|Home | About | Journals | Submit | Contact Us | Français|
Rhodoquinone (RQ) is an important cofactor used in the anaerobic energy metabolism of Rhodospirillum rubrum. RQ is structurally similar to ubiquinone (coenzyme Q or Q), a polyprenylated benzoquinone used in the aerobic respiratory chain. RQ is also found in several eukaryotic species that utilize a fumarate reductase pathway for anaerobic respiration, an important example being the parasitic helminths. RQ is not found in humans or other mammals, and therefore inhibition of its biosynthesis may provide a parasite-specific drug target. In this report, we describe several in vivo feeding experiments with R. rubrum used for the identification of RQ biosynthetic intermediates. Cultures of R. rubrum were grown in the presence of synthetic analogs of ubiquinone and the known Q biosynthetic precursors demethylubiquinone, demethoxyubiquinone, and demethyldemethoxyubiquinone, and assays were monitored for the formation of RQ3. Data from time course experiments and S-adenosyl-l-methionine-dependent O-methyltransferase inhibition studies are discussed. Based on the results presented, we have demonstrated that Q is a required intermediate for the biosynthesis of RQ in R. rubrum.
Rhodospirillum rubrum is a well-characterized and metabolically diverse member of the family of purple nonsulfur bacteria (29, 61). R. rubrum is typically found in aquatic environments and can adapt to a variety of growth conditions by using photosynthesis, respiration, or fermentation pathways (28, 70). In the light, R. rubrum exhibits photoheterotrophic growth using organic substrates or photoautotrophic growth using CO2 and H2 (15, 70). In the dark, R. rubrum can utilize either aerobic respiration (70, 73) or anaerobic respiration with a fumarate reduction pathway or with nonfermentable substrates in the presence of oxidants such as dimethyl sulfoxide (DMSO) or trimethylamine oxide (15, 58, 73). R. rubrum can also grow anaerobically in the dark by fermentation of sugars in the presence of bicarbonate (58). The focus of this work was the biosynthesis of quinones used by R. rubrum for aerobic and anaerobic respiration.
Rhodoquinone (RQ; compound 1 in Fig. Fig.1)1) is an aminoquinone structurally similar to ubiquinone (coenzyme Q or Q [compound 2]) (44); however, the two differ considerably in redox potential (that of RQ is −63 mV, and that of Q is +100 mV) (2). Both RQ and Q have a fully substituted benzoquinone ring and a polyisoprenoid side chain that varies in length (depending on the species; see Fig. Fig.11 for examples). The only difference between the structures is that RQ has an amino substituent (NH2) instead of a methoxy substituent (OCH3) on the quinone ring. While Q is a ubiquitous lipid component involved in aerobic respiratory electron transport (9, 36, 60), RQ functions in anaerobic respiration in R. rubrum (19) and in several other phototrophic purple bacteria (21, 22, 41) and is also present in a few aerobic chemotrophic bacteria, including Brachymonas denitrificans and Zoogloea ramigera (23). In these varied species of bacteria, RQ has been proposed to function in fumarate reduction to maintain NAD+/NADH redox balance, either during photosynthetic anaerobic metabolism (12, 15-18, 64) or in chemotrophic metabolism when the availability of oxygen as a terminal oxidant is limiting (23). Another recent finding is that RQH2 is capable of inducing Q-cycle bypass reactions in the cytochrome bc1 complex in Saccharomyces cerevisiae, resulting in superoxide formation (7). If RQ/RQH2 coexists in the cytoplasmic membrane with Q/QH2 in R. rubrum, it might serve as both a substrate for and an inhibitor of the bc1 complex (47).
RQ is also found in the mitochondrial membrane of eukaryotic species capable of fumarate reduction, such as the flagellate Euglena gracilis (25, 53), the free-living nematode Caenorhabditis elegans (62), and the parasitic helminths (65, 66, 68, 72). Similar to R. rubrum, these species can adapt their metabolism to both aerobic and anaerobic conditions throughout their life cycle. For example, most adult parasitic species (e.g., Ascaris suum, Fasciola hepatica, and Haemonchus contortus) rely heavily on fumarate reduction for their energy generation while inside a host organism, where the oxygen tension is very low (30, 65, 72). Under these conditions, the biosynthesis of RQ is upregulated; however, during free-living stages of their life cycle, the helminth parasites use primarily aerobic respiration, which requires Q (30, 65, 72). The anaerobic energy metabolism of the helminthes has been reviewed (63, 67). Humans and other mammalian hosts use Q for aerobic energy metabolism but do not produce or require RQ; therefore, selective inhibition of RQ biosynthesis may lead to highly specific antihelminthic drugs that do not have a toxic effect on the host (35, 48).
