Use of Oral Citrulline Administration as a Mitigation Strategy for Fescue Toxicosis in Sheep

Background: Gestating ewes consuming ergot alkaloids, from endophyte-infected (E+) tall fescue seed, produce smaller lambs suffering from intrauterine growth restriction. Arginine (Arg) supplementation has been shown to increase birth weight and oral citrulline (Cit) administration is reported to increase arginine concentrations. Two experiments were conducted to: 1) evaluate if oral supplementation with Cit or water, to ewes consuming E+ fescue seed, increases lamb birth weight and 2) determine the effectiveness of Cit and citrulline:malate as an oral drench in elevating circulating levels of Cit to determine dosing levels and frequency of dose. In Experiment 1, gestating Suffolk ewes (n = 10) were assigned to one of two treatments [oral drench of citrulline-malate 2:1 (CITM; 81 mg/kg/d of citrulline) or water (TOX)] to start on d 86 of gestation. In Experiment 2, non-pregnant Suffolk ewes (n = 3) were assigned to either oral citrulline (CIT; 81 mg/kg/d), citrulline-malate 2:1 (CITM; 81 mg/kg/d of citrulline), or water (CON) drench in a Latin Square design for a treatment period of 4 d with a washout period of 3 d. On d 4, blood samples were collected at 0, 0.5, 1, 2, 3, 4, 6, 8, 10, 12, and 18 h post drench. Results: with citrulline-malate (CITM) or water (TOX) (Experiment 1). interactions citrulline-malate treatment and day of gestation were non-signicant (P 0.40) and results are presented as main effects.

greater in order to observe consistent elevation of Cit/Arg concentrations to determine its effectiveness in mitigating fescue toxicosis.

Background
Administration of arginine (Arg) to rats with induced hypertension has been found to increase vasodilation response when compared to hypertensive untreated rats (Kitazono et al., 1996). Wu et al. (2016) reviewed the effects of Arg administration on the dam and developing fetus and noted the importance of arginine to the developing embryo, as well as, the relative safety of large dose supplementation. Intravenous arginine administration has also been show to increase systemic vasodilation (Lorente et al., 1993) and lamb birth weight (Lassala et al., 2010). Extra cellular Arg is essential for the formation of nitric oxide, a vasorelaxant, in mesenteric artery tissue (MacKenzie and Wadsworth, 2003), and increasing circulating Arg concentrations induces vasodilation in humans and sheep (Lorente et al., 1993;Bode-Boger et al., 1998). In rats, blocking the production of nitric oxide by inhibiting nitric oxide synthase causes symptoms of preeclampsia and IUGR (Yallampalli and Gar eld, 1993). Unfortunately, Arg is almost completely degraded by rumen bacteria (Chalupa, 1976). Citrulline (Cit) may be a viable alternative to arginine, as it is part of the urea cycle and can be bio-transformed into Arg (Ah Mew et al., 2003). Citrulline has been con rmed to survive rumen degradation in vitro and in vivo with sheep (Gilbreath et al., 2020a) and with cattle (Gilbreath et al., 2020b). Additionally, Cit has been shown to be more effective at raising Arg concentrations when administered intravenously (Lassala et al., 2009).
Tall fescue [Lolium arundinaceum (Schreb.) Darbysh], a common pasture grass in the southeastern United States, is normally infected with an endophytic fungus (Epichloe coenophiala; previously Neotyphodium coenophialum) (Ball et al., 2007). The fungus allows the plant to have an overall hardiness that is desirable compared to other forage species in the area. Unfortunately, the fungus also produces ergot alkaloids (Stuedemann and Hoveland, 1988), which have been linked to decreased weight gain (Schmidt et al., 1982), reproductive problems (Peters et al., 1992;Porter and Thompson, 1992), and vasoconstriction (Dyer, 1993;Klotz et al., 2010;Aiken and Flythe, 2014). Feeding endophyte-infected (E+) fescue to pregnant livestock has been reported to cause intrauterine growth restriction (IUGR) that is evident at d 133 of gestation in sheep (Greene et al., 2019) and at birth in sheep and cattle (Watson et al., 2004;Duckett et al., 2014a). Reductions in fetal weight caused by ergot alkaloid exposure are speculated to be due to placental insu ciency caused by increased vasoconstriction (Britt et al., 2019). Additionally, Klotz et al. (2019) found that ovine dams dosed with ergot alkaloids from tall fescue seed had increased vasoconstriction in umbilical arteries. Little research has been conducted to determine if the oral administration of citrulline may improve lamb birth weight or IUGR associated with dam consumption of ergot alkaloids. The two objectives of this study were to: 1) evaluate if oral supplementation with Cit or water, to ewes consuming E + fescue seed, increases lamb birth weight and 2) determine the effectiveness of citrulline and citrulline:malate as an oral drench in elevating circulating levels of Cit to determine dosing levels and frequency of dose.

