Project Topics
Get Complete Project Material File(s) Now! »
Study Summaries
This thesis is comprised of four papers in which both genomic and molecular biology approaches were used to identify novel genes involved in pigmentation and plumage variation in birds and describe their function. In paper I, a major driver of melanin distribution on individual feathers was studied using a chicken backcross and various molecular and computational methods. In paper II, the mutations associated with the sex-linked barring pattern in chickens were investigated using molecular methods to describe pattern formation. In paper III, a 4.5 mega base (Mb) inversion was found to be associated with alternate reproductive morphs in the ruff and in paper IV the effect of MC1R alleles on plumage variation in the ruff was studied in more detail.
Autosomal barring in chicken is strongly associated with segregation at the MC1R locus (Paper I).
Background
In the domestic chicken, two different types of barring pattern on individual feathers have been described (Smyth Jr, 1990): autosomal and sex-linked barring. Autosomal barring presents as a semi-dominant, autosomal inherited trait, which adds a dark eumelanin bar on a brown or depigmented background (Smyth Jr, 1990). Initially it was proposed that this type of barring pattern is the result of an interaction between two loci, namely Pg and Db (Smyth Jr, 1990, Moore and Smyth, 1972, Carefoot, 1999, Carefoot, 1984) (Table 1). Db is considered a restrictor of eumelanin as it keeps melanin away from certain body parts such as the breast and belly. In females Db results in an orange tan colouration over the entire body, except for the tail. The male phenotype is more variable and chickens shows a bright orange breast to various degrees (Gunnarsson et al., 2011). An 8.3 kilo bases (kb) deletion upstream of the transcription factor SOX10 on chromosome 1 is common to all Db individuals (Gunnarsson et al., 2011). Although no further exploration of the mode of action of this mutation has been carried out, the authors proposed that based on similar mutations in mouse, the deletion removes a cis-regulatory element, which might affect SOX10 expression in different body regions differently, leading to differences in melanogenesis in the breast and tail. Another theory suggests was that SOX10 is enhancing the expression of MC1R. As a consequence of the deletion, SOX10 is down regulated in Db birds and so is MC1R. This might cause a dosage-dependent down-regulation of MC1R resulting in the default production of pheomelanin on the body and breast but not in the tail. The authors did not elaborate on the potential function of SOX10 in feather follicles or the effect of the deletion on secondary patterns such as autosomal barring, but it is very likely that it does alter melanogenesis or related pathways. Hutt (1949) pointed out that autosomal barring could also be looked at as a trait, which restricts black eumelanin to a bar on an individual feather, just as dark pigment in Db birds is restricted to the tail (Hutt, 1949).
The second locus implicated in barring is Pg, for which the actual location has not been firmly established. Crossing experiments suggested that it should be found 20 cM from Db and 10 cM from Ml on chromosome 1 (Moore and Smyth, 1972). According to traditional Mendelian genetics, Pg is considered genetically homogenous and ubiquitously responsible for various within-feather pattern formations across different chicken breeds (Smyth Jr, 1990) (Table 1).
The Fayoumi is an old Egyptian chicken breed, which is believed to have originated from semi-rural chickens inhabiting the coastal area of the Nile and imported red jungle fowl chickens about 3,000 years ago. Fayoumis are always barred on a silver background. The barred plumage is typically visible in various degrees on the entire body, except head and neck. Fayoumis carry a rare form of the Birchen allele, E*R(Fay), at MC1R, which differs from the common Birchen E*R allele in its amino acid composition (Ling et al., 2003). E*R(Fay) is characterized by an amino acid exchange at position 133 (L133Q), whereas the more common Birchen E*R allele carry, among others, E92K, the mutation previously shown to be associated with black coat colour in mice (Robbins et al., 1993).
The goal of study I was to identify the genetic variant underlying the autosomal barring phenotype in the Fayoumi breed and gain some functional insights into how the autosomal barring pattern is formed.
