Thy1-eGFP-M (GFP-M; The Jackson Laboratory, Bar Harbor, ME, USA; RRID:IMSR_JAX:007788; Feng et al., 2000), p53^+/+ (WT), and p53^−/− (KO, p53KO; Zhang et al., 2017) mice with C57BL/6 genetic background were used in the present study. To obtain GFP-M:p53KO mice for the in vivo imaging experiments, GFP-M mice were first crossed with p53^−/− mice, then the double GFP-M:p53 heterozygous offspring was further intercrossed to generate p53^+/+:GFP-M and p53^−/−:GFP-M mice. The Thy1-eGPF-M mouse line primarily labels projection neurons (Feng et al., 2000). The mice genotypes were determined by PCR analysis of the genomic DNA extracted from tail snips. Both male and female mice aged 3–5 months were used for in vivo two-photon excitation (2PE) imaging experiments. Male mice aged 2–3 months were used for the rest of the experiments. The mice were housed in temperature-controlled facilities, maintained at 24 ± 1°C and 50%–60% humidity, with a 12:12-h light/dark cycle. Food and water were available ad libitum. Given that p53KO mice develop tumors mostly between 4–5 months of age (Jacks et al., 1994), we did not use mice older than 5 months old in this study. All mice used were examined to ensure no tumor presence at the time of the experiment. Experiments were conducted in accordance with the National Institutes Health “Guide for the Care and Use of Laboratory Animals” and were approved by the Institutional Animal Care and Use Committees at Tulane University School of Medicine and Nanchang University.

Six p53^+/+ mice and six p53^−/− mice were deeply anesthetized with urethane (1.5 g/kg; i.p.) and transcardially perfused with 30 ml of ice-cold normal saline. Brains were removed, weighed, and immersed in 10% formalin for fixation for at least 24 h. For Nissl staining, six consecutive coronal paraffin-embedded brain sections (3 μm) were stained with cresyl violet (G1430; Solarbio, Beijing, China) and imaged using a biological navigator (FSX100; Olympus, Tokyo, Japan). The cortex thickness and the number of neurons were measured using the Cell Counter tool in ImageJ software (version 1.50i, RRID: SCR_003070¹) within three random regions of interest (200 μm × 150 μm) in L5 of S1BF. The average number of neurons from three to five sections was taken as the final value for each mouse.

Acute coronal brain slices for electrophysiology were prepared as described in our previous study (Wu et al., 2021). Mice were deeply anesthetized with urethane and transcardially perfused with 20 ml ice-cold carbonated (95% O₂ and 5% CO₂) dissection solution containing (mM): 240 sucrose, 2.5 KCl, 3.5 MgCl₂, 0.5 CaCl₂, 1.25 NaH₂PO₄, 0.4 ascorbic acid, 2 sodium pyruvate, 25 NaHCO₃, and 1 kynurenic acid. Brains were quickly removed after decapitation and put into a chilled dissection solution. Coronal slices containing S1BF (350 μm thick) were obtained using a Leica VT1000S (Wetzlar, Germany) vibratome and then transferred to oxygenated artificial cerebrospinal fluid (aCSF) containing (mM): 117 NaCl, 3.6 KCl, 1.2 NaH₂PO₄, 2.5 CaCl₂, 1.2 MgCl₂, 25 NaHCO₃, 11 D-glucose, and 2 sodium pyruvate for 30 min at 32°C to recovery. Afterward, slices were maintained for at least 1 h at room temperature (RT, 22–24°C) in aCSF prior to being moved to the recording chamber.

