Microbiology (2003), 149, 601–609
DOI 10.1099/mic.0.25955-0
Reconstruction of C3 and C4 metabolism in
Methylobacterium extorquens AM1 using
transposon mutagenesis
Stephen J. Van Dien,1 Yoko Okubo,1 Melinda T. Hough,2
Natalia Korotkova,1 Tricia Taitano2 and Mary E. Lidstrom1,2
Departments of Chemical Engineering1 and Microbiology2, University of Washington, Seattle,
WA 98195, USA
Correspondence
Mary E. Lidstrom
[email protected]
Received 20 August 2002
Revised
1 October 2002
Accepted 18 November 2002
The growth of Methylobacterium extorquens AM1 on C1 compounds has been well-studied, but
little is known about how this methylotroph grows on multicarbon compounds. A Tn5 transposon
mutagenesis procedure was performed to identify genes involved in the growth of M. extorquens
AM1 on succinate and pyruvate. Of the 15 000 insertion colonies screened, 71 mutants were found
that grew on methanol but either grew slowly or were unable to grow on one or both of the
multicarbon substrates. For each of these mutants, the chromosomal region adjacent to the
insertion site was sequenced, and 55 different genes were identified and assigned putative
functions. These genes fell into a number of predicted categories, including central carbon
metabolism, carbohydrate metabolism, regulation, transport and non-essential housekeeping
functions. This study focused on genes predicted to encode enzymes of central heterotrophic
metabolism: 2-oxoglutarate dehydrogenase, pyruvate dehydrogenase and NADH : ubiquinone
oxidoreductase. In each case, the mutants showed normal growth on methanol and impaired growth
on pyruvate and succinate, consistent with a role specific to heterotrophic metabolism. For the first
two cases, no detectable activity of the corresponding enzyme was found in the mutant, verifying the
predictions. The results of this study were used to reconstruct multicarbon metabolism of
M. extorquens AM1 during growth on methanol, succinate and pyruvate.
INTRODUCTION
Methylotrophic bacteria are organisms capable of growth
using C1 compounds such as methanol as the only carbon
and energy source and therefore, are of interest as
biocatalysts for the conversion of methanol to useful
products (Lidstrom, 1992). Since bacteria are not optimized
to make products under large-scale commercial process
conditions, metabolic engineering of methylotrophs will be
required to develop strains with more-desirable process
characteristics. Ideally, for the development of economically
viable bioprocesses, cellular-wide resources should be
redirected to synthesizing a specific product from methanol
at high efficiency and yield. To achieve such a goal, it is
necessary to understand and manipulate central metabolism. One of the most widely studied of the methanolutilizing bacteria is the pink-pigmented facultative
methylotroph Methylobacterium extorquens AM1. This
organism is capable of growth on a variety of C2, C3 and
Abbreviation: TCA, tricarboxylic acid.
The GenBank accession numbers for the M. extorquens AM1
pdhABCD, sucABC and fumA sequences reported in this paper are
AF497851, AF497852 and AF497854, respectively.
0002-5955 G 2003 SGM
C4 compounds as well as methanol and methylamine,
assimilates methanol at the level of formaldehyde by the
serine cycle and has served as the primary model system for
the study of methylotrophic metabolism and methylotrophic enzymes (Chistoserdova, 1996; Lidstrom, 1992).
Methylotrophy in M. extorquens AM1 is well-understood at
the physiological and genetic levels and tools are available
for genetic manipulations, making it an attractive system for
the development of metabolic engineering techniques in
methylotrophs. Eighty-six genes have been characterized,
most of which are involved in the methanol assimilation
and oxidation pathways unique to methylotrophic bacteria
(Chistoserdova, 1996). In contrast, little is known about the
growth of M. extorquens AM1 on C3 and C4 compounds, or
of the metabolism of multicarbon compounds once they
are produced from formaldehyde via the serine cycle. This
unexplored region of heterotrophic central metabolism
includes parts of the tricarboxylic acid (TCA) cycle,
anapleurotic pathways, gluconeogenesis and the pentosephosphate pathway. Past work has shown that TCA cycle
enzymes are present at higher levels in cells grown on
multicarbon substrates than in cells grown on C1 substrates
(Salem et al., 1973; Taylor & Anthony, 1976). In addition,
mutants in 2-oxoglutarate dehydrogenase (Taylor &
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
Printed in Great Britain
601
S. J. Van Dien and others
Anthony, 1976) or pyruvate dehydrogenase (Bolbot &
Anthony, 1980) were capable of growth on C1 compounds
but not multicarbon compounds, indicating the specificity
of these reactions for heterotrophy. However, some of the
mutants were leaky and the site of the mutations was not
characterized.
The recent availability of the M. extorquens AM1 genome
sequence along with the advent of genetic tools for the
construction of insertion mutants (Chistoserdov et al.,
1994) and the overexpression of genes (Marx & Lidstrom,
2001) now enable the application of genomic approaches to
the understanding of metabolic pathways in this organism.
In a previous study, five genes predicted to be involved in
growth on succinate or pyruvate encoding citrate synthase,
succinate dehydrogenase, malic enzyme, phosphoenolpyruvate synthase and phosphoenolpyruvate carboxykinase
were identified in the genome sequence. Mutants were
generated and characterized, providing initial information on pathways of C3 and C4 metabolism (Van Dien &
Lidstrom, 2002).
