Differential digital display analysis performed on murine tissues provided a list of 101 potentially oocyte-specific murine genes 252627. Bioinformatic analyses were performed on this list of genes in order to find potentially oocyte-specific chicken orthologs. Genes with a blast score higher than 100 were localized using mapview 28 and blatsearch 29 tools. The syntenic regions were checked: chromosome localization of murine genes and chicken homologs were compared, in order to obtain the correct chicken ortholog of mice genes. Only genes with sufficient homology or whose localization was in accordance with syntenic regions were selected. Forty-one chicken genes were eliminated because of their poor homology with murine genes and 32 other genes were eliminated because they were localized outside the syntenic regions. Among the remaining 28 chicken genes the transcript of only 18 genes could be correctly amplified using real time or classic RT-PCR, of which the detailed study of two genes bmp15 and gdf9 has previously been reported 30. Thus 16 avian genes were finally retained (Table 1). The homology with murine genes was strong for 11 of these genes (blast score between 288 and 2149) and was weak for 5 (blast score between 104 and 132). The last five chicken genes were nevertheless considered as potential orthologs of murine genes and kept as candidate genes because of their correct localizations with respect to the syntenic region. Eleven of the selected genes were localized in the expected syntenic region (btg4, chkmos, msh4, mtprd, mcmip, znfingerRIZ, discs5, trans fact 20, wee, zar1 and ktfn). Three other genes were localized in the vicinity of the expected syntenic region (dazL, fbox and mark3). Two genes were localized in the unexpected syntenic region, but they were identified with the same name as murine genes (zpA and zpC).

The samples studied were: F1 GCs, corresponding to granulosa cells (GCs) of the largest follicles before ovulation (F1); F1 GDR, corresponding to the germinal disc region (GDR) of F1 follicles, and Ov GDR, corresponding to the germinal disc region of ovulated oocytes (Ov) (Fig. 1). Apart from the stage of maturation the main difference between these samples involved the presence of granulosa cells. They were not present in GDR from ovulated oocytes, they were slightly present in F1 GDR, and they constituted a major component of F1 GCs (see Materials and Methods). Statistical analysis of data obtained after Affymetrix microarray hybridization provided lists of genes differentially expressed in three comparisons: F1 GDR and Ov GDR, F1 GDR and F1 GCs, and Ov GDR and F1 GCs (accession number GSE7805). There were only a few differentially expressed genes in each comparison (fewer than 500 out of 28000 genes on the Chip) (Table 2). Indeed, the first comparison, between the F1 GDR and Ov GDR, showed 92 genes over-expressed at the F1 stage, including one of our in silico identified genes (zpC). In the second comparison, between F1 GDR and F1 GCs, 342 differentially expressed genes were identified. These genes involved 104 genes over-expressed in F1 GCs and 238 genes over-expressed in F1 GDR. Five of our in silico identified genes btg4, chkmos, dazL, zpA and ktfn were found among the latter genes over-expressed in F1 GDR (Table 2). The third analysis compared the expression of genes between F1 GCs and Ov GDR. We obtained a set of 448 genes that were differentially expressed between F1 GCs and Ov GDR, of which 392 genes were over-expressed in F1 GCs and 56 genes were over-expressed in Ov GDR. We found 1 of our in silico identified genes among these genes, (btg4) (Table 2). The Venn diagram (Fig. 2) shows overlapping differentially expressed genes between different dataset comparisons. Eighty-five differentially expressed genes were common for two comparisons (F1 GCs and Ov GDR, and F1 GCs and F1 GDR). Only one gene was common for the comparisons between F1 GDR and F1 GCs, and F1 GDR and Ov GDR and 1 gene was common for all three comparisons. F1 GDR samples contained a quantity of granulosa cells in contrast to Ov GDR samples which were free of granulosa cells. Analysis of the redundancy of over-expressed genes between different comparisons was therefore performed in order to distinguish between genes found over-expressed in F1 GDR samples due to granulosa cell contamination and those really over-expressed in the oocyte (Fig. 3). This analysis revealed that 85 of 92 genes over-expressed in F1 GDR were also over-expressed in F1 GCs, both compared to Ov GDR, indicating that these genes characterized granulosa cell expression rather than variation in the expression between F1 GD and Ov GD. Thus only 7 genes should be considered as overexpressed in F1 GD compared to Ov GD. Moreover, of the 104 genes overexpressed in F1 GCs compared to F1 GDR, 36 were also overexpressed in F1 GCs compared to Ov GDR. Consequently, taking into account the latter redundancy, a total of 460 genes was overexpressed in granulosa cells compared to oocytes. On the other hand 49 genes were redundant among genes overexpressed in the Ov GDR and F1 GCs, and F1 GDR and F1 GCs comparisons.

