Total Expression and Dual Gene-regulatory Mechanisms Maintained in Deletions and Duplications of the Pcdha Cluster*

The clustered protocadherin-α (Pcdha) genes, which are expressed in the vertebrate brain, encode diverse membrane proteins whose functions are involved in axonal projection and in learning and memory. The Pcdha cluster consists of 14 tandemly arranged genes (Pcdha1–Pcdha12, Pcdhac1, and Pcdhac2, from 5′ to 3′). Each first exon (the variable exons) is transcribed from its own promoter, and spliced to the constant exons, which are common to all the Pcdha genes. Cerebellar Purkinje cells show dual expression patterns for Pcdha. In individual Purkinje cells, different sets of the 5′ genes in the cluster, Pcdha1–12, are randomly expressed, whereas both 3′ genes, Pcdhac1 and Pcdhac2, are expressed constitutively. To elucidate the relationship between the genomic structure of the Pcdha cluster and their expression in Purkinje cells, we deleted or duplicated multiple variable exons and analyzed the expression of Pcdha genes in the mouse brain. In all mutant mice, transcript levels of the constant exons and the dual expression patterns were maintained. In the deletion mutants, the missing genes were flexibly compensated by the remaining variable exons. On the other hand, in duplication mutants, the levels of the duplicated genes were trimmed. These results indicate that the Pcdha genes are comprehensively regulated as a cluster unit, and that the regulators that randomly and constitutively drive Pcdha gene expression are intact in the deleted or duplicated mutant alleles. These dual regulatory mechanisms may play important roles in the diversity and fundamental functions of neurons.

damental question is how individual neurons obtain a unique cellular identity. In one model, individual neurons of the same anatomical type express a combinatorially distinct subset of genes, to generate diverse neural circuits. For example, in the olfactory sensory system, the axons of olfactory epithelium cells that express one odorant receptor project to the same glomerulus (1) and gene regulation play a critical role in the receptor choice of the olfactory epithelial cells (2). Another example is the distinct subsets of clustered protocadherin (Pcdh) genes, which are expressed combinatorially in individual neurons (3)(4)(5)(6)(7). The clustered Pcdh-␣ (Pcdha), Pcdh-␤, and Pcdh-␥ family genes encode cadherinrelated transmembrane proteins (4, 8 -10). In mouse, the Pcdh proteins are widely expressed in neurons (4,11,12), and localize to axons and synaptic junctions (4,(11)(12)(13)(14)(15). Their loss of function in mice revealed that the Pcdhs play important roles in neuronal survival (13), synaptic connectivity (16), axonal convergence (14), learning and memory (17), and axonal projections of serotonergic neurons (18). These results highlight the Pcdhs as candidate molecules for specifying neuronal identity in the nervous system. Thus, elucidating the regulatory mechanisms for the expression of the genes in Pcdh clusters could lead to better understanding of the development and function of neuronal identity.
In the mouse Pcdha cluster, the variable region encodes multiple first exons (variable exons) for 14 different Pcdha genes (Pcdha1-Pcdha12, Pcdhac1, and Pcdhac2) (19). Each variable exon (a1-a12, ac1, and ac2, from 5Ј to 3Ј) is transcribed from its own promoter and spliced to the constant region exons (CR1-CR3), which are shared by all the Pcdha genes (8,19,20) (see Fig. 1A). Almost all of the promoters contain a conserved sequence element (CGCT) (9). Single cell analysis shows that one or two of the Pcdha1-Pcdha12 genes, located 5Ј in the cluster, are independently and randomly expressed from both alleles in individual Purkinje cells of the cerebellum (5). On the other hand, Pcdhac1 and Pcdhac2, located at the 3Ј end of the cluster, are constitutively expressed from both alleles in individual Purkinje cells (6). These findings suggest that there are at least two kinds of expression regulations, random and constitutive, in the Pcdha cluster. In addition, the expression levels of Pcdha1-Pcdhac1, but not of Pcdhac2, are enhanced at the neu-ronal differentiation stage in embryonic stem (ES) 2 cells, by a cis-regulatory element consisting of five DNase I-hypersensitive sites (HS5-1) (21) (see Fig. 1A). Thus, random gene expression may be regulated by HS5-1. However, almost nothing is known about the dual regulatory mechanisms for the random and constitutive expression of the Pcdha genes.
