Supplementing branched-chain volatile fatty acids in dual-flow cultures varying in dietary forage and corn oil concentrations. III: Protein metabolism and incorporation into bacterial protein

Some cellulolytic bacteria cannot transport branched-chain AA (BCAA) and do not express complete synthesis pathways, thus depending on cross-feeding for branched-chain volatile fatty acid (BCVFA) precursors for membrane lipids or for reductive carboxylation to BCAA. Our objective was to assess BCVFA uptake for BCAA synthesis in continuous cultures administered high forage (HF) and low forage (LF) diets without or with corn oil (CO). We hypothesized that BCVFA would be used for BCAA synthesis more in the HF than in LF diets. To help overcome bacterial inhibition by polyunsaturated fatty acids in CO, BCVFA usage for bacterial BCAA synthesis was hypothesized to decrease when CO was added to HF diets. The study was an incomplete block design with 8 dual-flow fermenters used in 4 periods with 8 treatments (n = 4) arranged as a 2 × 2 × 2 factorial. The factors were: HF or LF (67 or 33% forage, 33:67 alfalfa: orchardgrass pellets), without or with supplemental CO (3% of dry matter), and without or with 2.15 mmol/d (5 mg/d 13 C) each of isovalerate, isobutyrate, and 2-methylbutyrate for one combined BCVFA treatment. The flow of bacterial BCAA increased by 10.7% by supplementing BCVFA and 9.14% with LF versus HF; similarly, dosing BCVFA versus without BCVFA increased BCAA by 1.98% in total bacterial AA, whereas LF increased BCAA by 1.92% versus HF. Additionally, BCVFA supplementation increased bacterial AA flow by 16.6% when supplemented in HF − CO and 12.4% in LF + CO diets, but not in the HF + CO (−1.5%) or LF − CO (+6.7%) diets (Diet × CO × BCVFA interaction). The recovery of 13 C in bacterial AA flow was 31% lower with LF than with HF. Of the total 13 C recovered in bacteria, 13.8, 17.3, and 30.2% were recovered in Val, Ile, and Leu, respectively; negligible 13 C was recovered in other AA. When fermenters were dosed with BCVFA, nonbacterial and total effluent flows of AA, particularly of alanine and proline, suggest decreased peptidolysis. Increased ruminal outflow of bacterial AA, especially BCAA, but also nonbacterial AA could potentially support posta-bsorptive responses from BCVFA supplementation to dairy cattle.


