Engineering the serine/threonine protein kinase Raf-1 to utilise an orthogonal analogue of ATP substituted at the N6 position

One key area of protein kinase research is the identification of cognate substrates. The search for substrates is hampered by problems in unambiguously assigning substrates to a particular kinase in vitro and in vivo. One solution to this impasse is to engineer the kinase of interest to accept an ATP analogue which is orthogonal (unable to fit into the ATP binding site) for the wild-type enzyme and the majority of other kinases. The acceptance of structurally modified, gamma-(32)P-labelled, nucleotide analogue by active site-modified kinase can provide a unique handle by which the direct substrates of any particular kinase can be displayed in crude mixtures or cell lysates. We have taken this approach with the serine/threonine kinase Raf-1, which plays an essential role in the transduction of stimuli through the Ras-->Raf-->MEK-->ERK/MAP kinase cascade. This cascade plays essential roles in proliferation, differentiation and apoptosis. Here we detail the mutagenesis strategy for the ATP binding pocket of Raf-1, such that it can utilise an N(6)-substituted ATP analogue. We show that these mutations do not alter the substrate specificity and signal transduction through Raf-1. We screen a library of analogues to identify which are orthogonal for Raf-1, and show that mutant Raf-1 can utilise the orthogonal analogue N(6)(2-phenethyl) ATP in vitro to phosphorylate its currently only accepted substrate MEK. Importantly we show that our approach can be used to tag putative direct substrates of Raf-1 kinase with (32)P-N(6)(2-phenethyl) ATP in cell lysates.     

Synthesis of a reactive ATP analog and its preliminary application as an affinity labeling reagent

A reactive ATP analog, N6-(6-bromoacetamidohexyl)-AMP.PCP, was synthesized in an attempt to covalently label the binding sites for adenine nucleotides, especially ATP, of various enzymes which utilize adenine nucleotides as substrates, cofactors, inhibitors or allosteric effectors. This reagent rapidly inactivated rabbit muscle glyceraldehyde 3-phosphate dehydrogenase (GPD), myokinase (MK), and creatine kinase (CK) under very mild conditions. Adenine nucleotide substrates prevented the inactivation. In the case of GPD, complete inactivation was observed when 1 mol of the reagent per mol of enzyme subunit was incorporated into the enzyme. These results indicate that the present ATP analog may be useful as an affinity labeling reagent for various adenine nucleotide-dependent enzymes.     

Src-Abl tyrosine kinase chimeras: replacement of the adenine binding pocket of c-Abl with v-Src to swap nucleotide and inhibitor specificities

Engineered protein kinases with unnatural nucleotide specificity and inhibitor sensitivity have been developed to trace kinase substrate targets. We first engineered unnatural nucleotide specificity into v-Src by mutating one residue, isoleucine 338, to alanine. This position is highly conserved among all kinases in the sense that it is always occupied by either a large hydrophobic residue or threonine. Because of the conservation of this residue and the highly conserved fold of the kinase family, we have attempted to generalize the engineering of all kinases on the basis of our success with v-Src. Although many kinases can be similarly engineered using v-Src as a blueprint, we encountered one kinase, c-Abl, which when mutated, does not display the ability to accept unnatural ATP analogues. To overcome this failure of the engineered c-Abl (T315A) to accept unnatural nucleotides, we developed a new strategy for introducing unnatural nucleotide specificity into kinases. We generated a chimeric kinase in which regions of the kinase domain of c-Abl were swapped with the corresponding regions of v-Src (I338A). Specifically, we engineered two chimeras in which the N-terminal lobe of the SH1 domain of c-Abl was swapped with that of v-Src. These kinase chimeras were found to have the same unnatural nucleotide specificity as that of v-Src (I338A), while retaining the peptide specificity of c-Abl. Thus, these chimeric kinases are suitable for identifying the direct substrates of c-Abl. These engineered chimeric enzymes provide a new strategy for constructing kinases with tailor-made ligand binding properties.     

