Impaired Terminal Differentiation of Hippocampal Granule Neurons and Defective Contextual Memory in PC3/Tis21 Knockout Mice

Neurogenesis in the dentate gyrus of the adult hippocampus has been implicated in neural plasticity and memory, but the molecular mechanisms controlling the proliferation and differentiation of newborn neurons and their integration into the synaptic circuitry are still largely unknown. To investigate this issue, we have analyzed the adult hippocampal neurogenesis in a PC3/Tis21-null mouse model. PC3/Tis21 is a transcriptional co-factor endowed with antiproliferative and prodifferentiative properties; indeed, its upregulation in neural progenitors has been shown to induce exit from cell cycle and differentiation. We demonstrate here that the deletion of PC3/Tis21 causes an increased proliferation of progenitor cells in the adult dentate gyrus and an arrest of their terminal differentiation. In fact, in the PC3/Tis21-null hippocampus postmitotic undifferentiated neurons accumulated, while the number of terminally differentiated neurons decreased of 40%. As a result, PC3/Tis21-null mice displayed a deficit of contextual memory. Notably, we observed that PC3/Tis21 can associate to the promoter of Id3, an inhibitor of proneural gene activity, and negatively regulates its expression, indicating that PC3/Tis21 acts upstream of Id3. Our results identify PC3/Tis21 as a gene required in the control of proliferation and terminal differentiation of newborn neurons during adult hippocampal neurogenesis and suggest its involvement in the formation of contextual memories.


Introduction
PC3, also known as Tis21 or BTG2 (in rat, mouse and human, respectively; see for review [1,2]), originally isolated as gene induced by nerve growth factor, negatively controls a cell cycle checkpoint at the G1 to S phase transition in fibroblasts and neuronal cells by direct inhibition of the activity of cyclin D1 promoter [3,4,5]. A number of studies in vivo have shown that PC3/Tis21 expression is associated with the neurogenic asymmetric division in neural progenitor cells [6,7,8,9,10]. Moreover, when overexpressed in neural cells, PC3/Tis21 promotes their differentiation [11,12]. Such pro-differentiative action appears to be consequent not only to inhibition of cell cycle progression but also to a PC3/Tis21-dependent activation of proneural genes in neural progenitor cells [12]. PC3/Tis21 has been shown to regulate and associate with the promoters of cyclin D1 and RARb, suggesting that it acts as a transcriptional cofactor, being part of transcriptional complexes [13,14].
The expression of PC3/Tis21 has also been detected in the dentate gyrus of adult hippocampus, in type-2 progenitor cells as well as in differentiated neurons [15]. Recently, by means of a transgenic mouse over-expressing PC3/Tis21 in adult hippocampal progenitor cells we have shown that PC3/Tis21 accelerates their differentiation, without affecting the final number of differentiated neurons. We found that the synaptic plasticity in the dentate gyrus and the performance in different hippocampus-dependent spatial learning and memory tests was severely reduced. This suggested that the time the young neurons spend in different states of neuronal differentiation is critical for their ultimate function in learning and memory [16].
The hippocampus is known to be required in the processing of spatial and contextual memories [17,18]. A specific role in this context appears to be played by the new neurons continuously generated during adulthood from progenitor cells in the subgranular zone of the dentate gyrus. In fact, adult hippocampal neurogenesis appears to be necessary for hippocampus-dependent learning and memory, as indicated by severe cognitive deficits following ablation of neurogenesis in mice models by toxins, x-ray irradiation or virus-activated pro-drugs [19,20,21]. As suggested by several studies, the new adult neurons might be used for storing new memories, thus protecting older memories from interference, or for encoding temporally proximal events [22,23,24,25]. Nonetheless, the ablation of new neurons did not affect hippocampusdependent tasks in a number of studies [26,27], raising questions about the precise role of new neurons in learning and memory. Moreover, the molecular mechanisms that coordinately govern in newborn neurons the processes of proliferation, differentiation and integration into the memory circuits are still incompletely understood.
Thus, in this study we wished to assess whether PC3/Tis21 exerted a physiological role in the control of neurogenesis. In fact, during brain development PC3/Tis21 is expressed in the proliferating neuroblasts of the ventricular zone of the neural tube, and to a lower extent in the differentiating neuroblasts of the mantle zone; postnatally it is expressed in cerebellar precursors mainly in the proliferating regions of the neuropithelium (i.e., in the external granular layer), and in the hippocampus in proliferating and differentiating progenitor cells [6,12,16,15]. Hence, given that PC3/ Tis21 appears to be a pan-neural gene that marks neural progenitors and induces their differentiation, its genetic ablation is expected to interfere with the ongoing neurogenesis. Indeed, we found that the terminal differentiation of adult-generated hippocampal granule neurons is specifically dependent on PC3/Tis21 expression, being impaired in mice ablated of PC3/Tis21. Such impairment was accompanied by a deficit of contextual memory in fear-related tests, suggesting that terminal differentiation of new neurons can be critical for associative memory.

PC3/Tis21-Null Mice and Genotyping
The PC3/Tis21 knockout mice had been generated previously, as described [28]. Mutant mice were of the C57BL/6 (B6) strain and had a replacement of the entire exon II of the Tis21 gene (referred to as PC3/Tis21 throughout this paper), which spans from aa 49 to 158 and includes the boxes A and B, with the neomycin resistance cassette. Genotypization of mice was routinely performed by PCR, using genomic DNA from tail tips. Two pairs of primers were used to identify mice carrying the different alleles PC3/Tis21 2/2 , PC3/ Tis21 +/2 or PC3/Tis21 +/+ , one complementary to the neo cassette and another complementary to the targeted exon II, and were amplified together in the PCR reaction to obtain different patterns of amplification specific for each of the three combinations of alleles: neo(+) 59-AGGCTATTCGGCTATGACTG-39, neo(2) 59-GT-CAAGAAGGCGATAGAAGG-39 (712 bp amplification); exII(+) 59-CATCCAAGGTTCTGGCTATC-39, exII(2) 59-GCCATCA-CTAGTTCTTCGAG39 (270 bp amplification, i.e., the whole exon II).
Mice were maintained under standard specific-pathogen-free conditions, and underwent behavioral testing during the second half of the light period (between 2:00 and 5:00 p.m.) in soundinsulated rooms. All animal procedures were completed in accordance with the Istituto Superiore di Sanita' (Italian Ministry of Health) and current European (directive 86/609/ECC) Ethical Committee guidelines.

