Methylotrophy in Methylobacterium extorquens AM1 from a Genomic Point of View

Methylotrophy is defined as the ability to “grow at the expense of reduced carbon compounds containing one or more carbon atoms but containing no carbon-carbon bonds” (3). It is an intriguing example of microbial metabolic agility, with the use of a class of chemicals disregarded by the majority of organisms. Even though the ability to grow methylotrophically was first discovered in the early 1900s (cited in reference 3), it was not until the 1960s to 1970s that an understanding of the biochemical nature of this capability started to emerge. Fascination with methylotrophy in those years was fueled by the commercial interest in single-cell protein production, and as a result, the specific details of the biochemistry of methylotrophy began to be revealed. Enzymes for the primary oxidation of C₁ substrates such as methanol dehydrogenase and methylamine dehydrogenase were characterized, and distinct modes of C₁ assimilation, such as the ribulose monophosphate cycle and the serine cycle were discovered. The biochemical processes involved in methylotrophy that were known by 1982 are described in detail in the now classic book Biochemistry of Methylotrophs by Christopher Anthony (3). In the 20 years following the publication of Biochemistry of Methylotrophs, a few additional methylotrophy biochemical pathways have been discovered, such as the pathway for C₁ transfer linked to methanopterin and methanofuran, which solved the long-standing mystery of formaldehyde oxidation in many methylotrophs (15, 53), and novel pathways for primary C₁ oxidation, such as the pathways for degradation of chlorinated methanes and methanesulfonic acid (21, 50).

The knowledge concerning the biochemistry and physiology of methylotrophic organisms accumulated over the past three decades suggests a new framework for understanding methylotrophy as a novel metabolic mode. In this framework, methylotrophy is envisioned as a set of specific metabolic functional modules, with different combinations of such modules being present in different methylotrophs (Fig. 1 for the methylotrophic metabolic modules in Methylobacterium extorquens AM1). However, until recently, a number of important details of these modules were missing, and so the picture remained incomplete. The availability of two unfinished genome sequences for the important model organisms M. extorquens AM1 (http://www.integratedgenomics.com/genomereleases.html#list6)and Methylococcus capsulatus Bath (http://tigrblast.tigr.org/ufmg/) is transforming our understanding of methylotrophy. Annotation of these two genomes combined with functional analysis will delineate the set of genes and functions that is both sufficient and necessary to define a methylotroph. Expanding genomic analyses to include other groups of methylotrophs will in turn provide clues to the origins of methylotrophy and the evolution of various methylotrophic pathways. In this publication, we summarize the existing knowledge of the genes involved in methylotrophic pathways in M. extorquens AM1, analyze its yet unfinished genome with respect to location and clustering of methylotrophy genes, and present a comprehensive list of methylotrophy genes and enzymes known at this time in M. extorquens AM1 (Table 1).

FIG. 1.

Methylotrophy metabolic modules in M. extorquens AM1. Known genes are in italic. For simplicity, redox reactions in the assimilatory pathways are not indicated. For details, refer to the references given in Table 1.

TABLE 1.

