Loss of Calretinin in L5a Impairs the Formation of the Barrel Cortex Leading to Abnormal Behaviors

The rodent whisker-barrel cortex system has been established as an ideal model for studying sensory information integration. The barrel cortex consists of barrel and septa columns that receive information input from the lemniscal and paralemniscal pathways, respectively. L5a is involved in both barrel and septa circuits and play a key role in information integration. However, the role of L5a in the development of the barrel cortex remains unclear. Previously, we found that Calretinin is dynamically expressed in L5a. In this study, we analyzed Cr KO mice and found that the dendritic complexity and length of L5a pyramidal neurons were signicantly decreased after Cr ablation. The membrane excitability and excitatory synaptic transmission of L5a neurons were increased. Consequently, the organization of the barrels was impaired. Moreover, L4 spiny stellate cells were not able to properly gather, leading to abnormal formation of barrel walls as the ratio of barrel/septum size obviously decreased. Cr KO mice exhibited decits in exploratory and whisker-associated tactile behaviors as well as social novelty preference. Our study expands our knowledge of L5a pyramidal neurons in the formation of barrel walls and deepens the understanding of the development of the whisker-barrel cortex system.


Introduction
Rodents use their whiskers to explore the presence and location of objects when moving through a nocturnal environment and evolved a whisker-barrel cortex system [1][2][3][4][5]. As the information-receiving region, the barrel cortex plays a crucial role in integrating information resources and coordinating the movement of the whiskers and determines the function of the entire system [2,[6][7][8][9]. The barrel cortex consists of barrel and septa columns that receive various input signals through distinct pathways. The lemniscal pathway transmits whisker-speci c signals to homologous barrel columns, and the paralemniscal pathway transmits multiwhisker signals to both barrel and septa columns [9][10][11][12][13]. The integration of information from both the lemniscal and paralemniscal pathways in the barrel cortex is a prerequisite for precise object recognition [6,14].
As the main target of the paralemniscal pathway, L5a is involved in both barrel and septa circuits and is considered to be an important component of information integration for the lemniscal and paralemniscal pathways [10,[15][16][17]. L5a pyramidal neurons directly accept the information inputs from the paralemniscal pathway and indirectly accept the information inputs from the lemniscal pathway. By neuronal tracer techniques, it has been observed that dendrites of L5a pyramidal neurons have a larger span and can establish synaptic connections with L2/3 and L4 neurons [18][19][20]. The earliest input to L5a pyramidal neurons is provided by L4 spiny stellate cells, which is then followed by an asynchronous L2/3 input [20]. The information from L4 spiny stellate cells is strong direct, monosynaptic [21]. Meanwhile, as the main information input layers of the paralemniscal pathway, L1 and L5a are closely related to each other [6,18,22]. However, the effects of L5a pyramidal neurons on the development of the barrel cortex and related behaviors need to be further elucidated.
Calcium ions (Ca 2+ ) participate in a series of physiological functions, including gene transcription, enzyme activity, ionic channel permeability, and neurotransmitter release [23][24][25]. The precise control of Ca 2+ concentration and spatial location is a prerequisite for accomplishing these tasks. This control is a function of the balance of Ca 2+ transport mechanisms across the plasma membrane, the storage and mobilization of Ca 2+ from intracellular stores, and the actions of a multitude of calcium-binding proteins (CaBPs) located throughout the cytoplasm [25]. Calretinin (CR) [26], which belongs to the EF-hand Ca 2+binding protein family [27,28], is a well-known Ca 2+ buffer in uencing spatiotemporal Ca 2+ transients within the cytosol [29]. Previous studies have reported that CR, as a Ca 2+ sensor [28], is also required for signaling cascades in response to intracellular Ca 2+ transients.
Previously, we showed that CR is dynamically expressed in L5a pyramidal neurons in the developing barrel cortex and displays a unique predominantly serrated pattern that mirrors the presynaptic projection pattern of the posterior medial nucleus (Pom) to L5a [10,30]. In this study, by analyzing Cr KO mice, we found that loss of Cr results in a reduced complexity of L5a pyramidal neuron dendrites, which leads to abnormal formation of the barrel wall and subsequently impairs the barrel and septa microcircuits. Cr KO mice exhibit abnormal exploratory and tactile sensation behaviors. Our results provide evidence that L5a pyramidal neurons direct the formation of the barrel wall during the development of the barrel cortex.