Whole-cell voltage- and current-clamp recordings were obtained as previously described (Popescu et al., 2017; Wu et al., 2018). In brief, one brain slice was moved to the recording chamber and continuously perfused with aCSF at RT with a perfusion rate of 2 ml/min. Recording electrodes were pulled from borosilicate glass (1.5 mm OD, 1.12 mm ID; World Precision Instruments, Shanghai, China) with a micropipette puller (P-97; Sutter Instrument, Novato, CA, USA). The typical resistance of the pipette was 3–6 MΩ when filled with intracellular solution containing (mM): 130 K-gluconate, 5 KCl, 10 Na₂-phosphocreatine, 0.5 EGTA, 10 HEPES, 4 Mg-ATP, and 0.3 Li-GTP (pH = 7.3 adjusted with KOH, 295 mOsm). The barrel field was visualized at low magnification under bright-field illumination and a target region in L5 directly below an L4 barrel was selected for recording. Individual L5 pyramidal neurons were visually identified at 40× magnification using an infrared differential interference contrast camera (IR-1000; Dage-MTI, Michigan City, IN, USA) with an upright microscope (BX51WI; Olympus). Signals were amplified with an EPC-10 amplifier and Patchmaster software (RRID: SCR_000034; HEKA Electronik, Reutlingen, Germany). The series resistances were typically 10–30 MΩ and were monitored throughout the recording period. Data were excluded if the series resistance changed by >20%. To assure the clear identification of each neuron in the anatomy post hoc analysis, only one or two cells were patched in each slice. Only well-labeled neurons with stable resting membrane potentials (V_rest ≤ −50 mV) were included in the final analysis. Data were not adjusted for a liquid junction potential of −15 mV. All the reagents used for the electrophysiological experiment were obtained from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise mentioned.

Three major parameters of L5 pyramidal neurons were analyzed here, passive intrinsic properties, active intrinsic properties by intracellular current injection, and miniature excitatory postsynaptic currents (mEPSCs). Membrane resting potential (V_rest; in mV) was defined as the membrane potential within 30 s after establishing the whole cell configuration with no current applied (I = 0 pA). Input resistance (R_in; in MΩ) was measured from the current response to a 10-mV hyperpolarizing step in voltage clamp. Active intrinsic properties were examined by injecting a series of depolarizing and hyperpolarizing current steps (2 s duration, 20 pA increments). Rheobase current (in pA) was determined as the smallest current step to elicit at least one action potential (AP). AP threshold (AP_thre; in mV) was defined as the inflection point during spike initiation of the first AP. AP amplitude (AP_amp; in mV) was defined as the difference between the firing threshold and its maximum positive peak. AP half-width (HW; in ms) was calculated at 50% of AP amplitude. To investigate the sag, neurons were given a series of 2 s hyperpolarized current pulses from V_rest. The sag ratio was calculated and expressed as a percentage by dividing the steady-state potential by the peak potential at −280 pA. Miniature EPSCs (in pA) were recorded at a holding potential of −70 mV in voltage clamp after 10-min application of 0.5 μM tetrodotoxin (TTX; Tocris, Bristol, UK). Membrane properties were analyzed using Clampfit 10.7 (pClamp, RRID: SCR_011323; Molecular devices, Hercules, CA, USA). Amplitudes and frequencies of mEPSCs were analyzed using the MiniAnalysisProgram (RRID: SCR_002184; Synaptosoft).

The S1BF of the cerebral cortex from WT and p53KO mice was dissected in a cold RIPA buffer. For Western blot assays, samples (200 μg protein for GluA1 and 40 μg protein for β-actin) were transferred onto a polyvinylidene fluoride (PVDF) membrane (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The membranes were blocked for 1 h in Tris-Buffered Saline with Tween 20 (TBST) buffer with 5% bovine serum albumin (BSA) and were incubated with primary antibodies against GluA1 (1:1,000; Cat# 13185, RRID:AB_2732897; Cell Signaling Technology, Danvers, MA, USA) and β-actin (1:1,000; Proteintech, Rosemont, IL, USA) overnight at 4°C. Membranes were then incubated with goat anti-rabbit Ig-G secondary antibody (1:1,000; Thermo Fisher Scientific, Waltham, MA, USA) for 6 h at 4°C. An imaging system (iBright FL1000, Thermo Fisher Scientific) was used to detect the immunocomplexes by Chemiluminescence (ECL) solution (Millipore Bioscience Research Reagents, Temecula, CA, USA). The band intensity of the detected proteins was determined using the Gels tool in ImageJ software (version 1.50i¹).

