Whole genome sequencing and metabolomics analyses reveal the biosynthesis of nerol in a multi-stress-tolerant Meyerozyma guilliermondii GXDK6

Background Nerol (C10H18O), an acyclic monoterpene, naturally presents in plant essential oils, and is used widely in food, cosmetics and pharmaceuticals as the valuable fragrance. Meanwhile, chemical synthesis is the only strategy for large-scale production of nerol, and the disadvantages of chemical synthesis greatly limit the production and its application. These defects drive the interests of researchers shift to the production of nerol by eco-friendly methods known as biosynthesis methods. However, the main technical bottleneck restricting the biosynthesis of nerol is the lacking of corresponding natural aroma-producing microorganisms. Results In this study, a novel multi-stress-tolerant probiotics Meyerozyma guilliermondii GXDK6 with aroma-producing properties was identified by whole genome sequencing and metabolomics technology. GXDK6 showed a broad pH tolerance in the range of 2.5–10.0. The species also showed salt tolerance with up to 12% NaCl and up to 18% of KCl or MgCl2. GXDK6 exhibited heavy-metal Mn2+ tolerance of up to 5494 ppm. GXDK6 could also ferment with a total of 21 kinds of single organic matter as the carbon source, and produce abundant aromatic metabolites. Results from the gas chromatography–mass spectrometry indicated the production of 8–14 types of aromatic metabolites (isopentanol, nerol, geraniol, phenylethanol, isobutanol, etc.) when GXDK6 was fermented up to 72 h with glucose, sucrose, fructose, or xylose as the single carbon source. Among them, nerol was found to be a novel aromatic metabolite from GXDK6 fermentation, and its biosynthesis mechanism had also been further revealed. Conclusion A novel aroma-producing M. guilliermondii GXDK6 was identified successfully by whole genome sequencing and metabolomics technology. GXDK6 showed high multi-stress-tolerant properties with acid–base, salty, and heavy-metal environments. The aroma-producing mechanism of nerol in GXDK6 had also been revealed. These findings indicated the aroma-producing M. guilliermondii GXDK6 with multi-stress-tolerant properties has great potential value in the fermentation industry.


Background
Nerol (C 10 H 18 O), an acyclic monoterpene, naturally presents in plant essential oils and is used widely in food, cosmetics and pharmaceuticals as the valuable fragrance [1,2]. Meanwhile, nerol possesses antifungal activity against Aspergillus niger [3] and Aspergillus flavus [4] and it can be used as a potential antifungal agent for food preservation. The global annual demand for nerol is more than 5,000 tons per year, but the global production capacity of nerol is only about 3,000 tons, which is still in greater demand for improvement. At present, chemical synthesis is the only way for large-scale production of nerol, but the disadvantages of chemical synthesis such as high cost, complicated process, serious pollution, low yield, and more by-products greatly limit the production and its application. These defects drive the interests of researchers shift to the production of nerol by ecofriendly methods, which are often known as biosynthesis methods. However, the main technical bottleneck restricting the biosynthesis of nerol is the lacking of corresponding natural aroma-producing microorganisms, especially the aroma-producing probiotics.
Aroma-producing probiotics has been demonstrated to be greatly beneficial to hosts and ecological balance. In the past few decades, well-known aroma-producing probiotics had been identified with yeasts, Lactobacillus, and Bacillus subtilis [5]. These species can produce abundant beneficial metabolites through their own metabolic regulations, which showed extremely high practical application. However, most probiotics exist in specific environments, resulting in the interspecies variations among the aroma substances often produced by different probiotics. Most probiotics cannot be fermented for a long time because of the unfavorable conditions, such as the acidification of carbohydrates [6].
To date, known aroma-producing probiotics are generally condition-dependent microorganisms [7]. Few probiotics can ferment different nutrient substrates and generate beneficial aromatic metabolites under special environments (e.g., high salt, extremely strong alkaline, or strong acidic condition), and sustain prolonged aroma production [8,9]. The screening of aroma-producing probiotics with multistress tolerance has become increasingly urgent [10].
M. guilliermondii is known for the production of flavour compounds in fermented food products, has been widely used in food fermentation in recent years [11,12]. For example, Coda et al. [11] used M. guilliermondii to ferment wheat bread, results showed that many active compounds could be synthesized by M. guilliermondii, and the shelf life of wheat bread was effectively extended by 14 days. M. guilliiermondii had also been demonstrated to belong to a kind of probiotic with great potential, which can produce various health bioactive compound, such as the efficient production of isoflavone aglycone [13,14], the overproduction of vitamin B2 (riboflavin) [15]. Herein, a novel aroma-producing M. guilliermondii GXDK6 with multistress-tolerant properties was isolated and identified from subtropical mangrove sediments mainly on the basis of whole genome sequencing. Metabolomics technology based on gas chromatography-mass spectrometry (GC-MS) was also performed to investigate the production of aroma. Among the aromatic productions, nerol was found as a novel aromatic metabolite from GXDK6 fermentation, and its biosynthesis mechanism had also been further revealed. To the authors' knowledge, M. guilliermondii GXDK6 is the first multi-stress-tolerant probiotics from subtropical marine mangrove sediments. This research shows the novel strategies of natural biosynthesis of nerol and also provides new materials and theoretical references for the industrial production of aroma-producing species.