BrdU Treatment of Mice and Sample Preparation for Immunohistochemistry
Postnatal day 60 (P60) PC3/Tis21 2/2 and PC3/Tis21 +/+ mice were analyzed after treatment with five daily injections of bromodeoxyuridine (BrdU; 95 mg/kg i.p.) to detect dividing adult neurons in dentate gyrus, in the subventricular zone, and in the olfactory bulb [16]. To detect new progenitor cells entering in Sphase in the dentate gyrus, P60 mice were analyzed after 2 hours treatment with BrdU, while P14 mice after one hour treatment, according to previous protocols [29,30]. New progenitor cells entering in S-phase in the subventricular zone were detected after 2 hour treatement with BrdU, following previous protocols [31]. Brains were collected after transcardiac perfusion with 4% paraformaldehyde (PFA) in PBS -DEPC and kept overnight in PFA. Afterwards, brains were equilibrated in sucrose 30% and cryopreserved at 280uC.

Quantification of Cell Numbers and Volumes
Stereological analysis of the number of cells was performed on one-in-six series of 40-mm freefloating coronal sections (240 mm apart), which were analyzed by confocal microscopy to count cells expressing the indicated marker throughout the whole rostrocaudal extent of the dentate gyrus, of the subventricular zone and of the olfactory bulb. To obtain the total estimated number of cells within the dentate gyrus, positive for each of the indicated markers, the average number of positive cells per section was multiplied by the total number of 40-mm sections comprising the entire dentate gyrus, as described [32,33,34,16]. Three animals per group were analyzed. The same procedure was used to measure cell numbers in the subventricular zone and in the olfactory bulb.
Stereological analysis of the volumes and of the absolute number of granule cells in the dentate gyrus was performed analyzing every eighth section in a series of 40-mm coronal sections, thus spaced 320 mm. Total cell number was obtained according to the optical disector principle, by systematic sampling of counting frames of 15-mm side in each section. Nuclei (stained with Hoechst 33258) that appeared in the different focal planes of the frame were included in the count, while nuclei in the uppermost focal plane of each section and intersecting the exclusion boundaries of the counting frame were excluded, as defined by the optical disector principle [35]. Total cell number (N) was calculated using the formula N = Nv x Vref, where Nv is the average cell number per disector volume (corresponding to 15615640 mm3) and Vref (reference volume) is the total volume of the dentate gyrus. The reference volume was obtained multiplying the sum of the traced areas of the dentate gyrus or hippocampus by the 320 mm distance between sections analyzed [35]. Labeled cells and areas were measured by computer-assisted analysis using the I.A.S. software (Delta Systems, Rome, Italy).
Images of the immunostained sections were obtained by laser scanning confocal microscopy using a TCS SP5 microscope (Leica Microsystem). Analyses were performed in sequential scanning mode to rule out cross-bleeding between channels.

Chromatin Immunoprecipitation (ChIP)
Regarding ChIP assays, purification of nuclei, preparation of chromatin by micrococcal nuclease digestion (to obtain fragments of predominantly one to five nucleosomes) and immunoprecipitation with anti-PC3/Tis21 antibody, were performed as described previously, with minor modifications [37]. The anti-PC3/ Tis21 antibody used for immunoprecipitations was A3H [4], while normal rabbit serum was used as control.
Approximately 4610 7 tet-off-PC3 PC12 cells, conditionally expressing PC3 and harvested at the stages of differentiation indicated, were used for each immunoprecipitation. The immunoprecipitated DNA and 1/200 dilution of the input DNA were analyzed in triplicate samples by real-time PCR using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA) and a 7900HT System (Applied Biosystems). The immunoprecipitated DNA promoter was quantified using the 2 -Ct formula and was calculated as input percentage (ratio of the average value of the DNA detected in immunoprecipitated samples to the average value of the DNA present in input lysates; as described by Heard et al., [38]). For each cell treatment we calculated in parallel the input percentages of DNA immunoprecipitated by immune and by normal rabbit serum. The amount of DNA promoter immunoprecipitated was finally expressed as fold enrichment, meant as ratio of input percentage of DNA immunoprecipitated by A3H antibody to input percentage of DNA immunoprecipitated by normal rabbit serum.

RNA Extraction and Real Time-PCR
Total cellular RNA was extracted from tet-off-PC3 PC12 cells and reverse-transcribed as described previously [5]. Real-time PCR was performed with a 7900HT System (Applied Biosystems) using SYBR GreenI dye chemistry in duplicate samples. Relative quantification was performed by the comparative cycle threshold method [39]. The expression values of PC3 and Id3 in PC12 cells and in the dentate gyrus were normalized to TATA-binding protein RNA, which was used as endogenous control.

In Situ Hybridization
Preparation of sections and hybridization were performed as reported previously [12]. Antisense riboprobe detecting Id3 mRNAs was synthesized by SP6 polymerase from the 39 UTR region of mouse Id3 cDNA [40], a 335 bp long sequence part of the third exon of Id3 and devoid of any cross-homology to Id1 and Id2 sequences, previously cloned by us into the pcDNA3 vector, and restricted by HindIII. The pcDNA3-Id3 construct was obtained by cloning the 39 UTR Id3 cDNA region -amplified using genomic mouse DNA as template -into the EcoRI59-HindIII39 sites of the pcDNA3 vector, and was checked by sequencing. Riboprobes were labeled with digoxigenin-UTP (Transcription kit; Roche Products), following the protocol of the manufacturer. No signal was detected by sense probes.

Behavioral Tests
Male mice (2 months of age; 12 per group of each genotype) were used for behavioral evaluation. The Morris water maze [41] was carried out in a circular swimming pool of 1.3 m in diameter, filled with opaque water at the temperature of 2561uC and located in a room containing prominent extra-maze cues. A hidden 15-cm-diameter platform was used. The training consisted of 24 trials (4 trials per day, lasting a maximum of 60 sec, with an intertrial interval of 30 min), with the platform left in the same position. A probe test (60 sec) was carried out 24 h after the last trial by removing the platform from the pool. Mice behavior was evaluated by EthoVision software (Noldus Information Technology, Wageningen, NL). Contextual fear conditioning [42], was carried out in a conditioning chamber (26622618 cm) made of transparent Plexiglas with a grid metal floor, located in a soundinsulated box lighted by a white tensor lamp (60 W). After an acclimatizing period lasting 180 sec, a foot-shock was delivered (unconditioned stimulus, US; 0.7 mA, 2 sec). Mice were left in the conditioning chamber for a further period of 60 sec and then returned into their home cage. Contextual test (5 min) was performed 24 h after training, in the same chamber. Freezing was defined as the complete absence of motion, except for respiratory movements.
Step-through inhibitory avoidance (passive avoidance) was tested in an apparatus previously described [43]. A straight alley was divided into two compartments, one 7.5 cm long and the other 14 cm long. The floor was 2.5 cm wide and the top 10 cm wide. The smaller compartment was made of white Plexiglas. The larger one was made of black Plexiglas and was equipped with a removable cover of the same material to allow the compartment to be in darkness. The two compartments were separated by a sliding door. A tensor lamp (60 W, positioned 80 cm above the apparatus) illuminated the small compartment. The floor of the larger compartment consisted of two oblique stainless-steel plates separated by 0.5 cm at the bottom and 10 cm at the top, through which scrambled constant current could be delivered. The shape of the electrified floor ensured that the mouse made contact with both plates simultaneously in order to receive the shock. On the training day, each mouse was placed in the light compartment, facing away from the dark compartment. When the mouse turned around, the door leading to the dark compartment was opened. When the mouse had stepped with all four paws into the dark side, the door was closed, the step-through latency was recorded, and two foot shocks (0.4 mA, 50 Hz, 2s, 5s ISI) were delivered. Twenty-four hours after training mice were tested in the same apparatus with a similar procedure to that of training, except that no foot shocks were administrated. A maximum step-through latency of 240s was allowed in the test sessions.