Total mRNA was extracted from the S1BF of the cerebral cortex from WT and p53KO mice using RNAiso^TM Plus (9108; Takara Bio, Kusatsu, Japan) as previously reported (Feng et al., 2019). Reverse transcription (37°C for 15 min and 85°C for 5 s) from total RNA (0.5 μg) was carried out with an ABI Veriti PCR machine (Applied Biosystems, Waltham, MA, USA). cDNA was synthesized using PrimeScript^TM RT reagent kit (RR037A; Takara Bio) and stored at −20°C until use. qRT-PCR reactions were performed in triplicate using the SYBR^® Premix Ex Taq^TM Π kit (RR820A; Takara Bio) on a StepOnePlus thermocycler (Applied Biosystems). Expression levels of mRNA were normalized to the reference gene (β-actin) and quantified using the 2^−ΔΔCT method. Reactions without any template were used as negative controls. The primers used are shown in Table 1.

For intracellular labeling of L5 pyramidal neurons, neurobiotin (0.1%, Vector Laboratories, Burlingame, CA, USA) was added to the intracellular solution, and the recording was maintained for at least 25 min. Subsequently, the patch pipette was slowly withdrawn. Then, slices were transferred to a container filled with 4% PFA at RT for 1 h and then at 4°C overnight for fixation. The next day, the slices were rinsed in PBS and incubated in rhodamine Red^TM-X-conjugated streptavidin (1:1,000, Jackson ImmunoResearch Inc., West Grove, PA, USA) in PBS for 6–8 h at 4°C. After three washes in PBS, the slices were mounted and imaged with a confocal microscope (LSM 700, Zeiss, Oberkochen, Germany). Low magnification image stacks (5×, 1.5-μm z step) were acquired for 3D reconstruction and subsequent analysis of cell morphology. Dendritic image stacks (within the first 100 μm from the pia mater) were acquired at high magnification (63×, 0.5-μm z step). ZEN 2010 (Zeiss) and the Measure utility in ImageJ/FIJI (version 1.53c²) were used for the acquisition and quantification of the images, respectively.

Morphological data were obtained as previously described (Oswald et al., 2013). The cortical thickness was defined as the distance from the pia through the center of the soma to the border of L6 and white matter. Cell depth was measured from the soma center to the outermost boundary of the pia in line with the axis of the apical dendrite. To avoid variability due to small differences in slice angle and facilitate comparisons across slices, absolute cell depth (distance from pia) was normalized (cell depth/cortical depth; Yamawaki et al., 2014). Soma size was measured as the average of height and width of the cell body. Apical dendrite length was measured from the line connecting the base of the apical dendrite and the orthogonal projection to this line of the most superficial point in the tuft. The dendritic tuft height and width were measured as the rectangular dimensions defined by the point at which the apical dendrite bifurcated and the most superficial and widest points in the tuft. Morphological data were not corrected for potential tissue shrinkage effects induced by fixation. The images of dendrites from fixed tissue were analyzed to calculate the density and to classify dendritic spines according to their morphology. Spine density was calculated by dividing the total number of spines by the total length of dendrite analyzed from two to five segments per neuron. Spines were classified as mature (stubby or mushroom) or immature (thin or filopodia) according to the spine length (L), the diameter of the head (dh), and the diameter of the neck (dn) using the Measure utility in ImageJ/FIJI (version 1.53c²) software as previously described (Tyler and Pozzo-Miller, 2003; Risher et al., 2014; Liu et al., 2017). The groups were defined as follows: stubby (dn ≈ L), mushroom (dh >> dn, dh ≥ dn), thin (L >> dn), filopodia (L > 2 μm).

Cranial window surgery was performed as previously described (Mostany and Portera-Cailliau, 2008; Mostany et al., 2013; Voglewede et al., 2019). To prevent brain swelling and inflammation mice were injected with carprofen (5.0 mg/kg; s.c.; Zoetis Inc., Parsippany-Troy Hills, NJ, USA) and dexamethasone (0.2 mg/kg; s.c.; MWI, Boise, ID, USA) after anesthesia induction and before any incision was made. Mice were anesthetized with isoflurane (5.0% for induction, 1.5%–1.7% for maintenance), placed in a stereotaxic frame, and a 4-mm craniotomy was performed with a pneumatic dental drill over S1BF, centered at 3 mm lateral to the midline and 1.95 mm caudal to bregma. A 5-mm glass coverslip (#1; Electron Microscopy Sciences, Hatfield, PA, USA) was placed over the intact dura and secured using cyanoacrylate glue and dental acrylic (Lang Dental Mfg. Co, Inc., Wheeling, IL, USA) to the skull. A custom-made titanium bar (9.5 mm × 3.2 mm × 1.1 mm) was cemented within the dental acrylic for securing the mouse to the microscope’s stage. A 3-week recovery time was allowed before in vivo imaging.

