2Ecole Polytechnique Fédérale de Lausanne, Swiss Institute for Experimental Cancer Research, Lausanne, Switzerland
3Indiana University School of Medicine, Department of Medicine, Indiana Center for Biological Microscopy, Indianapolis, IN
Clinical approaches for tumor treatment often rely on combination therapy where a DNA damaging agent is used in combination with a DNA repair protein inhibitor. For this reason, great efforts have been made during the last decade to identify inhibitors of DNA repair proteins or, alternatively, small molecules that specifically alter protein stability or trafficking. Unfortunately, when studying these drug candidates, classical biochemical approaches are prone to artifacts. The apurinic/apyrimidinic endonuclease (APE1) protein is an essential component of the base excision repair (BER) pathway that is responsible for repairing DNA damage caused by oxidative and alkylating agents. In this work, we combined conditional gene expression knockdown of APE1 protein by RNA interference (RNAi) technology with re-expression of an ectopic recombinant form of APE1 fused with the photoconvertible fluorescent protein (PCFP) Dendra2. Dendra2 did not alter the subcellular localization or endonuclease activity of APE1. We calculated APE1 half-life and compared these results with the classical biochemical approach, which is based on cycloheximide (CHX) treatment. In conclusion, we combined RNAi and in vivo confocal microscopy to study a DNA repair protein demonstrating the feasibility and the advantage of this approach for the study of the cellular dynamic of a DNA repair protein.
Investigation of protein dynamics is essential for understanding cellular regulation. Visualization of protein subcellular distribution and co-localization with other proteins is generally achieved using labled antibodies. Antibodybased techniques require fixation and permeabilization of the cell in order to guarantee reagents access to all subcellular compartments. Although fluorescence microscopy approaches are routinely used, the question of how well protein localization is preserved in cells after fixation and immunostaining procedures is still a matter of debate. In a recent paper, Schnell et al. highlighted how fixation and permeabilization procedures can affect epitope accessibility and therefore may create artifacts (1).
Here we combined a specific conditional gene expression knockdown of APE1 protein by RNA interference (RNAi) technology with the re-expression of an ectopic recombinant form of APE1 in fusion with the photoconvertible fluorophore (PCFP) Dendra2. Through this approach, we were able to (i) study the protein dynamics of APE1 in a physiological rather than over expression system, (ii) discriminate between protein newly synthesized or already present within the cell, and (iii) calculate APE1 half-life.
An alternative approach is to express a recombinant version of the protein of interest fused to a fluorescent protein (FP), a technique that revolutionized cell biology by allowing visualization and tracking of proteins in real time at high spatio-temporal resolution (2). However, this approach can also lead to erroneous results because fusing the FP with the protein of interest may alter its folding, localization or interaction with other proteins or nucleic acids, thereby affecting its function. Furthermore, the ectopic fusion protein is expressed in addition to the endogenous native protein, resulting in overexpression of the protein of interest.
The dynamic and kinetic behavior of a fluorescent fusion protein is typically determined using techniques such as FRAP or FLIP (5). In both cases, a small patch of the viewable area of the cell receives high intensity illumination, which causes the fluorescence lifetime of the FP to quickly elapse. Destroying the FP and watching the repopulation into the bleached area can reveal information about organelle continuity, protein trafficking, diffusion coefficient, and mobile fraction of a protein (6). However, GFP-tagged proteins are continuously synthesized, folded, and degraded within the cell (2), thus rendering the measurement of protein turnover, the analysis of the temporal expression pattern, and the assesment of protein function difficult, if not impossible. Therefore, to extract information on protein degradation or movement within a cell, one would need to block the synthesis of new GFP containing molecules. To overcome this problem, Terskikh et al. designed a FP with an emission spectrum that changes with time. This type of FP, called a timer proteins or a fluorescent timer (FTs), allows estimation of the age of fusion proteins by measuring the ratio of green and red fluorescence (7). FTs permit spatial and temporal observation of protein behavior in living cells.
