PERC up your CRISPR delivery

…using amphiphilic peptides and NLS-rich CRISPR enzymes! This page aims to provide user-friendly info for anyone interested in adapting our protocols for PERC (peptide-enabled RNP delivery for CRISPR engineering). PERC is described in our NBME paper [PDF][SI], which resulted from a collaboration between the Wilson, Eyquem, and Marson labs. We will be updating this page from time to time, and feedback is welcome: get in touch with Ross at rosswilson@berkeley.edu

Our manuscript reporting the use of PERC for T cell engineering has coincided with an exciting complementary report from Junwei Shi and colleagues. They describe PAGE, a protocol that also uses amphiphilic peptides and NLS-rich CRISPR enzymes for editing of primary T cells (and more). Together, the two studies establish the substantial promise of this new approach to CRISPR delivery.

Peptide

Overview
Peptide A5K (GLFEKIEGFIENGWEGMIDGWYGYGRKKRRQRR; 4033.6 Da) is derived from the scaffold INF7-TAT (described here), and it was the main workhorse for our manuscript (Foss et al.).

Procuring peptide
The A5K peptide we used was synthesized (via custom order) by CPC Scientific, who has now added peptide A5K as a catalog item. This makes it more affordable to purchase small amounts of peptide, and will prevent wait times associated with custom synthesis (typically at least a month). If you ever need large amounts of peptide, be aware that it becomes more and more affordable as scale increases.

Preparing peptide
A5K will arrive as a powder, and a 5 mg (1.2 μmol) aliquot can be dissolved in 124 μL DMSO to make a 10 mM stock. This stock should be stable in the freezer indefinitely, although we haven’t performed extensive testing to validate this. We typically use 1 part 10 mM A5K stock (in DMSO) and mix it with 9 parts water to create a 1 mM stock containing 10% DMSO (typically 10 μL). We tend to make these 10% DMSO stocks fresh and don’t store them for re-use.

Protein

Overview
PERC works best with NLS-rich CRISPR proteins, but you should see some activity regardless of the protein you use. For optimal results, we recommend employing NLS-rich protein.

Cas9
There are a few options on this front, and the most appealing will be listed at top. The three protein constructs highlighted here performed best in our comparison (S-Fig. 5).
• triNLS: This construct was reported by Scot Wolfe and colleagues, and is referred to as “3xNLS”. We avoid that name since it suggests 3xSV40, while this construct has a trio of different NLS sequences. This protein tends to purify with very high yield, and its activity is quite good with PERC (based on lots of data not in the manuscript). The UC Berkeley MacroLab has purified ready-to-ship aliquots of this protein. To request this material, email Chris Jeans (c.jeans@berkeley.edu) and tell him you would like some Cas9-triNLS.
• TrueCut V2: This is a commercially-available Cas9 construct from Invitrogen.
• 6×NLS: This “self-delivering” Cas9 construct was reported by Jennifer Doudna and colleagues, and contains six SV40 NLS sequences (4 N- & 2 C-terminal). It is sometimes referred to as “4xNLS” in the initial manuscript. Although we used this protein for almost all the Cas9 experiments in the PERC manuscript, it can be challenging to express/purify in high yield, so we do not recommend working with it.

Cas12a
Our team has not explored Cas12a (formerly Cpf1) design space extensively, so the entries below are in no particular order. Only the top entry is commercially available; the other two have expression plasmids available via AddGene (below)
• Ultra variants: As reported by Christopher Vakulskas and colleagues at IDT, point mutations in Cas12a can activate the enzyme without sacrificing specificity. On the PERC manuscript, we had great results with A.s. Cas12a Ultra (from IDT) and in unpublished work we also saw reasonable editing rates of editing using the L.b. version. Interestingly, these proteins only contain a single SV40 NLS, suggesting room for improvement (in the context of PERC).
• Cas12a-T8N: This construct was used in the PAGE manuscript and features 6x N-terminal Myc NLS and 2x C-terminal SV40.
• NLS-rich Cas12a: Scot Wolfe and colleagues reported a handful of NLS-rich Cas12a constructs with improved activity. Although there are no reports of these constructs being used in peptide-mediated delivery yet, they may work well.

