Overproduction of l-Lysine from Methanol by Methylobacillus glycogenes Derivatives Carrying a Plasmid with a Mutated dapA Gene

Abstract

The dapA gene, encoding dihydrodipicolinate synthase (DDPS) partially desensitized to inhibition by l-lysine, was cloned from an l-threonine- and l-lysine-coproducing mutant of the obligate methylotroph Methylobacillus glycogenes DHL122 by complementation of the nutritional requirement of an Escherichia coli dapA mutant. Introduction of the dapA gene into DHL122 and AL119, which is the parent of DHL122 and an l-threonine producing mutant, elevated the specific activity of DDPS 20-fold and l-lysine production 2- to 3-fold with concomitant reduction of l-threonine in test tube cultures. AL119 containing the dapA gene produced 8 g of l-lysine per liter in a 5-liter jar fermentor from methanol as a substrate. Analysis of the nucleotide sequence of the dapA gene shows that it encodes a peptide with an M_r of 30,664 and that the encoded amino acid sequence is extensively homologous to those of other organisms. In order to study the mutation that occurred in DHL122, the dapA genes of the wild type and AL119 were cloned and sequenced. Comparison of the nucleotide sequences of the dapA genes revealed that the amino acid at residue 88 was F in DHL122 whereas it was L in the wild type and AL119, suggesting that this amino acid alteration that occurred in DHL122 caused the partial desensitization of DDPS to the inhibition by l-lysine. The similarity in the amino acid sequences of DDPS in M. glycogenes and other organisms suggests that the mutation of the dapA gene in DHL122 is located in the region concerned with interaction of the allosteric effector, l-lysine.

Methanol, a compound easily synthesized from natural gas, is an attractive raw material for microbial industries. Using methanol as a carbon source, production costs could be greatly reduced and purification and waste treatment processes could be simplified. A number of production processes for useful compounds with methylotrophs (methanol-utilizing microorganisms) have been studied. Production of single-cell protein by a gram-negative methylotroph, Methylophilus methylotrophus, was extensively studied in the 1970s and finally industrialized (1). Efficient production systems for recombinant proteins were also constructed with the methanol-utilizing yeast Pichia pastoris (4).

Many attempts have also been made to use methanol in amino acid production; however, successful studies were limited. Lee et al. (9) reported the production of 47 g of l-lysine per liter by a gram-positive methylotroph, Bacillus methanolicus. Izumi et al. (7) reported efficient conversion of glycine to l-serine by a gram-negative methylotroph, Hyphomicrobium methylovorum. Breeding of amino acid-producing mutants requires isolation of mutants desensitized in the feedback regulation of the biosynthetic enzymes for a desired amino acid and blocked in metabolic pathways of by-products. Isolation of mutants from methylotrophs is usually difficult, due to unknown reasons, preventing the use of methanol in amino acid production.

We isolated l-glutamic acid-hyperproducing mutants from the obligate methylotrophs Methylobacillus glycogenes ATCC 21276 and 21371 and subsequently derived l-lysine- and l-threonine-producing mutants from l-glutamic acid-producing mutants (13). An l-threonine-producing mutant, AL119, was isolated among mutants resistant to both dl-α-amino-β-hydroxy-valeric acid and l-lysine, and an l-lysine- and l-threonine-coproducing mutant, DHL119, was isolated among 2,6-diamino-4-hexenoic acid hydrochloride-resistant mutants derived from AL119. Enzymatic analysis revealed that aspartokinase, the common regulatory enzyme in the biosynthesis of l-lysine and l-threonine, was insensitive to the feedback inhibition by l-lysine in AL119 and DHL122 and that dihydrodipicolinate synthase (DDPS), the key regulatory enzyme located at the branch point of l-lysine and l-threonine biosynthesis, was partially insensitive to the feedback inhibition by l-lysine in DHL122 (14). We concluded that desensitization of these regulatory enzymes led to the production of l-threonine and l-lysine.

