Crystal structure of the anti‐CRISPR, AcrIIC4

Abstract

Clustered regularly interspaced short palindromic repeats (CRISPRs)‐CRISPR‐associated protein systems are bacterial and archaeal defense mechanisms against invading elements such as phages and viruses. To overcome these defense systems, phages and viruses have developed inhibitors called anti‐CRISPRs (Acrs) that are capable of inhibiting the host CRISPR‐Cas system via different mechanisms. Although the inhibitory mechanisms of AcrIIC1, AcrIIC2, and AcrIIC3 have been revealed, the inhibitory mechanisms of AcrIIC4 and AcrIIC5 have not been fully understood and structural data are unavailable. In this study, we elucidated the crystal structure of Type IIC anti‐CRISPR protein, AcrIIC4. Our structural analysis revealed that AcrIIC4 exhibited a helical bundle fold comprising four helixes. Further biochemical and biophysical analyses showed that AcrIIC4 formed a monomer in solution, and monomeric AcrIIC4 directly interacted with Cas9 and Cas9/sgRNA complex. Discovery of the structure of AcrIIC4 and their interaction mode on Cas9 will help us elucidate the diversity in the inhibitory mechanisms of the Acr protein family.

Short abstract

PDB Code(s): 7F7P;

1. INTRODUCTION

The clustered regularly interspaced short palindromic repeats (CRISPRs)‐CRISPR‐associated protein (Cas) system is a conserved adaptive immune system that protects bacteria from viral invasion. ¹ , ² Bacteria contain CRISPR genetic material derived from previous infections with viruses and use these CRISPR genes to detect and destroy the DNA or RNA of similar viruses encountered in subsequent infections. ³ , ⁴ , ⁵ Cas proteins form a complex with a CRISPR RNA (crRNA, also called guide RNA) fragment encoded by the host CRISPR sequence, target the DNA sequences of the invading virus complimentary to the crRNA, and finally degrade viral DNA or RNA to protect the bacteria from viral invasion. ¹ , ⁶ The mechanism of action of CRISPR‐Cas systems has been studied extensively due to their utilization in basic biological research to edit genes in organisms and in clinical gene editing for the treatment of various human diseases. ⁷ , ⁸

CRISPR‐Cas systems can be classified into two broad classes based on the composition of Cas proteins that are responsible for foreign DNA digestion. ⁶ Class 1 systems comprise a multi‐subunit Cas protein complex. Whereas the class 2 system comprises large and multi‐functional Cas proteins, such as Cas9, which participate in a wide range of processes ranging from DNA targeting to cleavage. ⁶ , ⁹ These systems can be subclassified into six types (type I–type VI) based on their phylogeny and mechanism of action. Subtype I, III, and IV belong to class 1, and subtype II, V, and VI belong to class 2. ² , ⁶

To overcome the bacterial CRISPR‐Cas immune system, viruses have developed specific inhibitors, called anti‐CRISPRs (Acrs). ¹⁰ , ¹¹ , ¹² , ¹³ Genetic analyses have indicated that approximately 60 acr genes are present in phage and viral genomes. ¹⁰ , ¹⁴ Various bacteria also contain Acrs in mobile genetic elements (MGEs) derived from infected viruses and plasmids. ¹⁵ This bacterial Acrs play critical roles in persisting MGEs despite targeting by host CRISPR‐Cas system. ¹⁵ Due to the low sequence homology, Acrs are classified solely based on their target CRISPR‐Cas systems. ⁹ , ¹⁰ , ¹⁴ According to this classification, Acr proteins that inhibit Type IF CRISPR‐Cas systems are termed AcrIF and those that inhibit the Type IIA CRISPR‐Cas system are termed AcrIIA.

The inhibition of Type II CRISPR‐Cas systems (Cas9) by Acr was initially discovered in studies on the MGEs of Neisseria meningitidis (Type IIC) and prophages of Listeria monocytogenes (Type IIA). ¹⁶ , ¹⁷ , ¹⁸ At present, five different AcrIIC proteins (AcrIIC1–AcrIIC5) have been identified and characterized. ⁹ , ¹⁹ Structural and biochemical studies have shown that AcrIIC1 and AcrIIC3 directly interact with the HNH nuclease domain in Cas9 and inhibit the cleavage of target DNA, whereas AcrIIC2 prevents the loading of sgRNA into Cas9 (Table 1). ²⁰ , ²² Although the inhibitory mechanisms and structural data of AcrIIC1, AcrIIC2, and AcrIIC3 have been revealed, the inhibitory mechanisms of AcrIIC4 and AcrIIC5 proteins have not been fully understood and structural data are unavailable.

