Identification, chromosomal arrangements and expression analyses of the evolutionarily conserved prmt1 gene in chicken in comparison with its vertebrate paralogue prmt8

Nine protein arginine methyltransferases (PRMTs) are conserved in mammals and fish. Among these, PRMT1 is the major type I PRMT for asymmetric dimethylarginine (ADMA) formation and is the most conserved and widely distributed one. Two chicken prmt1 splicing variants were assembled and confirmed by RT-PCR experiments. However, only two scaffolds containing single separate prmt1 exon with high GC contents are present in the current chicken genome assembly. Besides, prmt1 exons are scattered in separate small scaffolds in most avian species. Complete prmt1 gene has only been predicted from two falcon species with few neighboring genes. Crocodilians are considered close to the common ancestor shared by crocodilians and birds. The gene arrangements around prmt1 in American alligator are different from that in birds but are largely conserved in human. Orthologues of genes in a large segment of human chromosomal 19 around PRMT1 are missing or not assigned to the current chicken chromosomes. In comparison, prmt8, the prmt1 paralogue, is on chicken chromosome 1 with the gene arrangements downstream of prmt8 highly conserved in birds, crocodilians, and human. However, the ones upstream vary greatly in birds. Biochemically, we found that though prmt1 transcripts were detected, limited or none PRMT1 protein was present in chicken tissues. Moreover, a much higher level of PRMT8 protein was detected in chicken brain than in mouse brain. While PRMT8 is brain specific in other vertebrate species studied, low level of PRMT8 was present in chicken but not mouse liver and muscle. We also showed that the ADMA level in chicken was similar to that in mouse. This study provides the critical information of chicken PRMT1 and PRMT8 for future analyses of the function of protein arginine methyltransferases in birds.

Introduction slow [23], crocodilians should be close to the common ancestor shared by crocodilians and birds. To trace prmt1, we compared the chromosomal neighbors of prmt1 and prmt8 in human, avian species including chicken, falcons and ground tit, as well as reptiles including lizard and crocodilians. We further analyzed the syntenic gene arrangement of prmt1 and prmt8 in typical vertebrate branches for the evolution of gene arrangements in vertebrates.
In addition, in this study, we also investigated the level of RNA and protein products of both prmt1 and prmt8, the level of ADMA-containing proteins, as well as the distribution of the PRMTs in subcellular fractions of chicken brain extracts. Basic biochemical analyses can provide more information to evaluate the chicken PRMTs.

Materials and methods
Establish the syntenic arrangement of the genes encoding PRMT1 and PRMT8 in representative vertebrate species The chicken (Gallus gallus) prmt1 coding sequence was assembled manually from available transcriptome databases. The sequence was used to BLAST avian databases for orthologues. Syntenic gene arrangement around prmt1 and prmt8 in human (Homo sapiens), American alligator (Alligator mississippiensis), green anole (Anolis carolinensis), and avian species including chicken (Gallus gallus), saker falcon (Falco cherrug), peregrine falcon (Falco peregrinus) and Tibetan ground tit (Pseudopodoces humilis) were retrieved from NCBI.

Animal tissue extract preparation
Mouse tissues were from mice sacrificed for a study of female hormones on neuromuscular transmission with the permission from the Institutional Animal Care and Use Committee, Chung Shan Medical University. Chicken tissues were freshly prepared from chicken slaughtered in a local market. Fresh tissues were stored at −80˚C. All tissues were weighed, washed two times with phosphate buffered saline (PBS), cut into small slices, and then resuspended into 5× (vol/wt) TRIzol reagent (Molecular Research Center, Inc.) for RNA preparation or in 3× (vol/wt) lysis buffer [50 mM Tris-HCl, pH7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10 mM NaF, 1x complete protease inhibitor (Roche)] for protein extract preparation. The protein concentration in each fraction was determined using the BCA reagent (Pierce) with bovine serum albumin as the standard protein.

Isolation of mouse and chicken tissue histones
Nuclei were collected by centrifugation at 17, 530 xg at 4˚C for 20 min. Histones were acid extracted by shaking in 0.2 M sulfuric acid for 1 h at 4˚C. After centrifugation, histones were precipitated with ethanol at −20˚C overnight, washed once with ethanol, and resuspended in distilled water.

