Predicted Gene Sequence C10orf112 is Transcribed, Exhibits Tissue-Specific Expression, and May Correspond to AD7

Abstract

Case-control and prospective longitudinal studies have revealed an interaction of the anonymous D10S1423 234bp allele with the APOE4 allele in determining the age-specific risk of Alzheimer's Disease (AD). The D10S1423 polymorphism resides within intron 10 of open reading frame C10orf112, whose predicted product resembles a low-density lipoprotein receptor (NCBI Build 35.1). These observations suggest that the D10S1423 234bp allele may be in linkage disequilibrium with a C10orf112 gene variant whose product interacts with the apoE4 lipoprotein. Our initial exploration of this hypothesis focused on validating the C10orf112 gene model. RT-PCR amplification from human hippocampal mRNA confirmed that 34 of the predicted 39 exons of C10orf112 were expressed in this brain region. Northern blots revealed 1.2 kb and 3.2 kb mRNA species that hybridize to a cDNA probe consisting of contiguous exons 23-26. Expression of these C10orf112 mRNA species was limited to a subset of brain regions and heart tissue.

INTRODUCTION

Typical, later-onset forms of Alzheimer's Disease (AD) appear to be influenced by multiple susceptibility loci, combinations of which contribute to the development of this disorder [1]. We previously reported the results of a systematic survey of the human genome for the identification of highly informative DNA polymorphisms (SSTRPs) that target new AD risk genes [2]. In addition to the APOE locus, our survey detected five new candidate susceptibility loci for AD, including D10S1423. An association of the D10S1423 234bp allele with AD has been reported in three independent samples of AD cases and controls (Boston, Pittsburgh, Bonn) [2-4]. Data from our case-control studies suggest a strong synergistic interaction between the D10S1423 234bp and APOE E4 risk alleles (234bp carrier: odds ratio = 2.5, 95% confidence interval = 1.4 to 4.5; E4 carrier: odds ratio = 8.3, 95% confidence interval = 4.3 to 15.8; both alleles: odds ratio = 23.1, 95% confidence interval = 5.3 to 99.5). Patients with autopsy-confirmed AD who carry the D10S1423 234bp allele have also been reported to manifest substantially lower levels of cortical dopamine and relative preservation of cortical norepinephrine levels compared to noncarriers, a neurochemical finding that further supports a role for this anonymous genotype in the pathophysiology of AD [5].

These results have been confirmed by a prospective, longitudinal, double-blind assessment of the age-specific risk of AD encountered by 325 asymptomatic first-degree relatives of AD probands who carried the D10S1423 234bp allele, the APOE E4 allele, or both, after 11.5 years of systematic follow-up [4]. A total of 18 incident cases of AD were detected during the first 3379 subject-years of this longitudinal study. The effects of carrying either or both of the D10S1423 234bp and APOE E4 alleles on the age-specific risk of developing AD were determined using Kaplan-Meier survival analysis [6-8]. The age-specific risk of developing AD was the greatest for individuals who carried both alleles (Mantel-Cox statistic = 20.12, df = 3, p = 0.0002; Breslow statistic = 13.36, df = 3, p = 0.004). Cox proportional hazards models were developed to estimate the risk ratios for each genotype, controlling for the potential effects of age at recruitment, sex, and years of education [6-8]. In the resulting best fitting model, only individuals who carried both risk alleles exhibited a risk ratio that differed significantly from 1 (risk ratio = 16.2, p = 0.008, 95% confidence interval = 2.1 to 128.3). In 2001, McKusick named the AD risk locus identified by the D10S1423 polymorphism AD7 in Online Mendelian Inheritance in Man [9].

The molecular mechanism for the interaction of the anonymous D10S1423 234bp allele and the APOE4 allele that influences the age-specific risk of AD is unknown. As described in the National Center for Biotechnology Information human genome database (NCBI Build 35.1; http://www.ncbi.nlm.nih.gov/), the D10S1423 SSTRP resides within intron 10 of predicted open reading frame C10orf112. At its location deep within intron 10, it is unlikely that the D10S1423 SSTRP would directly affect the expression or properties of the predicted protein product of C10orf112. The C10orf112 gene model includes 39 predicted exons that encode seven low-density lipoprotein (LDL) receptor class a domains that mediate binding of LDL receptors to lipoproteins, as well as nine MAM domains that are common components of membrane receptors. These observations raise the possiblility that the D10S1423 234bp allele may be in linkage disequilibrium with a regulatory or structural C10orf112 gene variant whose product interacts with the apoE4 lipoprotein to increase the age-specific risk of AD. To initially explore this hypothesis, we determined whether the predicted C10orf112 exons were expressed in human brain tissue, the size(s) of the mRNA species encoded by C10orf112, and its tissue-specific expression.

