Integrative mapping analysis of chicken microchromosome 16 organization

Sharing the characteristics of a classical avian karyotype, the chicken genome, as standardized by Ladjali-Mohammedi et al in 1999 [1], is composed of 9 macrochromosomes including the Z and W sex chromosomes and 30 pairs of microchromosomes. These very small chromosomes range in size between 3.5 and 23 Mb [2, 3] and are remarkably gene rich [4]. The draft sequence assembly, comprising the macrochromosomes and the largest microchromosomes, has an overall high quality, apart for numerous gaps and minor local rearrangements [5–7]. However, despite international efforts towards a complete sequence, 9 microchromosomes still remain absent from the current assembly. This can be caused by several factors, such as the difficulty to identify them by cytogenetic approaches or/and problems for cloning and sequencing these (G+C)-rich DNA fragments [4, 8]. The contigs which could not be assigned to a chromosome are arranged in a virtual chromosome (Chromosome Unknown, chrUn) and we have previously demonstrated that some of these sequences could be assigned to small microchromosomes [9, 10]. Moreover, we have also shown that many chicken cDNA and EST (Expressed Sequence Tag) sequence contigs are absent from the assembly, including in the chrUn fraction, suggesting the total absence of large amounts of the corresponding genomic DNA [10].

GGA16 is the chromosome bearing the NOR (Nucleolus Organizing Region, cluster of genes for ribosomal RNA 18S, 5.8S and 28S) and illustrates perfectly the sequencing and mapping problems encountered for avian microchromosomes. Its sequence is far from being complete, with only 433 kb of sequence presented in the current assembly whereas its estimated physical size is between 9 and 11 Mb [11]. Even when taking the rRNA genes that cover around 6 Mb and represent 40 to 70% of the chromosome into account [12], a large proportion of GGA16 remains to be sequenced.

Interestingly, the chicken MHC (Major Histocompatibility Complex), the role of which is central in the immune response, has also been mapped to GGA16. It is composed of two loci, named the B and Y complexes [13]. The B complex was first identified as a polymorphic blood group system in chicken [14], leading to the first method to identify MHC genotypes by serology. Later, the locus was associated with graft rejection [15] and with genetic control of the immune response [16–18], i.e. classical functions of MHC locus in vertebrates. The B region has been strongly associated with resistance to infectious diseases such as Marek's disease [19].

The first mapping of the B complex was conducted by Bloom and Bacon [20] who assigned it to a single microchromosome pair together with the NOR. Dominguez-Steglich and her co-authors [21] confirmed this result using non-isotopic in situ hybridization. The Y complex was discovered later and determined as genetically independent [22, 23]. Despite this independence, the two complexes are located together on the same microchromosome, referenced as the pair number 16 [24, 25]. The GGA16 genetic linkage group contains only a few microsatellite markers, organised in two independent groups which are separated by 60 cM [26]. Curiously, the recent high-density SNP-based linkage map has no SNP on GGA16 [7]. Finally, no Radiation Hybrid map was constructed for this microchromosome [10]. By comparison with the canonical organization of vertebrate MHC, the annotation of the chicken MHC-B region shows that only the main small class I and class II core sub-regions are represented in the genome assembly. The extended class I and class III sub-regions are only very partially represented [27, 28], and many genes usually present in mammalian MHC seem absent. Therefore, despite many attempts at sequencing GGA16 by whole genome or chromosome-specific mapping and sequencing approaches, including very early contig maps built with cosmid clones [29], this chromosome still remains badly covered. Amongst other consequences, this prevents from performing proper Quantitative Trait Loci (QTL) detection in a region which is important for the immune response.

Due to the difficulties encountered in obtaining information on the very specific organization of GGA16 by the sequencing and/or physical BAC chromosome walking strategies, we decided to undertake an integrated approach focussed specifically on this microchromosome. Recently, the order of the NOR, MHC-Y and MHC-B complexes was defined by FISH mapping on mitotic prometaphase and meiotic prophase I pachytene chromosomes [30]. To confirm this organization and build precise genomic markers maps, we integrated three alternative mapping strategies, namely Radiation Hybrid (RH) mapping, genetic mapping, and high resolution FISH mapping on giant lampbrush chromosomes from growing oocytes, allowing the estimation of recombination rates along the chromosome.

