Investigation of the prevalence and catalytic activity of rubredoxin-fused alkane monooxygenases (AlkBs)

Abstract

Interest in understanding the environmental distribution of the alkane monooxygenase (AlkB) enzyme led to the identification of over 100 distinct alkane monooxygenase (AlkB) enzymes containing a covalently bound, or fused, rubredoxin. The rubredoxin-fused AlkB from Dietzia cinnamea was cloned as a full-length protein and as a truncated protein with the rubredoxin domain deleted. A point mutation (V91W) was introduced into the full-length protein, with the goal of assessing how steric bulk in the putative substrate channel might affect selectivity. Based on activity studies with alkane and alkene substrates, the rubredoxin-fused AlkB oxidizes a similar range of alkane substrates relative to its rubredoxin domain-deletion counterpart. Oxidation of terminal alkenes generated both an epoxide and a terminal aldehyde. The products of V91W-mutant-catalyzed oxidation of alkenes had a higher aldehyde-to-epoxide ratio than the products formed in the presence of the wild type protein. These results are consistent with this mutation causing a structural change impacting substrate positioning.

Graphical abstract

1. INTRODUCTION

Alkane monooxygenase (AlkB) catalyzes the conversion of medium- to long-chain n-alkanes to primary alcohols.¹ Its ability to selectively activate a C-H bond and oxidize energetically rich -- although frequently toxic -- alkanes to ecologically less harmful alcohols has made AlkB an attractive candidate for biocatalysis and bioremediation of oil in the environment.^2–5 In addition to its function as an alkane monooxygenase, AlkB has been shown to catalyze the further oxidation of primary alcohols to aldehydes and carboxylic acids.^{6, 7} These overoxidation reactions have been investigated for their utility in biocatalysis.^{8, 9} In addition to alkanes, AlkB is active on terminal alkenes, producing epoxides and aldehydes.^10–12

Obtaining a high-resolution structure of AlkB has been a long-coveted goal, as it would enable a greater understanding of the molecular basis of its reactivity. Although no such structure has been obtained, analyses of the protein function and sequence have been performed through comparing the activity of mutants to wildtype proteins. These analyses have revealed clues to the protein’s overall structure. AlkB is an integral membrane protein that contains a non-heme dinuclear iron active site coordinated by three regions of conserved histidines.^13–16 It is related to a diverse group of enzymes that includes fatty acid desaturases, fatty acid hydroxylases, and decarbonylases.¹⁷ Members of this group contribute not only to bioremediation of oil, but also to human pathologies such as diabetes,¹⁸ obesity,¹⁹ cancer,¹⁹ development problems,²⁰ and attention-deficit/hyperactivity disorder.²¹ AlkB itself may also be relevant to human health. It is expressed in pathogenic bacteria such as Mycobacterium tuberculosis, Legionella pneumophila, and Pseudomonas aeruginosa.^{17, 22, 23} Its role in these bacteria is poorly understood.

The catalytic activity of AlkB depends on the presence of two essential electron transfer partners, a rubredoxin (AlkG) and a rubredoxin reductase (AlkT).^24–27 However, an alternative domain structure of AlkB has a fused rubredoxin domain located at the C-terminus of the protein.²⁸ Previous studies have identified fewer than 30 proteins with this fused domain structure.^{17, 28–30} Functional analyses have indicated that these rubredoxin-fused AlkBs enable cells to grow on long-chain alkanes.^{17, 29, 31} Previous functional studies have been conducted in vivo and have been limited in scope to proteins from only two genera, although genes encoding rubredoxin-fused AlkBs exist in at least 28 genera. To our knowledge, the function of rubredoxin-fused AlkBs has not been explored in vitro. Consequently, little is known about their mechanism or the products they generate.

