Abstract
The multifunctional proteasome activator PA28γ is involved in proteolytic pathways. Not surprisingly, upregulation of PA28γ contributes to tumor growth and progression. Therefore, CRISPR/Cas9 technology was used to specifically knockout the PSME3 gene encoding PA28γ in S462 tumor cells and HEK 293 cells. RNA guide sequences specific for PSME3 were designed and cloned into the lentiCRISPRv2 expression vector. Initially, successful Cas9-mediated gene-editing was verified by T7 endonuclease I DNA-mismatch assay. After the clonal selection, several clones showed no detectable PA28γ expression on protein level. Subsequently, a complete knockout of PA28γ was confirmed on genomic level by sequencing.
Initial functional analysis of effector caspase activity in PA28γ knockout cells showed a higher basal activity and an increased sensitivity towards doxorubicin/ABT-737-induced apoptosis. Generally, this confirms the role of PA28γ as an anti-apoptotic regulator. To our knowledge, this is the first study addressing the role of proteasomal regulators in malignant peripheral nerve sheath tumor (MPNST) derived S462 cells.
Background
The proteasome activator PA28γ is encoded by the PSME3 gene belonging to the proteasome activator complex subunit (PSME) gene family which modulates the 20S proteasome function. Recently, PA28γ has been established as anti-apoptotic key regulator in the ubiquitin- and ATP-dependent and -independent proteasomal pathway [1]. Moreover, PA28γ has been demonstrated to be involved in cell-cycle progression [2, 3], tumor growth [4] and apoptosis [5].
The knockout of proteasomal regulators by gene-editing is a useful tool to develop cancer cell models for anti-cancer drug development. Thus, clustered regularly interspaced short palindromic repeats/CRISPR-associated (CRISPR/Cas) technology was chosen for this technical note to explain in detail how a gene knockout is generated. Here, we knockout PA28γ in malignant peripheral nerve sheath tumor (MPNST) S462 cells.
Selection of the Cas expression system
The Cas9 enzyme of Streptoccocus pyogenes (SpCas9) is the most popular endonuclease used as genome-editing tool requiring a single guide RNA (sgRNA) to create specific double strand breaks (DSBs) at the target locus. These DSBs induce cellular DNA damage and repair mechanism mainly the error-prone non-homologous end joining (NHEJ) pathway, leading frequently to a frameshift and subsequently to a gene knockout. An optimal activity of the RNA-guided Cas9 endonuclease is critical for successful gene editing. A huge variety of Cas9-delivery systems are available ranging from expression vectors and lentiviruses (including selection cassettes e.g. puromycin or GFP) to synthetic sgRNAs in combination with recombinant Cas9 protein or mRNA. If you want to avoid random DNA integration, synthetic sgRNAs and Cas9 mRNA or protein can be used. The appropriate Cas9-delivery system is dependent on cell type and e.g. whether you want to enrich for Cas9-expressing cells or to establish a stable cell line. Delivery of the Cas9 protein has the advantage, that potential off-target effects are minimized since the nuclease is active only for a period of time. For gene-editing in difficult-to-transfect cells use a lentiviral-based system [6]. Here, a plasmid-based Cas9 expression system is selected using the lentiCRISPRv2 expression plasmid (Addgene plasmid #52961) allowing the co-expression of sgRNA and Cas9 as well as a puromycin selection cassette to enrich edited cells.
Selection of the target sequence
In the first instance, potential splice variants of the gene of interest are compared to identify the first constitutively expressed exon of the gene to avoid that the target sequence is spliced out. Moreover, N-terminal sgRNA targeting is recommended to decrease the risk of truncated protein expression.
Additionally, sequence variations of the target gene between cell lines should be considered, sequencing of the target region is recommended. Moreover, cancer cells are often polyploid resulting in allele amplification which in turn can affect the knockout efficiency [7]. The splice variants of the targeted PSME3 gene are identified using the UCSC Genome Browser. Among all known splice variants of the PSME3 gene, exon 3 is identified as the first common exon being subsequently used as sgRNA/Cas9 target region (Fig. 1A).
