Idea Transcript
Chapter 6 Molecular and biochemical characterization of a phosphoglycerate mutase isoenzyme from the methylotrophic actinomycete Amycolatopsis methanolica
A. M. C. R. Alves, G. J. W. Euverink and L. Dijkhuizen
Submitted to Gene.
99
Chapter 6
618-bp open reading frame (ORF2) located downstream from the prephenate A dehydratase gene in the actinomycete Amycolatopsis methanolica has been sequenced and characterized. The deduced amino acid sequence of ORF2 showed highest similarity with the putative 2,3-diphosphoglycerate (2,3-DPG) dependent phosphoglycerate mutase protein from Saccharomyces cerevisae (PGM2). Expression of ORF2 in E. coli confirmed that it encoded a functional 2,3-DPG dependent phosphoglycerate mutase (PGM2). Its calculated subunit Mr of 22,000 is clearly lower than that of the PGM1 enzyme (Mr of 28,000), previously purified from A. methanolica (Alves et al., 1994. J. Bacteriol. 176:6827-6835). Also the N-terminal amino acid sequences of PGM1 and PGM2 showed significant differences. The data thus provide evidence for the presence of PGM isoenzymes in A. methanolica. This is the first report that provides clear evidence for the presence of PGM isoenzymes in a Gram-positive bacterium.
INTRODUCTION Actinomycetes are abundant producers of secondary metabolites, including many antibiotics (Vining, 1992; Edwards, 1993). Only a few genes encoding enzymes involved in the primary metabolism of actinomycetes have been characterized (Angell et al., 1992; White et al., 1992; Bramwell et al., 1993; Kormanec et al., 1995). A detailed understanding of the regulation of central metabolism in these organisms is considered to be crucial for the further improvement of industrial processes for secondary metabolite production. Aims of our studies are the characterization of enzymes involved in glucose utilization in actinomycetes. Previously, we have reported the biochemical and molecular characterization of phosphofructokinase enzymes from the actinomycetes A. methanolica and Streptomyces coelicolor A3(2), involving purification of the proteins and cloning and sequencing of the genes (Alves et al., 1994, 1996b). PGM proteins (EC 5.4.2.1) comprise a family of glycolytic enzymes which catalyse the interconversion of 2- and 3-phosphoglycerates, involving the formation of a phosphohistidine intermediate (Fothergill-Gilmore and Watson, 1989). Various types of PGM proteins have been characterized that are kinetically and structurally distinct, but still have many features in common (Fothergill-Gilmore and Watson, 1989; Fothergill-Gilmore and Michels, 1993). Basically, two types PGM can be distinguished: Type I PGM is dependent on the presence of 2,3-DPG for its full activity, type II PGM does not require 2,3-DPG. The 2,3-DPG dependent PGM enzymes have been found in vertebrates, Bacteria and fungi. Only four bacteria type I PGM enzymes, from Zymomonas mobilis (Pawluk et al., 1986), A. methanolica (Alves et al., 1994) S. coelicolor A3(2) (White et al., 1992) and E. coli (D'Alessio and Josse, 1971), thus far have been characterized. The primary sequence of this S. coelicolor A3(2) protein showed high similarity with the 2,3-DPG dependent PGM enzymes from Eukarya (White et al.,
100
PGM isoenzymes in A. methanolica
1992). The PGM1 enzyme purified from A. methanolica (Alves et al., 1994) had similar kinetic properties as the S. coelicolor A3(2) PGM; these enzymes were characterized as monomeric and tetrameric proteins with a subunit Mr of 28,000 and 29,000 respectively. The 2,3-DPG independent PGMs of type II have been found in plants, Bacteria and lower Eukarya. Bacillus species contain a 2,3-DPG independent PGM enzyme that requires Mn2+ ions for activity (Vazquez-Leyva and Setlow, 1994). Based on sequence similarities and not on actual PGM enzyme activities, several putative PGM isoenzymes have been found in both Bacteria and Eukarya (Vandenbol et al., 1994; Tatusov et al., 1996). In E. coli both types of PGM enzymes appear to be present (D'Alessio and Josse, 1971; Burland et al., 1995). Previous studies of aromatic amino acid biosynthesis in A. methanolica involved characterization of its prephenate dehydratase enzyme (PDT) (Euverink et al., 1995; Vrijbloed et al., 1995b). Sequencing of the pdt gene revealed the presence of a second open reading frame (ORF2) directly downstream from pdt. Its deduced N-terminal amino acid sequence showed similarity with the enzyme PGM. In this paper we describe the molecular and biochemical characterization of ORF2 encoding a second PGM protein in the actinomycete A. methanolica. This is the first report that provides clear biochemical evidence for the presence of PGM isoenzymes in a Gram-positive bacterium.
MATERIAL AND METHODS Bacterial strains and plasmids. Bacterial strains and plasmids used are listed in Table 1. Media and growth conditions. A. methanolica was grown on mineral medium as described previously (Alves et al., 1994) supplemented with glucose (10 mM) or methanol (60 mM). E. coli strains DH5", BL21(DE3) and BL21(DE3)pLysE were grown on Luria-Bertani (LB) medium at 37oC and 30oC respectively. When necessary the following supplements were added (in micrograms per millilitre): Ap and Cm, 100; IPTG, 0.4 mM. Agar was added for solid media (1.5% [wt/vol]). DNA manipulations. Preparation of plasmid DNA was done by the alkaline lysis method (Birnboim and Doly, 1979). DNA-modifying enzymes were obtained from commercial sources and used as recommended by the manufacturer. All other DNA manipulations were done according to standard protocols (Sambrook et al., 1989). Nucleotide sequencing. Double stranded DNA was sequenced with the T7 DNA polymerase, with unlabelled primers and fluorescein-labelled ATP (Zimmermann et al., 1990; Voss et al., 1992). Construction of the expression vector pETPGM101. In order to obtain pgm2 under control of the T7 promoter, two restriction sites (NdeI and BamHI for translational fusion) were introduced by the polymerase chain reaction (PCR) in pNAT115. A NdeI restriction site was introduced at the start codon and a BamHI site downstream of pgm2. The PCR was done with a GeneAmp PCR system 2400 (Perkin Elmer, California) using the standard protocol, except that DNA denaturation
