Development of Next-Generation Probiotics by Investigating the Interrelationships between Gastrointestinal Microbiota and Diarrhea in Preruminant Holstein Calves

Simple Summary The present study investigated the relationship between gastrointestinal microbiota and diarrhea in preruminant calves by using immune-related markers and further isolating specific bacterial strains, enriched in clinically healthy individuals, for potential next-generation probiotics. The gathering of microbiomic data strongly indicated the possible beneficial effects of Bifidobacterium longum subsp. longum. With further screening and isolating with immunomodulatory and antagonistic effects, two Bifidobacterium longum subsp. longum strains might be expected to emerge as next-generation probiotics. The finding here might provide a solution for preventing gastrointestinal disorders for preruminant calves without sustained periods of administration through inhibiting the infectious bacteria, immunomodulatory effect and possible modulating microbiota. Abstract (1) Background: We aimed to isolate and identify potential next-generation probiotics (NGP) by investigating the interrelationships between gastrointestinal microbiota and diarrhea in preruminant Holstein calves. (2) Material and methods: Twenty preruminant Holstein calves were divided into healthy and diarrheic groups after the combination outcomes of veterinary diagnosis and fecal scores. The fecal microbiome, plasma cytokines, plasma immunoglobulin (Ig) G and haptoglobin were analyzed. The potential probiotic bacteria were identified by comparing the microbiota difference between healthy and diarrheic calves and correlation analysis with fecal scores and inflammatory markers. The identified bacteria were also isolated for further evaluation for antimicrobial activities and immunoregulatory effects. (3) Results: Microbiota analysis suggested that Ruminococcaceae_UCG_014, Bifidobacterium and Pseudoflavonifractor positively correlated with bovine IgG and negatively correlated with fecal score; inflammatory factors, bovine HP, and IL-8 were classified as beneficial bacteria contributing to the health of the calves. The alternation of gut microbial composition also induced changes in the functional gene enrichment of gut microbiota in calves. The gathering of microbiomic data strongly indicated the possible beneficial effects of Bifidobacterium longum subsp. longum, expected to develop as NGP. After isolation and evaluation of the potential functionality in vitro, two specific bifidobacterial strains demonstrated antimicrobial activities and immunoregulatory effects. (4) Conclusions: The results provide a new probiotic searching approach for preventing gastrointestinal disorders in preruminant calves. Further animal study is necessary to verify the results.


Introduction
Calf diarrhea, characterized by the frequent removal of soft feces, mostly impacts individuals under 6 weeks of age and their subsequent development [1]. The causes involve infectious agents, including viruses (bovine coronavirus, rotavirus, bovine viral diarrhea virus), bacteria (Escherichia coli, Salmonella spp.), and protozoa (Eimeria zuernii), which are mainly transmitted from the feces of infected animals to the mouths of susceptible animals [2]. The non-infectious parameters, such as stress, the nutritional and immunological conditions, and the production systems of young calves are also crucial for the incidence of diarrheic calves [2]. The treatment usually involves oral rehydration products and antibiotics. Accumulating evidence has indicated that the use of antibiotics in farm animals is associated with many adverse effects [3].
Several pieces of evidence suggested that the reconstitution of a healthy microbial community is an effective approach to prevent or treat gastrointestinal disorders [4]. The use of beneficial probiotics has been recognized to prevent the dysbiosis of intestinal microbiota and the establishment of pathogenic microbial populations. Many studies have highlighted that health and growth were the most positively pretentious responses to probiotic supplementation [5][6][7][8]. The most commonly used probiotic fed to young calves is live yeast, mainly Saccharomyces cerevisiae [9], and bacterial-based probiotics such as Lactobacillus spp., Enterococcus spp., and Bacillus spp. [10]. The feeding dosage varies depending on the probiotic strains and animal conditions, but sustained periods of administration are needed due to the fecal colonization of the probiotics decreasing with time.
However, a consensus regarding whether probiotics can in fact reduce the incidence of gastrointestinal diseases in young calves has not been reached. Alawneh et al. (2020) [11] indicated that, after a systematic review, there were no sufficient data to conclude that probiotic supplementation could provide a significant health benefit in the immunoregulatory effect and maintain gastrointestinal microbial balance. Most studies also specified that the exact mechanisms of the beneficial effects of probiotics on young calves still needed further investigation to precisely define and maximize the benefits that may be derived. Kim et al. (2021) [4] demonstrated that intensive multi-donor fecal microbial transplantation (FMT) could change the gut microbiota of diarrheic calves with alterations in fecal microbial metabolite concentrations, suggesting that FMT may be a promising treatment for calf diarrhea beyond antibiotic-based therapies. However, since changes in the gastrointestinal microbiota are dynamic, defining a healthy calf's microbiota is a difficult task. Probiotics isolated from preruminant Holstein calves according to the results obtained from next-generation sequencing (NGS) and bioinformatics analysis, which are the so-called next-generation probiotics (NGP), might provide a solution. Thus, in the present study, we aimed to investigate the interrelationships between gastrointestinal microbiota and diarrhea in preruminant Holstein calves, and identify and isolate the microbial biomarkers, associated with the healthy calves, as potential NGP. The antimicrobial activities and immunoregulatory effects of the isolated probiotic strains were also evaluated. Selecting probiotics originally from young calves, which might colonize the gastrointestinal track, could provide a solution for long-term health benefits without sustained periods of administration.

