ABSTRACT
Genomics has become the primary way to explore microbial diversity, because genetic tools are currently difficult to develop in non-model organisms. Here, we demonstrate that -omics can be leveraged to accelerate genetic tool development for the basidiomycete yeast Xanthophyllomyces dendrorhous CBS 6938, the sole biotechnologically relevant organism in the Tremellomycete family. First, we sequence the genome. Then, we perform transcriptomics under a variety of conditions, focusing on light and oxidative stress. This data not only reveals novel photobiology and metabolic regulation, it also allows derivation of constitutive and regulated gene expression parts. Our analysis of X. dendrorhous photobiology shows for the first time that a complex system of white-collar and cryptochrome homologs mediate response to ultraviolet light (UV). Our analysis of metabolic regulation shows that UV activates DNA repair, aromatic amino acid and carotenoid biosynthesis and represses central carbon metabolism and the fungal-like apoptotic pathway. Thus, X. dendrorhous shows a dynamic response toward biosynthetic pathways for light-absorbing compounds and survival and away from energy production. We then define a modular cloning system, including antibiotic selections, integration sites, and reporter genes, and use the transcriptomics to derive strong constitutive and regulated promoters. Notably, we discover a novel promoter from a hypothetical gene that has 9-fold activation upon UV exposure. Thus, -omics-to-parts workflows can simultaneously provide useful genomic data and advance genetic tools for non-model microbes, particularly those without a closely related model organism. This approach will be broadly useful in current efforts to engineer diverse microbes.
INTRODUCTION
Yeasts are a diverse group of fungi characterized by unicellular growth, which is useful in industrial fermentations. The group is not monophyletic, rather the yeast state has emerged several times across the fungal kingdom. Therefore, yeasts have vastly different genome architectures, metabolic capabilities, and stress tolerances that are not only due to the ecological niches to which they are adapted, but also due to their different lineages. The impact this biodiversity might have on industrial biotechnology has only just begun to be explored.1 However, the different lineages limit the transferability of genomics and genetic tools between yeasts, particularly for the basidiomycete yeasts that are not closely related to any model organism.2 Here, we demonstrate how modern genomics and transcriptomics can be leveraged to simultaneously gain new insight into nonconventional yeast biology and derive a collection of genetic parts for future studies. We demonstrate this for the basidiomycete yeast Xanthophyllomyces dendrorhous, the only biotechnologically relevant organism from the Tremellomycete family of Basidiomycota.3
Historically, the model ascomycete Saccharomyces cerevisiae has been the host of choice for yeast metabolic engineering.4 While valuable for converting hexoses to ethanol, engineering S. cerevisiae to consume other carbon sources and produce other economically valuable compounds can require extensive reprogramming, which imposes metabolic and gene expression burden.5 This is particularly true of tolerance, which can be due to many genes interacting in currently unpredictable ways. This has led to a growing body of literature that investigates other yeasts, termed nonconventional simply because they are not S. cerevisiae, as potential hosts for metabolic engineering since they are already adapted to consume desired carbon sources, produce desired classes of molecules, and tolerate different fermentation conditions, thus less burden is incurred when engineering them.6,7 Methods for developing genetic tools are needed for these organisms.
Developing genomics and genetic tools in nonconventional yeasts is challenging. This is because a high-quality genome, gene regulatory data, species-specific DNA parts for genetic manipulation, and efficient transformation procedures are often missing. The process of developing these elements, termed “onboarding,” can be accelerated for nonconventional ascomycete yeasts closely related to S. cerevisiae because the genomics and genetics are similar, permitting transfer of similar promoter and plasmid designs. Onboarding organisms from relatively unknown classes, particularly basidiomycete yeasts, does not have this advantage because there is not a closely related model organism with a complete, contiguous genome and well-developed genetics.
Xanthophyllomyces dendrorhous is a nonconventional yeast that has gained attention in the aquaculture and nutraceutical industries due to its natural ability to produce astaxanthin, a vibrant red pigment with potent antioxidant properties. It is currently the only Tremellomycete with biotechnology applications. While gene knock-out and upregulation techniques have been employed to modify X. dendrorhous, particularly strain CBS 6938, reprogramming this yeast remains a challenge because several key tools in the onboarding toolkit are missing, including a high-quality genome sequence, a modular gene expression parts collection, and efficient transformation methods.6 Furthermore, astaxanthin is only overproduced in response to light and oxidative stress, which complicates and limits the scale of bioprocesses.7–9
Here, we demonstrate an omics-driven onboarding workflow including genome sequencing, transcriptomics, high efficiency transformation, and modular gene expression parts collection development for X. dendrorhous. This workflow generates new insights into basidiomycete photobiology and metabolism, constitutive promoters, light-inducible promoters, and efficient genome integration. Therefore, this “omics-to-parts” approach can rapidly deliver a suite of data and tools for onboarding nonconventional yeast that have no closely related model host.
