Non-Destructive Sampling of Ancient Insect DNA

Background A major challenge for ancient DNA (aDNA) studies on insect remains is that sampling procedures involve at least partial destruction of the specimens. A recent extraction protocol reveals the possibility of obtaining DNA from past insect remains without causing visual morphological damage. We test the applicability of this protocol on historic museum beetle specimens dating back to AD 1820 and on ancient beetle chitin remains from permafrost (permanently frozen soil) dating back more than 47,000 years. Finally, we test the possibility of obtaining ancient insect DNA directly from non-frozen sediments deposited 3280-1800 years ago - an alternative approach that also does not involve destruction of valuable material. Methodology/Principal Findings The success of the methodological approaches are tested by PCR and sequencing of COI and 16S mitochondrial DNA (mtDNA) fragments of 77–204 base pairs (-bp) in size using species-specific and general insect primers. Conclusion/Significance The applied non-destructive DNA extraction method shows promising potential on insect museum specimens of historical age as far back as AD 1820, but less so on the ancient permafrost-preserved insect fossil remains tested, where DNA was obtained from samples up to ca. 26,000 years old. The non-frozen sediment DNA approach appears to have great potential for recording the former presence of insect taxa not normally preserved as macrofossils and opens new frontiers in research on ancient biodiversity.


Introduction
Most ancient genetic studies have focused on vertebrates, plants and to a lesser extent microbes revealing aDNA research as a powerful tool for testing hypotheses in biology [1,2]. Although insects are the most diverse animal group on Earth with more than 1 million described species, aDNA studies on insects have so far been limited and restricted largely to museum specimens of historical age, up to ca. 100 years [3,4,5,6], or to geologicallyancient amber-entombed specimens millions of years old (e.g. [7,8,9]). While the former have produced exciting results relating to events in the near past, the latter have proved a classical example of how a lack of appropriate contamination controls in aDNA research may produce false positive results [10,11].
Only three studies appear to have investigated insect DNA survival between these two extreme time-ranges: [12] studied grasshoppers from glacial deposits in Wyoming deposited ca. 400 years ago; [13] investigated beetle remains from ca. 20,000-year-old packrat middens from Texas; [14] studied 450,000-to 800,000-yearold silty-ice from the base of a Greenland ice core. All three studies gave positive results for the presence of insect DNA, which encourages further research on the possibilities of obtaining insect aDNA in other contexts.
Intriguingly, a major constraint on the use of historical, and particularly ancient, insect specimens in aDNA research is the destructive nature of the sampling procedure [15]. Obviously, this is a problem related to many aDNA sources, but is of particular concern with small specimens, such as insects, where even limited sampling may destroy important morphological characters. All the above insect ancient genetic studies have suffered from such destructive sampling procedures. One potential solution is the application of an extraction protocol that uses digestion buffers designed to enable the recovery of DNA from insect remains without causing visual external morphological damage to the material [16]. This method has been used successfully on museum specimens of beetles collected between 1952 and 2002.
Here, we report the results of a study that tested the potential of obtaining authentic ancient insect mtDNA using this nondestructive extraction procedure on historical museum beetle specimens dating back to AD 1820 and on ancient chitin from beetle macrofossils from permafrost dating back more than 47,000 years (DNA obtained from samples up to ca. 26,000 years old). Additionally, encouraged by the findings of insect aDNA in the Greenland silty-ice, we explored non-frozen sediments from New Zealand laid down between 3280 and 1800 years ago as a direct source of ancient insect DNA, even though no visible fossil insect remains were present. This non-frozen sediment DNA approach is interesting in the current context, as it holds the potential of obtaining insect aDNA without the destruction of valuable specimens, as well as providing data on former biodiversity in the absence of macrofossils and unobtainable in any other way.

