INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Halophilic archaea are aerobic chemo-organotrophs that grow on a variety of carbon sources. The central carbon metabolisms of some species are relatively well explored (15, 50), while comprehensive investigations of amino acid metabolism have so far been pursued only for organisms belonging to other phylogenetic groups (40) within the domain of the archaea (76), i.e., various methanogens (22, 24, 25, 61) and the anaerobic, extremely thermophilic Thermoproteus neutrophilus (54). These studies indicated that most amino acids in thermophilic and methanogenic archaea are synthesized via pathways that had previously been described for bacteria and eucarya (27, 46, 68, 69). An extension of such studies to halophilic archaea is thus of interest for obtaining new insights into the evolution of carbon metabolism in general. Furthermore, organisms living under extreme environmental conditions (extremophiles) are gaining increasing importance for biotechnological applications (17) and the analysis of their metabolism constitutes a prerequisite for possible future metabolic engineering (2).

In this paper we investigated amino acid biosynthesis in the halophilic archaeon Haloarcula hispanica. H. hispanica was selected for its potential biotechnology interest, since it can efficiently use glycerol for amino acid synthesis (31). We primarily employed biosynthetically directed fractional ¹³C labeling (53, 58, 63-66, 77) with glycerol as the sole carbon source, combined with two-dimensional (2D) ¹³C,¹H correlation nuclear magnetic resonance (NMR) spectroscopy for the analysis of the resulting nonrandom ¹³C-labeling patterns. In this approach, contiguous carbon fragments arising from a single carbon source molecule are traced through a cellular bioreaction network. Since the patterns of intact carbon fragments observed for a given metabolite are very often different depending on which pathway is employed for its synthesis, we are able to analyze both the topological structure of the bioreaction network, i.e., the locations of nodes at which one substance is either a substrate for two branching reactions or a product of two converging reactions, and the relative contributions of alternative pathways to the generation of amino acids (63, 64, 66).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Labeling strategy. The biosynthetic pathways were explored by biosynthetically directed fractional ¹³C labeling of the proteinogenic amino acids (53, 58, 63-66, 77), with glycerol as the sole carbon source. Fractional ¹³C labeling was achieved by growing H. hispanica in a minimal medium containing approximately 10% uniformly ¹³C-labeled glycerol and 90% glycerol containing ¹³C at natural abundance. Incorporation of intact two- or three-carbon fragments from the uniformly ¹³C-labeled carbon source leads to nonrandom ¹³C-labeling patterns in the amino acids. These are identified from ¹³C---¹³C scalar-coupling fine structures in 2D ¹³C,¹H correlation spectroscopy (COSY) (11). With a set of probabilistic equations (63), the observed ¹³C fine structures then yield the relative abundances of intact glycerol carbon fragments in the carbon skeletons of the amino acids. This approach allows for a comprehensive analysis of the bioreaction network (63, 64, 66). Cells were grown in a batch culture and harvested in the mid-exponential and the early stationary phases in order to assess possible changes of the metabolic state during growth. Since all relevant peaks of the individual amino acids are well resolved in the 2D NMR spectrum, a separation of the amino acids prior to NMR analysis is not required (63-66, 77). To verify the threonine pathway for isoleucine biosynthesis, the labeling experiment was repeated with a growth medium containing [¹³C₄]threonine instead of [¹³C₃]glycerol, which allowed us to directly trace the incorporation of ¹³C---¹³C units from threonine into isoleucine.