Methods
All animal and experimental protocols were approved prior to start by the Clemson University Institutional Animal Care and Use Committee (AUP2017-061 and AUP2018-050). All live animal research was conducted at the Clemson University Small Ruminant Facility.

Experiment 1
Experimental design. Suffolk ewes were obtained from the Clemson University Small Ruminant Facility's ock. Prior to breeding, ewes (n = 22, 86 ± 24 kg) were weighed and evaluated for body condition score (1-5; with 5 = heaviest condition; 1 = thinnest condition). Ewes were divided into 2 groups (n = 11/group) for estrus synchronization. Over a 2-w interval, each group was synchronized using an intravaginal water) prior to feeding each morning. Citrulline:malate 2:1 was chosen rather than pure Cit because it dissolved into a solution with a signi cantly smaller volume of water. TOX ewes (n = 5) were given a drench of ultrapure water of equal volume prior to feeding each morning.
Maternal Sample Collection. Blood samples were collected from ewes on d 85, 99, 113, 120, 127, 130, 134, and 140. Fasted blood samples were collected prior to morning drenching and feeding. Whole blood samples were collected by jugular venipuncture into collection tubes. Plasma samples were immediately placed on ice following collection and plasma was separated by centrifugation at 537 x g for 20 min at 4 °C. Plasma samples were stored at -20 °C for subsequent analysis. Whole blood was allowed to clot for 30 min at room temperature, stored at 4 °C for 24 h, then serum was separated by centrifugation at 537 x g for 20 min at 4 °C. Serum was stored at -20 °C for future analysis.
Doppler Ultrasound. Doppler ultrasound was used to examine blood ow in the carotid artery of all ewes on d 76 ± 4 of gestation, prior to treatment, and d 112 ± 4 of gestation, 27 ± 4 d after the start of treatment. Color Doppler ultrasound images of left carotid artery cross-sections were collected using a Classic Medical TeraVet 3,000 Ultrasound Unit (Classic Universal Ultrasound, Tequesta, FL) with a 12L5-VET (12 MHz) linear array transducer. Each imaging session was started at approximately 0900 h and was completed within 1 h. The ewes were accustomed to regular handling and being in their individual feeding pens prior to the start of the study to reduce excitability when being imaged. Ewes were placed in individual feeding pens for imaging, with each individual imagining session taking no more than 5 min.
Each ewe was clipped 1 d prior to scanning at the ultrasound site prior to each imaging session using surgical clippers. Scan depth was set at 4 cm. Cross-sectional images were collected for each artery using a frequency of 5.0 MHz and a pulse repetitive frequency that ranged between 2.5 and 3.0 kHz.
Following freezing of an individual scan, frames stored in the cine memory of the unit were searched to store the image exhibiting the maximum ow signal, assumed to be at peak systolic phase. The ow signal was traced to estimate lumen area (Aiken et al., 2009). Results are expressed as the difference between the pretreatment (d 76 ± 4) and post endophyte-infected tall fescue treatment (d 112 ± 4) measures.
Lamb Harvest and Sample Collection. Within 12 h of birth, birth weight and crown to rump length were measured for all lambs. Then the rst-born male lamb was harvested. The attending veterinarian euthanized the lamb with an intravenous over-dose of pentobarbital. Following euthanasia, the hide, head, feet, and tail were removed and the carcass was eviscerated and a weight of the carcass was obtained. From the viscera and carcass, weights of the fat masses (kidney fat, heart fat, and scapular fat) were collected. Samples of the brain, heart, kidney, and liver were taken and stored at -20°C for later fatty acid analysis. Scapular and heart fat were snap frozen in liquid nitrogen and stored at -80°C for RNA extraction. An aliquot of liver tissue was snap frozen in liquid nitrogen and stored at -80°C for Western Blot analysis. From the right side of each lamb, all fat and muscle were removed for total body proximate composition analysis and fatty acid analysis.