TABLE 1.

Features of the AcrIIC family of proteins

Proteins (PDB ID)	Source	Target proteins	Mechanism of action	References
AcrIIC1 (5VGB)	Neisseria meningitidis	nmCas9, hpCas9, smCas9	Inhibits DNA cleavage by interacting with the HNH nuclease domain of Cas9‐sgRNA‐dsDNA complex	17
AcrIIC2 (6J9K, 6JD7)	Neisseria meningitidis	nmCas9, hpCas9, smCas9	Prevents single‐guide RNA loading	20, 21
AcrIIC3 (6JE9, 6JHV)	Neisseria meningitidis	nmCas9, hpCas9, smCas9	Interacts with HNH nuclease domain in multiple Cas9 conformations	21, 22

The aim of this study is to elucidate the molecular mechanism underlying AcrIIC4‐mediated Cas9 inhibition for improving our understanding of the inhibitory process of Acr protein and enabling the fine regulation of Cas9 activity. Here, we demonstrate a molecular basis of Cas9 inhibition by AcrIIC4 by providing the 2.03 Å high‐resolution crystal structural data of AcrIIC4 derived from Haemophilus parainfluenzae.

2. RESULTS AND DISCUSSION

2.1. Characteristics of AcrIIC4

To understand the inhibitory mechanism of AcrIIC4 against Type IIC CRISPR‐Cas systems via structural and biochemical analysis, the AcrIIC4 protein (residues 1–88) was purified using a rapid two‐step chromatography method comprising affinity chromatography followed by SEC. Based on the fact that AcrIIC4 eluted between myoglobin (size 17 kDa) and vitamin B12 (size 1.350 kDa) in SEC, we speculated that AcrIIC4 existed in a monomeric state in the solution (Figure 1a). Since the functional stoichiometric characteristics of Acrs, whether as a monomer or dimer, are important for understanding the molecular mechanism of action of Acrs, MALS was performed for confirming the stoichiometry of AcrIIC4 by determining the absolute molecular mass of the particle in solution. MALS analysis showed that the molecular mass of AcrIIC4 in solution was 14.452 kDa (3.7% fitting error) with a polydispersity value of 1.002 (Figure 1b). This result provided slightly bigger size than we expected, since the theoretical molecular weight of monomeric AcrIIC4 including the C‐terminal His‐tag was 11.035 kDa. However, if we consider the error range of MALS and the SEC result, it was highly possible that AcrIIC4 existed as monomer rather than dimer in solution.

FIGURE 1.

Overall structure of AcrIIC4 derived from Haemophilus parainfluenzae. (a) SEC profile of AcrIIC4. SDS‐PAGE gel showing the protein eluted at the peak position was provided. (b) MALS profile corresponding to the main peak of the SEC. The experimental MALS data (red line) are plotted as the SEC elution volume (x‐axis) versus absolute molecular mass (y‐axis) distributions on the SEC chromatogram (black) at 280 nm. (c) Illustration of the structure of two molecules of AcrIIC4 presented with a crystallographic asymmetric unit. (d) Tentative dimeric structure of AcrIIC4 derived from the symmetry analysis of crystal packing. (e) Table summarizing the interaction details of the two types of interfaces analyzed using the PISA server. (f) Illustration of monomeric AcrIIC4 structure. The color of the chain from the N‐ to the C‐termini gradually changes through the spectrum from blue to red. (g) Superimposition of the structures of two molecules, A and B, detected in one asymmetric unit. (h) Topological representation of AcrIIC4. (i) B‐factor distribution in the structure of AcrIIC4. The structure is presented in a putty representation and rainbow colors from red to violet in B‐factor value order. The highest B‐factor region including the α2–α3 connecting loop is indicated. SEC, size‐exclusion chromatography; SDS‐PAGE, sodium dodecyl sulfate‐polyacrylamide gel electrophoresis; MALS, multi‐angle light scattering. (j) Surface electrostatic potential of AcrIIC4. The respective surface electrostatic distributions are represented. The scale ranges from −6.3 kT/e (red) to 6.3 kT/e (blue)