Animals
The CR-CreER line, which was designed to both abolish the Cr gene and express an inducible site-speci c Cre recombinase (stock number: 013730), was purchased from The Jackson Laboratory. The AI9-RFP (stock number: 007905) line was introduced to trace CR + L5a neurons [31]. Cr KO mice were obtained by intercrossing CR CreER/+ mice. WT and heterozygous mice were used as controls. The day of birth was de ned as postnatal day P0. Tamoxifen induction was performed at P5 and P8. Mice were bred in the animal facility at Southeast University. All of the experiments were performed according to the guidelines approved by Southeast University.
Images of dendritic arbors were captured under a 40x objective lens with a confocal microscope (Olympus, FV1000) in Z-stack mode [34]. Neuron morphology was traced manually using the NeuronJ plugin in ImageJ software. Standard morphometric analysis (Sholl analysis) was conducted as described earlier. Signi cance was determined by a two-way repeated-measures analysis of variance (RM 2-ANOVA; genotype and circle radius as factors).

Western blot
Somatosensory cortex homogenates were collected at P8 and prepared as described previously [34].
Protein samples were run on SDS-PAGE and transferred to cellulose acetate membranes. After incubation in TBS containing 5‰ Tween 20 and 5% milk for 1 h at room temperature (RT), the membranes were incubated with primary antibody (rabbit anti-Calretinin, Millipore, AB5054, 1:2000) at 4°C overnight. After washing in 5‰ Tween 20 in TBS for 30 min, the membranes were incubated with secondary antibody (HRP-linked anti-rabbit IgG, Cell Signaling Technology, 7074S, 1:5000) in TBS buffer for 1 h at room temperature, and immunoreactive bands were visualized with an ECL kit (Thermo Scienti c). Quantitative analysis was performed with ImageJ software. The intensity of bands was normalized to the intensity of the corresponding β-tubulin band. An unpaired Student's t-test was used to determine the signi cance.

Nissl staining
The brain slices were removed from the ultralow temperature refrigerator and allowed to dry at room temperature. Then, the slices were immersed in distilled water. After 2 min, the slices were removed and placed in Nissl staining solution for 10 min. After washing with distilled water for 10 min, the slices were immersed in 95% alcohol for 5 seconds. The slices were dipped in xylene for 5 min and sealed with rhamsan gum [35].

