Bright and stable monomeric green fluorescent protein derived from StayGold.

Molecular cloning and mutagenesis The StayGold gene was synthesized using polymerase chain reaction (PCR) with overlapping primers listed in Supplementary Table 9 and cloned as a BglII/EcoRI fragment in the pBAD/HisB or pBAD/HisB-Sumo vectors. For PCR amplification, we used a C1000 Touch Thermal Cycler (Bio-Rad, USA). Random mutagenesis of StayGold gene was performed using PCR in the presence of Mn 2+ ions in the conditions to achieve 2–3 random mutations per 1000 bp according to the Diversify PCR Random Mutagenesis Kit User Manual (Clontech, USA). The mutagenized StayGold PCR fragment was further cloned in pWA21cBP-mKate2-mTagBFP vector at PstI/BglII restriction sites to swap mKate2 gene and get StayGold in frame with AraC DNA gene and the library was transformed in BW25113 bacterial cells. Next, the obtained bacterial library was spread on LB/agar Petri dishes supplemented with 100 μg/ml ampicillin, 0.02% arabinose and 0.02% rhamnose and after incubation for 24 h at 37°C and 2-4 h at room temperature about twenty thousand of bacterial colonies were imaged using fluorescent stereomicroscope Leica M205FA (Leica, Germany) in blue (405/40 nm excitation and 450/40 nm emission) and green channels (480/40 nm excitation and 535/40 nm emission). About 60-70 of the brightest green and dimmer blue fluorescing colonies were further analyzed on bacterial streaks. After each round the ten brightest green/dimmer blue fluorescent variants were further purified from 5 ml of LB medium supplemented with 100 μg/ml ampicillin and 0.004% rhamnose using Ni-NTA resin and their oligomeric state was assessed using FPLC chromatography. Directed mutagenesis of mBaoJin was performed using PCR with overlapping fragments. The PCR products after PCR with overlapping fragments were inserted into pBAD/HisB vector at BglII/EcoRI restriction sites. Mammalian plasmid construction In order to construct the pAAV- CAG -NES-GFPs-P2A-mCherry plasmids, the GFPs genes were PCR amplified as BglII-EcoRI fragments and swapped with the NCaMP7 gene in the pAAV- CAG -NES-NCaMP7-P2A-mCherry vector. In order to construct the pAAV- CAG -mCherry-P2A-GFPs plasmids, the GFPs genes were PCR amplified as SpeI-EcoRI fragments and swapped with the EGFP gene in the pAAV- CAG -mCherry-P2A-EGFP vector. In order to construct the pAAV- CAG -dMito-mBaoJin plasmid, the mBaoJin gene was PCR amplified as XhoI-EcoRI fragment and swapped with the mCherry gene in the pAAV- CAG-d Mito-mCherry vector. In order to construct the pAAV- CAG -H2B-mBaoJin plasmid, the mBaoJin gene was PCR amplified as BglII-HindIII fragment and swapped with the B-GECO1 gene in the pAAV- CAG -H2B-B-GECO1 vector. In order to construct the pLU-Vimentin-mBaoJin plasmid, the mBaoJin gene was PCR amplified as BamHI-BsrGI fragment using PCR with overlapping fragments to delete BsrGI restriction site and swapped with the NeonOxIrr gene in the pAAV- CAG -Vimentin-NeonOxIrr vector 17 . To construct the pEMTB-mBaoJin plasmid, the mBaoJin gene was PCR amplified as BamHI-NotI fragment and swapped with the mNeonGreen gene in the pEMTB-mNeonGreen vector (Addgene Plasmid #137802). In order to construct the pmBaoJin-Keratin plasmid, the mBaoJin gene was PCR amplified as KpnI-NotI fragment and swapped with the mKate2 gene in the pmKate2-Keratin vector (Evrogene, Moscow, Russia). In Piatkevich’s group, synthetic DNA oligonucleotides used for cloning were synthesized by Tsingke Biotechnology Co., Ltd. or Zhejiang Youkang Biological Technology Co., Ltd., China. PrimeStar Max master mix (Takara, Japan) was used for high-fidelity PCR amplifications. Restriction endonucleases were purchased from New England BioLabs (USA) and used according to the manufacturer’s protocols. DNA ligations were performed using OK Clon DNA Ligation Kit II from Accurate Biotechnology (Hunan) Co., Ltd, Changsha, China. The ligation products were chemically transformed into the TOP10 E. coli strain (Biomed, China) and cultured according to the standard protocols. Sequencing of bacterial colonies and purified plasmids were performed using Sanger sequencing (Zhejiang Youkang Biological Technology