De novo synthesis of amino acids by the ruminal anaerobic fungi, Piromyces communis and Neocallimastix frontalis.

Anaerobic fungi are an important component of the cellulolytic ruminal microflora. Ammonia alone as N source supports growth, but amino acid mixtures are stimulatory. In order to evaluate the extent of de novo synthesis of individual amino acids in Piromyces communis and Neocallimastix frontalis, isotope enrichment in amino acids was determined during growth on (15)NH(4)Cl in different media. Most cell N (0.78 and 0.63 for P. communis and N. frontalis, respectively) and amino acid N (0.73 and 0.59) continued to be formed de novo from ammonia when 1 g l(-1) trypticase was added to the medium; this concentration approximates the peak concentration of peptides in the rumen after feeding. Higher peptide/amino acid concentrations decreased de novo synthesis. Lysine was exceptional, in that its synthesis decreased much more than other amino acids when Trypticase or amino acids were added to the medium, suggesting that lysine synthesis might limit fungal growth in the rumen.


Introduction
The anaerobic fungi play a signi¢cant, sometimes crucial, role in the degradation of plant cell walls in the rumen [1]. They are highly cellulolytic [1,2], and are particularly numerous in the rumen of animals given high ¢brecontaining diets [3]. Ruminal fungi, unlike ruminal cellulolytic bacteria, are proteolytic [4,5], which may enable the fungi to disrupt the proteinaceous layer that limits the access of the cellulolytic bacteria to the secondary cell wall [6].
The nutritional requirements of ruminal fungi are relatively simple. Lowe et al. [7] reported that they grew in medium lacking amino acids, implying that ammonia can be used to form all amino acids. Orpin and Greenwood [8] also demonstrated that Neocallimastix patriciarum grew in a de¢ned medium, and that growth was stimulated by amino acids, particularly glutamate, serine and methio-nine. However, little is known concerning amino acid metabolism in ruminal fungi, or the extent to which fungi may incorporate amino acids [9]. The present study was therefore undertaken to determine which amino acids are formed de novo by anaerobic fungi from the rumen and how di¡erent nitrogen sources a¡ected de novo amino acid synthesis by these organisms.

Fungal strains and culture conditions
The fungal species used in this study were Piromyces communis strain P, a gift from C.G. Orpin, and Neocallimastix frontalis strain RE1, isolated from a sheep at the Rowett Research Institute. The fungi were maintained on the liquid form of medium M2 [10]. 15 N experiments were carried out using the basal de¢ned medium of Hungate and Stack [11], which contains only cysteine in addition to NH 3 as N source, with 0.6% (w/v) cellobiose (Sigma, Poole, Dorset, UK) added as ¢lter-sterilised solution after autoclaving, and 0.05 g l 31 of vitamins B 1 and B 2 . Part (40%) of the NH 4 Cl on the minerals solution was replaced by 15  Becton Dickinson Microbiology Systems, Cockeysville, MD, USA), or amino acids [12] were added at 1 or 10 g l 31 before autoclaving.
Fungi were inoculated (5% by volume) from fresh 48-h cultures grown on M2 medium into cellobiose-containing media and M2 medium and incubated for 72 h at 39 ‡C. Fungi were harvested by centrifugation (15 000Ug, 20 min) and pellets were washed once with ice-cold water, then the pellets and supernatants were frozen at 320 ‡C. 15 N enrichment in total precipitable N was measured by isotope ratio mass spectrometry as described by Barrie and Workman [13]. Samples were processed and analysed for 15 N enrichment in amino acids by gas chromatography/ mass spectrometry (GC/MS) of derivatised amino acids [14] as described previously [15]. Ammonia was measured by an automated phenol^hypochlorite method [16]. Determination of 15 N enrichment in ammonia was carried out according to the procedure by Nieto et al. [17]. Calculations were described previously [15]. Protein was hydrolyzed using HCl, which results in the breakdown of glutamine, asparagine and tryptophan, and the GC/MS method did not detect cysteine adequately. Thus, the enrichment of these amino acids was not determined.

Statistical analysis
Results are all means derived from the analysis of trip-licate cultures derived from a single inoculum culture. The data were compared by analysis of variance with di¡erent cultures used as blocking factor. To compare the e¡ects of treatments on ammonia uptake into amino acids, individual amino acids were considered as a subplot within the design. All analysis was carried out using GENSTAT 5 (Lawes Agricultural Trust, Rothampsted Experimental Station, Harpenden, Herts, UK).

