Journal list menu
Probing the interaction of a quercetin bioconjugate with Bcl-2 in living human cancer cells with in-cell NMR spectroscopy
Abstract
In-cell NMR spectroscopy has emerged as a powerful technique for monitoring biomolecular interactions at an atomic level inside intact cells. However, current methodologies are inadequate at charting intracellular interactions of nonlabeled proteins and require their prior isotopic labeling. Herein, we describe for the first time the monitoring of the quercetin-alanine bioconjugate interaction with the nonlabeled antiapoptotic protein Bcl-2 inside living human cancer cells. STD and Tr-NOESY in-cell NMR methodologies were successfully applied in the investigation of the binding, which was further validated in vitro. In-cell NMR proved a very promising strategy for the real-time probing of the interaction profile of potential drugs with their therapeutic targets in native cellular environments and could, thus, open a new avenue in drug discovery.
Abbreviations
STD, saturation transfer difference
Tr-NOESY, transfer NOESY
In-cell NMR spectroscopy has been rapidly developed to a powerful technique for the study of protein–protein and ligand–protein interactions directly in the intracellular environment of living cells [1, 2]. Other experimental techniques, such as X-ray crystallography, hydrogen-deuterium exchange mass spectrometry, and in vitro NMR spectroscopy can provide information about the changes in structure and/or dynamics between the free and bound forms in vitro. However, all these physical methods require optimized experimental conditions that may not reflect the protein state in vivo [3, 4]. Instead, in-cell NMR is a noninvasive technique, which reveals structural and conformational information at the atomic level, under physiological conditions. Native complex environments affect several cellular processes such as protein interactions, conformational changes or changes induced upon binding, establishing in-cell NMR as a more reliable technique to detect and effectively delineate biological events inside intact cells [5, 6].
This methodology has been successfully applied in the analysis of 15N isotope-labeled proteins, which are overexpressed in Escherichia coli bacteria, where 1H–15N HSQC in-cell NMR experiment is performed directly in intact cells to study protein conformation, dynamics, protein–protein, and ligand–protein interactions [7-9]. However, prokaryotic cells exhibit a limited range of biological activities and most of the cellular functions that determine significant aspects of modern biological research are absent [9]. Therefore, much effort has been devoted to extending in-cell NMR studies in eukaryotic cells and better describes biological processes such as protein maturation, post-translational modification, and redox reactions, which alter the structural and functional properties [10-12]. Isotope-labeled proteins are initially produced in E. coli and then introduced into eukaryotic cells either by microinjection to Xenopus laevis oocytes [13-15] or by endocytotic transport mediated by cell-penetrating peptides [11, 16] or by diffusion through pore forming toxins and electroporation [17, 18]. However, delivering isotope-labeled proteins into eukaryotic cells has been a major drawback of this technique hampering the direct observation of proteins, which are overexpressed and modified in their natural environment. Only recently, protein isotopic labeling was achieved for the first time in human embryonic HEK293T cells with in-cell NMR studies revealing information on the intracellular conformation of the protein [2].
Real-time monitoring of small molecules that can disrupt protein–protein interactions in living cells has been one of the most valuable contributions of in-cell NMR spectroscopy in the drug discovery field, allowing for the screening of lead compounds and the characterization of their binding to receptors in their native environments [10, 19, 20]. The behavior of a drug in vivo may depend upon various factors, as for example its ability to penetrate the cell membrane, the fast metabolism, its binding to other cellular components with higher affinity or putative conformational differences of the target protein between in vitro and in vivo conditions. In-cell NMR spectroscopy is an ideal tool offering new exciting possibilities to gain insight into the interaction profile of a drug with its pharmaceutical target inside their cellular environments [21, 22].