Results and discussion
To verify the proposed position of the Pg locus on chromosome 1 and identify a candidate gene, we first generated a backcross by mating five Fayoumi females (Pg/Pg Db/Db) with two inbred Light Brown Leghorn males (pg+/pg+ db+/db+). Twelve homogeneous barred F1 females (Pg/pg+ Db/db+) were crossed again with another Light Brown Leghorn male. We then examined the 365 offspring progeny both phenotypically at hatch and at 12 weeks of age, and genotyped them for the Db mutation. If the assumed model of inheritance of autosomal barring (with two independent dominant mutations) was correct, we should observe this phenotype in 25% of the backcross progeny, which indeed was the case. A total of 102 chickens exhibited the characteristic autosomal barring phenotype, 203 did not have any pattern and were classified as wild-type, while 60 offspring did not fit either category. Among this set of 60 animals, 28 were neither plain nor clearly barred, and 32 additional males were initially phenotyped as either wild-type or autosomal barred at hatch. At 12 weeks of age, however, they had developed a reddish taint in the body region with no visible pattern anywhere on the body. This suggested that there was another pigmentation locus segregating in our pedigree, which appeared to act epistatically to autosomal barring. Those individuals were therefore excluded from further analysis.
Additional genotyping revealed that contrary to our expectations not all chicken classified as autosomal barred carried the Db deletion as expected. As many as 36 of the autosomal barred chicken (about 1/3) were Wild-type at this locus. We also had a closer look at the group of offspring, which was neither properly barred nor plain and discovered that this group almost entirely was db+/db+ as well (25 out of 28 offspring). The observation that Db apparently was not required to develop a regular autosomal barring but at the same time was also lacking in the group of chicken that showed irregular pattern, implied that Db is not required for autosomal barring as proposed in the literature, but is contributing (possibly among other unknown loci or genetic variants) to a more defined patterned phenotype. It appears likely that in the past breeders selected chickens with the most pronounced barring phenotype, most of them carrying the variant allele at Db, leading to the assumption that this locus is required for the pattern formation.
Next we used next generation sequencing (NGS) to pinpoint the location of Pg in the genome of our backcross progenies. We used the Fixation index (FST) and 50 kb sliding windows to look for regions with high genetic differentiation between autosomal barred and non-barred chicken. The highest differentiation between wild-type and patterned chicken was found on chromosome 11 covering a region between 18.2 – 18.9 mega bases (Mb) (the end of the chromosome in galGal4). This region contains a number of genes involved in pigmentation or melanocyte biology with MC1R being the most obvious candidate. To obtain a better resolution of the region on chromosome 1 and 11, we genotyped the entire pedigree for a total of 100 SNPs located on chromosome 1, 11 and a few individual high FST SNPs on chromosome 2. We were not able to detect a completely shared haplotype for either phenotype category on chromosome 1 or 2. On chromosome 11, however, all backcross progeny exhibiting either autosomal barring or an irregular pattern, shared the same haplotype that showed no recombination from approximately 18.2 Mb until the end of the chromosome. This haplotype included the Fayoumi allele at MC1R (E*R(Fay)), which was inherited to all patterned offspring, including those which showed an irregular pattern. Not a single wild-type or plain chicken was a carrier of this haplotype. The unusual low rate of recombination in this region might suggest that a structural variant such as an inversion could be causative of this observation. Although we did not specifically test our material for an inversion, we believe that this possibility is rather unlikely since others have reported similar low recombination rates in this region (Groenen et al., 2009).
A non-recombining region as observed in our pedigree makes the detection of candidate genes and mutations challenging, since the entire interval is statistically equally likely to carry the causative change(s). We used the UCSC Variant Annotation Integrator (VAI) to find mutations of interest within the non-recombined interval. We detected 42 non-synonymous mutations, which affect nine genes located within the interval of interest. However, PROVEAN, another online tool, which examined protein sequences based on their similarity, predicted that none of these mutation have a deleterious effect on the respective gene function, except for L133Q, the defining mutation for the E*R(Fay) allele at MC1R. As over 4500 non-coding SNPs were detected within the non-recombining region, it was challenging to pinpoint candidate variants, which could have functional implications. We therefore decided to evaluate the gene expression pattern in a subset of the in total 29 genes in the non-recombining region, which previously have been described to affect melanocyte biology and pigmentation including NAD(P)H quinone dehydrogenase 1 (NQO1), cadherin 1 (CDH1), ww domain-containing protein 1 (WWP) and MC1R. Except for WWP1, all investigated loci were found to be up regulated in growing, autosomal barred feathers as compared to the wild-type. If the elevated expression was the result of cis-regulatory mutation, then allelic imbalance in favour of the Fayoumi allele should be detected. This however was not the case. NQO1, MC1R and CDH1 were expressed in equal proportions from both alleles in heterozygous, autosomal barred chicken feathers.