RESULTS

Description of C10orf112

Salient features of the predicted open reading frame C10orf112 were derived from the NCBI database and are presented in Table 1 (http://www.ncbi.nlm.nih.gov/; Build 35.1). This model describes a large gene containing 39 exons and 38 introns, totalling 712720 bps. The 39 exons total 6885 bps that are sufficient to encode a protein containing 2295 amino acids. As shown in Table 1, six of the predicted exons encode a total of seven LDL-receptor class a domains. Typical LDL receptors include seven such domains that are tandemly repeated within the amino terminal region of the receptor and mediate binding to lipoproteins. The predicted amino acid sequence of C10orf112 also includes 9 MAM domains that appear in meprin, A5, receptor protein tyrosine phosphatase mu, as well as other cell surface receptors, and participate in cell-cell interactions. The midpoint of the PCR product used in our previous studies [2,4,5] to genotype the D10S1423 SSTRP is located at gene postion 45472, within predicted intron 10.

Table 1.

Description of C10orf112

	Gene Numbering		Exon	Intron	Transcript Numbering		Amino Acid Range		Domains (amino acid range)
Exon	Start	Stop	Size (bp)	Size (bp)	Start	Stop	Start	Stop	LDLR a	MAM
1	1	45	45	321	1	45	1	15
2	365	461	97	19418	46	142	16	47
3	19878	19979	102	932	143	244	48	81
4	20910	21056	147	3552	245	391	82	130		86---->
5	24607	24773	167	3240	392	558	131	186		<---176
6	28012	28145	134	1944	559	692	187	231
7	30088	30142	55	101	693	747	232	249
8	30242	30334	93	2627	748	840	250	280
9	32960	33167	208	9418	841	1048	281	349	313-345
10	42584	42730	147	8732	1049	1195	350	398		355--->
11	51461	51558	98	10466	1196	1293	399	431		<-------->
12	62023	62196	174	9399	1294	1467	432	489		<-------->
13	71594	71714	121	28401	1468	1588	490	529		<---511
14	100114	100266	153	429	1589	1741	530	580		533--->
15	100694	100799	106	486	1742	1847	581	616		<-------->
16	101284	101651	368	4004	1848	2215	617	738	706-738	<---690
17	105654	106066	413	48005	2216	2628	739	876		744--->
18	154070	154157	88	22277	2629	2716	877	905		<---899
19	176433	176609	177	2797	2717	2893	906	964	929-964
20	179405	179567	163	40769	2894	3056	965	1019		969--->
21	220335	220491	157	3459	3057	3213	1020	1071		<-------->
22	223949	224083	135	3673	3214	3348	1072	1116		<-------->
23	227755	227968	214	16190	3349	3562	1117	1187	1151-1182	<--1143
24	244157	244404	248	3989	3563	3810	1188	1270		1194-->
25	248392	248683	292	35232	3811	4102	1271	1367		<--1348
26	283914	284159	246	1411	4103	4348	1368	1449	1370-1404	1408-1444
27	285569	285748	180	91	4349	4528	1450	1509		1495-->
28	285838	285995	158	60699	4529	4686	1510	1562		<-------->
29	346693	346876	184	25610	4687	4870	1563	1623		<--1618
30	372485	372566	82	12781	4871	4952	1624	1651
31	385346	385497	152	2407	4953	5104	1652	1701
32	387903	388011	109	6802	5105	5213	1702	1738
33	394812	395032	221	32549	5214	5434	1739	1811		1767-->
34	427580	427737	158	36152	5435	5592	1812	1864		<-------->
35	463888	464089	202	27492	5593	5794	1865	1931		<--1928
36	491580	491843	264	12321	5795	6058	1932	2019	1941-1976 1985-2019
37	504163	504288	126	207652	6059	6184	2020	2061
38	711939	712048	110	83	6185	6294	2062	2098
39	712130	712720	591		6295	6885	2099	2295
		Total:	6885

LDLR a; Low Density Lipoprotein Receptor Class A domain. MAM; an extracellular domain in meprin, A5, receptor protein tyrosine phosphatase mu found in many membrane receptors. The midpoint of the PCR product used in our previous studies [1,3,4] to genotype the D10S1423 simple sequence tandem repeat polymorphism is located at gene postion 45472, within predicted intron 10.