R. rubrum is an excellent facultative model system for the study of RQ biosynthesis. The complete genome of R. rubrum has recently been sequenced by the Department of Energy Joint Genome Institute, finished by the Los Alamos Finishing Group, and further validated by optical mapping (57). The 16S rRNA sequence of R. rubrum is highly homologous to cognate eukaryotic mitochondrial sequences (46). Due to the similarities in structure, the biosynthetic pathways of RQ and Q have been proposed to diverge from a common precursor (67). Proposed pathways for RQ biosynthesis (A to D), in conjunction with the known steps in Q biosynthesis, are outlined in Fig. Fig.11 (31, 34, 60). Parson and Rudney previously showed that when R. rubrum was grown anaerobically in the light in the presence of [U-14C]p-hydroxybenzoate, 14C was incorporated into both Q10 and RQ10 (50). In their growth experiments, the specific activity of Q10 was measured at its maximal value 15 h after inoculation and then began to decrease. However, the specific activity of RQ10 continued to increase for 40 h before declining. These results suggested that Q10 was a biosynthetic precursor of RQ10, although this was not directly demonstrated using radiolabeled Q10; hence, the possibility remained that the labeled RQ10 was derived from another radiolabeled lipid species. We have done this feeding experiment with a synthetic analog of Q where n = 3 (Q3) and monitored for the production of RQ3. The synthesis and use of farnesylated quinone and aromatic intermediates for characterization of the Q biosynthetic pathway in S. cerevisiae and Escherichia coli has been well documented (4, 5, 38, 52, 59). The other proposed precursors of RQ shown in Fig. Fig.11 were also fed to R. rubrum, and the lipid extracts from these assays were analyzed for the presence of RQ3, i.e., demethyldemethoxyubiquinone-3 (DDMQ3; compound 9), demethoxyubiquinone-3 (DMQ3; compound 10), and demethylubiquinone-3 (DMeQ3; compound 11).
In S. cerevisiae and E. coli, the last O-methylation step in Q biosynthesis is catalyzed by the S-adenosyl-l-methionine (SAM)-dependent methyltransferases Coq3 and UbiG, respectively (26, 52); this final methylation step converts DMeQ to Q. Using the NCBI Basic Local Alignment Search Tool, an O-methyltransferase (GeneID no. 3834724 Rru_A0742) that had 41% and 59% sequence identity with Coq3 and UbiG, respectively, was identified in R. rubrum. S-Adenosyl-l-homocysteine (SAH) is a well-known inhibitor of SAM-dependent methyltransferases (13, 24). Because SAH is the transmethylation by-product of SAM-dependent methyltransferases, it is not readily taken up by cells and must be generated in vivo (24). SAH can be produced in vivo from S-adenosine and l-homocysteine thiolactone by endogenous SAH hydrolase (SAHH) (37, 71). A search of the R. rubrum genome also confirmed the presence of a gene encoding SAHH (GeneID no. 3836896 Rru_A3444). It was proposed that if DMeQ is the immediate precursor of RQ, then SAH inhibition of the methyltransferase required for Q biosynthesis should have little effect on RQ production. Conversely, if Q is required for RQ synthesis, then inhibition of Q biosynthesis should have a significant effect on RQ production. Assays were designed to quantify the levels of RQ3 produced from DMeQ3 and Q3 in R. rubrum cultures at various concentrations of SAH.
Wild-type R. rubrum (ATCC 11170) was cultivated under lighted anaerobic conditions at 32°C in an Innova 4430 incubator shaker (New Brunswick Scientific, Edison, NJ) equipped with a full-spectrum fluorescence lamp with a light output of 365 lx (Verilux Full Spectrum F20T12VLX; Veriflux, Inc., Waitsfield, VT). Cultures were grown in screw-cap medium bottles (600 ml) filled to capacity for 4 to 6 days. The medium recipe was adapted from reference 49 and contained NH4Cl (37 mM), dl-malic acid (30 mM), yeast extract (2 g/liter; total nitrogen, 9.8%; amino nitrogen, 5.1%, amino N/total N ratio, 0.52; sodium chloride, 0.3%; Research Products International Corp.), morpholinepropanesulfonic acid buffer (0.5 g/liter), MgSO4 (1.4 mM), CaCl2 (0.7 mM), H3BO3 (0.5 mM), Na2EDTA · 2H2O (54 μM), biotin (1 μg/liter), and the trace elements FeSO4 · 7H2O (16 μM), Na2MoO4 · 2H2O (4.5 μM), MnCl2 (1.5 μM), ZnSO4 · 7H2O (0.5 μM), NiSO4 · 6H2O (0.1 μM), and CoCl2 · 6H2O (0.1 μM) solubilized with HCl (60 μM). The final pH was adjusted to 7 with NaOH, and prior to inoculation, KH2PO4 buffer (pH 7, 0.15 mM) was added. All of the values in parentheses are final concentrations.