This study describes the bioinformatic analysis of rubredoxin-fused AlkBs and the subsequent investigation of one homologue from D. cinnamea through analysis of its activity on both alkane and alkene substrates. The activity of the full-length protein was examined, as was the activity of its truncated form without the fused rubredoxin domain and the activity of the full-length protein with a point mutation of a residue hypothesized to be in the substrate channel.

2. EXPERIMENTAL

2.1. Materials and methods

Chemicals and solvents were purchased from Millipore Sigma, Fisher Scientific, and Anatrase. Genes were synthesized by Integrated DNA Technologies (IDT). PCR was performed using a C1000 Series Touch Thermal Cycler (Bio-Rad), and Sanger sequencing was performed by Genewiz.

Low-speed spins were done using a Beckman Coulter Avanti J-26S XPI centrifuge with a JLA-8.1000 rotor, using a Beckman-Coulter Allegra X-30 centrifuge with a C0650 rotor, or using a Fisher Scientific AccuSpin Micro 17 centrifuge. High-speed spins were done with a Thermo Scientific Sorvall wX + Ultra Series centrifuge with an F50L-8 × 39 rotor. Cells were lysed using a French Press G-M.

Gas chromatography mass spectrometry (GCMS) analyses were performed on an Agilent 7820A GC with an Agilent HP-5 column and an Agilent 5977E MSD.

2.2. Bioinformatics.

The NCBI database was searched for genes coding for rubredoxin-fused AlkBs. These were filtered by cross-referencing with the SILVA database for “Type” strains to remove any redundancy.³² The programs RAxML, Clustal Omega, and ETE 3were used to generate a cladrogram.^33–35 All codes used have been deposited in a public repository available at https://github.com/ajlopatkin.

2.3. Genetic transformation of E. coli

The bacterial strains, plasmids, and oligonucleotide primers used in this study are described in the supplemental material (Supplemental Table 1). G-blocks encoding the desired AlkB proteins were codon optimized for E. coli expression and ordered from IDT (Supplemental Table 2). Genes were inserted into the pET-15b plasmid using a modified fast-cloning method.³⁶ Correct gene insertion was confirm using Sanger sequencing of the entire open reading frame (Genewiz). The miniprepped DNA was used to transform chemically competent BL21-DE3-star cells. These were plated on Luria-Bertani (LB)/agar plates with ampicillin.

2.4. Cell growth

For all cell cultures, a single colony was selected from the LB/agar plate and used to inoculate 50 mL sterile liquid LB media with 100 μg/mL ampicillin and grown overnight while shaking at 37 °C. The overnight inoculum was then transferred to 500 mL sterile liquid LB media with 100 μg/mL ampicillin and incubated at 37 °C until OD₆₀₀ reached 0.5. At this point, isopropyl β- D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 200 μM to induce alkB expression. FeCl₃ (50 μM) was also added, and the cells were incubated for an additional 6–8 h. Cells were harvested by centrifugation at 7500 ×g for 15 minutes and stored at –80 °C.

2.5. Collection of AlkB

Frozen cell pellets were thawed and resuspended in 1 mL lysis buffer (20 mM Tris-HCl, 150 mM NaCl, pH 7.8) per 1 g cell pellet. DNase (1.6 units/mL) and phenylmethylsulfonyl fluoride (PMSF, 0.4 mM) were added. Cells were lysed by French Press at a pressure of 10,000–12,000 psi. Unbroken cells in the crude lysate were removed by centrifugation at 7000 ×g for 20 minutes. The supernatant was saved as cell free extract.

2.6. Measuring AlkB activity

AlkB-containing cell free extract (0.5 mL) was diluted in lysis buffer to a final volume of 1 mL. It was incubated in the presence of ferrous ammonium sulfate hexahydrate (1 mM), sodium hydrosulfite (1 mM), NADH (12.7 mM), and purified substrate (2 μL) for 30 min at 37 °C.^{37, 38} The reaction products were extracted with methylene chloride (100 μL) and quantified by GCMS. For AlkB-containing cells expressing the rubredoxin-deletion mutant, 125 uL of a buffered solution containing AlkG and AlkT from P. putida GPo1 was added.³⁹ Control experiments in which all reagents were added, including cell free extract from E. coli not expressing AlkB, were done.