Workflow: Generation of PSME3 knockout cells.
For knockout approaches, Cas9 of S. pyogenes is mainly used. This Cas9 requires a Protospacer Adjacent Motif (PAM) directly upstream of the recognition sequence, more specifically NGG. In general, a 20 nucleotide (nt) target sequence is sufficient but, with respect to off-target effects within the genome, truncated sgRNAs have been shown to reduce these off-targets [8].
Many online tools are available to assist the design of sgRNAs based on algorithms to determine e.g. binding capacity as well as on- and off-target activity [9]. Off-targeting occurs due to imperfect complementarity between the sgRNA and the genome whereby the tolerance for mismatches grows with distance to the PAM sequence [10].
Choose more than one sgRNA as the efficiency of targeting is quite different depending on the target region and the cell line used. Use the CCTop-CRISPR/Cas9 online predictor for the sgRNA design and chose the length of the target site to 20 nt. The off-target prediction is set to 4 mismatches resulting in a list of potential off-target binding sites. Additionally, the human species is selected as the target is within the human cell lines HEK293 and S462. Finally, the in vitro transcription method is set to U6 (this depends on the chosen expression plasmid, here lentiCRISPRv2) which automatically adds the required 4 nt restriction site overhangs for cloning (underlined within the sgRNA sequence in section 3.5). The U6 promoter prefers a G as first transcribed nucleotide. Therefore, an additional G is added to sgRNA sequence if not already starting with one to make the transcription as efficient as possible. Out of the resulting list, select 2 or 3 sgRNAs with the lowest number of potential off-targets in an exonic region and with more than 3 mismatches. For the PSME3 gene: sgRNA T2 (fwd 5’ Order the complementary oligonucleotide and dilute to 100 μM stocks.
Design of DNA-mismatch detection primers
To validate successful gene-editing, the genetic target locus needs to be investigated. A fast and cost-efficient method is to PCR amplify genomic DNA and then subject the re-annealed PCR product to a DNA-mismatch cleavage assay e.g. by T7 Endonuclease I (T7EI). The primer pairs should be designed using the following criteria: i) the PCR product yields cleavage products between 200 and 800 bp to allow clear visualization by agarose gel electrophoresis [11] and ii) be specific to the target region. Design primers flanking each sgRNA target site. Use the online tool Primer Blast and select primer pairs that amplify PCR products with an optimal size of about 750 bp and without predicted side products. Use a proof-reading DNA polymerase e.g. high-fidelity Phusion (ThermoScientific) to amplify gene-targeting locus as described by manufacturer’s instructions. Perform a gradient PCR around the calculated annealing temperature. As both sgRNAs are located within one exon, one primer pair was designed and used at 500 nM: fwd 5’ATAAGCTATGAGGCTGAGGAAGG3’ and rv 5’TTCAGGTGGGGGTACTTGGGG3’. To each reaction add 3 % DMSO (v/v) and 0.1 μl genomic DNA extract (20 μl quick extract per 96-well). The PCR program is as followed: 98°C for 3 min, 38 cycles: 98°C for 15 sec, 67.7°C for 30 sec, 72°C for 1 min, and final elongation at 72°C for 10 min.
