Metabolic engineering of Rhodococcus ruber Chol-4: A cell factory for testosterone production

Rhodococcus ruber Chol-4 is a potent steroid degrader that has a great potential as a biotechnological tool. As proof of concept, this work presents testosterone production from 4-androstene-3,17-dione by tailoring innate catabolic enzymes of the steroid catabolism inside the strain. A R. ruber quadruple mutant was constructed in order to avoid the breakage of the steroid nucleus. At the same time, an inducible expression vector for this strain was developed. The 17-ketoreductase gene from the fungus Cochliobolus lunatus was cloned and overexpressed in this vector. The engineered strain was able to produce testosterone from 4-androstene-3,17-dione using glucose for cofactor regeneration with a molar conversion of 61%. It is important to note that 91% of the testosterone was secreted outside the cell after 3 days of cell biotransformation. The results support the idea that Rhodococcus ruber Chol-4 can be metabolically engineered and can be used for the production of steroid intermediates.


Introduction
Steroids are a family of terpenoid lipids widely distributed in nature that present a common structure formed by four fused alicyclic rings called gonane (S1 Fig). They play important biological roles in eukaryotic cell membrane stabilization (e.g. structural steroids such as cholesterol) and the regulation of cellular processes such as proliferation and differentiation (e.g. steroid hormones such as progesterone or testosterone, TS).
Steroid-based drugs are widely used as anti-inflammatory, diuretic, anabolic, contraceptive, anti-androgenic, progestational and anticancer agents and they are therefore a highly valuable resource for the pharmaceutical industry with an annual global market of over $10 billon [1]. Steroids can be chemically produced in a variety of ways, but the most common initial substrates for their production are cost-effective phytosterols, derived from plants [2]. However, in the recent years, there has been an increasing effort to move from the multistep chemical process to a single step microbial bioconversion process that could reduce costs and production yield losses, and would also be more eco-friendly [1]. All this led to an increasing interest PLOS  in the use of recombinant bacteria as cell biofactories. These compounds are synthesized and found predominantly in eukaryotic cells while some bacteria, that rarely generate them, are able to use these compounds as growth substrates [3]. Currently, this objective is much more achievable thanks to recent advances in our understanding of steroid bacterial catabolism, which led to the development of bacterial strains that can convert precursors (e.g. phytosterols from cheap agricultural plant waste) into high-value steroids [4,5]. Therefore, in this work, we present Rhodococcus as a promising candidate for steroid production.
The cholesterol-degrading R. ruber strain Chol-4, isolated from a sewage sludge sample, was chosen [6] as a model organism in this work for the following reasons: i) R. ruber strain Chol-4 is able to catabolize steroids; ii) many of its enzymes involved in the steroid catabolism have been characterized in the past years and iii) an important mutant strain collection listed in S1 Table is now available [7][8][9][10]. However, so far, there are only a few studies on the use of engineered Rhodococcus as a cell system for pharmaceutical steroid production [11][12][13].
A mutant with impaired Ksh and KstD activities is not able to degrade AD, favoring its modification into any molecule of interest with the appropriate enzymatic activity. Rhodococcus, just like other actinobacteria, has many isoforms of these enzymes. Fortunately, all R. ruber KshA isoforms (KshA1, KshA2, KshA3) combine with only one KshB, therefore Ksh activity is absent by knocking down the kshB gene [9]. R. ruber displays three KstD isoforms (KstD1,2,3) and therefore it is necessary to knock down all of them to impair KstD activity [8,10].
To study the biotechnological capabilities of a R. ruber system, a biotransformation of AD to testosterone, a natural steroid hormone from the androgen group, was conducted using the fungal enzyme 17-ketosteroid reductase. As the degradation pathway of AD is well characterized in Rhodococcus ruber, this substrate was chosen for the study. Assays were not considered on other steroids such as cholesterol or phytosterols as initial substrates because R. ruber displays more than one steroid pathway, whose major steps remain unknown and do not consequently yield AD as intermediate in this strain [8,9].

