3. Results
Organic matter truly fermented was higher in HSBM vs. CGM and decreased linearly from 58.0 to 51.0, and from 55.1 to 48.5 % in HSBM and CGM, respectively (
Table 3). Degradation of aNDFom was also higher in HSBM vs. CGM (34.0 vs. 29.0 %), but was not affected by the level of inclusion (
Table 3). Concentrations of total VFA (average of 95.0 mM) and the molar proportions (mol/100 mol) of acetate (average of 59.3), propionate (average of 21.8) and butyrate (average of 10.5) were not affected by treatments (
Table 3). Branched-chain VFAs (BCVFA) were affected by a protein source by the level of inclusion interaction, where the linear reduction was faster in HSBM (ranging from 5.0 to 3.13 mM) than in CGM (ranging from 4.9 to 4.15 mM).
The effect of treatments on N metabolism in fermenters is summarized in
Table 4. Protein supplements had no effect in any of the measurements except for the efficiency of microbial protein synthesis. As the inclusion level of HSBM and CGM increased, ammonia N concentration showed a quadratic effect, but the coefficient for the quadratic term, although significant, was very small and the overall effect was almost linear (ammonia-N, mg/dl = 46.2 − 0.445x + 0.0012 x
2). The flow of ammonia N followed a linear reduction as the level of protein supplements increased (
Table 4). Nonammonia N flow showed a quadratic effect characterized by a saturation shape as the level of inclusion of proteins increased, and was due to the quadratic effect of dietary N flow. Changes in ammonia N concentrations and flows of ammonia and non-ammonia N (NAN) reflected the ruminal degradation of dietary proteins. Accordingly, the degradation of dietary protein decreased quadratically as the level of HSBM or CGM increased (
Table 4). Microbial N flow showed a protein supplement by level of inclusion interaction, where it increased linearly in HSBM (from 1.03 to 1.27 g/d) but had no effect in CGM (average of 1.06 g/d;
Table 4). The efficiency of microbial protein synthesis (g of N/kg of organic matter (OM) truly digested) of a protein supplement was obtained by measuring the level of inclusion interaction, where it increased linearly in HSBM (from 24.1 to 35.2) but had no effect in CGM (average of 28.9;
Table 4). These differences were due to the greater efficiency of microbial protein synthesis in HSBM-67 and HSBM-100 compared with HSBM-0 and HSBM-33, resulting from the combination of an increase in bacterial N flow and the lower OM digestion observed in these treatments.
The AA compositions of the two supplemental protein sources used in this study are presented in
Table 2. Flows of total, essential and non-essential AA, Glu, Ser, Phe and Tyr were higher, and Lys was lower in CGM diets, and all increased linearly in both protein supplements as the level of supplemental protein increased (
Table 5). Significant protein supplementation by level of inclusion interactions were observed only for the flows of Asp, Ala, Arg, Pro and Leu. These changes reflect CP degradation, the AA profile of each supplement, and the degradation rates of individual AAs. Models for the degradation of individual AAs were developed using the relationship between the amount of each AA supplemented in HSBM and CGM, and its flow, and are presented in
Table 6. All models were linear, where the coefficient of the linear term reflects the degradability of each individual AA. Values less than 1 indicate that the AA was degraded more extensively than the total AA fraction. Values greater than 1 indicate that the AA was degraded less extensively than the total AA supplied by the supplement. The degradation of essential AA (EAA) was higher, and that of nonessential AA (NEAA) was lower in CGM. Similar trends were observed in HSBM, but differences were not significant. Degradation of Ile, Lys and Met were higher, and the degradation of Ala, Asp Glu, Gly, Pro and Tyr were lower than the average AA in CGM. Differences in the degradabilities of individual AAs in HSBM were lower, being significant only for Lys, which was more degraded, and Asp, Pro and Tyr, which were less degradable than the average AA in HSBM.
