Background
G72 and G30 are overlapping genes transcribed from opposite strands on chromosome 13q33. They were initially identified within a 65 kb region containing markers associated with schizophrenia in two independent disease cohorts [
1]. Genetic association between the G72/G30 locus and both schizophrenia and bipolar disorder has subsequently been reported in several studies and is supported by a recent meta-analysis [
2]. Thus, the G72/G30 locus may represent a common susceptibility region for both disorders. Whilst the combined evidence for association with both disorders can be considered robust, associated alleles are not consistent across studies and there may not be a distinct haplotype correlated with disease susceptibility [
2].
Multiple transcripts for each gene have been cloned by reverse transcription-PCR from human brain, spinal cord and testis cDNA libraries. The longest G72 open reading frame is predicted to encode a putative 153 amino acid protein, isolated from amygdala, caudate nucleus, spinal cord and testis libraries [
1]. No homology was found between any of the putative protein products from the transcripts of either gene and any known protein. Putative LG72 orthologues were identified
in silico in rhesus monkey and ape genomes leading to the suggestion that it might represent a rapidly evolving primate-specific gene.
In vitro transcription/translation assays with the cloned G72 and G30 candidate transcripts yielded a protein product for LG72 only, which generated a 24-kDa translation product (pLG72) that was reported to localise at the Golgi apparatus in transiently transfected cells [
1]. A more recent study, however, demonstrated mitochondrial localisation of overexpressed G72 [
3]. Yeast two-hybrid screening identified D-amino acid oxidase (DAO) as a putative protein interactor of pLG72 and recombinant pLG72 was demonstrated to behave as an activator of DAO
in vitro [
1]. On the basis of this observation, which has subsequently not been reproduced [
3], G72 has been renamed D-amino acid oxidase activator (DAOA). DAO catabolises D-amino acids including D-serine which is a co-agonist at the N-methyl-D-aspartate (NMDA) receptor [
4]. As NMDA receptor hypofunction has been implicated in the pathophysiology of schizophrenia, the report of a functional interaction between G72 and DAO suggests a pathway whereby G72 could modulate DAO activity, D-serine levels and NMDA receptor activity and, thus, contribute to the disease pathology.
To date, however, there have been no convincing reports of expression of native G72 (or G30) protein and accordingly no confirmation of the G72-DAO interaction
in situ. Moreover, reports of native G72/G30 mRNA expression are limited. In addition to the original study by Chumakov
et al. [
1], cDNAs have also been amplified from human and testis libraries by Hattori
et al. [
5] although only after ≥ 40 cycles of PCR. Korostishevsky
et al. [
6] have subsequently utilised real-time PCR to detect and quantify G72 and G30 mRNA in human post-mortem dorsolateral prefrontal cortex, reporting a tendency towards overexpression of G72 but not G30 transcripts in schizophrenic versus control brain. In order to learn more about the potential role of this schizophrenia susceptibility gene, we have analysed the gene structure and genomic context of G72/G30 and investigated expression of G72 in human tissues at the mRNA and protein level.
Methods
Cloning of human G72
Full-length human G72 cDNA (identical to published GenBank sequence AY138546, which results in the longest predicted protein) was amplified from human testis Marathon cDNA (Clontech) by nested PCR with forward primer 5'-GACCCAAAATGCTGGAAAAGCTGAT-3' and reverse primer 5'-CATCAGAAGGATTGGCTGGGAAGAAT-3' and cloned into pCDNA3.2/GW/D-TOPO. In order to obtain N-terminally epitope-tagged G72 in mammalian expression vectors, G72 cDNA was PCR-amplified from above template vector with forward primer 5'-CACCATGCTGGAAAAGCTGATGGGTGC-3' and reverse primer 5'-TCATTCAGCTTTGGTAGAAGTTATTTCCTTGTGG-3'. The PCR product was cloned into pENTR/D-TOPO entry vector using pENTR Directional TOPO Cloning Kit (Invitrogen), following manufacturer's instructions. Resulting entry clones were used in LR recombination reactions (Invitrogen) with pDEST26 (Invitrogen) and pDEST12.2-FLAG (kind gift from Jan Kopf) Gateway destination vector, to obtain G72 cDNA in frame with the respective N-terminal epitope tag (6 × His or FLAG-tag).
