NARG2 encodes a novel nuclear protein with (S/T)PXX motifs that is expressed during development Naoaki Sugiura*, Vladimir Dadashev† and Roderick A. Corriveau Department of Cell Biology and Anatomy, LSU Health Sciences Center, New Orleans, LA, USA We previously identified a partial expressed sequence tag clone corresponding to NA RG2 in a screen for genes that are expressed in developing neurons and misexpressed in transgenic mice that lack functional N-methyl- D -aspartate receptors. Here we report the first characterization of the mouse a nd human NARG2 gen es, cDNAs a nd the p roteins that they encode. M ouse and human NARG2 consist o f 988 and 982 amino acids, respectively, and share 74% identity. NARG2 does not display s ign ificant homology to other known genes, and lower organisms such as Saccharomyces cerevisiae, Drosophila melanogaster and Fugu rubripes appear to lack NARG2 orthologs. In vitro translation of the mouse cDNA yields a 150 kDa protein. NARG2 localizes to the nucleus in transfected cells, and deletion of a canonical basic nuclear localizatio n signal suggests that t his and other sequences in the protein cooperate for nuclear targeting. NARG2 consists of 16 exons in both mice and humans, 11 of which are ide ntical i n length, and alternative splicing is evi- dent in both species. Exon 10 i s the largest, and exhibits a much higher rate of nonsynonymous nucleotide substitution than the others. In addition, NARG2 contains (S/T)PXX motifs (11 in m ouse NARG2, six in human NARG2). Northern blot analysis and RNase protection d emonstrated that NARG2 is expressed at relatively high levels in dividing and immature cells, and that it is down-regulated upon ter- minal differentiation. The results indicate that NARG2 encodes a novel (S/T)PXX motif-containing nuclear protein, and suggest that NAR G2 may play an important role in the early d evelopment of a number of different cell types. Keywords: human; mouse P19 embryonic carcinoma c ells; cDNA; nuclear protein; SPXX. The N-methyl- D -aspartate (NMDA) receptor, a glutamate- gated ion channel that is p ermeable to Ca 2+ , plays an important role in brain d evelopment by regulating n euronal survival [1,2], migration [ 3], p roliferation [4] and the formation of precise neural circuits [5–7]. Programs of gene expression are also critical for brain development [8–10]. In an earlier study we used cDN A microarray analysis of m ice that lack NMDAR1, the obligatory subunit for NMDA receptor function, to screen for genes that are abnormally expressed in the developing brain i n the absence of NMDA receptors. A group of three genes was identified (termed NMDA receptor-regulated genes): NARG1, NARG2 and NARG3 [11]. These genes lack homology w ith one another, but all three are e xpressed at t he highest levels i n the neonatal brain and fail to b e appropriately down-regulated in NMDAR1 knockout animals. NARG1 ( now termed mNAT1) combines with its evolutionarily conserved cosubunit, mARD1, to form a functional acetyltransferase that may facilitate e ntry into the G 0 phase of the cell cycle [12,13] in higher animals, as it does in yeast. T he significance of NARG3 is unknown, as NARG3 cDNAs corresponding to the l ongest NARG3 transcript on Northern blots lack an open reading frame ( N. Sugiura and R. Corriveau, unpub- lished observations). Here we report the cDNA sequence and e xon–intron structure o f the mouse and human NARG2 genes, and provide evidence that NARG2 encodes a nuclear protein that i s expressed early in the development of a number of different cell types. Moreover, NARG2 contains repeats of (S/T)PXX, a putative DNA-binding motif that is found in many gene regulatory proteins including Kruppel, Hunchback and Antennapedia [14]. T he results suggest that NARG2 is a regulatory protein that i s present in the nucleus of dividing cells and then down-regulated as progenitors exit the cell c ycle and be gin to differentiate. Experimental procedures cDNA library screening Isolation of mouse NARG2 cDNA by cDNA microarray analysis originally identifie d NARG2, an EST ( AA472833) that is expressed at higher than normal levels in the developing brain of NMDAR1 knockout mice [11]. PCR primers w ere d esigned based on the sequence o f t his EST. Because embryonic mouse brain cDNA libraries are not readily available, and because the testis expresses significant levels of NARG2 relative to other adult tissues [11], w e used Correspondence to R. A. Corriveau, Department of Cell Biology and Anatomy, LSU Health Sciences Center, 1901 Perdido Street, New Orleans, LA 70112, USA. Fax: +1 504 568 4392, Tel.: +1 504 568 2011, E- mail: rcorri@lsuhsc.edu Abbreviations:NMDA,N-methyl- D -aspartate; EST, expressed sequence tag; NLS, nuclear localization signal; h,human;m,mouse. Present addresses: *Department of B ioc hemist ry, St. Marianna University, Sugao, Miyamae-ku Kawasaki, Kanagawa, Japan; Department of Neurological Surgery, Emory University School of Medicine, Atlanta, GA, US A. (Received 1 0 August 2004, revised 29 September 2004, accepted 4 October 2004) Eur. J. Biochem. 