101
Chapter 6
Table 1. Bacterial strains and plasmids used in this study. Strains and plasmids Strains E. coli DH5" BL21(DE3) BL21(DE3)pLysE
A. methanolica WVI
Relevant genotype or characteristicsa
supE44 )lacU169 (N80lacZ) M15) hsdR17 recA1 endA1 gyrA96 thi-1-relA1 F-ompT hsdSB (rB-mB-) gal dcm (DE3) F-ompT hsdSB (rB-mB-) gal dcm (DE3)pLysE(Cmr) pMEA300-free strain
Plasmids pNAT115
15-kb fragment containing pdt and pgm in pWV138 pAMPGM101 618-bp PCR fragment cloned in EcoRV site of pBluescript KS+ (Apr) pETPGM101 618-bp NdeI-BamHI fragment cloned in pET-3b restricted with NdeI-BamHI pET-3b 4639-bp expression vector (PM10, TM), with a NdeI restriction site in the ATG start codon (Apr) + pBluescriptKS Apr, phagemid derived from pUC18, lacZ a Cmr, chloramphenicol resistant; Apr, ampicillin resistant
Source or reference
Bethesda Research Laboratories Novagene Novagene Vrijbloed et al. 1995a
Vrijbloed et al. 1995b This study This study Novagene
Stratagene
was done at 95oC for 6 min, primer annealing was done at 60oC during the first 5 cycles and at 57.5oC during the following 25 cycles. The primers used were pPGMS (5'-GCGCATATGAAACTGTACCTGGTCCGG, NdeI site underlined and ATG start codon in bold) and pGMAS (3'- CGCGGATCCCTACATGGGCAAACCGGCCCA, BamHI site underlined). The PCR product corresponding to the size of pgm2 (618-bp), was excised from the agarose gel and ligated into pBluescriptKS+ cut with EcoRV, resulting in pAMPGM101 (Table 1). This construct was then cut with NdeI and BamHI, and the 618-bp band subsequently cloned in pET-3b cut with NdeI and BamHI, resulting in pETPGM101. This construct was subsequently transformed into the E. coli host strains BL21(DE3) and BL21(DE3)pLysE. The PCR- amplified 618-bp DNA fragment was fully sequenced, confirming that no errors had been introduced by enzymatic amplification. Expression of phosphoglycerate mutase in E. coli BL21(DE3) and BL21(DE3)pLysE strains. E. coli strains transformed with pETPGM101 were grown overnight in LB supplemented with Ap at 37°C for BL21(DE3) or with Cm at 30oC for BL21(DE3)pLysE until an OD660 of 1 was reached, after which 0.4 mM IPTG was added. Growth was allowed to continue for 3 h. Cells were harvested as described below. Preparation of extracts and enzyme assay. Cells were washed in buffer containing 50 mM Tris-HCl pH 7.5 and disrupted by passing the cells three times through a French Pressure cell at 140 MPa. Unbroken cells and debris were removed by centrifugation of the lysate at 40,000xg for
102
PGM isoenzymes in A. methanolica
30 min and the supernatant was used as cell-free extract. PGM (EC 5.4.2.1) was assayed in a reaction mixture containing 50 mM Tris-HCl pH 7.5, 5 mM MgCl2, 30 mM KCl, 1 U enolase, 2 U pyruvate kinase and 2 U lactate dehydrogenase, 0.15 mM NADH, and limiting amounts of protein in a final volume of 1 ml at 37°C. The reaction was started by addition of 2.5 mM 3-phosphoglycerate. The 2,3-DPG-dependent PGM activity was assayed in the same reaction mixture supplemented with 0.2 mM 2,3-DPG. One unit of activity is the amount of enzyme that oxidizes 1 µmol of NADH per minute. Protein characterization. PGM2 protein in cell-free extracts of cultures containing pETPGM101 was partially purified in a System Prep 10 liquid chromatography system (Pharmacia LKB Biotechnology Inc.). Protein determination and estimation of subunit size. Protein concentrations were determined with a Bio-Rad protein determination kit using bovine serum albumin as a standard (Bradford, 1976). SDS-PAGE electrophoresis was performed (Laemmli and Favre, 1973) with the following pre-stained marker proteins (Biolabs): maltose-binding protein fused with ß-galactosidase (175,000), maltose-binding protein fused with paramyosin (83,000), glutamic dehydrogenase (62,000), aldolase (47,500), triosephosphate isomerase (32,500), ß-lactoglobulin (25,000), lysozyme (16,500), and aprotinin (6,500). N-terminal amino acid sequence analysis. Purified PGM1 protein (Alves et al., 1994) was applied to a Pro-spin cartridge (Applied Biosystems), containing a polyvinylidene difluoride membrane. Sequencing was performed on an Applied Biosystems model 477A/120A automated gas-phase sequencer equipped with on-line high-pressure liquid chromatography for detection of the phenylthiohydantoin amino acid derivatives (Eurosequence, Groningen, The Netherlands). Sequence alignment. Amino acid sequences were aligned with Clustal W using the Blosum matrix with the following parameters: Gap open penalty: 10; Gap extension penalty: 0.05 (Thompson et al., 1994). The following groups of amino acids were considered to represent conservative changes: P, A, G, S, T; Q, N; E, D; R, K; I, L, V, M; F, Y, W, and C; H. Phylogenetic analysis. PGM alignments were made with Clustal W (Thompson et al., 1994). The programs supplied in the PHYLIP 3.5c package were used to determine phylogenetic relationships (Felsenstein, 1985). Distance matrices were calculated with PROTDIS. A phylogenetic tree was constructed via the neighbour-joining method (Saitou and Nei, 1987) implemented in the NEIGHBOR program (100 trees). A consensus tree was constructed using CONSENCE. Reliability of phylogenetic tree branches was tested via bootstrapping (Felsenstein, 1985). Accession number. The nucleotide sequence presented in this paper was entered into Genbank under accession number U73808.
