do not necessarily reflect the views of UKDiss.com.
Our genome is constantly exposed to many different harmful stimuli and a large number of DNA lesions take place in our cells daily. DNA double strand breaks (DSBs) are a common form of DNA damage in which both strands of the DNA double helix are broken. They are highly toxic and cause genomic instability, although they also have physiological functions, such as allowing the generation of a wide variety of antibodies in the immune system. Therefore, cells have evolved a complex machinery of proteins that meticulously maintain genome integrity and regulate how the repair of these DNA lesions. In fact, this regulation is essential, since its defect is not only an important promotor of cancer, but also responsible for developmental, immunological and neurological diseases. DSBs are normally repaired by two different pathways: homologous recombination (HR) and non-homologous end joining (NHEJ). Homologous recombination uses the sister chromatid as a template to recover the information that was lost and it is mostly error-free, although it just can be used during S and G2 phases of the cell cycle. In contrast, non-homologous end joining could cause the gain or the loss of some nucleotides, being error-prone (Chapman et al., 2012).
REV7 (also known as MAD2L2) was primarily identified in a screen for yeast mutants that could reverse the mutagenesis induced by ultraviolet light (Lawrence et al., 1985). It is a small adaptor protein of 211 amino acids that has been reported to have many different roles in cellular homeostasis. It is part of the DNA polymerase ζ (Polζ), an error-prone polymerase involved in replicating damaged DNA opposite a lesion during translesion DNA synthesis (TLS). In this process, there are two sequential steps: insertion and extension. REV7 coordinates it by bridging interactions between REV1, an insertion polymerase, and REV3, the catalytic subunit of DNA polymerase ζ (figure 1) (Makarova & Burgers, 2015).
Figure 1. Role of REV7 in TLS. REV7 (here referred as MAD2L2 acts as an adaptor protein between REV1 and REV3 (Sale, 2015)
Years later, REV7 was rediscovered as MAD2L2 (or MAD2B) due to its homology with MAD2L1 (or MAD2), a component of the spindle assembly checkpoint (SAC) (Cahill et al., 1999). REV7 prevents premature activation of the anaphase promoting complex/cyclosome (APC/C) through CDH1 sequestering (Listovsky & Sale, 2013).
In addition, some proteomic studies have shown that REV7 is part of more complexes with different proteins, such as isoforms of the heterochromatin-binding protein HP1 (CBX3 and CBX5) and two zinc finger proteins (POGZ and ZNF828) (Vermeulen et al., 2010).
Furthermore, REV7 has been claimed to have a role on epigenetic reprogramming and maintenance of pluripotency, as REV7-deficient mice are viable yet show an important defect in the development of primordial germ cells (Pirouz et al., 2013; Watanabe et al., 2014).
Recent studies have hinted towards a function of REV7 in DNA repair. REV7 inactivation by a point mutation (REV7 V85E) has recently been described to produce a Fanconi anemia phenotype, confirming its role in the maintenance of genome integrity (Bluteau et al., 2016).
Interestingly, REV7 has been reported to participate in DSBs repair pathway choice, where 53BP1 is a key regulator. By recruiting RIF1, 53BP1 blocks the activity of nucleases inhibiting the accumulation of ssDNA intermediates and promoting NHEJ. In contrast, BRCA1 prevents RIF1 recruitment and induces 5’ end resection during S/G2, allowing the strand invasion by the sister chromatid, necessary for homologous recombination (figure 2) (Sale, 2015).
Figure 2. Role of REV7 in NHEJ. During non-homologous end joining, 53BP1 recruits RIF1 and limits DNA end resection. REV7 is likely to act downstream these two proteins, but its mechanism is not clearly understood. In homologous recombination, BRCA1 recruits phosphorylated CtiP and avoids RIF1 – 53BP1 function. Therefore, DNA strands are processed and homologous recombination can proceed (Sale, 2015)
BRCA1-deficient cells are sensitive to inhibitors of the polyADP-ribose polymerase (PARPi), such as olaparib. This drug avoids the repair of DNA single strand breaks (SSBs), that after the S phase are converted into DSBs. The lack of BRCA1-induced HR leads to inefficiently repaired DSBs causing cell death. Consequently, they are used for the treatment of HR-deficient tumours. However, the restoration of homologous recombination is a cause of PARPi resistance and failure of chemotherapy. This can happen by reversion mutations of BRCA1 or by deletion of 53BP1, that impedes RIF1 recruitment allowing end resection and HR to proceed (Bunting et al., 2010). Recently, a screen for factors that could promote loss of sensitivity to PARPi in BRCA1-deficient cells found REV7 as an important hit, meaning that it collaborates downstream of 53BP1 and RIF1 in promoting NHEJ (Xu et al., 2015).
At the same time, an investigation about DNA repair at telomeres showed that the loss of Trf2, a protein that protects chromosomes ends, produces telemere fusions and the activation of DNA damage response. Knockdown of REV7 promoted survival in the absence of telomere capping and caused elongated 3’ telomeric overhangs, revealing again an important function at inhibiting 5’ end resection (Boersma et al., 2015). Both studies concluded that REV7 collaborates with 53BP1 and RIF1 but independently of REV1 and REV3 in preventing end resection and promoting NHEJ.
Strikingly, apart from promoting NHEJ, REV7 seems to be necessary for the last steps of DNA synthesis during HR as part of DNA polymerase ζ (Sharma et al., 2012).