Co., Ltd., China). Small-scale isolation of plasmid DNA was performed with commercially available Mini-Prep kits (Tiangen, China); large-scale DNA plasmid purification was done with Midi-Prep kits (Qiagen, Germany). The gene of StayGold was de novo synthesized to substitute phiLOV3 in pAAV-CAG-phiLOV3-P2A-FusionRed plasmid by Tsingke Biotechnology Co., Ltd., China, based on the DNA sequences reported on Genbank ( https://www.ncbi.nlm.nih.gov/nuccore/2204333803 ). The genes of mBaoJin, td8ox2StayGold, AausFP1 and mClover3 were de novo synthesized by Synbiob Gene Technology Co., Ltd., China. To clone pAAV-CAG-mBaoJin-P2A-FusionRed plasmid, the synthesized DNA were PCR amplified with KpnI/AgeI flanking sites and swapped with the StayGold gene in the pAAV-CAG- StayGold-P2A-FusionRed. To clone pAAV-CAG-StayGold(E138D)-P2A-FusionRed plasmid, overlap PCR with site mutation of E138D was performed with KpnI/AgeI flanking sites and swapped with the StayGold gene in the pAAV-CAG-StayGold-P2A-FusionRed. To construct pAAV-CAG-GFPs-P2A-mCherry, the genes of FusionRed in the construct of pAAV-CAG-GFPs-P2A-FusionRed were swapped with the mCherry gene flanking with SpeI/EcoRI sites. To construct CytERM (cytoplasmic end of an endoplasmic reticulum signal anchor membrane protein) fusions, StayGold, StayGold(E138D) and mBaoJin were PCR amplified with AgeI/NotI flanking sites and swapped with the mScarlet gene in the pCytERM-mScarlet-N1 (Addgene plasmid #85066). To add more control of OSER assay, EGFP, Venus, td8ox2StayGold, mClover3, mEGFP, mGreenLantern were also PCR amplified with AgeI/NotI flanking sites and swapped with the mScarlet gene in the pCytERM-mScarlet-N1. To construct plasmids for expression of structural protein fusions, mBaoJin was PCR amplified and swapped with the corresponding FP genes in pActin-Electra1 (Addgene #184941), pEB3-mScarlet-I (Addgene plasmid #98826) and pTubulin-Electra1 (Addgene #184929) plasmids. For comparison on SIM, StayGold was PCR amplified and swapped with the Electra1 gene in pTubulin-Electra1 (Addgene #184929) plasmid. For long-term super-resolution imaging on SIM, mBaoJin, EGFP, and mNeonGreen were cloned into pLifeActin-N1 plasmid. The mammalian plasmids generated in the course of this study are available from the WeKwikGene plasmid repository at Westlake Laboratory, China ( https://wekwikgene.wllsb.edu.cn/ ).

Protein purification and characterization

For protein expression, the genes of proteins were PCR amplified as BglII/EcoRI fragments and inserted into the pBAD/HisB (Invitrogen, USA) or pBAD/HisB-Sumo vectors at the BglII/EcoRII restriction sites, and the generated plasmids were transformed into BW25113 bacteria. The bacterial cultures were grown in 200 mL of LB medium supplemented with 0.004% arabinose and 100 μg/ml ampicillin overnight at 37°C and 180 rpm. The cultures were then centrifuged at 4648 g for 10 min. The cell pellets were resuspended in PBS buffer supplemented with 10 mM Imidazole, 300 mM NaCl, lysozyme (100 μg/ml final concentration), lysed by sonication on ice, centrifuged at 20000 rpm for 10 min, 4 °C and further purified using Ni-NTA resin (Qiagen, USA) and dialyzed for 12–16 h against PBS buffer. For preparative mBaoJin protein purification from 2.6 L of LB medium, the BW25113 bacterial cells expressing the HisB-small ubiquitin-like modifier (SUMO)-mBaoJin protein were centrifuged for 20 min at 4648 g and 4 °C using the Avanti J-E centrifuge (Beckman Coulter, USA), cells were disrupted by sonication and the cell extract was centrifuged for 30 min at 28,000× g , 4 °C. The supernatant with protein was further purified using a 5 mL Ni-NTA Superflow column (Qiagen, Hilden, Germany). Next, His-SUMO-tag was cleaved using SUMO protease and the cleavage mix was applied to a Ni-NTA Superflow column (Qiagen, EU). The concentrated protein was further purified using a 1 mL ResourceQ column (GE Healthcare, Sweden) and finally concentrated until a 10 mg/mL concentration. Extinction coefficients were determined by alkaline denaturation method as performed in ref . 