INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The extent to which ammonia is used for protein synthesis has important implications for the efficiency of ruminal fermentation (4, 23, 29, 32). Early nutritional studies with pure cultures of cellulolytic bacteria indicated that N incorporation into bacterial protein was equivalent to the disappearance of ammonia N from the medium, and it was therefore concluded that ammonia is the main source of N for growth of these species (9). Low incorporation levels of ¹⁴C-protein hydrolysate into Ruminococcus flavefaciens in the presence of ammonia (1) tended to support this view. A similar conclusion was made by Hobson et al. (22) with the noncellulolytic species Ruminobacter (then Bacteroides) amylophilus, which formed >90% of its protein N from the (¹⁵NH₄)SO₄ present in the medium even if protein hydrolysate was present. The survey by Bryant and Robinson (10) of a larger number of rumen bacterial species showed that the pattern was much more variable: many cultures incorporated large amounts of ¹⁴C-algal protein hydrolysate, while others did not. Russell et al. (32, 33) concluded that mixed bacteria fermenting soluble carbohydrate derived 66% of their N from peptides and 34% from ammonia when both were available.

In the mixed rumen microbial population, the extent to which microbial protein is synthesized de novo varies enormously (2, 27, 29-31, 34), depending on the relative amounts of energy and N available for microbial growth (6, 34). Some amino acids are synthesized de novo to a much greater extent than others (34). The aims of the present study were to provide information on which amino acids were formed de novo by pure cultures of different noncellulolytic bacterial species from the rumen and how the availability of different nutrients affected de novo protein synthesis by these species. It emerges that the average rate of de novo synthesis of microbial protein conceals a wide variation in de novo synthesis between individual amino acids and individual species and that the extent of incorporation of amino acids from peptides and amino acids in the medium is highly concentration dependent.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and growth conditions. The bacteria used in this study were Prevotella bryantii B₁4 (11) (a gift from J. B. Russell), Selenomonas ruminantium HD4 (ATCC 35018), and Streptococcus bovis ES1 (a strain prototrophic for amino acids isolated from a sheep at the Rowett Research Institute). The cultures were maintained on medium M2 (21).

¹⁵N and N analyses. ¹⁵N enrichment was measured by isotope ratio mass spectrometry as described by Barrie and Workman (5). Total cell N was measured by a Kjeldahl procedure (19). Freeze-dried pellets were prepared for analysis of ¹⁵N enrichment in amino acids by hydrolysis in 6 M HCl at 110°C overnight. Samples were evaporated at 70°C and then resuspended in 2 ml of 0.1 N HCl. The samples were centrifuged again, and 1 ml of supernatant was applied to a Bio-Rad AG 50W-X8 column (0.8 ml). The column was washed twice with 2 ml of water, and the amino acids were eluted with 2 ml of 2 M ammonia followed by 1 ml of water. The eluted liquid was freeze-dried and stored at 20°C. Tertiary butyl dimethysilyl derivatives of the amino acids were prepared, and the ¹⁵N enrichment in individual amino acids was determined by gas chromatography-mass spectrometry as described by Calder and Smith (13). Ammonia was measured by an automated phenol-hypochlorite method adapted from the work of Whitehead et al. (39). ¹⁵N enrichment in ammonia was determined by incorporating ¹⁵NH₃ into norvaline by using glutamate dehydrogenase and 2-oxopentanoate as described by Nieto et al. (28). Amino acids in spent medium were analyzed by ion-exchange chromatography following HCl hydrolysis (37).

Statistical analysis. Results are all means derived from the analysis of triplicate cultures. The data were compared by analysis of variance, with different cultures used as a blocking factor. To compare the effects of treatments on ammonia uptake into amino acids, individual amino acids were considered as a subplot within the design. All analysis was carried out by using the GENSTAT 5 statistical program (Lawes Agricultural Trust, Rothamsted Experimental Station, Harpenden, Hertfordshire, England).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

All species grew well in the general-purpose rumen fluid-containing medium M2 (see Tables 1, 3, and 5). Little net production or utilization of ammonia production occurred. The ¹⁵N enrichment in ammonia in the spent medium decreased for all species, however, indicating that some of the ammonia in spent medium was originally derived from other compounds; this effect was particularly marked for P. bryantii (Table 1). When bacteria were grown in medium M2, very low proportions of amino acids (see the equation on the previous page) were derived from ammonia in cultures of P. bryantii (0.03) (Table 1), S. ruminantium (0.05) (Table 2), and S. bovis (0.01) (Table 3). A slightly larger proportion of total cell N was derived from ammonia (0.09, 0.08, and 0.12, respectively).

In defined medium, with no added amino acids or Trypticase, virtually all amino acids were formed from ammonia (Tables 1 to 6). The recovery was sometimes apparently greater than 100%, for reasons which were not clear but which must presumably stem from an inaccuracy in the measurement of nitrogen or ¹⁵N in ammonia and cells, the data which were used to calculate the final proportions derived from ammonia. Ammonia enrichment seems likely to have been the measurement causing the problem, since both microbial N and amino acid N data appeared to be affected in similar ways. The degree of error is small, and the anomalous values have little influence on the overall conclusions, however.