Methylotrophy genes in M. extorquens AM1

Gene	Function	Cluster	Essential for C1 growtha	Metabolic moduleb	Previous name	Accession no.	Source or reference
Primary oxidation
mxaW	Unknown	1	−	MOX	moxW	M31108	41
mxaF	Methanol dehydrogenase large subunit	1	+	MOX	moxF	M31108	1, 2, 26
mxaJ	Unknown	1	+	MOX	moxJ	M31108	1, 2, 26
mxaG	Cytochrome c1	1	+	MOX	moxG	X07856	1, 2, 26, 30
mxaI	Methanol dehydrogenase small subunit	1	+	MOX	moxI	X15792	1, 26, 30, 31
mxaR	Unknown	1	+	MOX	moxR	Y07864	1, 26, 44
mxaS	Unknown	1	+	MOX	moxS	L41608	1, 26, 44
mxaA	Essential for Ca2+ insertion into MDH	1	+	MOX	moxA	L41608	1, 26, 29
mxaC	Essential for Ca2+ insertion into MDH	1	+	MOX	moxC	L41608	1, 26, 29, 44
mxaK	Essential for Ca2+ insertion into MDH	1	+	MOX	moxK	L41608	1, 26, 29
mxaL	Essential for Ca2+ insertion into MDH	1	+	MOX	moxL	L41608	1, 26, 29
mxaD	Essential for Ca2+ insertion into MDH	1	−	MOX		L41608	1, 26, 29, 44
mxaE	Unknown	1	+	MOX		AF017434	41
mxaH	Unknown	1	NM	MOX		AF017434	41
mxaB	Transcriptional regulator	1	+	MOX	moxB	AF017434	41
mxbD	Sensor kinase	2	+	MOX	moxD	AF032114	42
mxbM	Transcriptional regulator	2	+	MOX	moxM	AF032114	42
pqqA	PQQ synthesis	2	−	PQQ-S	pqqD, moxT	L25889	26, 45, 46
pqqB	PQQ synthesis	2	+	PQQ-S	pqqG, moxO	U72662	26, 45
pqqC/D	PQQ synthesis	2	+	PQQ-S	pqqB, pqqC, moxV, moxT	L25889/U72662	26, 45
pqqE	PQQ synthesis	2	+	PQQ-S	pqqA, moxC	U72662	26, 45
mxcQ	Sensor kinase	3	+	MOX	moxQ		26, 42
mxcE	Transcriptional regulator	3	+	MOX	moxE		26, 42
pqqF	PQQ synthesis	4	+	PQQ-S	pqqE, moxH	L43135	26, 43
pqqG	PQQ synthesis	4	+	PQQ-S	pqqF, moxU	L43135	26, 43
mauF	Unknown	5	+	MAU		L26406	5
mauB	Methylamine dehydrogenase large subunit	5	+	MAU		L26406	5
mauE	Essential for small subunit maturation	5	+	MAU		L26406	5
mauD	Essential for small subunit maturation	5	+	MAU		L26406	5
mauA	Methylamine dehydrogenase small subunit	5	+	MAU		L26406	5
mauC	Amicyanin	5	+	MAU		L26406	5
mauJ	Unknown	5	−	MAU		L26406	5
mauG	Unknown	5	+	MAU		L26406	5
mauI	Unknown	5	+	MAU		L26406	5
mauM	Unknown	5	−	MAU		L26406	5
mauN	Unknown	5	−	MAU		L26406	5
Formate oxidation
fdh1A	Tungsten-dependent formate dehydrogenase, α subunit	6	−	FOX1		AF489516	25
fdh1B	Tungsten-dependent formate dehydrogenase, β subunit	6	−	FOX1		AF489516	25
fdh2A	Molybdenum-dependent formate dehydrogenase, α subunit	7	−	FOX2		AY183757	Unpublished
fdh2B	Molybdenum-dependent formate dehydrogenase, β subunit	7	−	FOX2		AY183757	Unpublished
fdh2C	Molybdenum-dependent formate dehydrogenase, γ subunit	7	−	FOX2		AY183757	Unpublished
fdh2D	Molybdenum-dependent formate dehydrogenase, δ subunit	7	−	FOX2		AY183757	Unpublished
fdh3A	Cytochrome-linked formate dehydrogenase, α subunit	8	−	FOX3		AY181035	Unpublished
fdh3B	Cytochrome-linked formate dehydrogenase, β subunit	8	−	FOX3		AY181035	Unpublished
fdh3C	Cytochrome-linked formate dehydrogenase, γ subunit	8	−	FOX3		AY181035	Unpublished
C1 transfer
dmrA	Dihydromethanopterin reductase	9	+	H4MPT-T		AY093431	28
fae	Formaldehyde activating enzyme	2	+	H4MPT-T	orf18	AF032114	52
mtdB	Methylene-H4MPT dehydrogenase	2	+	H4MPT-T	orfX	AF032114	15, 18
mch	Methenyl-H4MPT cyclohydrolase	2	+	H4MPT-T	orfZ	AF032114	15, 36
fhcA	Formyltransferase/hydrolase complex, α subunit	2	+	H4MPT-T	orf1	AF032114	15, 34, 35