Results
The non-destructive DNA extraction procedure was tested on two types of samples: i) Twenty museum specimens (representing five different species) of beetles collected between AD 1820 and AD 2006 (the oldest historical museum insect remains from which DNA survival has been investigated), and ii) fourteen beetle macrofossils (chitin) from the late Pleistocene (ca. 47,600-20,100 14 C years BP) and late Pleistocene-early Holocene (ca. 10,595-7,110 14 C years BP). These macrofossils were recovered from permafrost sediments in Chukotka (Siberian Far East) and central Alaska in 2004 and 2005, respectively.
All twenty specimens of museum beetles produced amplifiable and authentic COI mtDNA sequences between 77-204 -bp in size. These were from the ground beetle Harpalus latus (Linnaeus, 1758), the pill beetle Byrrhus pilula (Linnaeus, 1758), the leaf beetle Chrysolina polita (Linnaeus, 1758) and two weevils, Otiorhynchus sulcatus (Fabricius, 1775) and Curculio pyrrhoceras Marsham, 1802 (Table 1). Of the 14 permafrost-preserved beetle chitin macrofossils, only three yielded successful COI or 16S mtDNA amplification products; a weevil Lepidophorus thulius (Kissinger, 1974) (ca. 10,595 14 C years BP, dated by association with wood from the sample), a ground beetle Amara alpina (Paykull, 1790) and a rove beetle Tachinus brevipennis Kiesenwetter, 1850, both with radiocarbon ages of ca. 26,000 14 C years BP, estimated from a sedimentation rate based on overlying and underlying radiocarbon dated samples of plant macrofossils ( Table 2). The amplification products from the macrofossil remains were between 91-159-bp in size. An inverse relationship between amplification strength and length typical of aDNA supports the authenticity of the findings as does sequence identification in agreement with the morphological based taxonomic affiliation of the specimens (Tables 1 and 2). Importantly, none of the insect specimens subjected to DNA extraction seemed to have undergone any visible physical alterations after the extraction procedure ( Figure 1). In addition to the above, we examined the potential for longterm survival of insect DNA in temperate sediments. Two insect COI mtDNA sequences of 96-bp in length were obtained from one of the two non-frozen sediment samples from New Zealand. The sediments were laid down between 3280 and 1800 years BP [17]. The sequences were identified as being from a beetle and a moth/butterfly, respectively (Table 3).

Discussion
The 100% success rate on the beetle specimens from museum collections dating back 188 years suggests that the non-destructive extraction procedure tested has considerable potential for sampling historical insect material, even when more than 100 years older than the specimens originally tested with the method [16]. It may be worth exploring if similar success can be obtained on insect groups other than beetles, such as Lepidoptera, Diptera and Hymenoptera, whose chitinous exoskeleton is not as thick and resilient as that of beetles. However, we see no obvious reason why the procedure should not work on a variety of taxa. The result is significant in that museum insect specimens have already proved to be an important resource for e.g. identifying recent bottlenecks [4] or the development of traits such as insecticide resistance [5] etc. In particular, the non-destructive extraction procedure appears to have removed the need for destructive sampling.
The limited success of 3/14 (ca. 21%) on the truly ancient beetle chitin remains may result from either the remains no longer containing amplifiable endogenous DNA despite preservation in ideal frozen conditions for most of the preservation period (e.g. see [18]), or the extraction procedure not being efficient enough to retrieve DNA from truly ancient remains even where destructive sampling could have been successful. The possibility of a lower extraction efficiency is supported by the results of a similar nondestructive DNA extraction protocol for mammalian teeth [19]: only specimens that had been in museum collections for relatively short times yielded DNA using the non-destructive sampling method and remains that had been in collections for much longer periods gave products only with destructive sampling strategies. It appears that only limited success can be expected using the method of [16] on truly ancient insect specimens.
Interestingly, DNA from a beetle and a moth/butterfly was obtained from one of the two New Zealand temperate sediment samples, even in the absence of visible macrofossil material. The failure to obtain insect DNA from one of the two samples could result from spatial differences in the distribution of DNA source material. The success of the New Zealand non-frozen sediment DNA compared to the permafrost preserved macrofossils is surprising in that, although the sediment samples were several thousands of years younger than the macrofossils examined, it is generally believed that it is the temperature of preservation rather than the age itself that determines the level of DNA degradation [1]. The source of insect DNA preserved in the sediments may include material other than macrofossil remains of adults, such as eggs or larvae, additional to that of harder, chitinous materials. The results from the sediments are important because this is the first time insect DNA has been retrieved directly from non-frozen sediments. The approach may have wide applications. Ancient sediment-preserved DNA studies could reveal the former presence of taxa not normally preserved in the fossil record such as soft-bodied insects. Although the non-frozen sediment DNA approach involves destructive sampling, it has the advantage that the material is the sediment itself, which is usually abundant, and normally not too valuable to process.

