**3. Discussion**

This study is the first attempt to constitute an epidemiological baseline regarding the geographical distribution patterns and prevalence of four honeybee viruses in wild bumblebee samples across the territory of Croatia. Results showed that a higher prevalence of BQCV and DWV was determined in tested wild bumblebee samples collected in the continental part of country where the honeybee colonies' density is higher, in comparison with locations at the Adriatic coast and on islands. In particular, high BQCV prevalence was determined at all locations, except in samples originating from the Dalmatian islands of Ugljan and Pašman. Similar results were presented by Alger et al. (2019) and Toplak et al. (2020), which provided support to the observations that honeybee viruses are probably spilling over from managed honeybee colonies to wild bumblebees through visits to the same floral sources [10,11].

Since the current study does not include bumblebee species differentiation, we are not able to discuss possible differences in the species-specific vulnerability of bumblebees to tested honeybee viruses. However, according to Toplak et al. (2020) the different species of bumblebees tested in two consecutive years showed high variability in prevalence for different viruses and species. In addition, their results showed high variability for different species of bumblebees in a short period in the same geographic area [11]. Nevertheless, in our study, the important first data were obtained regarding the prevalence of four honeybee viruses on bumblebees from our region.

This research confirmed that genetically identical or closely related honeybee strains of BQCV and DWV were also identified among tested bumblebees, collected from the same location (geographic area). The high genetic identity with previously determined Slovenian BQCV and DWV strains is not surprising because of the neighboring area and historically long tradition of honeybee pasturing during spring and summer time. Furthermore, those two countries have a good beekeeper connection via trade. The identification of a group of 17 genetically closely related BQCV-positive samples, which were collected throughout territory of Croatia, is probably the result of recent years' transmission events. The opposite observation with the identification of relatively high diversity of BQCV strains among bumblebees is the result of the persistence of this virus for a long period in the population and was like that observed in a previous study in Slovenia [10]. Although honeybee positive samples from Croatia were not included in this study, the genetically very similar strains of BQCV and DWV than identified in bumblebees could be expected in honeybees from the same territory. This was confirmed with the identification of DWV-positive sample Bombus-Mir/2018 (MW488261), which has 98.09% nucleotide identity with the two closely related DWV strains YU4 (JF346630) and YU5 (JF346631), collected more than 30 years ago in Croatia, both DWV-positive samples were collected in *A. mellifera*, from a location near Zagreb city [19].

Interestingly, ABPV and CBPV were detected only in one sample from one location each, with a very low viral load copy number. Due to low positive samples of both ABPV and CBPV, the sequencing and phylogenetic analyses were not possible in this study. In contrast, ten examined bumblebees' samples were DWV-positive (37.40%), which is significantly higher infection prevalence than results presented in previously published studies, where the range was 2.70 to 11% positive bumblebees' samples [9–11,15]. However, general observation for both BQCV- and DWV-positive samples showed that low copy numbers were identified in each pool of five bumblebees' positive sample, suggesting that these viruses are present in bumblebees, but they may have limited impact on bumblebee pathology. Tehel et al. (2020) reported higher viral titers of BQCV, DWV genotype A, and DWV genotype B in bumblebees after experimental inoculation of a pathogen by injection

in comparison with oral inoculation [20]. Namely, among more than 30 honeybee-infecting known viruses [21,22], three are characterized by specific clinical symptoms: CBPV, DWV, and SBV [23]. BQCV and ABPV with its belonging complex can show alterations in the morphology of developmental stages of honeybees and adults' behavior. Furthermore, some can be present as inapparent or subclinical infections [24].

It is not yet completely ascertained if field viruses' strains are able to cause clinical manifestation in bumblebees. It is also not clear which environmental stressors are promoting factors for converting an asymptomatic infection into symptomatic and overt. In this study, all collected bumblebees were without visible morphological or behavioral changes, so they were considered clinically healthy. Although the virus presence could also be detected in individual bumblebee tissues or organs, this approach was not applied in our study and was assessed firstly to define the prevalence and diversity of four tested viruses. However, for further research it would be good to use individual bumblebee tissues for the estimation of individual virus tropism for specific tissue and for evaluating the possible effects that those viruses may have on their bumblebee hosts. In addition, according to Manley et al. (2019), honeybee parasitic mite *V. destructor* drives DWV prevalence and titer in honeybees and wild bumblebees [25]. Similarly, experimental injections of DWV under laboratory conditions or natural direct inoculation through *V. destructor* host feeding into insect body haemocoel causes an increase in prevalence and virulence in honeybees [26,27].