271, 4629–4637 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04414.x these primers to screen a m ouse testis cDNA library (Origene, Rockville, MO) for full-length NARG2 cDNA. The screen yielded two clones, clone 3H (construct pCMV6-NARG2–3H) and clone 2C (construct pCMV6- NARG2–2C). The coding region of 3H was sequenced on both strands and t he cDNA sequence was registered in the GenBank database (accession number AY244558). pT7-NARG2, the construct used for in vitro translation, was generated by replacing the lucife rase cDNA in Lucif- erase T7 control plasmid DNA (Promega, Pittsburgh, P A) with an Eco RI(blunted)/Bpu10I(blunted) NARG2 fragment from pCMV6-NARG2–3H. An epitope-tagged NARG2 construct, pCS2+MT-NARG2, was generated by ligat- ing an EcoRI/Bpu10I(blunted) NA RG2 fragment from pCMV6-NARG2–3H into the EcoRI and SnaBI sites of pCS2+M T [15,16]. The nuclear localization signal (NLS)-deleted NARG2 mutant constr uct, pCS2+M T- NARG2DNLS, was prepared as described below. A 326 bp NARG2 fragment was generated by PCR using pCMV6-NARG2–3H as a template and the follow- ing primers: 5 ¢-GCTTTTAAAACCAGTTTCCAGG-3¢, and 5 ¢-GAAATTGTCTTCGCGTGGTCTCGTTTCTAC CCT-3¢. The latter primer consists of a fusion of the sequences from both adjacent sides of the 21 bp cDNA encoding the NLS and t herefore the r esulting PCR f ragment lacks the NLS sequence. This fragment was cloned into pBluescript I I S K and the sequence was verified. The 2 00 bp fragment containing the m utated site was e xcised by Msc I/ BbsI digestion and used to replace the corresponding wild- type s equence of pBS-NARG2-Pst930, which contains the 930 bp PstI fragment of p CMV6-NARG2–3H. Subse- quently, the 519 bp mutant EcoNI/SanDI fragment was excised f rom t he resulting plasmid and used t o r eplace the corresponding fragment of pCS2+M T-NARG2 to generate pCS2+MT-NARG2DNLS. In vitro translation of NARG2 was performed in the pre- sence o f Redivue [ 35 S]methionine (Amersham, Pitscataway, NJ) using a TNT T7 Quick coupled transcription/translation system (Promega) and pT7-NARG2. Post-translational modifications take place in this r eticulocyte-lysate based system, including acetylation [17], isoprenylation [18], myristolation [19,20], O-linked glycosylation [21] and phos- phorylation [22]. The translation product w as resolved on by 7% SDS/PAGE and analyzed by autoradiography. Analysis of genomic and cDNA sequences For genomic analysis we used the mouse and human genomic database at NCBI/NIH (http://www.ncb i.nlm. nih.gov). EST and open reading frame a nalyses, as well as Saccharomyces ce revisiae and Drosophila melanogaster genomic analyses, were also carried out using the NCBI/ NIH website. T he Caenorhabditis e legans, Fugu rubripes (pufferfish) and zebrafish genomic databases are available at http://www.sanger.ac.uk, genome.jgi-psf.org, and zfBl- astA.tch.harvard.edu, respectively. Ot her p rograms u sed f or data analysis inc lude nucleotide a lignment, CLUSTALW (http://www.clustalw.genome.jp); amino acid align- ment, CLUSTALW and MULTIPLE ALIGN SHOW (http:// www.ualberta.ca/stothard/javascript/); protein sequence analysis, PREDICTPROTEIN (http://www.embl-heidelberg.de/ Predictprotein/); N LS analysis, PREDICTNLS (http://cubic. bioc.columbia.edu/predictnls/). Exon–intron boundar ies were determined by genomic DNA and cDNA sequence comparisons, coupled with the identification of conserved GT:AG nucleotides of intron splice sites. Synonymous and nonsynonymous substitution rates between the human and mouse NARG2 cDNAs, as well as insertions and deletions, were calculated based on the method of Nei and Gojobori [23] using the synonymous/ nonsynonymous analysis program ( SNAP ; http://www.hiv. lanl.gov/content/hiv-db/SNAP/WEBSNAP/SNAP.html). Proline usage in Mus musculus proteins was found at http:// bioinformatics.weizmann.ac.il/blocks/help/CODEHOP/ codon.html. Cellular localization of NARG2 Rat NRK fibrobla st cells (ATCC, CRL-6509) were main- tained in Dulbecco’s modified Eagle’s medium containing 5% (v/v) fetal bovine serum, and 3 · 10 5 cells were replated on a 35 mm dish 24 h before transfection. One microgram of pCS2 + MT-NARG2 DNA was transiently transfected into the cells using F uGENE6 transfection reagent (Roche, Florence, SC). The cells were fixed with 3.7% (v/v) formal- dehyde/NaCl/P i for 10 min and then blocked and permea- bilized in 0.1% (v/v) Triton X-100 and 10% (v/v) goat serum in NaCl/P i for 10 m in. For immunostaining, 9E10 mouse anti-(c-Myc) monoclonal ascites fluid ( Sigma, St Louis, MO) was u sed at a dilution of 1 : 1000, followed by Alexa 488- conjugated goat anti-(mouse I gG) I gG at a dilution of 1 : 1000 (Molecular Probes, Eugene, OR). Samples were examined using a Nikon Eclipse TE2000-S microscope, and acquired images were subject to analysis of average pixel intensity in t he cytoplasm and the nucleus using METAMORPH software (Universal Im aging Co rporation, Marlow, B uck- inghamshire, UK). For each cell (11 for w ild-type NARG2, and 19 for NARG2 lacking the NLS), a ratio of average pixel intensity i n t he cytoplas m d ivided by the average pixel intensity in the nucleus was calculated, and statistical comparison was carried out using a two-tailed t-test. The background average signal intensities were negligible com- pared to the signal in transfected cells (<1%), were not included in the calculations, and do not significantly impact upon the numbers reported. Our staining p rotocol did not detect endogenous c-Myc p rotein in untransfected cells. P19 cell culture and RNase protection Mouse P19 embryonic carcinoma cell culture was carried out as described previously [13,24]. Briefly, monodispersed P19 cells were seeded in bacterio logical grade c ulture dishes (Asahi Techno Glass Corp., Funabashi, Japan) at 1 · 10 5 cellsÆmL )1 in the presence of 1 l M retinoic a cid. These aggregate cultures were maintained for 4 days, trypsinized, and then replated on t issue culture dishes in the absence of retinoic acid. Two days after replating the medium was replaced with fresh medium containing 5 lgÆmL )1 cytosine arabinoside ( Sigma); c ultures were then maintained for up to another six days for a total of eight days after retinoic acid treatment. Total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA). RNase protection was carried out as described previously using a NARG2 antisense p robe corresponding to nucleotides 102–365 of AA472833 [11,25]. 