key: cord-0290956-cegw80t1
authors: Méndez-Mancilla, Alejandro; Wessels, Hans-Hermann; Legut, Mateusz; Kadina, Anastasia; Mabuchi, Megumu; Walker, John; Robb, G. Brett; Holden, Kevin; Sanjana, Neville E.
title: Chemically modified guide RNAs enhance CRISPR-Cas13 knockdown in human cells
date: 2021-05-14
journal: bioRxiv
DOI: 10.1101/2021.05.12.443920
sha: 8cff0842b56715a89ae73a6b4c57eda6f267bc96
doc_id: 290956
cord_uid: cegw80t1

RNA-targeting CRISPR-Cas13 proteins have recently emerged as a powerful platform to transiently modulate gene expression outcomes. However, protein and CRISPR RNA (crRNA) delivery in human cells can be challenging and knockdown can be transient due to rapid crRNA degradation. Here we compare several chemical RNA modifications at different positions to identify synthetic crRNAs that improve RNA targeting efficiency and half-life in human cells. We show that co-delivery of modified crRNAs and recombinant Cas13 enzyme in ribonucleoprotein (RNP) complexes enables transient gene expression modulation in primary CD4+ and CD8+ T-cells. This system represents a robust and efficient method to transiently modulate transcripts without genetic manipulation.

The CRISPR-Cas13 enzyme family has shown remarkable versatility in basic biology and biotechnology applications 1 , including RNA knockdown [1] [2] [3] [4] , transcript modifications like RNA editing 4,5 , live imaging 6, 7 and diagnostics [8] [9] [10] . Importantly, CRISPR-based transcriptome modulation has been proposed to offer considerable therapeutic potential for a wide spectrum of RNA-mediated diseases 1, 11, 12 . In all cases, a CRISPR RNA (crRNA) guides Cas13 to its target RNA by RNA-RNA hybridization of a short 23 nt spacer sequence to the target site 13, 14 . Cas13based effectors act on the RNA level and do not result in permanent changes to the genome.

However, without a continuous source of crRNA expression, effects are short-lived due to rapid crRNA degradation by endogenous RNA nucleases and regeneration of the cellular steady-state by continuous target RNA expression.

Although continuous crRNA expression can be achieved via genetic manipulation, such as viral vectors expressing the crRNA and/or Cas13 2,3,15 , these methods are less desirable for more physiologically-relevant or therapeutic settings, such as targeting of immune checkpoints for immuno-oncology or differentiation of autologous stem cells prior to transplantation 16, 17 . One of the main challenges for transcriptome manipulations is to achieve efficient delivery of CRISPR systems for robust RNA manipulation without modifying host DNA sequence. Recently, others have demonstrated chemical activation of crRNAs for Cas13a cleavage in vitro 18 and delivered recombinant Cas13d into zebrafish embryos 19 . However, it remains unclear if a similar approach in human cells can lead to lasting effects. Inspired by guide RNA modifications for DNA-targeting Cas9 and Cas12a nucleases [20] [21] [22] [23] [24] [25] [26] , we sought to test if chemically modified synthetic crRNAs can be delivered into human cells to enhance Cas13-mediated transcript knockdown. First, we assessed the degree of target RNA knockdown efficiency upon exogenous delivery of unmodified and chemically-modified crRNAs. We synthesized crRNAs with different chemical modifications at three uridine nucleotides (3xU) at the 3′ end of a 23 nt spacer sequence or with an inverted thymidine (invT) capping the 3′ end (Figure 1a) Figure 1a) . For unmodified crRNAs, we noticed that protein knockdown for each of the three target transcripts was barely detectable relative to non-targeting crRNAs, suggesting that indeed unmodified crRNAs get rapidly cleared in human cells and cannot yield lasting knockdown effects (Figure 1b, c, Supplementary Figure 1b) . On the other hand, all of the chemically modified crRNAs improved target knockdown but did so to varying degrees (one-way ANOVA, p < 10 -4 ).

