Methods

Mouse strains and husbandry
CAST/EiJ (stock #000928) mice were purchased from the Jackson Laboratory. The hg mutation was originally identified in a selected outcross line [22] and has been introgressed onto a B6 background via nine backcrosses to create the HG strain. HG mice used in this experiment were from the 17th or later generation of inbreeding. Mice were housed in polycarbonate cages under controlled conditions of temperature (21°C ± 2°C), humidity (40–70%) and lighting (14 h light, 10 h dark, lights on at 7 AM), and managed according to the guidelines of the American Association for Accreditation of Laboratory Animal Care (AAALAC).

Genotyping
DNA for genotyping was isolated from 1.0–2.0 mm tail clips by digesting with Proteinase K (Fisher) at 55°C in a buffer composed of 0.45% NP40 (Sigma), 0.45% Tween 20 (Fisher) and 1X PCR buffer (Promega). The product of this digestion was diluted (1:10) in sterile H2O and used for genotyping without further purification. Microsatellite genotyping was performed using standard PCR and gel electrophoresis protocols. Reaction conditions for each marker are listed in Additional File 1 and 2.
MMU2 congenic mice were genotyped for hg using a two primer genotyping assay. One primer set (HG-F, ctcctgtctgggctgtgag and HG-R, caaaggcagaagtggggtaa) spanned the hg deletion producing a 447 bp product in hg/hg and +/hg mice. The other set (CRADD3a.F, gtccatcagcattcctgaaa and CRADD3.R, tgtccagcaacagcattgtc) amplified a 232 bp Cradd amplicon (located within the hg deletion) in +/+ and +/hg mice [54]. The PCR annealing temperature was 55°C and the MgCl2 concentration was 1.5 mM.

Development of B6.CAST and HG.CAST MMU2 speed congenic strains
All speed congenic strains were developed starting with an initial cross between a CAST male and HG females (Figure 1) [5,6]. Male F1 mice were then backcrossed to HG females. All agouti (the dominant nonagouti (a) locus is located at 154.8 Mbp on MMU2) N2 males were genotyped for 79 microsatellite markers (Additional File 1). These markers were evenly spaced across the genome, except in regions previously identified as harboring QTL [1], which were more densely screened. The "best" N2 agouti male with the lowest level of genome-wide unwanted heterozygosity while maintaining CAST alleles for all MMU2 markers was selected for breeding. This selection scheme was used at each generation until a N4 male was identified as homozygous HG for all markers typed outside MMU2. After an additional backcross to HG females, recombinant males were identified providing the foundation for the four overlapping donor regions. Selected recombinant males were then backcrossed to both B6 and HG females to create strains, which were B6 (+/+) or HG (hg/hg) and heterozygous congenic. These mice were intermated to produce homozygous founders for each strain. This novel breeding scheme created four identical founder congenics on two backgrounds B6 (+/+) and HG (hg/hg), which formed the basis for our examination of interactions caused by the presence of the hg deletion. MMU2 speed congenic strains were maintained through brother-sister mating. Once each congenic was stabilized, 20 additional microsatellite markers were used to refine the position of each congenic recombinant end point (Additional File 2).

Development of HG.CAST speed congenic strains for MMU1, 5, 8, 9, 11 and 17
All black N2 males from the first two crosses described above were genotyped for 12 markers (D1Mit432, -480, D5Mit353, -311, D9Mit60, -262, D11Mit5, -67, D8Mit234, -211 and D17Mit28 and -142), two spanning each of the six QTL harboring regions (MMU1, 5, 8, 9, 11 and 17; Table 1 and Additional File 1). Markers were selected to capture, at a minimum, the 2-LOD support interval. Two N2 males were selected to propagate the N3 generation; one heterozygous for QTL on MMU1 and 9 and the other heterozygous for QTL on MMU5, 8, 11 and 17 (Figure 1). Both males were homozygous for HG alleles at all other known QTL. These males were backcrossed to HG females and two of the resulting N3 males inheriting the same sets of QTL as their sire were selected for breeding. These males were subsequently backcrossed to HG females and three N4 males were identified heterozygous for the following regions: 1) MMU1 and 9; 2) MMU5 and 11; 3) MMU8, 11 and 17 (Figure 1). Starting at N4 and continuing through N6, the "best" male with the lowest percent of unwanted donor alleles was selected after performing a genome scan using the remaining 67 genome-wide markers (79 total markers minus the 12 markers genotyped in the first two backcrosses spanning the know QTL intervals). At N5 a distinct strain was created for each of the six individual donor regions and heterozygous mice were intermated (Figure 1). Homozygous HG.CAST speed congenic strains were maintained through brother-sister mating. Once each congenic was stabilized, 19 additional microsatellite markers were used to refine the position of each congenic recombinant end point (Additional File 2).

Development of B6.CASTC and HG.CASTC control strains
HG is a strain in which the hg deletion has been introgressed onto a B6 background, therefore the only genetic differences between the strains would be the hg locus, tightly linked alleles from the outbred strain on which hg arose and contaminating alleles remaining after the nine backcrosses and fixed during inbreeding. Instead of using parental B6 and HG strains as controls for phenotypic comparisons with each speed congenic, we choose to develop independent control strains originating from the same cross as the congenic panels. Separate B6.CAST control (B6C) and HG.CAST control (HGC) strains were developed using mice from the MMU2 experiment. Mice from the last backcross inheriting only B6 or HG MMU2 alleles at markers spanning MMU2 were intermated to serve as the basis for each control. Both control strains were subsequently maintained through brother-sister mating.
The control strains were coisogenic with the parental B6 or HG strain with the exception of mutations that arose during congenic construction and a small percentage of contaminating donor alleles missed after 6 backcrosses. Therefore, since the congenics and controls were developed through the same selection scheme and possibly share common contaminating regions, the B6C and HGC strains are the most ideal control to compare each congenic.

