Search
Close this search box.

Bioorthogonal photocatalytic proximity labeling in primary living samples – Nature Communications

Ethical statement

All animal studies were approved by the Institutional Animal Care and Use Committee of Peking University (CCME-ChenP-3). All protocols and experiments using human blood samples from Oricells (Shanghai OriBiotech Co., Ltd) were approved by the ethics committee of Shanghai Zhaxin Hospital (No. LP202006).

Materials

General considerations

All expression media, buffers, and antibiotics were prepared using purified H2O (Mili-Q Reference) and autoclaved or filter-sterilized, as appropriate. Photocatalysts were commercially purchased from J&K Scientific® or Energy Chemical®. Protein and nucleic acid concentrations were determined by NanoDrop 2000 spectrophotometer (Thermo). Images of protein gels including nucleotide agarose gel, coomassie SDS-PAGE gel, and western blotting membranes were taken on ChemiDoc XRS+ (Bio-Rad) with Image Lab Touch software (v2.3.0, BioRad). UV-vis spectrometry for cell samples was measured on a Synergy H4 microplate reader (Bio-Tek) with Gen5 software (v3.11.19, BioTek). Confocal microscopy images were obtained on an LSM 700 laser scanning confocal microscope (Zeiss) with ZEN software (v2.1, ZEISS). Flow cytometry was performed on LSRFortessa cell analyzer (BD biosciences) supplemented with BD FACSDiva (v7.0, BD biosciences) software. Figures with cartoons were created with BioRender.com.

Cell culture

HeLa (catalog#1101HUM-PUMC000011), HEK293T (catalog#1101HUM-PUMC000091), and K562 (catalog#1101HUM-PUMC000039) cells were purchased from Cell Resource Center, Peking Union Medical College, China. HeLa and HEK293T cells were grown in DMEM (Dulbecco’s modified Eagle’s medium, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco) at 37 °C in 5% CO2. K562 cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum (FBS, Gibco) at 37 °C in 5% CO2.

Mouse model

C57BL/6 J (B6) mice (female, 6–8-week, Cat# 219) were purchased from Beijing Vital River Laboratory Animal Technology (Beijing), China. db/db obese-diabetic mice (male, 7–8-weeks) and m/m nondiabetic mice (male, 7–8-week) were purchased from Cavens Laboratory Animal Technology (Changzhou), China. The three db/db mice (41.1, 40.5, and 46.6 g) and three m/m mice (26.5, 27.1, and 26.7 g) were weighed before execution. All the mice were housed at 18–24 °C with 40–70% humidity and a 14-h light/10-h dark cycle.

Human sample

Human blood samples from three healthy donors (male; aged 30, 21, and 27, respectively) were purchased from Oricells (Shanghai OriBiotech Co., Ltd). The human PBMCs were isolated from the blood sample using an isolation kit (Solarbio, Cat#P8610) according to the manufacturer’s protocol.

Molecular cloning

Molecular cloning operations were all carried out according to the manufacturer’s protocols. Primers were ordered from GENEWIZ Biotech (Suzhou). Human cDNA was synthesized by using Hifair® 1st strand cDNA synthesis superMix (YEASON, 11141ES) and human RNA extracted from HEK293T cells by TRIzol reagent. PCR was performed using Phanta Max Super-Fidelity DNA polymerase (Vazyme, P505-d1) in 50 µL reactions. In vitro, homologous recombination was carried out using Exnase MultiS (Vazyme, C113-02) in 20 µL reactions (with ~120 ng linearized vector and ~60 ng insert). Digestion of template was performed using DpnI (Thermo, ER1701) in Tango buffer.

In vitro labeling assays

Model protein labeling

Photocatalyst (5 μM) was combined with PAB-caged QM/thioQM probes (100 μM), NADH (500 μM), and bovine serum albumin (BSA, 1 mg/mL) in PBS (pH 7.4) to afford reaction mixtures with total solution volume of 100 µL. The samples were placed on a 96-well plate (100 μL/well) and irradiated by mild blue LED (450 nm, 4 mW/cm2) (unless otherwise noted) at room temperature for 5 min (unless otherwise noted). Eighty microliter of each sample was taken and combined with 20 µL 5× SDS-PAGE non-reducing loading buffer (Beyotime). Immunoblotting was performed with 4–15% gradient SDS-PAGE and 0.2 μm PVDF membrane. Streptavidin-HRP (Solarbio, SE068, 1:1,000 dilution) was used to detect the biotinylated bands. For validation of photocatalytic labeling by SF2, groups without Ir1 or without light irradiation were used as controls.

For UV-triggered labeling, SF2UV (100 μM) was combined with BSA (1 mg/mL) in PBS (pH 7.4) to afford reaction mixtures with a total solution volume of 100 µL. The samples were placed on a 96-well plate (100 μL/well), irradiated by UV light (CL-1000 UVP Ultraviolet Crosslinker) for 10 min at room temperature, and analyzed by immunoblotting as described above.

Identification of labeling site by LC–MS/MS

Protein labeling and isolation

The reaction was performed by combining Ir1 (10 μM), SF2-alkyne (100 μM), NADH (2 mM), and BSA (2 mg/mL) in PBS buffer (pH 7.4) with a total solution volume of 2 mL. The reaction system was placed in a 6-well plate and irradiated by mild blue LED (4 mW/cm2) for 15 min at room temperature. The BSA protein was then precipitated by mixing the solution with cold methanol and chloroform (solution:methanol:chloroform = 4:4:1, v/v/v), followed by >30 min of freezing at −80 °C. The sample was centrifuged (10,000 g, 20 min, 4 °C) and washed by cold methanol twice to yield a white pellet of protein.

