Production of carotenoids from aromatics and pretreated lignocellulosic biomass by Novosphingobium aromaticivorans

The N. aromaticivorans DSM 12444 genome predicts that this bacterium contains genes that encode previously uncharacterized proteins with 52% to 74% amino acid sequence identity to known enzymes that lead to nostoxanthin production (Table S1). To test the prediction that N. aromaticivorans uses these previously uncharacterized gene products to produce carotenoids, we grew cells in the presence of vanillate and analyzed acetone:methanol (lipophilic) extracts of the cells by liquid chromatography mass spectrometry (LC-MS). Our analysis of these extracts from the parent N. aromaticivorans strain (12444Δ1879) was consistent with nostoxanthin being a major carotenoid (compound 1 in Fig. 2B): compound 1 had absorbance maxima at 453 and 480 nm (Fig. S2B) and an m/z peak of 600 (Fig. S2C), both characteristic of nostoxanthin (22, 23). Two other compounds in these extracts had slightly longer retention times than compound 1: compound 2 had absorbance maxima at 339, 446, and 474 nm (Fig. S2B) and m/z peaks of 600 and 639 (Fig. S2C), and compound 3 had absorbance maxima at 453 and 481 nm (Fig. S2B) and an m/z peak of 584 (Fig. S2C). Although we were not able to identify compound 2, the properties of compound 3 are consistent with caloxanthin (23), a predicted precursor to nostoxanthin (Fig. 1) known to be present in some sphingomonads that produce nostoxanthin (15, 23). However, commercial standards of pure nostoxanthin or caloxanthin were not available to estimate the compounds’ abundances or provide further proof of their chemical identity. Another compound detected in the N. aromaticivorans lipophilic extracts had the same retention time and absorbance spectrum as a commercial CoQ₁₀ standard (Fig. 2A and B; Fig. S1 and S2B), consistent with previous work showing that CoQ₁₀ is the major quinone in this bacterium (16).

To further test whether N. aromaticivorans uses the predicted carotenoid biosynthesis pathway (Fig. 1), we generated a set of mutants with in-frame deletions of genes predicted to encode proteins involved in the pathway. We grew these strains with vanillate as the sole carbon source and analyzed the compounds present in lipophilic extracts from these mutants, using commercial standards when available to aid in the identification and quantification of the lipophilic compounds.

Deletion of Saro_1814 (encoding a putative CrtB homolog) resulted in a strain (12444ΔcrtB) that formed nonpigmented colonies on solid media (Fig. 3A), as expected given the predicted role of this gene product in phytoene synthesis (Fig. 1). We also found that strain 12444ΔcrtB only contained CoQ₁₀ as a major lipophilic compound (Fig. 2C; Fig. S3). Deletion of Saro_1817 (encoding a putative CrtY homolog) resulted in a strain (12444ΔcrtY) that formed light pink colonies (Fig. 3A) and contained the predicted pathway intermediate lycopene (compound 4) and CoQ₁₀ in its lipophilic extract (Fig. 2D; Fig. S4). Deletion of Saro_0236 (encoding a putative CrtG homolog) resulted in a strain (12444ΔcrtG) that formed yellow colonies (Fig. 3A) and contained the predicted pathway intermediate zeaxanthin (compound 5) and CoQ₁₀ as the major components of its lipophilic extract (Fig. 2E; Fig. S5). Finally, to test if the pathway intermediate β-carotene could be accumulated, we constructed a strain (12444ΔcrtGZ) in which both Saro_0236 and Saro_1168 (encoding a putative CrtZ homolog were deleted, as deletion of Saro_1168 alone would be expected to also produce 2,2ʹ-dihydroxy-β-carotene (Fig. 1). Strain 12444ΔcrtGZ formed yellow colonies (Fig. 3A), and the main components of its lipophilic extract were β-carotene (compound 6) and CoQ₁₀ (Fig. 2F; Fig. S6).

In sum, the compounds present in the lipophilic extracts of each of these mutants were consistent with predictions from the annotated carotenoid biosynthetic pathway in the N. aromaticivorans genome (Fig. 1). These experiments illustrate that one or more gene deletions can result in N. aromaticivorans strains that produce industrially valuable carotenoids when grown in the presence of vanillate. Furthermore, the lipophilic extracts from all strains tested also contain the electron carrier CoQ₁₀.

