|Home | About | Journals | Submit | Contact Us | Français|
Misidentification of the chicken leptin gene has hampered research of leptin signaling in this species for almost two decades. Recently, the genuine leptin gene with a GC-rich (~70%) repetitive-sequence content was identified in the chicken genome but without indicating its genomic position. This suggests that such GC-rich sequences are difficult to sequence and therefore substantial regions are missing from the current chicken genome assembly.
A radiation hybrid panel of chicken-hamster Wg3hCl2 cells was used to map the genome location of the chicken leptin gene. Contrary to our expectations, based on comparative genome mapping and sequence characteristics, the chicken leptin was not located on a microchromosome, which are known to contain GC-rich and repetitive regions, but at the distal tip of the largest chromosome (1p). Following conserved synteny with other vertebrates, we also mapped five additional genes to this genomic region (ARF5, SND1, LRRC4, RBM28, and FLNC), bridging the genomic gap in the current Galgal5 build for this chromosome region. All of the short scaffolds containing these genes were found to consist of GC-rich (54 to 65%) sequences comparing to the average GC-content of 40% on chromosome 1. In this syntenic group, the RNA-binding protein 28 (RBM28) was in closest proximity to leptin. We deduced the full-length of the RBM28 cDNA sequence and profiled its expression patterns detecting a negative correlation (R = − 0.7) between the expression of leptin and of RBM28 across tissues that expressed at least one of the genes above the average level. This observation suggested a local regulatory interaction between these genes. In adipose tissues, we observed a significant increase in RBM28 mRNA expression in breeds with lean phenotypes.
Mapping chicken leptin together with a cluster of five syntenic genes provided the final proof for its identification as the true chicken ortholog. The high GC-content observed for the chicken leptin syntenic group suggests that other similar clusters of genes in GC-rich genomic regions are missing from the current genome assembly (Galgal5), which should be resolved in future assemblies of the chicken genome.
The online version of this article (doi:10.1186/s12863-017-0543-1) contains supplementary material, which is available to authorized users.
The chicken leptin gene was recently identified, more than 20 years after the first identification of mammalian leptin . An extensive search for chicken leptin [2–5] led to erroneous identification, characterization and mapping ([4, 6–10]. While the erroneous mapping was retracted , the erroneous sequences of chicken, turkey and duck leptins were not withdrawn [GenBank: AF012727, AAC32381 and AAT38807], respectively. The high GC content of avian leptins [74 ± 2 (SE) %] is significantly higher than that in leptins from non-avian vertebrates [52 ± 3 (SE) %]  and contains repetitive and palindromic sequence elements. This observation pointed to the possibility that the avian leptins may be located on microchromosomes, which are known to contain these types of sequence elements and, in addition, are underrepresented in avian genomes assemblies [12, 13]. For example, the average GC-content in chicken chromosome 1 is 40.3% compared with an average of 56.5% for chromosomes 25 to 33 [GenBank: GCA000002315.3]. High-quality genome assemblies for chicken and other birds revealed a large proportion of missing genes compared to other vertebrates [14, 15]. The presumed loss of the protein-coding genes was estimated to be as high as 30% . These missing genes often form synteny groups, which are conserved among vertebrates. Therefore, it has been suggested that these genes were lost from avian genomes during the evolutionary process when microchromosomes were formed in the avian lineage [14, 15]. However, technical difficulties in identification of some avian genes, due to their sequence characteristics and high sequence divergence, have been suggested as an alternative explanation [1, 16–18]. Recently, the submission of a new version of the chicken genome assembly (Galgal5) revealed 1811 new protein-coding genes, however a fairly large number of genes belonging to clusters of genes showing conserved synteny in other species are still missing in the latest assembly . To better understand this phenomenon, it is important to characterize novel genes that were considered missing, especially those like leptin and its syntenic genes, most of which are missing in Galgal5. The mapping of leptin and characterization of its synteny group are essential; both as a final proof for the identification of chicken leptin and for shedding light on possible technical barriers that hamper identification of genes in the chicken genome and in other genome assemblies.
