Introduction

The ergosterol biosynthesis pathway has been proposed as a target for identifying and designing new antifungals. In a rate-limiting step of the biosynthetic pathway of ergosterol in fungi and cholesterol in humans, the enzyme 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGR) catalyzes the conversion of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) to mevalonate1. This pathway also generates other bioactive molecules, including a coenzyme (Q10) that participates in the transport of electrons in mitochondria2.

$$ {\text{(S)-HMG-CoA}}\, + \,{\text{2NADPH}}\, + \,{\text{2H}}^{ + } \to {\text{(R)-mevalonate}}\, + \,{\text{2NADP}}^{ + } \, + \,{\text{CoASH}} $$

Phylogenetic analyses have revealed two types of HMGR: class I of eukaryotic organisms and some archaea and class II of prokaryotes and some archaea. Whereas the first class has three domains (catalytic, linker, and membrane-anchor), most members of the second one contain only the catalytic domain1.

The HMGR enzyme is the target of lipid-lowering drugs such as statins3. These drugs have been described as competitive inhibitors because they dispute the active site of HMGR with the substrate HMG-CoA1. Since statins can induce myopathy and hepatotoxicity4, new bioactive principles derived from natural sources (e.g., alpha-asarone) have been synthesized. In a murine model, alpha-asarone decreased cholesterol and triglyceride levels. Moreover, its inhibitory effect was tested on the recombinant human HMGR (hHMGR) enzyme and an enriched enzyme extract from Schizosaccharomyces pombe5,6.

The catalytic domain of HMGR of eukaryotic organisms is highly conserved and harbors the dimerization site, the substrate (HMG-CoA) binding site, and the cofactor (NADPH) binding site1. Hence, our group has proposed fungal HMGR enzymes as a model for evaluating compounds with antifungal and hypocholesterolemic activity7. One of the fungi used for this purpose has been Candida glabrata because of its high incidence as an opportunistic pathogenic yeast in hospitalized patients with compromised immunity and its resistance to multiple conventional drugs, including azoles and echinocandins. Such multi-drug resistance has led to treatment failures8,9.

In previous research, the catalytic domain of the HMGR enzyme from C. glabrata (CgHMGR) was cloned, expressed and characterized, demonstrating its enzymatic activity as well as its inhibition by simvastatin and compounds derived from alpha asarone. Consequently, it has been proposed as a possible antifungal target10,11.

The recombinant CgHMGR enzyme is available and its catalytic activity is still under study. Therefore, the aim of the current contribution was to generate, characterize, and overexpress five point mutants of CgHMGR in order to explore the role of certain amino acid residues in the functional capacity of the catalytic domain of CgHMGR. Three highly conserved regions within the catalytic domain were mutated: the dimerization site, the substrate binding site, and the cofactor binding site12. The investigation of point mutations in HMGR has been limited, to the best of our knowledge, to humans, Pseudomonas mevalonii and Syrian hamsters. They have not been performed on the HMGR of fungi, much less in an opportunistic pathogenic yeast12,13,14,15,16.

Results

Conservation of the sites of interest in the catalytic domain of HMGRs from different organisms

A multiple alignment was carried out on the Clustal Omega program for the amino acid sequence of the soluble fraction of HMGR from plants, bacteria, insects, mammals (including Homo sapiens), and opportunistic pathogenic yeasts (e.g., C. glabrata) (Supplementary Fig. S1). The access codes of the corresponding sequences can be consulted in the Materials and Methods section. The motif sequences in the sites of interest in the HMGR enzymes are highlighted in Fig. 1: the dimerization site (ENVIG), the substrate binding site (EGCLVAS), and the cofactor binding site (DAMGMN). The amino acid sequences encoding the soluble fraction of the HMGR enzymes of different organisms are highly conserved, which emphasizes the importance of examining some of the amino acid residues in these sites in order to evaluate their role in the catalysis of the enzyme. Such an analysis has been performed in studies on P. mevalonii.

