BLINCAR: a reusable bioluminescent and Cas9-based genetic toolset for repeatedly modifying wild-type Scheffersomyces stipitis

ABSTRACT Scheffersomyces stipitis is a yeast that robustly ferments the 5-carbon sugar xylose, making the yeast a valuable candidate for lignocellulosic ethanol fermentation. However, the non-canonical codon usage of S. stipitis is an obstacle for implementing molecular tools that were developed for other yeast species, thereby limiting the molecular toolset available for S. stipitis. Here, we developed a series of molecular tools for S. stipitis including BLINCAR, a Bio-Luminescent Indicator that is Nullified by Cas9-Actuated Recombination, which can be used repeatedly to add different exogenous DNA payloads to the wild-type S. stipitis genome or used repeatedly to remove multiple native S. stipitis genes from the wild-type genome. Through the use of BLINCAR tools, one first produces antibiotic-resistant, bioluminescent colonies of S. stipitis whose bioluminescence highlights those clones that have been genetically modified; then second, once candidate clones have been confirmed, one uses a transient Cas9-producing plasmid to nullify the antibiotic resistance and bioluminescent markers from the prior introduction, thereby producing non-bioluminescent colonies that highlight those clones which have been re-sensitized to the antibiotic and are therefore susceptible to another round of BLINCAR implementation. IMPORTANCE Cellulose and hemicellulose that comprise a large portion of sawdust, leaves, and grass can be valuable sources of fermentable sugars for ethanol production. However, some of the sugars liberated from hemicellulose (like xylose) are not easily fermented using conventional glucose-fermenting yeast like Saccharomyces cerevisiae, so engineering robust xylose-fermenting yeast that is not inhibited by other components liberated from cellulose/hemicellulose will be important for maximizing yield and making lignocellulosic ethanol fermentation cost efficient. The yeast Scheffersomyces stipitis is one such yeast that can ferment xylose; however, it possesses several barriers to genetic manipulation. It is difficult to transform, has only a few antibiotic resistance markers, and uses an alternative genetic code from most other organisms. We developed a genetic toolset for S. stipitis that lowers these barriers and allows a user to deliver and/or delete multiple genetic elements to/from the wild-type genome, thereby expanding S. stipitis’s potential.

T he need for a renewable, carbon-neutral liquid fuel has energized the field of lignocellulosic ethanol production research. Developing technologies that allow yeasts to efficiently ferment the sugars found in wood, leaves, and agricultural stover into ethanol in a cost-effective manner may provide alternatives to gasoline and diesel (1). Although Saccharomyces cerevisiae is one of the most researched yeasts for ethanol fermentation and has an extensive set of molecular tools that have been developed

Bioluminescence gene selection
The gene for click beetle green luciferase "CBG99" (Promega, Madison, WI) was chosen as the bioluminescence reporter gene for this work because the protein's green emission was robust when produced by yeast on rich, solid media and can be visually distin guished from colonies producing red-emitting luciferases should the need for dual-color assays arise (16). A codon-optimized version of Promega's CBG99 gene was developed for expression by S. stipitis and called coCBG for this work.

Plasmid construction
All plasmids (see Table 1; Table S1) were constructed using a traditional restriction enzyme-based cloning approach. See supplement for a full description of construction methods; but briefly, heterologous sections of DNA (inserts) were either cut directly out of synthesized plasmids or amplified from S. stipitis genomic DNA or template plasmids by PCR using primers (see Table 2) that added specific restriction sites to the ends of the PCR product. Once a heterologous segment (insert) and plasmid vector were digested with the appropriate restriction enzyme(s), they were run on 1% agarose gel and bands of appropriate sizes were isolated, DNA purified, and ligated together. Ligation products were transformed into competent TOP10 E. coli. Each piece of the constructed plasmid was sequentially added in this way.

YUM1 Cas9 target design
The 23 base YUM1 sequence AAAGTTCTACTTAGTCCCCGGGG (underlined is the PAM) was designed as a Cas9 targeting sequence based on the sequence's absence in the S. stipitis and S. cerevisiae genomes. The 3′ portion of YUM1 sequence "CCCCGGGG" was reported to not exist in the genome of S. cerevisiae (19) and also provided a PAM for the intended Cas9 targeting. We used that sequence as a seed for the keeSeek algorithm (20) to find 23-mers that were nonexistent in the S. cerevisiae genome. The Yeast Unique Motif 1 (YUM1) was chosen as the top candidate based on it also having smaller strings of nucleotides that were also nonexistent or rare in yeast genomes, and it not causing hairpins when included in primers 38, 39, 49, 50, 54, and 55 (see Table 2).