RESULTS
High quality genome of X. dendrorhous CBS 6938
A high-quality CBS 6938 genome was obtained through Prymetime sequencing (Figure 1a).10 Prymetime assembled the genome into sixteen contigs, and relative lengths of the contigs are depicted in the figure. Subsequent benchmarking universal single-copy ortholog (BUSCO) analysis revealed that the CBS 6938 genome contains 94.0% complete and single copy BUSCOs (Figure 1b). This genome sequence was the foundation for the subsequent work.
(a) schematic of Prymetime next generation sequencing workflow. Illumina and Nanopore reads are combined for a higher quality genome. (b) BUSCO analysis of the X. dendrorhous genome.
Foundational genetic tools for onboarding X. dendrorhous
In addition to obtaining an accurate genome sequence, we performed a set of experiments to make X. dendrorhous easier to engineer. We first examined possible selection markers. Rather than seek to create synthetic auxotrophy, which has a variety of metabolic impacts,11 we screened for antibiotic sensitivity. We measured X. dendrorhous sensitivity to several common antifungals via minimum inhibitory concentration (MIC) assay (Methods). The most effective antifungal was found to be nourseothricin with 11.6% cell survival at a concentration of 10 µg/mL (Figure 2a).
(a) Minimum inhibitory concentration assay data from testing geneticin, zeocin, hygromycin, and nourseothricin against X. dendrorhous. (b) Cloning workflow adapted from CIDAR for homologous recombination into the crtYB site. (c) Depiction of high-efficiency transformation procedure. (d) Overlapped chromatograms illustrating the lack of astaxanthin at a retention time of 0.998. Wild type is shown in red and the crtYB knockout is shown in black. (e) Comparison of luminescence between two strong constitutive promoters for act and adh4, the crtYb knockout, and CBS 6938 wild type.
We then designed a modular cloning system for rapidly assembling X. dendrorhous genetic elements (Figure 2b). This system uses the same enzymes and scars as previously published modular cloning systems.12 There are two significant differences in this red yeast modular cloning system. The first is that the lacZ element in the destination plasmids is replaced with the lethal gene ccdb to shift from blue/white screening of correct clones to a system where only correct clones are able to grow on a selection plate. The second is that the Level 2 (L2) integrative plasmid has homology arms compatible with the X. dendrorhous genome. We used our genome sequence to select homology arms flanking that crtYB gene, the gene that catalyzes production of the colorful carotenoid pigments. This creates a facile red/white screening method for correctly integrated clones.
We then designed an initial integration cassette using the known glutamate dehydrogenase promoter (Pgdh) to drive expression of a codon-optimized nat gene which confers nourseothricin resistance. We used this initial integration cassette to optimize the efficiency of transformation. The method we found to be most efficient is described in Methods. In brief, it consists of the culture of CBS 6938 cells until the end of exponential growth phase is reached, combination with the integrative pathway, and application of an electric shock (Figure 2c). Ultra-high-performance liquid-chromatography (UPLC) was then used to demonstrate successful crtYB knock-out through measurement of astaxanthin. Astaxanthin peaks were detected in wild type but not knock-out extractions (Figure 2d). Thus, with the design of the integration cassette and optimization of the transformation protocol we are able to achieve efficient and targeted integration of DNA into the X. dendrorhous genome using homologous recombination.
We then established a luminescent reporter, which is common in fungal engineering, for measuring gene expression strength. We cloned the NanoLuc enzyme13 into our modular cloning scheme, integrated it into the genome, and measured expression via luciferase assay (Figure 2e). NanoLuc-producing cells created approximately 1,000-fold greater luminescence compared to the crtYB knockout strain, and nearly 10,000-fold greater luminescence compared to CBS 6938. The higher luminescence observed from the crtYB knockout strain may be attributed to biological autoluminescence, a phenomenon that has been found to be directly related to oxidative stress.14
Together with the genome sequence, these experiments enabled the development of advanced genetic tools for X. dendrorhous.