Figure 1
figure 1

Sequence logos of the conserved motifs found in the catalytic domain of the HMGRs of different organisms. Multiple alignments were performed with Clustal Omega and the consensus logos were generated with WebLogo. Group 1 (Candida auris, Candida haemulonii, Clavispora lusitaniae ATCCC 42720, Debaryomyces hansenii CBS767, Meyerozyma guilliermondii ATCC 6260, Candida parapsilosis, Candida orthopsilosis Co 90-125, Candida tropicalis MYA-3404, Candida albicans SC5314, and Candida dubliniensis CD36); group 2 (C. glabrata CBS138, Candida kefyr, Kluyveromyces lactis NRRL Y-1140, Saccharomyces cerevisiae S288c HMG1, and S. cerevisiae S288c HMG2); group 3 (Yarrowia lipolytica CLIB122, S. pombe, and Ustilago maydis); group 4 (Drosophila melanogaster, Mus musculus, H. sapiens, and Rattus norvegicus); and group 5 (P. mevalonii).

Full size image

A phylogenetic tree was constructed by examining the amino acid sequences detected in the catalytic domain of the twenty-four organisms included in the alignment test (Fig. 2). The phylogenetic analysis of CgHMGR demonstrated that it belongs to the clade of the yeast species of the Saccharomycotina subphylum and the Saccharomyces genus, not to the clade containing C. albicans.

Figure 2
figure 2

The phylogenetic tree of the catalytic domain of HMGR enzymes in different organisms. The percentage of replicated trees in which the associated taxa are grouped in the 100 bootstrap replicates is shown next to the branches. The phylogenetic tree is drawn to scale, with the length of the branches representing the corresponding evolutionary distances according to the MEGA6 program, the maximum likelihood method, and the WAG + G model. The development of the species in the clade that encompasses Candida glabrata (in red) was based on whole genome duplication (WGD). In the clade containing Candida albicans (in green), the species developed based on genetic code transition (GCT).

Full size image

Molecular modeling of wild-type and mutant CgHMGR proteins

The homology modeling of the five mutants herein generated was carried out in the Modeller program, using the crystal structure of hHMGR (PDB: 1DQA) as the template. Of the fifteen models of each mutant provided, the one with the lowest DOPE score (the most thermodynamically stable) was selected. The selected models are portrayed in Fig. 3, highlighting in red the amino acids that were changed in the transition from the wild-type (Fig. 3a) to the mutant proteins (Fig. 3b–f): for E680Q, Glu66 to Gln66; for E711Q, Glu97 to Gln97; for D805A, Asp191 to Ala191; and for M807R, Met193 to Arg193.

Figure 3
figure 3

Structural modeling of the conformational differences between wild-type and mutant HMGR proteins of Candida glabrata (CgHMGR), in flat ribbon representation. (a) Wild-type CgHMGR, (b) mutant CgHMGRE680Q, (c) mutant CgHMGRE711Q, (d) mutant CgHMGRD805A, (e) mutant CgHMGRM807R, and (f) mutant CgHMGRE680Q-M807R. The α monomer is illustrated in pink, the β monomer in cyan, and the mutation in stick representation (red). Made using Discovery Studio software.

Full size image

The models were validated with the PROCHECK program, which quantifies the loss of residues in the allowed regions of a Ramachandran plot (Supplementary Fig. S2). In all the models, over 90% of the residues are found in the most favorable regions, with no residues in the disallowed regions (Supplementary Table 1). All the models constructed had Ramachandran statistics similar to those reported in the wild-type CgHMGR model. These results confirmed the high quality and reliability of the models generated.

Docking study of ligand binding to wild-type and mutant HMGR proteins

An in-silico analysis was carried out for the interaction of mutant and wild-type CgHMGR proteins with two of their inhibitors (simvastatin and alpha-asarone) and their substrate (HMG-CoA) (Fig. S5). Each mutant showed a lower binding energy for each ligand than the wild-type CgHMGR (Fig. S5, Table 2). The mutant protein E711Q displayed the lowest value (Fig. S5). Glu97, Asp307, Lys321 and His403 are the residues most frequently involved in binding, being located in the catalytic domain of CgHMGR. However, not all of these amino acids participate in the binding of the ligands with each of the mutant peptides (Table 2).