All in vivo assays were prepared in 60-ml screw-cap glass centrifuge tubes in a Herasafe biological safety cabinet (Thermo Fisher Scientific, Waltham, MA) under aerobic conditions. Upon the addition of substrate and other assay components, tubes were immediately filled to capacity (with culture or medium) and sealed. Separate experiments using a Traceable Digital Oxygen Meter (accurate to ±0.4 mg/liter; Control Company, Friendswood, TX) showed that approximately 1.5 h was required for cultures of the same optical density (OD) to become anaerobic (<0.2 mg/liter oxygen) after sealing.
To quantify the biosynthesis of RQ3 over time, R. rubrum assays were prepared in duplicate using the substrates DMeQ3 and Q3, which were synthesized as previously reported (38, 52). Stock solutions of each substrate were prepared in absolute ethanol (Pharmco-Aaper, Brookfield, CT), and concentrations were determined using a UV/Vis spectrophotometer (Agilent 8453; Agilent Technologies, Foster City, CA) at 275 nm with molar extinction coefficients of 15,900 and 15,100 M−1 cm−1, respectively (36). Each substrate was added at a 1 μM final concentration to separate 650-ml R. rubrum cultures (each with a starting OD at 660 nm [OD660] of 1.7). Each 650-ml culture was divided into 10 60-ml tubes and filled to capacity (final volume per tube, 63 ml). The tubes were incubated at 32°C under tungsten illumination and harvested at 0, 4, 8, 12, and 24 h.
Control assay mixtures were prepared as described above, with RQ3 (1 μM) as the substrate, and incubated for 0, 4, 8, 12, and 24 h. RQ3 was synthesized as previously reported (7), and concentrations were determined in ethanol using a molar extinction coefficient of 10,300 M−1 cm−1 (36). Additional control assay mixtures containing a 1 μM final concentration of DMeQ3, Q3, or RQ3 were prepared with dead R. rubrum cells or no cells and incubated in the light at 32°C for 0 or 24 h. Dead cells were produced by heating for 50 min at 131°C in an autoclave (Tomy ES-315; Alfa Medical, Westbury, NY). Assay mixtures containing no substrate were also prepared using live R. rubrum culture and dead cells and incubated in the light at 32°C for 0 or 24 h.
To inhibit the methyltransferase involved in Q3 biosynthesis from DMeQ3, R. rubrum assay mixtures were prepared in duplicate using 60-ml screw-cap glass centrifuge tubes filled with 50 ml of bacterial culture (OD660 of 2.2). SAH was generated in vivo as follows. S-Adenosine (≥99% pure; Sigma-Aldrich, St. Louis, MO) was dissolved in DMSO (99.9% pure; molecular biology grade; Sigma-Aldrich, St. Louis, MO) at concentrations of 0.5, 0.1, and 0.02 M, and 0.6 ml of each was added to assay mixtures to give final concentrations of 5, 1, and 0.2 mM; the same volume of DMSO (0.6 ml) was added to the 0 mM SAH controls. dl-Homocysteine thiolactone hydrochloride (99% pure; Sigma-Aldrich, St. Louis, MO) was dissolved in water at 1.0, 0.2, and 0.04 M concentrations, and 0.6 ml of each was added to assay mixtures to give final concentrations of l-homocysteine thiolactone of 5, 1, and 0.2 mM. Prior to the addition of substrate, tubes were sealed and preincubated under lighted conditions for 10 min at 32°C to allow the generation of SAH. Following preincubation, an ethanolic solution of each substrate was added to give a final concentration of 1 μM DMeQ3 or 1 μM Q3; the same net volume of ethanol (125 μl) was added to all assay mixtures. Tubes were then filled to capacity with additional medium to ensure anaerobic growth. Controls were also prepared without DMeQ3 and Q3 at final SAH concentrations of 5 and 0 mM. All assay mixtures were incubated under lighted conditions at 32°C for 4 h.
