Introduction

Magnetoreception is a sense allowing an animal to detect the earth’s magnetic field to establish regional magnetic maps for navigation and positioning. Numerous studies demonstrate that honey bees have magnetoreception behavior. Applying an extra magnetic field changes the bees’ comb building and orientation and homing behaviors1,2,3. Gluing magnets to the anterior dorsal abdomen interface interferes with magnetic discrimination in choice experiments4. Free-flying honey bees can detect small static intensity fluctuations at the level of 26 nT against the earth-strength magnetic field5 and alternating fields of 430 μT at a frequency of 10 and 60 Hz6. In addition, the bees can be trained in discrimination assays to respond to local anomalies in a magnetic field7,8.

The magnetoreception of honey bees is further demonstrated by the finding of superparamagnetic magnetite in iron granules (IGs)9,10. Applying a magnetic field to the bee causes the IGs to shrink in the direction paralleled to the magnetic field and to elongate in the vertical direction to the magnetic field10. The changes in the IG’s confirmation by magnetic fields is posited to generate Ca2+ signals and signal transduction via fluctuation of cytoskeletons on iron deposition vesicles (IDVs)10. These events can allow honey bees to establish a magnetic map during orientation flights to facilitate future foraging behavior10. IGs, therefore, are proposed to act as the magnetoreceptor in honey bees and crucial to the bees’ orientation and homing10.

IGs are present in the IDVs of trophocytes which are located in the abdomen of honey bees11. IGs are derived from the aggregation of 7.5-nm diameter iron particles in the center of IDVs. 7.5-nm diameter iron particles are formed in the cloudy area beneath the inner membrane of IDVs11. IDVs are double-membrane vesicles with 10 nm wide space between the outer and inner membrane11. IGs are deposited from about 5 days to 25 days after adult emergence. When IGs are mature, they are filled to the inner membrane, leaving a 10 nm space between IGs and the outer membrane of IDVs11. An actin-myosin-ferritin transporter system participates in the formation of IGs12. Actin, myosin, ferritin 2, and ATP synthase are involved in this system12.

RNA interference (RNAi)-mediated gene knockdown has been used to study gene functions in different honey bee developmental stages, including embryos13,14, larvae15,16, pupae17,18, and fully developed adult honey bees19,20,21,22. Genes such as E30 homeobox, CSP5, vitellogenin, insulin receptor substrate, am-tra2, am18w, ultraspiracle, octopamine receptor, and DNA methyl-transferase have been successfully knocked down in honey bees by double-stranded RNA (dsRNA) injection13,14,17,19,22,23,24,25,26.

Ferritin is a hollow globular protein that stores iron and releases it. Each ferritin can store about 4500 iron (Fe+3) ions27. Ferritins are composed of heavy chains and light chains. Heavy chains are important for Fe+2 oxidation and light chains assist in core formation27. In honey bees (Apis mellifera), the heavy chain is called ferritin1 or Fer1HCH and the light chain is called ferritin2 or Fer2LCH. The amino acids sequences of ferritin2 (XP_624076.1) and ferritin1 (XP_016771737.1) is 23% degree of identity. In this study, we knocked down the genes for ferritin2 and ferritin1 to test whether ferritins RNAi can be used to inhibit or change IGs formation. The technique developed here may be used to further explore magnetoreception in honey bees.

Methods

The preparation of dsRNA toward ferritin2, ferritin1, and green fluorescent protein (GFP, control)