Fractionation of mouse and chicken brain extract
Brain washed with PBS was cut into small slices, re-suspended in 3× (vol/wt) homogenization buffer (0.32 M sucrose, 20 mM Tris, pH 7.3, 5 mM MgCl 2 , 1 mM PMSF, 1 mg/ml aprotinin, 1 μg/ml pepstatin, 1 μg/ml leupeptin) and homogenized with a Teflon pestle with about 10 strokes. The fractionation scheme was illustrated in S1 Fig following the procedures of Hung et al [24]. Briefly, the homogenate (S0) was centrifuged at 800 ×g for 10 min and the pellet was P1. The supernatant S1 was further centrifuged at 9,200 ×g for 15 min. The supernatant and pellet fractions were designated as S2 and P2. Ultracentrifugation of the S2 fraction at 130,000 ×g for 1 h resulted in the supernatant S3 (cytosolic) and pellet P3 (membrane) fraction. Part of the P3 was re-suspended in lysis buffer with 0.5 M KCl and subjected to ultracentrifuge again. The supernatant and the pellet fractions were S4 (high-salt-stripped soluble) and P4 (high-saltstripped membrane pellet), respectively.

Assembly of chicken prmt1 transcripts from the database
To investigate PRMT1 in birds, we first conducted tBLASTn with human PRMT1 sequence against the chicken genome and no targets could be identified. Nevertheless, prmt1 genes are present in peregrine falcon (Falco peregrinus) and saker falcon (Falco cherrug), whose genomes are assembled with more than 100-fold coverage and have about 16,200 genes predicted [25]. The predicted proteins share high sequence identity with human PRMT1. The avian genome is of a high evolutionary stasis at the level of nucleotide sequences, gene synteny and chromosome structures [18]. It is thus likely that the prmt1 orthologue is present in chicken but missed in the current chicken genome assembly. We then used the human or falcon prmt1 as the query to conduct nucleotide BLAST against the chicken or avian NCBI nr database. Even though there were few target hits, among them were two chicken cDNA entries (CR524297 and BX930603). We further assembled full-length chicken cDNA sequences with available EST and TSA sequences to fill the ends. The predicted sequences are consistent with two splicing variants with or without an alternative exon at the 5' region between the 1 st and 2 nd constitutive exons (Fig 1A). Six ESTs support the variant without the alternative exon and two ESTs support the other variant with the alternative exon. Due to alternative splicing of the 5' exons, there are at least seven different human PRMT1 transcripts reported, with v1 and v2 the most abundant ones [26]. The assembled chicken variant 1 and variant 2 with the accession number BK010274 and BK010275 are similar to the human v1 and v2 variants, with the first chicken exon corresponding to human e1d, the alternative exon corresponding to human e2 and the 2 nd constitutive exon corresponding to exon 4. In comparison, in zebrafish only one prmt1 transcript had been identified [14]. The illustration of the gene structure that is also conserved in the paralogous prmt8 gene is shown in Fig 1B. We then validated the expression of chicken prmt1 by RT-PCR analyses. We prepared RNA from chicken tissues and designed primers. The first primer set to amplify the constitutive exons (prmt1-1) could detect signals in brain, liver, and muscle. Besides, with the primer set encompassing the alternative exon (prmt1-2), two RT-PCR products corresponding to the two splicing variants could be detected (Fig 1C). The signals of the variant without the alternative exon appeared to be stronger, consistent with more supporting ESTs. We also showed the expression of mouse prmt1 transcripts in three tissues.
We then designed primers according to the assembled sequences to amplify the whole coding region of prmt1 and confirmed the coding sequences of v1 (accession number MF630622). A T/C single nucleotide polymorphism is present at nucleotide from the ATG start codon. We also used the chicken prmt1 cDNA sequence to conduct nucleotide BLAST of the chicken genome. In chicken, two hits corresponding to exon 1 (chrUn_Scaffold23446) and exon 8 (chrUn_Scaffold23978) were identified from the newly released Gallus_gallus-5.0. They are separate small scaffolds (2017 bp and 1863 bp) of unknown chromosomes with high GC content (0.7297 and 0.6334) and long GC stretches.