Id3 Promoter Activity
Tet-off-PC3 PC12 cell cultures expressing PC3 or not, seeded in 35-mm polylysine-coated dishes containing 5610 5 cells, were transfected the following day with the pGL3-Id3-prom/-1592 reporter construct using the Lipofectamine reagent, and harvested after 48 hours, being treated in parallel with NGF for the indicated times. Luciferase assays were performed by the Luciferase assay system (Promega) according to the manufacturer's instructions as previously described [12]. Luciferase activities were measured in luciferase units per microgram of protein normalized to the activity of the coreporter pRSV-b-galactosidase present in each extract, as a measure of the efficiency of transfection.
The pGL3-Id3-prom/-1592 construct was obtained by cloning in the 59SacI-39BglI site of pGL3-basic the PCR-amplified region of the Id3 promoter (1592 nt before transcrition start), using genomic mouse DNA as template. The construct was checked by sequencing.

Laser Capture Microdissection and RNA Extraction
One-in-ten series of 30-mm freefloating coronal sections (300 mm apart) comprising the entire hippocampus, obtained from 4% PFA-fixed brains, following dehydration in 100% ethanol and re-hydration in descending alcohols, were incubated in 0.02% Cresyl Violet, sampled on Rnase-free membrane mounted metal frame slides (Molecular Machines & Industries GmbH, Eching, Germany) and dried in a fume hood (1 h) at room temperature, in order to be processed by laser capture microdissection according to a described protocol [44]. Histological slides were examined with an inverted Nikon microscope (Nikon Eclipse TE2000-S) coupled with a UV laser (337 nm) unit and controlling software (SL Microtest GmbH, Jena, Germany) and a CCD camera (Sony, Japan). The dentate gyrus was visualized at 4-20X magnifications and the laser path was directed to the Nissl-stained granule cell layer outline in all the processed sections. The dissection of an equal number of bilateral areas from both wild type and knockout mice was checked by examination of the adhesive lid at the end of every laser capture microdissection session. Tubes were stored at 280uC until RNA extraction.
The Absolutely RNA FFPE kit (Agilent Technologies-Stratagene, CA, USA) was used to isolate RNA from laser capture microdissection samples, following manufacturer's protocol for formalin fixed tissues, excluding any deparaffinizing procedure. The eluted, DNase treated RNA was further precipitated by adding 1 mg/ml glycogen and finally dissolved in 10 ml of DEPC water.

Statistical Analysis
One-way ANOVA was used to analyze the levels of freezing in the contextual fear conditioning and in the inhibitory avoidance. Morris water maze results were analyzed by two-way ANOVA. Student's t-test was used to analyze number of neurons and hippocampal volumes mRNA levels and Id3 promoter activity.

The Development of Adult Hippocampal Granule Precursors Is Arrested before Terminal Differentiation in the Absence of PC3/Tis21
We sought in first place to assess whether the development of hippocampal progenitor cells is dependent on PC3/Tis21 expression. To this end, we used a loss-of-function approach, analyzing in mice ablated of PC3/Tis21 the maturation of progenitor cells of the dentate gyrus, in the adult hippocampus (P60), or in the developing hippocampus (at P14), when the population of differentiating new progenitor cells is massive [45].
Thus, mice with a null mutation of PC3/Tis21, generated by deletion of the second exon that carries the homology boxes A and B, i.e., the active domains of the protein [28,1,2], were treated at P55 with five daily injections of BrdU, and immediately after the different cell populations of the dentate gyrus were analyzed. Such BrdU treatment allows to detect new progenitors and neurons 1-to 5-day-old. The process of adult hippocampal neurogenesis from putative neural stem cells to post-mitotic granule neurons has been divided in six developmental stages [46]. The cell populations of the dentate gyrus are believed to originate from the putative neural stem cells with radial glial-like morphology identified by the expression of GFAP in their processes [47], expressing also nestin or Sox2 and defined as type-1 cells [48,49,50,51].
As a whole, this indicates that the development of the dentate gyrus neurons undergoes an arrest at the the end of stage 5, so that only few neurons attain terminal differentiation and stage 6. Consistently, we observed in PC3/Tis21-null mice a significant decrease of hippocampal granule neurons expressing NeuroD1, which is required for their terminal differentiation ( [55]; total NeuroD1 + neurons: about 15% decrease, p = 0.0011, Figure S1A; BrdU + /NeuroD1 + neurons: about 22% decrease, p = 0.020, Figure S1B; BrdU + /NeuroD1 + /NeuN + neurons: about 35% decrease, p = 0.0027, Figure S1C; a representative image is shown in Figure S1D). No difference was observed in glia, identified as GFAP + cells ( Figure 1J), indicating that the impairment of neuron differentiation did not result in altered gliogenesis.
As a next step, we checked whether the impairment of differentiation was detectable also in older neurons, of 28 days of age. Thus, we analyzed the hippocampi of P60 mice, four weeks after treatment with five daily injections of BrdU.
Altogether, the neurons of 4 weeks of age in PC3/Tis21-null hippocampi showed a delay in attaining terminal differentiation, as observed at 1-5 days of age, indicated by the increase of the population of neurons expressing the immature markers Calretinin and also DCX, in concomitance with a decrease of the terminally differentiated Calr 2 /NeuN + or Calb + /NeuN + neurons.