In this study, we apply a more-global approach to the
metabolic reconstruction of central heterotrophic metabolism in M. extorquens AM1. It has recently been demonstrated in our laboratory that a mini-Tn5 derivative,
ISphoA|hah-Tc (D’Argenio et al., 2001), could be used
successfully in M. extorquens (Marx et al., 2003). We describe here the generation of a pool of random ISphoA|hahTc insertion mutants, and the screening of these mutants for
defective growth on non-C1 substrates. Through identification and further characterization of the gene interruptions
responsible for these growth phenotypes, a more-complete
understanding of M. extorquens AM1 heterotrophic and
methylotrophic central metabolism has been obtained.
METHODS
Growth of M. extorquens AM1. Cultures were grown aerobically
at 30 ˚C either in liquid or on agar plates using mineral salts
medium (Attwood & Harder, 1972) containing either 0?5 % methanol, 0?4 % succinate or 0?4 % pyruvate as the growth substrate. To
generate growth curves, 30 ml of methanol-grown culture in exponential phase was harvested and resuspended in medium without
carbon source to a final OD600 value of 6?0. One millilitre of this
suspension was used to inoculate 30 ml cultures of different carbon
sources, and the OD600 value of these cultures was measured as a
function of time for approximately 30 h. Each growth experiment
was performed twice with reproducible results, so only one curve is
shown in each case. Doubling times, when given, represent means
over the curves for both experiments.
Generation and screening of transposon mutants. Transposon
mutagenesis of M. extorquens AM1 was performed using the
ISphoA|hah-Tc delivery plasmid pCM639 (D’Argenio et al.,
2001; Marx et al., 2003). pCM639 was introduced into wild-type
M. extorquens AM1 by biparental mating using Escherichia coli SM10
lpir (Miller & Mekalanos, 1988). Recombinants were selected on
minimal salts medium agar plates containing methanol as the growth
substrate, 10 mg tetracycline ml21 and 50 mg rifamycin ml21 for selection. Individual colonies were purified by streaking on fresh plates of
the same composition.
602
After growth for 3 days on plates containing methanol, mutants were
tested for C3 and C4 growth phenotype by streaking on minimal salts
medium agar plates containing pyruvate or succinate as the carbon
source and 10 mg tetracycline ml21. Mutants were allowed to grow for 3
days, at which time those with a visible growth deficiency were selected.
These strains were then retested on all three types of minimal plates
to verify phenotype. After 3 days of growth, mutants were assigned
a phenotype on each substrate based on their observed growth
characteristics: normal (++), if growth appeared similar to wild-type
on the screening medium; slow (+), if after 3 days it was possible to
observe colony growth but it was less than wild-type; minus (2), if after
3 days there was almost no observable growth on the screening
medium.
PCR amplification of interrupted chromosomal region. The
chromosomal region adjacent to the transposon insertion in each of
the mutant strains was amplified using a semi-random, two-step
PCR protocol (Chun et al., 1997; Marx et al., 2003). The products
were purified using Qiaquick spin columns (Qiagen), and sequence
analysis was performed by the University of Washington Sequencing
Facility.
Prediction of interrupted gene function. Identity searches
were performed to locate the sequences in the M. extorquens
AM1 partial genome sequence (Integrated Genomics; http://www.
integratedgenomics.com/genomereleases.html#list6). Putative gene
function was assigned by BLAST search of the corresponding translated sequence against the NCBI database (http://www.ncbi.nlm.nih.
gov/BLAST).
Generation of directed mutations. Data from the M. extorquens
AM1 genome project were used to design PCR primers specific for
regions of the genome containing candidates for multicarbonspecific genes. Putative genes encoding the alpha subunit of
the pyruvate dehydrogenase E1 component, the B subunit of
NADH : ubiquinone oxidoreductase and fumarase were identified in
the M. extorquens AM1 genome sequence and amplified by PCR
using chromosomal DNA as a template. Products of the expected
sizes were obtained and isolated, cloned directly into pCR2.1
(Invitrogen) and subcloned into pUC19 (Promega) as either EcoRI–
EcoRI or XbaI–KpnI fragments. Unique blunt restriction sites located
near the beginning of the gene were found and used for the insertion
of a 1?4 kb HincII fragment from pUC4K (van der Oost et al., 1989)
containing a kanamycin resistance (KmR) cassette. Orientation was
chosen so that the KmR gene was transcribed in the same direction
as the M. extorquens AM1 gene. The interrupted gene was subsequently removed and cloned into the suicide vector pAYC61
(Chistoserdov et al., 1994); the resulting plasmid was transformed
into E. coli S17-1 (Simon et al., 1983). The resulting strains were
used as donor strains in biparental matings with wild-type
M. extorquens AM1, and KmR TcS progeny were obtained on minimal
medium agar plates containing methanol as described previously
(Chistoserdov et al., 1994). In all cases, the identity of the doublecrossover mutants was confirmed by PCR using chromosomal DNA
as a template and the gene-specific primers described above.
Overexpression of genes in M. extorquens AM1. Genome
sequence data obtained as described above were used to design PCR
primers for the amplification of the wild-type gene. Products of the
expected sizes were obtained and isolated, cloned directly into
pCR2.1 and subcloned downstream of the PmxaF promoter in the M.
extorquens AM1 expression vector pCM80 (Marx & Lidstrom, 2001)
as either EcoRI–EcoRI or XbaI–KpnI fragments.
DNA manipulations. Plasmid isolation, E. coli transformation,
restriction enzyme digestion and ligation were carried out by standard protocols (Sambrook et al., 1989). The chromosomal DNA of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
Microbiology 149
Methylobacterium extorquens C3 and C4 metabolism
M. extorquens AM1 was isolated by the procedure of Saito & Miura
(1963). Biparental matings between E. coli and M. extorquens AM1
were performed as described previously (Chistoserdov et al., 1994).