Protein degradation is an important aspect of cellular physiology, helping to maintain the proper concentration of a protein in a living cell. Protein half-life is commonly determined by radioactive pulse-chase analysis (8). This method is laborious and requires radiolabeling procedures. An alternative biochemical approach to estimate protein half-life involves time course treatment with cycloheximide (CHX) to inhibit protein synthesis, coupled with immunoblotting. CHX acts by binding the E site of the large ribosomal subunit, thus blocking elongation (9). The main advantages of this technique are the reversibility of CHX treatment and the low cost. However, CHX treatment indiscriminately inhibits protein biosynthesis of all primary transcripts present within the cell and has been reported to induce apoptosis in human cell lines (10). The absence of selectivity toward a specific protein and cellular toxicity are the main limitations of this method for the determination of protein half-life. An alternative fluorescence microscopy-based approach for studying protein lifetimes uses photoactivatable fluorescent proteins (PAFPs) that display a steady state minimal fluorescence that is increased by radiation with a specific wavelength, thereby allowing monitoring of turnover and degradation of the fusion protein (11). In addition, neither pulse-chase nor CHX treatment allows real-time measurements at the single cell level.
The state of the art for investigating protein dynamics and lifetime relies on photoconvertible fluorescent proteins (PCFPs). This class of fluorescent molecules undergoes a switch (photoconversion) in its spectral emission properties in response to UV-violet (350–420 nm) or blue light (e.g., 488 nm) irradiation. Kaede, mKikGR, mEos2, and Dendra2 are the most frequently used PCFPs; they mature as green-fluorescent molecules and are irreversibly photoconverted to a stable brightred fluorescent state that can be tracked for several hours without photobleaching (12, 13).
Dendra2 is a human codon-optimized variant of the octocoral Dendronephthya sp. fluorescent protein, Dendra, which has been engineered for faster maturation and brighter fluorescence both before and after photoconversion. Dendra2 matures efficiently at both 20°C and 37°C, and undergoes irreversible photoconversion from green to red in response to intense irradiation at both 405 nm and 488 nm. While non-photoconverted (green) Dendra2 possesses excitation/emission maxima at 409/507 nm, photoconverted (red) Dendra2 possesses excitation/emission maxima at 553/573 nm. Neither toxicity nor photobleaching was reported using Dendra2 (12). Here we propose the use of Dendra2 activation within a whole cell to monitor APE1/Ref-1 protein degradation.
Apurinic apyrimidinic endonuclease/redox effector factor 1 (APE1/Ref-1; henceforth referred to as APE1) is a master regulator of cellular response to oxidative stress and plays a central role in the maintenance of the genome stability (14). Indeed, APE1 is the major human apurinic/apyrimidinic (AP) endonuclease, exerting a pivotal role in the base excision repair (BER) pathway. APE1 is able to recognize an AP site and cleave the phosphodiester bonds 5′ to the AP site, leaving a 3′-hydroxyl group and a 5′-deoxyribose phosphate (dRP) termini flanking the nucleotide gap (15). The C terminus exerts the enzymatic activity on the abasic sites of DNA while the N terminus, containing the nuclear localization signal sequence, is principally devoted to the redox transcriptional co-activation activity (16). Through its redox domain, APE1 can act as a redox-regulatory protein to maintain several transcription factors (Egr-1, NF-κB, p53, HIF-1α, AP-1, and Pax proteins) in an active reduced state (17). Interestingly, this protein is able to repress its own transcription (18).
APE1 is a multifunctional protein and its physiological relevance is underscored by the fact that nullizygous mice for APE1 show early stage embryonic lethality (14). Moreover, down-regulation of APE1 expression levels in human cells through RNAi leads to AP site accumulation, reduced cell proliferation, and triggering of apoptosis (19). Increased expression of APE1 is associated with several tumorigenic processes (20). As a fundamental component of the BER pathway, APE1 is an intriguing candidate for novel therapeutic strategies aimed at inhibiting DNA repair as an effective adjuvant treatment in cancer therapy (21).
Under normal conditions, APE1 mainly localizes within the nuclear compartment and accumulates in the nucleoli, but it was also shown to localize in mitochondria where it is involved in mtBER activity (22). Importantly, a higher intracellular expression and a robust cytoplasmic localization of APE1 have been described in lung, ovarian, thyroid, and breast cancer to correlate with higher tumor aggressiveness (20). The possible causal role played by this particular distribution in tumor progression is, however, unknown. It is therefore important to investigate the mechanisms regulating APE1 localization and subcellular distribution and to acquire more accurate information about its protein synthesis rate and lifetime.
