A Reconstruction of the Metabolism of Methanococcus jannaschii from Sequence Data

Evgeni Selkov*, Natalia Maltsev*, Gary J. Olsen**, Ross Overbeek*, and William B. Whitman***

* Mathematics and Computer Science Division, Argonne National Laboratory, 9700 South Cass Ave., Argonne, IL 60439-4844
** Microbiology Department, University of Illinois, Champaign-Urbana, IL 61801
*** Department of Microbiology, University of Georgia, Athens, GA 30602

Published in: Gene Compbios 197, issue 1/2, pp. 10-25, 1997

1. Introduction

The complete genome of Methanococcus jannaschii was placed in the public databases in late August 1996, just as the event was announced in Science[Bult et al., 1996]. The actual sequence, along with an emerging estimate of the genes and their functionality, is maintained by The Institute for Genome Research, which did the sequencing. Recognition of the significance of this event was almost immediate [Gray, 1996; Fox, 1996]. The availability of the first complete archaeal genome is certainly a major event in the history of microbiology. More archaeal genomes will follow in quick succession, along with many more bacterial and eukaryotic genomes. We are rapidly reaching the point where a goal as ambitious as "characterizing unicellular life" can be openly discussed without inviting scorn.

The work presented in this article is a direct outgrowth of our efforts to accurately identify the coding regions in Methanococcus jannaschii. A number of the authors participated in the initial attempt to determine the coding sequences and establish estimates of the function associated with the corresponding protein. It was decided that the development of a metabolic reconstruction for the organism was needed. Evgeni Selkov, working with a team at Argonne National Laboratory, had developed such reconstructions for Haemophilus influenzae and Mycoplasma genitalium , the first two prokaryotic genomes that were completely sequenced [Fleischmann et al., 1995; Fraser et al., 1995] . We decided to formulate an initial metabolic reconstruction that would integrate the sequence data with the known biochemical and phenotypic data.

What emerges is a reconstruction in which much of the metabolism revealed by sequence analysis is in close agreement with the known biochemistry. In these areas of agreement, we believe that the careful depiction of the pathways, labeled with EC numbers and connected to the actual coding sequences corresponding to these functional roles, will be of value to others exploring this genome. However, there is more to be said:

• The metabolic reconstruction represents an attempt to formulate a model reconciling the sequence data with known biochemistry. This model goes beyond asserting what can be reliably deduced from the sequence data. It includes assertions that must be viewed as hypotheses to be tested. It also includes numerous assertions of pathways for which some enzymes have not yet been identified in the sequence data. Each such assertion is a judgment that must continually be reconsidered as more data becomes available.
• An accurate understanding of this organism will ultimately arise from many sources, and we believe that this effort is advanced by making the initial reconstruction publicly available, rather than waiting for the experimental evidence required to confirm or reject some of these conjectures. Indeed, one of the central roles of a metabolic reconstruction of the sort we present is to focus experimentation on specific questions of central importance.
• Many aspects of the metabolism cannot at this time be resolved. Questions relating to the roles of specific transport proteins, whether the Calvin cycle is actually present, and a number of other issues must remain open at this point.

2. The Environment of Methanococcus jannaschii

M. jannaschii strain JAL-1 was isolated from surface material collected at a "white smoker" chimney at a depth of 2600 meters in the East Pacific Rise near the western coast of Mexico [Jones et al., 1983]. Two similar strains were isolated from hydrothermally active sediments in the Guaymas Basin at a depth of 2000 meters [Zhao et al., 1988; Jones et al., 1989].

Cells of M. jannaschii are irregular cocci [Jones et al., 1983]. The cell envelope is composed of a cytoplasmic membrane and a protein surface layer [Nusser and Konig, 1987]. Polar bundles of flagella are also present. This morphology is common among the methanococci.

The characteristics of the source material for these isolates suggest that M. jannaschii possesses adaptations for growth at high temperature and pressure as well as moderate salinity. The water chemistry of the sites suggests an environment rich in sulfide, H₂, CO₂, Fe⁺², and Mn⁺² [Jannasch and Mottl, 1986]. This anaerobic environment would be well suited for a H₂-utilizing methanogen that reduces CO₂ to methane. Fixed nitrogen, either as NH₃ or NO₂^-, is not abundant [Baley, et al., 1984]. In addition, small amounts of CO are present. Thus, it is possible that CO could be used as an electron donor in place of H₂.