Enzyme assays. Enzyme activities were determined in wild-type
or mutant M. extorquens AM1 crude extracts obtained by passing
cells through a French pressure cell at 1?26108 Pa, followed by centrifugation at 15 000 g. All measurements were done aerobically at
room temperature in a total volume of 1 ml using published
methods as follows: pyruvate dehydrogenase (Thissen et al., 1986)
and glucose-6-phosphate isomerase (Schreyer & Bock, 1980). Fumarase
activity was determined in the malate to fumarate direction using
the technique of Flint (1994). However, due to interference from
the crude cell extracts at 240 nm, the formation of fumarate was
monitored by following the increase in absorbance at 300 nm.
2-Oxoglutarate dehydrogenase assay was performed by monitoring
the release of 14CO2 from [1-14C]2-oxoglutarate in the presence of
NAD+ and Coenzyme A (Green et al., 2000). For all assays, a milliunit (mU) of activity is defined as 1 nmol substrate reacted min21.
RESULTS
Transposon mutagenesis using ISphoA|hah-Tc (D’Argenio
et al., 2001) was performed on wild-type M. extorquens AM1
using the procedure outlined above to create a bank of
random, tetracycline-resistant mutants capable of growth
on methanol. Fifteen thousand insertion strains were
screened for growth phenotype on pyruvate and succinate
as representative C3 and C4 growth substrates, respectively.
Since the M. extorquens AM1 genome contains approximately 7000 open reading frames, this theoretically
represented slightly over two times coverage and thus was
not intended to be saturating. Seventy-one mutants with
impaired growth on one or both substrates were identified
and subjected to further characterization (Table 1). The
chromosomal region adjacent to the transposon insertion of
mutant strains was amplified using a semi-random, twostep PCR protocol (Chun et al., 1997), and the resulting
products were sequenced. The amplification was successful
with 63 of the mutants and a total of 55 genes were
identified, after accounting for duplications. Putative gene
functions could be assigned to many genes based on BLAST
searches of the translated sequence in all reading frames
against the NCBI non-redundant sequence database (http://
www.ncbi.nlm.nih.gov/BLAST). Fifteen genes either
showed no significant identity with any genes of known
function or could only be identified as ‘hypothetical
proteins’ (Table 1). Most of the remaining 40 genes are
predicted to belong to one of several categories: those
involved in central carbon metabolism, those involved in
carbohydrate metabolism, putative regulators, putative
transporters and non-essential housekeeping genes. A
number of genes in these last three categories were identified
only as belonging to broad groups of proteins. A few of these
genes are of special interest, including the putative transcriptional regulators and a putative C4-dicarboxylate transport protein. These and the genes of unknown function
represent a pool of functions yet to be identified in C3 and C4
metabolism and will be analysed in separate studies. The
gene functions given in Table 1 are those of the protein with
highest identity to the mutated M. extorquens AM1 gene,
http://mic.sgmjournals.org
and are not necessarily the function of the M. extorquens
AM1 gene itself. In each case, experimental evidence will be
required to confirm function. However, in those cases in
which the identity is high, it is likely that the functions are
similar.
Since the aim of this study was the reconstruction of
multicarbon metabolism in M. extorquens AM1, we focused
our initial experimental verification studies on genes
predicted to encode known metabolic functions that were
of interest for filling in knowledge gaps and confirming
model predictions (Van Dien & Lidstrom, 2002). These
include 2-oxoglutarate dehydrogenase (M38-24), pyruvate
dehydrogenase (M106-62) and NADH : ubiquinone oxidoreductase (M06-14), all of which are predicted to not be
required for growth on C1 compounds (Taylor & Anthony,
1976; Bolbot & Anthony, 1980; Van Dien & Lidstrom, 2002),
and glucose-6-phosphate isomerase (M02-75), which is predicted to be required for all growth conditions (Van Dien
& Lidstrom, 2002). A mutant was also obtained in a
gene predicted to encode NAD-dependent malic enzyme
(M121B-13), which catalyses the reversible decarboxylation
of malate to pyruvate. However, an insertion mutant in this
gene having the same growth phenotype was already
constructed by a directed approach and characterized in a
previous study (Van Dien & Lidstrom, 2002), so this strain
was not investigated further.
2-Oxoglutarate dehydrogenase
Strain M38-24 contains a transposon insertion in a gene
predicted to encode the E1 component of 2-oxoglutarate
dehydrogenase (EC 1.2.4.2), approximately 400 bp downstream of the 59 end of the gene. The identity of this gene
is supported by the location of putative genes encoding E2
and dihydrolipoamide dehydrogenase components of the
enzyme immediately downstream of the gene for the E1
subunit. These genes were named sucA, sucB and sucC.
2-Oxoglutarate dehydrogenase activity in the wild-type
strain increased during growth on succinate as compared to
methanol as shown previously (Taylor & Anthony, 1976),
and the activity was non-detectable in the mutant strain
(Table 2). This mutant does not grow appreciably on succinate or pyruvate either on agar plates or in broth culture, but
has nearly a wild-type growth rate on methanol (Fig. 1).