Various approaches have been taken to elucidate the expressional regulation of clustered genes, including searching for a cis-regulatory element (22)(23)(24), deleting and duplicating clustered genes (25,26), and inverting clustered genes (26 -29). However, no study to date has reported how the expression of Pcdha genes is influenced by deletion or duplication of the variable exons. Here, we used the targeted meiotic recombination (TAMERE) method (30) to generate mutant mice in which the variable exons, with their promoters, were deleted or duplicated. We compared the expression patterns of the Pcdha genes within the mutant cluster (in the mutant mice) with those of the wild-type (WT) cluster, in WT mice, by quantitative real time-PCR (qRT-PCR), in situ hybridization, and single cell RT-PCR and single nucleotide polymorphism (SNP) analyses of Purkinje cells.
In the mutants, the expression levels and distribution patterns of the CR transcript were essentially unaltered. Interestingly, in the deletion and duplication mutants, the dual regulatory patterns of Pcdha gene expression were maintained. The single cell RT-PCR and SNP analyses, which can distinguish the mRNAs expressed from individual alleles, suggested that the regulators of these dual expression patterns exist outside the variable region. These dual regulatory systems may provide a clue to the role of Pcdha genes in neuronal functions.

EXPERIMENTAL PROCEDURES
Animals-All animals were maintained in a specific pathogen-free space under a 12-h light/dark regimen. Experimental procedures were in accordance with the Guide for the Care and Use of Laboratory Animals of the Science Council of Japan and approved by the Animal Experiment Committee of Osaka University.
Production of Mice Carrying the G1loxP Allele-The G1loxP mice, in which a loxP site was inserted into the genomic region 3-kb upstream from the CR1 exon, were described previously (14).
Production of Mice Carrying the G16Neo Allele-A genomic DNA library made from TT2 ES cells was screened with Pcdha1 cDNA, and a 13-kb genomic DNA fragment containing the a1 and a2 exons was obtained. This fragment was subcloned into the XhoI site of pMC1DT-A (31), and the Sleeping Beauty cassette (-loxP-IR/DR-L-loxP-PGK-neo-loxP-IR/DR-R) (32) was inserted into the NheI site between the a1 and a2 exons, to construct the targeting vector. The vector was linearized by NotI and introduced into TT2 ES cells by electroporation (33). Using standard methods, we obtained targeted recombinants and their chimeric offspring. The G16Neo mutant mice were backcrossed with the C57BL/6 (B6) strain (see supplemental Fig. S1 for details).
Production of Mice Carrying the 11R Allele-We constructed a targeting vector to insert, in-frame, a GAP43-HcRed gene cassette (encoding the N-terminal peptide, MLCCMRRTK, of GAP43 (34), the HcRed protein, which is a far-red fluorescent protein (35), and a stop codon), a floxed PGK-neo, an internal ribosome entry site (36), a Kozak sequence (37), and then the a11 coding sequence, between the 7133 bp of the 5Ј homologous arm (amplified with the 11R5ЈF and 11R5ЈR primers (sequences available under supplemental "Experimental Procedures") and 2542 bp of the 3Ј homologous arm (amplified with the 11R3ЈF and 11R3ЈR primers (sequences available in supplemental Table S1)). An MC1DT-A cassette (31) was inserted at the end of the 5Ј arm to allow selection. The 11R-targeting vector was linearized by NotI and introduced into TT2 ES cells by electroporation. We obtained targeted recombinants and their chimeric offspring using standard methods. The 11R mutant mice were backcrossed with the B6 strain (see supplemental Fig. S2 for details).
Synaptonemal Complex Protein 1 (Sycp1)-Cre Transgenic Mice-For efficient trans-allelic recombination (30,38), we generated Sycp1-Cre transgenic mice, which expressed Cre recombinase specifically in the testis. The transgene vector contained the Ϫ737 to ϩ87 promoter region of the Sycp1 gene and the Cre recombinase gene, with a nuclear localization signal and a polyadenylation signal, inserted into pBluescript II. The fragment, digested by SalI and NotI, was microinjected into fertilized eggs derived from B6 mice. Three transgenic lines were generated. They were maintained on the B6 genetic background, and expression of the Cre transgene in the germ cells was confirmed by crossing them with CAG-CAT-EGFP mice (39). For this study, we used one of three lines, in which Cre was expressed specifically in the testis.