3. RESULTS AND DISCUSSION

3.1. Diversity and prevalence of rubredoxin-fused AlkBs

A search of the NCBIdatabase revealed 345 genes coding for rubredoxin-fused AlkBs. After filtering remove redundancy,³² 110 results from high-quality genomes of distinct species were found to encode rubredoxin-fused AlkBs. These results came from 28 genera in the phyla proteobacteria and actinobacteria and included both gram-positive and gram-negative species. A cladogram was constructed to compare these rubredoxin-fused AlkB genes.^33–35.

The cladogram (Figure 1) demonstrates a genetic diversity that marks a significant increase from the fewer than 30 rubredoxin-fused AlkBs from only two genera that had been previously identified, several of which came from the same species.³⁰ It was found that of the 110 species identified to contain genes for rubredoxin-fused AlkBs, over 70 of them were from genera with organisms known to produce biosurfactants (highlighted in Figure 1).^40–47 Given the prevalence of rubredoxin-fused AlkBs and the observation that many organisms encoding them produce biosurfactants, we hypothesize that AlkB may play a role in biosynthetic pathways in addition to pathways that breakdown alkanes as an energy source. This hypothesis is supported by an analysis of the gene environment of 24 rubredoxin-fused AlkBs, showing that approximately ¾^th are near genes associated with cell wall biosynthesis (see SI). In order to better understand the biochemical capabilities of rubredoxin-fused AlkBs, the catalytic activity of AlkB from Dietzia cinnamea (Dc AlkB) was studied.^{48, 49} It was cloned both to encode the full-length protein and to encode the rubredoxin-deletion mutant (Dc nulRd AlkB).

Figure 1.

A cladogram prepared with RAxML, Clustal Omega, and ETE 3 comparing the similarity of rubredoxin-fused AlkB genes from 110 distinct bacterial species. Color-coded by genus. Labels indicate unique species, and the names in boxes indicate species known to produce surfactants.

3.2. Fused rubredoxin does not affect reactivity on alkane substrates

AlkBs can oxidize a variety of alkanes from C3 to C30, although each individual AlkB that has been examined shows a much narrower substrate range. The first AlkB isolated (and still the most widely studied), the AlkB from P. putida GPo1, oxidizes alkanes from propane to dodecane with a marked preference for nonane and octane.^{26, 50–53} Previous work has found that the presence of a rubredoxin-fused AlkB enables organisms to consume as their sole carbon source longer-chain alkanes than can be consumed by organisms bearing AlkBs that lack the rubredoxin fusion.³⁰ This work suggests that the rubredoxin-fused AlkB plays a role in long-chain alkane degradation. This role was investigated by examining the activity of Dc AlkB and comparing it to the activity of the rubredoxin deletion mutantDc nulRd AlkB. The activity was studied in cell lysate to which NADH (the electron source) and purified alkane substates were added. (Ferrous iron and dithionite were also added to ensure that the protein was fully metallated.³⁷) In experiments with the rubredoxin deletion mutant, the electron transfer proteins AlkT and AlkG from P. putida Gpo1 were also added. Prior work has shown that rubredoxins from one AlkB-expressing organism can complement AlkBs from other organisms.^{17, 54–56} Control experiments showed no transformation of any alkane in the absence of cell lysate from cells expressing alkB genes.

Contrary to the initial hypothesis, the presence of the fused rubredoxin did not yield increased overall activity on longer substrates (Figure 2). This result indicates that the previously reported in vivo results are likely influenced by factors other than the biochemical constraints of the AlkB protein.²⁹

Figure 2.

The relative substrate preference of Dc AlkB compared to that of Dc nulRd AlkB. Error bars +/− 1 se for 4 replicates.

