Chalkbrood samples of Apis mellifera L. were collected from the apiary severely infected with chalkbrood disease in Xinjiang province (N43°19′16.71, E84°01′3.22), China during April 2020. The chalkbrood mummies were collected from the bottom board or beehive entrance, while the surface of the mummies was covered with white filamentous and black spores. Then samples with typical symptoms from 10 colonies were selected, and all samples were stored in a refrigerator at −80°C until used for testing. Aspergillus niger ATCC16404 was purchased from China General Microbiological Culture Collection Center (CGMCC).

Fungi were isolated and purified as previously described with modifications (Burnside, 1981; Jensen et al., 2013).

Method 1 (hyphal tip isolation): Mummies were soaked in 10% NaClO for 10 min and washed twice in sterile water for 2 min. Then, the surface-sterilized chalkbrood mummies were cut into small pieces, which were further transferred to PDA (Coolaber, Beijing, China) medium with 100 μg/ml of ampicillin and incubated for 3–6 days in the dark at 30°C. The fungi were purified using hyphal-tipping technique under microscope for two rounds and cultured on PDA medium without antibiotic.

Method 2 (single-spore separation): Surface sterilization of chalkbrood mummies was conducted as above described. Then honey bee mummies were chopped up with sterile scalpel and put into new tube containing sterile water. After vortex, spore suspension was gradiently diluted to 10⁻²–10⁻⁴ and then spread on the PDA medium containing 100 μg/ml of ampicillin, which were incubated at 30°C for 2–3 days in darkness. Under the microscope, a single-spore colony was marked and the corresponding fungal plug was taken to new antibiotic-free PDA medium for further cultivation at 30°C for 4–6 days.

The hyphae block of purified fungi with a diameter of 0.5 cm was cut at the edge of the colony and placed to the center of PDA medium covered by cellophane membrane. After cultivating at 30°C for 2–3 days, the marginal hyphae were cut and observed by microscope. Furthermore, the conidiophore and mature ascoma were gently picked with sterile tweezers and placed on a glass slide with two drops of sterile water, which were directly observed using microscope (Ningbo Sunny Instruments, Zhejiang, China).

Fungi with different morphology were further identified by amplification of conserved genes containing internal transcribed spacer (ITS) and β-tubulin (for Aspergillus; Glass and Donaldson, 1995). Briefly, fungi were cultivated on PDA plates covered by sterilized cellophane membranes at 30°C for 6 days. Then 100 mg mycelia were collected for genomic DNA extraction, which was performed using Fungal Genomic DNA Extraction Kit (Solarbio, Beijing, China) according to the manufacturer’s instructions. DNA amplifications were performed in a 50 μl reaction volume, including 0.5 μM (each) primer (Supplementary Table S1), approximately 100 ng of genomic DNA and 25 μl of 2 × Es Taq MasterMix (CW Biotech, Beijing, China). The A. niger amplifications were used as positive controls. The PCR amplification was carried out in a thermocycler (Bioer, China) under the following conditions: 2 min at 94°C, followed by 30 cycles at 94°C for 30 s, 55°C or 58°C for 30 s, and 72°C for 30 s, and 72°C, 5 min for extra extension. PCR products were electrophoresed in 1% agarose gel containing 1 mg/ml GoldView (SBS Genetech Corp. Ltd., Beijing, China) and observed under gel imaging analysis system instrument (MiuLab, GIS-130, Beijing, China). The target gene fragments were purified by gel extraction (CW Biotech, Beijing, China) and verified by DNA sequencing (Sangon Biotech, Shanghai, China). The resulting sequences were analyzed by BLAST at https://blast.ncbi.nlm.nih.gov/Blast.cgi.

When the mycelium of purified fungi grew to about 2–3 cm at 30°C, the hyphal plugs was obtained and inoculated as the above described. The inoculated petri dishes were sealed and cultivated at 30°C, and a straight line was drawn across the center of the bottom of the petri dishes to measure and record the colony diameter every 24 h with three replicates in each group.