As a key player of NHEJ, REV7 has an important role in immunoglobulin class switch recombination (CSR), that is the process whereby mature B lymphocytes change the immunoglobulin isotype they express. To that end, B cells rearrange their immunoglobulin heavy-chain locus (IgH) through a deletion-recombination reaction. This reaction starts with the generation of programmed DNA DSBs by the activation-induced cytidine deaminase (AID), which are then resolved by several DNA repair pathways, such as non-homologous end joining between donor and acceptor regions (Matthews et al., 2014).
REV3, as the catalytic subunit of DNA polymerase ζ, is necessary for the embryonic viability of mammalian cells (Bemark et al., 2000; Lange et al., 2016). However, conditional REV3-inactivation reduced CSR efficiency in B cells, suggesting that REV7 could contribute to CSR through Polζ function (Schenten et al., 2009). In contrast, Xu et al. and Boersma et al. determined that REV7 played an important role in CSR independently of its interaction with REV3.
As previously introduced, REV7 forms a complex with REV3 and REV1 during translesion DNA synthesis. The study that identified the interaction between REV7 and REV3 used the yeast two-hybrid assay and limited the REV7-binding domain of REV3 to the region comprised between the residues 1776 and 2195 (Murakumo et al., 2000). A subsequent study delimited this region to the aminoacids 1847-1892 (Murakumo et al., 2001).
Further investigations showed that REV7 binds to a 9-aminoacid sequence in REV3 (1877-ILKPLMSPP-1885) by using the same type of assay. In this motif, when two of the three proline residues were changed to alanine, the interaction was lost (Hanafusa et al., 2010). Another study also identified an additional motif of interaction (1993-2003) and concluded that the consensus sequence for REV7-interaction was φφxPxxxxPSR (φ represents an alyphatic residue), sequence that is conserved in vertebrate REV3 proteins. Inmunoprecipitation (IP) assays with tagged-REV3 demonstrated that just the mutation of both REV7-binding sites (referred in this study as RBD1 and RBD2) eliminates REV3-REV7 interaction, whereas the modification of one of the REV7-binding sites alone does not disrupt the interaction. This suggested that REV7 could form an homodimer, being this dimerization favoured by REV3, and allowing more REV7-binding partners to interact the complex. Moreover, they demostrated that both binding sites were crucial for DNA damage tolerance (Tomida et al., 2015).
An interesting study showed the structure of human REV7 in complex with a human REV3 fragment (1847-1898) that just contains RBD1. Unfortunately, RBD2 was discovered after the co-crystal analysis and no crystal structure of RBD2 in complex with REV7 is available. In this study, REV7 showed that REV3-binding is similar to the interaction observed for Mad2 and Mad1. REV7 forms a seatbelt around REV3 that is formed by the residues 153-211. In similarity with Mad2, REV7 also adopts different conformations according to REV3-binding. In the absence of REV3, the conformation of REV7 is closed. However, in the presence of this protein, REV7 experiences a structural change of the seatbelt that makes two residues responsible for the interaction (Y63 and W171), as shown in figure 3. These aminoacids correspond to Y64 and W167 in MAD2. In vitro binding assays confirmed that the point mutations Y63A and W171A in REV7 significantly reduced the interaction with REV31847-1898 and in vivo co-immunoprecipitation retained a light affinity for REV31776-2044.
Figure 3. REV7 in complex with RBD1. RBD1 is shown in yellow and REV7 is coloured blue-green. The C-terminal region of REV7 (residues 153–211) corresponds to the seatbelt and is coloured green. Y63 and W171, responsible for the interaction with RBD1, are shown as stick models (Hara et al., 2010)
Interestingly, another mutation, REV7 R124A, stabilized the closed conformation of REV7 and enhanced REV3-binding.
The structural change in the seatbelt of REV7 upon REV3 binding provides a surface for REV1 interaction. The residues L186, Q200 and Y202 in REV7 were determined to be crucial for the interaction with REV1 (Hara et al., 2010)
Another study found a REV7 mutation in mice (C70R) that also causes the loss of the interaction with REV31847-1898 by in vitro interaction assays. The replacement of this cysteine residue with an arginine introduces a positive charge in the core of the protein that impedes the normal structure of the safety belt and REV3-binding (Khalaj et al., 2014).
This project is part of an ongoing research in the group, which is trying to decipher the role of REV7 in CSR. It is now clear that REV7 is a downstream effector of 53BP1 on inhibiting end resection and promoting CSR, but the mechanisms are not completely understood yet. In order to elucidate the minimum molecular requirements for the role of REV7 during CSR, a researcher in the laboratory performed a mutants’ screen for their ability to rescue CSR in the context of a REV7Δ background. Interestingly, one of the mutations that had been predicted to block REV7-REV3 interactions (REV7 Y63A) failed to rescue CSR whereas the mutation that stabilised the interaction (REV7 R124A) enhanced this process, questioning the idea of REV7 being independent of REV3 for CSR. Nevertheless, other REV3-interaction blocking mutations (REV7 W171A and REV7 C70R) showed little or no effect on CSR. Furthermore, the mutant of an ultra-conserved lysine (REV7 K129A) had not been previously identified and severely affected CSR. For that reason, our aim was to investigate the real effect of this mutants on REV7-REV3 interactions by using the yeast two hybrid with the intent of gaining understanding into the potential role of REV3 in CSR.