Intrinsic optical signal (IOS) imaging was performed to precisely locate S1BF as previously described (Alexander et al., 2018; Voglewede et al., 2019). Mice were anesthetized with isoflurane (5.0% for induction, ≤1.0% for imaging) and secured to a custom-built IOS microscope using the titanium bar. IOS imaging was performed through the cranial window preparation. An image of the vasculature was taken for reference under a green (535 nm) LED array. The imaging plane was focused 250–350 μm below the dura for the imaging. IOSs from S1BF were collected using red (630 nm LEDs) illumination and a fast camera (Pantera 1M60; Dalsa, Waterloo, Canada), frame grabber (64 Xcelera-CL PX4; Dalsa), and custom-written MATLAB (RRID:SCR_001622; MathWorks, Natick, MA, USA) routines. For the stimulation, whiskers were bundled and attached with dental wax to a glass capillary tube controlled by a piezo bender actuator (Physik Instrumente, Karlsruhe, Germany). An imaging session consisted of 30 trials of whisker stimulation for 1.5 s in the rostrocaudal direction at 10 Hz with 20-s breaks. The response signal (frames 1.5 s after the onset of the stimulation) from each stimulation was divided by the baseline signal (0.9 s before stimulation) and the individual values from the 30 trials were summed to obtain the signal map. To identify the activated cortical area, an outline of the IOS map was placed over the image of the vasculature.

In vivo 2PE imaging was done as previously described (Mostany et al., 2013; Voglewede et al., 2019; Davidson et al., 2020). Mice were anesthetized with isoflurane (5.0% for induction, 1.0%–1.5% for imaging) and secured to the microscope stage. The system was a custom-built 2PE microscope using a Ti: sapphire laser (Chameleon Ultra II; Coherent, Santa Clara, CA, USA) tuned to 910 nm, a 40× 0.8 NA water-immersion objective (Olympus), and ScanImage software (RRID: SCR_014307; Pologruto et al., 2003) written in MATLAB. At the start of the chronic 2PE imaging cells were chosen from within the IOS activity map and their locations and individual dendritic fragments for a cell of interest were recorded using a coordinate system. L5 pyramidal neurons were verified by a cell body depth of 400–700 μm below the dura, and five-eight dendritic fragments of the apical tuft located in L1 were imaged chronically. High magnification image stacks (0.14 × 0.15 mm/pixel; 1.5 μm apart) were collected for the analysis of dendritic spines. Dendritic fragments were imaged chronically over imaging intervals of 4 days for a total of 5 imaging sessions. Three sessions were done before (d0, d4, d8) and two after (d12, d16) sensory-evoked synaptic plasticity induction by whisker stimulation.

To induce synaptic plasticity in vivo mice were subjected to rhythmic whisker stimulation as previously described (Alexander et al., 2018; Voglewede et al., 2019). Briefly, mice were anesthetized with isoflurane (5.0% for induction, <1.0% during stimulation), and their whiskers carefully bundled and attached to a piezo bender actuator, which oscillated rostrocaudally at 8 Hz continuously for 10 min. Stimulation was performed daily for four consecutive days starting on d8, and immediately after the 2PE imaging session.

Spine density and dynamics from 2PE images were determined using spine analysis software written in MATLAB. All visible spines were annotated, including those on the z-axis that clearly protruded beyond the noise of the dendritic shaft. We analyzed dendritic spines from n = 10 cells (N = 8 WT mice) and n = 8 cells (N = 7 p53KO mice) and tracked a total of 5,789 and 4,393 distinct dendritic spines from WT and KO mice, respectively. For display purposes only, best projections of the dendritic segments were obtained, where the best focal plane is identified and overlaid in Adobe Photoshop CC (Adobe Systems Inc., San Jose, CA, USA), preserving all the elements in the segment, and a median filter (radius of 1) was applied. We defined the density of dendritic spines as the number of spines annotated per unit length (μm) of a dendritic segment. Spine turnover rate (TOR) was defined as the combined number of gained and lost spines per μm over a 4-day period. The fold-changes in spine density and TOR were defined as the density and TOR calculated the days after the whisker stimulation divided by the average spine density and TOR computed during the steady-state imaging regime, respectively. We defined persistent spines as those dendritic spines present at all the five imaging time points (d0 – d16). New persistent spines (NPS) were defined as newly-formed spines identified at day 4 (d4) that were still present at day 16 (d16). The survival functions of all the dendritic spines present at day 0 and of NPS were obtained by fitting their survival fraction plots to a single exponential decay curve. Survival fraction and half-life were calculated as follows: survival fraction = plateau + unstable fraction × e^−t/τ, where t is the time (days) and τ (tau) is the time constant. Rate constant (K) = 1/τ.