fhcB	Formyltransferase/hydrolase complex, β subunit	2	+	H4MPT-T	orf2	AF032114	15, 34, 35
fhcC	Formyltransferase/hydrolase complex, γ subunit	2	+	H4MPT-T	orf3	AF032114	15, 34, 35
fhcD	Formyltransferase/hydrolase complex, δ subunit	2	+	H4MPT-T	ffsA	AF032114	15, 34, 35
orf4	H4MPT biosynthesis	2	+	H4MPT-T		AF032114	15, 38
orf5	Unknown	2	−	H4MPT-T		AF032114	15
orf7	Cofactor biosynthesis	2	+	H4MPT-T		AF032114	15
orf9	Cofactor biosynthesis	2	+	H4MPT-T		AF032114	15
orf17	Cofactor biosynthesis	2	+	H4MPT-T		AF032114	15
orf19	Cofactor biosynthesis	2	+	H4MPT-T		AY117134	Unpublished
orf20	Cofactor biosynthesis	2	+	H4MPT-T		AY117134	Unpublished
orf21	Cofactor biosynthesis	2	+	H4MPT-T		AY117134	Unpublished
orf22	Cofactor biosynthesis	2	+	H4MPT-T		AY117134	Unpublished
orfY	Unknown	2	−	H4MPT-T		AF032114	15
mtdA	Methylene-H4MPT/Methylene-H4F dehydrogenase	2	+	H4F-T/H4MPT-T		L27235	9, 16, 17, 51
fch	Methenyl-H4F cyclohydrolase	2	+	H4F-T	orf4	L33465	36
ftfL	Formyl-H4F ligase	10	+	H4F-T			28
folK	2-Amino-4-hydroxy-6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase	2	E	H4F-S	folA	U72662	13
folB	Dihydroneopterin aldolase	2	E	H4F-S	folB	U72662	13
folP	Dihydropteroate synthase	2	E	H4F-S	folC	U72662	13
folE	GTP cyclohydrolase	11	E	H4F-S		AY183759	Unpublished
folC	Dihydrofolate synthase	12	E	H4F-S		AY183760	Unpublished
dyr	Dihydrofolate reductase	13	NM	H4F-S	dfrA	AY093432	28
C1 assimilation
sga	Serine glyoxylate aminotransferase	2	+	QSC	sgaA	L27235	9
hpr	Hydroxupyruvate reductase	2	+	QSC	hprA	L27235	7, 8, 9
ppc	PEP carboxylase	2	+	QSC	ppcA	U72662	4
mtkA	Malate thiokinase, α subunit	2	+	QSC/GRC		L33465	10
mtkB	Malate thiokinase, β subunit	2	+	QSC/GRC		L33465	10
mcl	Malyl-CoA lyase	2	+	QSC/GRC	mclA	U72662	4, 13
glyA	Serine hydroxymethyltrasferase	14	+	QSC		L33463	11
gck	Glycerate kinase	15	+	QSC	gckA	U87316	14
eno	Enolase	16	E	QSC/GNG		AY181039	Unpublished
mdh	Malate dehydrogenase	17	E	QSC/TCAC		AY181040	Unpublished
qscR	Serine cycle transcriptional regulator	18	+	QSC		AF516903	20
phaA	β-ketothiolase	19	+	GRC/PHBC		AF287907	22, 23, 24
phaB	Acetoacetyl-CoA reductase	19	+	GRC/PHBC		AF287907	22, 23, 24
phaR	Regulator of PHB synthesis and acetyl-CoA flux	19	+	GRC/PHBC	orf3	AF287907	23, 24
croR	Crotonase	20	+	GRC		AF416777	22
pccA	Propionyl-CoA carboxylase, α subunit	21	+	GRC		AY181038	22
pccB	Propionyl-CoA carboxylase, β subunit	22	+	GRC	pccA	L48340	12, 22
crr	Crotonyl-CoA reductase	22	+	GRC	adhA	L48340	12, 22
meaA	A mutase	22	+	GRC		U28335	12, 22, 40
meaB	Essential for methylmalonyl-CoA mutase reaction	22	+	GRC		AF416776	22
meaC	Unknown	23	+	GRC			28
meaD	Unknown	24	+	GRC			28
ibd2	Isobutyryl-CoA dehydrogenase	25	+	GRC		AY054980	22
mcmA	Methylmalonyl-CoA mutase, α subunit	26	+	GRC		AY181037	22
mcmB	Methylmalonyl-CoA mutase, β subunit	27	+	GRC			22
epm	Epimerase	28	+	GRC		AY183758	Unpublished
sdhA	Succinate dehydrogenase flavoprotein subunit	29	EG	GRC/TCAC		AY181036	48, 49
sdhB	Succinate dehydrogenase iron-sulfur binding subunit	29	NM	GRC/TCAC		AY181036	48, 49
sdhC	Succinate dehydrogenase cytochrome B556 subunit	29	NM	GRC/TCAC		AY181036	48, 49
sdhD	Succinate dehydrogenase membrane anchor subunit	29	NM	GRC/TCAC		AY181036	48, 49
fumA	Fumarase	13	EG	GRC/TCAC		AF497854	49