REV3 is a large protein of more than 3000 aminoacids and 350 kDa, fact that makes it difficult to find a suitable antibody. To overcome this limitation, previous interaction-assessment experiments used a fragment of REV3 for yeast two-hybrid assays and in vitro interaction assays, or the tagged version of the protein for immunoprecipitation (IP).
Figure 4. Yeast two-hybrid system. A) GAL4 DNA binding domain (BD) recognises a specific DNA sequence and GAL4 activation domain (AD) recruits the transcription machinery to trigger the expression of the genes under control of this sequence. B) If the fusion proteins interact, the two separable domains of GAL4 are close enough to activate transcription. C) If the fusion proteins do not show an interaction, the transcription machinery cannot be recruited and the expression of the reporter genes does not occur.
The technique used in this study is the yeast two-hybrid assay (Y2H), which is a general method for the identification proteins that interact with other proteins in vivo. This technology takes advantage of the fact that the GAL4 protein of the yeast Saccharomyces cerevisiae is a transcriptional activator with two functional domains that need to be in proximity to activate gene expression, but not necessarily in direct contact. It was originally developed by Stanley Fields in 1989 (Fields & Song, 1989). The N-terminal domain binds to a specific DNA sequence (UASG) and is named the DNA-Binding Domain (BD), while the C-terminal domain is necessary to activate transcription and is known by transcription Activation Domain (AD). These two separable domains can be fused to the proteins under investigation. Normally, the BD is fused to a protein called “bait” and the AD is fused to another protein referred as “prey”. Both fusion proteins are co-expressed in the same yeast cell and if they show an interaction, transcription of a gene under control of UASG will occur. This gene is a reporter gene and it will allow yeast to grow under certain conditions (Rajagopala, 2015). The use of a genetically engineered strain of yeast that lacks the biosynthesis of certain nutrients allows the selection of the modified cells as well as the identification of the reporter gene expression. In this case, we use a HIS3 reporter gene, that contains the imidazoleglycerolphosphate (IGP) dehydratase, the enzyme responsible for histidine biosynthesis, and the β-galactosidase, which produces a colorimetric reaction in the presence of ONPG (o-nitrophenyl-b-D-galactopyranoside) (figure 4). Analysis of yeast two-hybrid interactions can be done in plates lacking leucine, triptophane and histidine, as well as through a colorimetric assay.
- AD: GAL4 activation domain
- AmpR: ampicillin resistance
- att: attachment sites
- BD: GAL4 DNA-binding domain
- CARB: carbenicillin
- CSR: class-switch recombination
- DNA: deoxyribonucleic acid
- EB: ethidium bromide
- EV: empty vector
- HR: homologous recombination
- Ig: immunoglobulin
- IGP: imidazoleglycerolphosphate
- IP: immunoprecipitation
- GST: glutathione-S-transferase
- KAN: kanamycin
- KanR: kanamycin resistance
- LiAc: lithium acetate
- NHEJ: non-homologous end joining
- ONPG: o-nitrophenyl-b-D-galactopyranoside
- PARP: polyADP-ribose polymerase
- PARPi: polyADP-ribose polymerase inhibitors
- PCR: polymerase chain reaction
- PEG: polyethylene glycol
- Polζ: DNA polymerase ζ
- RBD: REV7-binding domain in REV3
- rpm: revolutions per minute
- TLS: translesion DNA synthesis
- YPD: yeast extract peptone-dextrose
- Y2H: yeast two-hybrid
- WT: wild-type
- 3AT: 3-aminotriazole
The first aim of this study was the optimization of the yeast two-hybrid technique as a tool to evaluate REV7-REV3 interactions.
Our second objective was the identification of novel point mutations in the REV7-binding domain of REV3 and to direct CRISPR-Cas9 mutagenesis studies. This would allow the generation of REV3 mutant B cells which lack the interaction with REV7.
The final aim was to correlate protein-protein interactions between an array of REV7 mutants and REV3 with the CSR phenotypes previously identified in mouse B cells.
DH5α and Stbl3 Escherichia coli were used for plasmid amplification.
- Liquid LB (Luria Bertani): 20 g of LB Broth Lennox powder (Sigma L3022) in 1L of water.
- Solid LB: 40 g of LB Broth with agar Miller powder (Sigma L3147) in 1L of water.
Autoclave for 15 minutes at 121ºC to sterilize. Allow to cool down before adding antibiotics: CARB stock solution (100 mg/ml 1000x) and KAN stock solution (50 mg/ml 1000x).
Bacteria were grown at 37ºC, with shaking (220 rpm) in the case of liquid cultures.
Saccharomyces cerevisiae strain: YF 2175(also known as PJ69-4A), with genotype MATa trp1-901 leu2-3,112 ura3-52 his3-200 gal4Δ gal80Δ LYS2::GAL1-HIS3 GAL2-ADE2 met2::GAL7-lacZ, originally used by James et al.,1996.
- Liquid YPD: 25 g of YPD Broth (Formedium CCM0205) in 500 ml of distilled water.
- Solid YPD: 35 g of YPD Agar (Formedium CCM0105) in 500 ml of distilled water.
- Liquid -Trp-Leu: 13.45 g of SD Broth / 2% glucose (Formedium CSM0205) and 320 mg of Complete Supplement Mixture Drop-out –LEU –TRP (Formedium DCS0561) in 500 ml of distilled water.
