Dose Effect of Drinking Water Nitrate on Health, Feed Intake, Rumen Fermentation and Microbiota, and Nitrogen Excretion in Holstein Heifers for a Sustainable Water Use

Abstract

The present study aimed to evaluate the potential hazardous effects of NO₃⁻ concentration in drinking water on health, feed intake, rumen fermentation and microbiota, and nitrogen excretion of Holstein heifers fed a high-concentrate diet for a sustainable water use. Twenty-four Holstein heifers were individually allocated and assigned to one of four treatments with increasing drinking water NO₃⁻ concentration: CTR, without NO₃⁻; LOW, with 44 mg NO₃⁻/L; MOD, with 110 mg NO₃⁻/L; and HIGH, with 220 mg NO₃⁻/L. The entire study lasted 168 days. Fortnightly water NO₃⁻ concentration and daily feed and water intake were recorded. Blood parameters, rumen pH, volatile fatty acids, NO₃⁻ and NO₂⁻ concentration, microbiota, and apparent total tract digestibility were determined at the beginning and at the end of the study. Most of the analyzed parameters were similar among treatments. Denitrifying bacteria population, estimated as nosZ gene copies, were greater in HIGH animals than in CTR animals at the end of the study. In conclusion, drinking water NO₃⁻ concentration up to 220 mg/L has no detrimental effect on health, feed intake, rumen fermentation, nor N excretion in dairy beef cattle for periods up to 168 days; moreover, denitrifying bacteria population increased, which are related with the neutralization of the greenhouse gas N₂O.

1. Introduction

Monitoring and managing water quality for livestock systems are essential to maximise the use of a precious natural resource like water and minimise the risk of adverse effects related to potential toxic compound content that may limit animal production [1]. Overuse of fertilizers in agriculture is one of the origins of high water NO₃⁻ content [2,3,4]. Nitrates per se are not toxic for ruminants, but they become potentially hazardous for their health when ruminal microbes convert them to NO₂⁻ and N₂O by means of partial dissimilatory denitrification [5]. In addition, NO₃⁻ and NO₂⁻ can be a source of N that can be transformed by rumen microbiota to NH₃ that is converted to microbial protein, thereby affecting animal protein metabolism and N excretion. In the case that there is an excess of NO₂⁻, it is absorbed across the rumen wall and converts blood haemoglobin (Hb) into methaemoglobin (MtHb), a molecule incapable of transporting oxygen to the tissues [6]. This can cause acute or chronic poisoning [7,8,9]. Recently it has been described that consuming feeds rich in NO₃⁻ may reduce rumen CH₄ production [10,11,12]; whether drinking water NO₃⁻ exhibits the same effects on rumen CH₄ production as feed NO₃⁻ has not been documented.

Scientific literature on ruminant drinking water is extensive, mainly focused on how ruminants are impacted by water quality [5,13,14], water disinfection [15,16,17], and water availability [18,19,20]. However, there is currently only one published reference that provides a guideline for the drinking water NO₃⁻ threshold in cattle to avoid its potential hazard effects [21]. The maximum safe concentration of water NO₃⁻ is 44 mg/L and, if diet ingredients are low in NO₃⁻ content, drinking water concentrations between 44 and 132 mg/L of NO₃⁻ may be acceptable [21]. In ruminants, NO₃⁻ can be metabolized in the rumen and its metabolism and microbiota is dependent on the feed consumption and the type of diet (feed NO₃⁻ content and fermentability of the diet); as such, those drinking water NO₃⁻ threshold references may be questioned. Moreover, the effect of the exposure time (days of consumption) to NO₃⁻ intake on animal health is not well established. We hypothesise that the capability of cattle fed high-concentrate diets to metabolize in the rumen the end products of NO₃⁻ (non-protein N sources; [22]) may reduce the potential hazardous effects (less rumen NO₃⁻ and NO₂⁻ available for potential absorption) for a long exposure time (over 5 months), thereby promoting a sustainable water use. This hypothesis is supported by a 10-year overview where drinking water quality was analysed in 36 batches of fattening farms (Guissona, Lleida, Spain), showing a mean value of 20 mg/L of NO₃⁻, with a maximal value of 75 mg/L and punctual events of 100 and 210 mg/L, highlighting that no detrimental effects on health and/or performance were recorded. Therefore, the novelty of the current study is to provide evidence for this hypothesis and evaluate the potential hazardous effects of NO₃⁻ concentration in drinking water on animal health, feed and water intake, rumen fermentation and microbiota, and N excretion in dairy beef heifers fed a high-concentrate diet for a long exposure time for a sustainable water use. An abstract that reported preliminary results from this study has been published [23].

2. Materials and Methods

2.1. Animals, Housing, Experimental Design, and Treatments

Twenty-four Holstein heifers (161 ± 19.9 kg of body weight (BW), and 183 ± 25.3 days of age) were allocated in individual slatted floor pens (1.9 × 3.3 m) at the Nial experimental farm of the Corporación Alimentaria Guissona, S.A.—bonÀrea Agrupa (Guissona, Lleida, Spain) between January and July 2020. The experimental farm was naturally ventilated and illuminated, and each pen was equipped with two feeders (0.6 × 1.2 × 0.3 m each) and a water trough. As urine sampling is easier and safer with heifers than with intact bulls, heifers were used in the present study even though intact bulls are the main type of animal used in the dairy beef production system in our region. The study had a randomised balanced design with covariance adjustment, in which animals were randomly assigned to one of four treatments with increasing drinking water NO₃⁻ concentration: CTR (n = 6), without addition of NO₃⁻; LOW (n = 6), with 44 mg/L of NO₃⁻ (maximal safe dose defined by National Research Council (NRC, 2001) guidelines [21]); MOD (n = 6), with 110 mg/L of NO₃⁻ (2.5-fold maximal safe dose); and HIGH (n = 6), with 220 mg/L of NO₃⁻ (5-fold maximal safe dose). The study lasted 168 days divided into 12 periods of 14 days. Drinking water used in this study was collected from the Segarra-Garrigues waterway irrigation system, and acidification (0.65 mL/L H₃PO₄ with 236 g/L (Tashia S.L., Artesa de Segre, Lleida, Spain)) and chlorination (0.14 mL/L NaClO 15% with 36 g/L (Tashia S.L., Artesa de Segre, Lleida, Spain)) were applied to disinfect it. The addition of NO₃⁻ was achieved by using KNO₃ (bonÀrea Agrupa, Guissona, Lleida, Spain) at 22,000 mg/L, which was dosed in the water supply system through an automatic dosage system (Grundfos Holding A/S, Bjerringbro, Denmark) at 2 mL/L in LOW treatment, 5 mL/L in MOD treatment, and 10 mL/L in HIGH treatment.