On the basis of the in silico and microarray approaches, 17 genes were retained for further study. Among these, 16 genes were found using the in silico approach as described above (Table 1), of which 6 genes (chkmos, dazL, btg4, zpA, ktfn and zpC) were also found using the microarray approache (Table 2) and 1 gene (zpD) was found only in the microarray approach. The latter gene was chosen because of its strong involvement in the fertilization process 31. In contrast to zpA and zpC, the zpD gene was not found in the mouse 32, explaining why it was absent from the list of murine potentially oocyte-specific genes. Four other genes already known for their involvement in final follicular maturation or early embryo development in the chicken or in other vertebrates were added (foxL2 33, igf2 3435, hsf1 36, and cvh 23) (Table 3). Finally 21 chicken genes were further studied. Real time RT-PCR performed on 11 adult tissue samples (total ovary, spleen, intestine, gizzard, liver, heart, skin, brain, pectoralis muscle, lung and pituitary gland) revealed differences in the specificity of their tissular expression patterns. Tissular expression profiles of some of these genes are presented on Fig. 4. Seven of these genes (dazL, wee, zar1, zpA, btg4, zpC and chkmos) were specifically expressed in the ovary (Fig 5A). The specificity of zpC and chkmos has previously been described and our results concerning these genes (data are not shown) were in accordance with the literature (37 and 38, respectively). Eight genes were preferentially and strongly expressed in the ovary (Fig. 5B). Three genes from last group were slightly expressed in another tissue: fbox and zpD in the pituitary gland and ktfn in the muscle. We also confirmed preferential ovarian expression of cvh 23 and foxL2 33, as well as a low expression of the latter in the pituitary gland (data not shown). Three other genes from this group (igf2, mark3 and znfingerRIZ) were slightly expressed in other tissues, in addition to the ovary. The last 6 genes (trans fact 20, msh4, mtprd, mcmip, discs5 and hsf1) were expressed as highly in the ovary as in other tissues (data not shown).

Analysis of total RNA, assessed with Agilent RNA nano chips (Fig. 6A), showed an atypical state of RNA from GDR between ovulation and oviposition. In our conditions, rRNA seemed degraded from ovulation until oviposition. With three different RNA extraction methods (tri reagent (Euromedex), RNeasy kit (QIAGEN) and MasterPure™ RNA Purification Kit (Epicentre Biotechnologies)), the RNA profile of GDR from the ovulation stage and from the following embryonic stages remained degraded (data not shown). Moreover, we assessed rRNA 18S and 28S by real time PCR in order to confirm this atypical RNA state (Fig. 6B and 6C). These results confirmed our previous observations; 18S and 28S rRNA expression showed a huge decrease from ovulation untill the oviposition. We then performed labelled reverse transcription to investigate whether mRNA was also degraded. We compared three samples, i.e. the RNA from GDR of F1 stage, from a whole adult ovary that had a normal rRNA profile and GDR of ovulation stage with had degraded rRNA. Electrophoresis profiles of labelled cDNA in denaturing agarose gel were almost identical for the three samples investigated and the smear corresponding to the reverse transcribed mRNA was still present in all these samples (Fig. 6D).

A detailed study of the expression profiles during follicular maturation and early embryo development using real time RT-PCR was performed for 21 genes, for which tissue specificity of the expression was characterized. First unsupervised hierarchical clustering of our data was performed in order to confirm the biological appropriateness of the selected genes and samples (Fig 7). Two other previously studied genes (gdf9 and bmp15) were added to this analysis in order to facilitate the clustering process because of their already known oocyte-preferential localization 30. This analysis discriminated two major groups of samples. The first included all granulosa samples from F6 to F1 (correlation threshold 0.58). The second group corresponded to GDR from F6 to the ovulation stage and to all embryo stages (correlation threshold 0.60). Several subgroups could be distinguished within each group (correlation threshold 0.69–0.94). Each subgroup corresponded to a different physiological state, indicated on Fig 6A, suggesting that the genes selected were pertinent. Moreover, unsupervised clustering arranged samples in the perfect chronological order.

