Biotechnology Letters 23: 917–921, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.
917
Over-production of lactate dehydrogenase from Plasmodium falciparum opens a route to new antimalarials Dilek Turgut-Balik1,∗ , Debbie K. Shoemark2 , Kathleen M. Moreton2, Richard B. Sessions2 & J. John Holbrook2 1 University
of Fırat, Faculty of Science and Arts, Department of Biology, Elazıg, Turkey
2 Molecular Recognition Centre and Department of Biochemistry, University of Bristol School of Medical Sciences,
Bristol BS8 1TD, UK ∗ Author for correspondence (Fax: +90 0424 2330062; E-mail:
[email protected] edu.tr) Received 2 February 2001; Revisions requested 12 February 2001; Revisions received 3 April 2001; Accepted 4 April 2001
Key words: APAD, lactate dehydrogenase, malaria, Plasmodium falciparum, Shine–Dalgarno
Abstract Over-production of lactate dehydrogenase (Pf LDH) from Plasmodium falciparum from E. coli TG2 cells transformed with a pKK223-3 plasmid containing the wild type gene isolated by Bzik DJ, Fox BA, and Gonyer K (1993) Mol. Biochem. Parasit. 59, 155–166, gave mostly an inactive protein after isolation. Sequencing the N-terminus of the over-produced protein showed that the major product commenced at an internal methionine. Truncation of the protein occurred due to the inappropriate priming from a Shine–Dalgarno (SD) sequence upstream of Met 35. Silent mutations of this SD sequence to remove the purine-rich region allowed over-production of the full length Pf LDH up to 15 mg protein l−1 broth. The purified protein exhibited biochemical properties of an authentic LDH enzyme. However, high activity with 3-acetylpyridine adenine dinucleotide as well as with the natural cofactor, NAD, was also observed. The high-resolution X-ray structure obtained from the recombinant enzyme has provided the opportunity for the development of inhibitors specific to Pf LDH. Abbreviations: LDH, L-lactate dehydrogenase; Pf LDH, lactate dehydrogenase from Plasmodium falciparum; APAD, 3-acetylpyridine adenine dinucleotide. Introduction The growing resistance of Plasmodium falciparum to established anti-malarials highlights the need for alternative drug regimens. Lactate dehydrogenase (LDH) catalyses the final step in glycolysis vital to P. falciparum during the anaerobic erythrocytic stages of its life cycle. Pf LDH is a good target for drug design as its inhibition results in parasite death within cultured red blood cells (Royer et al. 1986, Dr S. Croft, London School of Tropical Medicine and Hygiene, pers. comm.) and because the enzyme differs significantly from the human LDH isozymes. The major differences are key residue changes at the active site, Leu 163 shown to confer activity with 3-acetylpyridine adenine dinucleotide (APAD), and a five-residue inser-
tion in the substrate-specificity loop which is believed to be responsible for the selective action of gossypol derivatives (Sessions et al. 1997). Over-production of the target protein is a key step in drug design studies, providing material for the crystallographic and kinetic studies at the heart of this approach. The cDNA encoding an LDH-like protein was isolated from the P. falciparum strain Honduras1 by Bzik et al. (1993) who transformed E. coli cells with this gene inserted into a PK223-3 plasmid. The authors reported Pf LDH over-production but, in our hands, this type of construct gave a truncated, inactive protein as the major over-expression product. In this present study, a novel method was used to overproduce the complete LDH from P. falciparum strain K1, leading to the high-resolution X-ray structure of
918 the Pf LDH protein (Dunn et al. 1996). This, in turn, has led to a Medicines for Malaria Venture program to identify novel antimalarial drugs via a collaboration between Bristol University, Glaxo Wellcome Smith Kline Beecham and the London School of Tropical Medicine and Hygiene.