Cell culture and transient transfection
HEK-293 cells were cultured in Minimum Essential Medium (MEM, Gibco) supplemented with 10% fetal calf serum and non-essential amino acids (Gibco) at 37°C and 10% CO2. Cells were transiently transfected using Fugene-6 (Roche) according to the manufacturer's instructions. Cells were plated one day before transfection at 4 × 106 cells per 10 cm-plate, transfected with 5–7 μg plasmid DNA and incubated with transfection mix (complex of plasmid-DNA and Fugene 6 at a ratio of 1:3) for ~40 h prior to preparation of protein lysates.
RNA isolation and Northern Analysis of G72 mRNA
Total RNA was isolated from HEK-293 cells (mock-transfected or overexpressing His-G72) using the Qiagen RNeasy protocol, according to manufacturer's instructions. The optional on-column DNase digestion step was performed using the Qiagen RNase-free DNase set. RNA was eluted in 40 μl of RNAse-free water and quantified using a spectrophotometer.
For northern gels, 10 μg total RNA in NorthernMax Formaldehyde sample buffer (Ambion) were loaded per lane. Northern blotting was performed as described in "Promega protocols and applications guide", third edition (1996). Briefly, a formaldehyde-containing 1.2% agarose gel was prepared in 1× MOPS buffer. The gel was run in 1× MOPS buffer at 5 V/cm with regular buffer mixing. After completion, the gel was washed three times for 10 min in RNAse-free dH2O, soaked 15 min in 0.05 M NaOH, neutralised 15 min in 100 mM Tris-HCl pH 7.5 and soaked for 30 min in 20× SSC before capillary transfer (over-night) onto Hybond-N+ nylon membrane (Amersham). RNA was UV-crosslinked to the membrane (UV Stratalinker) and dried for at least 2 h prior to further use.
Northern blots were probed with a 390 bp NotI-XbaI fragment of G72 cDNA (corresponding to G72 clone AY138546 cDNA without the last 3'-terminal 85 nucleotides) or with beta-actin control probe (BD Biosciences). Probes were 32P-dCTP-labelled using Rediprime II Random Prime Labelling System (Amersham) according to manufacturer's instructions and purified with spin columns (Amersham). Prehybridisation and probe hybridisation were performed with Rapid-hyb buffer (Amersham) at 70°C, according to manufacturer's instructions. Stringency washes were performed according to the Rapid-hyb protocol (Amersham). Briefly, membranes were washed for 20 min at room temperature in washing buffer 1 (2× SSC, 0.1% (w/v) SDS) and twice for 15 min at 65°C in washing buffer 2 (0.1× SSC, 0.1% (w/v) SDS). Signals were detected using a sensitive phosphorimager (Typhoon Trio, Amersham), with exposure times between 2 h and over-night.
Alternatively, commercial human brain multiple tissue Northern (MTN) blots (Clontech/BD Biosciences MTN blot II and blot V) were used for analysis of G72 expression in human brain. Blots (nylon membrane) contained approximately 2 μg of poly A+ RNA per lane from eight (MTN blot II) or six (MTN blot V) different brain regions. Hybridisation conditions were essentially as described above, but less stringent conditions (washing buffer 2 at 55°C instead of 65°C) and longer exposure times of up to three days (phosphorimager, Typhoon Trio, Amersham) were used in order to detect even weak signals. Following G72 probing, membranes were stripped with boiling 0.5% SDS and reprobed with a beta-actin-probe to test for equal loading and good RNA quality.
Alternatively to radioactive labelling, northern blots were probed with biotinylated dCTP, using the "SpotLight Random Primer Labeling Kit" according to manufacturer's instructions (BD Biosciences, Clontech). Hybridisation and chemiluminescent detection were carried out using the "SpotLight Chemiluminescent Hybridisation and Detection Kit" (BD Biosciences, Clontech). This method appeared to be less sensitive compared to radioactive detection.