The chicken genome assembly reveals some conserved synteny between human chromosome 7 and chicken chromosome 1, including genes flanking leptin and a cluster of closely linked genes on both sides on human chromosome 7 . However, other gene orthologs within this region on chicken chromosome 1 map to human chromosome 22. This break in synteny gave ground to the hypothesis that leptin, together with flanking genes, could have been translocated onto another chicken chromosome, possibly a microchromosome. Here, we demonstrate using radiation hybrid (RH) mapping that leptin and six other closely-linked genes are in fact located in a highly GC-rich region on the distal tip of chicken chromosome 1p.
We mapped leptin using the well-established chicken RH panel in Wg3hCl2 cells, prepared by fusing chicken embryonic diploid fibroblasts with hypoxanthine-guanine phosphoribosyl transferase (HPRT)-deficient hamster cells [20, 21]. This analysis demonstrated that chicken leptin is located between markers SEQ0557 (428,679 bp) and SEQ559 (487,706 bp) at the distal tip of chromosome 1p (Fig. (Fig.1;1; Additional file 1: Table S1).
To corroborate this mapped position of leptin, we explored its co-localization with the orthologs of human genes located in the near vicinity of leptin. Using a BLASTN search, we identified poorly annotated chicken orthologs of RBM28, SND1, LRRC4, and FLNC in short unplaced scaffolds in Galgal5, which we annotated in a parallel study . miR-129a was found by blast in NT474180, and ARF5 was annotated in a short unplaced scaffolds in Galgal5. These leptin syntenic genes and their genomic scaffolds were found to have high GC-content of ~60% (Table (Table1).1). This signifies that these sequences have a lower complexity and a higher incidence of repetitive and palindromic sequences, all of which contribute to the technical difficulty of identification and mapping of GC-rich sequences in Galgal5.
Based on the sequence information gathered for these genes, PCR primers were designed (Additional file 1: Table S1) and the resulting amplicons were RH mapped. The analysis confirmed the close localization of these genes and leptin, which is consistent with co-localization of the human orthologs on human chromosome 7 (Fig. 1a-b). CPT1B, previously mapped to chicken chromosome 1 , was used as an additional positive control for the RH mapping and the synteny break on HSA22 (Fig. (Fig.1b1b).
In vertebrate genomes for which leptin and its syntenic group of genes are annotated, RBM28 is mapped adjacent to leptin. In most cases, it is in a tail-to-tail orientation (3′ to 3′; e.g. peregrine falcon, Falco peregrinus [23, 24]; fugu, Takifugu rubripes [GenBank Gene ID: 548,631]; green sea turtle, Chelonia mydas [GenBank Gene ID: 102,932,266]; and mouse, Mus musculus [GenBank Gene ID: 16,846]). Moreover, we have previously shown that leptin is very closely linked to RBM28 in peregrine falcon and fugu (1.8 and 0.4 kbp, respectively ). Since such close proximity and the conserved 3′ to 3′ orientation suggest a local regulation of transcription , we further characterized RBM28 and its expression pattern. We deduced the full-length cDNA of RBM28, using genomic and RNA-seq data available in GenBank (Fig. (Fig.2;2; Additional file 2: Fig. S1 and Table S2).