Lipid extractions were performed on all R. rubrum assay mixtures (excluding the no-cell assay mixtures) to isolate the quinones for analysis. Once the assay mixture incubation period was complete, cultures were transferred to separate 250-ml plastic centrifuge bottles. Cells were harvested by centrifugation at 2,000 × g for 30 min at 4°C (Avanti J-E high-performance centrifuge with a JA-10 rotor; Beckman-Coulter, Fullerton, CA). Cell pellets were resuspended in 5 ml of deionized water, transferred to 10-ml glass centrifuge tubes, and collected by centrifugation at 2,000 × g for 30 min at 4°C (Dupont Sorvall RT 600B tabletop centrifuge with a Sorvall H100B rotor; Block Scientific, Bohemia, NY). After pouring off the supernatant, the tubes were flushed briefly with N2 gas and sealed. The masses of the wet pellets were obtained before they were stored at −85°C. Prior to extraction, the pellets were thawed and Q2 (0.5 pmol/10 μl injection volume; Sigma Aldrich, St. Louis, MO) was added as an internal standard. The pellets were extracted with 2 ml methanol (Optima grade; Fisher Scientific, Pittsburgh, PA)-2 ml petroleum ether (J. T. Baxter, Phillipsburg, NJ) containing 10 μM 2,6-di-tert-butyl-4-methylphenol (BHT; Sigma-Aldrich, St. Louis, MO) and 110 μl deionized water. After thorough vortex mixing, the suspensions were separated into layers by centrifugation at 1,000 × g for 5 min at 4°C. The etherial layers were transferred to 5-ml glass centrifuge tubes, and the methanol layers were reextracted with another 2 ml of petroleum ether solution (containing 10 μM BHT). To prevent decomposition, sample tubes were stored on ice during the extraction procedure and kept away from direct light. The combined etherial layers were dried under a steady stream of N2 gas. The resulting dark blue residue from each assay mixture was immediately resuspended using 20 μl of hexanes (Optima grade; Fisher Scientific, Pittsburgh, PA) and 80 μl of absolute ethanol for liquid chromatography (LC)-mass spectrometry (MS) analysis.
Due to the absence of a pellet, extractions were performed immediately following incubation using a 125-ml separatory funnel. The assay medium was transferred to the separatory funnel with 10 ml brine and 18 ml petroleum ether (containing 1 μM BHT). The ether layer was transferred to a 100-ml round-bottom flask and concentrated in vacuo using a rotary evaporation (Büchi Rotovapor R-205; Büchi Labortechnik, Flawil, Switzerland). A second extraction was performed using 18 ml of petroleum ether solution. The second ether layer was added to the residue in the 100-ml flask and concentrated as before. The combined residue was then resuspended in 4 ml of petroleum ether solution, transferred to a 5-ml glass centrifuge tube, and dried under N2 gas. The extracts were stored at −85°C and eventually resuspended in 20 μl hexanes and 80 μl ethanol for LC-MS analysis as in the other extractions.
LC-MS standards were prepared containing Q3 (0.05, 0.25, 1, 2.5, 5, or 10 pmol/10-μl injection), RQ3 (0.015, 0.075, 0.15, 0.75, 1.5, or 3 pmol/10-μl injection), and Q2 (0.5 pmol/10-μl injection) in absolute ethanol. Lipid extractions were performed on the standards following the same procedure used for the cell pellets.
LC-MS analysis was performed at the UCLA Molecular Instrumentation Center in Los Angeles, CA. Authentic standards were prepared in the high-performance liquid chromatography (HPLC) running buffer at 1 to 2 pmol/μl and infused to determine optimum conditions for quantifying transition ions by multiple-reaction monitoring (MRM). For Fig. Fig.2,2, the compounds were infused at 2 pmol/μl in 80:20, ethanol-hexanes (Optima grade; Fisher Scientific, Pittsburgh, PA) under the conditions described below, except that the collision energy was increased to produce the spectra as shown. The collision energy ranged from 35 to 55 V. Quinone separation was accomplished with an HPLC system (Agilent 1200 Binary Pump SL; Agilent Technologies, Foster City, CA) and an autosampler containing thermostat-controlled tray holders and sample stack (PAL system; LEAP Technologies, Carrboro, NC), which was maintained at 4°C. Chromatography was performed using a pentafluorophenyl propyl column [Luna PFP(2), 50 by 2.00 mm, 3 μm, 100 Å; Phenomenex, Torrance, CA]. Farnesylated and geranylated quinones were eluted between 1 and 3 min using an isocratic mobile phase of acetonitrile-water (7:3, vol/vol) containing formic acid (0.1%, vol/vol). The acetonitrile used was Optima LC-MS grade (Fisher Scientific, Pittsburgh, PA), the formic acid was >99% pure and packaged in sealed 1-ml ampoules (Thermo-Scientific Pierce Protein Research Products, Rockford, IL), and the water was doubly distilled. All HPLC runs used a flow rate of 0.5 ml/min and an injection volume of 10 μl. All injections were performed in duplicate. Quantitation was accomplished using a triple-quadrupole mass spectrometer with a Turbo electrospray ionization source in positive mode (AP 4000 QTrap; Applied Biosystems, Foster City, CA) with MRM of singly charged ions. Q1 and Q3 were operated at single-unit resolution. Analyst 1.4.1 software was used for data acquisition and processing. Linear slopes were calculated using peak areas with a bunching parameter of 3 and three smoothing functions. Standard error was determined at a 95% confidence interval with a value of n = 4. The following global conditions were used for MS/MS analysis of all compounds: dwell time, 100 ms; entrance potential, 10.00 V; curtain gas pressure, 20 lb/in2; nebulizer gas pressure, 50 lb/in2; turbo gas pressure, 60 lb/in2; collision gas, medium; nebulizer voltage, 20 V; temperature, 450°C. Nitrogen gas was used for all applications and was obtained from the boiloff from a bulk liquid nitrogen storage tank. Additional quinone-specific parameters are listed in Table Table11.