Electrophysiology
Slice preparation P18 to P20 brain slices were used for electrophysiological experiments. Brie y, mice were anesthetized by inhalation of iso urane, and the brains were quickly removed and immersed in precooled arti cial cerebrospinal uid (ACSF) containing (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH 2 PO 4 , 26 NaHCO 3 , 1 CaCl 2 , 6 MgCl 2 , and 10 glucose. Coronal slices at a thickness of 350 μm were obtained using a vibrating microtome (Leica Microsystems, VT1200s) [36]. The slices were incubated in a chamber at 35°C for 30 min and then maintained at room temperature (22°C) for at least 1 h before recording.
Electrophysiological recording Electrophysiology was performed as described previously [35,37]. Brie y, brain slices were placed in the recording chamber and completely submerged in ACSF (bubbled with 95% O 2 /5% CO 2 ). Whole-cell recordings were performed on RFP + neurons in L5a. The RFP + neurons were detected by a uorescence microscope. The recording was aided with infrared optics using an upright microscope equipped with a ×40 water-immersion lens (Olympus BX51W1, Japan) and an infrared-sensitive CCD camera. The pipette (input resistance: 3-6 MΩ) solution contained the following (in mM): 125 potassium D-gluconate, 8 NaCl, 0.2 EGTA, 10 HEPES, 2 Mg-ATP, 0.3 Na-GTP. Patch pipettes were pulled on a horizontal pipette puller (P-97, Sutter Instrument). Series resistances were usually 15-30 MΩ upon break-in and were compensated by 70%. RFP + neurons with stable series resistance (20% change throughout the recording) were used for analysis. Data were recorded by an Axon patch 700B ampli er (Molecular Devices), low-pass ltered at 2 kHz and digitally sampled at 10 kHz online and analyzed o ine with Clamp t software (Molecular Devices). To characterize the intrinsic membrane properties of neurons, spiking patterns were recorded in the current-clamp con guration by injecting a series of current pulses (400 ms duration, -50 to 300 pA intensity with an increment of 50 pA). The following parameters were measured to characterize neuronal membrane properties: the resting membrane potential was recorded immediately after the rupture of the neuronal membrane; the action potential current threshold was de ned as the rst 400 ms rectangular current injection that elicited a spike; the input resistance was determined by measuring the voltage change in response to a hyperpolarizing current pulse; the amplitude of afterhyperpolarization (AHP) was measured as the distance between the threshold and the most negative membrane potential following the spike of the rst action potential evoked by the rst current step evoking action potentials; and the spike width was measured at half the height between the threshold and peak action potentials. To isolate miniature EPSCs (mEPSCs), tetrodotoxin (TTX, MCE, 1 μM, to block sodium current) and bicuculline (BMI, Sigma-Aldrich, 14343, 10 μM, to block GABA receptors) were added to the bath solution. mEPSCs were analyzed using the Mini Analysis Program (Version 6.0.3, Synaptosoft), and all events were detected above a threshold of 5 pA.

Behavioral tests
All behavioral tests were performed using groups of 2-to 3-month-old littermate male mice [38]. The mice were handled for 5 days before the beginning of behavioral testing and left to acclimate in the testing rooms for at least 30 min before the experiments. All tests had an interval of at least 1 day between each other. Open-eld tests, novel object investigation tests, elevated O-maze tests, elevated plus maze tests, S curve tests, texture discrimination tests and social behavior tests were conducted under red lighting; sticky paper tests and gap crossing tests were conducted under infrared lighting [39]. All videos were taken using high-resolution digital cameras and analyzed by EthoVision software (Noldus) in a doubleblinded manner.

Open-eld test
Mice were introduced into the center of the chamber (40 × 40 × 40 cm) at the beginning of the test. Their movements were recorded with a video camera for 30 min.

Novel object investigation test
Mice were placed into the open-eld box for 10 minutes of locomotor activity. Then, a novel object was placed in the center of the open-eld box. The mice were allowed to explore for 10 min, and their locomotor activity was recorded and analyzed [40].

S curve test
The size of the S curve was 6 × 120 × 15 cm; one end was closed, and the other end was open. Mice were introduced into the closed end. The time they spent reaching the destination was recorded as incubation, and the time before they left the curve was recorded as total time.
Texture discrimination test Similar to the novel object investigation test, two glass bottles (3 cm in diameter and 5 cm in height), one of which was wrapped in sandpaper, were placed in the opposite corners of the open-eld box. Mice were then introduced into the center of the box and allowed to explore different textures with their whiskers for 10 min.

Sticky paper test
We performed the sticky paper test in the home cage [41]. Two adhesive-backed papers (0.5 cm in diameter) were placed on the palmar surface of the hind paws, and the latency of the rst reaction to the stimulus was recorded as incubation.

Gap crossing test
The gap crossing test consisted of a series of trials during which mice were required to cross gaps of variable distance [42]. The mice were placed on an elevated lane (6 cm in diameter). The lane connected to a safe platform. The distance between the lane and the platform was changed from 0 to 7 cm in trials by 0.5 cm increments. The distances that the mice were able to cross were recorded.

