In bacteria and archaea, an important part of their immune systems are the clustered regulatory interspaced short palindromic repeats (CRISPRs) ^[1]. These are utilized to protect the host organism from invading viruses and plasmids. Within this system, the nucleic acids of intruders are silenced by specific small ribonucleic acids (RNAs) originating from the host organism itself ^[2]. Over recent years, scientific advancements have transformed this system into a practical tool for (epi)genome editing, organismal studies, and the exploration and combat of diseases, particularly hereditary diseases ^[3]. Playing a pivotal role in the CRISPR system are the CRISPR-associated proteins (Cas) ^[4]. Among them, Cas9 is the most widely used ^[5]. It functions as an endonuclease, capable of recognizing specific double-stranded DNA (dsDNA) and, in turn, silencing it through cleavage ^[2]. Moreover, in scientific applications, after dsDNA cleavage, a new sequence can be inserted to achieve designated genome editing ^[3]. However, low editing efficiency and unwanted off-target effects largely limit clinical applications of the CRISPR/Cas9 system ^[6].

2. Small Molecules Modulate Wild-Type Cas9 Protein

Through high-throughput screening, Maji et al., identified several small molecules that could disrupt Cas9 binding to DNA, preventing DNA double-strand breaks (DSBs). Testing eGFP further confirmed that some of these Cas9 inhibitors, such as BRD0539 (1), worked reversibly [37,38]^[7][8]. Recently, using a high-content fluorescence-based approach, the researchers identified valproic acid (2) (VPA) as a Cas9 degrader from a chemical library consisting of nearly 300 drug-like compounds and natural products [39]^[9]. VPA, a well-known histone deacetylase inhibitor (HDACi), demonstrated significant Cas9^WT degradation under hyperthermia conditions generated either in the presence of a photothermal agent, indocyanine green, upon irradiation by a near-infrared laser or heating with an external heat bag. This degradation effect was independent of its HDAC inhibitory effect. However, the off-target effects of BRD0539 or VPA remain unknown. For a clearer understanding of off-target effects, Yang et al. searched for small molecules that inhibit Cas9^WT. In their studies, the most effective Cas9 inhibitor was shown to be SP24 (3), with an IC₅₀ for Cas9^WT of approximately 14 µM and for the Cas9-sgRNA complex of about 7 µM. This was shown during an FP assay, where SP24 significantly decreased fluorescence polarization [38]^[8]. While there are very much more potent inhibitors (nM range) [40]^[10], the SP inhibitors showed higher IC₅₀ values for Cas9^wt than the previously known BRD0539 inhibitor with an IC₅₀ of 22 µM [37]^[7]. And even though the IC₅₀ may not be optimal, the advantage is that the inhibitor can be applied to Cas9^wt. The upcoming discussions will emphasize the often-severe modifications of Cas9 to make it addressable by small molecules. Furthermore, SP24 and SP2 (4) were proven to enhance the precision of Cas9-mediated genome editing (Figure 61) [37,38]^[7][8].