Figure Figure22 shows the mass spectra of all five of the farnesylated quinone substrates used in these experiments with their major tropylium product ions highlighted. All MRM analyses monitored for the mass transition from each quinone precursor ion ([M+H]+) to its respective tropylium product ion.
The DMeQ3 feeding assays revealed that R. rubrum is capable of converting DMeQ3 to both Q3 and RQ3 in vivo (Fig. (Fig.3A),3A), with the amount of Q3 produced being between 100 and 400 times greater than that of RQ3 from 24 to 4 h, respectively. A small amount of Q3 (0.07 ± 0.02 pmol/mg wet pellet weight) was already formed from DMeQ3 at the 0-h time point, since about 5 min of processing time was required for the transfer of cultures to centrifuge bottles; however, no RQ3 was detected. RQ3 was first detected after 4 h, when the amount of Q3 present had increased by about 36 times (2.5 ± 0.8 pmol/mg wet pellet weight). The amount of Q3 formed from DMeQ3 reached a maximal level after 12 h (8.3 ± 1.8 pmol/mg wet pellet weight), while the amount of RQ3 continued to increase throughout the 24-h growth period, reaching a maximum of 0.10 ± 0.01 pmol/mg wet pellet weight. Figure Figure3B3B shows a logarithmic comparison of the accumulation of Q3 versus RQ3 produced from DMeQ3 in 24 h.
Results from the Q3 feeding assays also showed the formation of RQ3 over time. As shown in Fig. Fig.4,4, there was a very small amount of RQ3 detected from Q3 at the 0-h time point (<1 fmol/mg wet pellet weight). The amount of Q3 recovered at the 0-h time point was about 2 orders of magnitude greater than the amount of Q3 that had formed from DMeQ3 at the same time point (data not shown). After 4 h, the amount of RQ3 increased to about twice that produced from the DMeQ3 assay in the same time period. However, by the 12-h time point, the amounts of RQ3 produced from DMeQ3 and Q3 were equivalent (within the standard error). The amount of Q3 recovered from the Q3 feeding assay after 12 h was roughly equal to the amount of Q3 formed from the DMeQ3 assay (~10 pmol/mg wet pellet weight; data not shown). Other, unknown, species with the same mass transition as DMeQ3 but with longer retention times were also detected in the Q3 assays, and the concentration of these species appeared to decrease over time (data not shown).
To test the reversibility of the conversions of DMeQ3 and Q3 to RQ3, a control experiment was performed with RQ3 as the substrate, and lipid extracts were analyzed for Q3 and DMeQ3 after 0, 4, 8, 12, and 24 h. As shown in Fig. 5, a small amount of Q3 (0.03 ± 0.01 pmol/mg wet pellet weight) was present at the 0-h time point; however, the amount decreased over time. Another unknown species with the same mass transition as Q3 but with a shorter retention time was also detected (data not shown); however, no DMeQ3 was detected in these assays. Control experiments were also performed under the same growth conditions and with the same concentrations of DMeQ3, Q3, and RQ3 with dead R. rubrum cells and with growth medium (no cells) to monitor for degradation or nonenzymatic conversion to RQ3, Q3, or DMeQ3. Assays were performed for 0 and 24 h, and lipid extracts were analyzed for the presence of farnesylated products. No RQ3 was detected from DMeQ3 or Q3 in the control assays (data not shown). However, there was partial degradation of Q3 to DMeQ3 after exposure to light for 24 h in the no-cell assays; this degradation was not observed in the dead-cell assays (due to the absence of a pellet, the data are not directly comparable to those obtained from assays with live cultures). A small amount of Q3 was detected from the DMeQ3 and RQ3 no-cell assays; however, Q3 was not detected in the dead-cell assays (again, due to the absence of a pellet, data are not directly comparable to those obtained from assays with live cultures). No farnesylated products were observed in R. rubrum control assays in the absence of synthetic substrates.
