Programmed Self-Assembly of an Active P22-Cas9 Nano Carrier System

Abstract

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) RNA-guided endonucleases are powerful new tools for targeted genome engineering. These nucleases provide an efficient and precise method for manipulating eukaryotic genomes; however, delivery of these reagents to specific cell-types remains challenging. Virus-like particles (VLPs) derived from bacteriophage P22, are robust supramolecular protein cage structures with demonstrated utility for cell type-specific delivery of encapsulated cargos. Here, we genetically fuse Cas9 to a truncated form of the P22 scaffold protein, which acts as a template for capsid assembly as well as a specific encapsulation signal for Cas9. Our results indicate that Cas9 and a single-guide RNA are packaged inside the P22 VLP and activity assays indicate that this RNA-guided endonuclease is functional for sequence-specific cleavage of dsDNA targets. This work demonstrates the potential for developing P22 as a delivery vehicle for cell specific targeting of Cas9.

Graphical Abstract

Introduction

CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) loci and their associated cas gene (CRISPR-associated) are essential components of an RNA-guided immune system in bacteria and archaea that targets foreign DNA for degradation.^1–5 Immunity to viruses and plasmids is acquired by integrating short fragments of invading DNA into the CRISPR locus and CRISPR-derived transcripts are used to guide dedicated nucleases to complementary sequences in foreign targets that are introduced to the cell during an infection. Recently, CRISPR RNA-guided endonucleases have been repurposed for targeted genome editing ^6–10 and the Cas9 protein (CRISPR-associated protein 9) from Streptococcus pyogenes (SpCas9) was one of the first proteins to be repurposed for precise genome engineering in human cells.^11–13 Target DNA recognition by SpCas9 occurs in two steps. First, two arginines interact with the nucleobase of two consecutive guanines. Detection of this motif, referred to as a protospacer-adjacent motif (PAM) facilitates RNA-guided strand invasion of the adjacent DNA duplex.^14–17 Base pairing between the CRISPR RNA (crRNA)-guide and the complementary strand of the target DNA triggers a conformational change that allosterically activates the Cas9 nuclease domains, resulting in near simultaneous cleavage of both strands in the target duplex.¹⁶ The simplicity and efficiency of programming these nucleases has been exploited for creating sequence-specific genome modifications with significant potential for curing genetic diseases.^18–20 However, cell-type specific delivery of these programmable nucleases remains challenging.

Current systems for delivering Cas9 include electroporation, transfection, microinjection, or viral transduction,⁶ but each of these delivery methods have their own set of limitation and none of these systems provide a generalizable platform for cell type-specific delivery. To overcome these limitations, we have developed virus-like particles derived from the bacteriophage P22 as a delivery vehicle for the RNA-guided Cas9 nuclease from S. pyogenes. Self-assembly of the P22 coat protein into a 58 nm capsid relies on coat protein interaction with a helix-turn-helix motif on the scaffold protein (SP), while the rest of the scaffold protein can be significantly truncated with little or no effect on assembly.^{21, 22} A variety of active enzyme cargoes have previously been genetically fused to a truncated form of the scaffold protein.^23–31 Here, we utilize this strategy to genetically fuse Cas9 to a truncated form of the P22 scaffold protein, which acts as a template for capsid assembly as well as a specific encapsulation signal for Cas9 (Figure 1).

Figure 1.

General scheme for P22 virus-like particle (VLP) assembly and packaging of cargo. Two plasmids, one expressing SP-Cas9 (pMJ809) and the other expressing a sgRNA and the P22-CP, where cotransnformed into E. coli (BL21 DE3) cells. Transformed cells were grown at 37°C for approximately 3 hours. Then the temperature was decreased to 30°C and expression of SP-Cas9 was induced using IPTG. P22-CP expression was induced using arabinose at 7 hour. The sgRNA is constitutively produced. Cells are harvested after a total of 9 hours and the P22-encapsulated SP-Cas9:sgRNA is purified (far right).