Animals were fed, ad libitum, a commercial pelleted concentrate formulated to cover their nutritional requirements [24]. The ingredient composition of the concentrate was: 39.9% corn, 17.9% barley, 15.0% corn gluten feed, 10.0% wheat bran, 6.59% wheat, 4.42% soybean, 2.98% beet pulp, 1.37% calcium carbonate, 0.80% by-pass fat, 0.50% sodium bicarbonate, 0.44% palm oil, and 0.10% salt. The nutrient composition of the concentrate on dry matter (DM) basis was: 3.63 Mcal/kg metabolizable energy, 13.8% crude protein (CP), 3.93% ether extract, 5.41% ash, 19.1% neutral detergent fibre (NDF), and 57.8% non-fibre carbohydrates. The NO₃⁻ and NO₂⁻ content of the concentrate were 11 and <5 mg/kg, respectively. Animals had ad libitum access to wheat straw (3.88% CP, 83.15% NDF, and 5.04% ash: DM basis) and water. The NO₃⁻ and NO₂⁻ content of the straw was 182 and 47 mg/kg, respectively.

2.2. Water Quality Measurements

Fortnightly determinations of the water NO₃⁻ concentration of each treatment were used to verify the correct application of KNO₃. For this purpose, water samples were collected from each treatment central tap at the beginning of the water supply line, with a 1-1.5 L non-sterile bottle. Subsequently, in the bonÀrea Agrupa analytical laboratory (Guissona, Lleida, Spain), the water samples were mixed with a NO₃⁻ reagent kit based on a colorimetric analysis method with a measuring range of 0.0 a 30.0 mg/L NO₃⁻N, and photometer readouts were performed (HI83399; HANNA Instruments S.L., Eibar, Gipuzkoa, Spain). In addition, water samples from the main central tap were collected on days 0, 70, and 168 for physicochemical analysis (conductivity, pH, hardness, water dry residue, calcium, magnesium, chloride, sulphate, NO₃⁻, and NO₂⁻), and microbiological analysis (total coliform, Escherichia coli, faecal enterococci, and Clostridium perfringens). Water samples were collected with a 1–1.5 L non-sterile bottle for the physiochemical analysis, and with a 350 mL sterile bottle containing sodium thiosulfate (VWR, Part of Avantor, Radnor, PA, USA) for the microbiological analysis, and sent at 4 °C within the next 24 h to the bonÀrea Agrupa analytical laboratory (Guissona, Lleida, Spain) for analysis.

2.3. Water and Feed Intake

Water intake was automatically recorded by a water trough measuring system (iPERL sensor, SENSUS, Morrisville, NC, USA) with a water meter based in volume (m³) and speed (L/h), recording L of water consumed per animal in 1-h intervals, and summarised in daily intake. Weighed amounts of concentrate and straw were separately offered every day in the morning, and their refusals were weighed every 14 days using a scale (GRAM Group, Hospitalet del Llobregat, Barcelona, Spain), summarising the daily intake of each feed. Animal BW was recorded with a scale (MOBBA INDUSTRIAL CATALUNYA, S.A., Badalona, Barcelona, Spain) every 14 days to the end of the study.

2.4. Potential Hazardous Effects Measurements

Health status (skin and mucous membranes colour, breathing rate, coughing, nasal or ocular discharge, ears and head position, faecal consistency, abdominal bloat, and fever) was recorded daily. Haematological and biochemical parameters determinations related with the potential hazardous effects on animal health (white blood cell, red blood cell, haemoglobin, MtHb, haematocrit, mean corpuscular volume, mean corpuscular haemoglobin concentration, red cell distribution width, haemoglobin distribution width, absolute segmented neutrophil, absolute lymphocyte, absolute monocyte, absolute eosinophil, absolute large lymphocyte, absolute basophil, platelets, mean platelet volume, mean platelet concentration, platelet distribution width, albumin, alanine aminotransferase, amylase, aspartate aminotransferase, total bilirubin, calcium, creatine kinase, chloride, cholesterol, creatinine, alkaline phosphatase, phosphorous, γ-glutamyl transpeptidase, glucose, lactate dehydrogenase, magnesium, potassium, total protein, sodium, urea, and haptoglobin) were carried out according to the European Food Safety Authority guidelines for tolerance studies in animal feeding [25], collecting blood samples from each animal on days 0, 14, and 168. Bulls were moved into a squeeze chute (Priefert Ranch Equipment, S01-Model 91, Austin, TX, USA) for blood samples collection via jugular venipuncture using a vacutainer and an 18 G needle. For haematological analysis, 4 mL of blood was collected in ethylenediaminetetraacetic acid (EDTA) vacutainer tubes (BD, Franklin Lakes, NJ, USA), inverted, and stored at 5 °C until analysis. For MtHb analysis, 4 mL of blood was collected in heparin vacutainer tubes (BD, Franklin Lakes, NJ, USA). For biochemical analysis, 10 mL of blood was collected in spray-dried clot activator vacutainer tubes (BD, Franklin Lakes, NJ, USA). For ammonia analysis, 4 mL of blood was collected in EDTA dipotassium salt vacutainer tubes (BD, Franklin Lakes, NJ, USA). For glucose analysis, 4 mL of blood was collected in sodium fluoride and potassium oxalate vacutainer tubes (BD, Franklin Lakes, NJ, USA). The vacutainer tubes for biochemical and glucose analysis were then centrifuged at 1500× g at 4 °C for 15 min, and the serum from each tube was divided equally between three Eppendorf tubes. Samples were frozen at −20 °C for further analyses, except MtHb and ammonia samples, which were analysed within 24 h after sampling, maintaining MtHb samples refrigerated at 4 °C, and ammonia samples frozen in liquid N₂.