Regarding the binding of simvastatin, alpha-asarone, and HMG-CoA to the active site of the CgHMGR enzyme, the docking simulations gave insights into the binding energies as well as the residues, polar interactions, and hydrophobic interactions involved (Table 1).

Table 1 Results of docking simvastatin, alpha asarone, and HMG-CoA at the active site of HMGR of Candida glabrata (CgHMGR).
Full size table

Generation and verification of the point mutants of HMGR genes

PCR reactions were carried out with specific oligonucleotides designed to obtain the following mutants: CgHMGRE680Q, CgHMGRE711Q, CgHMGRD805A, CgHMGRM807R and CgHMGRE680Q-M807R. The plasmid Rec-MBP-CgHMGR was used as the template DNA (where MBP refers to the maltose-binding protein). The amplified oligonucleotides had a size of 7.9 kbp (Supplementary Fig. S3a), which corresponds to the size of the plasmid pMAL-C2X (6.6 kbp) plus the gene encoding the soluble fraction of CgHMGR (1.3 kbp). The full-length gel original is included in Supplementary Information (Fig. S3c). Additionally, the plasmids afforded by the transformation of each of the amplified products were subjected to double digestion (Supplementary Fig. S3b), finding the expected molecules of 6.6 and 1.3 kbp. The full-length blot is also in Supplementary Information (Fig. S3d).

Likewise, the mutations generated by sequencing were corroborated. The sequences obtained were aligned with the coding sequence of wild-type CgHMGR to confirm that the expected mutations had indeed been introduced. For CgHMGRE680Q and CgHMGRE711Q, the triplet changed from GAA to CAA, though in a distinct position in each case. The GAT triplet changed to GCT in CgHMGRD805A, and CAT to CCT in CgHMGRM807R. For the double mutant CgHMGRE680Q-M807R, TTC changed to TTG and CAT to CCT (Fig. S6). The sequencing revealed that the point mutations were introduced successfully.

Expression, detection, and purification of CgHMGR mutant proteins

The protein expression of the CgHMGRE680Q, CgHMGRE711Q, CgHMGRD805A, CgHMGRM807R and CgHMGRE680Q-M807R mutants was induced with IPTG. Samples taken before and after induction were visualized with SDS-PAGE. The gel in Fig. 4b shows the overexpression of the proteins of interest. The approximate molecular mass of 90 kDa corresponds to the fusion of CgHMGR + MBP (Fig. 4a). The pMAL vector favored the expression of the recombinant proteins, which fused to the soluble MBP. The latter facilitated the isolation of the enzyme after its overexpression.

Figure 4
figure 4

Graphic representation, overexpression, and detection of the recombinant CgHMGRE680Q, CgHMGRE711Q, CgHMGR805A, CgHMGRM807R and CgHMGRE680Q-M807R proteins. (a) Cartoon illustration of the p-MBP-CgHMGR (where MBP refers to the maltose-binding protein), which was used to obtain and express mutant CgHMGR proteins. (b) SDS-PAGE of the expression of the mutated proteins: Lane 1, molecular weight marker; lanes 2, 4, 6, 8 and 10, extract of cells before induction; lanes 3, 5, 7, 9 and 11, extract of cells after induction with IPTG. The full-length gel is included in Supplementary Information (Fig. S4a). (c) Example of the verification of protein identity by Western blot based on anti-MBP antibodies: Lane 1, molecular weight marker; lane 2, uninduced cell extract; lane 3, induced cell extract. The full-length blot is also shown in Supplementary Information (Fig. S4b).

Full size image

Hence, the mutants were overexpressed and the presence of MBP fused to the mutants was verified. The mutants found were the same as those previously identified by using anti-MBP antibodies (Fig. 4c). Subsequently, the mutants were purified easily (in a few steps) by means of amylose affinity chromatography. In all cases, a single band appeared at the approximate molecular mass of 90 kDa, exhibiting the same behavior as wild-type CgHMGR.