18 by addition of the 2M NaOH water solution to protein solution in PBS buffer till 1M final concentration and assuming that GFP-like chromophore has extinction coefficient equal to 44,000 M −1 cm −1 in 1 M NaOH 18 . The absorption spectra were registered using the NanoDrop 2000c spectrophotometer (Thermo Scientific, USA). Fluorescence quantum yields were determined by the absolute method using Quantaurus-QY multichannel detector (Hamamatsu, Japan). For fluorescence measurements, the samples were diluted to have optical densities less than 0.1. Fluorescence emission and excitation spectra were measured with an LS55 spectrofluorometer (Perkin Elmer) controlled by FLWinLab ™ software. The quantum yield was obtained at a set of excitation wavelengths across the absorption band nm with the step of 5 nm. In the cases where quantum yield depended on excitation wavelength because of overlapping absorption of the red (possibly anionic) and blue (possibly neutral) forms of chromophore, the quantum yield of the red form was calculated by averaging the numbers at the longest wavelength part, where the dependence reached a plateau. Each measurement was repeated 3 times and the average was accepted as final value. Quantum yields of purified proteins mBaoJin and mNeonGreen were determined in the PBS buffer by the tangent of the slope of the dependence of integral fluorescence values in the range of 480-700 nm excited at 470 nm. The EGFP protein with a quantum yield of 0.7 was used as a reference. Steady-state excitation and emission spectra were recorded using a CM2203 spectrofluorometer (Solar, Belarus) controlled by the “Universal” program software. Fluorescence lifetimes were measured on dilute solutions with a Digital Frequency Domain system ChronosDFD (ISS) appended to a PC1 (ISS, Champaigne, IL) spectrofluorometer. The excitation of a laser (ISS) at 445 nm, selected with a 440-460 nm filter, was modulated with 30 harmonics in the range of 5–150 MHz. The fluorescence was collected at 90° through a 535/50 filter. The modulation ratio and phase delay curves were fitted to model functions corresponding to a single-exponential fluorescence decay with Vinci 3 software (ISS). Coumarin 6 in ethanol (fluorescence lifetime 2.5 ns) was used as a reference standard for obtaining instrument response function. For determination of the maturation rates, GFPs were expressed in bacterial cells from the pBAD/HisB vector as previously described. 19 Briefly, protein expression was induced by adding 0.2% arabinose (final concentration) to 15 ml of overnight culture in a 15-ml tube filled to the brim and sealed tightly to prevent oxygen access. Protein expression lasted for 3 h at 37 °C, 220 rpm in no oxygen environment. Then cells were centrifuged at 3000 g, 10 min and the pallet was resuspended on ice in 800 μl of PBS buffer supplemented with chloramphenicol (10 μg/ml) to block protein translation and sonicated for 40 sec at 20% power using Sonics vibra cell VCX130 sonicator and CV18 tip (Sonics&Materials Inc., Newtown, CT, USA). Sonicated cells were centrifuged at 46,090 g for 2 min at 0°C. 100 μl of the supernatant was added to 2.9 ml of PBS buffer pre-warmed at 37°C and supplemented with chloramphenicol (10 μg/ml), and green fluorescence (480 ex/530 em; 5 nm slits) was recorded using a CM2203 spectrofluorometer (Solar, Belarus). Photostabilities of purified proteins (50 μM concentration) in PBS buffer were determined in microdroplets in mineral oil. 10 μl of oil was mixed with 1.5 ul of protein solution and placed between coverslips. Protein droplets were photobleached under continuous wide-field illumination using Zeiss Axio Imager Z2 microscope (Zeiss, Germany) equipped with a X-Cite 200DC XCT200 200 W mercury arc lamp (Lumen Dynamics, Canada), a 63x 1.4 NA oil immersion objective lens (PlanApo, Zeiss, Germany), a 470/40BP excitation filter, a FT 495 beam splitter, and 525/50BP emission filter or a Nikon Ti2-E widefield fluorescence microscope equipped with Spectra III Light Engine (LumenCore, USA), 470/28BP excitation filter (Semrock), the ORCA-Flash 4.0 V3 sCMOS camera (Hamamatsu, Japan), 40x NA1.15 objective lenses (Nikon, Japan). Light power densities were measured at the focal plane of the objective lens and reported for each experiment in figure legends. The acquired