Expression of C10orf112 Exons in Human Brain

Reverse transcription PCR (RT-PCR) was performed using RNA isolated from human hippocampus to determine whether the predicted C10orf112 exons were expressed in human brain. Hippocampal RNA was used because the hippocampus undergoes early, characteristic, and extensive degenerative changes that contribute to memory and learning impairments that are cardinal clinical features of AD. As a first step, cDNA was synthesized from hippocampal RNA by extension of oligo(dT)₂₀ primers with reverse transcriptase. Following digestion of the residual RNA, the first strand cDNA was used as template for PCR that was performed using the primer pairs described in the Materials and Methods. Primer pairs for β-actin were included as a positive control. Substitution of the human hippocampal RNA for the first strand cDNA product as template resulted in no amplification (blank gels not shown), confirming the absence of contaminating genomic DNA that could confound the RT-PCR results.

As shown in Figure 1, the results of RT-PCR revealed that 34 of the 39 predicted exons in the C10orf112 model are expressed in human hippocampus. There was no evidence of expression of exons 1, 7, 22, 25, or 27. Exclusion of these five exons would result in a reduction of the predicted mRNA transcript length to 6175 bps, and a predicted protein product that retained all 7 LDL-receptor class a domains and 6 of the 9 MAM domains. PCR amplification of all five missing exons from the first-strand cDNA was reattempted with at least one additional primer pair without success. Furthermore, the successful amplification of exon 2 from the first strand cDNA preparation excluded incomplete reverse transcription of the C10orf112 mRNA as the basis for the failed amplification of exons 7, 22, 25 and 27.

Figure 1.

Electropherograms of RT-PCR products resulting from amplification of predicted C10orf112 exons using cDNA template prepared from human hippocampus total RNA. The methods employed for RT-PCR, agarose electrophoresis, and visualization of RT-PCR products are described in the Materials and Methods. Primer pairs are listed in Table 1. Each of the three electropherograms includes thirteen numbered lanes corresponding to the RT-PCR products for the predicted exon of the same number, followed by the RT-PCR products for β-actin as a positive (+) control. Substitution of the original human hippocampus total RNA for the first strand cDNA product as template resulted in no amplification (blank gels not shown). RT-PCR products were flanked by molecular weight standards (MW, 100 bp ladder).

Preparation of cDNA sequence for interrogating Northern blots

A cDNA probe consisting of exons 23 through 26 was synthesized to interrogate Northern blots of size fractionated mRNA to characterize the predicted C10orf112 mRNA and its tissue specific expression. This cDNA sequence was chosen because it covered multiple contiguous exons of the predicted transcript without being unwieldy (sum of exons 23 through 26 = 946). Its position within the midportion of the predicted gene allowed for revisions to occur in the endpoints of the C10orf112 model without confounding the results of the Northern blots by hybridizing to mRNA species from adjacent genes. Although exons 23 and 26 encode LDL-receptor class a domains, this was not expected to result in nonspecific hybridization to LDL receptor mRNAs because the domains are conserved at the amino acid level, not at the nucleotide sequence level. As shown in Figure 2, the RT-PCR amplification of the predicted contiguous exons 23 through 26 was performed using cDNA template prepared from human hippocampus total RNA. Amplification of contiguous exons 23 through 26 was performed using the forward primer within exon 23 and the reverse primer within exon 26, resulting in a product of the predicted size (lane 3). Substitution of genomic brain DNA for the cDNA template in the last experiment resulted in no amplification, as expected from the size of intervening introns (lane 4). These findings are consistent with the prediction that exons 23 though 26 are sequentially encoded in C10orf112 mRNA from the hippocampus.

Figure 2.

