The TRIM21-overexpressed Piranha^TM HEK293T cells GFP/Puromycin Stable Cell Line (#PTPD600A-1) was purchased from System Biosciences (CA, United States). We cultured HEK293T cells and Piranha^TM HEK293T cells in Dulbecco’s modified essential medium (DMEM) high-glucose medium (#041-30081; Fujifilm) supplemented with 10% fetal bovine serum (FBS), 2 mM of GlutaMax (#35050-079; Gibco, Grand Island, NY, United States), and 100 U/mL of penicillin and 100 μg/ml of streptomycin (#15140-122; Gibco) in 5% CO₂ at 37°C. We cultured HeLa cells in DMEM (#05915; Nissui Pharmaceutical, Tokyo, Japan) supplemented with 10% FBS and 100 U/mL of penicillin and 100 μg/ml of streptomycin in 5% CO₂ at 37°C. HeLa-GFP-LC3 cells and HeLa-GFP-CD63 cells, which continuously express GFP-LC3 and GFP-CD63, respectively, were grown in the same medium with the addition of 500 μg/ml of Geneticin (#11811023; Gibco). The human renal proximal tubular epithelial cell line HK-2 (ATCC CRL-2190) was purchased from the American Type Culture Collection (Manassas, VA, United States). We cultured HK-2 cells in DMEM and Ham’s F-12 medium (DMEM/F-12) (#11320082; Gibco) containing 10% FBS and 100 U/ml of penicillin and 100 μg/ml of streptomycin (#15140-122; Gibco) in 5% CO₂ at 37°C. For details of the reagents used in this study, see the Supporting Information.

TRIM21 protein was purified according to methods described previously with slight modification (Clift et al., 2018). E. coli C41(DE3) cells were transformed with HLTV-hTRIM21 plasmids (Addgene, plasmid no. 104973), and His-Lipoyl-TRIM21 protein expression was induced with 0.1 mM IPTG overnight at 22°C. Cell pellets were sonicated in lysis buffer (300 mM KCl, 50 mM KH₂PO₄, 5 mM imidazole, pH 8.0) supplemented with complete protease inhibitors (Roche), followed by the addition of 0.5% Triton X-100 and rotation for 30 min. Following 15,000 rpm centrifugation for 20 min, the supernatant was filtered through a 0.22 µm filter, and His-Lipoyl-TRIM21 protein was purified using a 5 ml Bio-Scale Mini Profinia IMAC cartridge (Bio-rad) and the Profinia Protein Purification System according to standard Bio-rad protocols. Protein was further purified in a HiLoad 16/600 Superdex 200 PG size exclusion column using an AKTA pure purification system (GE Healthcare). Peak fractions were pooled and concentrated using Amicon Ultra-4 50 kDa (Millipore), and the buffer was exchanged to PBS. Proteins were analyzed by 5%–20% gradient sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) and Coomassie Brilliant Blue staining. His-Lipoyl-TRIM21 protein aliquots were frozen and stored at −80°C.

We prepared cytosol from murine lymphoma L5178Y cells as described previously (Kano et al., 2012).

Cytosol from HEK293T cells, HeLa cells, and HK-2 cells was obtained by freeze-thawing. First, HeLa cells or HEK293T cells grown on a 10 cm dish were trypsinized and resuspended in complete medium. The cells were washed twice with transport buffer (TB) [25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid-potassium hydroxide (pH 7.4), 115 mM potassium acetate, 2.5 mM MgCl₂, and 2 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N,N-tetra-acetic acid], and then resuspended in 100 μl TB. Following three freeze-thaw cycles and 12,000 rpm centrifugation for 10 min, the supernatant was collected as the cytosol and stored at −80°C.