Real-time RT-PCR
Five different primer sets were designed within G72 exons that are common to most G72 splice variants (Exons 2, 4 and 7). Sequences are shown in table
1; G72 #2 corresponds to the primer pair used by Korostishevsky
et al. [
6]. All primer sets were tested by SYBR Green PCR (and TaqMan PCR, if applicable) on human genomic DNA (Clontech; 0.1 ng – 0.1 μg per reaction), human adult total brain, amygdala, testis and fetal brain cDNA (Marathon Ready cDNA, Clontech; 0.1 – 0.5 ng cDNA per reaction) and cDNA from HEK-293 cells overexpressing recombinant His-G72 cDNA (see below). Total RNA was isolated from HEK-293 cells (mock-transfected or overexpressing His-G72) as described above. Samples were reverse transcribed (in triplicate) into cDNA using the Omniscript RT-PCR kit (Qiagen). Briefly, for each 20 μl reverse transcription reaction, 300 ng RNA in a total volume of 12 μl was added to 8 μl RT-PCR mastermix (final concentrations/amounts: 1 μM Oligo-dT primer (Invitrogen), 0.5 mM each dNTP, 1× RT Buffer, 10 units RNase inhibitor (Invitrogen), and 4 units Omniscript Reverse Transcriptase). Negative controls were prepared as above, but without Reverse Transcriptase. Reactions were incubated for 60 min at 37°C. The masterplate was aliquoted into 20 optical 96-well plates using the Hydra-96 robot (Robbins Scientific, GRI). Taqman PCR was performed using an ABI 7700 Sequence Detection system (PE Applied Biosystems) as previously described [
7]. Briefly, the PCR reaction mastermix contained a final concentration of 200 nM of each of the forward and reverse primers (see table
1; Invitrogen), 200 nM FAM-TAMRA probe (Qiagen, Operon), and 1× Quantitect master mix (Qiagen). For SYBR Green chemistry, 1× SYBR Green mastermix (Perkin Elmer) was used, and no probe was added to the reaction. The PCR amplification was performed for 40 cycles, consisting of 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s and 60°C for 1 min. Standards were analysed on each plate using dilutions of genomic DNA. All measurements were performed in triplicate. The housekeeping gene cyclophilin was run as a control to ensure comparable amounts of cDNA in all wells. Ct values were converted into arbitrary units according to manufacturer's instructions (Applied Biosystems).
Table 1
Primer and probe sequences for G72 and cyclophilin
G72 #1 (Exon 4) | F: 5'-GGATGGAAGAGAAGGCATGAGG-3' |
| R: 5'-TCTGCATAGGGCTGAGGAAGG-3' |
| P: 5'-ATTTGGAAATGGCACAGAGG-3' |
G72 #2 (Exon 7) | F: 5'-GCTGAATGAGTTTGGAAGCA-3' |
| R: 5'-TGGGTCCCAGACACAGAGT-3' |
G72 #3 (Exon 2) | F: 5'-CATTGGGTAAAATCTACTTCATAGG-3' |
| R: 5'-GAGTTTAGAGAGTTTTCAGATTTGC-3' |
G72 #4 (Exon 4) | F: 5'-GCATGAGGACGGCTATTTGG-3' |
| R: 5'-TCTGCATAGGGCTGAGGAAGG-3' |
G72 #5 (Exon 7) | F: 5'-CCGCAGGCAGCCTCTAGAAC-3' |
| R: 5'-TCAGAAGGATTGGCTGGGAAG-3' |
| P: 5'-CCAGCAAAAAGACCAGTCATGCAACCA-3' |
Cyclophilin | F: 5'-TGAGACAGCAGATAGAGCCAAGC-3' |
| R: 5'-TCCCTGCCAATTTGACATCTTC-3' |
| P: 5'-AGCACCAATATTCAGTACACAGCTTAAAGCTATAGGTT-3' |
Generation of rabbit polyclonal antibodies against G72
Rabbit polyclonal antibodies were generated by Cambridge Research Biochemicals (Billingham, UK) against two peptides within the predicted G72 variant AY138546 protein sequence (antibodies #1410 and #1411: antigenic peptide VTRKEGWKRRHEDGY-acid, with an N-terminal cysteine added for coupling; antibodies #1412 and #1413: antigenic peptide SKDRRQPLERMWTC-amide). Crude antisera were purified by affinity chromatography on Thiopropyl Sepharose 6B derivatised with the antigen. The concentration of purified antibodies in glycine eluates was as follows: #1410: 1.69 mg/ml, #1411: 0.96 mg/ml, #1412: 0.64 mg/ml and #1413: 0.33 mg/ml.
Preparation of protein lysates
Transfected HEK-293 cells were washed twice with cold PBS and lysed in 500 μl – 1 ml ice-cold Triton-X-100 lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5% Glycerol, 1% Triton-X-100, supplemented with "Complete" Protease Inhibitor Cocktail (Roche) and phosphatase inhibitor cocktail (Sigma)) per 10 cm-plate. Samples were rotated for 15 min at 4°C and centrifuged 20 min at 14 000 g, 4°C. Supernatants were stored in aliquots at -80°C.