Construction of ΔkshB-kstD1,2,3 R. ruber strain Chol-4 mutant
AD is a substrate of both KstD and KshAB activities in R. ruber. The KstD1,2,3 R. ruber mutant accumulates 9OH-ADD from AD while the KshB R. ruber mutant accumulates ADD [8,9]. Therefore, the first step to obtain steroid derivatives from AD in R. ruber, is to block its AD catabolic pathway by building a mutant strain in which the activity of these enzymes is impaired (Fig 1). Besides avoiding the AD consumption, the quadruple mutant ΔkshB-kstD1,2,3 has also blocked the testosterone catabolism.
In order to get the quadruple mutant, the ΔkstD1,2,3 strain [8] (S1 Table) was used for kshB gene deletion, according to Material and methods section. The ΔkshB-kstD1,2,3 quadruple mutant was checked by PCR. The growth experiments proved that it was not able to grow on 2 mM AD, ADD or testosterone while it kept growing on 24 mM sodium acetate (S2 Fig). The slight growth seen on cholesterol is due to the consumption of the side chain of this substrate [9]. Therefore, the quadruple mutant was suitable to be used as a host for the 17β-hsd overexpression for AD biotransformation.

Construction of an expression vector for R. ruber Chol-4 and 17β-hsd gene cloning
Attempts to transform R. ruber Chol-4 with expression vectors (pTIP-QC1, pNIT-QC1, pTNR-KA, pTNR-TA, etc.) were unsuccessful. The biotechnological potential use of R. ruber was restricted by the need for an appropriate expression vector.
To create an inducible expression vector for R. ruber, we combined the pNV119 vector, which was able to replicate in this strain, and an artificial regulon (GenBank: FjI73069) [36] that contains a gene encoding a regulatory protein (NitR) under the control of the inducible promoter PnitA and a separate cistron with a second PnitA promoter followed by a multiple cloning site (MCS). The native ribosome binding site of the nitR gene was preserved in the MCS. This arrangement of regulatory elements creates a positive-feedback loop in which the expression of both the regulatory protein and the target gene are simultaneously induced (Fig  2A). The inductor is the nitrile derivative ε-caprolactam [36]. This cassette, called NIT-1, was synthesized for this work by Invitrogen. The pNV119 derived vector containing this cassette was called pNVNIT (Fig 2A). R. ruber was successfully transformed with this vector, providing a way to check its potential as a functional expression vector for this strain.
Similarly to M. smegmatis mc2155 [5], we were not able to identify any 17β-HSD homologous enzyme in R. ruber Chol-4 by blasting the protein sequence against its genome (Gen-Bank::NZ_ANGC00000000.2). Therefore, a synthetic variant of the fungal enzyme 17β-HSD codon-optimized for actinobacteria and efficient in the actinobacteria Mycobacterium [5] was chosen to be cloned and overexpressed in R. ruber.

