Construction of a Set of Novel Transposon Vectors for Efficient Silencing of Protein and lncRNA Genes via CRISPR Interference

In recent years, CRISPR interference (CRISPRi) technology of gene silencing has emerged as a promising alternative to RNA interference (RNAi) surpassing the latter in terms of efficiency and accuracy. Here, we describe the construction of a set of transposon vectors suitable for constitutive or tetracycline (doxycycline)-inducible silencing of genes of interest via CRISPRi method and conferring three different antibiotic resistances, using vectors available via Addgene repository. We have analyzed the performance of the new vectors in the silencing of mouse Adam10 and human lncRNA, NORAD. The empty vector variants can be used to efficiently silence any genes of interest.


Introduction
Over the years, RNA interference (RNAi) was one of the most influential technics applied to study gene functions despite quickly acquired awareness of its serious shortcomings [1]. The attempts to improve this technology, made in the absence of alternatives, did not eliminate the disadvantages resulting from the very essence of natural RNAi phenomenon. The major drawbacks of RNAi technology harnessing designed siRNA or shRNA molecules, which mimic miRNA or pre-miRNA molecules, include frequent off-target effects, possible cytotoxicity, and interference with miRNA pathways [2][3][4][5][6]. We also have encountered very unfavorable effects of RNAi technology in our research. While studying the impact of shRNA-mediated silencing of expression of ADAM (A Disintegrin and Metalloproteinase) family members, we found that some shRNAs targeting Adam17 affected also the expression of Adam10 and a number of Adam targeting shRNAs revealed that anti-inflammatory activity did not correlate with gene silencing [7].
What is more, we demonstrated that a control, non-targeting shRNA for MISSION® shRNA library exerts potent offtarget effects. It silences SNRPD3, which strongly affects cell viability and thus may lead to misinterpretation of RNAi results [8]. In recent years, serious pitfalls of RNAi-based analyses with far-reaching consequences have also been demonstrated by Lin et al. [9,10].
The discovery of bacterial CRISPR-Cas9 (Clustered Regularly Interspaced Short Palindromic Repeats-CRISPRassociated protein 9) system and its applicability in eukaryotic cells resulted in the development of alternative methods of gene silencing [11][12][13]. Initially, CRISPR-Cas9 was used for genome editing, a laborious, long-term procedure, which is not completely free from off-target effects resulting from DNA breaks randomly introduced by Cas9 enzyme [3]. Almost a decade ago another technic based on the CRISPR-Cas9 system, namely, CRISPR interference (CRISPRi) was developed, in which enzymatically inactive Cas9 (dead Cas9, dCas9) targeted to a gene of interest by a properly designed single guide RNA (sgRNA) inhibits the binding of RNA polymerase to the gene promoter or blocks the progress of transcription, thus effectively reducing the level of its expression [14]. The method has been constantly improved via, among others, optimizing selection of best sgRNA sequences [15][16][17][18] and generating fusion proteins in which dCas9 is complexed with known transcriptional repressors [19][20][21]. Although not completely flawless, CRISPRi outcompetes RNAi in terms of reducing off-target effect [22][23][24].
Here, we describe the design of a novel CRISPRi vector, which we constructed using non-commercial Addgenedeposited plasmids. We generated the vector variants that knockdown mouse Adam10 expression constitutively or upon induction with doxycycline with three different antibiotic resistances. Their empty counterparts can be used to target the expression of any gene of interest upon simple insertion of an appropriate gene targeting sequence into sgRNA backbone. Both empty vectors as well as these targeting Adam10 and NORAD (Non-coding RNA Activated by DNA Damage) are available via Addgene repository (plasmid numbers for empty vectors: 196074, 196076, 196078, 196080, 196082, 196084, for vectors targeting Adam10: 196075, 196077, 196079, 196081, 196083, 196085, for the vector targeting NORAD: 196086).
Three different sgRNA sequences targeting sites near the transcription start site of mouse Adam10 gene were designed using GPP sgRNA Designer (presently CRISPick, Broad Institute) ( Table 1). A single sgRNA sequence targeting NORAD comes from Elguindy et al. [26] ( Table 1). The oligonucleotides containing compatible ends with the SapIdigested plasmid were annealed in Taq buffer and cloned into the vector in one restriction-ligation reaction using SapI and T4 DNA ligase. The resulting vectors were named pSBbi-Pur-dCas9-KRAB-MeCP2-hU6-A10 (variant 1, 2, and 3) or pSBbi-Pur-dCas9-KRAB-MeCP2-hU6-NORAD. The proper insertions of the gene targeting sequences were confirmed by sequencing (Genomed, Warsaw).