Feeding assays were also performed using a 5 μM starting concentration of each of the proposed precursors DDMQ3 (compound 9) and DMQ3 (compound 10). No RQ3 was detected from either assay after 4 days of growth (data not shown). However, a small amount of Q3 was detected in both assays (~0.02 pmol/mg wet pellet weight), and DMQ3 was also formed from DDMQ3 (0.2 pmol/mg wet pellet weight). Since the starting concentrations of DDMQ3 and DMQ3 in these experiments were five times greater than in the reported DMeQ3 and Q3 assays, the data are not directly comparable.
The results presented here suggest that R. rubrum can synthesize RQ3 when provided either Q3 or DMeQ3. To investigate whether RQ3 could be produced directly from DMeQ3 in the absence of Q3, S-adenosine and l-homocysteine thiolactone were added to incubation mixtures in order to generate SAH, a competitive inhibitor of the O-methyltransferase-mediated conversion of DMeQ3 to Q3 (Fig. (Fig.1).1). Experiments were first performed to optimize the range of concentrations of SAH necessary for partial and complete inhibition of the O-methylation reaction of DMeQ3 to form Q3 (0.2 to 5 mM SAH). The average yield of Q3 from DMeQ3 was reduced by 85% with 0.2 mM SAH compared to the control (0 mM SAH), and O-methyltransferase inhibition was greater than 99.9% with 1 and 5 mM SAH (Fig. (Fig.6).6). A similar trend was observed with the amount of RQ3 formed from DMeQ3; the average yield of RQ3 was reduced by 89% with 0.2 mM SAH, and no RQ3 was observed with either 1 or 5 mM SAH (Fig. (Fig.66).
Since SAH was expected to inhibit other SAM-dependent methyltransferases that may indirectly have an effect on RQ3 biosynthesis, another control experiment was performed. Q3 was used as the substrate, and the production of RQ3 was monitored at the same three concentrations of SAH. Under 0.2 mM SAH inhibition conditions, the amount of RQ3 produced from Q3 was reduced by 31% compared to that produced by the 0 mM SAH control. The average yield of RQ3 was reduced by 81% at 1 mM SAH and by 91% at 5 mM SAH. A comparison of the levels of RQ3 produced from DMeQ3 versus Q3 at various concentrations of SAH is shown in Fig. Fig.7.7. Even though the yields were reduced, RQ3 was produced under all of the assay conditions with Q3 as the substrate.
The time course studies presented here show that DMeQ3 is a precursor of both Q3 and RQ3 in R. rubrum. However, formation of RQ3 was not detected in assays from the other proposed Q biosynthetic intermediates (compounds 9 and 10). The results from the SAH methyltransferase inhibition assays eliminated DMeQ3 as the immediate precursor of RQ3. These assays indicated that no RQ3 is produced from DMeQ3 in the absence of Q3. This observation further supports the results from the time course assay where Q3 from DMeQ3 was detected prior to that from RQ3. Our results favor pathway D for the biosynthesis of RQ (Fig. (Fig.1);1); however, this transformation may occur in more than one step.
The synthetic transformation of Q10 to RQ10 (and iso-RQ10) has been observed by treatment of Q10 with NH4OH in diethyl ether and ethanol (11, 42). It is therefore conceivable that this conjugate addition/elimination reaction could also occur in vivo. However, a methoxyquinone is not a typical substrate for transamination by an amido- or aminotransferase; instead, an alcohol or ketone functional group is usually required, respectively (51, 56). It is possible that there is another unstable intermediate formed during the interconversion of Q to RQ. Using the more sensitive technique of LC-MS with MRM analysis, we have observed several unknown farnesylated products that were generated from the Q3 feeding assays that have the same MRM transition as DMeQ3 (373.2 > 183.2 m/z) but with longer retention times than the standard (1.8 to 2.0 min versus 1.64 min). We have yet to fully characterize these compounds, as they degrade rapidly and the sample sizes are too small for other methods of structural elucidation, such as nuclear magnetic resonance analysis. In order for these unknown compounds to possess the same mass transition as DMeQ3, they can differ only in the regiochemistry of substituents or in oxidation state. Photodegradation of Q10 to DMeQ10 (and iso-DMeQ10) has been previously reported (27, 43), and this analogous degradation was observed in our no-cell time course assays with Q3. In the Q3 no-cell assays, DMeQ3 was detected by MS (presumably both the 2- and 3-hydroxyquinone regioisomers) with the same retention time as the standard. A small amount of Q3 was also detected from RQ3 and DMeQ3 in the no-cell assays. It is unlikely that this photodegradation occurs in vivo, as no degradation was observed in the dead-cell assays. In the presence of cells, most light is absorbed by pigments from the light-harvesting antennae of phototrophic bacteria (15, 70).