Transformations
Chemically competent E. coli (TOP10) were transformed using a standard chemical transformation procedure. S. stipitis was transformed using a modified lithium ace tate/PEG procedure. This consisted of growing a diluted culture of S. stipitis in 10 mL of YPD at 30°C shaking (180 rpm) until it reached OD 600 of 0.8-1.0, pelleting the cells at <3,000 rcf, and washing the cells with 1 mL 1× TE, 0.1 M lithium acetate. The cells were pelleted and resuspended in 100 µL of 1× TE, 0.1 M lithium acetate, and incubated at 30°C rolling for 1 h. Then, 5-10 µg of linearized or circular plasmid DNA (in 20 µL) was added, along with 15 µL of 10 mg/mL herring sperm DNA (Promega, Madison, WI), and 700 µL of 40% PEG 8000 , 1× TE, 0.1 M lithium acetate. For the case of ura3 gene deletion, 120 µg of linearized plasmid DNA (in 20 µL) was used. This mixture of DNA, cells, and PEG/TE/LiOAc was rolled at 30°C for 30 min, and afterward, the mixture was heat shocked at 42°C for 5 min, centrifuged for 10 s at 10,000 rcf, and the supernatant was removed by

PCR and PCR screening
Phusion high-fidelity DNA polymerase (Thermo Fisher, Waltham, MA) was used for PCR when CDSs were amplified for cloning purposes. For all other PCR, GoTaq (Promega, Madison, WI) was used. PCR was used to confirm S. stipitis genomic modification by BLINCAR. Introduction of pWR101, pWR102, and pWR103 (see Table 1) was confirmed using primer 57 paired with primers 69, 32, and 70, respectively (see Table 2). Targeting BLINCAR to replace the URA3 gene was confirmed upstream by pairing primer 71 (which targets a genomic sequence upstream of the introduced element) with primer 72 (which targets a portion of the coKanMX CDS within BLINCAR). BLINCAR replacement of URA3 was confirmed downstream by pairing primer 73 (which targets a portion of coCBG within BLINCAR) with primer 74 (which targets a genomic sequence downstream of the introduced element). Positive control PCR to confirm the presence of S. stipitis genomic DNA used either primers 75 paired with 76 to give a 1 kb product or primers 33 paired with 34 to give a 2 kb product.

Imaging
Bioluminescent and brightfield imaging of yeast grown on solid media in 10 cm Petri dishes was performed using a ChemiDoc MP (Bio-Rad, Hercules, CA). Bioluminescent imaging was recorded using 4 × 4 binning for exposure times of 1-5 min. Microscopic imaging was performed using an Olympus BX60 fluorescent microscope and images were captured from a DP72 color CCD camera (Olympus, Tokyo, Japan). Fluorescence was achieved from a U-MWB filter cube (450-480 nm band-pass excitation filter, 500 nm dichroic mirror, and 515 nm long-pass emission barrier filter).

Construction of a versatile bioluminescence reporter plasmid for S. stipitis
We desired a plasmid with room to sequentially clone 10 or more genetic elements [three coding sequences each with its own S. stipitis promoter and terminator, plus a plasmid-maintenance feature like an autonomous replicating sequence (ARS) or integrating element]. Since long, heterologous elements often contain problematic restriction sites that limit cloning designs, we realized the multiple cloning sites of common cloning vectors like pUC18 and pBluescript were too limited to provide the extent and flexibility we sought for our cloning platform, so we synthesized pAllet, a cloning vector with an extensive multiple cloning site (MCS) containing over 50 unique restriction sites (Fig. 1A). In our design of pAllet's MCS, we sought to maintain the order of some of the common restriction sites found in the pUC and pBluescript series of or integration locus) was added for many of the plasmids used in this work.
Research Article mSphere plasmids so that elements which have been previously cloned by others into these common vectors could more seamlessly be transferred to pAllet if needed. Namely, the order of SalI, XbaI, BamHI, KpnI, and SacI was preserved from pUC18, while the order of ApaI, XhoI, SalI, HindIII, EcoRI, PstI, SacII, and SacI was preserved from pBluescriptII SK+. Another problem with common cloning vectors is that often useful restriction sites are immediately adjacent to one another or even overlapping in the MCS thereby not allowing efficient use of both sites simultaneously. To alleviate this problem in pAllet and extend the range of useful cloning sites, we interspersed less common sites and blunt-cutting sites between common staggered-cutting sites that are often used for cloning.
