Sphingobium sp. KIB was stored in the China General Microbiological Culture Collection Center (CGMCC No.12394), which was isolated from Kunming Institute of Botany (Liu et al., 2019). Sphingobium sp. KIB and its mutants and engineered strains were cultured in basal salt medium (BSM) at 28°C with shaking (220 rpm). BSM medium contains 4 g of (NH₄)₂HPO₄, 10 g of glucose, 0.585 g of MgSO₄, 13.3 g of KH₂PO₄, 1.86 g of citric acid monohydrate, and 10 ml of trace element solution per liter. The trace element solution contains 15 mg of CuCl₂·2H₂O, 25 mg of CoCl₂·6H₂O, 25 mg of NaMoO₄·2H₂O, 30 mg of H₃BO₃, 84 mg of disodium EDTA·H₂O, 80 mg of Zn (CH₃COO)₂·2H₂O, 123 mg/L of MnCl₂·2H₂O, and 600 mg of ferric citrate. The pH value was adjusted with KOH. For testing of the carbon sources or nitrogen source utilization, the glucose or (NH₄)₂HPO₄ was replaced by 10 g/L of other carbon sources or 4 g/L of other nitrogen sources. E. coli JM109 was cultured at 37°C in LB medium and was used to clone expressing constructs for various carotenoids. Ampicillin (100 µg/ml) and/or chloramphenicol (34 µg/ml) was used when required.

Sphingobium sp. KIB was cultivated to a cell density of 0.7–0.8 measured by OD600. The bacterial cells (20 ml) were pelleted by centrifugation at 4,000 × g for 3 min and washed twice with 20 ml of PBS buffer. The cells were resuspended in 20 ml of LB medium and then treated with N-methyl-N-nitro-nitrosoguanidine (MNNG) in a final concentration of 3.5 mM for 1 h in the dark. Following the mutagenesis treatment, cells were washed with the growth medium four times to dilute mutagen and then incubated 6 h on a shaker (28°C, 150 rpm). Subsequently, cells were plated on LB plates, and colonies were developed at 28°C for 3 days. Colonies were screened based on their pigmentation (Li et al., 2017). Carotenoids appear to be different in colors, e.g., the colorless phytoene, red lycopene, and yellow zeaxanthin. Thus, mutants could be isolated by visual color screening. Colonies with different colors from wild-type strain were selected for further analysis of pigment composition.

DNA isolation were carried out according to Ma et al. (2016). Based on the genome sequence (accession No. SRR8864026) of Sphingobium sp. KIB, primers (Supplementary Table S1) were designed to amplify the full length of CrtB, CrtI, CrtY, and CrtG genes. The PCR programs were list in Supplementary Table S1. PCR products were gel purified and sequenced. The Bio Edit software was used for sequence alignment.

Exponential-phase cells were inoculated into 50 ml of culture medium in 250-ml Erlenmeyer flasks. Cultures were incubated at 28°C and 220 rpm in an orbital shaker. Samples were collected at an interval of 6 h for the determination of dry cell weight (DCW). For this, 2 ml of culture was washed twice with distilled water and then filtered through preweighted 0.22-µm membrane filters. Filtered cells were dried to a constant weight in an oven at 65°C.

Recombinant DNA techniques were performed using standard methods. The plasmids and primers used in this study are listed in Supplementary Table S1. Relevant structures of the plasmids are listed in Supplementary Figure S1. The full length of CrtB, CrtI, CrtY, and CrtG genes were cloned, respectively, using Sphingobium sp. KIB genome DNA as a template and cloned into the plasmid pACYC184 to create plasmid pACCARcrtEB, pACCARcrtEBI, pACCARcrtEBIYZ, or pACCARcrtEBIYZG. E. coli JM109 carrying plasmid pACCARcrtEB, pACCARcrtEBI, pACCARcrtEBIYZ, or pACCARcrtEBIYZG displayed a white, red, yellow, or deep yellow phenotype due to the synthesis of phytoene, lycopene, zeaxanthin, or nostoxanthin.