Confirming that astaxanthin biosynthesis responds to light through transcriptional regulation
We next sought to more fully understand X. dendrorhous biology by examining genetic response to light and oxidative stress. While previous work has shown that visible and ultraviolet (UV) light increases carotenogenesis in X. dendrorhous, specific genes and mechanisms involved in this regulation have not been investigated. Therefore, we first sought to confirm that the carotenogenesis pathway was transcriptionally regulated. Before conducting transcriptomics, we used the genome sequence to design quantitative polymerase chain reaction (qPCR) probes to target genes in the terpenoid (MVK, PMVK, MVD, IDI, FPS, and crtE) and carotenogenic (crtYB, crtI, and crtS) pathways. Cells were grown under either dark or UV-light conditions, RT-qPCR was performed, and log2 fold-change was calculated (Methods). Upstream genes MVK, PMVK, MVD, IDI, FPS, and crtE exhibited little to no expression response. However, crtYB, crtI, and crtS expression was upregulated by UV light exposure. Notably, the crtI gene was the most impacted with log2 fold-changes of 0.90 (Figure 3a). These results demonstrate that light, particularly UV, transcriptionally upregulates carotenogenesis, not the upstream terpenoid pathway, in X. dendrorhous.
(a) RT-qPCR results for enzymatic steps in the mevalonate and carotenogenesis pathway. Upstream steps are not affected by UV exposure, but downstream genes crtYB, crtI, and crtS are considerably induced. (b) RNA isolation process for transcriptomic sequencing. Six treatments were implemented including four light wavelengths, a positive control (H2O2), and a negative control (dark).
Transcriptomics to elucidate X. dendrorhous photobiology
We then designed a transcriptomics experiment to investigate the genome-wide response of X. dendrorhous across the light spectrum including UV, blue, green, and red light. We included a positive control condition of oxidative stress by hydrogen peroxide exposure, and growth of cells in the dark was used as a negative control (Methods) (Figure 4a). We extracted the RNA, which was then sequenced and processed by the DOE Joint Genome Institute through their Community Science Program. We defined a noteworthy log2 fold-change as any greater than 0.50.
(a) Regulation of mevalonate and carotenogenesis genes by light and H2O2. (b) Regulation of putative photoreceptors. Numbers are arbitrarily assigned. (c) Top ten UV induced and repressed genes. (d) Top ten H2O2 induced and repressed genes.
The data corroborated our previous evidence that exposure to UV light leads to upregulated expression of crtI with a log2 fold-change of 0.90. Further, the upstream terpenoid pathway again had no significant changes, indicating upstream terpenoid synthesis to the diterpenoid intermediate, geranylgeranyl pyrophosphate, was unaffected by light. In contrast, oxidative stress upregulated genes in both pathways (MVK, PMVK, IDI, FPS, crtYB, and crtI), with IDI having the greatest increase (log2 fold-change of 1.79) (Figure 4b). Induction of oxidative stress may be a method to overcome the bottleneck that IDI imposes on the mevalonate pathway.15 Overall, the results corroborate the previous evidence that UV light upregulates the final three steps of carotenogenesis in X. dendrorhous and show that oxidative stress broadly upregulates terpenoid biosynthesis.
We then looked for possible light response systems in X. dendrorhous. There are several known classes of fungal light receptors, particularly white collar and cryptochrome proteins in ascomycete and two receptors in basidiomycetes, BLUF, which senses to blue light, and rhodopsin, which senses green light.16 We searched our X. dendrorhous CBS 6938 genome with BLAST for putative fungal photoreceptor proteins within the BLUF, rhodopsin, white collar, and cryptochrome families. Photoreceptor proteins predicted to be in the same family were differentiated with arbitrarily assigned numbers. We found that the putative cryptochromes increased transcription most dramatically in UV and blue light exposure. To our knowledge, this is the first evidence of cryptochrome light response systems in basidiomycetes. Additionally, putative white collar and rhodopsin genes increased expression. Analysis of the oxidative stress positive control corroborated that these genes were responding specifically to light because most putative photoreceptors were downregulated or unaffected by hydrogen peroxide (Figure 4c).
Finally, we filtered the whole genome transcriptomic data for the genes most up- or downregulated by UV light. We identified ten upregulated and six downregulated genes with log2 fold-changes of 1.50 or greater (Figure 4d). These genes were from different functional classes, from cell survival to energy storage. Endonuclease III, MSS4-like protein, and short-chain dehydrogenases are directly involved in response to UV stress through DNA repair, cell survival through cell cycle regulation and actin cytoskeleton formation, and induction of light-regulated pathways, respectively.17–19 These functions collectively correspond to an expected stress response from UV exposure.20 Conidiation-specific protein 6, a protein involved in the reproductive cycle, is also upregulated after blue-light exposure in Neurospora crassa.21 Unexpectedly, UV light upregulates expression of glycogenin glucosyltransferase, an enzyme that initiates glycogen nucleation. Downregulated genes were found to be involved in transcriptional regulation, fungal apoptotic-like cell death, and the electron transport chain. Also of note, four of the ten upregulated genes were predicted to result in hypothetical proteins, indicating that our understanding of light response mechanisms is incomplete. Genes most up- or downregulated in response to hydrogen peroxide exposure were found to not be affected by light treatments, further supporting the notion of separate response pathways (Figure 4e). In summary, it appears that X. dendrorhous activates DNA repair, survival, and reproductive mechanisms while downregulating programmed cell death. The metabolic response is complex, with some pathways upregulated and others downregulated.