2.2. Overall structure of AcrIIC4 derived from H. parainfluenzae

Since data on the structural homologs of AcrIIC4 are not available, the molecular replacement phasing method could not be applied to determine the structure of AcrIIC4. Instead, the ab initio phasing method using ARCIMBOLDO_BORGES software, ²³ which analyzes data on small helix and sheet fragments available in the PDB, was used to solve the phasing issue. The structure of AcrIIC4 was determined and the final structural model was refined to R _work = 22.5% and R _free = 27.9%. Diffraction data and refinement statistics are summarized in Table 2. Two molecules, molecule A and B, were observed in a crystallographic asymmetric unit (Figure 1c). Although crystallographic packing analysis using the PISA server found several different potential dimers including the most probable symmetric dimer configuration (Figure 1d), the symmetrical molecules were not optimal dimeric molecules based on the PISA score (Figure 1e). Based on the SEC‐MALS data and crystallographic packing analysis, we concluded that AcrIIC existed as a monomer in solution.

TABLE 2.

Structural data and refinement statistics of AcrIIC4

Structural data
Space group	P 2 1
Unit cell parameter a, b, c (Å)
a, b, c (Å)	32.79, 52.11, 54.39
α, β, γ (°)	90, 95, 90
Resolution range (Å) a	29.26–2.03
Total reflections	80,594
Unique reflections	11,895
Multiplicity	6.8 (7.0)
Completeness (%) a	99.87 (99.83)
Mean I/σ(I) a	11.5 (1.5)
R merge (%) a , b	0.090 (1.359)
Wilson B‐factor (Å b )	43.72
Refinement
Resolution range (Å)	29.255–2.03
Reflections	11,893
R work (%)	22.5 (19.0)
R free (%)	27.9 (22.8)
No. of molecules in the asymmetric unit	2
No. of nonhydrogen atoms	1,533
Macromolecules	1,500
Solvent	33
Average B‐factor values (Å b )	48.22
Macromolecules	49.74
Solvent	39.27
Ramachandran plot:
Favored/allowed/outliers (%)	98.9/1.1/0
Rotamer outliers (%)	0
Clashscore	6.0
RMSD bonds (Å)/angles (°)	0.007/0.941

Abbreviation: RMSD, root‐mean‐square deviation.

^a Values for the outermost resolution shell in parentheses.

^b R _merge = ∑ _h ∑ _i |I(h) _i  − <I(h)>|/∑ _h ∑ _i I(h) _i , where I(h) is the observed intensity of reflection h, and <I(h)> is the average intensity obtained from multiple measurements.

The final model of molecule A contained the full‐length AcrIIC4 protein (residues 1–88) with extra LEHHHHHH residues at the C‐terminus, which were derived from expression constructs, whereas the final model of molecule B contained full‐length AcrIIC4 without extra construct‐derived residues (Figure 1f,g). The structures of molecule A and B were almost identical with a root‐mean‐square deviation (RMSD) value of 0.32 Å (Figure 1g). The overall structure of AcrIIC4 comprised an α‐helical bundle fold composed of four α‐helices (α1–α4) (Figure 1f,h). B‐factor analysis showed that the region of the α2–α3 connecting loop had a relatively higher B‐factor although the overall structure was rigid with a low B‐factor (average 30.22 Ų). This indicated that AcrIIC4 contained a relatively flexible α2–α3 connecting loop, although a major part of the protein was rigid and solid in solution (Figure 1i).

To elucidate the mechanism of action of AcrIIC4, we initially performed a structural homology search using a DALI server. ²⁴ The selected top five matches from the DALI server were FAP1 (PDB id: 3RGU), uncharacterized protein (PDB id: 5F3O), MCCA (PDB id: 4RKM), EzrA (PDB id: 4UXV), and Talin‐1 (PDB id: 6R9T) (Table 3). The decreasing Z‐scores ranging from 7.1 to 6.0, increasing RMSD values ranging from 2.4 to 3.1 Å, and low sequence identity ranging from 7 to 16%, strongly indicated that the structural similarity of the top matches with AcrIIC4 was fairly low and the structural homologs identified using the DALI server may not be functionally relevant to AcrIIC4 (Table 3).