It was demonstrated by a coupling study that the substrate (HMG-CoA) and probable inhibitors (simvastatin and alpha-asarone) of CgHMGR did not recognize a specific site in MBP. Based on the resulting low binding energy values, ​​these compounds seem to have very limited affinity towards MBP. Moreover, HMG-CoA selectively binds to HMGR, as has been observed for hHMGR. The active site and the residues involved in catalysis have been reported (Andrade-Pavón et al., 2017). Since CgHMGR displayed strong enzymatic activity, and the clone of E. coli with an empty plasmid did not show HMGR activity, the presence of MBP did not affect the activity of this enzyme10.

Reduced enzymatic activity of the mutants compared to wild-type CgHMGR

The enzymatic activity of the CgHMGRE680Q, CgHMGRE711Q, CgHMGRD805A, CgHMGRM807R and CgHMGRE680Q-M807R proteins was measured spectrophotometrically by reading the decrease in absorbance of NADPH at 340 nm during 10 min. For each mutant, this parameter was determined in triplicate and expressed as the percentage of activity of the wild-type protein (considered as 100%) (Table 2).

Table 2 The enzymatic activity and binding energies of the point mutations made in the motifs of the dimerization site, substrate binding site, and cofactor binding site of CgHMGR.
Full size table

Compared to wild-type CgHMGR, all five mutants displayed a lower enzymatic activity. The smallest percentage of residual activity (23%) was observed for E711Q. The least affected enzyme was E680Q (mutated at the dimerization site), with 88.2% residual activity. The double mutant E680Q-M807R exhibited the second lowest activity (44%), while the residual activity of the corresponding single mutants, E680Q and M807R, was substantially higher (82% and 79%, respectively). The enzymatic activity of each mutant was significantly less than that of the wild-type protein, according to a paired Student's t-type analysis.

Kinetic parameters of the mutants

For the determination of kinetic parameters (KM and Vmax), activity was assayed in a reaction mixture containing increasing amounts of HMG-CoA, ranging from 1 to 64 μM. Enzymatic activity was obtained from monitoring the conversion of HMGCoA-dependent oxidation of NADPH and taking the linear fraction for each of the substrate concentrations to obtain a Michaelis–Menten plot (Fig. S7). The data were evaluated with the Lineweaver–Burk and Hanes-Woolf models (Fig. S7). The Hanes-Woolf model was used because of explaining the results better than the other. Moreover, the Hanes-Woolf data analysis led to an outcome comparable to that described in a previous report on the wild-type protein10. The present data for the KM and Vmax was as follows: 8.31 µM and 0.28 µM min−1 for CgHMGRE680Q, 18 µM and 0.70 µM min−1 for CgHMGRE711Q, and 11.84 µM and 0.43 µM min−1 for CgHMGRD805A.

Discussion

HMGR is a highly conserved enzyme anchored to the endoplasmic reticulum membrane of eukaryotes and to the plasma membrane of prokaryotes. It is responsible for the synthesis of cholesterol in mammals, ergosterol in fungi, and isoprenoids in bacteria1. Intense research efforts have focused on hHMGR as the target of inhibitors such as statins and fibrates, which diminish the synthesis of cholesterol and reduce the risk of cardiac arrest in patients with hyperlipidemia. Consequently, hHMGR has been examined at many levels: transcription, translational regulation, synthesis, inhibition, crystallization, and tertiary structure. However, reports on point mutations in non-human HMGR are limited, to the best of our knowledge, to P. mevalonii and Syrian hamsters. The current investigation is the first to make point mutations in the HMGR enzyme of a yeast of medical interest.

The alignment presently carried out with the amino acid sequences of HMGRs from various species demonstrated that the motif sequences of the dimerization site (ENVIG), the substrate binding site (EGCLVAS), and the cofactor binding site (DAMGMN) are highly conserved, regardless of whether they are from fungal, mammalian, plant or bacterial proteins. Hence, these amino acid residues appear to have an essential function in the catalytic reaction of the enzyme and/or in its conformation and tertiary structure, which makes them candidates for the evaluation of point mutants in key sequences.