Spore cysts were produced between two opposing mating types of A. apis using mating tests, which were performed as previously described (Jensen et al., 2013). Then the suspension of ascospores was obtained according to the method of Tejerina and Benitez-Ahrendts (2021) with several modifications. Briefly, spore cysts were scraped from the surface of the colony using sterilized inoculating loop and collected in tubes containing sterile water and glass beads. The tubes were vortexed for 30 s (Vortex-Genie 2, Scientific Industries, United States) and centrifuged for 1 min at 15,294 g, then discarded the supernatant, and re-suspended the spore precipitate and repeated the steps three times. For obtaining conidia solutions, the procedures were followed according to previously described (Wu et al., 2012). Briefly, when conidia presented in the colonies, they were scraped from the surface of the colony with a sterilized inoculating loop. The conidia mixtures were pooled and filtered through four-layer cheesecloth to remove mycelial fragments. Spores were counted using hemocytometer and the suspensions of spores were adjusted to a concentration of 5 × 10⁶ spores/ml. The spore mixture of A. apis and A. tubingensis was obtained by mixing spore suspension of A. apis and A. tubingensis in a ratio of 1:1. New spore solutions were made prior to each experiment. The honey bee larvae (Apis mellifera L.) with 1 day old were collected from three random colonies located in the institute of apicultural research of Beijing, China during September and they were fed in sterile 48-well plate as proposed by Schmehl et al. (2016). Spore suspensions (10 μl) were added to the plate well containing 5 days instar larvae and sterile water was used as the control. Each replicate contained four larvae and three replicates in each group. The larvae were cultured at 35°C and 90% humidity, and the larval morphology was observed and recorded every day until 7 days after inoculation.

In order to detect several normal honey bee viruses in fungi, RNA extraction and reverse transcription PCR were conducted. For total RNA extraction, mycelial plugs were inoculated on PDA plates covered by sterilized cellophane membranes at 30°C for 6 days in the darkness. Around 100–200 mg mixture of mycelia and spores were collected and ground into a fine powder in liquid nitrogen. RNA extraction was conducted using ZN Fungal/Bacterial RNA Prep (TIANMO Biotech, Beijing, China) according to the manufacturer’s instructions. The resulting RNA was stored at −80°C before needed. Synthesis of cDNA was performed with 1 μg of total RNA using HiFiscript cDNA Synthesis Kit (CW Biotech, Beijing, China) in the presence of primer mixture (consist of Oligo dT and random primer) according to the manufacturer’s instructions. Then RT-PCR reaction mixture contained 1 μl of forward and reverse primers (Supplementary Table S2), respectively, 10 μl of 2 × Es Taq Master Mix (CW Biotech, Beijing, China), 1 μl of cDNA, and 7 μl of ddH₂O. PCR was performed with cycling conditions as follows: 94°C for 2 min, followed by 30 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s, and 72°C, 5 min for extra extension. Detection, sequencing, and analysis of the expected DNA products were performed as the above description. The sequence data from this study were submitted to and obtained accession numbers from GenBank.

Total RNA was extracted from A. tubingensis and A. apis using ZN Fungal/Bacterial RNA Prep (TIANMO Biotech, Beijing, China), and cDNA was synthesized using tagged specific primer (Supplementary Table S3) with super RT cDNA Synthesis Kit (CW Biotech, Beijing, China). The cDNA synthesis mixture was incubated at 42°C for 15 min, followed by 85°C for 5 min to terminate reaction. Then RNase H (Solarbio, Beijing, China) was added into the mixture, which was incubated at 37°C for 30 min to degrade RNA.

In order to quantify the positive and negative-strand RNAs, we first plotted their standard curves. For DNA fragment amplification, 1 μl of the above dilute cDNAs (1:50) were used as templates. The other PCR reagents included non-tagged primer and tag primer (0.5 μM of each, see Supplementary Table S4), 10 μl of 2 × Es Taq MasterMix and ddH₂O to 20 μl. Thermal cycling protocol was one cycle of 94°C for 2 min followed by 30 cycles of 94°C for 3 s, 57°C for 30 s, 72°C for 30 s, and 72°C for 5 min for extra extension. DNA fragments were recovered using Zymoclean Gel DNA Recovery Kit™ (CW Biotech, Beijing, China), and the purified fragments were cloned into plasmid pMD-18T using pMDTM-18T Vector Cloning Kit (TaKaRa, Japan). Then the constructed plasmids were transformed into competent cells Escherichia coli DH5α (Promega, Madison, WI, United States). The recombinant plasmids were extracted using the PurePlasmid MiniPrep Kit (CW Biotech, Beijing, China) and verified by sequencing (Sangon Biotech, Shanghai, China). All the above reactions were performed according to the manufacturer’s recommended instructions. The resulting plasmids for RT-qPCR were diluted to 10⁻²–10⁻⁸ and reacted with specific forward or reserve primers and tag primer (Supplementary Table S4). Reactions were subjected to 94°C, 3 min followed by 40 cycles of 95°C for 3 s; 60°C for 30 s in the presence of KAPA SYBR FAST qPCR kit (Sigma-Aldrich, United States). The real-time qPCR was performed on LineGene 9600 Plus real-time PCR instrument and the data was analyzed using software version 9600 Plus Software. The forward and reverse primers were designed in NCBI primer blast online,¹ based on the sequences of ABPV and CBPV submitted to NCBI Gen bank accession number: ABPV (MZ683936) and CBPV (OL321881) of A. apis, ABPV (OL321878) and CBPV (OL321879) of A. tubingensis in this study. Primer sequence of tag was non-homologous to the honey bee virus as previously described (Yue and Genersch, 2005).