4630 N. Sugiura et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Northern blot analysis A human poly(A) + RNA blot (OriGene HB-2010) con- taining RNA from 12 adult tissues, and a Fetal Multiple Human Tissue Blot ( Clontech 7756–1, Palo Alt o, C A), each using 2 lg poly(A) + RNA p er lane, w ere p rocessed in parallel. The intactness of the RNA samples and equival- ence in amounts from l ane to lane were verified by denaturing gel electrophoresis with ethidium bromide staining, and by Northern blot analysis for human b-actin controls performed by OriGene and Clontech. Hybridiza- tion was carried out with 1 · 10 6 cpmÆmL )1 of 32 P-labeled cDNA probe in ULTRAhyb buffer (Ambion, Austin, TX) at 4 2 °Cfor18 h.TheNARG2 probe used here corresponds to a 531 bp PstI fragment from the human NARG2 EST AL549015. The highest stringency wash was in 0.1· NaCl/ Cit and 0.1% SDS at 42 °C for 15 min. Results NARG2 cDNA and protein EST AA472833 was originally identified by cDNA micro- array a nalysis as one of a group of three independent genes that are expressed at higher than normal levels in the developing brain of NMDAR1 knockout mice [11]. We termed the corresponding gene mNARG2. To investigate the significance of m NARG2 further, we screened a mouse cDNA library and obtained two full-length clones, 2C and 3H. S equence analysis of c lone 3H indicated that this clone contains an open reading frame of 2964 bp encoding 988 amino acids with a predicted molecular mass of 109 880 (Fig. 1 A). The open reading frame of clone 2C lacks a 130 b p s equence that is present in 3H. In vitro translation performed using the 3H cDNA yielded a 150 kDa protein, as determined by SDS/PAGE and autoradiography. This migration is slower than p redicted, but may b e explained b y post-translational modification (see Experimental proce- dures) or the relatively high proline content of NARG2 (8.1% vs. the Mus musculus average of 6.0%). High proline levels in other proteins, for example in a s1 -casein B (8.6%), have been reported to slow their migration in SDS/PAGE, limiting the accuracy of this method for size determination of such proteins [26]. Genomic organization of NARG2 Analysis of EST, cDNA and genomic sequences available in public databases w as performed to investigate both t he origin of the d ifference i n sequence b etween mouse c lones 2C and 3H, a s well as to begin to e valuate the evolutionary significance of N ARG2. Human NARG2 (hNARG2) cDNA was identified in the GenBank database (AL832046, AK055752), a nd hNA RG2 shares 74% amino acid identity with mouse NARG2 (mNARG2) (Fig. 1A). However, neither mNARG2 nor hNA RG2 have significant similarity with other known genes. Only one gene encoding NARG2 was identified in both human and mouse: mNAR G2 is p resent on chromosom e 9D, a nd hNARG2 is present on chromosome 15q21.3. The NARG2 gene is present o n chromosomal regions that conserve human– mouse synteny. Three putative pseudogenes with significant homology to NARG2 were identified in human: two are tandem repeats on chromosome 4 that c orrespond to 2 kb of the 3¢ end of the cDNA, and a third is present o n chromosome 3 that corresponds to th e full-length cDNA. All three putative pseudogenes display high nucleotide homology to NARG2 cDNA, lack open reading frames long enough to encode NARG2 (< 2 50 amino acids), and lack intervening sequences that would correspond to introns. N o NARG2 ortholog was identified in lower organisms such as S. cerevisiae, C. elegans, D. m elanogas- ter, pufferfish and zebrafish. The human and mouse NAR G2 genes have highly conserved exon–intron structures including 16 exons, 11 of which are identical in size i n t he two species (Fig. 1B, Table 1 , Table 2). A comparison of cDNA and genomic sequences revealed that the 130 bp sequence present in mouse clone 3H bu t a bsent i n clone 2C corresponds to exon 12, and suggests that this exon can be eliminated by alternative splicing. In addition to the alte rnative splicing that yields the d ifferent mouse mRNAs described a bove, the human NARG2 gene generates an a lternatively spliced mRNA that lacks bot h exon 4 and e xon 5 (e.g. BE814990). Exon 10 is considerably larger than the other exons, and most of the d ifferences between the mouse a nd human gene products are localized to this exon (Fig. 1A). To quantify the differences between mouse and human NARG2, non- synonymous (amino acid altering), synonymous (silent), and insertion/deletion nucleotide d ifferences per c odon [23] were plotted along the a mino acid sequence (Fig. 2). Based on the slope, it i s clear that nonsynonymous