Animals and Sample Collection
All calves of the Holstein breed at a commercial dairy farm were housed in standard pens (1.5 m 2 ) and administered maternal colostrum for the first three days after birth. From 4th day to 42nd day, the calves were equally fed with whole milk twice at 10% of the initial BW daily, 1/15th during 7th and 8th week via bottles. At day 14, all calves were gradually given solid starter feed (Calf Starter, Yi-Chen Ltd., Yi-Lai. Taiwan).
Ten healthy calves and 10 diarrheic calves between 30 and 51 days old (prior to weaning) were selected (Table S1) by the outcomes of veterinary diagnosis and fecal scores. Blood and fecal samples were collected 2 h after the morning feeding for two consecutive days. Feces were collected per rectum from suckling calves by the veterinary team, and the samples were stored in a −80 • C freezer until analysis. Each sample was accurately weighed. Blood samples were collected from the jugular vein of the calves using a vacutainer tube. All animal experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the National Taiwan University (Approval No: NTU108-EL-00160).

Analyses of Fecal Score, Plasma Cytokines, Plasma IgG, and Haptoglobin
The fecal score was evaluated by the veterinary team using a calf health scoring guide written by the University of Wisconsin-Madison School of Veterinary Medicine [12]. Tumor necrosis factor (TNF)-α, interleukin (IL)-6 and IL-8 levels in the plasma were measured using commercial enzyme-linked immunosorbent assay kits (Bovine TNF-alpha, IL-6 and IL-8; DuoSet ELISA, R&D system, Minneapolis, MN, USA) according to the manufacturer's instructions. For serum immunoglobulin (Ig) G and haptoglobin (HP) analysis, commercial enzyme-linked immunosorbent assay kits (Bovine IgG kit, R&D system Inc., Mckinley, MN, USA; Bovine HP, haptoglobin ELISA Kit, Fine test, Wuhan Fine Biotech Co., Ltd., Wuhan, China) were used.

Microbiota Analysis
The microbiota analysis adopted the process described by Huang et al. (2021) [13]. Briefly, the V3-V4 regions of the 16S rRNA gene were amplified with barcodes and sequenced by the Illumina MiSeq paired-end sequencing platform after extracting total genomic DNA. The Silva v.132 database, QIIME v1.7.0, and R v2.15.3 software [14] were used to analyze taxonomic annotation of each operational taxonomic unit (OTU), alpha diversity (Chao1 and Shannon), and beta diversity (principal component analysis, PCA) [15], respectively. The biomarkers were identified by linear discriminant analysis (LDA) effect size (LEfSe) algorithm. The phylogenetic investigation of communities was performed by reconstruction of unobserved states (PICRUSt v1.1.1) using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database as reference genomes [16] to predict functional genes from the abundances of 16S rRNA sequencing data.

