Shifts in bacterial communities in the rumen, vagina, and uterus of beef heifers receiving different levels of concentrate

Abstract

This experiment investigated the effects of diet composition on rumen, vaginal, and uterine microbiota of beef heifers. Fifteen rumen-cannulated, pubertal Angus-influenced heifers were used in a replicated 3 × 3 Latin square design (28-d periods and 21-d washout intervals). Dietary treatments included diets based on (as-fed) 100% grass hay (HF), 60% grass hay + 40% corn-based concentrate (INT), or 25% grass hay + 75% corn-based concentrate (HG). Treatments were offered individually to heifers once daily at 2% body weight. Rumen, vaginal, and uterine samples were collected on days 0 and 28 of each period. Data were analyzed using orthogonal contrasts (linear and quadratic), using results from day 0 as independent covariates and heifer as the experimental unit. Ruminal pH on day 28 decreased linearly (P < 0.01) as concentrate inclusion increased. Uterine and vaginal pH on day 28 were not affected by treatments (P ≥ 0.35). Within the rumen samples, Bacteriodetes was the most abundant phylum and its relative abundance linearly decreased (P ≤ 0.01) with the inclusion of concentrate. Prevotella was the most abundant genus within the rumen but was not affected by treatments (P ≥ 0.44). Genera with relative abundance ≥1% (average across treatments) in the rumen that were impacted by treatments (P ≤ 0.01) included Bacteroides, Pedobacter, Dysgonomonas, Caloramator, and Ruminococcus. Firmicutes was the most abundant phylum in the vagina and uterus, but it was unaffected by treatments (P ≥ 0.16). Prevotella was the most abundant genus in the vagina, and its relative abundance increased (P < 0.01) with the inclusion of concentrate. Other genera with relative abundance ≥1% that were significantly affected (P ≤ 0.05) by treatments were Clostridium, Pedobacter, Roseburia, Oscillospira, Faecalibacterium, Caloramator, Paludibacter, Rhodothermus, and Porphyromonas. In uterine samples, Prevotella was the most abundant genus but was unaffected by treatments (P ≥ 0.29). Genera with relative abundance ≥1% in the uterus that were significantly affected (P < 0.01) by treatments were Caloramator, Paludibacter, and Thalassospira. Collectively, inclusion of concentrate in the diet altered the bacterial composition within the rumen as well as shifting bacterial populations within the vagina and uterus. Research is warranted to further understand the impacts of these diet-induced microbiota changes on reproductive function and performance of beef heifers.

Feeding beef heifers differing levels of concentrate not only alters the rumen microbiota, but also influenced both the vaginal and uterine microbiota. Further research is warranted to understand how altering the vaginal and uterine microbiome environment can affect reproductive efficiency in cattle.

Introduction

Cow–calf operations serve as the foundation for the beef industry and regulate the number of cattle available for beef production. Replacement heifers are vital to this production system and poor reproductive performance can result in significant economic losses for the producer. There are many management strategies that have been developed to maximize the reproductive potential of beef heifers such as nutritional management, identification of reproductive maturity by physiological and morphological indicators (e.g., reproductive tract score and pelvic area), and the implementation of an estrous synchronization program (Moorey and Biase, 2020). These have been shown to significantly improve heifer reproductive performance, although some heifers still fail to cycle properly, become pregnant, and calve by 2 yr of age.

Several heifer development programs have been established for cattle producers to ensure heifers are receiving adequate nutrition. Appropriate nutritional status is key for reproductive success in cattle, and heifers experiencing higher levels of nutrition and adequate weight gain prior to the first breeding season experience increased reproductive success in their first and subsequent calving seasons (Milagres et al., 1979; Fleck et al., 1980; Day et al., 1984). However, dietary management may also impact traits within the animal that play a role in reproductive efficiency, such as altering the microbiota of the reproductive tract (Ault et al., 2019a).

The largest and most diverse population of microbes in cattle is found in the rumen, and increasing the starch content of diets alters the taxonomic composition and diversity of rumen microbiota in cattle (Fernando et al., 2010; Khafipour et al., 2016; Tun et al., 2020). The direct (e.g., endogenous) and indirect contribution (e.g., feces) of ruminal microbes to the reproductive tract microbiome is unknown. Dietary energy and protein content have direct impact on reproductive function, but diet-induced changes in ruminal flora and its contributions (positive or negative) to vaginal and uterine microbiome require investigation. The objective of this study was to investigate this relationship using bacterial genome sequencing in beef heifers receiving different dietary profiles (i.e., forage-based to grain-based diets). The hypothesis is that heifers receiving increased amounts of a starch-based concentrate will experience a shift in ruminal bacteria profile, which will impact the vaginal and uterine microbiota.