photobleaching curves were presented with no corrections and with normalization by the method introduced by Shaner et al. 8 For pH titrations in citric/borax buffer, 5 μl of proteins (till 50 nM final concentration) in PBS buffer were added to 200 ul of buffer (30 mM citric acid, 30 mM borax, and 30 mM NaCl) with pH values (adjusted with HCl or NaOH) ranging from 2 to 11 in 0.5 pH units interval in a 96-well black clear bottom plate (Thermo Scientific, USA). For pH titrations in citric acid/Na-citrate and monobasic/dibasic phosphate buffers, 5 μl of proteins (till 50 nM final concentration) in 50 mM sodium monobasic phosphate buffer titrated with NaOH to pH 7.20 were added to 200 ul of buffer A (100 mM citric acid titrated with NaOH to pH 2.5-6.0) and buffer B (100 mM NaH 2 PO 4 titrated with NaOH to pH 5.5-9.0) with pH values ranging from 2.5 to 9.0 in 0.5 pH units interval in a 96-well black clear bottom plate (Thermo Scientific, USA). To reveal NaCl influence on pKa values, 300 mM NaCl was also added to buffers A and B. After proteins incubation for 20 min at room temperature in described buffers, the fluorescence values were measured using a ModulusTM II Microplate Reader (TurnerBiosystems, USA) equipped with fluorescence optical kit ex 490/em 510-570. For comparison of stabilities of proteins in 6M guanidinium hydrochloride (GdnHCl), the 5 ul of proteins (till 50 nM final concentration) in PBS buffer were added to 200 ul of 30 mM HEPES, pH 7.80 buffer supplemented with 6M GdnHCl. After incubation of proteins at room temperature for 24 hours, the green fluorescence was registered using a ModulusTM II Microplate Reader (TurnerBiosystems, USA) equipped with fluorescence optical kit ex 490/em 510-570. The oligomeric state of the proteins was characterized using ÄKTA prime plus and ÄKTA explorer 100 systems (GE Healthcare, Sweden) and a Superdex 75 10/30 GL column (GE Healthcare, Chicago, IL, USA) equilibrated with 20 mM Tris-HCl, pH 7.5, 200 mM NaCl. Characterization in cultured mammalian cells HEK293FT (Invitrogen), HeLa (ATCC CCL-2), and Cos7 cells (ATCC CRL-1651) cells were authenticated by the manufacturer using STR profiling, reauthenticated in our lab by inspecting stereotypical morphological features under widefield microscope and tested negative for mycoplasma contamination to their standard levels of stringency. Authentication by morphology was performed every time before transient transfection. Cos7 cells were cultured in DMEM medium supplemented with 10% fetal bovine serum in an incubator at 37 °C with 5% CO 2 . HEK293FT and HeLa cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% Fetal Bovine Serum (FBS) and 1% penicillin/streptomycin (PS), and seeded in 24-glass bottom well plate (P24-0-N Cellvis, USA) or glass bottom dish (MatTek, USA) after Matrigel (356235, BD Biosciences, USA) coating. Cells were transfected at 80-90% confluency using Hieff Trans Liposomal Transfection Reagent (Yeasen Biotechnology, 40802ES02) according to manufacturer’s protocol, and imaged 24-48 h post transfection. Cell were imaged using a Nikon Ti2-E widefield fluorescence microscope equipped with Spectra III Light Engine (LumenCore, USA), the ORCA-Flash 4.0 V3 sCMOS camera (Hamamatsu, Japan), 10x NA0.45 and 20x NA0.75 objective lenses (Nikon, Japan) controlled by NIS-Elements AR 5.21.00 (Nikon Japan) or using a laser spinning-disk Andor XDi Technology Revolution multi-point confocal system (Andor Technology, Belfast, UK). To measure the brightness of mBaoJin in live cells under widefield microscope, HEK cells were transfected with the P2A coexpression plasmids and were imaged in FITC (excitation 475/28 nm from SpectraIII LumenCor; emission 594/40 nm) and TRITC (excitation 555/28 nm from SpectraIII LumenCor; emission 535/46 nm) channels. To obtain statistically significant datasets, we performed 2 independent transfections and ROIs were determined using auto-detect function of NIS Elements software limiting the ROI area to 50 μm 2 as a minimal size of HEK cells. The mean fluorescence intensity in FITC and TRITC channels for ROIs was extracted, and the FITC-to-TRITC ratio were calculated after background subtraction for each channel, which