2.5. Rumen Fermentation and Microbiota

Rumen fluid sampling was conducted between 1 and 2 h after the morning feed offer on days 14 and 168. Rumenocentesis was the procedure used to obtain the rumen fluid, inserting a 14 cm 14 G needle (Abbocath-T; Hospira, Madrid, Spain) into the ventral sac of the rumen, around 15–20 cm caudal and ventral to the costochondral junction of the last rib. Rumen fluid pH was immediately measured after sampling with a portable pH meter (model 507, Crisson Instruments SA, Barcelona, Spain). After rumen pH measurement, rumen fluid was preserved based on [26] and as described [27] to the subsequent determination of volatile fatty acid (VFA), NO₃⁻ and NO₂⁻, and microbial population analysis (gene quantifying by quantitative polymerase chain reaction (qPCR) assays).

2.6. N Excretion and Apparent Total Tract Digestibility

Apparent total tract digestibility and N excretion were sampled at two rounds, between day 7 and 14 and between day 161 and 168. For the nutrient digestibility determination, chromium oxide (1 g/kg DM) was added to the concentrate as an indigestible marker. During the 7 days of each sampling round, concentrate offered and refusals samples from each animal were collected every day, and faecal grab samples were collected from the rectum on the last 3 days of each round of sampling. Faecal grab samples were dried at 100 °C for 48 h and composited by animal and sampling round on an equal DM basis. In parallel with faecal grab samples collection, a urine spot sample (100 mL) was obtained by perivaginal massage and frozen at −20 °C to determine creatinine and N contents.

2.7. Chemical Analyses

Physicochemical and microbiological water quality were analysed in the bonÀrea Agrupa analytical laboratory (Guissona, Lleida, Spain). The physiochemical water parameters analysed and analytical methodology used were as follows: conductivity, measured by a conductometer (Crison GLP 32, Hach Lange Spain S.L.U., Barcelona, Spain); pH, measured by a pH meter (Titrando 836, Metrohm AG, Herisau, Switzerland); calcium, magnesium, and chloride, measured by volumetric measurement (Titrando 836, Metrohm AG, Herisau, Switzerland); hardness, calculated by the calcium and magnesium equation ((Ca × 2.49) + (Mg × 4.11)) using Tiamo^TM software 3.0 (Titrando 836, Metrohm AG, Herisau, Switzerland); sulphate, measured by colorimetric test (VISOCOLOR Macherey-Nagel, Düren, Germany); and NO₃⁻ and NO₂⁻, measured by colorimetric test with 10–500 mg/L NO₃^– and 0.025–0.500 mg/L NO₂^–, respectively (Merck KGaA, Darmstadt, Germany). The microbiological water parameters analysed and analytical methodology used were as follows: total coliform, E. coli, faecal enterococcus and C. perfringens, measured as established in [28].

Concentrate and straw sampling were conducted monthly for nutrient composition determination. Samples were dried for 24 h at 103 °C for DM determination (method number 925.04; [29]), and for 4 h at 550 °C for ash determination (method number 642.05; [29]). The Kjeldahl method was used on samples for CP determination (method number 988.05; [29]). Neutral detergent fibre determination was based on the method of [30], specifically procedure A with the addition of sodium sulphite. Ether extract content was determined using a Soxhlet apparatus following acid hydrolysis preparation (method number 942.05; [29]). Total starch content was analysed using the polarimetric method for feed analyses according to [31]. In addition, NO₃⁻ in concentrate and straw samples were extracted using borax buffer solution and purified with C18 cartridges, and the NO₃⁻ concentration present in this extract was determined using a High Performance Liquid Chromatography (HPLC) method (Agilent technologies, Barcelona, Spain), using an anionic column with a diode array detector at 210 nm. Based on the same NO₃⁻ extract, the NO₂⁻ concentration present was determined by adding sulphanilic α-naphthylamine acid and measuring with a spectrophotometer (Jasco, Madrid, Spain).

Rumen NO₃⁻ and NO₂⁻ concentrations were analysed by ion chromatography (Metrohm 861 Advanced Compact IC), using a Metrohm Metrosep A Supp 4 column and pre-column ametrosep A Supp 4/5 Guard. Rumen VFA concentration was analysed with a semicapillary column (15 m by 0.53 mm i.d. and 0.5 μm film thickness; TRB-FFAP; Teknokorma, Barcelona, Spain) composed of 100% polyethylene glycol esterified with nitroterephthalic acid, using a CP-3800 Gas Chromatograph (Varian Inc., Walnut Creek, CA, USA) based on [25].

Regarding apparent total tract digestibility, each sample was subdivided into duplicates weighing 0.5 g prior to chromium oxide digestion. Chromium oxide digestion was divided into two parts. The first digestion step mixed the sample with 4 mL of HNO₃ at 220 °C for 15 min in a microwave oven (Ultrawave model, Milestone, Sorisole, Italy), resulting in colourless solutions with a green solid at the bottom of the digestion tube. This solid is attributed to Cr₂O₃(s). In the second step, 3 mL of H₂SO₄, 0.5 mL of HClO₄, and 2 mL of hydrofluoric acid were added to the same digestion tube, starting a new digestion procedure at 260 °C for 15 min. Finally, the Cr concentration was measured by inductively coupled plasma optical emission spectrometry (model Optima 4300D, Perking-Elmer, Shelton, CT, USA). The recovery rate of chromium oxide was not measured in the present study as 100% recovery was assumed in agreement with Cr₂O₃ recovery rates measured by [32].