Growth of the organism and sample preparation. By following the protocol developed in reference 49, H. hispanica (31) was grown in a medium containing, per liter, 200 g of NaCl, 36 g of MgSO₄ · 7H₂O, 6 g of Tris base, 4 g of KCl, 1 g of CaCl₂ · 2H₂O, 2 ml of FeSO₄ · 7H₂O (0.4% in 1 mM HCl), 2 ml of K₂HPO₄ (5% in distilled water), and 5 ml of NH₄Cl (20% in distilled water). Glycerol (20 ml, 25% in water) was added, and the pH was adjusted to 7.5 with HCl prior to sterilization (filter pore size cutoff, 0.45 µm). In the standard experiments, 10% of the glycerol was uniformly ¹³C labeled. In the experiment with ¹³C-labeled threonine, no labeled glycerol was used and 307 mg of [¹³C₄]threonine per liter was added under otherwise identical conditions. Cells were grown in six 35-ml cultures that were shaken at 100 rpm in 100-ml Erlenmeyer flasks for 7 days at 40°C until the early stationary phase was reached. Cells from cultures in the mid-exponential growth phase were harvested after 3 days. After centrifugation at 3,000 × g, the cells were taken up in 10 ml of water and frozen in liquid nitrogen. No DNase was added in order to prevent contamination of the halobacterial proteins with unlabeled amino acids. After being warmed to 0°C, the slurry was centrifuged at 20,000 × g and the protein in the clear supernatant was precipitated with 65% ethanol at a temperature of 20°C overnight. After centrifugation, the pellet was lyophilized and hydrolyzed after addition of 3 ml of 6 M HCl at 80°C for 2 days in a sealed Pyrex tube, yielding about 70 mg of dried biomass.

NMR spectroscopy. NMR experiments were performed at 40°C. Proton-detected 2D ¹³C,¹H-heteronuclear single-quantum correlation spectra were recorded with the pulse sequence devised by Bodenhausen and Ruben (11), which ensures that ¹H---¹³C scalar couplings do not affect the ¹³C---¹³C scalar-coupling fine structure along ₁(¹³C) (Fig. 1 in reference 65). Pulsed-field gradients were employed for coherence pathway rejection (6, 75), and a 2-ms spin-lock pulse (47) was used to purge the magnetization arising from ¹²C-bound protons and the residual ²HOH signal. ¹³C decoupling during t₂ was achieved with the composite-pulse decoupling scheme GARP (59), and quadrature detection in ₁ was accomplished with States-TPPI (41). For the samples obtained from the cultures grown with [¹³C₃]glycerol, two spectra were recorded, i.e., one focused on the aliphatic carbons, with the ¹³C carrier set to 42.5 ppm relative to the chemical shift of 2,2-dimethyl-2-silapentane-5-sulfonate sodium salt (DSS), and one for the aromatic rings, with the ¹³C carrier set to 125.9 ppm. For the sample obtained from the culture grown with [¹³C₄]threonine, only the spectrum focusing on the aliphatic carbons was recorded. The spectra for the aliphatic resonances were folded along ₁(¹³C), with a sweep width of 33.8 ppm.

Data analysis. The individual multiplet components of the ¹³C---¹³C scalar-coupling fine structures were integrated with the program XEASY (4), and the observed relative multiplet intensities were used to calculate the relative abundances of intact carbon fragments (63). By following the definitions of references 63 and 66, f⁽¹⁾ represents the fraction of molecules in which the observed carbon atom and its neighboring carbons originate from different source molecules of glycerol and f⁽²⁾ represents the fraction of molecules in which the observed carbon atom and at least one neighboring carbon originate from the same source molecule. For a central carbon in a C₃ fragment that exhibits different ¹³C---¹³C scalar-coupling constants with the two attached carbons, f⁽²⁾ represents the fraction of molecules for which the central carbon and the carbon with the smaller coupling come from the same source molecule, while f^(2*) is used if the carbon with the larger coupling comes from the same source molecule as the observed carbon. f⁽³⁾ denotes the fraction of molecules in which the observed carbon atom and both neighbors in the C₃ fragment originate from the same glycerol molecule.

(i) Biosynthesis of isoleucine. C₂---C of Ile derives from pyruvate in all pathways considered (see Fig. 2 and 3B), and Ile-₂ is therefore not used in the calculations.

(1)

(2)

(3)

(4)

(5)

(6)

(ii) Biosynthesis of lysine (Fig. 3C). (a) Diaminopimelate pathways. Pyruvate and aspartate are assessed via Ala- and Asp- or Asp-, respectively.

(7)

(8)

(9)

(10)

(b) -Aminoadipate pathway.