3. Small Molecules Modulate Engineered Cas9 Protein

While only a few small molecules have been identified to function with Cas9^WT, the primary Cas9 chemical modulators interact with engineered Cas9. A way to activate Cas9 with a small molecule was demonstrated by Davis et al. [41]^[11]. As shown in Figure 72, Buskirk et al. successfully evolved inteins that could only self-remove in the presence of 4-hydroxytamoxifen (4-HT) (5) [42]^[12]. Upon binding to the intein, 4-HT induces the self-splicing of the intein [42]^[12]. In the case of Cas9, it is inactivated when the intein is attached to it. However, treatment with 4-HT led to the self-splicing of the intein from Cas9, reactivating its functionality (Figure 83) [41]^[11]. Davis et al. modified spCas9 at 15 different positions and expressed the modified Cas9 variants in HEK293-GFP cells with sgRNA targeting the EGFP locus. They determined the loss of function in expressing GFP after treatment with 1 µM 4-hydroxytamoxifen in 8 cases. In a more in-depth analysis with Cas9 variants modified at S219 and C574, respectively, in comparison to Cas9^wt, it was demonstrated that the modified variants exhibited higher specificity at similar on-target cleavage rates. For instance, the C574-modified Cas9 and Cas9^wt had similar on-target DNA cleavage rates of 6.4%. However, the modified variant resulted in a fourfold lower frequency at the four critical off-target sites [41]^[11]. By precisely activating Cas9, the precision of Cas9-mediated genome editing could be significantly enhanced. A similar approach was undertaken by Wu et al., where Cas9 was modified with a small molecule-assisted shut-off (SMASh) tag. Cas9 was fused with a degron from the hepatitis C virus (NS4A) and a protease domain. Under non-treatment conditions, the protease self-cleaves the SMASh tag, converting Cas9 into an active species. However, upon adding the protease inhibitor Asunaprevir (ASV) (6), the protease activity is inhibited, preventing the cleavage of the SMASh tag. This results in the recognition of the degron by the proteasome or lysosome, leading to the degradation of Cas9 and rendering it inactive. With this system, the gene editing specificity in comparison to Cas9^wt could be increased by 1.4-fold in the lowest case (EMX1) up to 8.7-fold in the highest case (VEGFA) while targeting different genes with an application of 20 µM of ASV. The mode of action of this system, as well as ASV, are shown in Figure 94 [43]^[13]. The working principle is exactly inverted in comparison to the system described earlier. In the first case, the addition of a small molecule activates Cas9 through cleavage, as demonstrated by Davis et al. [41]^[11]. In the second case, the addition of the small-molecule inhibitor prevents self-cleavage and keeps Cas9 inevitably inactive because it is degraded by the proteasome or lysosome [43]^[13]. Notably, the removal of ASV by washing with uncontaminated media allowed the newly expressed Cas9 to become active again since self-cleavage was not hindered anymore. This reversibility is valuable to prevent Cas9 from re-editing previously edited loci [44]^[14]. Often, the editing of multiple genes is necessary [45]^[15], and in such cases, Cas9 can be inactivated after editing a particular gene and re-activated when editing the next gene. Such a light-switch-like system is extremely useful for controlling the effects of Cas9, making it safer and more efficient [3,43]^[3][13]. Speaking of light, it represents a very useful tool for spatial control of the activity of small molecules [46]^[16]. Based on photoactivable protecting groups (PPGs), Manna et al. designed a “fused” system of the ones described by Davis et al. and Wu et al. [41,43]^[11][13]. Cas9 was modified with destabilized domains (DDs) of dihydrofolate reductase (DHFR). Through these unstable domains, the fused Cas9 is recognized and degraded rapidly by the proteasome, making it inactive, similar to the activated SMASh tag [43]^[13]. However, if treatment with trimethoprim (TMP) (7) is applied, the DDs get stabilized, averting the degradation of the fused Cas9, thus keeping it active [47]^[17]. This is like the system of Davis et al. in the way that by binding a small molecule, Cas9 is activated [41]^[11]; however, in this case, it occurs through the inhibition of degradation of Cas9 so that it can provide its nuclease activity [43]^[13]. In such systems, only the dosage and timing of Cas9 can be controlled using the concentration and temporal exposure of the small molecules, e.g., ASV or TMP [47]^[17]. To add spatial controllability to the Cas9 activity, two PPGs were added to TMP [32]^[18]. The PPGs were introduced at the amine groups because, in the co-crystal structure of TMP and DHFR (PDB: 7R6G), it is visible that the amine groups of TMP are buried in the binding pocket of DHFR [48]^[19]. In an eGFP disruption assay with a sgRNA Plasmid targeting the eGFP gene and DHFR-fused Cas9, treatment with PPG-modified TMP 7 a or 7 b did not induce an observable loss of fluorescence. This leads to the conclusion that protected TMP indeed cannot bind to the DDs. Modified Cas9 is therefore left unstable and is quickly degraded by the proteasome. However, after an irradiation treatment of just a couple of minutes, loss of fluorescence was observable, meaning that the deprotection of the protected TMP into free TMP was possible and the binding ability to the DDs was restored. Following irradiation, inhibition of proteasomal degradation was gained, and Cas9 was kept active. Apart from controlling double-strand breaks and silencing genes with DHFR-modified Cas9, the expression of IL1RN could be influenced. For that, a dCas9 was modified with DHFR, and the transcriptional activator domains (VP64 and PP7) were attached. When treatment with protected TMP was applied, no leverage of expression was observable, meaning that protected TMP could not bind to the modified dCas9, leading to dCas9 being degraded. After treatment with light for 12 min, however, unprotected TMP was formed, which led to its binding to the modified dCas9. By binding, the degradation of dCas9 was prevented, resulting in an increase in IL1RN expression proportional to light exposure and compound concentration [32]^[18]. In total, the findings of Manna et al. provide a similar control of Cas9 activity to the system of Wu et al. by turning Cas9 active through treatment with a small molecule [43]^[13]. However, the system of Manna et al. not only provides temporal and dosage control but also spatial control through the inclusion of a light-activable pathway. Furthermore, they did not only gain control of double-strand breaks but also of dCas9-mediated gene activation [32]^[18]. The structure of TMP and its PPG-modified version are shown in Figure 105a, the co-crystal structure of TMP and DHFR in Figure 105b, and the working principle in Figure 105c [32,48]^[18][19]. In the systems described so far, rather substantial sequence modifications were needed to make Cas9 targetable by a small molecule. Recently, the researchers modified Cas9 with an amino acid sequence consisting of phenylalanine, cysteine, proline, and phenylalanine (FCPF) [49]^[20]. This so-called π-clamp leads to a specific reactivity of the cysteine, which then reacts with perfluoro aromatic moieties [50]^[21]. With the FCPF modification of Cas9 (Cas9^FCPF), small molecules could precisely recognize Cas9^FCPF [50]^[21]. This was used for labeling strategies, but most importantly for a proteolysis targeting chimera (PROTAC) [49]^[20]. PROTACs are hetero-bifunctional molecules with a ligand that binds an E3-ligase. The E3 ligand is connected via a linker to another ligand on the other side that can bind to the protein of interest (POI). By binding on both sides simultaneously, they can catalyze the transfer of ubiquitin (Ub) onto the POI. Ub is transferred from an E2 ligase, bound to the E3 ligase, which itself is bound to the POI. PROTAC serves as an enhancer of the binding between the E3 ligase and POI. Through the ubiquitination of the POI at a lysine residue or the N-terminal, it is marked for the 26S proteasome and is then degraded by the ubiquitin-proteasome system (UPS) [51]^[22]. The researchers generated a perfluoro derivative conjugated with lenalidomide, a ligand of the E3-ligase Cereblon (CRBN), called PROTAC-FCPF (8). 8-FCPF-Cas9 could be degraded in HeLa cells at a concentration of 10 µM after 6 h [49]^[20]. Via the T7E1 [52,53]^[23][24] assay, it was further proven that the biologic activity of Cas9^FCPF was comparable to unmodified Cas9. Degradation was further proven for dCas9^FCPF, Cas12^FCPF, and Cas13^FCPF [49]^[20]. In short, a similar system to that of Wu et al. was established. In both systems, Cas9’s stability is controlled by small-molecule-induced proteasomal degradation [43,49]^[13][20]. However, instead of introducing two domains as a SMASh-tag [43]^[13], only a peptide consisting of four amino acids was needed for Cas9^FCPF [49]^[20]. The mode of action of PROTAC-FCPF and its structure are shown in Figure 116 [49,51]^[20][22].