So far, the in-cell NMR strategies that have been developed, allow for the screening of interactions of small molecules with unlabeled cell surface receptors, such as integrins, based on the fact that the ligand is selective for the target protein [23, 24]. Ligand-observing in-cell NMR techniques, including transfer NOESY (Tr-NOESY) and saturation transfer difference (STD), have been used to rapidly probe binding of ligands to membrane receptors on the surface of the cells [23, 25]. STD NMR is applied in the characterization of the binding epitope of the ligand revealing the closest bound moieties to the receptor, whereas Tr-NOESY NMR defines the conformational changes of the bound ligand [24]. STD NMR experiment is based on the intermolecular transfer of magnetization from the irradiated receptor to the bound ligand resulting in the identification of protons which are in proximity to the protein (≤ 5 Å) [26]. Tr-NOESY relies on the fact that a small ligand bound to a large-molecular weight receptor behaves as a part of the large molecule and adopts the corresponding NOE behavior, demonstrating thus, Tr-NOEs that reflect the bound conformation [27]. Interestingly, monitoring the interaction between a ligand and a nonlabeled intracellular target protein, remains a rather challenging issue for in-cell NMR spectroscopy.
Natural products have always provided a rich source of lead compounds in the field of drug discovery in cancer [28]. Previously, we determined the basis of the pro-apoptotic activity of quercetin in human T-leukemia cells to be due to direct interaction with Bcl-2 antiapoptotic proteins [29]. Specifically, quercetin exhibited BH3-mimetic properties and could directly bind to the BH3 domain of Bcl-2 and Bcl-xL proteins with a Kd of 1.1 μm, suggesting it as a potential mechanism for driving cancer cells to apoptosis [29]. However, the main drawbacks of quercetin are its limited water solubility and cell selectivity that hamber its administration as a chemo-preventer [30, 31]. Based on this, we synthesized a bioconjugate of quercetin with an amino acid, which has shown significantly improved water solubility and cell permeability compared to the parent molecule [32]. We found that the quercetin-alanine bioconjugate exhibited an enhanced intracellular delivery and cytotoxic activity in different cancer cell lines [33]. The binding site of the quercetin-alanine bioconjugate to the Bcl-xL interface was localized also to the hydrophobic BH3-binding groove as demonstrated by 1H–15N HSQC NMR spectroscopy and docking calculations. Quercetin-alanine demonstrated only a slight decrease in the binding affinity for Bcl-2 family proteins with a Kd value of 8.7 μm [33]. Although we described the binding mode of the bioconjugate in vitro with Bcl-xL, its interaction profile in the native protein environment inside living cancer cells remained elusive. In-cell NMR spectroscopy was employed to explore the direct binding of quercetin-alanine to the unlabeled Bcl-2 protein in living human cells. Bcl-2 is an intracellular protein localized to the outer membrane of mitochondria and the study of the interaction with a ligand was, therefore, rather challenging.
Materials and methods
Synthesis of the quercetin-alanine bioconjugate
The bioconjugate of quercetin was synthesized by conjugating the amino acid alanine to the B-ring of quercetin, according to the synthetic procedure first reported by Kim et al. and then slightly modified by our group (Scheme S1), which resulted in considerable higher water solubility (26-fold increase with respect to quercetin) [32, 33].
Cell culture and cell preparation for NMR analysis
Human T-cell leukemic Jurkat Puro and Jurkat Bcl-2 cells were stably generated by retroviral infection of Jurkat cells, using either a control puromycin retrovector or the same retrovector carrying a human Bcl-2 cDNA, respectively [29]. Cells were cultured as previously reported [34]. The relative quantitation of Bcl-2 protein levels in the two cancer cell lines is presented in Fig. S1. image lab Software 6 from Biorad (Bio-Rad Laboratories, Inc., Hercules, CA, USA) was used to perform this analysis.
For each NMR experiment 5 × 106 cells were centrifuged at 250 g for 6 min and the pellet was re-suspended in 500 μL of deuterated 0.01 m PBS (pH 7.4) with the addition of 4% of DMSO-d6 to improve the solubility of the ligand and then transferred into a 5 mm NMR tube. For the STD experiment performed to compare Jurkat Puro and Jurkat Bcl-2 cells 2.5 × 106cells were used. Cells number was adequate to achieve both suspension homogeneity and efficient cell density in the acquisition window.
Western blot analysis of Bcl-2
Jurkat Bcl-2 and Jurkat Puro cells were lysed in lysis buffer containing a cocktail of complete protease inhibitors (Roche Diagnostics, Mannheim, Germany) and the protein concentration was determined by the BCA method. Five hundred microgram of the cell lysates were analyzed by SDS/PAGE electrophoresis followed by immunoblotting. The nitrocellulose membrane was incubated with the primary antibody overnight at 4 °C with a mouse monoclonal Bcl-2 primary antibody (Santa Cruz Biotechnology; Santa Cruz, CA, USA, sc-509, dilution 1 : 1000). The membrane was then incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature and the immunoreactive bands were detected using ECL detection reagent (GE Healthcare).