In order to group genes with similar expression patterns supervised clustering was performed with samples arranged first according to the nature of the sample (GCs or GDRs) and then according to chronological order (Fig. 8). Five clusters (threshold > 0.64) were identified (C1 to C5). A representative example of the expression profile characterizing each cluster is shown on Fig. 9. The graphical representation of expression profiles of all other genes with the corresponding statistical analysis of variation of their expression is supplied in additional files (see Additional file 1, 2, 3 and 4). The chkmos, btg4, wee, zpA, dazL, hsf1, fbox and bmp15 genes, forming cluster C1, were preferentially expressed in GDR. These genes showed a significant increase in mRNA expression in GDR during follicular maturation and a steady decrease after ovulation, becoming nearly undetectable at 36 h after fertilization. The cvh, ktfn and zar1 genes included in Cluster C2 were also preferentially expressed in GDR but, in contrast to genes from cluster C1, they were constantly expressed in the early embryo. Moreover, ktfn and zar1 genes displayed a significant increase in expression from 24 to 48 hours after fertilization. The genes from cluster C3 were expressed in GCs, in GDRs and in the embryo. Moreover, mark3, igf2, gdf9 and transfact20 genes showed a significant decrease in expression during the last stages of follicular maturation in both GDR and GCs. Cluster C4, including discs5, znfingerRIZ, foxL2 and mtprd genes, was characterized by expression that was fairly similar to that of cluster C3. The difference consisted of less pronounced variations in gene expression in GCs and more pronounced variations in gene expression in GDR for cluster C4 compared to genes in cluster C3 through the stages investigated. Genes belonging to cluster C5 were expressed in GCs and GDR but, in contrast to the clusters C3 and C4, their expression dropped dramatically at ovulation, especially for zpC and zpD whose transcripts showed increasing expression in GCs during the last stages of follicular maturation and a less marked increase in the expression in GDR. In contrast, the levels of mcmip decreased progressively during the same period. The last gene (msh4) was not clustered. Its expression profile was fairly similar to that of genes belonging to cluster C3, but its expression increased significantly 36 h after fertilization.

We performed in situ hybridization on mature and immature ovarian sections. All the probes were assessed on the two stages, but only the most significant results are shown on Fig. 10. For all the genes studied we detected homogeneous signals of mRNA expression in oocytes similar to those we have previously described for bmp15 and gdf9 30. The mRNA of four of eight genes localized by in situ hybridization (btg4, dazL, cvh and fbox) were detected with a high intensity signal in the oocytes from follicles of 300 μm – 600 μm from both immature (Fig. 10A,B,C and 10D) and mature (data not shown) ovaries. No significant expression was detected in somatic cells. For four other genes (chkmos, zpA, zpC and zpD) a signal was found in oocytes from the largest follicles of 500 μm-6 mm of mature ovaries (Fig. 10E,F,G and 10H). The signal was particularly high for chkmos and zpA (Fig. 10E and 10F), whereas zpC and especially zpD had weaker signals in both oocyte and somatic cells (Fig. 10G and 10H). No significant signal was detected with the corresponding sense probes.

Laying breed hens aged 60–70 weeks (ISA Brown, egg layer type, Institut de Selection Animale, Saint Brieuc, France) were housed individually in laying batteries with free access to feed and water and were exposed to a 15L:9D photoperiod, with lights-on at 8.00 pm. Individual laying patterns were monitored daily. For in situ hybridization, these hens and younger ones of (10 weeks old) were used to provide mature and immature ovaries, in order to study follicles at each stage. Hens used to provide fertilized eggs were bred in the same conditions and inseminated once a week.

Hens aged 60 weeks were used. Tissue samples were collected from the ovary, spleen, intestine, gizzard, liver, heart, skin, brain, pectoralis muscle, lung and pituitary gland. Germinal disc regions (GDR) and granulosa cells (GCs) surrounding the GDR (Fig. 1) were collected from different preovulatory follicles (F1 to F6), just ovulated oocytes and early embryos at 6.5 h, 12 h, 24 h, 36 h and 48 h post ovulation. The GDR and GCs surrounding the GDR were carefully dissected in the same way under a binocular microscope using fine forceps and scissors (World Precision Instruments) as previously described for quail oocytes 60. After washing in phosphate buffer saline (PBS, Gibco, Cergy Pontoise, France) GDR and GCs were frozen in liquid nitrogen and then were stored at -80°C until use. For the last two stages, eggs were incubated at 37.8°C for 12 and 24 h, respectively. The 6.5 h, 12 h, and 24 h stages of embryo development correspond to stages I, V and X of the Eyal-Giladi and Kochav classification, respectively 61. The 36 h and 48 h stages correspond to stages 3 and 6 of the Hamburger-Hamilton classification, respectively 62. During follicular maturation the germinal disc is closely associated with its overlying granulosa cells (GCs) and forms a structure called the germinal disc region (GDR) (Figure 1). The GDR from F6-F1 follicles used for these studies consisted not only of the germinal disc but also of the overlying layer of GCs, because GD and overlying GCs cannot be completely separated 606364 and the number of GCs in GDR preparations could not be counted.