To more clearly analyze the metabolic response, genes were clustered into functional groups using the ERGO™ Bioinformatics platform from Igenbio. Upregulation of functional groups is depicted in Figure 5a. Generally, DNA repair and amino acid metabolic functions were upregulated. Interestingly, aromatic amino acid degradation and modification were upregulated by both UV light and hydrogen peroxide, but amino acid uptake was only upregulated in UV-exposed cells. Furthermore, electron transport, glycolysis, gluconeogenesis, and carbohydrate metabolism were downregulated (Figure 5b). These results support the general trends observed in the analysis of individual genes. We synthesized all of these observations into a qualitative description of the CBS 6938 metabolic shift in response to UV (Figure 5c). With this view, the transcriptional response indicates that X. dendrorhous likely redirects flux from central carbon metabolism to carotenoids and aromatic amino acids upon UV exposure. Aromatic amino acids absorb UV and are also precursors to other UV light absorbing molecules like melanin and mycosporine compounds.22–24 Yet, the activation of aromatic amino acid pathways in response to UV has not been previously observed in basidiomycetes. These results indicate that the response to UV in X. dendrorhous is more extensive and complex than acting on carotenogenesis alone. Furthermore, the significant fraction of hypothetical proteins hints at additional unknown mechanisms that also contribute to the response to UV exposure.
(a) Overall up- and downregulation of metabolic processes by UV light or H2O2. (b) Summary of metabolic processes regulated by UV light. Generally, metabolic flux is being drawn from central carbon metabolism to synthesis of aromatic amino acids.
Transcriptomics-driven part selection allows simultaneous derivation of constitutive and regulated genetic parts
We then leveraged the transcriptomics data to derive functional genetic elements for use in our established modular cloning scheme. Since the data covered a variety of conditions, we reasoned that it should be possible to derive putative constitutive and regulated promoters. We identified strong constitutive promoters by sorting each transcriptomics dataset by fragments per kilobase of exon per million mapped fragments (FPKM) count, choosing the genes that had the highest and consistent FPKM counts across conditions (Figure 6a), and using our genome sequence to obtain the sequences upstream of those genes. We also obtained the sequences upstream of the genes in Figure 4c that were the most activated or repressed by UV light (Figure 6a). Promoter lengths were chosen with the assistance of an analysis conducted by JGI. These putative promoters were then cloned into our modular cloning system to drive expression of nanoluciferase and integrated into the genome (Figure 6b).
(a) Putative consitutive promoters are shown. (b) Putative nducible promoters are shown. (c) Putative epressible promoters are shown. (d) Depiction of putative inducible promoters with log2 fold-change. PHP3 has the greatest induction to UV light at a log2 fold-change of 3.01. (e) Depiction of putative repressible promoters with log2 fold-change. None were found to be repressible.
We measured expression of each construct via a luciferase assay in the dark and under UV exposure (Methods). As Figure 6c shows, the constitutive promoters result in high luminescence that does not change between the dark and the light. In contrast, the putative UV activated promoters showed increased luminescence in response to UV (Figure 6c). However, most of these increases were not large. The two strongest inducible promoters were found to be hypothetical proteins HP3 and HP2 with log2 fold-changes of 3.01 and 1.47, respectively. This means that the strongest UV-inducible promoter results in a 9-fold change in gene expression. The putative UV repressed promoters did not show any significant repression. We reasoned that the long half-life of luciferase was confounding measurement of repression. Thus, we designed a qPCR experiment, and the results indicated no change. Therefore, it is possible our promoter lengths were insufficient to capture the regulatory elements in these promoters.
Taken together, our results show that transcriptomics driven part discovery can consistently derive strong constitutive promoters, but inconsistently derive regulated promoters due to a variety of possible factors. Even so, we were able to discover two promoters that can be activated by UV. Thus, our part discovery efforts generated a collection of constitutive and regulated promoters that are compatible with our modular cloning workflow for X. dendrorhous CBS 6938.