Through the course of this work, we used pAllet to construct over 100 derivative plasmids, drawing on a modular approach. Figure 1B shows modules 1, 2, and 3, where we introduced up to three heterologous genes. Adjacent arrows in each module point to restriction sites in pAllet (from Fig. 1A), where a gene element (promoter, CDS, or terminator) was introduced. SacI (bottom arrow in Fig. 1B) was used to provide the plasmid with a replication or integration element for stability in S. stipitis. We chose the restriction sites BsrGI, KpnI, and FseI as promoter-CDS junction sites for modules 1-3, respectively. These sites were chosen because their 6 bp recognition sequences end with three nucleotides that are commonly found immediately before the start codon of S. stipitis CDSs. Given that the nucleotide sequence near the start codon influences efficient translation in yeast (21), we surveyed sequences of 60 S. stipitis genes suspected to be regularly expressed in common growth environments to help inform our selection of these junction sites (see Table S2).
We constructed a basic versatile bioluminescent reporter episomal shuttle vector, pWR9 (see Fig. 2A), for S. stipitis by introducing native regulatory elements into pAllet to control the expression of the HPH hygromycin resistance gene in module 1 and the green bioluminescent reporter gene CBG99 in module 2, both of which had been and replica-plated onto nutritive agar media containing 2% glucose, 2% galactose, or 2% xylose. Bioluminescent imaging for 5 min exposure (top right and lower panels) reveals that P TDH3 promotes transcription of the coCBG reporter on all three sugars, but P XKS1 only does so on xylose, and P GAL1 only does so on galactose.
Research Article mSphere codon optimized for S. stipitis to bypass the CUG codon restriction. We also added a 1.1 kb ARS found on S. stipitis Chromosome 1 (first reported as ARS2 by Yang et al. in 1994) for purposes of replication and maintenance as an episomal plasmid in S. stipitis (7). The unmodified CBG99 sequence from Promega's pCBG99-basic plasmid (in which S. stipitis translates the CUG codons as serine instead of leucine) was insufficient to produce bioluminescence in S. stipitis (see Fig. S1). Therefore, codon-optimization for 18 CUG codons to UUG codons was necessary for bioluminescence in S. stipitis.
Once we confirmed transformation, selection, and bioluminescence of pWR9 (see Fig.  S1) we made derivatives of the basic reporter plasmid by swapping various elements to compare antibiotic selectable markers, promoter strength, replication elements, and plasmid stability (see Table 1). In addition to pWR9, which used coHPH to confer hygromycin resistance, we also created pWR10, pWR23, and pWR24 to compare plasmid selection using codon-opti mized versions of ShBle (resistance to zeocin), KanMX (resistance to G-418), and Nat (resistance to nourseothricin). Nourseothricin was found to be quite toxic to S. stipitis and coNat was unable to provide any resistance/selection regardless of the concentration of antibiotic. And although coShBle and coKanMX provided sufficient selection for early identification of transformed candidates in the first day or two of growth, an inherent natural resistance/tolerance to zeocin and G-418 allowed some false positive colonies to form and eventually overtake the plate (data not shown). Hygromycin was found to be sufficiently toxic to S. stipitis with little to no natural tolerance, and the resistance provided by coHPH produced a clear selection differential in transformants (no false positives) which persisted for days to weeks of growth after inoculation.
We used the coCBG gene in module 2 of pWR9 derivatives as a reporter gene to evaluate the relative strength of seven S. stipitis promoters whose homologs in S. cerevisiae have been shown to have a strong constitutive expression (22). Figure 2B shows bioluminescence and quantification from patches of colonies transformed with pWR9, pWR54, pWR55, pWR56, pWR77, pWR78, and pWR81, which used the promot ers of TDH3, CCW12, PGK1, HHF1, TEF1, ADH1, and CYC1 (respectively) to drive coCBG expression. The TEF1 promoter produced the strongest bioluminescence. Promoters from TDH3, CCW12, ADH1, and CYC1 all produced about the same moderate amount of bioluminescence, and promoters from PGK1 and HHF1 produced the least biolumines cence.
We also tested bioluminescence from S. stipitis transformed with pWR11 and pWR22, which were pWR9 derivatives that used the sugar-inducible promoters of GAL1 and XKS1 (respectively) to drive coCBG expression. As shown in Fig. 2C, bioluminescence (at a moderate level) was only produced from the GAL1 promoter reporter when cells were grown on media containing galactose; and similarly, bioluminescence was only produced from the XKS1 promoter reporter when cells were grown on media contain ing xylose. Taken together, these data show that pWR9 and its derivatives are useful reporters for evaluating promoter strength and conditional gene regulation of S. stipitis grown on solid media.