Materials and Methods

Transformation, expression, and purification

P22-Cas9: pBAD (P22 CP under control of the AraC promoter; sgRNA expression controlled by constitutive promoter BBa_J23119) and pMJ806 pET-based vectors (SP-Cas9, promoter: T7) were co-transformed into BL21 (DE) E.coli and grown in 1L cultures inoculated with 10 mL starter culture (37°C, 200rpm). After 3 hours (OD₆₀₀ = 0.6), the cultures were induced with 0.05 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) and grown for 4 hours (30°C, 200rpm), followed by induction with 0.2 % arabinose (2 g per 1 L culture) and grown for 2 more hours (30°C, 200rpm). Cells were harvested by centrifugation at 5000 ×g for 15 min (SLA-4000, Sorvall). Cell pellets were suspended in 100 mM phosphate, 50 mM sodium chloride, pH 7.0 buffer and the cell slurry was sonication on ice (Branson Sonifier 250, Danbury, CT, power 4, duty cycle 30%, 3 × 5 min with 5 min intervals). Bacterial cell debris was removed via centrifugation at 12000 ×g for 45 min. The supernatant was then loaded on a 35% sucrose cushion and centrifuged at 235,000 xg for 50 min in an ultra-centrifuge (50.2Ti ultra centrifuge rotor). The resulting virus pellet was suspended in 100 mM phosphate, 50 mM sodium chloride, pH 7.0 buffer. Protein was further purified over a S-500 70 HR sephacryl size exclusion column (GE Healthcare Life Sciences). Wild type P22 was purified as described previously.³² SP-Cas9 and WT Cas9 were N-terminally His tagged and purified using a nickel column, followed by further purification using a sephacryl S-200 HR column. Purified proteins were analyzed using denaturing SDS-PAGE, transmission electron microscopy (TEM, Leo 912 AB), and size-exclusion chromatography coupled to multi-angle light scattering (Wyatt Technologies Dawn8+, Optilab T-rEX, Agilent Technologies 1200 HPLC).

Multi-Angle Light Scattering of P22-Cas9 and WT Cas9

P22-Cas9 and WT P22 were injected using an Agilent 1200 autosampler with 100 mM phosphate, 50 mM sodium chloride, pH 7.0 buffer. The buffer was degassed using an inline degasser. Samples were run over a WTC-100S5G guard column (Wyatt Technology Corporation) and a WTC-100S5 SEC column designed specifically for MALS (Wyatt Technology Corporation). The eluant was monitored using an in-line UV-Vis detector on the Agilent system as well as a Dawn Heleos 8 MALS detector and an Optilab T-rex RI detector (Wyatt Technology Corporation). All data were analyzed using ASTRA software from Wyatt. Samples were stored in the autosampler at room temperature, and the sample chamber in the RI detector was held at 25°C to reduce thermal drift. Molecular weights were determined from MALS and RI signals using the ASTRA software and dn/dc values of 0.185 was used for all proteins. The Cas9/P22 ratio was determined by subtracting the molecular weight of P22-Cas9 from P22 WT (see Figure 2c) and dividing by the molecular weight of Cas9.

Figure 2.

Programmed encapsulation of Cas9 in the P22 VLP. (a) Multi-angle light scattering coupled to size exclusion chromatography of P22-Cas9 (red) compared to WT-P22 (blue) (b) Transmission electron micrographs of WT-P22 (left) and P22-Cas9 (right) indicate encapsulation of Cas9 does not alter the morphology of P22 capsids. Scale bar 200 nm. Samples were stained with 2% uranyl acetate. (c) Approximately 20 Cas9 per P22 were encapsulated in the P22-VLP. This was determined by subtracting the molecular weight of P22-Cas9 from WT-P22 divided by the molecular weight of Cas9. (d) SDS-PAGE analysis comparing the P22 encapsulated Cas9 and P22 WT as a negative control. Coomassie stained protein gel showing WT-Cas9 with N-terminal fusion of 50 kDa maltose binding protein (MBP) (201.9 kDa;), SP-Cas9 (180.5 kDa), and P22-SP-Cas9 (180.5 kDa). (e) Poly-acrylamide gel comparing nucleic acids extracted from WT-P22, SP-Cas9, and SP-Cas9, and WT-Cas9. Cas9 was programmed with a 102-nucleotide sgRNA, which is observable in P22-Cas9, SP-Cas9, and WT-Cas9, but not WT-P22.