(11)

(12)

(iii) Biosynthesis of the aromatic ring of tyrosine. In the biosynthesis of the aromatic ring of tyrosine through the common shikimate pathway (7, 68, 69), the ring is assembled from phosphoenolpyruvate and erythrose-4-phosphate. Phosphoenolpyruvate is assessed via Phe- and Tyr-. Erythrose-4-phosphate is assumed to include an intact C₂---C₃---C₄ fragment derived from a first glycerol molecule via the C₃ pool of glycolysis and a C₁ atom that originates from a second glycerol molecule (see Fig. 1) (50). Due to the symmetry of the aromatic ring of tyrosine, the Tyr-_x and Tyr-_x carbons give rise to only one ¹³C fine structure. Therefore, f⁽ⁱ⁾{Tyr-} and f⁽ⁱ⁾{Tyr-} are calculated as averages of the f values predicted for Tyr-₁ and Tyr-₂ or Tyr-₁ and Tyr-₂, respectively (63).

(13)

(14)

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The 2D ¹³C,¹H COSY spectra were analyzed as described by Szyperski (63) (Table 1). Virtually identical ¹³C scalar-coupling fine structures were observed for samples taken from the mid-exponential growth phase and the early stationary phase, and hence there are no significant differences in the resulting relative abundances of intact carbon fragments, as collected in Table 1. This suggests that there are no major differences in flux ratios through the central metabolic network when the two growth periods are compared (63-66). During hydrolysis, cysteine and tryptophan were oxidized and could thus not be evaluated and asparagine and glutamine were deamidated to aspartate and glutamate (65, 77). The ring carbons of phenylalanine were not evaluated because of strong-coupling effects (63). The evaluation of the observations for all other carbon positions (Table 1) showed that except for isoleucine, lysine, and the aromatic ring of tyrosine, the proteinogenic amino acids in H. hispanica are synthesized according to the pathways commonly found in both bacteria and eucarya (27, 46, 68, 69).

Amino acid synthesis from glycolytic intermediates. Identical values for intact carbon fragments (f values) are observed for phenylalanine and tyrosine, where f⁽³⁾{Phe-} approximately equals f⁽³⁾{Tyr-}, which approximately equals 1 (Table 1). This finding is in agreement with the shikimate/chorismate pathway (7, 68, 69), where the C---C---C' fragments of phenylalanine and tyrosine are both derived from phosphoenolpyruvate, which is itself expected to derive from glycerol without cleavage of carbon-carbon bonds (50). Serine appears to be synthesized from 3-phosphoglycerate, which is also directly derived from glycerol. However, reversible interconversion into glycine and a C₁ unit leads to cleavage of C---C connectivities, so that f⁽¹⁾{Ser-} approximately equals 0.55 in both preparations (Table 1). Moreover, the fact that [f^(2*) + f⁽³⁾]{Ser-} approximately equals [f⁽²⁾]{Gly-} (Table 1) shows that glycine is nearly exclusively derived from C---C' of serine.

2

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Synthesis of erythrose-4-phosphate from glycerol via reversal of glycolytic reactions and the action of transketolase (EC 2.2.1.1). Thick lines indicate carbon-carbon connectivities arising from a single source molecule of glycerol (see the text).

Amino acid synthesis from intermediates of the tricarboxylic acid cycle. The f values observed for Asp- and Asp- coincide with those observed for Thr- and Met- (Asp-) and Thr- (Asp-). Moreover, virtually identical f values were found for Glu- and Pro-; for Glu-, Pro-, and Arg-; for Glu- and Pro-; and for Pro- and Arg- (Table 1). This is consistent with the well-known pathways in which aspartate, threonine, and methionine are derived from oxalacetate and in which 2-oxoglutarate serves for the biosynthesis of glutamine, proline, and arginine (63). Furthermore, the following observations demonstrate that 2-oxoglutarate is formed by irreversible condensation of oxalacetate with acetyl-CoA in the citric acid cycle (63): (i) the equation f⁽³⁾{Glu-}  f⁽³⁾{Pro-}  f⁽²⁾{Glu-}  f⁽³⁾{Glu-}  f⁽³⁾{Pro-}  0 shows that intact C₃---C₄ connectivities in 2-oxoglutarate are absent; (ii) the f values observed for Asp- (and also Thr-) are equal to those observed for Glu- (and Pro-), which is in agreement with the fact that the C₁---C₂---C₃ segment of 2-oxoglutarate is derived from C₂---C₃---C₄ of oxalacetate; and (iii) the equation f^(2*){Glu-}  f^(2*){Leu-}  f⁽²⁾{Pro-}  f⁽²⁾{Arg-} shows that C₄---C₅ of 2-oxoglutarate is derived from acetyl-CoA (Table 1).

