This is a protocol to study intracellular protein-protein interactions based on the biotin-avidin pull-down system with the novelty of combining cell-penetrating sequences. The main advantage is that the target sequence is incubated with living cells instead of cell lysates and therefore the interactions will occur within the cellular context.
Here we present a protocol to study intracellular protein-protein interactions that is based on the widely used biotin-avidin pull-down system. The modification presented includes the combination of this technique with cell-penetrating sequences. We propose to design cell-penetrating baits that can be incubated with living cells instead of cell lysates and therefore the interactions found will reflect those that occur within the intracellular context. Connexin43 (Cx43), a protein that forms gap junction channels and hemichannels is down-regulated in high-grade gliomas. The Cx43 region comprising amino acids 266-283 is responsible for the inhibition of the oncogenic activity of c-Src in glioma cells. Here we use TAT as the cell-penetrating sequence, biotin as the pull-down tag and the region of Cx43 comprised between amino acids 266-283 as the target to find intracellular interactions in the hard-to-transfect human glioma stem cells. One of the limitations of the proposed method is that the molecule used as bait could fail to fold properly and, consequently, the interactions found could not be associated with the effect. However, this method can be especially interesting for the interactions involved in signal transduction pathways because they are usually carried out by intrinsically disordered regions and, therefore, they do not require an ordered folding. In addition, one of the advantages of the proposed method is that the relevance of each residue on the interaction can be easily studied. This is a modular system; therefore, other cell-penetrating sequences, other tags, and other intracellular targets can be employed. Finally, the scope of this protocol is far beyond protein-protein interaction because this system can be applied to other bioactive cargoes such as RNA sequences, nanoparticles, viruses or any molecule that can be transduced with cell-penetrating sequences and fused to pull-down tags to study their intracellular mechanism of action.
Protein-protein interactions are essential for a great variety of cellular processes. To fully understand these processes, methods for identifying protein interactions within the complex intracellular environment are required. One of the most used methods to identify interaction partners of a protein is to use that protein or a mimetic peptide of a part of that protein as bait in affinity pull-down experiments followed by detection of binding proteins. The avidin-biotin system is frequently used because of the high affinity, specificity and stable interaction between avidin and biotin1,2. Usually, biotin is covalently bound to the bait (protein or peptide) and after a period of incubation with the cell lysates to allow the establishment of interactions, the biotinylated bait bound to its intracellular partners is pulled down with avidin or avidin derivatives conjugated with support beads. Then, bait-protein interactions are detected after washing, elution, and analysis by denaturing electrophoresis followed by Western blot. One of the problems of this technique is that the interactions between the protein of interest and its intracellular partners are taking place outside of the cellular context. This is especially important for the interactions involved in signal transduction pathways because they take place in specific intracellular locations, they are transient and they are typically carried out by not abundant proteins. Therefore, within the cell lysates these interactions can be masked by other more abundant proteins or by proteins that usually are not in close proximity.
Cell-penetrating peptides (CPPs) are short peptides (≤40 amino acids), composed mostly by cationic amino acids that are capable of transporting a wide range of molecules into almost any cell3. Cargoes such as proteins, plasmid DNA, siRNA, viruses, imaging agents, and various nanoparticles have been conjugated to CPPs and efficiently internalized4,5. Because of this transporting ability they are also known as protein transduction domains (PTDs), membrane translocating sequences (MTSs), and Trojan peptides. Among the CPPs, the TAT peptide from the HIV transactivator protein TAT6 has been one of the most widely studied 7,8,9. TAT is a nonapeptide that contains 6 arginine and 2 lysine residues and consequently is highly cationic. Substitution studies have demonstrated that the net positive charge of TAT is necessary for electrostatic interactions with the plasma membranes of eukaryotic cells and its subsequent internalization10. Similarly to other CPPs, positively-charged TAT strongly binds electrostatically to the various negatively-charged species present at the extracellular surface of cell membranes, including lipid head groups, glycoproteins and proteoglycans3,10. The bioactive cargoes transported by TAT become immediately free in the cytosol to reach their intracellular partners.