Click reaction

The BSA pellet was redissolved in PBS containing 1.2% SDS (200 μL solution per sample) under sonication and heated to 90 °C for 5 min, followed by dilution with PBS to reach the point of 0.2% SDS. The solution was then mixed with 2 mM Cu-BTTAA (1:2, prepared by mixing CuSO4 and BTTAA in H2O), 3 mM sodium ascorbate, and 2 mM DADPS (cleavable biotin-azide, BBBD-19, Confluore Biological Technology). After the Click reaction at 37 °C for 1 h with shaking, the protein in solution was purified by methanol-chloroform precipitation protocol as described above.

Streptavidin enrichment

The protein pellet precipitated from the Click reaction was redissolved in 1.2% SDS-PBS and then diluted with PBS to reach the point of 0.2% SDS. High-capacity streptavidin resin beads (Thermo, 20357) were equilibrated by washing three times with 1 mL PBS. Bead slurry (100 μL) was added and the mixture was incubated for 3 h at room temperature with gentle rotation. After the incubation, the beads were pelleted by centrifugation (1,400 g, 3 min), washed by 0.2% SDS-PBS (5 mL, 10 min incubation) once, and further washed by PBS (5 mL) five times.

Thiol blocking and trypsin digestion

The protein-bound streptavidin beads were pelleted by centrifugation (1,400 g, 3 min) and the supernatant was removed, followed by the addition of 500 µL 6 M urea in PBS and 25 µL 200 mM dithiothreitol (DTT). The beads were resuspended by vortexing and incubated at 37 °C for 30 min in a ThermoMixer (Eppendorf). Twenty-five microliter of 400 mM iodoacetamide (IAA) was then added, and the resulting mixture was incubated again at 37 °C for 30 min in the dark. After incubation, beads were washed with 0.1 M trimethylammonium bicarbonate (TEAB) buffer (1 mL, four times) and pelleted. Two-molar urea–PBS buffer was added to the beads, followed by the addition of 2 µg trypsin. The beads were incubated at 37 °C for 16 h in a ThermoMixer to digest the proteins.

Isolation of modified peptides by acidic cleavage

After on-bead trypsinization, the supernatant was removed, and the beads were washed by PBS (1 mL, three times) and H2O (1 mL, three times). The beads were then incubated with 2% formic acid (200 µL) in a ThermoMixer for 1 h at room temperature, and the supernatant was collected. The procedure was repeated once. The beads were further washed with 50% MeCN/H2O (with 1% formic acid) twice, with the supernatant collected. All the collected supernatant fractions were combined and evaporated to dryness in a vacuum concentrator.

LC–MS/MS data acquisition and analysis

Peptides were dissolved in 0.1% formic acid and analyzed on Orbitrap Fusion Lumos LC–MS (Thermo) coupled to a Dionex Ultimate 3000 RPLC nano system (Thermo), with Thermo Xcalibur (v4.1.50) software. Samples were loaded onto a loading column (100 µm × 2 cm) and a C18 separating capillary column (75 µm × 15 cm) packed in-house with Luna 3-μm C18(2) bulk packing material (Phenomenex, USA). The HPLC solvent A was 0.1% formic acid in H2O, and the solvent B was 0.1% formic acid in 80% MeCN. The samples were run using a 120-min gradient method (0 min 2% B; 8 min 2% B; 9 min 10% B; 123 min 44% B; 128 min 99% B; 138 min 99% B; and 139 min 2% B) with a flow rate of 300 nL/min. The Fusion Lumos was operated in data-dependent mode with a cycle time of 3 s. The MS1 was performed with a full-scan m/z range from 350.00 to 1,600.00 and a mass resolution of 60,000. The maximum injection time was set to 50 ms with an AGC target of 4e5. The MS/MS fragmentation was performed using a quadruple isolation window of 1.6 m/z and the HCD collision mode (30% energy), with the orbitrap resolution set to 15,000, maximum injection time set to 30 ms, and AGC target set to 5e4. Collected data were analyzed using pFind (version 3.1.5) software. Mass spectra were searched against the BSA (UniProt ID P02769) protein sequence, with mass tolerance set to 20 ppm and max missing cleavage (by trypsin) number set to 3. The minimum peptide length was set to 7. Carbamidomethylation at Cys (+57.0215 Da) was set as a fixed modification. Oxidation at Met (+15.9949 Da) and putative SF2-QM-IAA (+509.2108 Da) modification at all nucleophilic residues (Lys, Arg, Ser, Thr, Tyr, Asp, Glu, Asn, Gln, Trp, Met, His, and Cys) were set as variable modifications. The putative arylazide-labeling (+575.2378 Da) modification was also searched. Identified peptides were manually reviewed. The Open Search mode was turned off.

Imaging of photocatalyst subcellular location

Approximately 10,000 HeLa cells were seeded in a LabTek-II 8-well glass chamber with 200 µL DMEM (+10% FBS) and incubated at 37 °C with 5% CO2 for 24–48 h. The medium was removed and the cells were rinsed with PBS. Ir1 (0.5 μM) was combined with MitoTrackerTM Deep Red (0.1 μM, Thermo) in 100 µL DMEM and the cells were incubated with the mixture for 30 min at 37 °C. After washing by PBS three times, the cells were immediately analyzed o an LSM 700 laser scanning confocal microscope (Zeiss). The fluorescent images were captured in the Hoechst channel (λex = 405 nm) for photocatalyst luminescence and the Cy5 channel (λex = 639 nm) for MitoTracker using a 63× oil lens. Processing of images were carried out using ZEN 3.2 blue edition (ZEISS). Pearson’s R values for colocalization were calculated for individual cells using the Coloc2 tool embedded in Fiji-Image J (v1.53k).