Most avian prmt1 genes are poorly assembled but assembly of other avian prmt1 revealed high sequence conservation We used the assembled chicken prmt1 cDNA as the query to BLAST avian genomes to detect more avian prmt1 genes. Forty-three hits from twenty-two avian species could be identified with the default parameters. The identities of the nucleotide sequences are mostly higher than 90%. The only target with the sequence identity lower than 85% (83%, corresponding partial exon sequence) is from prmt8. Of the rest forty-two targets, thirty correspond to a single exon of prmt1 and twenty-one are in scaffolds shorter than 1000 bp, indicating that they are from poorly assembled chromosomal fragments. Among the hits, 17 contain exon 8 and none contain exon 4, the smallest constitutive exon (S1 Table). A scaffold that contains six prmt1 exons (2,5,7,8,9,10) was detected in common canary Serinus canaria. Though exon 3, 4 and 6 may be missed in the scaffold, partial sequences of exon 3 can be identified in another scaffold. We then manually assembled the canary prmt1 based on these sequence. We also assembled and re-annotated the prmt1 from four scaffolds (with exon 3, exon 5 and 7, exon 8 and 9, exon10) from American crow Corvus brachyrhynchos and three scaffolds from Tibetan ground tit Pseudopodoces humilis (exon3, exon 5, and exon 7,8,9) together with available predicted mRNA sequences (S1 File). Sequence alignments of the predicted protein-coding region of the assembled avian prmt1 sequences showed high conservation (S2 File). We also aligned the four avian PRMT1 sequences assembled in this study with predicted falcon PRMT1 sequences as shown in Fig 2. Though some of the sequences are incomplete, it is clear that they are highly conserved with few residue variations. The N-terminal sequence of PRMT1 encoded by exon 1 varies in different species and may be misannotated by automatic prediction. Almost perfect alignments of the N-termini from the predicted chicken and ground tit PRMT1 with that from human and alligator PRMT1 suggest that the sequences are highly reliable.

Comparison of the gene arrangements close to prmt1 in avian species and in sauropsids
As indicated in the previous sections, most of the avian prmt1 genes are now fragmented in small scaffolds from unknown chromosomes. We could obtain syntenic gene arrangements around prmt1 only in saker falcon with 21 genes in the scaffold containing prmt1 (Fig 3A). We then tried to trace the gene arrangement neighboring prmt1 through evolution. We used American alligator (Alligator mississippiensis; a reptile as a non-avian sauropsid), a crocodilian with the best assembled genome, to trace the syntenic block arrangement in the common ancestors of archosaurs. Fifteen genes are in the same assembly with prmt1 in alligator as shown in Fig 3B. Among them, only rcn3 is a direct prmt1 downstream neighbor with the same opposite direction in falcons.
Human PRMT1 is on chromosome 19, the chromosome that has the highest gene density and contains the majority of the missing blocks in the assembled avian genomes. The prmt1 gene in alligator clusters with seven orthologous genes directly upstream of human PRMT1 in the NOSIP-PRRG2-PRR12-RRAS-SCAF1-IRF3-BCL2L12-PRMT1 order ( Fig 3C). The rcn3 gene that is directly downstream of prmt1 in alligator rearranged upstream of the cluster in human. Human orthologues of the five prmt1 downstream genes cpt1c-tsks-ap2a-fuz-med25 in alligator also clustered the same way on chromosome 19.