3.3. Point mutation V91W in Dc modestly affects substrate preference

Previous studies have also identified the amino acid residue at location 55 in AlkB from P. putida as a determinant of substrate preference.⁵⁷ This position is occupied by a bulky tryptophan in AlkBs that oxidize smaller substrates but is replaced by a smaller residue in systems that oxidize larger substrates (Fig. 3).^{56, 57} A sequence alignment of AlkB from P. putida and D. cinnamea reveals that this residue lies at position 91 in the hydroxylase portion of Dc AlkB (Fig 3). Because most rubredoxin-fused AlkBs, including Dc AlkB, contain a smaller residue such as valine at this position, its relevance for substrate preference in rubredoxin-fused AlkBs was tested. The residue V91 in Dc AlkB was mutated to a tryptophan in Dc V91W AlkB, and the activity of the protein was examined.

Figure 3.

Alignment of region linked to substrate specificity for the AlkB from D. cinnamea and P. putida GPo1, along with two other AlkBs: one (the AlkB from A. borkumensis⁵⁶) that oxidizes shorter chain alkanes and one (the AlkB from P. aeruginosa²²) that oxidizes longer chain alkanes.The red block (corresponding to residue 91 in D. cinnamea) marks the position of interest. Created with Geneious version 2021.0 (created by Biomatters. Available from https://www.geneious.com

The results (Figure 4) demonstrate that in the case of the rubredoxin-fused AlkB, Dc AlkB, the point mutation V91W had a modest shift in substrate preference. The mutant, which contained a bulky tryptophan residue at the location 91, retained activity over all surveyed substrates on which the wildtype protein was active, but with slightly more activity on shorter substrates and slightly less activity on longer substrates, consistent with the hypothesis that this residue points into the substrate channel at a position that can impact substrate positioning.

Figure 4.

The relative substrate preference of Dc AlkB compared to that of Dc V91W AlkB. Error bars +/− 1 se for 3 replicates.

3.4. Point mutation V91W in Dc AlkB changes rubredoxin-fused AlkB reactivity on alkenes

The point mutation V91W also changed the activity of the rubredoxin-fused AlkB on alkenes. AlkB yields two products when catalyzing the oxidation of alkenes: a terminal epoxide and a terminal aldehyde.¹² Extensive work has determined that these products stem from two unique reaction pathways (Scheme 1).^{17, 58}

Scheme 1.

Proposed mechanism for the AlkB-catalyzed oxidation of alkenes to form A. epoxide and B. aldehyde products.

In the epoxidation pathway, one electron from the terminal carbon of the alkene attacks the oxygen in the iron-oxo moiety, reducing one of the irons in the di-iron site. Published work showing the loss of olefin configuration during epoxidation led May and coworkers to propose a step-wise mechanism for AlkB-catalyzed epoxidation.⁵⁹ The attack at the terminal carbon generates a radical at the C2 position, which is then quenched by the oxygen after rotation of the incipient bond to facilitate ring closure. May and coworkers demonstrated that the AlkB-catalyzed epoxidation of a cis-alkene (deuterated at the terminal position to create the cis-configuration) produced 70% trans epoxide while the same reaction done with the trans-alkene produced 70% cis epoxide. This result is consistent with preferential attack of the metal oxo from the re face and then ring closure from the si face of C-2 to avoid steric clashes with the methylene chain binding region as depicted in the mechanism above.¹² The aldehyde forms by the creation of a C-O bond on the terminal carbon as well; however this occurs when both electrons from the alkene form a bond with the oxygen in the iron-oxo intermediate, reducing the iron and creating a carbocation on C2. Hydrogen migration to C2, previously demonstrated with deuterium-labeling studies¹², displaces the carbocation to C1 which can then be quenched by collapse of the iron-oxygen bond to make an anti-Markovnikov aldehyde.