As is the case for other gram-negative methylotrophs, isolation of mutants from M. glycogenes was not easy. In order to efficiently enhance amino acid production, we tried to employ a recombinant technology. The hom and thrC genes, which encode homoserine dehydrogenase and threonine synthase in the l-threonine biosynthesis pathway, respectively, were cloned from M. glycogenes (15) and introduced into l-threonine-producing mutants. The strains amplified with these genes produced more l-threonine than their parents (16). We reasoned that amplification of the dapA gene of DHL122, which encodes desensitized DDPS in DHL122, might greatly elevate l-lysine production. We report here the effects of the dapA gene of DHL122 on amino acid production from methanol by M. glycogenes and the analysis of the nucleotide sequence of the dapA gene.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and plasmids used in this study are listed in Table 1.

TABLE 1.

Bacterial strains and plasmids used in this study^a

Strain or plasmid	Genotype or relevant characteristics	Phenotype	Source or reference
M. glycogenes
ATCC 21276		Wild type
AL119	l-Threonine-producing mutant isolated among AHVrl-Lysr mutants derived from ATCC 21276; AKis	AHVrl-Lysr	13
DHL122	l-Threonine- and l-lysine-coproducing mutant isolated from DHLr mutants derived from AL119; AKis DDPSis	AHVrl-Lysr DHLr	13
E. coli
AT997	dapA thi-1 relA spoT1	Dap−	20
DH5α	DH1 Δ(argF-lacZYA)U169 hsdR17 φ80dlacZΔM15
S17-1	recA; chromosomally integrated RP4 derivative	Pro− Tpr	19
Plasmids
pUC19	Vector, 2.7 kb	Ampr	Takara Shuzo
pMFY42	Broad-host-range plasmid, 10.9 kb	Tcr Kmr	5
pMW119	Vector, 4.2 kb	Ampr	Nippon gene
pDYO1	Derivative of pUC19 containing 3.9-kb DNA fragment carrying the dapA gene cloned from M. glycogenes DHL122	Ampr DDPS	This study
pDYO2	Derivative of pUC19 containing 1.7-kb EcoRI DNA fragment carrying the dapA gene	Ampr DDPS	This study
pDYO4-1, pDYO4-2	Derivative of pUC19 containing 2.4-kb PstI DNA fragment carrying the dapA gene	Ampr DDPS	This study
pDYO6	Derivative of pDYO1 lacking the 0.4-kb EcoRV DNA fragment of pDYO1	Ampr	This study
pDYOM4-2	Derivative of pMFY42 containing 2.4-kb PstI DNA fragment carrying the dapA gene	Tcr DDPS	This study

^aAbbreviations: AHV^r l-Lys^r, resistance to both dl-α-amino-β-hydroxy-valeric acid and l-lysine; AK^is, aspatokinase insensitive to the feedback inhibition by l-lysine and l-threonine; Amp^r, ampicillin resistance; Dap⁻, diaminopimelic acid auxotrophy; DDPS, conferring the dihydrodipicolinate synthase activity; DDPS^is, dihydrodipicolinate synthase partially insensitive to the feedback inhibition by l-lysine; DHL^r, resistance against 2,6-diamino-4-hexenoic acid hydrochloride; Km^r, kanamycin resistance; Pro⁻, proline auxotrophy; Tc^r, tetracycline resistance; Tp^r, trimethoprim resistance. 

Media and culture methods.

M. glycogenes ATCC 21276 and the derived mutants were cultivated as described by us previously (13). Methanol was used as a sole carbon source. Escherichia coli cells were cultivated as described previously (10). M9S5 medium (M9 medium [10] supplemented with 50 mg each of 19 amino acids [except l-lysine] per liter) was used for complementation testing of E. coli AT997. Tetracycline (10 mg/liter) or ampicillin (100 mg/liter) was used to supplement liquid or agar media to cultivate M. glycogenes and E. coli strains containing plasmids.

DNA manipulation.