4. Discussion
Almost all research conducted to date investigating the rumen-degradability of individual AAs within feeds used the in situ technique [
9,
10,
26], where most estimates were obtained after a single-point incubation, microbial colonization was not corrected in residual feed after incubation and the dynamic effects of the rumen were not considered, which may bias the estimates [
12]. The slope technique is an alternative method to determine individual AA degradation. It is a robust experimental design that prevents some of the problems of the in situ technique. Diets were designed to achieve a similar flow of AA from the basal diet plus microbes, as previously suggested [
13,
14]. The hypothesis was that changes to the flow of AA within protein source out of fermenters could be attributed specifically to the increasing supply of AA from each level of protein supplement.
The linear reduction in OM truly fermented as the level of inclusion of protein supplements increased was expected due to the lower degradation of the protein fractions of HSBM and CGM compared with the extensive degradation of urea and tryptone in HSBM-0 and CGM-0 diets. Other authors also reported a decrease in OM degradation when rumen-protected protein sources were used instead of highly degradable protein supplements [
14,
27,
28]. This linear reduction in OM degradation was parallel to the linear reduction in total VFA, although the differences did not reach significance. The lack of effect of the treatments on fiber degradation and the proportions of major individual VFAs indicated that the overall fermentation was similar among treatments, and that energy and protein availability did not limit microbial activity. The BCVFA (valerate, 2-methylbutyrate and isobutyrate) result from the deamination of the branched-chain AA (Leu, Ile and Val) and the changes observed reflect the content and degradation of these AAs in HSBM and CGM [
28,
29,
30]. Within each supplement, the reduction in the concentration of BCVFA was also paralleled by a reduction in ammonia N concentration, and reflects the fact that BCVFA and ammonia N derive from the deamination of branched-chain AAs.
The effects of treatments on N metabolism in the rumen are shown in
Table 4. Ammonia N concentration showed a significant quadratic effect, although the numerical impact of the coefficient of the quadratic term was very small, and the practical effect was almost linear (ammonia-N, mg/dl = 46.2 − 0.445x + 0.0012 x
2). The reduction in ammonia N concentration was expected, and in all cases was well above the 5 mg/dl suggested to maximize microbial growth, as expected [
31]. In fact, the diets were designed so that N available for microbial growth would not be limiting. Similarly, other authors observed a reduction in ammonia N concentration when diets contained a rumen-protected protein source compared with the use of rumen-degradable protein in vivo [
32] and in vitro [
14,
15,
33]. The increase in NAN flow as the level of inclusion of HSBM and CGM increased was associated with a greater dietary N flow (
Table 4). Clark et al. [
34], using results of eight trials in which different sources of supplemental CP were fed, concluded that protein supplements with low ruminal degradability fed at high concentrations in the diet increased the passage of NAN to the small intestine compared with feeding soyabean meal (SBM), because of increased passage of dietary N. Changes in ammonia N concentration and flows of ammonia and NAN reflected ruminal degradation of dietary proteins, and agreed with previous in vivo and in vitro reports [
13,
15,
29]. However, Clark et al. [
34] reported that the increase in non-ammonia N flow was lower than expected based on the increase in the flow of dietary N, and was due to the parallel reduction in microbial N flow. In the present study, microbial N flow did not decrease as the level of supplemental protein increased. This was likely the result of the experimental design, where diets were formulated to provide, even at the highest inclusion level of RUP, sufficient degradable protein to maximize microbial protein synthesis. However, there was a significant protein source–level of protein inclusion interaction, where microbial N flow increased with an increasing level of inclusion in HSBM, but not in CGM. Because microbial N flow in CGM diets was constant and closer to the lower inclusion rates of the HSBM diets, the data suggest that microbial N synthesis was stimulated in the higher inclusion levels of HSBM. The reason is not clear, because ammonia N was sufficient to guarantee microbial growth and the amount of OM fermented in the rumen decreased in the diets with the high proportions of protein supplements. Titgemeyer et al. [
13] suggested that if a significant percentage of dietary purines escaped ruminal degradation, bacterial N flows to the duodenum would be overestimated. However, McAllan and Smith [
35] demonstrated that pure nucleic acids are rapidly degraded in the rumen. Moreover, Calsamiglia et al. [
36] reported, using N
15 as a marker, that dietary purines from HSBM and CGM were almost completely degraded by ruminal microbes in continuous culture, regardless of the total amount of purines in diets, and the escape of feed purine N seemed to be a minor factor affecting calculations of microbial nitrogen flow. This higher microbial N flow may compromise the underlying assumptions required for the calculations of the degradation of individual AAs from HSBM, which tend to be overestimated.