Rat tissue was dissected from embryonic, young or adult rat (CD, Charles River), ground to a powder in liquid nitrogen and stored at -80°C. Total protein lysates were prepared from tissue powder using RIPA lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP40, 0.5% Na-Deoxycholate, 0.1% SDS, supplemented with "Complete" Protease Inhibitor Cocktail (Roche) and phosphatase inhibitor cocktail (Sigma)). Samples were tip sonicated briefly, rotated at 4°C for 15 minutes and centrifuged for 20 min at 14 000 g, 4°C. Supernatants were collected and protein concentration was determined using the BCA Protein Assay Reagent Kit (Pierce Biotechnology).
Immunoprecipitation
Cell lysates from transiently transfected HEK-293 cells were prepared as described above, using 750 μl to 1 ml Triton-X-100 lysis buffer per 10 cm plate. Supernatants were incubated with 3 μg of anti-G72 antibody (#1410 or #1411) for 2 h at 4°C, followed by incubation with protein-G-sepharose (Sigma) for another 2 h at 4°C. Immunocomplexes were washed twice with lysis buffer and once with 1× TBS before elution with 40 μl of 2× Laemmli loading buffer containing β-Mercaptoethanol. Eluates were analysed by SDS-PAGE and Western blotting as described below.
SDS-PAGE and Western blotting
Total lysates from transfected HEK-293 cells (lysate preparation see above; 10 – 20 μg total protein per well), rat tissue lysates (40 – 50 μg per well) or human protein medleys (50 μg per well; Clontech/BD Biosciences: SDS-solubilised proteins prepared from whole human tissues, prepared under conditions designed to ensure maximal representation of tissue-specific proteins) were separated on 4–20% Novex Tris-Glycine gels (Invitrogen). Proteins were transferred to nitrocellulose (Amersham) or PVDF (Millipore) membrane by wet-blotting. For conventional Western analysis, membranes were blocked with 5% non-fat milk powder in TBST (25 mM Tris pH 7.5, 150 mM NaCl, 0.1% Tween-20) for 1 h at room temperature, incubated overnight at 4°C with primary antibody (anti-G72 #1410, #1411, #1412 or #1413 at 1–5 μg/ml; anti-His H-15, goat polyclonal antibody, Santa Cruz, at 0.5 μg/ml; anti-actin, mouse monoclonal, ascites fluid, Chemicon, 1:20 000 dilution) in blocking buffer, washed five times in TBST, incubated with goat-anti-rabbit-IgG-POD (Santa Cruz, 1:10 000 in blocking buffer) for 1 h at room temperature, washed again and developed using the ECL Plus (Amersham), ECL Advance (Amersham) or Visualizer (Upstate) detection system. Alternatively, Western analysis was performed using the LI-COR Odyssey® Infrared Imaging System. Experiments were performed according to manufacturer's instructions. Briefly, proteins were separated as described above alongside Odyssey pre-stained molecular weight markers (Licor, fluorescence in 700 nm channel) and transferred onto PVDF membrane (Millipore) by wet blotting. Membranes were blocked with Odyssey blocking buffer (LI-COR), diluted 1:1 with 1 × PBS. For primary antibody incubations, the blocking buffer was supplemented with 0.1% Tween-20, and for secondary antibody incubations, 0.1% Tween-20 and 0.01% SDS were added. Primary antibody concentrations and incubation times were as described above. Blots were washed five times with PBST (1 × PBS, 0.1% Tween-20) prior to a 1 h incubation (room temperature, protected from light) with secondary antibodies (IRDye800CW goat-anti-rabbit, LI-COR, 1:10 000; Alexa Fluor® 680 donkey-anti-goat, Molecular Probes 1:15 000). Membranes were washed again and visualised with LI-COR Odyssey® Infrared Imaging System (700 nm and 800 nm channels).