3. Methanogenesis

From its growth characteristics and what little is known about its biochemistry, M. jannaschii appears to be typical of H₂-utilizing, autotrophic methanogens. These archaea perform anaerobic respiration with CO₂ as the terminal electron acceptor according to the general equation

 
      4H2 + CO2 ---> CH4 + 2H2O.

So far, all methanogens isolated appear to be obligate methanogens and do not possess additional sources of energy capable of supporting growth. As expected, M. jannaschii does not grow in a rich heterotrophic medium in the absence of H₂ [Jones et al., 1983] and related methanococci do not metabolize glucose or most amino acids. However, the current evidence does not exclude alternative but minor pathways of energy metabolism. For instance, M. jannaschii produces glycogen as an intracellalar storage material [Konig et al., 1985]. Presumably, it also possesses the pathways to utilize this carbohydrate [Yu et al., 1994].

The pathway of methanogenesis from CO₂ is complex and requires five unique coenzymes: methanofuran, tetrahydromethanopterin (H₄MPT), coenzyme M (HS-CoM), 7-mercaptoheptanoylthreonine phosphate (HS-HTP), and coenzyme F₄₂₀ (for reviews, see [Thauer et al., 1993; Muller et al., 1993]). Simply, the pathway involves the stepwise reduction of CO₂ with H₂ as the ultimate electron donor. It contains three coupling sites to the proton motive force (PMF). In the first, the PMF is utilized to drive the endergonic reduction of CO₂ to the formyl level. The second and third coupling sites generate the PMF by coupling exergonic steps in CO₂ reduction to proton or sodium pumps. Each of the three coupled reactions is catalzed by a membrane protein complex. In addition, the methylreductosome is a large complex attached to the interior of the cytoplasmic membrane that contains at least one "soluble" enzyme of the pathway.

There is substantial sequence evidence for the existence of formate dehydrogenase, which suggests that this organism is capable of utilizing formate in place of H₂. This property is widespread among H₂-utilizing methanogens. Although M. jannaschii does not grow on formate, cell extracts appear to have the ability to oxidize formate, and a closely related isolate grows with formate [Jones et al., 1983, 1989].

Electron carriers for many of the reactions in methanogenesis are not known with certainty. It is likely that Fe/S proteins are utilized for many steps. For some reactions, coenzyme F₄₂₀, a deazaflavin that was discovered in methanogens but subsequently found in the bacteria, is utilized. Methanococci also contain NAD(P)H and flavins, although cytochromes and ubiquinone or menaquinone are believed to be absent. The proton motive force generated during methanogenesis is utilized for ATP synthesis, transport, motility, and other cellular functions. In the related archaeon Methanococcus voltae, the sodium motive force is probably the major component of the membrane potential [Jarrell and Sprott, 1985]. It is coupled to ATP synthesis by a Na⁺-translocating ATPase and to the proton gradient by a Na⁺/H⁺ antiporter [Dybas and Konisky, 1992; Carper and Lancaster, 1986; Chen and Konisky, 1993]. Similarly, transport is dependent on sodium [Dybas and Konisky, 1989; Ekiel et al., 1985]. Presumably, other bioenergetic processes in methanococci such as motility will prove to be coupled to the sodium motive force.

4. Carbohydrate Metabolism

M. jannaschii grows autotrophically and there is little evidence that it assimilates organic compounds. Thus, it must biosynthesize all its cellular components from CO₂. In the related methanogen Methanococcus maripaludis, CO₂ is assimilated via a modified Ljungdahl-Wood pathway of acetyl-CoA biosynthesis [Shieh and Whitman, 1988; Ladapo and Whitman, 1990]. In this pathway, the methyl carbon of acetyl-CoA is derived from methyl- H₄MPT, an intermediate in the pathway of methanogenesis. The carboxy carbon is derived from CO₂ via reduction to CO. These reactions are catalyzed by an enzyme complex named acetyl-CoA decarbonylase/synthase. Because the complex also oxidizes CO, it is sometimes called carbon monoxide dehydrogenase (EC 1.2.99.2). Both of its subunits were identified in M. jannaschii.