The antibodies used for resealing were ultrafiltered to remove cytotoxic preservatives, such as sodium azide, as described previously (Clift et al., 2018). We developed the cell resealing protocol optimized for intracellular antibody introduction by modifying the resealing protocol described previously (Kano et al., 2012, 2017). HeLa cells and HK-2 cells were grown on 96-well dishes (#655090; Greiner Bio-One), and HEK293T cells and Piranha^TM HEK293T cells were grown on poly-D-lysine-coated 96-well dishes (#354640; Corning). The cells were washed with PBS, after which they were incubated with ∼0.4 μg/ml of SLO (#01-531; Bioacademia, Osaka, Japan) for 5 min on ice. Note that the SLO concentration should be routinely optimized because differences in cell number, nutrition conditions, or lot-to-lot variation in SLO activity affect permeabilization efficiency. After another wash, the cells were incubated with TB for 5 min at 37°C. Membrane-permeabilized cells were washed once with TB, after which they were incubated with the antibodies of interest and 3.0–6.0 mg/ml of cytosol supplemented with an ATP-regenerating system (1 mM ATP, 50 µM creatine kinase, and 2.62 mg/ml creatine phosphate), 1 mg/ml of glucose, and 1 mM GTP for 10 min at 37°C. For resealing, the cells were incubated with 1 mM CaCl₂ and 3.0–6.0 mg/ml of cytosol supplemented with an ATP-regenerating system for 5 min at 37°C. The resealed cells were further incubated with growth medium in a 5% CO₂ incubator at 37°C for the time indicated in the figure legends. Note that although the supplementation of cytosol is preferred for the resealing step for higher retention rate of the introduced antibody, cytosol in the antibody introduction step is optional, as the absence of cytosol does not affect resealing efficiency.

Two-step resealing was performed based on the protocol described above, except for the additional TRIM21 protein introduction step. Briefly, cells were treated with ∼0.4 μg/ml of SLO on ice for 5 min and incubated with prewarmed TB at 37°C for 5 min. Then, the semi-intact cells were incubated with antibody supplemented with an ATP-regenerating system for 10 min, followed by incubation with 6.5 mg/ml TRIM21 protein supplemented with an ATP-regenerating system for 5 min. Cells were then resealed by incubation with 1 mM CaCl₂ and 3.0–6.0 mg/ml of cytosol supplemented with an ATP-regenerating system for 5 min at 37°C and allowed to recover in the incubator at 37°C.

For resealing experiments using fluorophore-conjugated antibodies, the cells were imaged with a ×20 or ×40 objective using Cytation 5 reader and Gen5 software. Whole-well images were obtained by stitching the images.

For resealing experiments using a proteasome inhibitor, the cells were pretreated with 25 µM MG132 (Fujifilm; #139-18451) for 1–3 h and subjected to resealing procedures. MG132 was supplemented during the resealing steps and during incubation after resealing.

Piranha^TM HEK293T cells grown on poly-D-lysine-coated 96-well dishes (#354640; Corning) or HeLa cells grown on 96-well dishes (#655090; Greiner Bio-One) were electroporated using a Super Electroporator NEPA21 Type II (Nepagene, Chiba, Japan) and cell-culture-plate electrode (#CUY900-5-2-3; Nepagene) with 2 mm between electrodes. Cells were first washed with Opti-MEM, after which they were electroporated in 50 μl of PBS containing the antibodies of interest. The parameters for electroporation were as follows: the polarity exchanged poring pulse: 125–225 V, 2.5 m, two pulses with intervals of 50 m, and a 10% attenuation rate; the following polarity exchanged transfer pulse: 30 V, 50 m, five pulses with intervals of 50 m, and a 40% attenuation rate. This electroporation program was performed twice, and the electrode was rotated 90° the second time. Cells were incubated with growth medium in a 5% CO₂ incubator at 37°C for the time indicated in the figure legends.

HeLa cells grown on 6-cm dishes were washed with Opti-MEM, and 1 × 10⁵ cells were resuspended in 20 μL of PBS containing the antibodies of interest. The cells were then transferred to a cuvette with a 2-mm gap (#EC-002S; Nepagene) and electroporated using the following settings: poring pulse: 50–150 V, 2.5 m, two pulses with intervals of 50 m, and a 10% attenuation rate; the following polarity exchanged transfer pulse: 20 V, 50 m, five pulses with intervals of 50 m, and a 40% attenuation rate. After 600 μl of growth medium was added to the cuvette, 100 μL of the cell suspension was seeded in a 96-well dish. Cells were incubated in a 5% CO₂ incubator at 37°C for the time indicated in the figure legend.

siRNA against human IKKα (#SASI_Hs01_00206,922), siRNA against human TRIM21 (#SASI_Hs01_00060144) were purchased from Sigma Aldrich. Negative control siRNA (Silencer Negative Control 1 siRNA; #AM4635) was obtained from Ambion, Piranha HEK 293T cells were transfected using Lipofectamine^TM RNAiMAX (Invitrogen, CA, United States) according to the manufacturer’s instructions.

