Biochemical and molecular characterization of the NADþ-dependentisocitrate dehydrogenase from the chemolithotroph

Acidithiobacillus thiooxidans

Hiroyuki Inoue a;1, Takashi Tamura a, Nagisa Ehara a, Akira Nishito a,Yumi Nakayama a, Makiko Maekawa a, Katsumi Imada b, Hidehiko Tanaka a,

Kenji Inagaki a;�

a Department of Bioresources Chemistry, Faculty of Agriculture, Okayama University, 1-1-1 Tsushima-naka, Okayama 700-8530, Japanb Namba Protonic NanoMachine Project, ERATO, Japan Science and Technology Corporation, 3-4 Hikaridai, Seika, Kyoto 619-0237, Japan

Received 1 June 2002; received in revised form 6 July 2002; accepted 11 July 2002

First published online 1 August 2002

Abstract

An isocitrate dehydrogenase (ICDH) with an unique coenzyme specificity from Acidithiobacillus thiooxidans was purified andcharacterized, and its gene was cloned. The native enzyme was homodimeric with a subunit of Mr 45 000 and showed a 78-fold preferencefor NADþ over NADPþ. The cloned ICDH gene (icd) was expressed in an icd-deficient strain of Escherichia coli EB106; the activity wasfound in the cell extract. The gene encodes a 429-amino acid polypeptide and is located between open reading frames encoding a putativeaconitase gene (upstream of icd) and a putative succinyl-CoA synthase L-subunit gene (downstream of icd). A. thiooxidans ICDH showedhigh sequence similarity to bacterial NADPþ-dependent ICDH rather than eukaryotic NADþ-dependent ICDH, but the NADþ-preference of the enzyme was suggested due to residues conserved in the coenzyme binding site of the NADþ-dependent decarboxylatingdehydrogenase. : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Isocitrate dehydrogenase; Decarboxylating dehydrogenase; Coenzyme speci¢city; Enzyme puri¢cation; Acidithiobacillus thiooxidans

1. Introduction

Isocitrate dehydrogenase (ICDH) catalyzes the oxidativedecarboxylation of D-isocitrate to form 2-oxoglutarate,coupled with the reduction of a dinucleotide cofactor(NADþ or NADPþ). The enzyme has been known as akey enzyme in the tricarboxylic acid (TCA) cycle and alsoplays an important role in glutamate synthesis. ICDH and3-isopropylmalate dehydrogenase (IPMDH) with a sub-unit of between 40 and 57 kDa belong to a class of proteinfamily, decarboxylating dehydrogenases, that lack a typi-cal LKL nucleotide-binding fold which is commonly

present in most dehydrogenases [1^3]. Although these en-zymes show a high sequence homology and a similarity oftheir 3-D structure, their coenzyme speci¢cities are strict :Escherichia coli ICDH has a 6900-fold preference forNADPþ [4], and IPMDH generally shows NADþ-depen-dency. These enzymes also provide an attractive modelsystem to study the coenzyme recognition [4^6].Some chemolithotrophs, that possess a 2-oxoglutarate

dehydrogenase-de¢cient TCA cycle, contain NADþ-de-pendent ICDH (NAD-ICDH) activity [7,8], though mostbacteria contain NADPþ-dependent ICDH (NADP-ICDH) activity. It has been reported that the incompleteTCA cycle functions for CO2 ¢xation [9] as well as gluta-mate synthesis [8]. Thus, the bacterial NAD-ICDH is ex-pected to be metabolically and structurally (especially thecoenzyme recognition) unique, but little is known aboutthe properties of the enzyme.To understand the coenzyme preference of bacterial

NAD-ICDH, we report here the puri¢cation and charac-terization of NAD-ICDH from an acidophilic chemolitho-

0378-1097 / 02 / $22.00 : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 3 7 8 - 1 0 9 7 ( 0 2 ) 0 0 8 5 7 - 1

* Corresponding author. Tel. : +81 (86) 251 8299;Fax: +81 (86) 251 8299.

E-mail address: kinagaki@cc.okayama-u.ac.jp (K. Inagaki).

1 Present address: Institute for Marine Resources and Environment,National Institute of Advanced Industrial Science and Technology, 2-2-2Hiro-suehiro, Kure, Hiroshima 737-0197, Japan.

