Apis mellifera caucasica Gorb. or the gray mountain Caucasian bee was first described in 1916 [1]. This bee has attracted the attention of beekeepers and scientists for its productivity, gentleness, and ability to pollinate red clover [2]. The habitat Apis mellifera caucasica is found in the Caucasus [1–3]. On the territory of Russia, it lives in the regions of the Southern and North Caucasus federal districts. The Krasnopolyansk experimental station, created in 1963, carries out its scientific and economic activities on breeding this subspecies of the honey bee [4]. At the beginning of the 20th century mass export began of A. m. caucasica to other regions of Russia and abroad [2]. The genome of A. m. caucasica was deciphered in 2020 by Chinese scientists using Pacbio sequencing technology (Genome assembly ASM1384120v1, https://www. ncbi.nlm.nih.gov/datasets/genome/GCA_013841205.1/). In addition, the complete mitochondrial genome A. m. caucasica is recorded in GenBank (GenBank ID: MN714160).

Based on morphometric and genetic differences, all known subspecies were divided into four evolutionary lineages: African (A), Western and Northern European (M), Eastern European (C), and Western Central Asian (O) [3, 5, 6]. Based on this classification A. m. caucasica belongs to the evolutionary lineage O along with A. m. anatoliaca, A. m. remipes, A. m. macedonica, A. m. cecropia and A. m. cypria [3, 6]. In 1993, L. Garneri et al. found that with the use of polymorphism of the intergenic locus tRNAleu-COII (or COI-COII) mtDNA can differentiate bees from evolutionary lineages A, M, and C [7]. This method, called the DraI test, is based on analysis of the lengths of restriction fragments of the intergenic locus tRNAleu-COII. The subspecies A. m. caucasica according to morphometric and genome-wide data belongs to lineage O, but at the same time has common haplotypes tRNAleu-COII with subspecies from the evolutionary lineage C (A. m. ligustica, A. m. carnica) [8]. Therefore, by analyzing this locus it is impossible to differentiate A. m. caucasica from subspecies from lineage C. However, there are studies indicating that subspecies from lineage C differ in haplotype frequencies tRNAleu-COII [9, 10].

The purpose of this study is to identify the genetic characteristics of the population of the gray mountain Caucasian bee using the analysis of microsatellite loci and the intergenic mtDNA locus tRNAleu-COII.

To analyze the polymorphism of microsatellite loci (Ap243, 4a110, A24, A8, A113, A88, Ap049, A28, and A43) and intergenic mtDNA locus tRNAleu-COII we formed three samples. The sample of A. m. caucasica (N = 90) is represented by bee colonies selected at the Krasnopolyansk experimental beekeeping station (Krasnodar krai, Sochi district, Krasnaya Polyana village) in 2008–2020. The sample of A. m. mellifera (N = 93) was selected in the Burzyansky district of Bashkortostan, as well as in the Perm krai. The sample of A. m. carnica (N = 118) was selected from apiaries in the Republic of Adygea, Orenburg oblast, Uzbekistan and Kazakhstan. From the studied samples, we selected 20 individuals of each subspecies to evaluate the haplotypes of the locus tRNAleu-COII using sequencing.

DNA was isolated from the thorax muscles of worker bees using a DNA-EXTRAN-2 kit (SINTOL LLC, Moscow). The quality and quantity of total DNA were analyzed on an Implen N60 spectrophotometer. The PCR mixture for ten samples with a total volume of 150 μL included 120 μL of distilled water, 15 μL of magnesium buffer, 3 μL of the DNTP mixture (concentration, 10 μm each), 5 μL of F-primer and R-primer (concentration, 10 pmol/μL) and 3 µL Taq polymerase. The PCR mode was 3 min 94°C, then 30 cycles with denaturation of 30 s at 94°C, annealing for 30 s at 49°C (for locus tRNAleu-COII) and 55°C (for microsatellite loci), elongation for 60 s at 72°C, and final elongation for 3 min at 72°C. To visualize the amplification products, electrophoresis in 8% polyacrylamide gel (PAGE) was used, followed by detection in the Gel Doc™ XR+ photosystem (BioRad, United States).

Data on microsatellite loci were used to determine the genetic structure of samples using the Structure 2.3.4 program with a given number of clusters from 1 to 10. The numbers of expected groups (K) were calculated in Structure Harvester [11, 12]. The analysis was performed using the Admixture model with information about the geographic localization of samples (LocPrior) with a Burn-In Period and MCMC equal to 10 000 and 100 000 replicates, respectively. The analysis results were processed in CLUMPP 1.1.2 using the FullSearch algorithm. The standard genetic distance Nei [13] was calculated in POPULATION ver.1.2.32.

Sequencing of 60 tRNAleu-COII PCR products was performed at Syntol LLC (Moscow). DNA sequences were edited and trimmed manually using MEGA software to obtain consensus sequences, which were then aligned with previously published sequences tRNAleu-COII using the Clustal W algorithm. An in silico tRNAleu-COII DraI sequence test was performed using Unipro UGENE ver. 36.

At the first stage of the work, we analyzed the polymorphism of nine microsatellite loci (Ap243, 4a110, A24, A8, A113, A88, Ap049, A28, and A43) in samples of A. m. caucasica, A. m. carnica, and A. m. mellifera. Figure 1 presents the results of cluster analysis of the studied samples. DeltaK, used to calculate the optimal number of clusters, peaked at K = 2 (deltaK = 1479.4) and at K = 3 (deltaK = 276.6). At K = 2 samples of A. m. caucasica and A. m. carnica entered into one cluster. Differentiation of three subspecies was observed when K = 3. The Nei standard genetic distance between A. m. caucasica and A. m. carnica was 0.269. In the study of S. Nikolova et al. the Nei genetic distance between these two subspecies was 0.384 based on polymorphism analysis of nine microsatellite loci [14].