Use of a chemical genetic technique to identify myosin IIb as a substrate of the Abl-related gene (Arg) tyrosine kinase

Abl family kinases have been implicated in the regulation of cell morphogenesis and migration, but the molecular mechanisms through which they operate are not fully elucidated. We applied the bump-hole technique, pioneered by Shokat and colleagues, to identify direct substrates of Abl and the Abl-related gene (Arg) kinases. This technique required the engineering of Abl/Arg to utilize an unnatural ATP analogue as a phospho-donor. Mutation of T334A and T361A in Abl and Arg, respectively, altered their nucleotide specificity and allowed them to utilize N6-benzyl-ATP as a phospho-donor. These mutations did not affect the catalytic activity or protein substrate specificity of Abl and Arg. An unexpected high level of background labeling necessitated further optimization of this approach. Dialysis, pretreatment with a broad-spectrum Ser/Thr kinase inhibitor, K-252a, and purification of phosphotyrosine-containing proteins allowed for definitive identification of putative substrates. Using mass spectrometry, we identified eight putative substrates. One of these putative substrates, myosin IIB, can be phosphorylated in vivo by Arg. Our results indicate that the bump-hole technique can be used to identify Abl family kinase substrates and suggests that myosin IIB may be regulated by tyrosine phosphorylation.     

Identification of ChChd3 as a novel substrate of the cAMP-dependent protein kinase (PKA) using an analog-sensitive catalytic subunit

Due to the numerous kinases in the cell, many with overlapping substrates, it is difficult to find novel substrates for a specific kinase. To identify novel substrates of cAMP-dependent protein kinase (PKA), the PKA catalytic subunit was engineered to accept bulky N(6)-substituted ATP analogs, using a chemical genetics approach initially pioneered with v-Src (1). Methionine 120 was mutated to glycine in the ATP-binding pocket of the catalytic subunit. To express the stable mutant C-subunit in Escherichia coli required co-expression with PDK1. This mutant protein was active and fully phosphorylated on Thr(197) and Ser(338). Based on its kinetic properties, the engineered C-subunit preferred N(6)(benzyl)-ATP and N(6)(phenethyl)-ATP over other ATP analogs, but still retained a 30 microm K(m) for ATP. This mutant recombinant C-subunit was used to identify three novel PKA substrates. One protein, a novel mitochondrial ChChd protein, ChChd3, was identified, suggesting that PKA may regulate mitochondria proteins.     

An ATP analog-sensitive version of the tomato cell death suppressor protein kinase Adi3 for use in substrate identification

Adi3 is a protein kinase from tomato that functions as a cell death suppressor and its substrates are not well defined. As a step toward identifying Adi3 substrates we developed an ATP analog-sensitive version of Adi3 in which the ATP-binding pocket is mutated to allow use of bulky ATP analogs. Met385 was identified as the "gatekeeper" residue and the M385G mutation allows for the use of two bulky ATP analogs. Adi3(M385G) can also specifically utilize N(6)-benzyl-ATP to phosphorylate a known substrate and provides a tool for identifying Adi3 substrates.     

Protein kinase C and its substrates in tumor promoter-sensitive and -resistant cells

Calcium- and phospholipid-dependent protein kinase C activity and substrates were characterized in cell lysates of preneoplastic JB6 cells, a model system of genetic variants for sensitivity to tumor promoter-induced neoplastic transformation. Protein kinase C activity was similar for sensitive and resistant variants, as measured by calcium- and phospholipid-dependent phosphorylation of an exogenous substrate (histone HIII). Of 13 endogenous protein kinase C substrates, identified by labeling proteins with [gamma-32P] ATP, at least two (80 and 23 kDa) are potential candidates for mediating events on the pathway for promotion of transformation. 32P incorporation into the 80-kDa protein kinase C substrate was stimulated by tetradecanoylphorbol acetate and correlated with phenotype: the highest incorporation was found in promotion-insensitive cells, an intermediate level in promotion-sensitive cells and the lowest in the transformed cells. The phosphorylation of an 80-kDa protein, found by labeling intact cells in monolayer growth with [32P]orthophosphate, was also stimulated by tetradecanoylphorbol acetate and correlated inversely with phenotype. The 80 kDa protein kinase C substrate from cell lysates and the 80-kDa phosphoprotein from intact cells appear to be identical, as indicated by peptide mapping with protease V8 from Staphylococcus aureus. This finding suggests that the 80-kDa substrate is relevant to promoter-induced signal transduction in the intact cell. The 23-kDa protein kinase C substrate exhibited a band shift in sodium dodecyl sulfate gels in response to another transformation promoter in JB6 cells, the calcium analog, lanthanum (Smith, B. M., Gindhart, T. D., and Colburn, N. H. (1986) Carcinogenesis 7, 1949-1956). In summary, there are no unique substrates that distinguish the variants. Quantitative differences in certain substrates or their phosphorylation may, however, account for the difference in promotion sensitivity among the variants.     