Pyruvate dehydrogenase
The transposon insertion in strain M106-62 is located
approximately 300 bp upstream of the 39 terminus of a
gene predicted to encode the E2 component of pyruvate
dehydrogenase (EC 2.3.1.12). The gene is preceded by genes
with identity to both subunits of the E1 component and is
followed by a possible dihydrolipoamide dehydrogenase
gene. This gene cluster was named pdhABCD. M106-62
grows slowly on succinate and pyruvate as compared
to the wild-type strain, and the pyruvate dehydrogenase
activity in this strain is approximately one-third that of the
wild-type during methanol growth (Table 3). Furthermore,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
603
S. J. Van Dien and others
Table 1. List of insertion mutants for which sequence data were obtained, growth phenotypes on methanol (MeOH), pyruvate
(Pyr) and succinate (Succ) plates, and predicted gene function determined by BLAST search (see text)
‘No significant homology’ indicates that the insertion site could be located in the genome sequence, but the translated ORF has no homology with any protein in the NCBI database with E-value less than e-4. Mutants chosen for further study are shown in bold type. GenBank
accession numbers are provided for the sequence used in the functional assignment. Numbers beginning with NP are predictions from
other genome sequences; all others are sequences of functionally characterized products.
Mutant
name
MeOH
M01-20
M01-67
M02-41
M02-75
M03-10
M03-92
M06-14
M06-72
M09-34
M09-91
M10-64
M14-86
M15-30
M15-44
M17-54
M18-44
M18-47
M19-78
M23-54
M24-55
M25-42
M25-62
M26-12
M26-63
M28-54
M31-97
M36-84
M38-24
M39-38
M40-55
M40-86
M41-07
M42-53
M47-31
M52-50
M52-59
M52-60
M52-69
M52-70
M53-32
M53-82
M56-76
M58-33
M60-63
M68-64
M70-06
M71-36
M71-77
++*
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
+
++
++
++
++
++
++
++
+
+
++
++
+
++
++
++
++
++
++
++
++
++
+
++
++
++
+
++
++
++
604
Pyr
+
+
++
+
2
+
2
2
++
2
2
+
+
+
+
2
2
++
2
++
2
+
+
+
++
+
+
2
+
+
2
2
++
+
++
2
+
2
++
++
2
+
+
+
2
+
2
2
Succ
++
++
2
+
2
++
++
++
2
2
++
++
++
++
++
+
+
+
2
+
2
+
++
+
+
++
++
2
++
++
+
2
+
+
+
++
++
++
+
+
2
++
++
++
++
++
2
++
Predicted gene function of known protein
with smallest e-value
Acriflavin resistance protein B
Hypothetical cytoplasmic protein (198 aa)
ATP-dependent DNA helicase
Glucose-6-phosphate isomerase
2-Oxoglutarate dehydrogenase, E1 component
Biotin synthesis protein
NADH : ubiquinone oxidoreductase, Chain B
No significant homology
No significant homology
C4-dicarboxylate transport protein
Plasmid stabilization protein
Multidrug/solvent efflux protein
Integration/recombination protein
Penicillin-binding protein
No significant homology
Putative monooxygenase
Possible regulator containing HD-GYP domain
No significant homology
Putative haemoglutanin
Hypothetical protein (520 aa)
Pseudouridine synthase I, large subunit
Replication initiator and transcription repressor
Hypothetical protein (178 aa)
Co transport protein
Mannosyltransferase
No significant homology
No significant homology
2-Oxoglutarate dehydrogenase, E1 component
Component of Type IV secretion system
No significant homology
D-Alanyl-D-alanine carboxypeptidase
Glucose-1-dehydrogenase
Mannosyltransferase (same gene hit as M28-54)
DMSO/TMAO sensor kinase
No significant homology
No significant homology
LuxR-related regulatory protein
No significant homology
Osmotically inducible protein C
Hypothetical protein (126 aa)
Mn transport protein
Cobalt-dependent methionine amino peptidase
Biotin synthesis protein (same gene hit as M03-92)
Glutamine synthetase II
(p)ppGpp synthetase (RelA/SpoT homologue)
Oligoendopeptidase
2-Oxoglutarate dehydrogenase, E2 component
Multidrug efflux protein
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
E-value
Identity
(%)
1e-180
1e-48
7e-90
3e-16
<e-199
1e-60
3e-68
46
37
52
26
70
53
82
NP_386568
NP_540941
NP_295013
S15937
AAC44748
NP_539100
AAC12755
1e-152
3e-34
2e-64
3e-79
9e-95
74
30
42
75
32
A33597
NP_109544
AAD39553
S42585
NP_539972
1e-79
6e-34
57
43
NP_641207
NP_621747
2e-43
1e-126
3e-98
3e-28
3e-5
1e-42
1e-15
27
41
55
32
38
35
26
NP_639535
NP_108030
NP_102152
S65577
NP_634583
AAA67612
BAA92237
<e-199
3e-48
70
29
AAC44748
BAA97443
8e-59
7e-25
1e-15
1e-55
49
34
26
37
NP_420964
AF453501
BAA92237
AAB94870
3e-19
27
AAA27541
1e-33
7e-5
1e-129
3e-38
1e-60
1e-155
<1e-199
<1e-199
1e-111
<1e-199
52
32
67
33
53
85
59
64
51
39
S17652
NP_273937
NP_642362
3MATA
NP_539100
AAD11279
AAF04327
NP_540989
AAC45482
NP_107696
BLASTP
Accession no.
Microbiology 149
Methylobacterium extorquens C3 and C4 metabolism
Table 1. cont.
Mutant
name
MeOH Pyr
Succ
M73-53
M73-65
M74-50
M75-58
M76-27
M77-45
M90-53
M101-91
M102-41
M106-62
M116-45
M121A-17
M121B-13
++
++
++
++
++
++
+
++
+
++
+
++
++
++ +
2 ++
+
2
+
+
2
2
+
2
++ 2
+ ++
++ +
+
+
+ ++
+
+
++ +
M123A-25
+
2
2
M125B-04
++
2
++
Predicted gene function of known protein
with smallest e-value
BLASTP
E-value
Identity
(%)
Accession no.