^a+, mutation impairs growth on a C₁ compound; −, mutation does not impair growth on a C₁ compound; NM, no mutant available: involvement in methylotrophy implicated by regulation, location, or phenotype of mutant in other subunit(s); E, essential for both methylotrophic and heterotrophic growth; EG, growth on C₁ compounds rescued by glyoxylate.

^bMetabolic module designations: MOX, methanol oxidation; PQQ-S, PQQ biosynthesis; MAU, methylamine utilization; H₄F-T, H₄F-linked C₁ transfer; H₄F-S, H₄F biosynthesis; H₄MPT-TS, H₄MPT-linked C₁ transfer and H₄MPT biosynthesis; FOX1, FOX2, and FOX3, three nonhomologous formate dehydrogenases; QSC, (Quayle) serine cycle; PHBC, PHB cycle; TCAC, TCA cycle.

FROM GENETICS TO GENOMICS

M. extorquens AM1 is the most-well-studied methylotroph to date. In the 1960s and 1970s, this organism served as a model to characterize the reactions of the serine cycle for C₁ assimilation, and methanol- and methylamine dehydrogenases from M. extorquens AM1 were among the first primary C₁ oxidation enzymes characterized (3). Genetic studies of M. extorquens AM1 began with the isolation and characterization of chemically induced C₁-negative mutants (3). Later, mutants with defects in methanol oxidation were isolated via a specific allyl alcohol selection (32, 33). The availability of these C₁-negative mutants allowed identification and isolation of the corresponding DNA regions encoding key methylotrophy genes, and it became evident that some of these genes are clustered together (2, 4, 5, 6, 8, 9, 11). As sequencing became a routine technique, these clusters were analyzed and expanded via chromosomal walking (5, 12, 15, 27). Concurrently, directed mutagenesis techniques were developed and applied to the analysis of the genes present in these methylotrophy islands (5, 8-15). By the end of 1990s, classical genetic approaches in combination with chromosomal walking and directed mutagenesis had resulted in characterization of about 70 genes involved in methylotrophy in M. extorquens AM1, and these were localized to eight regions on the chromosome (27). One methylotrophy island spans about 60 kb and contains a number of tightly linked genes enabling different methylotrophy metabolic modules: most of the reactions of the serine cycle, most of the formaldehyde oxidation reactions, and some functions involved in methanol oxidation (27). Most of the methanol oxidation genes were found in a different location on the chromosome, where they formed a large operon (1). All of the genes enabling methylamine oxidation were found in one location and tightly linked (5). However, some important methylotrophy genes were not parts of methylotrophy islands (i.e., gck and glyA) (11, 14), suggesting that further expansion of the existing methylotrophy clusters via chromosomal walking promised limited opportunity for finding new methylotrophy genes. Meanwhile, a number of essential methylotrophy genes were still missing from the picture. These included (i) two of the serine cycle genes assumed to also participate in multicarbon metabolism, encoding malate dehydrogenase and enolase; (ii) formate dehydrogenase genes; and (iii) genes for the novel glyoxylate regeneration pathway involving propionyl coenzme A (CoA) as an intermediate (12). In addition, no regulators were known for assimilatory C₁ metabolism. In order to locate and study the missing methylotrophy genes in M. extorquens AM1, a whole-genome sequencing project was initiated in 1998, in collaboration with the Human Genome Sequencing Center at University of Washington and later Integrated Genomics, Inc. (Chicago, Ill.). At the time of this writing, a gapped sequence of the M. extorquens AM1 genome (6.5× coverage) is available (http://www.integratedgenomics.com/genomereleases.html#list6). Data mining began at the very early stages of the sequencing project, via BLAST analysis and key word searches against the partial genome database. At these early stages, many genes of interest were isolated and sequenced from a cosmid library of M. extorquens AM1. At the later stages when analysis of larger fragments became possible, genes of interest were PCR amplified from the chromosome and subjected to mutation analysis. A random (transposon-induced) mutagenesis approach was also employed, and sites of insertions resulting in a C₁-negative phenotype were identified (28). The whole-genome-based gene-mining approach outlined above has resulted in identification of about 30 new genes involved in methylotrophy. The major outcomes of the whole-genome-analysis approach have been twofold: (i) filling in existing metabolic gaps in our knowledge of methylotrophy in serine cycle facultative methylotrophs and (ii) providing, for the first time, comprehensive knowledge on the suite of necessary genes as well as the suite of sufficient genes for enabling methylotrophy in a given organism. Our data at this time suggest that we have accounted for the majority of methylotrophy genes in M. extorquens AM1. Only a few genes still remain unidentified: namely genes involved in the yet unsolved reactions of the glyoxylate regeneration cycle and a few regulatory genes participating in C₁ oxidation, C₁ assimilation, or both. Below, the current information about each of the methylotrophic metabolic modules is presented in more detail.