RESULTS Sequence of the DNA fragment downstream from pdt The last 4 bases (GTGA) of pdt are the beginning of a second ORF (ORF2) of 618-bp (Fig. 1). Frame analysis of the DNA sequence also predicted the same ORF, with a G or a C in the third position in 92.7% of the codons reflecting the typical coding usage of actinomycetes (Wright and Bibb, 1992). The GC content of the DNA was 71.5%,
103
Chapter 6
corresponding well with the GC content of other genes of A. methanolica and other actinomycetes. ORF2 starts with GTG as a start codon and was terminated by a TAG codon. ORF2 encodes a putative protein of 205 amino acids with a calculated subunit Mr of 22,000. This was later confirmed by SDS-PAGE analysis (see below). A search of the nonredundant protein database at the National Centre for Biotechnology Information with the predicted protein sequence as a query using BlastP (Altschul et al., 1990) revealed highest similarities with hypothetical PGM proteins from Saccharomyces cerevisiae PGM2 (accession number 1279707, 48%) and E. coli PGM2 (accession number 548530, 46%), the hypothetical proteins from Mycobacterium leprae (protein B2126-C1-148, accession number 466995, 45%), and E. coli (PhpB, similar to Salmonella typhimurium
1
61 121 181 241 301 361 421 481 541 601
. . . . . . gtgcagagaggggagcaggcGTGAAGCTGTACCTGGTCCGGCACGGGCAGACCGCGTCCA v q r g e q a * V K L Y L V R H G Q T A S . . . . . . ACGTCGCGAAGAAGCTGGACACCGCGCTGCCCGGCCCGCCGCTCACCGAGCTGGGCCACG N V A K K L D T A L P G P P L T E L G H . . . . . . AGCAGGCGCGGCAGCTGGCCGAGAAGCTCGCGACCGAGCCGGTGGAGGCGGTCTACGCGT E Q A R Q L A E K L A T E P V E A V Y A . . . . . . CGCACGCGACCCGTGCCCAGCAGACGGCCGCCCCGCTCGCCCAGGCGCTGGGCATGACGG S H A T R A Q Q T A A P L A Q A L G M T . . . . . . TGAAGCGGGTCGAGGGTGTGCACGAGATCGTCGTCGGCGACCTGGAGGGCCGCCACGACC V K R V E G V H E I V V G D L E G R H D . . . . . . GGGAGGCGATCGAGCACTACCTGACGGTGCTCAGCCACTGGACGCGGGGCGAGCTGCACG R E A I E H Y L T V L S H W T R G E L H . . . . . . TCCCGATGCCCGGCGGGGAGACCGGCGAGCAGGCCAGGGCGCGGTTCACCGGTGCGATCG V P M P G G E T G E Q A R A R F T G A I . . . . . . CCGGTCTGGCCGAGCGTCACGACCTCACCCGATCGGACGGTGTCGTGGTGCTGGTCTCGC A G L A E R H D L T R S D G V V V L V S . . . . . . ACGGTGGGCTGATCCGCATCGGCGCGGAGTGGCTGGCCCCGAATGTTCGGCCCGAGCTGG H G G L I R I G A E W L A P N V R P E L . . . . . . CCGATCAGGGCCTGATCCCGAACACCGGCATCGTCGAGCTGGAGATCGCCGCGGATGGTG A D Q G L I P N T G I V E L E I A A D G . . . GCTGGCAATGCCTGAACTGGGCCGGTTTGCCCATGTAG G W Q C L N W A G L P M *
60
120 180 240 300 360 420 480 540 600 638
Figure 1. Nucleotide sequence of A. methanolica pgm2. The last 24 nucleotides of pdt are shown in front of the GTG start codon of pgm2. The deduced amino acid sequence of PGM2 is shown below. The predicted ribosome binding site is shown in bold.
104
PGM isoenzymes in A. methanolica
CobC, accession number 727431, 43%). Somewhat lower similarity was found with the alpha-ribazol-5'-phosphate phosphatase from S. typhimurium (CobC, (O'Toole et al., 1994) 38%) and the PGM protein from Schizosaccharomyces pombe ((Nairn et al., 1994), 37%). In the 2,3-DPG dependent PGM enzymes, the cofactor 2,3-DPG donates a phosphogroup to histidine-8 in an initial priming reaction; it is this phosphoenzyme that participates in the reaction (Watson, 1982). The consensus pattern is: [LIVM]-x-R-H-G-[EQ]-x(3)-N. Since part the phosphohistidine signature is present in the deduced N-terminal amino acid sequence of ORF2 (Fig. 2), we identified it as a possible PGM (PGM2). From the 13 residues identified as being part or near the active site of the S. cerevisae PGM1 enzyme (Watson, 1982; Fothergill-Gilmore and Watson, 1989; White and Fothergill-Gilmore, 1989) 8 residues were found to be identical in the A. methanolica PGM2 (Arg-7, His-8, Gly-9, Thr-20, Arg-59, Glu-86, His-181 and Gly-182); 3 residues were conserved substitutions (Ser-11 in the S. cerevisae enzyme to Thr, Gly-21 to Ala and Ala-180 to Ser) and two residues were different (Trp-22 to Leu and Ala-179 to Val). Recently, PGM1 was purified from glucose grown cells of A. methanolica (Alves et al., 1994); at that time no evidence was obtained for the presence of PGM isoenzymes. The first 20 amino acids from the N-terminal amino acid residues of the purified PGM1 were determined. The initial yield of the first amino acid was between 50-65%. Surprisingly, quite a number of differences were found when compared with the deduced N-terminal amino acids of PGM2. Both N-terminal amino acid sequences show homology with other PGMs (Fig. 2).
Expression of pgm2 under control of a T7 promoter The pgm2 gene was put under the control of a T7 promoter in the expression vector pET-3b, resulting in pETPGM101. Expression of the gene was tested in two E. coli host strains. The BL21(DE3) strain has the advantage of lacking ompT outer membrane protease, so the expressed proteins will not be degraded. The E. coli BL21(DE3)pLysE strain is used to suppress basal expression of T7 polymerase prior to induction and in this way it stabilizes pET-3b recombinants that encode proteins that can be toxic to the E. coli host strain (Studier et al., 1990). PGM activity in of BL21(DE3) (pETPGM101) did show an increase in PGM activity after IPTG induction, when compared with the same strain with pET-3b (Table 2). In strain BL21(DE3) pLysE (pETPGM101), even a better expression of pgm2 was achieved: a 3.2-fold increase in PGM activity was found after induction with IPTG (Table 2) compared with the control (strain BL21(DE3)pLysE(pET-3b). The activity found in cell-free-extracts of BL21(DE3)pLysE(pETPGM101) was enhanced 2 fold by adding 250 µM of 2,3-DPG.