substitutions have accumulated faster in exon 10 than i n the other coding exons. The ratio dn/ds, a measure of the relative pressure of evolutionary selection based on the rate of nonsynonymous substitutions (dn) divided by the rate of synonymous substitutions (ds) [23,27], was calculated for exon 10 and the rest of the coding region outside of exon 10. T he overall dn/ds ratio for NARG2 is < 1, indicating that, on balance, this protein i s under evolutionary pressure to r esist amino acid substitutions, a nd that it is likely to h ave a functional importance in multiple species. However, the dn/ds ratio is higher for exon 10 (0.40) than f or the other exons (0.13). As the ds value is similar for exon 10 (0.62) and the other e xons (0.56), this difference in ratio is predominantly attributable to dn, i.e. nonsynonymous substitutions (exon 10, 0.25; other exons, 0.07), confirming the observation that nonsynonymous substitutions are more frequent in exon 10, a nd raising the possibility that parts of e xon 10 are under positive selection (see Discussion). Further examination of exon 10 revealed the presence of a number of repeats of (S/T)PXX, a putative DNA-binding domain that is presen t in many transcription factors including Kruppel, Hunchback and Antennapedia [14], as well as other DNA-binding proteins [28,29]. There are 11 repeatsof(S/T)PXXinmNARG2(seveninexon10),and six in hNARG2 (four in exon 10). Cellular localization of NARG2 hNARG2 and mNARG2 both contain a possible nuclear localization signal (NLS) that consists of a c anonical stretch of basic amino acids near the C-terminus of the protein (KKRKKIRR, amino acids 764–771 of the mouse sequence Ó FEBS 2004 NARG2, a novel nuclear (S/T)PXX protein (Eur. J. Biochem. 271) 4631 in Fig. 1A; [30]). A search of the PredictNLS database did not reveal any other known NLS (http://cubic. bioc.columbia.edu/predictnls [31]). To determine whether NARG2 l ocalizes to the nucleus, and the role o f the putative NLS, we examined NRK fibroblast cells transfected with Myc epitope-tagged NARG2 (pCS2+MT-NARG2), and pCS2+MT-NARG2DNLS, in which the NLS has been deleted by mutagenesis. Results were visualized by immu- nofluorescent staining u sing a p rimary antibody against c-Myc, followed by a n Alexa 488-conjugated secondary antibody. Wild-type NARG2 localized almost exclusively to the nucleus in mostcells (Fig. 3B). A few cells expressed NARG2 at abnormally high levels, in which case it was localized to the cytoplasm a nd excluded f rom the nucleus (< 20%; data not shown). This latter result m ay be an artefact resulting from excessive and pathological accumulation of the protein, a s moderate expression of a short NLS–MT fusion protein results in nuclear localization w hile very high levels of expression result in exclusion from the nucleus (data not shown). Alternatively, it is possible that under biological conditions NARG2 localization may be regulated by its concentration. Finally, NARG2 that lacks the NLS displays nuclear localization similar to that of wild-type, although significantly more than the typical wild-type levels of cytoplasmic NARG2 are observed (Fig. 3C; ratios of Fig. 1. Human a nd mouse N ARG2. (A) A lignmen t of dedu ced amino ac id sequences of h NARG2 and mN ARG2 . Identities and conserved amino acid substitutions are in black and grey sh aded back grounds, resp ectively. T he most divergent regio n of the protein, e ncoded by exon 1 0, is indicated between arrowheads. The cano nical NLS is und erlined. (S/T)PX X repeats are ind icated by asterisks (p resent in both mN ARG2 and hNA RG2), filled circles (mouse-specific), and o pen circles (human-specific). (B) Exon–intron structure of hNARG2 and mNARG2. Exon n umbers and sizes ( bp) are indicated. Coding regions are indicated by fi lled boxes (above) or vertical bars (below). 4632 N. Sugiura et al.(Eur. J. Biochem. 271) Ó FEBS 2004 cytoplasmic to nucle ar signals w ere 0.37 ± 0.0 7 and 0.49 ± 0.10 for wild-type NARG2 and for N ARG2 lacking the NLS, respectively, P < 0.005). The results indicate that NARG2 is usually locali zed to t he nucleus, t hat the canonical NLS p lays a s upporting role in nuclear localization, and that there is p robably an N LS in NARG2 that is not currently represented in the PredictNLS database [31]. Expression of NARG2 in human tissues Previous studies in mice demonstrated that, in the brain, NARG2 is expressed at the highest levels in neonates, and is subsequently down-regulated [11]. In the adult mouse, NARG2 is expressed at v ery low levels in all t issues examined, with the most appreciable levels of expression observed in the kidney, testes, liver and brain [11]. N orthern blot analyses demonstrate a similar e xpression pattern for NARG2 in humans (Fig. 4). S ignificant NARG2 expression was d etected in fetal k idney, liver, lung and brain (Fig. 4A) but little or no expression was observed in adult kidney, liver, brain or a number of other tissues (Fig. 4B). A small amount of expression was detected in adult lung, and, as previously reported in mouse [11], significant e xpression w as present in adult testes (Fig. 4B). Taken together, these Table 1. Exon–intron boundary sequences of the mouse NARG2 . Exon sequences are shown in uppercase, introns in lowercase. Exons of the same size in human and mo use are underlined. No. Exon size (bp) Boundary sequences Intron