Methylenetetrahydrofolate dehydrogenase-cyclohydrolase from Photobacterium phosphoreum shares properties with a mammalian mitochondrial homologue

e t B i o p h y s i c a AEta Biochimica et BiophysicaActa 1296 (1996) 47-54

ELSEVIER

Methylenetetrahydrofolate dehydrogenase-cyclohydrolase from Photobacterium phosphoreum shares properties with a mammalian mitochondrial homologue Peter D. Pawelek, Robert E. MacKenzie

*

Department of Biochemistry, McGill University, 3655 DrummondStreet, Montreal, QuebecH3G 1Y6, Canada Received 19 January 1996;revised 19 March 1996; accepted 27 March 1996

Abstract

The marine bioluminescent bacterium Photobacterium phosphoreum expresses a bifunctional methylenetetrahydrofolate dehydrogenase-cyclohydrolase with dual cofactor specificity. An investigation of the kinetic parameters of the P. phosphoreum enzyme indicate that its utilization of dinucleotide cofactors shares similarities with the human mitochondrial dehydrogenase-cyclohydrolase. Both enzymes exhibit dual cofactor specificity and the NAD+-dependent dehydrogenase activities from both enzymes can be activated by inorganic phosphate. Furthermore, an analysis of multiply aligned dehydrogenase-cyclohydrolase sequences from 11 species revealed that bacterial and mitochondrial enzymes are more closely related to each other than to the dehydrogenase-cyclohydrolase domains from eukaryotic trifunctional enzymes, and that the bacterial and mitochondrial enzymes share a common point of divergence. Since the NADP + cofactor is kinetically favoured by a factor of 18 over NAD +, and is therefore likely to be the preferred in vivo cofactor, we propose that the P. phosphoreum enzyme and the human mitochondrial enzyme evolved from a common ancestral dehydrogenasecyciohydrolase with dual cofactor specificity, but that cofactor preference in these two enzymes diverged in response to different metabolic requirements. Keywords: Methylenetetrahydrofolate;Phosphate activation; NAD; NADP; Bifunctionality;(P. phosphoreum)

1. Introduction

Folate-mediated metabolism is an essential process, providing one-carbon unit precursors for purine, thymidylate, and formylmethionyl tRNA fmet biosynthesis in both prokaryotic and eukaryotic organisms. The major onecarbon unit donors are 5,10-methyleneH 4 folate and 10-formylH4folate. Two enzymatic activities facilitate interconversion between these folate donors: 5,10-methyleneHafolate dehydrogenase and 5,10-methenylH4folate cyclohydrolase. As well, 10-formylH4folate can be synthesized by 10-formylHafolate synthelase, which incorporates formate into H4folate in an ATP-dependent reaction. MethyleneH4folate dehydrogenases can occur as monofunctional, bifunctional, oi: trifunctional enzymes. Monofunctional dehydrogenases are predominantly bacterial, in-

* Corresponding author. Fax: [email protected].

+ 1 (514) 3987384; e-mail:

cluding the enzymes from A. woodii and C. formicoacetium [1,2], although a monofunctional cytoplasmic dehydrogenase has also been identified in S. cerevisiae [3]. The dehydrogenase activity can also occur as bifunctional dehydrogenase-cyclohydrolases having subunit sizes approximately the same as the monofunctional enzymes [ 1 3]. Substrate channeling has been shown to occur between the dehydrogenase and cyclohydrolase activities in these enzymes, and recent evidence indicates that these two activities in the human cytoplasmic bifunctional domain share an active site [4]. Bifunctional dehydrogenasecyclohydrolases can exist as discrete enzymes or as a catalytically independent domain in trifunctional dehydrogenase-cyclohydrolase-synthetases which occur primarily in the eukaryotic cytoplasm. While monofunctional dehydrogenase enzymes can be either NAD +-, or NADP+-dependent, the bifunctional domains and enzymes are NADP+-dependent. Exceptions to this are the mammalian mitochondrial dehydrogenasecyclohydrolases. The preferred cofactor for these enzymes

01674838/96/$15.00 Copyright:© 1996 Elsevier Science B.V. All rights reserved PII SO 167-4838(96)00052- 9

48

P.D. Pawelek, R.E. MacKenzie / Biochimica et Biophysica Acta 1296 (1996) 47-54

is NAD +, although the human enzyme has been shown to be able to utilize NADP + with significantly less efficiency than NAD ÷ [5]. Preliminary assays of bacterial extracts conducted in our laboratory indicated that the marine bioluminescent species Photobacterium phosphoreum expresses both NAD+-dependent and NADP+-dependent dehydrogenase activities as well as synthetase activity (Narciso Mejia, unpublished). In this paper we investigate the nature of the dehydrogenase activities. We establish that the dehydrogenase activities are the properties of a single dehydrogenase-cyclohydrolase enzyme with unusual dual cofactor specificity. We cloned the gene encoding the P. phosphoreum dehydrogenase-cyclohydrolase to express a recombinant form of the enzyme so that we could more closely examine its kinetic properties and how they relate to those of the mammalian mitochondrial enzymes.

which synthesis of uridine-rich DNA is enhanced. Expression of recombinant enzymes was done in E. coli strain K38, containing the pGP1-2 plasmid encoding a temperature-sensitive T7 repressor; this strain was a gift from Dr. Charles C. Richardson and it is described in Tabor and Richardson [11 ]. Complementation screening was performed in the S. cerevisiae strain YA3-1 (a, ade3, leu2, ura3), which was obtained by crossing the haploid strains T9-3C (MAT a ura3 his3 leu2) and STX24-1B (alpha ade3 ural gal2) as performed by Dean Hum (Ph.D. thesis).