4. Small Molecules Regulate DSB Repair Mechanisms

Aside from regulating the activity of engineered Cas9, there is the possibility of modulating the CRISPR/Cas9 system by regulating the DNA repair system. Li et al. investigated three compounds regarding their ability to enhance HDR or down-regulate NHEJ towards a more precise editing [17]^[25]. In their study, Scr7 (9), L755507 (10), and Resveratrol (11) were found to inhibit DNA ligase IV and thereby reduce NHEJ-mediated repair [17]^[25], showing that the efficiency of knock-ins can be enhanced in the presence of chemical inhibitors targeting NHEJ-influencing factors [54]^[26]. The compounds are illustrated in Figure 127. A similar approach was undertaken by Bermudez Cabrera et al. Multiple small molecules were investigated, with a focus on the Ataxia-telangiectasia mutated (ATM) inhibitor KU60019 (12), as illustrated in Figure 138 [55]^[27]. ATM, a serine/threonine kinase, plays a determining role in the initiation of the DSB repair mechanism by recruiting necessary repair factors and determining whether HDR or NHEJ occurs [56]^[28]. During a study, a dose-dependent downshift from MH deletions (down to 0.74-fold) in the presence of various chemical compounds, including 11, was observed. Also, an increase in editing efficiency was noticeable [55]^[27]. However, as the main repair mechanism of DSB, downregulating NHEJ is associated with increased potential for tumorigenesis [57,58,59]^[29][30][31]. To avoid side effects, Zhang et al. screened 722 small molecules and identified farrerol (13), which increased knock-in efficiency by up to 2.9-fold at 5 µM, while NHEJ was clearly not affected. Also, single-strand annealing (SSA), which is said to cause loss of genome integrity, could be downregulated by up to 3.3-fold at a concentration of 10 µM (Figure 149) [60,61]^[32][33].

5. Small Molecules Regulate sgRNAs

Given the important role of sgRNAs in CRISPR/Cas9 ^[2], chemicals capable of controlling the activity of the sgRNA offer another possibility to regulate the process of genome editing [41,43]^[11][13]. Aptamers are short nucleic acids that can specifically bind ligands and are commonly employed to regulate nucleic acids, such as sgRNA [62]^[34]. Iwasaki et al. applied this concept to develop so-called aptamer-sgRNA (agRNA), in which the theophylline (14) or 3MX (15) aptamer was used and introduced into sgRNA at different positions, leading to an endonuclease-active Cas9 only upon treatment with 14 or 15, respectively. An up to 10⁴-fold increase in transformants in comparison to wild-type sgRNA was observed (Figure 150) [63]^[35]. Based on the theophylline aptamer, Bingqian et al. designed small-molecule-activated allosteric aptamer-regulating (SMART)-sgRNAs [66]^[38]. Instead of inhibiting the nuclease activity of Cas9, as demonstrated by Iwasaki et al. [63]^[35], they used a blocking motif to inhibit the binding of Cas9 in the first place. Further, a triggering motif at the sgRNA 3′ end was attached, containing the theophylline aptamer. In the absence of theophylline, the blocking and triggering motifs form a loop to block Cas9 from binding. After treatment with theophylline, the structure is changed, which allows Cas9 to bind to the sgRNA, leading to the DSB (Figure 116) [66]^[38]. By employing this system, the cleavage of EGFP could be reduced in the absence of theophylline. Significantly so, in comparison to unmodified sgRNA. Moreover, the system was successfully applied to the firefly luciferase and TurboRFP genes by modifying the sgRNA while retaining the blocking and triggering motif. After treatment with theophylline, a SMART-sgRNA activation of up to 61% was observed. This system has been demonstrated to be effective both in vitro and in vivo, providing a versatile tool for activating sgRNA and thereby controlling Cas9 [66]^[38].