Isoleucine biosynthesis. A variety of pathways are known for isoleucine biosynthesis in microorganisms. Most common is the threonine pathway, which uses threonine and pyruvate as precursors (1, 68, 69) (Fig. 2 and 3B). In the pyruvate pathway, isoleucine biosynthesis proceeds from pyruvate and acetyl-CoA (14) (Fig. 2 and 3B), whereas in the glutamate pathway, glutamate and pyruvate serve as precursors (36, 48) (Fig. 2 and 3B). In rare cases, isoleucine synthesis has also been found to proceed from homoserine (26, 71), propionate (23, 45, 52), or 2-methylbutyrate (23, 45, 51). In all pathways except the one starting from 2-methylbutyrate, -ketobutyrate serves as an intermediate and pyruvate yields the C₂---C fragment of isoleucine.

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Three routes for isoleucine biosynthesis (see the text). Pyruvate and acetyl-CoA are the precursors for Ile synthesis via the pyruvate pathway (route 1) (14), threonine and pyruvate are the precursors for Ile synthesis via the threonine pathway (route 2) (1, 68, 69), and glutamate and pyruvate are the precursors for Ile synthesis via the glutamate pathway (route 3) (36, 48). -Ketobutyrate is a common intermediate in these three pathways.

View larger version (34K): [in this window] [in a new window]   FIG. 3.   Incorporation of oxalacetate or threonine, 2-oxoglutarate, pyruvate, and acetyl-CoA, for which carbon skeletons are schematically shown (A), into the carbon skeletons of isoleucine (B) and lysine (C), according to the biosynthetic pathways shown in Fig. 2 and 6. Note that threonine is synthesized from oxalacetate without rearrangement of the carbon skeleton (see the text). The notation of the carbon atoms follows IUPAC-IUB recommendations (30), i.e., O2C(4)---C(3)H3---C(2)O---C(1)O2 for oxalacetate, O2C(5)---C(4)H2---C(3)H2---C(2)O---C(1)O2 for 2-oxoglutarate, C(3)H3---C(2)O---C(1)O2 for pyruvate, and C(2)H3---C(1)O---SCoA for acetyl-CoA.

2

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Decomposition of the experimental 13C---13C scalar-coupling fine structures for the Ile- and Ile- carbons into contributions from the threonine and pyruvate pathways (Fig. 2; Table 2). The stick diagrams represent the fine structures that would be expected if only a single pathway were operational, and the experimental cross sections on the right were taken along 1(13C) from the 2D 13C,1H COSY spectrum recorded with biomass that was harvested in the early stationary phase (see the text). Fifty-six and 44% of isoleucine are synthesized via the threonine and pyruvate pathways, respectively. The carbon chemical shifts are relative to those of DSS (2,2-dimethyl-2-silapentane-5-sulfonate sodium salt).

1

2

CC

C2C'

1

View larger version (24K): [in this window] [in a new window]   FIG. 5.   Cross sections taken along 1(13C) from the 2D 13C,1H COSY spectrum recorded with the biomass labeled with [13C4]threonine. (A to C) 13C---13C scalar-coupling fine structures of the precursors used for isoleucine biosynthesis via the threonine pathway (Fig. 2), i.e., with Thr-, Thr-2, and Ala- representing C3 of pyruvate. (D to F) 13C---13C fine structures observed at Ile-, Ile-, and Ile-2 (Ile-1 yields the same information as Ile- [63]). The fine structures of Ile- and Ile- are dominated by a doublet, which proves that the fragments C---C1 and C---C' originate from [13C4]threonine. The 2 carbon exhibits a 13C---13C fine structure similar to that of Ala-, which is consistent with the fact that the C2---C fragments of Ile are derived from pyruvate. The additional small splittings of the doublet components of Ile- and Ile- (D and E) and the broadening of the Thr-2 doublet lines (B) arise from the vicinal scalar couplings 3JCC in isoleucine and 3JC2C' in threonine (35) (see the text). The carbon chemical shifts are relative to those of DSS. The asterisk in panel E indicates an impurity.

