandry, gastrointestinal
   dysfunction and even diarrhea in juvenile ruminants are frequent, which
   negatively affects growth performance and even leads to the death.
   According to Steele et al., several bacteria have been alternating
   between weak and dominating communities, resulting in gut diseases
   ([155]72). As a result, we hypothesized some inevitable associations
   between the immature gut microbiome and the susceptibility to
   intestinal diseases. This approach might explain why juvenile ruminants
   were more vulnerable than adults. Altered abundances of some
   microorganisms contributed to the unique microbial outcomes in
   LAB-MR-fed calves. Most prominently, the microbes associated with
   digestion and absorption of protein or saccharolytic (Bacteroidetes,
   Senegalimassilia), intestinal environment health (Deferribacteres,
   Patescibacteria, Tenericutes, Candidatus_Saccharimonas, Barnesiella,
   Bacteroides, Phascolarctobacterium, and Blautia), and SCFAs producers
   (Oscillospira, Odoribacter, Parabacteroides, and Anaerostipes) were
   noticed ([156]73, [157]74). A recent study supported the potential use
   of probiotics in routine husbandry practices ([158]75).

   The most common varieties of microorganisms that are incorporated into
   husbandry include Lactobacillus spp., Bifidobacterium spp.,
   Streptococcus spp., Bacillus spp., and Enterococcus spp. Normally, the
   prophylactic application is based on the conceded mechanism of
   Lactobacillus with the restoration or establishment ability of a
   beneficial microbiota in the gastrointestinal tract of young animals. A
   meta-analysis investigated the effect of probiotic administration on
   the fecal microbiota and health of calves. According to the results,
   the microorganisms that threaten health were reduced ([159]76). These
   reports support our findings in this study, suggesting that shifts in
   these functional gut bacteria are related to the better feeding
   management (LAB-MR). Moreover, the dynamic distribution, interaction of
   the gut microbiome, and metabolites are still needed to be evaluated
   under the effect of Lactobacillus administration in calves.