A possible structure of one of the unknown compounds formed in the Q3 assay is the unstable orthoquinone of DMeQ3 (o-DMeQ3, compound 14), which could participate in a transamination with a PLP-dependent aminotransferase to form RQ3 (Fig. (Fig.8).8). Interestingly, these unknown compounds are not detected in the DMeQ3 assays, where, for example, a direct tautomerization to the orthoquinone seems more feasible. We are currently investigating the synthesis of the proposed tautomer o-DMeQ3 for use as a standard to further characterize these unknown products. A possible method for generation of the orthoquinone in vivo is through an O-demethylation reaction of the Q hydroquinone (QH2, compound 12), followed by oxidation of the catechol product (compound 13). This is perhaps more plausible than a nonenzymatic degradation process. It has been shown that degradation of Q often involves destruction of the polyprenyl tail (6, 10). In our experiments, we have found that the farnesyl tail remains intact and only selective modifications of the head group are observed. Furthermore, as demonstrated in our feeding control experiments, R. rubrum cannot produce farnesylated quinones de novo.
Aromatic O-demethylases have been characterized in several species, such as Pseudomonas maltophilia (8, 20), Moorella thermoacetica (45), and Sphingomonas paucimobilis (1). The O-demethylation reactions have been proposed to occur through several different pathways. The dicamba O-demethylase from P. maltophilia was recently reported to consist of a three-component enzyme complex involving a reductase, a ferredoxin, and an oxygenase (8, 20). The vanillate O-demethylase from the anaerobic bacterium M. thermoacetica has also been reported to contain three components and is tetrahydrofolate dependent (45), as is the O-demethylase of S. paucimobilis (1). Other O-demethylation reactions have been reported that involve a cytochrome P450-dependent monooxygenase (69) or peroxidase (39). A blastp screening of the R. rubrum genome identified candidates with sequence similarities to the vanillate O-demethylase oxidoreductase (YP_425621.1, expect value of 9 × 10−5) and the vanillate O-demethylase oxygenase subunit (YP_426436.1, expect value of 7 × 10−2). It is possible that an O-demethylase required for demethylation of Q is part of an enzyme complex that acts in concert with an aminotransferase, therefore not requiring DMeQ as a direct substrate.
As proposed in Fig. Fig.8,8, it is conceivable that reversible interconversion between intermediates is possible and could be regulated by environmental conditions (e.g., oxygen availability). It has been shown in the helminth parasites that the amount of RQ present is dependent on whether or not the parasite is living inside the host organism (35, 67). A similar phenomenon is observed with R. rubrum. When the bacteria are grown aerobically, the levels of RQ10 are substantially lower than the levels of Q10; however, when R. rubrum is grown anaerobically, the amount of RQ10 surpasses that of Q10 (54, 55). To investigate the possible reversibility of the interconversion between Q and RQ, RQ3 was fed to R. rubrum in a control experiment and monitored for the production of Q3. This experiment did, in fact, show the production of Q3, as well as a new compound with a slightly shorter retention time but with the same mass transition as Q3. The level of Q3 decreased over time, possibly due to the decrease in available oxygen. Assays were prepared under aerobic conditions and required about 1.5 h to become fully anaerobic. It is possible that Q3 was converted back to RQ3 due to the reduction in oxygen. Further experiments with controlled oxygen levels must be performed to investigate the dependence of RQ biosynthesis on oxygen availability.