DISCUSSION
In this study, a promising nonconventional yeast was subjected to a streamlined set of genomics and genetics experiments that transformed it into a powerful platform for genetic engineering. This was done by obtaining a high-quality genome, increasing transformation efficiency, developing advanced modular genetic tools, measuring gene regulation, and deriving a genetic parts collection from transcriptomic data. Now, it is possible to interrogate the unique metabolism and photobiology of X. dendrorhous in a similar manner to other nonconventional yeasts.
Our analysis of X. dendrorhous photobiology revealed a number of novel insights into the biological mechanisms of basidiomycete light response. We found evidence of multiple different fungal light response systems, including the first indication of cryptochromes in basidiomycetes. We determined for the first time that light transcriptionally regulates the carotenoid, or tetraterpenoid, pathway but not the upstream terpenoid pathway. We learned that X. dendrorhous engages a variety of survival responses in response to UV, including activation of DNA repair and reproduction pathways and repression of the fungal apoptotic-like pathway. We also observed a complex transcriptional response in metabolism, where central carbon metabolism is repressed in favor of aromatic amino acid and tetraterpenoid biosynthesis. Taken together, these data paint a picture of an extensive and coordinated response to light. This complex system is likely key to survival of X. dendrorhous in its ecological niche of fallen trees at high altitudes, which would include extended periods of intense UV exposure.
It is interesting to note that our genome integration approach results in efficient on-target insertion of constructs via homologous recombination (HR). Our findings support the results of others that X. dendrorhous CBS 6938 has an active HR pathway, which is incredibly beneficial for future strain engineering efforts. This challenges the dogma that nonconventional yeasts favor the non-homologous end joining (NHEJ) mechanism, and argues that other yeasts besides S. cerevisiae favor HR.25,26
Leveraging transcriptomics allowed us to derive constitutive and regulated gene expression parts. Transcriptomics has been previously used to identify regulated promoters in other yeasts. We built on this prior evidence and coupled it with a modular cloning standard to create an -omics-to-parts workflow that can yield needed constitutive and inducible promoters for construction of genetic devices. This -omics-to-parts workflow can be generalized to derive genetic parts collections for a number of nonconventional microbes.
Development of novel production platform hosts requires an understanding of genome composition and genetic regulation, which also is essential for genetic tool and genetic part development. This work demonstrates how an integrated genomics and genetics approach can simultaneously deliver new biological understanding and onboard a nonconventional organism, particularly a nonconventional yeast with no closely related model organism. This integrated approach, consisting of a small number of experiments, promises to accelerate organism onboarding efforts.
MATERIAL AND METHODS
Strains and media
The X. dendrorhous type strain CBS 6938 (ATCC 96594) was cultured at 21 °C in YPD (YEP, Sunrise Science, 1877-1KG plus 20 g/L glucose, Alfa Aesar, A16828) media for all experiments. For all light plate apparatus experiments, the X. dendrorhous seed cultures were shielded from any external light by wrapping the flasks with aluminium foil. X. dendrorhous transformants were selected using YPD with 30 mg/L nourseothricin (Jena Bioscience, AB-101-10ML). Solid media plates were made using 20 g/L agar (Sunrise Science, 1910-1KG). Golden gate (Type-IIS) cloning was conducted using chemically competent Escherichia coli DH5α cells (NEB, C2987H). Construction of destination vectors with the toxic selection marker ccdB was done with chemically competent One Shot ccdB Survival 2 T1R E. coli cells (Invitrogen, A10460). Both E. coli strains were grown in 25 g/L LB Miller broth (Fisher Scientific, BP1426-2) at 37°C. Antibiotic selection was carried out using 100 mg/L carbenicillin (Alfa Aesar, J61949), 25 mg/L chloramphenicol (Alfa Aesar, B20841), and 50 mg/L kanamycin (Alfa Aesar, J61272) for level 0, level 1, and level 2 assemblies, respectively.
gDNA isolation and sequencing
Isolation of gDNA from X. dendrorhous and next-generation sequencing was performed as described by Collins et al.10
Construction of a light plate apparatus for various light wavelengths
The light plate apparatus (LPA) was assembled following a user’s manual published by the Tabor lab at Rice University.27 The LPA is an instrument capable of shining two individual LED lights on cell cultures in a 24-well plate. It is comprised of a 3D-printed shell surrounding a soldering board with 48 LED light sockets oriented below a 24-well plate with a clear bottom (AWLS-324042, Arctic White). The LPA allows each culture to be exposed to two unique LED lights without disrupting neighboring cultures.