Proximate Composition. From the right side of each lamb, total muscle and fat samples were chopped and homogenized (Blixer ® 3 Series D, Robot Coupe Inc., Ridgeland, MS). Afterward, an aliquot was removed for moisture content analysis. The remaining sample was frozen at -20°C, lyophilized (VirTis, SP Scienti c, Warminster, PA), ground (Blixer ® 3), and stored at -20°C. Samples were analyzed in duplicate for nitrogen content by the combustion method utilizing a Leco FP-2000 nitrogen analyzer (Leco Corp., St. Joseph, MI). Nitrogen content was multiplied by 6.25 to determine the crude protein content of the sample. Moisture content was determined by weight loss after drying samples for 24 h at 100°C. Ash content was determined by ashing samples for 8 h at 600°C.
Fatty Acid Analysis. Fatty acid analysis was conducted on the brain, liver, kidney, heart, and total lean tissue collected from the right side of each lamb and follows the methods previously established by Duckett et al. (2014a). Tissue samples were freeze dried and transmethylated following the protocol of Park and Goins (1994). Fatty acid methyl esters (FAME) were examined using a gas chromatograph (Agilent 6850, Agilent, San Fernando, CA) in combination with an automatic sampler (Agilent 7673A, Hewlett-Packard, San Fernando, CA). Separations were achieved with a 120-m TR-FAME (Thermo Fisher, Greenville, SC) capillary column (0.25 mm i.d. and 0.20 µm lm thickness). Column oven temperature rose from 150 to 160°C at a rate of 1°C/min, then from 160 to 167°C at a rate of 0.2°C/min, and then from 167 to 225°C at a rate of 1.5°C/min. Once 225°C was reached, the column temperature was maintained for 16 min. Both the injector and ame-ionization detector were maintained at 250°C.
Samples were injected at a volume of 1 µL. Hydrogen was used as the carrier gas, with a ow rate of 1 mL/min. Each sample was run twice for analysis, rst with a split ratio of 100:1 for trans C18:1 and long chain-fatty acids. Then again, with a split ratio of 10:1 for omega-3 fatty acids and conjugated linoleic acid. Fatty acids were identi ed by evaluating the retention times with known standards (Sigma, St. Louis, MO; Supelco, Bellefonte, PA; Matreya, Pleasant Gap, PA) Measurement of fatty acids in each sample was accomplished by adding an internal standard, methyl tricosanoic (C23:0), during methylation, and expressed as a percentage of the total fatty acid weight. Total lipid content was analyzed using an Ankom XT-15 Extractor (Ankom Technology, Macedon, NY), with hexane as a solvent and samples were run in duplicate.
Blood Analyte Analysis. Serum prolactin levels were measured throughout late gestation by RIA using the procedures of Bernard et al. (1993). Average intra-assay variance was 4.71% and inter-assay variance was 5.75%. Plasma glucose levels were evaluated with colorimetric assay using a hexokinase reagent (Pointe Scienti c, Canton MI). Average intra-assay variance was 9.19% and inter-assay variance was 8.45%. Plasma insulin was measured using a Mercodia Ovine Insulin Kit (Mercodia, US, NC). Average intra-assay variance was 7.59% and inter-assay variance was 4.57%. Plasma glucose and insulin assays were validated for use in ovine samples previously (Duckett et al., 2014b;Long et al., 2014;Volpi-Lagreca and Duckett, 2017). Serum non-esteri ed fatty acids (NEFA) were quanti ed using a MaxDiscovery NEFA Assay (Bioo Scienti c, Austin, TX). Average intra assay variance was 7.37% and inter-assay variance was 10.48%. Serum NEFA enzyme-linked immunosorbent assay (ELISA) kits were validated by comparison to chloroform-methanol extraction and GLC analyses of non-esteri ed, free fatty acids. Insulin sensitivity, RQUICKI, was calculated according to Holtenius and Holtenius (2007), using glucose, insulin, and NEFA concentrations.
