Isolation of a dihydrofolate reductase-deficient mutant of Escherichia coli

A rugged yet easily navigable fitness landscape

Fitness landscape theory predicts that rugged landscapes with multiple peaks impair Darwinian evolution, but experimental evidence is limited. In this study, we used genome editing to map the fitness of >260,000 genotypes of the key metabolic enzyme dihydrofolate reductase in the presence of the antibiotic trimethoprim, which targets this enzyme. The resulting landscape is highly rugged and harbors 514 fitness peaks. However, its highest peaks are accessible to evolving populations via abundant fitness-increasing paths. Different peaks share large basins of attraction that render the outcome of adaptive evolution highly contingent on chance events. Our work shows that ruggedness need not be an obstacle to Darwinian evolution but can reduce its predictability. If true in general, the complexity of optimization problems on realistic landscapes may require reappraisal.

Development of antibacterial compounds that constrain evolutionary pathways to resistance

Antibiotic resistance is a worldwide challenge. A potential approach to block resistance is to simultaneously inhibit WT and known escape variants of the target bacterial protein. Here, we applied an integrated computational and experimental approach to discover compounds that inhibit both WT and trimethoprim (TMP) resistant mutants of E. coli dihydrofolate reductase (DHFR). We identified a novel compound (CD15-3) that inhibits WT DHFR and its TMP resistant variants L28R, P21L and A26T with IC₅₀ 50–75 µM against WT and TMP-resistant strains. Resistance to CD15-3 was dramatically delayed compared to TMP in in vitro evolution. Whole genome sequencing of CD15-3-resistant strains showed no mutations in the target folA locus. Rather, gene duplication of several efflux pumps gave rise to weak (about twofold increase in IC₅₀) resistance against CD15-3. Altogether, our results demonstrate the promise of strategy to develop evolution drugs - compounds which constrain evolutionary escape routes in pathogens.

In Vivo Titration of Folate Pathway Enzymes

How enzymes behave in cells is likely different from how they behave in the test tube. Previous in vitro studies find that osmolytes interact weakly with folate. Removal of the osmolyte from the solvation shell of folate is more difficult than removal of water, which weakens binding of folate to its enzyme partners. To examine if this phenomenon occurs in vivo, osmotic stress titrations were performed with Escherichia coli Two strategies were employed: resistance to an antibacterial drug and complementation of a knockout strain by the appropriate gene cloned into a plasmid that allows tight control of expression levels as well as labeling by a degradation tag. The abilities of the knockout and complemented strains to grow under osmotic stress were compared. Typically, the knockout strain could grow to high osmolalities on supplemented medium, while the complemented strain stopped growing at lower osmolalities on minimal medium. This pattern was observed for an R67 dihydrofolate reductase clone rescuing a ΔfolA strain, for a methylenetetrahydrofolate reductase clone rescuing a ΔmetF strain, and for a serine hydroxymethyltransferase clone rescuing a ΔglyA strain. Additionally, an R67 dihydrofolate reductase clone allowed E. coli DH5α to grow in the presence of trimethoprim until an osmolality of ∼0.81 is reached, while cells in a control titration lacking antibiotic could grow to 1.90 osmol.IMPORTANCEE. coli can survive in drought and flooding conditions and can tolerate large changes in osmolality. However, the cell processes that limit bacterial growth under high osmotic stress conditions are not known. In this study, the dose of four different enzymes in E. coli was decreased by using deletion strains complemented by the gene carried in a tunable plasmid. Under conditions of limiting enzyme concentration (lower than that achieved by chromosomal gene expression), cell growth can be blocked by osmotic stress conditions that are normally tolerated. These observations indicate that E. coli has evolved to deal with variations in its osmotic environment and that normal protein levels are sufficient to buffer the cell from environmental changes. Additional factors involved in the osmotic pressure response may include altered protein concentration/activity levels, weak solute interactions with ligands which can make it more difficult for proteins to bind their substrates/inhibitors/cofactors in vivo, and/or viscosity effects.