- Solid -Trp-Leu: 22.45 g of SD Agar / 2% glucose (Formedium CSM0105) and 320 mg of Complete Supplement Mixture Drop-out –LEU –TRP (Formedium DCS0561) in 500 ml of distilled water.
- Solid -Trp-Leu-His: 22.45 g of SD Agar / 2% glucose (Formedium CSM0105) and 310 mg of Complete Supplement Mixture Drop-out –HIS –LEU –TRP (Formedium DCS0971) in 500 ml of distilled water.
All of them were autoclaved before use.
- Solid -Trp-Leu-His + 3-AT: add 3-aminotriazole adjusting the concentration to -Trp-Leu-His agar after it has been autoclaved. 1 M 3-AT stock solution: 0.84 g 3-amino-1,2,4-triazole in 10 ml of ultrapure water. Filter sterilize. Store at -20ºC.
Yeast were grown at 30ºC, with shaking (200 rpm) in the case of liquid cultures.
The region that specifically binds REV7 in REV3 was amplified (REV31775-2200). The PCR product was amplified with attB-modified custom primers so that they were suitable for gateway cloning:
The PCR protocol was done in the absence of polymerase (negative control) and in two conditions: with GC enhancer and without GC enhancer, according to manufacturer’s instructions. PCR thermocycling conditions were set up on the BIORAD T100TM Thermal cycler.
Table 1. PCR conditions for REV7-binding domain of REV3 amplification
The PCR product obtained with GC enhancer was purified from a TBE gel.
- GatewayTM pDONR™221 Vector (Life Technologies 12536017). The map is shown in figure 1 of the appendix.
- pAD-DEST: Y2H destination vector GAL4 activation domain. TRP yeast marker gateway compatible. The map is shown in figure 2 of the appendix.
- pBD-DEST: Y2H destination vector GAL4 DNA-binding domain. LEU yeast marker gateway compatible.
To make the RBDm-p221DONR, the Q5® site-directed mutagenesis kit (NEB E0554S) was used. This kit is an efficient and fast method for site-specific mutagenesis of double stranded plasmid DNA. It consists of a PCR with designed mutagenic primers that will create insertions, deletions or sustitutions. The primers were designed by the online tool NEBaseChangerTM.
RBD1m (P1880A P1885A)
RBD2m (P1996A P2001A)
The procedure for the Q5®-site directed mutagenesis is the following, according to manufacturer’s intructions:
- Exponential amplification (PCR). Mix the following reagents in a PCR tube and set the following cycling conditions in the:
Table 2. PCR conditions for the Q5 site directed mutagenesis
- KLD treatment: the PCR product was added to a Kinase-Ligase-DpnI (KLD) enzyme mix for its circularization and removal of the original template. Mix the following components and incubate for 5 minutes at room temperature: 1 μl PCR product, 5 μl 2x KLD Reaction buffer, 1 μl 10x KLD Enzyme Mix and 3 μl of nuclease-free water.
- Transformation with 5 μl of KLD mix.
These plasmids were then verified by sequencing.
Gateway Cloning technology is a very fast and efficient way of cloning that is based on site-specific recombination properties and does not require restriction enzymes. Recombination takes place between attachment sites of DNA, abbreviated att. There are two reactions in gateway cloning: the BP reaction uses attB and attP sites and the LR reaction uses attL and attR.
The gene of interest (PCR product) flanked by attB sites was inserted into a Gateway entry vector (GatewayTM pDNORTM221) with attP sites. The resulting entry plasmid contained the gene of interest flanked by attL sites.
The procedure for transferring the PCR product to a Gateway entry clone (Gene-pDONR221) was performed by mixing the following components to a 1.5 ml tube at room temperature: 100 ng attB-PCR product, 100 ng donor vector, 1 µl BP Clonase ™ II enzyme mix (Life Technologies 11789020) and TE buffer (10 mM Tris and 1 mM EDTA) pH 8.0 up to 4 µl. The reactions were incubated at 25ºC for an hour.
The entry clone was flipped into a destination vector containing attR sites. The resulting plasmid was the expression clone.
The reaction for transferring the gene from a Gateway entry clone (Gene-pDONR221) to a destination vector was performed by mixing the following components to a 1.5 ml tube at room temperature: 100 ng Gene-pDONR221 , 100 ng destination vector (either AD or BD), 1 µl LR Clonase ™ II enzyme mix (Life Technologies 11791020) and TE buffer pH 8.0 up to 4 µl. The reactions were incubated at 25ºC for an hour.
A chemical transformation was performed by using 5x KCM (0.5 M KCl, 0.15 M CaCl2, 0.25 MgCl2), 2 μl of each plasmid and 50 μl of E. coli competent cells. After incubation on ice for 20 minutes and 10 minutes at room temperature, 950 μl of SOC medium was added. They were grown for an hour at 37ºC with shaking (450 rpm). The resulting pellet was plated in a LB agar plate containing the selection antibiotic (Kanamycin for the BP reaction and Carbenicillin for the LR reaction).
