A Modular Cloning Toolkit Including CRISPRi for the Engineering of the Human Fungal Pathogen and Biotechnology Host Candida glabrata

The yeast Candida glabrata is an emerging, often drug-resistant opportunistic human pathogen that can cause severe systemic infections in immunocompromised individuals. At the same time, it is a valuable biotechnology host that naturally accumulates high levels of pyruvate—a valuable chemical precursor. Tools for the facile engineering of this yeast could greatly accelerate studies on its pathogenicity and its optimization for biotechnology. While a few tools for plasmid-based expression and genome engineering have been developed, there is no well-characterized cloning toolkit that would allow the modular assembly of pathways or genetic circuits. Here, by characterizing the Saccharomyces cerevisiae-based yeast molecular cloning toolkit (YTK) in C. glabrata and by adding missing components, we build a well-characterized CgTK (C. glabrata toolkit). We used the CgTK to build a CRISPR interference system for C. glabrata that can be used to generate selectable phenotypes via single-gRNA targeting such as is required for genome-wide library screens.


Materials
Phire Green Hot Start II PCR Master Mix was used for all PCR reactions and was purchased from Thermo Fisher (#F126L). Restriction enzymes BsaI-HF®v2 (#R3733S) and BsmBI-v2 (#R0739S) and T7 ligase (#M0318S) were used for the Golden Gate reactions and were obtained from New England Biolabs (NEB). Media components were obtained from BD Bioscience and Sigma-Aldrich. Primers and synthetic DNA (gBlocks) were obtained from Integrated DNA Technologies (IDT); Primers used in this study are listed in Supplementary Table 4. Plasmids were cloned and amplified in E. coli DH5α. Clear, round bottom 96-well microtiter plates (Costar) were used for culturing. Black, clear-bottom 96-well microtiter plates (Costar) were used for fluorescence measurements. Fluorescence and optical density measurements were performed in a SynergyMx (Biotek) plate reader at 630 nm (optical density), 561ex/610em (red fluorescence) and 488ex/530em (green fluorescence).
E. coli was grown in Luria Broth (LB) media. To select for E. coli plasmids with drug-resistant genes, ampicillin (Sigma-Aldrich) or kanamycin (Sigma-Aldrich) were used at final concentrations of 75-200 µg/mL and 50 µg/mL, respectively. Agar was added to 2% for preparing solid yeast and bacterial media. Copper inductions were performed in SD media with 0-2 mM copper (II) sulfate added.
Methionine/cysteine repressions were performed in SD media lacking methionine in the drop-out mix, with 0-10 mM methionine or cysteine added.

Yeast transformation
C. glabrata and S. cerevisiae cells were transformed with plasmid DNA using the lithium acetate transformation protocol described before. 4

Cloning of new parts into the entry vectors (pYTK01) and Golden Gate assembly
Primers with overhangs for BsmBI-based cloning into the entry level vector pYTK01 were designed as instructed in Lee at al. 5 Golden Gate reactions were performed as instructed in Lee at al. 5 In brief: 0.5 μL of each DNA insert or plasmid, 1 μL T4 DNA Ligase buffer (NEB), 0.5 μL T7 DNA Ligase (NEB), 0.5 μL restriction enzyme, and water to bring the final volume to 10 μL. The restriction enzymes used were either BsaI or BsmBI (both 10 000 U/mL from NEB).
Reaction mixtures were incubated in a thermocycler according to the following program: 25 cycles of digestion and ligation (42 °C for 2 min, 16 °C for 5 min) followed by a final digestion step (60 °C for 10 min), and a heat inactivation step (80 °C for 10 min). In some cases, where noted in the text, the final digestion and heat inactivation steps were omitted.
On a technical note; Golden Gate-based cloning of dCas9-MxiI was less reliable and at least 5 white colonies needed to be picked after green white screening and their plasmids control digested in order to identify one correct assembly. This is in contrast to the usually highly efficient YTK-based Golden Gate assembly were almost 100% of white colonies carry correct assemblies.  6 Linear regression was used to derive the conversion factor.

Growth measurements
Growth curves were recorded in sterile, transparent round-bottom 96-well plates using 200 μL total culture volume, cultured at 30°C in a SynergyMx plate reader (high orbital shaking). Cells were seeded at an OD 630 of approximately 0.03 and culture turbidity (OD 630 ) was recorded every 30 minutes for 20 to 24 h. For gRNA depletion experiments, cells were grown in 50 mL shake flasks and since the later optical density values were outside the linear range of the photodetector, all optical density values were first corrected using the following formula to calculate true optical density values: where OD meas is the measured optical density, OD sat is the saturation value of the photodetector (2.568 for our instrument, as experimentally determined) and k is the true optical density at which the detector reaches half saturation of the measured optical density (2.075 for our instrument, as experimentally determined). For gRNA depletion experiments shown in Figure 2C, cells were grown in 50 mL shake flasks and diluted 1:10 every 7 to 10 h. Optical density was measured in 1 mL cuvettes in a spectrophotometer (Nanospec, Amersham Bioscience). For selections in the presence of HM-1, HM-1 was concentrated from an in-house production strain and used as 0.4x diluted supernatant in the final assay.