Glycogen Metabolism

The following enzymes participating in metabolism of glycogen were found in the sequence data: glycogen synthetase (EC 2.4.1.11), glycogen phosphorylase (EC 2.4.1.1), UDPglucose pyrophosphorylase (EC 2.7.7.9), and phosphoglucomutase (EC 5.4.2.2)

Although we could not locate the glycogen branching (EC 2.4.1.18) and debranching (EC 3.2.1.33/2.4.1.25) enzymes, which are required to support glycogen metabolism, we believe that further analysis will locate these enzymes in the genome.

Entries in the tables of assignments that have no sequence represent enzymes for which no sequence is available for any organism. Since our assignments of function are based on similarity to known, characterized sequences, no attempt could be made to locate sequences within M. jannaschii corresponding to these functions. On the other hand, enzymes characterized as missing (which occur in tables below) represent functions for which representative sequences do exist in the databases.

Embden-Meyerhof pathway

Six of nine enzymes of the Embden-Meyerhof pathway (EMP) catabolizing glucose-6-phosphate to pyruvate and lactate were found in the sequence data, although three important enzymes of glycolysis (6-phosphofructokinase (EC 2.7.1.11 or EC 2.7.1.90), fructose-bisphosphate aldolase (EC 4.1.2.13), and phosphoglycerate mutase (EC 5.4.2.1)) have not been located. A glucokinase (EC 2.7.1.2 or EC 2.7.1.63), which phosphorylates glucose at the expense of ATP or polyphosphate, has not been identified. However, this enzyme would not be required if glycogen was the major carbohydrate metabolized. Recent results [Kengen et al., 1994; Kengen et al., 1995] show that P. furiosus uses novel ADP-dependent (AMP-forming) forms of glucokinase and 6-phosphofructokinase. The ADP-dependent versions appear more appropriate to high-temperature environments. This is a most remarkable development and strongly suggests that a similar situation may exist in M. jannaschii. We suspect that the divergence of these two enzymes from the more common forms is substantial enough to make detection difficult.

Although we cannot yet verify the existence of ADP-dependent versions of these key enzymes, we believe that the possible implications are worth considering, should their presence be confirmed. In the more common versions of glycolysis, the ADP generated by the early stages is immediately phosphorylated in the later steps. If, instead, AMP is produced from ADP, recycling AMP becomes an issue. The most probable way of recycling AMP uses adenylate kinase (EC 2.7.4.3):

	AMP + ATP <-> 2 ADP

The adenylate kinase reaction here is far from equilibrium: to maintain stationary turnover of AMP, it must have a velocity twice as high as the glucose consumption rate. Therefore, we expect the adenylate kinase found in this organism to have a high affinity for AMP and ATP and a very high specific activity with respect to glucokinase and 6-phosphofructokinase.

We have found solid sequence evidence in favor of NAD-dependent GAP dehydrogenase (EC 1.2.1.12). It must be noted that NADP-dependent GAP dehydrogenase, as well as an ATP-dependent version of 6-phosphofructokinase (EC 2.7.1.11) have been reported in M. maripaludis [Yu et al., 1994]. These differences may reflect the considerable evolutionary distance that separates the mesophilic and hyperthermophilic methanococci. The presence or absence of the NADP-dependent GAP dehydrogenase is an issue that directly relates to the presence or absence of the oxidative portion of the pentose-phosphate shunt (see below).

Phosphonopyruvate decarboxylase (EC 4.1.1.-) potentially links glycolysis with a largely unknown metabolism of phosphonates.

No enzymes involved in the nonphosphorylated Entner-Doudoroff pathway were detected in the sequence data. This result agrees with the known biochemical evidence [Yu et al., 1994; Kengen et al., 1995].

Gluconeogenesis

Hexoses are made by gluconeogenesis [Fuchs et al., 1983]. Phosphoenolpyruvate biosynthesis for gluconeogenesis is catalyzed by pyruvate, water dikinase. Seven of nine enzymes of this pathway have been reliably identified. The three that have not are the phosphoglycerate mutase and aldolase, mentioned above, and fructose-bisphosphatase (EC 3.1.3.11). All the enzyme activities of the pathway have also been detected in M. maripalidus [Shieh et al., 1987; Yu et al., 1994).