Determination of the Targeted Allele in Mutant Mice-The TT2 ES cell line was established from fertilized F1 eggs from a cross between B6 and CBA mice. To identify which of these alleles was targeted, we used SNPs of Pcdha1 (rs24884904) and/or Pcdha10 (rs13498878) between the B6 and CBA strains. For Pcdha1, the PCR was performed as follows: 94°C for 5 min, 35 cycles of 94°C for 30 s, 57°C for 30 s, 72°C for 30 s, followed by 72°C for 7 min. The primers were Pcdha1SNPF and Pcdha1SNPR (sequences available in supplemental Table S1). The PCR product was digested with BmgT120I (Takara) yielding 202-and 139-bp bands from the B6 allele and a 341-bp band from the CBA allele (see supplemental Fig. S3I). For Pcdha10, the PCR was performed as follows: 94°C for 5 min, 35 cycles of 94°C for 30 s, 57°C for 30 s, 72°C for 30 s, and then 72°C for 7 min. The primers were Pcdha10SNPF and Pcdha10SNPR (sequences available in supplemental Table S1). Digestion with Sau3AI yielded 202-and 139-bp bands from the B6 allele and a 341-bp band from the CBA allele (see supplemental Fig. S3G).
Quantitative Real Time-PCR-cDNAs were prepared by reverse transcription from mRNAs extracted from postnatal day 21 (P21) male mice (see supplemental "Experimental Procedures"). Three or four WT, heterozygous, and homozygous littermates were used for each experiment. Each Pcdha1 to Pcdhac2 gene was amplified by its specific forward primer in the first exon and a common reverse primer in the CR1 exon. A forward primer in the CR1 exon and a reverse primer in the CR3 exon were used to amplify a region common to all of the Pcdha transcripts, to give their total transcript level. The primer sequences are listed in supplemental Table S2. All reactions were performed in duplicate with 20-l reaction volumes, and included 1 l of cDNA, 0.5 l of each of the forward and reverse primers (each at 10 M), 10 l of SYBR Green PCR master mix (Applied Biosystems), and 8 l of H 2 O. The product was analyzed using the 7900HT Sequence Detection System (Applied Biosystems). To quantify the number of molecules, five serial 10-fold dilutions of linearized plasmid vector of each gene were constructed and analyzed simultaneously to obtain standard curves. The plasmid vector inserts were full-length cDNAs of Pcdha1 to Pcdhac2, and partial cDNAs of the ␤-actin (Actb) genes. The mean quantity of each amplified Pcdha gene was normalized to the quantity of Actb. The value was divided by the mean amount of the WT transcript, and statistical analysis was performed. These qRT-PCR results for the WT, heterozygous, and homozygous mice were analyzed using a one-factor analysis of variance with Sheffe's F test.
In Situ Hybridization-The in situ hybridization was performed essentially as described previously with small modifications (40). The details are provided under supplemental "Experimental Procedures".
Split Single Cell RT-PCR and SNP Analysis-The single cell RT-PCR was performed essentially as described previously, with small modifications (42). The details are provided under supplemental "Experimental Procedures".

Deletions and Duplications of the Pcdha Cluster-
Little is known about the relationship between the genomic structure of the clustered Pcdha genes and their gene expression. Therefore we deleted or duplicated the clustered genes in mice using the TAMERE system. Before using the TAMERE system, we individually inserted three loxP sites into the variable region of the Pcdha cluster ( Fig. 1, B-D). First, the "G1loxP" allele (14) was generated by inserting a loxP site between the ac2 exon and the first exon of the constant region (CR1) in the Pcdha cluster (Fig. 1B). Second, a loxP site was inserted into the sequence between the a1 and a2 exons, to generate the "G16Neo" allele ( Fig. 1C and supplemental Fig. S1). Finally, a loxP site was inserted between the promoter and coding region of a11, to generate the "11R" allele ( Fig. 1D and supplemental Fig. S2).
To delete or duplicate the sequence between the loxP site of the 11R allele and that of the G1loxP allele by the TAMERE system, we obtained male mice that possessed the 11R and G1loxP alleles (11R/ G1loxP), and the Sycp1-Cre transgene, which elicits Cre recombinase expression specifically in the testis. The male mice were crossed with WT female mice, and the genotypes of the F1 pups were analyzed by PCR. The minority of F1 pups carried the del (11-c2) allele, in which exons a11 to ac2 were deleted (Fig.  1E), or the dup(12-c2) allele, in which exons a12 to ac2 were duplicated (Fig. 1H). F1 pups carrying these deletion or duplication alleles were obtained at 12.7% (17 of 134 pups). All of these heterozygous and homozygous mutant mice survived to adulthood and were fertile.