Under our experimental conditions, in which an alcohol dehydrogenase is present in the cell lysates, the aldehyde was immediately and completely reduced to the primary alcohol, which is observed by GC-MS, consistent with prior work.^8,41

The Dc AlkB-catalyzed oxidation of octadiene yieldsa ratio of aldehyde to epoxide of approximately 1:2. When Dc V91W AlkB catalyzes the oxidation of octadiene, however, the ratio of the aldehyde to epoxide is approximately 1:1. This represents a major shift in the distribution of products in favor of the production of the aldehyde. The shift towards the aldehyde product is also observed when Dc V91W AlkB oxidizes the longer substrate decadiene (Figure 5, with more detail provided in SI). To our knowledge, this is the first report of any AlkB homologue that produces a comparable amount of aldehyde relative to epoxide when oxidizing an alkene and provides an enzymatic route to the production of anti-Markovnikov products.

Figure 5.

Product distribution of Dc AlkB and Dc V91W AlkB with diene substrates. Alcohol concentration plotted as a proxy for aldehyde produced. Error bars +/− 1 standard deviation for three experimental replicates.

These results lead us to propose a model in which the tryptophan mutation at position 91 orients the alkene substrate in the active site. In the wild-type Dc AlkB, the absence of this amino acid gate allows for more lateral movement of the substrate, which enables the substrate to move closer to the diiron active site and facilitates the ring closure necessary for epoxide formation. In contrast, a tryptophan side chain at this position prevents the lateral movement of the substrate intermediate towards the active site and thereby inhibits the formation of the epoxide. Favorable interactions between the aromatic side chain and the substrate π bond may also play a role in limiting substrate motion.⁶⁰ Because the aldehyde formation mechanism does not require the active site to bridge both carbons, it is preferred when the substrate has less mobility the active site (as illustrated in Figure 6). This model reflects the knowledge that subtle structural changes that impact substrate position relative to high valent metal oxo intermediates can alter reactivity as has been seen with Ole-T⁶¹ and the OG-dependent Fe enzymes⁶² and is also informed by an intriguing early report that when AlkB oxidizes a diene twice to produce a diepoxide, the stereochemistry of the epoxide formed first impacts the stereochemical outcome of the second epoxide⁶³. More structural information is required to build a more precise model, information we hope in later work to provide.

Figure 6.

Illustration of how the bulky aromatic amino acid at position 91 could influence the reaction trajectory by making the diene substrate less mobile in the active site, which as illustrated in scheme 1, would favor the formation of the aldehyde.

4. CONCLUSION

Rubredoxin-fused AlkBs are far more ubiquitous in nature than previously appreciated. This study reveals the presence of rubredoxin-fused AlkB-encoding genes in at least 110 distinct species across 28 genera from both gram-positive and gram-negative bacteria. It is likely that these genes have been conserved because their encoded proteins provide useful and distinct reactivity. One such protein, Dc AlkB was expressed as the full-length protein, the rubredoxin-deletion mutant, and with a point mutation V91W. The rubredoxin-deletion mutant did not shift the substrate range towards shorter substrates for the range of alkanes tested. The point mutation V91W had a modest effect on alkane substrate preference and changed the protein’s reactivity on alkene substrates, leading to the increased generation of an anti-Markovnikov aldehyde product. Because they do not require an exogenous rubredoxin, these rubredoxin-fused AlkBs are good candidates for biocatalysis. Further studies will investigate the evolutionary selection for AlkBs, including the extent to which horizontal gene transfer may have spurred the dispersion of rubredoxin-fused AlkBs throughout so many bacterial species and the potential role of these proteins in biosynthetic processes in vivo.

Highlights.

• Alkane monooxygenases with fused rubredoxin widely distributed in nature

• Rubredoxin deletion mutant has similar substrate range to wild type

• Single point mutation in presumed substrate domain impacts reactivity and leads to an enzyme with substantial anti-Markovnikov activity on alkenes