2.8. DNA Extraction and Bacterial Gene Quantification

To quantify the rumen microbial population, quantitative PCR assays were conducted on defrosted samples at 4 °C except in MOD treatment samples, which were not included in this determination. Total DNA was extracted from each rumen sample (0.25 g) through the DNeasy PowerSoil Pro Kit (Qiagen, Valencia, CA, USA), according to the manufacturer’s instructions. Regarding the quantitative assessment, gene copy numbers of 16S rRNA (total bacteria), mcrA gene (methanogenic archaea), aprA gene (sulphate reducing prokaryotes (SRP)), amoA gene (ammonia-oxidising bacteria (AOB), and ammonia-oxidising archaea (AOA)), the clade I nosZ gene (typical complete denitrifying bacteria (clade I) linked to the transformation of N₂O to N₂) and hzo (anaerobic ammonium oxidation (ANAMMOX) bacteria) were quantified by quantitative real-time PCR (qRT-PCR). Samples were analysed in triplicate, employing 6 independent DNA extracts (6 animals per treatment at days 14 and 168). The analyses were accomplished using the Brilliant II SYBR ^®Green qPCR Master Mix (Agilent, Santa Clara, CA, USA) in a real-time PCR System MX3000-P (Stratagene, La Jolla, CA, USA) as described in the literature [33,34,35]. Ten-fold serial dilutions from synthetic genes were submitted to duplicate qPCR assays showing a linear range between 101 and 108 gene copy numbers per reaction, generating standard curves. Reactions of qPCRs fitted to quality standards of MIQE Guidelines [36]: efficiencies were between 90 and 110% and R² above 0.985.

2.9. Calculations and Statistical Analysis

Intake of NO₃⁻ through concentrate, straw, and water was estimated by multiplying concentrate, straw, and water intake by their corresponding NO₃⁻ content. Apparent total tract digestibility data were calculated using total faecal output, which was further estimated as the ratio of chromium intake to chromium concentration in faeces. The total volume of urine was estimated assuming 883 µmol-creatinine per kg-metabolic BW and day [37].

Each animal was considered an experimental unit. The univariate procedure of SAS^® software 9.4 (Statistical Analysis Systems, Cary, NC, USA) verified that all data were normally distributed, except gene copy data, which were analysed by the Shapiro-Wilk test to determine normal distribution data.

Data, including water and feed intake, blood parameters, apparent total tract digestibility and N excretion, and rumen parameters, were analysed using a mixed-effects model with repeated measures using SAS^® software 9.4 (Statistical Analysis Systems, Cary, NC, USA). The model included initial BW as a covariate and treatment, period, and interactions between them as fixed effects. The model was also tested for linear and quadratic effects of treatment. Analysis of water and feed intake, blood parameters, apparent total tract digestibility and N excretion, and rumen parameters included animal as a random effect. Period was considered a repeated measure in the analysis of water and feed intake, blood parameters, apparent total tract digestibility and N excretion, and rumen parameters, and animal was subjected to four variance-covariance structures, which included compound symmetry, variance components, autoregressive order one, and heterogeneous autoregressive order one. The variance-covariance structure that minimised Schwarz’s Bayesian criteria was considered the most suitable analysis. Rumen microbiota data were processed by MxPro™ QPCR Software v4.1D (Stratagene, La Jolla, CA, USA) and treated statistically by means of SigmaPlot and XLSTATS software 11.0. An ANOVA was performed for each universal and functional gene. Subsequently, pairwise comparisons (Fisher’s least significant difference) were applied to test differences between treatments. For all analyses, significant differences were accepted if p ≤ 0.05 and tendencies were discussed at 0.05 < p ≤ 0.10.

3. Results

No health incidences were observed during the whole study and no animal was removed from the study for statistical analysis.

3.1. Water Quality

Water used for treatment application had the following water quality characteristics (mean ± standard error mean (SEM)): 271 ± 3.7 μs/cm of conductivity, 6.44 ± 0.012 of pH, 4 ± 0.6° of hardness, 173 ± 14.5 mg/L of dry residual, 19 ± 4.3 mg/L of calcium, 17 ± 4.2 mg/L of magnesium, 18 ± 0.8 mg/L of chloride, 31 ± 3.1 mg/L of sulphate, 0 ± 0.0 mg/L of NO₂⁻, and 0 ± 0.0 colony-forming units/100 mL in all microbiological parameters. The mean ± SEM of water NO₃⁻ concentration of CTR, LOW, MOD, and HIGH treatments was 2.4 ± 0.30 mg/L, 49.3 ± 2.01 mg/L, 123.9 ± 6.13 mg/L, and 244.6 ± 9.66 mg/L, respectively. The gap between theoretical and observed water NO₃⁻ content increased by around 10% in all observed values of the treatments with some NO₃⁻ dosage (i.e., LOW, MOD, and HIGH treatments). Ultimately, the observed concentrations were close to the theorical concentrations.

3.2. Water and Feed Intake

Concentrate, straw, and water intake were not affected by treatment (

Table 1. Feed intake and water and NO₃⁻ intake of dairy beef heifers during the 168 days of the drinking water NO₃⁻ study.

Items	Treatment 1					p-Value
	CTR	LOW	MOD	HIGH	SEM	T	P	T × P
n	6	6	6	6
Feed intake (kg DM/day)
Concentrate	6.49	6.61	6.02	6.27	0.283	0.21	0.01	0.87
Straw	0.61	0.65	0.66	0.70	0.065	0.58	0.01	0.03
Total	7.10	7.26	6.69	6.97	0.285	0.26	0.01	0.48
Water intake (L/day)	28.4	27.0	24.4	25.9	1.99	0.27	0.01	0.75
NO3− intake (mg/day)
Concentrate	71	73	66	69	-	-	-	-
Straw	111	118	120	127	-	-	-	-
Water	67	1330	3023	6335	-	-	-	-
Total	249	1521	3210	6532	-	-	-	-

Abbreviations: CTR = 0 mg/L of NO₃⁻ in drinking water; LOW = 44 mg/L of NO₃⁻ in drinking water; MOD = 110 mg/L of NO₃⁻ in drinking water; HIGH = 220 mg/L of NO₃⁻ in drinking water; SEM = standard error mean; T = treatment effect; P = period effect; T × P = interaction effect between treatment and period; DM = dry matter. ¹ Drinking water NO₃⁻ concentration

). A significant treatment × period interaction (p = 0.03) was observed in straw intake, being lower in CTR animals than in MOD and HIGH animals in period 3 (0.51 vs. 0.67 and 0.67 ± 0.078 kg DM/day) and being lower in CTR animals than in HIGH animals in period 10 (0.66 vs. 0.83 ± 0.081 kg DM/day). Although not statistically significance, water intake was numerically reduced in MOD and HIGH treatments. Intake of NO₃⁻ through concentrate, straw, and water is presented in Table 1.

