A Complete Reference DNA Barcode Library for Austrian Bumblebees

Abstract

Bumblebees (Bombus spp.) are essential pollinators in natural and agricultural ecosystems but face increasing threats across Europe from habitat loss, climate change, and intensive land use. Austria hosts 42 recognized bumblebee species, yet comprehensive molecular data have been lacking. Here, we present the first complete DNA barcode reference library for the Austrian bumblebee fauna, generated as part of the Austrian Barcode of Life initiative. This reference library includes 586 partial mitochondrial COI sequences. DNA barcoding successfully identified all species, with distinct Barcode Index Numbers (BINs) and no BIN sharing observed, demonstrating its reliability as a complementary method to traditional morphology-based identification. Intraspecific genetic diversity was generally low, though B. jonellus exhibited notable mitochondrial structure with a complex biogeographic pattern. Our results underscore the value of DNA barcoding as a straightforward tool for accurate species identification and biodiversity monitoring, even for non-experts, while also highlighting cryptic genetic variation within widely distributed species. This reference library provides a robust framework for taxonomic, ecological, and conservation research, and supports future metabarcoding-based monitoring efforts in Austria and beyond.

1. Introduction

Europe hosts about 10% of the world’s bee diversity and the most prominent and diverse family of bees, the Apidae, consists of 561 species across Europe, including 68 species of bumblebees (Bombus spp.) [1]. According to the latest European red list of bees 23.6% of Bombus spp. are considered as threatened and 45.6% are showing a decreasing population trend [1]. Like other Apidae, bumblebees are pollinators and only few animal pollinators achieve such a numerical dominance as flower visitors. This makes them an important functional group for natural environments as well as for agricultural crops (e.g., [2,3,4]). Even though some species do specialize on certain plants, most bumblebees are quite unspecialized foragers. Bumblebees tend to be more abundant and diversified in habitats such as high mountains, boreal and arctic biomes. Due to their adaptations to cold climates, bumblebee species have been able to recolonize regions depopulated during the Ice Ages [5]. Therefore, they show a relatively high vulnerability as their preferred habitats are severely impacted by climate change [6,7].

For Austria, 42 species of bumblebees are currently recognized, and three historically documented steppe species (B. armeniacus, B. fragrans and B. laesus) are considered extinct. The most recent Red List of Austria considered 14 of these 42 species as threatened due to the constant decline of their preferred habitat [8]. Similar to other regions, species diversity of bumblebees in Austria reaches its maximum in higher altitudes such as montane and alpine zones. In addition to habitat loss due to climate change, especially for cold-adapted species (e.g., B. alpinus, B. mendax), agricultural land use seems to be a major threat to bumblebee species in Austria. Therefore, according to the latest red list of Austrian bumblebees, a high proportion of lowland species—which are often bound to open land habitats—are showing negative population trends [8]. On the other hand, the generalist species B. semenoviellus recently expanded its distribution from its core region in Eastern Europe and Western Asia towards the west, but can rarely be found in Austria [9]. Whether B. semenoviellus can establish itself in the Austrian fauna remains uncertain at this time. Bombus haematurus is another species that recently colonized eastern Austria from its native range in southeastern Europe [10] and is meanwhile considered an established part of the local bumblebee fauna in eastern Austria.

Due to the increasing public interest in the conservation of the genus Bombus—considering their status as important pollinators—several phylogeographic studies have been performed in recent years, uncovering cryptic diversity, lineage divergence, and phylogeographic structure within and among bumblebee species (e.g., [11,12,13,14,15,16,17]). These findings suggest that traditional morphological identification may underestimate true species diversity, underscoring the need for molecular tools in bumblebee taxonomy and conservation. DNA barcoding has become a widely accepted standard tool for species identification at the molecular level since its introduction for animals by [18]. Over the past two decades, numerous national and international projects under the umbrella of DNA barcoding have targeted a wide range of taxonomic groups, collectively contributing to the development of national and global reference sequence libraries essential for accurate species identification (e.g., [19,20,21,22]). As of September 2025, the Barcode of Life Data System (BOLD; https://www.boldsystems.org (accessed on 22 September 2025); [23]) contains approximately 10 million DNA barcodes. More than 50 countries and numerous national barcoding initiatives—such as the Austrian Barcode of Life initiative (ABOL, http://www.abol.ac.at (accessed on 22 September 2025))-contribute to this global database, with the aim of documenting native diversity and supporting taxonomic and general biodiversity research.