Reductive TCA

Biochemical evidence strongly supports the hypothesis that the reductive branch of the tricarboxylic acid cycle is utilized to make 2-oxoglutarate and glutamate from oxaloacetate [Shieh and Whitman, 1987; Sprott et al., 1993]. We were able to locate four of the five required enzymes (EC 1.1.1.37/1.1.1.82, EC 4.2.1.2, EC 1.3.99.1, and EC 6.2.1.5); the sequence of the fifth, 2-oxoglutarate synthase (EC 1.2.7.3), has not yet been identified in any organism. The alternative would require the existence of a portion of the oxidative TCA cycle. We doubt the presence of the three enzymes from the oxidative portion of the cycle leading to 2-oxoglutarate (citrate synthase, aconitase, and isocitrate dehydrogenase), although both the aconitase and isocitrate dehydrogenase were listed in Bult et al., 1996. The similarities between MJ1596 and MJ0720 and known versions of both isocitrate dehydrogenase and isopropylmalate dehydrogenase (EC 1.1.1.85, which is used in leucine biosynthesis) are very strong. MJ0499 is very similar to 3-isopropylmalate dehydratase (EC 4.2.1.33, which also is utilized in leucine biosynthesis) and less so to aconitase.

We have found membrane-bound, Na-dependent oxaloacetate decarboxylase (EC 4.1.1.3), which converts pyruvate into oxaloacetate. Pyruvate is formed by reductive carboxylation of acetyl-CoA catalyzed by pyruvate oxidoreductase (EC 1.2.7.1) [Shieh and Whitman, 1987] or by the glycolytic system. Based upon N-terminal sequence information for the pyruvate oxidoreductase from M. maripalidus [Yang and Whitman, unpublished data], four genes encoding subunits of the pyruvate oxidoreductase (EC 1.2.7.1) have been identified.

Pentose Biosynthesis

Two pathways have been proposed for pentose biosynthesis in the methanococci. In one proposal, a nonoxidative pathway is composed of transketolase, transaldolase, ribose-5-phosphate epimerase, and ribulose-5- phosphate isomerase [Yu et al., 1994]. In the second proposal, erythrose-4-phosphate is formed via carboxylation of dihydroxyacetone phosphate instead of transketolase [Choquet et al., 1994b].

Sequence analysis has identified genes that encode enzymes of the nonoxidative pentose-phosphate shunt; they are used to produce ribose-phosphate for nucleotide biosynthesis. The two dehydrogenases (EC 1.1.1.49 and EC 1.1.1.44) required for the oxidative part of the shunt have not yet been found and thir activities are not detectable in M. maripalidus [Yu et al., 1994]. Isotope labeling of M. jannaschii provides additional evidence, that the oxidative pentose phosphate pathway is absent [Sprott et al., 1993].

CO₂ fixation

The large subunit of RuBisCo (EC 4.1.1.39) has been identified, which raises the question "Is the entire Calvin Cycle actually present?" The answer to this question will hinge on whether phosphoribulokinase (EC 2.7.1.19) is present; it has not yet been identified.

One conjecture is that the phosphoribulokinase, which is normally a two-subunit enzyme (neither subunit of which has been located), might be ADP-dependent. Such coenzyme substitutions have been proposed in the glycolytic pathway, and they often make recognition of the enzyme from sequence data difficult or impossible. Another possibility is that protein MJ1235 is only paralogous to RuBisCo and has a different metabolic function yet to be identified.

Inositol Metabolism

Di-myo-inositol-1,1-phosphate (DIP) is an abundant osmolyte in M. igneus, a hyperthermophile related to M. jannaschii [Ciulla et al., 1994]. A gene encoding one of the enzymes necessary for inositol biosynthesis from glucose-6-phosphate has been found.

Other Pathways of Carbohydrate Metabolism

The presence of glycerol dehydrogenase (EC 1.1.1.6) appears clear. This would imply the presence of glycerone kinase (EC 2.7.1.29), since the only apparent way to consume glycerone is by conversion to glycerone-phosphate (a glycolytic intermediate).

It was believed until now that methanogenic archaea known to accumulate glycogen do not synthesize cyclic 2,3-biphospho-glycerate [Konig et al. 1985]. Nevertheless, in this organism both storage mechanisms seem to exist, since 2-phospho-glycerate-kinase (EC 2.7.2.-) has been clearly identified.

5. Amino Acids and Polyamine Metabolism

On the basis of labeling and enzymatic data, the biosynthesis of the most amino acids, nucleosides, and hexoses in methanogens appear to occur by pathways common in bacteria (for a review see Simpson and Whitman, 1993). Some noteworthy features are described below. Nearly all of the biosynthetic pathways for amino acids (including selenocysteine) have been detected, although a few of the required enzymes have not yet been found. The one main exception is the biosysthesis of cysteine; we have been unable to locate the enzymes of cysteine biosynthesis.