We assessed whisker-dependent perceptual learning using a novel texture discrimination task following a previously described method with slight modifications (Wu et al., 2013). Mice were transported to the testing room where they were allowed to acclimate for at least 1 h. On days 1 and 2 (habituation days) mice were placed into the open arena used for the testing (a 50 cm × 50 cm × 25 cm chamber carpeted with 2 cm of standard laboratory bedding) for 10 min to habituate to the arena and promote exploration. On the 3rd day (testing day), mice were individually placed in the arena containing two identical target objects, a and b (4 cm width, 15 cm high, 0.5 cm thickness, and covered with sandpaper, of either 80 or 100 grit), and allowed to freely explore the objects for 5 min (learning phase). Then, the mice were removed and held in the transport cage for 5 min and the two objects were replaced. Replacement objects were identical in size, shape, and color and included one identical in texture to those used in the learning phase, c, and a second of a different grit, d. Mice were then placed back into the arena for the test phase, and allowed to explore for 3 min. The learning and test phases were recorded and manually analyzed with a video tracking system (Kinoscope software³). The amount of time mice spent actively investigating the objects was recorded and was defined as directing the nose towards the object within 2 cm. General exploratory activity was defined as the total time the mouse spent investigating the two identically textured objects during the learning phase. A discriminative activity was defined as the percentage of time spent investigating the novel texture vs. the total time spent investigating the familiar textured objects during the testing phase. Learning phase exploration index, phase in which objects a and b are identical, was calculated as (t_b−t_a)/(t_a+t_b), where t_a and t_b are the times spent exploring objects a and b, respectively. Testing phase exploration index, phase in which object c is identical to the objects from the learning phase, and object d is different from the rest of objects only in its texture, was calculated as (t_d−t_c)/(t_d+t_c), where t_c and t_d are the times spent exploring object c and d, respectively (Michaelson et al., 2018). The arena was cleaned with 70% ethanol after each test to remove olfactory cues. Behavioral analysis was performed blinded to experimental conditions.

Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA) and the Statistics Toolbox in MATLAB (version 9.4.0.813654). The Shapiro-Wilk test was used to assess the normal distribution of data. The Levene test was used to test the homogeneity of variance. The comparison of two groups was determined by paired, unpaired student’s t-test, or the Mann-Whitney test. The comparison of multiple groups was assessed by one-way or two-way repeated measure (RM) analysis of variance (ANOVA). Tukey post-hoc test was used following ANOVA test. Wilcoxon signed ranks test was used to compare the fold-change in spine density. The extra sum-of-squares F test was used to compare the best-fit values for the survival function parameters, i.e., plateau and rate constant K. All data are given as the mean ± SEM, unless otherwise stated. Significance was set at p < 0.05.

We first examined macroscopic and microscopic brain features. Whole brain comparison did not reveal apparent differences between WT and p53KO mice (Figure 1A). The body weight (t₍₁₀₎ = 0.4577, p = 0.6570), brain size (t₍₁₀₎ = 0.6984, p = 0.5008), brain weight (t₍₁₀₎ = 0.7757, p = 0.4559), and the brain/body weight ratio (t₍₁₀₎ = 1.275, p = 0.2313) of p53^−/− mice were comparable to those of p53^+/+ mice (unpaired t-test; Figure 1B). We also compared the cortical thickness of S1BF of these mice (unpaired t-test; Figures 1C,D). Neither the cortical thickness (t₍₁₀₎ = 0.9902, p = 0.3454; unpaired t-test; Figure 1D) nor the laminar organization (F_{(1, 16)} = 3.066, p = 0.0861; Two-way ANOVA; Figure 1E) displayed alterations between the two groups of animals. We next tested whether p53 deletion may affect L5 neuronal density in S1BF. In agreement with previous findings (Filichia et al., 2015), neuronal density in L5 did not show any difference between p53^+/+ and p53^−/− mice (t₍₁₀₎ = 0.4211, p = 0.6826, unpaired t-test; Figure 1F). These results suggest that p53 is not essential for the gross morphology of mouse brain.