On the other hand, the phylogenetic analysis of the HMGR proteins from different yeasts revealed a classification in accordance with their taxonomy. The majority are of the Phylum Ascomycota, which includes the Candida and Saccharomyces genera. CgHMGR is grouped with the yeast proteins of the WGD clade, encompassing K. lactis and S. cerevisiae17,18, and not with those of the CTG clade, containing C. albicans and other Candida species. Therefore, C. glabrata is a species phylogenetically closer to S. cerevisiae than to C. albicans. The HMGRs of Y. lipolytica, S. pombe and U. maydis (the latter being a fungus of the phylum Basidiomycota) belong to branches completely independent of Ascomycota. The HMGRs of other eukaryotes (e.g., plants, insects and mammals) are grouped into an independent clade and the HMGR of the bacterium P. mevalonii functions as an external group, further validating the previous results1.

hHMGR, purified and crystallized in the year 2000, was used as a template for the molecular modeling of proteins19. Although the three-dimensional structure of the catalytic domain of hHMGR (426–888 aa) is a tetramer, it has been suggested that the protein dimer might be able to bind to the HMG-CoA substrate19. The molecular modeling of the five CgHMGR with point mutations generated proteins made up of two identical subunits (dimers). The structures of the mutated proteins provided an a priori approximation to the 3D structure of the corresponding peptides, a necessary step for their in-silico analysis11.

According to kinetic studies on hHMGR, statins compete for the HMG-CoA substrate without affecting binding to the NADPH cofactor20. Hence, the substitution of an amino acid residue in the conserved sequences could plausibly alter the affinity of the enzyme for its substrate HMG-CoA as well as for simvastatin and alpha-asarone. To explore such a possibility, a molecular docking study was carried out.

This is the first report, to our knowledge, of an evaluation of molecular coupling between mutant proteins of CgHMGR and reference inhibitors. Information is herein provided on the substantial effect of mutations to specific amino acids on the ligand-receptor interaction.

The amino acids Glu97, Asp307, Lys321 and His403 are part of the catalytic domain of CgHMGR and their plausible role in the ligand-HMGR interaction has been proposed1,11. According to previous docking studies, the amino acids affected by the mutations likely influence the recognition of the substrate/inhibitor and/or alter the native structure of the protein21.

Statins and alpha-asarone are competitive inhibitors of HMGRs, blocking access to the HMG-CoA substrate1,3,10,11. The CgHMGR mutant in the substrate binding motif, E711Q, displayed the lowest binding energy for the ligands (simvastatin, alpha-asarone, and the HMG-CoA substrate) compared to the other mutants herein examined. The limited enzymatic activity and binding energy ​​of this mutant was expected, considering that the three ligands currently tested interact with the EGCLVAS motif (of the substrate binding site)3,10,11. Glutamic acid in the EGCLVAS region seems to have an important function in the catalysis of CgHMGR, as previously reported12.

CgHMGR-M807R was mutated at the cofactor binding site, specifically at the methionine position 807 of the DAMGMN sequence12. The change was from methionine, an amino acid with a neutral and nonpolar charge, to arginine, one with a positive charge. Although both amino acids participate in protein methylation22, the positive charge of arginine could be detrimental to the binding of the cofactor to the enzyme. Additionally, arginine contains a guanidinium group, which when ionized has a lower charge density than other amino acids. While methionine is a weak nucleophile and cannot be protonated22, the positive charge of arginine might be able to alter the α-helix or β-folded structure of the protein and consequently its enzymatic activity. Finally, the larger size of arginine than methionine may modify the structure of the protein. It is possible that these factors hindered the proper binding of the enzyme to its cofactor, which would contribute to the reduced enzymatic activity.

The mutation made in the dimerization site caused glutamic acid to be replaced by glutamine in position 680. Although both amino acids are very similar in size, glutamic acid has a negative charge and glutamine a polar neutral charge. Glutamic acid has carboxylate (COO–) side chains that are potential proton acceptors, forming hydrogen bonds and thus the secondary structure of the protein (β-folded sheets or α-helix)23. This amino acid could possibly have a similar function as the Glu83 residue, which is highly conserved and participates in catalysis by transferring protons to Lys267. When glutamic acid carries out the decomposition of mevaldyl-CoA, its protonation and deprotonation confers distinct levels of energy to the reaction, and these are necessary for the enzymatic activity that gives rise to sterols24. Glutamine is a glutamic acid amide afforded by the replacement of the hydroxyl of glutamic acid with an amine group24. Since it has a polar neutral charge, however, it might have affected the interaction with the amino acids belonging to the ENVIG sequence. Therefore, the proper formation of the dimers may not have taken place, which would lead to the inadequate binding of the enzyme to the cofactor and consequently to limited enzymatic activity.