2.3. Enzyme assays

2. Materials and methods

Standard spectrophotometric assays for the dehydrogenase and cyclohydrolase activities were performed according to Tan and MacKenzie [12]. All enzyme assays were quantified using a Beckman DU-640 spectrophotometer. One unit of activity is defined as 1 Ixmol of product formed per minute.

2.1. Materials

2.4. Cloning and selection by complementation

All chemicals used in this study were of analytical grade. All solid chemicals were obtained from Sigma or Fisher unless otherwise noted. Restriction endonucleases were obtained from New England Biolabs, Gibco BRL, and Boehringer Mannheim. Radioisotopes used in DNA sequencing were obtained from Amersham, as was the Sequenase kit used to perform the procedure. (6R, S)H4 folate was prepared according to the procedure of Drury et al. [6] and (6R, S)-methenylH4folate was prepared as described previously [7]. Unless specified otherwise, all common molecular biological techniques were performed according to Sambrook and Maniatis [8].

P. phosphoreum genomic DNA was partially digested with Sau3A and size fractionated over a Sephacryl 500-S HR column (Pharmacia). DNA corresponding to fragments between 2 and 5 kb was subeloned into the BamHI site of the yeast expression vector pVT-102U (containing the URA3 marker gene). The library was then transfected into MAX Efficiency DH5-alpha E. coli cells (Gibeo BRL) and amplified. The amplified P. phosphoreum genomic library was transformed into competent YA3-I cells using the protocol of Schiestl et al. [13]. Successful transformants were selected by their ability to form colonies on minimal plates lacking uracil. These transformants were then screened for their ability to complement the purine auxotrophy of the YA3-1 strain by replica plating onto minimal plates lacking both adenine and uracil.

2.2. Strains and plasmids The phagemid pBluescript-KS + (Stratagene) was used as the common vector for routine procedures such as subcloning, oligo-directed mutagenesis, exonuclease III deletions, and sequence analysis. The E. coli expression vector, pBKe-DC, was modified from the vector pBKeHB 1, constructed by X.-M. Yang as described in Murley et al. [9]; it contains a T7 promoter, T7 enhancer and an E. coli ribosomal binding site to ensure efficient protein expression. The S. cerevisiae expression vector, pVT102-U, contains a 2Ix origin of replication, a constitutive alcohol dehydrogenase promoter, a URA3 marker gene, and the beta-lactamase gene encoding for ampicillin resistance ( AmpR); it was a gift from Dr. T. Vernet, and its construction is described in Vemet et al. [10]. The E. coli strain DH5-alpha was used for most common molecular biological procedures. Oligo-directed mutagenesis was performed in the strain CJ236 (dut-, ung-) in

2.5. Analysis of clones Positive colonies were grown in YEPD cultures and plasmid DNA was obtained by disrupting cell pellets with 0.2 g glass beads in a buffer consisting of 100 mM NaC1, 10 mM Tris (pH 7.5), 1 mM EDTA, 0.1% SDS. The aqueous phase was extracted with phenol/chloroform (1:1) and DNA was precipitated with isopropanol. The redissolved DNA fraction was then passed over a Magic Miniprep column (Promega) to purify plasmid DNA. In order to obtain larger quantities of the pVT plasmid containing the positive clones, the DNA obtained from yeast was transformed into E. coli DH5-alpha cells and larger DNA preparations were prepared from these transformants. To prepare nested deletions, a 1.1 kb PstI-HindlII fragment was excised from the pVT-S1 clone and ligated

P.D. Pawelek, R.E. MacKenzie / Biochimica et Biophysica Acta 1296 (1996) 47-54

into PstI-HindlII-digested KS + . Exonuclease III digestion of the construct resulted in nested deletions of approx. 200 bp increments which were analysed by Sanger dideoxy sequencing. Both strands of the insert were sequenced to ensure accuracy.