The degradation of dietary protein was not different between the HSBM and CGM diets, although it was numerically higher in the HSBM diets (41.2 vs. 37.3 %). Other authors found a trend for diets containing predominantly CGM to have lower protein degradation than diets containing treated SBM [
2,
15,
30]. In fact, the NRC [
2] recognizes a higher RUP level in CGM compared with HSBM. Within each protein supplement diet, there was a decrease in dietary protein degradation as the level of HSBM or CGM increased, according to the changes observed in the ammonia N concentration and the flows of ammonia and NAN.
The efficiency of microbial protein synthesis was within the ranges reported by Stern and Hoover [
26]. There was a significant protein source–level of inclusion interaction, where it increased quadratically in HSBM, but remained constant in CGM. The greater efficiency of microbial protein synthesis of HSBM-67 and HSBM-100 resulted from the combination of an increase in bacterial N flow and the numerically lower OM digestion observed in these treatments. Coomer et al. [
37] and Keery et al. [
38] reported an increase in the efficiency of bacterial protein synthesis in steers fed diets supplemented with RUP compared with steers fed diets supplemented with untreated SBM. Cecava et al. [
29] attributed changes in the efficiency of bacterial protein synthesis to differences in ammonia-N, AA and peptide availability for microbes. However, in the present experiment, the basal mix contained urea as a source of ammonia N, tryptone as a source of readily available peptides and AA, and the diets were fed semi-continuously every 10 min, providing N, AA and peptides on a constant basis throughout the day.
The increase in AA flows as the level of HSBM and CGM increased (
Table 5) agrees with other studies, which reported that feeding low degradable protein supplements resulted in an increase in total AA flow [
14,
15,
30]. The addition of CGM resulted in greater increases to flows of essential (EAA) and nonessential AAs (NEAA) compared with HSBM. Blake and Stern [
30], in a continuous culture study, also reported higher EAA and NEAA flows with diets containing CGM than with diets supplemented with extruded whole soyabeans. Santos et al. [
39] observed an increase in dietary AA flow when CGM was used as a source of supplemental protein, reflecting its higher degree of rumen undegradability [
2]. Flows of Glu, Ser, Phe and Tyr were higher and those of Lys were lower for diets containing CGM compared with HSBM diets. However, there were no differences in Met flows between treatments. Blake and Stern [
30] reported similar differences when comparing diets containing CGM or SBM. Similarly, Calsamiglia et al. [
15] found that fermenters fed diets containing lignosulfonate-treated SBM had higher flows of Lys than fermenters receiving CGM-supplemented diets, but the differences in Met flows between both treatments were not significant. These results suggest that, although CGM provided large amounts of total AA, some potential limitations (low Lys) should be considered, and feeding combinations of protein supplements could improve the AA profile reaching the duodenum [
13,
15,
29]. These results also suggest, that in spite of the higher content of Met in CGM diets, its flow was similar to HSBM, probably due to the extensive degradation of Met in CGM as compared with HSBM.