Isolation and Identification of Bifidobacterium longum subsp. longum
The strain isolation used the protocol described by Tanaka and Masahiko (1980) [17], with minor modifications. Briefly, 1 g fecal sample was homogenized. After 10-fold dilution with saline solution, 0.1 mL aliquots were cultured at 37 • C for 3 days anaerobically in modified Lactobacilli MRS agar (Acumedia, Neigen, Lansing, MI, USA) plates with 0.05% (w/v) L-cysteine hydrochloride monohydrate (Sigma-Aldrich Inc., St. Louis, MO, USA) and 0.01% cycloheximide (Sigma-Aldrich). The colonies with distinct morphologies were isolated. The Harrison's disc method [18] was also used for colony selection. The isolates were then purified on modified MRS agar plates by streaking at least three times. Strains were kept at −80 • C in 15% (w/w) glycerol for further analysis.

Antimicrobial Activity by an Agar Spot Test
Five pathogenic strains were used in this study. Bacillus cereus and Salmonella enterica were plated on Nutrient Broth (NB) agar (Fisher Scientific Ltd., Loughborough, Leicestershire, UK), while Staphylococcus aureus, Escherichia coli, and Vibrio parahaemolyticus were cultured using Tryptone Soy Broth (TSA; NEOGEN Corporation, Lansing, MI, USA), all at 37 • C with constant agitation. The antimicrobial activity was determined according to the process described by Tejero-Sariñena et al. (2012) [25].

Cytokine Production in RAW 264.7 Cells
The murine macrophage cell line (RAW 264.7, ATCC TIB71, American Type Culture Collection, Manassa, VA, USA) was cultured according to the procedure described by Hong et al. (2009) [26] with minor alterations. RAW 264.7 macrophages at a density of 5 × 10 5 cells/mL were seeded in 1 mL medium and co-cultured with 10 6 CFU/mL of each Bifidobacterium strain. A total of 50 ng of lipopolysaccharide (LPS; Merck-Sigma-Aldrich, Burlington, MA, United States) was added to 1 mL medium as the positive control. The levels of the cytokines (TNF-α and IL-10) in cell culture supernatants were measured by mouse cytokine ELISA kits (Invitrogen, Grand Island, NY, USA) after 24 h incubation at 37 • C.

Statistical Analysis
All phenotypic and next-generation sequencing (NGS) data were analyzed by a nonparametric Mann-Whitney U test to identify significant differences between groups. Bacterial networks and the correlation between bacterial biomarkers and diarrhea/inflammatory markers were constructed by Spearman's correlation analysis. A p-value less than 0.05 is statistically significant. All statistical analyses were conducted by Statistical Analysis System v9.4 (SAS Institute Inc., Cary, NC, USA) and R software. 2.1 (R Core Team, Vienna, Austria).