All of the enzymes involved in the common biosynthetic pathway leading from aspartate to diaminopimelate and then to lysine and methionine (including the multifunctional enzyme aspartokinase I (EC 2.7.2.4)/homoserine dehydrogenase I(EC 1.1.1.3)) were found. The identified methionine synthase (EC 2.1.1.14) has a high similarity score to a known cobalamine-independent counterpart in M. thermoautotrophicum [Vaupel,M., 1996]. This enzyme catalyzes the synthesis of methionine from homocysteine using (we believe) 5-methyl-tetrahydromethanopterin, rather than 5-methylmethyltetrahydrofolate, as the donor of the required methyl group. Lysine is made by the diaminopimelic pathway. There is biochemical data that in Methanobacteria isoleucine is synthetized from pyruvate and acetyl-CoA via the citramalate pathway [Eikmanns 1983; Ekiel et al., 1984]. Enzymes participating in the citramalate pathway have not been sequenced in any organism yet, so it is impossible to confirm it's existence in M. jannascii from the sequence data. All enzymes of arginine biosynthesis via ornithine carbamoyltransferase were found, which agrees with [Meile, L. and Leisinger 1984]. Other amino acids appear to be derived using well-known common pathways [Ekiel et al., 1983]. It is likely that polyamine biosynthesis begins with spermidine synthase (EC 2.5.1.16), which has been located.

6. Nucleotide Metabolism

Although pyrimidines and purines appear to be derived from common pathways [Choquet et al., 1994a] , C1 groups may be also contributed from methanogenesis [Ekiel et al., 1983]. The entire set of reactions for interconversions between nucleotides and their reduced forms listed below is present in M. jannaschii. This organism uses anaerobic nucleoside-triphosphate reductase (probably B12-dependent) to generate the precursors of DNA. Both thioredoxin and glutaredoxin are present and could be used by the reductase.

7. Lipid Metabolism

Like other archaea, M. jannaschii contains isoprenoid-based ether lipids (for a review see Koga et al., 1993). In addition to the common archaeol (2,3-di-O-phytanyl-sn-glycerol diether) and caldarchaeol (2,2',3,3'-diphytanyl-sn-diglycerol tetraether), M. jannaschii contains a unique macrocyclic diether (2,3-di-o- cyclic-biphytanyl-sn-glycerol). The polar lipids contain phosphoethanolamino, 6-(aminoethylphospho)glucosyl, glucosyl, and gentiobiosyl residues. Mevalonate is a precusor for the isoprenoid groups, as expected from common pathways [Sprott et al., 1993).

M. jannaschii must have the whole set of enzymes required to generate membrane lipids from glycolytic intermediates. However, since few sequences exist for this metabolism, few similarities were detected, and very little can be inferred directly from the sequence data. Even so, the key enzymes from the mevalonate pathway (EC 1.1.1.34 and 2.7.1.36) can be clearly recognized; this is the central pathway of archaeal lipid de novo biosynthesis. The end product of this pathway is isopentenyl pyrophosphate, which must be polymerized to forms of prenyl-pyrophosphates. We have located the trifunctional protein that polymerizes the isopentenyl-pyrophosphate to geranylgeranyl pyrophosphate and farnesyl pyrophosphate (EC 2.5.1.10, EC 2.5.1.29 and EC 2.5.1.1). The fatty-acid synthase complex, which occurs in both bacteria and eukaryotes, is absent.

The reliable identification of UDP-N-acetylglucosamine--dolichyl-phosphate-N-acetylglucosaminephosphotransferase indicates that dolichol biosynthesis from farnesyl-diphosphate is also present. The presence of acetyl-CoA carboxylase indicates that malonyl-CoA is probably used as a building block for complex lipids. We were able to reliably identify only a few enzymes dealing with metabolism of phospholipids. In particular, CDP-diacylglycerol--serine O-phosphatidyltransferase, omega-3 fatty acid desaturase, and phospholipase C were identified.