Physicochemical characterization of GXDK6
As shown in Fig. 1a, the colony of GXDK6 was milky white, round, convex, and smooth. The SEM results showed that GXDK6 was similar to the typical yeast species of Pichia anomala Y197-13 [16]. In addition, the cell size was between 2-12 μm, and the cell morphology was oval or semi-oval with a smooth surface (Fig. 1b).

Multi-stress-tolerant properties of GXDK6
As shown in Fig. 1c, GXDK6 could be incubated continuously in the pH range from 2.5 to 10.0. This result indicated that GXDK6 had a strong resistance to acids and alkalis. However, when GXDK6 was incubated in an salty, and heavy-metal environments. The aroma-producing mechanism of nerol in GXDK6 had also been revealed. These findings indicated the aroma-producing M. guilliermondii GXDK6 with multi-stress-tolerant properties has great potential value in the fermentation industry. Keywords: Meyerozyma guilliermondii, Whole genome sequencing, Metabolomics technology, Aroma-producing mechanism, Biosynthesis acidic condition at pH 2.5 or pH 3.0, the relative biomass GXDK6 was higher than that at pH 9.5 or 10.0, respectively. This result suggested that GXDK6 showed better resistance to an acidic environment.
The salt-tolerance results of GXDK6 indicated that the species was a probiotic with tolerance to high-salt concentrations (Fig. 1d). GXDK6 could be incubated continuously up to 12% NaCl, with the relative biomass of approximately 10% with that in 1% NaCl, and up to 18% KCl or MgCl 2 , with the relative biomass of near 90% with that with no salt condition. These findings indicated that the growth of GXDK6 decreased drastically with increasing concentration of NaCl up to 12% but showed no significant impact with increasing concentration of KCl or MgCl 2 up to 18%. As shown in Fig. 1e, GXDK6 also showed good tolerance to the seven heavy metals (i.e., Cd 2+ , Cu 2+ , Mn 2+ , Co 2+ , Ni 2+ , Cr 2+ , and Zn 2+ ). The maximum tolerated concentration for Ni 2+ was 5.8 ppm, while Mn 2+ was 5494 ppm. The tolerance for Cu 2+ was also more than 1000 ppm, but the maximum tolerance for Co 2+ , Cr 2+ , and Zn 2+ was more than 250 ppm. This characteristic shows the potential of the species for the deodorization and bioremediation in heavy metal environments. Similar results were reported by Fernández et al. [17] and Bhakta et al. [18]. The temperature sensitivity of GXDK6 was also investigated (Fig. 1f ). GXDK6 could be incubated at 25-45 ℃, and the optimal incubation temperature was 30-37 ℃. The relative biomass decreased drastically to 25.22% at 45 ℃ compared with that at 30 ℃. In addition, when the incubation temperature increased to 50 ℃, GXDK6 stopped to reproduce. A similar result was reported by Ndubuisi et al. [19], who found that the growth of heat-resistant Pichia was significantly inhibited by increasing temperature, and the ethanol production efficiency of Pichia was lowered.