When additions of Trypticase (Tables 1, 3, and 5) or amino acids (Tables 2, 4, and 6) were made to the same medium, the proportions of cell N and cellular amino acids derived from ammonia decreased as the Trypticase or amino acid concentration increased. With all species, the proportion of total cell N derived from ammonia was greater than the average proportion of amino acids derived from ammonia, indicating that non-amino-acid cell N was derived predominantly from ammonia; this difference was most pronounced for S. bovis (Tables 5 and 6).

Differences between species were evident when the effects of adding peptides were compared with the effects of adding amino acids. With P. bryantii, increasing the concentration of peptides decreased ammonia uptake more than increasing the concentration of amino acids did (Tables 1 and 2), indicating a preference for peptides over amino acids in protein synthesis. In contrast, ammonia uptake by S. ruminantium and S. bovis was influenced to approximately the same extent by peptides and amino acids (Tables 3 to 6). For all three species, the incorporation of ¹⁵N from ammonia into cell N was greater in the basal medium plus 10 g of Trypticase per liter (0.30, 0.30, and 0.23, respectively) than in medium M2 (0.09, 0.08, and 0.12, respectively), which contains the same concentration of Casitone, a similar pancreatic extract of casein, but also yeast extract and clarified rumen fluid.

The patterns of de novo synthesis among individual amino acids were similar for all media and for all species in that glutamate and aspartate were always the most highly enriched. Serine was the third most highly enriched amino acid for P. bryantii, while alanine was the third most highly enriched amino acid for the other species. Proline biosynthesis was inhibited to the greatest extent by the addition of Trypticase or amino acids, followed by phenylalanine biosynthesis for P. bryantii and S. bovis and valine biosynthesis for S. ruminantium.

The amino acid content of the spent medium from cultures containing peptides or amino acids at 1 g/liter was determined following acid hydrolysis of the medium (Table 7). No individual amino acid was exhausted, although the concentrations of the aromatic amino acids, phenylalanine and tyrosine, were much lower (0.05 to 0.13 mM) than that of other amino acids (0.15 to 1.06 mM) in the spent medium containing 1-g of Trypticase per liter. In the medium containing 1 g of amino acids per liter, the lowest concentrations of amino acids in spent medium were for phenylalanine in cultures of P. bryantii (0.09 mM) and glycine in cultures of S. ruminantium (0.03 mM) and S. bovis (0.09 mM). Other than methionine, which was added to the media at a high concentration (1 g/liter; 7 mM), glutamate was the most abundant amino acid in uninoculated medium, followed by proline and leucine (Table 7). Glutamate was metabolized extensively, particularly when presented as the free amino acid, by all three species. Significant increases in alanine concentrations were observed in the spent medium with S. ruminantium.

The cell density also increased as the concentration of Trypticase and amino acids increased. Trypticase at 30 g/liter increased the yield by 88, 35, and 164% for P. bryantii, S. ruminantium, and S. bovis, respectively, while amino acids had little influence on P. bryantii and increased the yields of S. ruminantium and S. bovis by 204 and 191%, respectively. The lowest concentration, 1 g/liter, of Trypticase resulted in increases of 34, 22, and 57% for the three species, respectively. Utilization of sugars was not measured in these experiments, so molar growth yields cannot be calculated.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The aim of this study was to investigate factors influencing de novo synthesis of different amino acids in three species of ruminal bacteria. Our results describe the extent to which individual amino acids are synthesized de novo by these bacteria. What also emerges, however, is that de novo synthesis of total bacterial N and amino acids in the present experiments differs significantly from the results of other studies. It is therefore important to explore these differences in order to understand how amino acid incorporation by rumen bacteria may be influenced by growth conditions in vivo.

De novo amino acid synthesis varied according to the concentrations of peptides and amino acids in the medium: there is no simple on-off switch for de novo amino acid synthesis. This variation occurred with both peptides and amino acids, and in neither case was any individual amino acid exhausted, even at the lowest concentration of amino acids used. Thus, the observed effects appear not to be a consequence of the removal of an essential or stimulatory single amino acid. Cotta and Russell (17) reported increases in the cell yields of the same and other species of ruminal bacteria when the concentration of amino acids was increased, again noting that no amino acid was exhausted at the lower concentrations. Because certain amino acids, including asparagine, glutamine, cysteine, and tryptophan, are destroyed by acid hydrolysis (7) and were therefore not measured, it is possible that one of these amino acids may limit growth rate or efficiency. However, the experiments of Argyle and Baldwin and Maeng et al. (3, 25, 26) would indicate that no single amino acid is limiting in the sense that is used, for example, to balance the amino acid composition of dietary formulations fed to farm animals. Instead, it was considered possible that groups of amino acids may relieve energetic or metabolic limitations on growth rate or growth yield. Although the practical implications of identifying groups of this nature could be significant, to date no defined group of amino acids appears to provide the benefits of the full array of amino acids (3, 26). It is also worth noting, as was found by Cotta and Russell (17), that only a small proportion of the amino acids added to the medium is incorporated; here, the highest incorporation (26%) was by S. ruminantium in medium with 1 g of amino acids/liter, and most conditions gave much lower incorporation. Increasing the utilization of available amino acids would decrease the quantity available for catabolism and therefore contribute to improving N retention by ruminants (23).