Astaxanthin is a valuable carotenoid that is not predicted to be produced by N. aromaticivorans, since the genome of this organism lacks a crtW gene. While several bacteria, including some other members of the Sphingomonadales order [Sphingomonas astaxanthinifaciens (24) and Sphingomonas taxi ATCC 55669 (25)], naturally produce astaxanthin (26), none of these are known to metabolize aromatic compounds present in pretreated lignocellulosic biomass. To test if we could engineer N. aromaticivorans to produce astaxanthin, we placed a recombinant crtW gene from S. astaxanthinifaciens or S. taxi separately into the crtG locus of 12444ΔcrtG to generate strains 12444SastaW and 12444StaxiW, respectively. The difference in colony colors of the 12444ΔcrtG, 12444SastaW, and 12444StaxiW strains (Fig. 3) suggested that the insertion of each of these crtW genes into the N. aromaticivorans genome resulted in altered carotenoid profiles.

To test this hypothesis, we analyzed the lipophilic extracts from vanillate-grown 12444SastaW and 12444StaxiW cells. This analysis showed that 12444SastaW produced astaxanthin (Fig. 2G; compound 7), as well as small amounts of three other putative carotenoids (compounds 8, 9, 10) and CoQ₁₀ (Fig. 2G; Fig. S7). The lipophilic extract of 12444StaxiW also contained astaxanthin (compound 7) and CoQ₁₀ (Fig. 2H; Fig. S8), although its predominant carotenoid (compound 8, which is also a minor component of the 12444SastaW extract) appears to be adonixanthin, based on its measured mass (m/z peak = 583; Fig. S8C). The 12444StaxiW lipophilic extract also contained two additional compounds, one of them identified as zeaxanthin (compound 5; Fig. 2H; Fig. S8). From this, we conclude that introducing the CrtW protein from S. astaxanthinifaciens into N. aromaticivorans generates a strain that is more effective at accumulating astaxanthin than a strain using the CrtW protein from S. taxi. Therefore, in subsequent experiments, we used cells containing the S. astaxanthinifaciens crtW gene (strain 12444SastaW) as a chassis for an astaxanthin-producing strain of N. aromaticivorans.

The dissolved O₂ concentration of a culture can affect carotenoid levels in various organisms in different ways (27). For example, lower O₂ tensions have been shown to increase CoQ₁₀ production in some bacteria (21), while other microbes increase carotenoid production at high O₂ tensions presumably since carotenoids can provide protection against reactive oxygen species (27). Therefore, we tested whether bubbling N. aromaticivorans cultures with a gas containing 5%, 10%, or 21% O₂ would lead to significant changes in carotenoids and CoQ₁₀ levels when using vanillate as a carbon source.

Carotenoid levels of N. aromaticivorans strains grown at different O₂ tensions are shown in Fig. 4A through E. Nostoxanthin levels in 12444Δ1879, which are reported as HPLC peak area due to lack of a standard for quantification, showed only modest changes with oxygen [ranging from 1.0 to 2.5 × 10⁵ area units per mg dry cell weight (dcw)]. Under the same growth conditions, lycopene levels in 12444ΔcrtY ranged from 0.24 to 0.75 µg/mg dcw, zeaxanthin in 12444ΔcrtG ranged from 0.25 to 0.54 µg/mg dcw, β-carotene in 12444ΔcrtGZ ranged from 0.29 to 0.64 µg/mg dcw, and astaxanthin in 12444SastaW ranged from 0.12 to 0.25 µg/mg dcw. Single-factor ANOVA tests (alpha = 0.05) of carotenoid levels suggested that only strain 12444ΔcrtGZ contained significantly different carotenoid levels at the O₂ concentrations tested (Fig. 4A through E). CoQ₁₀ levels for each strain at different O₂ tensions (Fig. 4F) ranged from 0.08 to 0.2 µg/mg dcw. ANOVA tests also showed that, for each of the strains, there was no significant difference in CoQ₁₀ production at the three O₂ concentrations tested (Fig. 4F). Furthermore, ANOVA tests showed there was no significant difference in CoQ₁₀ production between the different strains at any of the individual O₂ concentrations. Therefore, we conclude that O₂ availability does not have a consistently significant impact on N. aromaticivorans carotenoid or CoQ₁₀ levels over the concentration range tested.