Whole genome sequencing analysis of GXDK6
The whole genomic sequencing of GXDK6 was performed using the whole genome shotgun method reported by Marcel et al. [20]. The number of reads of GXDK6 was 8, 319, 572 items, the total bases were 2, 477, 073, 673 bps, of which the GC content accounted for 38.88% (Additional file 1). The fuzzy bases were only 0.001%, Q 20 % was 94.60%, and Q 30 % was 86.73%. Genome of GXDK6 was ~ 15,000 bps. The results showed that the extracted DNA was a clear single band, indicating that the extracted DNA was suitable for subsequent analysis (Fig. 2a). Then, a single-base mass distribution map of the sequencing results was constructed (Fig. 2b). The abscissa is the base position of the reads (5ʹ-3ʹ), while the ordinate is the base Q value statistics of all reads. The red and blue lines represent the median and average value of the reads. The yellow line shows that the reads are in the 25% -75% interval, and the tentacles represent that the reads are in the 10-90% interval. The results showed that the bases in the middle had higher base quality. The average quality of the data filtered was also reliable (Fig. 2c).
The sequence length distribution of the third-generation sequencing data is also shown in Fig. 2d, which was mainly used to reflect the average mass distribution of the sequencing data [21]. From Fig. 1e, the whole genome sequence length was good, while the ratio of the uncertain bases to the length of the splicing sequence was 0. This result indicated that the sequence could be used for subsequent gene splicing.
Sequencing and analysis of the ITS domain indicated that the strain with higher consistency basically belong to Meyerozyma sp. The evolutionary distance between GXDK6 and other species has been showed in the species evolutionary tree clearly (Fig. 2e). GXDK6 has a high affinity with 88% confidence level with M. guilliermondii KAML05, M. guilliermondii IFM6377, etc., instead of P. guilliermondii ATCC6260. Therefore, GXDK6 was classified as yeast Meyerozyma guilliermondii.
A complete sequence comparison of the genome sequences was conducted by using the online software BUSCO (http://busco .ezlab .org, v3.0.2) to obtain the percentage of single-copy genes in the total single-copy genes [22]. As shown in Fig. 2e, the spliced genome data of the species were relatively complete. According to the alignment results, GXDK6 belonged to the Pichia genus of the Saccharomyces family and showed the closest relation to M. guilliermondii. Therefore, the yeast Pichia was further confirmed as M. guilliermondii GXDK6.

Ability for single-organic matter fermentation
As shown in Fig. 3a, 21 organic matters (i.e., glucose, sucrose, fructose, xylose, xylan, sorbitol, raffinose, mannose, trehalose, cellulose, maltose, arabic candy, inulin, mannitol, sorbose, D-galactose, cellobiose, wheat bran, ethanol, succinic, and L-rhamnose) as the sole carbon source could be fermented by GXDK6. This result indicated that GXDK6 showed a strong ability to utilize organic matter, such as pentose and hexose. Figure 3b showed that the GXDK6 grew best when used with sorbitol as the sole carbon source, with a dry cell weight of ~ 1.499 g/L, but grew slowest with L-rhamnose, with a dry cell weight of ~ 0.437 g/L. Therefore, the growth rate results showed significant differences in the rate by which GXDK6 utilized diverse organic matters. The order of utilization could be summarized as follows: sorbitol > xylan > raffinose > mannose > sucrose > fruct o s e > t r e h a l o s e > i n u l i n > m a l t o s e > a r a b i c candy > mannitol > glucose > sorbose > D-galactose > cellobiose > xylose > cellulose > ethanol > wheat bran > succinic > L-rhamnose.
Many types of aroma-producing yeasts, such as Hansenula, Candida, S. cerevisiae, Pichia pastoris, and Sphaeropsis sphaeroides, had been reported [23]. The different yeasts also presented diverse aroma-producing characteristics, such as floral fragrance, fruity, delicateness, sweetness, and wine aroma. However, an aromaproducing yeast that can ferment various organic matters, produce abundant aromatic beneficial metabolites, and possess a strong multi-stress tolerance to various environments has not been reported yet. Therefore, M. guilliermondii GXDK6 will be a potential probiotic with important application value.
The metabolites produced by GXDK6 with glucose, sucrose, fructose, or xylose as sole carbon sources was detected with GC-MS method, results showed that 14, 20, 16, and 26 peaks were detected in the samples (Fig. 3a-d). Among them, when GXDK6 was fermented with six carbon sugars (glucose, sucrose) as a single carbon source for 72 h, the top five substances with peak areas were 3,5-ditert-butylphenol, sec-butyl cyclohexyl sulfide, isoamyl alcohol, 3,5-dimethylbenzaldehyde, and ethanol, respectively. However, when GXDK6 was fermented for 72 h with pentose (fructose, xylose) as a single carbon source, the top five substances in peak area were 3,5-ditert-butylphenol, sec-butyl cyclohexyl sulfide, 3,5-dimethylbenzaldehyde, nerol, and isoamyl alcohol (or 2-ethyl hexanol), respectively. Thus, GXDK6 could ferment with different kinds of organic matter as the sole carbon source to produce aromatic metabolites. However, the distinct metabolic regulation networks and mechanisms remain unknown [24,25], which still need further study.