The primers were designed based on the nucleotide sequences available in GenBank: ferritin2 (Fer2LCH) (XM_624073.4): forward 5′-ATTTTTGGCAACTGCCTCTG-3′, reverse 5′-ATTCTCGAACACGGTCTGCT-3′; ferritin1 (Fer1HCH) (XM_016916248.1): forward 5′-CCCCGTCGATTAAAGTACGA-3′, reverse 5′-GCATGTTCTCTTTCTTCTGTAGCA-3′; GFP: forward 5′-GAGATACCCAGATCAT-3′, reverse 5′-GATGATATTCACCACTT-3′. Primers were fused with the T7 promoter sequence. Total RNA was isolated from the trophocytes and oenocytes of three worker bees at 3 days after emergence using TRIzol (15596018; Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions. The complementary DNA (cDNA) was synthesized by using Superscript III First-Strand Synthesis System for RT-PCR (18080-051; Invitrogen, Carlsbad, CA, USA). The cDNA was transferred into E. coli by using Topo TA cloning Kit for sequencing (450030, Invitrogen, Carlsbad, CA, USA). The plasmid was isolated with QIAprep Spin Miniprep Kit (27104, Qiagen, Valencia, CA, USA). The double-stranded (dsDNA) was produced by PCR using the T7 primers and purified by QIA Quick Gel Extraction Kit (28704, Qiagen, Valencia, CA, USA). Finally, dsRNA was synthesized and purified by using AmpliScribeTM T7-FlashTM Transcription Kit (ASF3257, Epicentre Biotechnologies, Madison, WI, USA), and diluted with nuclease-free water to a final concentration of 5 μg/μl.

The microinjection of ferritin2, ferritin1, and GFP dsRNA

The microinjection of ferritin2, ferritin1, and GFP dsRNA into hemolymph was followed as Amdam’s method with minor modification19. Newly emerged worker honey bees were immobilized on a disc of bee wax with two crossed metal needles at room temperature. The bees were injected with 1 μl nuclease free water or 1 μl dsRNA solution (5 μg/μl) with a microinjector (Femtojet, Eppendorf). Microinjection was performed on the dorsum of the abdomen between the 1st and 2nd abdominal segment with glass needles. Care was taken to minimize mechanical damage to the bees. Individuals showing hemolymph leakage after microinjection were discarded. Successfully injected bees were housed in a cage (15 × 10 × 12 cm3) for 1 h before moving into an incubator set to 34 °C (NK system, Nippon, Japan). The worker bees were fed honey and fresh pollen grains mixed with honey (3:1) every day28.

Quantitative real-time polymerase chain reaction (qPCR) analysis

Trophocytes and oenocytes were isolated from two worker bees 6 days after emergence. Total RNA was extracted from these cells using Trizol® Reagent (15596018; Invitrogen, CA, USA). RNA concentration and quality were determined using a SynergyTM HT multi-mode microplate reader (7091000; BioTek). The cDNA synthesis was performed using an iScript™ cDNA synthesis kit (170-8891; Bio-Rad Laboratories, CA, USA). Each reaction contained 1 μg of total RNA in a 20 µl reaction volume. The qPCR was performed using a CFX connect RT-PCR detection system (Bio-Rad Laboratories) and each reaction contained 0.5 µl of 10 µM of each primer, 12.5 µl of SYBR Green (170-8882; Bio-Rad Laboratories), 1 µl of cDNA, and 10.5 µl of ddH2O in a final volume of 25 μl. Primer sequences were noted above. The actin gene was used as reference gene. Ten replicates were performed, and 20 worker bees in total were used in each group.

Western blotting

Trophocytes and oenocytes were isolated from two worker bees, homogenized in 100 μl of radioimmunoprecipitation (RIPA) buffer containing protease inhibitors (11697498001; Roche Applied Science, Indianapolis, IN, USA), and centrifuged at 5,000 g for 10 min at 4 °C. The protein concentration of resulting supernatant was determined using a protein assay reagent (500-0006; Bio-Rad Laboratories, Hercules, CA, USA). Proteins (30 μg) from the supernatant were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS/PAGE) on 10–15% polyacrylamide gels and transferred to polyvinylidene fluoride (PVDF) membranes. After blocking for 1 h at 25 °C, membranes were first incubated with primary antibodies against ferritin2 (1:1,000; produced in-house) or tubulin (ab6046, 1:10,000; Abcam, Cambridge, MA, USA) and then probed with the appropriate horseradish peroxidase-conjugated secondary antibody (1:10,000). Immunoreactive proteins were detected using a chemiluminescence method (PerkinElmer, Covina, CA, USA) and analyzed using Image J software (NIH, Bethesda, MA, USA). The protein production levels were normalized to tubulin28. Ten replicates were performed, and thus 20 worker bees in total were used in each group.

