Animals

All animal procedures were approved by the National Institutes of Natural Sciences Animal Care and Use Committee and were performed under the relevant guidelines and regulations. All dissociated cultures were prepared from C57BL/6N mice (SLC). All slice cultures were prepared from Wistar rats (SLC or Charles River). This study used dissociated and slice cultures from both male and female pups. For in vivo imaging, the experiments were performed using male and female adult (2 months) C57BL/6N mice (Charles River).

Reagents

Anisomycin and N-methyl-D-aspartic acid (NMDA) were purchased from Sigma-Aldrich (St. Louis, MO, USA). AAV1-CaMKII-Cre was purchased from the University of Pennsylvania (U Penn; Philadelphia, PA, USA). Cycloheximide was from Abcam (Cambridge, UK). MNI-caged L-glutamate (MNI-caged glutamate) was from Tocris Bioscience (Bristol, UK).

Plasmids

Plasmids containing CaMKIIα, Cdc42/RhoA/Pak, SEP-GluA1/2, Rhotekin, WPRE, and CaMKII0.4 promoter genes are gifts from Y. Hayashi, M. Matsuda, S. Soderling, G. Bokock, K. Kobayashi, and M. Ehlers, respectively. GCaMP6f, Clover, LOV2(SD), CAMKIIβ, Cre, WPRE3 genes, and pAAV-hSyn-DIO-EGFP plasmid were gifts from D. Kim, M. Lin, K. Hahn, T. Meyer, C. Cepko, B.K. Kaang, and B. Roth (Addgene plasmid #40755, #40255, #81033, #21227, #13775, #61463, #50457), respectively. The synthesized gene encoding codon-optimized LOV2_408–546 gene was purchased from Integrated DNA Technologies (Coralville, IO, USA). pAAV-RC-DJ (AAV2/DJ) and pAAV-MCS/pAAV-Helper were purchased from Cell Biolabs (San Diego, CA, USA) and Agilent Technologies (Santa Clara, CA, USA), respectively.

paCaMKIIα was constructed by inserting the codon-optimized LOV2_408–546 sequence between the kinase and association domains of rat CaMKIIα as described in the main text. The introduction of mutations was carried out by using the QuikChange Site-Directed Mutagenesis Kit (Agilent Technologies). The CMV-tdTomato-P2A-paCaMKIIα plasmid was constructed by inserting tdTomato⁶³ and paCaMKIIα, together with P2A⁶⁴ sequences ATNFSLLKQAGDVEENPGP into the modified pEGFP-C1 vector by replacing EGFP.

The CMV-mEGFP-paCaMKIIα and CMV-Flag-paCaMKIIα plasmids were constructed by inserting mEGFP (EGFP with A206K monomeric mutation)⁶⁵ or Flag with paCaMKIIα into the modified pEGFP-C1 vector by replacing EGFP. The CMV-mEGFP-paCaMKIIα-ShadowG was constructed by inserting mEGFP, paCaMKIIα, and ShadowG⁶⁶ into the modified pEGFP-C1 vector by replacing with EGFP.

For construction of the Cdc42 FRET sensor (CMV-ShadowY-CBD-P2A-Clover_T154M/F223R-Cdc42), we fused ShadowY (a.a. residues 1–232)⁵⁰ as an acceptor fluorescent protein to the N-terminus of the activated CBD of Pak3 (a.a. residues 60–113 with two mutations: S74A and F84A)⁴⁸ via the linker peptide RSRG. Subsequently, Cdc42 fused to Clover_T154M/F223R as a donor fluorescent protein via the linker peptide SGLRSRG was fused to the C-terminus of the acceptor protein via the P2A sequence so that the CBD and Cdc42 parts were translated into different polypeptides within the cell.

The RhoA FRET sensor (CMV-ShadowY-RBD-P2A-Clover_T154M/F223R-RhoA) was constructed similarly to the Cdc42 FRET sensor using an activated RhoA-binding domain (RBD) of rhotekin (a.a. residues 8–89).

The following plasmids pAAV-CaMP0.4-DIO-Clover-WPRE3 and pAAV-CaMP0.4-Flag-His×6-paCaMKIIα were constructed by inserting the respective components in the pAAV-MCS.

shRNAs were prepared using the custom pSuper vector (Oligoengine). The following target sequences were used for shRNA (5′-3′): CAGTCACAGTTATGATTGGT for Cdc42 (rat, mouse, and human), and ATCGATCGAAAGATCGCTC for control. shRNA-resistant Clover-Cdc42 was prepared by introducing six silent mutations in the targeted sequences.