Immunocytochemistry
HEK-293 cells were seeded on 13 mm glass coverslips (coated with poly-D-lysine, Sigma), transiently transfected with FLAG-G72 or His-G72 cDNA and fixed with 4% paraformaldehyde in PBS (pH 7.2) for 15 min at room temperature 36 h post-transfection. Cells were blocked for 1 h at room temperature using PBS with 0.1% Triton X-100, 10% normal goat serum (NGS) and 1% bovine serum albumin (BSA). Following one wash with 1× PBS, cells were incubated overnight at 4°C with primary antibodies in PBS with 5% NGS and 0.5% BSA. Rabbit polyclonal anti-G72 #1410 was used at a concentration of 3 μg/ml, #1411 at 5 μg/ml and anti-FLAG M2 mouse monoclonal antibody (Sigma) at 0.3 μg/ml. After four 10 min washes with 1× PBS, cells were incubated for 1 h at room temperature with Alexa Fluor®488 goat anti-rabbit IgG and Alexa Fluor®594 goat anti-mouse IgG at 5 μg/ml (Molecular Probes). Cells were washed four times with 1× PBS, mounted with ProLong® Gold antifade reagent (Molecular Probes) and analysed using an Olympus microscope and Image-Pro Plus imaging software (MediaCybernetics).
Gene sequence information was obtained from the National Center for Biotechnology Information (NCBI) Genbank database. Variants AY138546 and AY138548 were used to represent G72 and G30 respectively. In early 2007, NCBI replaced AY138456 with a longer sequence, NM_172370.3, which is extended at the 5'UTR by 228 bp. This study will refer to AY138546 as the G72 variant; however, analyses have been repeated on NM_172370.3 for completeness.
G72 gene environment
The region around G72 was assessed using the UCSC human genome browser (March 2006; NCBI build 36.1). All known genes (with a Genbank RefSeq transcript) in a specified region can be displayed using this viewer.
Promoters were predicted for G72 and G30 by collating and analysing results from several promoter prediction programs. Each of the features predicted by the programs adds a layer of confidence to the promoter prediction. The programs report CpG rich regions using EMBOSS newcpgseek (threshold CpG score of 17) & cpgplot, and CpGProD [
1,
8,
9], nucleosomal binding sites using recon [
10], transcription start sites using eponine [
11], human transfac transcription factor site clusters using Cister [
12], and EMBOSS matrix/scaffold attachment regions using marscan [
13].
Identification of Functional motifs
Secondary structure elements and previously identified motifs present in a number of motif databases were scanned for using the protein sequences of G72 and G30. Twelve structure predictions were performed including transmembrane helix prediction TMHMM [
14], and prediction/location of signal sequences SignalP [
15]. Six motif databases were searched including a database of protein motif regular expressions and profiles PROSITE [
16] and a database of protein fingerprints, each fingerprint usually consisting of multiple motifs PRINTS [
17].
EST searches
An extensive EST search was carried out using BlastN run through the following subsets of Genbank: Genbank EST, Genbank cDNA Genbank Human EST and Genbank Rat EST against more than fifteen species. The default BLAST Expectation thresholds used were lowered to 100 to show weaker matches. In addition, using the NCBI blast tool, G72 (both AY138546 and NM_172370.3) and G30 transcript sequences were run through BLASTN against all of the available EST and cDNA databases.
Ortholog searches
To search for evidence of G72 or G30 orthologues, several Genbank databases were searched. The nucleotide transcripts sequences were run through BLASTN against all the (more than fifteen) species specific EST and cDNA GenBank subdivisions. The protein sequences were run through BLASTP against fourteen species specific GenBank protein databases. The sequences were also run through BLASTN against over fifteen general primate and rodent genomic sequence databases.
Identification of RNA instability motifs
The G72 UTRs were searched for predicted AU-rich elements (AREs) using the motif ATTTA [
18].
Identification of miRNAs
Two miRNA prediction methods, miRANDA [
19] and RNA hybrid [
20] were run on the G72 transcripts to search for evidence that the gene is regulated by miRNAs. The thresholds for free energy of binding for miRANDA was -15, and RNA hybrid is -25. P values for both methods were < = 0.05.
Gene prediction
All of the gene prediction algorithms available as tracks in the UCSC (May 2004 version, NCBI build 35) were run on G72 and G30 sequences. The majority of the methods used to predict genes are based on gene models reconstructed solely from mRNA and EST evidence. Therefore in this analysis we have focused on prediction methods that also use evidence from other sources.