In cellulo CAT-S for miscellaneous assays

Photocatalytic labeling in living cells

Adherent cells

Cells (e.g., HeLa, HEK293T) were cultured to >90% confluence in a 12-well plate (approximately 106 cells/well). After removal of medium and washing with PBS (1 mL/well), cells were incubated with photocatalyst (Ir1 or other, 100 nM in 1 mL DMEM) for 30 min at 37 °C with 5% CO2, followed by washing with PBS once. Fresh DMEM (1 mL/well) was added, and the cells were incubated for 15 min at 37 °C with 5% CO2. After removal of the medium, a labeling probe (SF2 or other, 100 μM in 0.5 mL DMEM) was added and the cells were incubated for 30 min at 37 °C with 5% CO2, followed by mild blue LED irradiation (~4 mW/cm2, 12 min) at room temperature and incubation in the dark for 10 min at 37 °C. After further washing with PBS twice, the cells were ready for subsequent assays.

Suspension cells

The protocol was adjusted with additional transfer and centrifugation steps. Cells (e.g., K562) were cultured to near saturation in the T25 or T75 flask, with approximately 2 × 106 cells used for each sample. After removal of the medium by centrifugation (300–500 g, 3 min, 4 °C) in a tube, the cells were washed with PBS (1 mL/tube), followed by resuspension in RPMI-1640 containing photocatalyst (Ir1 or other, 50 nM in 1 mL medium. Note: 20–50 nM catalyst is sufficient for suspension cells, due to their higher uptake efficiency than adherent cells). The culture was transferred to a 12-well plate and incubated at 37 °C with 5% CO2 for 30 min (all the incubation steps herein were performed in a well plate or culture dish). Then the cells were transferred back to the tube for centrifugation to remove media, washed with PBS, and resuspended in RPMI-1640 (1 mL/sample), followed by incubation for 15 min at 37 °C with 5% CO2. After removal of the medium, the cells were resuspended in RPMI-1640 with a labeling probe (SF2 or other, 100 μM in 0.5 mL medium), incubated for 30 min at 37 °C with 5% CO2, followed by mild blue LED irradiation (~4 mW/cm2, 12 min) at room temperature and incubation in the dark for 10 min at 37 °C. After further washing with PBS twice, the cells were ready for the subsequent assays.

Immunoblotting assay

For sample preparation, adherent cells after photocatalytic labeling were directly scrapped from plate using RIPA strong lysis buffer (120 μL/well) (CWBIO, CW2333S) supplemented with 1× protease inhibitor cocktail (Bimake, B14001) and transferred to 1.5-mL tube, while suspension cells were pelleted by centrifugation (300–500 g, 3 min, 4 °C) and lysed in the same buffer. The samples were incubated on ice for 10 min, followed by the addition of methanol and chloroform (lysate:methanol:chloroform = 4:4:1, v/v/v) to precipitate proteins for purification. After cooling the mixture for >30 min at −80 °C, the proteins were pelleted by centrifugation (10,000 g, 20 min, 4 °C), carefully washed with cold methanol (200 μL) twice, and redissolved in 1% SDS-PBS buffer (80 μL). Five times SDS-PAGE loading buffer (20 μL) (Beyotime) was added to the sample, followed by heating at 95 °C for 15 min. Immunoblotting was performed with Tris-glycine SDS-PAGE and 0.2 μm PVDF membrane. Mouse anti-biotin mAb (Santa Cruz, sc-101339, 1:1,000 dilution) and rabbit anti-HSP60 mAb (Abcam, ab45134, 1:2,000 dilution) were used as primary antibodies, while HRP-linked anti-mouse IgG (Cell Signaling Technology, 7076S) and HRP-linked anti-rabbit IgG (Cell Signaling Technology, 7074S) were used as secondary antibodies (1:5,000 dilution).

Immunofluorescence assay

For photocatalytic labeling, cells (HeLa, HEK293T) were seeded on a LabTek-II 8-well glass chamber (pre-coated with Matrigel, Corning 356234), cultured to 50–80% confluency, and subjected to the CAT-S photocatalytic labeling procedure as described above. The incubation and washing steps were performed with a liquid volume of 200 μL/well. Upon completion of labeling, cells were washed with PBS (200 μL/well, twice), and fixed with 4% formaldehyde for 15 min at room temperature. After one PBS wash, cells were permeabilized by 0.2%Triton-X100 in PBS for 20 min, washed by PBST solution (0.1% Tween-20 in PBS pH 7.4) three times, and blocked by 3% BSA (PBST solution) for 1 h at room temperature. With three PBST washes between each incubation step, the cells were incubated with primary antibody rabbit anti-AIF mAb (Abcam, ab32516, 1:500 dilution) at 4 °C overnight, followed by incubation fluorescent antibodies goat anti-Rabbit-AlexaFluor546 (Invitrogen, A-11010, 1:500 dilution) combined with streptavidin-AlexaFluor488 (Invitrogen, S11223, 1:500 dilution) and Hoechst 33342 (2 μg/mL) at room temperature for 1.5 h. After three PBST washes, the samples were subjected to microscopy analysis on an LSM 700 laser scanning confocal microscope (Zeiss). The fluorescent images were captured in Hoechst channel (λex = 405 nm), AF488 channel (λex = 488 nm), and AF546 channel (λex = 555 nm). Processing of images were carried out using ZEN 3.2 blue edition (ZEISS).