Next, we examined the distribution of transcripts in the Pcdha del(11-c2)/del (11-c2) brain at P21 by in situ hybridization, using variable exon-specific probes for Pcdha1 to Pcdhac2 and the CR probe common to all the Pcdha genes. CR probepositive cells were observed throughout the WT brain (Fig. 2,  B-D). A similar CR staining pattern was seen in the Pcdha del(11-c2)/del(11-c2) brain (Fig. 2, B-D). Pcdha1-Pcdha12 transcripts are randomly expressed (see supplemental Note 2), and Pcdhac1 and Pcdhac2 transcripts are constitutively expressed in Purkinje cells of the WT cerebellum (5,6). This expression pattern was also observed in the cells of the cerebral cortex and in the pyramidal cells of the hippocampal CA3 region of P21 WT mice (data not shown). In these regions, in Pcdha del(11-c2)/del (11-c2) mice, the Pcdha1-Pcdha9 transcripts   (11-c2) mice at P21, by in situ hybridization. Serials sections of the WT or Pcdha del(11-c2)/del (11-c2) cerebellum were probed, and the positive-cell frequency of each randomly expressed gene varied widely in each section (see supplemental Note 2). Anterior is to the left, posterior to the right. The Pcdha1-Pcdha9 genes showed a random expression pattern in the Purkinje cells of WT and Pcdha del(11-c2)/del (11-c2) mice, but the expression pattern of Pcdha10 changed from a weak random expression in WT mice to a strong constitutive expression in Pcdha del (11-c2)/del (11-c2) mice. The Pcdha7/8 probe recognized the Pcdha7 and Pcdha8 transcripts. The Pcdha9 probe did not give a clear signal. Asterisks indicate signal-positive Purkinje cells except for Pcp2. Scale bar, 100 m.  (11-c2) (CBA) and WT (JF1) mice, the first filial generation (F1) mice, namely Pcdha ϩ/del (11-c2) , were generated. After reverse transcription of the RNAs of a single Purkinje cell isolated from the cerebellum neurons, the cDNA was split into three tubes. In each tube, PCR was performed using primers for the specific genes. B, electrophoresis results of the second-round PCR products by the split single cell RT-PCR for the Pcdha and Pcp2 genes in individual Purkinje cells. #1-17 numbers designate individual cells. 1-3, tubes into which the cDNA from an individual Purkinje cell was divided; independent PCRs were performed for each tube. C, after sequencing the PCR products, SNP analysis was used to distinguish between Pcdha transcripts from the del (11-c2) allele and those from the WT allele. Transcripts from the WT and the del (11-c2) alleles are shown as blue and red circles, respectively. The Pcdha6 gene has no SNP between the B6 and JF1 strains, and is undistinguishable. Transcripts that were undistinguishable or not determined are shown as plus signs. Nonspecific bands are shown as minus signs. Pcdha10 was clearly expressed from the del(11-c2) allele in all the cells examined. NOVEMBER 13, 2009 • VOLUME 284 • NUMBER 46

Deletions and Duplications in the Pcdha Cluster
were also randomly expressed. However, the number of cells expressing Pcdha10 and its expression level had dramatically increased in the cerebral cortex and hippocampus (Fig. 2, B-D). This result was consistent with the qRT-PCR analysis.
To quantify the expression frequency of Pcdha10, we examined Purkinje cells of the cerebellum. In the Purkinje cells, the expression frequency of Pcdha10 was higher in Pcdha del (11-c2)/del (11-c2) than in WT mice (middle of Fig. 3A), whereas the expression frequency of the CR was similar between WT and Pcdha del(11-c2)/del (11-c2) mice (top of Fig. 3A). We examined the ratio of Pcdha10-positive cells to total Pur-kinje cells, using a specific marker for Purkinje cells, Pcp2. Only 12% of the WT Purkinje cells were Pcdha10-positive, whereas 84% of the Pcdha del(11-c2)/del (11-c2) Purkinje cells were Pcdha10-positive (Fig.  3B). In addition, the expression frequency of Pcdha10 was close to that of Pcdhac1 (77 Ϯ 21%) and Pcdhac2 (100 Ϯ 10%). The expression patterns of the other Pcdha genes did not dramatically change in the cerebellum (Fig. 4). No significant change was detected in the number of Pcp2-positive Purkinje cells between the WT and Pcdha del(11-c2)/del (11-c2)  Finally, to examine expression of Pcdha genes from the single del(11-c2) allele, we performed single cell RT-PCR and SNP analysis of the Purkinje cells of Pcdha ϩ/del (11-c2) mice at P21. We were able to distinguish the del(11-c2) allele (CBA) from the WT allele (JF1) by SNP analysis (Fig. 5A). All of the Purkinje cells analyzed expressed Pcdha10 from the del(11-c2) allele (Fig. 5, B and C). Cells expressing Pcdha1 to Pcdha9 from the del(11-c2) allele were rare (Fig. 5, B and C). In this experiment, we could not detect Pcdha10 from the WT allele, and found Pcdha6 at high frequency. These results indicated that from the del(11-c2) allele, the expression pattern of Pcdha10 changed from random to constitutive, but the expression pattern of the Pcdha1-Pcdha9 genes remained random (only Pcdha10 from the del(11-c2) allele in the 9 of 17 cells). In addition, deletion of the a11-ac2 exons altered the expression from the del(11-c2) allele but not from the WT allele. Therefore, these results suggest that the original expression regulator for the Pcdhac1 and Pcdhac2 genes in the WT allele regulated Pcdha10 in the del (11-c2) allele. In addition, the dual expression pattern of random and constitutive expression was reallocated among the Pcdha genes of the del(11-c2) allele, suggesting that regulators of the dual expression were conserved for the del(11-c2) allele.