AAV production and purification

Serotype DJ AAVs were produced and purified as described previously⁶⁷. Briefly, HEK293 cell culture was maintained on 15-cm plates in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% fetal bovine serum (FBS) with no antibiotics at 37 °C and 5% CO₂. Six 15-cm dishes at 70% confluency were prepared for polyethylenimine (PEI) transfection. For PEI-based transfection, plasmids at a ratio of 1:1.6:1 (45 µg of a transgene in pAAV, 72 µg of pAAV-Helper, and 45 µg of pAAV-RC-DJ [AAV2/DJ]) were used for six dishes. The PEI–DNA complex at the ratio of 4:1 (w/w) (PEI “MAX” [648 µg, Cosmo Bio; Tokyo, Japan] and DNA [162 µg]) was incubated in 6 ml of serum-free DMEM for 10 min at RT, making DMEM–PEI–DNA mixture. Subsequently, the mixture was diluted with 150 ml of serum-free DMEM. The culture medium in the dishes was decanted and replaced with the DMEM–PEI–DNA mix (25 ml per dish), subsequently, the dishes were incubated at 37 °C and 5% CO₂ for 96 h. Before the collection of supernatants, NaCl was added to 500 mM to increase the solubility of AAV particles and further incubated 30 min at 35 °C and 5% CO₂.

The collected culture medium (~150 ml) was centrifuged at 4000 × g for 20 min to remove cell debris, and the supernatant was filtered through a 0.22 µm pore. The clarified supernatant containing AAV was concentrated by the cross-flow cassette (Vivaflow 50, 100,000 MWCO, Sartorius; Göttingen, Germany) to about 15 ml. The AAV solution was further concentrated by Amicon Ultra-15 (100,000 MWCO, Merck; Kenilworth, NJ, USA) to about 4 ml, and incubated with benzonase at 250 U/ml at 37 °C for 1 h. Iodixanol step gradients were performed as described by Addgene (homepage section: AAV Purification by Iodixanol Gradient Ultracentrifugation). The buffer solution of the virus was exchanged with phosphate-buffered saline (PBS) during the concentration process.

The titer of AAVs was determined by quantitative PCR (qPCR) using THUNDERBIRD qPCR Mix (Toyobo; Osaka, Japan) and LightCycler 96 (Roche; Basel, Switzerland) according to the manufacturers’ protocols. The primers used for qPCR were as follows: 5′-acctctggattacaaaatttgtgaaag-3′ and 5′-aaccaggatttatacaaggaggagaaaatg-3′, which anneal to both WPRE3 and WPRE. The resultant virus titers typically ranged between 2 × 10⁹ and 2 × 10¹⁰ genome copies/µl in a total volume of 400 μl.

Purification of tdTomato and the estimation of tdTomato concentration in neurons

His-tagged tdTomato was inserted into pRSET bacterial expression vector (Invitrogen; Carlsbad, CA, USA). Protein was overexpressed in Escherichia coli (DH5α) and purified with a Ni⁺-nitrilotriacetate column (HiTrap, GE Healthcare; Chicago, IL, USA), and desalted with a desalting column (PD10, GE Healthcare) equilibrated with PBS. The concentration of the purified protein was measured by the absorbance of the fluorophore (tdTomato, A₅₅₄ = 138,000 cm⁻¹ M⁻¹)⁶³. The concentration of tdTomato in neurons was estimated by measuring the fluorescence intensity of tdTomato in thick apical dendrites (3–6 µm in diameter) relative to that of purified tdTomato (10 µM) under a two-photon microscope.

HeLa cell culture and transfection

HeLa cells were cultured in DMEM supplemented with 5% FBS at 37 °C in 5% CO₂. The cells in 3-cm dishes were transfected with the plasmids using Lipofectamine 3000 (Invitrogen), followed by incubation for 16–22 h in the absence of serum. 2pFLIM-FRET imaging was conducted in HEPES-buffered artificial cerebrospinal fluid (HACSF; 30 mM HEPES, 130 mM NaCl, 2.5 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 1.25 mM NaH₂PO₄, 25 mM glucose, pH 7.3) at 24–26 °C.