Cell proliferation assay

The cells after photocatalytic labeling were harvested in the 1.5-mL tube (for adherent cells, Trypsin-EDTA digestion was performed for disassociation). The cells were pelleted by centrifugation (300–500 g, 3 min, 4 °C), and resuspended in PBS (1 mL). 20 μL cell suspension was added to 100 μL culture media (DMEM or RPMI-1640) in a well in 96-well plate, performed in triplicate. The cells were further cultured at 37 °C with 5% CO2 for 48 h before analysis. MTS assay was carried out using CellTiter 96 Aqueous One Solution kit (Promega, G3581) according to the manufacturer’s protocol. In brief, 10 μL of One Solution Reagent was added to each sample in a 96-well plate, and the plate was incubated at 37 °C with 5% CO2 for 1 h. The absorbance at 490 nm of each well was recorded on a Synergy H4 microplate reader (Bio-Tek). Controls were made by omitting photocatalysts (non-phototoxic group) during photocatalytic labeling. The proliferation data were normalized to the non-phototoxic control (100%) and medium (0%). GraphPad Prism 8.0 software was used to create the graphs and perform unpaired two-tailed Student’s t test.

MMP assay

The cells after photocatalytic labeling were harvested as described above. Samples omitting photocatalysts were used as healthy control, while samples with a five-fold concentration of Ir1 were used as phototoxic control. After resuspension in PBS (1 mL), 200 μL cell suspension was added to a well in a V-bottom 96-well plate. After pelleting by centrifugation (400 g, 3 min, 4 °C), the cells were resuspended in 200 μL culture medium containing 2 μg/mL JC-1 MMP probe (Bioss, D-9113) and incubated for 30 min at 37 °C with 5% CO2. Then the cells were washed with culture medium (200 μL) twice by centrifugation, resuspended in culture medium (200 μL), and immediately analyzed on BD LSRFortessa cell analyzer. JC-1 monomer was detected using the FITC channel, while JC-1 aggregate (forming in the presence of MMP) was detected using the PE channel. Data was further analyzed using FlowJo v10 software (FlowJo, Tree Star).

CAT-S mitochondrial proteomics for general cells

Photocatalytic labeling in living cells

This labeling protocol for proteomics was the scale-up version of the procedure described in “In cellulo CAT-S for miscellaneous assays”, and applied generally for living cells.

Adherent cells

Cells (e.g., HeLa, HEK293T) were cultured to >90% confluence in a 10-cm dish (approximately 107 cells per sample). After removal of medium and wash with PBS (5 mL per dish), the cells were incubated with Ir1 photocatalyst (100 nM in 10 mL DMEM, for HeLa and HEK293T) for 30 min at 37 °C with 5% CO2, followed by wash with PBS. Fresh DMEM (10 mL per dish) was added, and the cells were incubated for 15 min at 37 °C with 5% CO2. After removal of the culturing medium, an SF2 probe (100 μM in 5 mL DMEM) was added and the cells were incubated for 30 min at 37 °C with 5% CO2, followed by mild blue LED irradiation (~4 mW/cm2, 12 min) at room temperature and incubation in the dark for 10–15 min at 37 °C. After two washes with PBS (5 mL per dish), the cells were harvested using trypsin-EDTA (Gibco), resuspended in fresh medium (10 mL), and transferred to a 15-mL centrifuge tube placed on ice. The cells were washed by PBS, pelleted by centrifugation (400 g, 3 min, 4 °C) and placed on ice. The cell pellets could be either directly lysed or stored at −80 °C. An additional sample without a photocatalyst was used as a negative control.

Suspension cells

Cells (e.g., K562) were cultured to near saturation in a T75 flask, with 10–15 mL saturated culture used for each group (approximately 1–3 × 107 cells). After removal of the medium by centrifugation (300–500 g, 3 min, 4 °C) in a 15-mL tube, the cells were washed with PBS (10 mL per tube), followed by resuspension in RPMI-1640 containing Ir1 photocatalyst (50 nM in 10 mL medium. Note: 20–50 nM catalyst is sufficient for suspension cells, due to their higher uptake efficiency than adherent cells). The culture was transferred to a 10-cm dish and incubated at 37 °C with 5% CO2 for 30 min (all the incubation steps herein were performed in the culture dish). Then the cells were transferred back to the tube for centrifugation to remove media, washed with PBS, and resuspended in RPMI-1640 (10 mL per sample), followed by incubation for 10–15 min at 37 °C with 5% CO2 in a 10-cm dish. After removal of the medium by centrifugation, the cells were resuspended in RPMI-1640 with SF2 probe (100 μM in 5–10 mL medium), incubated for 30 min at 37 °C with 5% CO2 in a 10-cm dish, followed by mild blue LED irradiation (~4 mW/cm2, 12 min) at room temperature and incubation in the dark for 10 min at 37 °C. The cells were washed by PBS, pelleted by centrifugation, and placed on ice. The cell pellets could be either directly lysed or stored at −80 °C. An additional sample without a photocatalyst was used as a negative control.