The day before the transformation, S. cerevisiae was inoculated into 50 ml of YPD media and grown overnight at 30ºC. Transformation efficiency is increased if the cells are in logarithmic phase of growth, so yeast were grown up to a OD600 of 0.6 (Eppendorf BioPhotometer® D30). Centrifuge them at 3500 rpm for 3 minutes and resuspend pellet in 25 ml of miliQ water. Centrifuge them again and resuspend pellet in 1 ml 1x TE/1x LiAc. 50 μl of chemically competent cells were used for each transformation in 600 μl of a PEG/LiAc/TE solution (480 μl 50% PEG 3280, 60 μl of 10x TE and 60 μl of 10x LiAc) and 100 ng of each fusion plasmid, both AD and BD. Incubate at 30ºC for 30 minutes with shaking (600 rpm) and heat shock at 42ºC for 15 minutes with shaking (600 rpm). Then place the tubes on ice for 3 minutes and spin them down at 4000 rpm for 3 minutes. Resuspend cells in 100 μl of nuclease-free water and plate them with beads in -Trp-Leu plates. Let them grow for 3 days at 30ºC and pick two transformed colonies.
The PCR product was gel-extracted following the “DNA extraction from agarose gels” protocol (Nucleospin® Gel and PCR Clean-up – Macherey Nagel 740609.50).
The Nucleospin® plasmid kit (Macherey Nagel 740499.250) was used following the “Isolation of high-copy plasmid DNA from E. coli” protocol.
Plasmid DNA concentration was analysed with a spectrophotometer (NanoDropTM 1000).
In order to check the plasmids, a diagnostic digestion is performed. Check the plasmid map on SnapGene and choose the suitable restriction enzymes that will cut in both the backbone and the inserted region. 300 ng of DNA were used with 0.5 μl enzyme(s) in 10x FastDigest Green Buffer (Life Technologies B72). The reaction was incubated for 20 minutes at 37oC.
Digestion product and PCR product were loaded in a 1% agarose – 1:50 ethidium bromide electrophoresis gel. After running at 120 V for 30 minutes (PowerPac™ Basic Power Supply – BioRad), gel was visualized in the UV transilluminator (ChemiDoc MP Imaging System – BioRad). Digestion band patterns were analysed to verify the reactions.
Two days before, transformed clones were inoculated in 10 ml of -Trp-Leu media and incubated in the 30ºC shaker. After two days, spin the tubes down (5 minutes at 3500 rpm), discard supernatant and resuspend in 5 ml of miliQ water. Spin them down again (same conditions), discard supernatant and resuspend in 5 ml of miliQ water.
Measure OD600 of 1:20 dilutions and normalize the number of cells to have 2×107 per 200 μl in the first column of a 96-well plate. Do a serial dilution in each row according to the conditions of plating.
Yeast are seeded in plates lacking tryptophan and leucine (-Trp-Leu), which were used as a loading control, and in plates lacking also histidine (-Trp-Leu-His), which was the readout of Y2H interactions. Plates with a certain concentration of 3-aminotriazole were also used. This molecule is an inhibitor of the enzyme responsible for histidine synthesis, so that just the strong interactions will allow yeast to grow and the background can be removed. Add a 20 μl drop to each plate (after mixing very well). Let them grow for approximately 48 hours and scan the plates. The process is shown in the Appendix (figure 3).
The first experiment was carried out with 10-fold dilutions of yeast number, starting with 2×107 cells in 200 μl. Yeast cells were transformed with REV3 (full-length) together with REV7 Y63A mutant, REV7 WT as a positive control or GST as a presumably negative control, in both orientations. The experiment showed that the technique was working just in one orientation (AD REV7 – BD REV3). As shown in figure 5, the Y63A mutant of REV7 shows a slightly weaker interaction with REV3.
Figure 5. Yeast two-hybrid analyses for interactions between REV3 and REV7 WT, GST or REV7 Y63A using 10-fold dilutions. The plate lacking tryptophan and leucine (-Trp-Leu) was used as a loading control whereas the plate that also lacks histidine (-Trp-Leu-His) was the readout of the interactions. AD REV7 WT – BD REV3 was the positive control of the experiment and AD GST – BD REV3 was considered as a negative control. Representative of 2 biological repeats.
As there was an important difference from one dilution to the next one, the same experiment was performed with 5-fold dilutions, showing equivalent results to the first one (figure 6).
Figure 6. Yeast two-hybrid analyses for interactions between REV3 and REV7 WT, GST or REV7 Y63A using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His plate was the readout of the interactions. AD REV7 WT – BD REV3 was the positive control of the experiment and AD GST – BD REV3 was considered as a negative control. Representative of 2 biological repeats.
Since the technique was showing some differences between the mutant and WT, the same experiment was carried out with more mutants of REV7 (C70R, W171A, K129A and R124A) just in the right orientation. The dilutions in this case started with 4×106 cells in 200 μl. The different mutants showed subtle differences in REV3-binding (figure 7), but they were not as clear as previously published experiments.
Figure 7. Yeast two-hybrid analyses for interactions between REV3 and REV7 WT, C70R, W171A, R124A, K129A or GST using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His plate was the readout of the interactions. AD REV7 WT – BD REV3 was the positive control of the experiment and AD GST – BD REV3 was considered as a negative control. Representative of 2 biological repeats.
For the previous experiments, the GST was considered as a negative control. However, it was still able to grow in the -Trp-Leu-His plate, showing an interaction between GST and REV3. In order to remove the background and find the right negative control, we carried out another experiment trying increasing concentrations of 3-aminotriazole (3AT) and empty vector (EV) alone (either AD or BD). As shown in figure 8, the BD EV was the right negative control, as it did not grow in the -Trp-Leu-His plate, even without inhibitor.
