Feathers as a source of DNA CR0202:

The isolation and testing of microsatellite markers from some Birds of Prey

A Report to the Department of Environment, Transport and the Regions

David T Parkin & Nicola J Peck, Institute of Genetics, University of Nottingham.

INTRODUCTION

This report begins by detailing the methods and rationale behind an attempt to isolate microsatellites from target species of birds of prey (raptors): Peregrine Falcon, Goshawk, Red Kite, Golden Eagle and Saker Falcon. The laboratory work was initiated by Dr Jon Wetton (salary financed by the Leverhulme Trust), but he left for a career with the Forensic Science Service a few months into the project. The work was continued by Nicola Peck as part of her post-graduate studies (salary funded by the European Commission and latterly by Forest Enterprise). After successfully completing her Ph D, Dr Peck also left for a career with the Forensic Science Service. Further research applying these microsatellites to additional species (Merlin, Peregrine, Saker and Gyr Falcons) has been undertaken by Amy Marsden (Ph D student financed by NERC) and Caroline Metcalf (M Phil student financed by the European Commission). Some of the Goshawk primers have been successfully applied to Sparrowhawks by Arnold van den Burg (Ph D student funded by University of Nottingham). Shortly after the successful identification of two highly variable loci from Peregrine, plus several others that remain unquantified, we learned that a Norwegian researcher (Dr Marit Nesje) was undertaking an exactly similar study with that species (Nesje et al. 2000). We agreed to transfer our attention to the other species. Exchange of methods was agreed, and we would test her micros on Merlin, Gyr Falcon and Saker.

The identification of microsatellite loci in Goshawk was successful. Golden Eagle and Red Kite proved much more difficult. Other groups have had similar difficulty with eagles (and indeed Goshawks). It seems possible that there are fewer microsatellite loci in eagles and kites: the success with Peregrine and Goshawk (and House Sparrow, African Grey Parrot and Red Squirrel) indicates that the method works in our hands, and perhaps the material is less tractable.

Microsatellites developed for individual species were tested across the rest of the targets. Primer sequences are given in detail, and protocols are provided for the amplification of all primer sequences against all variable loci. The microsatellites that were identified were tested against known families to confirm their inheritance. Selected results are given in more detail to show the potential of these for the recognition of parentage and provenance of individual birds.

METHODS

The simplest way of identifying microsatellites is by using previously described sequences published in the literature or posted on data bases such as GenBank. This is popular method for species that have been extensively sequenced (eg Drosophila, mice and humans) but is less applicable when sequence data are not available for the species of choice. Although microsatellites can occasionally be used cross-species, it is usually necessary to isolate microsatellites for each species of interest.

Until recently, microsatellites have been isolated by probing whole genome DNA libraries. However, this is a long, laborious and inefficient process that often results in only a handful of markers being isolated due to their low frequency in the genome. To improve the efficiency of the technique, libraries are now ‘enriched’ for microsatellites. There are several techniques for enrichment that differ slightly in detail. However, as the frequency of microsatellites was known to be low in avian genomes (Primmer et al. 1997), we attempted a fast, and usually efficient technique enabling microsatellites to be isolated from even the most deficient genomes (Armour et al. 1994; Refseth et al. 1997).

The enrichment procedure can be split into four main stages:

  • Preparation of genomic DNA
  • Preparation of the oligonucleotide probes
  • Enrichment of library
  • Constructing of DNA library from enriched DNA

DNA Extraction

Extraction from blood

When blood samples were available, standard extraction techniques were used. Avian erythrocytes are nucleated and therefore a prime source of DNA. For this reason, the yield of DNA per volume of blood is much higher from avian than from mammalian blood (which contains non-nucleated red blood cells). A standard phenol/chloroform extraction process is used to recover the maximum yield of high molecular weight DNA devoid of protein and other restriction enzyme inhibitors (Sambrook et al, 1989).