Social behavior test
The social behavior test consisted of three phases [34,38]. In the rst phase, the test mouse was placed in the middle of a three-chambered box (40 × 20 cm) with open middle sections on each of the transparent dividing walls. There were two small containers in the left or right chamber. The mouse was allowed to explore all three chambers for 5 min for habituation. After 5 min, a stranger male mouse was introduced into one of the two containers, the test mice were placed in the middle and allowed to explore the new environment freely for 10 min. In the last phase, a new stranger male mouse was placed in the last empty container, and the test mice were also placed in the middle and allowed to explore the environment freely for 10 min. We measured the exploration time in each side of the three-chamber box during the three phases. All apparatus chambers were cleaned with 75% alcohol and dried with a pledget between trials.

Statistical analysis
All data are presented as the mean ± standard error of the mean (SEM) and were analyzed using GraphPad Prism 8.0 software. Student's two-tailed t-tests were used for analysis of two experimental groups. One-way ANOVA with Tukey's post hoc test was used when more than two groups were compared. Statistical signi cance was de ned at P < 0.05 and is presented as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Results
Loss of Cr results in decreased dendritic complexity of L5a pyramidalneurons P0-P15 is a key time window for the experience-dependent development of the barrel cortex [22,[43][44][45][46].
Interestingly, we previously observed dynamic expression of Cr in L5a pyramidal neurons during this period [30]. We then employed Cr KO mice to explore the role of CR in the development of L5a pyramidal neurons. We rst con rmed the disruption of CR in the developing barrel cortex at P8 by immunohistochemical staining (Fig. 1A) and Western blotting (Fig. 1B). Since an inducible Cre recombinase (CreER) was targeted to the Cr locus, the AI9-RFP line was next introduced to trace CR + L5a pyramidal neurons (Fig. 1C). We analyzed the dendritic morphology of RFP + neurons at P20 and found that control mice exhibited a typical morphology of L5a pyramidal neurons with an apical dendrite extending to L1 and several basal dendrites stretching to the sides and deeper [20]; however, the complexity of basal dendrites was signi cantly reduced in Cr KO mice (Fig. 1D). We measured the dendritic length of RFP + neurons and detected signi cant reductions in the apical, basal and total dendritic lengths in Cr KO mice compared with those in control mice (Fig. 1E). The apical dendritic length decreased by approximately 25%, and the basal and total dendritic lengths decreased by approximately 35%. We also observed a marked decrease in dendritic complexity in Cr KO mice (Fig. 1F). These results suggest that ablation of Cr led to abnormal dendritic development of L5a pyramidal neurons.

Impaired organization of barrels and abnormal formation of the barrel wall in Cr KO mice
To investigate the contribution of L5a to the development of the barrel cortex, we carefully pro led the morphological changes of L5a pyramidal neurons during the time window of P0-P30 by double immunostaining of CR and vGlut2, a marker commonly used to label cortical barrels [47]. At P4, the CR + L5a pyramidal neurons displayed a serrated pattern of alignment underneath the barrels, with their dendrites starting to extend towards the intervals between barrels (Fig. 2A). From P8 to P15, the serrated alignment pattern of the CR + L5a cell bodies became more distinct, and more dendrites were observed in the intervals and formed septa-like structures (Fig. 2B, C) [48]. From P15-P30, as the maturation of the barrel cortex proceeded, the expression level of Cr in L5a gradually decreased. Until P30, the barrels in L4 were well developed (Fig. 2D). The expression of Cr was decreased to a very low level both in L5a cell bodies and in septa-like structures (Fig. 2D).
L4 is the main recipient layer of the whisker-barrel cortex system, whereas L5a is the main output layer, and the two layers are monosynaptically connected [2,6,49]. The connections of L4 to L5a form a "short circuit" between afferent signals to the cortex and efferent signals that leave the barrel cortex from L5a [8]. To assess the effects of abnormal L5a dendrite on the development of barrel/septum microcircuitry, we examined the barrel cortex at P8. As shown in Figure 2E, in contrast to the regularly organized barrels at L4 in control mice, we found barrels with disrupted organization in Cr KO mice, with some barrels deviating from their intrinsic level.
During the development of the barrel cortex, thalamocortical axons (TCAs) reach L4 at around P1, and clusters of axons can be detected in the barrel cortex at P3, re ecting an increase in the complexity of axonal endings [22,50]. Subsequently, L4 spiny stellate cells reorganize around TCA axonal clusters to form barrel walls [45]. At P7, barrel walls are clearly visible. L5a pyramidal neurons are reported to preferentially connect with barrel walls, but whether L5a is required for the formation of barrel walls remains unclear. In our control mice, consistent with previous reports, at P8, L4 spiny stellate cells gathered to form a clear barrel wall between barrels, as shown by Nissl staining. Interestingly, we found that barrel wall-like structures were completely missing in the entire Cr KO barrel cortex (Fig. 2F). These data indicate that in addition to the TCAs endings, the L5a CR + dendritic tree is also very important for the organization of barrels and the formation of barrel walls.