There are a number of chicken breeds that are classified to carry Pg forming all sorts of patterns, including spangling or lacing. Since our analysis left us with only one candidate mutation, L133Q, in one gene, MC1R, we were wondering how transferable our findings are to other Pg breeds. The L133Q mutation, which is defining the Fayoumi Birchen E*R(Fay) allele, has not been found in other chicken breeds so far. Surprisingly, we found that the Fayoumi flock we used for our experiments was not fixed for E*R(Fay) either. The founders of our backcross were homozygous for the mutations but the entire flock was segregating for the MC1R E92K missense mutation as well, without any obvious heterogeneity in the autosomal barring phenotype. We further performed whole genome pooled sequencing of chicken breeds, which were described as carrying Pg. This approach would have made it possible to identify an identical-by-decent (IBD) haplotype with the Fayoumi breed, in case there is one, and would make it possible to screen the entire genome for any other fixed region between those breeds. We were not able to detect any shared haplotype, either on chromosome 11 or in the entire genome. This was surprising as it contradicts the long-standing hypothesis that Patterning is homogenous among Pg breeds. The most interesting finding however was, that all investigated breeds carried an activating mutations at MC1R (either E*R or E*B). Functional receptor assays of different MC1R alleles in chicken have revealed that E92K present in the MC1R E*E, E*R and E*B allele lead to an constitutively active receptor (Ling et al., 2003). The authors were not able to demonstrate the same effect for L133Q in E*R(Fay) but these experiments were conducted in mammalian cells, which might not perfectly reflect the conditions in an avian feather follicle resulting in misleading conclusions if extrapolated to birds. The experimental Fayoumi flock used for our crossing experiment has been kept and routinely monitored for over 30 years with no strong deviation or variation in the autosomal barring pattern. It is therefore likely that the two MC1R alleles (E*R(Fay) and E*R) have very similar effects on the autosomal barring pattern in this breed. Recent findings are supporting this hypothesis as they indicate that interactions between MC1R and its antagonist, ASIP, occur within the feather follicle (Lin et al., 2013) and have been found to be strongly associated with pigmentation in chicken (Takeuchi et al., 2000) and Japanese quail (Coturnix coturnix japonica) (Zhang et al., 2013) as well as in the golden winged (Vermivora chrysoptera) and blue winged warbler (Vermivora cyanoptera) (Toews et al., 2016). However, it needs to be taken into consideration that other chicken breeds carry activating MC1R variants, such as E92K, but do not exhibit autosomal barring or patterning. The mutations presented in this study might therefore not be sufficient to create feather patterning on just any genetic background and further studies are necessary to understand their actions on a molecular level.
In summary, this study suggests that MC1R has a major effect on pigment patterning both across the avian body as well as for within-feather pattern formation. We show that autosomal barring in Fayoumi is not dependent on the variant allele at Db but that this locus is contributing to a more refined phenotype.
Table of contents :
1 Introduction
1.1 Pigment patterning in birds
1.1.1 Pigment and colouration forms in birds
1.1.2 Function of pigmentation in birds
1.1.3 The bird feather
1.1.4 The feather follicle
1.1.5 Theory of natural pattern formation
1.2 Melanin-based pigmentation
1.2.1 The melanocyte
1.2.2 Melanogenesis
1.2.3 The Melanocortin 1 receptor
1.3 The domestic chicken (Gallus gallus domesticus)
1.3.1 Domestication history of the chicken
1.3.2 The chicken as a model for avian pigmentation
1.3.3 Pigment pattern genes in chickens
2 Aims of the Thesis
3 Study Summaries
3.1 Autosomal barring in chicken is strongly associated with segregation at the MC1R locus (Paper I).
3.1.1 Background
3.1.2 Results and discussion
3.1.3 Future prospects
3.2 Sex-linked barring is the result of both regulatory and missense mutations in the CDKN2A tumour suppressor gene (Paper II).
3.2.1 Background
3.2.2 Results and discussion
3.2.3 Future prospects
3.3 An inversion is associated with variant male reproductive strategies in the ruff (Philomachus pugnax) and lightly coloured ornamental feathers in the satellite morph (Paper III and IV).
3.3.1 Background
3.3.2 Results and discussion
3.3.3 Future prospects
4 Conclusion