DNA restriction enzyme digestion, separation of DNA fragments by gel electrophoresis, and transformation of E. coli strains were performed by standard methods as described previously (10). Restriction enzymes and DNA ligase were supplied by Takara Shuzo Co. Ltd. (Kyoto, Japan). Southern hybridization and colony hybridization were done with a DNA labeling and detection kit (Boehringer Mannheim) according to the method recommended by the supplier. Chromosomal DNAs of M. glycogenes strains were prepared by the method described by Marmur (11).

Cloning of the dapA gene.

The dapA gene of DHL122 was isolated by complementation of E. coli AT997 with a lesion in dapA as follows. The chromosomal DNA of M. glycogenes DHL122 was partially digested with Sau3AI and separated by agarose gel electrophoresis. The 2- to 6-kb DNA fragments purified from the gel were ligated into the BamHI site of pUC19 to construct the gene library and transform E. coli AT997. The cells were plated onto M9S5 medium supplemented with 100 mg of ampicillin per liter and incubated overnight at 37°C. Plasmids isolated from the resulting colonies were further analyzed. The dapA genes of the wild type and AL119 were isolated by colony hybridization with the dapA gene of DHL122 as a probe. The chromosomal DNAs of M. glycogenes ATCC 21276 and AL119 were completely digested with EcoRI and separated by 10 to 40% sucrose density gradient centrifugation. The 2- to 4-kb DNA fragments were collected, ligated into the EcoRI site of pMW119, and transformed into E. coli DH5. The cells were plated onto L broth (10) supplemented with 100 mg of ampicillin per liter, and the resulting colonies were examined by colony hybridization with the 1.7-kb EcoRI fragment containing the dapA gene of DHL122. Plasmids isolated from the positive colonies were analyzed.

Subcloning of plasmids.

The 1.7-kb EcoRI fragment of pDYO1 was introduced into the EcoRI site of pUC19 to construct pDYO2. The 2.4-kb PstI fragment of pDYO1 was inserted into the PstI site of pUC19 to form pDYO4-1 and pDYO4-2, in both orientations. pDYO6 was formed by the self-ligation of the 6.3-kb EcoRV fragment of pDYO1. The 2.4-kb PstI fragment of pDYO4-2 was inserted into the PstI site of pMFY42 to construct pDYOM4-2.

Conjugation.

Introduction of plasmids from E. coli S17-1 to M. glycogenes by conjugation was done as described previously (16).

DNA sequencing and analysis.

DNA sequencing was performed by the dideoxy chain termination method on both strands with an Applied Biosystems model 373A sequencer. DNA sequences analyses and homology alignments of amino acid sequences were carried out using GENETYX MAC version 9.0 (Software Development Co. Ltd., Tokyo, Japan).

Analysis.

Bacterial growth was measured by the increase in absorbance at 660 nm, and amino acids in culture supernatants were measured by high-pressure liquid chromatography as described previously (13). Cell extracts of the M. glycogenes transconjugants were prepared and DDPS activities were measured as described previously (14).

Nucleotide sequence accession number.

The nucleotide sequence of the 1.7-kb EcoRI fragment encoding DDPS of M. glycogenes ATCC 21276 will appear in the EMBL, GenBank, and DDBJ nucleotide sequence data libraries under accession no. AB038266.

RESULTS

Cloning of the dapA gene of M. glycogenes DHL122.

We attempted to isolate the dapA gene of M. glycogenes DHL122 by complementation of an E. coli DDPS-deficient mutant, E. coli AT997, which has a lesion in dapA and requires l-diaminopimelic acid for growth (20). This strain was transformed by the gene library of M. glycogenes DHL122 in a vector plasmid, pUC19, and the cells were inoculated onto M9S5 medium that did not contain l-diaminopimelic acid. A plasmid with a 3.9-kb insert, designated pDYO1, was obtained from a colony that appeared on the M9S5 medium. It complemented the nutritional requirement of AT997 by retransformation and was analyzed further. Subcloning of pDYO1 and complementation analysis with AT997 revealed that the dapA gene was localized in the 1.7-kb EcoRI fragment (Fig. 1). Both pDYO4-1 and pDYO4-2, which had the same 2.4-kb PstI fragment in both orientations, complemented E. coli AT997, suggesting that the promoter of the dapA gene was included in the fragment and was functional in E. coli.