Electropherogram of RT-PCR products from amplification of predicted exons 23 (lane 1), 26 (lane 2), and the contiguous exons 23 through 26 (lane 3), using cDNA template prepared from human hippocampus total RNA. The methods employed for RT-PCR, agarose electrophoresis, and visualization of RT-PCR products are described in the Materials and Methods. Primer pairs are listed in Table 1. The RT-PCR amplification of contiguous exons 23 through 26 was performed using the forward primer within exon 23 and the reverse primer within exon 26, resulting in a product of the predicted size (946 bp, lane 3). Substitution of genomic DNA for the cDNA template in the last experiment resulted in no amplification due to the size of the intervening introns, as expected (lane 4). RT-PCR products for β-actin served as a positive (+) control. RT-PCR products were flanked by molecular weight standards (MW, 100 bp ladder).

Identification of C10orf112 mRNA from brain

An autoradiogram of a Northern blot of size fractionated human brain poly A+ mRNA interrogated by the C10orf112 probe consisting of contiguous exons 23 through 26 (Figure 2, lane 3) is shown in Figure 3. The sizes of mRNA species shown in the Nothern blot were extrapolated from molecular weight standards included in the electropherogram. The radiolabeled probe hybridized to a prominent 1.2 kb mRNA species and a less prominent 3.2 kb mRNA species.

Figure 3.

Autoradiogram of Northern blot of size fractionated human brain poly A+ mRNA interrogated by a probe consisting of contiguous exons 23 through 26. The probe was produced by RT-PCR amplification of contiguous exons 23 through 26 using the forward primer within exon 23 and the reverse primer within exon 26 (see Figure 2, lane3). The methods employed for preparing and probing the Northern blot are described in the Materials and Methods. The sizes of mRNA species shown in the Nothern blot were extrapolated from molecular weight standards included in the electropherogram. The radiolabeled probe hybridized to a prominent 1.2 kb mRNA species and a less prominent 3.2 kb mRNA species.

Tissue specific expression of C10orf112

An autoradiogram of a high-throughput Northern blot of size fractionated poly A+ mRNA from human tissues was interrogated by the same C10orf112 probe consisting of contiguous exons 23 through 26 (Figure 2, lane 3). As shown in Figure 4, the probe predominantly hybridized to mRNA from three brain regions including hippocampus, occipital lobe, and parietal lobe. Hybridization to heart mRNA also exceeded background levels. In contrast, hybridization of the probe to mRNA from several other brain regions, including the frontal and temporal lobes and the cerebellum, and numerous other tissues did not exceed background levels. Placental mRNA served as a positive control. The sizes of the mRNAs detected were in the 1.2 kb to 3.2 kb range detected in Figure 3.

Figure 4.

Autoradiogram of high-throughput Northern blot of size fractionated poly A+ mRNA from human tissues interrogated by a probe consisting of contiguous exons 23 through 26. The probe was produced by RT-PCR amplification of contiguous exons 23 through 26 using the forward primer within exon 23 and the reverse primer within exon 26 (see Figure 2, lane3). In addition to placenta (positive control), the probe predominantly hybridized to mRNA from three brain regions (hippocampus, occipital lobe, parietal lobe). Hybridization to heart mRNA also exceeded background levels. Descriptions of the tissue sources of mRNA fractionated in each lane are indicated across the top of the blot. The size ranges (kb) of the fractionated mRNAs are shown for alternate rows.

DISCUSSION

The principal finding of this study is that nearly all (34 of 39) of the predicted exons of the C10orf112 model are selectively expressed in human brain regions. Like LDL receptors, the C10orf112 gene includes seven LDL-receptor class a domains, suggesting that its biological functions include binding to apolipoproteins. The presence of six encoded MAM domains, a common feature of membrane receptors, is consistent with the possibility that the C10orf112 gene product may reside at an internal or external cell membrane location with the opportunity to interact with apolipoproteins. These observations raise the possiblility that the D10S1423 234 bp allele may be in linkage disequilibrium with a regulatory or structural C10orf112 gene variant whose product interacts with apoE4 to increase the age-specific risk of AD. This hypothesis suggests an initial molecular mechanism for the elevated age-specific risk of AD among individuals who carry both the D10S1423 234 bp allele and the APOE4 allele. While our results also indicated modest expression of C10orf112 exons in heart tissue, it is unknown whether the D10S1423 234bp allele or variants of C10orf112 interact with APOE4 to mediate the risk of heart disease.