The Fecal Microbiota in Healthy and Diarrheic Preruminant Calves
Healthy and diarrheic calves (10 each) were selected according to the combinative results of clinical, systematic, and hematological examinations by the veterinarian and fecal score (>2). The diarrheic calves demonstrated significantly higher fecal scores ( Figure 1A) than the healthy counterparts, as we expected. The significantly higher bovine HP and IL-8 and lower IgG observed in the diarrhea group, as compared with those of the healthy group (p < 0.05) ( Figure 1B), indicate that the diarrheic calves were also under inflammatory status.
The results of the 16S rRNA analysis in the healthy calves and those with diarrhea show that a total of 871,579 and 871,107 effective tags were obtained from 1,042,854 and 1,041,882 raw paired-end reads, respectively. The alpha diversity, Chao1 richness estimator, and Shannon's diversity index show no significant difference (p > 0.05) between the healthy calves and calves with diarrhea ( Figure 2A). However, both groups could be clearly discriminated by the partial least squares discriminant analysis (PLS-DA) plot ( Figure 2B). Thus, we further analyzed the dominant taxa at the family, genus, and species levels and found that the abundances of the family Bifidobacteriaceae and genera Ruminococcaceae_UCG_014, Bifidobacterium and Pseudoflavonifractor in the healthy calves were significantly higher than that of the counterparts with diarrheic conditions (p < 0.05) (Table S1). Other genera, such as Faecalibacterium and Prevoltella_9, were also upregulated in the healthy group. Contrastingly, the abundances of the family Marinifilaceae and genera Alloprevotella, Subdoligranulum, Ruminococcaceae_UCG_005, Ruminococcus_2, Lachnoclostridium, Odoribacter, Ruminiclostridium_9, and Lachnospiraceae_NK4A136_group in the healthy calves were significantly lower than that of the diarrheic counterparts (p < 0.05). The abundance of the genus Escherichia_Shigella was also increased in the diarrheic calves. Regarding the top 20 species, the abundance of Bifidobacterium longum subsp. longum in the healthy group was signifi-cantly higher than that of the group with diarrhea (p < 0.05). While the level of bacterium_ ic1277 in the healthy group was lower compared with those of the group with diarrhea (p < 0.05). The number of Clostridium_sp_K4410MGS_306 increased in the diarrheic calves (p < 0.10) as well. We noticed that Corynebacterium_sp, Bacteroidaceae_bacterium_DJF_B220 and Corynebacterium_amycolatum were only observed in the diarrheic calves with diarrhea. The results of the 16S rRNA analysis in the healthy calves and those with diarrhea show that a total of 871,579 and 871,107 effective tags were obtained from 1,042,854 and 1,041,882 raw paired-end reads, respectively. The alpha diversity, Chao1 richness estimator, and Shannon's diversity index show no significant difference (p > 0.05) between the healthy calves and calves with diarrhea ( Figure 2A). However, both groups could be clearly discriminated by the partial least squares discriminant analysis (PLS-DA) plot (Figure 2B). Thus, we further analyzed the dominant taxa at the family, genus, and species levels and found that the abundances of the family Bifidobacteriaceae and genera Rumi-nococcaceae_UCG_014, Bifidobacterium and Pseudoflavonifractor in the healthy calves were significantly higher than that of the counterparts with diarrheic conditions (p < 0.05) (Table  S1). Other genera, such as Faecalibacterium and Prevoltella_9, were also upregulated in the healthy group. Contrastingly, the abundances of the family Marinifilaceae and genera Alloprevotella, Subdoligranulum, Ruminococcaceae_UCG_005, Ruminococcus_2, Lachnoclostridium, Odoribacter, Ruminiclostridium_9, and Lachnospiraceae_NK4A136_group in the healthy calves were significantly lower than that of the diarrheic counterparts (p < 0.05). The abundance of the genus Escherichia_Shigella was also increased in the diarrheic calves. Regarding the top 20 species, the abundance of Bifidobacterium longum subsp. longum in the healthy group was significantly higher than that of the group with diarrhea (p < 0.05). While the level of bacterium_ ic1277 in the healthy group was lower compared with those of the group with diarrhea (p < 0.05). The number of Clostridium_sp_K4410MGS_306 increased in the diarrheic calves (p < 0.10) as well. We noticed that Corynebacterium_sp, Bac-teroidaceae_bacterium_DJF_B220 and Corynebacterium_amycolatum were only observed in the diarrheic calves with diarrhea.