INTRODUCTION
Fermentation of branched-chain AA (BCAA; Val, Ile, Leu) begins with their deamination or transamination to their corresponding α-ketoacids (Dehority et al., 1958;Kaneda, 1977;Annous et al., 1997).The α-ketoacids are rapidly decarboxylated to the acyl-CoA derivatives by branched-chain α-keto acid dehydrogenase using NAD + as an electron carrier in nonrumen model bacteria, but this enzyme is poorly represented in genomes of sequenced ruminal bacteria (Roman-Garcia et al., 2021b).Instead, those authors described the likely role of ferredoxin as a cofactor for decarboxylation in another keto acid oxidoreductase complex awaiting characterization.Allison et al. (1984) reasoned that ferredoxin was the important cofactor for reductive carboxylation of branched-chain VFA (BCVFA), so bacteria such as the highly abundant Prevotella ruminicola could both decarboxylate α-ketoacids and reverse this process by reductively carboxylating BCVFA (isobutyrate, 2-methylbutyrate, and isovalerate).The BCVFA-CoA derivatives can be converted to BCVFA by acyl-CoA hydrolase or a CoA transferase, as shown for Megasphaera elsdenii (Schulman and Valentino, 1976), which would allow re-esterification of CoA to BCVFA.Both uptake of preformed BCAA and recycling between BCAA and BCVFA are important in mixed ruminal microbes (Atasoglu et al., 2004).Thus, amylolytic bacteria can degrade BCAA to provide BCVFA for cellulolytic bacteria requiring them, but amylolytics also can outcompete cellulolytics as starch increases and RDP becomes more limiting.Other amylolytic bacteria including Streptococcus bovis, Selenomonas ruminantium, and Ruminobacter amylophilus, and the hemicellulolytic/pectinolytic Butyrivibrio fibrisolvens do not produce BCVFA even when provided with BCAA (Allison, 1978;Stackebrandt and Hippe, 1986).Either these bacteria cannot transport BCAA, or those strains lacked 1 or more enzymes to catabolize BCAA.Stickland reactions convert BCAA to BCVFA using ferredoxin, particularly by clostridia (Neumann-Schaal et al., 2019).However, when Hino and Russell (1985) discussed this potential route, neither they nor authors citing their studies using ruminal isolates considered the critical role of ferredoxin.Moreover, subsequent work documented the role of the hyperammonia producers that are primarily clostridia (Bento et al., 2015).Because Leu is uniquely capable of serving as both oxidized and reduced AA in the Stickland reaction, including pairing with itself (Neumann-Schaal et al., 2019), it might be the BCAA most susceptible to the hyperammonia producers unless Leu is directly assimilated into protein.
Regardless of the process, the formation of BCVFA leads to cross-feeding by fibrolytic bacteria that consume the BCVFA to develop an efficient consortium in the rumen (Moraïs and Mizrahi, 2019); consequently, efficiency of bacterial growth should be decreased if BCVFA become insufficient in concentration.We therefore hypothesized that BCVFA supplementation would increase bacterial AA flow in both high forage diets (higher abundance of cellulolytics) and low forage diets (lower abundance of cellulolytics but more competition for BCVFA).When PUFA are added, the competition by amylolytics might be further exacerbated because there is less forage on which lipids can adsorb and be biohydrogenated by adherent bacteria (Jenkins et al., 2008); therefore, PUFA should be more toxic to the adherent cellulolytics in lower forage diets.
Our objective was to investigate the usage of BCVFA supplementation on bacterial AA metabolism under varying dietary conditions in dual-flow continuous cultures.We expected that supplemental BCVFA would increase bacterial AA outflow because BCAA, especially Leu and Ile, are particularly needed by cellulolytic bacteria (Bach et al., 2005).Studies cited therein dosed the BCAA and did not assess BCVFA, whereas cycling between BCVFA and corresponding BCAA also was considerable (Atasoglu et al., 2004).We hypothesized that increased forage: concentrate would increase the incorporation of supplemental BCVFA into bacterial AA by increasing the abundance of cellulolytic population that require BCVFA.Finally, supplementing PUFA in the diet was hypothesized to decrease BCVFA usage for bacterial AA synthesis in higher forage diets because PUFA was assumed to inhibit cellulolytic bacteria.

Experimental Design and Sample Collection
The experiment was a 2 × 2 × 2 factorial arrangement of 8 treatments of high (67:33) or low (33:67) forage: concentrate (100 g/d of DM split in 2 feedings), without or with 3% of DM as corn oil (CO), and without or with dosed BCVFA supplementation (2.15 mmol/d each of isobutyrate, isovalerate, and 2-methylbutyrate).Dietary conditions are described in Mitchell et al. (2023d).Two Jersey cows were housed according to Institutional Animal Care and Use Committee standards at the Waterman Dairy in Columbus, Ohio.Briefly, there were 8 anaerobic dual-flow continuous culture systems administered these 8 treatments in 4 periods in a randomized incomplete block (n = 4).Background samples were taken on d 5 to measure the natural enrichment of 13 C in the dual-flow cultures.Samples were also taken during the collection period, which was d 9 through 12.After background sampling (d 5) was completed, 13 C-labeled 2-methylbutyrate (which was racemic), isobutyrate, and isovalerate replaced a portion of the unlabeled BCVFA dose through d 12 to provide equivalent 5 mg/d 13 C doses from each BCVFA (Roman-Garcia et al., 2021c), and effluent samples were collected and partitioned into respective aliquots, including bacterial pellets, as described by Mitchell et al. (2023d).A 250-mL subsample of combined effluent from each 12-h sampling period was freeze-dried and then pooled over the collection period for AA analysis.

