Seasonal dynamics of seven viruses and Varroa destructor in Apis mellifera colonies: a 4-year survey in Tokyo

Abstract

Long-term monitoring of viruses and Varroa destructor within the same honey bee colonies is rare in Japan. Thus, we conducted a 4-year fixed-point survey of three Apis mellifera colonies in Tokyo with monthly screening for seven viruses (using polymerase chain reaction) and Varroa infestation. Deformed wing virus was detected year-round, whereas sacbrood virus and black queen cell virus peaked in spring and summer. Israeli acute paralysis virus appeared sporadically, without seasonal trends. Varroa infestation rates (monitored during the latter 2.5 years) did not correlate with the number of detected virus types and showed no seasonality. Although viral load was not quantified, this long-term dataset provides valuable baseline data on local virus dynamics, providing baseline information on seasonal virus dynamics in managed honey bee colonies in Tokyo.

The Western honey bee (Apis mellifera) is an essential global pollinator that plays a key role in crop farming [1, 2, 16]. However, the maintenance of honey bee colonies has become increasingly challenging in recent years due to the combined impact of the ectoparasitic mite Varroa destructor and viral infections [8, 14, 15]. Varroa parasitism of honey bees has been associated with increased viral loads of several viruses, particularly deformed wing virus (DWV), which causes colony decline [8, 14, 15]. Previous studies conducted in other regions have demonstrated that long-term monitoring of honey bee viruses and associated pathogens provides valuable insights into seasonal dynamics, co-infection patterns [4, 5]. However, long-term continuous monitoring of both viruses and Varroa within the same colonies is extremely limited in Japan. Since honey bee viral dynamics are heavily influenced by local climate and beekeeping practices, such data are crucial for understanding the region-specific seasonal patterns and infection dynamics of viruses that infect honey bees.

We performed monthly monitoring of seven major viruses in three fixed colonies of A. mellifera maintained at the same apiary under the same management conditions in Tokyo for 4 years (January 2022–December 2025). Colony #1 was maintained on the rooftop of a building in Tokyo, where 3–6 colonies were typically managed. Colonies #2 and #3 were located in a wooded area approximately 600 m from Colony #1, where 6–10 colonies were typically maintained. The same colonies were monitored throughout the study period. Queens were routinely replaced approximately once per year as part of standard colony management. However, no queens or colonies were introduced from external apiaries during the study period, and no colony merging was performed, allowing continuous monitoring of the same colony lineages throughout the study. We also assessed Varroa infection rates during the latter 2.5 years. We aimed to clarify the detection frequency, seasonality, and coinfection patterns of these viruses, as well as the relationship between the number of detected virus types and Varroa infestation levels.

For each sampling event, 30 adult worker bees were collected from comb surfaces within the hive. When brood was present, bees were sampled primarily from brood frames. In the absence of brood, bees were collected from approximately the third comb from the outer side of the hive. The bees were homogenized in 3 mL of phosphate-buffered saline (PBS) using a mortar and pestle. Following homogenization, total RNA was extracted from the supernatant using the RNeasy Mini Kit (Qiagen, Hilden, Germany), following the manufacturer’s protocol. Five microliters of the extracted RNA were reverse transcribed using the PrimeScript IV 1st Strand cDNA Synthesis Mix (Takara Bio Inc., Kusatsu, Japan). The resultant cDNA was used as the template for multiplex polymerase chain reaction (PCR), performed with the Multiplex PCR Assay Kit Ver. 2 (Takara Bio Inc.) according to previously published procedures [3, 13, 18]. The primer sequences, expected amplicon sizes, and references are provided in Supplementary Table 1. The following viruses were detected using previously reported primer sets: DWV [18], sacbrood virus (SBV) [18], black queen cell virus (BQCV) [18], Israeli acute paralysis virus (IAPV) [18], acute bee paralysis virus (ABPV) [18], and chronic bee paralysis virus (CBPV) [18]; Kashmir bee virus (KBV) [3]. The honey bee EF-1α gene was used as an internal control to confirm successful RNA extraction and reverse transcription [13]. The PCR conditions consisted of an initial denaturation at 94°C for 1 min, followed by 35 cycles of 94°C for 30 sec, 56°C for 30 sec, and 72°C for 45 sec, with a final extension at 72°C for 10 min. PCR products were identified based on the expected amplicon sizes reported in the original studies [3, 13, 18]. Due to colony reduction, Colony #1 was not sampled in October 2023. Furthermore, Colony #1 collapsed in October 2025 and could no longer be sampled. Prior to colony collapse, increased Varroa infestation and a decline in colony population were observed. No obvious virus-associated clinical signs, such as deformed wings or paralysis, were noted during routine colony inspections. Robbing by Apis cerana japonica was observed during the period immediately preceding collapse of Colony #1.