Materials and methods Bacterial strains and growth media The host bacterial strain used to prepare DNA for mutagenesis and sequencing in pUC-18 (Pharmacia Biotech, Uppsala, Sweden) was Escherichia coli TG2. The wild type Pf LDH protein and all of the mutant Pf LDH proteins were sub-cloned into pKK233-3 (Pharmacia Biotech), and expressed in the same host E. coli strain. E. coli TG2 was cultured in double strength yeast tryptone (2xYT) broth. Where appropriate ampicillin (100 µg ml−1 ) was used in media for the selection and growth of transformants. The competent cells for transformation were grown on minimal medium agar according to Sambrook et al. (1989). Cloning of WT PfLDH from strains K1 and PF-FCBR Two oligonucleotide primers, complementary to the forward-reverse strands of the P. falciparum LDH gene, were made: Pf1: 5 -EcoRI site ATGGCTCCAAAAGCAAAAATCG3
Pf2: 5 -PstI site GAGAATGAAGGCATTAGCTTAA-3. The amplification used the DNA polymerase from Thermus aquaticus (Taq). The reaction mixture consisted of: 5 µl Taq buffer (supplied with enzyme), 5 µl stock dNTPs (10 µl of each 100 mM dNTPs and 10 µl H2 O), 2.5 µl each forward and reverse primers (at 20 pmol), 1 µl (0.5 µg µl−1 ) genomic DNA, 2.5 units Taq DNA polymerase and 33.5 µl H2 O in a final volume of 50 µl. PCR was carried out at 94 ◦ C for 1.5 min, 55 ◦ C for 2 min, and 72 ◦ C for 2 min for 20 cycles. The resulting 1 kb product was purified from a 1% agarose gel, digested using EcoRI and PstI, and ligated into similarly digested pUC-18. The DNA was sequenced from both directions using a Sequenase Version 2.0 kit from US Biochemicals, Cleveland, OH. The DNA was then digested using EcoRI and PstI, ligated into similarly cut pKK223-3, and transformed into CaCl2 -competent E. coli cells for expression.
Mutagenesis to construct M35L and PfLDH(SDK1) All the mutants were made by site direct mutagenesis using the polymerase chain reaction on Pf LDH strain K1 plasmid DNA. All PCRs were carried out in the presence of 5 µl Pfu buffer, 5 µl stock dNTPs, 2.5 µl each 20 pmol µl−1 oligonucleotides, 2.5 units Pfu, the DNA polymerase from Pyrococcus furiosis. All DNA modifying enzymes were from Boehringer Mannheim or New England Biolabs, and were used according to the supplier’s instructions. The mutagenesis had two stages. Firstly, mutant oligonucleotides, together with Pf LDH forward and reverse primers Pf1 and Pf2, were used to generate two overlapping Pf LDH gene fragments. The PCR was carried out at temperatures 94 ◦ C for 1.5 min, 53 ◦ C for 2 min, 72 ◦ C for 2 min for 20 cycles. Secondly, the two overlapping fragments were joined by overlap extension after purification from a 1% agarose gel. The reaction consisted of 7 cycles of PCR with the following parameters: 2 min at 94 ◦ C, 1 min at 50 ◦ C, 2 min at 72 ◦ C prior to amplification for 20 cycles, after addition of the end primers, 1.5 min at 94 ◦ C, 1 min at 55 ◦ C, 2.5 min at 72 ◦ C. The following oligonucleotides for mutagenesis were synthesized using the Millipore Expedite Nucleic Acid Synthesis System. 5 -GGA GGA GTA TTA GCT ACC M35L: TTA ATT-3 , 5 -C AAT TAA GGT AGC TAA TAC TCC TCC-3 ;
Pf LDH(SDK1): 5 GGT ATG ATT GGT GGT GTA ATG GCT ACC-3 , 5 GGT AGC CAT TAC ACC ACC AAT CAT ACC-3 .