2.6. Enzyme expression and purification A BspHl site was created at position - 2 relative to the start codon by a C to T mutation. A BspH1-HindlII fragment containing the gene was then ligated in frame into a pBKe expression vector digested with NcoI/ClaI to form the expression construct pBKe-DC. The HindlII end of the insert was filled in with Klenow to facilitate bluntend ligation at the ClaI site (BspH1 and NcoI have compatible overhangs). Oligo-directed mutagenesis was performed according to Kunkel et al. [14], using oligonucleotides synthesized by GSD Oligos (Toronto, Canada). Expression constructs were transformed into competent E. coli K38 cells and gene expression was induced by rapidly heating a log phase culture to 42°C. The cells were then incubated at 42°C for a further 15 minutes at which time Rifampicin (Boehringer Mannheim) was added to a final concentration of 15 rag/1. The culture was incubated for a further 3 h at 37°C. After induction, cells were centrifuged at 5000 rpm (Sorvall RC3, GS3 rotor) :for 10 min and frozen at - 8 0 ° C overnight. Unless otherwise noted, all subsequent purification procedures were performed at 4°C. The frozen cells were resuspended in sonication buffer (50 mM potassium phosphate (pH 7.3), 0.18 m g / m l PMSF, 0.16 m g / m l benzamidine, 36 mM 2-mercaptoethanol) and disrupted (Blackstone Ultrasonic Model SS2 sonicator) for 2.5 min in 15 second intervals. Glycerol was added to 20% and the crude extract was centrifitged at 25 000 × g in an SS-34 rotor (Sorvall) for 45 min. Protamine sulfate was added to the supernatant to a final concentration of 0.22 mg/ml. The solution was allowed to stand on ice for 30 min and then centrifuged at 25 000 × g for 30 min. Cleared supernatant solution was applied to a Matrex Orange-A column (Amicon, 2.5 cm × 2.5 cm) which had been previously equilibrated with Buffer A (50 mM potassium phosphate (pH 7.3), 2 mM Na2EDTA, 36 mM 2-mercaptoethanol, 0.17 m g / m l PMSF, 0.16 m g / m l benzamidine, 20% glycerol). The loaded column was washed with 10 column volumes of Buffer A followed by an additional wash of 5 colamn volumes of Buffer A / 1 0 0 mM KC1. The enzyme was eluted with a linear gradient (0.1 M to 1 M KC1 in Buffer A; flow rate 0.5 ml/min) and the peak containing enzyme activity was dialyzed against Buffer A and loaded onto a Blue-A Sepharose (Pharmacia) column (2.5 cm × 2.5 cm, pre-equilibrated with Buffer A) which was then washed with 10 column volumes of Buffer A and eluted with a linear gradient (0 to 2 M KCI in Buffer A; flow rate 0.5 ml/min). Fractions containing enzyme activ-

49

ity were determined by monitoring dehydrogenase activity using standard assay conditions. Purified enzyme was dialysed against Buffer A and concentrated 25-fold with a Centricon- 10 concentrator (Amicon). Purified proteins along with samples from all purification steps were analysed by SDS-PAGE (12%) using the conditions of Laemmli [15]. All protein concentrations were determined using the protocol of Bradford, with bovine serum albumin as standard [16].

2.7. Kinetic characterization and phosphate activation Initial rates were determined under conditions in which no more than 10% of the limiting substrate was consumed over the time-course used to calculate initial rates. Kinetic constants were determined using five substrate concentrations per plot of initial velocity vs. [substrate], ranging from 0.25 × to 5 × K m for the substrate. Velocities at each concentration were assayed in triplicate. The data were fitted to the Michaelis-Menten equation using the nonlinear regression analysis program ENZ~.-~ER® (R.J. Leatherbarrow, Biosoft, Cambridge, UK). Kinetic parameters presented in this paper are the average _ S.D. of those determined from three to five curves. Initial rates of the NAD+-dependent dehydrogenase activity were assayed at potassium phosphate concentrations ranging from 0 to 60 mM in a buffer containing 25 mM Mops (pH 7.3), 2.5 mM NAD ÷, 28 mM formaldehyde, 10 mM (6R, S)-H4folate, and 15 mM 2-mercaptoethanol. The differences between the rates in potassium phosphate and the average initial rate at 0 mM potassium phosphate were calculated and these data were fitted to the MichaelisMenten equation using the computer program ENZ~rrr~R with robust weighting. The Kac t value for phosphate is the average of three independent experiments. The experiment was also repeated using 100 IxM NADP + in place of 2.5 mM NAD +. To control for anionic specificity and for possible activity changes due to increases in potassium levels, the experiments were repeated substituting potassium sulfate for potassium phosphate. In this case, initial rates were assayed in duplicate over a range from 0 to 60 mM potassium sulfate.

2.8. Multiple sequence alignment Eleven amino-acid sequences encoding dehydrogenasecyclohydrolase enzymes or domains (P. phosphoreum, E. coli [17], S. typhimurium [18], S. cerevisiae cytoplasmic [19], S. cerevisiae mitochondrial [20], human cytoplasmic [21], human mitochondrial [22], D. melanogaster mitochondrial [23], mouse mitochondrial [24], rat cytoplasmic [25], and Spodoptera frugiperda cytoplasmic [26]) were aligned using the program WinDNASIS (Hitachi Instruments) and refined manually. The refined alignment was analyzed by the program CLUSTAL-W (J. Thompson, T.

50

P.D. Pawelek,R.E. MacKenzie/ Biochimica et BiophysicaActa 1296 (1996) 47-54 3'

St

N 'I

~ I

P I

B L

X '

~

H I

K II

Fig. 1. Restriction map of the 3.3 kb pVT-S1 clone. Flanking sequences correspondingto the pVT-102-Uexpression vector are indicated as open boxes. The arrow represents the folD open reading frame. Abbreviations:B, BamHl; H, Hindlll; K, KpnI; N, NotI; P, Pstl; Pv, PvulI.