HEK293T, HeLa, and HK-2 cells grown in 96-well dishes were lysed with 40 μL of ice-cold radioimmunoprecipitation assay buffer [1% Triton X-100, 0.1% sodium dodecyl sulfate, 1% sodium deoxycholate, 150 mM NaCl, 50 mM Tris-HCl (pH 8.0)] supplemented with 1 mM PMSF. After the addition of 40 μl 2× Laemmli’s sodium dodecyl sulfate sample buffer, the lysates were boiled for 5 min and subjected to 5%–20% gradient sodium dodecyl sulfate polyacrylamide gel electrophoresis. Following electrophoresis, proteins were transferred to a PVDF membrane using a Criterion blotter (Bio-Rad) at 70 V for 120–180 min, after which the membrane was blocked with 5% (w/v) bovine serum albumin. The membrane was probed with primary antibodies at 4°C overnight and incubated with horseradish peroxidase-conjugated secondary antibodies at room temperature for 1 h. Finally, the proteins were detected with a Western Lightning Plus-ECL system (PerkinElmer) using LAS4000 mini (Fujifilm).

HEK293T, HeLa, and HK-2 cells grown in 96-well dishes were fixed in 4% paraformaldehyde for 15 min, after which they were permeabilized using 0.2% Triton X-100 in PBS for 5 min at room temperature. After the cells were blocked with 3% bovine serum albumin for 30 min, they were incubated with primary antibodies in blocking buffer for >2 h, washed three times with PBS, and further incubated with fluorophore-conjugated secondary antibodies and Hoechst 33342 in blocking buffer for 1 h at room temperature. The cells were washed three times with PBS, and images were obtained using a ×40 objective and a Nikon A1 confocal laser scanning microscope (Nikon, Tokyo, Japan). During automated image acquisition using NIS-Elements software, Hoechst 33342 was used for autofocusing.

After the first round of imaging, the fluorophores of the secondary antibodies were bleached using a mixture of 4.5% H₂O₂ and 25 mM NaOH in PBS for 1 h at room temperature with exposure to light from a table lamp (Lin et al., 2015, 2016). The cells were then washed three times with PBS, and a second round of immunofluorescence was performed using fluorescently labeled primary antibodies. The labeled antibodies were prepared using a Zenon Labeling Kit according to the manufacturer’s instructions (Invitrogen, CA, United States). Labeled antibodies were diluted with blocking buffer and used immediately. After the cells had been incubated with fluorescently labeled primary antibodies for 1 h, they were once again imaged with the Nikon A1 confocal laser scanning microscope. The same position in the well was imaged for all rounds of cyclic immunofluorescence, and the Hoechst images were used as references to generate registration.

After the second round of imaging, the fluorophores of the Zenon-labeled antibodies were bleached using a mixture of 3% H₂O₂ and 20 mM NaOH in PBS for 1 h at room temperature with exposure to light from a table lamp. The same procedure was performed after the third cycle of immunofluorescence.

We analyzed the microscopic images using NIS-Elements software Ver5.11 (Nikon, RRID:SCR_014329), performing maximum intensity projection for z-stack images. Cyclic immunofluorescence images acquired from different rounds of imaging were aligned and combined using the Hoechst images from every round as references.

Single-cell analysis was performed by first detecting the areas of the nucleus using binary images in the Hoechst channel and then detecting the cell area by the thickening of the nucleus area. We measured the mean or sum fluorescence intensity of the cell area for all channels. Intensity data were passed to R version 4.0.5 with RStudio (version 2021.09.1 + 372; RStudio Inc., MA, United States) for further analysis and visualization. The R packages “ggplot2” version 3.3.5 and “GGally” version 2.1.2 were used to generate pairwise plots. The dimensionality reduction algorithm t-SNE was implemented using “Rtsne” version 0.15.