The data from Fig. 2 show that the ARS sequence from Chromosome 1 is sufficient to maintain the plasmid in replicating cultures under selection pressure; however, to test whether the plasmid can be maintained in the absence of antibiotic selection and whether genomic integration of the plasmid could allow maintenance of transgenic elements without selection, an experiment was conducted with episomal and integrated plasmids. pWR16 is a derivative of pWR9 in which we swapped the ARS element for a copy of the 2.7 kb S. stipitis genomic region containing the URA3 gene (which also included a unique EcoRV site in the middle of this region that was used for plasmid linearization). Survival and bioluminescence from S. stipitis harboring the pWR9 episomal plasmid was compared to that of S. stipitis harboring the pWR16 integrated plasmid after growing overnight in YPD broth with and without hygromycin antibiotic. Figure  S2 shows that the pWR9 episomal plasmid maintained by a chromosomal ARS was rapidly lost (within 12 generations) during culture if not kept under selective pressure; however, the pWR16 integrated plasmid was stably maintained in absence of selection. A linearized plasmid containing homology to an S. stipitis genomic sequence is capable of stably integrating into the host's chromosome by way of HR. The difference in plasmid stability between these two maintenance elements is a phenomenon that we exploit for Cas9-based strain alteration later in this work.
Bioluminescence to report Cas9 activity A functioning Cas9 system requires properly produced Cas9 protein and small guide RNAs to be at sufficient concentrations to act on an intended DNA target. If any step in the production, assembly, or targeting phases of this system is insufficient, then Cas9 efficiency suffers. Two common approaches to demonstrate successful Cas9 activity are to use it to (i) nullify a detectable gene product by cutting a target gene that is briefly retracted before repair by NHEG, or (ii) repair a nonfunctioning gene to produce a detectable gene product by cutting a target gene that is then repaired by HR using a provided repair template. Both approaches have obstacles in organisms that are difficult to transform (like S. stipitis). Detecting a rare loss of function (especially if subtle or heterogeneous in a colony) can require screening hundreds of colonies, and often requires sequencing to confirm the loss of function was due to Cas9-targeted cuts and not spontaneous mutations. And repairing a nonfunctional reporter by repair template requires a successful co-transformation of the repair template along with the delivery of the other Cas9-dependent genetic elements.
To get around these obstacles, we augmented our bioluminescent reporter to create an integrating zeocin-selectable positive reporting system for Cas9 activity that requires only single selectable/detectable transformation events. The bioluminescence-based Cas9 reporter, pWR58, can be stably integrated into the S. stipitis genome and selected by growth on zeocin-containing media. In its initial state, pWR58 does not produce a bioluminescent product because its luciferase gene (coCBG) contains an interruption cassette (i.e., the zeocin selectable marker) nested within a coCBG internal duplication where the middle third of the coCBG CDS is duplicated on both sides of the interruption cassette (see Fig. 3A). If functional Cas9 is present and contains the necessary sgRNA sufficient to target a designed artificial 20 bp DNA target (yeast unique motif, "YUM1") which is natively absent in the S. stipitis genome that flanks the interruption cassette, then Cas9 can cut the integrated reporter at either (or both) of these target sequences, thereby spurring HR to repair the break by recombining the coCBG within its duplicated middle third and forming an intact, complete coCBG CDS (see Fig. 3B).
To provide S. stipitis, the necessary components to affect the integrated pWR58 reporter plasmid (namely a CUG-yeast-optimized Cas9 gene and a sgRNA encoding sequence), we created pWR63 (see Fig. 3C). This versatile episomal plasmid contained the CaCas9 CDS (18) optimized for C. albicans (which bypasses the CUG codon restriction in CUG group yeast like S. stipitis) driven by the CWW12 promoter from S. stipitis determined in this work to be moderately strong and constitutive (see Fig. 2B). Guide RNA expression for many engineered Cas9 systems is driven from POL III-type promoters like the SNR52 promoter, but confirming the boundaries and strengths of these types of promoters can be challenging and labor-intensive since their RNA products are not translated. To make a versatile sgRNA expression system that uses RNA expressed from S. stipitis POL II-type promoters (like those tested and confirmed in Fig. 2), we adopted a sgRNA production system published by Ng and Dean, where the transcribed targeting crRNA/ tracrRNA element is preceded by a tRNA sequence and followed by a hepatitis delta virus (HDV) ribozyme sequence (23). Once transcribed by POL II, the tRNA portion of the RNA is cleaved and removed by native RNAse Z (24) and the HDV portion is cleaved and removed by self-cleavage (25), leaving a functional sgRNA with precise (and uncapped) 5′-and 3′-ends. Dimerized oligos bearing the desired 20 nucleotide sgRNA-encoding sequence and three base overhangs can be added to pWR63 by cutting the plasmid with SapI, removing a small "stuffer" sequence and producing corresponding three-base overhangs for directional cloning of the dimerized oligo (see Fig. 3C). For this work, we introduced the sgRNA encoding sequence necessary to target the artificial YUM1 target sequence. We called this derivative pWR63(YUM1).