FIG. 1.

Restriction maps of the subcloned DNA fragments containing the dapA gene isolated from M. glycogenes DHL122. The arrow indicates the direction of the ORF of the dapA gene as determined by the nucleotide sequence. Complementation activities conferred by each plasmid in E. coli dapA mutant AT997 are indicated by + or −. P lac and P km indicate the locations of the lac promoter of pUC19 and the promoter of the kanamycin resistance gene of pMFY42, respectively. B, BamHI; EI, EcoRI; EV, EcoRV; Ps, PstI; Sp, SphI.

Effects of the dapA gene of M. glycogenes DHL122 on amino acid production from methanol.

In order to examine the effect of the dapA gene on amino acid production from methanol, the 2.4-kb PstI fragment of pDYO4-2 was inserted into the PstI site of a broad-host-range plasmid, pMFY42. The plasmid thus constructed, pDYOM4-2, was introduced into E. coli S17-1, a strain which could mobilize plasmids from E. coli to other gram-negative microorganisms (19), and then E. coli S17-1 containing the plasmid was conjugated with M. glycogenes DHL122 and its parental l-threonine producer, AL119, to transfer the plasmid. The transconjugants containing the vector were constructed in the same manner. Restriction analysis of the plasmids isolated from the M. glycogenes transconjugants showed the same structures as in the E. coli transformants from which the plasmids were transferred (data not shown).

Introduction of the dapA gene elevated the DDPS activities about 20-fold and reduced the growth of DHL122 and AL119 in test tube cultures (Table 2). The l-lysine in the culture supernatants increased two- to threefold, whereas the accumulation of l-threonine and l-glutamic acid was reduced. We thought that the growth reduction of the strains with the dapA genes might be caused by a shortage of cofactors, such as ATP and NADPH, the compounds required for both l-lysine biosynthesis and microbial growth. The enhanced l-lysine biosynthesis caused by the elevation of the DDPS activity might consume a considerable amount of cofactors. As sufficient amounts of the cofactors might not be generated due to the limited oxygen supply in test tube cultures, this might cause the growth to be reduced.

TABLE 2.

Effects of the dapA gene on amino acid production by the M. glycogenes mutants in test tube cultures^a

Host	Plasmid	Insert	Optical density at 660 nm	DDPS sp actb	Production (g/liter) of the following amino acid:
					Glu	Thr	Lys
DHL122	pMFY42		12.7	0.030 (1.0)	6.0	0.6	0.6
	pDYOM4-2	dapA	6.3	0.580 (19.3)	4.7	0.3	1.2
AL119	pMFY42		8.7	0.055 (1.0)	6.3	0.9	0.4
	pDYOM4-2	dapA	5.8	1.200 (21.0)	5.1	0.1	1.1

^aCells were cultivated in test tubes at 30°C for 48 h. 

^bSpecific activities of DDPS expressed as change in A₅₄₀ per minute per milligram of protein. Numbers in parentheses show relative activities. 

To circumvent the drawback of test tube cultures, we examined the effect of the dapA gene by cultivating the strains in 5-liter jar fermentors, where the oxygen supply was more favorable than in test tube cultures. Figure 2 shows the time courses of cultivation of the AL119 and DHL122 strains with or without the dapA gene. The growth of the strains with the dapA gene was comparable to that of the strains without the dapA gene in 5-liter jar fermentors. l-Lysine accumulation was greatly enhanced by the introduction of the dapA gene and was accompanied by the reduction of l-threonine accumulation as in test tube cultures. DHL122 containing pDYOM4-2 and AL119 harboring pDYOM4-2 accumulated 5.3 and 8 g of l-lysine per liter, respectively, after 72 h of cultivation. All of the transconjugants constructed produced more than 30 g of l-glutamic acid per liter, which was much higher than the accumulation of l-lysine and l-threonine, in 5-liter jar fermentors.