NMR-binding experiments of quercetin-alanine to Bcl-2
1H–15N HSQC NMR in vitro experiment
Truncated Bcl-2 was expressed and purified for NMR experiments, as previously reported [35]. 2D 1H–15N HSQC NMR spectra were collected on a Bruker Avance II 800 spectrometer at the Korea Basic Science Institute. The 2D spectrum of 0.1 mm Bcl-2 was obtained at 25 °C without and with the addition of 0.1, 0.2, and 0.4 mm quercetin-alanine. The NMR samples comprised of 90% H2O/10% D2O and were prepared in 20 mm Tris-HCl (pH 7.8) and 5 mm DTT for Bcl-2. The NMR data were processed and analyzed using nmrpipe/nmrdraw [36] and sparky 3 (T.D. Goddard and D.G. Kneller, University of California, San Francisco, CA, USA) software. For the assignment of 2D 1H–15N HSQC of Bcl-2 we used the BMRB entry 19559 [37].
STD in-cell NMR experiment
To characterize 3′ quercetin-alanine-binding epitope to Bcl-2, STD NMR experiments were performed at 37 °C on a Bruker AV-500 spectrometer equipped with a TXI cryoprobe (Brüker BioSpin, Rheinstetten, Germany) at the Department of Chemistry, University of Ioannina, Greece. STD experiments were carried out in the presence of 5 × 106 cells in 0.01 m PBS (pH 7.4) in D2O with 4% DMSO-d6 adding an excess of 3′ quercetin-alanine (1.4 mm) to allow for specific ligand-receptor interactions. The percentage of the bioconjugate entering the cells was estimated with NMR spectroscopy approximately at 50% (0.7 mm). The selective inhibitor HA14-1 was firstly dissolved in DMSO-d6 and then diluted in 0.01 m PBS as suggested for maximum solubility and added in the mixture to perform competition experiments [38, 39]. The concentration of HA14-1 was approximately at 0.2 mm due to low solubility. The receptor concentration inside the NMR tube was in the nm range. The on-resonance irradiation of the protein was performed at a chemical shift of 0.98 ppm. Off-resonance irradiation was applied at 40 ppm, outside the spectra region of the 1H NMR resonances of the protein. Each Gaussian pulse required 50 ms separated by a delay of 1 ms. Saturation time was set at 2 s and total experimental time was ~1 h (128 scans) to ensure cell viability throughout the experiment. STD spectra of the ligands in the absence of cells did not show any signals. Water suppression was achieved by use of an excitation sculpting pulse sequence.
The same STD experiment was performed for the recombinant Bcl-2 protein (40 μm) in 50 mm sodium phosphate buffer 50 mm NaCl pH 6.5 in D2O with 2% DMSO-d6 adding an excess of 3′ quercetin-alanine (1 mm) to allow for specific ligand–receptor interactions. The selective inhibitor HA14-1 was firstly dissolved in DMSO-d6 and added in the mixture to perform competition experiments in two concentrations (0.1 and 0.2 mm).
Tr-NOESY in-cell NMR experiment
To determine the conformation of 3′ quercetin-alanine upon binding to Bcl-2 in living cells, Tr-NOESY NMR experiments were performed on a Bruker Avance 600 Ultra Shield TM Plus 600 MHz spectrometer equipped with triple resonance cryoprobe (TCI) and pulsed field gradients at S. Raffaele Scientific Institute, Milano, Italy. Cells were prepared using the same conditions as above. A short mixing time (100 ms) was set in order to obtain maximum Tr-NOE intensity. All experiments were acquired at 37 °C and free induction decays were acquired (32 scans) into 2k data block for 110 incremental values of the evolution time (States-TPPI). Acquisition parameters for Tr-NOESY experiments performed on cells were set to minimize total experiment time (2 h) to ensure cell viability throughout the experiment. Solvent suppression was achieved using an excitation sculpting scheme.