It is not surprising that the double mutant (CgHMGRE680Q-M807R) exhibited the second lowest enzymatic activity of the present mutants. The two mutations occurred at sites where amino acids are highly conserved: the sequence of the cofactor binding site (DAMGMN) and of the dimerization site (ENVIG).

There was a greater KM value found for each mutant than for the wild-type protein10. Hence, the data suggest that the affinity of HMG-CoA would be lower for the mutant versions than for the wild-type protein. Indeed, a marked decrease was detected in the enzymatic activity of the mutants herein examined compared to that of the wild-type protein.

C. glabrata has only one HMG1 gene, which is orthologous to the S. cerevisiae gene7. In the current contribution, the sensitivity of CgHMGR mutants to antifungals was not evaluated. According to a previous study, the mutation of this gene in Aspergillus fumigatus causes resistance to antifungal compounds of the triazole class25,26. The mutation of the enzymes HMG1 and HMG2 of S. cerevisiae leads to a great sensitivity to compactin, an HMGR inhibitor27. There are no reports, to our knowledge, on a double mutant cdr1∆ hmg1∆. However, these mutants have been assessed separately, observing an increase in sensitivity to azoles caused by cdr1∆ in C. glabrata as well as resistance to the same drugs produced by hmg1∆ in Aspergillus fumigatus25,26,28.

The insights gained from the presently mutated proteins can be applied not only to the improvement of C. glabrata as a model for research on drug resistance, but also to the investigation of the HMGR of additional species, such as C. auris and C. haemulonii, which have recently emerged as pathogenic multi-drug resistant strains29. The emergence of Candida infections associated with SARS-Cov2 gives greater emphasis to the urgency of developing antifungal agents for alternative targets30.

Conclusions

Five point mutants were made at three highly conserved sites of the catalytic domain of HMGR: the dimerization, substrate binding and cofactor binding sites. After the mutants were expressed, they were detected and purified to verify their integrity. All the mutants exhibited significantly lower enzymatic activity and ligand affinity than wild-type CgHMGR. The in vitro and in silico testing of the mutants gave insights into the function of specific amino acid residues in the catalytic site of the enzyme, especially glutamic acid. The current findings should certainly facilitate the creation of new mutants of HMGR to further explore the catalysis, recognition, conformation, and dimerization mechanisms involved in the activity of the enzyme and the viability of C. glabrata and other opportunistic pathogenic fungi.

Materials and methods

Multiple sequence alignment of HMGR enzymes from different organisms

The amino acid sequences were downloaded from the NCBI database (https://www.ncbi.nlm.nih.gov/) for the HMGR enzymes of the following organisms: P. mevalonii (WP_016714288), Arabidopsis thaliana (NP_177775), D. melanogaster (NP_732900), H. sapiens (XP_011541659), Mus musculus (NP_001347095), Rattus norvegicus (NP_037266), U. maydis 521 (XP_011389590), S. pombe (NP_588235), Y. lipolytica CLIB122 (XP_503558), C. auris (XP_028890531), C. haemulonii (XP_025340375.1), C. lusitaniae ATCC 42,720 (XP_002615608), M. guilliermondii ATCC 6260 (XP_001482757), D. hansenii CBS767 (also called Candida famata) (XP_458872), C. parapsilosis (KAF6043359), C. orthopsilosis Co 90–125 (XP_003868277), C. tropicalis MYA-3404 (XP_002550050), C. albicans SC5314 (XP_713636), C. dubliniensis CD36 (XP_002417024), C. kefyr (QGN16198), K. lactis NRRL Y-1140 (XP_451740), C. glabrata CBS138 (XP_449268), S. cerevisiae S288c HMG1(NP_013636) and S. cerevisiae 2288c HMG2 (NP_013555)31. Once the sequences corresponding to the catalytic domain of such enzymes were downloaded, they were selected and subjected to multiple alignment in the Clustal Omega program32. The motif domains (structural motifs) were located with the ExPASy server, PROSITE (https://prosite.expasy.org/)33, and visualized with the web server WebLogo (http://weblogo.berkeley.edu/logo.cgi).