Primary neuronal culture and AAV infection

Low-density cultures of dissociated embryonic cortical and hippocampal neurons were prepared as described previously²³. Briefly, hippocampi or cortices were removed from C57BL/6N mice at embryonic days 18 and treated with papain for 10 min at 37 °C, followed by gentle trituration. Mouse cortical or hippocampal neurons were seeded onto PEI-coated 3-cm dishes (2 × 10⁵ cells/dish) and cultured in neurobasal medium (Gibco, Thermo Fisher; Waltham, MA, USA) supplemented with B-27 and 2 mM glutamax (Gibco). At DIV 9–11, primary neuronal cultures were infected with AAV-DJ particles at the concentration of 2.5 × 106 genome copies/ml. After ~72 h, the biochemical assay was carried out.

Biochemical assay of autophosphorylation and oligomerization

For the paCaMKII autophosphorylation assay in HeLa cells, the culture medium was replaced with HACSF and incubated for 20 min at RT before the experiment. For cultured dissociated neurons, 1 µM of tetrodotoxin (TTX) was added to the culture medium and incubated for 30 min in the CO₂ incubator. Subsequently, the culture medium was replaced with HACSF containing 1 µM of TTX and incubated for 20 min at RT before the experiment. To induce autophosphorylation, the samples in 3-cm dishes were continuously illuminated with a light-emitting diode (LED) (M455L2-C1, Thorlabs; Newton, NJ, USA) at 3 mW cm⁻² for 2–5 min. The reactions were stopped at the indicated time by adding a lysis solution (50 mM Tris pH 7.5, 1% NP-40, 5% glycerol, 150 mM NaCl, 4 mM EDTA, 1 tablet/10 ml PhosSTOP (Sigma-Aldrich)). The samples were collected and centrifuged, the supernatant was dissolved in SDS sample buffer and subsequently analyzed by western blotting.

For NMDA-induced autophosphorylation assays, neurons were first washed with Mg²⁺-free artificial cerebrospinal fluid (ACSF; 127 mM NaCl, 2.5 mM KCl, 4 mM CaCl₂, 25 mM NaHCO₃, 1.25 mM NaH₂PO₄, and 25 mM glucose) containing 300 µM glycine, and 1 mM EDTA to remove Mg²⁺ bound to NMDA receptors. Subsequently, ACSF containing 20 µM NMDA and 300 µM glycine was applied and incubated for 2 min at RT. The cells were lysed with the lysis solution and the samples for western blotting were prepared as described above.

For the pull-down assay, dissociated cortical neurons expressing Flag-His×6-paCaMKII were lysed in lysis buffer (1% Triton X-100, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 4 mM EDTA, 5% glycerol) and centrifuged. The total lysate was pooled from the supernatants, and the remaining supernatant was incubated with Ni sepharose (GE Healthcare) for 1 h at 4 °C. Samples were washed three times with wash buffer (20 mM Tris-HCl, 150 mM NaCl, 2 mM MgCl₂, 50 mM imidazole, pH 7.5) and proteins were released with wash buffer containing 500 mM imidazole and dissolved in SDS sample buffer.

Western blotting was performed with the following antibodies: anti-Phospho-CaMKII (Thr286) (D21E4, Cell Signaling Technology); anti-CaMKIIα (6G9; Cell Signaling Technology); anti-CaMKIIβ (ab34703; Abcam); anti-GFP (M048-3; MBL); anti-Cdc42 (11A11; Cell Signaling Technology); anti-β-Actin (8H10D10; Cell Signaling Technology); anti-RFP (1G9; MBL) and HRP-anti-mouse and -rabbit (Jackson Laboratory; Bar Harbor, ME, USA). Uncropped western blot images are presented in Supplementary Fig. 7.

Organotypic hippocampal and cortical slices and gene gun transfection

Hippocampal and cortical slices were prepared from postnatal day 5–9 Wistar rats as described⁶⁸. Briefly, we deeply anesthetized the animal with isoflurane, after which the animal was quickly decapitated, and the brain removed. For hippocampal slices, the hippocampi were isolated and cut into 350 µm sections in ice-cold dissection medium (25 mM HEPES, 2 mM NaHCO₃, 4 mM KCl, 5 mM MgCl₂, 1 mM CaCl₂, 10 mM D-glucose, 248 mM sucrose). For cortical slices, the cortex was isolated and cut in coronal sections. The slices were cultured on the membrane inserts (PICM0RG50, Millipore; Darmstadt, Germany) placed on culture medium (50% MEM, 21% HBSS, 15 mM NaHCO₃, 6.25 mM HEPES, 10 mM D-glucose, 1 mM L-glutamine, 0.88 mM ascorbic acid, 1 mg/mL insulin, 25% horse serum), and incubated at 35 °C in 5% CO₂.