Systems-level response to point mutations in a core metabolic enzyme modulates genotype-phenotype relationship

Linking the molecular effects of mutations to fitness is central to understanding evolutionary dynamics. Here, we establish a quantitative relation between the global effect of mutations on the E. coli proteome and bacterial fitness. We created E. coli strains with specific destabilizing mutations in the chromosomal folA gene encoding dihydrofolate reductase (DHFR) and quantified the ensuing changes in the abundances of 2,000+ E. coli proteins in mutant strains using tandem mass tags with subsequent LC-MS/MS. mRNA abundances in the same E. coli strains were also quantified. The proteomic effects of mutations in DHFR are quantitatively linked to phenotype: the SDs of the distributions of logarithms of relative (to WT) protein abundances anticorrelate with bacterial growth rates. Proteomes hierarchically cluster first by media conditions, and within each condition, by the severity of the perturbation to DHFR function. These results highlight the importance of a systems-level layer in the genotype-phenotype relationship.

The former annotated human pseudogene dihydrofolate reductase-like 1 (DHFRL1) is expressed and functional

Human dihydrofolate reductase (DHFR) was previously thought to be the only enzyme capable of the reduction of dihydrofolate to tetrahydrofolate; an essential reaction necessary to ensure a continuous supply of biologically active folate. DHFR has been studied extensively from a number of perspectives because of its role in health and disease. Although the presence of a number of intronless DHFR pseudogenes has been known since the 1980s, it was assumed that none of these were expressed or functional. We show that humans do have a second dihydrofolate reductase enzyme encoded by the former pseudogene DHFRP4, located on chromosome 3. We demonstrate that the DHFRP4, or dihydrofolate reductase-like 1 (DHFRL1), gene is expressed and shares some commonalities with DHFR. Recombinant DHFRL1 can complement a DHFR-negative phenotype in bacterial and mammalian cells but has a lower specific activity than DHFR. The K(m) for NADPH is similar for both enzymes but DHFRL1 has a higher K(m) for dihydrofolate when compared to DHFR. The need for a second reductase with lowered affinity for its substrate may fulfill a specific cellular requirement. The localization of DHFRL1 to the mitochondria, as demonstrated by confocal microscopy, indicates that mitochondrial dihydrofolate reductase activity may be optimal with a lowered affinity for dihydrofolate. We also found that DHFRL1 is capable of the same translational autoregulation as DHFR by binding to its own mRNA; with each enzyme also capable of replacing the other. The identification of DHFRL1 will have implications for previous research involving DHFR.

6-pyruvoyltetrahydropterin synthase paralogs replace the folate synthesis enzyme dihydroneopterin aldolase in diverse bacteria

Dihydroneopterin aldolase (FolB) catalyzes conversion of dihydroneopterin to 6-hydroxymethyldihydropterin (HMDHP) in the classical folate biosynthesis pathway. However, folB genes are missing from the genomes of certain bacteria from the phyla Chloroflexi, Acidobacteria, Firmicutes, Planctomycetes, and Spirochaetes. Almost all of these folB-deficient genomes contain an unusual paralog of the tetrahydrobiopterin synthesis enzyme 6-pyruvoyltetrahydropterin synthase (PTPS) in which a glutamate residue replaces or accompanies the catalytic cysteine. A similar PTPS paralog from the malaria parasite Plasmodium falciparum is known to form HMDHP from dihydroneopterin triphosphate in vitro and has been proposed to provide a bypass to the FolB step in vivo. Bacterial genes encoding PTPS-like proteins with active-site glutamate, cysteine, or both residues were accordingly tested together with the P. falciparum gene for complementation of the Escherichia coli folB mutation. The P. falciparum sequence and bacterial sequences with glutamate or glutamate plus cysteine were active; those with cysteine alone were not. These results demonstrate that PTPS paralogs with an active-site glutamate (designated PTPS-III proteins) can functionally replace FolB in vivo. Recombinant bacterial PTPS-III proteins, like the P. falciparum enzyme, mediated conversion of dihydroneopterin triphosphate to HMDHP, but other PTPS proteins did not. Neither PTPS-III nor other PTPS proteins exhibited significant dihydroneopterin aldolase activity. Phylogenetic analysis indicated that PTPS-III proteins may have arisen independently in various PTPS lineages. Consistent with this possibility, merely introducing a glutamate residue into the active site of a PTPS protein conferred incipient activity in the growth complementation assay, and replacing glutamate with alanine in a PTPS-III protein abolished complementation.