The ratio of barrel/septum size is decreased after Cr deletion
To further investigate the effect of Cr deletion in L5a on the maturation of the barrel/septum microcircuit, we explored the morphology of the barrel eld by immunostaining of vGlut2 at P30 when barrel cortex development was completed [43]. Control mice exhibited well-arranged barrels (Fig. 3A). Cr KO mice still displayed a disrupted arrangement pattern; furthermore, the individual barrel was more dispersed and indistinct (Fig. 3A). Nissl staining showed that at P30 in control mice, barrel walls clearly formed a "thin" wall. In Cr KO mice, barrel walls generally formed and were positioned in the expected area; however, it seemed that L4 spiny stellate neurons were distributed more broadly in Cr KO mice than in control mice (Fig. 3B), consistent with the widened intervals viewed by vGlut2 staining (Fig. 3A). This phenotype was further con rmed by DAPI staining (Fig. 3C).
We prepared attened tangential cortical slices to examine the holonomic barrel eld by double immunostaining of vGlut2 with DAPI [47,51], and the entire barrel eld was reconstructed and analyzed using ImageJ software (Fig. 3E, F). We calculated and compared the size of the entire barrel eld and found that there was no remarkable difference in the size of the entire barrel eld in Cr KO mice compared with control mice (Fig. 3G). We then explored the size of individual barrels and septa within the major mystacial whisker barrels (from A2 to E4). Our results showed that the barrel/septum ratio was signi cantly increased (Fig. 3H), indicating that the individual barrel columns were contractible and that the septa columns were correspondingly outstretched. This change was further con rmed by the decreased barrel/ (barrel + septum) ratio (Fig. 3I) and increased septum/ (barrel + septum) ratio (Fig. 3J). Taken together, these results indicate that CR is required for the normal formation of barrels and septum in the barrel cortex due to its regulation of the development of L5a pyramidal neuron dendrites.
Both membrane excitability and excitatory synaptic transmission are increased in Cr KO L5a pyramidal neurons As an intrinsic Ca 2+ buffer, CR plays an important role in the regulation of neuronal excitability and neurotransmitter release [25,52]. We next performed whole-cell patch-clamp recording on RFP + neurons using acute brain slices to investigate the effect of Cr deletion on the maturation and excitability of L5a pyramidal neurons. We investigated intrinsic cell electroresponsiveness through current-clamp recordings. The resting membrane potential was recorded immediately after perforating the cell membrane and was found to be comparable between control and Cr KO neurons (Fig. 4A) [37]. However, a signi cant decrease in the action potential current threshold was detected after Cr deletion (Fig. 4B). Moreover, the mean input resistance and amplitude of the AHP were signi cantly higher in Cr KO neurons than in control neurons (Fig. 4C, D). Although the action potential half-width showed a slight tendency to decrease, there was no statistical signi cance between control and Cr KO neurons (Fig. 4E). Unsurprisingly, in response to a series of suprathreshold depolarizing current injections with amplitudes ranging from −50 to 300 pA (with an increment of 50 pA), the number of action potentials recorded from Cr KO neurons was signi cantly higher than that of the control neurons (Fig. 4F, G).
To assess the functional consequences of Cr deletion on synaptic transmission, we next tested the basic synaptic transmission of L5a pyramidal neurons and compared them with the amplitude and frequency of spontaneous miniature EPSCs (mEPSCs) (Fig. 4H) [53]. BMI and TTX were applied to block GABA receptor-mediated inhibitory currents and action potential-dependent synaptic transmission, respectively. We found that the mean amplitude of mEPSCs was unaffected (Fig. 4I), while the mean frequency was strongly increased in Cr KO neurons compared with control neurons (Fig. 4J). Collectively, these results suggest that loss of Cr leads to increased membrane excitability and excitatory synaptic transmission of L5a pyramidal neurons.