A differential digital display analysis has already been performed with mouse ESTs 252627, providing a list of murine oocyte-specific genes. Using this murine gene list, we systematically searched for chicken orthologs of these genes in international public databases pubmed 65. Blast bit scores higher than 100 were retained. Moreover, the physical localization of genes identified on chicken chromosomes was retrieved from both mapview 28 and from blat search 29. We also verified that chicken homologs were localized in the syntenic genomic regions conserved with that of mouse species to have a better chance that true orthologs were studied with ensembl 66.

Total RNA was extracted from GDR of F1 and ovulated oocytes, and from GCs of F1 follicles as described above. We thus had 3 samples with a biological replicate of each sample. The RNeasy Mini Kit (QIAGEN, Hilden, Germany) was used according to the manufacturer's instructions. The tissues (GDR or GC) from 25 hens were pooled for each stage investigated to obtain enough RNA for probe synthesis. Two such a pools were constituted for each sample in order to achieve two biological replicates of microarray hybridization. All RNA samples were checked for their integrity on the Agilent 2100 bioanalyzer according to Agilent Technologies guidelines (Waldbroon, Germany). Two micrograms of total RNA were reverse transcribed with the One-cycle cDNA synthesis kit (Affymetrix, Santa Clara, CA), according to the manufacturer's procedure. Clean up of the double-stranded cDNA was performed with Sample Cleanup Module (Affymetrix, Santa Clara, CA) followed by in vitro transcription (IVT) in the presence of biotin-labelled UTP using GeneChip^® IVT labelling Kit (Affymetrix). The quantity of the cRNA labelled with RiboGreen^® RNA Quantification Reagent (Turner Biosystems, Sunnyvale, CA) was determined after cleanup by the Sample Cleanup Module (Affymetrix). Fragmentation of 15 μg of labelled-cRNA was carried out for 35 minutes at 94°C, followed by hybridization for 16 hours at 45°C to Affymetrix GeneChip^® Chicken Genome Array, representing approximately 32,773 transcripts, corresponding to over different 28,000 Gallus gallus genes. After hybridization, the arrays were washed with 2 different buffers (stringent: 6X SSPE, 0.01% Tween-20 and non-stringent: 100 mM MES, 0.1 M [Na+], 0.01% Tween-20) and stained with a complex solution including Streptavidin R-Phycoerythrin conjugate (Invitrogen/molecular probes, Carlsbad, CA) and anti Streptavidin biotinylated antibody (Vectors laboratories, Burlingame, CA). The washing and staining steps were performed in a GeneChip^® Fluidics Station 450 (Affymetrix). The Affymetrix GeneChip^® Chicken Genome Arrays were finally scanned with the GeneChip^® Scanner 3000 7G piloted by the GeneChip^® Operating Software (GCOS).

All these steps were performed on Affymetrix equipement at INRA-URGV, Evry, France.

The .raw CEL files were imported in R software for data analysis. All raw and normalized data are available from the Gene Expression Omnibus (GEO) repository at the National Center for Biotechnology Information (NCBI) 67, accession number GSE7805. Gene Ontology annotations were performed with NetAffx.

Total RNA was extracted from whole adult tissues (ovary, spleen, intestine, gizzard, liver, heart, skin, brain, pectoralis muscle, lung and pituitary gland) using Tri-reagent (Euromedex, Mundolsheim, France) according to the manufacturer's procedure. RNA quality and quantity were then assessed by using RNA nano chips and Agilent RNA 6000 nano reagents (Agilent Technologies, Waldbronn, Germany) according to the manufacturer's instructions. Samples were stored at -80°C until use. Reverse transcription (RT) was performed to test the expression of candidate genes in different tissues and at different stages of follicular maturation and embryo development using polymerase chain reactions (PCR). One microgram of total RNA extracted from tissues or GDR was digested by RQ1 DNase (Promega, Madison, WI, USA) and reverse transcribed to first-strand cDNA using Moloney Murine Leukemia Virus reverse transcriptase I with an oligo dT-random primer mix (Promega, Madison, WI, USA) according manufacturer's instructions.