A genome sequence survey of the mollicute corn stunt spiroplasma Spiroplasma kunkelii

The mollicute corn stunt spiroplasma (Spiroplasma kunkelii) is a leafhopper-transmitted pathogen of maize. Sequencing of the approximately 1.6-Mb genome of S. kunkelii was initiated to aid understanding the genetic basis of spiroplasma interactions with their plant and leafhopper hosts. In total, 144712 nucleotides of non-redundant, high-quality S. kunkelii genome sequence were obtained. Sequence tags were searched against the Mycoplasmataceae and Bacillus/Clostridium databases. Results showed that, in addition to spiroplasma phage SpV1 DNA insertions, spiroplasma genomes harbor more purine and amino acid biosynthesis, transcription regulation, cell envelope and DNA transport/binding genes than Mycoplasmataceae genomes. This investigation demonstrates that survey sequencing is an efficient procedure for gene discovery and genome characterization. The results of the S. kunkelii sequencing project are available at the Spiroplasma WebPage at http://www.oardc.ohio-state.edu/spiroplasma/genome.htm.

Growth properties of afolAnull mutant ofEscherichia coliK12

In Escherichia coli, dihydrofolate reductase is required for both the de novo synthesis of tetrahydrofolate and the recycling of dihydrofolate produced during the synthesis of thymidylate. The coding region of the dihydrofolate reductase gene, folA, was replaced with a kanamycin resistance determinant. Unlike earlier deletions, this mutation did not disrupt flanking genes. When the mutation was transferred into a wild-type strain and a thymidine- (thy) requiring strain, the resulting strains were viable but slow growing on rich medium. Both synthesized less folate than their parents, as judged by the incorporation of radioactive para-aminobenzoic acid. The derivative of the wild-type strain did not grow on any defined minimal media tested. In contrast, the derivative of the thy-requiring strain grew slowly on minimal medium with thy but exhibited auxotrophies on some combinations of supplements. These results suggest that when folates are limited, they can be distributed appropriately to folate-dependent biosynthetic reactions only under some conditions. Key words: dihydrofolate reductase, Escherichia coli, biosynthesis, folates, one-carbon metabolism.

A biochemical and genetic model for parasite resistance to antifolates. Toxoplasma gondii provides insights into pyrimethamine and cycloguanil resistance in Plasmodium falciparum

We have exploited the experimental accessibility of the protozoan parasite Toxoplasma gondii and its similarity to Plasmodium falciparum to investigate the influence of specific dihydrofolate reductase polymorphisms known from field isolates of drug-resistant malaria. By engineering appropriate recombinant shuttle vectors, it is feasible to examine mutations by transient or stable transformation of T. gondii parasites, in bacterial and yeast complementation assays, and through biochemical analysis of purified enzyme. A series of mutant alleles that mirror P. falciparum variants reveals that the key mutation Asn-108 (Asn-83 in T. gondii) probably confers resistance to pyrimethamine by affecting critical interactions in the ternary complex. Mutations such as Arg-59 (T. gondii 36) have limited effect in isolation, but in combination with other mutations they enhance the competitive ability of folate by increasing the speed of product turnover. Val-16 (T. gondii 10) confers low level resistance to cycloguanil but hypersensitivity to pyrimethamine. This mutation precludes Asn-108, probably because compression of the folate binding pocket introduced by this combination is incompatible with enzyme function. These studies permit detailed biochemical, kinetic, and structural analysis of drug resistance mutations and reconstruction of the probable phylogeny of antifolate resistance in malaria.