Construction of a sgRNA Cas9 expression vector
Depending on the chosen expression vector, sgRNA oligonucleotides were ordered with the according restriction overhangs. Firstly, hybridization and phosphorylation of the sgRNA oligonucleotides was performed following a combined digestion and ligation reaction to insert sgRNA upstream of the sgRNA scaffold. Since plasmids such as px458 (Addgene plasmid #48138) and lentiCRISPRv2 have type IIS restriction sites, a sequential digestion and ligation is almost 100 % efficient with increasing cycling number. Prepare a 10 μl hybridization mixture containing each 1 μl sgRNA top and bottom of a 100 μM stock, 1 μl T4 Ligation Buffer (10x), 6 μl ddH2O, and 1 μl T4 Polynucleotide Kinase (5 U). Incubate samples in a PCR thermo cycler at 37°C for 30 min; 95°C for 5 min and then ramp down to 25°C at 5°C min-1. For cloning, mix 100 ng lentiCRISPRv2 plasmid, 2 μl sgRNA oligo duplexes, 2 μl T4 Ligation Buffer (10x), 1 μl BsmBI restriction enzyme (10 U), 1 μl T4 Ligase (5 U) and adjust to 20 μl with ddH2O. Incubate samples in a PCR thermo cycler for 6 cycles of digestion at 37°C for 10 min and ligation at 25°C for 5 min. Transform 5 μl ligation reaction into 50 μl chemically competent E. coli Stbl3. Plate cells on lysogeny broth (LB) agar supplemented with 100 μg/ml ampicillin and incubate overnight at 37°C upside down. Next day, pick 2–4 colonies to perform a colony PCR with Taq polymerase using the human U6 primer (5’GATACAAGGCTGTTAGAGAGATAATT3’) and as reverse primer the corresponding bottom sgRNA. PCR conditions are as follows: 95°C for 3 min, 35 cycles of 95°C for 30 sec, 72°C for 2 min, 72°C for 30 sec, and final elongation at 72°C for 10 min. For each sgRNA, perform plasmid isolation of 2 positive clones and verify by sequencing using the human U6 primer.
Delivery of Cas9 and sgRNA by Lipofectamine transfection
Different approaches are available to deliver DNA into mammalian cells. In general, lentiCRISPRv2 plasmid can be transduced as lentivirus leading to a higher efficiency but this requires virus production. Alternatively, plasmids can be electroporated e.g. using the Nucleofector but it requires the instrument. Lipofectamine transfection has the advantage that it is easy to use and does not require any specific instruments. Take into account that the transfection efficiency is inversely proportional to the size of the transfected DNA molecule, as CRISPR plasmids are generally very large (10–15 kb), low transfection efficiencies can be expected. However, due to the subsequent clonal selection, the transfection efficiency is secondary. Nonetheless, px458 plasmid should be used to monitor the transfection efficiency. Grow S462 and HEK293 cells in high-glucose DMEM (Biochrom) with 10 % FCS superior (v/v) (Biochrom) until they reach 80–90 % confluence. Plate cells in 24 wells according to manufacturer’s instructions. S462 and HEK293 cells were seeded as 104 cells/well to achieve 80 % density. As transfection efficiency control use px458 plasmid and as negative control for the DNA-mismatch assay use the empty lentiCRISPRv2 plasmid. Perform transfection. For S462 and HEK293 cells, 1 μg DNA and 2 μl Lifpoectamine®2000 achieved the highest transfection efficiency. Incubate cells for 24 h, analyze by DNA-mismatch assay (see section 8) and subsequently treat with puromycin as described below.
Verification of sgRNA-guided gene targeting in cell pools
The DNA-mismatch assay is used to detect Cas9-induced insertions/deletions (indels) with T7E I. The T7E I assay is a good indicator for indel formation but has its limitations as it rarely detects single nucleotide mutations resulting in an underestimation of cleavage efficiency of Cas9 [13]. Here, your designed mismatch primers will be used. Prepare genomic DNA by adding 20 μl QuickExtract-DNA Extraction Solution (Epicentre) to a 24 well and incubate for 2 min at RT. Mix the solution in the well and transfer viscous content into a PCR tube and incubate for 15 min at 56°C followed by 15 min at 98°C in a thermo cycler. As control, use DNA extracted from unedited cells. Perform a 25 μl the Phusion PCR of the target region using the genomic extracts (see section 3). Confirm correct PCR amplicon using 5 μl of the PCR reaction by agarose gel electrophoresis. Add 2 μl NEB 2 buffer to the remaining 20 μl sample, then split into 2 tubes. For heteroduplex formation, the PCR products are initially denatured at 95°C for 5 min followed by a re-hybridization step, ramping down –2°C/sec to 83°C and –1°C/10 sec to 23°C in a thermo cycler. Add 0.2 μl of T7E I (2 U) to one PCR tube per condition and incubated for 30 min at 37°C. Visualize fragmented PCR products by agarose gel electrophoresis (as shown in Fig. 1C).