N-SCAN [
21] combines biological-signal modelling in the target genome sequence along with information from a multiple-genome alignment to generate de novo gene predictions. Gene predictions were taken from the SGP program, which predicts genes using mouse/human homology [
22]. The exoniphy program identifies evolutionarily conserved protein-coding exons in a multiple alignment using a phylogenetic hidden Markov model phylo-HMM [
23], a statistical model that simultaneously describes exon structure and exon evolution. Retrogenes are predicted by showing processed mRNAs that have been inserted back into the genome since the mouse/human split [
24]. RetroGenes can be either functional genes that have acquired a promoter from a neighboring gene, non-functional pseudogenes, or transcribed pseudogenes. The ExonWalk program merges cDNA evidence together to predict full length isoforms, including alternative transcripts. The Vertebrate Genome Annotation (VEGA) database build 30 is designed to be a central repository for manual annotation of different vertebrate finished genome sequence [
25]. Finished genomic sequence is analysed on a clone by clone basis using a combination of similarity searches against DNA and protein databases as well as a series of
ab initio gene predictions using GENSCAN [
26]. In addition, comparative analysis using vertebrate datasets such as the Riken mouse cDNAs and Genoscope
Tetraodon nigroviridis Ecores (Evolutionary Conserved Regions) are used for novel gene discovery.
The coding region of the chimpanzee G72 was predicted from chimpanzee genomic sequence, using human G72 protein sequence (NP_758958) as a template and the GENEWISE gene prediction tool [
27].
Discussion
The G72/30 locus has received considerable attention due to replicated disease association data for schizophrenia and bipolar disorder. Furthermore, it has been proposed that the G72 peptide is a direct modulator of DAO activity and as such has mechanistic relevance to schizophrenia.
In order to better understand the function of G72, we investigated expression of both the G72 mRNA and putative protein in a series of human CNS and peripheral tissue samples. We were unable to detect the G72 message using either Taqman or northern analysis (13 brain regions) in any of our human tissue samples including cerebral cortex, amygdala, caudate nucleus, spinal cord and testis, i.e. tissues from which G72 had previously been cloned [
1,
5]. We also analysed over 25 control and 25 schizophrenic patient BA10 samples and could not detect the G72 transcript in the control or disease samples. Furthermore, we generated two specific and sensitive antibodies to a G72 peptide (amino acids 55–69 of LG72). Although the recombinant protein was readily detected by western blotting and immunocytochemistry, native protein was not observed in different tissues reportedly expressing G72.
Several gene detection methods that combine DNA/protein database searches and comparative analysis of conserved vertebrate DNA sequences predict the G72 gene at the locus on human chromosome 13q.33. Our in silico analysis, however, indicates that the G72 gene lacks features of a robustly transcribed gene and moreover contains an instability sequence that is generally found in unstable mammalian mRNAs [
31]. There is also evidence for the presence of miRNA cognate sites, which regulate the expression of target genes by binding to their mRNAs [
20]. These findings suggest that the G72 transcript may be rapidly removed if or when it is expressed. Furthermore, endogenous RNA interference between G72 and G30 (which is transcribed in the opposite direction) may also account for very low or no detectable expression, as suggested previously [
2].
There is some evidence that G72 is only present in higher primates and not in rodent species [
1]. Extensive genomic sequence analysis identifies some regions of homology to human G72 in
Pan troglodytes (chimpanzee),
Gorilla gorilla (gorilla),
Pongo pygmaeus (orangutan),
Hylobates sp. (gibbon),
Macaca mulatta (rhesus monkey) and
Callithrix jacchus (marmoset) genomic sequences. There are, however, no full-length primate orthologues and for example, there are two stop codons in the predicted chimpanzee G72 transcript sequence, suggesting this could in fact be a pseudogene and furthermore that AY138546 (and NM_172370.3) is unique to humans.
Although we (and others: Dr. Phil Burnet, University of Oxford, personal communication) have been unable to detect G72 mRNA, it has been possible to clone G72 from several libraries [
1,
3,
5]. Indeed, we have cloned G72 cDNA from a testis cDNA library, but nevertheless failed to detect appreciable levels of G72 from the same testis cDNA sample using TaqMan or SYBR Green RT-PCR. Since these methods can detect tens of copies of a gene, expression levels below this detection limit are indicative of either no expression within any of the investigated tissues, or of an extremely localised or tightly regulated expression.