Identification and cloning of the Mycobacterium avium folA gene, required for dihydrofolate reductase activity

Dihydrofolate reductase is an essential bacterial enzyme necessary for the maintenance of intracellular folate pools in a biochemically active reduced state. In this report, the Mycobacterium avium folA gene was identified by functional genetic complementation, sequenced, and expressed for the first time. It has an open reading frame of 543 bp with a G + C content of 73%. The translated polypeptide sequence shows 58% identity to the consensus sequence of the conserved regions from eight other bacterial dihydrofolate reductases. Recombinant M. avium dihydrofolate reductase was expressed actively in Escherichia coli, and SDS-PAGE analysis revealed a 20 kDa species, agreeable with that predicted from the polypeptide sequence:

Dihydropteridine reductase from Escherichia coli exhibits dihydrofolate reductase activity

E. coli Dihydropteridine reductase, known to have a pterin-independent oxidoreductase activity with potassium ferricyanide as electron donor, has now been shown to possess also dihydrofolate reductase activity. The kinetic parameters for dihydrofolate reductase activity have been determined. The ratio of the three activities, dihydropteridine reductase, dihydrofolate reductase and pterin-independent oxidoreductase activity is 1.0, 0.05 and 4.3, respectively. The enzyme, a flavoprotein which is unstable in the presence of dithiothreitol, was shown to be a monomer with a molecular mass of 25.7 kDa. The apparent lack of discrimination between hydride transfer from the pyridine nucleotide to N5 of the pterin in the dihydropteridine reductase reaction and C6 of folate in the dihydrofolate reaction suggested that the FAD prosthetic group may be involved in the hydride transfers. The flavoprotein inhibitor N,N- dimethylpropargylamine inhibited the dihydropteridine reductase and oxidoreductase reactions differently and did not affect the dihydrofolate reductase activity however.

Purification and properties of recombinant Pneumocystis carinii dihydrofolate reductase

Pneumocystis carinii dihydrofolate reductase (DHFR) expressed in Escherichia coli was purified to homogeneity in a single step using methotrexate-Sepharose affinity chromatography. The purified enzyme migrated as a single 24-kDa protein on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The sequence of the first 26 amino acids from the N-terminus of the purified enzyme was in accord with that predicted from the DNA sequence. The enzyme showed a broad pH optimum with maximum activity over the pH range 6 to 7. The enzyme was activated by salts, with maximal twofold activation at 50 to 150 mM KCl and 50 to 200 mM NaCl. Urea at 2.5 M also increased the enzyme activity twofold. Kinetic analysis of the purified enzyme revealed that the Km values for dihydrofolate and NADPH were 1.8 and 1.4 microM, respectively, and that the kcat was 70 s-1. Inhibition studies verified that trimethoprim and pyrimethamine were poor inhibitors of P. carinii DHFR and showed little selectivity over the human DHFR. Trimetrexate and piritrexim were much more potent inhibitors of the P. carinii enzyme, but these inhibitors are also potent inhibitors of human DHFR.

Stability of mutant type II dihydrofolate reductase proteins in suppressor strains

In the course of study of a (Glu-58 to Gln-58) mutant type II dihydrofolate reductase (DHFR), it was found that the altered DHFR was poorly produced in vivo. Investigations with several common laboratory Escherichia coli strains including htpR and lon strains bearing plasmids expressing the Gln-58 DHFR indicated a correlation of rapid degradation with the presence of a sup+ phenotype. The supo strain MC1061(p3) was transformed with a series of plasmids containing the Gln-58 DHFR gene with and without an additional supF gene, and expression levels were compared. The supF+ constructs exhibited little accumulation of the Gln-58 DHFR, while reasonable levels were found in the supo cases. Experiments with extracts of plasmid-free sup+ and supo strains showed rapid degradation by certain strains compared to MC1061(p3) and this degradation was not dependent upon ATP. In another route to increasing the stability of labile DHFR derivatives, mutagenesis of a strain bearing a N-terminally shortened Gln-58 DHFR was performed. Selection and analysis of a trimethoprim-resistant stable mutant showed that this DHFR gene contained a triple repeat of leu-pro-ser in the enzymatically non-essential N-terminal portion of the protein.