In the Absence of PC3/Tis21 Dentate Gyrus Progenitor Cells Show Enhanced Proliferation and Shortened Duration of G1-Phase
The progenitor cells in S-phase, identified by incorporation of a short BrdU pulse of 2 hours [29,30], as well as the total number of dividing progenitor cells, identified by the proliferation marker Ki67 [56], increased significantly in mutant mice (BrdU + , 33% increase, p = 0.001, Figure 3A; Ki67 + , 25% increase, p = 0.0009, Figure 3B). Representative images of BrdU + and Ki67 + cells are shown in Figure 3I. This indicated that the absence of PC3/Tis21 led to an increase of proliferation of neural progenitor cells. Furthermore, the fraction of progenitor cells in S-phase (BrdU + ) among those actively dividing (Ki67 + ) increased significantly in mutant mice (BrdU + vs Ki67 + ratio, 23% increase, p = 0.008, Figure 3C); this ratio is inversely proportional to the length of the cell cycle [57,58] and indicated that the G1 phase of dentate gyrus progenitor cells was shortened in mutant mice. In contrast, the fraction of progenitor cells in G2/M-phase (positive to the mitotic marker antiphospho-histone H3 [PH3]; [59]) among those actively dividing (Ki67 + ), did not differ in mutant and WT mice, confirming that the decrease of cell cycle length concerned the G1 phase (PH3 + vs Ki67 + ratio, p = 0.84, Figures 3D and 3J).
In conclusion, we observed that in PC3/Tis21-null mice -in parallel with an accumulation of dentate gyrus neurons not terminally differentiated (stage 5, up to 4-week-old) and with a reduction of terminally differentiated neurons (stage 6) -the type-3 progenitor cells, that originate stage 5 neurons, showed an increase of proliferation and number. This indicates a delayed exit from cell cycle of progenitor cells and a delay of terminal differentiation, consistent with the ability of PC3/Tis21 to inhibit cell cycle and promote differentiation [4,12,16]. Furthermore, we analyzed dentate gyrus progenitor cells and neurons at an earlier developmental stage (P14), when the population of proliferating progenitors is still expanding and begins to increase in the subgranular zone of dentate gyrus, while decreases in the tertiary matrix [61]. We used Sox2 to identify progenitor cells, where it is expressed during embryogenesis and in adult brain [50,62], and Ki67 to label the whole population of dividing progenitor cells. We observed a significant increase of progenitor cells entering in S-phase, labeled by BrdU (a short pulse of 1 hour) and by Sox2 (BrdU + / Sox2 + ; 1.6-fold increase, p = 0.002; Figure 4A, E), and of the dividing progenitor cells labeled by Ki67, either alone (Ki67 + ; 18% increase, p = 0.01; Figure 4B) or with Sox2 (Ki67 + /Sox2 + ; 1.5-fold increase, p = 0.02; Figure 4C). In contrast, the whole population of terminally differentiated dentate gyrus neurons decreased significantly (BrdU + /NeuN + ; 20% decrease, p = 0.02; Figure 4D). Moreover, a significant decrease in the P14 dentate gyrus of PC3/Tis21-null mice was observed also in the total number of neurons expressing NeuroD1 (NeuroD1 + , PC3/Tis21 2/2 : 4306563998; PC3/Tis21 +/+ : 8428562395; 3 mice analyzed per genotype; p = 0.0000, Student's t test).
Thus, the delay of terminal differentiation of neurons and the increase of progenitor cell proliferation observed in adult dentate  gyrus was not restricted to adulthood, being detectable also at an earlier developmental stage.
We also evaluated the survival of cells in the dentate gyrus, and observed that in PC3/Tis21-null mice the total number of apoptotic cells was greater than in WT, as detected by positivity to Caspase-3 (1.6-fold increase; p = 0.03; Figure 5A). Such increment, analyzed as percentage of cells undergoing apoptosis, was evident in the progenitor cell types whose proliferation increases significantly, i.e., type-2b and type-3 progenitor cells positive for DCX (Caspase-3 + /DCX + /NeuN 2 ; p = 0.0001; Figure 5B, E). On the other hand, we observed a not significant increase, of about 26%, of the percentage of stage 5 neurons undergoing apoptosis (Caspase-3 + /Calret + /NeuN + ; p = 0.41; Figure 5C). In stage 6 neurons, i.e., cells that had terminally differentiated, apoptosis was instead significantly less frequent in Ti21-null mice than in WT mice (Caspase3 + /Calr 2 /NeuN + ; 54% decrease; p = 0.01; Figure 5D, E). This suggests that the stage 6 neurons of PC3/Tis21-null mice, being in lower number than in WT (about 38% less than WT, Figure 2C), underwent a less competitive selection for survival [63].
Thus, an evident increase in the frequency of apoptosis occurred only in type-2b/type-3 proliferating progenitor cells, whose proliferation in PC3/Tis21-null mice was more pronounced than in control mice.
Furthermore, to test whether nonspecific changes occurred in mutant mice, we conducted a stereological analysis of the hippocampus volumes and cell numbers. No difference was observed between PC3/Tis21-null and WT mice at P60 and P14 in the volumes of the dentate gyrus or of the whole hippocampus, or in the total cell number of the dentate gyrus (see Table 1).
Additionally, we were interested in assessing whether the ablation of PC3/Tis21 affected the maturation of the neurons of the subventricular zone, the other adult neurogenic brain region [64]. We observed that the number of new progenitor cells and neurons of the subventricular zone in P60 mice entering the Sphase, identified as BrdU + cells after a BrdU treatment of 2 hours, increased significantly in Tis-21 null mutant (PC3/Tis21 2/2 : 128726605; PC3/Tis21 +/+ : 104696510; 3 mice analyzed per genotype; p = 0.002, Student's t test). In contrast, we observed a considerable reduction of the final number of differentiated neurons generated in the subventricular zone, as detected by measuring BrdU + /NeuN + cells 4 weeks after their birth in the olfactory bulb, i.e., in their final migratory destination (PC3/Tis21 2/2 : 3272561346; PC3/Tis21 +/+ : 3841061269; 3 mice analyzed per genotype; p = 0.007, Student's t test). This indicated that PC3/Tis21 was required for the control of proliferation as well as differentiation of adult progenitor cells in the subventricular zone, similarly to what observed in dentate gyrus.

PC3/Tis21-Null Mice Show Normal Spatial Memory while Are Impaired in Contextual Fear Conditioning
As a further step, we sought to ascertain whether the delayed differentiation of dentate gyrus newborn neurons affected the hippocampus-dependent memory. Locomotor abilities and anxiety levels were preliminarily analyzed by submitting 2-month-old PC3/Tis21-null and WT mice to open field and plus maze tests [65,66]. No statistically significant differences were found among the groups in both tasks ( Figure S3). Then, spatial learning was tested in the Morris water maze, a task which is largely dependent on an intact hippocampus [41,67].
In this task, mice learn across daily sessions to find a hidden escape platform using extra-maze visual cues. PC3/Tis21 2/2 and control mice performed equally in the acquisition phase ( Figure 6A). Statistical analysis (two-way ANOVA) showed a We then investigated the effects of the ablation of PC3/Tis21 in the contextual fear conditioning task, which involves mainly the hippocampus [68,42]. In this task, immobility (freezing), a natural reaction elicited in mice by aversive stimuli, was recorded and considered as a measure of fear memory. Animals were trained in the conditioning chamber, where a foot-shock was delivered (unconditioned stimulus, US; see Materials and Methods). During the training ( Figure 6C, left), no significant effect of the genotype was observed on the level of freezing (two-way ANOVA, F (1,22) = 1.54; p = 0.23), and all mice reacted alike to the US (F (1,22) = 89.37; p,0.0001). In the contextual test ( Figure 6C, right), performed 24 h after training, PC3/Tis21 2/2 mice spent a significantly smaller amount of time in freezing behavior, compared  to control mice (one-way ANOVA, F (1,22) = 5.02; p = 0.03).
Contextual memory was further investigated by submitting animals to an inhibitory avoidance task. In this task animals learn to avoid the dark compartment of an apparatus in which they receive an aversive stimulus [69]. During the training, before foot-shock administration ( Figure 6D, left), mice showed no significant differences in the latencies to step-through into the dark compartment (F (1,22) = 0.33; p = 0.57). In the test ( Figure 6D, right) carried out 24 h after training, PC3/Tis21 2/2 mice showed significantly shorter step-through latencies compared to that of control mice (F (1,22) = 7.65; p = 0.01).