Cr KO mice exhibit pronounced exploratory behavior de cits Rodents use their whiskers as multipurpose organs for behaviors ranging from object detection, including object localization, judgment of shape and texture, and discrimination, to movement coordination, such as detecting distance and motor coordination [1][2][3]6]. Since L5a plays a crucial role in integrating information resources and coordinating the movement of the whiskers [9,30], we performed a series of behavioral tests related to the barrel cortex. First, we conducted an open-eld test to evaluate spontaneous motor ability and found that the mean velocity and the total distance traveled within a 30min duration were comparable between Cr KO and control mice (Fig. 5A, B), indicating that locomotor activity was unaffected. However, within the rst ve minutes, the time spent in the center zone and the frequency of entering the center zone were obviously decreased in Cr KO mice (Fig. 5C, D). This result suggested two possibilities: increased anxiety or decreased desire to explore.
To further examine the level of anxiety, we conducted elevated O-maze and elevated plus maze tests [34,54]. As shown in Figure 5E and F, there were no signi cant differences in the time spent in the open arms between the control and Cr KO mice, demonstrating that loss of Cr had no in uence on anxiety levels. To test the exploratory behaviors, a novel object was introduced to the center of the open eld after the mice became fully familiar with the environment [40]. As shown in Figure 5G, the trajectories of control mice were more concentrated in the central area around the novel object than those of Cr KO mice. Although the frequency of entering the center zone was comparable (Fig. 5H), Cr KO mice spent signi cantly less time investigating the object than control mice (Fig. 5I). Taken together, these data indicate that deletion of Cr in L5a has no effects on spontaneous motor ability but impairs exploratory behaviors.
Cr KO mice display defects in whisker-associated tactile sensation behavior To further investigate exploratory behavior de cits, we performed an "S" curve test. This test is designed to simulate a narrow nocturnal environment. When the mice were introduced into an unfamiliar curve, they explored the curve with their whiskers while moving forward [6]. Cr KO mice spent more time going through the "S" curve to reach their destination (incubation) than the control mice (Fig. 6A); moreover, Cr KO mice also spent more time exploring the opening area near the exit before leaving the curve (Fig. 6A).
To measure tactile responses after Cr deletion in L5a, we performed a sticky paper test [39,41]. This test is designed to measure tactile responses to adhesive paper stuck on the palmar surface of mouse hind paws. Mice were required to detect the tape with the help of whiskers and then to remove it. The results showed that Cr KO mice spent more time nding the tape (incubation); accordingly, they spent more time removing the tape from their hind paws (Fig. 6B).
Next, we examined accurate recognition ability through a texture discrimination test [1,55,56]. Mice were rst allowed to move freely to explore different textures in the open eld box. Ten minutes later, two glass bottles, one of which was wrapped in sandpaper, were placed in the opposite corners of the box. The time that mice stayed in each corner in the next ten minutes was assessed. We found that the control mice spent more time in the corner where the unwrapped glass bottle was placed, while the Cr KO mice did not exhibit any preference for the two corners (Fig. 6C), suggesting that the texture discrimination of Cr KO mice was impaired. We also assessed the time that Cr KO mice stayed with two identical bottles and found that they had no position preference (Fig. 6D, E).
Finally, we performed the gap crossing test [39,42], a speci c test to detect the distance perception ability of cortical whiskers. It consists of a series of trials requiring the mice to accurately measure gaps of variable distances and to cross the gaps to reach a safe platform (Fig. 6I). The average gap distance crossed in Cr KO mice was signi cantly shorter than that in control mice (Fig. 6F). Weight data con rmed that the disparity was not related to body size (Fig. 6G). As the gap distances increased, the percentage of mice that were able to cross the gap gradually decreased (Fig. 6H). Collectively, our data suggest that tactile sensation was impaired in Cr KO mice.