FEMSLE 10601 19-8-02

FEMS Microbiology Letters 214 (2002) 127^132

www.fems-microbiology.org

troph Acidithiobacillus (formerly Thiobacillus) thiooxidans[7] and the nucleotide sequence of the structural gene (icd).

2. Materials and methods

2.1. Bacterial strains and growth conditions

A. thiooxidans ON107, our laboratory strain, was aer-obically grown on a medium containing 0.25% Na2S2O3,0.3% (NH4)2SO4, 0.05% K2HPO4, 0.05% MgSO4W7H2O,0.01% KCl, 0.001% Ca(NO3)2, and 0.001% FeSO4W7H2O(pH 5.0) at 30‡C for 60 h as described previously [10].E. coli EB106 (icd-11), which is a glutamate requirement,was used as a host to con¢rm the recombinant ICDHactivity. The strain was grown on a minimum mediumcontaining 0.02 mM thiamine, 3 mM L-glutamate, and1 mM L-tryptophan [11] at 37‡C, and kanamycin (20 Wgml31) was added to the media as a selection marker, whennecessary.

2.2. Preparation of cell extracts

Cell-free extracts were prepared by sonication and sub-sequently ultracentrifugation (105 000Ug, 1 h) as de-scribed previously [10]. The supernatant obtained was dia-lyzed at 4‡C against Bu¡er A [10 mM potassium phosphatebu¡er (pH 7.5) containing 0.01% 2-mercaptoethanol and10% glycerol]. Protein concentrations were determined bymeans of a protein assay kit (Bio-Rad, Hercules, CA,USA) with bovine serum albumin as a standard.

2.3. Puri¢cation of ICDH

All the puri¢cation procedures described below werecarried out at 4‡C. The cell-free extract from A. thiooxi-dans cells (76 g wet weight) was applied to a DEAE^Toyo-pearl 650M (Tosoh) column (5.0U18 cm) equilibrated inBu¡er A containing 50 mM KCl. The ICDH activity waseluted with the same bu¡er, and that fraction was subse-quently loaded onto a Q-Sepharose Fast Flow (AmershamBiosciences) column (3.0U16 cm) equilibrated with Bu¡erA. The active fractions were eluted with a linear gradientof 0.1^0.2 M KCl in Bu¡er A, then combined, dialyzed,and loaded onto the second Q-Sepharose Fast Flow(2.0U11 cm) equilibrated with Bu¡er A. The active frac-tions were eluted with a linear gradient of 0^0.2 M KCl inBu¡er A and brought to 25% saturation with ammoniumsulfate (pH 7.5). The enzyme solution was subjected toButyl^Toyopearl 650M (Tosoh) column chromatography(1.5U6.0 cm) and was eluted with a linear gradient (from25 to 0% saturated ammonium sulfate) in Bu¡er A. Theactive fractions were combined, dialyzed against Bu¡er A,and brought to 25% saturation with ammonium sulfate(pH 7.5). The enzyme solution was subjected to Phenyl^Toyopearl 650M (Tosoh) column chromatography

(1.4U2.5 cm), equilibrated with 20% saturated ammoniumsulfate in Bu¡er A, and eluted with 15% saturated ammo-nium sulfate in Bu¡er A. The active fractions were com-bined, dialyzed against Bu¡er A, and used as puri¢edICDH.

2.4. Molecular mass determination of ICDH

The molecular mass of the native ICDH was estimatedby gel-¢ltration chromatography through a SephacrylS-200 column (1.5U125 cm, Amersham Biosciences) equil-ibrated with Bu¡er A containing 0.2 M KCl. The calibra-tion curve was plotted using the following standard pro-teins: aldolase (Mr 158 000), bovine serum albumin (Mr

67 000), ovalbumin (Mr 43 000), chymotrypsinogen A (Mr

25 000), and ribonuclease A (Mr 13 700). The molecularmass of the subunit was determined by SDS^PAGE. AnLMW electrophoresis calibration kit (Amersham Bioscien-ces) was used for calibration.