For colocalization analysis, segmentation of the cytosol region was performed using NIS-Elements software, and Pearson’s correlation coefficient between the indicated channels was calculated using the pixel intensity of the cytosol region for each cell.

We assessed the statistical significance of differences from the control using Dunnett’s test as a multiple comparison test, and the differences between the two groups were detected using Welch’s t-test. In Figure 3C, we assessed the differences between individual sets of data using Tukey’s tests. p < 0.05 was considered statistically significant. Results are reported as means ± standard errors of the mean, and the number of samples used and analyzed is shown in each figure legend.

In our previous studies, we attempted to apply semi-intact cells or a cell resealing technique to deliver antibodies and inhibit their intracellular targets (Kano et al., 2000; Sonoda et al., 2021). However, the standard protocols used in these studies were not thoroughly optimized for antibody delivery, and the actual efficacy or retention of the delivered antibodies has not yet been explored in detail. Therefore, we first established a cell resealing protocol for antibody delivery using three cell lines: HEK293T, HeLa, and HK-2 (Figure 1A). As shown in Figure 1B, the antibodies were successfully introduced without excessive cell damage in all three cell lines. In these resealing experiments, the optimal SLO concentration was first determined for each of the three cell lines using single-cell analysis and microscopic images (Figure 1C, Supplementary Figure S1). For HEK293T cells, 0.2 μg/ml was the optimal SLO concentration, which successfully delivered antibodies in 74.6% of the cells (Figure 1C). We also confirmed that the delivered antibody could successfully bind to the intracellular target by resealing the anti-GFP antibody in HeLa cells expressing CD63-GFP or LC3-GFP (Supplementary Figure S2). Next, we tested the retention of the resealed antibodies for up to 24 h, confirming that almost half of the antibodies were retained in the cells at 24 h after resealing (Supplementary Figure S3A). The membrane resealing efficiency was determined for the three cell lines using the membrane-impermeable dye propidium iodide as an indicator. The results indicated that most cells efficiently resealed the pores immediately after the resealing step, whereas few cells (∼1%) exhibited incomplete resealing (Supplementary Figure S4). We also observed the resealed cells up to 48 h and confirmed that these cells exhibited normal shapes and survived for at least 48 h (Supplementary Figure S5). Taking advantage of the ability of this cell resealing technique to simultaneously introduce a variety of antibodies, we introduced three different antibodies (IgGs) into HEK293T cells using a single resealing procedure (Supplementary Figure S3B).

Then, we applied the cell resealing technique for Trim-Away-mediated protein degradation. Compared to capillary electroporation-mediated Trim-Away, which is the “gold standard” for protein degradation in bulk cell populations, cell resealing enables more rapid analysis because the cells can be kept adherent throughout the procedure. Given that TRIM21 is degraded together with the antibody–target complex, the ability to degrade the target is limited by the amount of endogenous TRIM21. Therefore, we used a HEK293T cell line constitutively expressing Piranha^TM, a modified version of TRIM21 protein, to ensure that sufficient TRIM21 was supplied for target degradation. First, we tested the IkB kinase (IKK) IKKα, which is efficiently degraded by the conventional electroporation-mediated Trim-Away method (Clift et al., 2017). IKKα is one of the two catalytic subunits of the IKK complex and is a target involved in the activation of the nuclear factor kappa-Β (NF-ĸB) pathway (Figure 2B) (Hu et al., 1999; Kiyoshi et al., 1999; Clift et al., 2017). We resealed anti-IKKα antibody (or the same amount of normal rabbit antibody as a negative control) into the TRIM21-overexpressed Piranha^TM HEK293T cells. As shown in Figure 2A, resealing the anti-IKKα antibody led to a 40% reduction in IKKα protein levels within 30 min, which was further reduced to 65% at 3 h. The degradation efficiency was high until 12 h after resealing; however, after 24 h of incubation, IKKα levels recovered to some extent, which occurred simultaneously with the decrease in resealed antibody levels (Supplementary Figure S6). Interestingly, there was only a slight decrease in IKKβ, which forms the IKK complex together with IKKα, demonstrating that specific degradation was achieved using the anti-IKKα antibody/Trim-Away system (Figure 2A). Consistent with the results of siRNA-mediated IKKα suppression (Supplementary Figure S7), IKKα degradation by Trim-Away inactivated downstream signaling, as indicated by reduced phosphorylation of NF-ĸB p65 at serine 536 (Figure 2A) (Jiang et al., 2003). We confirmed that the degradation was proteasome-dependent by treating the cells with the proteasome inhibitor MG132, which inhibited IKKα degradation (Supplementary Figure S8). Further, we confirmed that antibody resealing itself did not alter target protein levels in all three cell lines (Supplementary Figure S9).