The plasmids pSPRGelb-DXS/IDI/CHY/CrtW, pSPRGelb-DXS/IDI/CHY/BKT, pSPRGelb-DXS/IDI/CrtZ/CrtW, and pSPRGelb-DXS/IDI/CrtZ/BKT were constructed according to Liu et al. The gene CHY was from Haematococcus pluvialis, and the gene CrtZ was from Brevundimonas sp. SD212. These two genes were cloned and then inserted into plasmid pSPRGelb-DXS/IDI at the SalI site, respectively, to construct the plasmids pSPRGelb-DXS/IDI/CHY and pSPRGelb-DXS/IDI/CrtZ. The gene BKT was from Chlamydomonas reinhardtii, and the gene CrtW was from Brevundimonas sp. SD212, and these two genes were cloned and then inserted into plasmids pSPRGelb-DXS/IDI/CHY and pSPRGelb-DXS/IDI/CrtZ at the SalI site, respectively, to construct the plasmids pSPRGelb-DXS/IDI/CHY/CrtW, pSPRGelb-DXS/IDI/CHY/BKT, pSPRGelb-DXS/IDI/CrtZ/CrtW, and pSPRGelb-DXS/IDI/CrtZ/BKT. Then these four plasmids were transformed using electrotransformation into Sphingobium sp. KIB-zea (Liu et al., 2019).

Carotenoid analysis was performed according to Liu et al. (2019). Acetone was used to extract carotenoids. The acetone dissolved the carotenoids, which were blown to dryness with nitrogen gas and dissolved in 100 µl of acetone. Carotenoids were analyzed and quantitated using an Agilent Ultra-Performance Liquid Chromatography (UPLC) 1290 Infinity system equipped with an Agilent Eclipse Plus C18 RRHD 1.8-µm column (2.1 × 50 mm). The mobile phase consisted of solvent A (15% methanol, 60% acetonitrile, 5% isopropanol, and 20% water) and solvent B (15% methanol, 5% isopropanol, and 80% acetonitrile). The extracted carotenoids were eluted at a flow rate of 0.5 ml/min with the following process: 100% A for min; a liner gradient from 0 to 100% within 1 min; 100% B for 8 min. Compounds were detected at 454 and 280 nm, and the absorption spectra, retention times, and peak area of each carotenoid were compared with standards purchased from Sigma (China).

The standard chloroform/isopropanol method was used to extract genomic DNA from Sphingobium sp. KIB. Two primers, 27F (5′-AGA​GTT​TGA​TCC​TGG​CTC​AG-3′) and 1429R (5′-GGT​TAC​CTT​GTT​ACG​ACT​T-3′) were used for amplification of the 16S rRNA gene. The amplification product was purified and sequenced by Shenggong Bioscience Company (Shanghai, China). The 16S rRNA gene sequence of Sphingobium sp. KIB was submitted to GenBank (accession number MK088252). The phylogenetic tree was constructed using the neighbor-joining method via the MEGA 7.0 software.

A yellow pigmented bacterial strain was isolated from tissue culture plates at Kunming Institute of Botany (KIB), which was revealed to be a rod-shaped and Gram-negative bacterium. Phylogenetic analysis based on the 16S rRNA gene sequences showed that the strain had the highest similarity to Sphingobium sp. SA2 (KJ767657.1) (Figure 1A). We, therefore, named the strain as Sphingobium sp. KIB. Sphingobium sp. KIB was found to produce nostoxanthin, a zeaxanthin derivative, up to 93% of total carotenoids (Figures 1B,C), which was much higher than the levels reported in its relatives (Jenkins et al., 1979; Zhu et al., 2012). We hypothesized that Sphingobium sp. KIB was a promising carotenoid-producing strain.

Various carbon sources or nitrogen sources were tested as a potential nutrient for the fermentation of the Sphingobium sp. KIB. It was observed that most carbon sources and nitrogen sources were able to effectively support the growth of Sphingobium sp. KIB (Figure 2). Of the eight carbon sources tested, sucrose, glucose, mannose, and glycerol demonstrated the best carbon sources for carotenoid production and cell growth (Figure 2A). Of the 14 nitrogen sources investigated, tryptone, soya peptone, and okara were found to be the best one for carotenoid production, and (NH₄)₂HPO₄ was found to be the best one for cell growth (Figure 2B). These results demonstrated that the agro-industrial byproducts, such as glycerol, okara, and corn steep liquor, could be effectively used by Sphingobium sp. KIB to benefit cell growth and stimulate carotenoid accumulation. So, at the following studies, 10 g/L of glycerol was used as carbon source and 4 g/L okara was used as nitrogen source.