Gibson, EMBL, Heidelberg, Germany) to generate an unrooted phylogenetic tree. The tree was displayed and printed using the program DRAWTREE(J. Felsenstein, University of Washington). 2.9. Nucleotide sequence accession number The nucleotide sequence for the P. phosphoreum folD gene has been submitted to GenBank and has been assigned the accession number U34207.

competent YA3-1 cells and screened for clones that could grow in the absence of exogenous purines. Plasmids purified from colonies that could grow under these conditions were retransformed into YA3-1 to confirm that cell growth in the absence of exogenous purines was due to the presence of the plasmid and not to a reversion mutation. Using these conditions, approx. 10000 transformants were screened and two were obtained that could successfully complement the YA3-1 purine auxotrophy. The clones were found to be identical by restriction mapping. 3.2. Analysis of the pVT-S1 clone

3. Results 3.1. Complementation cloning of the P. phosphoreum folD gene Complementation of t h e purine auxotrophic S. cerevisiae strain YA3-1, containing a mutation in the ADE3 gene was used to obtain a clone of the P. phosphoreum gene encoding the bifunctional enzyme, PPDC. Expression of the cDNA encoding the rat tfifunctional enzyme has already been shown to be able to functionally complement the ADE3 locus in S. cerevisiae [25]. Furthermore, the purine auxotrophy in the YA3-1 strain was shown to be complemented by heterologous expression of the human DC301 domain (the bifunctional dehydrogenasecyclohydrolase portion of the human cytoplasmic trifunctional enzyme) upon transformation of YA3-1 cells with a yeast expression construct containing this gene (Dean Hum, Ph.D. thesis). We therefore reasoned that this would be an effective system for complementation cloning of the NAD +-, and NADP+-dependent dehydrogenase activities that were previously observed in P. phosphoreum extracts (Narciso Mejia, unpublished). A P. phosphoreum genomic library, constructed in a pVT102-U yeast expression vector, was transformed into

Table 1 Purification of recombinantPPDC Fraction Volume(ml) Crude extract 20 Protamine sulfate 20 Matrex Orange A 60 Blue A Sepharose 50

Total protein (mg) 90 88 12 7

The pVT-S1 clone contains a 3.3 kb insert (Fig. 1) and sequence analysis revealed that a complete open reading frame of 855 bp was present, encoding a polypeptide of 285 amino acids. Sequence comparison to the E. coli folD gene which encodes a bifunctional methyleneHafolate dehydrogenase-cyclohydrolase enzyme [17], showed that the predicted amino-acid sequence of the open reading frame has 73.4% identity to that of the folD sequence. This open reading frame was therefore determined to encode a P. phosphoreum dehydrogenase-cyclohydrolase enzyme, PPDC. No open reading frames encoding polypeptides homologous to enzymes involved in folate metabolism or purine biosynthesis were observed to occur within 250 bp upstream or downstream of the folD gene. 3.3. Expression and purification of recombinant PPDC The P. phosphoreum folD gene was ligated in-frame into an E. coli expression vector, pBKe, creating the construct pBKe-DC. Optimal expression of the enzyme was obtained by transforming E. coli K38 cells with the construct and inducing expression by rapidly heating the cells to 42°C. The recombinant enzyme, PPDC, was determined to comprise approx. 14% of total protein in crude

Total activity(units) 1229 1086 936 677

Specificactivity(units/mg) 13.7 12.4 77.2 97.4

Recovery(%) 100 88 76 55

P.D. Pawelek, R.E. MacKenzie /Biochimica et Biophysica Acta 1296 (1996) 47-54 1

Z

3

4

5

51

Table 2 Kinetic parameters of recombinant PPDC

Substrate

K m (mM)

kcat (s - I )

kcat/K m

(s- i/mM) 94 kDa

<

30 kDa

PPDC

NAD + NADP + Pi methyleneH4 folate (NAD ÷ ) methyleneH4folate (NADP ÷ ) methenylH4 folate

4.76+0.97 0.20+0.01 12.1+1.3 b 0.0165:0.004 0.011 5:0.002 0.0315:0.009

136+ 12 a 29 1135:8 a 562

53+7

1700

a Enzyme activity extrapolated to infinite dinucleotide concentration in the presence of 10X K m (6R,S)-methyleneH4folate. b kact assayed in a 25 mM Mops (pH 7.3) buffer containing 2.5 mM NAD.

20 kDa 14 kDa Fig. 2. SDS-polyacrylamide gel e]ectrophoresis of fractions from purification of recombinant PPDC. A 12% polyacrylamide gel used to resolve samples (10 tzg total protein per lane), was stained with Coomassie brilliant blue. Lane 1, crude K38 extract before heat induction; lane 2, crude K38 extract after heat induction; lane 3, protamine sulfate supernatant; lane 4, Matrex Orange A fraction; lane 5, Blue A Sepharose fraction; the migration positions of low molecular weight standards (Pharmacia) also are indicated or: the left edge of the figure.

(6R, S)-methyleneH4folate for the dehydrogenase activity; (6R, S)-methenylH4folate for the cyclohydrolase activity (Table 2). As can be seen from Table 2, the enzyme has very similar turnover numbers for the dehydrogenase activity regardless of whether NAD ÷ or NADP + is a cofactor in this reaction. The K m values for the two cofactors differ greatly, however, resulting in an 18-fold higher kcat//Km value for NADP + over NAD +.