Identification of the Critical Gastrointestinal Bacterial Biomarkers
Since the NGS results indicate that the health and diarrhea groups could be distinguished by gastrointestinal bacteria, the bacterial biomarkers were identified by the linear discriminant analysis (LDA) effect size (LEfSe) algorithm. A total of 19 influential taxonomic clades were recognized, including 1 order, 3 families, 13 genera and 2 species ( Figure 2C). The most critical biomarkers identified in the healthy group were one order (Bifidobacteriales), one family (Bifidobacteriaceae), three genera (Bifidobacterium, Ruminococ-caceae_UCG_014 and Pseudoflavonifractor), and one species (Bifidobacterium longum subsp. longum). Contrastingly, 1 family (Marinifilaceae), 11 genera ([Ruminococcus]_gauvreauii_group, Ruminiclostridium_9, Lachnoclostridium, Prevotellaceae_UCG_001, Odoribacter, Prevotellaceae_ NK3B31_group, Ruminococcus_2, Lachnospiraceae_NK4A136_group, Subdoligranulum, Ru-minococcaceae_UCG_005, and Alloprevotella) and 1 species (bacterium_ ic1277) were the most influential taxa in the group with diarrhea. The results of the relative abundances of the bacterial biomarkers associated with the group with diarrhea are consistent with the above findings ( Figure 2D). The relative abundances of all 12 bacterial biomarkers (genus and species levels) identified in the group with diarrhea were significantly higher than those of the healthy group (p < 0.05). Conversely, the relative abundance of all four bacterial biomarkers identified in the healthy group was significantly lower than that of the group with diarrhea (p < 0.05).

Correlation of Diarrhea and Inflammation Parameters with the Bacterial Biomarkers
The results of correlation ( Figure 3A) show that the species enriched in the healthy group, such as Bifidobacterium, Ruminococcaceae_UCG_014, and Pseudoflavonifractor, were positively correlated with the bovine IgG, but negatively correlated with the fecal score and inflammatory factors, bovine HP, and IL-8. Conversely, the species enriched in the group with diarrhea, including [Ruminococcus]_gauvreauii_group, Ruminiclostridium_9, Lachnoclostridium, Prevotellaceae_UCG_001, Odoribacter, Ruminococcus_2, Lachnospiraceae_NK4A136_ group, Subdoligranulum, Ruminococcaceae_UCG_005, and Alloprevotella, demonstrated negative correlations with the levels of IgG and positive correlations with the levels of fecal score, bovine HP, and IL-8. Spearman's correlation test between diarrheic biomarkers and calf gut microbial markers at genus and species level. Each cell was colored corresponding to Spearman's correlation results. Significant difference: * p < 0.05; ** p < 0.01. Bacterial networks between calf gut microbial markers at genus level and their correlation with (B) bovine IgG, (C) fecal score or (D) IL-8 (Continued). Each node represents a genus biomarker, and the size of it corresponds to relative abundance. Colored nodes show significant correlation with IgG, fecal score or IL-8 (p < 0.05).
In addition, we also clarified the co-occurrence patterns among bacterial biomarkers by constructing the bacterial network among 14 genera and two species, which was further correlated with the levels of IgG, IL-8 and fecal score. Among the genus bacterial network, [Ruminococcus]_gauvreauii_group, Lachnoclostridium, and Prevotellaceae_UCG_001, which negatively correlated with IgG (p < 0.05), were positively associated with five ( Figure 3B), four ( Figure 3C), and one ( Figure 3D) other bacterial biomarkers in the group with diarrhea, respectively. In terms of the genus bacterial networks for fecal score and IL-8, Ruminococcaceae_UCG_014 and Pseudoflavonifractor, the bacterial biomarkers identified in healthy calves, were positively correlated with each other and negatively correlated with five other diarrheic biomarkers (p < 0.05). Contrastingly, the bacterial biomarkers in the group with diarrhea, such as [Ruminococcus]_gauvreauii_group, Ruminiclostridium_9, Odoribacter, Ruminococcus_2, Lachnospiraceae_NK4A136_group, Subdoligranulum, Ruminococ-caceae_UCG_005, and Alloprevotella, were positively and negatively associated with the certain diarrheic and healthy biomarkers, respectively.