Isolation and nucleotide sequence of the hmp gene that encodes a haemoglobin-like protein in Escherichia coli K-12

In the course of an attempt to identify genes that encode Escherichia coli dihydropteridine reductase (DHPR) activities, a chromosomal DNA fragment that directs synthesis of two soluble polypeptides of Mr 44000 and 46000 was isolated. These proteins were partially purified and were identified by determination of their N-terminal amino acid sequences. The larger was serine hydroxymethyltransferase, encoded by the glyA gene, while the smaller was the previously described product of an unnamed gene closely linked to glyA, and transcribed in the opposite direction. Soluble extracts of E. coli cells that overproduced the 44 kDa protein had elevated DHPR activity, and were yellow in colour. Their visible absorption spectra were indicative of a CO-binding b-type haemoprotein that is high-spin in the reduced state. The sequence of the N-terminal 139 residues of the protein, deduced from the complete nucleotide sequence of the gene, had extensive homology to almost all of Vitreoscilla haemoglobin. We conclude that E. coli produces a soluble haemoglobin-like protein, the product of the hmp gene (for haemoprotein). Although the protein has DHPR activity, it is distinct from the previously purified E. coli DHPR.

Heterologous expression of active thymidylate synthase-dihydrofolate reductase from Plasmodium falciparum

The coding sequence of the bifunctional thymidylate synthase-dihydrofolate reductase (TS-DHFR) from a moderately pyrimethamine-resistant strain (HB3) of Plasmodium falciparum was assembled in a pUC expression vector. The coding sequence possesses unique Nco1 and Xba1 sites which flank 243 bp of the DHFR gene that include all point mutations thus far linked to pyrimethamine resistance. Wild-type (3D7) and highly pyrimethamine-resistant (7G8) TS-DHFRs were made from this vector by cassette mutagenesis using Nco1-Xba1 fragments from the corresponding cloned TS-DHFR genes. Catalytically active recombinant TS-DHFRs were expressed in Escherichia coli, albeit at low levels. Both TS and DHFR coeluted upon gel filtration and copurified upon affinity and anion exchange chromatography. Gel filtration and SDS-PAGE indicated that the enzyme was a dimer with identical 67-kDa subunits, characteristic of protozoan TS-DHFRs. Amino-terminal sequencing gave 10 amino acids which perfectly matched the sequence predicted from the nucleotide sequence. The recombinant TS-DHFR was purified to homogeneity by 10-formylfolate affinity chromatography followed by Mono Q FPLC. The inhibition properties of pyrimethamine toward the purified recombinant enzymes show that the point mutations are the molecular basis of pyrimethamine resistance in P. falciparum.

Isolation and expression of the Pneumocystis carinii dihydrofolate reductase gene

Pneumocystis carinii dihydrofolate reductase (DHFR; 5,6,7,8-tetrahydrofolate: NADP+ oxidoreductase, EC 1.5.1.3) cDNA sequences have been isolated by their ability to confer trimethoprim resistance to Escherichia coli. Consistent with the recent conclusion that P. carinii is a member of the Fungi, sequence analysis and chromosomal localization show that DHFR is neither physically nor genetically linked to thymidylate synthase. Expression of recombinant P. carinii DHFR in heterologous hosts provides an abundant source of the enzyme that may form a basis for the development of new therapies for this enigmatic pathogen. Studies with the recombinant enzyme show that trimethoprim is a very poor inhibitor of P. carinii DHFR and, in fact, is a more potent inhibitor of human DHFR.