Amino Acid Analysis and Computations
Feed samples, lyophilized effluent, and lyophilized bacterial pellets were sent to Cumberland Valley Analytical Services (Waynesboro, PA) for AA quantification.The samples underwent separate hydrolysis procedures for Asp, Thr, Ser, Glu, Pro, Gly, Ala, Val, Ile, Leu, Tyr, Phe, His, Lys, Arg (AOAC International, 1997;method 994.12, Gehrke et al., 1987) and for Cys and Met, (AOAC International, 2006;method 982.30 E[a,b,c]), which included performic acid protection against sulfur AA destruction.Both hydrolysates were analyzed by ion exchange chromatography (AOAC International, 1997;method 994.12).Alkaline hydrolysis was performed with barium hydroxide and 5-methyltryptophan as the internal standard (AOAC International, 2006;method 988.15), and Trp was ana-  (Landry and Delhaye, 1992).Correction factors from Lapierre et al. (2019) adjusted for differential recoveries of AA with 24-h hydrolysis, which was necessary for unbiased 13 C recovery calculations and reported in all tables.Lyophilized effluent and bacteria pellets were also analyzed for N by combustion (AOAC International, 2006;method 990.03).The results of feed AA analysis are showing in Table 1.
The 13 C isotope ratios [i.e., 13 C/( 12 C + 13 C)] of bacterial AA were determined after acid hydrolysis followed by N-acetyl isopropyl derivatization (Styring et al., 2012) on a GC (Trace 1300; Thermo Fisher Scientific, Waltham, MA) equipped with a isotope ratio MS.For the separation of AA with GC, a DB-1301 column (Agilent, Santa Clara, CA; 60 m × 0.25 mm × 1.0 µm) and a 5-m deactivated guard column were used.A 2-µL sample was injected, splitless, at an inlet temperature of 225°C.The carrier gas was He at 1.2 mL/min.The initial oven temperature of 70°C was held for 2 min before a 15°C/min increase to 140°C, which was held for 4 min.The temperature was then increased at 12°C/ min to 240°C and held for 5 min.The final temperature ramp was 8°C/min to 255°C and held for 35 min.The eluate was then combusted through a combustion reactor (NiO and CuO at 1,000°C) and introduced into isotope ratio MS (Delta V Advantage; Thermo Fisher Scientific) for 13 C enrichment of compounds (m/z 44, 45, and 46).Isotope ratio MS was calibrated for 13 C enrichment using standards with known 13 C enrichment (IAEA-600; Vienna, Austria; n18M; Indiana University, Bloomington, IN; USGS71 and USGS72; Reston Stable Isotope Laboratory, Reston, VA).
Total N and bacterial N flows were measured as described in Mitchell et al. (2023d).Respective N flows were converted to AA flows based on the following equation: N (g/d) × % AA of sample/% N of sample.Nonbacterial AA was total AA flow − bacterial AA flow.The AA-N flows were calculated by multiplying AA flows by the % N/100 of each AA.The flows of Asp/Asn and Glu/Gln used the average % N to account for the loss of 1 N atom during hydrolysis of Asn to Asp and Gln to Glu.The N-acetyl isopropyl derivatization adds a propyl group (3 C) per carboxyl group and an acetyl group (2 C) per amino group in the AA structure (i.e., Leu has 6 C, but derivatized Leu has 11 C).Consequently, 13 C enrichment of AA (corrected for background 13 C) is diluted by the unlabeled C from derivatization.Therefore, the 13 C recovery in derivatized AA in bacteria was calculated as its atom percent excess (i.e., corrected for background) of 13 C divided by 100 and multiplied by C flow of the derivatized AA; the total 13 C recovered in each AA was then divided by total 13 C dosed.The total 13 C recovered in the bacteria C outflows was reported in Mitchell et al. (2023d).This recovered 13 C was partitioned into 3 fractions: the BCAA as reported herein, the total 13 C recovered as lipid fractions in Mitchell et al. (2023a) from which the total 13 C was reported herein, and the remainder.

Statistical Analysis
Bacterial AA, total AA, and nonbacterial AA flows and profile measurements were analyzed using PROC MIXED in SAS 9.4 (SAS Institute Inc.) according to the model described in Mitchell et al. (2023d).Briefly, a mixed model had random effects of period and fermenter and fixed effects of diet, CO, BCVFA, and all their interactions.Recovery of 13 C dose was analyzed with the same model but had an additional covariate of BCVFA flow because LF and HF differed in calculated RDP and therefore basal BCVFA production (Mitchell et al., 2023d).