We found that M-modified crRNAs had the overall largest knockdown (80% knockdown of CD71) but that there was greater variability among targeted transcripts. In contrast, we found that the invT modification improved knockdown efficiency in a more consistent manner across the three cell surface markers (55%) (Figure 1c) . Surprisingly, the combination of M and S modifications did In addition to improving Cas13-mediated target knockdown, we wondered if chemical modifications can close the gap between low and high efficiency guide sequences. Previously we used a high-throughput crRNA screen to identify optimal Cas13 guide sequence design parameters and developed a scoring metric to predict crRNA efficiency 14, 28 . We chose one low-scoring guide sequence per target gene in addition to the high-scoring guide sequence used above (Supplementary Figure 2a) and found that chemically modified crRNAs do not generally improve the activity of low-scoring guide sequences (Supplementary Figure 2b) . This suggests that chemical modification may improve crRNA stability but does not improve guide sequence efficiency. Stability of the crRNA is a key element of efficient Cas13 knockdown: Using Mmodified crRNAs, we found that substantial knockdown was only possible when Cas13 was already expressed (Supplementary Figure 2c) , similarly to previously reports with Cas9 20 . This suggests that Cas13 protein must be available for efficient ribonucleoprotein (RNP) complex formation after crRNAs are introduced into cells. We found similar results when we introduced Encouraged by these results, we next examined whether alternative placement of modified bases could further improve crRNA stability and transcript knockdown. Specifically, we tested if more extensively M-modified crRNAs could increase crRNA stability without interfering with target knockdown, as previously shown for DNA-targeting Cas9 sgRNA [24] [25] [26] . We synthesized crRNAs containing a M-modification along all the bases in the crRNA (53 modified bases), all bases in the spacer sequence (23 modified bases) or in the direct repeat loop sequence (3 modified bases). For all three targeted genes, these more extensive M-modifications abrogated Cas13 knockdown compared to unmodified crRNAs, presumably by disrupting the Cas13-crRNA interaction (Figure 1d ). With Cas9 guide RNAs, multiple groups have found that moderate modification at the ends of the guide RNA tend to work best 22,23 , although one group reported an enhanced guide RNA with a majority of bases modified (~70%) for targeting hepatocytes in vivo 24 .

In contrast, our results show that a high degree of chemical modification yields less Cas13d knockdown.

Since partial modifications at the 3′ end boost knockdown, we decided to more systematically test the effect of partial modifications on the crRNA 5′ and 3′ ends. We tested crRNAs with S-modification at the first and last three nucleotides of the crRNA's 5′ and 3′ ends (5S and 3S, respectively) or using a 3xU 5′ or 3′ extension (5′Se and 3′Se, respectively). We also tested all combinations of 5′ and 3′ end modifications. We found marked improvement of target knockdown for crRNAs with phosphorothioate modifications at the 3' end of the spacer sequence ( Figure 1e) . Interestingly, 3′ phosphorothioate modifications improved knockdown efficiency to a greater degree when placed within the spacer sequence (3′S) compared to being placed in 3xU extensions (3′Se) (two-tailed t-test, p = 0.03). Modifications at the 5′ end of the crRNA alone, or in conjunction with 3' modifications did not improve target knockdown efficiency (one-way ANOVA, Tukey-corrected p = 0.99 for all comparisons with added 5′ modifications). These results suggest that exonucleases or degradation processes at the crRNA 3′ end are a major hurdle for increased Cas13 activity using synthetic crRNAs 29 .