Metabolomic analysis of GXDK6
As shown in Fig. 4, the metabolites of GXDK6 produced from the selected carbon sources can be classified as alcohols, esters, acids, hydrocarbons, aldehydes, ketones, sulfide, phenolics, and other organic substance (Fig. 4a). However, alcohols, lipids and organic acids are usually defined as the main aromatic metabolites. Based on this, the main aromatic metabolites produced by GXDK6 fermentation of glucose and sucrose are 9 and 13 types, (b) Growth curve of GXDK6 with xylose, L-arabinose, glucose and sucrose as the single-carbon source, respectively; c The peak spectrum of GXDK6 fermented with glucose; d The peak spectrum of GXDK6 fermented with sucrose; e The peak spectrum of GXDK6 fermented with fructose; f The peak spectrum of GXDK6 fermented with xylose respectively (Table 1), and the aromatic metabolites produced by fermentation of fructose and xylose are 8 and 14 types, respectively (Table 1).

Metabolic pathways of nerol and its biosynthesis mechanism
In order to further study the molecular mechanism of aroma production by GXDK6 fermentation, nerol (Additional file 2) was taken as a representative novel aromatic metabolite from M. guilliermondii, which was further compared and analyzed with the whole genome data of GXDK6 (Additional file 3), and the metabolic pathway was further elucidated. As shown in Fig. 5, the upstream sources of glycolysis or citric acid cycle 1-deoxy-d-xylu-lose5-phosphate geranyl diphosphate or linalool nerol, was firstly converted into final geranic acid or 8-Oxogeranial. In this process, the proteins involved in nerol synthesis were geranyldi phosphotase, monoterphenyl diphosphatase, and geraniol isomerase. The proteins involved in nerol metabolic transformation were geraniol 8-hydroxyase, alcohol dehydrogenase, and geraniol dehydrogenase. As reported by Zong et al. [10], nerol was biosynthesized in the metabolic engineered Escherichia coli from glucose with an accumulation of 0.053 ± 0.015 mg/L, and the biosynthesis mechanism had also been revealed. The truncated neryl diphosphate synthase gene tNDPS1 was expressed that catalyzed isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) form neryl diphosphate (NPP), and then the nerol synthase gene GmNES was co-expressed to synthesize the final product nerol from NPP. However, the accumulation of nerol of GXDK6 was ~ 2.740 mg/L (p < 0.05) when fermented with glucose as the substrate, which is higher than it was produced by E.coli fermentation [10], and the biosynthesis mechanism of nerol in native Meyerozyma guilliermondii has not been reported yet.
The structure and function of the six proteins were further investigated (Fig. 5), results showed that geranyl diphosphatase was a dimer protein with ~ 65 kDa, and its corresponding ligand was found as Mn 2+ (Fig. 5b), which suggested that geranyl diphosphatase could be bound and interacted with Mn 2+ , and promoted the catalytic reaction to produce more nerol [28]. Monoterpenyl diphosphatase was also a dimer protein with a molecular weight ~ 68 kDa, but its corresponding ligand had not been found yet (Fig. 5c), this indicated it should be a non-allosteric enzyme with auxiliary catalysis [29]. Geraniol isomerase was a pentameric protein with ~ 44 kDa (Fig. 5d), and its corresponding ligand was found as geraniol (ligand for nonmetallic ion not shown), indicating that the existence of geraniol was beneficial to the catalytic reaction [30]. In summary, these proteins were indispensable and directly participate in the regulation and biosynthesis of nerol.
The subsequent metabolism or transformation of nerol was catalyzed by geraniol 8-hydroxyase, alcohol dehydrogenase, and geraniol dehydrogenase. Among them, geraniol 8-hydroxyase was a monomeric protein of ~ 55 kDa [31], and its corresponding ligand had not been found yet (Fig. 5e). Alcohol dehydrogenase was also a monomeric protein with a molecular weight of ~ 39 kDa [32], its corresponding ligand was found as Zn 2+ (Fig. 5f ), indicating alcohol dehydrogenase could be bound and interacted with Zn 2+ , it would contribute to the formation of nerol. Geraniol dehydrogenase was a dimer protein with ~ 41 kDa [33], and its corresponding ligand had not been found (Fig. 5g), suggesting that it could be a non-allosteric enzyme with auxiliary catalysis. These evidences indicated that the generation of nerol was mainly attributed to the existence of corresponding enzymatic system and metabolic pathway in GXDK6. Furthermore, nerol was classified as a typical example of aromatic metabolites in GXDK6, why GXDK6 can maintain a long-term aroma production should be the contribution of various aromatic metabolites. Therefore, the biosynthesis mechanism of nerol will contribute to better understand of the aroma-producing mechanism of GXDK6.