3.5. Phosphate activation of PPDC extracts of induced K38/pBKe-DC cells (Table 1). PPDC was purified over two dye-affinity columns (Table 1) resulting in a homogeneous preparation as determined by SDS-PAGE analysis (Fig. 2). The molecular mass of the subunit is approx. 31 kDa as determined by SDS-PAGE, which agree~ with the value 30.8 kDa as determined from the predicted amino-acid sequence using the program PROSIS (Hitachi Instruments). Gel filtration analysis of recombinant PPDC indicated that the active enzyme has a dimeric quaternary structure, which is the case with all known bifunctional dehydrogenasecyclohydrolase enzymes (data not shown).

3.4. Kinetic parameters of PPDC Kinetic parameters under standard assay conditions were obtained for the following substrates: NADP +, NAD + and (a)

Determination of the K m values for the dinucleotide cofactors in the absence of potassium phosphate (assays in which potassium phosphate was replaced with 25 mM Mops (pH 7.3)) indicated that the K m for NAD ÷ increased significantly, from 4.2 mM to 30.3 mM. That the NAD+-dependent dehydrogenase activity of PPDC is specifically activated by inorganic phosphate is shown in Fig. 3(a). Plots of initial rates vs. [phosphate] follow saturation kinetics with a Kac t of 12.1 mM for phosphate. Sulfate was not observed to have an effect, showing that the activation is phosphate-specific. Conversely, phosphate was observed to have an inhibitory effect on the NADP ÷dependent dehydrogenase activity (Fig. 3b). However, since sulfate is able to inhibit this activity in a manner similar to phosphate, this inhibition is not phosphate-specific. (b)

2:8.9 25.7

i

19.3

19.3 16.1

i

12.9 9.7

12.9

--9

I

o

2o

4o anion

so

6.5

o

~ ~

4o anion

Fig. 3. (a) Effect of anions on the NAD+-dependent dehydrogenase activity of PPDC. (b) Effect of anions on the NADP+-dependent dehydrogenase activity of PPDC. ( ~ ) Potassium phosphate, ( [] ) potassium sulfate.

52

P.D. Pawelek, R.E. MacKenzie / Biochimica et Biophysica Acta 1296 (1996) 47-54

PPDC

ECDC STDC

HCDCS RCDCS SCDCS YCDCS

m...................................................... ...................................................... X. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • ..................................................... II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

u~xx a~kXX a~kXX ap~eXl aplqX 1 wep~j a~vl ha~qyh .. . . . . . . . . . . . . . . . . . . . . Xl

oaxx~.~m RtXa~q~R E~tXa~q~ll ~eXl~qiR nmtvvSaqi~ sldpvlesie D ~ a c a 0 q f~ ~rkla0si~

smmqk'~-a slWklk1~-& arl]mqVtql nll~tcp

aiadm~,auum lvai~,lUklm keqvl~ftl~ qeqvp~

r~lrdqTaea rsrwagfePr mnlJU~iEsl q q ~ v p g f 1 ~ n

~l~rlsll~n srafqqar~r iyrlky~ptv ekandeiqai klkhpnfkl~ Natvl~ptpp g~gvl~peiw rtsskaigir qsvqfefstr k i ~ q k p q k e v -tisnl~QXX D~EaXaOeVR t q l i t h e l K - ~ a e a & e y p k P h -e~vvl lerkla~qik QK'vqqe'Ve-e W a ~ r P h llal~Ell~al avrllr~thg chprlqpfhl aavrn . . . . . . . . . . . . . . . . . . . ~ v v l serkla~qik QK'Vrqeve-e w v a ~ V n • ...... s a l aarllq~ahs c~lrlrpfhl aavrn . . . . . . . . . . . . . . . . . . . . L~r/%--Pl IIMTUM~

AI~X~alIER ~IT~a~ER d~nllinvEl d~InlTinvEl d~nvTin~l d ~ a t ~ d~mtTWrl~l Idlekl~anl~ Uhs~Ir~t uhs~In~t

~ ~ v ~ Elal~liGika K~aqEiGika r~a~miGita E~al~aeiva ~kd~nvdc v1&Qr~i~ r~aa~in~ r~aavWeimB

r|yDld~tt~ luByDL~ett~ thik/~Prttt thirst ehiqLPrdit nfih~dee&t iiek~Paeit etkrLPa~tt etivkPallv~ etimkPa~i8

~ X M ~aeX~eLIDt •aeLX~XDt IMlev~MyIts Esevl~yvis EteLl~nXts |fev~dcyvDq • v~Xsd ~ L X a d Beeld~nsXrk BeeLLnLInk

M ~ M X

ECDC STD~ HCDCS RCDCS SCDCS YCDCS YI~)CS DNDC

r~a~Ti~X L~a~Ti~X LlleDl~f IAleDaTvl~f Llle~/vh~x ~ t h ~ X ilk~Dd~ihel I~F~tGX ~nv~Gl

MInDdnvDel

L~LPL--I~ L~X~--I~ L~Llq~sen i~Llq~sen iVQMPLdsvh i~dPL--I~ LiQLPL--Pr LV~Pv--Pe LV~.,PL--Pe LVQL,PL--pe