In our study, BQCV-positive samples were determined in very high prevalence (24 positive samples/27 total number of samples), which is contrary to the published findings of Dolezal et al. (2016), where same virus was detected extremely rarely in wild bees [28]. However, a recent publication from Slovenia, with the identification of genetically identical strains of ABPV, BQCV, SBV, and Lake Sinai virus (LSV) supports the observation of our study, that identical strains are present in honeybees and bumblebees [10]. Previously published data from Croatian honeybee samples originating from 82 apiaries located in 20 different districts showed a wide spread of honeybee viruses, with 9.75% of CBPVpositive samples, while ABPV, BQCV, and DWV were found in 10.97%, 40.24%, and 95.12% of tested apiaries, respectively [17]. Simultaneous infections with a maximum of two different viruses were detected in 10 (37.03%) of 27 bumblebee samples, and this was lower than the previously observed 64.6% of multiple infections among tested honeybees in Croatia [18].

#### **4. Materials and Methods**

#### *4.1. Field Sampling*

To determine the presence and the prevalence of four different honeybee viruses (ABPV, BQCV, CBPV, and DWV) in wild bumblebees (*Bombus terrestris*, *Bombus lapidarius*, *Bombus pascuorum*), 27 randomly selected bumblebee samples were collected from a total of 27 sampling locations from the territory of Croatia (Figure 4). Sampling was conducted during July and August 2018. At each site, five clinically healthy bumblebees were taken from flowers, representing one pool sample from each location. Each sample was marked with the number of the sampling location. Collected samples of bumblebees were stored under –70 ◦C until the start of molecular analyses.

#### *4.2. Molecular Analyses in Laboratory Conditions*

For RNA extraction purposes, each pooled sample consisting of five bumblebees' specimens was placed into Ultra-Turrax DT-20 tubes (IKA, Königswinter, Germany) with five mL of RPMI 1640 medium (Gibco, Paisley, UK) and incubated at room temperature for 30 min. Then, prepared samples were homogenized and centrifuged for 15 min at 2500× *g*. Two milliliters of supernatant were stored from each sample as a suspension for further extraction.

**Figure 4.** Sampling location sites (sample number) where a total of 27 samples of healthy bumblebees was collected during July and August 2018. The sample numbers represent different locations on the map of Croatian territory corresponding to the names of those locations, as indicated in accompanying table.

Primers, TaqMan probes, and quantification standards for ABPV [29], BQCV [30], CBPV [31], and DWV [32] were set with reagents according to previously published protocol.

Briefly, the total RNA was extracted from 140 μL of suspension from each sample by the QIAamp viral RNA mini-kit (Qiagen, Hilden, Germany) and recovered from a spin column in 60 μL of elution buffer. Reverse transcription with RT-qPCR assay was made in a single step using QuantiNova Pathogen + IC Kit (Qiagen, Hilden, Germany). The RT-qPCR mix consisted of 5 μL QuantiNova Master Mix, 2 μL 10× Internal (inhibition) Control (IC) Probe Assay, 1 μL IC (1:100), 4.5 μL deionized water, 1 μL forward primer (200 nM), 1 μL reverse primer (200 nM), and 0.5 μL probe (100 nM) and 5 μL of extracted RNA with a total of 20 μL final volume. Thermal cycling was performed on Mx3005P thermocycler (Stratagene, La Jolla, CA, USA) with the following conditions: 20 min 50 ◦C, 2 min 95 ◦C, followed by 45 cycles of 15 s 95 ◦C, 30 s 60 ◦C, and 30 s 60 ◦C. In each run, the positive control was included, prepared as a mixed suspension of previously determined positive field samples of four different viruses (ABPV, BQCV, CBPV, and DWV). The negative control was prepared and used in the same way as the positive, while each negative control consisted only of 160 μL of RPMI 1640 medium (Gibco, Paisely, UK) in aliquot.