Lysine biosynthesis. Lysine biosynthesis via the "diaminopimelate pathway" starts from aspartate and pyruvate, whereas the "-aminoadipate pathway" relies on 2-oxoglutarate and acetyl-CoA as precursors (9, 54) (Fig. 3C and 6A). Two variants of the diaminopimelate pathway differ in the reaction sequence used to convert L-¹-piperidine-2,6-dicarboxylate to DL-diaminopimelate (Fig. 6B). In the dehydrogenase variant, L-¹-piperidine-2,6-dicarboxylate is directly converted to DL-diaminopimelate by diaminopimelate dehydrogenase (EC 1.4.1.16) (42, 43, 74). In the acetylase/succinylase variant (8, 32, 62), L-¹-piperidine-2,6-dicarboxylate is first acetylated or succinylated and then converted to the symmetric intermediate LL-diaminopimelate, which is finally epimerized to DL-diaminopimelate. The dehydrogenase variant can readily be distinguished from the acetylase/succinylase variants in the fractional labeling experiment, because the involvement of the symmetric intermediate LL-diaminopimelate implies symmetrization of the ¹³C-labeling pattern about the C---C bond (Fig. 3C).

View larger version (27K): [in this window] [in a new window]   FIG. 6.   Lysine biosynthesis. (A) Synthesis via the -aminoadipate pathway (9, 54); (B) synthesis via the diaminopimelate pathway. In the acetylase/succinylase variant, L-1-piperidine-2,6-dicarboxylate is converted to DL-diaminopimelate via the symmetric intermediate LL-diaminopimelate (route 1) (8, 32, 62). In the dehydrogenase variant, L-1-piperidine-2,6-dicarboxylate is directly converted to DL-diaminopimelate by diaminopimelate dehydrogenase (route 2) (42, 43, 74).

Threonine cleavage. In the labeling experiment with [¹³C₄]threonine, the fine structure of Leu- (representing C₂ of acetyl-CoA [see Fig. 2 in reference 63]) contains a higher proportion of the doublet component than the fine structure of Ala- (representing C₃ of pyruvate) and a large proportion of ¹³C---¹³C' fragments is also found for Gly- (Fig. 7). This provides evidence that the exogenously supplied threonine is cleaved into glycine and acetaldehyde, which would be compatible with the assumptions that there is threonine aldolase activity in H. hispanica and subsequent conversion of acetaldehyde into acetyl-CoA (69). This indication of threonine aldolase activity in a halophilic archaeon complements data obtained with other organisms: L-threonine aldolases (EC 4.1.2.5) from Saccharomyces cerevisiae (GLY1 [37, 44]), Escherichia coli (ltaE [39]), and Pseudomonas sp. strain NCIMB 10558 (ltaP [38]) have been cloned and expressed, and L-threonine aldolase activity has been demonstrated for serine hydroxymethyltransferase (EC 2.1.2.1) in E. coli (55), although this enzyme serves primarily for the cleavage of serine to glycine and a C₁ unit. The serine hydroxymethyltransferase from the thermophilic archaeon Sulfolobus solfataricus has been shown to possess allo-L-threonine aldolase activity (18). Genome sequencing showed that serine hydroxymethyltransferase is also present in Methanobacterium thermoautotrophicum (60), in Methanococcus jannaschii (12, 57), and in the hyperthermophilic, sulfate-reducing archaeon Archaeoglobus fulgidus (34).