The assembled 2007 bp cDNA [ENA Project ID: PRJEB18741] was partially corroborated by a match to a single expressed sequence tag (EST, 838 bp, [GenBank: DR425791]). With 69% GC-content, this putative transcript was found to be capable of encoding a 668 aa polypeptide with 50–51% identity and 65% similarity to human and alligator RBM28 [GenBank: NP001159607; GenBank: XP006033669] (Fig. (Fig.2).2). Higher amino acid identity of close to 90% was observed within the 3 RNA Binding Motifs (RBMs), which aligned with RBMs 2–4 of the human ortholog, each of about 90 amino acids long. Chicken and alligator proteins had a structure layout consisting of only 3 RBMs suggesting that such layout is common to reptiles, turtles (e.g. green turtle [GenBank Gene ID: 102,932,266]) and to birds (e.g. peregrine falcon [GenBank Gene ID: 102,049,081], while in mammals it retained the 4 RBM composition that is present in fish (e.g. fugu, [GenBank Gene ID: 548,631]). The acidic region between RNA recognition motifs 2 and 3, reported for the human RBM28 , was found to be only partially conserved in the deduced proteins of chicken and alligator, but it was followed by an additional acidic region in chicken RBM28 (annotated 281–309, Fig. Fig.2).2). Among the 18 exon-intron junctions, 13 were found to be conserved between the chicken and the human genes (Fig. (Fig.22 and Additional file 2: Table S2).
We used a comprehensive RNA-seq experiment of a female and a male red junglefowl available in GenBank (Chickspress, [GenBank: PRJEB4677]) to characterize the RBM28 expression pattern. The broad expression pattern obtained by searching the data for tissues included in this experiment (Fig. (Fig.3a)3a) seemed compatible with the housekeeping role of RBM28 as a nuclear component of the spliceosomal ribonucleoprotein complex . Nevertheless, the variation in expression level between tissues suggests tissue-specific functional roles.
To explore the possibility that the conserved proximity of leptin and RBM28 relates to their transcriptional control, the expression profile of chicken leptin was explored using the same RNA-seq dataset. Comparison of the two patterns revealed that the tissues with the highest expression of RBM28 (adipose, lung, heart and spleen) had no significant leptin mRNA, whereas, tissues expressing leptin (cerebellum, cerebrum and hypothalamus) had only a basal level of RBM28 expression (Fig. 3a-b). Thus, a significant negative correlation (R = −0.7) was observed between leptin and RBM28 expression profiles, taking into account only tissues expressing an above average level of either RBM28 or leptin (RPKM >13.1 and 0.07, respectively). In humans, leptin and RBM28 are also in tail-to-tail orientation but at a distance of more than 40 kb and with no indication of reciprocal expression in the human RNA-Seq Atlas .
An interesting feature was the relatively high expression level of the chicken RBM28 in adipose tissue. As a first step to relate this expression to adipose tissue regulation, we explored the expression of RBM28 in a previously published RNA-seq study of abdominal fat in genetically fat and lean broiler-type chickens, which exhibit a 2.8-fold difference in visceral fat deposition at 7 weeks of age . The RNA-seq analysis of four lean and four fat individuals revealed a significantly higher level of RBM28 expression in the lean line (P = 0.03; Fig. Fig.4a).4a). Leptin was not significantly expressed in adipose tissues of the fat and lean chickens (~0.002 RPKM). Since layer chickens, bred for efficient egg production, are substantially different from broiler type chickens in various parameters related to control of energy balance, we compared the expression of RBM28 and leptin also between RNA-seq data from commercial strains of broiler (meat-type) and layer (egg-type) chickens . Significantly higher levels of RBM28 expression were observed in the commercial layer line (P = 0.02; Fig. Fig.4b).4b). Also in these two strains of chickens, leptin was not significantly expressed in visceral fat. For comparison, we searched RNA-seq dataset from obese and lean human [GenBank: SRX470443–45] and, as expected, leptin expression was higher in adipose tissue of obese individuals (P = 0.04; ). However, the expression of RBM28 was not significantly different between obese and lean humans (Fig. (Fig.4c4c).
We have mapped the chicken leptin gene together with five closely-linked genes to chicken chromosome 1p, and demonstrated that these genes form a syntenic block in both chicken chromosome 1 and human chromosome 7. The order of the distal scaffolds on chicken chromosome 1 in Galgal5 is not in a complete accordance with the order of markers obtained from our RH mapping. Future long fragment sequencing, will hopefully resolve the correct order of genes in this region.