Bacterial AA Flow and Profile
The total AA percentage of OM and total AA-N percentage of total N in bacteria both were greater (P < 0.01, Table 2) by 15.0 and 4.0%, respectively, with LF versus HF.However, supplementation of BCVFA tended to decrease total AA in OM in HF treatments but increase it in LF treatments (P = 0.09, diet × BCVFA).The total AA-N decreased by 2.25% of total N when BCVFA was supplemented to HF treatments but increased by 3.59% of total N with BCVFA in LF treatments (P = 0.02, diet × BCVFA).With CO supplementation, AA-N as a percent of total N tended to decrease in both diets, but the decrease tended to be more with LF than HF (P = 0.10, diet × CO).Efficiency of bacterial AA-N per units of truly degraded OM or truly degraded N both increased (P < 0.02; main effect) with BCVFA supplementation by 7.14 and 7.28%, respectively.
Total BCAA in the total AA was higher with LF versus HF due to an increase (P < 0.01) of Val and Leu, and a numerical increase in Ile (Table 2).The percentages of Cys, Pro, Gly, and Trp in total AA tended to be lower (P ≤ 0.07) with LF versus HF, whereas Lys was greater (P = 0.01) with LF versus HF.When BCVFA were supplemented to HF diets, Pro percentage of total AA did not change, but Pro decreased when BCVFA were supplemented to LF (P = 0.10, Diet × BCVFA).When BCVFA were supplemented, Gly and Ala tended to increase with HF but decrease with LF treatments (P < 0.07, diet × BCVFA).In HF treatments, BCVFA decreased Arg profile, but BCVFA increased Arg in LF treatments (P = 0.03, diet × BCVFA).Supplemental BCVFA increased Met and Cys without CO, but BCVFA decreased Met and Cys profiles with CO supplementation (P < 0.04, CO × BCVFA).Supplemental BCVFA increased (P = 0.02) total BCAA by 1.99% of total AA (data not shown).This change was a result of trends or numerical increases in Val (1.95%, P = 0.08), Ile (2.01%, P = 0.06), and Leu (1.99%, P = 0.12).

Isotope Recovery in Bacterial AA
Only bacterial BCAA were enriched significantly with 13 C in our study (Table 3).Recovery of dose was greater (P ≤ 0.02) by 45.8% in total BCAA, by 55.3% in Val, by 53.9% in Ile, and by 39.6% in Leu with HF diets compared with LF diets.These results explain why total 13 C recovery in bacteria was about 39% greater in HF than LF (14.5 and 10.4% of the 13 C dosed), as reported in Mitchell et al. (2023a).When calculated as a percentage of total recovered 13 C in bacterial C outflow in effluent, bacterial AA 13 C flow (65.2%), bacterial lipid 13 C flow (14.0%), or not recovered in those sinks (other, 20.6%) was not influenced by any main effects or interactions (Table 3).The distributions of recovered 13 C in individual fatty ands and aldehydes are reported in Mitchell et al. (2023a).

Total AA Flow and Profile
Supplementation of BCVFA increased (P ≤ 0.01) total effluent AA concentration from 12.6% of DM with BCVFA to 14.1% without BCVFA (SE = 0.8%, Supplemental Table S2, https: / / doi .org/ 10 .6084/m9 .figshare.22300801.v1,Mitchell et al., 2023b).The effluent flow of total AA was 12.6 g/d with HF diets, which tended to be greater (P = 0.07) than the 12.0 g/d total AA flow with LF diets (data not shown).Supplemental BCVFA increased (P < 0.01) total AA flow from 11.6 g/d without BCVFA to 13.1 g/d with BCVFA (data not shown).The flow of total BCAA also increased (P < 0.01) from 2.48 g/d without BCVFA to 2.80 with BCVFA (data not shown).Most of the flows of individual AA were greater (P ≤ 0.01) when BCVFA were supplemented (Supplemental Table S2).
The relative ratio of the total AA flows (sum of bacterial plus nonbacterial AA) in effluent with BCVFA/ without BCVFA are presented in Figure 1 for clarity.The greatest relative ratio was Ala (1.21) and Pro (1.17), whereas the remaining relative ratios ranged from 1.10 to 1.15.The relative ratio of Trp was only 1.05, and the flow of Trp did not increase with BCVFA supplementation (P = 0.11, Supplemental Table S2).The relative ratios of Cys, Met, and Arg are not presented in Figure 1 because there were interactions be- tween BCVFA and other main effects (Supplemental Table S2).When BCVFA was supplemented, flow of Cys tended to increase with HF diets but decrease with LF diets (P = 0.07, Diet × BCVFA).Both Met and Arg flows had tendencies (P ≤ 0.09) for CO × BCVFA interactions.Both Met and Arg flows increased when BCVFA was supplemented without CO, but Met flow decreased when BCVFA was supplemented with CO, whereas Arg flow still increased but to a lesser extent.
Additional CO in the LF diet decreased Met flow, but there was no difference in flow with HF.