Next, we tested whether endogenous TRIM21 was sufficient for IKKα degradation by Trim-Away. We subjected HEK293T cells without TRIM21 overexpression to the same resealing procedure, and whole cell lysates were harvested 3 h later for western blotting. As expected, although the degree of IKKα degradation was limited compared with that in TRIM21-overexpressed Piranha^TM HEK293T cells, 41% of IKKα was degraded, and downstream phosphorylation of NF-ĸB p65 was also reduced (Figure 2C, Supplementary Figure S10). We confirmed that the observed degradation in HEK293T cells without TRIM21 overexpression was attributable to endogenous TRIM21 by siRNA-mediated knockdown (KD) of TRIM21 (Supplementary Figure S11). We also observed reduced phosphorylation of NF-ĸB p65 in TRIM21 KD cells, in which IKKα was not degraded, indicating that the anti-IKKα antibody (clone Y463) inhibits IKKα activity by binding to IKKα. This observation indicates that in Trim-Away experiments, intracellularly delivered antibodies exert an additional inhibitory effect by binding to as well as degrading the target protein, and this effect likely depends on the characteristics of the antibody (e.g., binding site and clonality). Such phenomena are common, or often intrinsic in targeted degradation strategies such as PROTACs. We also tested the effects of IKKα Trim-Away knockdown in HK-2 cells and HeLa cells (Supplementary Figures S12A,B); IKKα was successfully reduced in both the cell lines, showing that they had sufficient endogenous TRIM21 for Trim-Away degradation of IKKα.

Thus far, we confirmed the ability of cell resealing-mediated Trim-Away to degrade the target and inhibit downstream signaling. To expand the versatility of the method, we further aimed to improve some of the factors that might hinder its wide use. First, we sought to develop a resealing-mediated Trim-Away method that is applicable for cells with low TRIM21 levels. Although endogenous TRIM21 was sufficient for IKKα degradation in the three cell lines, the degree of IKKα degradation in HEK293T cells was lower than that in TRIM21-overexpressed Piranha^TM HEK293T cells (Figures 2A,C). As the burden of obtaining stable cells overexpressing TRIM21 or achieving transient expression of TRIM21 by transfection limits the broad application of Trim-Away, we next attempted to introduce the antibody together with recombinant TRIM21 protein by cell resealing. We purified functional recombinant TRIM21 using the protocol by Clift et al. (2017) (Supplementary Figure S13) and resealed the protein together with normal rabbit IgG. As shown in Figure 3A, Supplementary Figure S14, simultaneous introduction of the antibody and the TRIM21 protein resulted in a substantial decrease in the levels of antibodies delivered into the cytosol, presumably because the antibody–TRIM21 complex formed and aggregated outside the cells, making it difficult for it to pass through the SLO pores. Therefore, we established a two-step resealing protocol for efficient introduction of the antibody (see Figure 3A “Two-step introduction”) and the TRIM21 protein, as outlined in Figure 3B.

We then tested if we could enhance the degradation capacity of IKKα Trim-Away by the two-step resealing protocol. As shown in Figure 3C, Supplementary Figure S15, two-step resealing of the a-IKKα antibody and the TRIM21 protein in HEK293T cells enhanced the degradation rate; a 59% reduction in the IKKα level was observed with TRIM21 compared to a 31% reduction without TRIM21 supplementation. Thus, by supplementing TRIM21 protein in HEK293T cells, we could degrade IKKα to a similar extent as that in the TRIM21-overexpressed Piranha^TM HEK293T cell line.