Plasma amino acid analysis was conducted at the USDA-ARS, Forage Production Research Unit, Lexington, KY. In brief, amino acid standards were acquired from VWR. Plasma samples (250 µL) were spiked with internal standard solution (20 µL of 100 µM/µL sample of -amino butyric acid) and proteins precipitated with the addition of cold methanol followed by centrifugation. The resulting solutions containing serum amino acids were derivatized with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate, AQC (Waters AccQ-Tag; Waters Corp, Milford, MA). Analyte concentrations were determined by liquid chromatography tandem mass spectrometry on a Waters Acquity tandem quadrapole detector (TQD) equipped with a Waters Acquity Sample Manager and Binary Solvent Manager. Chromatographic separation was obtained using a reverse-phase ultra-performance liquid chromatography (UPLC) column (Waters BEH C18, 1.7um, 2.1 mm x 150 mm). Column temperature was 40 °C. The mobile phase employed a mixture of water containing 0.1% formic acid (solvent A; Fisher) and acetonitrile (Fisher Optima) containing 0.1% formic acid (solvent B) in a linear gradient from 5% B to 75% B at a ow rate of 0.35 mL/min. The TQD tandem mass spectrometer was operated in positive ion multiple reaction monitoring (MRM) mode with electrospray ionization using nitrogen as the nebulizing gas and argon as the collision gas. Quantitation was done using QuanLynx software, linear calibration curve, internal standard method.
Real-time PCR. Total ribonucleic acid (RNA) was extracted from scapular and heart fat samples (100 mg of tissue) using Trizol reagent (Invitrogen, Thermo Fisher Scienti c, Waltham, MA) in accordance with the manufacture speci cations. Samples were cleaned of genomic deoxyribonucleic acid (DNA) using a DNA-free Kit (Ambion, Carlsbad, CA) according to the manufacturer speci cations. The RNA yield was measured using a NanoDrop 1 spectrophotometer (Thermo Scienti c, Thermo Fisher, Waltham, MA). Extracted RNA was stored at -80°C after isolation. Samples were checked for quality after DNase treatment using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). Two micrograms of RNA, along with qScript complementary DNA (cDNA) SuperMix (Quanta Bio, Beverly, MA), was used for cDNA synthesis according to the manufacturer instructions. Following conversion, cDNA was stored at -20°C.
Perfecta (Quanta Bio, Beverly, MA) SYBR green and an Eppendorf Realplex Mastercycler (Eppendorf AG, Hamburg, Germany) were used, according to manufacturer instructions, for quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) analysis of gene expression. The PCR program used was a preliminary hold at 95°C for 2 min immediately followed by 40 cycles of 95°C for 15 sec and 60°C for 30 sec. Primers were designed to span exon boundaries using Primer 3 software (Table 1). Uncoupling protein 1 (UPC1; Symonds, 2013) and myogenic factor 5 (MYF5; Seale et al., 2008) were used to con rm the presence of brown adipose tissue (BAT). Fatty acid binding protein 4 (FABP4) was used to assess the presence of differentiated adipocytes (Shan et al., 2013). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), -actin (ACTB), and cyclophilin B protein (CYC) were tested for stability using RefFinder (Xie et α β al., 2012) and the most stable housekeeping gene (CYC) was selected for normalization. Gene expression is displayed as the relative abundance from the control (TOX).  Kumar et al. (2017). Due to the ability to recognize several proteins, CYP3A subfamily was assessed rather than individual proteins (Acevedo et al., 2005;Damiri et al., 2012). In brief, 10% polyacrylamide gel was used to separate proteins by gel electrophoresis (SDS-PAGE) and transferred to a 0.45 µm nitrocellulose (Bio-Rad). Molecular weight markers were a pre-stained protein standard (Bio-Rad). Wild-type mouse liver, 20 µg, was used as a positive control for antibody reactivity. The blot was blocked using 1% skim milk/0.1% Tween 20 suspended in phosphate buffered saline. Equal loading of samples was accomplished using rabbit anti-mouse -actin (Sigma Aldrich, St. Louis, MO). The primary antibody was rabbit anti-rat CYP3A1 (Chemicon International, Temecula, CA) diluted 1:1000. Secondary antibody used for recognition of CYP3A subfamily was goat anti-rabbit IgG (Bio-Rad) alkaline-phosphatase diluted 1:500. A chemiluminescent kit (Bio-Rad) was used for visualization of the bands according to the manufacturer recommendations. Quanti cation of chemiluminescence was done using a Chemi-Doc system coupled with Image Lab software (Bio-Rad).