was used for the comparison of intracellular brightness under corresponding imaging conditions. The data were excluded from the analysis if cells were out of focus. To assess the brightness of mBaoJin in the cytosol of live cells under confocal microscope, HeLa cells cultured and seeded as described above were transfected with the NES-GFPs-P2A-mCherry or mCherry-P2A-GFPs plasmids using TurboFect transfection reagent (Thermo Scientific, Vilnius, Litva) and imaged in green (excitation 488 nm and emission 525/50 nm) and red (excitation 561 nm and emission 617/73 nm) channels using a laser spinning-disk Andor XDi Technology Revolution multi-point confocal system (Andor Technology, Belfast, UK) under identical imaging conditions (the same exposure times and laser power). To measure photostability in live HEK cells, HEK cells were transfected as described above and imaged under continuous FITC wide-field excitation. A 20x NA0.75 objective lens (Nikon, Japan) and SpectraIII LumenCor were used and set at 100% power. The illumination power at the focal spot was 50.5 mW/mm 2 . The photobleaching curves were calculated for each cell individually and reported as the mean photobleaching curve for each protein (averaged from all individual curves). Cells that detached or died during photobleaching experiments were excluded from data analysis. To evaluate the photostability under live-cell confocal imaging conditions, HeLa cells were transfected with pmBaoJin-Tubulin-N1 in glass-bottomed 24-well plates and imaged for one hour after 24 h transfection using CSU-W1 SoRa imaging setup of Nikon Spinning Disk Field Scanning Confocal System with 488 nm excitation using a 40× objective lens (power at object plane 5.7 mW/mm 2 ). To compare the brightness of mBaoJin with other proteins after PFA fixation, HEK cells were transfected as described above and imaged using a fluorescence wide-field microscope 36-48 h post-transfection. Cells were washed with PBS twice and fixed with 4% PFA (15714, Electron Microscopy Sciences, USA) in PBS at room temperature for 10 min. Fixed cells were gently washed with PBS twice and imaged under identical imaging settings for each protein. Image analysis was performed as described above for live cells. To obtain statistically significant datasets, 2 independent transfections were performed and ROIs were determined using the auto-detect function of NIS Elements software. To quantify monomeric state of GFPs in mammalian cells, OSER assay 4 was utilized as described before 20 . Briefly, two sets of experiments were conducted for different expression time length whereHeLa (ATCC CCL-2) cells were cultured and transfected with the corresponding plasmids. In the first set, 3 independent experiments were performed. Cells were seeded on a 12-well glass bottom plate after Matrigel (356235, BD Biosciences) coating for 30-45 min at 37°C. Cells were then transfectedat 70-90% confluency with 1 μgDNA per well using Hieff Trans Liposomal Transfection Reagent (Yeasen; 40802ES02) according to the manufacture’s protocol. Cells were imaged 12 h post-transfection using FITC channel (Nikon CSU-W1 SoRa; OBIS 488nm LX 100 mW laser; ORCA-FusionBT; 40x WI or 100xOIL; 525/50 filter). In the second set, cells were seeded on 24-well glass bottom plate after Matrigel (356235, BD Biosciences) coating and 3 independent transfections were performed with 500 ng DNA per well at 70-90 % confluency according to manufacturer’s instructions (Hieff Trans Liposomal Transfection Reagent; Yeasen; 40808ES03). Cells were imaged 22 h post-transfection with FITC channel. To collect enough cells from each well, large image mode and automated image stitching were used with 10% overlap in NIS elements software. Positive cells selected for analysis had overall similar fluorescence brightness, and cells that were significantly brighter were excluded (indications of unhealthy or highly stressed cells). Cells with non-spherical nuclei, ER sheet architectures, or condensed nuclei were regarded as stressed and excluded from normal cells fraction quantification. Cell cultures with more than 25% of stressed cells were excluded from quantification of normal cells fraction.