The pWR58 reporter, once affected by Cas9 [from pWR63(YUM1)], produced a bioluminescent product that could be visually detected in colonies by a chemilumi nescence imaging camera (see Fig. 3D). A strain of S. stipitis which had been stably transformed with the pWR58 reporter plasmid was subsequently transformed with the pWR63(YUM1) episomal plasmid to produce Cas9 and cut the YUM1 targeting elements in the pWR58 construct. After a 3-h recovery period post-transformation, these yeasts were plated on a YPD nutrient plate containing hygromycin [which selected for the pWR63(YUM1) plasmid] and luciferin (the substrate for the luciferase). Figure 3D shows Research Article mSphere that some of the transformed yeast grew into solid, intensely glowing colonies (e.g., Fig.  3D, red arrows), while other transformed colonies exhibited a sectoring phenotype or did not show detectable levels of bioluminescence at all (see Fig. 3D inset which shows the three colonies contained in the red box). Sectoring could result if colony-forming units were comprised of more than one cell in which only a subpopulation (or a single cell) of those cells underwent the Cas9-directed reporter gene modification. Alternatively, the Cas9-directed reporter gene modification could have occurred after plating in one cell of the early growing colony. Transformation with a plasmid that lacked either the YUM1 encoding sgRNA sequence or the CaCas9 gene resulted in colonies that survived selection but were not bioluminescent (see Fig. S3). This assay confirmed that Cas9 and the guide RNA were expressed in S. stipitis at sufficient levels to target and cut the artificial target sequence in the reporter and that HR repaired the broken chromosome in instances that were easily identifiable.

BLINCAR for iterative additions to the wild-type S. stipitis genome
Given that wild-type S. stipitis is only susceptible to the toxicity of a few antibiotics (e.g., hygromycin, zeocin, and G-418) for which codon-optimized resistance genes have been found to circumvent, there are not many selectable markers available if one desires extensive augmentation to the wild-type S. stipitis genome. However, the Cas9 actuated recombination that was shown between the coCBG repeats flanking a selectable marker (Fig. 3) highlighted a way the limited set of selectable markers could be reused for potentially unlimited genomic modification.
To demonstrate this potential, we developed pWR88, a versatile, reusable plasmid for integrating desired genetic payloads repeatedly to the S. stipitis genome (Fig. 4A). This plasmid consisted of a selectable and detectable element, comprised of a codonoptimized G-418 resistance gene (coKanMX) and the coCBG luciferase gene, flanked by two 0.6 kb heterologous repeats intended for Cas9 actuated recombination (the 0.6 kb repeats were designated "CAR" sequences). We chose three native S. stipitis genes (EGC3, GSC2, and CRF1) in which to target the plasmid's integration for this demonstration. These genes were chosen for integration targets because duplication of these genes (a consequence of this form of gene-targeted integration) was predicted to be well-toler ated and not cause adverse phenotypes. Moreover, they were chosen because they each possessed a single EcoRV restriction site that could be used for plasmid linearization. Sections of DNA of approximately 2 kb containing each of these genes (EGC3, GSC2, and CRF1) were PCR amplified from wild-type S. stipitis and individually cloned into pWR88 at its SacI site to create the integrating derivatives pWR89, pWR90, and pWR91, respectively (Fig. 4A). Genetic payloads can be added to these derivatives on either side of the CAR regions using one or more of the unique restriction sites remaining in the plasmid (indicated in bold in Fig. 4A). Figure 4B illustrates how pWR88 derivatives can be used repeatedly to add multiple genetic payloads to the S. stipitis genome. Once linearized and transformed, the entire plasmid (including payload) integrates into the S. stipitis chromosome through HR. And although rare, successful integration produces cells that exhibit G-418 resistance and bioluminescence (Fig. 4B, top). If a second, separate genetic payload needs to be added to the modified strain, the antibiotic resistance and bioluminescence markers from the first integration can be removed by Cas9 cutting YUM1 targets near the CAR regions of the integrated plasmid (Fig. 4B, middle). Once cut, HR between the CAR regions repairs the broken chromosome, excluding the G-418 selectable marker and bioluminescence gene while leaving in place one CAR element and the genetic payload (Fig. 4B, bottom).