Transmission Electron Microscopy of P22-Cas9 and WT P22

P22-Cas9 and WT P22 were imaged by transmission electron microscopy (Leo 912ab) by negatively staining the sample with 2% uranyl acetate on formvar carbon coated grids.

Protein Gel Electrophoresis

P22-Cas9 was analyzed for presence of Cas9 and P22 protein. Analytical denaturing gel electrophoresis was performed in 12% (w/v) polyacrylamide [acrylamide:bis-(acrylamide) 29:1] slab gels. Protein was visualized using Coomassie Blue.

Nucleic Acid Gel Electrophoresis

P22-Cas9 was analyzed for presence of RNA using denaturing gel electrophoresis. Total nucleic acids were extracted using a standard phenol:choloroform extraction procedure. Samples were separated on 12% (w/v) polyacrylamide [acrylamide:bis-(acrylamide) 29:1] gels containing 7 M urea and TBE buffer pH 8.3 (89 mM Tris–borate pH 8.3, 2 mM Na₂EDTA). Nucleic acids were stained with SYBR gold (Invitrogen).

Oligonucleotide cleavage assay

DNA oligonucleotides (44 nucleotides, 50 pmol) were radiolabeled by incubating with 20 units of T4 polynucleotide kinase (New England Biolabs) and ~35 pmol (~230 mCi) [γ−³²P]-ATP (Promega) in 5 μL 10X T4 polynucleotide kinase reaction buffer at 37°C for 46 min, in a 50 μL reaction. Reaction mixtures were purified through an Illustra MicroSpin G-25 column (GE Healthcare) to remove unincorporated label. Duplex substrates (100 nM) were generated by annealing labeled oligonucleotides with 5x molar excess of unlabeled complementary oligonucleotide at 95°C for 5 min, followed by slow cooling to room temperature. DNA cleavage reactions were initiated by the addition of 1 μl target DNA (2 nM) to 90 nM Cas9:sgRNA in a cleavage assay buffer containing 20 mM HEPES pH 7.5, 100 mM KCl, 5 mM MgCl₂, 1 mM DTT, 5% glycerol. The samples were incubated for 1 h at 37°C. Reactions were quenched by the addition of 20 μl of loading dye (5 mM EDTA, 0.025% SDS, 5% glycerol in formamide) and heated to 95°C for 5 min. Cleavage products were resolved on 12% denaturing polyacrylamide gels containing 7 M urea and visualized by phosphorimaging (Storm, GE Life Sciences).

Plasmid DNA cleavage assay

Plasmid DNA (8 nM) was linearized by restriction enzyme DraII, then incubated for 60 min at 37°C with purified P22-Cas9 protein, WT P22, SP-Cas9, and WT Cas9 (100 nM) in a Cas9 plasmid cleavage buffer (20 mM HEPES pH 7.5, 150 mM KCl, 0.5 mM DTT, 0.1 mM EDTA) with10 mM MgCl₂. The reactions were stopped with 5X DNA loading buffer containing 250 mM EDTA, and the products were resolved using 1% agarose gel electrophoresis followed by SYBR gold staining. For the Cas9 mutant cleavage assays, the reactions were stopped with 5X SDS loading buffer (30% glycerol, 1.2% SDS, 250 mM EDTA) prior to loading on the agarose gel.

Results and Discussion

Virus-like particles (VLPs) derived from bacteriophage P22, can be produced in Escherichia coli by co-expressing the coat protein (CP) and scaffold protein (SP).³⁴ CP and SP self-assemble into non-infectious protein capsids that are 58 nm in diameter.^35–37 To determine if the P22 VLPs could be programmed for encapsulation of the RNA-guided Cas9 protein from S. pyogenes, the truncated scaffold protein (SP) was genetically fused to either the N- or C-terminal domain of Cas9 and co-expressed with the P22-coat protein (P22-CP) and a single guide RNA (sgRNA) (Figure 1). This co-expression system allows for temporal de-coupling of cargo and capsid (Supplemental Figure S1). Temporal control facilitates proper assembly of the SP-Cas9:sgRNA cargo prior to capsid formation, which helps to retain activity of the cargo protein.³⁸ The P22-CP was induced 4 hours after expression of Cas9 protein, based on previous work done with enzyme encapsulation in P22 (Figure 1).^23–31