The Relative Abundance of PICRUSt Functional Prediction of Fecal Microbiota in the Preruminant Calves
The results of the functional profiles indicate that the relative abundances of functional pathways in energy metabolism (proteasome, photosynthesis, photosynthesis proteins), metabolism of cofactors and vitamins (thiamine metabolism and cysteine), and amino acid metabolism (methionine metabolism) were significantly higher in the healthy group as compared with those of the diarrheic counterparts ( Figure 4). Conversely, infectious diseases (Staphylococcus aureus infection), replication and repair (non-homologous end-joining), biosynthesis of other secondary metabolites (flavonoid biosynthesis), and xenobiotics biodegradation and metabolism (metabolism of xenobiotics by cytochrome P450, drug metabolism-cytochrome P450, chloroalkane and chloroalkene degradation, naphthalene degradation) and metabolism of terpenoids and polyketides (carotenoid biosynthesis) were significantly higher in the group with diarrhea compared with those of the healthy counterparts. These results suggest that the altered gut microflora not only influenced the healthy status in the preruminant calves, but it may also induce the increase in functional gene families in the gut of the calves.

Isolation and Identification of Bifidobacterium longum subsp. longum from the Healthy Preruminant Calves
First, 40 isolates were obtained from the feces of the preruminant calves using MRS medium and the Harrison disc method. After Gram staining and microscopic observation, nine Gram-positive isolates with similar phenotypic characteristics as compared with the type strain Bifidobacterium longum subsp. longum (BCRC14664) were selected for further identification ( Figure 5A). The results of the comparative 16S rRNA gene analysis show that all nine LAB isolates belonged to the genus Bifidobacterium (Table 1). Of the nine, seven isolates (HCF-4, HCF-12, HCF-14, HCF-19, HCF-22, and HCF-24) shared 97% similarity with Bifidobacterium longum subsp. Longum, while the other two isolates (HCF-28 and HCF-30) had 97% similarity with Bifidobacterium longum subsp. infantis. Figure 5B shows the phylogenetic tree based on the 16S rRNA gene sequence analysis, depicting the phylogenetic relationships among the nine Bifidobacterium strains and seven strains obtained from the GenBank. Escherichia coli (U5 41) was used as the outgroup. The six strains, HCF-4, HCF-12, HCF-14, HCF-19, HCF-22 and HCF-24, were grouped together and formed a monophyletic clade, which showed a bootstrap value of 98% with Bifidobacterium longum subsp. longum (LC612559.1 and MT268981.1). Another strain, HCF-27, formed a monophyletic clade with Bifidobacterium longum subsp. longum (LC612559.1 and MT268981.1) with a bootstrap value of 98%. Two strains, HCF-28 and HCF-30, belonged to the Bifidobacterium longum subsp. infantis, pantothenic acid, and anacardic acid (p < 0.05) ( Figure 5B), which are associated with purine metabolism and pantothenate and CoA biosynthesis ( Table 1).
The results of the immunoregulatory effects show that the strain HCF-24 could significantly stimulate the production of proinflammatory cytokine, TNF-α, without LPS treatment, unlike the normal control (NC) group (p < 0.05). The upregulatory effect on TNF-α was not observed in the other six strains ( Figure 6A). After LPS treatment, all seven isolated strains could significantly inhibit the levels of TNF-α compared with those of the LPS control (PC) (p < 0.05). For regulatory cytokines, IL-10, four strains, HCF-4, HCF-12, HCF-22, and HCF-27, could significantly stimulate the levels with or without LPS treatment compared with that of the LPS group ( Figure 6B). The strain HCF-27 could induce in vitro proinflammatory cytokine, TNF-α, and inhibit its level after LPS treatment. This strain also induced regulatory cytokine, IL-10, with or without LPS treatment, indicating that this isolated strain might possess an immunoregulatory function.