The higher concentrations of peptides tested here correspond to peptide concentrations employed routinely in many growth media in vitro. Between 0.58 and 0.70 of cell N was formed from peptides in medium with 5 g of Trypticase per liter, and 0.88 or more of the bacterial cell N was formed from peptides when bacteria were grown in the routine rumen fluid-containing medium M2 described by Hobson (21). However, peptide concentrations in ruminal fluid are only 1 g/liter or less (8, 14, 15, 37, 40), and at these concentrations, the bacteria in the pure cultures investigated here used peptides for only between 0.11 and 0.36 of their cell N synthesis. Thus, care should be taken in extrapolating from some in vitro experiments to the in vivo situation.

The low uptake of peptides from medium with 1 g/liter contrasts with the findings of Russell et al. (33), who used a medium containing casein at 1 g/liter for the growth of mixed ruminal bacteria on soluble carbohydrates; their results suggested that noncellulolytic bacteria formed 0.66 of their cell N from amino acids, with the remainder being derived from ammonia (33). The experimental conditions were different in the present study, with pure cultures being used here rather than mixed ruminal bacteria. It is, however, important to find the reason for the different incorporation ratios, because the possible influence of growth conditions on amino acid incorporation has important implications for nutritional modelling (32). Many factors other than peptide concentration may influence ammonia and peptide uptake, including growth rate, ammonia concentration, energy source, and availability of precursors for amino acid synthesis. The results of the present experiments do not explain, for example, why mixed ruminal microorganisms can, under some dietary circumstances, derive 50% or more of their cell N from N sources other than ammonia (27, 30), nor is there any explanation as to why medium M2 gave such different results in comparison with the 10-g/liter defined medium, which contains approximately the same concentration of peptides. These issues need to be resolved by further experimentation.

The influence of preformed amino acids on growth yield is an important feature of nutritional modelling (32) and other studies (3, 17, 25, 26). The present experiments have limited value in this respect. Residual carbohydrate concentrations were not measured, so molar growth yields could not be calculated, and yields were measured after overnight growth, when each culture would have been in stationary phase for a number of hours and therefore subject to lysis before being analyzed. Nevertheless, peptides at 1 g/liter increased the apparent yield by between 22 and 57% in the species tested, with even greater stimulation arising from higher concentrations. Similar increases in growth yields of pure cultures of ruminal bacteria have been observed previously (17), and increased yields of up to almost 150% were found in mixed cultures with 0.1 g of Trypticase per liter added (3).

The manner in which ammonia incorporation was affected by preformed amino acids depended on the bacterial species and on whether the amino acids were supplied as peptides or as free amino acids. The availability of preformed amino acids suppressed de novo synthesis more in S. bovis than in the other species, a finding consistent with Cotta and Russell (17). P. bryantii showed a clear preference for the utilization of peptides over amino acids, while S. ruminantium and S. bovis used peptides and amino acids similarly. The preference for peptides of P. bryantii is consistent with the findings of Ling and Armstead (24), who observed that peptides were preferred for growth by P. ruminicola. However, the apparent lack of preference of the other species contrasts with the preference for free amino acids found by Ling and Armstead (24) and a report that amino acid transport was much more predominant than peptide transport in S. bovis (38).

Total cell N was always formed de novo to a greater extent than amino acids. Glutamate and aspartate were the most enriched amino acids in bacteria, consistent with glutamate dehydrogenase being the main route of ammonia assimilation in most ruminal bacteria (12, 20, 36). Transamination with oxaloacetate and pyruvate would account for subsequent enrichment of aspartate and alanine in S. ruminantium and S. bovis, and the enrichment of serine and to a lesser extent glycine in P. ruminicola indicates an active synthesis from 3-phosphoglycerate, involving transamination from glutamate. Why biosynthesis of proline should be switched off so much more readily than that of other amino acids is not clear. Proline can be formed from either glutamate or ornithine, and its synthesis is not linked to the synthesis of other amino acids (18). A similar phenomenon, the switching off of proline synthesis when preformed proline becomes available, occurs in the mixed ruminal population (34).

In conclusion, the present study illustrates that de novo synthesis of amino acids varies in accordance with the amino acid, the bacterial species, and the concentration of amino acids and peptides in the medium. The significance of proline biosynthesis as a possible limitation for rumen bacteria merits further investigation. Future work should also explore the influence of specific growth rate, nutrient concentrations, and the effects of other precursors on amino acid biosynthesis, in order to explain why the proportion of amino acids synthesized de novo is much lower in some in vivo and mixed culture studies than would be predicted from the present experiments.