Super-resolution BaLM imaging of the cytoskeleton of cultured mammalian cells

Immediately before imaging, the cell medium was changed to a minimal essential medium (MEM, Sigma-Aldrich) supplemented with 20 mM HEPES. Single-molecule localization super-resolution imaging of living cells was performed using Nanoimager S (ONI, UK) microscope, equipped with Olympus UPlanSApo x100 NA 1.40 oil immersion lens, 488 nm laser, 560 nm on-camera beam splitter and Scope8 sCMOS camera. Imaging was performed using three imaging condition sets: 475 W cm −2 488 nm laser and 50 ms frame time (20 fps acquisition speed), 600 W cm −2 488 nm laser and 10 ms frame time (100 fps acquisition speed), 1.8 kW cm −2 488 nm laser, 5 ms frame time (200 fps acquisition speed). The difference between the signal-to-noise ratio of mBaoJin and mNeonGreen localizations was tested using the Kolmogorov–Smirnov test. Image acquisition and super-resolution reconstruction were performed using NimOS 1.18.3.15066 (ONI, UK). Image reconstruction was performed using default parameters. Data analysis was performed using FiJi ImageJ 1.53f51 (ref. 21 ) and custom Python 3.9 scripts. For blinking duration calculation, the “Tracking” tool of NimOS was used. Spatial resolution was calculated using decorrelation analysis 22 with default parameters. As a widefield image, the standard deviation of images stack was used. Super-resolution HIS-SIM imaging of nuclearpores and actin filaments in Cos7 cells The procedure for imaging was performed by following the previous report 23 . Briefly, Cos7 cells, cultured as described above, were seeded in in 12 well plate and transfected using Lipofectamine 2000 (ThermoFisher Scientific) according to the manufacturer’s instructions with 500 ng of the EGFP-Lifeact/mNeonGreen-Lifeact/mStaygold-Lifeact/ mNeonGreen-Nup98/ mStaygold-Nup98 plasmid per well. After 24 hours of transfection, cells were replated on coverslips (H-LAF 10L glass; reflection index, 1.788; diameter, 26 mm; thickness, 0.15 mm, customized) coated with 10% poly-l-lysine solution (Sigma) for ~24 h before seeding transfected cells. Live cells were imaged in HBSS (Gibco) at 37°C in a humidified live cell imaging workstation for live SIM imaging. Super-resolution imaging of nuclearpores and actin filaments were performed using commercialized HIS-SIM (High Intelligent and Sensitive Structured Illumination Microscope) provided by Guangzhou Computational Super-resolution Biotech Co., Ltd. Images were acquired using a 100×/1.5 NA oil immersion objective (Olympus) using 488 nm laser. To further improve the resolution and contrast in reconstructed images, sparse deconvolution was used by the previous report 13 . The software of HIS-SIM, IMAGER, was used for data collection and export. Data analysis was performed using Fiji 2.9.01/1.53t ImageJ.

Characterization of GFPs in mouse brain tissue

All animal maintenance and experimental procedures for mice were conducted according to the Westlake University Animal care guidelines, and all animal studies were approved by the IACUC of Westlake University, Hangzhou, China under animal protocol #19-044-KP-2. Procedures involving experimental animals are reported in accordance with Animal Research: Reporting of In Vivo Experiments (ARRIVE). Mice were maintained at strict barrier facilities with macroenvironmental temperature and humidity ranges of 20–26 °C and 40–70%, respectively. Food and water were provided ad libitum. The rooms had a 12 h light–12 h dark cycle. The housing conditions were closely monitored and controlled. C57BL/6 mice supplied by the Animal Facility at Westlake University were used without regard for sex. For expression of GFPs in the brain, the neonatal intraventricular injections of custom-made recombinant AAVs, serotype 2/9 (rAAV2/9-CAG-GFPs-P2A-FusionRed, titer: >10 12 v.g. ml −1 ; Shanghai Sunbio Medical Biotechnology,China) as described previously 20 . In brief, viruses were injected pan-cortically into pups at postnatal day 0 regardless of sex with a Hamilton microliter syringe. For each hemisphere, 0.5 μl virus solution (titer: >10 12 v.g. ml −1 ) supplemented with 0.1% FastGreen dye (Sigma-Aldrich) was injected under the skull manually. After injections in both hemispheres, the pups were returned to the home cages. The whole injection process for each pup was controlled within a few minutes to prevent cannibalism. For assessment of intracellular brightness, acute brain slices were obtained from 7-week old mice and at least 10 FOVs were imaged using Nikon wide-field microscope described above. To assess the performance of GFPs in fixed tissue and ExM, mice were deeply anesthetized and perfused using 4% PFA and at least two 50-μm brain slices from two mice at postnatal days 50 were imaged and processed. Fixed brain tissue samples were imaged using confocal microscope Zeiss LSM 980 (Germany; 10x NA 0.45, 40x NA 0.95) controlled by ZEN blue v3.5. For