A significant percentage of human mRNA transcripts are either not polyadenylated or are bimorphic [
32]. Although we failed to detect G72 transcripts in polyadenylated mRNA samples (northern analysis; real-time semiquantitative RT PCR analysis using cDNA from cell lines or Clontech "Marathon Ready cDNA") we also did not detect G72 transcripts in the human post mortem samples that were prepared with random hexamers, which excludes the possibility that our analysis may have failed to detect G72 transcripts based on the method of cDNA synthesis from isolated RNA. A possible reason why e.g. Korostishevsky et al. [
6] observed a G72 PCR product may be explained by the apparent lack of sample treatment to eliminate contaminating genomic DNA and the absence of RT minus PCR controls for signal detection. We used the same primer pair as described by that group (primer set #2) and furthermore confirmed activity on genomic DNA (Fig.
3). It would seem likely therefore that the discrepancy between that study and our study (and others) might be explained by amplification from contaminating genomic DNA.
It should be noted that a recent study also reports difficulty in identifying native G72 protein and failed to replicate the proposed modulation of DAO activity [
3]. The authors claim, however, to detect the native protein in a small (0.01%) sub-population of HeLa cells and in the membrane fraction of human amygdala. We have also analysed G72 expression (mRNA and protein) in the human amygdala and found no detectable G72 present. The antibody generated in the present study is very similar to that described by Kvajo
et al. [
3] in that we used the same epitope to generate the antibody and the cellular localisation of recombinant G72 is also punctuate/tubular rather that Golgi-restricted as suggested by Chumakov
et al. [
1] (Additional File
1). We also detected recombinant G72 by western analysis but not the native protein in tissue or cell extracts. When combining a very sensitive western blot detection method and long exposure times, a signal at around 20 kDa, i.e. just above the expected size of human G72 (18 kDa) was detected in human fetal brain lysate (Additional File
2). However, this signal was only observed with one of our two validated anti-G72 antibodies and a signal at the same apparent molecular weight was also detected in brain lysates from rat, which does not have a G72 gene, thus making it highly unlikely that this band represents G72. These results highlight the importance of suitable negative controls, as well as the danger of potentially misinterpreting bands if technical limits of the method are pushed (e.g. through loading of large protein quantities, long exposure times and high antibody concentration).
While further studies are required to understand the significance of the G72/30 locus to schizophrenia, we propose that if native G72 protein exists at all, it is expressed at such low levels that any physiological role is called in to question. We also conclude that the lack of demonstrable G72 expression in relevant brain regions does not support a role for G72 protein in modulation of DAO activity and the pathology of schizophrenia via a DAO-mediated mechanism. Following the original proposal for a direct G72 and DAO interaction [
1], the interaction has not been confirmed using an endogenous tissue source. More importantly, two recent reports have failed to confirm the previous findings [
3,
33], which is consistent with our own inability to reproduce the protein-protein interaction of G72 and DAO using recombinantly expressed proteins (data not shown).
Whilst we were unable to detect G72 expression in human fetal brain, it remains a possibility that G72 is developmentally regulated and plays a unique spatio-temporal role in human brain development, independent of an interaction with DAO. We have not had the opportunity to investigate this further.
Conclusion
Robust genetic findings from several independent studies have suggested the G72/G30 locus as a common susceptibility region for schizophrenia and bipolar disorder. This study investigated the mRNA and protein expression of the G72 gene product in human tissues and aimed to shed light on the physiological role of the predicted G72 protein, which has thus far not been resolved. Our experimental results, supported by an in silico analysis, do not provide any evidence of appreciable G72 expression in human brain or other human tissues at either the mRNA or the protein level. Thus, native G72 protein, if expressed at all, appears to only be present at extremely low levels that may not be physiologically relevant. This casts doubt on the proposed G72 protein modulation of DAO activity, which has been suggested to contribute to the pathology of schizophrenia. Further studies are required therefore to understand the significance of the G72/30 locus to schizophrenia.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
IB: Experimental design, conducted all experiments that contributed to Figures
1,
3,
4,
5 and Additional files
1 &
2, analysis of results, writing of manuscript, discussion of experimental results. JNCK: concept and interpretation of study, writing of "background" paragraph, manuscript revision. RV:
in silico analysis of G72 expression and writing of the respective manuscript sections (Bioinformatics methods and results). THC: Cloning of epitope-tagged G72 eukaryotic expression vectors. FK: QPCR analysis of CCHPC samples. JB and SH: Collection of CCHPC samples and discussion of experimental results. PRM: concept and interpretation of study, discussion of experimental results, writing of manuscript, manuscript revision.
All authors read and approved the final manuscript.