Materials and Methods

This study was conducted at Texas A&M University O.D. Butler, Jr. Animal Science Complex Nutrition and Physiology Center in College Station, TX from November 2020 to February 2021. All animals were cared for in accordance with acceptable practices and experimental protocols reviewed and approved by the Institutional Animal Care and Use Committee of Texas A&M University (IACUC #2018-0099).

Animals and treatments

Fifteen ruminally cannulated, pubertal, nulliparous, nonpregnant Angus-influenced heifers were assigned to this experiment. All heifers were around 15 mo of age and puberty status was determined by a pretrial reproductive tract score that was conducted by a trained technician. Any heifers that did not have a sufficient score of 4 or 5 were excluded from the entire trial. Heifer body weight (BW) was recorded to represent pretrial BW (333 ± 34 kg). Heifers were ranked by pretrial BW and assigned to three groups of five heifers each, in a manner that all groups had equivalent pretrial BW. Groups were enrolled in a 3 × 3 Latin square design containing 3 periods of 28 d, and a 21-d washout interval between periods. During each period (days 0–28), heifers were housed in an enclosed barn in individual pens (2 m × 4 m) with ad libitum access to water and fed their respective diet for each period (Table 1) at 2% of their BW recorded at the beginning of each period. Diet was fed daily at 0700 hours. Throughout the washout interval, heifers were housed in an outside paddock and received ad libitum access to water and grass hay. The last 3 d of the washout interval, heifers were returned to the individual pens to allow for a period of adaptation in the barn.

Table 1.

Ingredient composition (as-fed basis) of diets offered to heifers and nutrient profile of ingredients¹

		Diets2
Item	HF	INT	HG
Ingredients, % (as-fed basis)
Bermudagrass hay	100	60	25
Dried distillers’ grains	0	30	60
Cracked corn	0	10	15
Nutrient profile2 (dry matter basis)
Dry matter, %	92.7	91.5	90.5
Crude protein, %	14.2	14.2	14.2
Neutral detergent fiber, %	65.0	46.4	28.9
Starch, %	0.900	8.20	12.1
Net energy for maintenance, Mcal/kg	1.16	1.52	1.85
Net energy for growth, Mcal/kg	0.594	0.902	1.21
Ca, %	0.580	0.358	0.164
P, %	0.280	0.339	0.395

¹Heifers were fed daily at 0700 hours. Heifers also received 230 g/heifer daily a premix that delivered 0.5 mg/heifer daily of MGA, in addition to 60 g/heifer daily of a mineral–vitamin supplement containing 21% Ca, 0.01% P, 21% NaCl, 0.20% K, 0.10% Mg, 0.045% Cu, 0.001% Se, 0.280% Zn, 220,000 IU/kg of vitamin A, 19,800 IU/kg of vitamin D3, and 3,500 IU/kg of vitamin E (Anipro Xtraperformance Feeds, College Station, TX).

²Analyzed via wet chemistry procedures by a commercial laboratory (Dairy One Forage Laboratory, Ithaca, NY). Calculations for net energy for maintenance and growth used the equations proposed by the NASEM (2016).

At the beginning of each period (day 0), groups were assigned to receive one of three treatments as described in Table 1: (1) 100% grass hay (HF: n = 15), (2) 60% grass hay + 40% corn-based concentrate (INT: n = 15), or (3) 25% grass hay + 75% corn-based concentrate (HG: n = 15). Due to shifts in the abundance of bacterial communities throughout the estrous cycle, caused by varying amounts of pre-ovulatory progesterone and estradiol (Ault et al., 2019a, 2019b; Quereda et al., 2020), heifers received 0.5 mg/d of melengestrol acetate (MGA) to suppress ovulation. Dietary treatments were designed to be iso-nitrogenous to ensure that dietary treatments differed mainly on energy and starch content. Dietary protein also impacts ruminal flora with potential consequences to the vaginal and uterine microbiota (Ault-Seay et al., 2022), which deserves specific investigation and was not part of the main hypothesis of this experiment.

Sampling

Samples of feed ingredients were collected prior to the beginning of each period, pooled across all period, and analyzed for nutrient content by a commercial laboratory (Dairy One Forage Laboratory, Ithaca, NY). Samples were analyzed by wet chemistry procedures for concentrations of crude protein (method 984.13; AOAC, 2006), acid detergent fiber (method 973.18 modified for use in an Ankom 200 fiber analyzer, Ankom Technology Corp., Fairport, NY; AOAC, 2006), neutral detergent fiber using a-amylase and sodium sulfite (Van Soest et al., 1991; modified for use in an Ankom 200 fiber analyzer, Ankom Technology Corp.), and starch (YSI 2700 SELECT Biochemistry Analyzer; YSI Inc., Yellow Springs, OH). Calculations for net energy for maintenance and gain used equations from the NASEM (2016). Nutritional profile of the dietary treatments is described in Table 1.