Results
Ammonia concentrations in the medium following growth of both fungal species (Tables 1 and 2) indicated a small net incorporation of NH 3 at low added peptide and amino acids concentrations and a small net production at higher concentrations. A low catabolism of amino acids to NH 3 was indicated by similar 15 N enrichment in NH 3 before and after growth. Medium M2 contains 10 g l 31 casitone, equivalent to about 1.4 g N l 31 , yet only 0.09 mg NH 3 N l 31 was released by N. frontalis during growth on M2 (Table 1). Thus, the fungi produce only a small amount of NH 3 from amino acids. 15 N enrichment in cell N indicated that both species formed most of their cell N from ammonia when grown in Hungate and Stack medium with added 1 g l 31 trypticase or amino acids, with P. communis incorporating a little less amino acid N than N. frontalis (0.22 of cell N cf. 0.37). The proportion of cell N derived from NH 3 fell when the amino acids concentration in the medium in- creased to 10 g l 31 . The form of amino acids, whether free or predominantly in the form of peptides (Trypticase) had no signi¢cant e¡ect except at 10 g l 31 concentrations with N. frontalis, where amino acids were incorporated a little more extensively than peptides. There was considerable variation in the extent to which di¡erent amino acids were synthesised de novo by the fungi. Glutamate was the most enriched amino acid in N. frontalis, while aspartate was generally the most enriched amino acid in P. communis. Both fungal species had exceptionally low 15 N enrichment in lysine, 10^30% compared to the average in all amino acids of 21^73%.

Discussion
The aim of this study was to provide information on de novo synthesis of amino acids by the two important ruminal fungal species and to determine the in£uence of di¡erent N sources on amino acid synthesis by these species. The results revealed that, at 1 g peptides l 31 , similar to maximum peptides concentrations prevailing in the rumen [18^20], 63 and 78% of cellular N were derived from ammonia by N. frontalis and P. communis, respectively (Tables 1 and 2). In the rumen, therefore, it might be expected that on average of 70% or more of the cell N of ruminal fungi would be derived from NH 3 . Similar proportions were found with cellulolytic ruminal bacteria [12], in contrast to lower proportions for non-cellulolytic bacteria [15].
Higher concentrations of protein hydrolysates (10 g l 31 ) are used routinely in growth media such as M2 medium [10]. The present results show that a signi¢cantly greater proportion (32^61%) of the cell N of the fungal species was derived from ammonia compared to 1 g l 31 protein hydrolysate (59^75%). Thus, care must be taken when extrapolating from in vitro incorporation data to the ecosystem in vivo. Similar concentration-dependence occurs with cellulolytic bacteria [12]. There is also concern that local peptides and amino acids concentrations in the ¢bre-associated microenvironment could be much higher than in the rest of the digesta, particularly for penetrating structures such as the fungal rhizoid [2,3], resulting in an extent of amino acid incorporation similar to that observed here at the higher peptides and amino acids concentrations. A further complication is that, during the course of a 72-h incubation, some fungal lysis will take place, releasing amino acids available for reincorporation. However, the concentration of the released amino acids would be expected to be much lower than the concentrations of peptides and/or amino acids in the medium. Thus, although the results and conclusions are based on net growth, we would not expect the conclusions to be di¡erent to those that would be obtained measuring total growth.
The amino acids treatment was included to compare the e¡ects of provision of amino acids in di¡erent forms as additions to the growth medium since some of the bacterial species have preference for peptides over amino acids or vice versa [12,15,21^23]. Amino acids decreased NH 3 incorporation into amino acids to similar extents, except for a minor preference for amino acids at 10 g l 31 by N. frontalis, suggesting that the fungi have no strong preference for amino acids over peptides. Orpin and Greenwood [8] reported that the provision of amino acids greatly stimulated the growth of N. patriciarum. They reported that the stimulation was greatest with glutamate (290%), serine (24%), and methionine (216%), suggesting that this fungus could not adequately form these amino acids by synthesis. However, 15 N enrichment data of the present study in individual fungal amino acids have clearly revealed that glutamate was the most highly enriched amino acid, followed by aspartate in N. frontalis (Table 1), even when pre-formed glutamate and aspartate were available in the growth medium. This implies that, like in ruminal bacteria, GDH is the main pathway for ammonia uptake in this microorganism [24,25]. It also implies that the synthesis of glutamate or serine would not limit the growth in ruminal fungi. As the fungi have little deaminative activity, as shown here, the reasons for the Orpin and Greenwood results [8] are unclear.
Aspartate was the amino acid most highly labelled with 15 N, followed by glutamate and alanine in P. communis ( Table 2), indicating that transamination with oxaloacetate and pyruvate would most probably account for subsequent enrichment of aspartate and alanine in this species. In contrast, de novo synthesis of lysine was much less than that of the other amino acids in both species, indicating preferential utilisation of this amino acid. A similar phenomenon, except with di¡erent amino acids, was observed before: non-cellulolytic ruminal bacteria [15] and mixed ruminal microorganisms fermenting starch and soluble sugars [26] rapidly and preferentially closed down proline synthesis when trypticase was provided; in contrast, cellulolytic bacteria appeared to have a primary requirement for phenylalanine [12]. The apparent bene¢t of lysine by the fungi therefore illustrates further the variation in amino acid metabolism and requirements between di¡erent members of the rumen microbial community, and therefore the di⁄culty of predicting the amino acid requirements of the community as a whole.