In this study, we provide a comprehensive reference DNA barcode library for all 42 currently recognized Austrian bumblebee species as part of the ABOL initiative. Emphasis was placed on broad geographic sampling to capture intraspecific genetic variation. Specifically, we aimed at (i) contributing reference data to BOLD, (ii) enhancing the current understanding of the extant diversity and distribution of the Austrian bumblebee fauna, (iii) assessing the discriminatory power of DNA barcoding for species identification—important for its potential application in metabarcoding-based monitoring programs—and (iv) exploring and discussing notable intraspecific genetic patterns in more detail.

2. Materials and Methods

The present dataset consists of two sources of DNA barcodes: 532 newly generated COI-sequences (658 bp) as part of the ABOL initiative and second, and 54 barcodes of Austrian Bombus spp. available from BOLD [16,24,25]. In total, the dataset comprised 586 DNA barcodes from 185 sampling sites across Austria (Figure 1). Bumblebees were collected by hand during the years 2019–2025 under the collecting permits A4/NN.AB-10200-5-2019, A4/NR.AB-10080-10-2022, ABT13-53W-50/2018-2, ABT13-198250/2020-9, AB13-523696/2023-9, FE3-NS-2825/2020 (010/2020), SP3-NS-3742/2021 (005/2021), SP3-NS-4142/2023 (005/2023), HE-NS-1922/2024 (005/2024), KL-NS-1978/2018 (003/2018), N-2020-68581/4-Has, RU5-BE-1489/001-2018, RU-BE-1489/002-2021, RU5-BE-1722/001-2021, RU5-BE-1489/003-2023, 08-NATP-845/1-2019(007/2019), 20505-05RI/1241/10-2023 and 20505-05RI/1241/12/5-2025, and determined to species level using the identification keys in [26,27]. Bumblebees were euthanized using ethyl acetate. Specimens were photographed and tissue samples (two legs per specimen) were stored in pure ethanol at −20 °C. The voucher specimens, mounted dry, were deposited at the Universalmuseum Joanneum, Graz, Austria or the Natural History Museum Vienna, Austria (a single B. flavidus and a single B. pomorum are in the private collection of Miriam Öttl and the collection of the BOKU University Vienna, Austria, respectively).

DNA extraction followed a modified Chelex protocol [28]. One part of the samples was sequenced using the cost-efficient Oxford Nanopore sequencing technologies (ONT), as nanopore-based COI sequences were previously shown to be of the same quality and consistent with Sanger sequencing [29]. PCR, purification and sequencing on Flongle flow cells (version R10.4.1; FLO-FLG114) followed [30]. PCR was performed using the primer pair LCO1490_JJ2/HCO2198_JJ2 [31], tagged with individual index oligos [32], with an annealing temperature of 49 °C. The remaining samples were sequenced following a standardized Sanger sequencing protocol [29], using the primer pair LCO1490_JJ2/HCO2198_JJ2, and visualized on an ABI 3500xL capillary sequencer (Applied Biosystems, Waltham, MA, USA), or sent to CCBD (Canadian Centre for DNA Barcoding, University of Guelph, Guelph, ON, Canada) for sequencing using the C_LepFol primer cocktail [33].

For the data generated by nanopore sequencing, basecalling was performed with Dorado (ONT), de-multiplexing and consensus barcode generation with ONTbarcoder 2.0 [34]. Sequences generated by Sanger sequencing were checked and edited in MEGA v11.0 [35]. Sequences were aligned using MUSCLE [36] in MEGA v11.0. The final dataset including 583 samples with sequence lengths of 425–658 bp is available on BOLD (https://doi.org/10.5883/DS-ABOMBAT2 (accessed on 1 October 2025)).