Northern blots of size fractionated mRNA from human hippocampus interrogated with a cDNA probe consisting of C10orf112 exons 23 through 26 revealed a predominant 1.2 kb mRNA species and a rare 3.2 kb mRNA species. In comparison, the size of the predicted C10orf112 mRNA that includes the 34 expressed exons is substantially larger at 6.175 kb. Assuming that these 34 expressed exons are part of a single C10orf112 gene transcript, the 1.2 and 3.2 kb mRNA species could result from the splicing of a larger mRNA precursor that is not detected by the Northern blots shown in Figures 3 and 4. Alternatively, the C10orf112 gene model described in NCBI Build 35.1 may include more than one gene. These alternatives cannot be distinguished by our data.

In silico modeling of the C10orf112 gene has continued since its description in NCBI Build 35.1. In the current NCBI Build 36.3, the 39 predicted C10orf112 exons described in Build 35.1 have been distributed among C10orf112 and three additional predicted loci (LOCs). The most recent model of C10orf112 includes the previous exons 18-26, 28, 29, and 32, encodes seven LDLa and four MAM domains, and predicts a 2.298 kb mRNA transcript. The predicted functional domains of the most recent C10orf112 model continue to make it an attractive candidate for AD7. However, this revision would place D10S1423 approximately 100 kb proximal to C10orf112, within the adjacent gene model LOC729195. As a result, the hypothesis that the D10S1423 234 bp allele serves as a proxy for a C10orf112 risk allele with which it is in linkage disequilibrium would be less tenable.

It should be noted that the predicted features of the revised C10orf112 model are not entirely consistent with the empirical data presented in our study. Since the revised C10orf112 model includes the sequential exons 23 through 26 identified in the previous gene model, the cDNA probe used to interrogate the Northern blots in our study would be expected to identify the mRNA transcribed from the revised version as well, without hybridizing to mRNA species from adjacent genes. The observed 1.2 kb mRNA species is substantially smaller than the 2.298 kb transcript predicted by the revised gene model, but could arise from splicing. In contrast, the observed 3.2 kb mRNA species is substantially larger than could be accomodated by the revised C10orf112 model. These observations highlight the importance of integrating gene modeling approaches with empirical data to evaluate the predictions of the models.

The selective expression of C10orf112 mRNA in human brain tissue is consistent with its involvement in the pathophysiology of AD, the most common brain disorder in late life. Considerably lower expression was observed in heart tissue, and only trace (if any) expression was observed in a wide range of other tissues besides the placental control. Selective expression of C10orf112 was also noted within brain regions. High levels of expression were observed in the hippocampus, and the occipital and parietal lobes, brain regions that undergo degeneration and manifest both senile plagues and neurofibrillary tangles in AD. However, little or no expression was observed in the frontal or temporal lobes, cortical regions that also undergo the typical degenerative changes that occur in AD, or in the cerebellum that is largely spared by this condition. We have prevously suggested that the heterogeneous clinical, histopathological, neurochemical, and genetic features of AD more closely resemble a syndrome with multiple contributing etlogies than a disease with a unitary cause [10, 11]. The expression of C10orf112 in some, but not all, brain regions affected by AD is consistent with this hypothesis.

MATERIALS AND METHODS

Synthesis of cDNA from Human Hippocampal RNA

The cDNA first strand product was prepared from human hippocampus total RNA (BioChain Institute, Inc., Hayward, CA) using the SuperScript™ III First Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Briefly, 1 μg of total RNA was combined with 5 mM oligo(dT)₂₀ and 1 mM dNTP mix in a total of 10 μl. The oligo(dT)₂₀ was allowed to anneal to the RNA poly(A) tails at 65°C for 5 min, followed by incubation on ice for 1 min. An equal volume of DNA synthesis mixture was then added, consisting of 2 × RT buffer, 10 mM MgCl₂, 20 mM DTT, 40 U RNaseOUT™, and 200 U SuperScript™ III reverse transcriptase. After incubation at 50°C for 50 min to extend the primer, the enzyme was inactivated at 85°C for 5 min. The residual RNA was degraded with 2 U E. coli RNase H at 37°C for 20 min. The first strand synthesis product was used in subsequent PCR reactions without further purification.

