Copyright by the Brazilian Society of Genetics. Printed in Brazil

Brazilian poultry industry has reached a high level of development in both meat and egg production as a result of constant technological modernization. Further improvements can be achieved through genomics, but before this can be accomplished, a better understanding of gene expression profiles and nucleotide polymorphisms is necessary. Since animal physiology is directly or indirectly controlled by the pituitary and hypothalamus, the aim of the present work was to identify and analyze genes expressed in these tissues in chicken lines with different growth potential. Two pituitary and hypothalamus cDNA libraries from 21 day broiler (TT) and layer (CC) chickens lines were constructed and allowed identification of 3,074 unique sequences and 77 single nucleotide polymorphisms (SNPs). The collection of expressed sequence tags (ESTs) and SNPs identified in this study represents an important resource for future studies aimed at identifying genes responsible for growth in chicken.


Introduction
Over the last decades, the poultry industry has experienced a substantial increase in production efficiency.For example, there has been a threefold increase in egg production per chicken/year and a substantial decrease in the time necessary for broilers to reach 1.5 kg of live weight (Burt, 2002).Classic selection greatly contributed to this progress, since selected broiler lines grow 3 to 4 times more rapidly than their non-selected ancestor, red jungle fowl (Bulfield, 2004).Nevertheless, several phenotypic traits are difficult to improve through traditional breeding, such as those that are difficult or expensive to measure (carcass quality and composition, behavior and welfare) or have low heritability (reproduction and fitness) (Bulfield, 2004).In addition, unwanted characteristics can be indirectly selected with classic breeding (Burt, 2002).Modern genomic technologies can greatly impact selection of these difficult to target traits (Bulfield, 2004).
Important aspects of animal physiology are directly or indirectly controlled by the pituitary and hypothalamus, but the genetic mechanisms controlling processes such as metabolism, somatic growth and reproduction in chickens still remain largely unknown (Cogburn et al., 2003).Therefore, identification and study of genes expressed in the pituitary and hypothalamus can fill existing gaps in understanding the molecular pathways involved in several physiological processes, as well as provide tools for future animal breeding programs.
Among the methodologies available for gene identification, analysis of expressed sequence tags (ESTs) has proven to be very efficient.This methodology consists in partially sequencing the extremities of clones obtained from cDNA libraries and establishing groups of specifically expressed genes, as well as their transcription levels in determined tissue or cell types (Adams et al., 1991).This approach allows comparisons to be made between different tissues or species, and the identification of polymorphisms in intragenic regions (Adams et al., 1991;Hatey et al., 1998).Single nucleotide polymorphisms (SNPs) have emerged as a principal DNA marker class which greatly helps in developing high-density genetic maps for use in QTL identification through linkage disequilibrium analysis (Smith et al., 2002).
In previous studies, ESTs were generated from skeletal muscle precursor tissues (somites and neural tube), limbs and whole embryos (Jorge et al., 2004), and from young fowl pectoral musculature (unpublished results).The present work was developed with the scope of cataloguing genes expressed in chicken hypothalamus and pituitary gland and identifying distinct features possibly associated with growth.

Material and Methods
A total of about 120 eggs from broiler (TT) and layer (CC) lines supplied by Embrapa Swine and Poultry National Research Center were incubated at 37 °C in a humidity-controlled atmosphere.The TT line is a male line obtained from a cross of Cornish, Hampshire and Plymouth Rock breeds.This line has been selected for meat production since 1985.In the first stages of the breeding program the focus was upon weight gain and carcass traits; however, since 1992, males of this lineage have also been selected for feed conversion rate.CC line is a female line of White Leghorn selected initially (1989) for egg production and quality.After hatching, chicks were kept in a commercial broiler house at the Animal Science Department, ESALQ-USP.Chicks were initially given commercial broiler feed and water ad libitum.They were exposed to room temperature and continuous luminosity until 21 days of age.At this age, the pituitary and hypothalamus were extracted surgically and stored in liquid nitrogen.Total RNA was extracted from each line separately according to the protocol described by Chomczynski and Sacchi (1987), followed by poly(A)+RNA isolation using the Oligotex kit (GE HealthCare).cDNA libraries were constructed from 1-2 μg poly(A)+ RNA using the SuperScript Plasmid System kit (Invitrogen), according to manufacturer's protocol.Fractions containing cDNA larger than 500 bp were ligated into the SalI-NotI site of pSPORT1 vector (Invitrogen).Clones were sequenced from the 5'ends using the Big Dye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems) in conjunction with primer T7 (5'-TAATACGA CTCATATAGGG-3').
Comparative analysis of expression patterns was performed between the two lines and all other libraries constructed in the laboratory, according to the relative frequency of the ESTs.SNPs were identified analyzing nucleotide discrepancies between ESTs from the two lines.Only those SNPs that appeared at least twice in the same line/po-sition and showed PHRED quality equal or higher than 20 were considered as hypothetic line-specific SNPs.