Missing chicken genes were reported to be clustered in syntenic blocks in lizard and human. To follow the possible chromosome rearrangements near prmt1, we also analyzed the genes in green anole (Anolis carolinensis), another model reptile. Most of the orthologues of human PRMT1 neighbors are present on chromosome 6 of green anole. The upstream genes nosip-prrg2-prr12-rras-scaf1-irf3 are linked and five downstream genes cpt1c-tsks-ap2A-fuz-med25 are clustered together as in human (Fig 3B). The rcn3-pax6-adm arrangement is present in both alligator and green anole but not human or birds and is connected directly with the cpt1c to med25 block. However, the prmt1 gene in green anole is assigned to a separate scaffold of 222,408 bp (NW_003338834, Anolis carolinensis unplaced genomic scaffold, AnoCar2.0 chrUn0095, whole genome shotgun sequence) with tbc1d17, its human orthologue is a few genes downstream of the cpt1c-tsks-ap2A-fuz-med25 cluster. The bcl2L12 gene connected with prmt1 in both human and alligator is located upstream of the rcn3 and med25 clusters in near proximity in green anole. Comparison of the PMRT8 sequences and gene arrangements close to prmt8 in sauropsids In comparison, we performed similar analyses for prmt8, the vertebrate paralogous gene of prmt1. From "gene" search in PubMed, thirty-one prmt8 genes are annotated in birds and the prmt8 gene is assigned to chromosome 1 in five different species including chicken, Japanese quail (Coturnix japonica), Great Tit (Parus major), turkey (Meleagris gallopavo) and zebra finch (Taeniopygia guttata). The avian PRMT8s also share very high sequence identities (~99%). Basically, the prmt8 genes are in well assembled segments in avian genomes and the arrangements can be obtained in numerous species. We illustrated the syntenic gene arrangement around prmt8 in chicken (Fig 4A). The arrangements of genes at the downstream side are conserved in avian species. However, the genes directly upstream of prmt8 are completely different in chicken and falcons ( Fig 4A) and other avian species.
When compared with gene arrangement in alligator, cracr2A-parp11 downstream of prmt8 in alligator fused with a syntenic block of eight genes from tspan9 to fkbp4 as in falcon. However, tspan11, the gene directly upstream of alligator prmt8, is missing in the avian species. The cracr2A gene directly downstream of prmt8, and the nrip2 gene in the tspan9 to fkbp4 block, are missing in chicken but not in most other avian species (Fig 4B). A syntenic block of seven genes from ccnd2 to dyrk4 downstream of prmt8 in alligator, though is not directly neighboring prmt8 in chicken, is still present on the same chromosome around 2 Mbp upstream with dyrk4 missing. This block is completely or partially reserved in synteny with prmt8 in birds with different distances from prmt8.
We also examined the gene arrangements around PRMT8 in human (Fig 4C). For the 19 genes from ccnd2 to fkbp4, the only major differences between human and alligator are that tspan11 in alligator is moved to further downstream of PRMT8 in human, and the syntenic block of the seven genes from ccnd2 to dyrk4 in alligator flipped over in human.

Expression of chicken prmt8 at the RNA and protein levels
In mammals, PRMT1 is broadly expressed in different tissues and different PRMT1 isoforms function in cytosol and nucleus. In comparison, its paralogue PRMT8 is brain-specific and plasma membrane associated. As we confirmed the presence of chicken prmt1 transcripts in Fig 1C, we also analyzed the expression of prmt8 in chicken and mouse tissues by RT-PCR. As shown in Fig 5A, to our surprise, besides strong signals for prmt8 transcripts in chicken brain, weak signals could be detected in chicken liver and muscle. However, mouse prmt8 RNA was present in brain but not in liver or muscle as expected.
We then prepared brain, liver and muscle extracts from both chicken and mouse to examine the expression of PRMT1 and PRMT8 proteins. As shown in Fig 5B by western blot analyses, we detected weak signals in chicken brain and muscle but the signals detected in chicken muscle were lower than that of mouse PRMT1. Nevertheless, strong PRMT1 signals were detected in all mouse tissues. We also showed that the antibody (Abcam ab7027) could well recognize zebrafish PRMT1. Chicken PRMT1 shares 96% while zebrafish shares 91% sequence identity with human PRMT1. Thus the weak/no chicken PRMT1 signals are less likely to be due to the inability of the PRMT1 antibody to recognize the chicken PRMT1 protein.