Here we present a method that combines the TAT CPP with biotin to study intracellular interactions. The aim is to design cell-penetrating baits by fusing the target biomolecule to TAT and to biotin. The main advantage of this proposal is that the interactions between the bait and its partners will take place within its cellular context. To show the efficacy of this method we used as bait a small sequence of the protein Cx43 that has been reported to interact intracellularly with the proto-oncogene c-Src11,12,13. Cx43 is an integral membrane protein that is widely expressed in astrocytes14and is down-regulated in high-grade gliomas, the most common malignant tumor of the central nervous system15,16,17,18.It has been previously shown that the Cx43 region that interacts with c-Src (amino acids 266-283 in human Cx43; Pubmed: P17302) fused to TAT (TAT-Cx43266-283) inhibits the oncogenic activity of c-Srcin glioma cells and glioma stem cells (GSCs)19,20,21. To design the intracellular bait, Cx43266-283 has been fused to TAT at the N-terminus (TAT-Cx43266-283) and to biotin at the C-terminus (TAT-Cx43266-283-B). This strategy has been successfully used in the rat glioma C6 cell line to identify c-Src, c-terminal Src kinase (CSK) and phosphatase and tensin homolog (PTEN) as intracellular partners of this region of Cx4320. Here, we describe this method testing its efficacy in human GSCs, which are very relevant for glioma therapy but much harder to transfect than non-stem glioma cells.
All experimental procedures were carried out at the University of Salamanca.
1. Cells
2. Biotinylated CPPs
3. Tubes
NOTE: Prepare at least twelve 1.5 mL tubes per condition required in the Section 7.
4. Cellular treatment with the BCPPs
5. Buffers and solutions.
6. Protein extraction
NOTE: Protein extraction was performed as previously described20,23. Carry out this whole section of the procedure at 4 °C.
7. Pull-down
8. Western Blot
NOTE: Western blotting was performed as previously described24.
9. Problem Solving
10. Other Techniques Used in this Article
Before using BCPPs to study intracellular interaction, it is critical to compare the effects of BCPP vs CPP to validate the results obtained with BCPP. Consequently, to study whether the inclusion of biotin modifies the activity of the target sequence, we first analyzed the effect of TAT-Cx43266-283-B compared with TAT-Cx43266-283 on G166 GSCs morphology. To do so, we performed some immunofluorescence analyses of two cytoskeletal proteins, F-actin and α-tubulin after 24 h of treatment. Figure 1 shows that G166 GSCs in the presence of 50 µM TAT-Cx43266-283 or TAT-Cx43266-283-B acquire a more rounded shape compared to the elongated and expanded cellular prolongations shown in the controls (TAT or TAT-B). In fact, Figure 1b shows that actin filaments are mostly assembled as actin networks when the cells were treated with TAT-Cx43266-283 or TAT-Cx43266-283-B while they form more actin bundles in the control cells (treated with TAT or TAT-B)25. In contrast, α-tubulin distribution does not vary between the different conditions. These results showed that the presence of biotin did not modify the effect of the target sequence on the morphology of G166 GSCs. In previous studies20,21, we showed that TAT-Cx43266-283 reduced G166 GSCs proliferation. In this study, we investigated whether TAT-Cx43266-283-B exerts the same effects in the growth as TAT-Cx43266-283. To do so, we analyzed the G166 GSCs proliferation by MTT assay after 72 h of treatment. The MTT assay is a colorimetric assay for assessing cell metabolic activity. MTT is metabolized by NAD(P)H oxidoreductase enzymes in mitochondria reflecting the number of viable cells present. Figure 2 shows that the reduction in the G166 GSCs cell viability is not significantly different when cells were treated with 50 µM TAT-Cx43266-283 or 50 µM TAT-Cx43266-283-B. Indeed, both significantly diminished G166 GSCs proliferation as compared to the control, TAT or TAT-B.
Once we confirmed that the effect of our target sequence in G166 GSCs (TAT-Cx43266-283) was not modified by the inclusion of biotin at the C-terminus (TAT-Cx43266-283-B), we investigated the intracellular partners of this sequence following the protocol described in this study (Figure 3). Because caveolae have been involved in the mechanism of TAT internalization26, we analyzed the presence of caveolin-1 (Cav-1) in the pull-downs. Western blot analysis (Figure 4) showed that TAT-B and TAT-Cx43266-283-B interact with Cav-1. However, the ability of TAT-Cx43266-283-B to recruit c-Src, PTEN and CSK is stronger than that found with TAT-B. Focal adhesion kinase (FAK) is a substrate of c-Src that has not been shown to interact with Cx43. Indeed, FAK did not show any significant interaction with either TAT-B or TAT-Cx43266-283-B.