Nonbacterial AA Flow
The nonbacterial flows of total AA, total BCAA, Thr, Ser, Pro, Gly, Ala, Val, Ile, Phe, His, and Lys all experienced trends for 3-way interactions (P < 0.10, diet × CO × BCVFA; Table 4).Supplementation of BCVFA increased the flows mainly with the HF + CO treatment and to a lesser degree with HF + CO and LF − CO diets, whereas BCVFA supplementation with LF + CO usually decreased undegraded AA flows.Supplementation of BCVFA tended to increase nonbacterial Cys and Arg flows with HF diets but not LF diets (P < 0.09, diet × BCVFA).The flows of nonbacterial Met, Asp, Leu, and Trp were lower (P < 0.03) with LF diets when compared with HF diets.Supplementation of BCVFA tended to increase (P = 0.09) nonbacterial Met flow and increased (P ≤ 0.05) flows of nonbacterial Glu, Leu, and Tyr.

BCVFA Usage by Bacteria
Much more isotope was recovered in bacterial BCAA than bacterial lipids reported in Mitchell et al. (2023a) because the bacteria were about 73.7% AA but only 9.98% FA and 0.57% fatty aldehyde on OM bases, respectively.Therefore, recovery of 13 C is expected to be much greater in bacterial protein, especially because BCAA were >20% of total AA in our study.Approximately 79.3% of the recovered dose in bacteria C flow was distributed in bacterial AA or lipids.The source of the 20.7% of recovered 13 C that was not recovered in lipids or AA is unknown.Allison and Bryant (1963) noted that nucleic acids accounted for approximately 2% of total radioactivity, compared with 17 to 25% in lipids and 70 to 79% in protein, of ruminal contents incubated with 14 C-labeled isovalerate and isobutyrate.However, the mechanism is not clear how nucleic acids would be labeled unless they were metabolized to central intermediates for synthesis of precursors such as ribose, Gln, and Asp.There could be alternate pathways of Leu (Monticello and Costilow, 1982;Atasoglu et al., 2004) and therefore some isovalerate interconversion to other products.Some Ile is synthesized from Thr (Kaiser and Heinrichs, 2018), which had a very high flux rate in vitro (Atasoglu et al., 2004).However, the lack of 13 C recovery in the major FA other than BCFA (Mitchell et al., 2023a) argues against 13 C-BCVFA metabolism to central metabolites that would mix with 6 Branched-chain lipids (BCL) is the sum of iso even, anteiso odd, and iso odd fatty acids and aldehydes reported by Mitchell et al. (2023a).  Other is the balance of 13 C recovered in bacterial 13 C outflow that is not represented by BCAA or BCL.
acetyl-CoA as a major route for 13 C other than that in AA and lipids.Because pantothenic acid is synthesized from the ketoacid of Val, some isobutyrate is likely converted to pantothenic acid (Kaiser and Heinrichs, 2018).However, this source should contribute a negligible proportion to the unrecovered 13 C pool.Therefore, the remaining 13 C most likely are intracellular BCVFA but could be other minor intermediates of lipid or AA synthesis or BCVFA catabolism that are not recovered by our analytical methods.Although flow of bacterial BCAA was greater with decreasing forage, the isotope recovery decreased by approximately 40% for Leu, whereas recovery in both Val and Ile decreased by more than 54%.The prokaryotic profile shifted from a fibrolytic population that requires BCVFA primers, as explained in Mitchell et al. (2023d), whereas increasing amylolytics, which are the predominant proteolytics, probably directly incorporated BCAA precursors from RDP, which averaged 63.7 and 66.5% of CP in HF and LF, respectively (Mitchell et al., 2023d).Recovery of the label in Leu followed by Ile supports previous expectations (Bach et al., 2005).However, 2-methylbutrate conversion to Ile might be underestimated in this study because it was dosed in a racemic mix, and our recovery in Ile therefore could be considered