Cell Culture and Transfection
Mouse colon carcinoma MC38CEA cell line [8] and HeLa (human cervical adenocarcinoma, ATCC CCL-2) were cultured in DMEM (Biowest) supplemented with 10% heat-inactivated fetal bovine serum (FBS) or with tetracycline-negative FBS (Capricorn Scientific, Ebsdorfergrund, Germany) under standard conditions. The cell cultures were tested by PCR for mycoplasma contamination using mycoplasma rDNA-specific primers and carried out without Pen/ Strep to avoid unnoticed mycoplasma contamination.
The cells were grown in 6-well plates and when reached 80-90% confluency were transfected with 950 ng of a respective CRISPRi plasmid together with 50 ng of pCMV(CAT) T7-SB100 vector encoding SB100X transposase (a gift from Zsuzsanna Izsvak; Addgene plasmid # 34879) [27] using jetPRIME (Polyplus Transfection) for MC38CEA or Lipofectamine 3000 (Thermo Fisher Scientific) for HeLa cells. One day after transfection an appropriate selection antibiotic was added to the cell cultures at following concentrations: puromycin-5 µg/ml for MC38CEA and 1 µg/ml for HeLa cells, blasticidin S-3 or 5 µg/ml, or hygromycin B-300 or 500 µg/ml. The cells were cultured in the presence of a selection antibiotic for at least a week.

RT-qPCR
After antibiotic selection, the cells were grown in 12-well plates. The cells transfected with the vectors with doxycycline-inducible expression of dCas9-based repressor were cultured for 3 days in the absence (control) or presence of doxycycline (added daily to a concentration of 100 ng/ ml) prior to RNA isolation. RNA was isolated using Fenozol (A&A Biotechnology) and, after removal of DNA by TURBO DNase (Thermo Fisher Scientific), purified with Clean Up RNA Concentrator columns (A&A Biotechnology). Equal amounts of RNA (1 μg) were subjected to reverse transcription using M-MLV polymerase (Promega) and a mixture of oligo(dT)15 (Genomed) and random hexamer primers (Promega). The levels of gene expression were evaluated via qPCR on an Eco Real-Time PCR System (Illumina) using AceQ qPCR SYBR GreenMix (Vazyme Biotech) and specific primers listed in Table 1. The fold changes in gene expression were quantified using ΔΔCt method in relation to the expression of two reference genes Eef2 and Polr2b for mouse cells and EEF2 for human cells.

Western Blotting
The cells were lysed in ice-cold RIPA buffer enriched with Halt Protease Inhibitor Cocktail containing 5 mM EDTA (Thermo Scientific). Protein samples (20 μg) were subjected to tris-glycine SDS-PAGE containing trichloroethanol, which enables UV-induced protein visualization in the gel [28]. The proteins were transferred onto 0.45 μm PVDF membrane (Immobilon FL, Merck) and probed with rabbit anti-ADAM10 (ab1997, Abcam) at 1:5000 and then with HRP-conjugated secondary antibody (anti-rabbit, Cell Signaling Technology, at 1:10,000). Bands were developed with Immobilon Western Chemiluminescent HRP Substrate (Merck) and visualized using Fusion FX (Vilber Lourmat).  Fluorescence imaging procedures were essentially performed as described previously [29]. Concisely, HeLa cells were plated on glass coverslips (#1.5H, Menzel Gläser) in 12-well culture plates at a density of 25,000 cells/well and cultured for 3 days in the absence (control) or presence of doxycycline (added daily to a concentration of 100 ng/ml). Next, the cells were washed twice in RNasefree PBS (Thermo Fisher Scientific) and fixed for 15

Additional Information
Data analysis was performed using Microsoft Excel (Excel 2016) or GraphPad Prism v. 9 and all graphs were created using GraphPad Prism v. 9 (GraphPad Software) and Corel-DRAW 2020 (Corel).