After 7–12 days in slice culture, neurons were transfected with a gene gun (Scientz Biotechnology; Ningbo, China) using 1.6 µm gold particles coated with plasmids and imaged after 2–5 days. For making bullets, gold particles (2–4 mg) and DNA (4–16 µg) were used for a 30 cm long tube. For shRNA experiments, neurons were transfected with control shRNA (shCtrl, black), shRNA against Cdc42 (shCdc42, blue), and shRNA against Cdc42 with shRNA-resistant Clover-Cdc42 (Rescue, magenta). CA1 pyramidal neurons of cultured rat hippocampal slices (DIV 8–10) were transfected with biolistic gene transfer. For making bullets for a 30 cm long tube, DNA containing tdTomato-P2A-paCaMKII (20 µg) with shCtrl (10 µg) or shCdc42 (10 µg) plasmids was used. For rescue experiments, shRNA-resistant Clover-Cdc42 (5–7.5 µg) plasmid was further added. After 5 days of transfection, the experiments were carried out.

Ca²⁺ imaging in spines

Ca²⁺ imaging in hippocampal slice cultures was performed using a custom two-photon microscope. A Ti:sapphire laser (Spectra-Physics; Santa Clara, CA, USA) tuned to 1000 nm was used for the excitation of both GCaMP6f and tdTomato. Fluorescence signals of GCaMP6f and tdTomato collected with a ×60, NA1.0 objective lens (Olympus; Tokyo, Japan) were divided with a dichroic mirror (565DCLP, Chroma; Taoyuan City, Taiwan). GCaMP6f fluorescence was detected by a photomultiplier tube (H7422-40p, Hamamatsu; Hamamatsu, Japan) through an emission filter (FF01-510/84; Chroma). The fluorescence of tdTomato to monitor spine volume change was detected by a photomultiplier tube (R3896; Hamamatsu) through an emission filter (FF01-625/90, Semrock; Rochester, NY, USA). The acquired signals were processed using a data acquisition board (PCI-6110, National Instruments; Austin, TX, USA) and ScanImage software⁶⁹. For image acquisition, 128 × 32 pixels were acquired at 15.6 Hz. To induce sLTP at single spines, bath-applied 2 mM MNI-caged glutamate was uncaged by a second Ti:sapphire laser at a wavelength of 720 nm (30 trains, 0.5 Hz, 6 ms duration/pulse, 6–8 mW) near a spine of interest. paCaMKII was uncaged by a second Ti:sapphire laser at a wavelength of 900 nm (30 trains, 1 Hz, 80 ms duration/pulse, 4 mW) in a spine of interest. Laser power was measured under the objective lens. Two-photon glutamate uncaging was carried out in ACSF containing no MgCl₂, 4 mM CaCl₂, 1 µM TTX, and 2 mM MNI-caged L-glutamate aerated with 95% O₂/5% CO₂ at 24–26 °C.

Two-photon paCaMKII uncaging

To activate paCaMKII in single or clustered spines with two-photon excitation, a second Ti:sapphire laser tuned at a wavelength of 820 or 900 nm was used with 30 trains (0.5 Hz or 1 Hz, 40–80 ms duration/pulse, 4 mW) in a spine of interest. Since the focal plane of imaging (1000–1010 nm) and uncaging (820–900 nm) lasers were different (0.5–1.0 µm) due to chromatic aberration in the microscope, it was compensated by moving sample stage in z-axis (0.5–1.0 µm) with piezo stages (PKVL64F-100U, NCS6101C, Kohzu; Kawasaki, Japan) during the light activation of paCaMKII.

For global paCaMKII activation, photoactivation was done by raster scan (15 × 15 µm² region; three planes with 2-μm z-step were stacked and each plane was scanned 24 times) at 900 or 920 nm (laser power 4 mW, scan speed 7.5 µm/ms). For the experiment, ACSF containing 2 mM MgCl₂ and 2 mM CaCl₂ was aerated with 95% O₂/5% CO₂ at 24–26 °C.