PRIMARY C₁ OXIDATION

M. extorquens AM1 possesses two primary oxidation metabolic modules for methylotrophy, which function in the oxidation of methanol and methylamine, respectively. All 11 of the known methylamine utilization genes (mauFBEDACJGIMN)—i.e., the genes for the catalytic subunits for methylamine dehydrogenase, the specific electron acceptor amicyanin, and the specific cofactor (TTQ) biosynthesis enzymes—are located in a single 8.4-kb gene cluster with all genes transcribed in the same direction (5), except the regulatory gene or genes, which remain unknown.

The genes enabling methanol oxidation in M. extorquens AM1 are found in three different locations on the chromosome. One 12.5-kb cluster (cluster 1) contains 14 genes (mxaFJGIRSACKLDEHB), all transcribed in the same direction. These genes encode the structural polypeptides of methanol dehydrogenase, the specific cytochrome c that accepts electrons from methanol dehydrogenase, the proteins essential for calcium insertion into the apoprotein, one regulatory protein, and a few proteins whose functions are still unknown (1, 2, 27, 29-31, 41, 44). One gene (mxaW) is located immediately upstream of this cluster, transcribed in the opposite direction by a methanol-inducible promoter, but its function is unknown (41). One pair of genes involved in transcriptional regulation of MeDH (mxbMD) are a part of the 60-kb methylotrophy island (cluster 2) (42), and another pair (mxcQE) are located elsewhere on the chromosome. The six genes for PQQ biosynthesis are located in two different clusters: one of them (pqqABC/DE) is located in the large methylotrophy island immediately downstream of mxbMD (45), while another cluster (pqqFG) is not linked to other methylotrophy genes (cluster 2) (43). In this work, we view PQQ biosynthesis as a separate metabolic module, because it is a cofactor of other dehydrogenases in M. extorquens AM1, based on genomic predictions.

C₁ TRANSFER BETWEEN FORMALDEHYDE AND FORMATE

Two distinct metabolic modules operate in M. extorquens AM1 for transferring C₁ units between the oxidation levels of formaldehyde and formate, both linked to folate cofactors. One module uses tetrahydrofolate (H₄F) as a cofactor. The enzymes converting methylene-H₄F to formyl-H₄F, methylene-H₄F dehydrogenase, and methenyl-H₄F cyclohydrolase are encoded by genes (mtdA and fch) (9, 36, 51) unique to methylotrophs, while in most known bacteria, both reactions are performed by a bifunctional enzyme encoded by folD. While mtdA and fch are cotranscribed and are a part of the serine cycle gene cluster (9, 20, 36), the gene encoding formyl-H₄F ligase (ftfL) is located elsewhere on the M. extorquens AM1 chromosome, and the encoded polypeptide shows high similarity to known formyl-H₄F ligases (28). mtdA and fch are regulated coordinately with the serine cycle enzymes, suggesting a potential link between the H₄F-linked C₁ transfer module in M. extorquens AM1 and C₁ assimilation (20).