Varroa infestation rates were measured monthly (April 2023–December 2025) using the sugar-roll method [7]. Briefly, approximately 100 adult worker bees were collected into a jar and coated with 7 g of powdered sugar. After 1 min, the jar was inverted and shaken for 1 min to dislodge mites together with the powdered sugar. The number of fallen mites was counted, and the infestation rate was calculated as the number of mites divided by the number of bees ×100. Chemical control of Varroa was mainly performed in summer and late winter using amitraz strips (Apivar; Véto-pharma, Palaiseau, France). Amitraz treatment was applied to all three colonies, although the timing occasionally differed by a few days among colonies. In addition, integrated pest management practices, including drone brood removal and queen caging, were applied. Due to colony reduction, Colony #1 was not sampled in October 2023 and after September 2025.

Among the 141 sampling events (Colony #1: 12, 11, 12, and 10 months over four years; Colonies #2 and #3: 12 months each year), DWV was the most frequently detected (137/141, 97.2%), followed by SBV (45/141, 31.9%), BQCV (44/141, 31.2%), and IAPV (23/141, 16.3%). CBPV was only detected once, and neither ABPV nor KBV was detected. Regarding coinfection, no viruses were detected in 2/141 (1.4%) samples, 1 in 69/141 (48.9%), 2 in 34/141 (24.1%), 3 in 31/141 (22.0%), and 4 in 5/141 (3.5%). There were no sampling events where ≥5 viruses were detected (Table 1).

Table 1. Distribution of the number of codetected virus types per colony during the study period.

Number of virus types	DWV	SBV	BQCV	IAPV	CBPV	Number of samples	Frequency (%)
0	-	-	-	-	-	2	1.4
1	+	-	-	-	-	68	48.2
	-	-	+	-	-	1	0.7
2	+	+	-	-	-	10	7.1
	+	-	+	-	-	12	8.5
	+	-	-	+	-	11	7.8
	-	+	+	-	-	1	0.7
3	+	+	+	-	-	24	17.0
	+	+	-	+	-	5	3.5
	+	-	+	+	-	1	0.7
	+	-	-	+	+	1	0.7
4	+	+	+	+	-	5	3.5
Total						141	100

DWV, deformed wing virus; SBV, sacbrood virus; BQCV, black queen cell virus; IAPV, Israeli acute paralysis virus; CBPV, chronic bee paralysis virus.

Figure 1 presents monthly counts of colonies testing positive for DWV, SBV, BQCV, and IAPV (the four most frequently detected viruses), along with six-month centered moving averages shown as line graphs. Following one-way analysis of variance (ANOVA) and Tukey’s multiple comparison test, seasonal categorization−growth (spring, March–May), peak (summer, June–August), pre-overwintering (autumn, September–November), and overwintering (winter, December–February)−was applied (Fig. 2). Significant seasonal differences were observed for SBV and BQCV, with higher detection rates in spring and summer (P<0.05). For IAPV, detection rates were significantly higher in autumn than in winter (P<0.05). No significant seasonal patterns were observed for DWV.

Fig. 1.

Monthly number of virus-positive colonies for the four most frequently detected viruses, along with six-month centered moving averages shown as line graphs. (A) DWV, (B) SBV, (C) BQCV, and (D) IAPV. Data represent monthly monitoring results from January 2022 to December 2025 across three fixed colonies. Each panel shows the count of colonies that tested positive for the respective virus in each month. Data points marked with an asterisk indicate a sample size of 2 colonies. DWV, deformed wing virus; SBV, sacbrood virus; BQCV, black queen cell virus; IAPV, Israeli acute paralysis virus.

Fig. 2.

Seasonal variation in virus prevalence. Seasons were categorized as growth (spring, March–May, n=36), peak (summer, June–August, n=36), pre-overwintering (autumn, September–November, n=34), and overwintering (winter, December–February, n=35). Different letters indicate significant differences among seasons within each virus (P<0.05). Error bars represent 95% confidence intervals. DWV, deformed wing virus; SBV, sacbrood virus; BQCV, black queen cell virus; IAPV, Israeli acute paralysis virus.