Identifying specific kinase substrates through engineered kinases and ATP analogs

Intracellular signaling by protein kinases controls many aspects of cellular biochemistry and physiology. Determining the direct substrates of protein kinases is important in understanding how these signaling enzymes exert their effect on cellular functions. One of the recent developments in this area takes advantage of the similarity in the ATP binding domains of protein kinases, where a few conserved amino acids containing large side chains come in close contact with the N-6 position of bound ATP. Mutation of one or more of these residues generates a "pocket" in the ATP binding site that allows the mutant kinase, but not other cellular kinases, to utilize analogs of ATP with bulky substituents synthesized onto the N-6 position. The use of such a mutated kinase and radiolabeled ATP analogs allows for the specific labeling of direct substrates of the kinase within a mixture of cellular proteins. We have recently reported the generation of "pocket" mutants of extracellular regulated kinase 2 (ERK2) and their use in the identification of two novel substrates of ERK2. In this report, we discuss the generation and characterization of ERK2 mutants that utilize analogs of ATP and describe the methodology used to identify ERK2-associated substrates. We also describe the direct labeling of ERK2 substrates in cell lysates. These methodologies can be adapted for use with other protein kinases to increase the understanding of intracellular signal transduction.     

Conformational restraint is a critical determinant of unnatural nucleotide recognition by protein kinases

This report describes the synthesis of N(4)-(benzyl) AICAR triphosphate, a conformationally restrained analogue of N(4)-(benzyl) ribavirin triphosphate. Both of these nucleotides were evaluated as phosphodonors for wild-type p38MAP kinase and T106G p38MAP kinase, a designed mutant with expanded nucleotide specificity. The conformationally restrained nucleotide, N(4)-(benzyl) AICAR triphosphate, is orthogonal to (not accepted as a substrate by) wild-type p38MAP kinase, in contrast to N(4)-(benzyl) ribavirin triphosphate. Furthermore, N(4)-(benzyl) AICAR triphosphate, is accepted as a substrate by T106G p38MAP kinase, in contrast to N(4)-(benzyl) ribavirin triphosphate. We hypothesize that the presence of an internal hydrogen bond in N(4)-(benzyl) AICAR and its absence in N(4)-(benzyl) ribavirin triphosphate is the main determinant for their differing structure-activity relationships.     

The binding of ATP and AMP to Escherichia coli adenylate kinase investigated by 1H and 15N NMR spectroscopy

[N6 15N]ATP and [N6 15N]AMP, complexed with E.coli adenylate kinase (AKe), were observed with 15N isotope-filtered NMR pulse sequences and 1H[15N] heterocorrelated experiments to determine differences between binding sites based on chemical shifts and competition by substrate analogs. The chemical shifts of the N6 amino proton and nitrogen signals changed significantly after mixing with adenylate kinase. Differences in chemical shifts between the bound ATP and AMP signals are slight. The response of these shifts to further addition of other substrates or Mg2+ supports the view that the unchelated nucleotides can bind to both the sites, whereas the metal complexed species are restricted to the MgATP/MgADP binding site.     

LRRK2 kinase activity is dependent on LRRK2 GTP binding capacity but independent of LRRK2 GTP binding

Leucine rich repeat kinase 2 (LRRK2) is a Parkinson's disease (PD) gene that encodes a large multidomain protein including both a GTPase and a kinase domain. GTPases often regulate kinases within signal transduction cascades, where GTPases act as molecular switches cycling between a GTP bound "on" state and a GDP bound "off" state. It has been proposed that LRRK2 kinase activity may be increased upon GTP binding at the LRRK2 Ras of complex proteins (ROC) GTPase domain. Here we extensively test this hypothesis by measuring LRRK2 phosphorylation activity under influence of GDP, GTP or non-hydrolyzable GTP analogues GTPγS or GMPPCP. We show that autophosphorylation and lrrktide phosphorylation activity of recombinant LRRK2 protein is unaltered by guanine nucleotides, when co-incubated with LRRK2 during phosphorylation reactions. Also phosphorylation activity of LRRK2 is unchanged when the LRRK2 guanine nucleotide binding pocket is previously saturated with various nucleotides, in contrast to the greatly reduced activity measured for the guanine nucleotide binding site mutant T1348N. Interestingly, when nucleotides were incubated with cell lysates prior to purification of LRRK2, kinase activity was slightly enhanced by GTPγS or GMPPCP compared to GDP, pointing to an upstream guanine nucleotide binding protein that may activate LRRK2 in a GTP-dependent manner. Using metabolic labeling, we also found that cellular phosphorylation of LRRK2 was not significantly modulated by nucleotides, although labeling is significantly reduced by guanine nucleotide binding site mutants. We conclude that while kinase activity of LRRK2 requires an intact ROC-GTPase domain, it is independent of GDP or GTP binding to ROC.     

