Introduction

Angelman syndrome (AS [MIM 105830]) is a neurodevelopmental disorder, characterized by mental retardation, epilepsy, seizures, frequent smiling and laughter, absence of speech, and abnormal gait, with an occurrence of approximately 1:15,000 live births (Williams et al. 1995). This syndrome is one of the best examples of human disease involved in genomic imprinting, caused by the absence of a normal maternal contribution to the human chromosome 15q11-q13 region. The E6-AP ubiquitin protein ligase gene, UBE3A, has been strongly implicated as the AS gene because of genetic mutation and tissue-specific imprinting with preferential maternal expression in human brain, fibroblasts, and lymphoblasts, and in specific regions in the mouse brain (Matsuura et al. 1997; Rougeulle et al. 1997; Albrecht et al. 1997; Herzing et al. 2002). However, the phenotypes of AS patients and a model mouse with a 15q11-q13 deletion are more severe than that with an UBE3A mutation (Jiang et al. 1998).

Recently, human ATP10C encoding a putative aminophospholipid translocase was identified from the adjacency to UBE3A, and also exhibited maternal expression in human brain and lymphocytes and lack of expression in AS patients, suggesting that it may contribute to AS phenotypes (Meguro et al. 2001; Herzing et al. 2001). Moreover, abnormal phenotypes, such as language delays and autism spectrum disorders, were observed with the maternal 15q11-q13 duplication, while the paternal duplication has no obvious phenotypes (Cook et al. 1997). It was also suggested that chromosome 15q11-q13 contains maternally expressed autism associated gene(s). Since ATP10C is presumed to function as an aminophospholipid-transporting ATPase, it may play a role in cell signaling in the central nervous system (CNS) (Halleck et al. 1999). Moreover, many human genetic disorders caused by loss of function of members of the P-type ATPase family display neurological symptoms (DiDonato and Sarkar 1997). Thus, it is possible that ATP10C may contribute to chromosome 15 associated neurological disorders including AS and autism.

A recent report demonstrated that Atp10c/pfatp, which is a mouse homologue of ATP10C, was associated with obesity, and maternal inheritance of the deletions resulted in increased body fat when compared with the inheritance of the deletion from the father, suggesting it was involved in genomic imprinting (Dhar et al. 2000). Although Atp10c was not imprinted in testis and adipose tissues, it is possible that it could be imprinted tissue specifically as with Ube3a. To determine whether mouse Atp10c is imprinted in the CNS, we examined the allele-specific expression and methylation status of Atp10c in brain tissues. Although the parent-of-origin-specific methylated region was not observed, mouse Atp10c was imprinted in a tissue-specific manner, with predominant expression of the maternal allele in hippocampus and olfactory bulb, overlapping with regions where Ube3a is imprinted.

Materials and methods

Strains and matings

Mouse strains used were C57BL/6J and JF1. (B6 x JF1)F1 mice were generated by mating a female B6 with a male JF1 mouse. (JF1 x B6)F1 mice were generated by mating a female JF1 with a male B6 mouse. These reciprocal crosses to generate F1 mice were used for allele-specific expression analysis. Each of the F1 mice was 30 weeks old.

Expression analysis

Total RNA was isolated from each tissue using the AGPC method. Total RNA (6 µg) was treated with DNase I (Wako Nippon Gene, Tokyo, Japan) and the reaction was subsequently used to synthesize first-strand cDNA with random primers (Roche Diagnosis Co., Indianapolis, IN), with or without reverse transcriptase (Invitrogen Co., Carlsbad, CA). RT-PCR was performed on the cDNA with Ampli Taq Gold (Roche Diagnosis Co., Indianapolis, IN) using a step-down protocol. The reaction parameters were as follows: an initial denaturation at 95°C for 10 min, three cycles of 95°C for 30 s, 64°C for 30 s, 72°C for 30 s, three cycles of 95°C for 30 s, 62°C for 30 s, 72°C for 30 s, three cycles of 95°C for 30 s, 60°C for 30 s, 72°C for 30 s, and 27 cycles of 95°C for 30 s, 58°C for 30 s, 72°C for 30 s. Primer sequences were as follows: Atp10cF2, 5'-GCAATGTGCATTGTTTCTTCA-3' and Atp10cR2:5'-TCAGACCCATGAGGTGAACTT-3'. PCR products were analyzed on a 2% agarose gel followed by SYBR green I (BioWhittaker Molecular Applications, Rockland, ME) staining.