Quantification of the positive and negative-strand RNA was determined using strand-specific RT-qPCR. The contents of mixture (20 μl) were similar to that of DNA amplification for standard curve. The reactions were performed on LineGene 9600 Plus real-time PCR instrument, and the results were analyzed using 9600 Plus Software and standard curves.

In order to test whether virus can be transmitted to offspring or other related virus-free fungi, we followed the protocol of Wu et al. (2012) with several modifications.

To study vertical transmission of viruses, fresh fungal plug was cultivated on PDA plate at 30°C for 5–7 days. When colonies produced conidia or ascospore, the solution of conidia or ascospore was prepared according to the above method. Then 100 μl of diluted spore suspension was spread on the PDA plate, which was incubated at 30°C for 2–4 days in darkness. The mycelia at the edge of a colony were collected for virus detection.

To study horizontal transmission of viruses, pairing culture technique was used. Viruses were transferred to virus-free strain of A. niger through hyphal anastomosis. In brief, two different strains were placed on the two sides of PDA plate, respectively, with 2 cm distance. Ascosphaera tubingensis and A. apis served as the donors, whereas virus-free strain A. niger served as recipient. After 17 days under dark condition at 30°C, three mycelium plugs from A. niger were taken and placed on the new PDA plates. The mycelia near the edge middle of A. niger colony were labeled JL1, mycelia below the edge of A. niger colony were labeled JL2, and the mycelia far from A. tubingensis colony were labeled JL3. Detection of the viruses in mycelia using the methods above described.

Hyphal tip isolation and single-spore separation were used to purify the fungi from chalkbrood. The surface-sterilized chalkbrood mummies were cut into small pieces and transferred to PDA for cultivation of 6 days, then colonies with gray white surface were observed (Figure 1A). When the plate was cultured at 30°C for 30 days or longer, black spores appeared and the hyphae were no longer fluffy but appressed on the surface of the medium (Figure 1B), while the spore morphology was different from that of common A. apis. Method of single-spore isolation was also adopted. After incubation for 2 days, a smaller special colony was observed (Figure 1C), which produced black spores similar to that in Figure 1B. In order to further isolate and purify fungi, the hyphal tip was picked from the edge of the colony in Figure 1A and cultured on PDA plates. As a result, two fungi with different morphology were isolated and labeled as HQS-1 and HQS-2, respectively (Figures 1D,E). The mycelia of HQS-1 and HQS-2 colony were fluffy and velour, respectively. Black spores were selected from the colonies in Figures 1B,C, diluted, and spread on PDA plate. The resulting fungi marked as HYS (Figure 1F), which had a typical morphology of Aspergillus with phialides conidiophores and was speculated to belong to genus Aspergillus.

Optical microscopy observations showed that most of the hyphae of the three strains were branched (Figure 2A). Besides, the conidia of strain HYS was spherical, and the mature ascoma of strains HQS-1 and HQS-2 were cracked, nearly spherical and hyaline, which contained several asci varied in diameter (Figure 2B). There were a large number of ascospores gathered in asci (Figure 2C). These results indicated that HQS-1and HQS-2 might belong to Ascomycete, and HYS might belong to genus Aspergillus.

The DNA fragments of ITS were amplified for molecular identification of fungal species (Figure 3). The ITS sequence alignments showed that strain HYS had 99.29% similarity with A. tubingensis isolate TF1 and A. niger isolate NRF9. In addition, the similarity of strain HQS-1 and HQS-2 to A. apis isolate SX and A. apis strain CBS534.69 was both up to 100%. In order to further identify strain HYS, β-tubulin was amplified and analyzed. The alignment showed that the strain HYS had 100% similarity to the first 16 A. tubingensis species. These results combined with morphological analysis suggested that strain HYS was A. tubingensis, strains HQS-1 and HQS-2 were A. apis.