  1. Resuspend approximately 18l of whole blood or 1cm2 chopped tissue in 650l of isotonic 1 x SET buffer (150mM NaCl, 50mM Tris, 1mM EDTA.Na2. pH8.0) in a 1.5ml eppendorf tube.
  2. Lyse the cells by addition of 7.5l SDS (sodium dodecyl sulphate solution) and add 15l 10mgml-1 proteinase K to inhibit activity of nucleases released from the cells. For tissue samples add double the amount of proteinase K.
  3. Aid lysis by applying mechanical action using a plastic pestle.
  4. Incubate overnight at 55C.
  5. Add 500l of phenol solution (Phenol liquefied washed in Tris Buffer – Fisher Scientific) to each sample and gently invert on a rotary mixer for 1 hour. This allows contaminating proteins to be removed by their differential solubility in aqueous and organic solutions.
  6. Centrifuge the samples at 8000g for 7 minutes to separate the phases. Carefully remove the viscous aqueous upper layer to a fresh tube using a wide aperture, 1ml disposable pipette tip to minimise shearing of the DNA. Care must be taken to not carry over the interface.
  7. Repeat phenol extraction (stages 5 and 6) until the aqueous layer is colourless (two phenol extractions are usually sufficient). Mixing time can be reduced to 40 minutes.
  8. Meanwhile, prepare fresh phenol/chloroform solution by mixing equal volumes of phenol solution and chloroform/iso-amyl alcohol (23:1, v:v) in a Duran flask. Mix well before allowing the two phases to separate.
  9. Repeat the extraction of the aqueous layer, substituting phenol/chloroform solution for phenol. Mixing time can be further reduced to 30 minutes.
  10. Perform a final extraction step using cholorform/iso-amyl alcohol (23:1) instead of phenol/chloroform solution. Mixing for 20 minutes is sufficient to remove the last traces of phenol.
  11. 500l of the aqueous layer is transferred to a fresh tube and the DNA precipitated by the addition of 2ml of cold absolute ethanol (-20C). Tubes are mixed on the rotor for 10 minutes or until the white precipitate of DNA can be clearly seen.
  12. Store the samples at -80C for 30 minutes as this has been shown to increase the DNA yield by up to 20%.
  13. Pellet the DNA by centrifugation in a microfuge for 15 minutes, then remove the ethanol from the pellet with a vacuum pump. Rinse the pellet in 2ml of 75% ethanol by vortexing the tubes before centrifuging the samples for 5 minutes. Remove the remaining ethanol with the vacuum pump.
  14. Dry the pellets overnight on the ‘bench’ with eppendorf lids open or alternatively in a 37C oven for 20 minutes, again with the eppendorf lids open.
  15. When dry, re-suspend the pellets, according to size, in 50-150l TE by overnight incubation at 55C. Store at 4C until required.

Extraction from feathers

We found that it is possible to extract good quality DNA from fresh down plumes or from growing feathers from the body of young birds. These contain live cells and are frequently pink (blood) coloured at the base. We did not use remiges (flight feathers) or retrices (tail feathers) since Home Office legislation prevented us from extracting these feathers. There is no doubt that these would provide a plentiful supply of DNA, but we had no wish to compromise the flying ability of fledgling birds by removing developing feathers at the nestling stage. Growing (‘pin’) feathers from adult birds were similarly well-endowed with DNA, but cast feathers were more problematic. Body feathers were unreliable: some provided DNA, others did not – with no obvious cause for success or failure. Wing and tail feathers, however, provided plenty of DNA and could be used for typing birds repeatedly. Amy Marsden’s Merlin study showed that genetic profiles could be generated from cast wing and tail feathers that had been lying on the moor for several weeks, and also from old feathers stored casually in drawers of field workers’ filing cabinets for up to 20 years.

For feather or down samples, DNA was extracted using a chelating resin approach (Morin et al, 1994) followed by an ethanol precipitation step to purify the DNA for long term storage and successful PCR. The ‘feather DNA extraction’ protocol is as follows:

  1. Add 50l of 75% ethanol to approximately 3mm of the body tip of feathers (1-2 feathers) or down (3-4 pieces of down) and chop finely with a sterile scalpel blade on a sterile glass plate.
  2. Once chopped, place feather/down tips in individual eppendorfs and add 200l Chelex chelating resin (5% in Tris, pH8.0) and 10l 1% Proteinase K to each tube.
  3. Incubate whilst rotating overnight at 55C.
  4. After incubation, vortex the eppendorf tubes for 10 seconds and pierce the tube lids before heating in a 100C waterbath for 10 minutes.
  5. Vortex the tubes for 10 seconds before centrifuging for 3 minutes (13000rpm).
  6. Remove the DNA containing supernatant (approximately 200l) to a fresh tube and further centrifuge for 2 minutes to ensure complete removal of feather debris and Chelex beads.
  7. Again, remove the supernatant (approximately 150l) to a fresh tube before adding 200l of cold (-20C) 100% ethanol and 5l 3M sodium acetate.
  8. Mix gently and chill for 45 minutes at -80C before centrifuging for 20 mins (13000rpm). Ensure consistent orientation of tubes in the centrifuge for ease of identification of the DNA pellet.
  9. Carefully remove the supernatant with a vacuum pump and wash the DNA pellet with 200l of chilled (-20C) 70% ethanol before centrifuging for 5minutes.
  10. Remove the ethanol once more and allow the DNA pellet to dry for 10 minutes at 55C.
  11. Add 75l of TE (10mM Tris, 1mM EDTA.Na2, pH8.0 with HCl) and incubate overnight at 55C to resuspend the DNA. Store at 4C until required.

Enrichment

Libraries were constructed for Golden Eagle, Red Kite and Goshawk. The individuals chosen for DNA extraction were unrelated, and preferably female to ensure that both sex chromosomes were included in the sample of the genome scanned for microsatellites.

In the initial libraries, tetra-nucleotide probes were utilised in preference to dinucleotides due to the reduction in artifacts of tetranucleotide microsatellite loci. This removes some ambiguity when scoring autoradiographs, as does the greater size difference imposed by a tetranucleotide compared with dinucleotide or trinucleotide repeats (Schlotterer and Tautz, 1992). Tetranucleotide probes are four base units repeated six times forming, what will henceforth be called a 24mer, eg (xxxx)6. However, problems with the low variability of the tetranucleotide microsatellites necessitated the use of dinucleotide oligonucleotides as well. An advantage of dinucleotides in that their shorter repeat length increases slippage errors thereby elevating the mutation rate, and hence the variability of the loci. The total length of the dinucleotide probes was equal to that of the tetranucleotide probes as they consisted of two base units repeated 12 times eg (xx)12. Furthermore, since it has been observed that longer microsatellites are more likely to be variable, it was also decided that longer probes would facilitate the isolation of longer arrays (Rico et al., 1994; Armour et al., 1994; Karagyasov et al., 1993). For this reason, libraries utilising the di and tetranucleotide probes concatenated to lengths of 200-700 base pairs were also constructed (see Table 1).

Table 1. Tetranucleotide probes

(TATC)6 / (TTTG)6 / (CCAA)6 / (TTGG)6
(GATA)6 / (TTTC)6 / (CCAT)6 / (GGAA)6
(GGAA)6 / (TTTA)6 / (AAAT)6 / (TTCC)6

All enrichment procedures used involve the hybridisation of size selected genomic DNA to synthetic ‘microsatellite-like’ oligonucleotides to isolate microsatellite enriched DNA. Two main types of enrichment procedure were used. The first method is a modification of the methods of Armour et al. (1994) and Karagyasov et al. (1993) methods which involve fixing the synthetic oligos to a nylon membrane before hybridising with genomic DNA. The second requires the synthetic oligos to be modified by attaching a biotin molecule to them, and then utilises the exceedingly strong biological interaction between biotin and streptavidin by subsequently attaching these ‘bio-oligos’ to streptavidin-coated magnetic particles. This allows the oligo probes to hybridise to the genomic DNA in solution and the magnetic properties of the beads also allows for quick and easy isolation of this selected DNA. The first library created using this biotin approach was based on the methods of Kijas et al. (1994) and Prochazka et al. (1996). Subsequent biotin libraries (see Tables 2 and 3) were constructed following Refseth et al. (1997).