View larger version (13K): [in this window] [in a new window]   FIG. 7.   Evidence for threonine cleavage in H. hispanica derived from the labeling experiment with [13C4]threonine. The 13C---13C scalar-coupling fine structure of Leu- (A), which represents C2 of acetyl-CoA, exhibits a significantly more intense doublet component than that of Ala- (B), which represents C3 of pyruvate. A strong doublet is also detected for Gly- (C). These data indicate that a fraction of the exogenously supplied threonine was cleaved into glycine and acetaldehyde, with subsequent conversion of acetaldehyde into acetyl-CoA (69). The carbon chemical shifts are relative to those of DSS.

Analysis of the ¹³C-labeling pattern in tyrosine indicates yet unknown biosynthetic pathways. Commonly, erythrose-4-phosphate and phosphoenolpyruvate serve for the biosynthesis of the aromatic ring of tyrosine via the shikimate pathway (7, 68, 69). To evaluate the ¹³C-labeling pattern of the tyrosine ring in light of this pathway, f values for this pathway were predicted with the assumption that histidine biosynthesis proceeds via ribose-5-phosphate (68, 69). The pentose pool must then be composed of molecules carrying intact C₁---C₂---C₃ and C₄---C₅ fragments (Fig. 1) (see above for f values obtained for His), so that erythrose-4-phosphate would be expected to bear a C₂---C₃---C₄ fragment from one source molecule and a C₁ carbon from a second glycerol molecule (Fig. 1). Significant deviations between the experimental data and the thus calculated values for a "pure" shikimate pathway (Table 4) cannot be explained within the framework of commonly known pathways. In particular, biosynthesis of the aromatic ring from 2-keto-3-deoxyarabino-heptulosonate-7-phosphate (DAHP), which itself arises from erythrose-4-phosphate and phosphoenolpyruvate via DAHP synthase (EC 4.1.2.15), predicts that 50% of the  carbons must not be connected to carbon atoms that arise from the same source molecule, i.e., f⁽¹⁾{Tyr-} equals 0.5 and f⁽²⁾{Tyr-} equals 0 (Fig. 8). Hence, the detection of intact C---C fragments (f⁽²⁾{Tyr-}  0.21) (Table 4) excludes the sole operation of the standard shikimate pathway. This result appears to be in line with data obtained previously for the methanogen Methanococcus maripaludis with a different ¹³C-labeling approach, where the standard shikimate pathway (67) could not account for all the observations. (Note that the operation of the oxidative part of the pentose phosphate cycle has not yet been observed for halophilic archaebacteria. In cell extracts of Haloferax mediterranei and Haloarcula vallismortis, 6-phosphogluconate dehydrogenase (EC 1.1.1.44) was not active at physiological salt concentrations whereas the observed transketolase and transaldolase activities ensured the formation of pentoses from hexoses (50). However, even if pentose biosynthesis in H. hispanica occurred via glucose oxidation, the same relative abundances of intact carbon fragments would be predicted for the erythrose-4-phosphate pool.

View larger version (17K): [in this window] [in a new window]   FIG. 8.   Scheme indicating the transfer of intact C2 and C3 fragments of glycerol via DAHP (2-keto-3-deoxyarabino-heptulosonate-7-phosphate) to the aromatic ring of tyrosine. Thick lines indicate carbon-carbon connectivities arising from a single molecule of glycerol. DAHP was assumed to be synthesized from erythrose-4-phosphate and phosphoenolpyruvate (compare with Fig. 1). One of the  carbons of tyrosine must then be bound to two carbons that originate from different glycerol molecules. The experimental detection of C---C fragments thus excludes the possibility that the standard shikimate pathway is the only pathway leading to tyrosine biosynthesis in this system (see the text and Table 4). The double line indicates the cyclization site of the Tyr ring. C1 of DAHP is lost as CO2.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In perspective with current knowledge on amino acid biosynthesis in bacteria and eucarya, the present data on isoleucine and lysine biosynthesis are particularly intriguing and expand knowledge accumulated for other organisms in the domain of the archaea. Thus, the identification of a split threonine/pyruvate pathway for isoleucine biosynthesis in H. hispanica is a novel finding, since the threonine pathway has not previously been reported for archaea. In archaebacterial species, such as methanobacteria (21-25, 61) and the thermophilic archaeon Thermoproteus neutrophilus (54), the pyruvate pathway is used. In methanogens, isoleucine synthesis can also proceed from propionate or 2-methylbutyrate (23). Notably, the threonine pathway is common in eucaryotic microorganisms and bacteria, although the pyruvate pathway has also been reported for such species (20, 33, 71, 73). However, constitutive, simultaneous operation of the threonine and pyruvate pathways has so far been observed only in a very few organisms, e.g., spirochetes of the genus Leptospira (73). In addition, split isoleucine synthesis pathways have been reported for eucaryotic and procaryotic microorganisms such as Serratia marcescens (threonine/pyruvate pathways) (33), E. coli Crookes and K-12 (threonine/glutamate pathways) (36, 48), Rhodopseudomonas sphaeroides (threonine/glutamate pathways) (16), and S. cerevisiae (threonine/pyruvate pathways or synthesis from homoserine) (71), but with these organisms a genomic mutation or special growth conditions are required to activate routes other than the threonine pathway.