The above results showed that wild-type and engineered N. aromaticivorans strains accumulate carotenoids and CoQ₁₀ when grown on vanillate, an aromatic compound predicted to be found in deconstructed lignocellulosic biomass. However, we also wanted to test whether carotenoids and CoQ₁₀ could be generated when cells were grown on the mixture of aromatics directly obtained from plant biomass. We therefore grew several strains in a sorghum APL (28), which contains a mixture of aromatic monomers and other organics (Fig. S9; Table S2). With a few exceptions, the N. aromaticivorans strains grown in sorghum APL (in shake flasks with atmospheric O₂ conditions) produced amounts of carotenoids and CoQ₁₀ (normalized by dry cell weight; open squares in Fig. 4) that were within the production ranges of the vanillate-grown cultures grown at different dissolved O₂ tensions (filled circles in Fig. 4). One exception was strain 12444ΔcrtB, which lacks detectable carotenoids in lipophilic extracts; this strain produced significantly more CoQ₁₀ when grown in APL compared to when grown on vanillate (Fig. 4F). In addition, strains 12444Δ1879 and 12444SastaW produced significantly less nostoxanthin and astaxanthin, respectively, when grown in APL compared to when grown on vanillate. Overall, these results show that it is possible to use N. aromaticivorans for producing carotenoids from a mixture of aromatic and other organic compounds derived from pretreated lignocellulosic biomass.

Previous work has shown that engineered N. aromaticivorans strains containing defined mutations in aromatic metabolism can convert the three major classes of biomass aromatics (syringyl, guaiacyl, and p-hydroxyphenyl) into PDC and secrete it into the medium (2, 7). We sought to test whether individual N. aromaticivorans strains could produce both extracellular PDC and intracellular lipophilic compounds (carotenoids and CoQ₁₀) as valuable products from biomass-derived media. To do this, we generated a set of strains that contained both the mutations needed to accumulate extracellular PDC and those needed to accumulate the carotenoids zeaxanthin, β-carotene, or astaxanthin. We found that these engineered strains produce extracellular PDC, as well as the expected carotenoid and CoQ₁₀ when grown in sorghum APL (Fig. 5). The levels of individual carotenoid species and CoQ₁₀ (normalized by dry cell weight) in the lipophilic cell extracts are comparable to or greater than the amounts produced by the respective non-PDC-producing strains grown in sorghum APL (Fig. 5A through D). In addition, the levels of extracellular PDC are equal to or greater than 100% theoretical yield, based on the measured concentrations of aromatic compounds in APL (Fig. 5E; Table S2). This observation is consistent with the stoichiometric conversion of aromatic monomers to PDC reported previously for the PDC-producing strain that was used in these studies (7).

Although other microbes can be used as sources of carotenoids and CoQ₁₀, our work is important for several reasons. First, we confirmed predictions from N. aromaticivorans genomic and physiological studies that this bacterium contains metabolic pathways that can produce the carotenoid nostoxanthin (13) as well as CoQ₁₀ (16). Our work also shows that minimal genomic modifications of N. aromaticivorans can lead to strains that accumulate valuable carotenoids such as β-carotene, lycopene, astaxanthin, or zeaxanthin. In addition, we demonstrate that N. aromaticivorans cells lacking crtB (Saro_1814) could simplify industrial production of CoQ₁₀, another important commodity chemical used in the pharmaceutical and cosmetics industries, because the crtB mutation prevents the accumulation of other acetone:methanol soluble materials.

We also found that changes in O₂ tension did not have a significant impact on N. aromaticivorans carotenoid or CoQ₁₀ levels, unlike in other microbes where O₂ tensions or the resulting oxidative stress can have a significant impact on the accumulation of these products (27). Effects of O₂ tension on carotenoid production can be variable; in some cases, lower O₂ tensions lead to higher production due to increased membrane synthesis (29), while in other cases, higher O₂ tensions lead to higher production (30) presumably due to the ability of these compounds to quench reactive oxygen species (27). Although our data suggests that N. aromaticivorans might not regulate carotenoid synthesis in response to changes in O₂ tension, additional studies are needed to test this hypothesis. The ability of N. aromaticivorans to synthesize comparable levels of carotenoids and CoQ₁₀ at different O₂ concentrations could be advantageous in industrial settings, due to the capital and operational costs associated with the aeration of large bioreactors.