105
Chapter 6
S.cer. pgm1 S.cer. pgm2 S.pombe E.coli pgm2 M.lep. pgm1 M.lep. pgm2 H.influenza S.coelicolor Z.mobilis E.coli pgm1 A.met. pgm2 A.met. pgm1
----------------PKLVLVRHG QS EWNEKNLF-TGW VD-VKLSAKGQQEAARAGELL MTKEVPYYCDNDDNNIIRLFIIRHG QT EHNVKKIL-QGH KD-TSINPTGEEQATKLGHYL ---------MTTEAAPNLLVLTRHG ES EWNKLNLF-TGW KD-PALSETGIKEAKLGGERL ---------------MLQVYLVRHG ET QWNAERRI-QGQ SD-SPLTAKGEQQAMQVATRA ----------MQQGNTATLILLRHG ES DWNARNLF-TGW VD-VGLTDKGRAEAVRSGELL ----------------MTVILLRHG RS TSNTAGVL-AGR ADGVDLDDRGREQAVGLIDRI ----------------MELVFIRHG FS EWNAKNLF-TGW RD-VNLTERGVEEAKTAGKKL -----------MADAPYKLILLRHG ES EWNEKNLF-TGW VD-VNLTPKGEKEATRGGELL ---------------MPTLVLSRHG QS EWNLENRF-TGW WD-VNLTEQGVQEATAGGKAL -------------MAVTKLVLVRHG ES QWNKENRF-TGW YD-VDLSEKGVSEAKAAGKLL ----------------VKLYLVRHG QT ASNVAKKLDTAL PG-PPLTELGHEQARQLAEKL -------------AELGTLVLLRHG QS TWNAEN . *** . * . . * .*
42 58 49 43 48 43 42 47 43 45 43 20
S.cer. pgm1 S.cer. pgm2 S.pombe E.coli pgm2 M.lep. pgm1 M.lep. pgm2 H.influenza S.coelicolor Z.mobilis E.coli pgm1 A.met. pgm2
KEKKVYPDVLYTSKLSR AIQTANIALEKADRLWIPVNRSWRLNE RHYGDLQGKDKAETLK RSRGIHFDKVVSSDLKR CRQTTALVLKHSKQENVPTSYTSGLRE RYMGVIEGMQIT-EAE KSRGYKFDIAFTSALQR AQKTCQIILEEVGEPNLETIKSEKLNE RYYGDLQGLNKDDARK KELGITH--IISSDLGR TRRTAEIIAQACG---CDIIFDSRLRE LNMGVLEKRHIDSLTE AEHNLLPDVLYTSLLRR AITTAHLALDTADWLWIPVRRSWRLNE RHYGALQGLDKAVTKA AELPIRA--VVCSPLLR CRRTINPLAETLC---LEPFIDDRLSE VDYGEWTSRSIGDLAK LDKGYEFDIAFTSVLTR AIKTCNIVLEESHQLWIPQVKNWRLNE RHYGALQGLDKKATAE KDAGLLPDVVHTSVQKR AIRTAQLALEAADRHWIPVHRHWRLNE RHYGALQGKDKAQTLA AEKGFEFDIAFTSVLTR AIKTTNLILEAGKTLWVPTEKDWRLNE RHYGGLTGLNKAETAA KEEGYSFDFAYTSVLKR AIHTLWNVLDELDQAWLPVEKSWKLNE RHYGALQGLNKAETAE ATEPVEA--VYASHATR AQQTAAPLAQALG---MTVKRVEGVHE IVVGDLEGRHDREAIE * * * . * *
102 117 109 98 108 98 102 107 103 105 98
S.cer. pgm1 S.cer. pgm2 S.pombe E.coli pgm2 M.lep. pgm1 M.lep. pgm2 H.influenza S.coelicolor Z.mobilis E.coli pgm1 A.met. pgm2
K--FGEEKFNTYRRSFDVPPPPIDASSPFSQKGDERYKYVDPNVLPETESLALVIDRLLP K--YADKHGEGSFRNFGEKSD-------------------------------DFVARLTG K--WGAEQVQIWRRSYDIAPP-------------------------NGESLKDTAERVLP E----EE--N-WRRQLVNGT------------VDGR--------IPEGESMQELSDRVNA R--YGEERFMAWRRSYDTPPPPIEKGSEFSQDADPRYTDIG--GGPLTECLADVVTRFLP EPLWQVVQAHPSAAVFPSGEG------------LAQVQ------VRAVAAIREYDRRFTS Q--YGDEQVHIWRRSYDISPPDLDPQDPNSAHNDRRYANIPSDVVPNAENLKLTLERALP E--FGEEQFMLWRRSYDTPPPALDRDAEYSQFSDPRYAMLPPELRPQTECLKDVVGRMLP K--HGEEQVHIWRRSYDVPPPPMEKGSKFDLSGDRRYDGVK---IPETESLKDTVARVLP K--YGDEQVKQWRRGFAVTPPELTKDDERYPGHDPRYAKLSEKELPLTESLALTIDRVIP H--YLTVLSH-WTRGELHVP------------------------MPGGETGEQARARFTG *
160 144 142 131 164 140 160 165 158 163 131
S.cer. pgm1 S.cer. pgm2 S.pombe E.coli pgm2 M.lep. pgm1 M.lep. pgm2 H.influenza S.coelicolor Z.mobilis E.coli pgm1 A.met. pgm2
YWQDVIAKDLLS-GKT-VMIAA -----HG NSLRGLVK-----HLEGISDADIAKL-NIPT CVEEEVAEASNE-GVKNLALVS -----HG AIRMILQ------WLKYENHQAHKII-VFNT YYKSTIVPHILK-GEK-VLIAA -----HG NSLRALIM-----DLEGLTGDQIVKR-ELAT ALES--CRDLPQ-GSR-PLLLS -----HG IALGCLVS-----TILGLPAWAERRL-RLRN YFTDVIVPDLRT-GRT-VLIVA -----HG NSLRALVK-----HLDEMSDDEVVGL-NVPT EHGGDTLWVACTHGDVIKAVIA DAFGMHL DSFQRVIADPGSVSVIRYTQLRPFVLHVNHT FWEDQIAPAMLS-GKR-VLVVA -----HG NSLRALAK-----HIIGISDAEIMDF-EIPT YWFDAIVPDLLT-GRT-VLVAA -----HG NSLRALVK-----HLDGISDADIAGL-NIPT YWEERIAPELKA-GKR-VLIGA -----HG NSLRALVK-----HLSKLSDEEIVKF-ELPT YWNETILPRMKS-GER-VIIAA -----HG NSLRALVK-----YLDNMSEEEILEL-NIPT AIAGLAERHDLTRSDGVVVLVS -----HG GLIRIGAE-----WLAPNVRPELADQGLIPN . . * . .
207 191 189 176 211 200 207 212 205 210 181
S.cer. pgm1 S.cer. pgm2 S.pombe E.coli pgm2 M.lep. pgm1 M.lep. pgm2 H.influenza S.coelicolor Z.mobilis E.coli pgm1 A.met. pgm2
GIPLVFELDENLKPS---------KPS-YYL-DPEAAAAGAAAVANQGKKSVTIVDYVKDSKQFI---------VR---RVGNTQHLGDGEFVVSDLRLRGVPIVYHLDKDGKYV---------SK---ELIDN----------------CSISRVDYQESLWLA---------SG--WVVETAGDISHLDAPALDELQRGIPLRYDLDADLRPV---------VPGGTYL-DPEAAAAVISQARP----GAQLSSALRAAMKPSGESHSDSCGERSGEIRNEPDAAVPSGDAVLGGSADGQPLVLKLDDKLNYV---------EH--YYL-------------------GIPLSYELNAEFKPL---------NPGGTYL-DPDAAAAAIEAVKNQGKKK GQPLVYELNDDLTPK---------DR--YFLNER----------------GVPLVYEFDENFKPL---------KR--YYLGNADEIAAKAAAVANQGKAK TGIVELEIAADGGWQ---------CLN--WAGLPM----------------
246 229 211 215 247 250 227 253 228 250 205
(44%) (48%) (43%) (46%) (42%) (45%) (46%) (34%) (40%) (41%) (100%)
Legends are shown on the next page
106
PGM isoenzymes in A. methanolica
Figure 2. Sequence alignment of 2,3-DPG dependent PGMs from S. cerevisae PGM1 (548534), S. cerevisae PGM2 (1279707), S. pombe (548532), E. coli PGM2 (548530), M. leprae PGM1(467056), A. methanolica PGM2 (73808), H. influenza (1220849), S. coelicolor A3(2) (417496), Z. mobilis (400804), E. coli PGM1 (400802), M. leprae PGM2 (466995) and the N-terminus of` the purified PGM1 from A. methanolica. Bold characters refer to residues identified as part of the active site of the 2,3-DPG dependent PGM1 from S. cerevisae (Fothergill-Gilmore and Watson, 1989; Watson, 1982; White and Fothergill-Gilmore, 1989). Alignment was done with Clustal W. , similar residues; *, identical residues. Accession numbers are given between brackets. Percentage of similarity of each PGM protein to PGM2 of A. methanolica is given at the end of the alignment.