Metabolism of Coenzymes and Prosthetic Groups

As was mentioned above, methanogens have a unique set of the coenzymes, including methanofuran, tetrahydromethanopterin (H₄MPT), coenzyme M (HS-CoM), 7-mercaptoheptanoylthreonine phosphate (HS-HTP), and coenzyme F₄₃₀ (for reviews, see [DiMarco et al., 1990]). Methanogenes also use a number of familiar coenzymes and cofactors participating in various metabolic processes (for a review, see [Jones et al., 1989]), such as thiamine, riboflavin, pyridoxine, cobamides, biotin, niacin, and panthotenate.

The autotrophic nature of M. jannaschii implies its capability to synthesize all coenzymes and prosthetic groups required for its metabolism. In many cases, however, the enzymes involved in these biosyntheses have not been thorough characterized in any organism. We found at least partial evidence for genes encoding the biosynthesis of the following enzymes: methanopterin, NAD, cobalamine, riboflavin, FMN, FAD, thiamine pyrophosphate, and biotin.

Like those for thiamine, niacin, and panthotenate, we believe that the M. jannaschii counterparts of some biosynthetic enzymes either have diverged too far from the bacterial or eukaryotic versions to be recognizable or are analogs, but not homologs, of them.

Biochemical evidence indicates that folic-acid levels are extremely low in methanogens [Leigh, 1983] and that tetrahydrofolate coenzymes are probably not present [Purwantini et al. 1996]. Our analysis of the sequence data also indicates no presence of enzymes using these coenzymes.

Some evidence exists that lipoic acid occurs in archaea [Noll and Barber, 1988]. Its main function is as a prosthetic group within the pyruvate dehydrogenase complex and the 2-oxoglutarate dehydrogenase complex. Of the five enzymes normally involved in these complexes, only the lipoate dehydrogenase (EC 1.8.1.4) has been located. This result leads to a puzzling situation in which there seems no apparent physiological function for lipoic acid, but the mechanism for reoxidizing it appears to exist. However, it has been recently shown [Bunik V, and H. Follman, 1993] that lipoate dehydrogenase can also use thioredoxin as a substrate, and thioredoxin may play a significant role in M. jannaschii.

8. Enzymatic Activities Coupled to Oxidation or Reduction of F₄₂₀

F₄₂₀ can act as a replacement for ferredoxin in some methanogens, including Methanococcus jannaschii. It functions as a low-potential two-electron acceptor. The following table summarizes the instances in which enzymatic activities using F₄₂₀ were detected:

9. Membrane Transport

Like many autotrophic methanogens, M. jannaschii has a limited capacity to assimilate organic molecules [Sprott et al., 1993]. Compounds assimilated well include leucine, isoleucine, phenylalanine, formate, pyruvate and malate. Compounds assimilated poorly include mevalonate, glycerol, and lysine. Compounds assimilated in very low amounts or not at all include serine, aspartate, citrate, glucose, and acetate. The inability to assimilate acetate is unusual for methanogens, and acetate kinase, phoshotransacetylase, and acetyl coenzyme A synthetase activities are not detectable.

Sequence data reveals a wide spectrum of membrane transport proteins, the substrates for which have not yet been identified.

Membrane transport is driven by both ATP-dependent and osmotic-potential-based mechanisms. The proton motive force is generated during methanogenesis and drives a classical H-ATPase (EC 3.6.1.34) for ATP biosynthesis; the key subunits have been reliably identified. A second H-ATPase (EC 3.6.1.35), more typical of plants and fungi is also present.

Summary

The interpretation of the Methanococcus jannaschii genome will inevitably require many years of effort. This initial attempt to connect the sequence data to aspects of known biochemistry and to provide an overview of what is already apparent from the sequence data will be refined.

Numerous issues remain that can be resolved only by direct biochemical analysis. Let us draw the reader's attention to just a few that might be considered central:

• We are still missing key enzymes from the glycolytic pathway, and the conjecture is that this is due to ADP-dependency. The existence of glycolytic activity in the cell-free extract should be tested.
• The issue of whether the Calvin cycle is present needs to be examined.
• We need to determine whether the 2-oxoglutarate synthase (ferredoxin-dependent) (EC 1.2.7.3) activity is present.
• The issue of whether cyclic 2,3-bisphosphate is detectable in the cell-free extracts needs to be checked. If it is, this result would confirm our assertion of the two pathways controlling synthesis and degradation of cyclic 2,3-bisphosphate.