Next, we examined the necessity of exogenous cytosol for antibody delivery and Trim-Away degradation, as the requirement of adding exogenous cytosol may be a major drawback of the cell resealing technique. As shown in Supplementary Figure S16A, the presence of exogenous cytosol enhanced the retention rate of the introduced antibody, indicating that the cytosol facilitated the efficient resealing of the membrane pores. In contrast, cells resealed in the absence of cytosol displayed higher susceptibility to cell death, as indicated by decreased cell counts and nuclear shrinkage, which may be due to inefficient resealing (Supplementary Figure S16A). Further examination revealed that efficient antibody resealing depended on the supplementation of the cytosol only during the resealing step (5 min incubation with calcium ions) and not during the antibody introduction step when SLO pores were present (Supplementary Figure S16B). Thus, although the conventional cell resealing protocol utilizes cytosol throughout the antibody introduction step and the resealing step, we concluded that cytosol was preferable only for the resealing step. As shown in Supplementary Figure S17, Trim-Away degradation of IKKα occurred without exogenous cytosol. However, we recommend supplementation with cytosol in the resealing step because efficient resealing helps the cells recover more rapidly and reduces cell death.

The conventional cell resealing protocol used cytosol taken from L5178Y cells, a murine lymphoma cell line that enables us to obtain a large amount of cytosol at once as they can be easily grown in suspension culture. However, it is preferable to use cytosol obtained from the same cell line. Thus, we established a new convenient protocol for obtaining cytosol by simple freeze-thaw and centrifugation. We tested the cell resealing efficiency using this cytosol, which showed similar efficiency for promoting cell resealing as the conventional L5178Y cytosol (Figure 3D and Supplementary Figure S18).

In the above experiments, we established an improved resealing protocol that efficiently delivers antibodies within 40 min with limited inflow of exogenous factors, followed by Trim-Away degradation, which enables rapid analysis within 3 h. The degradation capacity could be further increased by introducing recombinant TRIM21 protein with the two-step resealing protocol, reinforcing the versatility of the Trim-Away degradation system.

A major disadvantage of Trim-Away is that it requires highly specific antibodies, which must be able to bind to the protein of interest under non-denaturing conditions. Although antibodies are commercially available for almost all proteins, only a fraction of such antibodies are appropriate for Trim-Away, emphasizing the need for validation of antibodies. When using delivery methods such as microinjection or electroporation, only one antibody can be intracellularly delivered at a time; thus, validation becomes laborious as the number of antibodies increases. In contrast, when using the cell resealing technique, antibody screening is more efficient because many antibodies can be assayed at once using standard multi-well plates and multi-channel pipettes.

Here, we applied the cell resealing technique to screen and assess the applicability for Trim-Away degradation of five different antibodies against IKK. TRIM21-overexpressed Piranha^TM HEK293T cells were resealed with 0.125 mg/ml of the respective anti-IKK antibodies and incubated for 3 h. Subsequently, the efficiency and specificity of IKK degradation was tested by quantifying the degradation of IKKα and IKKβ using western blotting. As shown in Figure 4A, only one of the five antibodies significantly degraded IKKα. Although not significant, one anti-IKKβ–specific antibody (clone EPR6043) tended to reduce IKKβ levels. Next, the two antibodies that displayed the highest potential for degrading IKKα and IKKβ, respectively, were further analyzed for their concentration-dependent degradation activity and specificity (Figure 4B). The IKKα-specific antibody (clone Y463) degraded IKKα more efficiently than IKKβ, whereas the IKKβ-specific antibody (clone EPR6043) degraded IKKβ more efficiently than IKKα. Both antibodies showed a concentration-dependent ability to degrade IKKα and IKKβ. Although we anticipated that antibodies recommended for immunoprecipitation by the manufacturer would be suitable for Trim-Away, the immunoprecipitation-validated antibodies for IKKα+β (clone EPR16628) and IKKα (clone 14A231) degraded neither IKKα nor IKKβ. Thus, the validation and selection of antibodies prior to conducting Trim-Away with the cell resealing technique is highly recommended to ensure the success of Trim-Away.