For MNI-caged glutamate uncaging, a second Ti:sapphire laser tuned at a wavelength of 720 nm was used in extracellular solution with a train of 6 ms and 6–8 mW pulses (30 trains at 0.5 Hz) near a spine of interest. Experiments were performed in ACSF containing no MgCl₂, 4 mM CaCl₂, 1 μM TTX, and 4 mM MNI-caged L-glutamate aerated with 95% O₂/5% CO₂. Experiments were performed at 24–26 °C.

Two-photon fluorescence lifetime imaging

Details of two-photon FLIM-FRET imaging were described previously³⁷. Briefly, mEGFP or Clover_T154M/F223R in the FRET sensor was excited with a Ti:sapphire laser (Mai Tai; Spectra-Physics). The scanning mirrors were controlled with the ScanImage software⁶⁹. The green fluorescence photon signals were collected by an objective lens (×60, 1.0 NA; Olympus) and a photomultiplier tube (H7422-40p; Hamamatsu) placed after a dichroic mirror (565DCLP; Chroma) and emission filter (FF01-510/84; Semrock). Measurement of fluorescence lifetime was conducted using a time-correlated single-photon counting board (SPC-150, Becker & Hickl GmbH; Berlin, Germany) controlled with custom software³⁷. For the construction of a fluorescence lifetime image, the mean fluorescence lifetime in each pixel was translated into a color-coded image⁷⁰. Analysis of the lifetime change and binding-fraction change was conducted as described elsewhere⁷⁰.

To measure the light-dependent structural change of paCaMKII expressed in HeLa cells, the culture medium was replaced with HACSF before the observation under 2pFLIM-FRET. HeLa cells expressing mEGFP-paCaMKII-ShadowG or its mutants were imaged with a Ti:sapphire laser tuned to 920 nm at the power of 1 mW. To activate paCaMKII, the samples were continuously illuminated by a blue LED (470 nm LED; CoolLED) with a bandpass filter (FF01-469/35-25; Chroma) at 35 mW cm⁻² for 2 s.

To measure Rho GTPase (Cdc42/RhoA) activity in hippocampal slice culture, Cdc42 (CMV-ShadowY-CBD-P2A-Clover_T154M/F223R-Cdc42) or RhoA FRET sensor plasmids (CMV-ShadowY-RBD-P2A-Clover_T154M/F223R-RhoA) were cotransfected with CMV/CaMP0.4-tdTomato-P2A-paCaMKII plasmids by gene gun. To simultaneously activate Clover_T154M/F223R and tdTomato, a Ti:sapphire laser tuned to 1010 nm in a range of 0.7–1.7 mW was used.

Analysis of the fluorescence lifetime image

To generate fluorescence lifetime images, we acquired the mean fluorescence lifetime in each pixel by calculating the mean photon arrival time <t > using the following equation:

where t_o is obtained by fitting the whole image with single exponential or double exponential functions convolved with an instrument response function as described previously⁷⁰. Subsequently, the mean fluorescence lifetime in each pixel was converted to the corresponding color. FRET efficiency and the binding fraction (fraction of the donor fluorescent protein undergoing FRET) were calculated as in other studies^37,70.

Calculation of the AMPA receptor insertion

The fluorescence of SEP and tdTomato was monitored and compared before and after induction of sLTP at multiple spines using raster scanning (15 × 15 µm). Imaging SEP and tdTomato fluorescence and the induction of the sLTP by paCaMKII were simultaneously carried out by two-photon laser scanning at 920 nm (3–11 mW). To assess the incorporation of SEP fused AMPA receptor subunits GluA1/GluA2 into spines, the density of spine surface subunits as an enrichment value was also measured (Fig. 5c)⁴⁷. First, tdTomato fluorescence was converted to two-dimensional using the following equation by assuming that spine heads are spherical:

SEP fluorescence (G_{spine_surface}) was divided by R_{spine_surface} to measure the relative density of GluA1/GluA2. To compare across different cells, the spine enrichment value was further divided by the dendritic enrichment value (i.e., G_{dendrite_surface}/R_{dendrite_surface}), where R_{dendrite_surface} is calculated by the above equation by replacing R_{spine_surface} with R_{dendrite_surface} and R_{spine_volume} with R_{dendrite_volume}, respectively. Since the dendrites are not spherical structure, R_{dendrite_surface} is an approximate calculation.