size (kb) 5¢ boundary 3¢ boundary 1>79 CCTGAG gt gggc … cttc ag TTAACT 0.89 2 142 GAACTG gt gagt … ccac ag GGATGT 1.4 3 105 a TAACAG gt aata … ttcc ag ACGTAT 6.5 4 250 TATCTG gt atgt … taac ag GATATG 0.89 5 120 ACGGAG gt aaaa … tccc ag CAACTC 1.7 6 138 GTTTTG gt aaga … attt ag GGCAGT 0.52 7 117 CACTGT gt aagt … tttc ag GATATT 0.42 8 160 TGGAAG gt ttgt … ttta ag GTTCCA 0.74 9 182 GTGGAC gt atgt … tcca ag ATGCCC 3.0 10 1024 AAGGAC gt aagt … ttca ag GATTGC 0.57 11 176 AGAAGA gt aagt … ttgc ag CAATTT 4.5 12 130 ATGTTG gt gagt … tttt ag GGCATA 3.9 13 85 ACTCAA gt aagg … taat ag GATTTC 1.1 14 51 ATGTAG gt aagt … atgc ag CTTGCA 1.4 15 259 GCAAAG gt aaga … tcta ag GTTGGT 3.5 16 >600 Table 2. Exon–intron boundary sequences of th e hum an NARG2 . Exon sequen ces a re shown in u pperca se, introns in l owercase. Exons of the same size in human and mo use are underlined. No. Exon size (bp) Boundary sequences Intron size (kb)5¢ boundary 3¢ boundary 1 >120 CATGAG gt gggc … actc ag CTGAGT 0.93 2 133 GAATTG gt gagt … tcac ag GGATAT 1.8 3 105 CAACAG gt aata … ttcc ag ACGTAT 7.7 4 262 a TACTTG gt atgt … aatc ag GATATG 1.3 5 120 a ACAGAG gt aaaa … tctc ag AAAATT 9.9 6 138 GCATTG gt aagg … attt ag GGCAGT 1.2 7 117 CATTAT gt aagt … tttc ag GATATT 0.17 8 160 TTGCAG gt ttgt … ttta ag GTTCAA 1.2 9 182 ATGGAC gt atgt … cata ag ATGTCC 3.8 10 994 AAGGAC gt aagt … ttaa ag GATTGC 0.70 11 176 AGAAGA gt aagt … ttgc ag CAATTT 5.4 12 130 ATGTTG gt aagt … tttt ag GGCATA 6.3 13 85 ACTCAA gt aaga … tatc ag GATTTC 4.2 14 51 AAGTAG gt aagt … attt ag CTTGCA 3.2 15 259 CAAAAG gt aaga … tctt ag ATTGGT 4.7 16 >640 a Some ESTs (e.g. accession AL549015) lack part of exon 4; a few ESTs lack both exon 4 and exon 5 (e.g. accession BE814990). Ó FEBS 2004 NARG2, a novel nuclear (S/T)PXX protein (Eur. J. Biochem. 271) 4633 results suggest a trend of relatively abundant NARG2 expression during development followed by low expression in the adult that is conserved among mammals. Down-regulation of NARG2 in P19 cells undergoing neuronal differentiation Mouse P19 embryonic carcinoma cells are multipotential cells that can be induced to exit the cell c ycle and attain a neuron-like phenotype in vitro [24]. Following treatment with retinoic acid u nder s pecific culture c onditions, P19 cells start to express neuronal markers including glutamic acid decarboxylase, neural cell adhesion molecule, NMDA receptors and metabotropic glutamate receptors [13,32– 35]. We extracted RNA from P19 cells at various stages of differentiation a nd used RNase protection t o determine whether, as is the case in vivo (Fig. 4; [ 11]), NARG2 is down- regulated during cellular differentiation in vitro. NARG2 is expressed at the highest levels in P19 cells before the addition of retinoic acid, and is progressively down-regulated as the cells differentiate and acquire a neuron-like phenotype (Fig. 5). T his p attern of expression is opposite t o that found for neuronal markers. For example, RNase protection analyses performed on aliquots of the P19 R NA samples used here demonstrated strong up-regulation of NMDAR1 following treatment with retinoic acid [13]. Increased NMDAR1 expression with NARG2 down-regulation in differentiating P19 cells concurs with the pattern of regulation of thes e genes during neuronal Fig. 2. Comparison of mouse and human NARG2 coding sequences reveals a high rate of nonsynonymous nucleotide substitutions as well as insertions and deletions in exon 10. Cumulative indexes of synony- mous nucleotide subs titutions and non syn onymous su bstitutions p er codon, and insertions and deletions are plotted vs. the NARG2 amino acid sequence, starting at the N-terminus of the protein. The stretch of amino acids derived from exon 10 is indicated between two broken vertical lines. Th e rate of synonymous sub stit utions remains relatively c onstant in all exons , i ncluding in exon 10. Figure gener- ated by SNAP . Fig. 3. Localization o f NARG2. (A) Autoradiogram of in vitro trans- lated mouse NARG2 (lane 2), and vector without insert (negative control) (lane 1 ). (B,C) Cellular localization of myc-tagged wild-type (B, WT) and mutant (C, DNLS) mNARG2 in rat NRK fibroblast cells. Mutant mNARG2 lacks the putative nu clear localizatio n sequence. Cells were tra nsie ntly trans fected with N ARG2 con structs and the proteins were visualized by immunoflu orescence using a primary antibody (anti-Myc) followed by an Alexa 488-conjugated secondary antibody. Fig. 4. Autoradiograms of Northern blot analyses o f NARG2 expres- sion in human fetal and adult tissues. (A) Fetal tissues, by lane: 1, 19–23 week kidney; 2 , 18–24 week l iver; 3, 22–23 week lung; 4, 19–22 week brain. (B) Adult tissues, by lane: 1, brain; 2, colon; 3, heart; 4,kidney;5,liver;6,lung;7,muscle;8,placenta;9,smallintestine;10, spleen; 11, stomach; 12, testes. 4634 N. Sugiura et al.(Eur. J. Biochem. 