Given the many fundamental functions played by APE1 overexpression, as well as silencing, strategies to study these issues may give rise to artifacts. For this reason, we generated and characterized a cellular model introducing APE1 expressed as a fusion protein with Dendra2 in APE1-silenced cell clones, thus maintaining physiological expression levels of APE1 protein. We confirmed that the presence of Dendra2 expressed in fusion with APE1 did not alter its subcellular localization nor its enzymatic activity on abasic DNA. Then we used this cellular model to calculate APE1 half-life. Dendra2 has been used previously to follow the degradation of target proteins (23), but our approach is the first study combining RNAi knock down (KD) with the re-expression of a PCFP fusion protein. In this study, we focused our attention on a DNA repair protein, but our approach is potentially applicable to a variety of protein of interests. Materials and methods Plasmid preparation
For generating the APE1 KI clone, APE1 cDNA was sub-cloned from a pFLAG-CMV-5.1/APE1vector (Sigma Aldrich, Milan, Italy) into the pDendra2-N vector (Clontech, Mountain View, CA) in frame with the Dendra2 coding sequence to generate an APE1-Dendra2 fusion protein. To avoid degradation of ectopic APE1 mRNA by the specific siRNA sequence, two nucleotides (+180 and +183) of the APE1 cDNA coding sequence were mutated with site-directed mutagenesis kits (Stratagene, Milan, Italy), leaving the APE1 amino acid sequence unaffected (20). To generate cell clones stably expressing APE1-Dendra2 or Dendra2 proteins, vectors were linearized before transfection by digestion with HpaI restriction enzyme (Fermentas, Milan, Italy), purified through ethanol precipitation, and finally quantified with Nanodrop (Thermo Scientific, Milan, Italy). Cell clone generation and culture
The HeLa APE1 inducible siRNA clone was previously generated and characterized in our lab (24). Briefly, cells were grown in DMEM high glucose (PAA, Milan, Italy) supplemented with 2 mM stable L-Glutamine, 10% fetal bovine serum (both from Euroclone, Milan, Italy), 100 U/mL penicillin, 10 µg/mL streptomycin sulfate, 3 µg/mL Blasticidine and 100 µg/mL Zeocine (all from Invitrogen, Monza, Italy). To induce siRNA expression, 1µg/mL doxycycline (Sigma Aldrich) was added to the medium and cells were grown for the reported amount of time. Dendra2 and APE1-Dendra2 reconstituted clones were generated by transfecting the above-mentioned siRNA clone with 2 µg of linear pDendra2-N or pDendra2-N-APE1 expressing vectors, respectively. A day before transfection, 4 × 105 cells were seeded into a 6 well plate and transfection was carried out using Lipofectamine 2000 (Invitrogen) following manufacturer instructions. After a 48 h transfection period, cells were grown in the presence of 800 µg/mL G-418 (Invitrogen) for 14 days and selected for the acquired resistance. Single clones were isolated by using cell cloning cylinders (Sigma Aldrich), and expanded to 2 × 106 cells in the presence of selective antibiotics. For calculation of APE1 half-life, CHX (Sigma Aldrich) was resuspended in DMSO and added to the cell medium at a final concentration of 100 µg/mL. Confocal microscopy analysis
For confocal microscopy analysis of live cells, 4 × 105 cells were seeded onto glass bottom WillCo Wells Petri dishes (Ted Pella Inc., Redding, CA) and grown in DMEM without phenol red to avoid interference from medium auto-fluorescence. Images reported in Figure 1B and C, and Supplementary Figure sS1 and S3, were obtained using a Leica TCS SP confocal microscope equipped with a temperature control unit and 405 nm diode, 458/488 nm argon and 563/683 nm HeNe lasers, RSP 500 filter, and 63×/1.32 oil CS. Acquisition parameters were 488 nm excitation line at 15% laser power, emission filter BP (band pass) 500–535 nm, PMT at 760V and 543 nm excitation line at 70% laser power, emission filter BP 555–620 nm and PMT 820V were used for channel 1 (Ch1) and channel 2 (Ch2) accordingly. Images reported in Figure 3B, Supplementary Figure S2, and Supplementary Movie S1 were obtained using a Zeiss LSM 710 confocal microscope equipped with a full living-cell system and 405 nm diode, 458/488/514 nm argon and 633 nm HeNe lasers, FSet 10 filter, and 63×/1.40 oil DIC M27 objective. Acquisition parameters were BP 488 nm SP mirror, BP 515–565 nm with laser power 5%, and PMT 760V in Ch1; BP 555 nm SP mirror, BP 575–640 nm with laser power 70%, and PMT 880V in Ch2. All images were collected at 1024 × 1024 pixel resolution. A 405 nm diode laser was employed at 100% power for a single scan with a pixel dwell time of 1.27 µs for total photoconversion of Dendra2. To obtain partial photo-conversion of the fluorescent protein, laser power was reduced to 30%.