measuring the photostability of GFPs in fixed brain slice, two brain slices from two mice for each protein were imaged under continuous FITC wide-field excitation. A 20x NA0.75 objective lens (Nikon, Japan) and SpectraIII LumenCor were used and set at 50% power. The illumination power at the focal spot was 25.9 mW/mm 2 . The photobleaching curves were calculated for each neuron individually and reported as the mean photobleaching curve for each protein (averaged from all individual curves). Characterization of GFPs in C. elegans neurons The FPs were expressed using extrachromosomal arrays in C. elegans and assessed for intracellular localization, brightness and photostability in the green channel. For expression in C. elegans , the target genes were codon-optimized using C. elegans Codon Adapter application ( https://worm.mpi-cbg.de/codons/cgi-bin/optimize.py ) with insertion of two introns 24 and de novo synthesized (Tsingke Biotechnology Co., Ltd, China). The codon optimized genes of GFPs were cloned into pSF11 vector (Addgene plasmid #179485) under the pan-neuronal tag-168 promoter. The establishment of transgenic lines was done by SunyBiotech Co. Ltd (China) according to standard protocols. Briefly, wild-type N2 worms were co-injected with two plasmids encoding wGFPs::wmTagBFP2 with 10 ng/μl final concentration each. wmScarlet (Psur-5::sur-5::NLSwmScarlet) was used as a reporter marker for screening positive worms (provided by SunnyBiotech Co. Ltd, China). Transgenic lines were selected by green fluorescence and confirmed with sequencing of the target. Selected worms were maintained and grown on nematode growth medium (NGM) plates seeded with E. coli OP50-1 at 20°C following standard protocols. Positive worms with green fluorescence (used without regard to sex) at L4 stage of development were selected for further imaging. Worms were mounted on immunofluorescence assay slides (Jiangsu Shitai Zhenduan Jishu Co. Ltd, 80383-0209-01), immobilized with 25 mM levamisole and imaged using a wide-field Nikon microscope with 10x NA0.45 objective lens for quantification and confocal microscope Zeiss LSM 980 for structural imaging (Germany; 10x NA 0.45, 40x NA 0.95) controlled by ZEN blue v3.5. Individual neurons were selected manually, and mean fluorescence intensity was calculated using Nikon Elements software.

Expansion Microscopy of HEK cells and brain tissue

HEK293FT and HeLa cells, cultured as described above, were seeded onto Matrigel treated (356235, BD Biosciences, USA) glass-bottom dishes (MatTek, USA) and transfected at 40-50% confluency with the pAAV-CAG-GFPs-P2A-mChilada plasmids (mChilada is a new RFP described in the upcoming pre-print, gift from Nathan Shaner) and mixture of pN1-mBaoJin-mito and pN1-mTurquise2-H2B plasmids, respectively, using Lipofectamine 3000 (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. After 36-48 hours, cells were briefly rinsed with 1x PBS and fixed for 10 min with 4% PFA (Vendor) and imaged before undergoing ExM procedure. ExM was performed according to the optimized protocol called Octa-ExM 25 . Briefly, PFA-fixed mice brain slices and cultured mammalian cells were treated with 5% allyl glycidyl ether (TCI, Japan) dissolved in 100 mM NaHCO 3 for 2 hours at 37°C, followed by 3 brief washes with 1xPBS. The Stock8X monomer solution (8.624% (w/v) sodium methacrylate, 31.37% (w/v) N,N-dimethylacrylamide, 0.385% (v/w) trimethylolpropane propoxylate triacrylate, all from Sigma-Aldrich, United States) was mixed in water, adjusted to pH=6.5 with HCl, and stored at 4°C before use. Activated monomer solution were prepared by adding freshly prepared 10% (w/v) solution of ammonium persulfate (APS) initiator and tetramethyl-ethylenediamine (TEMED) accelerator (Sigma-Aldrich, United States) into Stock8x monomer solution to final concentration of 0.3% (w/v) and 0.2% (w/v), respectively. Pre-treated samples were place into glass-bottom dish (MatTek, USA) and incubated in activated Stock8x monomer solution for 30 minutes for brain sections and 10 minutes for cell culture at 4°C. Gelation was carried out in an anaerobic and humid environment for 3 hours for brain sections and for 2 hours for cell cultures at 37°C. The fully formed gel were then submerged with homogenization buffer containing 8 U/ml proteinase K (NEB, United States or YEASON, China). Sample were homogenized overnight at RT (brain sections) or for 4 hours at 37°C (cell cultures). Homogenized sample-hydrogel composites were expanded in doubly deionized water for 4-5 h and imaged. Protein crystallization An initial crystallization screening of mBaoJin was performed with a robotic crystallization system (Rigaku, Woodlands, TX, USA) and commercially available 96-well crystallization screens (Hampton Research, Aliso Viejo, CA, USA and Anatrace, Maumee, OH, USA) at 15°C using the sitting drop vapor diffusion method. The protein concentration was 8.3 mg/mL in the