Labelled RT was performed in order to assess mRNA quality in GDR of F1 stage and just ovulated oocytes. Ten μCi αP32 dCTP was added to the reverse transcription mix in order to label cDNA. Labelled cDNA was then separated on 1.2% denaturing agarose gel (50 mM NAOH, 1 mM EDTA) by electrophoresis. A storage phosphor screen (Amersham Biosciences, Bucks, UK) was placed on the gel in an exposure cassette (Amersham Biosciences, Bucks, UK). The signal was detected one hour later with a STORM 840 (Molecular Dynamics), a phosphor screen imaging system.

Specific sets of primer pairs (Sigma Genosis), designed using Vecteur NTI software to amplify fragments of 21 different transcripts, are shown in Table 4. Real time PCR reactions were carried out in 25 μl containing primers at a final concentration of 150 nm of each, 5 μl of the RT reaction diluted 1/30 and qPCR Mastermix Plus for Sybr Green I (Eurogentec) according to the manufacturer's instructions. Real time PCR was performed using an ABI Prism 7000 (Applied Biosystems). After incubation at 50°C for 2 min, and 95°C for 10 min, the thermal cycling protocol was as follows: 40 cycles at 95°C for 15 sec and 60°C for 1 min. The amounts of 18S rRNA and 28S rRNA were measured in the RT reactions diluted 1/5000 using 28S rRNA and 18S rRNA control kits (Eurogentec), respectively, according to the manufacturer's instructions. PCR amplification without cDNA was performed systematically as a negative control. 18S rRNA was used as a reporter gene in the study of the mRNA tissular expression pattern, whereas β actin,gapdh and ef1α were used as reporter genes in the mRNA expression study during follicular maturation and early embryo development, the expression of these three genes being similar in GCs and in GDR (data not shown). In both cases (tissular expression study or temporal dynamic study) the samples were analyzed in duplicate on the same plate for a given gene. Real time RT-PCR was performed in the temporal dynamic study with two biological replicates of each sample. Melting curve analysis was systematically performed for all genes in order to verify the specificity of the PCR product. Real time PCR efficiency (E) was measured in duplicate on serial dilutions of cDNA (pool of reverse-transcribed RNA from GDR and GC samples at the same stages as used for the temporal dynamic study) for each primer pair and was calculated using the following equation: E = (10^1/slope)^-1. In the tissular expression study the relative amounts of gene transcripts (R) were calculated according to the equation:

R = E gene − Ct gene E 18 S rRNA − Ct 18 S rRNA MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xI8qiVKYPFjYdHaVhbbf9v8qqaqFr0xc9vqFj0dXdbba91qpepeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGaeeOuaiLaeyypa0tcfa4aaSaaaeaacqqGfbqrdaqhaaqaaiabbEgaNjabbwgaLjabb6gaUjabbwgaLbqaaiabgkHiTiabboeadjabbsha0naaBaaabaGaee4zaCMaeeyzauMaeeOBa4MaeeyzaugabeaaaaaabaGaeeyrau0aa0baaeaacqaIXaqmcqaI4aaocqqGtbWucqqGGaaicqqGYbGCcqqGsbGucqqGobGtcqqGbbqqaeaacqGHsislcqqGdbWqcqqG0baDdaWgaaqaaiabigdaXiabiIda4iabbofatjabbccaGiabbkhaYjabbkfasjabb6eaojabbgeabbqabaaaaaaaaaa@5468@

where Ct is a cycle threshold. In the temporal dynamic study R was calculated using following equation:

R = E gene − Ct gene Mean ( Rc ef 1 α , Rc β actin , Rc gapdh ) MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xI8qiVKYPFjYdHaVhbbf9v8qqaqFr0xc9vqFj0dXdbba91qpepeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGaeeOuaiLaeyypa0tcfa4aaSaaaeaacqqGfbqrdaqhaaqaaiabbEgaNjabbwgaLjabb6gaUjabbwgaLbqaaiabgkHiTiabboeadjabbsha0naaBaaabaGaee4zaCMaeeyzauMaeeOBa4MaeeyzaugabeaaaaaabaGaeeyta0KaeeyzauMaeeyyaeMaeeOBa4MaeiikaGIaeeOuaiLaee4yam2aaSbaaeaacqqGLbqzcqqGMbGzcqaIXaqmcqaHXoqyaeqaaiabcYcaSiabbkfasjabbogaJnaaBaaabaGaeqOSdiMaeeyyaeMaee4yamMaeeiDaqNaeeyAaKMaeeOBa4gabeaacqGGSaalcqqGsbGucqqGJbWydaWgaaqaaiabbEgaNjabbggaHjabbchaWjabbsgaKjabbIgaObqabaGaeiykaKcaaaaa@62FE@