Results and Discussion
The first goal of our work was to generate a single vector for stable expression of both elements of CRISPRi system: dCas9-based repressor and sgRNA, using non-commercial resources available via Addgene repository. As a transcriptional repressor, we chose dCas9-KRAB-MeCP2, in which dCas9 is fused to bipartite repressor domain containing Krüppel-associated box (KRAB) and methyl CpG binding protein 2 (MeCP2) domains, known to play roles in blocking transcription [31,32]. The sequence contains two nuclear localization signals (NLS) directing dCasbased repressor to the nucleus. The vector coding for this fusion protein generated by Yeo et al. has been proved to efficiently inhibit expression of studied genes [21]. dCas9-KRAB-MeCP2 fusion protein alone is encoded by a sequence of more than 5300 bp and this size predestined the choice of a high-capacity transposon system for the construction of our vector. Transposons are natural, mobile genetic elements. They are flanked by terminal inverted repeat sequences (TIRs) recognized by transposase, an enzyme with excisase and integrase activities. The genetically modified transposons are used as efficient tools for stable incorporation of various DNA sequences into the genomes of target cell. A transposon system consists of two elements: a plasmid containing a transgene placed between TIR sequences and a compatible transposase delivered to the cells as a cDNA, mRNA, or protein. The activity of transposase empowers the efficiency of genomic integration comparable to that of viral vectors. Three transposon systems applicable for mammalian cells genetic modifications are currently available: Sleeping Beauty, piggyBac, and Tol2 [33]. We have chosen Sleeping Beauty system, which was reconstructed from the dormant salmonid Tc1/mariner-type transposon via elimination of inactivating mutations [34]. The major advantage of this system is close-to-random distribution of integration sites, not biased towards transcriptionally active sites, as observed for other transposon and viral vectors [33,35]. The comparison of SB characteristics with other commonly used vectors is presented in Table 2. As a scaffold for our vector, we used the version of SB plasmid redesigned by Kowartz et al. to simplify cloning procedure, introduce selection markers, and optimize promoters, enhancers, and polyA sequences [25].
The fragment directing sgRNA expression was derived from pX330-U6-Chimeric_BB-CBh-hSpCas9 (a gift from Feng Zhang; Addgene plasmid # 42230), which contains a sequence coding for a constant fragment of sgRNA required for recognition of sgRNA by dCas9 under U6 promoter and transcription termination signal for RNA polymerase III [11]. In the original plasmid, the site for the cloning of a gene targeting sequence into the sgRNA backbone was created by two sequences recognized by BbsI, a type IIS restriction enzyme, which generates non-palindromic overhangs ensuring proper cloning orientation. We replaced BbsI sites with the ones recognized by SapI, another type IIS enzyme, to prevent cleavage of dCas9-KRAB-MeCP2-coding sequence containing four BbsI sites [8]. The amplified DNA fragments (the scaffold, dCas9-KRAB-MeCP2, and U6 promoter together with sgRNA-encoding cassette) were assembled into pSBbi-Pur-dCas9-KRAB-MeCP2-hU6-SapI. Fig. 1a.