We will provide the current metabolic reconstruction, which will be updated as new interpretations and data emerge, via the WIT system, which is a Web application that can be accessed via the URL

	http://www.cme.msu.edu/WIT/

Our sincere hope is that others will find this initial model useful and will forward criticisms, corrections, and updates to Evgeni Selkov at the e-mail address Evgeni@mcs.anl.gov.

References

Baley, N.,R. Sparling, and L. Daniels. 1984. Dinitrogen fixation by a thermophilic methanogenic bacterium. Nature 312: 286-288.

Bult, Carol J., et al. 1996. Complete genome sequence of the methanogenic archeon Methanococcus jannaschii. Science 273: 1058-1073

Bunik V, and H. Follman. 1993. Thioredoxin reduction dependent on alpha-ketoacid oxidation by alpha-ketoacid dehydrogenase complexes. FEBS Lett 336(2): 197-200

Carper, S. W., and J. R. Lancaster, Jr. 1986. An electrogenic sodium-translocating ATPase in Methanococcus voltae. FEBS Lett. 200: 177-180.

Chen, W., and J. Konisky. 1993. Characterization of a membrane- associated ATPase from Methanococcus voltae, a methanogenic member of the archaea. J. Bacteriol. 175: 5677-5682.

Choquet, C. G., J. C. Richards, G. B. Patel, and G. D. Sprott. 1994a. Purine and pyrimidine biosynthesis in methanogenic bacteria. Arch. Microbiol. 161: 471-480.

Choquet, C. G., J. C. Richards, G. B. Patel, and G. D. Sprott. 1994b. Ribose biosynthesis in methanogenic bacteria. Arch. Microbiol. 161: 481-488.

Ciulla, R.A., S Burggraf, K.O., Steller, and M.F. Roberts. 1994. Occurence and role of di-myo-inositol-1,1-phosphate in M. igneus. Appl.Environ.Microbiol. 60: 3660-3664.

DiMarco, A.A., T.A. Babik, and R.S. Wolfe. 1990. Unusual coenzymes of methanogenesis. Annu. Rev. Biochem., 59: 355-394.

Dybas, M., and J. Konisky. 1992. Energy transduction in the methanogen Methanococcus voltae is based on a sodium current. J. Bacteriol. 174: 5575-5583.

Ekiel, I., K. F. Jarrell, and G. D. Sprott. 1985. Amino acid biosynthesis and sodium-dependent transport in Methanococcus voltae, as revealed by 13C NMR. Eur. J. Biochem. 149: 437- 444.

Ekiel, I., I.C.P. Smith, and G. D. Sprott. 1983. Biosynthetic pathways in Methanospirillum hungatei as determined by 13C nuclear magnetic resonance. J. Bacteriol. 156:316-326.

Ekiel, I., I.C.P. Smith and G. D. Sprott. 1984. Biosynthesis of isoleucine in methanogenic bacteria: a 13C NMR study. Biochemistry 23: 1683-1687.

Eikmanns, B., D. Linder, and R. K. Thauer. 1983. Unusual pathway of isoleucine biosynthesis in Methanobacterium thermoautotrophicum. Arch Microbiol 136: 111-113

Fleischmann, R. D., et al. 1995. Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 169: 496-512

Fox, Jeffrey L. 1996. Another genome, another kingdom, another set of surprises. Nature Biotechnology 14: 1211

Fraser, Claire M., et al. 1995. The minimal gene complement of Mycoplasma genitalium. Science 270: 397-403.

Fuchs, G., Winter, H., Steiner, I. and Stupperich, E.. 1983. Enzymes of gluconeogenesis in the autotroph Methanobacterium thermoautotrophicum. Arch. Microbiology 136: 160-162

Gray, Michael W.. 1996. The third form of life. Nature 383: 299-300

Jannasch, H. W., and M. J. Mottl. 1986. Geomicrobiology of deep-sea hydrothermal vents. Science 229: 717-725.

Jarrell, K. F., and G. D. Sprott. 1985. Importance of sodium to the bioenergetic properties of Methanococcus voltae. Can. J. Microbiol. 31: 851-855.

Jones, W. J., J. A. Leigh, F. Mayer, C. R. Woese, and R. S. Wolfe. 1983. Methanococcus jannaschii sp. nov., an extremely thermophilic methanogen from a submarine hydrothermal vent. Arch. Microbiol. 136: 254-261.