Fig. 1.
figure 1

Genetic structure of the studied samples at K = 2 and K = 3.

Full size image

Our next step was to establish the haplotypes of the tRNAleu-COII locus. Haplotype sequences with geographic data are available at https://doi.org/10.6084/ m9.figshare.22348063. Unique Sequences of tRNAleu-COII were deposited in GenBank under accession numbers OR761847–OR76187.

Subspecies of three evolutionary lineages live on the territory of Russia: М (A. m. mellifera), С (A. m. carnica and A. m. ligustica) and О (A. m. caucasica). By separating tRNAleu-COII locus amplification products in 8% PAGE it was found that all samples of A. m. caucasica and A. m. carnica have an allelic variant Q, characteristic of representatives of the evolutionary lineage C (571 bp fragment), while all samples A. m. mellifera had an 825 bp PQQ in size, characteristic of the M lineage. Although we cannot distinguish between the C and O evolutionary lineages using tRNAleu-COII locus analysis, we were interested to know whether there were differences in haplotype frequencies between A. m. caucasica and A. m. carnica. Sequencing of tRNAleu-COII locus amplifications showed that out of 20 colonies of A. m. caucasica 16 had the C2j haplotype and two colonies had variants of the C2j haplotype–C2jd and C2jf. The remaining two colonies had haplotypes C2c and C2l. In the A. m. carnica sample the predominant haplotype was C2c. In A. m. mellifera the haplotype M17j was dominant. Table 1 presents a list of identified haplotypes.

Table 1. Haplotypes tRNAleu-COII in the studied samples A. m. caucasica, A. m. carnica, and A. m. mellifera
Full size table

Sequences of tRNAleu-COII from GenBank of A. m. caucasica (Ap018404.1 from Russia, OP404074.1 and OP404073.1 from Turkey, and MN714160.1 from the United States) also belong to the evolutionary lineage C. The Ap018404.1, OP404074.1, and OP404073.1 sequences belong to the C2j haplotype. MN714160.1 differs from C2j by only one nucleotide substitution. Samples of A. m. caucasica from Turkey from the study of C. Tozkar et al. also belong to the C2j haplotype [15]. Thus, based on these limited data, it can be assumed that the characteristic haplotype for the population of the gray mountain Caucasian bee is C2j.

It was previously assumed that subspecies from lineage C differ in haplotype frequencies for tRNAleu-COII [9, 10]. For the Italian bee A. m. ligustica the dominant haplotype is C1 [10, 16–18]. For A. m. carnica, according to some data, the dominant haplotype is C2c [9]; according to others it is C2j [10]. In the first case, haplotype identification was based on DraI RFLP analysis of amplifiers for tRNAleu-COII. The bees were collected in Slovenia, in the natural habitat of this subspecies. In the second case, bees were selected from different states of the United States, where bees of different subspecies had been introduced since the 17th century [2, 17, 19]. The authors of this work note that the C2j haplotype differs from other haplotypes of the C2 haplogroup. In addition, it is indicated that bees with the C2j haplotype entered the common cluster with the samples of A. m. caucasica from GenBank [10].

In addition to A. m. caucasica, subspecies A. m. anatoliaca (Turkey), A. m. remipes (Armenia), A. m. macedonica (North Macedonia, Greece, Bulgaria), A. m. cecropia (Greece), and A. m. cypria (Cyprus) also relate to the evolutionary lineage O. The tRNAleu-COII sequences of A. m. anatoliaca and A. m. macedonica, which are available in GenBank, also belong to C haplogroup. For example, the A. m. anatoliaca ON933877.1 and MN701760.1–MN701763.1 sequences belong to the C2 haplotype and FJ357798.1 belongs to C1. The haplotypes of A. m. macedonica differ by one substitution from the C2 haplotype. For A. m. remipes, A. m. cecropia, and A. m. cypria we found no tRNAleu-COII sequences.

The tRNAleu-COII mitochondrial locus is an informative and easy-to-use genetic marker [7]. It has been used to analyze a large number of honey bee populations around the world [10, 20–24]. However, due to the lack of rules for classifying tRNAleu-COII haplotypes a large number of incorrectly designated haplotypes have accumulated. A study [25] appeared in 2017 that describes the rules for the haplotype nomenclature; it carried out a revision of haplotypes from previously published works. In our study, we used these nomenclature rules to classify tRNAleu-COII haplotypes.

Analysis of the genetic structure of representatives of three subspecies of honey bees that are widespread in Russia showed a significant level of their differentiation even when using a small set of microsatellite loci. Estimation of the prevalence of the tRNAleu-COII haplotype in the three studied samples showed that for A. m. caucasica the predominant haplotype is C2j. Genetic studies of honey bee populations are not only of fundamental importance for the preservation of unique gene pools, but are also used by breeding farms specializing in breeding certain subspecies of honey bees.

Differentiation of subspecies using genetic and morphometric markers is complicated by the fact that populations of honey bees have hardly survived within the boundaries of their natural distribution due to the export of bees. Data on population studies of honey bees of different subspecies are very scattered. To solve this problem, first of all, it is necessary to find native, or local, populations of bees within the boundaries of their natural distribution, and then conduct genome-wide studies of these populations and perform morphometric analysis and analysis of the tRNAleu-COII intergenic locus.