2.5. Measurement of enzyme activity

ICDH activity was routinely measured by following thereduction of NADþ (or NADPþ) at 340 nm. Reactionmixtures were incubated at 30‡C and contained 100 mMTris^HCl bu¡er (pH 9.0), 0.67 mM D,L-isocitrate, 6.7 mMNADþ (or NADPþ), 50 mM KCl, 0.5 mM MgCl2, andthe enzyme in a total volume of 1.5 ml.

2.6. PCR ampli¢cation of A. thiooxidans chromosomalDNA

Chromosomal DNA of A. thiooxidans was isolated bythe method of Saito and Miura [12]. Oligonucleotide prim-ers, primer-F, R1, and R2, were designed according to thealignment of the amino acid sequences of bacterialNADP-ICDHs [13]. The ¢rst PCR was performed accord-ing to the manufacturer’s protocol using EX-Taq polymer-ase (2.5 U, Takara Shuzo) in a 50-Wl mixture, containing0.2 Wg of genomic DNA. The mixture was subjected to astep-cycle program consisting of initial melting at 94‡C for1 min and 30 cycles of 94‡C for 0.5 min, 60‡C for 0.5 min,and 68‡C for 3 min. The second PCR was performedunder the same condition in a 50-Wl mixture containing0.5 Wl of the ¢rst PCR product as a template. The se-quences of the primers used were as follows: primer-F(sense direction), 5P-AA(AG)GG(ACGT)CC(ACGT)(TC)-T(ACGT)AC(ACGT)AC(ACGT)CC(ACGT)GT-3P basedon the conserved sequence K100GPLTTPV107 (the number-ing refers to E. coli ICDH); primer-R1 (anti-sense direc-tion), 5P-CC(ACGT)GC(AG)TA(CT)TT(ACGT)GG(AC-GT)GC(ACGT)GT(ACGT)CC-3P based on the sequenceG340TAPKYAG347 ; and primer-R2 (anti-sense direction),5P-GC(ACGT)GT(ACGT)CC(AG)TG(ACGT)GT(ACGT)-GC(TC)TC(AG)AA-3P based on the sequence F335EATH-GTA342.

FEMSLE 10601 19-8-02

H. Inoue et al. / FEMS Microbiology Letters 214 (2002) 127^132128

2.7. DNA hybridization, cloning, and sequencing

Southern blot hybridization and plaque hybridizationwere performed by the standard techniques. The 770-bpPCR fragment containing icd was used as a hybridizationprobe. To construct the A. thiooxidans genomic DNA li-brary, chromosomal DNA was digested with BamHI, andthe 4.4^9.4-kb DNA fragments were cloned into ZAP ex-press (Stratagene). A recombinant plasmid, pLWD1, wasobtained from a positive plaque according to the manu-facturer’s protocol using an ExAssist1 helper phase (Stra-tagene).Nucleotide sequencing reactions were performed with

double-stranded pLWD1 by the dideoxy chain-termina-tion method using a BigDye Terminator Cycle Sequencingkit (PE Applied Biosystems). To obtain overlapping se-quence on both DNA strands, 20-bp custom syntheticprimers were used. Sequencing products were analyzedwith an ABI PRISM 310 Genetic analyzer (PE AppliedBiosystems). Sequence analyses and searches for sequencesimilarities in the data bases were carried out with theGENETYX-Mac ver. 7.0.9 software (Software Develop-ment, Japan) and the program BLAST, respectively.

3. Results and discussion

3.1. Puri¢cation and characterization of ICDH

The ICDH from A. thiooxidans was puri¢ed 520-fold(Table 1), almost to electrophoretic homogeneity. The spe-ci¢c activity of the puri¢ed enzyme was 120 U mg31 withNADþ, and only 15% of the activity appeared with

NADPþ. In non-denaturing PAGE, a major band de-tected on the CBB-staining gel was consistent with eachprotein band stained for NAD- and NADP-ICDHactivity. Thus, the puri¢ed enzyme is a single proteinwith a dual coenzyme speci¢city. The optimum pH foractivity was 8.5 when determined over a pH range of6^10.2. The enzyme was relatively stable up to 55‡C,but no activity was detected after exposure to 70‡C for30 min.