Rumen fluid samples were collected via rumen cannula from all heifers on days 0 and 28 of each period 4 hours postfeeding (Rett et al., 2020). The pH of the rumen samples was immediately measured using the Orion Star pH portable meter (Thermo Fisher Scientific Inc., Waltham, MA). Rumen samples were strained through eight layers of cheesecloth for fluid extraction, which was stored into individual stainless-steel thermoses to maintain both temperature and an anaerobic environment and transported to the laboratory for further processing. A 5-mL subsample of each rumen fluid sample was transferred into individual falcon tubes containing 1 mL of 25% metaphosphoric acid and stored at −20 °C on the same day of collection. These samples were processed and analyzed for volatile fatty acid (VFA) profile as described by Cappellozza et al. (2013).

Uterine and vaginal flushes were collected on days 0 and 28 of each period 4 hours postfeeding as previously described (Clemmons et al., 2017; Ault et al., 2019a, 2019b). Briefly, the perineal area was cleaned and disinfected prior to flushing. To obtain vaginal flush samples 60 mL of 0.9% sterile saline was expelled by sterile syringe into the vagina and recovered via vaginal lavage. After vaginal flush, to obtain uterine flush samples, 180 mL 0.9% sterile saline was flushed through a sterile Foley catheter into the uterus by a single technician. A new sterile catheter was used for each heifer to ensure that there was no cross contamination of bacteria. Resulting uterine flush fluid was collected by rectal massage. The pH of the flush samples was measured immediately after collection (Orion Star pH portable meter, Thermo Fisher Scientific, Waltham, MA). Samples were snap-frozen in liquid nitrogen and stored at −80 °C until analysis.

DNA extraction and sequencing

Rumen and reproductive flush samples were sent to FERA Diagnostics and Biologicals Corp. (College Station, TX) for DNA extraction and 16S rRNA gene amplicon sequencing. Samples were transferred to a 96-well plate and DNA extraction was performed using Mag-Bind Universal Pathogen 96 Kit (Omega Bio-Tek, Norcross, GA). The 16S amplicons were amplified by PCR for individual metagenomic DNA samples according to Bicalho et al. (2017). The V4 hypervariable region of bacterial/archaeal 16S rRNA gene were amplified with 515F (5ʹ-GTGCCAGCMGCCGCGGTAA-3ʹ) and 806R (5ʹ-GGACTACHVGGGTWTCTAAT-3ʹ) primers using methods optimized for the Illumina MiSeq platform (Caporaso et al., 2012).

Statistical analysis

Data were analyzed as a 3 × 3 Latin square PROC MIXED of SAS 9.4 (SAS Inst., Cary, NC) using heifer as the experimental unit. The model included treatment and period as independent variables, and all results from day 0 were included as covariate for each respective analysis. All data were analyzed with heifer as the random variable. Orthogonal contrasts were used to partition-specific treatments effects. Contrast statements included were: (1) linear effect of treatment and (2) quadratic effect of treatment. All results are reported as covariately-adjusted least square means. Significance was set at P < 0.05 and tendencies were if 0.05 ≤ P ≤ 0.10.

RESULTS

Ruminal fermentation parameters and reproductive flush pH

Ruminal pH on day 28 linearly decreased (P ≤ 0.01) with the inclusion of concentrate (Table 2). Vaginal and uterine pH were not affected (P ≥ 0.35) by treatments (Table 2). Ruminal concentrations of propionate, butyrate, isovalerate, isobutyrate, and valerate linearly increased (P ≤ 0.03) with the inclusion of concentrate. Ruminal concentrations of acetate and total VFA were not affected (P ≥ 0.20) by treatments (Table 2).

Table 2.

Treatment effects on pH in rumen and reproductive tract, as well as volatile fatty acid (VFA) profile in the rumen of beef heifers¹

	Treatments				P-value
Item	HF	INT	HG	SEM	Linear	Quadratic
Rumen pH	6.805	6.628	6.380	0.049	<0.01	0.46
Vaginal flush pH	6.926	6.937	6.918	0.051	0.91	0.81
Uterine flush pH	6.567	6.507	6.457	0.084	0.35	0.98
Rumen VFA profile, mM
Acetate	107.72	106.60	109.29	9.33	0.90	0.86
Propionate	22.33	29.31	37.98	3.31	<0.01	0.57
Butyrate	11.44	12.75	15.60	1.21	0.02	0.58
Isovalerate	1.00	1.33	1.86	0.15	<0.01	0.49
Isobutyrate	1.17	1.41	1.50	0.11	0.03	0.61
Valerate	1.14	1.20	1.68	0.37	<0.01	0.17
Total	145.03	152.17	167.82	13.13	0.20	0.47

¹HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Rett et al., 2020), and results from day 0 used as independent covariate for each respective analysis.