3.3. Haematological and Biochemical Blood Parameters

For most haematological (

Table 2. Haematological parameters of dairy beef heifers during the 168 days of the drinking water NO₃⁻ study.

Items	Treatment 1					p-Value
	CTR	LOW	MOD	HIGH	SEM	T	P	T × P
n	6	6	6	6
WBC (K cells/µL)	8.83	8.40	8.23	8.83	0.832	0.84	0.79	0.49
RBC (×106 cells/µL)	8.85	9.03	9.15	8.82	0.403	0.83	0.01	0.98
Hb (g/dL)	12.0	11.9	12.0	12.0	0.49	0.99	0.01	1.00
MtHb (%)	2.42	2.40	2.46	2.42	0.154	0.97	0.01	0.17
HCT (%)	32.8	32.3	32.4	32.3	1.28	0.98	0.01	0.94
MCV (fL)	33.4	32.5	32.0	33.0	1.38	0.77	0.01	0.46
MCHC 2 (g/dL)	36.7 b	36.8 b	37.1 ab	37.4 a	0.23	0.03	0.01	0.99
RDW (%)	22.9	22.8	22.6	22.5	0.51	0.84	0.01	0.71
HDW (g/dL)	2.33	2.28	2.34	2.27	0.054	0.55	0.04	0.61
NEU (K cells/µL)	2.65	2.61	2.41	2.48	0.437	0.94	0.01	0.28
LYM (K cells/µL)	4.73	4.39	4.48	4.85	0.477	0.75	0.01	0.90
MONO (K cells/µL)	1.01	1.00	0.99	0.99	0.106	1.00	0.01	0.94
EOS (K cells/µL)	0.30	0.22	0.22	0.37	0.117	0.52	0.06	0.71
L-LYM (K cells/µL)	0.04	0.03	0.03	0.03	0.007	0.34	0.01	0.28
BASO (K cells/µL)	0.10	0.10	0.11	0.12	0.014	0.58	0.01	0.59
PLT (K cells/µL)	408	475	512	449	67.5	0.49	0.58	0.07
MPV 3 (fL)	7.03 a	6.64 b	6.41 b	6.50 b	0.163	0.01	0.01	0.85
MPC (g/dL)	24.4	23.7	24.2	23.7	0.61	0.53	0.04	0.25
PDW (g/dL)	6.53	6.53	6.57	6.49	0.108	0.91	0.01	0.48

Abbreviations: CTR = 0 mg/L of NO₃⁻ in drinking water; LOW = 44 mg/L of NO₃⁻ in drinking water; MOD = 110 mg/L of NO₃⁻ in drinking water; HIGH = 220 mg/L of NO₃⁻ in drinking water; SEM = standard error mean; T = treatment effect; P = period effect; T × P = interaction effect between treatment and period; WBC = white blood cell; RBC = red blood cell; Hb = haemoglobin; MtHb = methaemoglobin; HCT = haematocrit; MCV = mean corpuscular volume; MCHC = mean corpuscular haemoglobin concentration; RDW = red cell distribution width; HDW = haemoglobin distribution width; NEU = absolute segmented neutrophil; LYM = absolute lymphocyte; MONO = absolute monocyte; EOS = absolute eosinophil; L-LYM = absolute large lymphocyte; BASO = absolute basophil; PLT = platelets; MPV = mean platelet volume; MPC = mean platelet concentration; PDW = platelet distribution width. ¹ Drinking water NO₃⁻ concentration. ² p-value of the linear effect (p < 0.01) and of the quadratic effect (p = 0.41) of treatment. ³ p-value of the linear effect (p < 0.01) and of the quadratic effect (p = 0.04) of treatment. ^a,b Values within a row with different superscripts differ significantly at p ≤ 0.05.

) and biochemical blood parameters (

Table 3. Biochemical blood parameters of dairy beef heifers during the 168 days of the drinking water NO₃⁻ study.

Items	Treatment 1					p-Value
	CTR	LOW	MOD	HIGH	SEM	T	P	T × P
n	6	6	6	6
Ammonia (µmol/L)	58.7	56.3	57.9	59.1	5.94	0.81	0.01	0.02
Albumin (g/dL)	3.22	3.29	3.31	3.26	0.126	0.80	0.01	0.92
ALT (U/L)	19.0	17.8	18.5	18.4	1.81	0.89	0.52	0.47
Amylase (U/L)	126	125	124	124	23.1	1.00	0.84	0.22
AST (U/L)	78.8	71.2	74.3	72.8	8.49	0.73	0.01	0.38
Total bilirubin (mg/dL)	0.15	0.15	0.15	0.14	0.010	0.21	0.01	0.39
Calcium (mg/dL)	10.5	10.4	10.6	10.5	0.17	0.49	0.01	0.76
Creatine kinase (U/L)	179	188	185	182	121.8	1.00	0.35	0.74
Chloride (mmol/L)	98.3	98.0	97.5	97.6	1.11	0.79	0.01	0.81
Cholesterol (mg/dL)	123	129	123	123	15.8	0.94	0.02	0.86
Creatinine (mg/dL)	0.87	0.83	0.84	0.84	0.069	0.91	0.01	0.52
ALP (U/L)	285	214	274	240	44.0	0.27	0.01	0.37
Phosphorous (mg/dL)	9.03	9.10	8.85	8.93	0.429	0.79	0.05	0.32
GGTP (U/L)	19.6	20.9	17.4	18.6	3.94	0.68	0.01	0.68
Glucose (mg/dL)	92.1	90.7	88.2	88.2	5.36	0.57	0.01	0.61
LDH (U/L)	2856	2792	2774	2673	284.2	0.69	0.01	0.60
Magnesium (mg/dL)	2.86	2.58	2.71	2.54	0.235	0.38	0.01	0.73
Potassium (mg/dL)	4.65	4.48	4.45	4.49	0.152	0.20	0.01	0.48
Total protein (g/dL)	6.89	7.07	6.96	6.99	0.226	0.82	0.35	0.03
Sodium (mmol/L)	143	143	142	142	0.9	0.89	0.01	0.82
Urea N (mg/dL)	16.0	17.8	17.8	17.2	2.64	0.71	0.01	0.55
Haptoglobin (mg/mL)	0.27	0.19	0.20	0.20	0.164	0.56	0.67	0.17

Abbreviations: CTR = 0 mg/L of NO₃⁻ in drinking water; LOW = 44 mg/L of NO₃⁻ in drinking water; MOD = 110 mg/L of NO₃⁻ in drinking water; HIGH = 220 mg/L of NO₃⁻ in drinking water; SEM = standard error mean; T = treatment effect; P = period effect; T × P = interaction effect between treatment and period; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGTP = γ-glutamyl transpeptidase; LDH = lactate dehydrogenase. ¹ Drinking water NO₃⁻ concentration.