Nucleotide binding to nucleoside diphosphate kinases: X-ray structure of human NDPK-A in complex with ADP and comparison to protein kinases

NDPK-A, product of the nm23-H1 gene, is one of the two major isoforms of human nucleoside diphosphate kinase. We analyzed the binding of its nucleotide substrates by fluorometric methods. The binding of nucleoside triphosphate (NTP) substrates was detected by following changes of the intrinsic fluorescence of the H118G/F60W variant, a mutant protein engineered for that purpose. Nucleoside diphosphate (NDP) substrate binding was measured by competition with a fluorescent derivative of ADP, following the fluorescence anisotropy of the derivative. We also determined an X-ray structure at 2.0A resolution of the variant NDPK-A in complex with ADP, Ca(2+) and inorganic phosphate, products of ATP hydrolysis. We compared the conformation of the bound nucleotide seen in this complex and the interactions it makes with the protein, with those of the nucleotide substrates, substrate analogues or inhibitors present in other NDP kinase structures. We also compared NDP kinase-bound nucleotides to ATP bound to protein kinases, and showed that the nucleoside monophosphate moieties have nearly identical conformations in spite of the very different protein environments. However, the beta and gamma-phosphate groups are differently positioned and oriented in the two types of kinases, and they bind metal ions with opposite chiralities. Thus, it should be possible to design nucleotide analogues that are good substrates of one type of kinase, and poor substrates or inhibitors of the other kind.     

Enzymatic incorporation of ATP and CTP analogues into the 3' end of tRNA

Structural analogues of adenosine 5'-triphosphate and cytidine 5'-triphosphate were investigated as substrates for ATP(CTP):tRNA nucleotidyl transferase. Eight out of 26 ATP analogues and six out of nine CTP analogues were incorporated into the 3' terminus of tRNA. In general, for the recognition of the substrates the modification of the cytidine is less critical than is the modification of adenosine. An isosteric substitution on the ribose residue is possible in both CTP and ATP. The free hydroxyls of these triphosphates can be replaced by an amino group or hydrogen atom without loss of substrate properties. Modifications of positions 1, 2, 6, and 8 on the adenine ring of ATP are not allowed whereas modification on positions 2, 4 and 5 on the cytosine ring of CTP are tolerated by the enzyme. No differences can be observed in the substrate properties of ATP(CTP):tRNA nucleotidyl transferase isolated from different sources. Methods for preparation of tRNA species, which are shortened at their 3' end by one or more nucleotides, and analytical procedures for characterisation of these modified tRNAs are described.     

Analysis of MDR1 P-glycoprotein conformational changes in permeabilized cells using differential immunoreactivity

The reactivity of the ATP-dependent multidrug transporter P-glycoprotein (Pgp) with the conformation-sensitive monoclonal antibody UIC2 is increased in the presence of Pgp transport substrates, ATP-depleting agents, or mutations that reduce the level of nucleotide binding by Pgp. We have investigated the effects of nucleotides and vinblastine, a Pgp transport substrate, on the UIC2 reactivity of Pgp in cells permeabilized by Staphylococcus aureus alpha-toxin. ATP, ADP, and nonhydrolyzable ATP analogues decreased the UIC2 reactivity; this effect was potentiated by vanadate, a nucleotide-trapping agent. The Hill number for the nucleotide-induced conformational transition was 2 for ATP and ADP but 1 for nonhydrolyzable ATP analogues. The Hill numbers for ATP and ADP were decreased to 1 by mutations in one of the two nucleotide binding sites of Pgp, whereas mutation of both sites greatly diminished the overall effect of nucleotides. Vinblastine reversed the decrease in the UIC2 reactivity brought about by all the nucleotides, including nonhydrolyzable analogues; this effect of vinblastine was blocked by vanadate. These data indicate that UIC2-detectable conformational changes of Pgp are driven by binding and debinding of nucleotides, that nucleotide hydrolysis affects the Hill number for its Pgp interactions, and that Pgp transport substrates promote nucleotide dissociation from Pgp. These findings are consistent with a conventional E1/E2 model that explains conformational transitions of a transporter protein through a series of linked equilibria.     