271) Ó FEBS 2004 development in viv o, a nd is consistent with NMDA receptor function playing a role in the down-regulation of NARG2 [11]. T he finding that some NARG 2 expression re mains after 12 days of differentiation is probably due to the presence of significant numbers of cells that do not differentiate upon treatment with retinoic acid [24]. These results provide evidence that, a s is t he case in vivo [11], NARG2 expression is inversely related to the degree of differentiation in cell culture. Discussion In a p revious study we identified NARG2 as one of a group of three genes, NARG1 (now referred to as mNAT1), NARG2 and NARG3, which are expressed at higher than normal levels in the brain of NMDAR1 knockout mice [11]. We identified these genes by cDNA microarray analysis, and all three w ere previously uncharacterized in vertebrates. Moreover, they share regulatory properties including high levels of expression in the neonatal b rain, d ramatic down- regulation during early postnatal development, and high levels of expression in proliferating c ells. We now report that NARG2 is a novel gene that encodes a nuclear protein that is conserved in mammals, but appears t o be absent in lower organisms. NARG2 contains repeats of ( S/T)PXX, a motif present in many t ranscription factors as well as other regulatory proteins that bind to DNA such as histones and RNA polymerase II [14,2 8,29]. The classic monopartite N LS present n ear the C-ter- minus of NARG2 appears to cooperate with other regions of the protein for nuclear localization. However, sequence analyses did not reveal evidence of a second typical NLS. In this context, functional cloning has demonstrated a higher frequency of atypical amino acid sequences that target proteins to the nucleus than was previously appreciated [30]. To evaluate fully the signifi- cance of t he monopartite N LS, i t will be necessary to identify other amino acids in NARG2 that participate in nuclear localization. Examples of proteins that have two characterized nuclear targeting signals that contribute to nuclear localization i n different ways and to different degrees include E1a [36], hnRNP K [37], an d USF2 [38]. These proteins all contain at least one atypical NLS, as appears to be the case for NARG2. Although NARG2 shares regulatory features with NARG1/mNAT1 and NARG3, these three genes do not share sequen ce homology and probably have very different functions. For example, while NARG2 is a nuclear protein, NARG1/mNAT1 encodes a critical subunit of an N-terminal acetyltransferase that is localized to the cyto- plasm [13]. These findings illustrate that, in contrast to cDNA screens based on sequence homology, cDNA microarray screens are driven by similarities in the regula- tion of gene expression that are not necessarily reflected by similarities in structure or function. Nevertheless, groups of coregulated genes may p lay diverse roles i n determining a specific phenotype. It will be interesting to d etermine whether, in the absence of NMDAR1, t he increased le vels of NARG1, NARG2 and NARG3 each contribute in different ways to, for example, maintaining the cell in an undifferentiated state. NARG2 as a whole is well-conserved between human and m ouse, with 74% overall identity, suggesting t hat this protein has functional significance. Of particular interest is exon 10, the largest and most divergent of the 16 exons. When the amino acids from exon 10 a re excluded the identity between mouse a nd human NARG2 rises to 86%, indicating that mos t of NARG2 may already be fixed for function across mammals. However, the dn/ds ratio for exon 10 (0.40) is higher than that for the other e xons, which have a dn/ds value (0.13) that is similar to the average of that for m ouse–human 1 : 1 orthologs (0.115; [39]). In other words, a dn/ds ratio of about 0.115 is expected for mouse– human orthologs when functional domains are conserved and nones sential domains experience random drift in their amino acid c omposition. The dn/ds ratio o f 0 .40 f or exon 10 indicates an a ccumulation of nonsynonymous changes faster than in the rest of the coding sequence, and faster than would b e expected by random drift. This suggests that nonsynonymous su bstitutions are subject to positive selec- tion in exon 10. Although dn/ds for exon 10 is not high enough to be formally defined a s b eing subject to positive selection ( dn/ds > 1; [39]), w e s uggest that a subset of amino acids encoded by exon 10 may be under positive selection for substitution s as a result of a need for diversity in this domain, possibly for a species-specific function. The contribution to dn/ds by positively selected sites in exon 10 may be offset by other amino acids in this exon that are not under positive selection, i.e. b y amino acids that are randomly drifting or being actively conserved [40]. A classic Fig. 5. Down-regulation of NARG2 during n eu rona l differentiation of mouse P19 embryonic carcinoma cells. Neuronal differe ntiation o f P19 cells was induced with 1 l M retinoic acid. T he cells were h arvested at the indicated times and total RNA was extracted. Single-stranded antisense 32 P-labeled riboprobe complementary to mNARG2 was synthesized, isolated , a nd hybridiz ed with 5 lg of the in dicated t otal RNA sample, digested with RNases, and fractionated by gel electro- phoresis. An auto radiogram of th e g el is sho wn, with the sizes (bp) of the undigested probe (arrow) an d the protected species (arrowhead) indicated. As a control, 2 lg a liquots of the samples used for RNase protection were run in parallel o n an e th idium bromide-stained denaturing agarose gel. That e ach lane c ontains similar a mounts of good quality total RNA is indicated by the r elative signals and i ntact 28S and 18S ribosomal bands (inset). The t ¼ 0 sample was not treated with retinoic acid (–RA). For additional details see Experimental procedures. Probe , und igested p robe; t RNA, negative c on trol; d, d ay. Ó FEBS 2004 NARG2, a novel nuclear (S/T)PXX protein (Eur. J. Biochem. 