   Furthermore, the impact LAB-MR supplementation on the gut metabolome
   was assessed using an untargeted metabolomics technique, i.e., LC-MS.
   Metabolites drive the crucial cellular functions, for example, energy
   production and storage, which affects gut conditions. Metabolomics can
   detect subtle alterations in biological pathways because of its
   inherent sensitivity to investigate the potential mechanism of various
   physiological conditions. We performed an enrichment analysis to
   evaluate broad classes of metabolites that were significantly up- or
   downregulated. In total, a considerable number of metabolites were
   changed due to LAB-MR. The pathway enrichment analysis showed that
   metabolites, including pyrimidine metabolism, lipopolysaccharide
   biosynthesis, carbon fixation in photosynthetic organisms, ascorbate
   and aldarate metabolism, chlorocyclohexane and chlorobenzene
   degradation, estrogen signaling pathway, GnRH secretion, steroid
   biosynthesis, cAMP signaling pathway, axon regeneration, thermogenesis,
   and shigellosis were altered significantly between FC and FW. Finally,
   significant differential metabolites were identified by overlapping the
   significant metabolites and annotated metabolites in enriched pathways
   (P < 0.05; variable influence on projection [VIP], >1). Here, the
   ascorbic acid and orotic acid (OA) were significantly high in FW,
   whereas dihydrouracil and 4,4-dimethyl-5alpha-cholest-7-en-3beta-ol had
   the opposite trend. Since the discovery of ascorbic acid (vitamin C) in
   1920s, no other chemical has been proven to have an extraordinary
   effect ([160]77). It is an antioxidant, which efficiently scavenges
   reactive oxygen species (ROS) and toxic free radicals. OA is another
   essential and versatile molecule for regulating genes that involve the
   development of host cells and tissues. Early nutrition research
   identified it as vitamin B13, which is a precursor in the biosynthesis
   of pyrimidines ([161]78). Gastrointestinal toxicity has been attributed
   to dihydropyrimidine dehydrogenase (DPD) deficiency when fluorouracil
   is administered to an absolute and partial DPD deficiency population
   ([162]79), while 4-alpha-methyl-5-alpha-cholest-7-en-3-beta-ol has been
   reported to induce oxidative stress to the host ([163]80). It is well
   known that young yaks can easily face gastrointestinal disorders due to
   harsh living conditions (cold and hypoxia), stress response, unhygienic
   forage, or weather mutation ([164]81). We speculated that the
   metabolites that endanger host health can disturb intestinal
   microflora, which further increases their vulnerability to intestinal
   diseases. Moreover, we found that LAB-MR supplementation significantly
   increased the concentration of several metabolites associated vitamins,
   e.g., thiamine, riboflavin, and phylloquinone. Thiamine (vitamin B[1])
   is synthesized by plants, fungi, and microorganisms. Its derivatives
   have a nonenzymatic role in regulating stress response and signal
   transduction pathways, associating with harsh environmental factors
   ([165]82). Riboflavin belongs to the B vitamin family, and its intake
   probably has protective effects on a series of medical conditions, such
   as ischemia, while these biological effects have been investigated
   widely for their anti-inflammatory, antioxidant, and antinociceptive
   properties ([166]83). Phylloquinone (vitamin K1) plays a necessary role
   in bone and vascular metabolism ([167]84), while various clinical
   abnormalities occur under the conditions of biotin (vitamin H)
   deficiency, which include growth retardation and dermatological
   abnormalities ([168]85). In the FW group, we observed an increased
   level of metabolites associated with amino acids, i.e., l-isoleucine,
   l-proline, l-tyrosine, and l-phenylalanine, of which all have specific
   functions that make them beneficial to the host. In the
   gastrointestinal tract, amino acids, which are necessary precursors for
   synthesizing proteins and polypeptides, have been identified as markers
   of protein metabolism ([169]86). In the healthy gut, chenodeoxycholates
   are decomposed to secondary bile acids, contributing to the digestion
   of lipids ([170]87), while the ursodeoxycholic acid (formed from the
   transformation of deoxycholic acid) contributes to regulating lipid
   metabolism and intestinal barrier integrity ([171]88). Several
   molecular classes, e.g., nicotinic acid, novobiocin, and sodium
   deoxycholate, are associated with moderating intestinal health. These
   classes inhibit pathogenic growth that significantly depletes in FM
   compared with FW ([172]89, [173]90). The gut microbiome interacts with
   the host in a variety of ways, of which one is via a range of
   metabolites (created as end or intermediate products of microbial
   metabolism). The integrative analysis described the potential
   correlation between gut microorganisms and metabolites, driven by
   LAB-MR feeding strategy. The changes and potential relations, along
   with the previously mentioned beneficial amino acids and vitamins, are
   consistent with previously suggested host gut homoeostasis and health
   from the prospective of potential probiotic properties of Lactobacillus
   supplementation ([174]36).

Conclusion.

   Overall, the changes in final body weight suggested milk replacer-based
   Lactobacillus had a positive impact on growth performance. The feeding
   in yaks’ calves had the potential for improving serum antioxidant
   properties and serum biochemical parameters. The bacterial community of
   LAB-MR calves exhibited higher diversity and richer symbiotic
   microorganisms than that of the control groups. Additionally, the gut
   was colonized by a succession of microbiota that assembled into a more
   mature microbiome driven by LAB-MR. The gut metabolic profiling was
   also significantly improved after LAB-MR, i.e., the concentrations of
   metabolites and the metabolic pattern. The current study involved a
   conjunction of gut metabolomics and bacterial community analyses
   between functional bacteria and metabolites in yaks’ calves that were
   significantly prompted by a milk replacer-based Lactobacillus feeding
   strategy. Such integrative information contributed to the development
   of efficient, healthy, and ethical modern husbandry strategies for
   yaks’ industry.

MATERIALS AND METHODS

Field methods.

   This experiment was conducted at Maiwa Yak Breeding Base in Aba Tibetan
   and Qiang Autonomous Prefecture (>3,500-m altitude), in the eastern
   part of the Qinghai-Tibet Plateau Sichuan, China. All calves were born
   and enrolled at the experimental unit on the same day. After visual
   inspection, 18 competent calves (0 d, born within 24 h) with similar
   health status and immunization were recruited.