Tr-NOESY experiments were also performed for the recombinant Bcl-2 protein (40 μm) in 50 mm sodium phosphate buffer 50 mm NaCl pH 6.5 in D2O with 2% DMSO-d6 adding an excess of 3′ quercetin-alanine (1 mm) to allow for specific ligand–receptor interactions. A second Tr-NOESY spectrum was recorded upon addition of HA14-1. The selective inhibitor was firstly dissolved in DMSO-d6 and then added in the mixture to perform the competition experiment (0.2 mm).
Trypan blue cell viability assay
Cell viability throughout the total NMR experimental time was determined with the trypan blue exclusion assay. Briefly, an aliquot of the cell suspension was diluted 1 : 1 (v/v) with 0.4% trypan blue and cells were counted with a hemocytometer. The number of blue (B) and white (W) cells was counted and cell viability was calculated as W/(B+W).
Results and Discussion
In vitro study of the quercetin-alanine bioconjugate binding to Bcl-2 using 2D 1H–15N HSQC NMR
The direct interaction of the quercetin-alanine bioconjugate with the 15N labeled Bcl-2 protein in vitro was investigated using 2D 1H–15N HSQC NMR spectroscopy. Figure 1A illustrates the superposition of 1H–15N HSQC spectra of Bcl-2 in the absence or presence of the quercetin-alanine bioconjugate. We found significant chemical shift changes of several 1H–15N cross-peaks in Bcl-2 upon the addition of quercetin-alanine, indicating a direct interaction (Fig. 1 and Fig. S2). NMR chemical shift perturbations of Bcl-2 upon binding to quercetin-alanine bioconjugate were plotted against Bcl-2 amino acid sequence (Fig. 1B) and binding sites were mapped onto the three-dimensional structures (Fig. 1C). Residues exhibiting chemical shift perturbation values over 0.035 and 0.07 ppm were mapped on the protein structure in yellow and red colors, respectively. The NMR chemical shift perturbations induced by binding to quercetin-alanine were localized to the hydrophobic BH3-binding groove surrounded by the BH1, BH2, and BH3 domains of Bcl-2 (Fig. 1C and Fig. S3). These results are in good agreement with those previously reported for the binding of the parent compound (quercetin) to Bcl-2 (Fig. S4), thus proving a direct interaction between the quercetin-alanine bioconjugate and a specific interface of Bcl-2 protein [29].
Mapping the binding of quercetin-alanine to Bcl-2 inside living human T-leukemic cells with STD in-cell NMR
Having defined the binding interface between quercetin-alanine and Bcl-2 in vitro, we performed in-cell STD NMR experiments to reveal the binding sites of the bioconjugate in living human cancer cells overexpressing Bcl-2. For this, human T-leukemic Jurkat cells were stably transduced with a retroviral vector conferring puromycin resistance carrying a human Bcl-2 cDNA [29]. Τhe cellular uptake of quercetin by Jurkat cells, which is taking place by passive diffusion, is extremely rapid and effective and a remarkable accumulation of large amounts of the flavonoid in a very short time has been observed intracellularly [40]. Even larger amounts of quercetin were found to further accumulate in their mitochondria. In addition, the quercetin-alanine bioconjugate presented 2-fold increased cell permeability properties in a model cell line compared to quercetin [32, 33].
We first performed 1D 1H NMR control experiments with the free ligand (Fig. 2A), cells in the presence of quercetin-alanine (Fig. 2B) and the cells alone (Fig. 2C). The spectrum of intact cells demonstrated background signals of endogenous metabolites such as glucose in the region between 3–4 ppm and at 5.25 ppm attributed to the Ha proton, pyruvate at 2.2 and 2.4 ppm, alanine which is overlapping at 1.4 ppm etc. [41]. Figure 2A represents the 1H NMR spectrum of the quercetin-alanine bioconjugate, which was a mixture of the main 3′ quercetin-alanine analogue and the minor 4′ quercetin-alanine product (Scheme S2A), annotated with purple and green colors, respectively. Minor peaks in the spectrum were attributed to quercetin, which was also present in the mixture (8.3%) as a hydrolysis product of quercetin-alanine bioconjugate to quercetin, N-carboxy-alanine and alanine (Scheme S1). In-cell NMR studies were performed for the main analogue 3′ quercetin alanine. The multiplet peaks annotated with the asterisk at 3.28 and 3.79 ppm represent the Ha protons of N-carboxy-alanine and alanine (Scheme S1), which resulted from the hydrolysis of quercetin-alanine. Furthermore, the doublet peak annotated with an asterisk at 1.4 ppm is attributed to the hydrolyzed alanine methyl protons.