Discussion
In the present study, we first investigated the interrelationships between gastrointestinal microbiota and diarrhea in preruminant Holstein calves. The physiological condition and diarrhea diagnosis of the calves were the key factors affecting the outcome. Thus, besides clinical and hematological examinations by the veterinarian, three immune-related markers, Hp, IL-8, and IgG, were also used to verify the diarrheic inflammatory status. Hp, one of the important acute phase proteins in ruminants [27], was related to the severity and activity of inflammation. Hp level in calves could be utilized for diagnosing bacterial [28] and viral [29] diseases. IL-8, known as an inflammatory marker, is a strong chemoattractant for T lymphocytes and polymorphonuclear leukocytes [13]. Bovine IgG has been reported not only binding to a wide range of pathogenic bacteria and viruses, but also to various allergens [30]. The non-sick calves feeding with IgG from 2 to 14 days of age decreased the frequency of incidence of diarrhea [31]. Thus, the concentrations of Hp, IgG, and IL-8 can be used as prediction indices of calf health [32] or prognose the course of the disease during rearing [33]. Our result was paralleled with previous findings, demonstrating that upregulation in the serum Hp concentration and IL-8 in the diarrheic calves were associated with clinical symptoms of diarrhea.
The finding in 16S rRNA indicates that the taxa with significant differences in abundances between the healthy and diarrheic calves were paralleled with LEfSe, which were identified as the critical gastrointestinal bacterial biomarkers. The most abundant and important commensal bacterial genus in the healthy calves was Faecalibacterium, which was also found in previous studies. The body weight gain during the pre-weaning period was reported to be significantly associated with the prevalence of Faecalibacterium spp. during the first week of a calf's life [34]. The intervention with Faecalibacterium prausnitzii, a butyrate producer, reduced the incidence of severe diarrhea and related mortality rate in pre-weaned dairy heifers [35]. Our finding, accompanied with other previous studies, suggests a possible beneficial effect of Faecalibacterium spp. on the health and growth of young calves. Bifidobacterium has been reported to prevent gastrointestinal infections through competing for binding sites with pathogens and viruses on epithelial cells [36,37] and regulating the gastrointestinal immune system response [38]. Previous studies revealed that downregulation in the relative abundance of Bifidobacterium was involved in various dysbiosis-related intestinal diseases [39,40]. Our results also suggest a positive impact of Bifidobacterium on the prevention of gastrointestinal diseases in calves, with positive correlation with the bovine IgG and negative correlation with fecal score and inflammatory factors, bovine HP, and IL-8. Pseudoflavonifractor, positively correlated with the bovine IgG and negatively correlated with inflammatory factors in the present study, was reported to be beneficial to the immune homeostasis [41] and positively associated with the content of anti-inflammatory cytokines in yak calves [42]. The cell-free supernatant of Pseudoflavonifractor sp. AHG0008 showed the ability to suppress IL-8 secretion by peripheral blood mononuclear cells [43,44]. However, little was known about the role of Pseudoflavonifractor in the gastrointestinal sites.
Conversely, the diarrheic samples were categorized by a high abundance of sequences assigned to Escherichia_Shigella, the ileum mucosa-associated opportunistic pathogens. Shigella is phylogenetically distinct from several independent E. coli strains with four subgroups differentiated, which all cause shigellosis [45]. Thus, further identification of the ileum mucosa-associated Escherichia_Shigella, at species or strain level, is necessary. Other genera in the family of Rumminococcaceae (Rumminococcaceae UCG 005, Ruminococcus_2 and Ruminococcus_9) were also enriched in the diarrheic calves. Ruminococcaceae_UCG_005 was the most abundant genus at the post-weaning period [46] and at lactating Holstein dairy cows [47]. Ruminococcus_2 has been reported to significantly reduce in probiotic-treated calves [46] and is potentially related to intestinal permeability and hepatic inflammation in a rodent model [48]. However, no evidence indicated the detrimental effects of these genera in calves. This result supports the high functional diversity reported within the Ruminococcaceae family and the difficulty to infer functions based on amplicon sequencing data [49].
The changes in gut microbial composition induced the functional gene family enrichment of gut microbiota in calves. More genes responsible for the energy and nutrient metabolisms were downregulated in the diarrheic calves with upregulating infectious diseases (Staphylococcus aureus infection), which was consistent with previous studies in calves [4] and humans [40]. Kim et al. [4] indicated that the improper metabolism of amino acids was one of the most important factors associated with diarrhea, suggesting that the interfering metabolisms were strongly connected to dysbiosis-related disease.
Moreover, broad upregulation in the poorly characterized two-component system and xenobiotics biodegradation and metabolism (metabolism of xenobiotics by cytochrome P450, drug metabolism-cytochrome P450) were observed in the present study. Twocomponent systems such as QseBC were mainly allocated in Escherichia, which produces a barrier effect against enteropathogens [50]. Other two-component regulatory systems, KdpDE, LiaS-LiaR, and GlnK-GlnL, are involved in the use of glutamine to sustain the homeostasis of stress response and cellular energy [51]. Poor characterization in twocomponent systems, suggesting a depleted beneficial metabolism from Escherichia in diarrheic calves, might play a crucial role in the development of diarrhea. The importance of the cytochrome P450 in the metabolism of endogenous compounds and exogenous compounds (drugs, natural products) has been well documented [52]. The enhancement of xenobiotics biodegradation and metabolism of cytochrome P450 may be due to the therapeutic treatment of diarrheic calves. Nevertheless, PICRUSt could just predict metagenomic function. Other omics, such as metabolomic approaches, are suggested to be able to recognize real changes in the metabolic function of the microbiota of diarrheic calves.
During the analysis in the interrelationships between gastrointestinal microbiota and diarrhea in preruminant calves, the current findings regarding the possible beneficial effects of Bifidobacterium longum subsp. longum appear promising, which are expected to develop as NGP. However, we acknowledged that a correlation study was unable to illustrate any underlying relationships. Thus, hypothesis-driven experiments were performed to evaluate the potential functionality of Bifidobacterium longum subsp. longum in vitro. Infectious agents, such as Salmonella spp. and Escherichia coli, are the major causes for calf diarrhea [2]. The following prevention or treatments with antibiotics, leading to an imbalance in the intestinal microbial community, could activate immune responses, inflammation, and peristalsis in the host gut, which also causes diarrhea [4]. The specific bifidobacterial strains, HCF-19 and HCF-27, presented here, may provide a solution for diarrhea prevention through the inhibition of infectious bacteria, an immunomodulatory effect, and possible modulating microbiota. Several studies have reported that Bifidobacterium longum could inhibit the growth of Salmonella enterica, E. coli, Bacillus cereus, and Staphylococcus aureus [53,54] due to the production of organic acids and specific bacteriocins. Certain Bifidobacterium spp. have also been used as a supportive treatment for diarrhea in dairy calves [55,56], but few of them were Bifidobacterium longum subsp. lognum and directly isolated from calves. It is worth noticing that the probiotic effects of Bifidobacterium longum subsp. longum were strain specific. Ibrahim et al. (2003) [57] also indicated that the beneficial effects of Bifidobacterium seem strain specific, not species specific.