AddGene
Plasmids for recombinant expression of the non-commercial proteins above are available from AddGene. The top two Cas9 entries have cysteine-free plasmids available; if you are not interested in thiol-based conjugation, there is likely no practical difference between Cas9 with/without cysteine.
• triNLS: 2-Cys, 0-Cys (Wilson lab, adapted from Wolfe lab)
• 6×NLS: 2-Cys (Doudna lab), 0-Cys (Wilson lab)
• Cas12a-T8N (PAGE manuscript)
• NLS-rich Cas12a (Wolfe lab)

Protocols

Our materials/methods section covers the published work, and the text below elaborates further (note the linked shared doc at the bottom of this section).

We plan to provide additional context involving other things we have attempted (and how that worked out for us… and the cells), and therefore this section will continue to be updated.

Selection of guide RNA
Guides that result in high-efficiency editing with electroporation tend to also give high-efficiency editing with PERC. We purchase guides from Synthego or IDT with manufacturer-recommended standard chemical modifications. Resuspend sgRNAs in nuclease-free water to prepare a stock concentration of 100 μM. Pulse-vortex for 30 seconds after adding water to the lyophilized RNA and pipette to mix well. A 10 nmol vial can be resuspended in 100 μL DEPC-treated water. Store sgRNA at −20°C for several weeks, or -80C for longer.

Diluting and refolding guide RNA
Dilute sgRNA to 15 μM in ‘10× folding buffer’ (20 mM HEPES pH 7.5, 150 mM NaCl), then refold it by warming to 95°C for 5 min in a heat block and slow cool to room temperature for 25 min. We do this by removing the entire insert from the heat block (containing the Eppendorf tube) and placing it on the bench for 20–25 minutes. Once cool, briefly spin down the tube and add 2 mM MgCl2, pipette to mix.  

Preparing protein
Dilute protein in RNP buffer (20 mM HEPES pH 7.5, 150mM NaCl, 10% glycerol and 2mM MgCl2) to 10 μM. Bring diluted protein to room temperature before moving to the next step.

Assembling the RNP
The molar ratio of sgRNA:Cas protein when making RNP should be at least 1.2:1, with a volume ratio of 1:1. We typically make RNPs at a concentration of 5 μM. We do not recommend making RNPs at concentrations higher than 12.5 μM. Mix diluted Cas protein into the tube containing refolded sgRNA, making sure to pipette slowly while swirling the tip and simultaneously dispensing the protein. After addition, pipette 2–3 times to mix well. Incubate at 37°C for 15 minutes.
Repeating this since it’s important: mix the protein into the gRNA (not the other way around).

Peptide preparation
Peptides purchased from CPC Scientific are provided in lyophilized form. These should be stored in a desiccator at −20°C. Before resuspending, leave lyophilized vial at room temperature for ~20 minutes. Spin down the vial (while it’s wrapped in a kimwipe, inside of a 50 mL conical tube) at 2000 × g for 1 minute. Carefully break open the seal and resuspend the peptide in 100% DMSO to prepare a stock concentration of 10 mM. The molecular weight of INF7TAT-A5K is 4033.59 Da, so a 10 mM stock calls for 5 mg of peptide to be resuspended in 124 μL of 100% DMSO. Before adding peptide to RNP, dilute the 10 mM peptide stock (100% DMSO) to 1 mM in water by mixing 1 part of peptide stock with 9 parts of nuclease-free water. This results in a 1 mM working stock of peptide containing 10% DMSO.

Cell culture & editing via PERC
This shared document contains some of the text from the Foss et al. manuscript. It is being provided simply to make it easier to work with the text.