Expression of 17β-HSD recombinant protein and AD biotransformation
The wild-type (WT) and the knockout mutant strain (ΔkshB-kstD1,2,3) were transformed with two plasmids: pNVNIT (as control, Fig 2B) and pNVNIT-17βHSD expression vector (Fig 2C). Recombinants harboring the plasmids were used for a growing-cell biotransformation study. As a first approach to detect testosterone, a culture of each strain (wild-type and mutant strain transformed with both plasmids) were grown on LB. After 16-24 hours (0.8-1.0 DO 600nm ), the ε-caprolactam inductor was added and 24 hours later, 1 mg/mL of AD was added in powder form. Samples were taken at different times for analysis by Thin-layer chromatography (TLC). As it is shown in Fig 3A, line 1, the wild-type strain, harboring the pNVNIT plasmid, progressively converted the AD into ADD, as its metabolic machinery is kept intact. The quadruple R. ruber mutant, also containing the pNVNIT plasmid, but with the steroid catabolic pathway impaired, was unable to metabolize AD ( Fig 3B, line 1).
In the case of the quadruple mutant harboring the pNVNIT-17βHSD vector, the biotransformation of AD to TS can only be seen 6 hours after the addition of AD.
The wild-type strain expressing 17βHSD does not produce any TS detectable by TLC ( Fig  3). The WT strain can deploy its complete catabolic system to consume the AD entering into the cell. Therefore, it could be expected that the metabolic flow goes toward AD degradation rather than testosterone production. Moreover, if testosterone was produced, it could soon be degraded by the intact cell machinery and by no means detected by TLC or HPLC.
The TLC analysis of the quadruple mutant also showed two spots: one with the same RF than the ADD standard and another an unknown product.The presence of the last one in the TLC might be a LB medium metabolite since the spot disappears when the biotransformation takes place in minimal media. The product that was expected to be ADD by TLC (Fig 3B) was not found by HPLC, and therefore it is not ADD. The presense of this product could be explained due to the action of other enzymes different to KstD1,2,3 that might function as a ketosteroid dehydrogenase in R. ruber and that could alternatively yield an additional product. Minimum medium instead of LB medium was used for the rest of the experiments in order to avoid contamination with undesired metabolites.
These results suggest that the pNVNIT can be an excellent expression vector for R. ruber, a vector that can broaden the use of this bacteria for biotransformation processes.
The ability to produce TS from AD was verified by HPLC in the recombinant R. ruber cells supplemented with 1 mM AD (Fig 4). TS production observed was 410 μM 24 hours after the ε-caprolactam induction of the culture with a molar conversion rate of AD to TS of 48.2 ± 3.9%.
On the other hand, the activity of the enzyme 17β-HSD involves a nicotinamide cofactor [26,37,38]. The ratio between the oxidized and the reduced form of this type of cofactors plays a crucial role in microbial redox reactions and energy metabolism. The regeneration of this ratio is an important step to be taken into account in biocatalysis [39]. Moreover, in vivo, this kind of enzymes displayed a directional preference that depends on the relative affinity for nicotinamide cofactors [NAD(P)(H)] and existing cofactor gradients. For instance, ketosteroid reduction could be favored by keeping the NADPH/NADP + ratio high [37,38]. Therefore, the  ratios of NAD + :NADH or NADP + :NADPH would encourage the production of TS from steroids rather than the steroid concentration itself [37].
To restore the metabolic cofactor ratio, a supplemental carbon source must be added during the biotransformation studies [38]. Usually, cofactor regeneration is performed using glucose as a co-substrate because of its easy availability and low cost [40] and therefore, the effect of glucose was tested in the recombinant strain of this study.
The time course of TS production from AD was analyzed with and without glucose in the medium (Fig 5). By adding glucose 48h after the inoculation, cell density increased from 2.5 ± 0.3 to 8.6 ± 0.5. This fact raises the molar conversion rate of AD to TS after 24 hours of bioconversion from 48.2 ± 3.9% to 61.5 ± 3.0%. The cell culture grown with glucose in the medium, kept AD conversion constant over time in comparison with the culture grown without glucose (Fig 5).
This result is consistent with the effects of the supplemented carbon source detected in TS production in other strains [41][42][43]. However, the increase in the molar converstion rate at 24 hours was not as high as the one reported in Mycobacterium resting-cells (from around 35% in the absence of glucose to 90% molar conversion rate in a medium supplemented with glucose) [5] reinforcing the idea that this reaction could depend on the specific metabolic context of the cell.
The maximum conversion rate of 61% obtained for the recombinant R. ruber is consistent with the conversion range from AD to TS described in growing cells which vary from 27% (Saccharomyces cerevisiae [44]) to 93% (Zygowilliopsis sp. WY7905 [43]). TS productivities in other bacterial cell-factories are shown on S2 Table. Optimization of this process in the recombinant R. ruber to obtain a better yield will need a more detailed study.
Lastly, we looked where the testosterone produced by this biotransformation process remained in the cell. After the biotransformation experiment, recombinant R. ruber cells were centrifuged and pellet and supernatant were separately prepared and analyzed by TLC and HPLC (Fig 6). After 3 days of biotransformation, 91% of the testosterone appeared outside the cell, indicating that the recombinant strain whose testosterone catabolic pathway is blocked due to the lack of KstD and Ksh activities, could get rid of the compound to avoid its intracellular accumulation. After 5 days of biotransformation, this percentage decreases to 75%. The reduction of the testosterone found extracellularly from 91% to 75% is not easy to explain and requires further studies. It could be possible that providing enough time, a putative induction of testosterone-transporter gene expression and an increased re-uptake of testosterone could occur in the cell and in this way favor the entry of TS from the extracellular medium.
The fact that most of the product stayed in the supernatant within the first three days of the assay shows that there is an easy way to recover the TS after a biotransformation process. Some further improvements that need to be optimized may include the rational design of the enzyme itself to obtain a better testosterone yield. Some attempts to modify the fungal enzyme have been made [25] but they have not been tested so far in a biofactory system. Further enzymatic engineering of this activity together with an in-depth study via system biology to optimize the metabolic state of the cell would help to improve the whole production process.