Clonal selection of genome-edited cell lines
This step is crucial! Due to transfection stress, allow cells to recover for one day in normal growth medium. Depending on the plasmid used for Cas9 and sgRNA expression, cells can be selected either by puromycin (lentiCRISPRv2) or sorted using GFP (px458) to enrich genome-edited cells. In case of puromycin selection, cell-specific puromycin concentration needs to be determined beforehand. Depending on cell type e.g. S462 cells, the determined antibiotic concentration might be cut by half and/or discontinued to avoid clonal death. Treat cells with antibiotic selection medium and change every 3 days. A 7-days dose response curve in a range from 0.5 to 10 μg/ml puromycin revealed that the optimal concentration of puromycin is 2 μg/ml for S462 cells and 1 μg/ml for HEK293 cells (use the lowest antibiotic concentration which killed all cells after 7 days of selection). Isolate single cell clones and expand clones for analysis (Fig. 1B). Single cell expansion can be achieved either by limiting dilutions or sorting into 96 wells. Alternatively, cells can be plated sparsely on 10 cm dishes and later be transferred using cloning rings or trypsin discs. This strategy is useful for cell lines that will not survive single cell growth. Constantly monitor single colonies by microscopy until visible by eye. Then, transfer colonies using Trypsin/EDTA-saturated cloning discs. Place the discs on the marked colonies with sterile tweezers, incubate and remove by softly pressing to the dish.
Verification of knockout cells
Expanded clones can be tested again for indel formation using e.g. a DNA-mismatch assay. However, the evidential loss of the protein expression by Western Blot is essential to verify a functional knockout. If no protein expression is detectable, the complete knockout should be verified on genomic level. This can be achieved by sequencing which can be analyzed by various online tools e.g. TIDE to identify the exact mutation. Alternatively, allelic variations, especially for polyploid cells such as cancer cells, can be identified by cloning the PCR-amplified genomic target locus into the pJET blunt vector and sequencing. Finally, implement mutated sequence in the coding sequence to determine frame shifts leading to premature stop codons.
A. DNA-mismatch assay Optional: Extract genomic DNA and subject to PCR and T7E I assay as previously described in section 7. Analyze the PCR products including the wildtype DNA. “Side products” around the expected PCR product length in comparison to the wildtype indicate larger indel formation. Moreover, non-cleaved PCR products in the T7E I assay are not necessarily a negative result as this can also be the consequence of single nucleotide or homozygous mutations.
B. Western Blot For Western Blot analysis, harvest cells, lyse in RIPA buffer (50 mM Tris/HCl pH 7.6; 150 mM NaCl; 1 % NP-40 (v/v); 0.5 % sodium deoxcycholate (w/v), 0.1 % SDS (w/v); 1 x Protease inhibitor cocktail) and determine protein concentration with Micro BCA protein Assay kit (Pierce). Prepare 30 μg protein in 6x SDS loading buffer (250 mM Tris/HCl pH 6.8; 8 % SDS (w/v), 60 % Glycerol (v/v); 15 % β-Mercaptoethanol (v/v); Bromphenol blue), denature for 5 min at 95°C and separate proteins. Transfer proteins to a nitrocellulose membrane. After membrane blocking, incubate with primary antibody PA28γ (K58.4; [5]), diluted 1:5,000 in 2 % BSA/PBST (w/v) or β-actin (Cell Signaling) for 60 min while agitating. As secondary antibody use anti-rabbit antibody coupled to horseradish peroxidase, diluted 1:10,000 in 2 % BSA/PBST (w/v). Incubate membrane for 2 min in ECL-solution (10 ml of 100 mM Tris/HCl pH 8.5; 3.5 μl 30 % H2O2 (v/v); 20 μl 100 mM p-Coumaric acid in DMSO; 100 μl of 250 mM Luminol in DMSO) and detect the chemiluminescent signal.
As shown in Fig. 1C, no PA28γ expression was detectable in all tested S462 and HEK293 clones treated with sgRNA T8. While with sgRNA T2, PA28γ expression was only decreased in tested HEK293 clones (Fig. 1C) and not affected in S462 cells (data not shown).