), including MtHb (2.43 ± 0.154% on average), no differences were observed among treatments except for mean corpuscular Hb, mean platelet volume, ammonia, and total protein. Mean corpuscular Hb concentration was greater (p = 0.03) in HIGH animals than in CTR and LOW animals, with MOD animals being intermediate, with a significant linear effect (p < 0.01) of treatment. Mean platelet volume was greater (p = 0.01) in CTR animals than in the other treatments, with a significant linear (p < 0.01) and quadratic (p = 0.04) effect of treatment. A tendency (p = 0.07) in treatment × period interaction was observed in platelets, being greater in MOD animals than in the other treatments in period 12 (416, 423, 548, and 399 ± 55.2 K cells/µL, for CTR, LOW, MOD, and HIGH treatments, respectively). A significant treatment × period interaction (p = 0.03) was observed in platelet distribution width with no clear pattern between treatments in any period. Also, significant treatment × period interaction (p = 0.02) was observed in ammonia, being greater in LOW animals than in CTR animals in period 1 (45.2, 53.2, 50.4, and 51.2 ± 3.00 µmol/L for CTR, LOW, MOD, and HIGH treatments, respectively) and being lower in LOW animals than in CTR and HIGH animals in period 8 (69.6, 55.4, 60.2, and 67.4 ± 4.93 µmol/L for CTR, LOW, MOD, and HIGH treatments, respectively). A significant treatment × period interaction (p = 0.03) was observed in total protein, being greater in LOW animals than in CTR animals at the beginning of the study (6.67, 7.14, 6.83, and 7.07 ± 0.226 g/dL, for CTR, LOW, MOD, and HIGH treatments, respectively).

3.4. Rumen pH, Volatile Fatty Acids, NO₃⁻ and NO₂⁻, and Microbiota

Treatment did not affect rumen pH, rumen VFA, nor NO₃⁻ and NO₂⁻ concentrations (

Table 4. Rumen pH, volatile fatty acids, NO₃⁻, and NO₂⁻ of dairy beef heifers during the 168 days of the drinking water NO₃⁻ study.

Items	Treatment 1					p-Value
	CTR	LOW	MOD	HIGH	SEM	T	P	T × P
n	6	6	6	6
pH	5.71	5.51	5.56	5.57	0.251	0.86	0.25	0.63
Total VFA (Mm)	158	179	169	161	19.9	0.65	0.01	0.87
Acetic	72.1	83.3	75.9	73.6	8.86	0.52	0.01	0.82
Propionic	63.4	72.3	68.6	63.8	10.87	0.79	0.01	0.96
i-Butyric	0.44	0.46	0.41	0.45	0.093	0.95	0.01	0.82
n-Butyric	12.4	13.6	14.4	13.4	1.92	0.77	0.01	0.51
i-Valeric	0.56	0.72	0.65	0.75	0.202	0.77	0.01	0.49
n-Valeric	3.46	3.53	4.56	3.42	0.811	0.47	0.01	0.25
i-Caproic	3.95	3.95	3.67	3.96	0.346	0.81	0.01	0.91
n-Caproic	1.16	0.95	1.06	0.99	0.354	0.92	0.01	0.75
n-Heptanoic	0.17	0.15	0.18	0.15	0.058	0.94	0.01	0.62
Acetic:Propionic	1.30	1.29	1.35	1.43	0.237	0.92	0.01	0.69
NO3− (ppm)	1.75	1.67	1.18	1.42	0.390	0.47	0.21	0.31
NO2− (ppm)	22.2	17.7	16.1	21.2	6.33	0.75	0.01	0.79

Abbreviations: CTR = 0 mg/L of NO₃⁻ in drinking water; LOW = 44 mg/L of NO₃⁻ in drinking water; MOD = 110 mg/L of NO₃⁻ in drinking water; HIGH = 220 mg/L of NO₃⁻ in drinking water; SEM = standard error mean; T = treatment effect; P = period effect; T × P = interaction effect between treatment and period; VFA = volatile fatty acids. ¹ Drinking water NO₃⁻ concentration.

). Each rumen microbial population remained quite stable among treatments, with minor quantitative differences (Figure 1). Total bacterial populations were found at around 10¹¹ 16S rRNA gene copies/mL, while methanogenic archaea were between 10⁹ and 10¹⁰ mcrA gene copies/mL. Dissimilatory SRP were between 10⁷ and 10⁸ aprA gene copies/mL, one and two orders of magnitude below methanogenic archaea. Ammonia-oxidising bacteria (average 10⁷ amoA_AOB gene copies/mL) were in a similar SRP range but slightly greater than the denitrifying population, in which the nosZ gene was detected at between 10⁶ and 10⁷ gene copies/mL. Specifically, the nosZ gene was greater in HIGH animals than in CTR animals in the second sampling at the end of the study (1.74 × 10⁷ vs. 8.21 × 10⁶ ± 2.754 × 10⁶ gene copies/mL, respectively; p = 0.025). Ammonia-oxidising archaea (10⁵–10⁶ amoA_AOA gene copies/mL) were detected below the denitrifying population and AOB, while ANAMMOX bacteria were less represented in rumen material (around 10⁵ hzo gene copies/mL).

3.5. Nitrogen Excretion and Apparent Total Tract Digestibility

No differences were found between treatments in N excretion and apparent total tract digestibility. Only significant treatment by period interaction (p = 0.01) was observed in water N intake (

Table 5. N excretion and apparent total tract digestibility of dairy beef heifers during the 168 days of the drinking water NO₃⁻ study.