Expression Levels of Pcdha Genes from the dup(12-c2) Allele-We examined the effects of duplicating the a12 to ac2 exons using the Pcdha ϩ/dup (12-c2) and Pcdha dup(12-c2)/dup (12-c2) mice, and qRT-PCR analysis. The expression level of the CR transcript in the Pcdha ϩ/dup (12-c2) and Pcdha dup(12-c2)/dup (12-c2) brains was essentially the same as in the WT brain (Fig. 10A). Among the duplicated genes, the expression level of Pcdhac2 was significantly increased, but those of Pcdha12 and Pcdhac1 were not changed (Fig.  10A). Among the single genes, the expression levels of Pcdha6, Pcdha9, Pcdha10, and Pcdha11 were significantly decreased, but those of Pcdha1-Pcdha5, Pcdha7, and Pcdha8 showed no significant change. Although the expression levels of some genes were changed, that of the CR transcript showed no significant change, suggesting that expression of the Pcdha genes from the dup(12-c2) allele was reallocated.

DISCUSSION
In individual neurons, the Pcdha1 to Pcdha12 genes are expressed randomly, and Pcdhac1 and Pcdhac2 are expressed constitutively (5,6). This dual expression indicates that there are at least two kinds of gene regulation mechanisms for the Pcdha cluster. However, almost nothing is known about how this dual expression is regulated. To elucidate the relationship between the genomic structure of the Pcdha cluster and the dual expression patterns, we deleted or duplicated multiple variable exons and analyzed expression of the Pcdha genes in the brain.
We also examined the CR transcript in these mutants. Because all the spliced Pcdha transcripts contain the CR exons, the level of spliced CR transcript was assumed to be equal to the sum of all the spliced transcripts of Pcdha genes. We found that the expression level of the CR transcript in the engineered mutants was hardly influenced by the number of variable exons. The putative cis-elements may therefore determine the overall expression levels of the Pcdha genes in the brain.
Deletions and Duplications Suggest Mechanisms Involving Positional Effects and Random Promoter Choice in the Pcdha Cluster-In the WT mouse, the variable exons located in the most 3Ј position of the variable region (ac1 and ac2) are constitutively expressed (5,6). Likewise, in the mouse bearing the del(11-c2) allele, the most 3Ј variable exon (a10) was constitutively expressed. This result indicates that the constitutive expression of Pcdhac1 and Pcdhac2 requires their location at the most 3Ј position of the cluster, and suggests that the specific promoters of these genes may not be essential for their constitutive expression. Our results also indicate the existence of a regulatory cis-element for Pcdhac1 and Pcdhac2 outside of the variable exons, a11-ac2, because the constitutive expression was independently regulated by individual alleles.
In the HoxD cluster, serial deletions and duplications revealed that the Hoxd gene located at the 5Ј end of the cluster is preferentially expressed in digit development, and the regulation of expression is reallocated among the genes in the mutant HoxD cluster. The cis long-distance digit enhancer is located 5Ј upstream of the HoxD cluster (24). The preferential expression from the 5Ј end of the HoxD cluster is regulated by enhancer tropism (25,43). Similarly, the 3Ј end of the Pcdha cluster, namely Pcdhac2, was preferentially expressed among the Pcdha genes (data not shown). The Pcdha cluster might therefore be regulated by enhancer tropism like the HoxD cluster. In fact, DNase I-hypersensitive sites, HS7 and HS5-1, are located downstream of the clustered variable Pcdha exons (21) (see Fig. 11). HS5-1 is an enhancer for the Pcdha1-Pcdha12 and Pcdhac1 genes, but no enhancer for the Pcdhac2 gene has yet been found (21), although HS7 is a candidate enhancer.