The diagram of the fragment incorporated into DNA of transfected cells is presented in
We next cloned three different RNA sequences targeting mouse Adam10 into SapI sites. They were chosen from the sequences designated by GPP sgRNA Designer, presently CRISPick (Broad Institute) [15,18]. We selected the sequences with presumed very low off-target activity. The sequences are located in the first exon and comprise nucleotides: (1) 85-105, (2) 124-143, and (3) 171-190 from the transcription start site (TSS) and are followed, respectively, by GGG, CGG, and AGG PAM motifs. According to Gilbert et al., the window of DNA regions efficiently targeted by dCas9-KRAB comprised fragments between − 50 and + 300 bp relative to TSS with a maximum of inhibitory activity observed for the + 50-100 bp region [24]. Even broader optimal DNA window was indicated for dCas9-KRAB-MeCP2-executed repression [21]. All three sequences substantially inhibited expression of Adam10 in MC38CEA cells transfected with pSBbi-Pur-dCas9-KRAB-MeCP2-hU6-SapI_A10 variants and SB100X transposaseencoding vector (Fig. 1b). Their specificity towards Adam10 was indicated by the lack of impact on the expression of Adam17 (the closest relative of Adam10) (Fig. 1b). The decrease in Adam10 mRNA levels was accompanied by diminished levels of ADAM10 protein (Fig. 1c), which confirms our previous observations of a correlation between silencing of mouse Adam10 at the transcript and protein levels [7].
The expression of neither sgRNA induced the so-called interferon response since the expression of two interferoninducible genes, Oas1 and Ifit1 were not significantly stimulated in the transfected cells (Fig. 1d). The strongest inhibition of Adam10 expression was observed for the vector bearing gene targeting sequence #3. Initially, when lone dCas9 was used as a repressor in CRISPRi, the use of multiple sgRNAs potentiated gene expression knockdown [14]. However, if the repressor was dCas9-KRAB-MeCP2, the use of multiple sgRNAs did not silence gene expression more than a single, most potent sgRNA [21]. Therefore, the sequence #3 was chosen for all subsequent variants of CRIS-PRi vectors targeting Adam10.
We prepared two additional vectors constitutively targeting Adam10 and providing resistance to blasticidin S (BsdR) or hygromycin B (HygR) following the procedure described above and using pSB scaffolds with different antibiotic resistances. However, these vectors showed lower effectiveness in Adam10 silencing than the one conferring puromycin resistance (Fig. 1e). The low efficiency of the vector providing resistance to blasticidin S is in agreement with the recently observed correlation between type of selectable marker and the levels of expression of recombinant proteins [36]. The authors showed that the expression of genes of interest may be significantly lower when blasticidin S or G418 are applied as selection antibiotics instead of other commonly used such as zeocin, puromycin, or hygromycin. It is possible that in MC38CEA cells transfected with BsdRexpressing vector the levels of dCas9-KRAB-MeCP2 were too low to efficiently block Adam10 transcription.
To create doxycycline-inducible set of CRISPRi vectors targeting Adam10, we have used previously generated vector for doxycycline-inducible expression of genes of interest [8]. This vector with puromycin resistance (Fig. 2a) was used for generation of its BsdR and HygR variants. The Adam10 targeting sequence was cloned into all of them, MC38CEA cells were transfected with the vectors and after antibiotic selection the expression of dCas9-KRAB-MeCP2 was induced with doxycycline. In two independent experiments we observed a moderate or substantial decrease of Adam10 transcript levels (Fig. 2b) accompanied by diminished ADAM10 protein levels as analyzed for the PurR vector variant (Fig. 2c). In the case of this set of vectors, PurR and BsdR variants were more efficient than the one conferring resistance to hygromycin. The increased concentrations of blasticidin S and hygromycin B during selection process did not improve effectiveness of doxycycline-induced Adam10 silencing (data not shown).
We have also evaluated the applicability of our CRIS-PRi vectors, both with constitutive and inducible expression of dCas9-KRAB-MeCP2 to silence the expression of one of human long non-coding RNAs, namely NORAD. We have cloned the sequence encoding the oligonucleotide previously shown to efficiently target NORAD [26], into our CRISPRi vectors conferring puromycin resistance. The expression of NORAD in HeLa cells was evaluated by RT-qPCR and by counting NORAD molecules in the cells using smRNA FISH. The results were fully consistent and indicated that both vectors led to profound inhibition of NORAD expression in HeLa cells upon transfection (in the case of the vector for constitutive silencing, Fig. 3a) or upon transfection and switching on dCas9-KRAB-MeCP2 synthesis with doxycycline (in the case of the vector for inducible silencing, Fig. 3a, b).
The level of expression of this RNA in HeLa cells is high as the number of NORAD molecules exceeds a hundred (Fig. 3b) and the transcription of NORAD is driven by RNA polymerase II, as is the transcription of genes encoding proteins. The impressive, almost complete repression of NORAD transcription in HeLa cells supports our belief in excellence of the vectors we constructed to silence the expression of genes encoding both lncRNA and proteins. However, other factors such as selection of the best sgRNA sequence, efficiency of transfection of a given cell line, sensitivity of a given cell line to selection antibiotics may all influence the effectiveness of CRISPRi-mediated gene silencing, and thus every single experimental setting requires optimization for the satisfactory results. The summary of advantages and drawbacks of CRISPRi method in comparison to CRISPR-Cas9-mediated knockout (CRISPR ko) and to shRNA-mediated gene silencing (shRNAi) is presented in Table 3. were not treated with doxycycline were taken as 1 (white symbols). The lack of influence of the empty vector conferring puromycin resistance was documented previously [8]. c Western blot analysis of ADAM10 levels in MC38CEA transfected with pSBtet-Pur-dCas9-KRAB-MeCP2_hU6-A10 and cultured for 72 h in the absence or presence of doxycycline. Even protein loading is presented in the bottom panel. Representative image of two independent experiments is shown  Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.