We demonstrated repeated transgene delivery using a genetic payload of the mCherry gene codon-optimized for C. albicans (17) under the control of the S. stipitis CWW12 promoter. The coCherry gene was cloned into pWR89, pWR90, and pWR91 plasmids between their SbfI and SacII sites to create the derivatives pWR101, pWR102, and pWR103, respectively. Successful transformants of pWR101 were selected on YPD plates containing G-418 and luciferin; however, after 2 days of growth at 30°C, a background of non-transformed cells grew up around the transformed colonies (Fig. 4C). Bioluminescence from the transformants, however, allowed us to discriminate positive colonies and streak for single transformed colonies. This was done in case the initial sampling of the colony was inadvertently mixed with non-transformed cells (Fig. 4D). The single colony identified by the red arrow in Fig. 4D was subcultured for confirmation and continued work. The presence of the coCherry gene was confirmed in the strain's genomic DNA by PCR (data not shown) and evidence of red fluorescence was detec ted from the cells (Fig. 4E; Fig. S4). After confirmation of coCherry delivery, the select able and bioluminescence markers were removed from the genome by transforming the culture with pWR63(YUM1) and selecting for transformants on YPD plates contain ing hygromycin and luciferin. Successful removal of the G-418 resistance marker and coCBG bioluminescence gene by Cas9-actuated recombination appeared as non-glow ing colonies on the hygromycin selective plates (Fig. 4F, red arrow). The process was repeated for the introduction of pWR102 and pWR103 subsequently. Retention of each iteration's addition of coCherry was confirmed by PCR using a 5′ primer specific for An instance for which Cas9-actuated recombination removed the coCBG gene appears as a white colony (i.e., no bioluminescence) indicated by the red arrow.
This colony was chosen for further work. (G) PCR screen results (top) for a wild-type strain of S. stipitis and three rounds of sequential addition of the mCherry payload (diagramed bottom). Blue arrows in the diagram show primer binding sites to generate PCR products of varying sizes (approximately 1.5, 2, or 3 kb) depending on the integration element used for that round. A control set of primers (cont) was used to amplify a 1 kb portion of the URA3 gene from all candidate strains. The first round of treatment (1st) delivered pWR101, detectable by a 1.5 kb band. The second round (2nd) delivered pWR102, detectable by a 3 kb band.
And the third round (3rd) delivered pWR103, detectable by a 2 kb band.
Research Article mSphere the coCherry transgene and a 3′ primer specific to each iteration's integration element (either EGC3, GSC2, or CRF1) such that each iteration would produce a PCR product of a different size (Fig. 4G). The first iteration contained only coCherry delivered by pWR101. The second iteration contained that from pWR101 but also the one provided by pWR102. And the third iteration contained coCherry from all three (pWR101, pWR102, and pWR103). Red fluorescence from these modified strains appeared to increase in intensity with the addition of one and two copies of coCherry, but the addition of a third copy did not seem to increase fluorescence above that provided by two copies (see Fig.  S4).

BLINCAR for marker-less gene deletions in wild-type S. stipitis genome
Strain engineering goals often benefit from the removal of multiple genes from the microbe's genome, but established CRISPR Cas9 methods for repeated gene disruption are problematic in wild-type strains that (i) have only a few selectable markers, (ii) resist transformation, and (iii) predominantly repair double-stranded DNA breaks by rapid/ efficient NHEJ. We tested whether our BLINCAR approach could permit targeted gene disruption in conjunction with rapid identification of genomic modification followed by marker removal. For proof-of-concept, we targeted S. stipitis's URA3 gene for removal.
To accomplish this, we designed pWR95, a modified version of pWR88 where sections of upstream and downstream genomic homology (that flank an intended target gene) can be directionally introduced on either side of the BLINCAR element. Approximately 2 kb of URA3 upstream homology was added to pWR95 at the plasmid's unique XhoI and XbaI sites, and a similar length of URA3 downstream homology was added at the unique NdeI and SacII sites. EcoRV sites (which were otherwise absent) were added to the extremes of these homology portions for purposes of extracting the linear homologyflanked BLINCAR element from the plasmid for transformation into S. stipitis. Figure 5A shows pWR95(URA3), this pWR95 derivative with URA3 homology flanking the BLINCAR element.