We found that in addition to the high GC-content of the leptin coding sequence (68%), the neighboring genes and genomic sequences also had high GC-content (~70% and ~60%, respectively). It is likely that the high GC-content of this cluster of genes in chickens hampered previous identification of this syntenic group [4, 11] and could explain why the leptin locus is missing in chicken BAC libraries (Richard Crooijmans, Wageningen University, personal communication). This finding is compatible with the known higher GC content in macrochromosomes near telomeres .
The chromosomal region of the chicken leptin synteny group and CPT1B, which was the gene used as a positive control for mapping and placed telomeric to the leptin synteny group (Fig. (Fig.1b)1b) are closely linked to quantitative trait loci (QTLs) for performance and carcass traits [31–34]. Therefore, it is possible that leptin, RBM28 and CPT1B are positional candidate genes for QTLs related to growth and metabolic regulation. Therefore, reassessment of mapping of these QTLs using our improved map of this region will be important to uncover a possible effect of these candidate genes on production traits.
As a representative neighboring gene of leptin, we chose RBM28 for further characterization. RBM28 is a nucleolar protein associated with small nuclear ribonucleoprotein (snRNP) and ribosome biogenesis . Sequence similarity, conserved exon-intron boundaries (13 out of 18), and similarity in the domain structure with the mammalian and reptilian RBM28 proteins, confirmed the identification of chicken RBM28 and suggested at least partial functional conservation.
The pattern of expression obtained using red junglefowl RNA-seq data revealed its expression in all of the examined tissues, compatible with its apparent fundamental role. Nevertheless, the expression pattern also shows variability in level of expression, suggesting a tissue-specific role. Indeed, human RBM28 deficiency is not lethal and shows some specific phenotypes including alopecia, mental retardation, progressive motor decline, and hypopituitarism (ANE syndrome) [35, 36].
The high expression of RBM28 found in the visceral fat of chickens is intriguing due to the key role of its neighboring gene leptin, in the control mechanism of energy homeostasis, in mammals [1, 37]. This observation and apparent correlation of RBM28 expression with leanness and slow body growth, suggest evolutionary adaptations in adipose tissue of birds that recruited other signaling pathways than leptin. The unique challenges of some avian species imposed by migration and strong seasonal variations of feeding, may have required different adaptation of the control mechanism of feed intake and fat deposition than that in mammals.
Inverse correlation between leptin and RBM28 patterns of expression in red junglefowl tissue suggest opposite regulatory function such as that reported in mammalian hair follicle in mammals, where RBM28 stimulates and leptin inhibits anagen-phase development [36, 37]. This observation could help future studies of the interactions between leptin and RBM28. It is possible that their proximate chromosomal location contributes to their opposing expression pattern. In this respect, it is interesting to note that in the falcon, we have previously demonstrated that leptin and RBM28 are less than 2 kbp apart in a tail-to-tail orientation . An additional control mechanism could be through the activation of miR-203 by RBM28 , to which the leptin receptor is a potential target .
Mapping and characterization of genes in leptin’s synteny group suggested that its depletion from genome assemblies has been due to high GC-content. Detailed analysis of the gene adjacent to leptin, RBM28 and its mRNA expression pattern indicated increased transcription in adipose tissue of chicken breeds with lean phenotypes. Our observation of negative correlation between the expression patterns of chicken leptin and RBM28 could be related to their adjacent chromosomal positions.
PCR amplifications were carried out for each marker (Additional file 1: Table S1) in 15 μl reactions containing 25 ng DNA from the RH panel , 0.4 μM of each primer, 0.25 units Taq polymerase (GoTaq, Promega), 1.5 MgCl2, 0.2 mM dNTP, using the Applied Biosystems 2720 thermal cycler. The first 5 min denaturation step was followed by 35 cycles, of denaturation at 94 °C for 30 s, annealing at Tm for 30 s and elongation at 72 °C for 30 s. Each marker was genotyped twice and a third genotyping was performed in case of discrepancy between the first two determinations. The RH map was built as previously described  using the Carthagene software  and drawn with MapChart 2.0 .