Results and Discussion
Sequence analysis A total of 5,017 ESTs were obtained from the 5'end of the inserts cloned in the two cDNA libraries.By PHRED analysis, 2,133 reads from broiler line (TT) and 2,153 of the layer line (CC) were considered valid.Sequences referring to the CC line library were deposited in dbEST division of GenBank with accession numbers ranging from CO419474 to CO421626, and those referring to the TT line library received numbers ranging from CO421627 to CO423759.After clustering and assembly (CAP3 software, Huang and Madan, 1999), the TT line library presented 1,643 unique sequences (contigs + singletons).Of these, 1,477 were singletons and 656 were grouped into 166 contigs, whose sequence number varied from 2 to 59.The CC line library represented 1,649 unique sequences, with 1,475 singletons and 678 sequences grouped into 174 contigs, with 2 to 70 sequences per contig.This clustering indicated novelty rates of 77% for the TT library and 76.6% for the CC line library.Both library sequences were also clustered together, revealing 3,074 unique sequences and a novelty index of 71.1%.

Expression profile ("digital Northern")
Relative frequency of ESTs was compared using all G. gallus sequences obtained in the laboratory (a total of 13,521 ESTs), in a strategy known as "digital Northern" (Audic and Claverie, 1997).The clustering and assembly of all 13,521 ESTs resulted in 680 contigs formed by grouping TT and CC ESTs.Sequences present in 472 (69.4%) of these contigs were also identified in libraries constructed from somites, limbs (Jorge et al., 2004) and breast muscle, suggesting that these ESTs are coordinately expressed in all the different tissues studied.Sequences present in 94 contigs (13.8%) were identified as unique to the pituitary and hypothalamus libraries in both TT and CC lines, and were called library-specific contigs.The 114 remaining contigs (16.8%) were library and line-specific, since they were only encountered in the pituitary and hypothalamus libraries of the lines studied.
Among the ESTs identified as library and library/line specific, sequences coding for proteins known to participate in molecular pathways for growth and reproduction were identified.Some examples of genes represented preferentially in the CC line were Ca ++ /Calmodulin-dependent protein kinases proteins (CAMK1 and CAMK2), G protein-coupled receptors (GPRs), cAMP phosphodiesterase (LOC771318, LOC425199), phosphoinositide 3-kinase (PIK3), and N-myc downstream-regulated gene 1 protein (NDR1).

SNPs identification
Following clustering and assembly of the sequences from the pituitary and hypothalamus libraries with CAP3 software, ESTs grouped in the same contig were used to search for single nucleotide polymorphisms.Only those SNPs that appeared at least twice in the same line/position and showed PHRED quality equal or higher than 20 were considered as hypothetic line-specific SNPs.
Of the 389 contigs constituted by sequences from the pituitary and hypothalamus libraries, 28 (7.2%) presented 77 line-specific SNPs, corresponding to 52 TT-specific and 25 CC-specific SNPs (Table1).Most SNPs found in ESTs were related to the mitochondrial genome, to structural proteins, neuronal constituents, ribosomal proteins and iron binding proteins.SNPs were also observed in hypothetical proteins (proteins still lacking a defined biological function), calcium binding proteins, lipid metabolism related proteins and ESTs lacking similarity to any sequence in the database consulted.
PHRED quality of the SNP nucleotides varied from 20 to 68, with an average value of 50.1.Among the 28 contigs, 14 presented only one SNP, six presented two SNPs, three showed three SNPs, one contig presented five SNPs, one had seven SNPs, two contigs showed eight SNPs and one presented 14 SNPs.These 28 contigs were constituted by four to 112 ESTs each, presenting an average number of 21.2 ESTs per contig and a mean length of 1,441 bp.Minimum length was 746 bp and maximum length was 3,006 bp.Contig SNP density varied from 0.5 to 9.1 SNPs per kb, with an average density of 1.9 SNP/kb.This result is within the estimated polymorphism range observed by Smith et al. (2002).TT-specific SNPs showed greater density than the CC-specific ones.Average density values were 1.3 and 0.6, respectively.This two times higher density value for TT-specific SNPs is interesting since the total ESTs and putative line-specific SNPs 1079  61 (q: 43, 30, 37, 35) 62 (q: 22, 43, 37, 42, 42) 63 (q: 27, 53, 44, 42, 43) 64 (q: 54, 53, 53, 47, 50) 67 (q: 57, 57, 53, 57)
The EST collection obtained in this study allowed the identification of genes expressed in the pituitary and hypothalamus in two commercial chicken lines, thus providing an important resource for studies of growth physiology and animal breeding.In addition, we identified a series of linespecific SNPs.The SNPs identified in genes specifically expressed in this major control center of animal physiology hold great potential for selection studies, since SNPs are the most frequent form of genome variation and are currently considered a new generation of molecular markers.

Table 1 -
Relation of putative line specific SNPs found in 28 contigs.