Rumen bacterial community composition – phylum

The relative abundance of bacteria derived from each individual rumen fluid sample was assigned to 29 different phyla, of which Bacteriodetes, Firmicutes, Proteobacteria, Tenericutes and Spirochaetes were the most abundant (average across treatments ≥1%; Table 3). Relative abundance of Bacteriodetes and Proteobacteria linearly decreased (P < 0.01) whereas relative abundance of Firmicutes and Actinobacteria linearly increased (P ≤ 0.01) with the inclusion of concentrate. Relative abundance of Spirochaetes was quadratically affected (P ≤ 0.01) by concentrate inclusion, and relative abundance of Tenericutes were not impacted (P ≥ 0.68) by treatments. Phyla that were less abundant (average across treatments <1%) but were affected by treatments (P ≤ 0.05) included Actinobacteria, Fibrobacteres, Chlorobi, Verrucomicrobia, Chloroflexi, Synergistetes, Planctomycetes, and Thermotogae (Table 3).

Table 3.

Treatment effects on bacterial phyla (relative abundance, %) in the rumen of beef heifers¹

	Treatments				P-value
Item	HF	INT	HG	SEM	Linear	Quadratic
Above 1% relative abundance
Bacteroidetes	64.895	62.566	53.936	2.181	<0.01	0.14
Firmicutes	18.848	21.149	33.070	2.554	<0.01	0.02
Proteobacteria	10.065	9.330	7.394	0.673	<0.01	0.22
Tenericutes	2.295	2.387	2.120	0.358	0.73	0.68
Spirochaetes	1.171	1.745	1.155	0.187	0.98	<0.01
Below 1% relative abundance
Fibrobacteres	0.993	0.877	0.024	0.149	<0.01	<0.01
Actinobacteria	0.276	0.413	0.454	0.050	<0.01	0.39
Chlorobi	0.212	0.239	0.459	0.087	<0.01	0.21
Chloroflexi	0.201	0.288	0.068	0.031	<0.01	<0.01
Synergistetes	0.126	0.088	0.093	0.012	0.01	0.05
Verrucomicrobia	0.029	0.019	0.012	0.004	<0.01	0.80
Thermotogae	0.001	0.001	0.005	0.002	0.02	0.23
Planctomycetes	0.004	0.001	0.001	0.001	0.05	0.11

¹HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Rett et al., 2020), and results from day 0 used as independent covariate for each respective analysis.

Rumen bacterial community composition – genus

The relative abundance of bacteria derived from each individual rumen fluid sample was assigned to 873 different genera, of which 20 bacteria were significantly affected (P ≤ 0.05) by concentrate level inclusion (Table 4). Prevotella was the most abundant genera within the rumen but was not affected by treatments (P ≥ 0.44). Bacteroides was the second most abundant genera and concentrate inclusion quadratically affected (P ≤ 0.02) the relative abundance of this genera as well as Dysgonomonas. Pedobacter was the third most abundant bacteria and linearly decreased (P ≤ 0.01) with the inclusion of concentrate. Relative abundance of Caloramator was also linearly decreased (P = 0.04) whereas Ruminococcus linearly increased (P = 0.02) with the inclusion of concentrate. Other genera with relative abundance ≥1% (average across treatments) that were affected (P ≤ 0.05) by concentrate inclusion were Porphyromonas, Blautia, Oscillospira, Weissella, Treponema, Emticicia, Fibrobacter, Alkaliphilus, Thalassospira, and Olivibacter (Table 4). Genera with relative abundance <1% (average across treatments) that were affected (P ≤ 0.03) by treatments were Anaeroplasma, Succiniclasticum, Butyrivibrio, and Rhodothermus (Table 4).

Table 4.

Treatment effects on bacterial genera (relative abundance, %) in the rumen of beef heifers¹