In this study, we place the genes for H₄F biosynthesis into a separate metabolic module, because H₄F is involved not only methylotrophy functions, but also in general metabolism during growth on multicarbon compounds. fol genes (involved in folate synthesis) are therefore essential, and no null mutants can be isolated in these genes (13). Three fol genes (folKBP) are linked together, transcribed in the same direction, and are a part of a methylotrophy island (cluster 2); one gene (dyr) is loosely linked to the gene for fumarase (fumA; cluster 13); and two other genes (folC and folE) are not linked to other fol or methylotrophy genes.

The H₄MPT-linked metabolic module that involves “archaeal-like” genes and enzymes appears to be the pathway responsible for the majority of formaldehyde oxidation (15). Seventeen of the genes involved in this module (fhcCDAB-orf4-mtdB-orfY-mch-orf5-orf7-fae-orf17-orf9-[3 non-“archaeal” genes]-orf19-orf20-orf21-orf22) are clustered together in a 20-kb region on the chromosome, located at one end of the largest methylotrophy island (cluster 2). Some of these genes are in the opposite orientation with respect to others, and therefore the region must be composed of a number of transcriptional units. One gene involved in this module, dmrA, encoding putative dihydromethanopterin reductase, was found in a different location on the chromosome, and was not linked to any other known methylotrophy genes (28). One other gene has been identified as a putative methanopterin biosynthesis gene, orf4, encoding the first enzyme in the methanopterin biosynthesis pathway, β-ribofuranosylaminobenzene 5′-phosphate synthase (40). dmrA and orf4 remain, at this point, the only genes proposed to participate in archaeal cofactor (H₄MPT) biosynthesis in M. extorquens AM1. It is likely, however, that many of the archaeal-like genes of yet unknown function in the “archaeal-like” gene cluster are involved in biosynthesis of H₄MPT or methanofuran or are involved in regulation of these biosynthetic pathways. As such, we have assigned these genes to the H₄MPT-linked C₁ transfer metabolic module.

FORMATE OXIDATION

Until recently, the formate oxidation step was believed to be essential for energy generation during methylotrophic growth (3). However, no randomly generated mutants were available with lesions in formate oxidation. The whole-genome approach has revealed the presence of three gene clusters unlinked to each other, encoding three nonhomologous formate dehydrogenases in M. extorquens AM1 designated fdh1AB, fdh2ABCD, and fdh3ABC (25; L. Chistoserdova, M. Laukel, J. A. Vorholt, and M. E. Lidstrom, unpublished observations). In each case, the genes are transcribed in the same orientation with respect to each other. Mutation analysis has shown that the formate oxidation step is not essential for energy generation during growth on methanol or methylamine. It is essential, however, for growth on formate, but any of the three formate oxidation modules can fulfill this energy-generating function (Chistoserdova et al., unpublished).