Table 2. Specific activities of PGM in cell-free extracts of E. coli *. Plasmid and strain
Uninduced Induced Specific activity Specific activity (U. mg protein-1) (U. mg protein-1) pETPGM101 in BL21(DE3) 2.6 7.2 pETPGM101 in BL21(DE3)pLysE 5.0 12.0 pET-3b in BL21(DE3) 2.5 3.7 pET-3b in BL21(DE3)pLysE 2.5 3.7 * activities were determined in triplo with a standard deviation of " 0.2 U.mg protein-1. Addition of 0.2 mM 2,3-DPG to the reaction mixture gave a two-fold increase of the specific activity.
Previous kinetic studies of PGM1 showed a higher activation (4.5 fold) at the same 2,3-DPG concentration (Alves et al., 1994). From these results we conclude that A. methanolica pgm2 encodes a functional 2,3-DPG dependent PGM.
Characterization of PGM2 Extracts of BL21(DE3)pLysE(pETPGM101) (IPTG induced) were used for further characterization of the PGM2 enzyme. SDS-PAGE revealed the presence of an additional band migrating at about 22,000 (Fig. 3). Gel filtration revealed a single peak indicating that the Mr of PGM2 holoenzyme was 56,000 ± 4,000, suggesting a dimeric structure. When applying the extract of BL21(DE3)(pETPGM101) to the Mono Q anion exchange column only one peak containing PGM activity was found. This peak eluted at the same salt concentration as PGM1 from A. methanolica cells grown on glucose or on methanol. As a control an extract of BL21(DE3)pLys(pET-3b) was applied to the MonoQ column. No PGM activity was found in the eluent in this case.
107
Chapter 6
1
2
3
4
175,000 83,000 62,000 47,500
32,500
25,000
Figure 3. SDS-PAGE (12.5%) of cell-free extracts of IPTG induced E. coli BL21 (DE3)pLysE strain containing the vectors pET-3b and pETPGM101. 1, molecular weight markers; 2, E. coli BL21(DE3)pLysE containing pET-3b, 8 hours after IPTG induction; 3, BL21(DE3)pLysE containing pETPGM101, 8 hours after IPTG induction; 4: the same as 3, after 3 hours IPTG induction. Numbers at the left side refer to the molecular weight markers.
Phylogenetic tree of 2,3-DPG dependent PGMs For the construction of the phylogenetic tree the full PGM sequences depicted in Fig. 2 were used. The tree was constructed by the neighbor-joining method and revealed two clusters of PGM proteins (Fig. 4). PGM2 of A. methanolica is closest to the hypothetical PGM enzymes of S. cerevisae PGM2, E. coli PGM2 and M. leprae PGM2. In the other part of the tree the PGM enzymes of S. coelicolor A3(2), PGM1 of E. coli and of S. cerevisae and the PGM enzymes of Z. mobilis, H. influenza, S. pombe and the hypothetical PGM1 of M. leprae form a second cluster.
DISCUSSION The goal of this work was to identify the putative protein encoded by ORF2, a 618-bp DNA fragment downstream of pdt of A. methanolica. Sequence analysis revealed extensive similarity with (hypothetical)PGM proteins from various sources. In some bacteria glycolytic genes are clustered in an operon. In B. subtilis pgm is flanked by triose phosphate isomerase gene (tpi) and enolase (eno) (Vazquez-Leyva and Setlow, 1994). Analysis of the DNA sequence downstream of ORF2 did not show any similarity with a glycolytic gene, so pgm2 is not part of a glycolytic cluster of genes. The expression of ORF2 in E. coli clearly showed that it encodes a functional 2,3-DPG dependent PGM (PGM2). The availability of purified PGM1 protein from A. methanolica made it possible
108
PGM isoenzymes in A. methanolica
A. methanolica PGM2
H. influenza
M. leprae PGM2
44
29
Z. mobilis 74
100 41
E. coli PGM1
E. coli PGM2
41
S. cerevisiae PGM2
43
S. pombe
97
98
S. coelicolor A3(2)
S. cerevisiae PGM1
M. leprae PGM1
Figure 4. Phylogenetic tree of 2,3-DPG dependent PGMs. The tree is based on a distance analysis of the full primary sequence of the following proteins: S. cerevisae PGM1 (548534) 246 aa, S. pombe (548532) 211 aa, E. coli PGM2 (548530) 215 aa, M. leprae PGM1 (467056) 247 aa, A. methanolica PGM2 (73808) 205 aa, H. influenza (1220849) 227 aa, S. coelicolor A3(2) (417496) 253 aa, Z. mobilis (400804) 228 aa, E. coli PGM1 (400802) 250 aa , S. cerevisae PGM2 (1279707) 229 aa and M. leprae PGM2 (466995) 250 aa. Data bank accession numbers are given in brackets. Hypothetical PGM proteins are indicated with a grey circle.
to determine its N-terminal amino acid sequence. The differences between the N-terminal amino acids of PGM1 and the deduced N-terminal amino acid sequence of PGM2 were extensive. On a total of 20 amino acids only 7 were identical; still both N-terminal amino acid sequences were similar to PGMs of other organisms (Fig. 2). These data suggested the presence of PGM isoenzymes in A. methanolica. During the purification of PGM1 from glucose grown cells of A. methanolica only a single activity peak was detected throughout the protocol. The specific activity of PGM in A. methanolica methanol cells was much lower than in glucose cells (Alves et al., 1994). An alternative PGM thus might become expressed under these conditions. However, gel filtration (Mr of 36,000) and anion exchange revealed with extracts of both glucose or methanol cells only single peaks of PGM activity eluting at the same positions. In these cells pgm2 may not be expressed
109
Chapter 6
or PGM2 is less stable than PGM1. Lack of stability of A. methanolica PGM2 was not detected in extracts of E. coli containing the expressed protein. Surprisingly, the buffer (50 mM Tris-HCl pH 7.5) and/or the conditions used during chromatography were not suitable for the PGMs of the E. coli BL21(DE3)pLysE(pET-3b) control cells, resulting in complete loss of activity. Freshly prepared extracts of this E. coli strain showed quite high specific activities for PGM. The presence of PGM isoenzymes is not unusual. In E. coli three PGM isoenzymes appear to be present (Davies and Davidson, 1982; Sofia et al., 1994; Burland et al., 1995). In S. cerevisiae five PGM homologous were found in the genome (White and Fothergill-Gilmore, 1988; Vandenbol et al., 1994) (S. cerevisae 1363818, 1279707 and 603260 direct submissions to data bank). In the actinomycetes, S. coelicolor A3(2) (White et al., 1992) only one PGM has been identified. In the related bacterium M. leprae two hypothetical PGMs have been sequenced. (direct submissions 466995 and 467056). Relations between 2,3-DPG dependent PGMs as depicted in the phylogenetic tree, show that PGM2 from A. methanolica is more related to the hypothetical PGM proteins from E. coli PGM2, M. leprae PGM2 and S. cerevisae PGM2 than to the 2,3-DPG dependent PGMs from S. coelicolor A3(2), S. cerevisae PGM1 or E. coli PGM1 (Fig 4). N-Terminal amino acid alignments showed that A. methanolica PGM1 is more similar to the PGM enzymes of S. coelicolor A3(2) and S. cerevisae (PGM1), than A. methanolica PGM2 (Fig. 2). Cloning of the A. methanolica pgm1 gene is now possible via hybridization with a specific DNA probe, synthesized on the basis of the N-terminal amino acid sequence of PGM1. Inactivation of both genes can, in future experiments, lead to more information about the roles of both PGM1 and PGM2 in the metabolism of A. methanolica. REFERENCES References are listed on pages 129-140.