Furthermore, another antibody (Millipore/Upstate 07-404) that could well recognize zebrafish PRMT1 as shown in our previous study [14], detected strong PRMT1 signals in all three mouse tissues but no PRMT1 signals in chicken brain and liver, even when there were three times more chicken protein (75 μg) compared with mouse protein (25 μg). Again very weak signals detected in chicken muscle were lower than that of mouse PRMT1 (S1 Fig).
For different tissues, the amount of GAPDH and beta-actin varies as detected by the antibody and thus were difficult to be used as loading controls. Nevertheless, coomassie blue staining showed the similar amount of protein loading of the chicken, mouse or zebrafish samples. The same tissue showed similar protein expression patterns whether they were from chicken or mouse (Fig 5C).
For PRMT8, the western blot results were consistent with the RT-PCR results. Interestingly, the PRMT8 level in chicken brain was much higher than that in mouse brain. Besides, weak PRMT8 protein signals were detected in chicken liver and muscle (Fig 5D). In mouse, the PRMT8 signals could be detected in brain but not in liver or muscle.

Asymmetric dimethylarginine protein patterns in mouse and chicken tissues
Since PRMT1 is the major type I PRMT in other vertebrates but the expression level of chicken PRMT1 is low as shown in the previous section, we analyzed the levels of asymmetric dimethylarginine (ADMA) containing proteins formed by the catalysis of type I PRMTs in chicken. Identification, chromosomal arrangements and expression analyses of prmt1 and its paralogue prmt8 in chicken The levels of ADMA-containing proteins in chicken were close to that in mouse as detected by western blot analyses with ADMA-specific antibodies (Fig 6A). For brain and muscle extracts, the patterns of ADMA-containing proteins from the same tissue were similar whether they were from mouse or chicken. Nevertheless, two strong signals of the molecular mass of about 50 and 40 kDa were detected in chicken but not mouse muscle.
Birds retain type II and type III PRMTs catalyzing the formation of symmetric dimethylarginine (SDMA) and monomethylarginine (MMA). In comparison, we also analyzed the expression pattern of SDMA and MMA containing proteins with specific antibodies (Fig 6B  and 6C). In general, SDMA and MMA signals in the same tissues were of similar levels and patterns whether they were from mouse or chicken.
Specifically, an intensive signal of the molecular mass around 20 kDa was detected in chicken but not mouse brain by all three methylarginine-specific antibodies (Fig 6A, 6B and 6C). A major polypeptide signal can be detected by protein staining at this molecular mass, indicating that it is highly abundant in chicken brain. We subjected this signal to mass spectrometric analyses for its identity. Multiple polypeptides with more than five peptides with the MASCOT score suggesting identity were identified including myelin basic protein that has been reported to contain SDMA [27].
We further detected the arginine methylation status of specific PRMT1 substrates in chicken. PRMT1 is an epigenetic modifier catalyzing the methylation of the arginine 3 (R3) residue of histone H4 to form H4R3Me2a [12]. We prepared histone proteins from both mouse and chicken brain. Specific H4R3 modification could be detected in chicken as well as in mouse brain using the H4R3Me2a-specific antibody (Fig 6D).

Fractionation of the brain extract for subcellular distribution of type I PRMTs in chicken and mouse brain
We would like to determine the PRMT activities in chicken. By in vitro methylation reactions, the methylation signals catalyzed by the tissue extracts were very low, with brain extract contained the highest PRMT activity. In case that there were inhibitors or certain factors in the total extracts that might interfere the assays, we further used a scheme similar to that had been used to fractionate porcine brain [24] to fractionate mouse and chicken brain extracts, and used the subcellular fractions as the enzyme source for methyltransferase assays (S2A Fig). The pellet from the first low-speed centrifugation (P1) is the nuclear fraction, and the pellet from the second high-speed centrifugation (P2) consists of large organelles such as mitochondria. The pellet from the ultracentrifugation (P3) contains membranes and super complexes such ribosomes, while the supernatant of the ultracentrifugation (S3) is cytosol. We used a recombinant glycine and arginine rich (GAR) protein with the GAR sequences from a typical PRMT substrate fibrillarin as the substrate. As shown in Fig 7, the supernatant fractions contained higher activity than the pellet fractions and the S3 fraction representing cytosol appeared to have the strongest type I PRMT activity in both mouse and chicken brain extract.
