Health Status of Honeybee Colonies Differing in Genetic Intra-Colonial Diversity

The study was performed at the Research Institute of Horticulture, the Apiculture Division in Puławy, Poland. Sister queens originating from a Carnica commercial line (M) were reared and then instrumentally inseminated to obtain two levels of genetic diversity in the offspring within colonies. The queens from one group (SCS-single colony semen) were inseminated with semen collected from drones from one of thirty paternal colonies belonging to the same strain as the queen (M) and additionally from two unrelated strains (N) and (G), ten colonies each strain. This gave three sub-groups indicated below as SCS-M (n=17), SCS-N (n=16) and SCS-G (n=18), and each combination was represented by one or two queens. The queens from the second group were inseminated with semen collected and mixed from drones coming from each of the thirty colonies (the MCS - mixed colony semen, n=51). Preparation of the semen used for insemination is described in detail in the publication by Gerula et al. (2014). When oviposition started, the queens of the two groups were introduced into the newly created colonies. A total of 102 colonies were set in Dadant hives with wax foundation frames. The colonies of the two experimental groups were randomly placed in two apiaries (est. 2009) located in Wola Bukowska (W) 51°40′09″ N 22°21′22″ E and Sielce (S) 51°26′23″ N 22°04′46″ E, 50 km apart.

In late summer at the beginning of experiment in 2009 and in 2010–2011 Varroa destructor mites were controlled with amitraz or flumethrin strips for about eight weeks and oxalic acid was applied during the broodless period usually in the middle of November. Throughout the Varroa control period, dead mites on the bottom board were counted. Bee colonies were grouped into three classes according to mite infestation: 1- up to 200 mites, 2- from 200 to 400 mites and 3- over 400 mites found on the bottom boards during Varroa treatment. Data from 2009 is not included in the paper because later assessed queens were introduced to the colonies in late July 2009.

In the spring of each year from 2010 to 2012, the degree of infestation of bees by microsporidia Vairimorpha spp. was examined microscopically. Samples were taken from winter debris and the dead bees were counted. Then, in order to estimate the numbers of Vairimorpha spp. spores for every colony, by twenty whole bees were grounded in 20 ml of sterile distilled water and microscope samples for examination were prepared (Pohorecka et al., 2011). Three independent samples (twenty bees) for each experimental colony were prepared. The number of Vairimorpha spp. spores per bee was estimated through the use of Olympus BX 51 light microscopy and a haemocytometer (Cantwell, 1970; Human et al., 2013). For each spore suspension, the averages of three estimates of intensity were used. Based on the number of spores, bee infection in each colony was determined as low <1 mil/bee, medium 1–5 mil/bee and high above 5 mil/bee.

In the autumns of 2010 and 2011, bee samples (ca. 60) were collected for testing to identify bee viruses accompanying Varroa destructor infestations. The tests considered the presence of ABPV and DWV genetic material. The study was conducted by RT-PCR qualitative analysis using bee heads. For each colony, the heads of all workers were pooled, crushed in liquid nitrogen and further submitted to RNA isolation with Total RNA Mini Kit (A & A Biotechnology, Gdansk, Poland). According to manufacturer's protocol, the powder was suspended in phenosol. RNA was reverse transcribed to produce cDNA using Thermo scientific, Revert Aid First Strand cDNA Synthesis Kit. PCR analysis for the presence of viruses was done using Taq PCR Core Kit (Qiuagen) The thermal cycling profiles were set as follows: the mixture was heated for 3 min at 94°C, 40 PCR cycles at 94°C for min, 55°C for 1 min and 72°C for 1 min. The reactions were completed by a final elongation step at 72°C for 2 min and the next step was cooling down to 4°C. The PCR products were electrophoresed in a 1.5% agarose gel stained with ethidium bromide.

The multi-factor ANOVA model was used to evaluate interaction between the groups of colonies, locations, year and Varroa infestation. Number of mites in each colony was transformed according to (log10) formula. The LSD test was used to identify the homogeneous groups, and ANOVA was used to determine how Varroa infestation, Vairimorpha and viruses' infection influenced the number of dead bees in winter debris. Kruskal-Wallis's test was used to determine how Varroa infestation, Vairimorpha spp. and viruses' infection affected nest adjustment in the spring. Spearman's rank correlation coefficients were calculated to find the relationship between Varroa infestation and indicators of weakened colonies in the spring (number of dead bees, nest adjustment). Differences in the frequency of Vairimorpha infected colonies (scale 0–3) and colonies infected by the viruses in experimental group of colonies were analysed based on Pearson's χ² test. All statistical calculations figures were performed using Statistica 13 (TIBCO Software Inc. (2017).