110
Summary and concluding remarks
111
Summary and concluding remarks
Amycolatopsis methanolica is one of the few methanol-utilizing Gram-positive bacteria that have been characterized (de Boer et al., 1990a). It has been considered as an interesting strain for production of aromatic amino acids and secondary metabolites (de Boer, 1990; Euverink, 1995). In recent years studies of this actinomycete have resulted in characterization of the pathways involved in aromatic amino acid biosynthesis and utilization (Euverink, 1995) as well as the development of suitable plasmid vectors and transformation systems (Vrijbloed, 1996). At present, Streptomyces coelicolor A3(2) is the best studied actinomycete at the genetic level with respect to both production of antibiotics and morphological differentiation (Hopwood et al., 1985, 1995; Kieser et al., 1992; Chater and Hopwood, 1993; Redenbach et al., 1996) . Until some years ago, relatively little information was available about primary metabolism in actinomycetes. There was a general feeling that actinomycetes produce secondary metabolites as some kind of overflow metabolism largely due to a general lack of regulatory mechanisms in primary metabolism (Malik, 1980). Also it was clear that certain primary metabolites accumulated at the end of vegetative growth, especially those acting as starter units for secondary metabolites such as antibiotics. The analysis of glucose metabolism in A. methanolica, with emphasis on phosphofructokinase (PFK), is the main topic in this thesis. Studies in the actinomycete S. coelicolor A3(2) permitted a comparison of the characteristics of PFK enzymes in these two actinomycetes. The results presented in this thesis show that, at least in these two actinomycetes, glucose metabolism is clearly regulated. The characterization of these regulatory steps in primary metabolism may be important for further improvement of industrial actinomycete strains overproducing secondary metabolites (chapter 1).
Glucose metabolism in A. methanolica A. methanolica employs the Embden-Meyerhof-Parnas pathway (glycolysis) and the pentose phosphate cycle when growing on glucose as sole carbon and energy source. Normally some of the steps in the glucose utilization are regulated at the enzyme activity level and/or synthesis level. The glycolytic pathway may be controlled at the level of glucose transport, glucose kinase (phosphorylation of glucose to glucose-6-phosphate), phosphofructokinase (phosphorylation of fructose-6-phosphate to fructose-1,6-bisphosphate) or pyruvate kinase (conversion of phosphoenolpyruvate to pyruvate) (chapter 1). Analysis of glucose metabolism in A. methanolica revealed some unusual features (chapter 2). Instead of an allosteric ATP-dependent phosphofructokinase (ATP-PFK), normally present in aerobic bacteria (Fothergill-Gilmore and Michels, 1993), an unusual PPi-dependent phosphofructokinase
112
Summary and concluding remarks
(PPi-PFK) was detected in A. methanolica. In Bacteria PPi-PFK enzymes are generally found in anaerobic strains only. A good example of this is the anaerobic bacterium Propionibacterium shermanii, possessing several PPi-linked enzyme activities, such as phosphoenolpyruvate carboxytransphosphorylase, pyruvate orthophosphate dikinase and PPi-PFK. The presence of these enzymes may provide an energetic advantage (Mertens, 1991) increasing the growth efficiency of this fermentative organism, since in this case glycolysis is the only source of ATP (3 ATP instead of 2 from glucose)(Reeves, 1976; Mertens, 1991, 1993). In the aerobic bacterium A. methanolica the energetic advantage would be minimal (39 ATP instead of 38 ATP from glucose; glycolysis + TCA cycle). PPi accumulation in the cells is generally prevented by high cytosolic pyrophosphatase activities. Organisms possessing a PPi-linked glycolysis generally display a low activity or no activity of pyrophosphatase (Entamoeba histolytica, Thrichomonas vaginalis), suggesting the use of PPi-PFK to remove the excess of PPi in the cell (Mertens, 1993). Studies in E. coli have demonstrated that a basal level of pyrophosphatase expression is essential for growth, probably to counterbalance PPi synthesis. In cells with pyrophosphatase activity below 0.01 U.mg-1, PPi accumulates intracellularly to a concentration of 16 mM, resulting in nonviability of the cells (Chen et al., 1994). Characterization of the purified A. methanolica PPi-PFK enzyme showed some properties similar to those of other PPi-PFKs from bacterial sources, namely an absolute specificity for the substrate PPi, catalysis of a reversible reaction and lack of allosteric regulation. However, the tetrameric structure, the measured affinity constant for F-6-P, and the N-terminal amino acid sequence of this PPi-PFK enzyme indicated clear similarity with ATP-PFK enzymes from other sources. The A. methanolica PPi-PFK enzyme thus possessed characteristics from both groups of PFK enzymes. The cloning and characterization of the A. methanolica gene encoding PPi-PFK (pfp) revealed further interesting aspects (chapter 3). In A. methanolica the pfp gene is not flanked by other glycolytic genes, differing from the observations made for the ATP-PFK encoding gene in several other bacteria (Hellinga and Evans, 1985; Alefounder and Perham, 1989; Schwinde et al., 1993). The two flanking ORFs were tentatively assigned as aroA and chiA (on the basis of the high similarity with DAHP synthase- and chitinase-encoding genes). The deduced amino acid sequence from the pfp gene indeed revealed a high degree of similarity between the PPi-PFK of A. methanolica and bacterial ATP-PFKs and a somewhat lower degree of similarity with the other PPi-PFK enzymes described until now. Based on the amino acid residues identified for substrate binding in the allosteric ATP-PFK from E. coli, suggestions for the corresponding amino acid residues in the A. methanolica PPi-PFK enzyme were made. The 11 residues implicated in binding of F-6-P in E. coli ATP-PFK are conserved in the A. methanolica PPi-PFK enzyme. Regarding ATP binding in the E. coli ATP-PFK enzyme, the level of similarity
113
Summary and concluding remarks
was lower: 2 out of 10 residues are identical in the A. methanolica PPi-PFK enzyme (Gly-12 and Arg-73); four others are conserved substitutions (Tyr-42 to Trp, Arg-78 to Lys, Asp-104 to Glu and Ser-106 to Thr); the last four are different (Phe-74 to Thr, Gly-105 to Asp , Met-108 to Gly and Gly-109 to Val). Modelling of the PPi-PFK enzyme structure (collaboration with Dr. Daniel Ridgen), using the well characterized 3D-structure of the E. coli ATP-PFK protein (Evans et al., 1986; Shirakihara and Evans, 1988; Rypniewski and Evans, 1989; Scrimer and Evans, 1990), indicated that the primary change affecting substrate specificity at the PPi site is the replacement of Gly-105 to Asp (E. coli numbering). Gly-105 is completely conserved among the ATP-PFKs. The residue Asp-105 residue is also found in the PPi-PFK enzymes from P. freudenreichii, Giardia lamblia, Naegleria fowleri and potato beta chain (see PFK alignment in chapter 4). Another possibility is that the conserved substitution of Tyr-42 in the ATP-PFK of E. coli by Trp in the A. methanolica enzyme confers PPi specificity. In the PPi-PFK model, the Trp residue is very close to the PPi binding site preventing binding of ATP. Further studies, including determination of the 3D-structure of this PPi-PFK, comparison with the available structures of the ATP-PFK enzymes from E. coli (Shirakihara and Evans, 1988) and Bacillus stearothermophilus (Evans et al., 1986), and site directed mutagenesis of some of these residues supposed to be involved in substrate binding, may contribute to a clear understanding of structural factors determining