PCR Amplification of Predicted C10orf112 Exons

To confirm its presence in the C10orf112 transcript, each predicted exon was amplified in a PCR reaction using the cDNA first strand material prepared as described and a pair of custom synthesized (Invitrogen, Carlsbad, CA) oligonucleotide primers as shown in Table 2, which were designed using the software program Primer3 (http://frodo.wi.mit.edu/cgibin/primer3/primer3_www.cgi) based upon sequence derived from the NCBI database (http://www.ncbi.nlm.nih.gov/; Build 35.1). A reaction employing primers from β-actin served as a positive control for PCR amplification, and a reaction using RNA alone (prior to first strand synthesis) as template served as a negative control to rule out the presence of genomic DNA contamination. Each reaction consisted of 1 unit of Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA), 1 × buffer and 1.5 mM MgCl₂, both supplied with the enzyme, 0.2 mM of each dNTP (Amersham Pharmacia Biotech, Piscataway, NJ), 5 pmole of each oligonucleotide primer, and 2 μl of first strand synthesis product, in a total volume of 25 μl. Following an initial denaturation at 94°C for 5 min, the fragments were amplified by 40-50 cycles of denaturation at 94°C for 30 sec, annealing at 62°C for 30 sec, and extension at 72°C for 1 min, followed by a final extension at 72°C for 7 min to assure all transcripts are full length. Bands were resolved on 1.2% agarose (Invitrogen, Carlsbad, CA) gels, stained with ethidium bromide (Amresco, Solon, OH), and photographed.

Table 2.

Primers Used for PCR Amplification of C10orf112 Predicted Exons and β-Actin (ACTB)

Exon	Forward Primer	Reverse Primer	Predicted PCR Product Size(bp)
1*	ATGGAGCATGGATCCAAAAG	CAAGCAGCCTGAACTGAAAG	110
2	AAGGTCAAATGGCTTCAACG	CAAGCAGCCTGAACTGAAAG	77
3	TGGAATTTTACTGTGTCAAGAAGC	TTCATCTGTAGCATCAAATCTGC	91
4	TGTGAGTGGACGTCAGAAGC	TCATCATCACCTCCTTGATCC	113
5	TTATGTATGGGTAGGCGCTAAG	CAGTCTTACTCTCAGGACACTGC	151
6	GCGATCAAAAGTAGTGAAACTCAAC	GATGCCAGTCTATGCTTTTCTC	84
7*	ACCAAGACTTTAGATTTCCTAGTACTTT	GGAAATCTAAAGTCTTGGTAGTAAAATTG	128
8	TTTGGACATACAACATATCAACTCAC	AATGTCTTCAGATCTTCTGGTATTAAC	76
9	GGGACTCTTTTGAGCCAGAG	AGTCCTGCCGAGAGTCACAG	160
10	ACTTTGAGTCGGGTTTCTGC	TGATCAGCTGCAGGAAAGTG	109
11	CGTTTATTTATTTGGAGGCACAG	GACATGGGGTAGAGGCAGTG	93
12	TGCAGTTTTGGTATCATTTGTC	GTAGATTCTCCCGCTTCTGC	160
13	TGTTTTGTCGTCAAATGCTACC	GCTTTGTGTACTGGTCCTTGG	100
14	AAGCAAACAGCTGTGATTGG	AGCCTGGTGCTCAAAATCAG	101
15	GGCATTTTATGTTCATTCTGAAG	AAGGAAAGTATGCATCCAGGTC	103
16	ACAATCAGGGCAAACAATGG	CGATCACCACAGTCATCCAC	256
17	CACTTAACACAGGGCCAATG	CTGCTGGAAATGTTGAGGTG	301
18	CTTCAGTGGGAGATGGCTTC	CAGGGTAGAGGGTGCAGTCC	75
19	CTCCCAACTCCACCAGAAAC	TCTGATTTGTCAGGGCAGTC	146
20	TGGAAGTTTGCAGCTTTGAG	GATCCACTGATGGCCTGTG	153
21	TGGCTGGTACCTGTATGCTG	TGGCCCCATTCATATGTGTC	128
22	ATTGTGGGCTCAAACTGGAC	CCCATAACTACCTGTATTACACACC	103
23	TTCAGAGCCAAACGTGGTATC	GCTTGTCTTTGGCAATGCAG	151
24	AGTGGGCGCTGTGATTTC	AATTCTGGCAGCTCTCATGG	201
25	GGGCACTCACCTTAATGCAG	CAGTTCTTCTTGCACGGAATC	255
26	TGCTATAGGCTGGAACAAAGC	GAGAGCCAACCCCTAAAACC	157
27	TGATGCGACGGCATAAATATC	ATGATCTGAGCTGCCAAAAG	164
28	GGCACTTCATGTATCTGGAAGC	GAATCTGAATGGATCCTGTGG	150
29	AAAGGACTATCAAAAGTATGGCAAG	AGCTCCTCCTCCCAGGTC	153
30	GGATGATGGGTCGGAAAAG	GCTGAGATCCTTGGACCTTACTC	75
31	CTCCCAATCCGGTTCTATCC	CTTCCTCTTGGGCATTCTGG	130
32	CTGAGCTTTGTCCGGAAATC	AGTCACAAAGAAGGTGGGATG	75
33	GCAAATGTGATAGACATTTCCAG	CAGGAATTCTGCTGCCAATG	200
34	TGGTCAGCACTTCCTGTACG	ACACATTCCAAGGCTTGCTG	100
35	CTGGTGTGGTCAGTGATTGG	CACACTCTGGGGTAAAAGAGATG	160
36	CCATCACCCTGTGAAGCTG	GCACTCCATCGCATAGCAG	211
37	ATGGAGCTCTGGTGTGTGC	AGCTGGAAATCCATGCAGTC	88
38*	GGCTTCGCAGAGTCGGTAAC	ACAGGCCGGGGAGCTATC	303
39	ATAGCTCCCCGGCCTGTC	CCACACTTTGAGGAGGGAAG	302
ACTB	GCTCGTCGTCGACAACGGCT	CAAACATGATCTGGGTCATCTTCTC	353