With the complete dataset, a Neighbor-joining (NJ) tree based on the Kimura 2-parameter (K2P) distance model and 1000 bootstrap replicates was generated with MEGA XI. Maximum intraspecific distances and minimum interspecific distances were calculated using the “Barcode Gap Analysis” tool (K2P distance model, pairwise deletion of ambiguous characters or missing data, BOLD aligner) implemented in BOLD (https://www.boldsystems.org (accessed on 22 September 2025)).

As it turned out that our B. jonellus samples constitute two distinct clades (see Section 3), we looked deeper into the intraspecific structure of this species. Therefore, we downloaded all DNA barcodes of this species available on BOLD [37,38,39] and added them to our ten B. jonellus samples. Sequences were aligned with MUSCLE in MEGA v11.0 and trimmed to the same length, resulting in an alignment of 167 samples with a length of 605 bp. A statistical parsimony network [40] was inferred and drawn with default settings in popart 1.7 [41].

3. Results

3.1. Phylogeny and Species Coverage

This study represents a reference DNA barcode inventory for all 42 Austrian bumblebee species, including 532 newly generated and 54 previously published DNA barcodes with a length between 425 bp and 658 bp. For the species B. mesomales, B. distinguendus and B. semenoviellus only one specimen could be obtained. All morphologically identified species were distinguishable through their DNA barcodes and were represented by a single Barcode Index Number (BIN). No cases of BIN sharing were detected. A NJ tree (Figure 2) was generated based on COI sequences; this tree confirms unambiguous species identification of all species. Intraspecific diversity within the NJ tree resulted mostly from differences in sequence length (and applying the pairwise deletion of missing data option).

A clear barcode gap was present for all species with a mean intraspecific distance of 0.13% and a mean interspecific distance of 12.78%. The distances to the nearest neighbor (DNN) varied from 1.87% to 9.83%. The species B. ruderarius, B. sylvarum and B. veteranus showed low interspecific distance to their nearest neighbor of 1.87% but the max. intraspecific distance within these species was between 0.16% and 0.62% (

Table 1. List of the species analyzed, including the barcode index number (BIN), number of specimens (N), the maximum intraspecific (Imax) and the minimum distance to the nearest neighbor (DNN).

