The BAC library was constructed by cloning genomic DNA isolated from the white blood cells of a male giant panda, and partially digested with BamHI/HindIII, into the vector pBeloBAC11. It contains 198,844 clones, which were deposited in 518 384-well plates. Among these 198,844 clones, 145920 were from BamHI digestion and 52924 from HindIII digestion.

To evaluate the average size of inserted fragments in the library, DNAs were prepared from 261 randomly selected clones, cleaved with NotI and subjected to pulse-field gel electrophoresis with midrange PFG Marker II (New England Biolabs, USA). The insert size distribution is shown in Figure 1. The average insert size of these 261 clones is 108 kb, indicating that the library represents a 7-fold coverage of the giant panda haploid genome. Among the 261 clones, only 6 had no insert, suggesting that the percentage of non-recombination in the library is about 2.3%. Approximately 53% of the clones contain inserts larger than 100 kb.

To assess the quality of this library further, 8 microsatellite markers were used for screening. The number of positive superpools varied from 1 to 10 with an average of 4.6 (Table 1). At the same time, 15 giant panda genes (ACTIVIN, AGXT, BDNF, FSH, FSHR, IRBP, GHR, GpH, GDNF, NT-4, LH, SOX14, TTR, ZFX1 and ZFX2) were also screened in the library using PCR. As shown in Table 2, the number of positive BAC clones varied from 2 to 8 with an average of 4.7. The actual number of clones obtained was slightly lower than was calculated theoretically, but at least 2 positive BACs for each of the genes were identified from the library.

To estimate the fraction of chimeric clones in the BAC library, and furthermore to map some of the genes, a FISH approach was used. Ten BAC clones, containing AGXT, GHR, FSHR, IRBP, SOX14, TTR, BDNF, NT-4, LH and ZFX1 respectively, were mapped to 8 pairs of giant panda chromosomes. The specific locations of the 10 functional genes on the G-banded ideogram are listed in Figure 2. The band names of each chromosome are not given because no international standard karyotype has yet been established for the giant panda. The order of chromosomes was identified according to a previous study 12 by analyzing the size, G-band and centromere location. Fifteen metaphases were analyzed to map each gene.

The BAC vector pBeloBAC11 was purified by CsCl-ethidium bromide density gradient centrifugation, digested with the appropriate amount of HindIII or BamHI (New England Biolabs, USA) and treated with shrimp alkaline phosphatase (USB, USA). Linear vectors were recovered by electroelution as described by Osoegawa et al. 16. A DNA plug was prepared as described by Birren et al. 17 from the white blood cells of a male giant panda and partially digested using 5 units of HindIII or BamHI per μg DNA at 37°C for 45 min. Size-fractionation was performed in a CHEF III apparatus (Bio-Rad, USA) and DNAs were separated in three stages as described by Osoegawa et al. 16. DNA fragments with range of 150–250 kb were recovered by electroelution.

At an approximately 1:10 molar ratio of insert: vector, 150 kb–250 kb DNA fragments were ligated to 30 ng of linearized pBeloBAC11 vector in a 120 μl total volume and incubated at 12°C for 24–48 h. The ligation mixture was then dialyzed on a microdialysis filter (0.025 μm pore size; Millipore) against 0.5 × TE for 40 min. Two microlitres of the ligation product was used to transform 20 μl ElectroMAX DH10B competent cells (Gibico, USA) by electroporation on a Gene Pulser II (Bio-rad). The transformation parameters were as follows: resistance of pulse controller plus, 100 Ω; voltage gradient, 18 kv/cm; capacitance, 25 μF. In total, 87 electroporations were performed to make the panda HindIII and BamHI library. Transformed cells were frozen at -70°C prior to colony picking.

The frozen stocks of the primary clones were recovered and spread on large selection plates. Colonies with diameters >1.5 mm were picked manually and incubated overnight at 37°C in 96 deep well plates containing LB medium supplemented with 10% glycerol and 12.5 μg/ml chloramphenicol. Using MultiPROBE^® II automated liquid handling systems (PerkinElmer, USA), 80 μl of cultures from four 96 plates were transferred to 384 well plates and stored at -70°C. To establish the two-step PCR screening systems, the library was divided into 26 superpools and one superpool comprised 20 384-well plates. Cultures from every superpool were combined to make superpool DNA for the first step PCR screening. In each superpool, cultures from each plate (384 clones), row (24 clones × 20 plates) and column (16 clones × 20 plates) were combined respectively to make DNA for the second step screening. The combined cultures were centrifuged and the bacterial pellet was suspended in TE then boiled in a microwave oven. After centrifugation, the supernatants containing the pooled DNA were transferred and stored in 96 well plates. The stock DNA was diluted 20 times to the working concentration for PCR screening.

Two hundred and sixty-one recombinant BAC clones were randomly picked and grown in LB medium overnight at 37°C. DNA was prepared by alkaline lysis and digested with NotI, then separated by PFG electrophoresis. PFGE conditions were set with Bio-Rad CHEF III: 6.0 v/cm for 12 h, linear pulse time ramping from 0.1 to 40 s.

Primers for each gene and some microsatellite markers were designed according to the giant panda DNA sequences published in NCBI, and the other microsatellite markers were designed according to the sequences from our laboratory (A204, A066 and B043). BAC screening was performed by two-step PCR (superpools PCR and 4D-PCR) [see Additional file 1]. Positive BAC clones were confirmed by sequencing of PCR products.

For FISH analysis, chromosomes were prepared from a male giant panda fibroblast using a standard cytogenetic protocol. The FISH method was modified from the previous study 18. Briefly, chromosomes were performed G-bands and well-banded metaphases were photographed. The BAC clones of 10 functional genes (Table 3) were used as probes and were labelled by nick translation with biotin-14-dATP (Invitrogen, Carlsbad, CA92008, USA). The denatured probes were dissolved in hybridization solution to a final concentration of 50 ng/μl and prehybridized for 40 min at 37°C. Hybridization was performed for 17 hours at 37°C in a humid chamber. Probes were detected with FITC-conjugated avidin (Vector, Burlingame, CA94010, USA) and signals were amplified by biotinylated anti-avidin (Vector). Chromosomes were counterstained with 0.5 μg/ml propidium iodide. Images were obtained with an epifluorescence microscope equipped with a DP70 CCD camera (Olympus, Japan). G-banded ideograms were drawn with Video TesT-karyotype3.1 software (Video TesT Ltd, Russia, 2003).