It has been reported by Van Hellemond et al. that F. hepatica and Schistosoma mansoni synthesize both Q10 and RQ10 de novo (66, 68). These conclusions are based on the different number of isoprenoid units in the tail between the host's quinone (Q9) and that of the parasite (Q10 and RQ10). If Q is a precursor of RQ, the authors expected that RQ9 would be detected from catabolism of the host's Q9, and this was not observed. It was also demonstrated that F. hepatica could incorporate [2-14C]mevalonic acid into Q10 and RQ10, suggesting that each is synthesized de novo. The results reported by Van Hellemond et al. suggest that Q from the host is not catabolized by the parasites to form RQ; however, these experiments only address the host's Q supply. These results do not eliminate the possibility that the parasites are catabolizing their own supply of Q10. Because the tail lengths are the same, it is not possible to differentiate the two in this experiment. The authors also performed in vitro assays with Q9 but did not detect any RQ9; however, experimental conditions were not provided. It is possible that small quantities of RQ9 were not detected due to degradation or lack of sensitivity of the UV analysis method. We observed that low concentrations of RQ3 (1 to 10 nM) rapidly decomposed in the lipid matrix, and our analysis proved initially difficult. We were able to stabilize RQ3 in the lipid extracts using BHT during extraction (~400 μM final concentration in the LC-MS sample). Furthermore, the quantities of RQ3 generated from our in vivo feeding assays were in the low fmol/μl range and also required the sensitivity of a triple-quadrupole mass spectrometer with MRM for detection. We were unable to detect RQ3 from our in vivo assays using UV analysis.
Pathways for RQ biosynthesis may have evolved separately in eukaryotic and prokaryotic species. Evidence to support this hypothesis has come from experiments performed with the C. elegans clk-1 mutant, which exhibits slow developmental growth and behavior and has an increased life span compared to that of wild-type C. elegans (14). The clk-1 mutants are deficient in Q9 biosynthesis and require dietary Q from E. coli; in fact, the only detectable Q in the clk-1 mutants is Q8, the form produced by E. coli (32, 33, 40). E. coli does not make or require RQ, and no RQ8 is detected in the clk-1 mutant from the proposed catabolism of Q8. However, an interesting finding is that the clk-1 mutants are still capable of producing RQ9. In addition, these mutants produce increased levels of RQ9 compared to those produced by the wild-type species (32, 33). The clk-1 mutant has been found to accumulate the Q precursor DMQ9 (compound 10, Fig. Fig.2)2) in its mitochondria (32, 33). It is possible that an alternative hydroxylase is present in the worm that catalyzes the conversion of DMQ9 to DMeQ9 (or o-DMeQ9) under anaerobic conditions to allow the synthesis of RQ9. A precedent for bypass hydroxylase mechanisms has been shown in E. coli for the biosynthesis of Q under anaerobic conditions using mutants (ubiH, ubiF, and ubiB) blocked in hydroxylation reactions of the aerobic pathway (3). However, it is unlikely that the mutant worms are synthesizing RQ9 directly from Q9, since no Q9 is detected. In contrast, evidence suggesting that RQ biosynthetic pathways have evolved similarly in prokaryotes and eukaryotes has been reported by Powls and Hemming (53). It was shown with the single-celled eukaryote E. gracilis that the kinetics of labeling with [U-14C]p-hydroxybenzoic acid is consistent with a precursor-product relationship between Q9 and RQ9 (53). These results agree with our R. rubrum feeding experiments with farnesylated substrates, as well as the radiolabeling experiments performed by Parson and Rudney, where RQ10 appeared to be catabolized from Q10 (50).
In summary, we have clearly demonstrated that Q is a biosynthetic intermediate of RQ in R. rubrum. Results from our in vivo feeding experiments render it unlikely that RQ is derived directly from the products (DDMQ3 [compound 9], DMQ3 [compound 10], and DMeQ3 [compound 11]). Even though RQ was detected in assays from DMeQ3, it was not observed in the absence of Q3, as demonstrated in the SAH O-methyltransferase inhibition assays. Further characterization of pathway D (Fig. (Fig.1)1) is currently under way in our laboratories. Identification of new intermediates, the metabolic source of the amino group in RQ, and the enzyme which catalyzes the amino transfer, is the focus of our research. Complete characterization of the amination step may permit regulation of RQ and would provide a parasite-specific enzyme target for drug development.
This research was supported in part by grants to Gonzaga University from the Howard Hughes Medical Institute through the Undergraduate Science Education Program (award 52006297), the National Science Foundation CAREER Program (award CHE-0135091), and the CRIF-MU Program (award CHE-0741868) and grants to the University of California, Los Angeles, from the National Institutes of Health (award GM45952) and the National Center for Research Resources (grant S10RR024605).
Any opinions, findings, and conclusions or recommendations expressed in this material are ours and do not necessarily reflect the views of the National Science Foundation, NIH, or NCRR.
We thank Steve Clarke for advice on the SAH assays, Yuchen Shi for his assistance with lipid extraction and mass spectroscopy at UCLA, and Erin Dickson for her work in determining oxygen levels in R. rubrum cultures.
Published ahead of print on 20 November 2009.