All equations for the proportional flow of individual AAs from fermenters were linear. The coefficient of the linear term represents the ruminal escape of each AA from HSBM and CGM (
Table 6). Overall, NEAA were less extensively degraded than EAA in CGM. Stern et al. [
40] reported, in an in vivo study with cannulated cows, that the six most degradable AA in CGM was EAA. Particularly relevant was the higher rumen degradability of Lys in both protein supplements. The rumen degradability of Met was affected by an interaction (
p < 0.06) with the protein supplement, where rumen degradability was higher in CGM, but not affected in HSBM. Isoleucine was also more degraded in CGM, but was not affected in HSBM. This result agrees with previous reports conducted in vivo with cannulated cows [
40,
41], in situ [
7,
42], and in vitro [
15,
30], which observed that Lys was one of the most degraded AAs among the EAAs. Conversely, Titgemeyer et al. [
13] reported that the relative ruminal escape of Lys in CGM was lower than the total AA pool. Titgemeyer et al. [
13] suggested that these differences may be explained by an increased Maillard product formation during processing that protected Lys from ruminal degradation. Some reports have shown Met to have higher degradability than total AA in several feeds [
6,
13,
30]. However, data from other authors [
7,
43] have shown the degradation of Met to be dependent on the feedstuff. The degradation of branched-chain AAs was higher than total AA degradation in CGM, although some authors found these AAs to be more resistant to degradation in the rumen, depending on the protein source [
7,
12,
13]. Alanine, Asp, Glu, Gly, Pro and Tyr were more degradable than the average total AA in CGM, and only Asp, Pro and Tyr in HSBM. Chalupa [
44] found, in an in vitro system, that Tyr was degraded to a lesser extent when the fermentation system contained NEAA and EAA mixtures, and suggested that Tyr could be a degradative intermediate of Phe, therefore increasing the pool size of Tyr. Crooker et al. [
43] reported that the proportion of Tyr increased or tended to increase as a result of ruminal exposure in 5 out of the 7 feeds tested in situ. Titgemeyer et al. [
13] indicated that the rumen degradability of Tyr appears to be protein source-dependent, because it was degraded less than total AA in SBM and fish meal, but slightly more than total AA in CGM and blood meal. The lower ruminal degradation of Asp after the in situ ruminal exposure of several supplements was also reported previously [
6,
45]. Gonzalez et al. [
12] suggested that hydrophobic non-polar AAs were less degradable in the rumen compared with hydrophilic polar AAs. Other factors, such as the solubility of the protein itself and the location of the protein within the structure of the feed or the protein, may also affect the rumen-degradability of individual AAs.
If precision feeding of AAs is to be implemented in current feeding systems, differences within and among feeds of individual AA degradation in general, and EAA in particular, need to be considered. Rulquin and Vérité [
46] stated that the modifications produced by rumen fermentation on dietary AA profiles could vary broadly according to feedstuffs and the level and degradability of the protein. However, because of methodological limitations and the small number of available data, these authors suggested using feed AA profiles as a first guide to estimate undegraded protein AA profiles. Currently, the recent version of INRA [
3], the CNCPS system [
5] and the NRC [
2] used the AA profile of the original dietary proteins instead of that of the insoluble fraction.
The results from the present experiment should be interpreted with caution, particularly in the HSBM treatment. The increasing flow of bacterial N in HSBM compromises the underlying assumption required to test the hypothesis and may overestimate the degradation rates of individual AA from HSBM. However, calculations of differences in the degradation of individual AAs within each protein supplement were performed relative to the degradation of total AAs, and this would reduce some of the bias. In contrast, the similar flow of microbial N in CGM validates the use of the slope methodology to calculate the differential degradability of AAs.
This paper was designed as a robust alternative approach to evaluate the differential degradability of individual AAs within and between protein supplements. The results suggest that this occurs and affects particularly to EAA, including Lys and Met. If this is confirmed, flows of these AA would be overestimated in current feeding systems, and this effect may contribute to a limitation. The slope methodology is a robust design and the results were, in spite of the limitations of the methodology, consistent. Additional studies specifically designed to determine the differential degradability of individual AAs within and between proteins sources are necessary to advance theprecision feeding of AAs in dairy cattle diets because the magnitude of the differences is important.