Mutant double-stranded Pf LDH genes were purified from 1% agarose gel. They were digested using EcoRI and PstI, and were ligated into similarly digested pUC-18. The plasmids were sequenced in the region of mutation using the Sequenase version 2.0 DNA sequencing kit. The mutant genes were digested using EcoRI and PstI, ligated into similarly cut pKK223-3, and transformed into CaCl2 -competent E. coli cells for expression. N-Terminal sequencing Western blotting was performed as described by Sambrook et al. (1989). The membrane was air dried at the end of procedure, bands were cut out of the membrane
919 Table 1. Comparison of the relative rates of lactate oxidation by Pf LDH(SDK1) and human M4 -LDH using the coenzymes 3-acetylpyridine adenine dinucleotide (APAD) and NAD [A548 min−1 (APAD)/A548 min−1 (NAD)]. Ratio of rates: APAD/NAD
Lactate (M) Coenzyme (mM)
1 5 10
Pf LDH (SDK1) 0.06 0.3 0.6
Human M4 LDH 0.06 0.3 0.6
4.86 6.05 3.91
0.007 0.021 0.023
2.93 4.72 2.38
and submitted for N-terminal automated sequencing on an Applied Biosystem Model 477A sequencer. Purification of PfLDH protein The E. coli TG2 cells containing the SDK1 mutant Pf LDH gene, encoding full length wild type Pf LDH protein, were grown overnight, at 37 ◦ C, in 2 × YT broth (containing ampicillin). The cells were harvested by centrifugation at 4000 × g for 30 min. The cells were resuspended in minimum volume of 50 mM triethanolamine, pH 6. The cells were sonicated, and the debris was removed by centrifugation. Protein extraction and purification involving an oxamate affinity column was performed according to Barstow et al. (1986). APAD utilization This assay is based on the observation that the lactate dehydrogenase purified from Plasmodium falciparum has the ability to use rapidly 3-acetyl pyridine dinucleotide (APAD) as a coenzyme in the reaction leading to the formation of pyruvate from high concentrations of lactate (Makler & Hinrichs 1993). The reaction rates were determined for both the recombinant Pf LDH from the SDK1 construct and human-M4 LDH in the lactate to pyruvate direction using both APAD and NAD as coenzyme. The spectrophotometric assessment of LDH activity was facilitated by adding Nitro Blue Tetrazolium (NBT), and phenazine ethosulfate (PES) resulting in the formation of a blue formazan precipitate measurable at 548 nm (Dr A. Cortes, pers. comm.). The buffer used was 20 mM Bis-Tris/HCl (pH 8) containing 50 mM KCl. The three lactate concentrations used were 0.05, 0.2 and 0.5 M at three coenzyme concentrations, 1, 5 and 10 mM.
2.21 3.02 1.53
0.008 0.02 0.025
0.013 0.023 0.026
Results and discussion Expression of PfLDH in E. coli from DNA of P. falciparum strains K1 and PF FCBR The LDH genes from P. falciparum strains K1 and PF FCBR were cloned and fully sequenced revealing no variation between the sequences of these two strains. They are also identical to the LDH gene of the Honduras I strain (Bzik et al. 1993). Of course, there is no guarantee that all strains are identical (Turgut-Balik & Holbrook 2001, Turk. J. Biol., in press). Since the sequence data were the same for both strains, K1 and PF FCBR, of P. falciparum further experimental work was carried out on the strain K1 alone. Transformed E. coli containing the wild type Pf LDH K1 gene was assayed for both expression and LDH activity. The cells were grown to O.D600 0.6 prior to induction with 1 mM IPTG and then samples were taken after 1 h, 2 h, 3 h and 16 h. There was a prominent protein band of approximately correct molecular weight in the induced cell lines when the sample was run on 10% SDS-PAGE. A gel was run of the soluble fraction after sonication to check whether the protein was water soluble. The bulk of the over-produced protein was not soluble and was located in the pellet fraction following sonication and centrifugation of the cells. Weak LDH activity was detected in the supernatant, but this activity did not rise on induction, unlike the level of protein over-production. Inspection of the Coommassie Bluestained gel of the insoluble, over-produced protein showed two intense bands corresponding to proteins around 30 kDa with intensities around 5% and 95% of the total. The N-terminal sequences of the first 10 residues were obtained for both bands. The sequence of the minor band corresponded to the gene-derived sequence, while that of the major band showed that
920
Fig. 1. Amino-terminus amino acid sequences of the two protein bands of approximate molecular weight 30 kDa from Pf LDH K1. The gene-derived N-terminal amino acid sequence of P. falciparum strain K1 LDH is 15 residues shorter than that of dogfish LDH (White et al. 1976) whose sequence is used as a reference in LDH numbering. For example, the first methionine of Pf LDH is named Met 15 to correspond to the homologous residue in dogfish LDH.