SERINE CYCLE

The serine cycle is the pathway for formaldehyde assimilation during methylotrophic growth of M. extorquens AM1. Even though the role of the pathway, the net production of one C₃ molecule (phosphoglycerate) from two molecules of formaldehyde and one molecule of CO₂, is uniquely methylotrophic, little is unique about the 11 genes involved in this module. The first enzyme in the pathway, serine hydroxymethyltransferase (GlyA) is a traditional enzyme found in most known organisms, where it functions in supplying C₁ units in the form of methylene-H₄F for biosynthetic pathways, for instance, purine biosynthesis. Mutants in this enzyme are normally deficient in the biosynthesis of purines (37, 47). GlyA in M. extorquens AM1, however, is specialized to methylotrophy and is not required for growth on multicarbon compounds (11), so an alternative source of C₁ units must exist for purine biosynthesis. Homologs of other enzymes of the serine cycle are also found in nonmethylotrophic bacteria. It therefore seems that the functionality of the serine cycle must be determined by subtle substrate specificity adjustments for the enzymes involved and by common regulation. The glyA gene is not linked to other serine cycle genes. However, six of the serine cycle genes are clustered at one end of the large methylotrophy island (the end opposite to that containing the H₄MPT module genes), together with two of the H₄F-linked C₁ transfer module genes (mtdA and fch), and these are transcribed in two units, sga-hpr-mtdA-fch and mtkA-mtkB-ppc-mcl (20). The serine cycle enzymes encoded by this gene cluster are serine-glyoxylate aminotransferase (sga), hydroxypyruvate reductase (hpr), the two subunits of malate thiokinase (mtkAB), an acetyl-CoA-independent phosphoenolpyruvate (PEP) carboxylase (ppc), and malyl-CoA lyase (mcl). Another serine cycle gene, gck, encoding glycerate kinase, is not linked to other serine cycle genes. Most of the serine cycle genes are regulated coordinately, but so far, only one regulator is known, QscR (20), the gene for which has been discovered via random mutagenesis (28). QscR is a LysR-type regulator with high identity to CbbR, a regulator of autotrophy and photosynthesis in other bacteria (39). The location of qscR on the chromosome, adjacent to the fructose-1,6-bisphosphatase and phosphoribulokinase genes, might be indicative of a relatively recent acquisition from an autotrophic bacterium. While most of the reactions of the serine cycle are enabled by genes specific to the methylotrophic mode of metabolism, and mutants with mutations in these genes grow normally on multicarbon compounds, two enzymes in the pathway are directly borrowed from other metabolic modules and therefore have dual functions. The genome of M. extorquens AM1 contains only one gene for malate dehydrogenase (mdh), and therefore the same enzyme functions in the serine cycle during growth on C₁ compounds and in the tricarboxylic acid (TCA) cycle during growth on multicarbon compounds. Likewise, only one gene is found for enolase (eno); therefore, one and the same enzyme functions in both the serine cycle and gluconeogenesis. Mutant analysis has confirmed that both genes, mdh and eno, are essential, and no null mutants can be obtained on either C₁ or multicarbon substrates (L. Chistoserdova and M. E. Lidstrom, unpublished observations).

GLYOXYLATE REGENERATION CYCLE

The glyoxylate regeneration cycle (GRC) in serine cycle methylotrophs not containing isocitrate lyase has remained a mystery for three decades. The pregenomic efforts resulted in identification of four genes loosely linked on the chromosome (cluster 22); two genes encoding polypeptides of unknown function, MeaA (12, 40) and MeaB (23); and the genes encoding crotonyl-CoA reductase and propionyl-CoA carboxylase (crr, pccB) (12, 23). The genomic approach combined with gene-specific, as well as random mutagenesis and metabolite analysis, has resulted in the identification of most of the reactions of the GRC (23, 24). The pathway involves an elaborate series of reactions proceeding via the CoA derivatives of C₃, C₄, and C₅-carboxylic acids and involves two carboxylation and two decarboxylation reactions and at least two mutase reactions (23). A few pieces of the puzzle are still missing: i.e., the two enzymes participating in conversion of ethylmalonyl-CoA into isobutyryl-CoA (a putative mutase and a putative decarboxylase) are yet to be identified, and the substrate for MeaA, a putative mutase, remains unknown, as do the other enzymes involved in the conversion of β-hydroxyisobutyryl-CoA into propionyl-CoA (Fig. 1). Mutant analysis shows that early steps of the GRC overlap with the pathway for poly-β-hydroxbutyrate (PHB) biosynthesis (24), and the late steps overlap with late steps of the TCA cycle and the serine cycle (23). A total of 12 genes are known that are specific to this pathway (croR, crr, pccAB, ibd2, meaABCD, mcmAB, and epm), 3 that overlap with PHB biosynthesis (phaABR), 5 that overlap with the TCA cycle (sdhABCD, fumA), and 3 that overlap with the serine cycle (mtkAB, mcl). With the exception of the four genes that are part of cluster 22 and the sdh and the pha genes, the genes for the GRC are not linked to each other or to other known methylotrophy genes. Why such an elaborate pathway is employed by many serine cycle methylotrophs instead of the classic glyoxylate shunt remains unknown. However, it may in part reflect the need for low carbon flux through the initial steps of the TCA cycle during methylotrophic growth. The GRC is viewed here as a separate metabolic module, as opposed to a part of the serine cycle due to its role not only in C₁, but also in C₂ metabolism in M. extorquens AM1 and possibly in other bacteria. The presence of homologs for the GRC genes in other α-proteobacterial genomes points towards this pathway being widespread in non-methylotrophs, and at least in Streptomyces, the pathway has been shown to be involved in C₂ metabolism (19).