Species	BIN	N	Imax	Nearest Neighbor	DNN
Bombus alpinus	BOLD:AAN9551	2	0	Bombus humilis	8.08
Bombus argillaceus	BOLD:AGH5555	5	0.31	Bombus ruderatus	2.97
Bombus barbutellus	BOLD:AAF7051	11	0.31	Bombus quadricolor	9.83
Bombus bohemicus	BOLD:AAD2530	25	0.62	Bombus norvegicus	8.77
Bombus campestris	BOLD:AAD8221	22	1.08	Bombus quadricolor	7.94
Bombus confusus	BOLD:AAJ7886	7	0.31	Bombus subterraneus	6.56
Bombus cryptarum	BOLD:ACE4135	16	1.01	Bombus lucorum	3.61
Bombus distinguendus	BOLD:ACF4670	1	na	Bombus subterraneus	3.29
Bombus flavidus	BOLD:AEI1064	2	0	Bombus norvegicus	5.27
Bombus gerstaeckeri	BOLD:AAF0186	6	0.15	Bombus hortorum	6.25
Bombus haematurus	BOLD:ADK0974	7	0.15	Bombus pratorum	6.76
Bombus hortorum	BOLD:AAD2566	39	0.15	Bombus ruderatus	5.74
Bombus humilis	BOLD:ABY7210	21	0.48	Bombus ruderarius	2.33
Bombus hypnorum	BOLD:AAB3227	13	0.48	Bombus pratorum	6.49
Bombus inexspectatus	BOLD:AEI3825	4	0.15	Bombus veteranus	3.3
Bombus jonellus	BOLD:AAD4941	10	1.24	Bombus pratorum	5.9
Bombus lapidarius	BOLD:AAC6016	34	0.16	Bombus sichelii	9.65
Bombus lucorum	BOLD:AAB1060	51	1.01	Bombus cryptarum	3.61
Bombus mastrucatus	BOLD:AAD8243	31	0.31	Bombus semenoviellus	8.93
Bombus mendax	BOLD:AAF5994	4	0.15	Bombus confusus	7.72
Bombus mesomelas	BOLD:AAD9869	1	na	Bombus pomorum	4.11
Bombus monticola	BOLD:AAA8078	17	0.31	Bombus pratorum	6.75
Bombus mucidus	BOLD:ABW9027	16	0.46	Bombus veteranus	4.6
Bombus muscorum	BOLD:AAD8159	3	0	Bombus pascuorum	3.12
Bombus norvegicus	BOLD:AAD4756	3	0.31	Bombus quadricolor	4.45
Bombus pascuorum	BOLD:AAC4378	22	0.62	Bombus ruderarius	2.49
Bombus pomorum	BOLD:AEI7391	2	0	Bombus mesomelas	4.11
Bombus pratorum	BOLD:AAD4735	21	0.31	Bombus pyrenaeus	5.22
Bombus pyrenaeus	BOLD:AAF0421	17	0.15	Bombus pratorum	5.22
Bombus quadricolor	BOLD:AAF6417	3	0.62	Bombus norvegicus	4.45
Bombus ruderarius	BOLD:ABY7209	13	0.16	Bombus veteranus	1.87
Bombus ruderatus	BOLD:AAJ7737	4	0.15	Bombus argillaceus	2.97
Bombus rupestris	BOLD:AAD8184	16	0.92	Bombus quadricolor	8.52
Bombus semenoviellus	BOLD:AAF6479	1	na	Bombus mastrucatus	8.93
Bombus sichelii	BOLD:ACF1761	17	0	Bombus semenoviellus	9.27
Bombus soroeensis	BOLD:AAB1065	46	0.31	Bombus mastrucatus	12.35
Bombus subterraneus	BOLD:AAA8260	2	0.31	Bombus distinguendus	3.29
Bombus sylvarum	BOLD:AAD2551	13	0.62	Bombus veteranus	1.87
Bombus sylvestris	BOLD:AAB4516	20	0.31	Bombus quadricolor	4.45
Bombus terrestris	BOLD:AAB1062	27	1.35	Bombus cryptarum	6.39
Bombus vestalis	BOLD:AAI8745	8	0.46	Bombus norvegicus	9.29
Bombus veteranus	BOLD:AAM3090	3	0.46	Bombus sylvarum	1.87

).

3.2. Intraspecific Structure in B. jonellus

In this study, ten specimens of the B. jonellus were collected and analyzed from across Austria, revealing three distinct haplotypes that group into two mitochondrial lineages, with a maximum intraspecific divergence of 1.24% (Table 1). Despite this genetic structuring, all Austrian B. jonellus specimens fall within a single BIN, indicating that these lineages do not meet the threshold for BIN-level separation. The haplotype network with all available DNA barcode data on BOLD shows 40 haplotypes that fall into three main lineages (Figure 3). These lineages correspond to an exclusive European, an exclusive North American, and an intermediate cluster that includes both samples from the Austrian Alps and Canada, with haplotype sharing between these geographically distant regions (Figure 3 and Figure 4).

4. Discussion

This study shows that all currently recognized 42 bumblebee species of Austria can be reliably identified using COI-based DNA barcoding. No BIN sharing was observed, and all species comprised only a single BIN, confirming the effectiveness of DNA barcoding as a tool for accurate species-level identification in the Austrian bumblebees. The addition of newly generated barcodes to BOLD contributes valuable reference data, especially for species like B. haematurus, B. inexpectatus, B. pomorum, and B. veteranus, for which only very few DNA barcodes are available from other European countries, or B. semenoviellus, whose presence in Austria was first recorded only in 2009 [9]. In particular the case of B. semenoviellus underscores the importance of maintaining updated and geographically comprehensive barcode reference libraries, especially for mobile taxa with expanding or shifting distribution ranges.

Our results also demonstrate the usefulness of DNA barcoding in resolving taxonomically challenging groups, such as the B. lucorum-terrestris complex [42]. This complex consists—in Austria—of the species B. lucorum, B. cryptarum, and B. terrestris, whose phenotypic species identification remains difficult and may lead (or has led) to misidentifications of historical records [7,43,44]. Although DNA barcoding has limitations in detecting deeper evolutionary relationships of bumblebees, it is, as in other taxonomic groups, a powerful and accessible tool for routine species identification, especially in ecological monitoring, biodiversity assessments, and citizen science initiatives (e.g., [45,46,47,48]).