Next, to test whether other target proteins were also degraded by the cell resealing-mediated Trim-Away system, we targeted mTOR, a central kinase regulating nutrient signaling and cell growth (Figure 5A) (Laplante and Sabatini, 2009; Clift et al., 2017) for degradation. We resealed 0.5 mg/ml of anti-mTOR antibody to TRIM21-overexpressed Piranha^TM HEK293T cells, resulting in a 44% reduction in mTOR protein level at 12 h after resealing (Figures 5B,C “Cell resealing”). Treatment with an mTOR-specific inhibitor, Torin, resulted in the inactivation of downstream signaling, i.e., reduced phosphorylation of Akt at S473 and ribosomal protein 6 (S6rp) at Ser235/236 (Supplementary Figure S19). Therefore, we assessed the inactivation of downstream signaling using mTOR Trim-Away, which revealed the suppression of this signaling (Figure 5, Supplementary Figure S20). Notably, endogenous TRIM21 was not sufficient to degrade the substantial amount of mTOR in HEK293T cells, whereas cells with additional expression of exogenous TRIM21 (“TRIM21-overexpressed HEK293T cells” in Supplementary Figure S21) showed improved mTOR degradation.

Comparing the degradation efficiency of mTOR with IKKα, the extent of mTOR degradation was less than that observed for IKKα, despite a higher amount of antibody used and a longer incubation time for degradation (comparing Figure 5C “Cell resealing” with Figure 2A). As the Trim-Away degradation efficiency depends on the amount of antibodies delivered to the cytosol, the amount of TRIM21 protein, and the endogenous level of the target protein, the amount of antibody and TRIM21 protein may have been insufficient to degrade the high abundance of mTOR. Another possible reason for the lower degradation efficiency might derive from the inaccessibility of the antibody to the target. A subset of mTOR, which is known to localize to lysosomes (Betz and Hall, 2013), might have been difficult to target for degradation (Clift et al., 2017).

Here, we also tested adherent electroporation-mediated antibody delivery using electrodes designed for 96-well plates. As with cell resealing, adherent electroporation allows rapid analysis following antibody introduction by maintaining cell adherence throughout the procedure, in contrast to capillary electroporation or conventional cuvette electroporation. Therefore, we aimed to compare the efficiency of adherent electroporation-mediated Trim-Away with that of cell resealing-mediated Trim-Away. After determining the optimal pulse conditions for adherent electroporation-mediated antibody delivery (Supplementary Figure S22), we introduced anti-mTOR antibody through adherent electroporation and analyzed Trim-Away-mediated degradation using western blotting. We used 0.5 mg/ml of anti-mTOR antibody for adherent electroporation-mediated antibody delivery; this concentration was the same as that used in the resealing experiment. As shown in Figures 5B,C (“Adherent electroporation”), we did not detect significant degradation of mTOR or the inhibition of downstream signaling. This is likely because only the cells near the center of the well, which were located between the electrodes, were well electroporated, whereas the vast majority of cells at the periphery were not well electroporated (Figure 5D). In contrast to adherent electroporation, the cell resealing technique enables the delivery of antibodies into most of the cells in the well, whether they are at the center or the periphery (see the whole-well view in Figure 1C). Thus, in comparing the two antibody introduction methods for adherent cells, only the cell resealing technique was suitable for Trim-Away when analyzed by western blotting.

Considering the advantage of the cell resealing technique, which enables antibody introduction while keeping the cells adherent, microscopic image-based single-cell analysis is especially suitable for detecting and analyzing Trim-Away. In addition to the ability to detect the morphology of cells or the localization of proteins, another important advantage of microscopic analysis is the ability to compensate for the cell-to-cell variation in the amount of resealed antibody. As there are cell-to-cell differences in membrane cholesterol levels that affect the number of SLO molecules that can bind to the membrane cholesterol to form pores, cell-to-cell variation in the amount of resealed antibody is inevitable. Accordingly, in analyses such as western blotting, which use whole cell lysates, the results represent the average of a heterogenous cell population, i.e., in some cells antibodies are abundant, whereas in other cells they are scarce. In contrast, microscopic single-cell analysis enables a more sensitive and precise analysis by examining only the efficiently resealed cells or the correlation between the amount of antibody introduced and the degradation efficiency. We have employed single-cell analysis with immunofluorescence and confocal microscopy in our previous studies, and this technique enabled the successful analysis and phenotyping of various cells, including resealed cells (Kano et al., 2017; Kunishige et al., 2020; Noguchi et al., 2021). Therefore, we reasoned that single-cell analysis would allow us to detect the effects of Trim-Away more clearly.