Conclusions
There is a great interest in the reduction of 17-oxosteroids to 17β-hydroxysteroids as an important way of preparing many steroidal drugs and valuable intermediates. In this study, R. ruber has been metabolically engineered to effectively convert AD to testosterone as proof of pNVNIT-17βHSD (line 3 to 6). (B) TLC of the ΔkshB-kstD1,2,3 mutant harboring the pNVNIT-17βHSD plasmid (line 3 to 6). The control (line 1) in Figs A) and B) belongs to the strains WT or ΔkshB-kstD1,2,3 harboring the empty plasmid pNVNIT at 48 hours after the AD addition. M (line 2): markers AD, ADD, TS (testosterone) and 9OH-AD (1 μg/μL).
https://doi.org/10.1371/journal.pone.0220492.g003 its potential as a cellular factory. Genes involved in steroid ring breakage were knocked out to prevent AD and testosterone catabolism and the 17β-HSD enzyme from Cochliobolus lunatus was overexpressed using an expression vector specifically designed for R. ruber. The recombinant strain produces a yield of 61% testosterone from AD after the biotransformation studies. 91% of the testosterone was recovered extracellularly, broadening the chances of this strain to function as a steroid factory. The whole process still requires optimizing to achieve higher conversion yields, but this work validates the promising use of R. ruber for the biotechnological production of steroids such as testosterone.

Chemicals
AD was kindly given by Gadea Pharmaceutical Group. Chloroform, methanol and phenol of HPLC quality, were supplied by Scharlab S.L. Progesterone, testosterone, sodium acetate,  ΔkshB-kstD1,2,3 R. ruber harboring the pNVNIT-17βHSD construction was grown in minimal medium with sodium acetate as carbon source, at 30˚C and 250 rpm until 0.8 OD 600nm was reached. After 24 hours of the NIT-1 regulon induction (ε-caprolactam at 28 mM) the biotransformation started by adding AD at 1 mM with or without 1% w/v glucose. Samples were taken at different times, extracted by chloroform and analyzed by HPLC. ΔkshB-kstD1,2,3 R. ruber harboring the pNVNIT vector as control were also tested and no testosterone was detected (not shown).Testosterone and AD concentration were determined as indicated in Material and methods. Average and standard deviation of three biological replicates are shown.

Bacterial strains, culture conditions and DNA manipulation
The bacterial strains and plasmids used in this work are listed in Table 1. Escherichia coli cells were grown in Luria Bertani (LB) broth in an orbital shaker, at 250 rpm [45] or on LB plates containing the appropriate antibiotics at 37˚C. R. ruber and the mutant strains were routinely grown on LB or minimal medium (Medium 457 of the DSMZ, Braunschweig, Germany) containing the desired carbon and energy source under aerobic conditions at 30˚C in a rotary shaker (250 rpm) for 1-3 days. When necessary, antibiotics were added to the medium at 15 μg/mL nalidixic acid and at 50 μg/mL or 200 μg/mL kanamycin for E. coli or R. ruber respectively.
For the biotransformation experiments, an LB pre-grown cultures were washed twice with minimal medium prior to inoculation to 10 mL of fresh minimal medium (initial 0.05 DO 600nm ) supplemented with sodium acetate at 24 mM as the only energy and carbon source.
Competent and electrocompetent cells of E. coli were prepared and transformed as previously described [45]. Selection of transformed cells was carried out in LB agar plates supplemented with the appropriate antibiotics.
Electroporation of 200 μL of R. ruber cells was made with 1 μg DNA at 400 O, 25 mA, 2.5 μF, 10-11 milliseconds; the resulting cells were suspended in 800 μL of LB and kept for 6 min at 46˚C, and then for 5 hours at 30˚C without shaking. They were finally plated on LB Agar with 200 μg/mL kanamycin, 15 μg/mL nalidixic acid and kept at 30˚C.
The verification of the plasmid transformation of the R. ruber strains was made in several steps: first, plasmids were extracted from Rhodococcus; second, E. coli was transformed with the plasmid preparation and lastly, E. coli transformation was checked by standard methods.
All DNA manipulations were performed according to standard molecular cloning procedures [45] or following manufacturers' instructions (NZYMiniprep and NZYGelpure from NZYtech). DNA sequencing was performed with an ABI Prism 377 automated DNA sequencer (Applied Biosystems Inc.) at Secugen S.L. (Madrid, Spain).