Chicken NOR-bearing chromosome GGA16 was originally identified as microchromosome 17 using classical silver staining techniques [43] and re-numbered later as chromosome 16 [25]. Due to technical issues the size of this chromosome was overestimated and its real length is much smaller than chromosome 19 [8]. Past estimations of the total amount of DNA for this chromosome were between 9 and 11 Mb [11]. We have here new estimations for the size of GGA16. The first is deduced from the counting of chromomeres: the average amount of DNA per chromomere in an avian lampbrush microchromosome is between 1.4 and 1.5 Mb [39] and we identified 7 chromomeres in chicken LBC16 (Figure 4), resulting in a size estimation between 9.8 and 10.5 Mb. The second is deduced from our RH map: when mapping the other chicken chromosomes with the chickRH6 panel, we found an average value of 38.7 kb/cR [10], our GGA16 RH map is 244 cR (Figure 1), which suggests a coverage of 9.5 Mb. Therefore, the two size estimations agree together and with the most recently published one [11], and they suggest our RH map covers GGA16 almost entirely. We confirm therefore that approximately 95% of the GGA16 sequence is missing from the current genome assembly.

Three major gene complexes were assigned so far to GGA16: the NOR, MHC-B and MHC-Y. The fact that the NOR and MHC-B are linked on one microchromosome was discovered 25 years ago [20] and the assignment of MHC-Y more recently [24, 25].

The chicken MHC shows strong similarities with mammalian MHC with the presence of genes from class I, II, and III and extended class I regions, but its organization is very different, principally due to the existence of the two genetically independent complexes MHC-B and MHC-Y. The chicken MHC-B complex contains, among other genes, two genes homologous to classical mammalian class I genes: B-F1 and B-F2, and two genes homologous to classical mammalian class II genes: B-L1 and B-L2[13]. It was proposed that the central region of the B complex could be considered as minimal and essential, as the limited number of class I and class II genes are found in tight linkage with other genes having functions in the immune response [44]. The MHC-Y complex - initially named Rfp-Y (restriction fragment pattern Y) - was discovered later, and was defined as a polymorphic locus "MHC like" [22]. This second MHC region also contains expressed class I and class II genes [45]. The independent genetic assortment of the two complexes has been known for years [22, 23], first suggesting they were located on two different chromosomes. However, FISH mapping on chicken chromosomes clearly showed that these two loci both map to GGA16 at a short physical distance, given the small size of this chromosomes [24, 25].

For long, the order of the three loci along GGA16 was difficult to resolve as genetic maps were composed of two independent groups, as exemplified by our data (Figure 2) and FISH mapping to mitotic microchromosomes gave signals too close to one another for determining an order (Figure 3). The most commonly admitted model was that the NOR region should be situated between the two MHC complexes which were situated at the two extremities of the chromosome. In this model, the NOR was thought to act as a spacer allowing for meiotic recombination events to take place [25].

Our detailed RH map and high-resolution FISH mapping results on giant lamprush chromosomes from chicken oocytes clearly demonstrate a different order for the three complexe: centromere - NOR - Y - B (Figure 1). The orientation with respect to centromere is obvious both from FISH data in which the centromere is visible and from RH mapping data. For the latter, the position of the centromere on the map is deduced from the retention frequency values along the chromosome, which are usually higher near the centromere [35–37] (Figure 1). The same gene order was recently shown by Delany and her colleagues (2009), based on cytogenetic observations on elongated mitotic and meiotic chromosomes [30]. Thus, with no doubts we can say now that the elusive order of the NOR, Y and B complexes in GGA16 has been established.

The genetic map was enriched by developing polymorphism for markers selected in the RH map and integrating them with markers from the consensus linkage map. The average cR/cM ratio is about 2 when calculated over the whole map length. This relatively low value, as compared to the 4 cR/cM obtained for GGA7 [35] or 6.5 cR/cM for GGA5 [37], is influenced by the disparity in recombination rate observed along the chromosome and is a consequence of the systematic recombination event between the MHC-Y and MHC-B complexes.