Next, we sought to assess the temporal dynamics of Cas13 activity with synthetic crRNAs by comparing knockdown of CD46 over time for 3′ invT, 3′S and unmodified crRNAs relative to non-targeting crRNAs (Figure 2a) . All three targeting crRNAs, including unmodified crRNAs, yielded almost complete CD46 knockdown at 24 hours after nucleofection with around 95% protein loss for 3′S-modified crRNAs (Figure 2b) . While CD46 expression quickly recovered in cells targeted with unmodified crRNAs, 3′S crRNAs led to pronounced knockdown at 48 hrs after crRNA delivery (Figure 2c) . Even 4 days after nucleofection, the modified crRNAs resulted in In addition to delivery of chemically-modified crRNAs for extended knockdown, another major challenge is delivery of the effector protein into systems that cannot easily be genetically modified to continuously express Cas13. Therefore, we next sought to evaluate if we could preassemble Cas13 RNP complexes with synthetic crRNAs and deliver these RNPs into primary human cells. First, we confirmed that preassembled RNPs with recombinant RfxCas13d with nuclear localization signals (NLSs) on the N-and C-termini (rCas13d) could efficiently degrade different target RNAs in vitro (Figure 3a, Supplementary Figure 3a-c) . We found that purified rCas13d with or without an N-terminal MKIEE expression and solubility tag resulted in similar nuclease activity in vitro. By nucleofection of RNPs into human HEK293FT cells, we identified optimal buffer and temperature conditions for RNP assembly, achieving ~85% knockdown of CD46 at 24 hours post-nucleofection (Figure 3b, Supplementary Figure 3d) . At 48 hours postnucleofection, we found that complexing in a Tris-based buffer at an elevated temperature (37°C) led to greater knockdown than an alternative buffer and room temperature RNP formation (twoway ANOVA with Tukey-corrected p < 10 -4 for buffer and p < 10 -3 for temperature). We further optimized RNP delivery using different protein amounts, while holding the molar ratio of rCas13d to crRNA (1:2) constant. We found that 10µg of rCas13d protein yielded the strongest knockdown and that additional rCas13d does not lead to greater knockdown (two-tailed t-test, p = 0.7) (Figure   3c ). In agreement with our prior results using viral delivery of Cas13d, we find that using chemically-modified crRNAs also improves RNA knockdown with Cas13d RNPs (two-tailed paired t-test, p = 5 x 10 -3 ).

Using these optimized conditions for RNP formation and nucleofection, we isolated human T-cells from healthy donors, stimulated the cells with anti-CD3 and anti-CD28 antibodies, and asked whether we could knockdown CD46 expression in primary T-cells using these RNP complexes (Figure 3d) . We assessed CD46 expression in CD4+ and CD8+ populations separately 24 hours after nucleofecting activated T cells (Figure 3e, Supplementary Figure 3e) . For both T-cell populations we achieved a 60 -65% knockdown of CD46 with 3′S modified crRNAs (Figure 3f) . With unmodified crRNAs, we achieved only 40 -45% knockdown of CD46. Taken together, these experiments demonstrate that Cas13 RNPs can modify gene expression in primary immune cells and that modified crRNAs lead to greater gene knockdown than unmodified crRNAs.

Here, we show that chemically-modified crRNAs can be used for Cas13d RNA targeting and knockdown. By testing different chemical modifications and their placement within the crRNA, we identify a subset of optimal modifications for RNA targeting in cells and demonstrate that knockdown effects persist longer than with unmodified crRNAs. Furthermore, we demonstrate the use of Cas13d RNPs complexes together with chemically-modified crRNAs in human cell lines and primary immune cells. Although chemically-modified crRNAs improve Cas13 knockdown with Cas13 from an integrated transgene or from RNPs, the improvement is more modest for RNPs 

We cloned the TRS-Leader-SARS-CoV-2-d2eGFP plasmid (Addgene 171585) by introducing 75 nt of the SARS-CoV-2 Leader sequence into the NheI-digested EFS-EGFPd2PEST-2A-Hygro plasmid (Addgene 138152). We inserted the TRS-Leader immediately before the coding sequence by ligation of annealed oligos (see Supplementary Table 3 for sequences). We generated bacterial RfxCas13d expression vectors as follows. Codon-optimized cDNAs containing the effector protein RfxCas13d were synthesized (Genscript) and assembled into modified pET28 vectors by NEBuilder HiFi DNA assembly (E2621, NEB). activated with the ImmunoCult Human CD3/CD28 T Cell Activator (StemCell Technologies). T cells were cultured at 37°C, 5% CO2, and ambient oxygen levels.

The crRNAs were synthesized using solid-phase phosphoramidite chemistry (Synthego CRISPRevolution platform). Following synthesis, postprocessing and purification steps, we quantified each crRNA by UV absorption using Nanodrop (Thermo). We confirmed their identities and quality using an Agilent 1290

Infinity II high-performance liquid chromatography (HPLC) system coupled with an Agilent 6530B Quadrupole time of-flight mass spectrometry (Agilent Technologies) in negative ion polarity mode.