Despite the lack of apparent abnormalities at a gross anatomical level, it has been reported that p53KO mice suffer from several neurological problems, including learning deficiencies, anxiety, and motor deficits (Amson et al., 2000; Campana et al., 2003). Therefore, it is possible that p53 deletion may alter neuronal function without large-scale morphological changes. We tested whether p53 deficiency might affect the intrinsic properties of L5 pyramidal neurons in S1BF. L5 pyramidal neurons were categorized into regular spiking (RS) and intrinsically bursting (IB) neurons according to their firing pattern in line with previous reports (Schubert et al., 2001; Jacob et al., 2012). We recorded the responses of L5 pyramidal neurons to a series of depolarizing and hyperpolarizing current injections (Figure 2A). Membrane properties are summarized in Table 2. As indicated in prior studies, including ours (Popescu et al., 2017), and confirming the identity of the cells recorded, sag ratio, AP amplitude, and AP half-width were larger in IB neurons, whereas rheobase and AP threshold were larger in RS neurons (Supplementary Figure 1), and no significant differences were found in the V_rest of the two types of neurons between genotypes. Specific to the p53 deletion, the sag ratio was significantly decreased in RS neurons (Figure 2B). The f-I curves showed that the deletion of p53 significantly decreases the firing frequency in IB neurons. RS neurons displayed lower firing frequency rates in both genotypes (Figure 2C). These results suggest that p53 deficiency alters intrinsic electrophysiological properties of L5 pyramidal neurons in S1BF that lead to a hypoexcitable phenotype.

Next, to study the effect of p53 deletion on spontaneous glutamatergic synaptic transmission, we recorded mEPSCs in RS and IB neurons from p53^+/+ and p53^−/− mice (Figure 3A). The amplitude and frequency of mEPSCs were generally larger in IB than in RS pyramidal neurons (Supplementary Figure 2, Table 3). The deletion of p53 significantly reduced the amplitude of mEPSCs in IB neurons (Figure 3B) and the frequency of mEPSCs in IB and RS neurons (Figure 3C). No differences were observed in the decay time of mEPSCs between p53^+/+ and p53^−/− mice. These results suggest that the glutamatergic synaptic transmission is impaired in L5 pyramidal neurons when p53 is deleted.

The amplitude reduction of mEPSCs in p53KO mice may be due to a decreased expression of AMPAR subunit GluA1, which is a target of p53 signaling pathway (Jewett et al., 2015). We measured the expression levels of GluA1 but did not detect any significant changes in protein expression between the phenotypes (Figure 3D). Another major excitatory glutamate receptor is the NMDA receptor (NMDAR). Thus, we examined the mRNA levels of relevant NMDAR subunits and again did not find any difference in the mRNA levels of Grin1, Grin2a-2d, and Grin3a in the brains between p53^+/+ and p53^−/− mice. However, we found a significant increase in Grin3b mRNA levels in p53^−/− brains (Figure 3E), an increase that may account for the decreased glutamatergic synaptic transmission, as previously reported (Nishi et al., 2001; Matsuda et al., 2002), of L5 pyramidal neurons in these mice (Figures 3A–C).

We then tested if p53 is required for the dendritic development of L5 pyramidal neurons. To do so, L5 pyramidal neurons were recorded and simultaneously filled with neurobiotin for post hoc morphological analysis and comparison between p53^+/+ and p53^−/− mice (Figures 4A,B). Besides their firing pattern, L5 pyramidal neurons can also be classified into slender-tufted (ST) and thick-tufted (TT) neurons according to the morphology of their apical dendrite and dendritic tuft (Schubert et al., 2006; Hattox and Nelson, 2007; de Kock et al., 2007; Narayanan et al., 2015; Krieger et al., 2017; Popescu et al., 2017). Similar to our previous study (Popescu et al., 2017), RS and IB neurons corresponded to the ST and TT neurons, respectively. As shown in Figure 4C, RS (ST) neurons had a less complex and shallow apical tuft than IB (TT) neurons, regardless of the genotype.