Figure 8. Yeast two-hybrid analyses for interactions between REV7 WT-REV3, GST-REV3 and REV7 WT-EV with increasing concentrations of 3AT. Representative of 2 biological repeats.
Having found what was presumably the negative control, another transformation was done with the array of mutants and also AD EV together with REV3. As shown in figure , AD EV – BD REV3 also exhibited an interaction (figure 9). Therefore, we concluded that REV3 is probably a sticky protein due to its big size and there was a considerable background interaction, making the rest of the results not reliable.
Figure 9. Yeast two-hybrid analyses for interactions between REV3 and REV7 WT, Y63A, C70R, W171A, R124A or K129A using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His + 1 mM 3AT plate was considered the readout of the interactions. AD REV7 WT – BD EV and AD EV – BD REV3 were supposedly the negative controls of the experiment. Representative of 2 biological repeats.
As REV3 full-length is a very large and sticky protein, we amplified the region in REV3 that specifically binds REV7 (REV31775-2200), which presents the two REV7-binding motifs, each one containing the consensus sequence φφxPxxxxPSR. These binding sites are referred as RBD1 (P1880, P1885) and RBD2 (P1996, P2001) (figure 10). The following transformation was carried out using the combination of REV7 mutants together with the amplified fragment of REV3 (REV31775-2200), in the orientation that was described in the literature: AD REV3 – BD REV7. Neither AD EV – BD REV7 WT nor REV31775-2200 had an interaction, meaning that they were the right negative controls for the rest of the study (figure 11).
Figure 10. REV3 protein scheme with the amplified fragment (REV31775-2200), marked in orange, that corresponds to the REV7-binding domain of REV3. This domain contains both motifs of interaction: the first motif (RBD1) includes P1880 and P1885 whereas the second motif (RBD2) includes P1996 and P2001.
Figure 11. Yeast two-hybrid analyses for interactions between REV31775-2200 and REV7 WT, Y63A, C70R, W171A, R124A or K129A using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His + 1 mM 3AT plate was considered the readout of the interactions. AD EV – BD REV7 WT and AD REV31775-2200 – BD EV were the negative controls of the experiment. Representative of 2 biological repeats.
A higher concentration of 3AT (2 mM) was not able to make the differences more significant (figure 12). The fact that just little or no differences could be detected revealed a higher complexity in the interactions between these two proteins.
Figure 12. Yeast two-hybrid analyses for interactions between REV31775-2200 and REV7 WT, Y63A, C70R, W171A, R124A or K129A using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His + 2 mM 3AT plate was considered the readout of the interactions. AD EV – BD REV7 WT and AD REV31775-2200 – BD EV were the negative controls of the experiment. Representative of 2 biological repeats.
To further investigate the contribution of each motif to REV7-binding we mutated the two conserved prolines in the REV7-binding motifs of REV3. The RBD1 mutant (RBD1m) contains P1880A P1885A whereas the RBD2 mutant (RBD2m) presents P1996A P2001A. The RBD double mutant (DM) contains the four substitutions (P1880A P1885A P1996A P2001A). As shown in figure 13, when both REV7-interaction motifs were disrupted separately (RBD1m and RBD2m), the experiment showed no effect on REV7-binding, even in a high concentration of 3AT. In contrast, the RBD DM completely impaired the interaction. This result confirms that there are two different motifs of interaction in REV3 that redundantly interact with REV7, so that both of them need to be altered to completely block the interaction.
Figure 13. Yeast two-hybrid analyses for interactions between REV31775-2200, RBD1m, RBD2m or RBD DM and REV7 WT using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His + 10 mM 3AT plate was considered the readout of the interactions. AD EV – BD REV7 WT and AD REV31775-2200 – BD EV were the negative controls of the experiment. Representative of 2 biological repeats.
The previous experiment revealed that one of the motifs of interaction was enough to bind REV7. Therefore, the results of the first experiments, which showed that some REV3-binding blocking REV7 mutants were still binding the REV3 fragment, could be explained if these mutants (Y63A and W171A) just disrupted the interaction with the first motif of interaction (P1880 P1885), but were able to bind the second one (P1996 P2001). To prove that hypothesis, we designed an experiment where we combined the REV7 mutants with the RBD mutants.
Figure 14. Yeast two-hybrid analyses for interactions between REV31775-2200, RBD1m, RBD2m or RBD DM and REV7 WT, Y63A, W171A, Y63A/W171A, K120A, C70R or R124A using 5-fold dilutions. The -Trp-Leu plate was used as a loading control and the -Trp-Leu-His + 1 mM 3AT plate was considered the readout of the interactions. AD EV – BD REV7 WT and AD REV31775-2200 – BD EV were the negative controls of the experiment. Representative of 4 biological repeats (2 different clones from 2 different transformations). Another set of clones is shown in the Appendix (figures 4-9).
The results derived from this experiment (figure 14) reveal that:
- REV7 Y63A interacted with REV31775-2200 (WT) and with RBD1m but not with RBD2m, as shown in the plate containing 3AT. This means that this mutant is interacting with the REV3 fragment when the first motif (RBD1) is altered because it can still bind the second one (RBD2). However, when the second motif (RBD2) is also modified, then it cannot interact with either of them. Therefore, RBD2 is responsible for the interactions with REV7 Y63A, since this point mutation blocks RBD1-binding but does not affect RBD2-binding.