GiDnvkvX~ GiDnvkvX~ sinteevina sinteavina piDshaXtda hlDed~Itsdl hlDeTtItna hinertXcna hiDerkvcna hiDerllIcna

PPDC ECDC STDC RCDCS SCDCS ¥CDCS YMDCS DNDC ImDC ~DC

TPI~XXTLIdl T~reXvTLL~ TPr~XvTLIdl TIN~Gcle~ik ~le~ik T~aecveLik TA~rXXer~h T~2~cmkX~ TPiGvkr~I~ TPw~vweiik ~ i i k

ItT~IK'VlteEH lTNXdtfGln RTNZdty~ln e~xjvpi&er~ etq~qiaer~ ]~c~iae~n ka]~tieesr eahvkld~n h~kX~tl~rn Rt~Xptl~Kn ~qlptl**~,n

& ~ &WviG&~MI~ &WvieJ~XV &Wv~z~kX~ &Vv~Irlk3~ vVvler~rXV sWiGr~dXV &WlGr~sX~ &Wv~r~knV VVvaGrS)~mV v~vaerSknV

e~Nl2 ~-~.va, •8 1 q l ~ ~lllql~~'nTa. • alq~l..... avwa 4L~P~'~-dLL.L ~tPv~-lZ&k ~k~va -md~k ~nPia - s ~ k sl~ai-l&h ~i~-L~h ~ia~-LLh

&-G . . . . . . . . &-~ . . . . . . . . &-G . . . . . . . . wn........ wn . . . . . . . . we ........ sl ........ na ........ kdGkyatkem t~Gaherr~g tdeaherpgg

&TTTTC~RF cT%~EvtlIIIF cTTTvtlIRF n&TvTT~k n a ~ k h A ~ k nsTvTit~sk n~TvTv~sh d~TvTi~lly ~TvTi~ d~TvTi~

T--(~LmI T--knLrh~l T--kDLrtUlv T--ahLdeev T--~r~kev T--knL~eit T--rnlasyl T--rniaew TppkeLarXc TpkeqLk~Xt TpkeqLkkllt

ppDC ECDC STDC HCDCS RCDCS SCDCS YCDCS YMDCS DMDC MMDC RMDC

I~TVIIR~X ~ ...... I~aiVXDV~I ~ ...... t ~ I~aiVXDV~X ~ ....... Kv~OW, - ~ i ~ X I~vpdd-kkp ngrK~a -~VVXDcGX ~oddtk-p nffrEvv~a - e m V V l q ~ I Rpisdptkks - g q r L v ~ V a kktv~I~W~t Myva~p~kks - g f l c v ~ M ~ I~L~v~ID~X llyvpdiskks - g q K L v ~ V d -~J~qIDV~X l ~ i k d e ~ - - t g q f l ~ d l~aVI~X ~tvqdpv--t akpl~vg~Vd I~a~DW~I MRvhdpv--t akpl~v~D~d

iT~v~eev~

XT~a~LX~T LLaC~Um ~tLX~T LqaCveyhdp NT~tI~X~T ~ieyhdp IITVIK~q~T v e s a k r f l e k E~WUU,xq~J~ v e s a q r f l M k It'5'qm~m,"tST v~aarr~l~r Liaakr~aee ~ v ~ W v L L ~ U ~ r ~ fve X~V~ark~ f ~I ii~akkvlrp ii~akkvlr I

HCDCS

Fe~ak~T l~e~aa~sT ~elLl~

XTI~XqMr~ XTI~I~ XTl~-~i~vw

~e~u~

ZT~NeWW

yeeakvqvl~h ~ Pn-aikyvhl X T I ~ I ~ R ~ ~vkektsl X T ~ Feevr~vi~h X T I ~ v ~

Xal~e~ Xa~MEDW~I X a l ~ I vskll~l vla~e~ VDfkAKDWD~

LAwrMmsn~ ML~_.VlmdLJ LaLlcffem~ T,~Llc~e~-d L&Ivqveq~e ~llq~s~-P Lkiiq~larP LtaTiVge~B X ~ i L ~ X~iL~enP

~ X ~ -

- ~ X ~

~ c tsi~a~M~ar tsi~a~ar ntilleglvrv g~iEM~mk ~rTMa~ak nev~i~Rta-

-~a~rZal~ ~Indcfil~ ~llkdcfip~ ~dl~-fip~ }L~flp~ kg~k~yfipC - idm~anlpa

141

vsI~IIKDVI~ l i v e -

-idq~l~lpa

~ X L ~ V enlUDLLiVlkV eha.ILivav .~q.xLvvat .~ZLVVat kt~DXLV~Ikl h ~ X ~ i ~Xvia~c a~mxiv~av il~xvi~a il~Xvis~a

X~ ........ 222 IE ........ Ix . . . . . . . . ~ ..... ~p ....... Xlp ....... fltprdg~8~4 XX ........ v~p . . . . . . IX ........ IX ........