	Treatments				P-value
Genus	HF	INT	HG	SEM	Linear	Quadratic
Above 1% relative abundance
Prevotella	26.209	25.477	23.642	1.534	0.22	0.75
Bacteroides	13.862	11.741	14.118	0.747	0.85	0.02
Pedobacter	14.719	13.404	7.620	0.786	<0.01	<0.01
Dysgonomonas	3.569	4.360	2.929	0.298	0.10	<0.01
Oscillospira	1.719	2.728	5.201	0.471	<0.01	0.09
Paludibacter	3.508	3.138	2.395	0.727	0.28	0.81
Ruminococcus	1.501	2.132	3.962	0.952	0.02	0.45
Clostridium	2.354	2.556	2.214	0.218	0.55	0.14
Blautia	1.910	1.871	3.208	0.234	<0.01	0.01
Caloramator	2.298	2.137	1.648	0.231	0.04	0.52
Alkaliphilus	1.111	1.448	3.259	0.356	<0.01	0.06
Treponema	1.394	2.045	1.306	0.216	0.744	<0.01
Porphyromonas	2.131	1.632	0.927	0.288	<0.01	0.61
Butyrivibrio	0.880	0.832	2.893	0.380	<0.01	0.01
Succiniclasticum	0.975	1.092	2.169	0.440	<0.01	0.04
Paraprevotella	1.245	1.607	1.065	0.198	0.49	0.04
Weissella	1.623	1.604	0.616	0.293	<0.01	0.05
Below 1% relative abundance
Emticicia	1.286	1.130	0.454	0.110	<0.01	0.02
Olivibacter	1.071	1.069	0.524	0.111	<0.01	0.02
Anaeroplasma	0.991	1.001	0.504	0.175	0.03	0.16
Fibrobacter	1.184	0.991	0.039	0.175	<0.01	0.65
Thalassospira	1.087	0.865	0.096	0.105	<0.01	0.02
Rhodothermus	0.532	0.339	0.273	0.080	<0.01	0.34

¹HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Rett et al., 2020), and results from day 0 used as independent covariate for each respective analysis.

Reproductive bacterial community composition – phylum

The relative abundance of bacteria derived from each individual vaginal and uterine flush sample was assigned to 29 different phyla, of which Bacteroidetes, Firmicutes, Proteobacteria, Actinobacteria, Tenericutes, and Spirochaetes were the most abundant (average across treatments ≥1%; Figure 1). Table 5 also lists bacterial phyla within reproductive flushes with relative abundance <1% (average across treatments) that were affected by treatment.

Figure 1.

Relative abundance of bacterial phyla. Samples were grouped by treatment and uterus or vagina. Treatment group abbreviations: HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Clemmons et al., 2017; Ault et al., 2019a, 2019b), and results from day 0 used as independent covariate for each respective analysis.

Table 5.

Treatment effects on bacterial phyla (relative abundance, %) in vaginal and uterine flush samples of beef heifers¹

	Treatments				P-value
Genus	HF	INT	HG	SEM	Linear	Quadratic
Vaginal samples
Above 1% relative abundance
Firmicutes	46.042	42.763	45.800	5.854	0.95	0.49
Bacteroidetes	32.202	28.122	30.153	2.660	0.58	0.35
Proteobacteria	10.898	13.463	14.209	4.169	0.37	0.79
Actinobacteria	4.158	5.352	5.374	1.443	0.42	0.65
Tenericutes	2.271	3.582	1.871	1.736	0.87	0.45
Spirochaetes	2.386	1.840	1.436	0.502	0.09	0.90
Below 1% relative abundance
Euryarchaeota	0.502	0.470	0.311	0.066	0.05	0.40
Verrucomicrobia	0.710	0.179	0.096	0.105	<0.01	0.10
Chloroflexi	0.189	0.139	0.073	0.031	<0.01	0.71
Synergistetes	0.128	0.070	0.040	0.028	<0.01	0.60
Thermi	0.082	0.062	0.050	0.009	0.01	0.73
Deferribacteres	0.011	0.003	0.004	0.002	0.04	0.11
Caldithrix	0.006	0.001	0.001	0.002	0.03	0.11
Uterine samples
Above 1% relative abundance
Firmicutes	35.849	36.229	31.389	2.632	0.16	0.36
Proteobacteria	26.893	22.645	28.051	2.703	0.74	0.12
Bacteroidetes	22.062	27.921	24.233	3.394	0.59	0.25
Actinobacteria	8.519	7.661	7.847	1.595	0.69	0.75
Tenericutes	1.519	2.226	3.980	1.605	0.15	0.70
Spirochaetes	1.674	1.999	1.489	0.470	0.75	0.42
Below 1% relative abundance
Fibrobacteres	0.818	1.215	0.222	0.311	0.15	0.05
Chloroflexi	0.260	0.226	0.094	0.061	0.06	0.51
Verrucomicrobia	0.223	0.147	0.083	0.068	0.07	0.96
Synergistetes	0.103	0.083	0.045	0.025	0.09	0.74
Thermotogae	0.001	0.002	0.004	0.001	0.08	0.42

¹HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Clemmons et al., 2017; Ault et al., 2019a, 2019b), and results from day 0 used as independent covariate for each respective analysis.

Firmicutes was the most abundant bacterial phylum in the vaginal and uterine environments but was unaffected by treatment (P ≥ 0.16). There was a tendency (P = 0.09) for relative abundance of Spirochaetes to decrease as concentrate inclusion increased (Table 5), but no other treatment effects were noted for bacteria with relative abundance ≥1% in vaginal flush samples (Table 5). Uterine bacterial phyla greater than 1% relative abundance were also not affected (P ≥ 0.12) by treatments (Table 5). Phyla in vaginal samples that were less than 1% relative abundance and were affected by treatments (P ≤ 0.05) included Euryarchaeota, Verrucomicrobia, Chloroflexi, Thermi, Synergistetes, Deferribacteres, and Caldithrix (Table 5). Fibrobacteres was the only phyla in uterine samples with less than 1% relative abundance that was impacted (quadratic effect; P = 0.05), whereas tendencies (P ≤ 0.09) were noted for Verrucomicrobia, Chloroflexi, Synergistetes, and Thermotogae (Table 5).