Fig. 2. Construction of the Pf LDH(SDK1) mutant. The Shine–Dalgarno sequence is shown in bold, the third methionine is shown underlined.
Fig. 3. Detail of the interactions between the nicotinamide group of NADH and the protein in the crystal structures of the ternary complexes of (a) human muscle LDH and (b) Pf LDH. Atom colouring is: C (medium grey); N (light grey); O (dark grey). Hydrogen bonds involving the nicotinamide amide group are shown as dotted lines.
the sequence started at residue Met 35. Hence the first 19 amino acids were missing from the N-terminus of 95% of the Pf LDH produced from the K1 construct (Figure 1). It was concluded that the RNA was mostly being translated from the third methionine, hence most of the protein was in a truncated, insoluble and inactive form. This inactivity is expected since residues 15–35 contain the first strand of the β-sheet in the Rossmann-fold. Mutagenesis to obtain full-length active protein The initial strategy to obtain full-length protein was to remove the third methionine by a conservative mutation. Hence Met 35 was replaced with a leucine residue by site directed mutagenesis (changing the
ATG codon of methionine to TTA). The replacement of Met 35 stopped the production of the insoluble truncated protein, but did not improve the expression of the full-length protein. Inspection of the DNA sequence surrounding the third methionine codon in the Pf LDH gene revealed an upstream Shine–Dalgarno sequence (GGAGGA). Such a sequence has been shown to be vital to chain initiation and termination of translation in prokaryotes (Shine & Dalgarno 1974) and has been introduced into plasmids in prokaryotic expression systems to ensure correct initiation (see, for example, Kim & Pack 1993). In an eukaryotic system, the presence of a Shine–Dalgarno sequence upstream of an initiation codon would have no consequence for the translation
921 of the gene, but explains the N-terminal truncation in the prokaryote E. coli. The Shine–Dalgarno sequence was removed by making silent mutations in the gene, changing the GGA codon of two glycines to GGT (Figure 2). This resulted in the over-production of a soluble, full length, active P. falciparum LDH with wild-type protein sequence. This mutant gene construct is referred to as Pf LDH(SDK1).
compounds to combat the increasing problem of drug resistance encountered with conventional antimalarials. Crystallography of Pf LDH/ligand complexes will aid the development of these to effective antimalarial drugs, a program currently supported by the World Health Organisation.