As society looks for ways to produce commodity chemicals from abundant renewable resources, pretreated lignocellulosic biomass is an attractive material. N. aromaticivorans has natural converging pathways to catabolize major components found in pretreated lignocellulosic biomass, including the most abundant aromatic monomers found in lignin (syringyl, guaiacyl, and p-hydroxyphenyl aromatics) (2), some aromatic dimers (3, 5), and other organic compounds (31). Because carotenoids and CoQ₁₀ are produced from central metabolites, N. aromaticivorans could thus “funnel” mixtures of compounds found in pretreated lignocellulosic biomass into these commodity chemicals. This is in contrast to existing microbial hosts for producing carotenoids or CoQ₁₀ that typically use food-grade sugars as carbon sources (32). To date, relatively few of the microbes being considered for industrially producing carotenoids and CoQ₁₀ also have the native ability to metabolize any aromatic compounds (33–37). Thus, N. aromaticivorans could be an important microbial catalyst for the industrial production of carotenoids and CoQ₁₀ from renewable lignocellulosic carbon sources.

The method used to generate the sorghum APL feedstock is well known to solubilize easily cleavable aromatics from plant cell walls, without the breakdown of the carbohydrate and lignin polymers in the biomass (28). We showed that N. aromaticivorans can grow in the presence of sorghum APL alone, unlike previous studies which have supplemented lignocellulosic APL with minerals (38–43) and/or nitrogen/carbon sources (38, 40, 44, 45). We also found that N. aromaticivorans produces nearly the same amounts (normalized to dry cell weight) of carotenoids and CoQ₁₀ when grown on sorghum APL as when grown in a defined medium containing vanillate (Fig. 4), in addition to accumulating comparable amounts of cellular material (Fig. S10). These results suggest that N. aromaticivorans could produce these valuable isoprenoids from lignocellulosic substrates without the need to add other nutrients.

Our work illustrates the potential for N. aromaticivorans to produce carotenoids and CoQ₁₀ as intracellular lipophilic products. We also generated strains that concurrently produce these lipophilic products along with the soluble extracellular product PDC, which has several potential industrial uses (46). In strains that accumulate both intracellular (carotenoids and CoQ₁₀) and extracellular (PDC) products, the cellular and aqueous fractions can be separated and each used as a source of valuable products to increase the economic value derived from pretreated lignocellulosic biomass. Notably, the strains we engineered for simultaneous PDC and carotenoid production accumulate at least as much carotenoid as the non-PDC-producing strains, showing that the synthesis of two products does not have a significant negative impact on the overall output. The greater than 100% theoretical yield of PDC observed for these strains (Fig. 5E) likely reflects the conversion of aromatic compounds present in the sorghum APL that were not detected by our HPLC-MS/MS analysis, in addition to the complete conversion of detected aromatic monomers (Table S2).

In previous studies with strains that produce PDC from aromatics, a second carbon source (glucose) was added to cells, since the mutations that result in accumulation of PDC block the use of aromatics to support growth (2, 7). The growth of PDC-producing strains in sorghum APL reported here predicts that N. aromaticivorans will not need to be supplemented with other nutrients to produce extracellular and intracellular compounds from this and possibly other types of feedstocks derived from plant biomass.

In sum, this work adds to a growing body of evidence that N. aromaticivorans is a promising microbe for converting lignocellulosic biomass into valuable compounds because it is amenable to genomic modifications and can metabolize abundant aromatic components of this biomass. This work establishes N. aromaticivorans as a promising host for producing valuable carotenoids and CoQ₁₀ from both pretreated lignocellulosic biomass and purified aromatics. In addition, N. aromaticivorans has the ability to produce these intracellular lipophilic compounds concurrently with PDC, which could help to improve the economics of converting plant biomass into industrial commodities. Future work will focus on improving yields of these and other products under industrially relevant conditions.