To confirm the interaction between TAT-Cx43266-283-B and c-Src, G166 GSCs were incubated with 50 µM TAT-Cx43266-283-B for 30 min and their localization was followed with fluorescent streptavidin by confocal microscopy (Figure 5). Our results showed that the intracellular distribution of TAT-Cx43266-283-B is close to the plasma membrane (shown by phosphatidylserine staining) and matches with that of c-Src. In fact, co-localization analyses revealed some points of co-localization (white) between TAT-Cx43266-283-B and c-Src in the merge image. Consequently, confocal microscopy studies confirm the results obtained with the BCPP pull-down protocol described in this study.
Figure 1: Effect of BCPP and CPP on GSC morphology.
G166 GSCs were plated at a low density (2 x 104 cells / cm2) and after 24 h they were incubated with 50 µM control CPP (TAT), control BCPP (TAT-B), treatment CPP (TAT-Cx43266-283) or treatment BCPP (TAT-Cx43266-283-B). a) F-actin (red), α-tubulin (green) and merged + DAPI immunostaining of the same field showing G166 GSCs morphology. Bars = 50 µm. b) F-actin immunostaining showing the different distribution of F-actin in G166 GSCs after incubation for 24 h with 50 µM control CPP (TAT) or control BCPP (TAT-B) as compared with 50 µM treatment CPP (TAT-Cx43266-283) or BCPP (TAT-Cx43266-283-B). Bars = 10 µm. Please click here to view a larger version of this figure.
Figure 2: Effect of BCPP and CPP on GSC viability.
G166 GSCs were plated at 5500 cells/cm2 in 24-multiwell plates and incubated with 50 µM control peptides, CPP (TAT) or BCPP (TAT-B), or 50 µM treatment peptides, CPP (TAT-Cx43266-283) or BCPP (TAT-Cx43266-283-B). The cell viability was analyzed using a MTT assay after 72 h. The results are expressed as MTT absorbances and are the mean ± s.e.m. of at least 3 experiments (++ p˂0.01 vs control. ** p˂0.01, *** p˂0.001 vs TAT or TAT-B; one-way ANOVA withTukey post-test). Note that there are not significant differences between the effects of CPPs vs BCPPs. Please click here to view a larger version of this figure.
Figure 3: Protocol diagram.
Step by step graphical depiction of the procedure as described in the section "Protocol", from the incubation of the BCPPs until the eluted BCPPs and their interacting proteins were obtained.1) Incubate culture cells with BCPPs at the desired concentration for the required time. 2) During the incubation, the BCPPs are internalized and they interact with their intracellular partners. 3) Wash the cells three times on ice with ice-cold PBS. 4) Lyse the cells to extract proteins. 5) Transfer cell lysates to tubes. 6) Spin at 11000 x g for 10 min at 4 °C. 7) Transfer the supernatants to new tubes (A) and keep a small aliquot of the lysates to process as regular Western blot samples in tube (B). 8) Resuspend the NeutrAvidin Agarose beads and add 50 µL to each tube A using a cut pipette tip. 9) Incubate with gently shaking for 12 h at 4 °C to allow the NeutrAvidin agarose beads to interact with BCPPs and their partners. 10) Spin for 1 min at 3000 x g to pellet the beads with the biotinylated baits and their interacting proteins bound to them. 11) Transfer supernatants to new tubes (C) and keep them to use in case the pull-down need to be repeated. 12) Wash the pellet five times with fresh lysis buffer, resuspend by inversion, spin for 1 min at 3000 x g and discard the supernatant. 13) Remove all the supernatant carefully. 14) Add the desired volume of 4x Laemmli buffer and elute the proteins at 100 °C for 5 min. 15) Spin at 8200 x g for 30 s to pellet the beads. 16) Transfer the eluted proteins found in the supernatant with capillary tips to new tubes (D). 17) Load onto gels for Western blot analysis. Please click here to view a larger version of this figure.
Figure 4: Study of the intracellular interactions of TAT-Cx43266-283-B in G166 GSCs by pull-down followed by Western blot.