By high resolution mapping on LBC16 we showed that the segment between the MHC-B and MHC-Y complexes contains many copies of the highly repeated 41 bp tandem repeat, PO41. Until now, PO41 sequences were found to concentrate in the subtelomeric regions of GGA1, GGA2, GGAZ and GGAW [42] and in the current chicken genome assembly, sequences homologous to PO41 are only assigned to the virtual chromosome (chrUn), probably due to difficulties with the repeats in the assembly process. GGA16 is the first identified chicken microchromosome known to contain this repeat. PO41 repeats are (G+C)-rich (63%) and their presence on GGA16 fits the hypothesis of Delany et al.[30] that a GC-rich region separates the MHC-Y and MHC-B. This organization of the MHC complexes seems to be evolutionary conserved in birds, even if not yet demonstrated in Passerines [46]. For instance, a GC-rich satellite MgaSat2 as well as an ultra-long internal telomere array were found between genetically unlinked MHC-B and MHC-Y loci in turkey [47]. The long arrays of tandem repeats between the two MHC complexes in birds can explain their independent genetic segregation. Our data made us return to the idea of highly repeated sequences playing a major role in the formation of recombination hot spots. Indeed, inverted-repeated sequences are known to induce DNA double-strand breaks (DSB) during meiosis in yeast [48], as the first step in a series of events associated with meiotic recombination. Importantly, the same machinery responsible for normal recombination process is involved in palindrome-mediated DSB formation [48]. Although the PO41 repeated units themselves do not tend to form hairpins, they can be found in inverted orientation in the Chicken Genome sequence databases. Moreover, the existence of short tracks of inverted copies was demonstrated for PO41 using sense and antisense oligonucleotide FISH probes applied to chicken LBCs [42]. Thus, the inverted arrays of PO41 repeats can be responsible for DSB generation and consequently a high frequency of recombination events between the MHC-Y and MHC-B complexes in GGA16. Moreover, the fact that PO41 is intensely transcribed at the position of chiasma formation seems not to be accidental. The relationship between transcription and recombinations hot-spots is well-established (reviewed in [49]).

As it is difficult to obtain the precise distribution and dimension of the PO41 tracks, one can speculate about the degree of complexity of the region between two MHC clusters. For instance, part of it could be homologous to a recently cloned sequence from the zebrafinch germ-line restricted chromosome (GRC sequence) [50].

It is not excluded that the region of interest can also be interspersed with coding sequences which remain unsequenced due to compositional reasons. If so, the 12 EST found by examining sequences similar to GGA16 in the human genome and having no similarity with the chicken genome assembly (Additional File 1) could illustrate this case: they all map to the region between the two complexes, or near this portion of the map (Figure 1).

Comparative mapping with the human genome using the EST data reveals that in addition to the known and obvious HSA 6p21 region containing the human MHC (HLA), several other human chromosomes show limited homology or synteny conservation with GGA16 (Figure 1). The HSA01 terminal region and the HSA16 short arm region, which were already suspected to have a region of conserved synteny with GGA16 are confirmed here, both for a short genomic fragment. Although our data confirm the assignment of these regions to GGA16, the issue as to whether these fragments showing sequence similarities really correspond to evolutionary orthologous fragments remains unresolved. A larger conservation of synteny fragments is confirmed with HSA05 and a new human chromosome is shown to share synteny fragments with GGA16: HSA19, the most gene-rich human chromosome. The region of sequence similarity stands at the end of the HSA19 q arm, known to be nearly absent from the actual chicken sequence and genetic map [7]. This human region was already known to share synteny conservation with microchromosomes 30 and 32 [10] and chromosome 11 [51].

The GGA16 sequence coverage is still far from being complete but our addition of EST improves the knowledge on this microchromosome. One could have expected a better sequence coverage with the release of a novel Gallus gallus 12x 454 sequence assembly (http://genome.wustl.edu/genomes/view/gallus_gallus/). But a BLAST analysis of these new contigs with the available sequence assembly reveals only a partial coverage of the GGA16 sequences in this lastest sequence data source (Additional File 2).

Our data confirm and refine the GGA16 chromosome organization, by using three integrated mapping strategies: RH mapping, genetic mapping and high resolution FISH, especially on lampbrush chromosomes from growing oocytes which were shown here to provide a unique opportunity to study recombination hot spots and to analyse their molecular context. The markers developed here may be useful to improve the GGA16 sequence assembly, and for QTL mapping studies, for which this chromosome is usually missing. The high-recombining region bearing a systematic chiasma separating the two MHC complexes is open for further detailed research.

Several sources of DNA sequence fragments putatively belonging to GGA16 were used: chicken EST, chicken chrUn sequence and chicken BAC clones. Primer data and other information on the sequences used for marker development are given in Additional File 1. Primers for RH mapping or BAC screening were chosen using the Primer3 server (http://frodo.wi.mit.edu/primer3/).