An average of 410 dead mites were found on the bottom boards during the entire period of parasite control in bee colonies n=122 (Tab. 1). In the SCS group colonies, there were on average 352 mites, while in the MCS group colonies significantly more, 456 mites, were found (F_{(1, 114)}=4.151, p=0.043). Bee infestation was significantly lower in apiary W, as evidenced by the number of dead mites on the bottom boards than in apiary S (F_{(1, 114)}=12,10, p<0.001). In 2010, infestation was significantly higher than in 2011 (F_{(1, 114)}=4.64, p=0.033) and on average there were respectively 455 and 345 mites. ANOVA was used to note the interaction effect of the factors of the year of study and the location of the apiary on the bee infestation (F_{(1, 114)}=48.852, p=0.035) (Fig. 1).

Fig. 1.

When we divided the SCS group, with homogeneous mating, into SCS-M, SCS-N and SCS-G according to the origin of the paternal colonies, similar numbers of Varroa mites were found on the bottom boards after late summer - autumn treatments in each experimental subgroup and MCS group in 2010 and 2011 (F_{(3, 118)}=1.559, p=0.2) (Tab. 2).

Varroa infestation was not found to be associated directly with the weakening of bee colonies after winter. Bee colonies from each class was similar, in regard to the number of dead bees with the winter debris (F_{(2, 91)}=0.552, p=0.57) as well in terms to the Kruskal-Wallis nest adjustment (H_{(2, 91)}=5.36, p=0.68) (Tab. 3). Relationships between Varroa infestation and the number of dead bees in spring, defined as Spearman rank correlation coefficients, were not significant, while the relationship between Varroa infestation and nest adjustment was significant and unexpectedly amounted to −0.23.

In the spring of each year, a total of 182 bee samples were tested (Tab. 4) to examine the rate of bees infected with Vairimorpha spp. In 2010 and 2011, a lower percentage of colonies infected with Vairimorpha was found at 40% and 39%, respectively, compared to the year 2012 when the percentage increased dramatically to 80%. In 2010, in the slightly more SCS group colonies were observed to be free of the Vairimorpha (71%) than in MCS group colonies (54%). Colonies with high infection (above 5 mil/bee) in both groups accounted for 70%.

In the MCS group, only 46% out of thirteen positive samples were determined as highly infected, and in the SCS group, as many as 71% out of fourteen positive samples, were thus identified. In 2012, Vairimorpha-free colonies in the MCS group accounted for 18%, while only 9% in the SCS group. The percentages of colonies with a high infection level were 60% and 72%, respectively. Based on data from all years, the number of infected and healthy colonies was found to be similar in both experimental groups (Pearson χ²=3.581, df=3, p=0.31) (Fig. 2). In colonies, where Vairimorpha spp. infestation was tested at least twice, some colonies (n=58) had a recovery rate of 27.6% and new case of disease rate of 39.6%. The number of colonies that were always healthy or always sick were 18.9% and 13.8%, respectively. MCS colonies recovered more frequently than those from the SCS group, but there were no significant differences (Pearson χ²=3.721, df=3, p=0.29) (Fig. 3).

Fig. 2.

Fig. 3.

The winter loss of bees found on the bottom boards in both healthy and infected colonies did not differ significantly and averaged 1174 individuals F_{(3, 174)}=0.240, p=0.86) as well in terms of the nest adjustment Kruskal-Wallis (H_{(2, 91)}=5.36, p=0.68) (Tab. 5).

In 2010, out of seventy-one bee samples, ABPV was detected in twenty-three samples (nine in the SCS and fourteen in the MCS group), while DWV was detected in three samples (two in the SCS group and one in the MCS group). Neither virus was detected jointly in any bee sample. In 2011, out of forty-five bee samples tested, ABPV was detected only in one sample (SCS group), while DWV was detected in as many as twelve samples (five in SCS group and seven in MCS group) (Tab. 6). In one sample from a single colony, both viruses were identified, and in the previous year, no viruses were found in this colony. There was no interaction between the presence of the viruses and the experimental group (Pearson χ²=1.525, df=3, p=0.67) (Fig. 4), and the listed bee viruses were present in the colonies only temporally. None of the colonies which were infected by the viruses in 2010 were still infected in 2011. The virus had persisted only in three colonies over the two years of the study, ABPV infection in the first year while DWV infection in the second year.

Fig. 4.

Varroa infestation of bees in the period prior to sampling for testing did not differ significantly in either healthy colonies or those in which the viruses were detected, F_(2,108)=2.15, p=0.12 (Tab. 7). No relationship was found between the presence of ABPV virus in the bees' bodies and the loss of bees during the subsequent overwintering period. However, the number of bees died in winter, in colonies with DWV was observed to amount to 2047 bees and to be significantly higher than in healthy and ABPV-infected colonies, 1251 and 1035 bees respectively F_(2,91)=6.77, p=0.001 (Tab. 7). By analogy, when there were more bee deaths in winter, more combs were removed from nests in spring after the first inspection. Significantly more combs were removed from colonies with DWV (33.3%) than from healthy and ABPV-infected colonies, (11.2 and 7.7%, respectively) Kruskal-Wallis Test H_{(2, N= 85)}=9.05, p=0.011 (Tab. 7).