Conclusions
In the present study, significant alterations in microbiota structure between healthy and diarrheic calves with deviations in the predictive metagenomic function of the bacterial communities and strongly correlating with immune-related markers provided a novel insight regarding the interrelationships between gastrointestinal microbiota and diarrhea in preruminant calves. Based on the microbiota findings, further screening and isolation of Bifidobacterium longum subsp. longum strains with immunomodulatory and antagonistic effects was conducted to characterize the relationship with the possible amelioration of diarrheic diseases. Two Bifidobacterium longum subsp. longum strains might be expected to emerge as next-generation probiotics. To the best of our knowledge, this is the first study investigating the relationship between gastrointestinal microbiota and diarrhea in preruminant calves using immune-related markers and further isolating specific bacterial strains, enriched in clinically healthy individuals for potential next-generation probiotics. Our results also suggest the importance of Bifidobacterium longum subsp. longum in the overall health of calves.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ani12060695/s1, Table S1: The information of calf health status; Table S2: The relative abundances (%) of dominant taxa in 20 fecal samples at genus and species levels.  Institutional Review Board Statement: All animal experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the National Taiwan University (Approval No: NTU108-EL-00160).

Data Availability Statement:
The datasets used and/or analyzed in the current study are available from the corresponding author on reasonable request.