Cultures are typically grown at 37 °C for P22 coat protein expression³⁹ and 16 °C for optimal Cas9 expression.⁴⁰ In order to determine the optimal temperature for expression of both proteins, cultures were grown at a range of temperatures from 16 °C to 37 °C. The results indicated P22 coat protein expressed at 37 °C, but Cas9 did not (Supplemental Figure 2). Similarly, Cas9 expressed at 16 °C, but P22 coat protein did not. The best temperature for expression of both proteins was found to be 30 °C. After induction with IPTG (SP-Cas9) and arabinose (P22-CP), cultures were grown at 30 °C followed by purification of P22-Cas9 VLPs from E. coli lysates using sucrose cushion ultra-centrifugation and size-exclusion chromatography (SEC). Results indicate fusing SP to the N-terminus of Cas9 (SP-Cas9) did not disrupt capsid formation, while fusing SP to C-terminus of Cas9 (Cas9-SP) resulted in incomplete particle formation as determined by size-exclusion chromatography and electron microscopy (Supplemental Figure S3). Thus, P22 encapsulated with the SP-Cas9 fusion protein was further characterized to determine number of enzymes encapsulated per capsid and how it compares to wild type P22 in size and morphology.

SEC elution profiles of the P22-Cas9 VLPs were superimposable with control WT P22 VLPs made without Cas9 or the sgRNA (Figure 2a), and TEM images of these complexes reveal particles that are the same size and shape as unmodified WT P22 control particles (Figure 2b). In addition, multi-angle light scattering coupled to SEC revealed no change in particle size, a tight size distribution of the P22-Cas9 similar to WT P22, and the expected increase in molecular weight indicative of packaging Cas9 on the interior cavity of the capsid (Supplemental Figure S4). The molecular weight difference between P22-Cas9 and the WT P22-VLP was used to estimate the number of Cas9 molecules per P22 capsid, which is consistent with approximately 20 SP-Cas9 molecules per P22 capsid (Figure 2c, see Table S1 for calculation). The P22 coat protein and SP-Cas9 are evident in coomassie blue stained SDS-PAGE gels, with higher intensity for the CP band as compared to SP-Cas9. This difference is consistent with 420 subunits of the CP encapsulation of ~20 molecules of SP-Cas9 (Figure 2d). To determine if the P22-VLP encapsulated SP-Cas9 is associated with an sgRNA, we extracted total nucleic acids, separated these nucleic acids using electrophoresis, and imaged the samples using SYBR gold staining (Figure 2e). In addition to packaging the sgRNA, the P22-Cas9 sample also contains several other nucleic acids not observed when SP-Cas9 or the P22-VLP are purified separately. RNase and DNase treatment of these nucleic acids suggest that the higher MW nucleic acids contain DNA while the smaller nucleic acids (<500 nt) are RNA (Supplemental Figure S5).

To determine if the encaspsultated SP-Cas9 is capable of cleaving a specific double-stranded DNA (dsDNA) target, activity assays were conducted with a 44-nt dsDNA containing a PAM and a complementary target sequence. The P22 encapsulated SP-Cas9 (P22-Cas9) cleaves the 44-nt dsDNA target, but no cleavage activity was observed with an oligonucleotide of the same size that did not contain a complementary target sequence (Figure 3a,b, and Supplementary Figure S6). In addition to the 44-nt dsDNA target, a larger plasmid DNA target was used to probe activity of Cas9 (Figure 3c,d). In this case, cleavage was observed only for WT Cas9 and the free SP-Cas9 fusion, but not for P22-Cas9.

Figure 3.