Recombinant RfxCas13d proteins were expressed in E.coli NiCo21(DE3) (NEB C2529). Cells were grown in 1L of lysogeny (Luria-Bertrani) broth containing 40 mg/mL kanamycin with shaking at 30°C. Protein expression was induced by the addition of 0.4 mM IPTG and the temperature was reduced to 16°C for 16 hours. Pelleted cells were disrupted by sonication. Proteins were purified from the lysate supernatant using HiTrapDEAE FF Sepharose, HisTrap HP and HiTrap HeparinHP columns on an Akta Go instrument (Cytiva). To produce sTag RfxCas13d, the N-terminal His-MPB fusion partner was cleaved with Sumo protease and removed. Purified RfxCas13d proteins were stored in 20mM Tris-HCl, pH7.5, 500mM NaCl, 1mM EDTA, 1mM DTT and 50% glycerol (v/v). Protein concentration was determined by Bradford assay (Bio-Rad).

For nucleofection experiments in HEK293FT-RfxCas13d cells, we seeded three replicates of ~2x10 6 cells in D10 media with 1 ug/mL of doxycycline 24 hours before nucleofection. Using Nucleofections were performed in triplicate. Immediately after nucleofection T cells were plated in preequilibrated ImmunoCult-XF T Cell Expansion Medium supplemented with 10 ng µL-1 recombinant human IL-2 (equilibrated to 37°C and 5% CO2). Before flow cytometry analysis, nucleofected T cells were washed with D-PBS and stained with LIVE/DEAD Fixable Violet Dead Cell Stain (ThermoFisher) for 5 minutes at room temperature in the dark before proceeding with antibody stain.

We synthesized 200 bp target RNAs for CD46, CD55 and CD71 by PCR amplification of templates from total cDNA from HEK293FT cells and then performed T7 in vitro RNA transcription. To prepare the cDNA library, we extracted total RNA from HEK293FT cells using Direct-zol purification (Zymo) and reverse transcribed 1 ug of RNA into cDNA using Revert-Aid (ThermoFisher). Using the PCR templates of target RNAs, we performed in vitro RNA transcription using the HiScribe T7 High Yield RNA Synthesis Kit (NEB). After transcription, RNA was purified using a Monarch RNA Cleanup Kit (NEB) and quantified on a Nanodrop spectrophotometer (Thermo).

Similarly, RfxCas13d and eGFP messenger RNAs (mRNAs) were synthesized using a T7 RNA polymerase in vitro transcription with the HiScribe T7 ARCA mRNA Kit (with tailing) (NEB). The mRNAs were purified using a Monarch RNA Cleanup Kit (NEB). We amplified target RNAs and mRNAs using primers that include a T7 promoter sequence on the 5′ end; for the mRNAs, we also included the corresponding Kozak and start/stop codon sequences (Supplementary Table 2 ).

Purified RfxCas13d proteins and synthetic crRNAs were mixed (unless otherwise indicated) at 2:1 molar ratio in Buffer 1 (GenScript SC1841) or Buffer 2 2 (25mM Tris pH 7.5, 1mM DTT, 6mM MgCl2). The reaction was prepared on ice and incubated at 25°C or 37°C for 15 minutes prior to the addition of target RNA at 1:2 molar ratio relative to RfxCas13d protein. The reaction was incubated at 37°C for 45 minutes and quenched with 1 µL of enzyme stop solution 2 (10 mg/mL Proteinase K, 4M Urea, 80mM EDTA, 20mM

Tris pH 8.0) at 37°C for 15 minutes. The samples were denatured at 95°C for 5 minutes in 2X RNA Gel Loading Dye (ThermoFisher) and loaded onto 10% TBE-Urea gels (ThermoFisher). The gels were run at 180 V for 35 mins. After separation, gels were stained with SYBR Gold (ThermoFisher) prior to imaging via Gel Doc EZ system (Bio-Rad).