We also compared the distribution of PRMTs in the fractions (S2B Fig). PRMT1 signals were in all fractions of mouse brain extract but the strongest one was in the S4 fraction, the high salt stripped supernatant of the P3 fraction from ultracentrifugation. The peripheral association of mouse brain PRMT1 with membranes/ribosomes is similar to the PRMT1 distribution in the porcine brain [24]. Mouse and chicken PRMT8 showed similar expression pattern as mouse PRMT1.
We had examined the levels of prmt3 transcripts in both chicken and mouse and observed that high expression level is high in brain and low in liver and muscle. We also detected the expression of PRMT3 in these brain fractions. PRMT3 in mouse and chicken was mainly detected in the S4 fraction. The result is consistent with the reports that PRMT3 exists only in the cytosol of mammalian cells [28] and methylate ribosomal protein S2 [29]. The distribution of PRMT5, the predominant type II enzyme catalyzing the formation of SDMA, was also detected. PRMT5 was present in all fractions and the strongest signals were in the P1 and P3 fractions. The S4 and P4 fractions both from the P3 fraction also showed high levels of PRMT5.
We finally detected some markers to verify the fractionation. Fragile X mental retardation protein (FMRP) and heterogeneous nuclear ribonucleoproteins A2/B1 (hnRNP A2/B1) Identification, chromosomal arrangements and expression analyses of prmt1 and its paralogue prmt8 in chicken showed the highest expression in S4. Besides, histone H3 was only detected in the P1 fraction, consistent with the nuclear distribution. These results confirmed the successful fractionation of the brain extracts.

Discussion
In this study, we focused on the prmt1 gene encoding the most conserved PRMT in birds. At first, we considered that this gene is missing in birds because we could not identify the gene in chicken genome, the earliest assembled and annotated avian genome assembly. In a previous study of conserved syntenic clusters of protein coding genes missing in birds, a block with genes just downstream of human PRMT1 was recorded as a deletion group, and some genes just upstream of PRMT1 were listed as missing genes in chicken in close proximity to missing syntenic blocks [21]. We thus speculated that a whole block of genes with prmt1 was deleted in chicken. However, there are reports that genes previously considered missing in birds have been identified in avian species especially the recently sequenced and assembled genomes of saker falcon, perigrine falcon and ground tit [30]. Indeed, the prmt1 gene was annotated in two falcons. We thus carefully examined the avian and chicken nucleotide databases and identified a few cDNA and EST targets encoding PRMT1. We assembled the sequences and predicted two chicken prmt1 variants with high sequence identities with the human and alligator PRMT1. RT-PCR analyses confirmed the chicken prmt1 expression. Besides, we identified two separate small contigs for separate prmt1 exons, both with high GC contents. It is consistent with the study by Hron et al that a subset of avian genes with high GC content and long GC stretches are underrepresented in genomic databases and thus are hidden but not missing in birds [31]. Thus the prmt1 gene is present in chicken but genomic fragments containing the gene were poorly recovered and assembled.
Further BLAST analyses against avian genomes in the databases showed that the prmt1 gene is fragmented and only single and few exons are present in short scaffolds in most avian genomes. We also tried to assemble prmt1 in other avian species with the limited sequences available. Extremely high sequence identities within avian PRMT1s and with human or alligator PRMTs indicate that PRMT1 is highly conserved in amnions.
The slowly evolved crocodilians are proposed to be close to the common ancestor of the archosaurs group containing crocodilians, birds and the extinct dinosaurs. We compared gene arrangements neighboring prmt1 in birds (falcons), reptiles (American alligator and green anole) and mammals (human). Only the rcn3 gene linked to prmt1 in alligator is conserved in falcon. On the other hand, the gene arrangements near prmt1 in alligator is very close to that in human. A syntenic block of eight genes upstream and another syntenic block containing five genes downstream of PRMT1 are conserved in human and alligator. In human, the RCN3 gene downstream of PRMT1 in alligator moved upstream of the NOSIP-PRMT1 block and flipped its direction. Interestingly, most of the orthologues of alligator and human PRMT1 neighbors are present on chromosome 6 in green anole, but the prmt1 gene is assigned to a separate long contig. Besides, the bcl2l12 gene connected with prmt1 in both human and alligator is located upstream of the rcn3 and med25 clusters in near proximity in green anole. Thus the rcn3 and bcl2l12 genes together with prmt1 are the major breakpoints for chromosome translocations through evolution from the common ancestor to birds, mammals, and green anoles.