- REV7 W171A shows the same binding profile as the Y63A.
- Regarding REV7 Y63A/W171A, cells did not grow in any of the conditions, meaning that either this mutant does not interact with this REV3 fragment or it is unstable and it is not even expressed in the cells.
- Consistently, all four biological repeats of REV7 C70R exhibited a normal interaction with RBD1m but a slightly weaker interaction with RBD2m in the plate with 3AT, possibly because this modification minimally impairs RBD1-interaction.
- REV7 R124A and K129A did not show any significant difference compared to REV7 WT.
The yeast two-hybrid technique had been previously used to assess the interactions between REV3 and REV7 (Murakumo et al., 2000; Murakumo et al., 2001). None of them ever examined REV3 (full-length) possibly because its size facilitates non-specific interactions, as shown in our experiments, where it was able to bind GAL4 AD alone (figure 9).
As mentioned in the introduction, REV3 interacts with REV7 via two REV7-binding sites (residues 1877-1887 and residues 1993-2003). Both of them need to be altered to eliminate the interaction, as shown in IP experiments (Tomida et al., 2015). Contradictorily, the study that examined the interaction between REV31915-2196 and REV7 by yeast two-hybrid had determined that they did not show an interaction (Murakumo et al., 2001).
In our investigation, the examined REV3 fragment contained the domain enclosing residues 1775-2200, so that it includes both REV7-binding motifs (named RBD1 and RBD2), and the assay was optimised by finding the accurate negative controls of the experiment (AD EV – BD REV7 WT and AD REV31775-2200) as well as the positive control (AD REV31775-2200 – BD REV7 WT). The introduction of REV7 mutants into the system showed that those which have been previously reported to block the interaction with REV31847-1898, such as Y63A and W171A (Hara et al., 2010), did not have a relevant effect on REV31775-2200 binding. More interestingly, the C70R mutant of REV7, that had been reported to block the interaction with REV31847-1898 (Khalajet al., 2014), did not show any difference in the yeast two-hybrid interaction when compared to WT. Similarly, the R124A mutation, that was supposed to stabilise the complex (Hara et al., 2010), showed a hardly perceivable effect. These delicate differences as well as the discovery of two binding motifs could suppose a higher complexity in the interaction between these two proteins.
The experiment with the different RBD mutants proved that one of the motifs of interaction was enough to bind REV7, as previously shown by IP (Tomida et al., 2015), meaning that both RBD motifs redundantly interact with REV7. The model of interaction that they concluded was that REV7 has a surface of interaction that can bind either RBD1 or RBD2. Considering the results of the previous experiments, some REV3-binding blocking REV7 mutants (Y63A and W171A) might still interact with REV3 via the second motif if REV7 had another surface of interaction.
The combination of the mutants of both proteins finally proved this hypothesis. Both REV7 Y63A and REV7 W171A were able to bind RBD1m, but not RBD2m, meaning that RBD2 was responsible for the interactions with the mutants REV7 Y63A and REV7 W171A. This reveals that these surface aminoacids are responsible for RBD1-binding, and its modification blocks RBD1-interaction but has no effect on RBD2. Therefore, there is likely to be another interaction surface in REV7 responsible for RBD2-binding (figures 15 and 16).
Derived from the results of this experiment, this is the model of interaction that we propose:
Figure 15. Model of interaction between REV31775-2200 and REV7 WT. A) RBD1 (P1880 P1885) interacts with the Y63/W171 surface of REV7, whereas RBD2 (P1996 P2001) binds another unknown surface of REV7. B) RBD1m can interact with REV7 WT via RBD2. C) RBD2m can interact with REV7 WT via RBD1. D) RBD DM is not able to interact with REV WT since it does not present any suitable RBD.
Figure 16. Model of interaction between REV31775-2200 and REV7 Y63A. The binding profile is identical in the case of REV7 W171A. A) REV31775-2200 (WT) interacts with REV7 Y63A via RBD2 and the unknown surface of REV7. B) RBD1m can interact with REV7 Y63A also via RBD2. C) RBD2m cannot interact with REV7 Y63A because the interaction via RBD1 is blocked by the point mutation Y63A and RBD2-binding is prevented by the point mutations of the conserved prolines in RBD2. D) RBD DM is not able to interact with REV Y63A since it does not present any suitable RBD.
Surprisingly, REV7 Y63A and REV7 W171A present the same binding profile, but their ability for CSR was completely different, as the Y63A mutant showed a severely affected phenotype whereas the W171A just exhibited a slight reduction. This effect cannot be explained by a REV3-dependent effect. However, it seems that RBD1 specific interactions with REV7 located around Y63 are essential for REV7’s role in CSR, but the factor involved in this interaction could be distinct from REV3.
On the one hand, if the factor were REV3, a possible explanation is that the Y63 and the W171 mediate interactions with RBD1 but with a different biological significance (that the technique is not capable of detecting), being Y63 essential and W171 not required for CSR.
On the other hand, if REV7 were independent of REV3 for CSR, at least its RBD1-binding region is important for the process possibly by interacting with a distinct unknown factor, subject to Y63-interactions, but not W171-binding. This assumption is difficult to understand considering that both residues conform the dock site for RBD1. However, the W171 belongs to the moving seatbelt and the Y63 is part of the α helix. The crystallised structure was obtained with REV7 R124A, that shows the stable closed structure of the protein and, in this conformation, Y63 and W171 are very close to each other (figure 3). However, the open conformation could have a different structure where these two residues are more distant, allowing the interaction of a partner with Y63 but not with W171.