Cells were stained with the following antibodies: PE anti-human CD4 (clone RPA-T4, Biolegend 300507), FITC anti-human CD8a (clone RPA-T8, Biolegend 301006) and APC anti-human CD46 (clone TRA-2-10, Biolegend 352405). Antibody staining was performed for 20 minutes on ice. After staining the cells were washed 2X in D-PBS and analyzed via flow cytometry (Sony SH800S). A minimum of 10,000 viable events was collected per sample.

Data analysis and statistical testing was performed using GraphPad Prism 8 (GraphPad Software Inc.). Flow cytometry analysis and figure generation was performed using FlowJo v10 (BD). Specific statistical analysis methods are described where results are presented.

Metagenomic Discovery To Transcriptomic Engineering

Transcriptome Engineering with RNA-Targeting Type VI-D CRISPR Effectors

RNA targeting with CRISPR-Cas13

RNA editing with CRISPR-Cas13. Science (80-. )

Targeted mRNA demethylation using an engineered dCas13b-ALKBH5 fusion protein

Dynamic Imaging of RNA in Living Cells by CRISPR-Cas13 Systems

CRISPR-mediated live imaging of genome editing and transcription

Multiplexed and portable nucleic acid detection platform with Cas13, Cas12a and Csm6. Science (80-. )

Nucleic acid detection with CRISPR-Cas13a/C2c2. Science (80-. )

Deployable CRISPR-Cas13a diagnostic tools to detect and report Ebola and Lassa virus cases in real-time

CRISPR-Cas13 Precision Transcriptome Engineering in Cancer

Programmable Inhibition and Detection of RNA Viruses Using Cas13

Structural Basis for the RNA-Guided Ribonuclease Activity of CRISPR-Cas13d

Massively parallel Cas13 screens reveal principles for guide RNA design

Abrogation of PRRSV infectivity by CRISPR-Cas13b-mediated viral RNA cleavage in mammalian cells

Cancer immunotherapy using checkpoint blockade

In situ programming of leukaemia-specific t cells using synthetic DNA nanocarriers

Conditional control of RNA-guided nucleic acid cleavage and gene editing

CRISPR-Cas13d Induces Efficient mRNA Knockdown in Animal Embryos

Chemically modified guide RNAs enhance CRISPR-Cas genome editing in human primary cells

Chemically Modified Cpf1-CRISPR RNAs Mediate Efficient Genome Editing in Mammalian Cells

Synthetically modified guide RNA and donor DNA are a versatile platform for CRISPR-Cas9 engineering

Chemical modifications of adenine base editor mRNA and guide RNA expand its application scope

Structure-guided chemical modification of guide RNA enables potent non-viral in vivo genome editing

Improving CRISPR-Cas specificity with chemical modifications in single-guide RNAs

Incorporation of bridged nucleic acids into CRISPR RNAs improves Cas9 endonuclease specificity

In vitro-transcribed guide RNAs trigger an innate immune response via the RIG-I pathway

Transcriptome-wide Cas13 guide RNA design for model organisms and viral RNA pathogens

Minimal 2'-O-methyl phosphorothioate linkage modification pattern of synthetic guide RNAs for increased stability and efficient CRISPR-Cas9 gene editing avoiding cellular toxicity

Development of CRISPR as an Antiviral Strategy to Combat SARS-CoV-2 and Influenza

Integrated sample inactivation, amplification, and Cas13-based detection of SARS-CoV-2. bioRxiv

Treatment of influenza and SARS-CoV-2 infections via mRNA-encoded Cas13a in rodents

Coronavirus leader-RNA-primed transcription: An alternative mechanism to RNA splicing

The Architecture of SARS-CoV-2 Transcriptome

Amplification-free detection of SARS-CoV-2 with CRISPR-Cas13a and mobile phone microscopy

Clinical validation of a Cas13-based assay for the detection of SARS-CoV-2 RNA

Detection of SARS-CoV-2 with SHERLOCK One-Pot Testing

We thank the entire Sanjana laboratory for support and advice. N.E.S. is supported by NYU and NYGC 

Cas13 RNPs complexed with CD46-targeting synthetic crRNAs with no chemical modification or with a 3′S modification. Bars represent mean values ± s.d., n =3 biological replicate nucleofections.