following buffer: 50 mM Tris, 150 mM NaCl pH 7.5. The initial conditions were optimized by the sitting-drop vapor-diffusion method in 24-4 Intelli plates. The crystals were obtained within several days under the following conditions: 28% PEG3350, 0.2 M lithium sulfate, 0.1 M sodium acetate pH 4.6; 30% PEG8000, 0.1M ammonium acetate, 0.1M sodium cocadilate pH 6.5, and 27% PEG 3350, 0.2M lithium sulfate, 0.1 M Tris-Cl pH 8.5. Data collection, structure solution, and refinement Crystals were briefly soaked in a cryosolution containing precipitant supplemented with 20% Glycerol (Hampton Research, Aliso Viejo, CA, USA) immediately prior to diffraction data collection and flash-frozen in liquid nitrogen. The X-ray data were collected using 0.97861 Å wavelength from a single crystal (pH 6.5) at 100 K at beamline BL19U1 of the National Facility for Protein Science Shanghai at Shanghai Synchrotron (Shanghai, China). The X-ray data for other crystals (pH 4.6 and 8.5) were collected using 1.54 Å wavelength at 150K at Regaku XtalLAB Synergy-S laboratory system (The Woodlands, Texas, USA). The data were indexed, integrated, and scaled using the XDS program 26 ( Supplementary Table 10 ). Structure of mBaoJin crystallized at pH 6.5 (PDB ID: 8Q79) was solved using MOLREP program 27 with a subunit A of pdb 8BXT as a model 6 [ https://www.rcsb.org/structure/8BXT ]. For other structures (PDB ID: 8QBJ, 8QDD) mBaoJin-pH6.5 was used as a model. Refinement was carried out using the REFMAC5 program of the CCP4 suite 28 . The visual inspection of electron density maps and the manual rebuilding of the model were carried out using the COOT interactive graphics program 29 . The hydrogen atoms in fixed positions were introduced during the refinement. In the final model, an asymmetric unit of every structure contained two independent copies of the protein with chromophores and solvent molecules. The mBaoJin structures were deposited in the Protein Data Bank 30 with PDB ID accession codes 8QBJ (pH4.6), 8Q79 (pH 6.5) and 8QDD (pH 8.5). Ramachandran statistics showed 98% (mBaoJin pH4.6), 99 % (mBaoJin pH6.5) and 98% (mBaoJin pH8.5) of residues in the most favored region and 2% (mBaoJin pH4.6), 1% (mBaoJin pH6.5), and 2% (mBaoJin pH8.5) in the allowed regions. Data analysis, image processing, statistics, and data visualization To estimate the significance of the difference between two values, we used the Mann–Whitney Rank Sum Test and provided p -values calculated for the two-tailed hypothesis. We considered the difference as significant if the p value was < 0.05. The mean fluorescence intensity in the green and red channels for ROIs corresponding to the cytosol region was extracted, and after background subtraction for each channel, the green-to-red ratio was calculated and used to compare intracellular brightness. Fluorescence brightness and photostability data were analyzed offline using NIS-Elements Advance Research software (v5.21.00; Nikon Japan), Origin (OriginLab 2019b), Excel (Microsoft Office Professional Plus 2021). Cells were selected automatically using built in function of NIS-Elements and background regions were selected manually and fluorescence measurements were performed for each region of interest (ROI), and then fluorescence from an background region was subtracted from cell fluorescence to correct for background, normalized in Excel and ploted in Origin. The raw data for 2D-SIM images were acquired using Hessian-SIM, and subsequent image reconstruction was performed using Wiener deconvolution in HiS-SIM IMAGER software (version v1.2.3.d). The Wiener deconvolution utilized North filter with a 30-pixel setting and Defocus elimination set to 0.5. Sparse-SIM, as an additional processing step, involved Sparse deconvolution based on the 2D-SIM results. This step was carried out using MicroscopeX FINER software (version v1.0.9b). For actin, the sparse parameters included sample type (-actin), sparse level (level1), and specific parameters for Nuclearpore, such as sparse iteration times (200), image fidelity (800), t axial continuity (0.1), sparsity (10), deblurring method (Landweber), and oversampling method (Fourier Sampling). Boxplots with notches are used in Fig. 2 , 4 , 5 : narrow part of notch, median; top and bottom of the notch, 95% confidence interval for the median; top and bottom horizontal lines, 25% and 75% percentiles for the data; whiskers extend 1.5 × the interquartile range from the 25th and 75th percentiles; horizontal line, mean; outliers not shown but included in all calculations and available in the source datasets.

Supplementary Material Supplementary Information Suppl video 1 Suppl video 2 Supple video 3 Suppl video 5 Suppl vide 4 Suppl video 6 Suppl video 7 Suppl video 8 Suppl video 9 Suppl video 10 Suppl video 11 Suppl video 12 Suppl video 13 Suppl video 14 Source data Fig 1 Source data Fig 2 Source data Fig 3 Source data Fig 4 Source data Fig 5