Cell lysis and protein enrichment

For each labeling sample, 1 mL RIPA strong lysis buffer (pH 7.4, with 1% Triton X-100, 50 mM Tris, 150 mM NaCl, 1% sodium deoxycholate, etc.) (CWBIO, CW2333S) supplemented with 1× protease inhibitor cocktail (Bimake, B14001) was added to lyse the cell. The sample was sonicated to generate a clear lysate using Vibra Cell VCX Processor (SONICS). Then the protein was precipitated by the addition of methanol/chloroform (lysate:methanol:chloroform = 4:4:1) followed by >30 min of cooling at −80 °C. The proteins were pelleted by centrifugation (10,000 g, 20 min, 4 °C), and washed three times with ice-cold methanol (1 mL). The protein pellet was redissolved in 2% SDS-PBS (200 μL) under sonication and heated to 90 °C for 5 min. The sample was then diluted by PBS to the point of 0.2% SDS and transferred to a 15-mL centrifuge tube (Corning). A small fraction of the solution (20–100 μL) was taken from the sample, diluted to 100 μL, and used for BCA analysis against BSA standards, as well as immunoblotting quality control.

High-capacity streptavidin resin beads (Thermo, 20357) were equilibrated by washing three times with PBS (1 mL). Bead slurry was added according to the amount of the original cell lysate (normally 40 μL for 1.5–3 mg total protein). The mixture was incubated for 3 h at room temperature with gentle rotation. After the incubation, the beads were pelleted by centrifugation (1,400 g, 3 min), resuspended in PBS with 0.2% SDS, and incubated at room temperature for 10 min for washing. The beads were then washed with PBS (5 mL) for 6 times by resuspension and centrifugation.

Quality control

As a quality control measure, the remaining whole-cell lysate fraction (~80 μL) after BCA analysis was boiled at 95 °C for 20 min with 20 µL of 5× SDS-PAGE loading buffer added, to prepare “whole-cell” samples. Meanwhile, a fraction (~10%) of the post-enrichment beads was taken, resuspended in 80 µL PBS buffer, and boiled with 2 mM biotin and 20 µL of 5× SDS-PAGE loading buffer added, to prepare “pull-down” samples.

Immunoblotting was performed with 4–15% gradient Tris-glycine SDS-PAGE gel and 0.20 μm PVDF membrane. Mouse anti-biotin mAb (Santa Cruz, sc-101339, 1:1,000 dilution) and rabbit anti-HSP60 mAb (Abcam, ab45134, 1:2,000 dilution) were used as primary antibodies, while HRP-linked anti-mouse IgG (Cell Signaling Technology, 7076S) and HRP-linked anti-rabbit IgG (Cell Signaling Technology, 7074S) were used as secondary antibodies (1:5,000 dilution). Ruby staining of pull-down samples was performed using 4–15% gradient Tris-glycine SDS-PAGE gel, and SYPROTM Ruby protein gel stain kit (Invitrogen, S12001) according to the manufacturer’s instruction.

Protein digestion and dimethyl labeling

The protein-bound streptavidin beads were transferred to a 1.5-mL clear tube (Axygen, MCT-150-C) and pelleted, followed by the addition of 500 µL 6 M urea in PBS and 25 µL 200 mM DTT. The beads were resuspended by vortexing, and incubated at 37 °C for 30 min in a ThermoMixer (Eppendorf). Twenty-five microliter of 400 mM IAA was then added for thiol blocking, and the resulting mixture was incubated again at 37 °C for 30 min in the dark. After incubation, the beads were washed with 0.1 M TEAB buffer (1 mL, four times) and pelleted. After the addition of 0.1 M TEAB buffer (100 µL) and trypsin (1 µg), the beads were incubated at 37 °C for 16 h in a ThermoMixer to digest the proteins.

To each 100 µL of the digested peptides in 0.1 M TEAB buffer were added 4 µL of 0.6 M NaBH3CN and 4 µL of 4% (v/v) stable isotopic formaldehyde (normally CH2O for control (−) catalyst group and CD2O for experimental (+) catalyst group). The resulting mixture was vortexed immediately and incubated at room temperature for 30 min. After incubation, 16 µL of 1% (v/v) ammonia was added to quench the reaction, followed by the addition of 8 µL formic acid. After vortexing, heavy (CD2O)- and light (CH2O)-labeled samples were combined at a ratio of 1:1 (v/v) to generate duplex dimethyl-labeled samples, which were further desalted on C18 tips (Thermo, 87784) according to the manufacturer’s protocol and evaporated to dryness in a vacuum concentrator.

LC–MS/MS acquisition

Each biological replicate was analyzed as a single-shot sample. Peptides were dissolved in 0.1% formic acid and analyzed on Orbitrap Fusion Lumos LC–MS (Thermo) coupled to a Dionex Ultimate 3000 RPLC nanosystem (Thermo), with Thermo Xcalibur (v4.1.50) software. Samples were loaded onto a loading column (100 µm × 2 cm) and a C18 separating capillary column (75 µm × 15 cm) packed in-house with Luna 3-μm C18(2) bulk packing material (Phenomenex, USA). The HPLC solvent A was 0.1% formic acid in H2O, and the solvent B was 0.1% formic acid in 80% MeCN. The samples were run using a 120-min gradient method (0 min 2% B; 8 min 2% B; 9 min 10% B; 123 min 44% B; 128 min 99% B; 138 min 99% B; and 139 min 2% B) with a flow rate of 300 nL/min. The Fusion Lumos was operated in data-dependent mode with a cycle time of 3 s. The MS1 was performed with a full-scan m/z range from 350.00 to 1,600.00 and a mass resolution of 60,000. The maximum injection time was set to 50 ms with an AGC target of 4e5. The MS/MS fragmentation was performed using a quadruple isolation window of 1.6 m/z and the HCD collision mode (30% energy), with the orbitrap resolution set to 15,000, maximum injection time set to 30 ms, and AGC target set to 5e4.