Jones, W. J., C. E. Stugard, and H. W. Jannasch. 1989. Comparison of thermophilic methanogens from submarine hydrothermal vents. Arch. Microbiol. 151: 314-318.

Kengen, S.W.M., F.A.M. de Bok, N. van Loo, C. Dijkema, A.J.M. Stams, W.M. de Vos. 1994. Evidence for the operation of a novel Embden-Meyerhof pathway that involves ADP-dependent kinases during sugar fermentation by Pyrococcus furiosus. J. Biol. Chem. 269: 17537-17541.

Kengen, S.W.M., J.E. Tuininga, F.A.M. de Bok, A.J.M. Stams, W.M. de Vos. 1995. Purification and characterization of a novel ADP-dependent glucokinase from the hyperthermophilic archaeon Pyrococcus furiosus. J. Biol. Chem. 270: 30453-30457.

Koga, Y., M. Nishihara, H. Morii, and M. Akagawa-Matsushita. 1993. Ether polar lipids of methanogenic bacteria: structures, comparative aspects, and biosynthesis. Microbiol. Rev. 57: 164-182.

Konig, H., E. Nusser, and K. O. Stetter. 1985. Glycogen in methanolobus and methanococcus. FEMS Microbiol. Lett. 28: 265-269.

Ladapo, J., and W. B. Whitman. 1990. Method for isolation of auxotrophs in the methanogenic archaebacteria: Role of the acetyl-CoA pathway of autotrophic CO₂ fixation in Methanococcus maripaludis. Proc. Natl. Acad. Sci. U.S.A. 87: 5598-5602.

Leigh J. A. 1983. Levels of water-soluble vitamins in methanogenic and non-methanogenic bacteria. Appl. Environ. Microbiol. 45: 800-803.

Meile, L., and T. Leisinger. 1984. Enzymes of arginine biosynthesis in methanogenic bacteria. Experientia 40:899-900.

Muller, V., M. Blaut, and G. Gottschalk. 1993. Bioenergetics of methanogenesis, p. 360-406. In J. G. Ferry (ed.), Methanogenesis, Chapman & Hall, New York.

Noll, Kenneth M. and Barber, Ted. S. 1988. Vitamin Contents of Archaebacteria J. Bacter. vol 170, no. 9: 4315-4321

Nusser, E., and H. Konig. 1987. S layer studies on three species of Methanococcus living at different temperatures. Can. J. Microbiol. 33: 256-261.

Purwantini, E. and Daniels, L. 1996. Purification of a novel coenzyme F₄₂₀-dependent glucose-6-phosphate dehydrogenase from Mycobacterium smegmatis. J. Bacteriol, vol. 178: 2861-2866

Shieh, J. S., and W. B. Whitman. 1987. Pathway of acetate assimilation in autotrophic and heterotrophic methanococci. J. Bacteriol. 169: 5327-5329.

Shieh, J. S., and W. B. Whitman. 1988. Autotrophic acetyl coenzyme A biosynthesis in Methanococcus maripaludis. J. Bacteriol. 170: 3072-3079.

Simpson, P. G., and W. B. Whitman. 1993. Anabolic pathways in methanogens, pp. 445-472. In J. G. Ferry (ed.), Methanogenesis, Chapman & Hall, New York.

Sprott, G. D., I. Ekiel, and G. B. Patel. 1993. Metabolic pathways in Methanococcus jannaschii and other methanogenic bacteria. Appl. Environ. Microbiol. 59: 1092-1098.

Thauer, R. K., R. Hedderich, and R. Fischer. 1993. Reactions and enzymes involved in methanogenesis from CO₂ and H₂, p. 209- 252. In J. G. Ferry (ed.), Methanogenesis, Chapman & Hall, New York.

Vaupel,M., Dietz,H., Linder,D. and Thauer,R.K. Primary structure of cyclohydrolase (Mch) from Methanobacterium thermoautotrophicum (strain Marburg) and functional expression of the mch gene in Escherichia coli Eur. J. Biochem. 236 (1), 294-300 (1996)

Yu, J.-P., J. Ladapo, and W. B. Whitman. 1994. Pathway of glycogen metabolism in Methanococcus maripaludis. J. Bacteriol. 176: 325-332.

Zhao, H., A. G. Wood, F. Widdel, and M. P. Bryant. 1988. An extremely thermophilic Methanococcus from a deep sea hydrothermal vent and its plasmid. Arch. Microbiol. 150: 178-183.