We collected down feathers from the rump of 20 wild Kagus of a population in the Parc Provincial de la Rivière Bleue, south New Caledonia (22°3–12′S, 166°33–46′E), from 2004 to 2009. We plucked out about 10 feathers without touching their roots, placed them in hermetically sealed bags, and stored them at 4°C. Additionally, we sampled 2 Kagus from other populations: a feather sample from 1 wild-caught Kagu of a population near Poindimié in the northeast of New Caledonia (20°56′S, 165°20′E), and a tissue sample (stored in 95% alcohol) from 1 deceased captive-bred Kagu (parents were wild-caught) of the Parc Zoologique et Forestier in Nouméa, New Caledonia (22°15′S, 166°27′E).
We isolated DNA from the tissue sample following Hogan et al. (1994). The feather DNA was isolated according to Segelbacher (2002). The DNA of the tissue sample was subjected to 454 sequencing at the GS-FLX LAB (Eurofins MWG Operon, Ebersberg, Germany). A 1/16 plate generated 52,583 reads (13.7 Mb) with an average length of 261 bp. A single Fasta file containing all reads was screened for di-, tri- and tetra-nucleotide repeats using MSATCOMMANDER (Faircloth 2008) with at least six repeats for di-nucleotide and four for tri- and tetra-nucleotide microsatellites. We detected 349 loci (104 di-nucleotides, 175 tri-nucleotides, 70 tetra-nucleotides) suitable for primer design with Primer3 software (Rozen and Skaletsky 2000) using the default settings. Subsequently, we ignored loci with extremely too long and compound interrupted repeat stretches as described by Opgenoorth (2009). In order to detect identical sequences within the remaining set of potential microsatellites, we compared the sequences by similarity and additionally aligned them with the software GENEIOUS v4.7 (Drummond et al. 2009). Finally, out of 231 potential loci (65 di-nucleotides, 117 tri-nucleotides, 49 tetra-nucleotides), we randomly ordered 60 primer pairs flanking 25 di-, 25 tri- and 10 tetra-nucleotide from Biomers (Ulm, Germany). The fluorescent labeling was done according to Schuelke (2000) with a universal M13 Primer. We used gradient PCR-protocol (50–60°C) (Mastercycler Gradient, Eppendorf, Germany) to test the 60 primer pairs on DNA extracted from 3 individuals and obtained amplified products of expected size with low rates of intense stuttering for 30 of the 60 microsatellite markers.
For genotyping, we used the DNA of the 22 sampled individuals. Polymerase chain reactions (PCRs) were performed in a total volume of 15 μl with the following components: 25 ng of genomic DNA, 0.2 μM of each reverse primer, and the M13 universal primer (fluorescently labeled with 6-FAM), 0.05 μM of each forward primer, 0.2 mM of each dNTP (Solis BioDyne), 1.5–3 mM MgCl2 (Table 1), 1 × PCR buffer (Solis BioDyne), and 0.5 U Taq DNA Polymerase (FIREPol®, Solis BioDyne). For PCR amplification, we used a thermal cycler (Mastercycler Gradient, Eppendorf, Germany) with the following PCR profile: initial denaturation at 94°C for 5 min, 30 cycles of 30 s at 94°C, 30 s at the primer specific annealing temperature (Table 1), 45 s at 72°C, followed by eight cycles of 30 s at 94°C, 45 s at 53°C, 45 s at 72°C, and a final elongation step at 72°C for 10 min. PCR products were separated on 6% polyacrylamide gels on an ABI Prism 377 automated sequencer (Perkin Elmer) and scored in reference to a ROX standard (79–540 bp) by GENESCAN® 3.1.2 and GENOTYPER® 2.5 software (Applied Biosystems, Foster City, CA, USA). We repeated the PCR amplification and genotyping of DNA extracts from feathers three times as the use of this material as a source of DNA can lead to genotyping errors, mainly allelic dropout (Taberlet and Luikart 1999; Segelbacher 2002). We used GENEPOP 4.0 (Rousset 2008) to calculate the number of alleles, to generate allele frequencies, expected (H
E) and observed (H
O) heterozygosities, and to test for linkage disequilibrium and deviations from Hardy–Weinberg equilibrium (frequency of alleles in a large, interbreeding population characterized by random mating, mendelian inheritance, and the absence of migration, mutation, and selection). For calculating the polymorphic information content (PIC), we used the EXCEL MICROSATELLITE TOOLKIT (Park 2001). MICRO-CHECKER 2.2.3 (Van Oosterhout et al. 2004) was used to test the dataset for genotyping errors and for the presence of null alleles.