Insects are currently facing global decline1, to better understand their historic patterns and assess biodiversity without further depleting populations, there is a necessity to turn to museums. Genetic and genomic techniques and their applications have proved to be a useful complement to morphological taxonomy2. The array of molecular approaches is effective in delineating evolutionary boundaries, it does not replace the critical role of classic morphological taxonomy3. Indeed, classic taxonomy remains a fundamental discipline at the basis of biological sciences, which the advent of molecular approaches has expanded upon and enriched. For example, the molecular taxonomy has facilitated the description of delimitating species boundaries and for phylogenetic reconstruction4, understanding the species boundaries5,6 and reconstructing the species lineage79.

Insects represent 40% of all living species, this massive biodiversity has made museums a critical place for studying them, as they foster historic material on a global scale10. Molecular taxonomy, as mentioned above, has been identified as a valid instrument for museum material, making the scope of knowledge more accessible11. Recently, the molecular studies to support classical taxonomy have constantly increased but they are focused especially on individuals collected for genetic analysis and preserved in a specific way to not degrade the DNA (i.e. avoid the use of ethyl acetate)7. Unfortunately, this does not allow the use of rare or extinct species and type material. The current approaches to DNA extraction require the total or, at least, partial destruction of the individuals3,7,12, approaches that should not be used for unique specimens or specimens attributable to voucher collections. To defeat these limitations, in 2007 Gilbert and colleagues developed a method to extract DNA on museum specimens without external morphological damages13. This method is very successful for beetles because they are covered by a robust, hairless exoskeleton, but is unsuitable for bees, due to the treatment with digestion buffer which leads to depigmentation of hair, an important morphological character for bee taxonomy. Our project aimed to test the above-mentioned Gilbert et al.13 method of non-invasive DNA extraction on bee specimens, a group that has 20,000 species and is critical for biodiversity and ecosystem functions.


mtDNA was successfully amplified (Fig. 1) and sequenced from 8 of the 9 bee species investigated, regardless of when they were collected (Table 1). The only species we could not amplify DNA from was Ceratina cucurbitina (Rossi, 1792).

Figure 1
figure 1

Gel electrophoresis of PCR performed on DNA amplified from the investigated bee species. As a negative control double distilled RNase-DNase-free water. Light bands also could mean low DNA concentration.

Table 1 Details of bee individuals analyzed in this investigation.

No amplification was obtained from the negative control and no DNA sequence highlighted evidence of contamination or alteration, and their identity was confirmed through the BLAST analysis and the alignment with sequences deposited in the Barcoding of Life Data system (BOLD system) (

Sequences were deposited in GenBank (OL961135, OL966966, OL966999, OL967010, OL979169-OL979174, OL979211-OL979218, OL982531-OL982538, OL984023-OL984030; OL986022-OL986029).

Current non-destructive methods of total immersion of the individual in buffer (Fig. 2) alters the colouration of the bee's hair and diagnostic characters. Whereas our extraction method using swabs does not exhibit significant external damage or colour change, thus validating its use on important specimens’ the remainder is not necessary as is implied (Fig. 3). To confirm that, all investigated specimens retained all diagnostic characters for recognition at the species level, after careful post-analysis microscopic examination. This supports the potential of the proposed method.

Figure 2
figure 2

Dorsal and ventral comparison between the post-treatment in two Bombus terrestris individuals using the Gilbert et al.13 (A,B) and the here proposed (C,D) methods.

Figure 3
figure 3

Ventral pictures of four bee individuals before (AD) and after (a,b,c,d) the treatment using the swabs soaked with the digestion buffer. (A)/a Xylocopa violacea (L., 1758) (collected in Emilia-Romagna region in 2011), (B)/b Andrena flavipes (Panzer, 1799) (collected in Tuscany region in 1994), (C)/c Megachile centuncularis (L., 1758) (collected in Umbria region in 1996), (D)/d Andrena lagopus (Fabricius, 1775) (collected in Latium region in 1996).


Before this investigation, only a few studies showed evidence for the possibility of extracting DNA from bee specimens non-destructively. These methods required puncturing the exoskeleton, grinding body parts, or immersion in a buffer, all of which result in the destruction of the specimen in some capacity14, while the full immersion of the specimens in the digestion buffer13. This method did not confer any external damages to the specimens, but is not applicable for hairy insects, like bees, because the digestion buffer causes the depigmentation of hairs. We proposed here a non-invasive method that does not cause any external damages and can be done without removing the entomological pin, decreasing further damages. The method consisting of the full immersion of bee individuals in the buffer was just used to perform the DNA barcoding of a new Megachilidae species, Trachusa vietnamensis15. Although the extraction was successful, it was very complicated to restore the individual to an acceptable initial condition. Several drying processes were necessary to avoid changing the appearance and diagnostic characteristics, which took a total of 3 days. This makes the proposed method an important aid for the molecular identification of bee specimens.

As previously reported, in beetles the digestion buffer acts to liberate the DNA from the mouth, anus, spiracles, sclerites, ectodermal glands, and possibly broken setae, in pinned beetles, the man-made opening in the left elytron and pterothorax13. It remains unclear why it has not been possible to extract DNA from C. cucurbitina individuals, though it could be related to the size of bee individuals, as it is the smallest investigated species. The fragility along with size could also pose a problem, as in the attempt to not break the specimens, the microbiological swab was passed too gently which did not produce the required results. It is necessary to investigate this question on as many bee species as possible to know the feasibility of the proposed method across taxa.