Items	Treatment 1					p-Value
	CTR	LOW	MOD	HIGH	SEM	T	P	T × P
n	6	6	6	6
Intake
Concentrate (kg DM/day)	6.38	6.57	6.25	6.33	0.272	0.68	0.01	0.89
Straw (kg DM/day)	0.56	0.60	0.61	0.65	0.081	0.74	0.01	0.51
Total (kg DM/day)	6.95	7.17	6.86	6.98	0.259	0.68	0.01	0.87
CP (kg/day)	0.86	0.88	0.84	0.85	0.041	0.78	0.01	0.83
EE (kg/day)	0.25	0.25	0.24	0.25	0.011	0.78	0.01	0.76
OM (kg/day)	6.58	6.79	6.50	6.61	0.244	0.68	0.01	0.87
NDF (kg/day)	1.64	1.70	1.65	1.70	0.079	0.81	0.01	0.70
Starch (kg/day)	3.21	3.30	3.13	3.17	0.157	0.76	0.01	0.83
Faeces
Total (kg)	1.82	1.96	1.88	1.97	0.180	0.81	0.01	0.34
CP (kg/day)	0.28	0.30	0.29	0.30	0.025	0.94	0.01	0.47
EE (kg/day)	0.08	0.09	0.09	0.09	0.010	0.81	0.01	0.21
OM (kg/day)	1.65	1.77	1.70	1.78	0.163	0.85	0.01	0.34
NDF (kg/day)	0.97	1.09	1.04	1.07	0.109	0.72	0.01	0.30
Starch (kg/day)	0.10	0.08	0.09	0.10	0.022	0.83	0.01	0.35
Digestibility (%)
DM corrected	74.7	73.0	72.9	72.6	2.15	0.75	0.01	0.36
CP	67.4	66.7	66.5	65.9	2.62	0.95	0.01	0.46
CP corrected by water N intake	67.4	66.7	66.6	66.2	2.61	0.98	0.01	0.46
EE	67.8	65.7	64.8	64.5	4.63	0.89	0.01	0.33
OM	75.7	74.2	74.0	73.8	2.09	0.79	0.01	0.35
NDF	41.1	35.4	36.2	37.4	6.14	0.80	0.75	0.40
Starch	96.9	97.4	97.1	97.0	0.57	0.85	0.67	0.32
N balance (g/day)
N intake
Feed	138	142	135	137	6.6	0.77	0.01	0.83
Water 2	0.01 d	0.31 c	0.74 b	1.49 a	0.060	0.01	0.01	0.01
Total	138	142	136	138	6.6	0.82	0.01	0.82
N excretion
Faeces	45.5	47.4	45.6	47.4	4.01	0.93	0.01	0.43
Urine	32.5	39.6	40.8	35.8	5.37	0.41	0.03	0.96
Total	78.0	87.0	86.4	83.2	6.18	0.46	0.01	0.63

Abbreviations: CTR = 0 mg/L of NO₃⁻ in drinking water; LOW = 44 mg/L of NO₃⁻ in drinking water; MOD = 110 mg/L of NO₃⁻ in drinking water; HIGH = 220 mg/L of NO₃⁻ in drinking water; SEM = standard error mean; T = treatment effect; P = period effect; T×P = interaction effect between treatment and period; DM = dry matter; CP = crude protein; EE = ether extract; OM = organic matter; NDF = neutral detergent fibre. ¹ Drinking water NO₃⁻ concentration. ² p-value of the linear effect (p < 0.01) and of the quadratic effect (p < 0.01) of treatment. ^a–d Values within a row with different superscripts differ significantly at p ≤ 0.05.

), increasing from CTR animals to HIGH animals in the first sampling (0.01, 0.20, 0.48, and 1.10 ± 0.077 g/day; CTR, LOW, MOD, and HIGH treatments, respectively) and the second sampling (0.02, 0.43, 1.00, and 1.88 ± 0.077 g/day; CTR, LOW, MOD, and HIGH treatments, respectively).

4. Discussion

Drinking water NO₃⁻ content could harm the health of cattle [9]. As mentioned previously, depending on the enzymatic activity balance, the reduction sequence from NO₃⁻ to NH₃ can result in intermediate compound accumulation at any step, such as NO₂⁻, which is a potential health hazard [38]. Usually, the NO₃⁻ reduction to NO₂⁻ is faster than the reduction of NO₃⁻ to NH₃, and excess NO₂⁻ is readily absorbed across the rumen wall and oxidises blood Hb from the ferrous (Fe²⁺) to the ferric (Fe³⁺) form. The ferric form of Hb, MtHb, renders the molecule incapable of transporting oxygen to the tissues [6]. The resulting condition, methemoglobinaemia, is a state of general hypoxia, which, in mild cases, may depress animal performance, but in severe cases may be fatal [9]. In the present study, no signs of acute NO₂⁻ intoxication, such as hampered breathing, diarrhoea, polyuria, and/or anorexia, among others, were observed in any treatment. Supporting this absence of clinical signs, although some haematological and biochemical parameters were affected by treatment and/or period, all blood values were within the reference range by age in cattle [39,40]. Specifically, serum MtHb concentration, a parameter indicative of NO₂⁻ poisoning, was between 2.40% and 2.46% in all treatments of the present study. Moreover, these values were below the 10% upper threshold in healthy cattle [41]. Supporting the statement that rumen NO₃⁻ is rapidly converted into NO₂⁻, rumen NO₃⁻ content was low and did not differ among treatments. Rumen NO₂⁻ content was greater compared with rumen NO₃⁻ but also did not differ among treatments, and, if potentially absorbed across the rumen wall, it did not result in great serum MtHb concentrations, which explains the lack of clinical signs related to NO₂⁻ poisoning. However, drinking water intake and NO₃⁻ intake through drinking water differed numerically among treatments. In the present study, drinking water intake was numerically reduced to around 4 L/day in MOD animals and around 2.5 L/day in HIGH animals compared with CTR animals; therefore, NO₃⁻ intake through drinking water was also reduced in animals consuming water treatments with higher drinking water NO₃⁻ concentrations. These changes in water intake patterns could be an autoprotective mechanism against NO₂⁻ intoxication, which could have reduced the potential hazardous effects of drinking water with NO₃⁻ high levels. Drinking water and concentrate intake are positively related in cattle [42], so, probably because of numerical reduction of the drinking water intake in MOD and HIGH animals, their concentrate intakes were numerically reduced (reductions of 0.47 kg/day and 0.22 kg/day, respectively) compared with CTR animals. No statistically significant differences in drinking water intake between treatments were observed, probably due to the reduced number of animals used in the present study; nevertheless, reduced drinking water intake should be considered one of the first signs of poisoning. Cattle are sensitive to water palatability and prefer to drink water without contamination, such as salt, NO₃⁻, or faecal microorganism content, which are probably the most predictable factors reducing water palatability [13,14,43]. Decreasing water palatability may reduce drinking water intake and consequently decrease feed intake, and potentially impair animal performance [14,18,44]. In summary, no clear effects on health were observed when increasing drinking water NO₃⁻ concentration up to 220 mg/L; however, animals may have decreased their drinking water intake at high NO₃⁻ concentrations as a protective measure to avoid NO₃⁻ poisoning.