Plasmid stabilization protein (same gene hit as M10-64)
3e-34
30
NP_109544
Nitrate-binding ATP-dependent transporter
1e-47
49
AAB86903
2-Isopropylmalate synthase
1e-127
50
S52294
Pyruvate dehydrogenase, E3 component
1e-145
54
S57635
No significant homology
N-Carbamyl-L-cysteine amidohydrolase
5e-50
35
BAB78482
NADH-dependent flavin oxidoreductase
1e-170
49
NP_254085
NADH : ubiquinone oxidoreductase, Chain J
7e-26
45
P29922
UTP-glucose-1-phosphate uridyltransferase
2e-87
59
D49349
Pyruvate dehydrogenase, E2 component
1e-92
42
AAD46491
NADH : ubiquinone oxidoreductase, Chain E
2e-52
52
P29914
Low-pH-induced protein
2e-67
50
NP_539595
NAD-dependent malic enzyme
Identity of gene confirmed by assay
(Van Dien & Lidstrom, 2002)
2-Oxoglutarate dehydrogenase, E1 component
<e-199
70
AAC44748
(same gene hit as M38-24)
Putative aldehyde dehydrogenase subunit
4e-52
64
NP_436869
*++, Normal growth; +, slow growth; 2, no growth.
Table 2. 2-Oxoglutarate dehydrogenase activity in cell extracts of various strains
One milliunit of activity is defined as the reduction of 1 nmol NAD+ min21. Where error ranges are given, each value represents the
mean±SD of three measurements. ND, No activity detected in any of three independent measurements.
M. extorquens AM1 strain
Wild-type
Wild-type
M38-24 (2-oxoglutarate dehydrogenase mutant)
M106-62 (pyruvate dehydrogenase mutant)
Growth substrate
Activity [mU (mg protein)21]
Methanol
Succinate
Methanol
Methanol
2?06±1?10
6?10±2?02
ND
3?67±0?55
the activity in the wild-type strain increases more than twofold during growth on pyruvate or succinate, as described
previously (Salem et al., 1973). Finally, to demonstrate that
no sharing of the E1 and E2 subunits occurs between the
pyruvate and 2-oxoglutarate dehydrogenase complexes,
assays for both dehydrogenases were performed on the
strains mutant in each enzyme. 2-Oxoglutarate dehydrogenase activity in the pyruvate dehydrogenase mutant
M106-62 was comparable to that of the wild-type (Table 2).
Similarly, the 2-oxoglutarate dehydrogenase mutant M3824 did not exhibit reduced pyruvate dehydrogenase activity
(Table 3).
Fig. 1. Representative growth curves of M38-24 (2-oxoglutarate
dehydrogenase mutant) compared to those of wild-type
M. extorquens AM1. &, Wild-type on methanol; ., wild-type
on succinate; m, wild-type on pyruvate; %, M38-24 on methanol; #, M38-24 on succinate; n, M38-24 on pyruvate.
http://mic.sgmjournals.org
In the transposon mutant screen we also identified a strain
(M75-58) with an insertion in the putative dihydrolipoamide dehydrogenase gene upstream of the putative E2
subunit gene. This mutant had the same phenotype on
agar plates as M106-62. In an attempt to generate a nonleaky pyruvate dehydrogenase mutant, an insertion mutant
in the putative E1 subunit gene was generated by the
directed approach described in Methods. This new strain,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
605
S. J. Van Dien and others
Table 3. Pyruvate dehydrogenase activity in cell extracts of
various strains
One milliunit of activity is defined as the reduction of 1 nmol
NAD+ min21. Each value represents the mean±SD of at least
three measurements from multiple cultures. ND, No activity
detected in any of three independent measurements.
M. extorquens AM1 strain
Growth
substrate
Activity
[mU (mg protein)21]
Wild-type
Wild-type
Wild-type
M106-62 (pyruvate
dehydrogenase mutant)
M38-24 (2-oxoglutarate
dehydrogenase mutant)
AM1-PHD1
Methanol
Succinate
Pyruvate
Methanol
33?1±10?1
107?4±7?2
95?9±30?8
10?7±7?2
Methanol
115?6±18?1
Methanol
ND
contig and show high identity to genes of other organisms.
For example, the translated amino acid sequence of the
M. extorquens AM1 subunit B gene is 83 % identical to that
of Sinorhizobium meliloti and 79 % identical to that of
Agrobacterium tumefaciens. Unfortunately, all of our presumed NADH : ubiquinone oxidoreductase mutant strains
grew normally on pyruvate when revived from 280 ˚C
freezer stocks, while retaining tetracycline resistance. This
result suggests the mutants acquired a compensating
mutation that not only allowed growth on pyruvate, but
also exerted strong selective pressure during culturing on
methanol. In order to confirm the preliminary phenotype
assignments, it was necessary to regenerate a mutant by a
directed approach. A strain containing an insertion
mutation in the gene predicted to encode the B subunit
of NADH : ubiquinone oxidoreductase was constructed as
described in Methods. Subunit B was chosen because one of
the strains obtained in the transposon mutagenesis screening process contains an insertion in this gene and there are
no known paralogues of this gene in the M. extorquens
AM1 genome sequence. In contrast, putative paralogues for
several other subunits exist elsewhere on the chromosome.