These byproducts were highly underused energy sources, as a large proportion is dumped into incinerators and landfills (Lee et al., 2019; Kang et al., 2020). It would be desirable to find that Sphingobium sp. KIB could use these agro-industrial byproducts to produce high-value carotenoids. Hence, Sphingobium sp. KIB could serve as a starting strain for the efficient production of zeaxanthin (Liu et al., 2019) and possibly other carotenoids using agro-industrial byproducts via knocking out or introducing the key enzymes involved in the carotenoid biosynthetic pathway.

Sphingobium sp. KIB was treated with a chemical mutagen (MNNG) followed by a color-based screening process. About 100 colonies exhibiting different colors from wild-type strain were selected and subjected to pigment analysis. Four stable mutants, accumulating no carotenoid, phytoene, lycopene, or zeaxanthin, were successfully achieved based on their pigment profiles (Figure 3). Sphingobium sp. KIB-noc demonstrated a white color and accumulated no carotenoid due to a nonsense mutation in the CrtB gene that encodes phytoene synthase involved in catalyzing the first step of carotenoid formation (Figure 3). Sphingobium sp. KIB-phy also demonstrated a white color but predominantly accumulated the colorless carotenoid phytoene (up to 90% of total carotenoids) owing to one amino acid substitution (G476D) in the CrtI (Figure 3). Sphingobium sp. KIB-lyc exhibited a red color and mainly accumulated lycopene due to the nonsense mutation of CrtY, which is involved in converting lycopene to β-carotene (Figure 3). Sphingobium sp. KIB-zea was yellow and accumulated zeaxanthin up to 90% of total carotenoids, which consisted of an amino acid substitution (H127Y) in CrtG; the enzyme catalyzes zeaxanthin to nostoxanthin (Figure 3).

The four mutants showed similar growing status to their parent wild-type strain (Figure 4A). All strains reached their highest biomass at 36 h. Wild type and the mutants KIB-phy, KIB-lyc, and KIB-zea constitutively accumulated nostoxanthin, phytoene, lycopene, and zeaxanthin, respectively. The contents of the carotenoids in the strains increased over time, reaching the highest value of about 5.4 mg g⁻¹ of dry cell weight at 48 h (Figure 4B). This result indicated that Sphingobium sp. KIB and its mutants maintained a homeostasis of carotenoids, irrespective of the kinds of carotenoids.

For further confirmation of the catalytic efficiency of carotenoid biosynthetic genes in Sphingobium sp. KIB, a high-efficiency targeted metabolic engineering carotenoid biosynthesis platform was constructed in E. coli (Figure 5). According to bioinformatic analysis (Liu et al., 2019) and mutant analysis, six enzymes, namely, CrtE (geranylgeranyl pyrophosphate synthase), CrtB (phytoene synthase), CrtI (phytoene desaturase), CrtY (lycopene beta-cyclase), CrtZ (beta-carotene hydroxylase), and CrtG (2,2′-beta-ionone hydroxylase) (GenBank: no. SRR8864026), involved in the successive condensation process from IPP and DMAPP to nostoxanthin, were identified. These engineered E. coli strains, named as E-phy, E-lyc, E-zea, and E-nos, were able to accumulate 6.01 ± 0.034 mg of phytoene, 5.96 ± 0.058 mg of lycopene, 5.99 ± 0.083 mg of zeaxanthin, and 6.03 ± 0.057 mg of nostoxanthin g⁻¹ dry cell, respectively, when cultured in LB medium for 48 h (Figure 5B). From these analyses, we clarified the entire carotenoid biosynthesis pathway and found that the genes involved in carotenoid biosynthesis of Sphingobium sp. KIB, thus, provided more biosynthesis components for bioengineering of carotenoids.