Electrophysiology

Whole-cell patch clamping was performed with patch pipettes (9–12 MΩ). We used cells with series resistance lower than 40 MΩ for experiments. CA1 pyramidal neurons of cultured rat hippocampal slices (DIV 8–10) were transfected with biolistic gene transfer using gold beads coated with plasmids containing cDNA of tdTomato-P2A-paCaMKII (16 µg). To induce LTP with paCaMKII uncaging (920 nm, 30 trains, 0.5 Hz, 80 ms, 4 mW), we performed experiments in voltage-clamp mode using Cs⁺ internal solution (130 mM CsMeSO₃, 10 mM Na-phosphocreatine, 4 mM MgCl₂, 4 mM Na₂-ATP, 0.4 mM Na₂-GTP, 10 mM HEPES, and 200 μM Alexa Fluor 488 hydrazide (Invitrogen), pH 7.3) under a two-photon microscope with a 40 × 0.8 NA objective lens (Olympus), and measured two-photon glutamate uEPSC at a spine through the patch pipette using a patch-clamp amplifier (MultiClamp 700B, Molecular Devices; San Jose, CA, USA). Experiments were performed in a buffer (136 mM NaCl, 5 mM KCl, 0.8 mM KH₂PO₄, 20 mM NaHCO₃, 1.3 mM L-glutamine, 0.2 mM ascorbic acid, 2 mM CaCl₂, 2 mM MgCl₂, MEM amino acids solution (Gibco), MEM vitamin solution (Gibco), 1.5 mg/ml phenol red) containing 1 µM TTX and 2–3 mM MNI-caged glutamate aerated with 95% O₂/5% CO₂ at 24–26 °C.

Imaging and paCaMKII uncaging in vivo

For viral infection, mice were anesthetized with an intraperitoneal injection of ketamine (70 mg/kg) and xylazine (10.5 mg/kg) and were secured on a stereotaxic frame (Narishige; Amityville, NY, USA). To inject the viruses into the primary somatosensory cortex (S1), the skull above the S1 (0.5 mm caudal from bregma, 1.5 mm lateral from the middle) was thinned (1 mm diameter) with a drill for the AAV-containing glass pipette insertion. A glass pipette was slowly inserted to a depth of 350 µm (layer 2/3) from the surface of the cortex. Approximately 250 nl of a viral solution was injected at a rate of 25 nl min⁻¹. To sparsely label neurons with Clover, we used a DIO reading frame system in combination with a lower amount of Cre expression. The mixture of the following AAV vectors was infected with the genome titer ratio of 1:900:900, CaMKII-Cre (U Penn), CaMP0.4-Flag-His×6-paCaMKII, and CaMP0.4-DIO-Clover-WPRE3. After AAV injection, the incision was sealed with a surgical staple, and mice were returned to their home cage and housed until the imaging sessions started. We typically waited ~4 weeks after AAV injection until a sufficient level of Clover expression was obtained.

To construct a cranial window for imaging, we anesthetized mice with isoflurane (3% for induction, 1% for surgery), and a custom-made head plate was attached to the skull. A cranial window was made above the S1 region where AAV plasmids were injected, and the dura mater was removed. The exposed S1 region was covered with double cover glass (top: diameter 4.0 mm, bottom: 2 × 2 mm, Matsunami; Bellingham, WA, USA). The coverslips were secured with adhesive glue and dental cement.

For in vivo two-photon imaging, dendritic spines of cortical neurons in awake mice or mice anesthetized with 1% isoflurane were observed using a two-photon microscope (A1R MP, Nikon; Tokyo, Japan) with a water immersion objective lens (×25, 1.1 NA, Nikon). The imaging locations of dendritic spines were 100 μm below the cortical surface (layer 1). Images of dendritic spines of neurons expressing Clover were taken every 1 min (61 × 30 µm² rectangle region, 5–12 planes with 0.5 μm z-step were stacked, each plane was scanned once) at the wavelength of 1000 nm (laser power 8–15 mW). For global two-photon paCaMKII activation, 5–12 planes (61 × 30 µm² region, each plane was scanned once) with 0.5 µm apart to z-axis were scanned at 920 nm (laser power 12–29 mW) every 30 s for 10 min. Subsequently, the dendritic spines were observed at 1000 nm every 5 min for 25 min. The images were analyzed by ImageJ (National Institutes of Health; Bethesda, MD, USA).

Quantification and statistical analysis

Statistical analysis was performed using the Matlab or GraphPad Prism software. The types of statistical tests, the number of samples, and statistical significance are described in the figures or legends.

Reporting summary

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