In agreement with the phenotype of the original transposon
mutant M06-14, this putative NADH : ubiquinone oxidoreductase mutant (YO1) does not grow on pyruvate plates
but grows normally on succinate plates.
To further study the growth phenotype of the mutant strain
YO1, growth rates were measured in liquid mineral salts
medium with each of the three substrates and compared
with that of wild-type (Fig. 3). Although the pyruvate
culture grew, with a doubling time of 9?9 h compared to
4?7 h with the wild-type, there was a significant lag upon
transfer from methanol culture that did not occur in the
wild-type. A less-severe lag also occurred with the succinate
Fig. 2. Representative growth curves of AM1-PDH1 (pyruvate
dehydrogenase subunit E1 mutant) compared to those of
wild-type M. extorquens AM1. &, Wild-type on methanol;
., wild-type on succinate; m, wild-type on pyruvate; %,
AM1-PDH1 on methanol; #, AM1-PDH1 on succinate; n,
AM1-PDH1 on pyruvate.
AM1-PDH1, does not grow appreciably on succinate or
pyruvate on agar plates or in broth culture (Fig. 2) and has
no detectable pyruvate dehydrogenase activity (Table 3).
NADH : ubiquinone oxidoreductase
Several transposon mutants were obtained with insertions
in genes predicted to encode different subunits of the
NADH : ubiquinone oxidoreductase (EC 1.6.5.3), all with
a pyruvate-negative growth phenotype. This enzyme is
highly conserved among bacterial species and consists of
14 subunits. In the M. extorquens AM1 genome sequence,
putative genes for these 14 subunits are clustered on a single
606
Fig. 3.
Representative
growth
curves
of
YO1
(NADH : ubiquinone oxidoreductase subunit B mutant) compared to those of wild-type M. extorquens AM1. &, Wild-type
on methanol; $, wild-type on succinate; m, wild-type on pyruvate; %, YO1 on methanol; #, YO1 on succinate; n, YO1 on
pyruvate.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
Microbiology 149
Methylobacterium extorquens C3 and C4 metabolism
culture, and the final succinate doubling time was 5?4 h,
compared to 4?2 h with the wild-type.
Glucose-6-phosphate isomerase
An insertion (M02-75) was also obtained in a putative
glucose-6-phosphate isomerase gene; this mutant grew
slowly on both pyruvate and succinate on agar plates. The
existence of this transposon mutant is surprising, since
glucose-6-phosphate isomerase is expected to be a necessary
enzyme for all carbon sources, given the essential role it
performs in sugar metabolism. To assure correct prediction
of gene function, activity assays were performed. Glucose-6phosphate isomerase (EC 5.3.1.9) activity in extracts from
M02-75 is 15?8±1?67 mU (mg protein)21, as opposed to
408?5±3?06 mU (mg protein)21 in extracts from wild-type
M. extorquens AM1, both during growth on methanol,
indicating that this gene does affect the predicted activity.
It seems likely that the residual activity in the mutant is
sufficient to allow normal growth on methanol, providing a
possible explanation for why this mutant was obtained.
Directed mutagenesis of a putative fumarase
gene
A few genes that would be expected to function in central
metabolism during C3 and C4 growth were not identified
by this random mutagenesis procedure. One such gene of
interest is that encoding the TCA cycle enzyme fumarase
(EC 4.2.1.2) since it functions both in the TCA cycle during
heterotrophic growth and in the glyoxylate regeneration cycle during methylotrophic growth (Korotkova
et al., 2002). Two putative fumarase genes were identified
in the M. extorquens AM1 genome sequence. One of these
genes, named fumA, was cloned behind the PmxaF promoter
of pCM80 and the resulting plasmid mated into wild-type
M. extorquens AM1. The resulting strain exhibited a
fumarase activity of 2120±170 mU (mg protein)21, as
opposed to 140±39 mU (mg protein)21 in the wild-type
strain with no plasmid, thus confirming the predicted
enzyme activity. A double-crossover mutant in this gene
could not be obtained, suggesting that it is required for
growth on all substrates tested. These results suggest that
under these growth conditions the other putative fumarase
gene either does not encode fumarase or is not expressed at a
sufficient level to rescue the mutant.
Fig. 4. Summary of growth phenotypes for insertion mutants in
genes encoding enzymes of the TCA cycle and anapleurotic
pathways. (a) Methanol growth; (b) succinate growth; (c) pyruvate growth. Heavy solid line, mutant will not grow on substrate; heavy dashed line, mutant grows slowly; heavy dotted
line, mutant grows like wild-type; light solid line, not determined. Enzymes are as follows: 1, citrate synthase (Van Dien &
Lidstrom, 2002); 2, 2-oxoglutarate dehydrogenase (this work); 3,
succinate dehydrogenase (Van Dien & Lidstrom, 2002); 4,
fumarase (this work); 5, malate dehydrogenase (M.
Chistoserdova, unpublished data); 6, malic enzyme (Van Dien &
Lidstrom, 2002); 7, phosphoenolpyruvate carboxylase (Arps et
al., 1993); 8, phosphoenolpyruvate carboxykinase (Van Dien
& Lidstrom, 2002); 9, pyruvate kinase (Chistoserdova &
Lidstrom, 1997); 10, phosphoenolpyruvate synthase (Van Dien
& Lidstrom, 2002); 11, pyruvate dehydrogenase (this work).