Table 2. Tetranucleotide Biotin probes

Bio-(TGTC)6 / Bio-(TTTA)6 / Bio-(TTTG)6 / Bio-(TTTC)6
Bio-(TATC)6 / Bio-(TTCC)6 / Bio(CCAT)6

Table 3. Dinucleotide Biotin probes

5’-(AT)12GCCG-Bio / 5’-(CG)12GCCG-Bio
5’-(CT)12GCCG-Bio / 5’-(CA)12GCCG-Bio

The genomic DNA (now enriched for microsatellites) was cloned into a suitable vector. Two different types of vector have been used in this study: T-vector and pUC18 plasmid. Once the DNA insert has been successfully ligated into the chosen vector, the plasmid is transformed into suitable competent E.coli cells. These cells are grown up on agar plates and individual clones selected for arrangement in an ordered array genomic library. Clones that prove ‘positive’ when hybridised with the radiolabelled synthetic oligos are likely to contain microsatellite sequences. Once the clones containing the microsatellites have been isolated, the sequence of the specific microsatellite can be obtained by direct sequencing of the clone. From the sequence data, it is possible to design polymerase chain reaction primers that flank the repeat region and which can subsequently be used to amplify the microsatellite in samples of DNA. A summary of the type of library used for each species is shown in Table 4.

Table 4. Summary of types of library constructed

TETRANUCLEOTIDE / DINUCLEOTIDE
Concat / Biotin / Concat
Red Kite /  /  / 
Goshawk /  /  / 
Golden Eagle /  /  / 

Isolation and sequencing of microsatellites

Initially the enriched amplified DNA was ligated into the pGEM-T vector to take advantage of the deoxyadenosine molecule often added to the 3’ ends of PCR templates by the Taq polymerase during PCR. Lack of 3’-5’ exonuclease activity in Taq polymerase ensures this additional base is not subsequently removed. In this way it is not necessary to add another step to modify the ends of the target DNA for ligation into the vector as the PCR process will have already done so. The T-vector is manufactured to be complementary to these 3’-A overhangs by cutting the circular plasmid and adding terminal deoxythymidines to its 3’ ends.

A critical step in the efficiency of ligation is the ratio of vector to insert DNA concentration. There should be enough DNA to ligate into all available vectors but not so much that two separate DNA inserts ligate to each end of the linearised vector. To optimise the insert/vector molar ratios, a number of different ligation reactions were initially examined to gauge the optimum ratio. Initially three reactions were employed: 3 insert / 1 plasmid, 1 insert / 1 plasmid, and 1 insert / 3 plasmid. A simple equation can be used to calculate the appropriate amount of PCR product to include in each of the test ligation reactions.

Although ligation into the T-vector proved successful, it was noted in the first (tetra-24mer) library that many of the positive clones contained no flanking regions. Closer inspection of these revealed that they were identical with those of the original oligo probes. The results of the probing of the PCR product fraction revealed the presence of very heavy DNA that was not present in the either the input or output fractions. This probably represents a concatenation product of the residual 24mer probes that may have contaminated the target DNA elution product. Presence of these large concatamers also suggests smaller, partially concatenated probes would be present in the PCR product. These oligo probes will also have a 3’-A overhang added during the PCR and hence will also be available to ligate into the T-vector. The ‘microsatellites’ with no flanking regions we were finding were no doubt contaminating oligo probes accidentally cloned into the T-vector.

To overcome these problems, a system had to be found that discriminated between true target DNA and residual oligo probes. The answer was found in the characteristics of the linker/DNA join. The linkers are annealed to the DNA due to their MboI compatible overhangs. When the DNA and linkers have annealed they once again form a recognition site for the MboI enzyme. Restricting the PCR product with MboI will cut the linkers off the ends of the DNA again creating GATC overhangs. These DNA fragments can then be ligated into a vector with compatible ends to this MboI cut site. The oligo probe DNA does not have MboI recognition sites within the sequence, and so will not be cut and hence will not be compatible for ligation into the chosen vector. The vector chosen for ligation with MboI cut ends is pUC18 BamHI/BAP (Pharmacia biotech). This pUC18 vector contains an ampicillin resistance gene and the LacZ gene required for colour screening as well as being compatible with MboI cut DNA. The enriched DNA from the tetra-biotin library was ligated into pUC18 plasmid. It was found that a ratio of 3 vector / 1 insert worked best. This ratio also increased the likelihood of the genomic DNA fragments annealing to the vector instead of together, and hence ensured fewer double insert ligations.