In the recent Gallus_gallus-5.0 genome assembly with 1911 protein coding genes added, some human PRMT1 neighbors previously consider missing [21] have been re-assigned to separate scaffolds of unknown chromosomes or have been predicted in other avian species but not in chicken. Genes including RCN3 remained in the missing gene list [22]. Nevertheless, we showed that rcn3 exists in falcons just behind prmt1, thus is present in birds. Though the missing gene list in chicken is shrinking, it is apparent that genes near human PRMT1 in a large syntenic block on human chromosome 19 are underrepresented in chicken and many avian genomes currently. Human chromosome 19 has the highest gene density with 1,461 proteincoding genes and 321 pseudogenes [32]. Chicken microchromosomes have been considered to be gene rich but are poorly assembled even in the recent Gallus_gallus-5.0. They have higher GC content than macrochromosomes [22]. An attractive proposal is that chicken orthologues of human PRMT1 together with its neighbors on chromosome 19 are on microchromosomes that are still missing in the current assembly. However, more investigations are required for the proposal.
Besides PRMT1, CARM1 (PRMT4) is also on human chromosome 19. PRMT1 and CARM1 locate on chromosome 19q13.3 and 19p13.2 respectively and both locations are reported to have syntenic blocks deleted in birds [21]. CARM1 is critical in mouse embryogenesis and mouse with a targeted disruption of CARM1 are small in size and die perinatally, thus is essential in mammals [33]. As reported by the recent Gallus_gallus-5.0, carm1 can be predicted in eagle with correct synteny but does not BLAST align to chicken [22]. From gene search, carm1 is predicted in 50 different avian species but no in Galliformes. Nevertheless, we could amplify chicken carm1 cDNA by RT-PCR with the primer sequences conserved in other avian species. Thus carm1 is not missing in chicken and its expression and function in birds require further analyses.
We also performed similar analyses for the prmt1 paralogous gene prmt8. Unlike prmt1 with no chromosomal assignment, prmt8 is well assembled in alligator and some avian species. The gene is located on chicken chromosome 1 with neighbors annotated. Genes neighboring prmt8 in alligator are in a similar arrangement in a syntenic block downstream of prmt8 in the avian species we examined. However, the direct neighboring genes upstream of prmt8 are completely different in different avian species and alligator (Fig 4). The gene arrangements around prmt8 in human and alligator are highly conserved with one gene transposition and one syntenic block turn over.
The amniotes divided into the ancestral lineages of mammals and reptiles about 320 million years ago (MYA). Archosaurs containing crocodilians and birds is one of the two major clades in reptiles diverged about 280 MYA. Birds originated from a theropod lineage more than 150 MYA [18]. As alligator was proposed to be the closest to the common ancestor of archosaurs, through analyzing the human, crocodilian, chicken and a few avian genes on chromosomes, we suggested that prmt1 with its syntenic clusters might be specifically rearranged in the evolution of the avian lineage since 280 MYA. Though the data is limited, it appears that both prmt1 and prmt8 changes their neighbors (at least one side) in different avian lineages in the recent 150 million years. In contrast, their neighbors and arrangements between crocodiles and human are highly conserved since 320 MYA.
We conducted RT-PCR and western blot analyses to confirm the expression of PRMT1and PRMT8 in chicken as well as to compare the PRMT expression in chicken and mouse. To our surprise, it appears that the expressed protein level of PRMT1 in chicken is lower than that in mouse, but the level of PRMT8 is higher than that in mouse. While PRMT1 is the major type I PRMT in almost all vertebrate species, in chicken we could barely detect PRMT1 protein expression in the tissues we examined. The antibodies we used could recognize zebrafish PRMT1. Compared with chicken PRMT1, zebrafish PRMT1 is less conserved with human PRMT1 that is the immunogen of the antibodies. It is thus unlikely that the weak/none chicken PRMT1 signals was due to poor antibody recognition of chicken PRMT1. Because we could detect the mRNA expression of chicken prmt1, why and how prmt1 is regulated at the posttranscriptional or translational level requires further experiments.