Data analysis

All MS data were interpreted using MaxQuant v1.6.10 software. The quantification of heavy/light ratios was calculated with a mass tolerance of 20 ppm. For protein ID identification, MS/MS spectra were searched against a human UniProt database containing 20,378 proteins. The minimum peptide length was set to 7. Half-tryptic termini and up to two missing trypsin cleavages are allowed. Carbamidomethylation at cysteine (+57.0215 Da) and isotopic modifications (+28.0313 and +32.0564 Da for light and heavy labeling, respectively) at lysine/N-terminal were set as fixed modifications. Oxidation at methionine (+15.9949 Da), and acetylation of N-terminal (+42.0106 Da), were set as variable modifications. Each of the three biologically independent replicates was analyzed by MaxQuant separately. Contaminants and proteins identified as reverse hits were removed. Proteins detected with unique peptides <2 were also removed. Protein lists from replicates were combined, and only proteins detected over twice across biological triplicates were retained. Cutoff analysis based on the averaged heavy/light (experimental/control group, in which the control group was (−) catalyst) intensity ratio was further performed to filter off background noise (i.e., proteins with a ratio lower than the threshold were removed) while maximizing the retention of true positives. Proteins detected only in the (+) catalyst group were also retained. Taking mitochondrial specificity as the paramount criterion, a cutoff ratio was set to 5.0 for HeLa, HEK293T, and K562 experiments, based on the distribution profiles of proteins with and without mitochondrial annotation according to MitoCarta3.0 or UniProt (2022/10/13) database, yielding high-confidence protein lists with specificity level around 80%. GO enrichment analysis was performed by Metasacpe (https://metascape.org). Additional cutoff analysis with other criteria was also performed.

Discovery of mitochondrial proteins

Identification of “mito orphan” candidates

The HEK293T datasets (including log2(+/− catalyst) values from biological replicates for each protein) and combined datasets (including log2(+/− catalyst) values from HeLa, HEK293T, and K562 experiments for each protein) were z scored, and visualized by t-SNE analysis (T-distributed Stochastic Neighbor Embedding) in two-dimensional space, computed by t-SNE command from scikit-learn 1.1.3 Python package with perplexity set to 20 (Supplementary Software), available on GitHub at https://github.com/hefei8alex/CAT-S_scripts. According to the 2D plot, proteins “surrounded” by known mitochondrial proteins also without MAM (ER-mito contact site) localization evidence were picked as “mito orphan” candidates for further validation (listed in Supplementary Table 3).

Expression of orphan protein

HEK293T cells were seeded on a LabTek-II 8-well glass chamber (pre-coated with Matrigel, Corning 356234), and cultured to 20–60% confluency. Two hundred nanogram plasmid of C terminus V5-tagged orphan protein on a pcDNA3.1 vector was dissolved in 10 μL Opti-MEM and combined with 10 μL Opti-MEM containing 280 ng PEI. After incubation for 15 min, the mixture was combined with 180 μL DMEM (with 1% FBS), and incubated with the cells for 4 h for transfection. The media were then replaced by DMEM (with 10% FBS) and the cells were further cultured for 36 h before analysis.

Immunofluorescent imaging

After removal of media, the cells were washed by PBS, and fixed with 4% formaldehyde for 15 min at room temperature. After one PBS wash, cells were permeabilized by 0.2%Triton-X100 in PBS for 20 min, washed by PBST solution (0.1% Tween-20 in PBS pH 7.4) three times, and blocked by 3% BSA (PBST solution) for 1 h at room temperature. With three PBST washes between each incubation step, the cells were incubated with primary antibody rabbit anti-AIF mAb (Abcam, ab32516, 1:500 dilution) combined with mouse anti-V5 mAb (Biodragon, B1005, 1:1000 dilution) at 4 °C overnight, followed by incubation fluorescent antibodies goat anti-Rabbit-AlexaFluor488 (Invitrogen, A-11008, 1:500 dilution) combined with goat anti-Mouse-AlexaFluor555 (Invitrogen, A-21422, 1:500 dilution) and Hoechst 33342 (2 μg/mL) at room temperature for 1.5 h. After three PBST washes, the samples were subjected to microscopy analysis on an LSM 700 laser scanning confocal microscope (Zeiss). The fluorescent images were captured in the Hoechst channel (λex = 405 nm), AF488 channel (λex = 488 nm), and AF555 channel (λex = 555 nm). Processing of images were carried out using ZEN 3.2 blue edition (ZEISS). Pearson’s R values for colocalization were calculated for individual cells using the Coloc2 tool embedded in Fiji-Image J (v1.53k).

Mitochondrial proteomic profiling for living tissues

Processing of living tissue samples

Living tissues (kidney or spleen) were excised immediately after the mice (C57BL/6 J) were sacrificed. Each tissue/organ was transferred to a dish, washed by HBSS (or PBS), and cut into small pieces with media (RPMI-1640 with 10% FBS) added to prevent drying. The tissue pieces with media were then transferred to a 70-µm cell drainer fitted on a 50-mL tube, gently ground using a 2-mL syringe plunger, and pressed through the cell drainer, while media was continuously added during the process. The collected cells were pelleted by centrifugation (400 g, 3 min), resuspended in 0.3 mL HBSS buffer, and gently mixed with 1 mL RBC (red blood cell) lysis buffer (TIANGEN, RT122). The mixture was incubated at room temperature for 3 min to lyse red blood cells followed by the addition of 3 mL HBSS. The cells were then pelleted from the mixture by centrifugation, washed by HBSS (15 mL) once, and resuspended in media (10 mL) for subsequent applications.