Reproductive flushes bacterial community composition – genus

The relative abundance of bacteria derived from each individual reproductive flush sample was assigned to 873 different genera. A total of 14 vaginal and 3 uterine genera with relative abundance ≥1% (average across treatments; Figure 2) were affected by treatment (Tables 6 and 7). Bacterial genera within reproductive flushes with relative abundance <1% (average across treatments) that were affected by treatments are also listed in Tables 6 and 7.

Figure 2.

Relative abundance of bacterial genera. Samples were grouped by treatment and uterus or vagina. Treatment group abbreviations: HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Clemmons et al., 2017; Ault et al., 2019a, 2019b), and results from day 0 used as independent covariate for each respective analysis.

Table 6.

Treatment effects on bacterial genera (relative abundance, %) in vaginal flush samples of beef heifers¹

	Treatments				P-value
Genus	HF	INT	HG	SEM	Linear	Quadratic
Above 1% relative abundance
Prevotella	7.786	10.350	18.623	2.027	<0.01	0.22
Blautia	7.173	9.700	8.162	1.300	0.58	0.22
Ruminococcus	9.526	7.954	6.861	3.560	0.07	0.88
Bacteroides	9.016	6.400	6.965	1.038	0.15	0.21
Clostridium	5.379	3.940	3.648	0.794	<0.01	0.11
Pedobacter	6.073	4.246	2.172	0.677	<0.01	0.81
Succinivibrio	1.135	0.527	5.208	1.771	0.07	0.14
Roseburia	1.221	2.226	3.313	0.656	0.04	0.92
Oscillospira	2.810	2.124	1.805	0.288	0.02	0.65
Faecalibacterium	1.446	1.469	3.589	0.697	0.04	0.22
Treponema	2.719	2.077	1.604	0.565	0.09	0.90
Corynebacterium	2.366	1.919	1.251	0.540	0.13	0.82
Bifidobacterium	0.948	1.256	2.946	0.713	0.07	0.41
Caloramator	2.192	1.668	0.905	0.248	<0.01	0.58
Alkaliphilus	1.293	1.146	1.481	0.266	0.63	0.44
Paludibacter	1.422	1.875	0.329	0.345	0.04	0.02
Rhodothermus	1.730	1.095	0.473	0.160	<0.01	0.92
Porphyromonas	1.590	1.174	0.365	0.184	<0.01	0.30
Below 1% relative abundance
Paraprevotella	0.961	1.002	0.599	0.117	0.06	0.04
Luteibacter	1.043	0.502	0.208	0.141	<0.01	0.54
Mogibacterium	0.555	0.384	0.361	0.061	0.04	0.34
Sphingobacterium	0.614	0.367	0.107	0.123	<0.01	0.87

¹HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and 28 (Clemmons et al., 2017; Ault et al., 2019a, 2019b), and results from day 0 used as independent covariate for each respective analysis.

Table 7.

Treatment effects on bacterial genera (relative abundance, %) in uterine flush samples of beef heifers¹

	Treatments				P-value
Genus	HF	INT	HG	SEM	Linear	Quadratic
Above 1% relative abundance
Prevotella	8.364	12.513	11.650	2.431	0.29	0.40
Bacteroides	4.868	4.672	6.033	0.846	0.27	0.41
Blautia	3.776	4.103	3.944	0.692	0.78	0.69
Ruminococcus	3.616	3.892	3.970	0.714	0.69	0.91
Corynebacterium	4.261	2.996	3.815	0.772	0.62	0.26
Halomonas	3.755	3.531	3.332	0.991	0.74	0.99
Mesorhizobium	4.306	2.633	2.663	0.804	0.13	0.44
Pedobacter	3.368	3.738	2.368	0.443	0.12	0.12
Clostridium	2.756	2.918	2.321	0.327	0.33	0.34
Acinetobacter	3.420	1.801	2.742	0.709	0.45	0.16
Succinivibrio	1.237	1.145	4.584	1.575	0.13	0.37
Treponema	1.877	2.180	1.673	0.551	0.76	0.49
Caloramator	2.201	1.811	0.896	0.311	<0.01	0.47
Ruminobacter	1.748	1.760	1.389	0.916	0.66	0.78
Oscillospira	1.540	1.781	1.469	0.267	0.85	0.39
Bacillus	1.923	0.976	1.147	0.429	0.17	0.32
Thalassospira	0.788	2.568	0.539	0.542	0.79	0.01
Caldicellulosiruptor	1.974	0.858	1.013	0.398	0.06	0.27
Paludibacter	0.684	1.882	0.654	0.313	0.99	<0.01
Below 1% relative abundance
Dysgonomonas	0.918	1.203	0.588	0.193	0.21	0.05
Pseudomonas	1.388	0.682	0.587	0.282	0.03	0.39
Fibrobacter	0.927	1.360	0.243	0.355	0.15	0.05