Phylogenetic analysis of HMGR proteins from different organisms

A phylogenetic tree was constructed with the sequences of the catalytic domain of the aforementioned twenty-four HMGRs in the MEGA6 program34, utilizing the maximum likelihood method and the WAG + G model. Additionally, a bootstrap replication of 100 was used to evaluate the reliability of the phylogenetic tree.

Homology modeling of CgHMGR mutant proteins

Fifteen models were generated for each of the five CgHMGR mutant proteins with Modeller 9.2435. The models with the lowest discrete optimized protein energy (DOPE) score were selected and edited in Discovery Studio36, then subjected to a validation process to assess their stereochemical quality with the PROCHECK37 tool integrated in the Structural Analysis and Verification Server (SAVES) (https://servicesn.mbi.ucla.edu/SAVES/). This tool also quantified the amino acid residues in the available regions of the Ramachandran plots.

Docking studies

The ligands were downloaded from the PubChem server38 in 2D format and transformed into 3D mol2 format in the Open Babel GUI program39. Kollman charges were assigned to the hydrogen atoms of the protein. Preparation of the docking parameters and molecular coupling of the five mutants to the ligands (HMG-CoA, simvastatin, and alpha asarone) was carried out on AutoDock Tool 4.040. The grid dimensions were set at 128 × 88 × 64 Å3 with points separated by 0.375 Å, and the grid center was X = − 10.517, Y = 14.058 and Z = 22.162. One hundred docking runs were made with the Lamarckian genetic algorithm and various parameters were estimated. The results were compared between ligands based on their lowest free coupling energy. Interactions were visualized with Discovery Studio.

Microorganisms and plasmids

Transformation was achieved with the Escherichia coli DH10β strain and the overexpression of heterologous proteins with the E. coli BL21 strain. The plasmid pMAL-C2X-CgHMGR, which was previously obtained by our group and harbors the gene encoding the wild-type protein, served as the template DNA for generating the mutants10.

Culture media

E. coli harboring the plasmids was grown in liquid Luria Bertani (LB) medium (1.0% casein peptone, 0.5% yeast extract, and 0.5% NaCl) and solid LB medium with 1.5% bacteriological agar. These media were adjusted to pH 7.0 and supplemented with 100 μg/mL ampicillin. Culture media and antibiotics were acquired from Sigma Aldrich (St. Louis, MO, USA).

Mutagenic oligonucleotide design

The mutagenic primers were designed manually, taking as templates the sequences of the dimerization site, the substrate binding site, and the cofactor binding site within the catalytic domain of the HMGR enzyme. Consideration was given to 18 base pairs to the left and 18 base pairs to the right of the codon where the mutation was designed, resulting in a total of 39 nucleotides for each primer. Point mutations were made involving the following changes: glutamic acid to glutamine in the dimerization site (ENVIG) to provide E680Q, aspartic acid to glutamine in the substrate binding site (EGCLVAS) to afford E711Q, aspartic acid to alanine in the cofactor binding site (DAMGMN) to furnish (D805A), and methionine to arginine in the cofactor binding site to yield M807R. The double mutation of E680Q-M807R was formed by the respective single point mutations.

The oligonucleotides herein designed were synthesized at the Institute of Biotechnology of the National Autonomous University of Mexico (UNAM). The melting temperature of the oligonucleotides was calculated after their synthesis, using the bioinformatics program Oligo Analyzer from Integrated DNA Technologies (IDT) (https://www.idtdna.com/pages/tools/oligoanalyzer). Table 3 shows the original sequence (wild-type CgHMGR) and the mutagenic oligonucleotides. The codon of the amino acid to be mutated is illustrated in red, and the change made is underlined.