FIG. 2.

Time courses of amino acid production in 5-liter jar fermentors. (A) DHL122/pMFY42; (B) DHL122/pDYOM4-2; (C) AL119/pMFY42; (D) AL119/pDYOM4-2. ♦, optical density (OD) at 660 nm; ▪, Lys; □, Thr; ×, Glu.

Analysis of the nucleotide sequence of dapA and the location of the mutation in DHL122.

The 1.7-kb EcoRI fragment of pDYO2 was sequenced on both strands, and a single open reading frame (ORF) was found (Fig. 3). The ORF encodes a predicted peptide with an M_r of 30,664, initiating at the ATG codon (nucleotides 472 to 474) and terminating at the TGA codon (nucleotides 1342 to 1344). Neither distinct promoter-like nor distinct terminator-like sequences were found in the upstream and downstream regions of the ORF, respectively. To identify the mutation in DHL122, the 1.7-kb EcoRI fragments containing the dapA gene were cloned from the wild type, ATCC 21276, and AL119 and sequenced. The nucleotide sequences of the 1.7-kb EcoRI fragments from ATCC 21276 and AL119 were identical to that from DHL122 except at nucleotide 733. Nucleotide 733 is C in ATCC 21276 and AL119, whereas it is T in DHL122, which altered the amino acid residue L88 in ATCC 21276 and AL119 to F88 in DHL122. The mutation in this amino acid residue was considered to cause the partial desensitization of DDPS of DHL122 to feedback inhibition by l-lysine (14).

FIG. 3.

Nucleotide sequence of the 1.7-kb EcoRI fragment containing the dapA gene of M. glycogenes ATCC 21276. The putative amino acid sequence is shown below the nucleotide sequence. The start codon, ATG, and the stop codon, TGA, are underlined. The location of the mutation found in DHL122 (nucleotide 733) is indicated.

Comparison of the predicted amino acid sequences of the dapA genes.

The predicted amino acid sequence of the dapA gene of M. glycogenes was compared with those from other organisms (Fig. 4). Extensive amino acid sequence homology was found between M. glycogenes and other organisms. The amino acid sequence of M. glycogenes has identities of 52.4% (153 of 292 amino acid residues), 47.2% (137 of 290 amino acid residues), 35.5% (107 of 301 amino acid residues), 29.4% (96 of 326 amino acid residues), and 25.3% (96 of 380 amino acid residues) with those from E. coli, Bacillus subtilis, Corynebacterium glutamicum, Nicotiana sylvestris, and Zea mays, respectively.

FIG. 4.

Homology between the amino acid sequences of the DDPSs of M. glycogenes ATCC 21276 and those of other organisms. Conserved amino acid residues found in more than four species are shaded. Asterisks indicate the putative active sites of the DDPS. Amino acid residue 88, which was found to be mutated in DHL122, is indicated as F(DHL122) above the sequence. The amino acid residues whose mutations were reported to cause the desensitization of DDPSs of other organisms are indicated by plus signs: E. coli, A81→V and H118→Y (8); Z. mays, S157→N, E162→K, and A166→T,V (17); and N. sylvestris, N104→I (6). MG, M. glycogenes; EC, E. coli; BS, B. subtilis; CG, C. glutamicum; NS, N. sylvestris; ZM, Z. mays.