TABLE 3.

Structural similarity search using DALI

Proteins (PDB accession number)	Z‐score	RMSD (Å)	Identity (%)
FAP1 (3RGU)	7.1	2.7 (77/88)	9
Uncharacterized protein (5F3O)	7.0	2.5 (77/194)	16
MCCA (4RKM)	6.9	2.5 (72/620)	16
EzrA (4UXV)	6.4	2.4 (67/522)	7
Talin‐1 (6R9T)	6.0	3.1 (81/2185)	11

Abbreviations: PDB, Protein Data Bank; RMSD, root‐mean‐square deviation.

To determine the characteristics that may help elucidate the functional mechanism of AcrIIC4 in the absence of structural information of structural homologues, we next analyzed the electrostatic surface features of AcrIIC4. We found that the surface of AcrIIC4 had distinct negatively and positively charged areas (Figure 1j). Since positively charged patches are often involved in nucleotide binding, and several Acr proteins directly interact with sgRNA or target DNA during the inhibition process, we speculated that AcrIIC4 may directly interact with DNA or RNA for inhibiting the type II CRISPR‐Cas9 system. To test this speculation, we performed a simple gel‐shift assay and fount that neither DNA nor RNA caused a band‐shift of the AcrIIC4 protein (Figure S1a,b), indicating that AcrIIC4 may not directly interact with nucleotides during the inhibition of the CRISPR‐Cas system.

2.3. Binding of AcrIIC4 to Cas9 and inhibition of target DNA recruitment

Since AcrIIC4 did not bind to nucleic acids, and a previous biochemical study suggested that AcrIIC4 directly binds to Cas9, ¹⁹ we analyzed the direct interaction of AcrIIC4 with H. parainfluenzae‐derived Cas9 by performing Native PAGE. We found that the Cas9/AcrIIC4 mixture produced a new complex band on the gel, indicating that AcrIIC4 could directly bind to Cas9 in vitro (Figure 2a). Additionally, a Cas9/AcrIIC4/sgRNA mixture produced a Cas9/AcrIIC4 complex band on the gel, indicating that AcrIIC4 could bind to Cas9 regardless of the presence of sgRNA. This interaction was also analyzed by SEC. According to our SEC experiment, AcrIIC4 was co‐migrated with HpaCas9, indicating that AcrIIC4 directly binds to HpaCas9 (Figure 2b).

FIGURE 2.

Analysis of Cas9‐binding capability and function of AcrIIC4. (a) Native polyacrylamide gel electrophoresis. The loaded samples are indicated above the gel. Newly produced bands indicating Cas9/AcrIIC4 complex are indicated by a black arrow. (b) SEC profile of AcrIIC4/Cas9 mixture. SDS‐PAGE gel showing the protein eluted at the peak position was provided. (c) Electrophoretic mobility shift assay for analyzing the inhibition of DNA recruitment to Cas9 by AcrIIC4 in a concentration‐dependent manner. M, DNA size marker. The amount of protein added in each lane has been indicated. (d) Final model of the working process of AcrIIC4

Since AcrIIC4 interacts with Cas9/sgRNA complex, we performed EMSA to analyze the inhibition of CRISPR‐Cas9 system by AcrIIC4 in a concentration‐dependent manner. The addition of AcrIIC4 blocked the gel‐shift in a concentration‐dependent manner (Figure 2c), indicating that AcrIIC4 may interact with Cas9/sgRNA complex and prevent the recruitment of target DNA. Based on this observation, we concluded that monomeric AcrIIC4 may competitively bind to target DNA‐binding sites in Cas9 for inhibiting the activity of Cas9. When AcrIIC4 binds to the specific target DNA‐binding pocket, the target DNA cannot access the active site on Cas9 (Figure 2d). Further structural studies on AcrIIC4/Cas9 complex and studies on various strategies for Cas9 inhibition by different Acrs will be critical for understanding the various regulation processes of CRISPR‐Cas systems, and will enable the fine‐tuning of gene editing‐based therapeutic applications.