CAT-S mitochondrial proteomics

Photocatalytic labeling

Tissues (kidney or spleen) disassociated as described above were employed for CAT-S mitochondrial proteomics. One mouse was normally used for two labeling samples (i.e., experimental and control group). Photocatalytic labeling was performed according to the general protocol for suspension cells (described in “CAT-S mitochondrial proteomics for general cells”), with the following optimizations: (1) all the incubation steps and irradiation step were performed in a 6-cm dish with 5 mL media due to the smaller sample size; (2) Ir1 concentration was optimized to 100 nM for kidney samples and 50 nM for spleen samples (due to the higher uptake efficiency of splenocytes), respectively.

MS sample preparation

Cell lysis, protein enrichment, enzymatic digestion, and isotopic labeling processes were all performed according to the general protocol described in “CAT-S mitochondrial proteomics for general cells”.

LC–MS/MS acquisition

Samples were run under the same condition (120-min gradient method on Orbitrap Fusion Lumos LC–MS) described in “CAT-S mitochondrial proteomics for general cells”.

Data analysis

The obtained MS data were interpreted using MaxQuant v1.6.10 software as described in “CAT-S mitochondrial proteomics for general cells”, searched against a mouse UniProt database containing 17,107 proteins. Contaminants and proteins identified as reverse hits were removed. Proteins detected with unique peptides <2 were also removed. Protein lists from replicates were combined, and only proteins detected over twice across biological triplicates were retained. Cutoff analysis based on the averaged heavy/light intensity ratio (reflecting spatial accessibility to photocatalyst) was further performed to filter off background noise (proteins with a ratio lower than threshold were removed). Proteins detected only in (+) catalyst group were also retained. Taking mitochondrial specificity as the paramount criterion, the cutoff ratio was optimized to 2.0 for the kidney and 8.0 for the spleen in order to obtain high mitochondrial specificity (>60%) protein lists, based on the distribution profiles of proteins with and without mitochondrial annotation according to MitoCarta 3.0 or UniProt (2022/10/13) database. GO enrichment analysis was performed by Metasacpe (https://metascape.org). Additional cutoff analysis with other common criteria was also performed (Supplementary Fig. 16).

For comparison and validation, the proteins detected only in post-cutoff kidney and spleen proteomes (i.e., “kidney-only” and “spleen-only”) were cross-analyzed with the published datasets40 (https://www.ebi.ac.uk/pride/archive/projects/PXD030062) of high-coverage mitochondrial proteomics for mouse tissues by Mann group, which contained six samples for each tissue. Proteins were mapped to their datasets by UniProt ID. The data were normalized in the same way as reported2, in which the intensity value of each protein was log2 transformed, and then subtracted from the mean log2(intensity) value of all proteins. Proteins missing in their datasets were set as log2(intensity) = 0 during the analysis. The normalization was performed for each biologically independent dataset separately. Mean normalized intensities of “kidney-only” proteome, “spleen-only” proteome as well as proteins grouped by GO terms were therefore calculated. For heatmap analysis, the data were further z scored for visual interpretation. GraphPad Prism 8.0 software was used to create the graphs (box-whisker plots and heatmap).

Comparative mitochondrial proteomics for diabetic and healthy mice

Processing of living tissue samples

Living kidneys from 10 to 11-week male obese-diabetic mice (db/db) and nondiabetic mice (m/m) were excised and processed as described in “CAT-S mitochondrial proteomics for living tissues”. The processed cell sample from each mouse was adjusted to a volume of 15 mL with RPMI-1640.

CAT-S mitochondrial proteomics

Photocatalytic labeling

Tem microliter processed samples from db/db mice and m/m mice were used for comparative proteomics. Photocatalytic labeling was performed according to the same protocol for the kidney (described in “CAT-S mitochondrial proteomics for living tissues”), using 100 nM Ir1 photocatalyst and 100 µM SF2 probe. A negative control (Ctrl) sample for further background subtraction was made by omitting the Ir1 photocatalyst during the labeling using a 1:1 mixture of 5 mL processed samples from db/db and m/m mice.

MS sample preparation

Cell lysis, protein enrichment, and enzymatic digestion processes were all performed according to the general protocol described in “CAT-S mitochondrial proteomics for general cells”. For dimethyl labeling, the three samples (diabetic group, nondiabetic group, and negative control) were subject to triplex stable isotopic labeling to yield heavy (13CD2O and NaBD3CN, for diabetic group), medium (CD2O and NaBH3CN, for nondiabetic group) and light (CH2O and NaBH3CN, for (−) catalyst control)-labeled samples, which were combined at a ratio of 1:1:1 (v/v/v) to generate a triplex dimethyl-labeled sample. The sample was further desalted on C18 tips (Thermo, 87784) according to the manufacturer’s protocol and evaporated to dryness in a vacuum concentrator.

LC–MS/MS acquisition

The triplex sample was analyzed on Orbitrap Fusion Lumos LC–MS (Thermo) according to the general protocol (described in “CAT-S mitochondrial proteomics for general cells”), using a 180-min gradient method (0 min 2% B; 8 min 2% B; 9 min 10% B; 183 min 44% B; 188 min 99% B; 198 min 99% B; and 199 min 2% B; flow rate, 300 nL/min) to enhance the performance of data acquisition.