In the rumen, NO₃⁻ may be metabolised and can be a source of N for rumen bacteria and, therefore, affect rumen fermentation and N excretion. It also may modify microbial pathways, such as those related to greenhouse gas (GHG) production reducing enteric CH₄ gas emission [45]. In the present study, the drinking water NO₃⁻ supplementation had a small overall impact on total N entering the rumen as only 0–1% of the total N consumed was provided via water intake (Table 5). As mentioned previously, rumen NO₃⁻ content did not differ among treatments, and although rumen NO₂⁻ content was 15 to 22% greater than NO₃⁻ content, values were low compared with NO₃⁻ content consumed through water and concentrate. These low values (low N intake through drinking water in relation to total N intake, resulting in low rumen NO₃⁻ and NO₂⁻) may explain the neglible impact of drinking water NO₃⁻ concentration on rumen fermentation (pH and VFA concentrations), apparent total tract digestibility of all nutrients, and N excretion (N in faeces and urine), neither in the short (14 days) nor in long term (168 days) exposure time. Accumulation of NO₂⁻ in the rumen may alter rumen microbiota populations, inhibiting the growth and development of microorganisms that derive their energy from electron transport-mediated processes, among which are fibrolytic and cellulolytic bacteria [11]. In the present study, no effect of drinking water NO₃⁻ on rumen fibre digestibility was observed probably because low rumen NO₂⁻ concentrations. Crude protein apparent digestibility and N excretion data after the correction of the N consumed by drinking water were within reference values of dairy beef calves fed high-concentrate diets with 14% dietary CP [9,14] indicating that drinking water NO₃⁻ did not impair rumen fermentation nor digestibility or environmental N excretion (faeces and urine N).

Although the main objective of the study was to evaluate the detrimental effects of drinking water NO₃⁻ on animal health, also drinking water NO₃⁻ could have positive effects on environment reducing GHG emissions. Therefore, in the present study, the impact of drinking water NO₃⁻ on rumen microbiota was explored to support further studies where GHG emissions should be registered to confirm the hypothesis. The NO₃⁻ are inorganic anions with high redox potential that may act as an alternative sink for rumen dissolved H₂ [46]. As 8 electrons from H₂ are used to reduce both 1 mole of NO₃⁻ to NH₃ and CO₂ to CH₄, H₂ is a substrate required for both metabolism pathways, creating resource competition between them; as such, each mole of reduced NO₃⁻ diminishes CH₄ production by 1 mole. Therefore, it has been suggested that NO₃⁻ supplementation can be a good strategy to redirect H₂ toward alternative sinks for CH₄ production [47,48]. The potential effect of dietary NO₃⁻ supplementation on methanogenic inhibition and its potential use as feed additives has been previously evaluated in several studies [10,38,49]. In the present study, total NO₃⁻ intake was below those dosages used in these cited studies, and, in contrast to our preliminary expectations, no decrease in the total abundance of methanogenic archaea was observed, although this affirmation should be treated with caution as the expression of the mcrA gene was not assessed. Despite this, in the present study, the typical denitrifying population, encompassing the nosZ gene with the capacity of reducing N₂O to N₂, slightly increased in HIGH animals compared with the other treatments, suggesting a potential reduction in N₂O (a GHG) emissions from cattle eructation. Other pathways whereby the additional supply of NO₃⁻ to the rumen can promote competition mechanisms displacing methanogenic populations and reduce ruminal CH₄ generation are the stimulation of strategies of assimilative and non-assimilative denitrification [45]. In addition, a batch study with ruminal content showed that NO₃⁻ addition affected the rumen’s microbial community in a way that enriched species of anaerobic denitrifying bacteria and archaea that oxidise CH₄, the so-called denitrifying anaerobic CH₄ oxidation (DAMO; not analysed in the present study) [50]. In fact, DAMO and ANAMMOX bacteria, the latter analysed in the present study, usually coexist because DAMO oxidises CH₄ inside the rumen together with the denitrification pathway to N₂, and ANAMMOX bacteria oxidised NH₃ to N₂, also using NO₂⁻ generated during the denitrification pathway. The enrichment of this microbial consortia is advantageous for both animal health and environmental impact because it would avoid NO₂⁻ accumulation and decrease CH₄ emissions. In addition, AOB and AOA oxidise NH₃ to NO₂⁻, enhancing DAMO and ANAMMOX bacteria metabolism. As indicated previously, although CH₄ emissions were not recorded in the current study, the presence of AOB, AOA, and ANAMMOX bacteria populations was in a steady range in all treatments. The present study suggested that there is a potential impact of NO₃⁻ drinking water content on the metabolically active ruminal microbial consortia involved in GHG emissions that needs to be further studied; in these future studies, CH₄ and N₂O (i.e., GHG) emissions should be recorded. Moreover, the sampling frequency in relation to feeding time should be increased to provide a more detailed daily microbiota picture. In addition, microbiota population analyses, like DAMO, which are missing in the present study, are needed.

5. Conclusions

In conclusion, drinking water NO₃⁻ concentrations of up to 220 mg/L continuously for 5 months have no detrimental health effects in ruminants fed high-concentrate diets. Moreover, the increased rumen complete denitrifying bacteria population observed at the dose of 220 mg/L of NO₃⁻ in drinking water suggests a potential N₂O emission reduction; this potential GHG reduction should be further studied.