   Calves were assigned randomly for further treatments by a veterinary
   technician, who was unaware of the experiment. Three groups were
   allocated with 6 calves per treatment (half male and half female) at
   3 ± 1 day of age. Treatments were as follows: (i) control group (WTM
   and WFM), mother’s milk + milk replacer (MR) at 0.5 L/day MR per calf
   (06:00 h and 18:00 h); (ii) probiotics group (WTL and WFL), mother’s
   milk + milk replacer-based Lactobacillus (LAB-MR) at 10 g/day LAB
   product isovolumetric to MR (06:00 h and 18:00 h); and (iii) blank
   control group (WTC and WFC), mother’s milk. The LAB product was a
   water-soluble powder that provided 1 × 10^8 CFU/g of Lactobacillus
   reuteri. The potential probiotic properties of this strain had been
   described according to our previously published study ([175]14). The
   nonmedicated MR containing 20% CP, 17% fat, and 10% ash content was
   purchased from Chengdu Xingguang Quan Nutrition Food Co., Ltd. with
   producer no. Q/91510115794900448U.3-2020. The MR was reconstituted with
   warm water (46°C) to attain a solid concentration of 10%. Before
   animals were fed, the incubator was used at 37°C to ensure that MR or
   LAB-MR were homogeneous. Upon completion of supplement consumption, the
   bucket was rinsed and kept sterile every day. Numbered necklaces
   instead of ear tags were assigned to calves for providing natural
   habits of the calves before letting them graze with their dams. Other
   than the experimenter, the entries were limited to two herders and a
   staff veterinarian. The job of the two herders was to find the calves
   with necklaces twice/day (morning and afternoon) to execute the feeding
   protocol and then lift all the calves back into the herd. Calves were
   assumed to have fed from the same mother, which was an uncontrollable
   factor. Moreover, there is no record of calves’ grass intake, as it was
   impossible to track in the natural pasture. Furthermore, all the yaks
   drank water from the ranch’s mountain spring.

Body weight, rectal temperature, and blood sample collection.

   The body weight of all calves was measured before the morning feeding.
   Rectal temperature was measured four times per week (Monday, Wednesday,
   Friday, and Sunday) by using a handheld thermometer before morning
   feeding. Blood samples for hematology analysis were collected two times
   during the experiment (day 21 and day 42) by using a 21-gauge needle
   via jugular venipuncture into no-additive evacuated tubes. The
   centrifugation of all samples at 3,000 × rpm/min was performed just
   after 30 min at room temperature, and then serum samples were stored at
   −80°C in quadruplicates until hematology analysis. The analyzed
   indicators were alanine aminotransferase (ALT), aspartate
   aminotransferase (AST), glucose (Glu-G), albumin (ALB), triglycerides
   (TG), total cholesterol (TC), total protein (TP), creatinine (CREA-S),
   urea, calcium (Ca), phosphorus (P), and magnesium (Mg).

Fresh fecal samples collection.

   Fresh fecal samples were collected on day 21 (named blank control;
   control; and LAB-MR groups WTC, WTM, and WTL) and day 42 (named blank
   control; control; and LAB-MR groups WFC, WFM, and WFL) from each calf’s
   rectum via digital stimulation or fresh defecation. Then samples were
   placed immediately in liquid nitrogen and shipped on dry ice to the
   laboratory. We performed metagenomics profiling and untargeted
   metabolomics on three cohorts. The fecal samples were stored at −80°C
   prior to high-throughput sequencing or metabolomics profiling as
   described below.

Microbiome sample processing and sequencing.

   Total genomic DNA in fecal samples collected at days 21 and 42 was
   extracted using the QIAamp fast DNA stool minikit (Qiagen, Inc.),
   according to the manufacturer’s instructions. Then, a 1% agarose gel
   was used to monitored genomic DNA concentration and purity. To perform
   the subsequent pyrosequencing, the V3-V4 region of bacterial 16S rRNA
   gene was amplified using universal primers (338F,
   5′-ACTCCTACGGGAGGCAGCA-3′; and 806R, 5′-GGACTACHVGGGTWTCTAAT-3′). The
   PCR amplification was executed at following conditions: the annealing
   temperature was 57°C during the 25 PCR cycles. The PCR products were
   confirmed by 1% gel electrophoresis, cleaned, and normalized using the
   SequalPrep normalization plate kit (Life-Technologies, CA). According
   to the manufacturer’s instructions, purified amplification PCR
   productions were generated in a sequencing library using the Next Ultra
   DNA library prep kit (New England BioLabs [NEB], USA). After quality
   inspection using a bioanalyzer (Agilent Technologies, USA) and
   quantitative PCR (qPCR), only libraries with a single peak and
   concentration of more than 2 nM were kept and used for high-throughput
   sequencing. The qualified library was sequenced on the HiSeq 6000
   platform (Illumina, San Diego, CA), targeting the sequences with
   paired-end reads.