Details on all strains in this study can be found in Table 1. N. aromaticivorans 12444Δ1879 is a derivative of wild-type strain DSM 12444 (also called F199 [31, 47]), in which a putative sacB gene (Saro_1879 or SARO_RS09410) was deleted to create a strain amenable to genomic modifications using a sacB-containing plasmid (3, 48). We used 12444Δ1879 as the parent strain to generate strains 12444ΔcrtB (lacking crtB; Saro_1814 or SARO_RS09080), 12444ΔcrtY (lacking crtY; Saro_1817 or SARO_RS09095), 12444ΔcrtG (lacking crtG; Saro_0236 or SARO_RS01180), 12444ΔcrtGZ (lacking crtG and crtZ; Saro_0236 and Saro_1168 or SARO_RS01180 and SARO_RS05825), 12444StaxiW [replacing Saro_0236 with the gene for the CrtW protein from S. taxi ATCC 55669 (NCBI accession WP_038660513.1)], and 12444SastaW (replacing Saro_0236 with the gene for the CrtW protein from S. astaxanthinifaciens [NCBI accession WP_211248127.1)].

N. aromaticivorans 12444PDCΔdmtS is a derivative of 12444Δ1879 that was genetically modified to accumulate stoichiometric amounts of PDC from syringyl, guaiacyl, and p-hydroxyphenyl aromatic compounds (7). The strain 12444PDCΔdmtS has Saro_2819 (ligI), Saro_2864–5 (desCD), and Saro_1872 (dmtS) deleted from the genome. We used 12444PDCΔdmtS as the parent strain to generate strains PDCΔcrtG (lacking Saro_0236), PDCΔcrtGZ (lacking both Saro_0236 and Saro_1168), and PDCSastaW (replacing Saro_0236 with the gene for the CrtW protein from S. astaxanthinifaciens).

Genes for CrtW proteins were synthesized as gBlocks (Integrated DNA Technologies, Coralville, IA, USA). Plasmids for cloning were constructed with the NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs, Ipswich, MA, USA). Methods for constructing mutants [including PCR primers used (Table S5)] are contained in Supplementary Information.

E. coli strains used for plasmid cloning were grown in lysogeny broth and shaken at ~200 rpm at 30°C or 37°C. For routine manipulation, N. aromaticivorans cultures were grown in GluSis at 30°C. GluSis is a modification of Sistrom’s minimal medium (50) in which the succinate has been replaced by 22.6-mM glucose. The minimal medium used for N. aromaticivorans experiments was SMB (3) at an initial pH of 7.0. Where needed to select for the presence or absence of plasmids, media were supplemented with 100-µg/mL ampicillin, 50-µg/mL kanamycin, or 10% sucrose (wt/vol).

Sorghum APL (45) was prepared by mixing samples of milled 2014 GLBRC sorghum (2 g) with a sodium hydroxide solution (1% NaOH in H₂O, 20 mL) in sealed 125-mL Erlenmeyer flasks, before heating for 90 min in an oil bath at 90°C. The flask was then immediately placed in ice for 10 min, after which the biomass and aqueous phases were separated by centrifugation at 4,300 × g for 15 min and the supernatant was recovered as a source of soluble aromatics. The solid biomass was rinsed three times with ddH₂O (20 mL, 15 mL, and 15 mL), and the washes were recovered through centrifugation. The initial aqueous supernatant and washes were combined and adjusted to pH 7.0 using 1-M HCl. The solution was centrifuged at 20,000 × g for 1 h at 4°C and passed through a 0.2-µM surfactant-free cellulose acetate (SFCA) filter to remove any remaining insoluble material, yielding the APL used in further experiments.

Cultures of each N. aromaticivorans strain were initially grown in a 125-mL conical shake flask containing 10-mL SMB supplemented with 4-mM vanillate. Between 3 and 8 mL of this culture was combined with 480 mL of fresh SMB + 4-mM vanillate in a glass roux bottle. Roux bottle cultures were attached to a gas mixer using SideTrak 840 mass flow controllers attached to a FloBox 954 (Sierra Instruments, Monterey, CA, USA) in a 30°C temperature-controlled room. Gas was piped into the bottoms of the cultures and exhausted from the headspace through outlets in stoppers. The gas contained 5, 10, or 21% O₂, 1% CO₂, and N₂ as the remainder. Cell growth was monitored by periodically removing samples for analysis using a Klett-Summerson photoelectric colorimeter with a red filter. Cultures were grown until they reached late exponential growth or early stationary phase. For dcw determination, aliquots (~80 mL) were centrifuged in pre-weighed tubes (8,000 × g for 15 min), supernatants were removed, cell pellets were air-dried in a fume hood, and then the tubes were reweighed (Fig. S10). Aliquots (~160 mL) were also harvested (centrifuged at 8,000 × g for 15 min) for isolation of lipophilic compounds by extraction with acetone:methanol (see below).