In order to compare the layer profile of RS and IB neurons, we normalized the depth from the pia of each soma to the cortical thickness. S1BF L5 RS neurons occurred at a normalized depth of 0.55 ± 0.01 and 0.53 ± 0.01, n.s., in WT and p53KO mice, respectively. However, IB neurons in KO mice appeared to be more superficially located than those from WT mice (0.59 ± 0.01 vs. 0.55 ± 0.01, WT vs. KO, respectively; t₍₄₃₎ = 2.975, p = 0.0048; unpaired t-test; Figure 4D, Table 4). Morphologically, IB neurons were distinguished from RS neurons by their larger soma size and longer and wider dendritic tuft (Supplementary Figure 3). However, for a given neuron subtype, there were no differences between the two genotype groups (Figures 4E–G). Apical dendrite length was significantly shorter in KO-RS neurons than that in WT-RS neurons (Figure 4H).

Given that dendritic spines are the main input of excitatory transmission in pyramidal neurons, we postulated that the decrease in mEPSC frequency in KO mice may result from a reduction in spine density. To test this speculation, we compared the dendritic spine density of the neurons recorded from p53^+/+ and p53^−/− mice (Figure 5A). Spine density in p53KO mice was significantly decreased in both RS and IB neurons compared to those of WT mice (WT-RS, 0.25 ± 0.02 vs. KO-RS, 0.18 ± 0.01; WT-IB, 0.17 ± 0.01 vs. KO-IB, 0.13 ± 0.01; Figure 5B). We next classified the spines into mature (stubby or mushroom) or immature (thin or filopodia) spines (Figure 5C) as previously described (Fiala et al., 1998; Risher et al., 2014). The percentages of mature and immature spines were not significantly different between genotypes for either spine category (Figures 5D,E). Together, these data suggest that p53 may be required for the dendritic development of L5 pyramidal neurons of S1BF and involved in spinogenesis. The latter set up the motivation for the in vivo imaging studies, aimed to study the dynamics of dendritic spines.

To examine dendritic spine dynamics of L5 pyramidal neurons from both WT and p53KO mice during steady-state and after plasticity-inducing conditions, we used chronic in vivo 2PE microscopy through a cranial window (Figure 6A). We imaged the same dendritic segments of individually identified L5 pyramidal neurons for three baseline sessions (d0, d4, d8; steady-state) before the introduction of daily whisker stimulations (d8-d11) and after whisker stimulation (d12 and d16; stimulated state). Time-lapse imaging every 4 days over a total of 16 days showed the dynamic nature (gains and losses) of dendritic spines in both WT and KO mice, including persistent, newly formed, and eliminated spines (Figures 6B,C). The spine density was not significantly different across the duration of the experiment in either WT or KO mice (Figures 6D–F). However, the density of dendritic spines (F_{(1, 16)} = 10.65, p = 0.0049, RM two-way ANOVA), TOR/μm (F_{(1, 16)} = 8.406, p = 0.0105, RM two-way ANOVA), number of gained spines/μm (F_{(1, 16)} = 5.899, p = 0.0273, RM two-way ANOVA), and number of eliminated spines/μm (F_{(1, 16)} = 9.416, p = 0.0073, RM two-way ANOVA) were significantly decreased in the p53KO mice compared to WT mice (Figures 6F–I). Subsequent analysis at different time points found a significant lower spine density in KO compared with WT mice (p = 0.0064 for d0, p = 0.0177 for d4, p = 0.0154 for d8, p = 0.0364 for d12, p = 0.0246 for d16; Sidak’s multiple comparisons test; not shown), whereas no significant differences after post hoc analysis were computed for TOR/μm, gained spines/μm, and eliminated spines/μm at the different time points.