Data analysis

The obtained MS data were interpreted using MaxQuant v1.6.10 software searched against a mouse UniProt database containing 17,107 proteins. The minimum peptide length was set to seven. Half-tryptic termini and up to two missing trypsin cleavages are allowed. The quantification of heavy/light, medium/light, or heavy/medium ratios was calculated with a mass tolerance of 20 ppm. Carbamidomethylation at cysteine (+57.0215 Da) and isotopic modifications (+28.0313, +32.0564, and +36.0757 Da for light, medium, and heavy labeling, respectively) at lysine N-terminal were set as fixed modifications. Oxidation at methionine (+15.9949 Da) and acetylation of N-terminal (+42.0106 Da) were set as variable modifications. Contaminants and proteins identified as reverse hits were removed. Proteins detected with unique peptides <2 were also removed.

For each biologically independent experiment, cutoff analysis was performed for diabetic and nondiabetic protein lists to filter off background noise (proteins with a ratio lower than the threshold were removed), based on the heavy/light (diabetic/Ctrl) and medium/light (nondiabetic/Ctrl) ratios, respectively. To maximize the retention of captured mitochondrial proteins, a less stringent cutoff ratio set to 1.2 was applied to remove obvious non-biotinylated proteins (i.e., Filter#1), yielding the “biotinylated” protein lists. Then, proteins detected in both diabetic and nondiabetic “biotinylated” lists were retained, while the others were removed (i.e., Filter#2). Lists of three biologically independent experiments were further intersected to generate the final proteomic list. Volcano plot was generated by using GraphPad Prism 8, based on the mean heavy/medium ratios and p values. GO enrichment analysis was performed by Metasacpe (https://metascape.org). PPI network analysis was performed by STRING, with only high-confidence (score > 0.7) PPIs displayed.

Immunoblotting analysis of protein candidates

Whole-cell lysates and mitochondrial lysates were taken as small fractions from the samples during MS sample preparation steps, measured by BCA quantification, and mixed with 5× SDS-PAGE loading buffer to a final concentration of 1×. Samples of nondiabetic and diabetic mice were adjusted to the same protein concentration using 1× SDS-PAGE loading buffer based on the BCA quantification. Immunoblotting was performed with 4–15% gradient Tris-glycine SDS-PAGE gel and 0.20 μm PVDF membrane. Rabbit anti-CPT1B pAb (Abcam, ab134988, 1:1,000 dilution), rabbit anti-ACSM2A (Abcam, ab181865, 1:2,000 dilution), and rabbit anti-Aldh3a2 (Abcam, ab184171, 1:2,000 dilution) were used as primary antibodies while HRP-linked anti-rabbit IgG (Cell Signaling Technology, 7074 S, 1:5,000 dilution) was used as the secondary antibody.

Mitochondrial proteomic profiling for human primary T cells

Isolation of primary T cells from human PBMCs

Primary T cells were isolated from human PBMCs using EasySepTM human T cell isolation kit (STEMCELL technologies, 17951) according to the manufacturer’s protocol. In our cases, approximately 4.6 × 107 human T cells with 99% purity (CD3-positive) were isolated per 108 PBMCs.

CAT-S mitochondrial proteomics

Photocatalytic labeling

Approximately 3 × 107 freshly isolated primary T cells were used for each labeling group (Note: since human primary T lymphocytes are small cells with 8–10 μm diameter, each labeling group contained <0.5 mg total cellular protein). Photocatalytic labeling was performed according to the general protocol for suspension cells (described in “CAT-S mitochondrial proteomics for general cells”) with the following optimizations: (1) all the incubation steps and irradiation step were performed in a 6-cm dish with 5 mL RPMI-1640 due to the smaller sample size; (2) Ir1 concentration was optimized to 20 nM to reduce background labeling noise. A group using equivalent cells without Ir1 photocatalyst was used as a negative control.

MS sample preparation

Cell lysis, protein enrichment, enzymatic digestion, and isotopic labeling processes were all performed according to the general protocol described in “CAT-S mitochondrial proteomics for general cells”.

LC–MS/MS acquisition

Samples were all run under the same condition (120-min gradient method on Orbitrap Fusion Lumos LC–MS) described in “CAT-S mitochondrial proteomics for general cells”.

Data analysis

The obtained MS data were interpreted using MaxQuant v1.6.10 software as described in “CAT-S mitochondrial proteomics for general cells”, searched against a human UniProt database containing 20,378 proteins. Contaminants and proteins identified as reverse hits were removed. Proteins detected with unique peptides <2 were also removed. For each donor’s sample, cutoff analysis based on the averaged heavy/light intensity ratio was further performed to filter off background noise (proteins with a ratio lower than the threshold were removed). Proteins detected only in the (+) catalyst group were retained. Taking mitochondrial specificity as the paramount criterion, the cutoff ratio was set to 7.0 for all donor’s samples based on the distribution profiles of proteins with and without mitochondrial annotation according to MitoCarta3.0 or UniProt (2022/10/13) database, yielding a final proteome with specificity around 60%.

Statistics and reproducibility

Statistical analysis methods of the data were detailed in the subsections above. All CAT-S proteomic experiments were performed in three biological replicates. All other experiments based on cells or animal samples were performed in ≥2 biological replicates (with replicate number n denoted in figure legends). No statistical method was used to predetermine the sample size. All samples and cells were randomly allocated into experimental groups. The investigator was blinded when grouping diabetic and nondiabetic mice, and blinded to group allocation when imaging “mito orphans” and controls. For other experiments, the investigators were not blinded to sample identity, since the data was from objective quantitative methods so subjective bias was not relevant. No data were excluded from the analyzes.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.