allosteric properties and ATP/PPi substrate specificity of the PFK enzymes. From this analysis it appeared to be quite clear what determined ATP/PPi specificity in these PFK enzymes (but see below). It has been suggested that PPi-PFK is more suited for a role in anaerobic metabolism than its ATP-PFK counterpart and may have evolved from ATP-PFK at several independent occasions (Mertens, 1991; Fothergill-Gilmore and Michels, 1993). The phylogenetic analysis presented in chapter 3 shows that the PPi-PFK enzymes characterized until now form one monophylogenetic group (they form one part of the tree). These data suggested that both types of PFK evolved from a common ancestor. Since anaerobic metabolism already existed before aerobic metabolism evolved, the early PFKs must have been adapted to anaerobiosis. Interestingly, the simplest PFK proteins, PPi-PFK homodimeric enzymes lacking allostery, are encountered in P. freudenreichii (O'Brien et al., 1975), G. lamblia (Li and Philips, 1995) and E. histolytica (Reeves et al., 1974), all obligate anaerobes. Interestingly, there is a strong correlation between the structural and biochemical complexity of the PPi- and ATP-PFK proteins and the phylogenetic distance between these proteins and those from P. freudenreichii and E. histolytica. The phylogenetic analysis has placed the enzymes from P. freudenreichii and E. histolytica in the middle of the tree. On the basis of evolution correlated with complexity, we have proposed that ancestral PFKs resemble the enzymes found in P. freudenreichii and E. histolytica (chapter 3).
114
Summary and concluding remarks
Pyruvate kinase (PK) of A. methanolica was purified to homogeneity. This enzyme has a tetrameric structure as is generally the case for PK from bacterial sources (Fothergill-Gilmore and Michels, 1993); PK displayed positive cooperativity towards the substrate phosphoenopyruvate (PEP) as has been observed for the enzyme from various other sources. The PK enzyme activity was inhibited by ATP and phosphate. All these characteristics resemble those of PK enzymes from other organisms (Pawluk et al., 1986; Fothergill-Gilmore and Michels, 1993; Ponce et al., 1995). These data and PK enzyme levels in A. methanolica grown on different substrates suggest that this enzyme is regulated both at the synthesis and activity levels (chapter 2). Characterization of the single phosphoglycerate mutase (PGM1) detected in A. methanolica glucose grown cells showed that this enzyme activity is stimulated by 2,3-diphosphoglycerate (2,3-DPG), as has been observed for PGM of the only other actinomycete studied, S. coelicolor A3(2) (White et al., 1992). Chapter 6 describes the gene sequencing and gene expression in E. coli of a second PGM from A. methanolica (PGM2), encoded by ORF2 downstream of the prephenate dehydratase gene. Comparison of the N-terminal amino acid sequence of the previously purified PGM1 (chapter 2), with the deduced N-terminus of the PGM2 protein revealed only 35% identity indicating that two PGM isoenzymes are present in A. methanolica. Also the differences in subunit sizes of the PGM1 (Mr 28,000) and PGM2 (Mr 22,000) enzymes support this hypothesis. Cloning of the gene encoding PGM1 (e.g. via hybridization with a specific probe designed against its N-terminus) and determination of its full sequence, allowing further comparisons, remains to be done. Gene disruption studies may provide information on the physiological roles of these two PGM isoenzymes in A. methanolica. This is the first report that provides clear biochemical evidence for the presence of PGM isoenzymes in a Gram-positive bacterium.
Methanol metabolism in A. methanolica A. methanolica is able to utilize methanol as carbon and energy source and represents an interesting bacterium for studies on regulation of glucose and methanol metabolism. During growth on methanol, metabolic energy is generated by its oxidation via formaldehyde to CO2. Carbon assimilation is initiated by assimilation of formaldehyde in the RuMP pathway (Fructose bisphosphate aldolase/transketolase variant) (Anthony, 1982). Growth on methanol results in strongly increased levels of enzymes involved in both the RuMP cycle and glycolysis (PFK, FBP aldolase, transketolase) and the enzymes from the pentose phosphate pathway (Dijkhuizen and Sokolov, 1984). A screening of enzymes of glucose and methanol metabolism in A. methanolica cells has shown that some enzymes from the RuMP cycle (hexulose-6-phosphate synthase,
115
Summary and concluding remarks
hexulose-6-phosphate isomerase and transketolase) are clearly induced by C1 substrates, whereas the enzyme activities of PPi-PFK, FBPase, PGM and glucose kinase drastically decreased during growth on C1 compounds (chapter 2). Surprisingly, a second ATP-PFK enzyme activity was detected in A. methanolica cells grown on C1 substrates (chapter 5). This is the first observation that two PPi-dependent and ATP-dependent PFK isoenzymes coexist in the same bacterium. Coexistence of these two types of PFK only had been found in higher plants. Purification and characterization of this ATP-PFK from A. methanolica showed quite unusual properties, namely the inhibition of ATP-PFK activity by PPi, ADP and F-2,6-P2. This last metabolite and its physiological role in glucose metabolism has been extensively studied in yeast. This has led to the conclusion that higher intracellular concentrations of F-2,6-P2 enhance the glycolytic flux by activating ATP-PFK enzyme activity. The presence of this metabolite in Bacteria has never been reported. Further, in eukaryotic cells F-2,6-P2 only acts as an activator at micromolar concentrations, and not as an inhibitor as observed for the A. methanolica enzyme. Proof for the presence of F-2,6-P2 in A. methanolica, and subsequent experiments elucidating its possible physiological role in glucose and methanol metabolism, still have to be carried out. Screening of cells of A. methanolica grown on different substrates showed that the synthesis of ATP-PFK in A. methanolica is induced during growth on C1 compounds. This was also observed for the RuMP cycle enzymes, HPS and HPI (de Boer et al., 1990b). Growth on the mixed substrates glucose plus methanol gave only low activities of these enzymes during the first phase of glucose utilization and an increase of the activities when methanol started to be utilized. The continuous culture experiments described in chapter 5 showed clearly that addition of glucose resulted in a drastic decrease of ATP-PFK and the appearance of high levels of PPi-PFK activity, indicating that glucose or a metabolite derived from glucose utilization represses the synthesis of ATP-PFK. It is concluded that ATP-PFK and some enzymes of the RuMP cycle are subject to catabolite repression by glucose. In another facultative methylotroph, Arthrobacter P1 (Dijkhuizen et al., 1992), synthesis of HPS and HPI is solely induced by formaldehyde. The precise molecular mechanisms controlling synthesis of ATP-PFK as well as other RuMP pathway enzymes in A. methanolica remains to be elucidated. Attempts to isolate methanol-negative mutants blocked in the RuMP pathway yielded only many HPS- and /or HPI mutants (chapter 2). This might suggest the presence of isoenzymes for the steps common to both pentose phosphate pathway and the RuMP cycle, similar to the situation described for transaldolase in Arthrobacter P1 (Dijkhuizen et al., 1992). Mutant evidence for the presence of transketolase isoenzymes in A. methanolica is reported in chapter 2. In A. methanolica WVI glucose kinase activity could utilize ATP, GTP or polyphosphates as a phosphor donor, as observed in the actinomycetes Actinomyces naeslundii (Takahashi et al., 1995) and Mycobacterium tuberculosis (Hsieh et al., 1996a).