RNA isolation from the light plate apparatus
LPA cultures were started in 2 mL of YPD media at OD600 = 1 and shaken for 4 hours at 200 rpm while exposed to one red (λ = 660 nm), green (λ = 565 nm), blue (λ = 470 nm), white (visible light spectrum), or UV LED light (λ = 400 nm), or no LED light. For the transcriptome sequencing experiments, the hydrogen peroxide condition replaced the white light condition. The hydrogen peroxide condition consisted of YPD media with hydrogen peroxide at a concentration of 10 mM and had no LED light. 1 mL of X. dendrorhous culture was taken forward for RNA isolation, which used a cell homogenization and TRIzol-based method. X. dendrorhous cells were first resuspended in 200 μL of cell lysis buffer (0.5 M sodium acetate, 5% SDS, 1 mM EDTA) and transferred to pre-filled tubes with 400-micron zirconium beads (OPS diagnostics, PFMB 400-100-34). The cells were homogenized using a FastPrep-24 machine from MP Biomedical for two cycles at 6 m/s for 30 seconds each, with a 1-minute incubation on ice between each cycle. 800 μL of Trizol reagent (Invitrogen, 15596026) was then added to the cell lysate and incubated on ice for 10 minutes, followed by two more cycles on the FastPrep machine. On ice, 100 μL of 1– bromo–3–chloropropane (Sigma Aldrich, B9673-200ML) was added to the cell lysate, inverted to mix, and incubated for 5 minutes. The tubes were spun down at 4°C for 14,000 x g for 15 minutes. 400 μL of supernatant was transferred to 400 μL of 100% ethanol. The RNA was purified using the Monarch Total RNA Miniprep Kit (NEB, T2010S) following the supplier’s instructions.
qRT-PCR evaluation of carotenogenesis pathway activity in response to light exposure
RNA samples were converted to cDNA using the Protoscript II First Strand cDNA Synthetsis Kit (NEB, E6560S) following the supplier’s standard protocol instructions. RT-qPCR was then performed using IDT’s PrimeTime Gene Expression Master Mix (Integrated DNA Technologies Inc., Skokie, Illinois, 1055772) and PrimeTime qPCR pre-mixed assay with probes and primers. Sequences can be found in Supplementary Table 2. Probes targeting genes-of-interest were labeled with FAM fluorophores and probes targeting the actin housekeeping gene were labeled with HEX fluorophores. Primers were designed on Benchling (https://benchling.com/) and optimal probe placement was determined using PrimerQuest (https://www.idtdna.com/PrimerQuest). The QuantStudio 6 Flex system was used, and instructions provided by IDT for the standard cycling protocol were followed.
RNA transcriptome sequencing and differential gene expression analysis
cDNA prep and sequencing protocol JGI method for cDNA preparation and sequencing. Raw FASTQ file reads were filtered and trimmed using the Joint Genome Institute (JGI) QC pipeline.28 Filtered reads from each library were aligned to the reference genome using HISAT version 0.1.4-beta.29 featureCounts was used to generate the raw gene counts file using gff3 annotations.30 Raw gene counts were used to evaluate the level of correlation between biological replicates using Pearson’s correlation and determine which replicates would be used in the DGE analysis. DESeq2 (version 1.10.0) was subsequently used to determine which genes were differentially expressed between pairs of conditions.31 The parameter used to call a gene differentially expressed between conditions was a P-value < 0.05.
Gene set enrichment analysis
Annotated X. dendrorhous CBS 6938 genome and gene count table from mapped transcriptomic reads were uploaded to ERGO 2.04 (Igenbio, inc).32 Genes were functionally enriched according to ERGO groups using GAGE5.33
Polymerase Chain Reaction (PCR)
All PCRs were done using Q5 2X Master Mix (NEB, M0492L). Primers were designed on Benchling (https://benchling.com/) and the NEB Tm Calculator (https://tmcalculator.neb.com/). All primers were ordered from IDT. PCR reactions closely followed NEB instructions. Briefly, reactions were done in 50 μL total volume; 25 μL Q5 Master Mix, 2.5 μL of each primer, X μL of template DNA (1 ng plasmid DNA or 100 ng genomic DNA), and 20-X μL nuclease free water (VWR 02-0201-0500). Reactions were run on a thermocycler with the following settings:
98°C for 30 sec
30 PCR cycles:
98°C for 10 sec
annealing temp. for 15 sec
72°C for 20 sec per kbp
72°C for 2 min
10°C hold
Gibson cloning
Gibson assembly was used to construct destination vectors, which were designed for TypeIIS-based cloning. Reactions closely followed instructions from the NEBuilder HiFi DNA Assembly Master Mix (NEB, E2621S). Fragments were amplified with PCR and designed to have 20-30 bp overlaps. The DpnI enzyme was used to digest template plasmid or X. dendrorhous genomic DNA according to the manufacturer’s instructions (NEB, R0176S). The amplified fragments were purified using the Zymo Clean & Concentrator Kit (Zymo, D4005), followed by dilution to 0.2 pmols for 2-3 fragments or 0.5 pmols for 4 or more fragments. The fragments, 10 μL of HiFi master mix, and nuclease free water were combined in a PCR tube for a 20 μL total reaction volume. The mixture was run on a thermocycler at 50°C for 60 minutes, with a hold at 10°C.