Lysine biosynthesis via the dehydrogenase variant of the diaminopimelate pathway has previously been identified, for example, in Bacillus sphaericus (74) and in Corynebacterium glutamicum (29, 56, 72), and here we now present direct evidence that this pathway also exists within the domain of the archaea. It has previously been shown that lysine biosynthesis in the methanogens Methanospirillum hungatei (22), Methanococcus voltae (24), Methanothrix concilii (25), and Methanobacterium thermoautotrophicum (3) occurs via the diaminopimelate pathway, but the experiments performed in these earlier studies did not allow us to distinguish between the two variant pathways of Fig. 6B. For Methanobacterium thermoautotrophicum the enzymes that were assayed, i.e., dihydrodipicolinate synthase (EC 4.2.1.52) and diaminopimelate decarboxylase (EC 4.1.1.20), catalyze reactions that are common to both variants of the pathway. For Methanospirillum hungatei (22), Methanococcus voltae (24), Methanococcus jannaschii (61), and Methanothrix concilii (25), ¹³C-labeling experiments with [1-¹³C]acetate, [2-¹³C]acetate, or ¹³CO₂ were used in biosynthetic studies, but with these specifically labeled precursors both variants of the diaminopimelate pathway (Fig. 6B) yielded the same positional ¹³C enrichments in lysine (15). Indications for the operation of the acetylase/succinylase variant in methanogens and Archaeoglobus fulgidus were obtained from genome sequencing. The gene for LL-diaminopimelate epimerase (EC 5.1.1.7), which catalyzes the conversion of LL-diaminopimelate to meso-diaminopimelate in the acetylase/succinylase variant (Fig. 6B), was identified in Methanobacterium thermoautotrophicum (60), Methanococcus jannaschii (12, 57), and Archaeoglobus fulgidus (34), and the succinyl-diaminopimelate desuccinylase gene (EC 3.5.1.18) was found in Methanococcus jannaschii (12, 57) and Archaeoglobus fulgidus (34). The -aminoadipate pathway for lysine synthesis (Fig. 6A), which has previously been identified solely for lower eucarya such as fungi, algae, and yeast (9, 70), operates in the thermophilic archaeon Thermoproteus neutrophilus (54). Overall, with the new data presented in this paper the implication is that all currently known pathways for lysine biosynthesis in bacteria and eucarya exist also in the domain of archaea.

Organisms living under extreme environmental conditions, such as archaebacteria, have considerable interest for applications in biotechnology (17). Elucidation of their bioreaction networks is thus of interest as a basis for metabolic engineering (2). In this context the present investigation is another telling illustration of the power of biosynthetically directed fractional ¹³C labeling in combination with 2D ¹³C,¹H NMR techniques (58, 65) for support of flux analysis in central metabolism (53, 63, 64, 66). This approach revealed a remarkable diversity of pathways employed for isoleucine and lysine biosynthesis in a halophilic archaeon and showed that with the sole exception of the aromatic ring of tyrosine (and possibly phenylalanine), the proteinogenic amino acids are synthesized according to the common pathways (68, 69). Comparison with the amino acid metabolisms of eucarya and bacteria then also supports the notion that pathways for synthesis of proteinogenic amino acids were probably largely established before the divergence of these three domains (76).