Identification of polymorphism

Genomic DNA was prepared from both B6 and JF1 mice. DNA samples were amplified by PCR using primers corresponding to Atp10c 3' UTR (Atp10cF1:5'-AGGAAGCCAGAGGTACCAAA-3' and Atp10cR1:5'-GGACCCCCACTCTTCTTACC-3'). PCR products were purified and directly sequenced.

PCR-RFLP

Total RNA from hippocampus, olfactory bulb, cerebellum, cerebral cortex, and brain stem of adult (B6 x JF1) F1 and (JF1 x B6) F1 mice was used for PCR-RFLP analysis. RT-PCR was performed using the primers Atp10cF2 and Atp10cR2. PCR products were digested with MspI (Nippon Gene, Tokyo, Japan). The digested products were separated in a 5% polyacrylamide gel and stained with SYBR green I followed by quantification using a phosphoimager.

Intron-spanning RT-PCR

Intron-spanning RT-PCR analysis was performed using the primers spanning an Atp10c intron 20 (Atp10cF3:5'-TGTCTCATCGCACCTATTGC-3' and Atp10cR2). Genomic DNA was prepared from liver from (B6 x JF1)F1 and cDNA were prepared from CNS tissues from (B6 x JF1)F1. The PCR reaction parameters were as follows: an initial denaturation at 95°C for 10 min, three cycles of 95°C for 30 s, 66°C for 30 s, 72°C for 30 s, three cycles of 95°C for 30 s, 64°C for 30 s, 72°C for 30 s, three cycles of 95°C for 30 s, 62°C for 30 s, 72°C for 30 s, and 26 cycles of 95°C for 30 s, 60°C for 30 s, 72°C for 30 s. PCR products were analyzed on a 1.2% agarose gel followed by SYBR green I staining.

Methylation Southern blot analysis

Southern blot analysis was performed to verify the methylation status of the CpG island of Atp10c intron 1, predicted by EMBROSS/EMBL-EBI (http://www.ebi.ac.uk/emboss/cpgplot/). Genomic DNA was isolated from B6 (♂/10 weeks) various brain tissues and testis. The 10-µg genomic DNA was digested with or without 50 U methylation-sensitive enzyme NotI after 20 U ApaI digestion, separated on 0.8% agarose gel and transferred onto Hybond N+membrane (Amersham Pharmacia, Bucks., England). Radiolabeled PCR probes were generated using the following primers, Atp10cF4:5'-AAGCTGGAGGGTAGGGTGTT-3' and Atp10cR3:5'-TCAAAAACACTGCAGCAAGG-3'. Hybridizations were performed in ×5 SSPE, 0.5% SDS and 200 µg/ml salmon sperm DNA at 55°C and a final wash in ×0.1 SSC and 0.1% SDS at 55°C. Autoradiography was analyzed with a BAS-2500 phosphoimager (Fuji Film).

Results

The database searches permitted us to identify the sequence of the mouse ATPase class V type 10A gene, which was the mouse ortholog of human ATP10C. We first examined the tissue-specific expression of mouse Atp10c. It was expressed at high level in the CNS including cerebral cortex, cerebellum, hippocampus, olfactory bulb and brain stem as well as lung, kidney and testis, and at a low level in thymus, heart, liver, pancreas and muscle (Fig. 1). A faint signal was detected in spleen. These results of tissue-specific expression suggested that Atp10c had significant functions in the CNS. To determine the imprinting status of Atp10c, we searched for sequence polymorphisms between C57BL/6J and JF1 mice. By direct sequence analysis, we found a single nucleotide polymorphism within the 3' UTR region of the Atp10c gene, which abolished a MspI restriction site in JF1 that is present in B6 (Fig. 2A). Using this polymorphism, we verified the parental origin of the transcripts in reciprocal crosses between B6 and JF1 mice. RNAs from hippocampus, olfactory bulb, cerebellum, brain stem and cerebral cortex were subjected to RT-PCR and the PCR products were digested with MspI. Allelic expression bias with preferential maternal expression was shown in hippocampus and olfactory bulb with a small amount of paternal expression, but not in other brain tissues (Fig. 2B). Parent-of-origin-specific expression was not observed in any other tissues examined (data not shown). These results suggested that mouse Atp10c was imprinted in a region-specific distribution in the same manner as Ube3a.