Stage 5 Neurons of PC3/Tis21-Null Mice Show Selective Incorporation into Dentate Gyrus Memory Networks after Behavioral Tasks
Mice tested for hippocampal memory were then analyzed for c-fos expression in dentate gyrus, to correlate their performance with the incorporation of neurons in spatial memory networks. In fact, c-fos is induced in dentate gyrus neurons that, following spatial memory training, are recruited into spatial memory circuits [34,70].
1.5 hours after the end of the Morris water maze test, the activation of c-fos in the whole population of dentate gyrus neurons was equal in PC3/Tis21-null and WT mice, since no significant difference in number of c-fos + /NeuN + neurons emerged ( Figure 6E). In contrast, 1.5 hours after the end of the contextual fear conditioning, the whole population of c-fos + / NeuN + neurons significantly decreased in PC3/Tis21-null mice, of about 21% (p = 0.006, Figure 6F). Representative confocal images are shown in Figure 6I.
Thus, after the Morris water maze, where no cognitive deficit was observed, an equal number of activated neurons was recruited from the whole neuron population into memory circuits of PC3/ Tis21-null mice, despite the strong decrease of stage 6 terminally differentiated neurons.
As a further step, we evaluated whether the incompletely differentiated stage 5 neurons were integrated into memory circuits after each test, identifying them by c-fos, Calretinin and NeuN. Noteworthy, after the Morris water maze test and also after the fear conditioning test, we detected a significant number of activated stage 5 neurons in PC3/Tis21-null mice, while none, or very few, were present in WT mice (c-fos + /Calretinin + /NeuN + cells; after Morris water maze: p = 0.002, Figure 6G; after contextual fear conditioning: p = 0.000002, Figure 6H).
As a whole, this indicated that: i) the extent of incorporation in memory circuits of neurons from the whole neuron population of dentate gyrus was closely correlated with the performance in the memory tests; ii) the incompletely differentiated stage 5 neurons were surprisingly activated in PC3/Tis21-null mice, after both memory tests. Since stage 5 neurons of 4 weeks of age are present also in WT mice ( Figure 2B), this denotes a selective recruitment of this type of neurons in memory circuits of PC3/Tis21-null mice.