Each N. aromaticivorans strain was initially grown in a 125-mL conical shake flask containing 10-mL SMB supplemented with 10-mM glucose. In addition, 1-mL aliquots were centrifuged at ~7,000 × g for 5 min, the supernatant was removed, and the cell pellet was used to inoculate 18 mL of sorghum APL in a 125-mL conical shake flask. Cultures were shaken at ~200 rpm at 30°C until they reached the early stationary phase. Aliquots of cultures for extraction into acetone:methanol (10 mL) and dcw determination (5 mL) were harvested as described above (see Fig. S10 for dry cell weight measurements).

Care was taken to minimize O₂ and light exposure to acetone:methanol extracts, although samples were not handled anaerobically. Cell pellets were resuspended in water (950 µL for roux bottle samples and 100 µL for pellets from shake flask cultures) and then transferred into a 15-mL glass Sorvall centrifuge tubes. Extraction solvent (7:2 acetone:methanol solution; 5 mL or 1.5 mL, respectively, for roux bottle or shake flask samples) was added, and the samples were mixed by pipetting. The tube was centrifuged (10,000 × g for 20 min), and then the supernatant was transferred to a new 15 mL glass tube. The pelleted cells were extracted a second time, after resuspending cells in water (500 µL or 100 µL, respectively, for roux bottle or shake flask samples) followed by extraction solvent (4.5 mL or 1.5 mL, respectively, for roux bottle or shake flask samples). After centrifugation, the supernatants from both extractions were combined. The combined supernatants were partially dried under a stream of N₂ (to a final volume of ~1–4 mL) to concentrate materials before analysis by HPLC. The concentration of compounds in lipophilic extracts was calculated after correcting for dry cell weight, any dilution prior to extraction, and the final volume of the sample after drying under N₂.

For identification and quantification, the acetone:methanol lipophilic extracts were analyzed via reverse-phase HPLC using a Kinetex 2.6-µM PS C18 100 Å (150 × 2.1 mm) column (Phenomenex, Torrance, CA, USA) attached to a Shimadzu Nexera XR HPLC system. The mobile phase was a binary gradient (Fig. S11) of Solvent A (70% acetonitrile/30% water) and Solvent B (70% acetonitrile/30% isopropanol) flowing at 0.45 mL/min. Absorbance was measured between 200 and 600 nm using a Shimadzu SPD-M20A photodiode array detector. The following commercial standards were used to identify compounds in the lipophilic extracts: β-carotene (Sigma-Aldrich), lycopene [Pharmaceutical Secondary Standard, Certified Reference Material (7.2%), Supelco], zeaxanthin (United States Pharmacopeia Reference Standard), astaxanthin (Sigma-Aldrich), and coenzyme Q₁₀ (Sigma-Aldrich).

To identify compounds that were not commercially available for use as standards, the eluent from the HPLC was analyzed via mass spectrometry using a Shimadzu triple quadrupole mass spectrometer LCMS-8045. We used positive mode Q3 scans from 450 m/z to 700 m/z around the retention times of unknown HPLC peaks to obtain mass spectra of compounds eluting at such times (Fig. S2 and S8).

Extracellular media samples were prepared by centrifuging 1.5 mL of culture at 20,000 × g for 2 min before passing the supernatant through a 0.2-µM SFCA membrane filter. The filtered media was analyzed using the Shimadzu Nexera XR HPLC system with the photodiode array detector and LCMS-8045 described above. The mobile phase was a binary gradient (Fig. S12) of Solvent A (0.2% formic acid in water) and Solvent B (methanol) flowing at 0.4 mL/min. The stationary phase was a Phenomenex Kinetex F5 column (2.6-µM pore size, 2.1-mm ID, 150-mm length). Aromatic compounds were identified by multiple-reaction monitoring (MRM) using the transition ions specified in Table S3, which were obtained from analyzing pure standards as previously described (7). Aromatic compounds were quantified by comparing sample absorbance at specific wavelengths and retention times with known standards, as measured by the photodiode array detector (Table S4).