271) 4635 example of positive selection is the class I major histocom- patibility family of proteins, which have specific domains that require diversity for effective immune function [27]. Olfactory receptor molecules are a lso und er positive selec- tion [41]. In addition to the relatively high value of dn/ds for exon 10, it contains most of the ( S/T)PXX repeats found in both mNARG2 (seven of 11) and hNARG2 (four of six). The prevalence of this motif in NARG2 lends support to the conclusion that it is a nuclear protein, and raises the possibility that NARG2 is involved in regulating gene expression [14]. Future studies will address the specific role of the (S/T)PXX motifs, and whether NARG2 is a regulatory protein that binds to DNA. Numerous eukaryotic proteins are conserved from single cell organisms to higher mammals. However, about 22% of vertebrate proteins do not have obvious homologues in lower organisms [39,42]. Genes that are involved in immune and n ervous system function are particularly enriched in this group [ 42]. Here we propose that one such gene, NARG2, plays a role in development. Examples of vertebrate genes not represented i n lower organ isms that r egulate c ell g rowth and differentiation include dkk and krm. These genes encode proteins that functionally cooperate to block Wnt/ b-c atenin signaling [43]. From an e volutionary perspective it may be significant that dkk and krm are not essential for signaling, but rather modulate the transmission of signals through the Wnt/b-cateninpathwaybyregulatingproteins that are critical for signaling. We hypothesize that NA RG2, similar to the proteins encoded by dkk and krm,interacts with an d modulates the function o f e volutionarily conserved proteins. The findings reported here provide the primary characterization required for studies that will test this hypothesis and determine the function of this novel nuclear protein. Acknowledgements This work was supported by a Louisiana Board of Regents Research Competitiveness Subprogram Award to R.A.C. We thank R ajan Patel for expert technical assistance, and Dr Oliver Wessely for his careful reading o f the manuscript and helpfu l s uggestions. We are also grateful to Dr David Turner and S tacy DeRuiter for vectors, cells, protocols and a dvice, and to Dr Jeffrey Loeb, Dr Qunfang L i and Dr Kunio Takishima for assistance with preliminary experiments. References 1. Ikonomidou, C., Bosch, F., Miksa, M., Bittigau, P., Vockler, J., Dikranian, K., Tenkova, T.I., Stefovska, V., Turs ki, L. & Olney, J.W. (1 999) Blockade of NMDA receptors and apoptotic neuro- degeneration in the developing brain. Science 283, 70–74. 2. Adams, S.M., de Rivero Vaccari, J.C. & Corriveau, R.A. (2004) Pronounced cell death in the absenc e of NMDA re ceptors in the developing somatosensory thalamus. J. Neurosci. 24, 9441–9450. 3. Komuro, H. & Rakic, P. (1993) Modulation of neuronal m igra- tion by NMDA receptors. Science 260, 95–97. 4. Gould,E.,Cameron,H.A.&McEwen,B.S.(1994)Blockadeof NMDA receptors increases ce ll death a nd birth i n the d eveloping rat dentate gyrus. J. Comp . Neurol. 340, 551–565. 5. Goodman, C.S. & Shatz, C.J. (1993) Developmental mechanisms that generate precise p atterns of n euronal c onnectivity. Cell 72 (Suppl.), 77–98. 6. Constantine-Paton, M. & Cline, H.T. (1998) LTP and activity- dependent synaptogenesis: the more alik e they are, the mo re dif- ferent they become. Curr. Opin. Neurobiol. 8, 139–148. 7. Corriveau, R.A. (1999) Electrical activity and gene expression in the development of n eural circuits. J. Neurobiol. 41, 148–157. 8. Krumlauf, R., Marshall, H., Studer, M., Nonchev, S., Sham, M.H. & Lumsden, A. ( 1993) Hox homeobox genes and regionalization of t he nervous system. J. Ne urobiol. 24, 1328–1340. 9. Acampora, D. & Simeone, A. (1999) The TINS lecture: Under- standing the roles of Otx1 and O tx2 in the c ontrol of brain m or- phogenesis. Trends Neur osc i. 22, 116–122. 10. Puelles, L. & Rubenstein, J.L. (2003) Forebrain gene expression domains and th e evolving pro someric model. Tre nds Neurosci. 26, 469–476. 11. Sugiura, N., P atel, R.G. & Corrive au, R.A. (2001) NMD A receptors regulate a group of transiently expressed genes in the developing brain. J. Biol. Chem. 276, 14257–14263. 12. Mullen, J.R., Kayne, P .S., Moerschell, R.P., Tsunasawa, S., Gri- bskov, M., Colavito-Shepanski, M., Grunstein , M., Sherman, F. & Sternglanz, R. (1989) Identification and characterization of genes a nd mutants f or an N -terminal acetyltransferase f rom yeast. EMBO J. 8, 2067–2075. 13. Sugiura, N., Adams, S.M. & Corriveau, R.A. (2003) An e volu- tionarily conserved N-terminal acetyltransferase associated with neuronal development. J. Biol. Chem. 278, 40113–40120. 14. Suzuki, M. ( 1989) SPXX, a frequent sequence motif in gene reg- ulatory proteins. J. M ol. Biol. 207, 61–84. 15. Turner, D.L. & Weintraub, H. (1994) E xpression of achaete-scute homolog3inXenopus embryo s converts ectodermal cells to a neural fate. Genes Dev. 8, 1434–1447. 16. Rupp, R .A., Snider, L. & Weintraub, H. (1994) Xenopus embryos regulate the nuclear localization of XMyoD. Genes Dev. 8, 1311– 1323. 