These partners could be 53BP1 and RIF1, which were also examined by the yeast two-hybrid assay. Nonetheless, unfortunately the technique did not work for these two proteins (data not shown).
Regarding the C70 residue, it is a core aminoacid that, as introduced previously, contributes to the seatbelt conformation. Indeed, the point mutant C70R has been reported to block the interaction. The results show that it does not block RBD1-binding but it showed a subtle reduction in RBD1-interaction, while its CSR phenotype was slightly reduced, fact that could be explained by the REV3-dependent hypothesis.
The K129 was not identified before but it is a surface aminoacid on the αC helix of REV7 (figure 17). Its interaction as shown by the Y2H with REV31775-2200 is equivalent to REV7 WT while its CSR phenotype was severely compromised. This result reveals that its effect on CSR cannot be explained considering REV3 interactions, even though this residue may be responsible for an important interaction with another partner.
The R124A mutant of REV7 did not show a clear difference when compared to REV7 WT. Therefore, it cannot explain alone the large increase that it had for CSR.
Figure 17. REV7 representation in complex with RBD1. The lysine 129 is marked (PDB: 3ABE)
Table 3. Summary of the correlation between REV7-REV3 interaction and CSR phenotype. * Data obtained from previous experiments in the laboratory
|Mutation||Interaction (Y2H)||CSR phenotype*||Previous evidence|
|Y63A||Blocks RBD1-binding but does not affect RBD2-interactions||Severely compromised||Alters interaction with REV31847-1898 (Hara et al., 2010)|
|W171A||Blocks RBD1-binding but does not affect RBD2-interactions||Slightly impaired||Alters interaction with REV31847-1898 and REV1826-1251 (Hara et al., 2010)|
|C70R||Slightly reduces RBD1-binding||Slightly impaired||Disrupts interaction with REV31847-1898 (Khalaj et al., 2014)|
|R124A||Slight enhancement or no difference||Increased||Increases affinity for REV31847-1898 and REV1826-1251 (Hara et al., 2010)|
|K129A||Similar interaction to wild-type||Severely compromised||–|
During this investigation, we have found some evidence that supports two different hypotheses.
Further studies are required to determine whether the role of REV7 in class switch recombination is regulated by its DNA polymerase ζ function, in other words, if REV7 is REV3-dependent for CSR.
REV3, as the catalytic subunit of DNA polymerase ζ, has an essential role in proliferation of normal mammalian cells (Lange et al., 2016). Nevertheless, as previously explained, REV7 depleted cells are viable (Pirouz et al., 2013; Watanabe et al., 2014). The generation of a REV3 mutant that cannot bind REV7 could rescue the essential non-REV7 dependent functions while not disrupting other functions. For that purpose, the identified point mutations that block the interaction with REV7 in the REV7-binding domain of REV3 have helped to direct CRISPR-Cas9 mutagenesis studies in CH12 mouse B cells. These cells will then be avaluated for their ability to rescue CSR and the results will be compared to the interaction determined by the yeast two-hybrid.
More studies that assess the interaction of REV7 with other partners will help to answer this question, such us recruitment analysis to DSBs foci.
REV7’s role in the maintenance of genome integrity is so important that it has been described to have prognostic relevance in some type of tumours (Rimkus et al., 2007; Okina et al., 2015; Feng et al., 2016). A deeper understanding of these DNA repair pathways could uncover possible weaknesses of cancer cells that can be exploited to make therapeutic approaches more efficient in the long term.
This study has optimised the yeast two-hybrid assay for the characterisation of the interactions between REV7 and a fragment of REV3 (REV31775-2200).
The model of interaction derived from the results confirms that REV3 interacts with REV7 through two different binding motifs (RBD1 and RBD2). RBD1 is likely to rely on the Y63/W171 surface of REV7, whereas RBD2 is interacting with a distinct binding surface in REV7 yet to be determined.
The correlation between the profiles of interaction and the previously obtained CSR phenotypes reveals interesting conclusions.
On the one hand, there are some results that could be explained if REV7 were dependent of REV3 for CSR, such as the correlation between the REV7 C70R interaction and its CSR phenotype, both slightly impaired. The discrepancy between the Y63A and the W171A mutants of REV7 could be just a consequence of a limitation of the assay, that may not be sensitive enough to differentiate between a structural interaction and a biologically functional interaction. For all these reasons, we would not yet consider rejecting the REV3-dependent hypothesis until further CRISPR-Cas9 mutagenesis studies in mouse B cells are performed and give a convincing evidence.
On the other hand, the REV7 K129A result shows that its CSR phenotype cannot be explained considering REV3 interactions, even though this residue may be responsible for an important interaction with another partner.
Also, the fact that REV7 Y63A and REV7 W171A present a similar binding pattern as shown by the yeast two-hybrid the but a different effect on CSR could suggest that the Y63A phenotype is explained by a REV3-independent effect. This supports the hypothesis of REV7 being independent of REV3 for CSR although its REV3-binding region is important for the process possibly by interacting with a distinct unknown factor, which is subject to tyrosine 63-interactions, but not tryptophan 171-binding. Additional investigations are required to finally unravel the minimum molecular binding partners of REV7 for its role in immunoglobulin class switch recombination.