Phenotypic characterization
Trait data were collected on approximately 40 mice (20 of each sex) from each congenic and control strain. To eliminate parity and reduce litter size effects only progeny from uniparous dams were characterized and all litters were standardized to 5–7 pups/litter. Mice were weaned at 3 weeks of age. Feed (Purina 5008; 23.5% protein, 6.5% fat, 3.3 Kcal/g) and water were offered ad libitum. Mice were weighed to the nearest 0.1 g at 2WK, 3WK, 6WK, and 9WK of age. At 9WK ( ± 5 days) mice were anesthetized under isoflurane and nasal-anal length (NA) and nasal-tail (NT) were measured to the nearest mm. Tail length was calculated as NA minus NT. Anesthetized mice were then sacrificed by decapitation and exsanguinated. Femoral fat pad (FFP), gonadal fat pad (GFP), mesenteric fat pad (MFP) and retroperitoneal fat pad (RFP) were removed and weighed to the nearest mg.
Chemical compositional analysis was performed for HGC, HG11 and HG17 carcasses as previously described with slight modifications [24]. Briefly, after weighing each fat pad was returned to the carcass. The entire gastrointestinal (GI) tract was subsequently removed and carcasses were again weighed. This represented the empty carcass weight (ECW). Carcasses were labeled and secured in two layers of cheesecloth (Fisher) and frozen at -20°C until analysis. At this time, carcasses were freeze-dried for seven days and water content was determined by subtracting the freeze-dried weight from ECW. FAT was then extracted with ether for 7 days, followed by acetone for an additional 7 days in a Soxhlet apparatus. Carcass ash (ASH) was determined measuring the remains after a 16 hour incineration at 575°C. Carcass protein (PROT) was calculated as the remaining portion after carcass fat (FAT) and ASH were subtracted from ECW.

Statistical analysis
The MEANS and UNIVARIATE procedures of SAS were used to generate descriptive statistics and test normality assumptions for each trait [55]. All data were then analyzed using the GLM procedure of SAS [55] with a linear model that included the fixed effects of strain, sex and strain by sex interaction; dam's weight at breeding by strain was used as a covariate. A second linear model was used to test for strain by hg genotype (+/+ or hg/hg) interactions. This model included the fixed effects of donor region, sex and HG genotype and all possible two and three-way interactions. Choosing a nominal P value of 0.05 and applying the Bonferroni correction for multiple comparisons established significant differences in the ANOVA's. The critical P values used are indicated in each table.

Identification of candidate genes
Genes were identified by manual data mining of primary literature, reviews and book chapters. To organize and collate genomic and functional information for each gene we created a custom Gh signaling Gene Map Annotator and Pathway Profiler (GenMAPP) pathway [56] (Figure 5). Visualization and color-coding of genes using GenMAPP aided the selection of candidate genes mapping within QTL regions on MMU2, 9, 11 and 17.

Candidate gene sequencing
PCR amplicons covering the coding sequence and partial 5' and 3' untranslated regions for each selected candidate gene were amplified, purified and sequenced from the CAST strain using protocols outlined in [57] with slight modifications. Total RNA from brain, liver, spleen, lung and testis was isolated from an adult CAST male mouse using Trizol (Ambion). cDNA was produced from total RNA using standard procedures. PCR primer sets for each gene were designed using Primer3 [58] (Additional File 5). The initial amplification was performed in 10 μl PCR reactions using the Invitrogen Platinum TaqPCRx amplification system (Invitrogen). The reactions contained 1X PCR buffer, 1X Enhancer solution, 1.5 mM MgSO4, 0.17 mM dNTPs, 1 μM each primer, 0.1 unit of Platinum Taq (Invitrogen) and approximately 25 ng of cDNA. Reactions were incubated for 5 min at 95°C, then cycled for 45 s at 95°C, 45 s at 55°C, 1 min at 72°C for 35 cycles with a final 72°C extension for 10 min on a MJ Research PTC-200. The products were visualized on 1.5% agarose gels containing 0.06 μg/ml EtBr. Bands of the correct size were excised and incubated in 100 μl of sterile H2O at 80°C for 10 min to elute DNA. Reamplification reactions consisted of 1X PCR buffer, 1.5 mM MgCl2, 0.17 mM dNTPs, 1 μM each primer, 0.1 unit of Taq (Promega) and 10 μl of eluate in a total volume of 50 μl. The same cycling parameters were used for reamplification reactions. Products were visualized on 1% 0.5X TBE agarose gels containing 0.06 μg/ml EtBr, excised and purfied using PCR purification columns (Promega). To quantity each fragment 1 μl of each purified product was run on a 1.0% agarose gel containing 0.06 μg/ml EtBr, along with a DNA mass ladder (Invitrogen). In a 96 well plate, 15 ng of each PCR product was added to 5 pM of primer for sequencing. The College of Agriculture and Environmental Sciences (CAES) Genomics Facility at the University of California, Davis, performed bidirectional sequencing of each amplicon.

Sequence analysis
B6 mRNA sequences for each gene were downloaded from the May 2004 (mm5) UCSC (NCBI Build 33) genome assembly [28]. These sequences were imported along with all CAST sequence traces into the SeqMan sequence assembly program (DNASTAR) and manually curated for quality. Poor quality reads were resequenced. Individual contigs were created for each gene and polymorphisms were detected and curated by manual inspection.