G166 GSCs were incubated with 50 µM TAT-B or TAT-Cx43266-283-B. After 30 min the cells were lysed and TAT-B or TAT-Cx43266-283-B attached to their intracellular partners were pulled down with NeutrAvidin beads. The eluted proteins were loaded and analyzed by Western blot to study the levels of FAK, c-Src, CSK, PTEN and Cav-1. Note that Cav-1 interacts with both TAT-B and TAT-Cx43266-283-B, c-Src, PTEN and CSK interact preferentially with TAT-Cx43266-283-B and FAK did not show any interaction with either TAT-B or TAT-Cx43266-283-B. Please click here to view a larger version of this figure.
Figure 5: Confirmation of TAT-Cx43266-283-B intracellular interactions in G166 GSCs by confocal microscopy.
G166 GSCs were incubated with 50 µM TAT-Cx43266-283-B. After 30 min, cells were fixed and processed to localize TAT-Cx43266-283-B with Cy2-Streptavidin (green), c-Src by immunofluorescence (red) and phosphatidylserine with annexin V (blue). Note some points of co-localization (white) between TAT-Cx43266-283-B and c-Src close to the plasma membrane in the merge images. Please click here to view a larger version of this figure.
There are many methods to study protein-protein interactions. The method presented in this study is based on the widely used biotin-avidin pull-down system in which a biotinylated bait is incubated with cell lysates to allow the establishment of interactions. The modification presented in this study includes the combination of this technique with cell-penetrating sequences. We propose to design cell-penetrating baits that can be incubated with living cells instead of cell lysates and therefore the interactions found will reflect those that occurred within the cellular context.
Here we use TAT as the cell-penetrating sequence, biotin as the pull-down tag and the region of Cx43 comprised between amino acids 266-283 as the target to find intracellular interactions in human GSCs. The structural basis for the interaction between Cx43 and c-Src is well known11,12. This is an important interaction because it inhibits the oncogenic activity of c-Src in glioma cells24,13. In fact, CPPs containing this region (TAT-Cx43266-283) mimic the antioncogenic properties of Cx43 on glioma cells19,20,21. In rat glioma C6 cells, the mechanism by which TAT-Cx43266-283 inhibits c-Src includes the recruitment of c-Src together with its endogenous inhibitors CSK and PTEN20. It should be mentioned that GSCs are very interesting as a target in glioma therapy, because they constitute a subpopulation that is resistant to conventional treatments and therefore responsible for the recurrence of this malignant brain tumors27. Furthermore, they are hard-to-transfect cells and therefore the study of intracellular interactions becomes more difficult. CPPs are rapidly and efficiently internalized in GSCs19 favoring their use for the study of intracellular interactions. In this study, using CPPs fused to biotin we confirm the interaction of the Cx43266-283 sequence with c-Src together with its endogenous inhibitors CSK and PTEN in human GSCs.
This method is very powerful to study the intracellular mechanism of bioactive compounds. However, it is very important to confirm that the biological effect of the biotinylated cell penetrating bait is not different from that obtained with the non-biotinylated one. This step is required to associate the interactions found with the effect of the bioactive compound. In addition, the stability of the compound, its possible degradation by proteases as well as its possible toxicity, should be carefully tested and taken into account before planning the experiment. In the example presented, the anti-proliferative effect of TAT-Cx43266-283 on G166 human GSCs has been previously documented20. In this study, we confirm that the anti-proliferative effect of TAT-Cx43266-283-B and of TAT-Cx43266-283 is very similar. In addition, the analysis of cellular morphology revealed that α-tubulin and F-actin distribution is very similar in G166 GSCs treated with TAT-Cx43266-283-B or with TAT-Cx43266-283. Altogether, these results indicate that the inclusion of biotin at the c-terminus of TAT-Cx43266-283 did not modify the effects of this compound on human GSCs. However, if biotin would modify the effects of the bioactive molecule, other tags for protein purification can be tested, such as the FLAG octapeptide (DYKDDDDK)28, the human influenza hemagglutinin-derived tag HA (YPYDVPDYA) or glutathione S-transferase (GST)29. Similarly, if TAT does not target the cell population of interest, other cell penetrating sequences, such as penetratin, MPG (for a review, see30) or cell specific sequences can be used31.