Conclusions
A novel aroma-producing M. guilliermondii GXDK6 was identified successfully by whole genome sequencing and metabolomics technology. GXDK6 showed high multistress-tolerant properties with acid-base, salty, and heavy-metal environments. Furthermore, GXDK6 could be fermented with 21 organic matters (including pentose and hexose) as the sole carbon sources and could produce abundant aromatic metabolites, such as alcohols, esters, acids, etc. The aroma-producing mechanism of nerol in GXDK6 had also been revealed. These findings indicated the M. guilliermondii GXDK6 has great potential value in the fermentation industry. The chemical reagents were analytical grade. Glucose, fructose, maltose, sorbitol, lactose, sucrose, mannitol, soluble starch, bran, cassava flour, and bagasse were  Japan). The single colony without any chemical treatment was also scanned using a scanning electron microscope (TM3000; Olympus, Japan), with a scanning voltage set to 10 kV and magnification of 10, 000 folds.

Multi-stress-tolerant properties of GXDK6
The acid-base stress tolerance of GXDK6 was investigated in the pH range of 2.5-10.0. Salt tolerance was investigated by using NaCl, KCl, and MgCl 2 in the concentration range of 12%-18%. Heavy-metal tolerance was investigated using seven heavy metals (i.e., Cd 2+ , Cu 2+ , Mn 2+ , Co 2+ , Ni 2+ , Cr 2+ , and Zn 2+ ) in the concentration range of 0-5494 ppm. Temperature sensitivity was also investigated at 25-50 ℃. Three parallel experiments were performed, and the inoculation rate of GXDK6 was 2% (v:v). The growth of GXDK6 was evaluated using the turbidity method as reported by Irache et al. [34], with slight modifications.

Whole-genome sequencing analysis
GXDK6 was incubated by using an enrichment technique and GBM liquid medium (0.2% yeast extract, 0.2% beef extract, 0.5% polypeptone, 0.6% sucrose, pH 7.0). Both enrichment processes were carried out at 200 rpm and 30 ℃ for 8 h. The bacterial cells were then centrifuged at 8, 000 rpm and 4 °C for 10 min. Then, the cells were washed repeatedly with 0.1 mol/L PBS buffer. The whole genomic DNA of GXDK6 was extracted using the CTAB method as reported by Van et al. [35], with slight modifications. The purity of the extracted DNA was verified by PCR (Polymerase Chain Reaction) and agarose gel electrophoresis [36]. The ITS gene was amplified by PCR using ITS1 (5′-TCC GTA GGT GAA CCT GCG G-3′) /ITS4 universal primers (5′-TCC TCC GCT TAT TGA TAT GC-3′) [37]. The ITS sequencing data of GXDK6 were deposited to the National Microbiology Data Center database (http://nmdc.cn) under the Accession Number of NMDCN000022O.
The whole-genome sequencing and analysis of GXDK6 were performed by the BGI Genomics Co., Ltd. (Shenzhen, China). An online software FastQC (http://www.bioin forma tics.babra ham.ac.uk/proje cts/fastq c) was used for the quality control of the second-generation sequencing downtime data. To determine the percentage of single-copy genes in the total single-copy genes, an online software BUSCO (http://busco .ezlab .org, v3.0.2) was conducted to complete the sequence comparison of the genome sequences. Hmmscan software [38] was adopted to predict the presence of Carbohydrate-Active enzymes (http://www. cazy.org) [39] genes in the genome sequence. The whole genome sequencing data of GXDK6 were deposited to the National Microbiology Data Center database (http://nmdc. cn) under the Accession Number of NMDC60014229.

Metabolomic analysis of GXDK6
GXDK6 was incubated in the medium with glucose, sucrose, fructose, or xylose as the sole carbon source for 72 h. Then, the fermentations were separated by centrifugation at 12, 000 × g for 10 min. The supernatants were then freeze-dried to powder by a refrigerated centrifugal concentrator (Shanghai, China). The powder samples were then derivatized for 120 min by using a 0.1 mg/mL methoxyamine hydrochloride-pyridine solution (Reagents for GC level) and alkylated for 120 min by using trifluoroacetamide (Reagent for GC level). When completed, the samples were centrifuged at 12, 000 × g for 10 min. The aromatic metabolites were detected and analyzed using GC-MS according to Pautova et al. [40], with slight modifications.