up to 32.6 to 57.2 µg/mg of total 13 C dosed from the predominant stereoisotopic precursor instead of 16.3 to 29.2 µg/mg (Table 3), as justified by Robinson and Allison (1969).Recovery in Val was lower than expected, especially when considering that it was the second highest BCAA in bacterial AA profile after Leu.When nonlabelled isobutyrate was supplemented with radiolabeled glucose in cultures of Bacteroides fragilis and P. ruminicola, which do not require BCVFA precursors because they can synthesize BCAA from pyruvate, the radioactivity of Val did not decrease; in contrast, when nonlabelled isovalerate or 2-methylbutyrate were provided, radioactivity of Leu or Ile decreased, indicating that reductive carboxylation of these 2 BCVFA replaced biosynthesis of Leu and Ile from pyruvate (Allison et al., 1984).Potentially, bacteria that do not require BCVFA precursors still benefit from isovalerate and 2-methylbutyrate supplementation and spare pyruvate or other central metabolites for synthesis of other cellular constituents, whereas isobutyrate was less beneficial for sparing Val biosynthesis.An inability to downregulate Val biosynthesis might explain why it was so readily excreted from mixed rumen bacterial cells compared with the other 2 BCAA (Stevenson, 1978).If Val was not secreted, it would presumably allosterically inhibit the synthesis of its keto acid, which also branches to produce pantothenic acid and Leu (Amorim Franco and Blanchard, 2017).However, for bacteria that require BCVFA, isobutyrate  typically can replace isovalerate and can increase the response to 2-methylbutyrate (Dehority et al., 1967).Dosing 2-methylbutyrate without other BCVFA might inhibit rumen bacteria growth because supplementing only Ile at their lowest concentration inhibited growth of mixed rumen bacteria, apparently by inhibition of common biosynthesis pathways (Kajikawa et al., 2002).Isotope recovery in both bacterial Ile in the present study and anteiso odd-chain lipids (Mitchell et al., 2023a) was high relative to the other 2 BCVFA recoveries.Additionally, as discussed by Roman-Garcia et al. (2021a), supplementation of an iso fatty acid precursor may be necessary for eliciting the ruminal benefits of BCVFA.Even though 13 C recovery in Val was the lowest, phospholipids with an iso even branched-chain fatty acid preferentially formed plasmalogens (Mitchell et al., 2023a), indicative that benefits of isobutyrate supplementation may be more specific to bacterial membrane structure rather than bacterial protein.Previously discussed by Mitchell et al. (2023c) and Roman-Garcia et al. (2021a), the benefit from isovalerate supplementation may be limited because of the high Leu content already available in commercial diets.There was trivial recovery of label in AA other than in BCAA in our study.When labeled BCVFA or BCAA were dosed, a similarly small amount of label was recovered in other AA in addition to the corresponding BCAA (Allison et al., 1962;Allison and Bryant, 1963;Robinson and Allison, 1969).There was also negligible isotope recovery outside of branched-chain lipids reported by Mitchell et al. (2023a).
Our results indicate that BCVFA are being utilized as precursors for their specific BCAA or branchedchain lipids.Mitchell et al. (2023d) reported that 8.2 and 19.7% of the total 13 C dose was not recovered in the total effluent in HF and LF diets.Catabolism of BCVFA to acetate or propionate would be expected to be greater in Proteobacteria (Kazakov et al., 2009), which increased from 5.8% in HF to 18.7% in LF (Mitchell et al., 2023d), and could increase 13 C lost through straight-chain VFA and CO 2 .Such changes for decreasing dietary forage are likely at the extremes of expectations in dairy cows; consequently, BCVFA not used by bacteria are either absorbed or flow out of the rumen.