In addition to study proteins that specifically interact with the target sequence, ideally, the presence of proteins that interact with both the control and the target sequence and proteins that do not interact with them, as positive and negative controls, should be addressed. In this sense, we found Cav-1 in the control and treated situation, suggesting that the caveolae have been involved in the mechanism of internalization, as it has been previously shown26. Furthermore, FAK, which interacts with c-Src but is supposed not to interact with the Cx43 c-terminal, was absent in both the control and treated situation. These results reinforce the specificity of the interaction between TAT-Cx43266-283-B, c-Src, CSK and PTEN.To confirm the results obtained with this protocol, confocal microscopy can be used to visualize the distribution of the interacting proteins and to study their co-localization. Thus, we found that TAT-Cx43266-283-B and c-Src exhibit a similar intracellular distribution with some points of co-localization confirming the results obtained with the pull-down experiments. In fact, TAT-Cx43266-283-B is distributed close to the plasma membrane suggesting that the cargo, in this study Cx43266-283, directs the molecule to its intracellular partners.
One of the limitations of the proposed method is that the molecule used as bait could fail to fold properly and the expected effects would not be found. In this situation, the interactions found could not be associated to the effect. However, this method can be especially interesting for the interactions involved in signal transduction pathways because they are usually carried out by intrinsically disordered regions32 and therefore they do not require an ordered folding. In addition, one of the advantages of the proposed method is that the time course of the interaction can be followed, which is especially relevant for transient interactions. Furthermore, the relevance of each residue on the interaction can be easily studied. Indeed, it is possible to study the relevance of posttranslational modifications on protein-protein interaction, for instance, by phosphomimetic substitution of glutamate for serine or threonine. Similarly, substitution of serine or threonine for alanine or tyrosine for phenylalanine allows testing the effect of non-phosphorylatable serine, threonine or tyrosine. To mimic phospho-tyrosine, the most accurate way is the substitution of Tyr for p-Tyr33.
Finally, the scope of this protocol is far beyond protein-protein interaction because this system can be applied to other bioactive cargoes such as RNA sequences, nanoparticles, viruses or other molecules that can be fused to biotin and transduced with CPP to study their intracellular mechanism of action.
The authors have nothing to disclose.
We thank M. Morales and J. Bravo for their help with the design of CPPs and J.C. Arévalo for his help with the pull-down protocol. We are grateful for the technical assistance of T. del Rey. This work was supported by the Ministerio de Economía y Competitividad, Spain; FEDER BFU2015-70040-R, Junta de Castilla y León, Spain; FEDER SA026U16 and Fundación Ramón Areces. M. Jaraíz-Rodríguez and A. González-Sánchez are recipients of a fellowship from the Junta de Castilla y León and the European Social Fund.
G166 GSC line | BioRep | ||
RHB-A stem cell medium | Takara | Y40001 | |
Laminin Mouse Protein | Invitrogen, Life Technologies, ThermoFisher Scientific | 23017-015 | 10 µg/ml |
B-27 Serum free Supplement (50X) | Invitrogen, Life Technologies, ThermoFisher Scientific | 17504-044 | 2% |
N-2 Supplement (100x) | Invitrogen, Life Technologies, ThermoFisher Scientific | 17502-048 | 1% |