Dihydropteridine reductase from Escherichia coli

A dihydropteridine reductase from Escherichia coli was purified to apparent homogeneity. It is a dimeric enzyme with identical subunits (Mr 27000) and a free N-terminal group. It can use NADH (Vmax./Km 3.36 s-1) and NADPH (Vmax./Km 1.07 s-1) when 6-methyldihydro-(6H)-pterin is the second substrate, as well as quinonoid dihydro-(6H)-biopterin (Vmax./Km 0.69 s-1), dihydro-(6H)-neopterin (Vmax./Km 0.58 s-1), dihydro-(6H)-monapterin 0.66 s-1), 6-methyldihydro-(6H)-pterin and cis-6,7-dimethyldihydro-(6H)-pterin (Vmax./Km 0.66 s-1) when NADH is the second substrate. The pure reductase has a yellow colour and contains bound FAD. The enzyme also has pterin-independent NADH and NADPH oxidoreductase activities when potassium ferricyanide is the electron acceptor.

Construction of a fol mutant strain of Escherichia coli for use in dihydrofolate reductase mutagenesis experiments

A strain of Escherichia coli with the fol gene deleted and a kan gene inserted in its place was created for use in cloning and isolation of mutant dihydrofolate reductase. Southern blot analysis and dihydrofolate reductase enzyme assays confirmed the delta fol::kan genotype. A thyA mutation accompanied the fol deletion and is required for survival of a dihydrofolate reductase-deficient strain.

Construction of a dihydrofolate reductase-deficient mutant of Escherichia coli by gene replacement

The dihydrofolate reductase (fol) gene in Escherichia coli has been deleted and replaced by a selectable marker. Verification of the delta fol::kan strain has been accomplished using genetic and biochemical criteria, including Southern analysis of the chromosomal DNA. The delta fol::kan mutation is stable in E. coli K549 [thyA polA12 (Ts)] and can be successfully transduced to other E. coli strains providing they have mutations in their thymidylate synthetase (thyA) genes. A preliminary investigation of the relationship between fol and thyA gene expression suggests that a Fol- cell (i.e., a dihydrofolate reductase deficiency phenotype) is not viable unless thymidylate synthetase activity is concurrently eliminated. This observation indicates that either the nonproductive accumulation of dihydrofolate or the depletion of tetrahydrofolate cofactor pools is lethal in a Fol- ThyA+ strain. Strains containing the thyA delta fol::kan lesions require the presence of Fol end products for growth, and these lesions typically increase the doubling time of the strain by a factor of 2.5 in rich medium.

Mapping and sequencing of the dihydrofolate reductase gene (DFR1) of Saccharomyces cerevisiae

The dihydrofolate reductase gene (DFR1) from Saccharomyces cerevisiae has been mapped and sequenced. The gene was isolated on an 8.8-kb BamHI fragment from a yeast genomic library by screening of Escherichia coli transformants for resistance to trimethoprim. A 1.8-kb SalI-BamHI fragment which was able to confer methotrexate resistance in yeast also complemented an E. coli DHFR-deficient (folA) mutant. Nucleotide sequence analysis revealed that the yeast DFR1 gene encoded a polypeptide with a predicted Mr of 24230. The deduced sequence of 211 amino acid residues showed considerable homology with DHFRs from both bacterial and animal sources. The codon bias index of the DFR1 coding region is 0.0083, which indicates a random pattern of codon usage. The upstream region contains two consensus sequences required for binding of the yeast's positive regulatory factor, GCN4, suggesting that the DFR1 gene might be subject to the amino acid general control. Several potential 'TATA' boxes are located in the sequence 5' to the gene. Located in the 3' flanking region are homologies with several canonical sequences thought to be required for efficient transcription termination in yeast. We also mapped the DFR1 gene to a position 1.4 cM proximal to the MET7 locus on chromosome XV.