3. EXPERIMENTAL PROCEDURES

3.1. Cloning, overexpression, and purification

The acrIIC4 gene of H. parainfluenzae (accession code: WP_049372635.1) was synthesized by Bionics (Daejeon, Republic of Korea) and then cloned into a pET21a vector (Novagen, WI). NdeI and XhoI restriction sites were utilized for cloning. The resulting plasmid vector containing acrIIC4 (residues 1–88) was transformed into Escherichia coli strain BL21(DE3) competent cells. The cells (5 ml culture) were cultured at 37°C in 1 L of lysogeny broth (LB) containing 50 μg/ml kanamycin. When the optical density measured at 600 nm reached 0.7–0.8, the temperature was adjusted to 20°C and 0.5 mM isopropyl β‐d‐1‐thiogalactopyranoside was added for inducing expression of the target gene. The induced cells were further cultured for 18 hr at 20°C in a shaking incubator. Cultured cells were harvested via centrifugation at 2,000g for 15 min at 4°C. The harvested cells were then resuspended in a lysis buffer (20 mM Tris–HCl (pH 8.0) and 500 mM NaCl) and lysed via ultrasonication. The cell lysate was separated via centrifugation at 10,000g for 30 min at 4°C. The separated supernatant was mixed with Ni‐nitrilotriacetic acid affinity resins and incubated for 3 hr at 4°C. The incubated mixture was then loaded onto a gravity‐flow column (Bio‐Rad, Hercules, CA). The resin in the gravity column was washed with 50 ml of washing buffer (20 mM Tris–HCl (pH 8.0), 500 mM NaCl, and 30 mM imidazole) to remove nonspecifically bound proteins. After washing, the resin‐bound AcrIIC4 was eluted using 2 ml elution buffer (20 mM Tris–HCl (pH 8.0), 500 mM NaCl, and 250 mM imidazole). The purity of the AcrIIC4 protein sample was further increased via size‐exclusion chromatography (SEC) using a Superdex 200 10/300 GL column (GE Healthcare, Waukesha, WI) pre‐equilibrated with 20 mM Tris–HCl (pH 8.0) and 150 mM NaCl. Gel filtration standard, composed with Thyroglobulin (670 kDa), Gamma‐globulin (158 kDa), Ovabumin (44 kDa), Myoglobin (17 kDa), and Vitamin B12 (1.35 kDa), was used for size determination. The peak fractions were collected, pooled, and concentrated to 16.8 mg/ml for crystallization. Purity of the protein was visually assessed using sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE).

3.2. Multi‐angle light scattering (MALS) analysis

The absolute molecular weight of AcrIIC4 in solution was measured using SEC‐coupled MALS (SEC‐MALS). The protein solution was loaded onto a Superdex 200 Increase 10/300 GL 24 ml column (GE Healthcare) pre‐equilibrated with an SEC buffer (20 mM Tris–HCl (pH 8.0), 150 mM NaCl). The flow rate of the buffer was maintained at 0.4 ml/min and SEC‐MALS was performed at 20°C. A DAWN‐TREOS MALS detector was connected to an ÄKTA Explorer system (GE Healthcare). The molecular weight of bovine serum albumin was measured as a reference value. Data were processed and analyzed using ASTRA software (Wyatt Technology, Santa Barbara, CA).

3.3. Crystallization and X‐ray diffraction data collection

The AcrIIC4 protein sample was crystallized using the hanging‐drop vapor diffusion method at 20°C. Initial crystals were obtained by equilibrating a mixture containing 1 μl of protein solution (16.8 mg/ml protein in 20 mM Tris–HCl (pH 8.0) and 150 mM NaCl) and 1 μl of a reservoir solution containing 0.1 M CHES/sodium hydroxide (pH 9.5) and 30% (W/V) polyethylene glycol (PEG)‐3000, against 0.3 ml of reservoir solution. Crystals of the best quality developed in 12 days were obtained using optimized conditions comprising 0.1 M CHES/sodium hydroxide (pH 9.8) and 34% (W/V) PEG‐3000. A single crystal was selected and soaked in the reservoir solution supplemented with 40% (v/v) glycerol for cryo‐protection. X‐ray diffraction data were collected at −178°C using the beamline BL‐5C at Pohang Accelerator Laboratory (Pohang, Korea). Data processing including indexing, integration, and scaling, was performed using HKL2000. ²⁵