PC3 Is Recruited on the Id3 Promoter and Negatively Regulates Id3 Expression
The marked delay of terminal differentiation of stage 6 neurons occurring in the absence of PC3/Tis21 prompted us to investigate the underlying molecular mechanism. Given that PC3/Tis21 is a transcriptional cofactor, which we have shown to be recruited to the promoter of cyclin D1 and RAR-b [13,14], we asked whether PC3/ Tis21 could bind also to the promoters of neural genes. To verify this possibility, we used a PC12 cell line -able to differentiate into sympathetic neurons upon addition of nerve growth factor (NGF)expressing the tetracycline-regulatable tTA transactivator (PC12/tetoff). These cells were stably transfected with an inducible PC3 gene (rat sequence) under control of a tetracycline responsive element. This system allowed us to induce high levels of exogenous PC3 in the presence or in the absence of the differentiating stimulus, thus circumventing the limit presented by the narrow time-window of induction of endogenous PC3 by NGF (in the range of 1 hour after treatment; Figure 7A). The induction of PC3 mRNA was accompanied by induction of PC3 protein (data not shown). In these cells we analyzed, by chromatin immunoprecipitation (ChIP), the association of the PC3 protein to different neural gene promoters (namely, those of NeuroD1, Hes1, Mash1, Ngn1, Msx1, Id1, Id3; data not shown), observing that a significant association occurred only with the promoter of Id3, a gene that negatively regulates neural differentiation [71,72,73]. Indeed, we found that the amount of Id3 promoter sequences recovered in the PC3 immunoprecipitates increased significantly above basal level both before and after NGF treatment, in correlation with the induction of exogenous PC3 expression ( Figure 7B; the increase of binding above the condition without induction of exogenous PC3 was: 51% in the absence of NGF Figure 6. PC3/Tis21 2/2 mice show selective impairment of contextual memory and preferential recruitment of stage 5 neurons into dentate gyrus after both spatial and contextual behavioral task. (A) Escape latency (sec) throughout the Morris water maze behavioral test, to reach the hidden platform during the 6-day acquisition phase. (B) Time (sec) spent in the quadrants during the probe trial, when the hidden platform, located in the target quadrant (TG) during the acquisition phase, was removed. Both PC3/Tis21 2/2 and PC3/Tis21 +/+ mice spent a significantly greater amount of time in the TG quadrant, compared to the other quadrants. AR, adiacent right; AL, adiacent left; OP, opposite. (C) Contextual fear conditioning test: percentage of time spent in freezing behavior by PC3/Tis21 2/2 and PC3/Tis21 +/+ during both training (left) and test (right). During the training, no significant differences in the level of freezing between groups were observed both before (Pre-US) and after (Post-US) foot-shock administration. During the test, performed 24 h after training, PC3/Tis21 2/2 mice showed a significant reduction of freezing behavior compared to control mice. *, p,0.05, ANOVA. (D) Inhibitory avoidance test: mean latency to step-through into the dark compartment by PC3/Tis21 2/2 and PC3/Tis21 +/+ mice during both training (left) and test (right). During training, no significant differences were observed in the step-through latency between groups. During the test, performed 24 h after training, PC3/Tis21 2/2 mice showed a significant reduction of the step-through latency compared to that of control mice. *, p,0.05, ANOVA. (E) The activation into memory circuits of neurons within the whole dentate gyrus neuronal population (i.e., comprising stage 5 and 6) after the Morris water maze, measured as number of c-fos/NeuN-positive cells, was similar in PC3/Tis21-null and WT mice, while (F) after the contextual fear conditioning was significantly lower in PC3/Tis21-null mice. (G) The activation into memory circuits of stage 5 neurons, measured as number of c-fos/ Calretinin/NeuN-positive cells, occurred quite exclusively in PC3/Tis21-null mice after the Morris water maze test and (H) also after the contextual fear conditioning test (being no c-fos/Calretinin/NeuN-positive neuron detectable in WT mice after the Morris water maze, or only few after the contextual fear conditioning test). This suggests a preferential recruitment in spatial memory networks of stage 5 neurons of PC3/Tis21-null mice. Analyses shown in (E-H) were performed in PC3/Tis21 2/2 and WT mice 1.5 hours after the end of behavioral test, expected to activate responsive neurons inducing c-fos expression. Cell numbers are represented as mean 6 SEM of the analysis of three animals per group. **, p,0.01 or ***, p,0.001 vs. PC3/Tis21 +/+ dentate gyrus; Student's t test. (I) Representative confocal images showing stage 5 dentate gyrus neurons expressing c-fos, following behavioral training by the Morris water maze test, which are detectable only in PC3/Tis21-null mice (c-fos/Calretinin/NeuN-positive cells, in orange in the merged image, are indicated by arrowheads; single labeling is blue, red, green, respectively), or Calretinin/NeuN-positive c-fos-negative cells (indicated by arrows). Scale bar, 75 mm. doi:10.1371/journal.pone.0008339.g006 Figure 7. Binding of PC3/Tis21 to the Id3 promoter and corresponding decrease of Id3 mRNA. (A-D) A PC12 cell line carrying an inducible PC3 (rat) cDNA under control of the tet-off system (i.e., a tetracycline responsive element TRE-PC3, activated by a tetracycline-regulated transactivator, CMV-tTA, in the absence of doxycycline) was used to analyze the binding of PC3 to the Id3 promoter, by chromatin immunoprecipitation (ChIP). (A) PC12 cells without (white columns) or with expression of exogenous PC3 (activated a week before by withdrawal of 2 mg/ml of doxycycline; grey columns) were exposed to NGF (100 ng/ml) for 1 and 48 hours. The levels of exogenous PC3 increased considerably in the absence of NGF (t 0), and even more as the treatment with NGF proceeded. Endogenous PC3 was also transiently i51duced by NGF after 1 hour, as expected (Bradbury et al., 1991). (B) ChIP analysis of PC3 binding to the Id3 promoter in PC12 cells, treated as shown in (A). A scheme above the graph describes the Id3 gene and the promoter region analyzed, 700 nt 59 to the transcription start. The amount of Id3 promoter present in immunoprecipitates obtained using anti-PC3 antibody is quantified by real time PCR and is expressed as fold enrichment (ratio of the percentage of the PC3-immunoprecipitated amount of Id3 promoter detected in the input cell lysates to the percentage of the normal serum-immunoprecipitated amount detected in the input cell lysates). The binding of exogenous PC3 (grey columns) to the Id3 promoter increased in correlation with the increase of exogenous PC3 mRNA levels (panel A). [t 0 hour], p = 0.01; 96% increase 1 hour after NGF addition, p = 0.007; 119% increase 48 hour after NGF addition, p = 0.002). In contrast, the recruitment of the PC3 protein to the muscle creatin kinase (MCK) promoter, which is inactive in PC12 cells and used here as negative control, was not above the background levels ( Figure 7C). In parallel with the increased binding of PC3 to the Id3 promoter, we observed a significant decrease of Id3 mRNA levels upon induction of exogenous PC3, 1 and 48 hours after NGF treatment ( Figure 7D; 26% decrease 1 hour after NGF addition, p = 0.01; 49% decrease 48 hours after NGF addition, p = 0.04). Next, we asked whether PC3 could inhibit the Id3 promoter activity. Therefore, we generated the construct pGL3-Id3-prom/-1592, which carries the Id3 promoter region upstream of the luciferase reporter gene. This construct was transfected into PC12/tet-off cells, treated or not with NGF. Id3 promoter activity was significantly reduced upon induction of exogenous PC3, 1 and 48 hours after NGF treatment, as compared to the activity in the absence of exogenous PC3 ( Figure 7E; 34% decrease 1 hour after NGF addition, p = 0.01; 25% decrease 48 hours after NGF addition, p = 0.03).
Taken together, these data clearly indicate that PC3 can bind the Id3 promoter, and strongly suggest that it may inhibit the transcription of the anti-differentiative gene Id3 through direct binding to its promoter.
We also sought to ascertain whether the regulation of Id3 transcription by PC3/Tis21 observed in differentiating PC12 cells occurred in vivo. To this aim, we analyzed by in situ hybridization the expression of Id3 mRNA in the dentate gyruswhere it is normally expressed [74] -of WT or of PC3/Tis21 2/2 P14 mice, at a stage of high expansion of progenitor cells. We observed that in PC3/Tis21 2/2 mice, as compared to WT, not only Id3 mRNA increased markedly in the proliferative subgranular zone, but also it was highly expressed ectopically in the granule cell layer where normally Id3 is almost undetectable ( Figure 7F). A significant increase of Id3 mRNA in P14 PC3/Tis21 2/2 mice, with respect to WT mice, was observed also in the dentate gyrus microdissected by laser capture and analyzed by real time PCR (2.4-fold increase, p = 0.009; Figure 7G and Figure S4A-E). These data, confirming those acquired by in situ hybridization, are compatible with the results obtained in differentiating PC12 cells, and support the idea that PC3/Tis21 negatively controls the expression of Id3, by inhibiting the Id3 promoter activity.
We further sought to identify the cell populations of the dentate gyrus expressing the Id3 protein in PC3/Tis21 2/2 and WT mice, at P14. The total population of Id3-positive cells within the dentate gyrus of PC3/Tis21 2/2 mice increased considerably in number and in distribution, expanded over the whole granule cell layer up to the border with the molecular layer, consistently with the increase observed above for Id3 mRNA expression (1.31-fold increase; Id3 + , p = 0,0004; Figure 8A, F). New proliferating progenitor cells expressing Id3, identified by Sox2 and BrdU (1 hour pulse), were present in control mice but increased significantly in PC3/Tis21-null mice (1.56-fold increase; BrdU + /Sox2 + /Id3 + , p = 0.0008; Figure 8B, F). Moreover, the frequency (analyzed as percentage) of Id3positive cells among the new proliferating progenitor cells increased significantly, suggesting a tight link between the increase of Id3 expression and the increase in number of progenitor cells observed in PC3/Tis21-null mice (BrdU + / Sox2 + /Id3 + versus BrdU + /Sox2 + ; p = 0.016; Figure 8C). In contrast, we observed a dramatic decrease of the number of new differentiating neurons expressing Id3 (BrdU + /NeuN + /Id3 + ; 64% decrease, p = 0,00006; Figure 8D), and a significant decrease also of the whole population of terminally differentiated neurons co-expressing Id3 and Calbindin (Calbindin + /Id3 + ; 18% decrease, p = 0,04; Figure 8E). Notably, as shown in Figure 8G, in PC3/Tis21-null mice the population of Calbindin-positive cells, normally localized in the outer region of the dentate gyrus, decreased in number (Calbindin + , PC3/Tis21 2/2 : 5605665381; PC3/Tis21 +/+ : 7274061201; 3 mice analyzed per genotype; p = 0.016, Student's t test) and became more restricted toward the outer border of the dentate gyrus, correspondingly to the expansion of the area occupied by Id3-positive cells.
Altogether, this indicates that the increase of Id3 expression is associated to the increase in number of proliferating progenitor cells and is also concomitant with the reduction of neurons attaining terminal differentiation observed in PC3/Tis21-null mice. This is compatible with the idea that the impaiment of terminal differentiation observed in the absence of PC3/Tis21 is mediated by Id3.