Single cell RT-PCR analysis of Purkinje cells revealed that one or two gene(s) among Pcdha1-Pcdha12 are selectively expressed from the WT Pcdha gene locus (5,6). Likewise, in the present study, we showed that one or two gene(s) between Pcdha1 and Pcdha12 were selectively expressed from a Pcdha del (2)(3)(4)(5)(6)(7)(8)(9)(10)(11) gene locus at much higher frequency than that seen from the WT Pcdha gene locus. On the other hand, in mice bearing the dup(2-10) allele, each duplicated Pcdha gene appeared to be expressed at a lower frequency. These results indicate that only one or two of the Pcdha1-Pcdha12 genes is FIGURE 11. A phase diagram of the random and constitutive expression of Pcdha genes in the WT, deleted, and duplicated alleles. Genomic structures from top to bottom show the WT, del (11-c2), del (2)(3)(4)(5)(6)(7)(8)(9)(10)(11), dup (2)(3)(4)(5)(6)(7)(8)(9)(10), and dup (12-c2) alleles. Both random and constitutive regulation always occurred for all of these alleles. The variable exons located the most 3Ј in the variable region are expressed constitutively (red arrows). The variable exons located in the 5Ј portion of the variable region are expressed randomly (blue arrows). The arrows indicate the direction of transcription. HS5-1 and HS7 enhancers are shown as ovals. Open boxes, variable exons. Black bars and boxes, constant region exons. i, internal ribosome entry site. H, HcRed. expressed selectively, independent of the number of a1-a12 exons, although at least two genes among Pcdha1-Pcdha12 are necessary. This phenomenon can be explained by the idea that a putative cis-element selects one or two gene(s) within Pcdha1-Pcdha12. In other words, the cis-element may be competitively shared by all of the randomly regulated variable exons. In the ␤-globin and HoxD clusters, common enhancers are thought to regulate the differential transcription by promoter competition (25,44,45). Likewise, the differential expression of the Pcdha cluster may be controlled by a common enhancer, and our data are consistent with a competition mechanism, resulting in random promoter choice. The putative cis-element is thought to be located outside of the variable region, because the random expressions were maintained in mice bearing the del (2)(3)(4)(5)(6)(7)(8)(9)(10)(11) or dup (2)(3)(4)(5)(6)(7)(8)(9)(10) allele. The putative cis-element may be HS5-1.
Potential Functional Significance of the Random and Constitutive Pcdha Gene Expression in Neurons-The Pcdha cluster encodes 14 kinds of single pass transmembrane proteins. The a1-a12, ac1, and ac2 exons encode six cadherin-like extracellular domains, a transmembrane domain, and part of a cytoplasmic domain, and the CR1-CR3 exons encode the rest of the cytoplasmic domain (4,19,46). In amino acid sequence, the variable regions of Pcdha1-Pcdha12 are similar to each other, and are distinct evolutionarily from those of Pcdhac1 and Pcdhac2 (47,48). The extracellular 1 domains of Pcdha1-Pcdha12 have an Arg-Gly-Asp (RGD) sequence, which binds to integrin-␤1 (49), whereas the extracellular 1 domains of Pcdhac1 and Pcdhac2 have no RGD sequence (47). Thus, Pcdha1-Pcdha12 are functionally different from Pcdhac1 and Pcdhac2. Furthermore, the expression patterns of Pcdha1-Pcdha12 are different from those of Pcdhac1 and Pcdhac2 (5, 6). This expression difference may also reflect distinct functions of the Pcdhas.
In this study, we showed evidence for two independent regulatory mechanisms, one directing random expression and one directing constitutive expression of the variable exons in the Pcdha cluster (see Fig. 11). Thus, in the WT allele, the Pcdha1-Pcdha12 genes, and the Pcdhac1 and Pcdhac2 genes appear to be regulated independently. These dual expression mechanisms may reflect two different functions of the Pcdha genes in neurons. For instance, the randomly expressed Pcdha1-Pcdha12 may functionally contribute to the enormous diversity of neurons, whereas the constitutively expressed Pcdhac1 and Pcdhac2 may be essential genes for all neurons.