Statistical Analysis. The univariate procedure of SAS (SAS, version 9.4) was used to test all variables for normality. Serum prolactin concentrations were not normally distributed (Shapiro-Wilk test P < 0.01) and therefore, were log-transformed to preserve normality. Data were analyzed using the mixed procedure of SAS with citrulline treatment (CITM vs. TOX) in the model. Ewe(lamb) was the experimental unit for all analyses. For ewe plasma and serum analytes, repeated measures analyses with autoregressive covariance structure was used to evaluate treatment, day of gestation and the two-way interaction in the model and the Kenward-Roger method was used to generate degrees of freedom. Lamb number (single or twin) was included as a covariate when signi cant (P < 0.05). Least square means were generated and separated using a protected least signi cant difference test. Statistical signi cance was determined at P < 0.05 with trends at P < 0.10. β Experiment 2 Experimental Design. A 3 × 3 Latin Square design was used to determine the effects of oral citrulline and citrulline-malate 2:1 on circulating citrulline concentrations of non-pregnant ewes without the addition of E + fescue treatment. Suffolk ewes (n = 3, BW 42 ± 7 kg) were selected and administered either citrulline (CIT), citrulline:malate (CITM), or an equal volume of water (CON) as a drench immediately prior to feeding. Dose level administered to each CIT and CITM ewe consistent with Experiment 1. Ewes were fed once daily and maintained on the same TMR diet as Experiment 1 that was formulated to meet NRC requirements. The experimental period was 7 d, with 4 d of treatment drench and 3 d of washout. Blood samples were collected on the 4th d of treatment as a serial collection. Jugular cannulas were placed ~ 18 h prior to sampling and ushed every 4 h with sodium citrate until the start of sampling.
Blood Sampling. Whole blood was collected from preplaced cannulas with a syringe immediately prior to drenching (0 h) and at time 0.5, 1, 2, 3, 4, 6, 8, 10, 12, and 18 h post drenching. Whole blood was allowed to clot for 30 min at room temperature and then stored for 24 h at 4°C and then, serum was separated by centrifugation at 537 x g for 20 min at 4°C. Serum was stored at -20°C for future amino acid analysis. Serum amino acids were measured using the same procedures at Experiment 1. Plasma and serum Cit concentrations were evaluated from Experiment 1 and there was no difference (P = 0.74) between the levels of Cit in the plasma and serum samples.
Statistical Analysis. Data were analyzed as a 3 × 3 Latin Square using a repeated measure with PROC GLMMIX (SAS, version 9.4). Terms in the model included ewe, period, collection time, and treatment. The Kenward-Roger method was used to generate degrees of freedom. Area under the curve (AUC) was calculated using the trapezoid method (Le Floch et al., 1990) and analyzed using PROC GLM. The absorption and elimination coe cients (k a and k e ) for circulating citrulline were calculated using the serum citrulline concentrations according to Renwick (1994). The apparent half-life of elimination for citrulline was calculated using the formula t 1/2 = ln2/k e according to Moinard et al. (2008).
Ewe characteristics are presented in Table 4. Citrulline-malate treatment did not alter (P > 0.29) ewe body weight, average daily gain, dry matter intake (DMI), or gestation length (Data not shown). Carotid artery luminal area was measured on d 112 of gestation, 36 d after the E + fescue inclusion in the diet, and tended (P = 0.06) to be reduced when compared to pretreatment (d 76 of gestation) carotid artery luminal area regardless of Cit treatment (Data not shown). Citrulline-malate drench did not impact (P = 0.37) carotid artery luminal area difference between CITM and TOX ewes. Additionally, the number of lambs born, lamb birth weight, and average crown rump length per ewe did not differ (P > 0.50) by citrulline treatment. Table 4 Effect of dosing citrulline-malate (CITM) vs. water (TOX) to ewes exposed to endophyte-infected tall fescue on maternal characteristics in experiment 1.