Figure 3 presents temporal changes in Varroa infestation rates. Multiple comparisons using Games–Howell test did not reveal seasonal differences in Varroa infestation rates (P>0.05). Furthermore, Spearman rank correlation analysis found no significant association between the number of detected virus types and Varroa infestation rates (ρ=−0.0470, P=0.6619).

Fig. 3.

Monthly changes in Varroa destructor infestation rates (%) from April 2023 to December 2025. Varroa infestation rates measured using the sugar-roll method. Arrows indicate the timing of amitraz (Apivar) treatments.

Our 4-year fixed-point survey of three A. mellifera colonies in Tokyo showed a significantly higher detection rate of IAPV in autumn than in winter. This finding was inconsistent with a previous study conducted in China, which reported no clear seasonal pattern for IAPV [4]. However, the detection pattern in our study was sporadic, and the observed seasonal difference may have been influenced by a relatively high number of detections in autumn 2022 rather than reflecting a consistent seasonal trend. In contrast, SBV and BQCV showed distinct seasonal increases in spring and summer, supporting reports that these viruses tend to increase in warmer seasons. Our findings align with the summer increases in BQCV reported in China [4]. BQCV also shows summer increases in several European regions [5, 19], suggesting the conservation of this seasonal pattern across different climates. In contrast, SBV seasonality varies among regions. A study in China reported no apparent seasonality [4], whereas a European study reported increased SBV detection during summer [5], supporting our findings. These discrepancies suggest that local beekeeping practices and ecological factors influence SBV dynamics.

DWV was detected almost continuously in A. mellifera throughout the study period and did not exhibit any statistically significant seasonal patterns. This is consistent with the dominant DWV-A genotype (clade-I) in Japan [10, 12], which is known to show year-round persistence [20]. Although DWV genotyping was not performed in this study, the observed detection pattern may be consistent with the predominance of DWV-A previously reported in Japan. The two major DWV genotypes (DWV-A and DWV-B [clade-II]) interact differently with Varroa. For example, DWV-B actively replicates in the gut and salivary glands of mites, whereas DWV-A does not [9]. Thus, such genotype-specific interactions may contribute to differences in seasonality across regions. DWV-B has been increasing rapidly in Europe and North America [11, 17], and its potential introduction into Japan may significantly alter the current virus–Varroa relationship and seasonal trends. Therefore, continuous monitoring of viral genotypes is essential to prepare for potential shifts in disease dynamics.

No clear seasonal pattern in Varroa infestation was detected in this study, despite previous reports describing seasonal fluctuations [6]. This may be explained by the relatively stable infestation levels observed throughout the study period, possibly due to regular mite control practices.

No significant association was observed between Varroa infestation and the number of detected virus types. These findings may be attributed to several factors. First, we used conventional PCR, which is a qualitative method that cannot capture changes in viral load, the parameter most strongly influenced by Varroa parasitism. Second, the near-constant, high detection rate of DWV probably masked variations in the number of detected virus types, rendering this metric insensitive to Varroa fluctuations. Third, the long-term, low-amplitude fluctuations in Varroa infestation (no statistically significant seasonal differences using pairwise comparisons) suggest that periodic Apivar treatments limited substantial variations in Varroa pressure. Consequently, our study design provides limited capacity to infer a causal relationship between Varroa pressure and viral infection patterns. Quantitative analyses using real-time PCR may provide a more detailed understanding of the relationship between Varroa infestation and virus dynamics, particularly for DWV, which was detected in most samples throughout the study period.

In addition, this study was limited to three colonies within a single beekeeping operation, and the observed virus dynamics and seasonal patterns may not fully represent those of other apiaries, management systems, or regions. Furthermore, colony mortality and population dynamics were not systematically recorded, limiting our ability to assess the relationship between pathogen detection and colony performance. Future studies should also consider including mortality rates and population dynamics in long-term monitoring. Despite these limitations, our results represent a long-term dataset obtained from the same colonies. Such data are rare both domestically and internationally and provide valuable baseline information on local virus dynamics. The distinct patterns of clear seasonal increases in SBV and BQCV, year-round detection of DWV, and sporadic occurrence of IAPV offer important regional insights for pathogen monitoring and honey bee colony management. Expanding this approach across multiple sites and using quantitative PCR to assess viral load will further clarify infection ecology, including interactions with Varroa mites. In addition, investigating potential associations with surrounding insects and the native Japanese honey bee (Apis cerana japonica) may help to better characterize seasonal dynamics and provide further insights into virus transmission and management of A. mellifera colonies.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

Funding Statement

This work was supported by the Center of Education and Research for Agromedicine, Kitasato University.