AA Supply
Previously, we hypothesized that many of the benefits of BCVFA supplementation resulted from increased fiber degradation and microbial protein synthesis, but BCVFA in our study also specifically increased BCAA in bacterial AA profile and increased bacterial BCAA effluent flow.Increased MP supply resulting from increased postruminal AA flow would improve milk production, especially milk protein yield (e.g., Rius et al., 2010;Nichols et al., 2019).The increase in BCAA supply can also have a postruminal influence on milk protein synthesis (Kim, 2009;Appuhamy et al., 2012) and potentially milk fat synthesis (Bionaz et al., 2020) via mechanistic target of rapamycin signaling.
Supplementation of BCVFA increased bacterial protein synthesis beyond expectations from improved NDF degradation because efficiency of bacterial protein synthesis also increased (Mitchell et al., 2023d).If the greater outflow of BCAA in microbial protein is from BCVFA carboxylation to BCAA, then the BCAA would be more shielded from deaminating clostridia compared with dosing BCAA directly in the rumen.The greater intracellular BCAA in bacteria (Table 2) suggests that BCVFA stimulated bacterial protein synthesis by increasing concentration-dependent entry into cells and reductive carboxylation to BCAA.Such responses seem to be simultaneously decreasing peptidolysis and increasing efficiency of growth as mediated by transcriptional regulation, which will be discussed subsequently.

Decreased Proteolysis or Peptidolysis
In the BCVFA treatments, the decreased net production (i.e., daily dose subtracted from actual production per day) of the individual BCVFA (Mitchell et al., 2023d) coincides with increased nonbacterial AA outflow for supplementation of BCVFA (Table 4).The lack of difference in nonammonia-nonbacterial N observed in (Mitchell et al., 2023d) suggests greater AA:N ratios of the undegraded AA compared with degraded AA.Previously, supplementation of 2-methylbutyrate increased CP effective degradability in situ (Wang et al., 2018), whereas supplementation of isobutyrate (Wang et al., 2015) and isovalerate (Liu et al., 2009) decreased CP effective degradability in situ.All responses were linear, but only isovalerate lacked significant quadratic or cubic responses reflecting waning responses.In their studies, effective degradability responses were similar for ruminal protease activity, which used trichloroacetic acid and therefore likely precipitated not just protein but also larger peptides.Accumulation of peptides ranging between 3 and 10 kDa in omasal contents supports peptidolysis being the rate-limiting step for proteolysis in the rumen (Reynal et al., 2007).
The increased flows of total Ala and Pro in total effluent (but not for bacterial flow) for BCVFA supplementation (Figure 1) point toward decreased peptidolysis and therefore increased outflow of peptides because Ala and Pro are preferred hydrolysis sites for several of the characterized dipeptidyl peptidases (DPP), par-ticularly DPP-4 from genus Prevotella (Walker et al., 2005a).Genomics approaches have supported a major proteolytic and peptidolytic niche for genus Prevotella in vivo (Hart et al., 2018;Hartinger et al., 2018).As those authors explained, protozoa also contribute significantly to proteolysis, but they were absent in our system.
In general, increased proteolysis has been positively associated with increased microbial biomass or bacterial growth rate.For example, Griswold et al. (2003) noted that replenishing urea in buffer stimulated microbial N flow and efficiency of microbial protein synthesis in continuous culture, thus decreasing measured peptide concentration presumably because of assimilation; urea also decreased nonammonia-nonbacterial N, suggesting increased proteolysis resulting from the increased bacterial growth and therefore proteolytic capacity.Because an increase in bacterial protein flow in our study should be associated with increased proteolysis, feedback of peptidolysis seems the more likely explanation for our increased nonbacterial AA flow based on results from various batch and continuous culture systems (Bach et al., 2005).Increasing protein decreased proteolysis or peptidolysis in most pure cultures of proteolytic bacteria tested (Griswold et al., 1999;Sales et al., 2000).
Based on the importance of the prevotellas and their high (15-18%) abundance in our study (Mitchell et al., 2023d), we investigated their genomes for peptidase function.All of the prevotellas cataloged in the Hungate 1000 collection (Seshadri et al., 2018) have 2 copies of DPP-4 identified as K01278 in KEGG (Kanehisa et al., 2012).All have one or more (some as many as 4) copies of the Lrp/AsnC transcriptional regulator (K03718 and K03719) family that respond to Leu or Asn binding and have important regulation in Escherichia coli (Ziegler and Freddolino, 2021).One of the Lrp/AsnC regulators (K03718) is consistently flanked by a Zn-dependent dipeptidase (COG2355; Galperin et al., 2015).This enzyme appears to be the dipeptidase characterized in Prevotella albensis with activity against substrates including Ala peptides (Walker et al., 2005b).When intracellular Leu concentration increases in Gramnegative bacteria, Lrp inhibits AA biosynthesis and AA catabolism for central metabolites to be prioritized toward growth functions (Kaiser and Heinrichs, 2018).
Regulation of transcription by BCAA, which increased in bacteria when BCVFA were dosed, has not been well studied in rumen bacterial cultures.Kajikawa et al. (2005) suggested that Ile feedback-inhibited the synthesis of the other 2 BCAA, but they gave little consideration to transcriptional regulators, probably because little information was known at that time.Kim et al. (2017) reported that P. ruminicola repressed ammonium transporter expression when provided with peptides, noting that genus Prevotella transports small peptides rather than free AA, but they did not address BCAA per se.Isovalerate supplementation to Prevotella bryantii resulted in many transcriptional changes, including increased expression of BCAA synthesis enzymes (Trautmann et al., 2020).Because their measurements were several hours after the last transfer, isovalerate might have rescued growth responses in that study.Therefore, without more specific explanations, we hypothesize that BCVFA increased NDF degradability and efficiency of microbial protein synthesis, but an increased concentration of intracellular BCAA, particularly Leu arising from isovalerate, likely decreased peptidolysis to prioritize expression of genes related to growth rather than to using AA for maintenance or fuel.The importance of BCAA sensing in rumen bacterial growth and AA metabolism warrants future experimental confirmation.