Historical Museum Specimens
Four individuals each of five beetle species (a total of 20 specimens) were selected, to cover a historic period spanning from AD 1820 until today (Table 1). All specimens were collected in Denmark, and are held in the collection of the Natural History Museum, Copenhagen, Denmark. Sequences of the COI gene for all the five species were available on GenBank, which allowed the construction of species-specific primers (Table S1).
Co-ordinates for Chukotka samples: 64u069N, 171u119E. All samples were kept in 96% ethanol in the freezer until nondestructive DNA extraction. Before extraction, samples were dried overnight at 55uC for the ethanol to evaporate. New primers for insects were constructed and additional primers from the literature were used [14,20] to amplify COI and 16S mtDNA sequences, and the same sets of primers were used on all macrofossil samples (Table S1).

Sediments
DNA from two samples of cave sediment from the late Quaternary Hukanui Pool site, eastern North Island, New Zealand [17,21] was extracted and assayed as described in [22] and [23] ( Table 3). The site was deposited beneath large erratic limestone blocks, and contained sediment layers between layers of well-dated volcanic tephras originating from the Taupo Volcanic Zone, 100 km to the west. The sediment was deposited between the Waimihia eruption of ca. 3280 years BP [24], and the AD 1870 surface, and the specific sediment sample from this study is ca. 3280-1800 years old. Primers from [14] were used to amplify COI mtDNA sequences (Table S1).

DNA Extraction and PCR
DNA extraction and PCR setup was carried out in dedicated aDNA clean-laboratories [2]. DNA was extracted from museum specimens and macrofossils using the non-destructive method [16]: Whole specimens were placed in 2 ml Eppendorf Biopur tubes, fully immersed in digestion buffer (volume dependent on specimen size, 0.5-1.5 ml in this study), and incubated overnight at 55uC with gentle agitation. The buffer consisted of 3 mM CaCl 2 , 2% sodium dodecyl sulphate (SDS), 40 mM dithiotreitol (DTT), 250 mg/ml proteinase K, 100 mM Tris buffer pH 8 and 100 mM NaCl (final concentrations). After incubating with gentle agitation for 16-20 hours, specimens were removed from the buffer, placed in 100% EtOH for 2-4 hours to stop further digestion, air-dried, and replaced in their collections. Nucleic acids were purified from the digestion buffer using a Qiagen PCR purification kit (QIAquick).
DNA from the sediments was extracted using the procedure described in [22] and [23].
PCR reactions for all samples, except the sediments (see [22]), were the following: 1 ml DNA, 2.5 ml of each primer (10 mM), 2.5 ml 106HiFi Buffer, 2 ml BSA, 1 ml MgSO 4 , 0.2 ml dNTPs and 0.1 ml Platinum Taq HiFidelity Polymerase enzyme (invitrogen, Carlsbad, CA) and 13.2 ml ddH 2 O giving a total 25 ml PCR reaction. PCR conditions were: 94uC for 2 min. followed by 60 cycles of 94uC for 30 sec., 50-52 uC for 30 sec., 68uC for 40 sec., completed with a final 68uC for 7 min. PCR products were tested on 2% Agarose gels stained with ethidium bromide. The amplified PCR products were purified using an Invitek purification kit (PCRapace) and cloned with Invitrogen Topo TA cloning kit. All PCR products were cloned prior to sequencing in order to ensure sequence accuracy. New PCRs were performed on 8-24 E. coli colonies, using the primers M13F and M13R and amplified for 35 cycles with annealing temperature of 54uC. PCR products containing the inserted PCR extracts were purified using vacuum suction and commercially sequenced (Macrogen, Seoul, Korea). If sequence differences were obtained from individuals of the same species, the sequence results were replicated to test for miscoding lesion damage [25,26].

Sequence Identification
All sequences were identified by a method using Bayesian approach to statistical assignment [27]. The method has advantages compared to the online BLAST search tool, by including phylogenetic information and providing statistically meaningful measures of confidence (posterior probabilities) to the taxonomic assignment.

Radiocarbon dating
The age of the insect fossil remains were estimated from associated radiocarbon ages from in situ plant macrofossils from the same sampling horizons using Accelerator Mass Spectrometry (AMS) or radiometric (conventional) radiocarbon dating ( Table 2). All ages are reported in radiocarbon years BP, which are slightly younger than their corresponding calendar year ages. Chukotka sample ages were estimated from sedimentation rate based on the following ages (  The New Zealand samples investigated are from a series taken from sediments between independently dated tephra (volcanic ash) beds laid down ca. 3280-1800 years BP as described in [17] and [21].