To determine the role of p53 in structural plasticity, we used a 4-day interval and sampled the same dendritic fragments pre- and post-stimulation. Data from the steady-state imaging sessions were averaged to serve as a baseline (Pre) for comparison with the data obtained the days after the whisker stimulation (d12 and d16). KO mice displayed lower pre-stimulation spine density than WT mice did (Figure 6J, t₍₁₆₎ = 3.365, p = 0.0039; unpaired t-test). The TOR of dendritic spines was also significantly decreased in KO mice compared with WT mice (Figure 6K, t₍₁₆₎ = 2.809, p = 0.0126; unpaired t-test). While the number of gained spines per μm was not significantly different between genotypes (Supplementary Figure 4A; t₍₁₆₎ = 1.893, p = 0.0765; unpaired t-test), the linear density of eliminated spines was reduced in p53KO mice compared with WT mice (Supplementary Figure 4B; t₍₁₆₎ = 2.437, p = 0.0286; unpaired t-test). Whisker stimulation did not affect the density of dendritic spines (Supplementary Figure 4D). However, while the changes in TOR after the stimulation relative to the steady-state were significantly increased at d12 in WT mice (WT-pre vs. d12, t₍₉₎ = 3.737, p = 0.005; paired t-test; Figure 6L) and returned to normal levels at d16, similar to previous results (Voglewede et al., 2019), we could not detect significant changes in TOR compared to baseline in p53KO mice (p53KO-pre vs. d12, t₍₇₎ = 1.372, p = 0.2124; paired t-test; Supplementary Figure 4D). We did not detect statistically significant differences in the fold-change of gained and eliminated spines (Supplementary Figures 4D,E). Finally, we analyzed the survival functions of the dendritic spines identified on the first day of imaging and of new persistent spines (NPS) identified on d4. When all the spines scored at d0 were followed over time we did not find differences in the rate constant (K) of the exponential decay fits computed for both groups (WT, 0.042 ± 0.002; KO, 0.038 ± 0.002; F_{(1, 88)} = 1.996; p = 0.7179; mean ± SD; extra sum-of-squares F test; Supplementary Figure 4F). These K values translated into half-life values of 16.6 ± 0.6 and 18.3 ± 1.1 days (mean ± SD) for dendritic spines of WT and p53KO mice, respectively, and the fraction of persistent spines at day 16 for both groups was not different either (WT, 0.57 ± 0.02; KO, 0.60 ± 0.03; t₍₁₆₎ = 1.009, p = 0.3280; extra sum-of-squares F test; Supplementary Figure 4G). Regarding NPS, K values were significantly different (WT: 0.076 ± 0.006; KO: 0.154 ± 0.011; F_{(1, 70)} = 44.57; p < 0.0001; mean ± SD; extra sum-of-squares F test; Figure 6M), representing NPS half-life values of 9.1 ± 0.7 and 4.5 ± 0.3 days (mean ± SD) for WT and KO mice, respectively. In fact, the fraction of d4 NPS still present the last day of imaging was significantly reduced in p53KO mice (0.23 ± 0.04) compared to WT mice (0.44 ± 0.06; t₍₁₆₎ = 2.839, p = 0.0118; unpaired t-test; Figure 6N). These results indicate that p53 deficiency leads to the alteration of both steady-state and sensory-evoked dynamics of dendritic spines in L5 pyramidal neurons of S1BF.

To assess whether the impairment in plasticity observed in the p53KO group could lead to the impairment of whisker-mediated contextual learning, we performed a whisker-dependent novel texture discrimination task on p53^+/+ and p53^−/− mice (Figures 7A,B). During either the learning or the testing phase on day 3, the amount of time spent investigating the textured objects was not significantly different between two groups of mice (F_{(3, 30)} = 0.8077, p = 0.4996, one-way ANOVA followed by post hoc Tukey’s test; Figure 7C). During testing with the novel textured object, p53^+/+ mice were able to discriminate and spent significantly more time exploring the new textured object (paired t-test: t₍₉₎ = 4.206, p = 0.0023; one-sample t-test: WT learning, t₍₉₎ = 0.0602, p = 0.9533; WT testing, t₍₉₎ = 3.561, p = 0.0061, Figure 7D). In contrast, P53^−/− mice had a harder time discriminating between the novel and the familiar texture during the testing phase (paired t-test: t₍₆₎ = 3.646, p = 0.0108; one-sample t-test: KO learning, t₍₆₎ = 1.197, p = 0.2763; KO testing, t₍₆₎ = 2.374, p = 0.0552, Figure 7D). These data suggest that the deletion of p53 impairs whisker-dependent texture discrimination and contextual learning.