A. thiooxidans ICDH has a 78-fold preference forNADþ over NADPþ (Table 2). The apparent Km valuefor D,L-isocitrate was 0.12 mM when determined for theNADþ-linked reaction. It appears that the NADþ-depen-dency of the enzyme is similar to that of Saccharomycescerevisiae NAD-ICDH (Table 2), which is an octamercomposed of two non-identical subunits. However, SDS^PAGE analysis showed that the A. thiooxidans ICDH con-sisted of a single subunit of Mr 45 000. Because the nativeMr was estimated to be 88 000, the enzyme is homodimer-ic. These results suggest that the primary structure of theenzyme is similar to that of bacterial NADP-ICDH ratherthan eukaryotic NAD-ICDH.

3.2. Cloning, nucleotide sequence, and expression of theicd gene

A 770-bp fragment of icd from A. thiooxidans was am-pli¢ed using PCR primers that were designed from homol-ogous regions of bacterial NADP-ICDHs. Southern blotanalysis of A. thiooxidans genomic DNA probed with theampli¢ed fragment showed only one signal, suggestingthat icd of the homodimeric-type enzyme is single in thegenomic DNA. We cloned a 8-kb BamHI fragment con-

Table 1Puri¢cation of A. thiooxidans ICDH

Puri¢cation step Total protein (mg) Total activity (U) Speci¢c activitya (U mg31) Puri¢cation (fold) Recovery (%)

Crude extract 1800 430 0.23 (1) (100)DEAE^Toyopearl 530 370 0.69 3 85Q-Sepharose (1st) 12 150 13 56 35Q-Sepharose (2nd) 8.7 145 17 72 34Butyl^Toyopearl 0.51 62 120 520 14

aSpeci¢c activity is de¢ned as Wmol of NADH formed per mg of protein per min.

Table 2Kinetic parameters of ICDH toward NADP and NAD

Enzyme NAD NADP Preference (A)/(B)

Km(WM)

kcat(s31)

kcat/Km (A)(WM31 s31)

Km(WM)

kcat(s31)

kcat/Km (B)(WM31 s31)

A. thiooxidans ICDH 180 45.1 0.251 6570 21.3 0.0032 78S. cerevisiae NAD-ICDH [17] 210 40 0.190E. coli NADP-ICDH [4] 4700 3.22 0.00069 17 80.5 4.7 0.00015Engineered E. coli NAD-ICDH [4]a 99 16.2 0.164 5800 4.70 0.00081 202

aThe mutant enzyme engineered coenzyme speci¢city in ICDH from NADP to NAD.

FEMSLE 10601 19-8-02

H. Inoue et al. / FEMS Microbiology Letters 214 (2002) 127^132 129

taining the intact icd (Fig. 1a). The nucleotide sequence ofthe gene (the DDBJ/EMBL/GenBank accession no.AB062764) and the deduced amino acid sequence of theenzyme is shown in Fig. 1b. The gene encodes a 429-ami-no acid polypeptide. The predicted molecular mass (46 218Da) of the gene product was in good agreement with thatof a subunit of A. thiooxidans ICDH. An E. coli EB106/pLWD1 transformant showed NAD-ICDH activity in thecell-free extract (0.3 U mg31 protein), but NADP-ICDHactivity was not observed. In addition, it was found thatthe recombinant NAD-ICDH functions physiologically inE. coli cells, because the transformant complemented aglutamate requirement of the host strain. The recombinantICDH from E. coli EB106/pLWD1 was partially puri¢ed,and its physicochemical properties were con¢rmed to beconsistent with those of A. thiooxidans ICDH (data notshown).

A computer search with the 5P- and 3P-£anking regionsof icd revealed the presence of the 3P-part of the putativeaconitase gene (50% identity for 256 amino acids of theE. coli enzyme) and the 5P-part of the putative succinyl-CoA synthetase L-subunit gene (55% identity for 86 aminoacids of the E. coli enzyme), respectively (Fig. 1), suggest-ing that these genes form an operon. It has been reportedthat the NAD-ICDH gene of Streptococcus mutans is en-coded in an operon with citrate synthase and aconitasegenes and is related to glutamate synthesis [14]. AlthoughA. thiooxidans ICDH also can contribute to glutamatebiosynthesis [8], the structure of the icd operon containingthe putative succinyl-CoA synthetase gene is di¡erent fromthat of the icd operon of S. mutans. The complete analysisof the operon may reveal the physiological role of the2-oxoglutarate-de¢cient TCA cycle containing NAD-ICDH of A. thiooxidans.