To determine the binding interface of 3′ quercetin-alanine we employed STD in-cell NMR methodology. In the STD spectrum (Fig. 3A) we observed that all the protons of 3′ quercetin-alanine were involved in interactions with intracellular receptors. Binding of 3′ quercetin-alanine to receptors caused line-width broadening due to the large differences in molecular weight and correlation time (τc) between the receptors and the small molecule, thus, providing further evidence of binding. On the other hand, the STD intensities of the methyl protons of the hydrolysis products alanine and N-carboxy alanine were significantly reduced, although they demonstrated higher signal intensities in the 1H NMR spectrum in comparison to the respective protons of 3′ quercetin-alanine. The lower intensities of their STD signals indicated that alanine and N-carboxy alanine participate in very weak interactions with biomolecules inside the cells. This suggestion was further supported by the absence of the STD signals of the Ha protons at 3.3 and 3.7 ppm. Finally, the decreased STD signal of proton H–8 of 3′ quercetin-alanine is attributed to the hydrogen-deuterium (H → D) exchange effect that has been previously reported [42, 43].
Even though with this methodology we were able to identify the protons of 3′ quercetin-alanine participating in interactions with proteins within cells, we had still to prove the direct binding to Bcl-2. To monitor this direct interaction of the ligand with the specific target protein we utilized proper positive and negative controls. As a positive control, we used the established inhibitor HA14–1 (Scheme S2B), which selectively binds to the BH3-binding pocket of Bcl-2 (IC50 = 9 μm) [38, 39]. As a negative control, we used taxifolin, which is a natural polyphenol that differs by quercetin in C2–C3 double bond (Scheme S2C). Interestingly, taxifolin has previously demonstrated an apoptotic activity independent of Bcl-2 mRNA expression and has been shown to be devoid of interaction with Bcl-2 and Bcl-xL [29]. Therefore, HA14–1 selective binding to Bcl-2 should not interfere with taxifolin, which does not bind to the target protein. Furthermore, the cell line Jurkat Puro, which does not overexpress Bcl-2 protein was employed in order to investigate the difference in the bioconjugate interactions intracellularly between the two cancer cell lines. Finally, STD NMR experiments were performed with the recombinant target protein and the quercetin-alanine bioconjugate to detect the interaction mode in vitro.
Firstly, we recorded an in-cell STD spectrum of 3′ quercetin-alanine (1.4 mm) adding the selective inhibitor HA14-1 (Fig. 3B). The concentration of HA14-1 was approximately 10-fold less (0.2 mm) due to very low solubility. Therefore, we were able to observe the STD signal derived only from the methyl protons of the HA14-1 inhibitor. Interestingly, a significant decrease (~ 60%) of the signals in the STD spectrum of 3′ quercetin-alanine in the presence of the selective inhibitor was observed. Such effect could be explained by the fact that the selective inhibitor HA14-1, competes and replaces bound 3′ quercetin-alanine to the BH3 domain of Bcl-2, owing to its higher binding affinity and resulting in less saturation transfer from the receptor to 3′ quercetin-alanine. This result is an indirect, yet, strong indication of the direct binding of 3′ quercetin-alanine to Bcl-2. The remaining STD signal may be attributed to the interaction of the ligand with other intracellular receptors as it is known that quercetin also interacts with various other intracellular receptors [31]. Another interesting observation was that the STD signals of the methyl protons at 1.4 ppm of alanine and N-carboxy alanine did not present any decrease of their intensity upon addition of the selective inhibitor, thus, indicating that they do not interact with Bcl-2. Thus, they can be also exploited as negative controls for the interaction with Bcl-2.