Functional role of "N" (nucleotide) and "P" (phosphorylation) domain interactions in the sarcoplasmic reticulum (SERCA) ATPase

Experimental perturbations of the nucleotide site in the N domain of the SR Ca2+ ATPase were produced by chemical derivatization of Lys492 or/and Lys515, mutation of Arg560 to Ala, or addition of inactive nucleotide analogue (TNP-AMP). Selective labeling of either Lys492 or Lys515 produces strong inhibition of ATPase activity and phosphoenzyme intermediate formation by utilization of ATP, while AcP utilization and reverse ATPase phosphorylation by Pi are much less affected. Cross-linking of the two residues with DIDS, however, drastically inhibits utilization of both ATP and AcP, as well as of formation of phosphoenzyme intermediate by utilization of ATP, or reverse phosphorylation by Pi. Mutation of Arg560 to Ala produces strong inhibition of ATPase activity and enzyme phosphorylation by ATP but has a much lower effect on enzyme phosphorylation by Pi. TNP-AMP increases the ATPase activity at low concentrations (0.1-0.3 microM), but inhibits ATP, AcP, and Pi utilization at higher concentration (1-10 microM). Cross-linking with DIDS and TNP-AMP binding inhibits formation of the transition state analogue with orthovanadate. It is concluded that in addition to the binding pocket delimited by Lys 492 and Lys515, Arg560 sustains an important and direct role in nucleotide substrate stabilization. Furthermore, the effects of DIDS and TNP-AMP suggest that approximation of N (nucleotide) and P (phosphorylation) domains is required not only for delivery of nucleotide substrate, but also to favor enzyme phosphorylation by nucleotide and nonnucleotide substrates, in the presence and in the absence of Ca2+. Domain separation is then enhanced by secondary nucleotide binding to the phosphoenzyme, thereby favoring its hydrolytic cleavage.     

Chemically reprogramming the phospho‐transfer reaction to crosslink protein kinases to their substrates

The proteomic mapping of enzyme–substrate interactions is challenged by their transient nature. A method to capture interacting protein kinases in complexes with a single substrate of interest would provide a new tool for mapping kinase signaling networks. Here, we describe a nucleotide‐based substrate analog capable of reprogramming the wild‐type phosphoryl‐transfer reaction to produce a kinase‐acrylamide‐based thioether crosslink to mutant substrates with a cysteine nucleophile substituted at the native phosphorylation site. A previously reported ATP‐based methacrylate crosslinker (ATP‐MA) was capable of mediating kinase crosslinking to short peptides but not protein substrates. Exploration of structural variants of ATP‐MA to enable crosslinking of protein substrates to kinases led to the discovery that an ADP‐based methacrylate (ADP‐MA) crosslinker was superior to the ATP scaffold at crosslinking in vitro. The improved efficiency of ADP‐MA over ATP‐MA is due to reduced inhibition of the second step of the kinase–substrate crosslinking reaction by the product of the first step of the reaction. The new probe, ADP‐MA, demonstrated enhanced in vitro crosslinking between the Src tyrosine kinase and its substrate Cortactin in a phosphorylation site‐specific manner. The kinase–substrate crosslinking reaction can be carried out in a complex mammalian cell lysate setting, although the low abundance of endogenous kinases remains a significant challenge for efficient capture.     

Preparation and isolation of dansyladenylates as fluorescent substrates in reactions of the energy metabolism

A method was developed for preparation of dansylated derivatives of adenine nucleotides characterized by fluorescence when being irradiated with UV-light. The involvement of dansylated ATP, ADP and AMP as substrate analogues in energy metabolism is demonstrated in the ATPase, hexokinase, pyruvate kinase and adenylate kinase reactions. The kinetics of the reactions is discussed.     

Bovine spermatozoa incorporate 32Pi into ADP by an unknown pathway.