TypeIIS Cloning
A modular, hierarchical TypeIIS cloning scheme was used construct genetic designs for integration into X. dendrorhous. This scheme consisted of three levels; transcriptional parts (level 0), transcription units (level 1), and integrative pathways (level 2). Cloning reactions were based on the enzymes BbsI (Thermo Scientific, ER1011) or BsaI (NEB, R3733). Cloning reactions consisted of 1 μL of N parts, 1 μL of BsaI or BbsI, 1 μ L of 10X Ligase Buffer, 0.4 μL of T4 DNA ligase (Promega, M1794), and 7.9 - N μL of nuclease free water, where N is the number of parts. Reactions were run on a thermocycler at 37°C for 5 hours, 50 °C for 15 minutes, 80 °C for 20 minutes, and a hold at 10°C.
Parts and strain design
All genetic parts used in this study can be found in Supplementary Table 1.
Minimum Inhibitory Concentration Assay
X. dendrorhous cultures were diluted with YPD to an OD600 = 1.0 and 100 μL was pipetted into columns 2-11 of a 96-well plate (Corning, 3596). Column 11 served as a positive control of only X. dendrorhous culture. A negative control of 100 μL YPD was pipetted into column 12. The antifungals hygromycin (ThermoFisher, 10687010), zeocin (Jena Bioscience, AB-103S), geneticin (ThermoFisher, 10131-035), and nourseothricin (Jena Bioscience, AB-101L) were chosen for the experiment. Stocks of each antifungal were made in concentrations of 5120 μg/mL and 7680 μg/mL. Three replicates of antifungal concentration were made by pipetting 200 μL of an antifungal into three wells of column 1. A multichannel pipette was used to perform a serial dilution to column 10. At column 10, 100 μL of the culture and antibiotic mixture was pipetted and discarded, leaving 100 μL of mixture in every well. The 96-well plates were incubated at 21°C for 3 days. The OD600 was measured on a BioTek Synergy H1 plate reader with column 12 as a blank. Outliers were calculated using an interquartile range and were excluded in survival percent calculations.
X. dendrorhous transformation
Integrative pathway DNA for transformation was excised from level 2 vectors with a BsaI digestion. This digestion was comprised of 40 μL plasmid DNA, 5 μL 10X CutSmart Buffer, 2 μL BsaI, and 3 μL nuclease free water. The reaction was run at 37°C for 10 hours, 80°C for 20 minutes, and a hold at 10°C. The resulting DNA fragments were then purified with the Zymo Clean & Concentrator Kit and were ready for transformation. X. dendrorhous transformations were based on an electroporation method described by Visser et. al.34 X. dendrorhous was first streaked on a YPD agar plate and grown at 21°C for approximately 2 days. A single colony was used to inoculate 50 mL YPD media in a 125 mL Erlenmeyer flask. The cells were shaken at 21°C and 200 rpm for another 2 days. These cells were then used to inoculate 200 mL of YPD in a 1 L Erlenmeyer flask at OD600 = 0.02. The cells were grown at 21°C and 200 rpm until reaching OD600 = 1.2 (approximately 20 hours). From here, the cells were pelleted at 1500 x g for 5 min and resuspended in 25 mL of freshly made 50 mM potassium phosphate buffer (pH=7.0, Sigma-Aldrich, P8281-100G & P9791-100G) with 25 mM diothiothreitol (Acros Organics, 426380500). The cells were incubated at room temperature for 15 minutes, and then pelleted at 4°C for 5 minutes at 1500 x g. The cells were then washed with 25 mL of ice cold STM buffer (pH = 7.5, 270 mM sucrose, Millipore Sigma, 1.07651.5000, 10 mM Tris HCl, Alfa Aesar, J67233, 1 mM MgCl2, VWR, E525-100ML) and spun down again at 4°C for 5 minutes at 1500 x g. This wash step was repeated and followed by resuspension of the pellet in 500 μL of STM buffer. Now electrocompetent, the cells were divided into 60 μL aliquots and kept on ice until electroporation. 10 μL of transforming DNA was mixed with a 60 μL aliquot and transferred to an ice cold 0.2 cm electroporation cuvette (ThermoFisher, 21-237-2). Electroporation was performed with a Gene Pulser electroporator (Bio-Rad, 1998.018.1, 1998.018.2, & 1998.018.3) at 0.8 kV, 1000 ohms, and 25 μF. Immediately following the electric pulse, 500 μL of ice cold YPD was added to the cuvette. This mixture was then transferred to 4 mL of YPD in a 14 mL Falcon tube and grown overnight on a rotating drum. The next morning, the cells were spun down for 5 minutes at 1500 x g, resuspended in 500 mL of filtered water, and spread onto selection plates. The plates were incubated at 21°C until colonies appeared, which typically occurred in 2-3 days.