C. Allele-specific sequencing for heterozygous mutations PCR amplify the targeted genomic locus as described in section 4, ligate it into pJET1.2/blunt plasmid (ThermoScientific) and transform using chemically competent E. coli. Isolate the plasmids of at least 10 randomly picked colonies depending on the ploidity of the investigated cell line for allele-specific sequencing. The absence of wildtype sequence is a good indicator that all alleles are edited. The expected number of target copies can be evaluated e.g. by FISH analysis or karyotyping.
Functional analysis of the knockout: Reduced resistance towards apoptosis in PA28γ knockout cells
The expression level of PA28γ has been shown to modulate the cellular response towards apoptosis [5]. To proof this concept, cells were treated with apoptosis inducing anticancer drugs e.g. doxorubicin and analyzed by measuring caspase-3/7 activity. Per condition, seed 106 cells per well of a 6 well plate and incubate for 24 h at 37°C and 5 % CO2. Treat cells with apoptosis inducing anticancer drugs. For S462 cells, a treatment of 0.5 μg/ml doxorubicin alone or in combination with 1 μM ABT-737 (Bcl-2 inhibitor) was used for 24 h. Harvest and lyse cells for protein determination. Incubate 40 μg total protein together with 40 μM Ac-DEVD-AMC substrate for up to 4 h at 37°C. Detect the caspase-3/7 activity by measuring each hour the AMC released spectrofluorometrically using an excitation wavelength range of 340–360 nm and an emission wavelength range of 440–460 nm (CytoFluor).
As shown in Fig. 1D, the caspase-3/7 activity is significantly increased in cells lacking PA28γ compared to unedited control cells with or without apoptosis-inducing agents.
Discussion
The CRISPR/Cas9 gene-editing system is a fast and powerful method to knockout the PSME3 gene in S462 and HEK293 cells. The data indicate that the design of more than one sgRNA is advisable, since even with perfectly matched sequences the knockout efficiency can vary considerably. Here, T8 sgRNA led in all tested clones to a complete knockout in both cell lines while T2 sgRNA induced maximal a partial knockout (as shown in Fig. 1C). However, cancer cells as the S462 cell line are often polyploid resulting in a higher copy number of the target gene making gene-editing more challenging [7].
The detection of indels by DNA-mismatch assay is a good approach if only one sgRNA is inducing a DSB. A faster method to detect gene-edited cells is by designing two sgRNAs in close proximity to induce a partial deletion of the targeted genomic sequence. Consequently, the shortened PCR amplicons of this genomic locus are easy to detect. Alternatively, Cas9 nickase can be used to increase the specificity and thereby reduce potential off-target effects [10]. To introduce DSBs, two Cas9 nickases directed by different sgRNAs for the target locus have to be designed. With a deletion of more than 100 bp, a reliable loss-of function may be obtained whereas small mutations, introduced by one sgRNA-guided Cas9, may result in frameshifts that not always induce a loss of function of the protein of interest as hypomorphic alleles may be produced [15]. Nonetheless, two sgRNAs need to bind at once resulting in a lower knockout frequency [14].
In conclusion, the use of sgRNA/Cas9 expression vector combined with lipofectamine and puromycin selection is a valid method to generate knockouts in S462 and HEK293 cells.
Moreover, the functional knockout of PA28 was proven as the generated S462 PA28γ knockout cell line showed a correlation between loss of PA28γ and sensitivity towards apoptosis. Caspase-3/7 assay showed an enhanced activation of effector caspases in total lysates of PA28γ knockout cells compared to the wildtype (Fig. 1D). These data correlate with our previous finding that effector caspase activity was significantly increased in PA28γ silenced HT29 colon carcinoma cells [5].
Conflict of interest
The authors declare no conflicts of interest.
Footnotes
Acknowledgments
This work was supported by the Ministry of Science, Research and Culture (MWFK) Brandenburg (project number c90110631) and the Health Campus Brandenburg, Cluster ”Consequences of age-associated cell and organ functions“ (project number 92106360). We thank Dr. Anja Harder from the Brandenburg Medical School Theodor Fontane (MHB) who kindly provided the S462 cells.