Most species in our dataset exhibited low intraspecific divergence and high interspecific distances, consistent with expectation for ideal DNA barcoding performance. Only four species—B. campestris, B. jonellus, B. lucorum and B. terrestris—showed maximum intraspecific divergences (I_max) above 1% (Table 1). In the case of B. lucorum and B. terrestris, these values were largely attributable to variation in sequence length (several previously published DNA barcodes were considerably shorter than the 658 bp standard length). When only full-length (658 bp) DNA barcodes were considered, I_max dropped way below 1%, aligning with previous studies that reported limited genetic diversity in these species [15,44,49]. We observed comparatively high intraspecific diversity in B. campestris. Our data, however, do not show a clear geographic structure in Austrian B. campestris, suggesting that the high levels of intraspecific divergence seen in B. campestris are due to large population sizes. The retention of divergent haplotypes is a common feature of large and stable populations, where the effects of genetic drift are reduced and ancestral polymorphisms can persist over extended evolutionary timescales.

Notably, B. jonellus showed two distinct mitochondrial lineages within Austria, despite all individuals falling under the same BIN. Bombus jonellus is a circum-boreally distributed species, which is mostly associated with heathlands, mountain meadows and tundra habitats. This species shows an especially wide distribution for a non-managed bumblebee species and whilst it is common in the North of its distribution it is considered rare in southern parts of Europe, where its occurrence is restricted to higher altitudes [7,17]. When combining our data with all available BOLD data, this species formed three major haplogroups: (i) a clearly separated European lineage with samples from northern Europe and Central European lowlands (with the only exception of a single individual from the southeastern edge of the Alps), (ii) a North American lineage, and (iii) an intermediate lineage that include samples from both the Austrian Alps and Canada. This latter lineage even includes a haplotype shared between Austrian and Canadian bumblebees, suggesting a more complex biogeographic history than previously assumed.

Earlier studies proposed a relatively low mitochondrial variability in B. jonellus [38], but our findings (building upon the results of [17]) suggest a higher level of mitochondrial diversity and hint at historical population processes, such as postglacial recolonization or multiple glacial refugia [17]. The presence of shared haplotypes between Canada and Austria—especially in a species not known for human-mediated dispersal—raises intriguing questions about the origin and distribution of this particular lineage [17]. Using shorter COI fragments, we also identified specimens of the intermediate clade in western Siberia, alongside individuals of the European clade. Due to its specific habitat requirements, B. jonellus is largely absent from the lowlands of central and eastern Europe. As a result, its current distribution between the Alps and more eastern regions is patchy and may represent remnants of a once broader range in central and eastern Europe during glacial times. The haplotype network structure of the European clade suggests a long-term stable population, whereas the North American lineage shows a pronounced star-like structure, a pattern characteristic of a recent and rapid population expansion, likely following postglacial (re-)colonization events. Due to limited data, no definitive conclusions can be drawn regarding the intermediate lineage. To fully understand the intraspecific patterns of B. jonellus, future studies incorporating nuclear markers and more comprehensive geographic sampling, particularly in eastern Europe and Asia, are essential.

5. Conclusions

Bumblebees play a vital role as pollinators in both natural ecosystems and agriculture, and public interest in their conservation continues to grow. However, many species are experiencing significant population declines across Europe, particularly those adapted to colder environments, and climate change is expected to accelerate these declines by reducing suitable habitats and steppe species, whose habitat is increasingly being pushed back by the intensification of agriculture [1,6,7,50,51]. Addressing these challenges requires more than general awareness; it demands accurate species identification and a deep understanding of distribution patterns and genetic structure. This study contributes to these goals by providing a complete reference dataset of all bumblebee species recorded in Austria, supporting reliable identification and regional monitoring. More broadly, our findings highlight the need for larger-scale genetic research to uncover hidden population structures and evolutionary lineages within bumblebee species. Such insights are essential for effective conservation planning, particularly in the face of accelerating environmental change.