3.4. Structure determination and refinement

The AcrIIC4 structure was determined using ARCIMBOLDO_BORGES, an ab initio phasing software ²³ and by combining the fragment search using Phaser ²⁶ and density modification using SHELXE. ²⁷ The initial model was generated automatically using AutoBuild from the Phenix package. Subsequent model building with refinement was performed using Coot ²⁸ and phenix.refine. ²⁹ The structure quality and stereochemistry were validated using MolProbity. ³⁰ All structural figures were generated using the PyMOL program. ³¹

3.5. Native PAGE

The complex formation of AcrIIC4 and its various mutants with Cas9 was evaluated via native (nondenaturing) PAGE using PhastSystem (GE Healthcare) with precast 8–25% acrylamide gradient gels (GE Healthcare). Coomassie brilliant blue was used for the staining and detection of bands. Complex formation was evaluated based on the appearance of new bands on the gel. The expression construct of Cas9 from H. parainfluenzae was purchased from addgene (#121540) and used for Native PAGE study.

3.6. Cas9 inhibition assay

The expression construct of Cas9 from H. parainfluenzae was purchased from addgene (#121540) and purified by previously reported protocol. ¹⁹ sgRNA targeting pET21a plasmid vector was designed and synthesized by Bioneer (Dae‐jeon, South Korea). Digestion of target sequence in vitro using CRISPR sgRNA was performed in the reaction mixture containing 500 ng HpaCas9, 200 ng sgRNA, 80 ng pET21a plasmid, 1 μl NEB buffer 3 (10×), 1 μl BSA (10×), and 5 μl water. The reaction mixture was put in the 37°C heat block for 1 hr and treated with 4 μg of RNAse for 15 min. Finally, stop solution (30% glycerol, 1.2% SDS, 250 mM EDTA [pH 8.0]) was added to the reaction mixture and incubated for 15 min at 37°C. The mixture was analyzed on 2% agarose gel. Indicated amount of AcrIIC4 was added in the initial stage of the reaction buffer to analyze the inhibition activity of AcrIIC4.

3.7. Complex formation assay by SEC

Purified HpaCas9 (10 mg/ml) was incubated with purified AcrIIC4 (10 mg/ml) for 1 hr at 25°C. The mixture was then loaded onto a Superdex 200 Increase 10/300 GL 24 ml column (GE Healthcare) pre‐equilibrated with an SEC buffer (20 mM Tris–HCl (pH 8.0), 150 mM NaCl). The flow rate of the buffer was maintained at 0.4 ml/min and SEC was performed at 20°C. The peak fractions were analyzed by SDS‐PAGE.

3.8. Accession codes

Atomic coordinates and structure factors for the reported crystal structures have been deposited with the Protein Data bank under accession number 7F7P.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.

AUTHOR CONTRIBUTIONS

Hyun Ho Park designed and supervised the project. Gi Eob Kim performed cloning, expression, and protein purification. Gi Eob Kim and So Yeon Lee performed crystallization and collected X‐ray data. Gi Eob Kim and Hyun Ho Park analyzed the protein structure. Gi Eob Kim performed MALS and So Yeon Lee performed EMSA. Hyun Ho Park, Gi Eob Kim, and So Yeon Lee wrote the manuscript. All authors discussed the results and approved the manuscript.

Supporting information

Figure S1 AcrIIC4 did not bind to nucleotides. (a) EMSA for detecting the direct interaction between AcrIIC4 and DNA. Different concentrations of purified AcrIIC4 protein was pre‐incubated with 10 μg of plasmid DNA (Plasmid DNA1: pET21a, Plasmid DNA2: pET28a) at 25°C for 30 min in a final volume of 20 μl in buffer containing 20 mM Tris–HCl (pH 8.0) and 150 mM NaCl. The samples were then separated on 1% agarose gel. M indicates size marker. (b) EMSA for detecting the direct interaction between AcrIIC4 and RNA. Purified AcrIIC4 was incubated with 10 μg of synthesized RNA (RNA1: UUAAUACGACUCACUAUAGG, RNA2: UAGCUCUAACCACGUUAAAC) at 25°C for 30 min in a final volume of 20 μl in buffer containing 20 mM Tris–HCl (pH 8.0) and 150 mM NaCl. The samples were then separated on 3% agarose gel for DNA analysis and 3% agarose gel for 30 min at 100 V.

Kim GE, Lee SY, Park HH. Crystal structure of the anti‐CRISPR, AcrIIC4 . Protein Science. 2021;30:2474–2481. 10.1002/pro.4214