Also noted are enzymes for which mutants have not been
obtained but for which enzyme activities have been detected in
cell extracts (this work; M. Chistoserdova, unpublished data), (12)
aconitase, (13) isocitrate dehydrogenase and (14) succinyl-CoA
synthetase or succinyl-CoA hydrolase.
http://mic.sgmjournals.org
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
607
S. J. Van Dien and others
DISCUSSION
To further extend our understanding of central metabolic
pathways in M. extorquens AM1, we have undertaken a
transposon mutagenesis screen to isolate mutants defective
in growth on pyruvate and/or succinate. In addition, other
genes predicted to be involved in growth on multicarbon
compounds have been analysed by directed mutation.
With the results presented here in combination with those
from a recent paper (Van Dien & Lidstrom, 2002), we
now have a nearly complete genetic and biochemical
characterization of the TCA cycle and anapleurotic
pathways and can thus reconstruct this region of central
metabolism. The growth phenotypes of mutants in the genes
encoding the various steps of these pathways are summarized in Fig. 4. As predicted previously (Taylor & Anthony,
1976; Anthony, 1982), during growth on methanol a
complete TCA cycle is not required, as evidenced by the
wild-type growth rate of the 2-oxoglutarate dehydrogenase
mutant M38-24. The enzymes leading to the biosynthetic
precursor 2-oxoglutarate are necessary, as are those leading
from succinate to malate because they form part of the
essential pathway for the conversion of acetyl-CoA to
glyoxylate (Korotkova et al., 2002). Pyruvate can be formed
either from malate by the NAD-dependent malic enzyme or
from phosphoenolpyruvate via pyruvate kinase, so nullmutants in either of these enzymes have no growth defect on
methanol (Van Dien & Lidstrom, 2002). Likewise, acetylCoA is formed from the serine cycle (Anthony, 1982),
so pyruvate dehydrogenase is not required during
methylotrophic growth.
During growth on either succinate or pyruvate a functional
TCA cycle is required, as indicated previously (Taylor &
Anthony, 1976). The inability of the pyruvate dehydrogenase null-mutant AM1-PDH1 to grow on multicarbon
compounds confirms that pyruvate dehydrogenase is the
primary means of generating the acetyl-CoA required to
drive the TCA cycle (Bolbot & Anthony, 1980). With
succinate as the substrate, pyruvate must be formed either
by malic enzyme or from oxaloacetate by a combination of
phosphoenolpyruvate carboxykinase and pyruvate kinase.
Based upon the relative activities of these enzymes, we have
predicted malic enzyme to be the primary route for pyruvate
synthesis (Van Dien & Lidstrom, 2002). Labelling studies
have demonstrated that with pyruvate as the growth
substrate, significant flux occurs through the TCA cycle
(Salem et al., 1973). However, the anapleurotic enzymes are
necessary to replenish TCA cycle intermediates that are lost
to biosynthesis. This requirement can be fulfilled either by
phosphoenolpyruvate synthase followed by phosphoenolpyruvate carboxylase or by malic enzyme.
The final genes of interest detected in this study are
those predicted to encode the various subunits of the
NADH : ubiquinone oxidoreductase. This enzyme forms an
integral part of energy metabolism during heterotrophic
growth and is thus important to understanding the energy
608
and redox balance of the cell. The oxidation of NADH
by this enzyme complex is the first step of oxidative
phosphorylation, and is necessary for the conversion of
reducing power, in the form of NADH, to energy in the form
of ATP. A metabolic model of M. extorquens AM1 predicts
that the entry of NADH into oxidative phosphorylation is
important during growth on succinate and pyruvate, but
not on methanol (Van Dien & Lidstrom, 2002). According
to the model, methanol oxidation by methanol dehydrogenase (Lidstrom, 1992) produces sufficient reduced
cytochrome, which enters the oxidative phosphorylation
chain below NADH, so that NAD(P)H is more valuable to
the cell for biosynthetic needs than for energy production.
The growth phenotype data presented here agree with
the prediction. An insertional mutant in a putative
NADH : ubiquinone oxidoreductase gene grew normally
on methanol and showed impaired growth on succinate
and pyruvate. A similar heterotrophic phenotype of
NADH : ubiquinone oxidoreductase mutants has been
observed in Rhodobacter capsulatus, with impaired aerobic
growth on malate and succinate (Dupuis et al., 1998).
The definition of the steps in C3 and C4 metabolism and
their relationship to methylotrophy now provides a framework within which to assess growth on both multicarbon
and single carbon compounds in M. extorquens, a necessary
step for metabolic engineering of central metabolism. In
addition, the pool of transposon insertion mutants with
altered growth on C3 and/or C4 compounds is now available
for further in-depth analysis of central heterotrophic
metabolism in this bacterium.
ACKNOWLEDGEMENTS
This work was supported by a grant from the National Institutes
of Health (GM58933). We thank D. D’Argenio, L. Gallagher and
C. Manoil for providing the ISphoA|hah-Tc delivery strain.
REFERENCES
Anthony, C. (1982). The Biochemistry of Methylotrophs. London:
Associated Press.
Arps, P. J., Fulton, G. F., Minnich, E. C. & Lidstrom, M. E. (1993).
Genetics of serine pathway enzymes in Methylobacterium extorquens
AM1: phosphoenolpyruvate carboxylase and malyl coenzyme A lyase.
J Bacteriol 175, 3776–3783.
Attwood, M. M. & Harder, W. (1972). A rapid and specific enrichment
procedure for Hyphomicrobium spp. Antonie Van Leeuwenhoek 38,
369–377.
Bolbot, J. A. & Anthony, C. (1980). The metabolism of 1,2-
propanediol by the facultative methylotroph Pseudomonas AM1.
J Gen Microbiol 120, 245–254.