Fig. 1. a: Restriction map of the BamHI fragment of pLWD1 containing icd. Arrows indicate the positions and directions of the genes. Black parts inthe arrows show the sequence region (b) determined in this study. b: Nucleotide sequences of icd, the 3P-part of the putative aconitase gene (ORF1)and the 5P-part of the putative succinyl-CoA synthase L-subunit gene (ORF2). Putative ribosome binding sites (RBS) are underlined. Primer-F, R1, andR2 correspond to sequences of PCR primers used to amplify the icd fragment. B, BamHI; Sc, SacI ; K, KpnI ; Sm, SmaI; P, PstI.

FEMSLE 10601 19-8-02

H. Inoue et al. / FEMS Microbiology Letters 214 (2002) 127^132130

3.3. Sequence homology and determinant of coenzymespeci¢city

The deduced amino acid sequence of A. thiooxidansICDH showed a high homology to bacterial NADP-ICDHs (ex. 59% identity [E. coli ICDH]) and S. mutansNAD-ICDH (55% identity) rather than eukaryotic NAD-ICDHs (ex. 26.7% identity [S. cerevisiae ICDH2]) andNADþ-dependent IPMDH (ex. 23% identity [Acidithioba-cillus ferrooxidans IPMDH]). These results suggest that agroup of the bacterial dimeric-ICDH was divergent fromthe ancestral NAD-ICDH before speci¢city toward thecoenzyme [15].In a comparison of the deduced coenzyme binding sites

of decarboxylating dehydrogenases, we found that A. thio-oxidans ICDH possesses three conserved residues com-prising an NAD recognition site; Asp357 which forms adouble hydrogen bond with the 2P- and 3P-hydroxyls ofNADþ, Ile358, and Ala364 (Fig. 2). This result suggeststhat the coenzyme binding mechanism of A. thiooxidansICDH should be same as that of NADþ-dependent decar-boxylating dehydrogenase. A. thiooxidans ICDH has oneconserved residue (Arg305) which interacts with the 2P-phosphate of NADPþ [16]. The residue may be requiredfor the dual coenzyme recognition of the enzyme. Interest-ingly, a comparison of a seven-fold mutant ICDH fromE. coli which engineered preference for NADþ [3] andA. thiooxidans ICDH revealed that these enzymes havesimilar kinetic parameters (Table 2) and coenzyme bind-ing sites (Fig. 2). Further analysis of naturally occurringA. thiooxidans ICDH could ¢nd the structural constraintof the seven-fold mutant ICDH.

References

[1] Hurley, J.H., Thorsness, P.E., Ramalingam, V., Helmers, N.H.,Koshland Jr., D.E. and Stroud, R.M. (1989) Structure of a bacterialenzyme regulated by phosphorylation, isocitrate dehydrogenase.Proc. Natl. Acad. Sci. USA 86, 8635^8639.

[2] Imada, K., Sato, M., Tanaka, N., Katsube, Y., Matsuura, Y. andOshima, T. (1991) Three-dimensional structure of a highly thermo-stable enzyme, 3-isopropylmalate dehydrogenase of Thermus thermo-philus at 2.2 AU resolution. J. Mol. Biol. 222, 725^738.

[3] Imada, K., Inagaki, K., Matsunami, H., Kawaguchi, H., Tanaka, H.,Tanaka, N. and Namba, K. (1998) Structure of 3-isopropylmalatedehydrogenase in complex with 3-isopropylmalate at 2.0 AU resolu-tion: the role of Glu88 in the unique substrate-recognition mecha-nism. Structure 6, 971^982.

[4] Chen, R., Greer, A. and Dean, A.M. (1995) A highly active decar-boxylating dehydrogenase with rationally inverted coenzyme speci¢c-ity. Proc. Natl. Acad. Sci. USA 92, 11666^11670.

[5] Hurley, J.H., Chen, R. and Dean, A.M. (1996) Determinants of co-factor speci¢city in isocitrate dehydrogenase : structure of an engi-neered NADPþ^sNADþ speci¢city-reversal mutant. Biochemistry35, 5670^5678.