116
Summary and concluding remarks
Chapters 2 and 5 also describe experiments with a 2-deoxyglucose resistant mutant, strain G1. Strain G1 is a mutant that is unable to grow on glucose, but still can grow on fructose and methanol. Cells of this mutant grown on glycerol lacked glucose kinase and PPi-PFK activities. However, cells of this mutant grown on fructose did possess ATP-dependent fructose kinase and PPi-PFK activities. Cells of wild type A. methanolica WVI grown on fructose showed both glucose and fructose kinase activities. From these observations we have concluded that glucose kinase plays a role in glucose metabolism but not in fructose metabolism and that PPi-PFK is involved in fructose and glucose metabolism (chapters 2 and 5) The presence of two PFKs in A. methanolica is puzzling. The HPS, HPI and ATP-PFK enzyme activities are induced only by C1-compounds and repressed by glucose. The genes encoding these RuMP cycle enzymes, including the ATP-PFK enzyme, may be clustered together in an operon ensuring their coordinate regulation. The PPi-PFK is present in glucose grown cells but not in methanol grown cells. 31P-NMR studies showed that large amounts of PPi accumulated in glucose grown cells, whereas no PPi could be detected in methanol grown cells. Why such differences in PPi concentrations occur between these cells remains unclear. It has been suggested that PPi may accumulate in the cells by metabolic cycling between glycogen and G-1-P through the action of UDP-glucose pyrophosphorylase, glycogen synthase and glycogen phosphorylase (Takahashi et al., 1995). A big difference in glycogen accumulation in glucose and methanol grown cells of A. methanolica, however, was not observed. Also the levels of UDP-glucose synthase activity on these two growth substrates were not significantly different. Polyphosphate-dependent glucose kinase activity was detected in A. methanolica suggesting that PolyPn may play a role as energy source, as previously observed in other microorganisms (Dawes and Senior, 1973; Kulaev and Vagabov, 1983; van Veen, 1994). Further studies on polyphosphate and PPi metabolism in A. methanolica are required for a better understanding of the sources and roles of these compounds.
Glucose metabolism in Streptomyces coelicolor A3(2) In order to establish the major pathway(s) for glucose utilization in S. coelicolor A3(2) activities of enzymes from the glycolytic, Entner-Doudoroff and the pentose phosphate pathways were measured. These studies have indicated that glycolysis is the major pathway for glucose utilization in S. coelicolor A3(2) (Hobbs et al., 1992). Previous studies have shown that actinomycetes do not possess a complete PTS system for glucose uptake (Titgemeyer et al., 1995). In S. coelicolor A3(2), glucose represses the expression of many genes involved in the utilization of alternative carbon sources. It has been suggested that glucose kinase functions in mediating glucose repression through
117
Summary and concluding remarks
modification of a regulatory protein that interacts with promoter regions of glucose repressible genes (Angell et al., 1992). We have been able to purify and characterize the single ATP-PFK enzyme detected in glucose grown cells of S. coelicolor A3(2) (chapter 4). Biochemical characterization of this ATP-PFK showed that this enzyme is related to the allosteric ATP-PFKs from bacterial sources: it is inhibited by PEP and shows cooperativity for F-6-P. These data provide evidence that the glycolytic pathway of S. coelicolor A3(2) is also regulated at the activity level. Following the cloning of the gene encoding this ATP-PFK enzyme, analysis of the deduced amino acid sequence revealed to our initial surprise that it displays the highest similarity with the non-allosteric PPi-PFK from the actinomycete A. methanolica. Interestingly, based on current knowledge of the amino acid residues involved in the ATP binding site of the ATP-PFK of E. coli, the ATP binding site in S. coelicolor A3(2) ATP-PFK appears identical to the PPi binding site of the PPi-PFK from A. methanolica. It thus remains unclear what determines the ATP/PPi substrate specificity in these actinomycete enzymes.
Special features of glycolysis in the actinomycetes A. methanolica and S. coelicolor A3(2) The studies presented in this thesis provide evidence that glycolysis in these two actinomycetes is regulated. To our surprise a PPi-PFK enzyme activity could be detected in A. methanolica glucose grown cells. Phylogenetic analysis of this PPi-PFK enzyme showed clearly that ATP and PPi-PFKs form two distinct groups. This supports the view that PPi-PFK and ATP-PFK enzymes evolved from a common ancestor. Modelling of the PPi-PFK enzyme of A. methanolica compared with the known 3D-structure of the ATP-PFK enzyme from E. coli provided important information on the possible amino acid residues involved in PPi binding of the A. methanolica enzyme. Interestingly, information on the primary sequence of the ATP-PFK enzyme from S. coelicolor A3(2) showed that this enzyme has the same differences in the amino acid residues involved in ATP binding as the PPi-PFK from A. methanolica. It thus remains unclear what determines ATP/PPi specificity of PFK in these actinomycetes. In this thesis evidence is presented for the involvement of transketolase, PFK and PGM isoenzymes in A. methanolica. Growth of A. methanolica on C1 compounds (e.g. methanol) induced an ATP-PFK enzyme activity as well as high levels of HPS, HPI and transketolase, suggesting a common regulatory mechanism for the enzymes involved in the RuMP cycle.
118
Summary and concluding remarks
REFERENCES References are listed on pages 129-140.
119