Female broiler (Cobb) and White Leghorn (Lohman) chickens were purchased from commercial husbandries (Brown & Sons and Hasolelim, Israel, respectively) at the age of 1 day and grown according to recommended husbandry and feeding conditions (NRC 1994) with free access to food and water. At sexual maturation, as indicated by egg lay (about 4 month of age), samples of abdominal (visceral) fat were snap-frozen in liquid nitrogen after neck dislocation. Total RNA was prepared using a RNA isolation kit (miRNeasy, Qiagen).
cDNA libraries were prepared by the Uppsala sequencing Platform from 1 μg RNA using the TruSeq Stranded mRNA Sample Preparation Kit (Illumina, San Diego, CA) as described earlier . Briefly, libraries were prepared, from visceral adipose tissue of three individual birds of each chicken breed: broiler breeder and layer hens. The libraries were uniquely tagged and paired-end sequenced (2 × 124 bp) with Illumina HiSeq sequencers, producing about 33 million paired-end reads per library.
Sequence homology searches were carried out at NCBI [non-redundant nucleotide collection (NR), SRA, and WGS] using the BLAST family of programs. Relevant sequence entries were downloaded with their quality information (FASTQ format) and reassembled using either MIRA4  or GAP5 software . Sequences were aligned using CLUSTALW or MAFFT (http://www.genome.jp/tools/clustalw) with the default parameters and the GONNET matrix or E-INS-I strategy for amino-acids and nucleotides, respectively; and colored using the BOXSHADE program (http://www.ch.embnet.org/software/BOX_form.html).
Details about the PCR primers and amplicons used for the RH mapping, sequence statistic and position. (XLSM 14 kb)
Characterization of RBM28. Figure S1. cDNA and predicted protein sequences of RBM28. Table S2. RBM28 Exons in chicken alligator and human. (XLSX 13 kb)
We thank Richard Crooijmans (Wageningen University) for screening of leptin sequence in the chicken Wageningen BAC library.
The study was supported by the Israel Academy of Sciences 876/14, and Chief Scientist of the Israeli Ministry of Agriculture 0469/14 (to MFE and ES).
The annotated sequences are available in ENA under project ID: PRJEB18741. The sequence of RBM28 is also available in Additional file 2: Figure S1.
FP, SL, and MM performed the mapping of leptin and its neighboring genes. SB and LA identified some of the leptin neighboring genes, ES extended characterized the RBM28 predicted cDNAs and proteins. SM contributed to the bioinformatics search. MFE, EY, LAC, DB and LA, designed the experiments and wrote the manuscript. All authors read and approved the final manuscript.
All animal procedures were described before  and were carried out in accordance with the National Institutes of Health Guidelines on the Care and Use of Animals and Protocol IL536/14, which was approved by the Animal Experimentation Ethics Committee of the Agricultural Research Organization, Volcani Center, Israel.
All authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
The online version of this article (doi:10.1186/s12863-017-0543-1) contains supplementary material, which is available to authorized users.
Eyal Seroussi, Email: email@example.com.
Frédérique Pitel, Email: firstname.lastname@example.org.
Sophie Leroux, Email: rf.arni.esuoluot@xuoreL.eihpoS.
Mireille Morisson, Email: email@example.com.
Susanne Bornelöv, Email: firstname.lastname@example.org.
Shoval Miyara, Email: email@example.com.
Sara Yosefi, Email: li.vog.irga.inaclov@sifesoy.
Larry A. Cogburn, Email: ude.ledu.ga.sppa@nrubgoc.
David W. Burt, Email: firstname.lastname@example.org.
Leif Anderson, Email: email@example.comL.
Miriam Friedman-Einat, Email: firstname.lastname@example.org.