CONCLUSIONS
Increasing forage increased the incorporation of BCVFA into bacterial protein, but there was no appreciable label recovered outside of BCAA in this study.Previous research with BCVFA supplementation indicated that many of the benefits were due to increased NDF degradation and increased bacterial N flow.However, this study indicates that increased AA flow and BCAA flow is an additional benefit that would increase metabolizable AA and result in greater milk production or production efficiency in ruminants.

Figure 1 .
Figure 1.Relative ratio of total effluent (i.e., bacterial plus nonbacterial) AA flows (mg/d) of continuous cultures with branched-chain VFA (BCVFA) supplementation/total effluent flow of AA from continuous cultures without BCVFA supplementation.The main effect comparison of BCVFA on AA flows is reported as *P ≤ 0.01.
Mitchell et al.: BRANCHED-CHAIN VOLATILE FATTY ACIDS AND BACTERIAL PROTEIN METABOLISM lyzed with chromatography with fluorescence detection

Table 1 .
Mitchell et al.: BRANCHED-CHAIN VOLATILE FATTY ACIDS AND BACTERIAL PROTEIN METABOLISM Amino acid composition of the diets varying in forage and corn oil (CO) supplementation 1Treatments are high forage (HF, 67% forage) or low forage (LF, 33% forage) and − CO (no additional supplemented fat) or + CO (3% DM supplemented fat as corn oil).

Table 3 .
Mitchell et al.:BRANCHED-CHAIN VOLATILE FATTY ACIDS AND BACTERIAL PROTEIN METABOLISM Dose recovery in bacterial AA in continuous cultures that were administered with either high-or low-forage diets that varied in corn oil (CO) supplementation P-values for main effect means for diet (HF vs. LF), CO, and interactions.The model also included a covariate (BCVFA mmol/d production) for the differences between diet for RDP and feed AA profile.3The 13 C dose was provided as branched-chain VFA (5 mg/d 13 C from each isobutyrate, 2-methylbutyrate, and isovalerate).
2 4 Branched-chain amino acids (BCAA) is the sum of Val, Ile, and Leu. 5 Proportion of recovered dose was calculated by 100 × (mg 13 C recovered in each fraction/total mg 13 C recovered in bacterial C outflow).

Table 4 .
Flow of nonbacterial AA in continuous cultures that were administered with either high-or low-forage diets that varied in corn oil (CO) and branched-chain VFA supplementation