DISCUSSION

In our previous study we derived an l-lysine and l-threonine coproducer, DHL122, from an l-threonine producer, AL119 (13). However, DHL122 could produce only a small amount of l-lysine. We report here that the introduction of the dapA gene into DHL122 and AL119 greatly elevated the specific activity of DDPS and l-lysine production with concomitant reduction of l-threonine accumulation. This suggests that the carbon flow from l-aspartic acid to l-lysine was limited at the conversion of l-aspartic acid-β-semialdehyde to dihydrodipicolinate in both strains and was liberated by the amplification of the dapA gene with the redirection of the carbon flow from toward l-threonine to toward l-lysine. We found that DDPS of DHL122, from which the dapA gene was cloned, was partially desensitized to the feedback inhibition by l-lysine (14). The altered regulatory property of the enzyme was thought to contribute to the increased production of l-lysine. However, we believe that this effect was partial. DHL122, which had mutated DDPS, produced more l-lysine than AL119, which had wild-type DDPS, but the production of l-lysine by DHL122 was less than that of l-threonine, suggesting that the partial desensitization of DDPS was not enough to change the carbon flow from toward l-threonine to toward l-lysine (13) (Fig. 2A and C). The introduction of the dapA gene-containing plasmid drastically enhanced the enzyme activity, and the production of l-lysine greatly exceeded that of l-threonine (Fig. 2B and D). We speculate that the elevated DDPS activity mainly caused the overproduction of l-lysine in the strains with the dapA-containing plasmid.

AL119 and DHL122 were derived from an l-glutamic acid producer, iA111 (13). The transconjugants bearing the dapA gene produced much more l-glutamic acid than l-lysine and l-threonine. The metabolic flow from methanol is possibly limited at some steps in the biosynthesis of l-aspartic acid and directed toward the formation of l-glutamic acid. To further enhance l-lysine production, it is necessary to reduce the metabolic flow toward l-glutamic acid and direct it toward l-aspartic acid. In the course of breeding l-lysine-producing mutants from a C. glutamicum l-glutamic acid producer, pyruvate kinase deficiency (18), reduction of citrate synthase activity, and desensitization of phosphoenolpyruvate carboxylase (21) were reported to be effective to reduce the production of l-glutamic acid and enhance that of l-aspartic acid and l-lysine. These strategies may also be applicable for M. glycogenes. The isolation of mutants with lesions in the enzymes of the tricarboxylic acid cycle and those altered in the regulation of the enzymes of the biosynthesis of l-aspartic acid, together with the amplification of biosynthetic genes of l-aspartic acid, should be considered.

The dapA gene was cloned from M. glycogenes and sequenced, and the amino acid sequence was found to have extensive homology with those from other organisms. In E. coli (12) and N. sylvestris (2), DDPS was shown to consist of homotetramers by X-ray crystallography studies. The homology throughout the amino acid sequence between DDPS from M. glycogenes and those from E. coli and N. sylvestris suggests that DDPS of M. glycogenes has a similar structure. Blickling and Knäblein (3) proposed that in E. coli and other organisms K161, Y133, and R138 (numbering refers to the E. coli and M. glycogenes sequences) are involved in significant roles in catalysis—the formation of Schiff bases with pyruvate, proton shuttling during imine formation and transimination, and interaction with the carboxy group of l-aspartatic acid-β-semialdehyde, respectively. These residues are conserved in M. glycogenes and might have the same roles in catalysis as in other organisms. It was also suggested that in E. coli the residues H53, H56, Y106, Y107, N80, and E84 were involved in the interaction with an allosteric effector, l-lysine (3). Most of the mutations that have been reported to cause the desensitization of E. coli and plant DDPSs (marked by plus signs in Fig. 4) were found in the region between amino acid residues 79 and 88. The mutation in the DDPS of DHL122 (F88) is also located in this region, suggesting that the mutation that occurred in this region might alter the structure of DDPS, prevent the efficient interaction between l-lysine and other amino acid residues (possibly N80 and E84), and lead to the partial desensitization of DDPS to l-lysine. DDPS of DHL122 was shown to be inhibited by high concentrations of l-lysine (14). The DDPS might be further desensitized by alteration of other amino acid residues that were found to be effective for the desensitization of other DDPSs. The availability of more desensitized DDPSs constructed in such ways will be an important tool to improve l-lysine production from methanol by M. glycogenes.