¹HF: 100% grass hay, INT: 60% grass hay + 40% corn-based concentrate, HG: 25% grass hay + 75% corn-based concentrate. Treatments were provided to heifers for 28 d. Samples were collected on days 0 and28 (Clemmons et al., 2017; Ault et al., 2019a, 2019b), and results from day 0 used as independent covariate for each respective analysis.

Prevotella was the most abundant genus present in the vagina, and its relative abundance increased (P < 0.01) with the inclusion of concentrate. Other genera, with relative abundance ≥1%, that were significantly affected (P ≤ 0.05) by dietary treatments were Clostridium, Pedobacter, Roseburia, Oscillospira, Faecalibacterium, Caloramator, Paludibacter, Rhodothermus, and Porphyromonas, whereas tendencies were noted (P ≤ 0.09) for Ruminococcus, Succinivibrio, Treponema, and Bifidobacterium. Genera with relative abundance ≤1% that were affected (P ≤ 0.04) by treatments were Paraprevotella, Luteibacter, Mogibacterium, and Sphingobacterium.

Within the uterine environment, Prevotella was also the most abundant genera but its relative abundance was not affected by treatments (P ≥ 0.29). Relative abundance of Caloramator decreased (P < 0.01) with the inclusion of concentrate, and a tendency (P = 0.06) for the same outcome was noted for Caldicellulosiruptor. Relative abundance of Paludibacter was quadratically affected (P = 0.05) by treatments. Genera with relative abundance ≤1% that were affected (P ≤ 0.04) by treatments were Dysgonomonas, Pseudomonas, and Fibrobacter.

Discussion

Studies conducted in livestock have identified that the reproductive tract of healthy animals have a substantial diversity in bacterial species (Swartz et al., 2014; Clemmons et al., 2017; Ault et al., 2019a, 2019b; Quereda et al., 2020). Although these bacterial communities have many beneficial contributions to the animal, dysregulation of these communities (i.e., dysbiosis) concurrent with a change in bacterial abundances may have negative consequences for reproductive performance (Macklaim et al., 2013; Belizário and Napolitano, 2015; Green et al., 2015). In the current study, changes in the bacterial communities’ diversity in the rumen, due to varying levels of concentrate in the diet, potentially shifted bacterial abundances in the reproductive tract of beef heifers.

The experimental model and concentrate inclusion rates used herein were successful in altering ruminal fermentation parameters and microbial populations in the rumen. For ruminal pH, increasing concentrate levels decreased the pH of the rumen and increased ruminal concentrations of VFAs including propionate, butyrate, isovalerate, isobutyrate, and valerate. These outcomes were expected and are in agreement with previous research (Petri et al., 2012; Mao et al., 2013; Plaizier et al., 2018). The microbial diversity and relative abundance of significant bacteria (relative ≥1%) acted in a similar way as previous research, and reduced the richness and diversity of ruminal fluid microbiota (Fernando et al., 2010; Petri et al., 2012).

As shown in this study and in previous research, the most common phyla in the rumen are Bacteroidetes, Firmicutes, and Proteobacteria (Khafipour et al., 2009; Petri et al., 2013; Ramos et al., 2021). The phylum Bacteroidetes is known for degrading structural carbohydrates and can be sensitive to more acidic environments, which resulted the linear decrease in its relative abundance as concentrate inclusion increased (Kaoutari et al., 2013; Ramos et al., 2021). In accordance with previous research, Prevotella within the phylum Bacteroidetes, was the most abundant genus and is a well-known degrader of starch, β glycans, protein, pectin, and hemicellulose in the rumen (Henderson et al., 2015). This genus can use a variety of substrates which allows it to dominate the rumen environment in a wide range of diets. Bacterial genera within the Bacteroidetes phylum that were significantly affected by treatments also included Pedobacter, Porphyromonas, and Dysgononmonas. All three significantly decreased with concentrate inclusion likely due to reduced levels of structural carbohydrates within the rumen. The relative abundance of the phylum Firmicutes increased with concentrate inclusion, in accordance with previous research showing that concentrate intake promotes bacterial genera from this phylum in the rumen (Cristobal-Carballo et al., 2021). This included the genera Oscillospira, Blautia, Ruminococcus, and Butyrivibrio which had the greatest relative abundance in HG heifers. Firmicutes in gut microbiota are known to ferment carbohydrates to a variety of short-chain fatty acids and utilize dietary fiber more efficiently than Bacteroidetes, which could explain the decrease in the relative abundance of Bacteroidetes and increase in Firmicutes according to dietary concentrate inclusion. Contrary to previous research, Proteobacteria in this study decreased with increasing levels of concentrate, which could have been caused by the supplementation of MGA. Owens (2020) found that Proteobacteria were negatively correlated with serum progesterone concentration. Heifers received the synthetic progesterone, MGA, daily throughout the feeding period and could have altered the results for Proteobacteria.