Recombinant Human EGF | Peprotech | AF-100-15 | 20 ng/ml |
Recombinant Human b-FGF | Peprotech | AF-100-18B | 20 ng/ml |
PBS pH 7.4: In deionized water, 136 mM NaCl ; 2.7 mM KCl; 7.8 mM Na2HPO4·2H2O ; 1.7 mM KH2PO4 | |||
Accutase | Sigma | A6964 | |
Cryostor CS10 cryopreservation medium | StemCell Technologies | 7930 | |
TAT | GenScript | – | Custom made |
TAT-B | GenScript | – | Custom made |
TAT-Cx43266-283 | GenScript | – | Custom made |
TAT-Cx43266-283-B | GenScript | – | Custom made |
Alexa Fluor 594 Phalloidin | Molecular Probes, Life Technologies, ThermoFisher Scientific | A1275737 | 1/20 |
Monoclonal α-tubulin mouse antibody | Sigma-Aldrich | T9026 | 1/500 |
4',6-diamidino-2-phenylindole (DAPI) | Molecular Probes, Life Technologies, ThermoFisher Scientific | 1.25 mg/ml | |
Pierce™ NeutrAvidin™ Agarose | ThermoFisher Scientific | 29200 | |
Protein lysis buffer: 5 mM Tris-HCl (pH 6.8), 2% (w/v) SDS, 2 mM EDTA , 2 mM EGTA | |||
Protease Inhibitor Cocktail Set III. EDTA-Free | Calbiochem, Bionova | 539134 | 1/100 (v/v) |
Sodium Fluoride | PanReac AppliChem | 141675 | 1 mM |
Phenylmethanesulfonyl fluoride (PMSF) | Sigma-Aldrich | P7626 | 1 mM |
Sodium orthovanadate | Sigma-Aldrich | S6508 | 0.1 mM |
Laemmli buffer: (4x: 0.18 M Tris-HCl pH 6.8; 5 M glycerol; 3.7 % (w/v) SDS; 0.6 M β-mercaptoethanol or 9 mM DTT ; 0.04% (v/v) bromophenol blue (BB) . | 1X | ||
Xcell 4 SureLock Midi-Cell Electrophoresis System | Life Technologies, ThermoFisher Scientific | WR0100 | |
NuPAGE Novex Bis-Tris Midi-Gels 4-12% | Life Technologies, ThermoFisher Scientific | WG1402box | |
NuPAGE MOPS SDS Running Buffer (20X) | Life Technologies, ThermoFisher Scientific | NP0001 | 1X |
NuPAGE Transfer Buffer 20x | Life Technologies, ThermoFisher Scientific | NP0006 | 1X |
Precision Plus Protein Dual Color Standard | Bio-Rad | 161-0374 | |
iBlot 2 NC Regular Stacks (nitrocellulose membranes) | Life Technologies, ThermoFisher Scientific | IB23001 | |
iBlot 2 Dry Blotting System – Gel transfer device | Life Technologies, ThermoFisher Scientific | IB21001 | |
10% Ponceau S Solution (0.1% Ponceau (w/v) in 5% acetic acid (v/v)) in water | Sigma | P7170 | |
FAK polyclonal rabbit antibody | Life Technologies, ThermoFisher Scientific | AHO0502 | 1/500 |
Src polyclonal rabbit antibody | Cell Signalling (WERFEN) | 2108S | 1/500 |
Csk polyclonal rabbit antibody | Cell Signalling (WERFEN) | 4980 | 1/500 |
PTEN polyclonal mouse antibody | Cell Signalling (WERFEN) | 9552 | 1/500 |
Caveolin-1 polyclonal rabbit antibody | Abcam | ab2910 | 1/1000 |
Goat anti-mouse IgG-HRP antibody | Quimigen, Santa Cruz Biotechnology | Sc-2005 | 1/5000 |
Goat anti-rabbit IgG-HRP antibody | Quimigen, Santa Cruz Biotechnology | SC-2030 | 1/5000 |
Western Blotting Luminol Reagent | Santa Cruz Biotechnology | SC-2048 | |
Src polyclonal rabbit antibody | Cell Signalling (WERFEN) | 2108S | 1/500 |
Antibody solution: PBS, 10% FBS, 0.1 M lysine, 0.02% sodium azide | |||
Alexa Fluor 488 goat anti-mouse IgG | Invitrogen, Life Technologies, ThermoFisher Scientific | A11029 | 1/1000 |
Cy2-conjugated streptavidin | Jackson ImmunoResearch | 016-220-089 | 1/500 |
MicroChemi Luminescence system | DNA Bio-Imaging Systems | ||
Dead Cell Apoptosis Kit with Annexin V Alexa Fluor® 488 & Propidium Iodide (PI) | Molecular Probes, Life Technologies, ThermoFisher Scientific | V13241 | 1/500 |
SlowFade Gold antifade reagent | Life Technologies, ThermoFisher Scientific | S36936 | |
Thiazolyl Blue Tetrazolium Bromide (MTT) | Sigma-Aldrich | M2128 | |
Dimethyl sulfoxide for UV-spectroscopy, >=99.8% (GC) | Honeywell | 41641-1L | |
Appliskan 2001 | Thermo Electron Corporation, Thermo Scientific |