where Rc is corrected relative reporter gene expression calculated as explained below. In order to take into account only fold changes in the expression levels of reporter genes between samples but not the differences in the expression levels of reporter genes in the sample, the expression levels of reporter genes in each sample (sample_i) was adjusted against the relative amount of ef1α in the F1GDR sample according to the equation:

Rc reporter = E e f 1 α − Ct e f 1 α  in F1GDR × E reporter Δ Ct reporter MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xI8qiVKYPFjYdHaVhbbf9v8qqaqFr0xc9vqFj0dXdbba91qpepeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGaeeOuaiLaee4yam2aaSbaaSqaaiabbkhaYjabbwgaLjabbchaWjabb+gaVjabbkhaYjabbsha0jabbwgaLjabbkhaYbqabaGccqGH9aqpcqqGfbqrdaqhaaWcbaGaemyzauMaemOzayMaeGymaeJaeqySdegabaGaeyOeI0Iaee4qamKaeeiDaq3aaSbaaWqaaiabdwgaLjabdAgaMjabigdaXiabeg7aHjabbccaGiabbMgaPjabb6gaUjabbccaGiabbAeagjabbgdaXiabbEeahjabbseaejabbkfasbqabaaaaOGaey41aqRaeeyrau0aa0baaSqaaiabbkhaYjabbwgaLjabbchaWjabb+gaVjabbkhaYjabbsha0jabbwgaLjabbkhaYbqaaiabfs5aejabboeadjabbsha0naaBaaameaacqqGYbGCcqqGLbqzcqqGWbaCcqqGVbWBcqqGYbGCcqqG0baDcqqGLbqzcqqGYbGCaeqaaaaaaaa@7176@

where ΔCtreporter=Ctreporter in F1GDR−Ctreporter in samplei MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=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@7032@.

The hierarchical classification of data obtained using real time RT-PCR was performed with the Cluster 3.0 program using unsupervised single linkage or supervised complete linkage clustering in order to classify biological samples or to group together genes with a similar expression pattern, respectively 68.

Female chickens were sacrificed at different stages of sexual development. Two types of tissue were used, i.e. mature ovaries, containing follicles of different sizes (50 μm-7 mm) from 60-week-old hens (most follicles being larger than 300 μm) and immature ovaries, containing a majority of small follicles (25–500 μm) from 10-week-old hens (most follicles being smaller than 100 μm). Mature and immature ovaries were then collected and included in Tissue-Tek (Sakura Finetek Europe BV, Zoeterwoude, The Netherlands). Frozen ovaries were serially sectioned with a cryostat (thickness 10 μm) to perform in situ hybridization experiments using ³⁵S-labeled chicken gene cRNA. The gene antisense and sense constructs used for in situ hybridization were generated by inserting 700 – 800 bp fragments of chicken gene cDNA into the pGEM-T vector (Promega, Madison, WI, USA), and selecting a clone with the appropriate antisense or sense orientation. The gene cDNA fragments were generated by RT-PCR from chicken ovary mRNA using forward and reverse primers (Table 4). The in situ hybridization was performed as previously described 69. Hybridization specificity was assessed by comparing signals obtained with the cRNA antisense probe and the corresponding cRNA sense probe.

Data obtained after Affymetrix microarray hybridization analysis were normalized with the gcrma algorithm 70, available in the Bioconductor package 71. Differential analysis was performed with the varmixt package of R 72. A double-sided, unpaired t-test was computed for each gene between the two conditions. Variance of the difference in gene expression was split between subgroups of genes with homogeneous variance 72. The raw P values were adjusted by the Bonferroni method, which controls the Family Wise Error Rate (FWER) 73. A gene is declared differentially expressed if the Bonferroni-corrected P-Value is less than 0.05.

All other experimental data are presented as means ± SEM. One-way analysis of variance (ANOVA) was used to test differences. If ANOVA revealed significant effects, the means were compared by Fisher's test, with P < 0.05 considered significant. Different letters indicate significant differences.