Intact ejaculated bovine sperm incorporate 32Pi into ADP to a specific activity two to three times higher than into ATP. This contrasts with other cell types where ATP specific activity is higher than that of ADP. Predominant labeling of ADP may be partially due to compartmentation of ATP, but removal of cytosolic ATP does not change the relative labeling of ADP and ATP. Dilution of extracellular 32Pi following labeling resulted in loss of 70% of label from ADP but only 50% loss from gamma-ATP at 26 min. ADP was labeled in the absence of detectable ATP in the presence of rotenone plus antimycin. Fractionation of ejaculated sperm yielded midpieces that are depleted of adenylate kinase and have coupled respiration. ATP was labeled with 32Pi, but ADP was not in midpieces. Evidence for mitochondrial substrate level phosphorylation-supported incorporation of 32Pi into nucleotides was observed for intact sperm incubated with pyruvate and inhibitors of oxidative phosphorylation, but this activity did not occur in midpieces and does not appear to explain disproportionate labeling of ADP. We conclude that labeling of ADP in intact and permeabilized cells occurs by two pathways; one involves adenylate kinase, and the other is an unknown pathway which may be independent of ATP.     

Structural basis for substrate binding and the catalytic mechanism of type III pantothenate kinase

Pantothenate kinase (PanK) catalyzes the first step of the universal five-step coenzyme A (CoA) biosynthetic pathway. The recently characterized type III PanK (PanK-III, encoded by the coaX gene) is distinct in sequence, structure and enzymatic properties from both the long-known bacterial type I PanK (PanK-I, exemplified by the Escherichia coli CoaA protein) and the predominantly eukaryotic type II PanK (PanK-II). PanK-III enzymes have an unusually high Km for ATP, are resistant to feedback inhibition by CoA, and are unable to utilize the N-alkylpantothenamide family of pantothenate analogues as alternative substrates, thus making type III PanK ineffective in generating CoA analogues as antimetabolites in vivo. Previously, we reported the crystal structure of the PanK-III from Thermotoga maritima and identified it as a member of the "acetate and sugar kinase/heat shock protein 70/actin" (ASKHA) superfamily. Here we report the crystal structures of the same PanK-III in complex with one of its substrates (pantothenate), its product (phosphopantothenate) as well as a ternary complex structure of PanK-III with pantothenate and ADP. These results are combined with isothermal titration calorimetry experiments to present a detailed structural and thermodynamic characterization of the interactions between PanK-III and its substrates ATP and pantothenate. Comparison of substrate binding and catalytic sites of PanK-III with that of eukaryotic PanK-II revealed drastic differences in the binding modes for both ATP and pantothenate substrates, and suggests that these differences may be exploited in the development of new inhibitors specifically targeting PanK-III.     

Improving methionine and ATP availability by <Emphasis Type="Italic">MET6</Emphasis> and <Emphasis Type="Italic">SAM2</Emphasis> co-expression combined with sodium citrate feeding enhanced SAM accumulation in <Emphasis Type="Italic">Saccharomyces cerevisiae</Emphasis>

S-adenosyl-l-methionine (SAM), biosynthesized from methionine and ATP, exhibited diverse pharmaceutical applications. To enhance SAM accumulation in S. cerevisiae CGMCC 2842 (wild type), improvement of methionine and ATP availability through MET6 and SAM2 co-expression combined with sodium citrate feeding was investigated here. Feeding 6 g/L methionine at 12 h into medium was found to increase SAM accumulation by 38 % in wild type strain. Based on this result, MET6, encoding methionine synthase, was overexpressed, which caused a 59 % increase of SAM. To redirect intracellular methionine into SAM, MET6 and SAM2 (encoding methionine adenosyltransferase) were co-expressed to obtain the recombinant strain YGSPM in which the SAM accumulation was 2.34-fold of wild type strain. The data obtained showed that co-expression of MET6 and SAM2 improved intracellular methionine availability and redirected the methionine to SAM biosynthesis. To elevate intracellular ATP levels, 6 g/L sodium citrate, used as an auxiliary energy substrate, was fed into the batch fermentation medium, and an additional 19 % increase of SAM was observed after sodium citrate addition. Meanwhile, it was found that addition of sodium citrate improved the isocitrate dehydrogenase activity which was associated with the intracellular ATP levels. The results demonstrated that addition of sodium citrate improved intracellular ATP levels which promoted conversion of methionine into SAM. This study presented a feasible approach with considerable potential for developing highly SAM-productive strains based on improving methionine and ATP availability.