Immunofluorescence is a standard method for detecting the morphology of cells as well as the amount, modification state, and localization of proteins in a single-cell. However, conventional immunofluorescence imaging is limited to 4–6 channels, making it difficult to analyze the complex relationships between proteins in heterogeneous cell populations. As such, we adopted cyclic immunofluorescence, a method used to produce highly multiplexed imaging via a repeated staining and bleaching procedure, which can be conducted using standard reagents and microscopes (Lin et al., 2015, 2016; Radtke et al., 2020, 2022). Here, we established a cyclic immunofluorescence single-cell analysis system by optimizing the staining and bleaching protocol as well as the image acquisition procedure for obtaining the same field of view and the alignment of the images obtained from different cycles. By combining Trim-Away with cyclic immunofluorescence, we were able to simultaneously detect protein degradation and the resulting perturbation of downstream signal transduction at the single-cell level.

As outlined in Figure 6A, we first performed mTOR Trim-Away via cell resealing, and this was followed by fixation and three cycles of cyclic immunofluorescence (as described in the Experimental Procedures). The mean intensities of mTOR, phosphorylated Akt (pAkt), Akt, phosphorylated S6rp (pS6rp), S6rp, rabbit antibody (mTOR antibody or normal rabbit antibody), and Hoechst were quantified per cell using confocal images (Figure 6B).

For single-cell data analysis, we first created pairwise plots from the multivariate data, which enabled us to explore distributions and correlations (Supplementary Figure S23). A histogram of fluorescence intensity indicated that the anti-mTOR antibody resealing condition had lower mTOR and pAkt intensities. In addition, a scatterplot showing the fluorescence of resealed anti-mTOR antibody and the immunofluorescence-stained mTOR fluorescence revealed a weak negative correlation between the two, indicating that cells with higher amounts of mTOR antibodies were able to degrade more mTOR. The same tendency was detected for pAkt and the resealed anti-mTOR antibody. For comparison, we applied the cyclic immunofluorescence analysis framework described above for the single-cell analysis of adherent electroporation-mediated Trim-Away. In contrast to the western blotting results, we were able to detect Trim-Away degradation of mTOR and its downstream effects by microscopic image-based single-cell analysis, which enabled us to focus only on the well electroporated cells (Supplementary Figure S24).

In further analyses, we applied t-distributed stochastic neighbor embedding (t-SNE) (Van der Maaten and Hinton, 2008), a dimensionality reduction algorithm useful for visualizing high-dimensional single-cell data in a two-dimensional space (Figure 6C) (Amir et al., 2013; Lin et al., 2015, 2016). We analyzed cells in which negative control antibodies or anti-mTOR antibodies had been delivered, and we distinguished a domain on the left side of the plot consisting of anti-mTOR antibody-resealed cells (see the red arrowhead in Figure 6C “Ab condition”). However, some of the cells in the anti-mTOR resealed condition diffused outside this domain; presumably, these cells were insufficiently permeabilized (see the black arrowheads in Figure 6C “Ab condition”). Importantly, the anti-mTOR antibody-resealed cells clustered on the left side of the plot had relatively low mTOR and pAkt intensity levels (see the red arrowhead in Figure 6C “mTOR” and “pAkt” plots). Consequently, we reasoned that the cells with higher amounts of mTOR antibodies were able to degrade more mTOR and that the downstream phosphorylation of Akt and S6rp would likely be reduced. Accordingly, we filtered cells that were in the top 5%, 10%, 25%, and 50% by the fluorescence of the introduced antibody. Figure 6D includes density plots showing the mean intensity of mTOR and phosphorylated levels of Akt and S6rp. As expected, the difference in the density plots between anti-mTOR and the negative control became clearer as the data were narrowed down to the most efficiently resealed cells. The same tendency was observed for phosphorylated levels of Akt and S6rp (see Figure 6D “pAkt/Akt” and “pS6/S6rp”).

Thus, the cyclic immunofluorescence method and subsequent microscopic image-based single-cell analysis were found to be a suitable analysis system for cell resealing-mediated Trim-Away, where cell-to-cell variation in the amount of antibody had been introduced.