It was not possible to develop both forward and reverse primers within predicted exons 1 or 7 due to their small lengths (45 and 55 bps, respectively), or predicted exon 38 because of its high GC content. To assay for the expression of these predicted exons, a forward primer in exon 1 was paired with a reverse primer in exon 2; forward and reverse primers for exon 7 were paired with the reverse primer for exon 8 and the forward primer for exon 6, respectively; and a forward primer within exon 38 was paired with a reverse primer within exon 39. ACTB, β-actin.

Northern Blot Analysis of Human Brain mRNA

Preparation of probe

The DNA probe was a PCR product spanning exons 23 through 26. Primers 23F and 26R (Table 2) were used to amplify this region from 2 μl of cDNA first strand material in a 40 cycle PCR reaction as described above. The products of three such PCR reactions were combined and run on a 0.75% NuSieve GTG low melting agarose preparative gel (FMC BioProducts, Rockland, ME). The 946 bp band was excised with a scalpel blade, weighed, and the DNA was recovered using the Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI) according to the manufacturer's instructions. The DNA was eluted in 50 μl of nuclease free water, an aliquot was diluted 20 fold, and the concentration was determined spectrophotometrically using a Beckman Instruments (Fullerton, CA) DU-640 spectrophotometer.

Preparation of blot

2 mg of human brain poly A+ RNA (Stratagene, La Jolla, CA) was size fractionated along with an RNA molecular standard (Invitrogen, Carlsbad, CA) on a 1% agarose (Bio-Rad Laboratories, Richmond, CA) gel containing 1 × MOPS buffer (Amresco, Solon, OH) and 6.66% formaldehyde (EMD Chemicals, Gibbstown, NJ). After two 30 min washes in water to remove the formaldehyde, the gel was sliced between the sample and marker lanes. The marker portion was stained with ethidium bromide (Amresco, Solon, OH) and photographed under UV exposure along with a ruler to permit size determinations of bands on the subsequent Northern autoradiogram.

The sample RNA was transferred from the gel to Nytran SPC blotting membrane (Schleicher & Schuell, Inc., Keene, NH) by drawing 20 × SSC (Amresco, Solon, OH) from a sponge, through the gel and membrane, and into Quick Draw Blotting Paper (Sigma, St. Louis, MO) by capillary action. After transfer over night, the blot was rinsed in 2 × SSC for 5 min with gentle rocking, air dried, and then dried under vacuum at 80°C for 2 hours. The RNA was then crosslinked to the membrane by exposure on a UV transilluminator (Fotodyne, New Berlin, WI) for 1 min.

Prehybridization of blots

Prior to use, the blot membrane was wetted in 6 × SSC at room temperature for 5 min. It was then prehybridized in 5 ml of 5 × SSC (Amresco, Solon, OH), 5 × Denhardt's Solution (Amresco, Solon, OH), 50% formamide (Roche Diagnostics, Indianapolis, IN), 1% SDS (Amresco, Solon, OH), and 100 μg/ml denatured and quenched salmon sperm DNA (Invitrogen, Carlsbad, CA) at 42°C for 3 hours in a Fisher Scientific (Pittsburgh, PA) hybridization oven at 6 rpm.