[6] Miyazaki, K. and Oshima, T. (1994) Co-enzyme speci¢city of 3-iso-propylmalate dehydrogenase from Thermus thermophilus HB8. Pro-tein Eng. 7, 401^403.

[7] Hampton, M.L. and Hanson, R.S. (1969) Regulation of isocitratedehydrogenase from Thiobacillus thiooxidans and Pseudomonas £uo-rescens. Biochem. Biophys. Res. Commun. 36, 296^305.

[8] Smith, A.J., London, J. and Stanier, R.Y. (1967) Biochemical basis ofobligate autotrophy in blue-green algae and thiobacilli. J. Bacteriol.94, 972^983.

[9] Shiba, H., Kawasumi, T., Igarashi, Y., Kodama, T. and Minoda, Y.(1985) The CO2 assimilation via the reductive tricarboxylic acidcycle in an obligatory autotrophic, aerobic hydrogen-oxidizing bac-terium, Hydrogenobacter thermophilus. Arch. Microbiol. 141, 198^203.

[10] Kawaguchi, H., Inagaki, K., Matsunami, H., Nakayama, Y., Tano,T. and Tanaka, H. (2000) Puri¢cation and characterization of 3-iso-propylmalate dehydrogenase from Thiobacillus thiooxidans. J. Biosci.Bioeng. 90, 459^461.

[11] Eikmanns, B.J., Rittmann, D. and Sahm, H. (1995) Cloning, se-quence analysis, expression, and inactivation of the Corynebacteriumglutamicum icd gene encoding isocitrate dehydrogenase andbiochemical characterization of the enzyme. J. Bacteriol. 177, 774^782.

[12] Saito, M. and Miura, K. (1963) Preparation of transforming deoxy-ribonucleic acid by phenol treatment. Biochim. Biophys. Acta 72,619^629.

[13] Miyazaki, K. (1996) Isocitrate dehydrogenase from Thermus aqua-ticus YT1: puri¢cation of the enzyme and cloning, sequencing,and expression of the gene. Appl. Environ. Microbiol. 62, 4627^4631.

Fig. 2. Sequence alignments of coenzyme recognition site in the decarboxylating dehydrogenase. The sequence numbers of Ec-ICDH, At-ICDH and Ec-IPMDH are indicated. Putative residues interacting uniquely with NADþ or NADPþ are boxed. Highly conserved residues in NADþ-dependent en-zymes are shown with arrows. The conserved residues in all enzymes are shown with asterisks. Ec, E. coli (GenBank/EMBL accession number J02799[ICDH], P30125 [IPMDH]); Tt, Thermus thermophilus (M94317 [ICDH], K01444 [IPMDH]) ; Bs, Bacillus subtilis (U05257); At, A. thiooxidans(AB062764, this study); Sm, S. mutans (U62799); Sc, S. cerevisiae ICDH2 (M74131); Af, A. ferrooxidans (JX0286). NAD-ICDH Ec (e) indicates theseven-fold mutant enzyme from E. coli which engineered coenzyme speci¢city [4].

FEMSLE 10601 19-8-02

H. Inoue et al. / FEMS Microbiology Letters 214 (2002) 127^132 131

[14] Cvitkovitch, D.G., Gutierrez, J.A. and Bleiweis, A.S. (1997) Role ofthe citrate pathway in glutamate biosynthesis by Streptococcus mu-tans. J. Bacteriol. 179, 650^655.

[15] Dean, A.M. and Golding, G.B. (1997) Protein engineering revealsancient adaptive replacements in isocitrate dehydrogenase. Proc.Natl. Acad. Sci. USA 94, 3104^3109.

[16] Hurley, J.H., Dean, A.M., Koshland Jr., D.E. and Stroud, R.M.

(1991) Catalytic mechanism of NADPþ-dependent isocitrate dehy-drogenase: implications from the structures of magnesium-isocitrateand NADPþ complexes. Biochemistry 30, 8671^8678.

[17] Cupp, J.R. and McAlister-Henn, L. (1991) NADþ-dependent isoci-trate dehydrogenase. Cloning, nucleotide sequence, and disruption ofthe ICDH2 gene from Saccharomyces cerevisiae. J. Biol. Chem. 266,22199^22205.

FEMSLE 10601 19-8-02

H. Inoue et al. / FEMS Microbiology Letters 214 (2002) 127^132132