Encapsulated Cas9 retains RNA-guided nuclease activity. (a) Schematic of Cas9-mediated cleavage of 5’-radiolabeled dsDNA containing a PAM and a sequence complementary to sgRNA (upper panel), or a non-complementary dsDNA ((lower panel). (b) Cleavage of complementary (upper panel) or non-complementary (lower panel) dsDNA in the presence of various protein combinations. The complementary dsDNA is cleaved in lanes containing Cas9 (lanes 2, 3, and 4), while neither the P22 WT (lane 1) nor the buffer only samples (lane 5) result in nuclease activity. (c) Schematic of Cas9-mediated cleavage of a linearized plasmid harboring either a sequence complementary to sgRNA with a functional PAM, or a non-complementary plasmid. (d) The nuclease assays performed above were repeated with a linearized plasmid harboring a complementary (upper panel) or non-complimentary (lower panel) sequence. The plasmid is cleaved in lanes containing SP-Cas9 and WT-Cas9 (lanes 3 and 4). However, these large molecular weight targets are not cleaved by P22-Cas9 (lane 2).

P22 is a T=7 particle, with 2.5 nm pores at the 5-fold axis.⁴¹ The width and persistence length of dsDNA is approximately 2nm and 50 nm, respectively, thus the 44-nt DNA is expected to passively traverse through the pores with no conformational constraints inside the capsid. Our results indicate that P22 encapsulated Cas9 is functional and cleaves the target DNA, similar to wild-type (WT) Cas9 and the un-encapsulated SP-Cas9 fusion (Figure 3b) and that the P22 capsid serves as a molecular sieve that imposes a size limitation on DNA accessible to the encapsulated Cas9.

To determine if the P22 capsid stabilizes Cas9 by protecting it from protease cleavage, Cas9 and P22-Cas9 were subjected to proteolysis using trypsin. Results show Cas9 degraded within the first 15 minutes of digestion, whereas P22-Cas9 was not degraded after 24 hours. These results verify Cas9 is packaged within the P22-VLP and indicated that the P22-VLP provides protection for the Cas9 cargo (Supplemental Figure S7).

Conclusion

To develop a programmable delivery system, we fused the Cas9 protein to the scaffold protein of P22, which mediates self-assembly of the P22 coat protein to form a protein capsid in which the Cas9 is encapsulated. Using an inducible dual-expression system, the SP-Cas9 fusion protein was co-expressed together with an sgRNA and the P22 coat protein. In previously reported enzyme fusions, where the enzyme was encapsulated in P22, fusion were made the to the N-terminus of SP.^23–31 In this paper, we fuse Cas9 to the C-terminus of SP and show that “end swapping” does not compromise SP-mediated scaffolding of the P22-CP. This demonstrates the genetic versatility of SP-directed encapsulation and presents new opportunities for fusing cargos to either N- or C- terminal end of SP.

DNA cleavage assays showed that the P22 encapsulated sgRNA-Cas9 cleaves a 44-nt DNA target containing a PAM and a complementary sequence to the sgRNA, while a 44-nt non-target DNA or large linearized plasmid containing a PAM and a non-complementary sequence to the sgRNA could not be cleaved. This suggests that the P22 encapsulated sgRNA-Cas9 is selective for targets that contain a PAM and a target that is complementary to the sgRNA-guide. In addition, 2.5 nm pores in the P22 capsid serve as a molecular sieve that restricts access of large substrates to the Cas9 nuclease.

P22-VLPs may serve as a programmable delivery vehicle for Cas9, but further modifications that facilitate cell entry, and programmable release of the cargo will be required for optimizing in vivo delivery. However, nanoparticles are emerging as important vehicles for cell type specific delivery (REFS), and the versatility, genetic control, and programmability of VLPs make these systems attractive targets for optimization. The results presented here demonstrate programed encapsulation of Cas9 inside the P22-VLP, leaving capsid assembly unaltered and suggesting that P22 may be utilized as a potential system for delivery of nuclease active Cas9.

A LIST OF NONSTANDARD ABBREVIATIONS

VLP

virus like particles

CRISPR

Clustered Regularly Interspaced Short Palindromic Repeats

Cas9

(CRISPR-associated protein 9)

PAM

protospacer-adjacent motif

sgRNA

single guide RNA