~ I ~ X ~ ~l~J~xlqldll ~mq~n~ea eq~.~kee~ ~ e e ~ G r Q ~ GCl~eFvkG~ql Gi~qyvks~l e~l~IXt~dn Gilqrlxta~ Gil~IXta~

X q~x q~x ~p-

285

llap,s~psek-rks~X eelevfk~k qrgwatnX eerev Ik~k e l ~ a t r ~

Fig. 4. Alignment of the predicted amino-acid sequence of PPDC with those of ten other dehydrogenase-cyclohydrolase enzymes. Asterisks over sequences represent positions of conservation between only bacterial and mitochondrial enzymes. Abbreviations: PPDC, P. phosphoreum DC; ECDC, E. coli DC; STDC, S. typhimurium DC; HCDCS, human cytoplasmic DCS; RCDCS, rat cytoplasmic DCS; SCDCS, SF9 cytoplasmic DCS; YCDCS, yeast cytoplasmic DCS; YMDCS, yeast mitochondrial DCS; DMDC, D. melanogaster mitochondrial DC; MMDC, mouse mitochondrial DC; HMDC, human mitochondrial DC.

3.6. Multiple sequence alignment The predicted amino-acid sequence of PPDC was aligned to ten other dehydrogenase-cyclohydrolase amino-

Miuw3smdrialBO~mc6~m~

acid sequences using the program WinDNASIS. The computer-generated alignment was refined manually and analysed by CLUSTAL-W(Fig. 4). The following overall percent identities to the other sequences were calculated: E. coli,

Cycopaonic T

~

Yeast T

~

Bacter/a/BO~z6oncz/

Fig. 5. Unrooted phylogenetic tree of aligned dehydrogenase-cyclohydrolase amino-acid sequences; abbreviations are given in the legend for Fig. 4.

P.D. Pawelek, R.E. MacKenzie / Biochimica et Biophysica Acta 1296 (1996) 47-54

73.7%; S. typhimurium, 74.1%; S. cerevisiae (mitochondrial), 43.1%; S. cerevisiae (cytoplasmic), 46.7%; D. melanogaster (mitochondrial), 51.5%; SF9 (cytoplasmic), 40.9%; mouse (mitochondrial), 46.4%; rat (cytoplasmic), 42.3%; human (mitochondrial), 47.4%; human (cytoplasmic), 43.8%. A phylogenetic analysis of this alignment was performed using the program CLUSTAL-W.An unrooted phylogenetic tree was generated in which the bacterial and mitochondrial DC sequences were clearly seen to be distinctly grouped, yet less distant to each other than to the DC sequences from eukaryotic trifunctional enzymes (Fig. 5). This tree was found to be identical to one generated by a bootstrap analysis of the alignment using 1000 iterations (also performed by CLUSTAL-W),indicating that the grouping represented by the tree is significant.

4. Discussion Photobacterium phosphoreum expresses a bifunctional 5,10-methyleneH 4folate dehydrogenase-5,10-methenylH 4 folate cyclohydrolase enzyme (PPDC) which can utilize either NAD + or NADP ÷ as a dinucleotide cofactor in its dehydrogenase reaction, l)ehydrogenase enzymes which possess dual coenzyme specificity are not uncommon; isocitrate dehydrogenase from R. vannielii [27], glucose6-phosphate dehydrogenase from Pseudomonas C. [28], and glutamate dehydrogenase from A. brazilense [29], as well as a number of eukaulotic glutamate dehydrogenases can all utilize both NAD ÷ and NADP +. However, dual cofactor specificity in the methyleneH4folate dehydrogenase-cyclohydrolase enzymes is rare. The only other dehydrogenase-cyclohydrolase known to exhibit dual cofactor specificity is the bifunctional human mitochondrial enzyme. This enzyme has been shown to be able to utilize NADP + as a cofactor, but with a reduction in kca t to about 20% that of the NAD ÷-dependent activity [5]. That both PPDC and the human mitochondrial enzyme share the characteristic of dual cofactor specificity prompted us to further investigate the functional relationship between the two enzymes. It has already been hypothesized that the human mitochondrial enzyme evolved from an NADP+-dependent dehydrogenase-cyclohydrolase to more efficiently use NAD ÷ in the presence of inorganic phosphate and magnesium. This hypothesis was supported by evidence which showed that the addition of inorganic phosphate significantly reduces the NAD + K m [5]. It is interesting to note that another folate metabolic enzyme, methyleneHnfolate reductase, also has its K m for NADH reduced by inorganic phosphate [30]. In PPDC, we observed that a dramatic increase in the N A D ÷ K m occurred upon replacement of the standard potassium phosphate assay buffer with Mops, indicating that the presence of inorganic phosphate greatly enhances binding of the NAD + cofactor to the enzyme. This activa-