Previous analyses of PRMT1 knockdown/knockout animals or cells all showed a significant decrease of ADMA-containing proteins as detected by western blot analyses using ADMAspecific antibodies [14,34,35]. With the limited expression of PRMT1, the level of ADMAcontaining proteins detected in the chicken samples was of the same level as that in mice. Furthermore, arginine 3 of histone H4, a PRMT1 specific substrate, was methylated in chicken as detected by an H4R3me2a-specific antibody. These results indicated that the type I PRMT and PRMT1-like enzymatic activity function well in chicken. Furthermore, using a recombinant GAR protein as the substrate, we detected a similar level of type I PRMT activity in mouse and chicken brain. Nevertheless, in vitro methylation analyses revealed that fractions containing the highest type I PRMTs activity in mouse and chicken were not in accord with any PRMT expression patterns we examined. Though GAR with glycine and arginine rich sequences from human fibrillarin protein is a typical type I PRMT substrates, we could not exclude the possibility that type II and III PRMTs might still recognize the protein and catalyzed the reaction.
For other PRMT to complement low PRMT1 levels in chicken, it has to function in the same subcellular compartments as other vertebrate PRMT1. Fractionation of the brain extracts from either chicken or mouse showed highly concentrated PRMT3 in a specific fraction consistent with the previous reports of cytoplasmic distribution of PRMT3 [28] with the strong association ribosomes [29]. Thus PRMT3 is less likely to compensate for the function of other PRMTs in chicken for its restricted subcellular distributions.
PRMT8 is a minor brain-specific paralogue of PRMT1 in other vertebrates [15]. However, we detected high PRMT8 expression in chicken brain. Interestingly, PRMT8 protein in chicken brain is of a much higher level compared with that in mouse brain. Chicken PRMT8 was highly expressed in brain and distributed in similar sub-fractions as mouse PRMT1 or PRMT8. In addition, weak prmt8 RNA, as well as protein signals, could be detected in chicken liver and muscle, indicating putative de-repression of prmt8 expression in these chicken tissues. PRMT8 had been shown to methylate histone H4 in vitro [15,36]. Though membrane localization of PRMT8 due to N-terminal myristoylation had been shown [15], there was evidence of nuclear localization of endogenous PRMT8 by immunohistochemical analyses of mouse brain tissues or stem cell derived neuronal cells [37,38]. PRMT8 had been considered as a chromatin modifier in neuronal differentiation [39] and had been shown to be an epigenetic regulator for synaptic maturation by chromatin immunoprecipitation [40]. Thus it is possible that the H4R3me2 modification detected in chicken samples could be modified by the small amount of PRMT1 and/or also PRMT8. However, more experiments are required to examine the postulation that PRMT8 might at least partially complement low PRMT1 expression in chicken. PRMT8 knockout mice study showed that PRMT8 has the phospholipase activity to regulate Purkinje cell dendritic arborization and motor coordination [16]. We could not exclude the possibility that it is the phospholipase activity but not the PRMT activity that is critical for PRMT8 and upregulates PRMT8 expression in birds.
In summery, we determined the cDNA sequence of chicken prmt1 and showed that its coding sequence is highly conserved with that of other amniotic PRMT1. However, the chromosomal localization of chicken prmt1 remained undetermined and the PRMT1 protein expression level we detected was low. It is not known why chicken maintained the prmt1 gene but expressed a low level of this critical PRMT. Instead, prmt8 that is on chromosome 1 in chicken expressed at a high level. Chicken thus appears to be a unique natural system with prmt1 knocked down and PRMT8 overexpression. Our studies provide the basic bioinformatic as well as biochemical information for further investigation in birds to elucidate the critical functions of protein arginine methylation conserved and not conserved in vertebrates.