Fig. 1.
figure 1

Expression analysis of the mouse Atp10c gene in multiple tissues. RT-PCR with the specific primers for Atp10c showed a high level of expression in brain tissues, suggesting that it has significant functions in CNS. Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as a control

Full size image
Fig. 2A–C.
figure 2

Imprinted expression of Atp10c in the F1 mice of reciprocal crosses (JF1 x B6) and (B6 x JF1). A Schematic of the transcribed sequence polymorphisms in the ATP10c 3' UTR and the primers used for amplification of ATP10c. An asterisk denotes the polymorphic site. Arrowheads represent oligonucleotide primers. B Imprinted expression of Atp10c in brain tissues. The imprinted expression was assessed by PCR-RFLP analysis. PCR products were digested with MspI, producing the undigested 288-bp fragment (a) for the B6 allele and the digested 217-bp and 71-bp fragments (b) for the JF1 allele. Densitometric values are given below each lane. Maternal allele-specific transcription is seen for Atp10c in hippocampus and olfactory bulb, in contrast to biallelic expression in the other samples. C Intron-spanning RT-PCR analysis of Atp10c. Genomic DNA and cDNAs were amplified using the primers spanning the intron 20 (F3 and R2). Only the spliced transcripts were amplified, indicating that the antisense transcript does not lie in the 3' UTR region of Atp10c (RT+ reverse transcriptase-positive lane, RT−reverse transcriptase-negative)

Full size image

An antisense transcript extending to the Atp10c locus has not yet been identified, but it is tempting to speculate that it may also be associated with the imprinted expression of Atp10c. To study the mechanism of the imprinted expression of Atp10c, we performed RT-PCR analysis using primers spanning an Atp10c intron 20 (Fig. 2C). Only the spliced RNA was expressed in any tissues, suggesting the antisense transcript was not extending to this locus. Then, we analyzed methylation status in the CpG island of Atp10c intron 1. The NotI site (Fig. 3) and the SmaI site (data not shown) in this CpG island were unmethylated in any tissues, whether Atp10c was imprinted or not. While it is possible for other CpG dinucleotides in Atp10c to display allele-specific differential methylation, our result may be related to no DMR identified in Ube3a.

Fig. 3A, B.
figure 3

DNA methylation analysis of the CpG island of Atp10c intron 1. A Restriction map of the genomic fragment in the CpG island of Atp10c. The probe generated by PCR amplification is indicated by the black box. B Southern blot analysis of the Atp10c CpG island. Genomic DNA was digested with ApaI (A) or ApaI plus NotI (N). The 4.7-kb ApaI fragment was completely digested by the methylation-sensitive endonuclease NotI in all tissues, indicating absence of methylation on both alleles (M marker)

Full size image

Discussion

In the present study, expression analysis of Atp10c also showed higher levels in the CNS including cerebral cortex, cerebellum, hippocampus, olfactory bulb and brain stem. A previous study showed the localization of Atp10c in mouse CNS to subiculum, cerebellar granule cells, hippocampus, olfactory bulb mitral cells, hypothalamus by in situ hybridization (Halleck et al. 1999), and expression overlapped with regions where Ube3a was imprinted (Albrecht et al. 1997). These findings suggest that Atp10c had significant functions in the CNS.

UBE3A is the only gene where mutations have been found in AS patients, strongly supporting its causative role in AS. However, there is evidence to suggest that another gene may play a role either directly in AS or indirectly by regulating UBE3A. We previously reported that ATP10C was preferentially expressed from the maternal allele in human lymphoblasts and brain tissues (Meguro et al. 2001). A recent study demonstrated the preferential maternal expression of UBE3A in human fibroblasts, lymphoblasts and neural precursor cells by FISH, while RT-PCR analysis could not detect allelic expression bias (Herzing et al. 2002; Nakao et al. 1994). Here we demonstrated that mouse Atp10c was imprinted in a tissue-specific manner, with a predominant expression from the maternal allele in hippocampus and olfactory bulb, where mouse Ube3a also shows imprinted expression. This overlap suggests that the imprinted expression of these two genes is coordinately regulated in the CNS.

Recently, a paternally expressed Ube3a antisense transcript was demonstrated, which is under control of an imprinting center (IC) (Rougeulle et al. 1998; Chamberlain and Brannan 2001; Runte et al. 2001). Although the role of this antisense transcript is unknown, it may regulate the imprinted expression of Ube3a. However, the antisense transcript extending to the Atp10c locus has not been identified. The DMR also has not been identified in Atp10c, as well as Ube3a. Although the mechanism of the imprinted expression of Atp10c is unknown, it is possible that Atp10c may play an important role in CNS development, and the absence of maternally expressed Atp10c may be causative of the phenotypes of AS and autism.