Discussion
This report provides the first evidence of a requirement of PC3/ Tis21 for neurogenesis. Previous studies have traced the expression of PC3/Tis21 in neural cells, exploiting a phenotypically wild type Tis21 GFP knock-in mouse, where the GFP sequence was fused to the second exon carrying the active domains of PC3/Tis21 [8,9,15]. Here, in a PC3/Tis21 knock out mouse model, we observe impairment of the last stage of maturation of hippocampal neurons, associated to deficit of contextual fear memory and to reduced incorporation of neurons into memory circuits. Key issues raised from these observations concern the underlying cellular and molecular mechanisms.

Requirement of PC3/Tis21 for Terminal Differentiation of Dentate Gyrus Neurons
The stage 5 hippocampal neuron is a post-mitotic cell type that in normal conditions is present only for a limited period after its birth, since it rapidly evolves into the stage 6 neuron [54]. In fact, in wild type mice we find that stage 5 neurons (BrdU+/Calr+/NeuN+) of 1-5 days of age are a small fraction of the terminally differentiated stage 6 neurons (BrdU+/ Calr2/NeuN+), being less than one third. Moreover, at 28 days of age stage 5 neurons further decrease, their number becoming only 1/16 of stage 6 neurons. In mutant PC3/Tis21 mice we observe that the number of stage 5 neurons of 1-5 days of age increases strikingly (about 2-fold), in comparison with normal mice, and remains higher also at 28 days of age, when the number of stage 5 neurons increases to represent 1/4 of stage 6 neurons. A similar increase is observed also for stage 5 neurons of 28 days of age still expressing DCX (i.e., BrdU+/ DCX+/NeuN+).
Such major increase of stage 5 neurons, together with a significant reduction of terminally differentiated stage 6 neurons, clearly indicates that stage 5 neurons fail to mature into terminally differentiated neurons, and consequently accumulate. Therefore, stage 5 neurons of mutant mice are neurons whose final differentiation is almost indefinitely delayed by the absence of PC3/Tis21 activity.
A similar delay of differentiation is observed also for the other adult neurogenic niche, the subventricular zone, in which neural cells do not attain terminal differentiation in the absence of PC3/ Tis21, as the number of 28-day-old BrdU/NeuN-expressing neurons detected at P60 in the olfactory bulb, i.e., in their final migratory destination, decreases significantly.
As a whole, this suggests that PC3/Tis21 is essential for the process of neuronal maturation. decreased considerably, and (E) a significant decrease was observed also for the total population of terminally differentiated neurons co-expressing Id3 and Calbindin. Cell numbers are represented as mean 6 SEM of the analysis of three animals per group. *, p,0.05 or ***, p,0.001 vs. PC3/Tis21 +/+ dentate gyrus; Student's t test. (F, G) Representative confocal images of the dentate gyrus in P14 mice; (F) The new proliferating progenitor cells expressing Id3, identified also by BrdU and Sox2 (BrdU/Sox2/Id3-positive, indicated by arrowheads; single labeling is red, blue, green, respectively) increase significantly in PC3/Tis21-null mice. It is also evident that the total number of Id3-positive cells increases greatly, with a localization expanded to the whole area of the granule cell layer, up to its outer third (delimited by the white broken line). Scale bar, 45 mm. (G) The whole population of terminally differentiated neurons, identified by Calbindin and co-expressing Id3, in PC3/Tis21-null mice is restricted toward the outer border of the dentate gyrus, in correspondence with the expansion to that area of the Id3-expressing cells. The white broken line delimits the inner and outer borders of dentate gyrus. Scale bar, 45 mm. (H) A model proposed for PC3/Tis21 activity during neuronal differentiation (see Discussion). Delta1 is represented upstream of PC3/Tis21, whereas cyclin D1 downstream, in agreement with previous findings [78,12]. doi:10.1371/journal.pone.0008339.g008 The Control of G1 Phase and of Id3 Promoter Underlies the Role of PC3/Tis21 in Differentiation Moreover, the absence of PC3/Tis21 led to an increase of proliferation, indicated by the greater number of progenitor cells in S-phase -labeled by a short 2 hours pulse of BrdU [29,30] -and of the progenitor cells in active phases of cell cycle, identified by Ki67 [56,75]. Remarkably, also the ratio between progenitor cells in S-phase (BrdU + ) and cycling cells (Ki67 + ) increased, indicating that the cell cycle length was reduced. In fact, the ratio between BrdU-and Ki67-labeled cells accounts for G1 phase length because in mammalian cells the length of the S phase remains relatively constant while the length of G1 regulates proliferation [57,58]. Furthermore, we did not observe any difference in the number of PC3/Tis21-null dentate gyrus cells in G2/M phase, as indicated by the phospho-histone H3 antigen. This reduction of the length of the G1 phase indicates that the control of the G1 phase is a physiological function of the gene PC3/Tis21, possibly exerted, according to our previous findings, by direct inhibition of the activity of the cyclin D1 promoter [5,13].
A primary effect of the reduction of G1 phase length occurring in the absence of PC3/Tis21 may be a delay or inhibition of the neurogenic asymmetric division, to which the expression of PC3/ Tis21 is known to be associated in neuroblasts [6,7,1,10]. Indeed, inhibition of the neurogenic asymmetric division can be the cellular mechanism at the origin of the increased proliferation observed in progenitor cells of mutant mice, which plausibly would be driven to preferentially divide symmetrically, thus generating a greater number of dividing cells.
On the other hand, the inhibition of the neurogenic asymmetric division alone may not account for the impairment of the terminal differentiation of stage 5 neurons into stage 6 neurons, dependent on PC3/Tis21 ablation, given that stage 5 neurons are already post-mitotic. Without excluding the possibility that the action of PC3/Tis21 on the cell cycle and cell division may have an influence on differentiation protracted also after mitosis, other mechanisms, likely involving the activity of proneural genes, should be implicated in the delayed differentiation of dentate gyrus neurons in Tis-21null mice. In this regard, we have observed the direct binding of PC3/Tis21 to the promoter of Id3, which is an anti-differentiative gene [71], accompanied by down-regulation of Id3 mRNA expression and promoter activity, suggesting that PC3 may function as repressor of Id3. Accordingly, we observed an increase of Id3 mRNA levels in the dentate gyrus in the absence of PC3/Tis21, accompanied by an increase of Id3-expressing proliferating progenitor cells and a parallel decrease of Id3expressing differentiating neurons. Id3 has a helix-loop-helix (HLH) dimerization domain, but lacks a DNA-binding domain and acts by sequestering E proteins, thus preventing their association to proneural basic HLH transcription factors, which are consequently inactivated [71,72,73]. One of these is NeuroD1, which is required for the maturation of hippocampal granule progenitor cells in differentiated neurons [55]. Consistently, we have also observed that NeuroD1 expression decreases remarkably in dentate gyrus neurons of PC3/Tis21-null mice, compared to WT mice. Given that PC3/Tis21 has been shown to bind histone deacetylases (HDACs) and to act as an inhibitory transcriptional co-factor [13], we can speculate that PC3/Tis21 may impair the transcriptional activity of Id3 by recruiting these inhibitory molecules to the Id3 promoter, thereby facilitating the terminal differentiation of dentate gyrus neuron. Part of the transcriptional complex associated with the Id3 promoter might also be Smad1 and Smad8 proteins, known to bind PC3/Tis21 and to induce the levels of Id3 protein [28,76,77].
We have previously evidenced a control by PC3/Tis21 on the activity of Math1 promoter in cerebellum [13], but here we present the first demonstration of direct molecular association to the promoter of a neural gene, i.e., Id3.
Remarkably, the expression of endogenous PC3/Tis21, present in type-2/type-3 progenitor cells and in stage 6 (Calbindinpositive) terminally differentiated neurons, but absent in stage 5 neurons [15], follows a pattern fully compatible with the alterations emerging after ablation of PC3/Tis21 in the proliferation of progenitor cells and in the maturation of stage 5 neuron.
Altogether, our data suggest that PC3/Tis21 is required to control the differentiation of hippocampal neurons through a dual regulation of cell cycle and of the activity of neural genes.
Thus, PC3/Tis21 appears to be a non-redundant component of a pro-differentiative pathway, represented upstream by Delta1 [78] and, downstream, by Id3 and its targets, i.e. the basic HLH proneural genes ( Figure 8H).