Table 3 Oligonucleotides designed to obtain the following point mutants: E680Q, E711Q, D805A, M807R and E680Q-M807R.
Full size table

Generation of CgHMGR point mutants by PCR

The E680Q, E711Q and D805A point mutants were generated by PCR with a "Quick Change" commercial kit (Thermo Fisher Scientific, USA), as specified in the manufacturer’s instructions with some modifications. This procedure is based on a PCR reaction with two complementary and antiparallel oligonucleotides and the appropriate DNA polymerase to achieve the desired mutation. Each oligonucleotide hybridizes with one of the two strands of the plasmid to be amplified. The plasmid obtained led to the sequence to be mutated (pMAL-c2X-CgHMGR), which was used as the template DNA. Mutants M807R and E680Q-M807R were produced by inverse PCR, with plasmid pMBP-CgHMGR as the template DNA10,11. For the double mutant, the plasmid DNA of the rec-CgHMGR-E680Q mutant and the mutagenic oligonucleotide M807R served as the template DNA.

Once the reaction conditions were established and developed, the amplicon was digested with the Dpn-I enzyme (Thermo Fisher Scientific, USA). E. coli DH10β cells were then transformed with the digested PCR product. Subsequently, E. coli cells were plated on LB agar with ampicillin (100 µg/ml) and incubated at 37 °C for 24 h. The transforming colonies were selected for the extraction of plasmid DNA. Plasmids were purified with the ZymoPURE plasmid Miniprep Kit (Zymo Research, CA, USA) and confirmed by sequencing them according to the Sanger method, a procedure that demonstrated the presence of the desired mutations (Macrogen®, Seoul Korea).

Expression, purification and detection of wild-type and mutant recombinant proteins

Once the presence of the desired mutations was confirmed by sequencing the corresponding plasmids, the recombinant proteins were expressed. Briefly, E. coli BL21 cells were transformed with the mutated DNA by growing the cells containing the fusion plasmid in LB medium plus ampicillin (100 µg/mL) and 0.2% glucose until reaching an optical density of 0.6 at 600 nm. IPTG was added immediately (0.3 mM final concentration) to induce the expression of the protein, a process carried out at 37 °C for 4 h. Upon completion of this time, cells were harvested at 4000 × g for 20 min at 4 °C. The cells were mechanically broken by using glass beads and a vortex, and then the cell pack was dissolved in a column buffer (1 M Tris–HCl, pH 7.4, 11.7 g NaCl, 0.5 M EDTA, and 154 mg DTT) to obtain a cell-free extract. The cells were disrupted for 20 min, alternating every other 30 s on and off ice. The cell lysate was centrifuged at 18,000 g for 20 min and the supernatant (cell-free extract) was recovered. The protein was purified by affinity column chromatography with amylose resin (New England Biolabs, USA). The wild-type and mutant proteins were eluted with a buffer supplemented with 10 mM maltose. Fractions were collected to determine CgHMGR activity and protein concentration by the Lowry method. The expression and the successful purification of the mutant proteins was visualized by means of 10% SDS-PAGE stained with Coomassie blue. The mutated proteins were detected by Western blot with the anti-MBP monoclonal antibody (New England Biolabs, USA), according to the manufacturer’s specifications.

The HMGR enzymatic activity assay

The enzymatic activity of the five mutants was quantified by the method of Bischoff and Rodwell41, beginning with the elaboration of a reaction mixture containing 10 mM HMG-CoA, 50 mM Tris–HCl regulator, injectable water, 9 µg of the protein of the respective mutant, and 10 mM NADPH. The cofactor (NADPH) is oxidized by the catalytic subunit of HMGR in the presence of the substrate HMG-CoA, causing a decrease in absorbance. Thus, absorbance was monitored at 340 nm for 10 min. Reagents were purchased from Sigma Aldrich (St Louis, MO, USA). One unit of specific activity was defined as the amount of enzyme that oxidizes 1 µmol of NADPH to NADP + per minute per milligram of protein. All spectrophotometric readings were made on a BioSpectrometer® kinetic apparatus (Eppendorf AG, Germany).

Substrate specificity and kinetic parameters

The steady-state kinetic constants for mutants were established by means of the standard activity assay under the same conditions described in the previous section. Activity was determined by incubating reaction mixtures with substrate concentrations ranging from 1 to 64 µM. The KM and Vmax constants were obtained from Hanes-Woolf plots and expressed as the mean of three different experiments10,42.