Acknowledgements Use of the non-natural cofactor APAD by PfLDH One of the biochemical characteristics that distinguishes Pf LDH from human LDH is the ability of Pf LDH to use APAD as coenzyme in a reaction leading to the formation of pyruvate from lactate in vitro (Makler & Hinrichs 1993). While human red-bloodcell LDH can also use APAD instead of NAD, it does so at a much lower rate. This behaviour was examined using the recombinant proteins, rec-Pf LDH from the SDK1 construct and rec-human-M4 LDH (rec-hLDH). Table 1 shows the ratio of rates observed when recPf LDH and rec-hLDH catalyze the reaction in the lactate to pyruvate direction using APAD or NAD as a cofactor. The results indicate that Pf LDH is about 600 times more active with the synthetic coenzyme APAD than with NAD, while hLDH is 50–100 times more active with its natural coenzyme NAD than with APAD. The basis for this alteration in activity lies in the presence of Leu 163 in Pf LDH. This residue position is occupied by serine in all other known LDHs, and the side chain hydroxyl of S163 is involved in a hydrogen bonding network which includes the nicotinamide amide group in all known LDH/NADH crystal structures. APAD differs from NAD simply by having the NH2 of the nicotinamide amide group replaced by a methyl group. APAD cannot participate in this hydrogen-bonding network in hLDH hence the binding is weakened with respect to NAD. Recombinant Pf LDH from the SDK1 construct was crystallised as the ternary complex containing NADH and the substrate analogue oxamate, and the X-ray structure of the complex solved (Dunn et al. 1996). This structure showed that the methyl group of APAD could pack against this hydrophobic, bulky leucine residue (Figure 3). A backbone twist at this position also allows some backbone hydrogen-bonding interaction with the nicotinamide amide group of NAD. Overproduction of large quantities recombinant Pf LDH has paved the way for a screening program in collaboration to identify novel antimalarial lead
We would like to thank to Dr J. Hyde, UMIST, UK, for providing us Plasmodium falciparum strain K1 genomic DNA, and to Dr K. Lingelbach, Hamburg, Germany, for the PF FCBR genomic DNA.
References Barstow DA, Clarke AR, Chia WN, Wigley DB, Sharman AF, Holbrook JJ, Atkinson T, Minton P (1986) Cloning, expression and complete nucleotide sequence of the Bacillus stearothermophilus L-lactate dehydrogenase gene. Gene 46: 47–55. Bzik DJ, Fox BA, Gonyer K (1993) Expression of Plasmodium falciparum lactate dehydrogenase in Escherichia coli. Mol. Biochem. Parasit. 59: 155–166. Dunn CR, Banfield MJ, Barker JJ, Higham CW, Moreton, KM, Turgut-Balik D, Brady RL, Holbrook JJ (1996) The structure of lactate dehydrogenase from Plasmodium falciparum reveals a new target for antimalarial design. Nature Struct. Biol. 3: 912–915. Eszes CM, Sessions RB, Clarke AR, Moreton KM, Holbrook JJ (1996) Removal of substrate inhibition in a lactate dehydrogenase from human muscle by a single residue change. FEBS Lett. 399: 193–197. Kim JH, Pack MY (1993) Over-production of extracellular endoglucanase by genetically engineered Bacillus subtilis. Biotechnol. Lett. 15: 133–138. Makler MT, Hinrichs DJ (1993) Measurement of the lactate dehydrogenase activity of Plasmodium falciparum as an assessment of parasitemia. Am. J. Trop. Med. Hyg. 48: 205–210. Royer RE, Deck LM, Campos NM, Hunsaker LA, Vander Jagt DL (1986) Biologically active derivatives of gossypol. Synthesis and antimalarial activities of peri-acylated gossylic nitriles. J. Med. Chem. 29: 1799–1801. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Sessions RB, Dewar V, Clarke AR, Holbrook JJ (1997) A model of the Plasmodium falciparum lactate dehydrogenase and its implications for the design of new antimalarials. Protein Eng. 10: 301–306. Shine J, Dalgarno L (1974) The 3 -terminal sequence of Escherichia coli 16S ribosomal RNA: complementarity to nonsense triplets and ribosome binding sites. Proc. Natl. Acad. Sci. USA 71: 1342– 1346. White JL, Hackert ML, Buehner M, Adams MJ, Ford GC, Lentz PJ, Smiley IE, Steiner SJ, Rossman MG (1976) A comparison of the structures of apo dogfish M4 lactate dehydrogenase and its ternary complexes. J. Mol. Biol. 102: 759–779.