53

tion of the NAD+-dependent dehydrogenase activity of PPDC, like that of the human mitochondrial enzyme, is phosphate-specific since substitution of potassium phosphate with potassium sulfate was observed to have no effect on the NAD +-dependent dehydrogenase activity. There are two significant differences in this activation from that of the human mitochondrial enzyme, however. The kac t for phosphate is significantly higher for PPDC and we observed that Mg 2+ had no effect on phosphate binding. These differences suggest that phosphate activation of the NAD+-dependent dehydrogenase of PPDC has not been optimized over the course of the enzyme's evolution. This is consistent with our observation that the kcat//Km for NADP ÷ is approx. 20-times higher than for NAD +, suggesting that NADP + is the preferred in vivo cofactor. The similarities in kinetic properties between PPDC and the human mitochondrial enzyme is supported by our analysis of a multiple alignment of predicted amino-acid sequences of dehydrogenase-cyclohydrolases from 11 species. It revealed that the bacterial sequences share a higher degree of similarity with the eukaryotic NAD +-dependent mitochondrial enzymes than with the dehydrogenase-cyclohydrolase domains from eukaryotic NADP +dependent trifunctional enzymes. This was made even more apparent by a phylogenetic tree generated from this alignment which shows that the bacterial and mitochondrial enzymes likely share a common point of divergence. We therefore propose that bacterial and mitochondrial dehydrogenase-cyclohydrolase enzymes diverged from a common ancestral dehydrogenase-cyclohydrolase enzyme with dual cofactor specificity. Due to metabolic pressures specific to the cellular contexts of these enzymes, the human mitochondrial dehydrogenase-cyclohydrolase evolved to more efficiently utilize phosphate in the presence of magnesium resulting in a higher NAD +-dependent catalytic activity, while PPDC evolved to more efficiently utilize the NADP + cofactor, retaining a weak ability to bind NAD +.

Acknowledgements We thank Dr. E.A. Meighen for providing P. phosphoreum strain NCMB 844 as well as genomic DNA from this strain. This work was supported by Grant MT 4479 from the Medical Research Council of Canada.

References [1] Tanner, R.S., Wolfe, R.S. and Ljungdahl, L.G. (1978) J. Bacteriol. 134, 668-670. [2] Moore, M.R., O'Brien, W.E. and Ljungdahl, L.G. (1974) J. Biol. Chem. 249, 5250-5253. [3] West, M.G., Barlowe, C.K. and Appling, D.R. (1993)J. Biol. Chem. 268, 153-160.

54

P.D. Pawelek, R.E. MacKenzie / Biochimica et Biophysica Acta 1296 (1996) 47-54

[4] Pelletier, J.N. and MacKenzie, R.E. (1995) Biochemistry 34, 12673-12680. [5] Yang, X.M. and MacKenzie, R.E. (1993) Biochemistry 32, 1111811123. [6] Drury, E.J., Bazar, L.S. and MacKenzie, R.E. (1975) Arch. Biochem. Biophys. 169, 662-668. [7] Rios-Orlandi, E., Zarkadas, C.G. and MacKenzie, R,E. (1986) Biochim. Biophys. Acta 871, 24-35. [8] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. [9] Murley, L., Mejia, N.R. and MacKenzie, R.E. (1993) J. Biol. Chem. 268, 22820-22824. [10] Vernet, T., Dignard, D. and Thomas, D.Y. (1987) Gene 52, 225-233. [11] Tabor, S. and Richardson, C.C. (1985) Proc. Natl. Acad. Sci. USA 82, 1074-1078. [12] Tan, L.U.L., Drury, E.J. and MacKenzie, R.E. (1977) J. Biol. Chem. 252, 1117-1122. [13] Schiestl, R.H., and Gietz, R.D. (1989) Curr. Genet. 16, 339-346. [14] Kunkel, T.A., Roberts, J.D. and Zakour, R.A. (1987) Methods Enzymol. 154, 367-382. [15] Laemmli, U.K. (1970) Nature (London) 227, 680-685. [16] Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. [17] D'Ari, L. and Rabinowitz, J.C. (1991) J. Biol. Chem. 266, 2395323958.

[18] Rossolini, G.M., Muscas, P., Chiesurin, A. and Satta, G. (1994) FEMS Microbiol. Lett. 119, 321-328. [19] Staben, C. and Rabinowitz, J.C. (1986) J. Biol. Chem. 261, 46294637. [20] Shannon, K.W., and Rabinowitz, J.C. (1988) J. Biol. Chem. 263, 7717-7725. [21] Hum, D.W., Bell, A.W., Rozen, R. and MacKenzie, R,E. (1988) J. Biol. Chem. 263, 15946-15950. [22] Peri, K.G., Belanger, C. and MacKenzie, R.E. (1989) Nucleic Acids Res. 17, 8853. [23] Price, B.D. and Laughon, A. (1993) Biochim. Biophys. Acta 1173, 94-98. [24] Belanger, C. and R.E. MacKenzie. (1989) J. Biol. Chem. 264, 4837-4843. [25] Thigpen, A.E., West, M.G. and Appling, D.R. (1990) J. Biol. Chem. 265, 7907-7913. [26] Tremblay, G.B. and MacKenzie, R.E. (1995) Biochim. Biophys. Acta 1261, 129-133. [27] Leyland, M.L. and Kelly, D.J. (1991) Eur. J. Biochem. 202, 85-93. [28] Ben-Bassat, A. and Goldberg, I. (1980) Biochim. Biophys. Acta 611, 1-10. [29] Maulik, P. and Ghosh, S. (1986) Eur. J. Biochem. 155, 595-602. [30] Matthews, R.G. and Kaufman, S. (1980) J. Biol. Chem. 255, 60146017.