Plasmid/DNA extraction and qPCR
Two different protocols for extracting DNA or plasmids from C. glabrata were employed:  Table 4 for primer sequences) and water for a total volume of 25 µL was performed in white 96-well plates. Relative copy numbers and relative normalized copy numbers were determined in three biological replicates, which were measured in three technical replicates respectively. Standard cycling conditions as recommended by the manufacturer of the reaction mix were employed for data generation. For determining relative normalized expression of plasmid per cell, total DNA extract sample was added to the reaction mixture.

SUPPLEMENTARY TABLES
Supplementary Table 1. YTK parts that were used for vector assembly or re-characterized in C. glabrata.

Supplementary Figure 1: Overview on the CgTK.
Overview on the YTK-parts that were characterized in C. glabrata and newly added parts (indicated in bold). In total, the CgTK contains 25 constitutive promoters (19 YTK, 2 from C. glabrata and 4 minimal promoters), three inducible/repressible promoters (copper and methionine/cysteine), three degradation tags, and nine pre-assembled vectors with different selection markers (eight auxotrophic vectors as shown here and one with a nourseotricin marker, Supplementary Table 2

Supplementary Figure 4. Differences in expression level of Venus from the various vectors. A.
Overview of used vectors. B: Difference in Venus expression levels. Numbers indicate fold difference between the indicated vectors. Experiments were run in biological triplicates (three transformants) and error bars represent the standard deviation.

Supplementary Figure 5. qPCR analysis of relative cellular abundance of various vectors. A:
qPCR on plasmid extracts B: qPCR based on total DNA extract. In this case plasmid copy number was normalized to actin mRNA abundance (see Methods section for both protocols). In both (A and B), the abundance of the highest vector was set to 1 and the abundance of the other vectors is given relative to this number. Samples were measured in biological triplicates and each replicate was measured again in technical triplicate. Error bars represent standard deviation.

Supplementary Figure 6. Fold-difference in expression across various promoters when comparing the vector versions 1.1 and version 2.2.
The same data as generated for Figure 1A and B were used. Experiments were run in biological triplicates (three transformants) and error bars represent the standard deviation.  Table 3. MP1-4 showed low to medium expression strength in C. glabrata. Interestingly, also expression levels in S. cerevisiae were medium (Supplementary Figure 8C), although reported expression levels of these promoters should be close to the strongest promoters in S. cerevisiae. 8 It has been shown before that the YTK Golden Gate design leaves a BglII seam between the promoter and the ORF which is suboptimal for expression strength. This might explain why the MP design did not achieve the reported high expression levels in this specific set-up. 8 Optimization could thus be achieved by designing better Golden Gate seams. 11

Note 4. Performance of CuSO 4 -inducible promoters in C. glabrata and S. cerevisiae.
First, we noticed that C. glabrata cells were much more sensitive to CuSO 4 than S. cerevisiae cells (Supplementary Figure 13). For induction, we used 5-fold dilutions starting from 2 mM CuSO 4 and C. glabrata cells would only grow at concentrations lower than 16 μM (though reaching lower final OD 630 ) and would only reach full OD 630 at concentrations lower than 3 μM. In contrast, S. cerevisiae cells reached full OD 630 at the highest tested concentration of 2 mM CuSO 4 . Still, C. glabrata cells seemed metabolically active and produced high levels of Venus protein at concentrations above 16 μM. For comparison, we also characterized both promoters in S. cerevisiae (Supplementary Figure   17). Interestingly, the MT-1 promoter was not functional in S. cerevisiae (Supplementary Figure  17C).

Note 5: Degradation tags.
The YTK-derived degradation tags UBI-M (weak), UBI-Y (medium) and UBI-R (strong) were fused to the Venus protein cloned under the control of the TDH3 and ALD6 promoters. As before, expression levels were measured in bulk and normalized to fluorescein. Figure 1G shows that all three degradation tags were functional in C. glabrata reducing the fluorescence in cells with a TDH3-driven degradationtagged Venus to 56% (UBI-M), 39% (UBI-Y), and 37% (UBI-R) of the untagged Venus levels. For the ALD6p-driven degradation-tagged Venus, fluorescence was reduced to 52% (UBI-M), 33% (UBI-Y), and 20% (UBI-R). In comparison, in S. cerevisiae the tagging led to a remaining fluorescence of 63% (UBI-M), 30% (UBI-Y) and 15% (UBI-R) for a TDH3p-driven degradation-tagged Venus protein and