Measurement of the growth rate showed that A. apis HQS-1 grew faster than A. apis HQS-2, and both of them grew faster than A. tubingensis HYS. Furthermore, the growth rates of A. tubingensis HYS, A. apis HQS-1, and HQS-2 were 0.67, 2.08, and 1.65 cm/days, respectively (Figure 4A). Sexual spores were produced between the opposing mating types of A. apis HQS-1 and HQS-2 (Figure 4B). When A. apis HQS-1, HQS-2, and A. tubingensis HYS were co-cultured on PDA medium, the contact of hyphae could be observed after 4 days. While after 8 days of culture, mycelia of A. tubingensis HYS covered the surface of A. apis colonies (Figure 4C). Three groups of honey bee larvae were infected with the spores of A. apis, A. tubingensis alone or combination, the results showed that 33.33–41.67% of honey bee larvae in each group were infected (Supplementary Table S5). Although the mortalities were similar among the three groups, the characteristics of infected honey bee larvae differed significantly. There were many white mycelia and a small number of ascospores on the surface of honey bee larvae infected with A. apis. On the contrary, a large number of gray black conidiospores covered on the surface of honey bee larvae infected with A. tubingensis. However, after co-infection with spores of A. apis and A. tubingensis, there were a large number of white hyphae on the surface of honey bee larvae, and the number of spores was much more than that infected with A. apis alone, meanwhile, the infected larvae did not show symptom of A. tubingensis infection (Figure 4D). These results indicated that both A. apis and A. tubingensis could infect honey bee larvae and cause corresponding symptoms. While different from the co-culture in PDA medium, A. apis that colonized honey bee larvae had more growth advantages than A. tubingensis, or the growth of A. tubingensis was inhibited by A. apis in honey bee larvae gut. It might be further indicated that the presence of A. tubingensis in honey bee larvae could promote the spore production of A. apis and the prevalence of chalkbrood.

The results of virus amplification in A. tubingensis and A. apis were shown in Figure 5. We detected several honey bee viruses in both A. tubingensis and A. apis. ABPV (accession number: OL321878), CBPV (accession number: OL321879), IAPV (accession number: OL321884), KV (accession number: OL321882), and DWV-A (accession number: OL321883) were detected in A. tubingensis. Besides, ABPV (accession number: MZ683936) and CBPV (accession number: OL321881) were detected in both A. apis HQS-1 and HQS-2. There was no difference in virus species between A. apis HQS-1 and HQS-2.

As both ABPV and CBPV were detected in A. apis and A. tubingensis with very obvious bands, the replication and accumulation of which were tested in A. apis and A. tubingensis using strand-specific RT-qPCR. Hence, the presence of negative-strand RNA and positive-strand RNA for each virus would be determined when fresh fungal plugs were cultivated on PDA for 3, 6, and 9 days. The specificity of positive and negative chain amplification primers of ABPV and CBPV and the standard curves had been verified (Supplementary Figure S1). As it was showed, with the increase of culture time, ABPV titers increased in A. tubingensis, while the titers of CBPV were high on the third day then decreased slightly (Figure 6A). Meanwhile, the replication levels of ABPV and CBPV in A. tubingensis increased gradually, indicating that ABPV and CBPV could exist and replicate in A. tubingensis. For A. apis, ABPV titers of HQS-1 and HQS-2 both slightly decreased and then increased, and CBPV titers steadily increased from day 3 to day 9 (Figures 6B,C). Similarly, the replication levels of ABPV and CBPV also increased gradually with the exception that the copy number of ABPV in A. apis HQS-1 decreased to the lowest on the 6th day, but increased to the highest on the 9th day (Figures 6B,C). The above results suggested that ABPV and CBPV could exist and replicate in A. apis, and there was no significant difference in virus titer and replication level between A. apis HQS-1 and HQS-2.

Viruses can be transmitted vertically by spores. In order to study the transmission mode and ability of honey bee viruses in A. tubingensis and A. apis, we detected the virus distribution in the spores of A. tubingensis and A. apis. Virus detection in spores showed that in A. tubingensis, not all viruses could be transmitted to the next generation through spores. CBPV (19/19) could be detected in all spores, which may be beneficial to the long-term stable existence of the virus, while ABPV (12/19), IAPV (7/19), KV (5/19), and DWV-A (4/19) distributed in spores to varying degrees (Table 1) and the virus distribution details was listed in Supplementary Table S6. As for A. apis, CBPV (15/15) could be also detected in all spores, while ABPV (10/15) distributed in partial spores (Table 1) and the virus distribution details was listed in Supplementary Table S7. The above results indicated that not all of the five viruses were distributed in the offspring spores, and the vertical transmission ability was CBPV > ABPV > IAPV > KV > DWV-A.

Viruses also can be transmitted horizontally via hyphal anastomosis, so we detected virus distribution in the mycelia of Aspergillus niger anastomosed with that of A. tubingensis and A. apis, respectively. Virus detection in mycelia showed that only CBPV was detected in the mycelia of A. niger anastomosed with A. tubingensis (Figure 7; Supplementary Table S8). And as expected, no virus was detected in the mycelia of A. niger anastomosed with A. apis (Supplementary Table S8). Besides, in order to avoid the contamination of Aspergillus tubingensis, the strain JL2 was further identified using the primers of β-tubulin (Supplementary Figure S2A). The result sequence was aligned with β-tubulin sequence of standard strain Aspergillus niger, and the two sequences were the same, except for the unstable bases generated by sequencing at the two ends (Supplementary Figure S2B), which indicated that strain JL2 was Aspergillus niger. The above results suggested that only CBPV could break through the barrier of vegetative incompatibility and be transmitted from A. tubingensis to A. niger through hyphal anastomosis.