Further confirmation of this interaction was obtained when the STD experiment was performed for the quercetin-alanine bioconjugate in Jurkat Puro cells, which do not overexpress Bcl-2 protein since they were stably transduced with an empty retroviral vector conferring puromycin resistance. A decrease of approximately 35% of the bioconjugate STD signals was observed in the Jurkat Puro cells compared to the respective experiment in Jurkat Bcl-2 cells, which overexpress the protein (Fig. S5). This result is in excellent agreement with the difference in the expression levels of Bcl-2 between the two cancer cell lines (Fig. S1). Another positive control experiment was performed to verify the interaction of the ligand with the target protein in vitro. An STD NMR experiment was acquired for the 3′ quercetin-alanine bioconjugate with the recombinant Bcl-2 protein as illustrated in Fig. S6B. The ligand demonstrated a direct interaction with the protein in vitro with all the protons participating in the interaction. Then, the selective inhibitor of Bcl-2 protein, HA14-1, was added in the mixture in 0.1 and 0.2 mm and a decrease in approximately 40% of the quercetin-alanine STD signals was observed, whereas the STD signals of HA14-1 started to appear and increase in intensity (Fig. S6C,D).
To confirm this result, further in-cell STD experiments were performed using taxifolin as a negative control (Fig. 4A). In the STD spectrum (Fig. 4B) we observed that all protons of taxifolin were implicated in intracellular interactions, except H–3 for which we cannot reach any safe conclusion due to overlapping with the H2O peak. We then titrated the Bcl-2 selective inhibitor (0.2 mm) in the sample and as expected no reduction on the intensity of taxifolin STD signals was recorded (Fig. 4C). The addition of the Bcl-2 selective inhibitor, HA14-1 did not perturb the monitored intracellular interaction profile of taxifolin, indicating that taxifolin interacts with different partners than Bcl-2. This result is particularly important since it further corroborates to the hypothesis that intracellularly, 3′ quercetin-alanine is a direct partner of Bcl-2 protein. Finally, in the STD spectra as well as in the 1H NMR spectra a faster decrease in the intensity of proton H–6 of taxifolin in comparison to the H–8 of 3′ quercetin-alanine was observed, which is attributed to the higher rate of hydrogen-deuterium (H → D) exchange reactions for this proton [42, 43].
To ensure cell viability a trypan blue assay was performed and indicated that above 90% of the control cells and above 80% of the treated cells remained viable throughout the total experimental time (1 h 15 min) (Fig. S7) [23].
Mapping the intracellular bound conformation of quercetin-alanine with Tr-NOESY in-cell NMR
We then set out to investigate the conformational changes of 3′ quercetin-alanine induced upon binding to Bcl-2 in the intracellular environment and, thus, performed a series of NOESY NMR experiments for the free 3′ quercetin-alanine, for the intact cells and a Tr-NOESY experiment for the 3′ quercetin-alanine in the presence of cancer cells. The NOESY spectrum of the free 3′ quercetin-alanine showed a weak positive NOE (cross-peak with opposite sign with respect to the diagonal) between the aromatic protons H–5′ and H–6′ (Fig. 5A) and a second positive NOE between the methyl and Ha protons of the conjugate alanine, at a mixing time of 100 ms, as expected for small molecules of this size that have a faster tumbling rate compared to the bound forms.
Despite the high background signals of the cells, the Tr-NOE cross-peaks, derived from correlations of the 3′ quercetin-alanine protons in the cells (Fig. 5B), can be unequivocally identified. Upon binding to intracellular proteins, 3′ quercetin-alanine adopted the corresponding NOE behavior of the receptor and acquired a slow tumbling rate and a fast NOE build-up, with the maximum NOE intensity (cross-peaks with the same sign as the diagonal) at a short mixing time (100 ms) [23, 44, 45]. Therefore, strong negative Tr-NOEs (same sign as the diagonal) were observed for correlation peaks between the aromatic protons Η5′–Η6′, H2′–H5′ and H2′–H8 and between the methyl and Ha protons of the conjugate alanine inside the cells (Fig. 5B). Particularly noteworthy is the fact that new Tr-NOE cross-peaks were detected between protons H2′–H5′ and H2′–H8, due to protein binding, which were not present in the spectrum of the free 3′ quercetin-alanine. This suggests that the flexibility of 3′ quercetin-alanine was reduced, and the conformation was changed in the bound state [46]. At this point it should be noted that several crystal structures of quercetin in complexation with protein receptors have shown that ring B may adopt such orientation that protons H–8 and H–2′ can be located adjacent due to the rotation of ring B around the C2–C1′ bond [49, 47, 48]. The absence of significant cross-peaks in the Τr-NOESΥ of the major analytes in the intact cells demonstrate that the strong in-phase cross-peaks of Fig. 5B are not due to change in viscosity. It can be argued that the Tr-NOESY experiments reveal the binding epitope of 3′ quercetin-alanine as it was mapped by the STD NMR results. Both in-cell NMR experiments suggested that the edges of 3′ quercetin-alanine are engaged in strong binding with the Bcl-2 receptor.