Chistoserdov, A. Y., Chistoserdova, L. V., McIntire, W. S. &
Lidstrom, M. E. (1994). The genetic organization of the mau gene
cluster in Methylobacterium extorquens AM1: complete nucleotide
sequence and generation and characteristics of mau mutants.
J Bacteriol 176, 4052–4065.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
Microbiology 149
Methylobacterium extorquens C3 and C4 metabolism
Chistoserdova, L. (1996). Metabolism of formaldehyde in
M. extorquens AM1: molecular genetic analysis and mutant characterization. In Microbial Growth on C1 Compounds, pp. 16–24. Edited
by M. E. Lidstrom & F. R. Tabita. Dordrecht: Kluwer Academic.
Chistoserdova, L. & Lidstrom, M. E. (1997). Identification and
mutation of a gene required for glycerate kinase activity from
a facultative methylotroph, Methylobacterium extorquens AM1.
J Bacteriol 179, 4946–4948.
Chun, K. T., Edenberg, H. J., Kelley, M. R. & Goebl, M. G. (1997).
Rapid amplification of uncharacterized transposon-tagged DNA
sequences from genomic DNA. Yeast 13, 233–240.
D’Argenio, D. A., Gallagher, L. A., Berg, C. A. & Manoil, C. (2001).
Drosophila as a model host for Pseudomonas aeruginosa infection.
J Bacteriol 183, 1466–1471.
Dupuis, A., Darrouzet, E., Duborjal, H., Pierrard, B., Chevallet, M.,
van Belzen, R., Albracht, S. P. & Lunardi, J. (1998). Distal genes
identified by using transposon mutagenesis including a putative
dihydromethanopterin reductase. J Bacteriol 185, 669–673.
Miller, V. L. & Mekalanos, J. J. (1988). A novel suicide vector and its
use in construction of insertion mutations: osmoregulation of outer
membrane proteins and virulence determinants in Vibrio cholerae
requires toxR. J Bacteriol 170, 2575–2583.
Saito, H. & Miura, K.-I. (1963). Preparation of transforming
deoxyribonucleic acid by phenol treatment. Biochim Biophys Acta
72, 619–629.
Salem, A. R., Wagner, C., Hacking, A. J. & Quayle, J. R. (1973). The
metabolism of lactate and pyruvate by Pseudomonas AM1. J Gen
Microbiol 76, 375–388.
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning:
a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold
Spring Harbor Laboratory.
of the nuo operon of Rhodobacter capsulatus equivalent to the
mitochondrial ND subunits are all essential for the biogenesis of the
respiratory NADH-ubiquinone oxidoreductase. Mol Microbiol 28,
531–541.
Flint, D. H. (1994). Initial kinetic and mechanistic characterization
of Escherichia coli fumarase A. Arch Biochem Biophys 311, 509–516.
Schreyer, R. & Bock, A. (1980). Phosphoglucose isomerase from
Green, L. S., Li, Y., Emerich, D. W., Bergersen, F. J. & Day, D. A.
(2000). Catabolism of a-ketoglutarate by a sucA mutant of
Taylor, I. J. & Anthony, C. (1976). A biochemical basis for obligate
Bradyrhizobium japonicum: evidence for an alternative tricarboxylic
acid cycle. J Bacteriol 182, 2838–2844.
Korotkova, N., Chistoserdova, L., Kuksa, V. & Lidstrom, M. E.
(2002). Glyoxylate regeneration pathway in the methylotroph
Methylobacterium extorquens AM1. J Bacteriol 184, 1750–1758.
Lidstrom, M. E. (1992). The genetics and molecular biology of
methanol-utilizing bacteria. In Methane and Methanol Utilizers,
pp. 183–206. Edited by J. C. Murrell & H. Dalton. New York:
Plenum.
Marx, C. J. & Lidstrom, M. E. (2001). Development of improved
versatile broad-host-range vectors for use in methylotrophs and
other Gram-negative bacteria. Microbiology 147, 2065–2075.
Marx, C. J., O’Brien, B. N., Breezee, J. & Lidstrom, M. E. (2003).
Novel methylotrophy genes of Methylobacterium extorquens AM1
http://mic.sgmjournals.org
Escherichia coli K10: purification, properties and formation under
aerobic and anaerobic condition. Arch Microbiol 127, 289–298.
Simon, R., Priefer, U. & Puhler, A. (1983). A broad host range
mobilization system for in vivo genetic engineering: transposon
mutagenesis in Gram-negative bacteria. Bio/Technology 1, 784–791.
methylotrophy: properties of a mutant of Pseudomonas AM1 lacking
2-oxoglutarate dehydrogenase. J Gen Microbiol 93, 259–265.
Thissen, J., Desai, S., McCartney, P. & Komuniecki, R. (1986).
Improved purification of the pyruvate dehydrogenase complex from
Ascaris suum body wall muscle and characterization of PDHa kinase
activity. Mol Biochem Parasitol 21, 129–138.
van der Oost, J., van Walraven, H. S., Bogerd, J., Smit, A. B., Ewart,
G. D. & Smith, G. D. (1989). Nucleotide sequence of the gene
proposed to encode the small subunit of the soluble hydrogenase of
the thermophilic unicellular cyanobacterium Synechococcus PCC
6716. Nucleic Acids Res 17, 10098.
Van Dien, S. J. & Lidstrom, M. E. (2002). Stoichiometric model for
evaluating the metabolic capabilities of the facultative methylotroph
Methylobacterium extorquens AM1, with application to reconstruction of C3 and C4 metabolism. Biotechnol Bioeng 78, 296–312.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 18:55:24
609