Some noncarotenogenic bacteria and yeast have been engineered to accumulate astaxanthin by heterologous expression of a series of carotenogenic genes. However, the yield is less than unsatisfactory compared with other carotenoids (Jin et al., 2018). To further prolong the metabolic pathway, we engineered Sphingobium sp. KIB-zea to accumulate astaxanthin. A previous study suggested that the pathway from β-carotene to astaxanthin is the limited step of astaxanthin synthesis, which was catalyzed by β-carotene ketolase and β-carotene hydroxylase (Scaife et al., 2009). The combination of these two enzymes is critical to astaxanthin production. We combination express two β-carotene ketolase (Chlamydomonas reinhardtii BKT and Brevundimonas sp. SD212 CrtW) and two β-carotene hydroxylase (Haematococcus pluvialis CHY and Brevundimonas sp. SD212 CrtZ) to find a better combination for efficiently converting to astaxanthin in Sphingobium sp. KIB-zea. Among these four combinations, CrtZ and BKT was the best one with the highest titer of astaxanthin (10.75 mg/g DCW) and total carotenoids (11.6 mg/g DCW) (Figure 6). It should be pointed out that the high astaxanthin content in this strain represents 93% of total carotenoids, which will facilitate downstream processing to obtain a pure product.

Carotenoids are bioactive compounds with numerous biological functions. The increasing demand of various carotenoids has led to many methods to manufacture carotenoids. To date, commercial carotenoids are mainly produced through chemical synthesis. On an industrial scale, carotenoids can also be extracted from naturally producing organisms, e.g., astaxanthin from Haematococcus pluvialis (Lorenz and Cysewski, 2000), β-carotene from Dunaliella (Ben-Amotz, 1995), and Blakeslea trispora (Goksungur et al., 2002). In addition, heterologous biosynthesis of carotenoids in noncarotenogenic microorganisms, such as E. coli and Saccharomyces cerevisiae, has achieved great progress (Chen et al., 2013; Zhou et al., 2015; Shen et al., 2016). Bacterial cultivation is more convenient for large-scale production due to their unicellular nature and high growth rate (Silva et al., 2004; Shiloach and Fass, 2005). Hence, the noncarotenogenic bacterium E. coli has been engineered to produce significant amounts of carotenoids with different strategies. At least five heterologous genes were required to make E. coli produce zeaxanthin at a level of 1.1 mg g⁻¹ dry cell weight (Misawa et al., 1990). Strategies, including coordinating the expression of two or more enzymes (Li et al., 2015), codon optimization (Wang et al., 2011), gene copy number adjustment (Misawa, 2011; Zelcbuch et al., 2013), and integrated multifactor (Li and Huang, 2018) were used to improve carotenoid production. Generally, two or more plasmids consisting of a number of genes had to be introduced into the host (Zelcbuch et al., 2013; Ma et al., 2016), or alternatively, the expression cassettes were integrated into the genomes of the targeted organisms (Zelcbuch et al., 2013; Lu et al., 2017). However, the introduction of too many foreign genes generally resulted in impediment of host growth and production instability (Shen et al., 2016). As a result, no engineered bacteria have been used for carotenoid production on a large scale.

Thus, carotenogenic microorganisms, in particular, those that are generally regarded as safe, e.g., sphingomonads species (Liu et al., 2012b; Kim et al., 2014; Ma et al., 2016), are promising hosts for carotenoid production. Our strain Sphingobium sp. KIB produced nostoxanthin up to 93% of the total carotenoids, which was much higher than that from others (Silva et al., 2004; Asker et al., 2009; Zhu et al., 2012), and demonstrated that Sphingobium sp. KIB could effectively use agro-industrial byproducts, such as glycerol, okara, and corn steep liquor, to benefit cell growth and stimulate carotenoid accumulation. Furthermore, we demonstrated that this strain could be tailored to specifically produce various carotenoids at a level of over 5 mg g⁻¹ dry cell weight through chemical mutagenesis and produce astaxanthin at a level of 10 mg g⁻¹ dry cell weight through directed metabolic engineering. So far, Sphingobium species have not been reported to be animal pathogens. Moreover, Sphingobium was found to possibly play a role in preventing acute otitis media (AOM) (Chonmaitree et al., 2017). Hence, Sphingobium sp. KIB and its mutants could further serve as novel hosts for improved carotenoid productivity by metabolic engineering of the endogenous carotenoid pathways reported by us recently (Liu et al., 2019).