The in vitro Tr-NOESY spectrum of the bound 3′ quercetin-alanine (1 mm) to Bcl-2 protein (40 μm) was also performed with and without the addition of HA14-1 (0.2 mm) (Fig. S8). Interestingly, the intensity of the 3′ quercetin-alanine cross-peaks was significantly reduced upon HA14-1 addition, which is a strong indication of their competition for binding to BH3 domain. Although, the ratio of 3′ quercetin-alanine concentration to HA14-1 was 1 : 0.2 the relative intensities of the selective inhibitor Tr-NOE cross-peaks that appeared were almost equal to those of the quercetin-alanine (1 : 0.8). Figure S8 demonstrates the change in the intensity of the Tr-NOE cross-peaks of the 5′–6′ protons of 3′ quercetin-alanine as well as the aromatic protons of HA14-1 on meta-position.
To ensure cell viability a trypan blue assay was performed and indicated that above 90% of the control cells and above 75% of the treated cells remained viable throughout the total experimental time (2 h 15 min) (Fig. S7) [23].
Conclusion
Herein, we employed STD and Tr-NOESY in-cell NMR spectroscopy to evaluate the direct binding of the quercetin-alanine bioconjugate to the nonlabeled intracellular Bcl-2 protein in living human cancer cells overexpressing the target protein. All the aromatic protons of the ligand were found to interact with receptors intracellularly, whereas competition experiments with a selective inhibitor of Bcl-2 clearly indicated the direct binding of the bioconjugate to the BH3 domain of the protein. Tr-NOESY in-cell NMR was recorded to investigate the preferred conformation of bound quercetin-alanine. Two new Tr-NOE cross-peaks of the ligand inside the intact cells were detected between protons H–8 and H–2′ and protons H–5′ and H–2′, suggesting the adaption of a new conformation of the bioconjugate upon binding. In vitro STD and Tr-NOESY NMR experiments with the recombinant Bcl-2 protein and the quercetin bioconjugate further supported our in vivo results and proved a direct binding.
In conclusion, in-cell NMR methodology was successfully applied, for the first time to our knowledge, in the investigation of the direct binding of a quercetin bioconjugate with the antiapoptotic Bcl-2 protein in living human cancer cells without requiring prior isotopic labeling of the target protein. This approach has proved a very promising strategy for the real-time screening of the interaction profiling of drugs with their therapeutic targets in their native cellular environment in living eukaryotic cells, paving the way to the new field of intracellular rational drug design.
Acknowledgments
The research of Dr A. Primikyri was implemented with an ΙΚΥ fellowship from the State Scholarships Foundation of Greece, funded by the Act ‘Supporting Postdoctoral Researchers’ from the resources of the NF ‘Human Resources Development, Education and Lifelong Learning’ and co-funded by European Social Fund - ESF and the Greek State. This work was also supported by the NRF grant funded by the Korea Government (MSIT) (NRF-2017R1E1A1A01074403; to SWC). This work was co-financed by the European Union (European Social Fund ESF) and Greek national funds through the Operational Program “Education and Lifelong Learning” of the National Strategic Reference Framework (NSRF) - Research Funding Program: ARISTEIA II [grant number: 5199]. We are grateful to Prof T. Fotsis (Foundation of Research and Technology–Hellas, Institute of Molecular Biology and Biotechnology, Division of Biomedical Research, Ioannina, Greece) for kindly providing us the research infrastructure of his lab for the cell culture experiments.
Author contributions
AP, AGT, and IPG carried out project conception and preparation of the manuscript. AP, AGT, IPG, and GM involved in experimental design and data analysis of the manuscript. AP, GQ, KL, and SWC performed the NMR data acquisition. NS and EIV carried out the synthesis present in the paper.