Construction of the R. ruber ΔkshB-kstD1,2,3 mutant strain
The kshB gene was chromosomally deleted for this work using the R. ruber triple mutant ΔkstD1,2,3 as host and following the unmarked gene deletion method previously described [8] using the plasmid pK18(kshBU+D) [9] and the E. coli strain S17-1 ( Table 1). The construction of the deleted strain was verified by PCR using the primers CH564 (5'-CGCGTCTCTCCTGATGTGTCGG) and CH331 (5'-ACGTAGCCTGCCTCGAT GTCC), at Tm 55˚C, 1.5 min and 30 cycles, and further DNA sequencing. In all cases, the Expand High-Fidelity Taq DNA Polymerase from Roche was used for PCR reactions in a Mastercycler personal (Eppendorf). Restriction enzymes and DNA modifying enzymes were purchased to Takara Bio Inc. and New England Biolabs (UK).

Construction of an expression plasmid for R. ruber and 17β-hsd gene cloning
The Nocardial high-copy-number cloning vector pNV119 [48] was modified to be used as a R. ruber expression vector. The NIT-1 cassette (GenBank: FJ173069), an artificial bacterial regulon developed for mycobacteria [36], was synthesized (Invitrogen) and cloned into the KpnI-SphI restriction sites of the pNV119 plasmid. This regulon can use an inexpensive and nontoxic nitrile analog, ε-caprolactam, as an inducer [36]. The resulting plasmid pNVNIT was checked by restriction analysis and sequenced.
On the other hand, the 17β-hsd gene of C. lunatus with an optimized codon usage for Mycbacterium expression was obtained from digestion of the plasmid pUC57-17HSD [5] with NdeI-Hin-cII restriction enzymes and cloned into the pNVNIT expression plasmid previously digested with NedI-DraI yielding pNVNIT-17βHSD. This plasmid and a control vector without the 17β-hsd gene were used to electroporate both R. ruber wild-type and the ΔkshB-kstD1,2,3 mutant.

Monitoring of AD biotransformation into testosterone by analytical methods
The ΔkshB-kstD1,2,3 strain and R. ruber wild type were electroporated with either the empty plasmid pNVNIT as control or the plasmid harboring the 17β-hsd gen of C. lunatus (pNVNIT-17βHSD). The recombinant strains were grown in 10 mL of LB medium or minimal medium (MM) with sodium acetate (24 mM) as carbon and energy source with 200 μg/mL kanamycin and 15 μg/mL nalidixic acid, at 30˚C in a rotary shaker (250 rpm). Cells were grown up to 0.8 OD 600nm , then induced with 28 mM ε-caprolactam for 24h; at this point the cultures were in stationary phase with an OD 600nm 2.5 ± 0.3 (MM) or OD 600nm 9.0 ± 1.0 (LB). Afterwards, 1 mg/mL AD (LB cultures) or 1 mM AD with or without 1% (w/v) glucose (MM cultures) was added. Due to the low solubility of this steroid, the 1 mM AD was prepared from a 10 mM stock dissolved by sonication in 10% (w/v) tyloxapol. The stock solution was autoclaved and kept at room temperature prior to its addition to the minimal medium. Aliquots of 500 μL of the cell culture were taken at fixed times up to 120 hours.
For the analytical studies, the lipid fraction of the samples was obtained by double extraction with 1 mL chloroform and left to dry at 65˚C. Two analytical methods were used: the first approach by TLC and a second one by HPLC.
The AD in the TLC experiment was added as powder because the presence of cyclodextrins or tyloxapol damages the TLC plate. 50 μL of chloroform were added to every dried sample and an aliquot of 5 μL was applied onto aluminium TLC Silica gel 60 F254 sheets (Merck). Chromatography was performed using chloroform:acetone (9:1 v/v) as solvent and spots were revealed by UV exposure (λ 254nm ). Afterwards, the TLC plate was dunked into a sulphuric acid:methanol solution (1:9 v/v) followed by a drying step with warm air and heating 1 min at 100˚C. 1 μg of standard control samples (testosterone, AD, ADD, and 9OH-AD) were also included in the analysis.
For the reverse-phase HPLC analysis, samples were resuspended in 600 μL chloroform and filtered. Steroids were separated on a Teknokroma mediterranea TM Sea 18 column (15 cm x 0.46 cm; 5 μM) and UV detected at 245 nm at room temperature. The mobile phase was composed of methanol and water (70/30 v/v) at a flow rate of 1 mL/min. Progesterone was used as an extraction control and phenol as an internal standard on HPLC. AD, ADD, progesterone, and testosterone were used as steroid standards. The conversion rate of TS was calculated on the basis of AD measured into the sample in the resting-cell biotransformations.