Exp. 1: Lamb Characteristics
One male lamb (n = 10, 5 CITM, 5 TOX) was harvested from each ewe for body compositional analyses (Table 4). Lamb carcass weight did not differ (P > 0.20) by maternal citrulline-malate treatment. Total fat mass, and crude fat percentage were lower (P < 0.03) in lambs born to CITM ewes compared to TOX. Scapular fat (BAT) and kidney fat weight tended to be reduced (P < 0.07) in lambs from CITM ewes. When normalized to an empty body weight, kidney fat and total body fat percentages were lower (P < 0.05) for CITM compared to TOX lambs. Scapular fat mass also tended to be lower (P = 0.09) in CITM lambs on an empty body weight basis. Scapular and heart fat both expressed UCP1 and FABP4 but did not differ in relative abundance between treatments (P > 0.56) or tissues (P > 0.23, data not shown). In both tissues, MYF5 was present in low abundance. Additionally, Western blot analysis of lamb liver showed no difference (P = 0.17) for CYP3A presence in neonatal livers based on maternal citrulline-malate treatment (data not shown).
These results indicate that there are regional differences in the fatty acid composition of neonatal lamb tissues and that the brain contains the most n-3 fatty acids of the tissues examined. α fg Means in the same row differ (P < 0.10).

Exp. 2: Citrulline Concentrations
Oral drenching of citrulline and citrulline-malate 2:1 increased (P < 0.0001) citrulline levels during the 18 h sampling period (Fig. 1). Citrulline concentrations were elevated between 1 and 8 h after oral drenching of CIT when compared to water drench (CON). Drenching CITM elevated circulating Cit levels between 2 and 8 h after drenching when compared to water. Citrulline treatment, both CIT and CITM, response did not differ from each other at any time point. Both CIT and CITM drench took 2 h to increase circulating Cit concentrations above levels at 0 h, and then levels remained elevated until 6 h post drenching. The AUC for Cit was increased (P = 0.01) for both CIT and CITM when compared to water (CON; Table 5). The apparent k a and k e of Cit were estimated for each ewe and did not differ (P > 0.10) for CIT and CITM treatments. The average k a for CIT was 0.8577 µM/h and the average k e was 0.0817 µM/h. The average k a for CITM was 1.2787 µM/h and the average k e was 0.0667 µM/h. Apparent half-life of elimination for CIT and CITM were 8.484 h and 10.392 h respectively.

Discussion
The vasoconstrictive effects of ergot alkaloids are well documented in cattle (Foote et al., 2012;Aiken et al., 2013;Pesqueira et al., 2014) and sheep (Klotz et al., 2019). Experiment 1 ewes showed a reduction in carotid artery area when pre E + exposure scans was compared to scans taken after E + seed was incorporated into the diet. These results show that E + fescue did cause vasoconstriction, however there was no difference between CITM and TOX ewes. Reductions in maternal weight gain and average daily gain have been reported in steers (Matthews et al., 2005) and lactating ewes fed E + hay (Zbib et al., 2014) or gestating ewes fed E + seed (Britt et al., 2019). Ewes consuming E + fescue have been reported to have shorter gestation lengths when compared to ewes consuming nontoxic endophyte-free fescue (Duckett et al., 2014a;Britt et al., 2019). Additionally, exposure to ergot alkaloids from E + fescue causes a reduction in prolactin levels by interacting with the D2-dopamine receptor (Strickland et al., 2011;Klotz, 2015;Britt et al., 2019). In this study, all ewes were fed E + tall fescue seed and therefore differences in ewe measurements with E + exposure may not have been observed without a negative control, or ergot alkaloid-free, treatment.
Citrulline:malate treatment had no effect on circulating glucose and insulin concentrations during gestation. Glucose is a primary energy source for the placenta and the fetus during gestation (Hay, 2006) and can be a major limiting nutrient for fetal development (Battaglia and Meschia, 1988).
Citrulline:malate treatment tended to elevate NEFA concentrations for gestating ewes consuming E + fescue seed. During gestation, NEFA levels will increase because the dam is mobilizing fat tissue to keep up with both fetal growth and metabolic need (Luther et al., 2007). The RQUICKI values did not differ based on citrulline:malate treatment. As gestation progresses, especially near term, insulin secretion and sensitivity decrease and can potentially contribute to cases of pregnancy toxemia (Duehlmeier et al., 2013).