Although DNA degradation increases over time16,17, the mitochondrial DNA remain amplifiable through PCR for a long period13,17. For this reason, we were able to extract DNA from investigated individuals, promoting non-destructive methods for bees. On the other hand, the possible DNA degradation can be caused by chemical reactions with ethyl acetate or ethyl alcohol (compounds usually used to kill bees upon collection)1820, which seems to affect DNA integrity and its. This degradation might be probably also linked to no DNA sequences obtained from all investigated Ceratina individuals.

The efficacy of this proposed method is also improved by the results obtained from the sequence analysis. Each obtained sequence matched (from 95 to 100% Identity) with COI sequences deposited. None of these sequences amplified a Wolbachia sp., a problem recently highlighted for the barcoding of bees21, which further highlights the robustness of our method.

Given the demand for the application of molecular taxonomy on museum specimens, our method can be effective. The museum collections preserved rare species and individuals collected in habitats that have changed through time, which could give not only important taxonomic and phylogenetic information but also ecological and evolutionary data. The next approach will be to test the method with museum specimens sampled more than 30 years ago, to verify its feasibility for rare old materials. This method minimizes the risk of damaging the specimen, critical for the future of the field.

The proposed methods could increase the taxonomic information on bee individuals preserved in museums and historical entomological collections, as many of these materials are very rare and not yet investigated. Although we considered hairs as a possible source from which DNA was extracted, it cannot be ruled out that this method is also effective for even other arthropods. We have focused mainly on bees since that is our subject of study, but we hope that this method will be useful for the whole entomological research, implementing knowledge on species that are present in rare or even extinct. Due to the high heterogeneity of the insect class, it is impossible to define a single method for DNA extraction from preserved individuals in collections. This, therefore, makes it necessary to test its efficacy across taxa.


From the entomological collection of the Research Centre for Agriculture and Environment (CREA-AA), seventy-two dry bee specimens belonging to nine common and widely distributed species among 4 families, between 2021 to 1992, were selected for DNA analysis and sequencing (Table 1). All precautions have been taken to avoid environmental contamination, sterilizing the working tools and the worktop after each processed sample, and molecular works were performed in sterile conditions under a laminar flow hood. A specific digestion buffer was used for the analysis, modified from Gilbert et al.13, consisting of 5 mM CaCl2, 2% sodium dodecyl sulphate (SDS), 65 mM dithiothreitol (DTT), 450 µg/ml proteinase K, 150 mM Tris buffer pH 8 and 100 mM NaCl.

The procedures were schematized in Fig. 4. A sterile microbiological swab, previously soaked in the digestion buffer for 5 min, was gently rubbed twenty times (in a total of 2 min) over the sternites of each investigated bee (Fig. 5). Due to the removal of locality and identification tags resulted very complicated, it was chosen to work on the sternites because these areas usually have few diagnostic characters related to hair. This procedure makes the proposed method even more specifically to avoid visually damaging the specimens.

Figure 4
figure 4

Schematic description of the experimental procedures.

Figure 5
figure 5

Application of the proposed extraction method on the Bombus terrestris sternites.

Each swab was soaked in a 2 ml microtube, filled with 1 ml digestion buffer, and incubated for 18–22 h at 56 °C with gentle agitation. After the treatment, a sterile microbiological swab soaked with 100% EtOH was gently rubbed several times over the sternites to stop further digestion. Finally, the individuals were air-dried and replaced back in the collections.

Before DNA extraction, the swabs were removed from each 2 ml microtube, and DNA purification was performed using a phenol:chloroform extraction (Ultrapure™ Phenol:Chloroform:Isoamyl Alcohol, ThermoFisher Scientific, Waltham, MA, USA), following Gilbert et al.13. Briefly, 20 µg glycogen, 0.6 volumes 100% isopropanol and 0.1 volumes 3 M Sodium acetate (pH 5.2) were added, and the microtubes were immediately vortexed softly and centrifuged at room temperature at maximum speed (1400 g) for 30 min to collect DNA as a pellet. The supernatant was then removed, and the pellet was washed twice in 1.5 ml 85% ethanol, air-dried at 65 °C, and resuspended in 100 µl RNase-DNase-free water. The obtained DNAs were quantified using the spectrophotometer Infinite 200 PRO NanoQuant™ (TECAN Life Technologies, Männedorf, Switzerland) and placed at − 20° until the analysis. For all of these processes, double-distilled RNase-DNase-free water was used as the negative control.

The extracted DNAs were analysed by PCR to amplify the mtDNA region. Primers amplified a 710-bp fragment within the highly conserved region coding for the Cytochrome C oxidase subunit I (COI) gene: LCO1490 (5′-GGTCAACAAATCATAAAGATATTGG-3′) and HC02198 (5′-TAAACTTCAGGGTGACCAAAAAATCA-3)22. The PCR was performed in 25 µl of volume using HotStarTaq Polymerase (Qiagen, Hilden, Germany) following manufacturers’ instructions using 5 µl of DNA, forward and reverse probes (500 nM). The PCR assay was performed on Applied Biosystems® 2720 Thermal Cycler (ThermoFisher Scientific) and samples were amplified, after an initial activation at 95 °C for 15 min, through 35 cycles (1 min at 95 °C, 1 min at 40 °C, and 1.5 min at 72 °C), followed by a final extension at 72 °C for 7 min. All amplicons were visualized on a 1.5% agarose gel. The obtained amplicons were purified using ExoSAP-IT Express (ThermoFisher Scientific) and they were sequenced throughout the standard Sanger methodology (BMR Genomics, Padua, Italy). The obtained sequences were analysed using BioEdit23 to create the consensus one aligning forward and reverse sequences and BLAST (using megablast algorithm)24 and compared to sequences deposited in the Barcode of Life Data Systems (BOLD Systems)25.