Some yeasts like S. cerevisiae are much easier to transform than S. stipitis (requiring only a few micrograms of DNA), and efficiently undergo homologous recombination allowing much shorter arms of homology be used for their gene integration cassettes (in some cases as short as 40 bp). For such yeast, it is common to produce gene integration cassettes using single reaction PCR, or even overlap extension PCR. However, in our hands, S. stipitis only reliably transformed with gene-targeting integration when we used large amounts of DNA (60-120 µg) and much longer arms of homology (1500-2000 bp). Transformation and targeting frequencies correlated with greater amounts of DNA and longer homology arms used. This performance was consistent with reports by others (2). Because our linearized BLINCAR gene targeting cassettes were about 10 kb long, and because we required 100 µg or more of DNA per transformation, we chose to use a plasmid-based approach rather than PCR to construct and produce the gene integration cassettes. This allowed us to produce large amounts of our long constructs in E. coli and purify them at high concentrations with maxiprep techniques. Figure 5B shows how BLINCAR gene disruption is theorized to work. We linearized the construct and transformed it into wild-type S. stipitis where rare double HR events replaced the targeted gene (URA3) with the portion of the BLICNCAR construct between the regions of targeted homology (Fig. 5B, top). Successful transformants were G-418 resistant and bioluminescent (Fig. 5C). Eighteen candidate colonies were patched for clonal isolation (Fig. 5D, top) and were PCR-screened for targeted integration using primers that target portions of the chromosome outside of the homology paired with those that target regions within the BLINCAR element (Fig. 5B, middle). All candidates produced a positive control PCR product (2 kb) however only isolate #14 produced the 3.4 kb PCR product showing evidence of targeted integration (Fig. 5D). For isolate #14, PCR was used to confirm the integration of both upstream (3.4 kb product) and downstream (3.2 kb product) portions of the BLINCAR construct (Fig. 5D, bottom). Once gene replacement had been confirmed, the strain was transformed with the episomal plasmid pWR63(YUM1) which provided Cas9 and guide RNA that cut the BLINCAR construct once or twice (Fig. 5B, middle) and allowed HR between CAR regions to repair the chromosome (Fig. 5B, bottom). Candidate colonies that had likely undergone such Cas9-actuated recombination were identified by loss of bioluminescence (Fig. 5E, red arrows). Six candidate colonies were patched for isolation and confirmation of stable loss of bioluminescence (Fig. 5F, top, isolates 1-6) along with two colonies that retained their bioluminescence (Fig. 5F, top, isolates 7 and 8). To confirm the removal of the selectable marker and bioluminescence gene, genomic DNA from isolates was PCR-screened using a pair of primers that targeted central portions of the upstream and downstream URA3 homology (Fig. 5B, bottom and Fig. 5F, bottom), and sequenced to show a remaining CAR site where URA3 used to be. The sensitivity of S. stipitis to G-418, hygromycin, and uracil auxotrophy was tested for each of the three phases of the BLINCAR URA3 replacement process (see Fig. S5). Phase 1 was the wild-type strain after stable integration of pWR95(URA3). Phase 2 was the strain after transformation with the pWR63(YUM1) episomal plasmid. And phase 3 was the strain after the pWR63(YUM1) episomal plasmid was cured from the strain. A fivefold serial dilution for wild-type S. stipitis and that for each phase of treatment was plated on YPD where each grew normally. However, when the same array of yeast was plated on YPD containing antibiotics or minimal media lacking uracil; only yeast from phase 1 grew in the presence of G-418, only yeast from phase 2 grew in the presence of hygromycin, and only wild-type S. stipitis grew in the absence of uracil, even after 3 weeks of growth. When uracil was supplemented into the minimal media, all yeast grew; however, those with the URA3 gene removed grew very slowly, requiring several weeks to attain cell density observable on the plate. These yeasts exhibit growth as expected for each stage of modification, in agreement with the model that (i) a G-418 antibiotic resistance marker replaced the URA3 gene thereby conferring uracil auxotrophy, (ii) a Cas9 expression plasmid provided temporary hygromycin resistance while Cas9 targeted the G-418 resistance marker for removal, and (iii) the Cas9 expression plasmid was eliminated from the yeast resulting in a strain that was ura3 deficient and re-sensitized to G-418 and hygromycin. Finally, we demonstrated that this ura3 auxotrophic strain could be used for URA3 selection for pWR96, a derivative of pWR9 which was a biolumi nescence plasmid bearing a functional copy of URA3 (Fig. 5G).