Bacterial metagenome bioinformatics and statistical analysis.

   The raw reads from high-throughput sequencing were quality filtered to
   obtain effective reads through performing the following preprocedures.
   Primer sequences were trimmed to obtain clean reads by using cutadapt
   1.9.1 software, and then UCHIME v4.2 software was used to identify and
   remove chimera sequences. Furthermore, the effective sequence alignment
   performed with the SILVA database was clustered into operational
   taxonomic units (OTUs) using Quantitative Insights into Microbial
   Ecology (QIIME) software (Uparse v7.0.1001), with a threshold of ≥97%
   sequence similarity. The taxonomic assignment of all sequences and
   microbial composition were analyzed based on normalized output data.
   The calculations of alpha diversity (Chao1, ACE, Shannon, and
   PD_whole_tree) were performed in QIIME software. The rarefaction curve,
   Shannon curve, rank abundance curve, species accumulate curve, and
   Good’s coverage were visualized in R software (v3.6.0) to demonstrate
   species abundance and evenness and reflect whether the current
   sequencing depth covered the vast majority of species information.
   Principal coordinates analysis (PCoA) and unweighted pair-group method
   with arithmetic mean (UPGMA) investigated the similarities between
   groups or individuals. Metastats analyses (performed t test) were
   conducted at different taxonomic levels (phylum and genus) for
   assessing the difference in the relative abundance of intestinal
   microbiome members for finding biomarkers between groups of samples.

Stool sample processing and metabolite profiling analysis.

   Fecal sample metabolomics profiles were performed using liquid
   chromatography-tandem mass spectrometry (LC-MS/MS). Chromatographic
   analysis was executed on the Waters Acquity ultra-high-performance
   system (1290 UHPLC; Agilent), and high-resolution mass spectrum (HRMS;
   TripleTOF 5600; AB Sciex) enabled the nontargeted measurement of
   metabolites. Briefly, stool samples (about 50 mg) were triturated in
   precooled methanol (CNW Technologies) by using a bead mill
   (TissueLyser; Qiagen). Next, the mixtures were incubated successively
   at 0°C for 10 min and −20°C for 1 h and then were centrifuged at
   13,000 rpm at 4°C for 15 min. The sample supernatant (20 μL) was
   collected and mixed together to 200 μL as a quality control (QC) sample
   and then was injected into the LC-MS/MS system for further analysis.

   The mass spectrometry data were obtained from the AB 5600 TripleTOF
   mass spectrometer under the control of Analyst software (Analyst TF
   1.7; AB Sciex). Time-of-flight (TOF) parameters were set as follows:
   bombardment energy, 30 eV. The electrospray ionization (ESI) ion source
   parameters are set as follows: atomization pressure (GS1), 60 Psi;
   auxiliary pressure, 60Psi; air curtain pressure, 35 Psi; temperature,
   650°C; and spray voltage, 5,000 V (positive ion mode) or −4,000 V
   (negative ion mode). Raw LC-MS data were proceeded using Genedata
   Expressionist software (v9.0) to remove chemical noise. The procedure
   included chromatographic peak detection, integration, normalization,
   and alignment retention times between samples. The processed data were
   then used to execute principal-component analysis (PCA) and orthogonal
   to partial least-squares discriminate analysis (OPLS-DA) for
   visualizing the metabolic profiling differences between groups. Taken
   together, the significantly different metabolites were considered
   regarding variable influence on projection (VIP) of >1 and a P value of
   <0.05.

Statistical analysis.

   The SPSS (v21.0) software was used to perform statistical analyses.
   GraphPad Prism (v7.0) software was used to draw box plots. The values
   were presented as mean ± standard (SD). Statistical significance was
   identified as a P value of <0.05.

Ethics approval.

   The animal-specific procedures were approved by the animal ethical
   committee of Huazhong Agricultural University. The experiments did not
   involve any invasive operations on animals.

Data availability.

   The raw sequence reads were submitted to the NCBI public database (SRA)
   with accession no. [176]PRJNA818126.

Supplementary Material

   Reviewer comments
   [177]reviewer-comments.pdf^ (205.2KB, pdf)

ACKNOWLEDGMENTS