The Ablation of PC3/Tis21 Causes a Deficit of Contextual Fear Memory and a Selective Recruitment in Memory Circuits of Stage 5 Undifferentiated Neurons
In previous studies adult neurogenesis was suppressed by different approaches, such as irradiation, genetic manipulation or cytostatic drugs [79]. A different paradigm, recently developed by us, targeted new neurons by inducing them to differentiate faster without altering their final number [16]. The majority of these approaches led to a significant memory impairment. In the present model, terminal differentiation of new neurons is prevented, while the proliferation rate of progenitor cells is increased.
This cellular phenotype is associated to different performances in memory tests, being a deficit evident in the contextual fear conditioning and inhibitory avoidance, but absent in the Morris water maze. A possible explanation is that fear-related tests, as one-time associative tasks, are more sensitive to the large decrease, up to 40%, of terminally differentiated stage 6 neurons 1-to 4week-old, than the Morris water maze that, conversely, consists of repeated training trials. Consistently, the total number of c-fos activated neurons was equal in PC3/Tis21-null and WT mice only after the Morris water maze test, suggesting that a functional compensation of the decrease of terminally differentiated neurons occurred in memory circuits of PC3/Tis21-null mice during that test.
Notably, c-fos activation of stage 5 neurons (BrdU/Calretinin/ NeuN-positive) occurred quite selectively in PC3/Tis21-null mice, after both Morris water maze and contextual fear conditioning tests. In fact, stage 5 neurons are present also in WT mice, which suggests that such preferential recruitment in memory circuits of stage 5 undifferentiated neurons may act as a homeostatic mechanism, compensating for the large reduction of stage 6 differentiated neurons in PC3/Tis21-null mice. However, since the undifferentiated stage 5 neurons in PC3/Tis21-null mice are equally recruited after both tests, it is unlikely that these neurons are responsible for the normal performance of mutant mice in the Morris water maze test. Nonetheless, we cannot exclude that, under certain learning conditions such as those of the Morris water maze test in PC3/Tis21-null mice, stage 5 neurons may contribute to recruit a larger number of neurons within memory circuits.
More generally, our data reveal how critical for spatial/ contextual memory and recruitment in memory circuits of newborn neurons is the control of the neural differentiation process in dentate gyrus, not only at the stage of progenitor cells, as previously observed following an acceleration of their differentiation [16], but also for early post-mitotic neurons when their terminal differentiation is delayed, as observed here after genetic ablation of PC3/Tis21. Moreover, given the physiological role in the control of proliferation and differentiation of developing neurons highlighted here for PC3/Tis21, and considering the neuroprotective role assigned to this gene in several systems [80,81,11], future studies should aim at defining whether PC3/Tis21 plays a role in neurodegenerative diseases. In fact, a recent hypothesis implies the control of cell cycle in the onset of neurodegenerative diseases such as Alzheimer's disease, which impacts heavily on the hippocampal function [82]. Figure S1 Decrease of neurons expressing NeuroD1 in the dentate gyrus of PC3/Tis21-null mice. (A) In P60 PC3/Tis21-null mice the total number of NeuroD1-positive cells was significantly reduced, of about 15%. (B, C) Consistently, new neurons 1-to 5day-old expressing NeuroD1, identified either as (B) BrdU/ NeuroD1-positive, or as (C) BrdU/NeuroD1/NeuN-positive neurons, presented a significant decrease (of 22% and 35%, respectively). Immunoistochemical analysis was performed after five daily injections of BrdU. Cell numbers are represented as mean 6 SEM of the analysis of three animals per group. *, p,0.05, or **, p,0.01 vs. PC3/Tis21+/+ dentate gyrus; Student's t test. (D) Representative confocal images showing dentate gyrus neurons of 1 to 5 days of age positive to BrdU/ NeuroD1/NeuN (in orange in the merged image, indicated by arrowheads; single labeling is red, green, blue, respectively), which clearly decrease in PC3/Tis21-null mice. Scale bar, 40 mm. surrounded by walls (20 cm high). Animals were placed in the centre of the arena and allowed to explore it over a 8-min period. Elapsed the first 4 min, an object was inserted to the centre of the apparatus and mice were leaved into the arena for an additional 4min period. Behaviors were videotaped and the time spent in locomotion (Loc), no locomotion (No loc), rearing (Rear), freezing (Freez) and contact with object (Cont), as well as the time spent in sectors (external, Ext; middle, Mid and internal, Int) were analyzed by using the Observer software (Noldus Information Technology, Costerweg, NL). No significant differences were observed between PC3/Tis21+/+ and PC3/Tis212