The majority of fetal growth (80%) occurs during the last trimester in sheep (Rattray et al., 1974), for this reason we administered the citrulline drench during the late gestation (d 86 -parturition) time period in attempt to mitigate E + fescue effects. Britt et al. (2019) reported feeding E + tall fescue seed during late gestation (d 85 to 133) reduced fetal weight per ewe by 15% compared to control. Additionally, fetuses from ewes consuming E + fescue have reduced muscle weights when compared to fetuses from ewes consuming E-fescue (Duckett et al., 2014a;Greene et al., 2019). Supplementation with Arg has been shown to increase the number of live-born offspring and litter weight in ewes, pigs, and rats (Lassala et al., 2010(Lassala et al., , 2011Wu et al., 2013). Lambs from CITM ewes did not show increased weight or size when compared to TOX lambs. Oral administration of citrulline:malate also reduced Arg and Cit concentrations in gestating ewes. These results are in contrast to previous reports with intravenous Cit administration where Cit was more effective at elevating circulating levels of Arg and Cit over 240 min period compared to Arg infusion in pregnant ewes (Lassala et al., 2009). No previous work has been reported on oral drench administration of Cit to gestating ewes, and samples from the current study were collected 24 h after a single dose administration.
In Experiment 2, Cit concentrations were increased in 2 h when an oral drench was used to administer treatment. Research by Gilbreath et al. (2020a), shows that rumen samples incubated with 5 mM of Cit showed no degradation of Cit after a 4 h period. In sheep, feeding 8 g of Cit results in a 117% increase in serum Cit concentrations over a 4 h period. In the second experiment, Cit levels increased by 149% in 2 hours when compared to water. The increase in circulating Cit was 89% greater than baseline (0 h) levels.
These results indicate that a dosing of citrulline:malate would need to be done every 6 hours at this dose level to continuously maintain elevated citrulline levels in the blood. Human studies with Cit dosages have shown that there is a positive linear relationship between the half-life and Cit dose level (Moinard et al., 2008). Subjects administered an oral dose of 10 g of Cit dissolved in water had an apparent half-life was determined to be 1.01 h. Moinard et al. (2008) reported that Cit was at maximum concentration > 1 h of oral administration and sharply declined thereafter. In the current study, maximum Cit concentrations were seen between 2 and 4 h for CIT and CITM treatments. The differences observed between the two studies could be related to differences in human and sheep digestive systems.
Lamb body composition was altered by maternal CITM administration. Fat masses for lambs from CITM ewes were reduced when compared to lambs from TOX ewes. Duckett et al. (2014a) found that lambs at birth from ewes consuming E + fescue during gestation had reduced fat masses when compared to lambs from ewes fed E-fescue. In contrast, Greene et al. (2019) did not report a difference in fat masses for fetuses on d 133 of gestation from ewes fed E + or E-fescue seed. In the current study, lambs were allowed to be carried to term and this may have attributed to differences in fat mass accumulation as fetal lipid deposition occurs more near term (Symonds et al., 2016). In cases of IUGR, fetuses may be born with decreased fat deposition due to reduced nutrient transport to the fetus, however human studies have found IUGR fetuses have a higher incidence of obesity post maturity (Engelbregt et al., 2004;Fernandez-Twinn and Ozanne, 2006).
Fatty acid analysis of lamb tissues revealed that maternal citrulline:malate treatment is capable of altering fatty acid deposition. Duckett et al. (2014a) found that EPA concentrations of the total lean muscle were elevated for lambs from dams on E + fescue when compared to E-lambs. Additionally, polyunsaturated n-6 fatty acids were found in larger concentrations for lambs from ewes consuming E + fescue when compared to lambs from E-fescue. In experiment 1, lambs from CITM ewes had increased EPA concentrations and n-6 fatty acids, eicosadienoic and dihomo-γ-linoleic acid compared to lambs from TOX ewes. The eicosanoid (C20) family of fatty acids can be elongated and desaturated into downstream fatty acids such as DHA and DPA (Gurr et al., 2016). Additionally, eicosanoids can be biologically transformed into prostaglandins, thromboxanes, leukotrienes, and lipoxins. Maternal amino acid treatment may have altered the conversion of essential fatty acids to downstream long-chain n-3 and n-6 polyunsaturated fatty acids and/or other essential signaling molecules, warranting further research.