Astaxanthin extraction from wild-type and crtYB knockouts
X. dendrorhous CBS 6938 wild-type and crtYB knockout strains were grown for 3 days in the dark at 21°C and 200 rpm. 1 mL of culture was taken forward and centrifuged at 10,850 x g for 10 minutes. The cell pellet was washed with Milli-Q filtered water twice before resuspension in 1 mL of acetone (SigmaAldrich, 34850-1L). Cell and acetone mixtures were poured into pre-filled tubes of 400-micron zirconium beads (OPS Diagnostics, PFMB 400-100-34), then transferred to snap-cap microcentrifuge tubes (Eppendorf, 022363743). The Bullet Blender Storm Pro tissue homogenizer (NextAdvance, BT24M) was used at 12 m/s for 5 minutes to mechanically lyse the cells. Afterward, tubes were centrifuged for 10,850 x g for 10 minutes. 200 μL of supernatant was pipetted into a nylon syringeless filter (Whatman, UN203NPENYL) and transferred to a 2 mL chromatography vial (Agilent, 5181-3376) with a glass vial insert (5183-2085) and crimp top cap (Agilent, 8010-0051).
Ultra-high performance liquid chromatography detection of astaxanthin
UPLC analysis was performed using a modified method from Bohoyo-Gil et al.35 Modifications include an injection volume of 10 μL, usage of a Shimadzu Nexcol C18 1.8 μm column (Shimadzu, 220-91394-03), and analysis on the Nexera Series UPLC (Shimadzu; RF-20AXS, RID-20A, SCL-40, DGU-403, DGU-405, CTO-40C, SPD-M40, C-40 LPGE, LC-40D XS, SIL-40C XS).
Bioluminescence detection assay
LPA cultures were started in 2 mL of YPD media at OD600 = 1 and shaken for 24 hours at 200 rpm. X. dendrorhous was exposed to either a UV (λ = 400 nm) or dark (no LED light) condition. Bioluminescence was detected using the Nano-Glo Luciferase Assay System (Promega, N1110). Assay buffer and assay reagent were mixed in a 50:1 ratio to create to assay solution. Aliquots of 50 μL of X. dendrorhous cells were mixed with 50 μL assay solution in a 96-well plate (Corning, 3904). Bioluminescence was measured on a BioTek Synergy H1 plate reader.
FUNDING
This project is supported by a National Science Foundation CAREER award to EMY, award number 1944046. EET is supported by a National Science Foundation Graduate Research Fellowship.
The work (proposal 10.46936/10.25585/60001246) conducted by the U.S. Department of Energy Joint Genome Institute (https://ror.org/04xm1d337), a DOE Office of Science User Facility, is supported by the Office of Science of the U.S. The Department of Energy operated under Contract No. DE-AC02-05CH11231.
Development of Prymetime was supported by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA) under Finding Engineering Linked Indicators (FELIX) program contract #N66001-18-C-4507. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of ODNI, IARPA, or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation therein.
AUTHOR CONTRIBUTIONS
EET, JHC, and EMY conceived of the study. EET and JHC performed all experiments and analysis save the sequencing and initial processing of RNA-seq data. CBM and GTN constructed the light plate apparatus (LPA). KM acquired microscope images of red yeast. AL, SM, and IG performed transcriptomic sequencing and initial processing of RNA-seq data. EET and EMY wrote the manuscript.
* Emma E. Tobin and Joseph H. Collins contributed equally.
CONFLICT OF INTEREST
The authors declare that there are no conflicts of interest regarding this research study.