The relationship between diet induced changes in ruminal flora and its contribution to the reproductive microbiome has not previously been evaluated in cattle. The current study found significant associations of bacterial genera present in the vaginal environment with the different levels of concentrate added to the diet, as well as differences in ruminal flora. Pedobacter and Porphyromonas acted similarly in the vaginal and rumen environment with decreased amounts found in the cattle receiving higher levels of concentrate. Porphyromonas is a reproductive pathogen, but commonly inhabit the rumen and are excreted in feces (Appiah et al., 2020). Based on the current results, diet could have a direct (e.g., endogenous) impact on vaginal bacteria, though it cannot be ruled out as indirect contribution (e.g., feces). Paraprevotella was significantly affected by diet and acted quadratically with the INT cattle having the highest relative abundance both vaginal and rumen environments. Zhang et al. (2018) reported that cattle receiving an intermediate concentrate diet experienced lower fecal prevalence of Paraprevotella, suggesting that treatment differences noted herein for this genus were also resultant from a endogenous contribution of ruminal flora to the vaginal environment.

The current study also observed associations between dietary treatments and their effects on the prevalence of Prevotella and Clostridium within the vaginal environment. Previous research suggests that Prevotella is associated with uterine disease and infertility in cattle (Jeon et al., 2015; Ault et al., 2019b; Lima et al., 2019). Ault et al. (2019a) reported cattle that failed to become pregnant had a greater relative abundance of Prevotella in the uterus two days prior to artificial insemination. Clostridium is also a genus of bacteria that contains several pathogenic species that can be detrimental to reproduction (Williams et al., 2005; Wang et al., 2016; Kronfeld et al., 2022). Clostridium has also been associated with a decrease in reproductive success, causing dysbiosis of the bacterial community and increase in reproductive diseases (Kronfeld et al., 2022). Interestingly, concentrate inclusion linearly increased the relative abundance of Prevotella but linearly decreased the relative abundance of Clostridium in the vaginal environment of heifers evaluated herein. Caloramator was the only genus of bacteria with substantial abundance (≥1% average across treatments) that was affected treatments in the rumen, vagina, and uterus, and linearly decreased across environments with the inclusion of concentrate. This genus of bacteria has previously been associated with a healthy uterine environment in cattle (Jeon et al., 2015; Lima et al., 2019), suggesting that the HF diet yields in healthier vaginal and uterine environments. Nonetheless, all these results are novel and research is warranted to explore additional differences in both vaginal and uterine microbiota according to dietary profile, and subsequent impacts to heifer reproductive responses.

In conclusion, the taxonomic composition of the rumen and reproductive tract can shift in the relative abundance of bacteria due to concentrate inclusion in the diet. Previous research highlighted the changes in bacterial flora in the rumen due to dietary shifts (Petri et al., 2012; Mao et al., 2013; Plaizier et al., 2018), but how that contributed to the uterine and vaginal microbiome was previously unknown. This study demonstrated that concentrate inclusion can affect not only the rumen but can also subsequently affect the reproductive microbiome. Most changes occurred in the vagina, which could be due to the anatomy of cow’s reproductive tract and the constant introduction of these bacteria in the vagina from the feces. The uterus was observed to be less affected by treatment possibly due to the uterus being more restricted from external exposure compared to the vagina (Sheldon et al., 2002). Protection of the uterine environment from the external environment by cervical barrier may contribute to the decreased diversity present in the uterine environment (Clemmons et al., 2017). Future research is needed to evaluate the direct and indirect contributions that the ruminal flora has on the reproductive microbiome, whereas continued investigation of the reproductive microbiome is warranted to understand how altering that environment can affect reproductive efficiency in cattle.

Glossary

Abbreviations

BW

body weight

HF

high forage diet (100% grass hay)

HG

high grain diet (75% corn-based concentrate + 25% grass hay)

INT

intermediate diet (40% corn-based concentrate + 60% grass hay)

MGA

melengestrol acetate

VFA

volatile fatty acid

Conflict of Interest Statement

The principal investigator (Reinaldo F. Cooke) and all other authors of this manuscript have no conflict of interest to report.