Labeling of probe and hybridization

25 ng of gel purified exon 23-26 spanning PCR product was labeled with ³²P to a specific activity of 9.1×10⁸ cpm/μg using the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA) in a 50 μl reaction containing 20 μM each of dATP, dTTP and dGTP, 50 μCi of [α-³²P] dCTP at 3000 Ci/mmole (PerkinElmer Life Sciences, Boston, MA), 15 μl of Random Primers Buffer Mixture, and 3U of Klenow Fragment at 25°C for 2 hours. The reaction was terminated by the addition of 5 μl of Stop Buffer. Unincorporated nucleotides were removed using a Sephadex G-25 Quick Spin Column (Roche Diagnostics, Indianapolis, IN). The probe was denatured at 100°C for 5 min and quenched on ice. The entire quenched probe was added to the prehybridization solution, and the blot was hybridized over night at 42°C and 6 rpm.

Washing of blot and autoradiography

After draining off the hybridization solution, the blot was washed twice in the hybridization tube in 25 ml of 2 × SSC / 0.1 % SDS (Amresco, Solon, OH) for 5 min at room temperature at 12 rpm. Two additional 5 min washes were performed in 25 ml of 0.2 × SSC / 0.1 % SDS (Amresco, Solon, OH) at room temperature and 12 rpm. One final wash was performed in 200 ml of 0.2 × SSC / 0.1 % SDS (Amresco, Solon, OH) at room temperature in a plastic tray with moderate rocking. The membrane was patted dry, wrapped in Saran Wrap, and visualized by autoradiographic exposure of Kodak (Rochester, NY) BioMax MR film in a Kodak X-Omatic cassette with intensifying screens at -70°C for exposure times that varied from over night to 6 days.

High Throughput Northern Blot for Determining Human Tissue-Specific Expression

Prehybridization of blot

The High Throughput Human Northern Blot (BioChain Institute, Inc., Hayward, CA) was manipulated according to the manufacturer's instructions. The membrane was prehybridized in 3 ml of 5 × SSC (Amresco, Solon, OH), 5 × Denhardt's Solution (Amresco, Solon, OH), 50% formamide (Roche Diagnostics, Indianapolis, IN), 1% SDS (Amresco, Solon, OH), and 100 μg/ml denatured and quenched salmon sperm DNA (Invitrogen, Carlsbad, CA) at 42°C for 3 hours in a Fisher Scientific (Pittsburgh, PA) hybridization oven.

Labeling of probe and hybridization of blot

The DNA probe was labeled using the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA) with minor modifications from the manufacturer's instructions. 46 ng of gel purified exon 23-26 spanning PCR product was labeled with ³²P in a 50 μl reaction containing 20 μM dATP, 20 μM dTTP, 50 μCi each of [α-³²P] dCTP and [α-³²P] dGTP at 3000 Ci/mmole (PerkinElmer Life Sciences, Boston, MA), 15 μl of Random Primers Buffer Mixture, and 3U of Klenow Fragment at 25°C for 2 hours. The reaction was terminated by the addition of 5 μl of Stop Buffer. Unincorporated nucleotides were removed by passing over a Sephadex G-25 Quick Spin Column (Roche Diagnostics, Indianapolis, IN) according to the manufacturer's instructions. An aliquot of the final product was removed and counted in a Beckman Instruments (Fullerton, CA) LS-1801 Liquid Scintillation Counter, and the specific activity was determined to be 3.3×10⁹ cpm/μg. The probe was denatured at 100°C for 5 min and quenched on ice. The entire quenched probe was added to the prehybridization solution, and the blot was hybridized over night at 42°C and 6 rpm.

Washing of blot and autoradiography

After thoroughly draining the hybridization solution, the blot was washed in three changes of 10 ml of 2 × SSC / 0.05% SDS (Amresco, Solon, OH) at room temperature for 15 min with gentle rocking. The majority of the remaining counts were removed by one additional wash with 100 ml of 0.1 × SSC / 0.1% SDS at room temperature for 20 min with gentle rocking. The blot was patted dry, wrapped in Saran Wrap, and visualized by autoradiographic exposure of Kodak (Rochester, NY) BioMax MR film in a Kodak X-Omatic cassette with intensifying screens at -70°C for 72 hours.