17. Palmiter, R.D., Gagnon, J. & Walsh, K.A. (1978) Ovalbumin: a secreted protein without a t ransient hydrophobic l eader s eq uence. Proc.NatlAcad.Sci.USA75, 94–98. 18. Sanford, J., Codina, J. & Birnbaumer, L. (1991) c-Subunits of G p rotei ns, but no t their a-orb-subunits, are polyisoprenylat- ed. Studies on post-translational modifications usin g in vitro translation with rabbit reticulocyte lysates. J. Biol. Chem. 266, 9570–9579. 19. Heuckero th, R.O., Towler, D.A., Adams, S.P., Glaser, L. & Gordon, J.I. (1988) 11-(Ethylthio) undecanoic acid. A myristic acid analogue of altered hydrophobicity which is fun ctional for peptide N-myristoylation with wheat germ and yeast acyl- transferase. J. Biol. Chem. 263, 2 127–2133. 20. Deichaite, I., Casson, L.P., Ling, H.P. & Resh, M.D. (1988) In vitro synthesis of pp60 v–src : myristylation in a cell-free system. Mol. Cell. Biol. 8, 4 295–4301. 21. Starr, C.M. & Hanover, J.A. (1990) Glycosylation of nuclear p ore protein p62. Reticulocyte lysa te catalyzes O-linked N-acetylglucosam ine ad dition in vitro. J. Biol. Chem. 265, 6868– 6873. 22. Joshi, B., Cai, A.L., Keiper, B.D., Minich, W .B., Mendez, R., Beach, C.M., Step inski, J., Stolarski, R., Darzynkiewicz , E. & Rhoads, R.E. (1995) Phosphorylation of eukaryotic protein synthesis initiation f actor 4E at Ser-209. J. Biol. Chem. 270, 14597– 14603. 23. Nei, M. & Gojobori, T. (1986) S imp le meth ods f or estim ating th e numbers of synonymous and nonsynonym ous nucleot ide sub- stitutions. Mol. Biol. Evol. 3, 418–426. 24. Rudnicki, M.A. & McBurney, M.W. (1987) Cell culture methods and in duction of d ifferentiation of embryo nal carcinoma cell lines. In Teratocarcinomas and E mbryonic Stem Cells: a Practical Appr oach. (Robertson, E .J., ed.), pp. 19–49. IRL Press, Oxford. 4636 N. Sugiura et al.(Eur. J. Biochem. 271) Ó FEBS 2004 25. Corriveau, R .A. & Berg, D .K. (1993) Coexpression of multiple acetylcholine recep tor genes in neuro ns: quantificat ion of tran - scripts during d evelopment. J. Neurosci. 13, 2662–2671. 26. Noelken, M.E., Wisdom, B.J. Jr & Hudson, B.G. (1981) Estima- tion of the size of collagenous polypeptides by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Anal. Biochem. 110, 131–136. 27. Hughes, A .L. & Nei, M. (1988) Pattern of su bstitution at major histocompatibility c omplex class I loci reveals o verdominant selection. Nature 335, 167–170. 28. Suzuki, M. (1990) The heptad repeat in the largest subunit of RNA p olymerase II b inds by intercalating into D NA. Nature 34 4, 562–565. 29. Lindsey, G.G. & Thompson, P. (1992) S (T) PXX mo tifs promote the interaction be tween the extended N-terminal t ails of histone H2B with ÔlinkerÕ DNA. J. Biol. Chem. 267, 14622–14628. 30. Christophe, D., Christophe-Hobertus, C. & Pichon, B. (2000) Nuclear targeting of proteins, h ow many different sig nals? Cell. Signal. 12, 337–341. 31. Cokol, M., Nair, R. & Rost, B. (2000) Finding nuclear localization signals. EMBO Report 1, 4 11–415. 32. Bain, G ., Ram kumar, T.P., Cheng, J.M. & G ottlieb, D.I. (1993) Expression of the genes coding for g lutam ic a cid decarboxylase in pluripotent cell lines. BrainRes.Mol.BrainRes.17, 23–30. 33. Husmann, M., Gorgen, I., Weisgerber, C. & Bitter-Suermann, D. (1989) U p-regulation of e mbryonic NCAM in an EC cell line by retinoic acid. Dev. Biol. 136, 194–200. 34. Ray, W.J. & Gottlieb, D.I. (1993) Expression of ionotropic glu- tamate receptor genes by P19 em bryonal carcinoma cells. Biochem. Biophys. Res. Commun. 197, 1475–1482. 35. MacPherson, P.A., Jones, S., Pawson, P.A., Marshall, K.C. & McBurney, M.W. (1997) P19 c ells diffe rentiate into g lutamatergic and glutamate-respo nsive n eurons in vitro. Neuroscience 80 , 487–499. 36. Standiford, D.M. & Richter, J.D. (1992) Analysis of a devel- opmentally regulate d nucle ar localiz ation signal in Xe nopus. J. Cell Biol. 118, 991–1002. 37. Michael, W.M., Eder, P.S. & Dreyfuss, G. (1997) The K nuclear shuttling d omain: a novel signal for nuclear import an d nuclear export in the hnRNP K prote in. EMBO J. 16, 3587–3598. 38. Luo, X. & Sawadogo, M. (1996) Functional domains of the transcription factor USF2: atypical nuclear localization signals and c ontext-dependent transcriptional activation domains. Mol. Cell. Biol. 16 , 1367–1375. 39. Mouse Genome Sequencing Consortium ( 2002) In itial s equencing and comparative analysis of the mouse genome. Nature 420,520– 562. 40. Sharp, P.M. (1997) In search of molecular darwinism. Nature 385, 111–112. 41. Gilad, Y., S egre, D., Skorecki, K.,Nachman,M.W.,Lancet,D.& Sharon, D. (2000) D ichotomy o f single-nu cleotide po lymorph ism haplotypes in olfactory receptor genes and pseudogenes. Nat. Genet. 26, 221–224. 42. International Human Genome Sequencing Consortium (2001) Initial s equencing and analysis of the human genome. Nature 409, 860–921. 43. Mao, B., Wu, W., Davidson, G., Marhold, J., Li, M., Mechler, B.M.,Delius,H.,Hoppe,D.,Stannek,C.W.,Glinka,A.&Niehrs, C. (2002) Kremen proteins are Dickkopf receptors that regulate Wnt/b-catenin signaling. Nature 41 7 , 664–667. Ó FEBS 2004 NARG2, a novel nuclear (S/T)PXX protein (Eur. J. Biochem. 271) 4637 . gt atgt … aatc ag GATATG 1.3 5 120 a ACAGAG gt aaaa … tctc ag AAAATT 9.9 6 138 GCATTG gt aagg … attt ag GGCAGT 1.2 7 117 CATTAT gt aagt … tttc ag GATATT. CAATTT 5.4 12 130 ATGTTG gt aagt … tttt ag GGCATA 6.3 13 85 ACTCAA gt aaga … tatc ag GATTTC 4.2 14 51 AAGTAG gt aagt … attt ag CTTGCA 3.2 15 259 CAAAAG