PHB CYCLE

Serine cycle methylotrophs accumulate PHB as a reserve material. Metabolism of PHB in M. extorquens AM1 is intimately interlinked with its C₁ metabolism. The first two reactions of the PHB cycle (catalyzed by PhaA and PhaB) are also the first reactions of the GRC (23, 24). Besides the catalytic steps, there are common regulatory mechanisms controlling both PHB production and C₁ metabolism. These are not completely understood at this point, but one such common regulator, PhaR, seems to be involved in directing flows of acetyl-CoA between C₁ assimilation and PHB accumulation (23).

TCA CYCLE

During growth on multicarbon compounds, the TCA cycle plays its classical role in carbon and energy metabolism in M. extorquens AM1 (3). In contrast, it is not involved in energy generation during growth on C₁ compounds. Instead, α-ketoglutarate dehydrogenase is repressed, and the incomplete cycle plays a strictly assimilatory role (3, 48, 49). However, many of the TCA cycle enzymes are involved in C₁ assimilation. Besides malate dehydrogenase, a series of the TCA cycle reactions converting succinyl-CoA into malate form a part of the GRC. Genes encoding two enzyme systems capable of converting succinyl-CoA into succinate have been identified in the genome of M. extorquens AM1: for succinyl-CoA synthase (the genes scsA and scsB are linked to mdh) and keto acid succinyl-CoA transferase (kst). In addition, cell extracts contain a succinyl-CoA hydrolase activity, but the gene responsible for this activity is unknown (Chistoserdova and Lidstrom, unpublished). Null mutations in scsB and kst caused no effect on growth of M. extorquens AM1 on C₁ or multicarbon compounds (Chistoserdova and Lidstrom, unpublished). This result points toward either succinyl-CoA hydrolase being the essential enzyme for this conversion or the three systems being degenerate for this function. The genes for succinate dehydrogenase (SDH) are all linked on the chromosome (cluster 29), and SDH null mutants cannot be obtained either on methanol or on succinate (48, 49). However, they can be obtained on methanol supplemented with glyoxylate (N. Korotkova and M. E. Lidstrom, unpublished observations). This growth condition circumvents the GRC, confirming the hypothesis that SDH is involved in glyoxylate regeneration during growth on C₁ compounds. Two genes showing homology to fumarases (fumA and fumB) are present in the genome. One of the potential fumarase genes (fumA) has been subjected to mutagenesis, and, as in the case of SDH, null mutants could only be obtained on methanol supplemented with glyoxylate (Korotkova and Lidstrom, unpublished), confirming the dual function of this enzyme in C₁ and multicarbon metabolism. The role of the second homolog of fumarase remains unknown.

CONCLUSIONS

We have presented here the genome-based analysis of methylotrophic metabolism in one model facultative methylotroph, M. extorquens AM1. This analysis presents a comprehensive picture of the complex genetic and biochemical makeup of methylotrophy in a given organism. A little over 100 genes participate in C₁ metabolism in M. extorquens AM1, and these belong to a few specialized metabolic modules. Some genes involved in these modules are located in “methylotrophy islands,” while others are scattered around the chromosome and are present as singular entities. While some genes are specialized in methylotrophic metabolism, others are shared with nonmethylotrophic pathways. Such a complex framework of methylotrophy in M. extorquens AM1 may reflect a complex and nonlinear history of this metabolic capability. Because a nearly complete set of genes involved in the different steps of methylotrophic metabolism in M. extorquens AM1 is now defined, the stage is now set for studies directed at understanding how this complex network of genes and enzymes is coordinated and what are the important mechanisms for switching between C₁ and multicarbon metabolic modes in this organism.