Individual tissue analysis showed differences in the fatty acid concentration of various tissue examined. Brain tissue had a 352% more DHA than liver tissue and had 3889% more DHA than the total lean muscle tissue. Brain tissue is known to have a high concentration of DHA (Blank et al., 2002). The liver contained 318% more docosapentaenoic (C22:5 cis-4, 7, 10, 13, 16; DPA) than the total lean tissue, which is consistent with previous analysis of liver tissue compared to muscle in mature lambs (Malau-Aduli et al., 2016). -linolenic acid was only detected in liver samples in the current experiment. Malau-Aduli et al. (2016) showed that -linolenic acid is present in adipose, heart, and liver tissue in mature lambs but was not found in muscle or kidney tissue. Additionally, oleic acid was most abundant in the muscle tissue of lambs from the current study. Previous work has shown that oleic acid is found in abundance in adipose tissue with muscle tissue being the second largest amount in mature sheep (Malau-Aduli et al., 2016). The presence of BAT was examined by looking at the expression of biomarkers, UCP1 and MYF5 for lamb fat masses. Earlier reports have stated that lambs and calves are born with only BAT, which converts to white adipose tissue (WAT) within 3 weeks of birth (Thompson and Jenkinson, 1969;Gemmell et al., 1972;Casteilla et al., 1989). Lambs from CITM ewes showed no difference in the abundance of UCP1 or MYF5 when compared to lambs from TOX ewes. These ndings are in contrast to Satter eld et al. (2013), who reported that arginine administration to gestating ewes increased perirenal BAT accumulation. All fat masses analyzed showed the presence of FABP4 and this indicates that the tissue analyzed is adipose (Shan et al., 2013). Both WAT and BAT are capable of expressing FABP4. Additionally, previous work in our lab has shown that UCP1 and MYF5 are present in abundance in scapular fat at d 133 of gestation, about 2 weeks prior to parturition (Greene et al. 2019). Uncoupling protein 1 is used by BAT stores for thermogenesis and the loss or reduction of this ability may result in hyperthermia and death to the lamb (Clarke et al., 1996;Symonds, 2013). Brown adipose tissue originates from a myogenic line and expresses MYF5 as a cell marker (Seale et al., 2008). Expression of UCP1 in BAT peaks at or close to birth and drops signi cantly in the rst 7 d postpartum and BAT is replaced with WAT (Pope et al., 2012). Little research has been done to explain how BAT is replaced by WAT. Mueller (2014) reported on the discovery of beige fat, a "browned" WAT, an adipose type that produces UPC1 but is from the WAT cell lineage and not the myogenic lineage so does not produce MYF5. The ndings from the current study would indicate that the fat masses may be transitioning from BAT to WAT, due to the minimal expression of MYF5. Few studies have looked at the both UCP1 and MYF5 in fat masses from livestock animals, also few experiments have looked to describe the transition of BAT to WAT in adipose. Additional studies are needed to examine the cellular mechanisms and timeline for WAT replacement of BAT, and what, if any, role beige fat may play in the transition.

Conclusions
These experiments have demonstrated that the oral drenching of Cit and citrulline:malate can alter circulating Cit concentrations in nonpregnant ewes. Additionally, citrulline:malate supplementation to late gestation ewes consuming E + fescue and can alter fetal deposition of fat and fatty acid pro le; however, lamb birth weight was not affected when compared to lambs from TOX ewes when administered once daily. Citrulline:malate was more water soluble and determined to be equally effective at increasing circulating citrulline concentrations when compared to pure Cit. The dose level and frequency of Cit dosing may need to be more frequent, every 6 h, in order to observe consistent elevation of Cit concentrations over a 24 h period. Additional experiments are needed to determine if Cit, when maintained in an elevated steady state, is capable of mitigating fescue toxicosis in gestating ewes.

Declarations
Ethics: All animal and experimental protocols were approved prior to start by the Clemson University Institutional Animal Care and Use Committee (AUP2017-061 and AUP2018-050).