Deletion of Cr impairs social novelty preference
In addition to discriminating simple tactile properties, whiskers of rodents also play an important role in social interaction [3,57]. The three-chamber test showed that neither control nor Cr KO mice displayed a preference for either of the two empty chambers during the habituation phase (Fig. 7A), and they all spent more time with the rst stimulating mouse (Fig. 7B), indicating that the loss of Cr had no effect on social ability. In the social novelty test, control mice displayed a preference for the novel mouse, while Cr KO mice did not show any preference (Fig. 7E), suggesting that loss of Cr impairs social novelty preference. In summary, loss of Cr affects the normal formation of the barrel and septa columns in the barrel cortex so that tactile information cannot be processed properly, leading to related behavioral de cits.

Discussion
In this study, we show that deletion of Cr in L5a pyramidal neurons resulted in a decrease in dendritic complexity. Importantly, L5a dendritic de cits subsequently led to abnormal formation of the barrels and barrel walls. Moreover, the membrane excitability and excitatory synaptic transmission of L5a neurons were increased. Cr KO mice exhibited pronounced exploratory and whisker-associated tactile sensation behavioral de cits. Our results demonstrate that CR expressed in L5a is important for the development of the barrel cortex, including both the barrel and septa columns.

L5a and the formation of the barrel and the barrel wall
In the barrel cortex, L5a pyramidal neurons are involved in both the barrel and septa circuit and considered to be an important integration site of the lemniscal and paralemniscal pathways [15,17]. The development of the barrel cortex is an activity-and experience-dependent process [22,44,45]. L5a pyramidal neurons receive a strong direct input from L4 barrels and preferentially establish synaptic connections with cells in the L4 barrel wall region [14,16,18]. Previously, we reported that CR is dynamically expressed in L5a pyramidal neurons during the key developmental window of the barrel cortex [30]. In this study, we found that the length and complexity of L5a pyramidal neuron dendrites were signi cantly decreased in Cr KO mice. Moreover, in contrast to the dense barrel wall observed in the control mice, L4 spiny stellate cells could not properly gather, leading to undetectable barrel walls at P8 and thickened barrel walls at P30 in Cr KO mice. To our knowledge, this study is the rst to morphologically show changes in the accumulation of L4 spiny stellate cells, which may be a direct consequence of failed/fewer connections between L5a and L4 neurons. It will be interesting to explore how the information ow contributes to the formation of the barrel wall.
CR in the excitability of L5a pyramidal neurons Ca 2+ is required for numerous cellular functions [25]. As a Ca 2+ -binding protein, CR regulates intracytoplasmic Ca 2+ concentration though direct "binding"; moreover, as a Ca 2+ sensor, CR controls the distribution of Ca 2+ though its spontaneous activity, thus affecting a series of physiological activities [26][27][28]. Here, we show that deletion of Cr led to an increase in both the neuronal excitability and the excitatory synaptic transmission of L5a pyramidal neurons. Thus, our results are congruent with previous reports of increased neuronal excitability in cerebellar granule cells lacking CR.
Abnormal exploratory behavior and whisker-associated tactile sensation behavior in Cr KO mice As nocturnal animals, mice navigate in the dark and explore objects largely with their whiskers [2,3]. The whisker-barrel cortex system executes tasks such as estimating spatial orientation, object positioning and texture discrimination as well as gap distance measurement [4,42]. The afferent information from the lemniscal and paralemniscal pathways converges on L5a pyramidal neurons [9]. L5a pyramidal neurons also receive encoded information from other laminae and convey information to the secondary somatosensory (SII) cortex, primary motor (MI) cortex, contralateral barrel cortex and subcortical motor regions [6,9]. In this study, Cr KO mice exhibited exploratory behavior de cits and were unable to estimate gap distance accurately. They also show impaired tactile sensation. These ndings expand our knowledge of L5a pyramidal neurons in the whisker-barrel cortex system.