Conclusions
Although some approaches using CRISPR/Cas9 technology allow seamless genome editing that does not require selection markers to promote successful gene modification [e.g., see reference (14)], such approaches in S. stipitis have the drawbacks of low efficacy in wild-type strains or require specialized/engineered host strains to improve efficiency. Seamless/markerless gene alteration typically requires DNA sequencing to identify and confirm successful alteration, and if the genetic modification occurs at low frequencies, then researchers must sequence many candidates just to find one successful alteration. Our approach was to use antibiotic resistance genes as selectable markers for genetic alterations rather than seamless gene editing. We acknowledge that genetic alterations through this method may be rare in wild-type strains of S. stipitis, but for those wanting to work with a wild-type strain, antibiotic resistance markers offer three advantages: (i) selection for candidates that possess (or have lost) the marker thereby minimizing the impact for low frequency events by allowing those candidates to either outperform their counterparts or display a detectable phenotype, (ii) retention/maintenance of the genetic trans-element allowing the user to control the environment under which interruption cassettes or plasmids are retained or lost in the host strain, and (iii) limited protection of the culture from contaminating microbes that are not resistant to the antibiotics used.
However, using antibiotics and antibiotic resistance markers also comes with drawbacks. First, antibiotics themselves can be costly, and maintaining plasmids in S. stipitis requires constant selection pressure therefore constant antibiotic use. Second, without proper care in the laboratory, using antibiotics in research can contribute to antibiotic resistance in environmental pathogenic microbes which negatively impacts society. Furthermore, making complex genetic alterations to one's host strain can require the use of multiple antibiotics and can require the creation of a multi-drug-resistant microbe. Our work sought to gain the benefits of using antibiotic selectable markers to promote intentional alterations to wild-type strains of S. stipitis while minimizing the drawbacks by removing all selectable markers in the strain after the alterations were made and confirmed.
This work contributes to the growing pool of molecular tools the research community has for investigating and manipulating CUG group yeast like S. stipitis and C. albicans. And although our work focused specifically on S. stipitis, some of the products (like pAllet) are broadly applicable to molecular biology research in general, and many other products from this work can easily be made to function in other yeasts (like S. cerevisiae) by swapping out promoters or in some cases replication features. The modular approach we took in constructing the plasmids from this work makes these alterations possible using traditional restriction enzyme-based cloning methods.
In this work, we produced pAllet, a small, high-copy cloning vector with an MCS containing over 50 unique restriction sites. We used pAllet to build the diverse set of molecular tools in this work, but it can benefit many other researchers who need a cloning vector with an extensive MCS.
First, we produced pWR9, a versatile bioluminescence reporter for S. stipitis to evaluate three things. We evaluated the stability of episomal maintenance versus chromosomal integration and found that integration allows stable maintenance of the introduced plasmid without the need for selection; however, the episomal plasmid (maintained by ARS1) was rapidly lost if not kept under selective pressure. Second, we used pWR9 to test candidate antibiotic resistance genes in S. stipitis and found that the coHPH sequence provides strong and discriminative selection on hygromycin, while coKanMX provides moderate early selection that wanes after a few days of growth on plates. Third, we used pWR9 to evaluate candidate promoter strength and conditional expression, for which we go on to use those confirmed promoters to drive the expression of other heterologous genes in later plasmids.
We produced pWR58, an integrating, zeocin-selectable, positive bioluminescence reporter for Cas9 activity in S. stipitis. We used it to demonstrate the sufficient expression, activity, and cooperation of Cas9 and the Pol-II-derived sgRNA elements to target our artificial YUM1 sequence in S. stipitis. The pWR58 plasmid could be modified by others to test similar components and sgRNA targets of interest in S. stipitis (besides YUM1), as well as modified by others to function in other yeasts besides S. stipitis.
Finally, we produced BLINCAR, a two-component system where first one delivers an antibiotic selectable/bioluminescent element that can be used to either provide an exogenous genetic payload to the S. stipitis genome or can be used to replace a native (non-essential) genetic element thereby deleting the targeted sequence; then secondly, one introduces a temporary, episomal plasmid that produces Cas9 and sgRNAs to target the selectable/bioluminescent element for removal. Such a system can be used repeatedly to extensively modify the S. stipitis genome. We demonstrated this potential by repeatedly adding additional copies of coCherry to the genome, and later by deleting the URA3 gene and re-sensitizing the deleted strain to all antibiotics used.