Transmission electron microscopy (TEM)

Worker bees without RNAi (control group 1), with GFP RNAi (control group 2), and with ferritins RNAi (experimental group) at 6, 8, 10, 12, 14, and 20 days after emergence were fixed in 2.5% glutaraldehyde in a 0.1 M phosphate buffer containing 0.35 M sucrose (pH 7.4) for 30 min at 25 °C, and postfixed in 1% osmium tetroxide in a 0.1 M phosphate buffer with 0.35 M sucrose (pH 7.4) for 2 h. Trophocytes were dehydrated through an ethanol series and embedded in Spurr’s resin. Thin sections (60–90 nm in thickness) were cut with a diamond knife, stained with uranyl acetate and lead citrate, and then examined using a TEM system (JEOL JEM-2000EXII; Tokyo, Japan) operating at an accelerating voltage of 100 kV12.

IGs formation assay

IGs formation was evaluated according to the IG area/IDV area ratio from TEM images. The IG area and IDV area were analyzed using Photoshop (CS6)29. Five replicates were performed, and thus 5 worker bees in total were used in each group. Ten IDVs in each bee were used to evaluate IGs formation. Thus, the results for IGs formation were derived from 50 IDVs in each group.

Statistical analysis

Differences in the mean values among the three treatment groups were determined by one-way ANOVA and by Tukey’s HSD for pairwise comparisons. SPSS software was used for statistical analysis. A p-value of less than 0.05 was considered statistically significant.

Results

Ferritin2 RNAi inhibits the mRNA and protein production of ferritin2

To evaluate the effect of ferritin2 RNAi, we assayed the ferritin2 mRNA and protein levels in the trophocytes and oenocytes of worker bees at 6 days after ferritin2 or GFP dsRNA injection. We found that worker bees subjected ferritin2 RNAi had lower ferritin2 mRNA levels than the water control and the GFP RNAi control (n = 10, P < 0.05; Fig. 1A), indicating that ferritin2 RNAi suppressed ferritin2 mRNA production. Worker bees with ferritin2 gene knockdown also had lower ferritin2 protein levels than the water control and the GFP RNAi control (Fig. 1B). Statistical analyses revealed that ferritin2 protein levels were significantly different from the water control and the GFP RNAi control (n = 10, P < 0.05; Fig. 1C) indicating that ferritin2 gene knockdown also suppressed ferritin2 protein production.

Figure 1
figure 1

The production of ferritin2 mRNA and protein in worker bees at 6 days after ferritin2 RNAi. (A) The mRNA production of ferritin2 in trophocytes and oenocytes was measured via qPCR. Actin served as the loading control. The results were normalized to the water control and are shown as fold changes, representing the mean ± standard error of the means (SEMs) (n = 10). Fer2, ferritin2. (B) Ferritin2 protein was analyzed by western blotting. Tubulin served as the loading control. C, control; G, GFP dsRNA; F, ferritin2 RNAi. (C) The production levels of ferritin2 were normalized to the control. The results were expressed as percentages and presented as the means ± SEMs (n = 10). Fer2, ferritin2. Asterisk indicates a statistically significant difference (**P < 0.01; ***P < 0.001; one-way ANOVA).

Full size image

Ferritin2 RNAi inhibits the formation of IGs

To evaluate the effect of ferritin2 knockdown on IGs formation, we examined the morphology of IGs at the 6, 8, 10, 12, 14, and 20 days after dsRNA injection. We found that the IGs of worker bees with ferritin2 knockdown had decreased iron mineralization at 12, 14, and 20 days after ferritin2 RNAi (Fig. 2A,a–c), as compared to the GFP RNAi control (Fig. 2A,d–f) and the water control (Fig. 2A,g–i). Statistical analyses revealed that iron mineralization in the worker bees with ferritin2 knockdown was significantly different from the GFP RNAi control and the water control (n = 50, P < 0.05; Fig. 2B) indicating that ferritin2 RNAi inhibited the formation of IGs. The GFP RNAi control is also significantly different from the water control at 12 and 14 days.

Figure 2
figure 2

The morphology of IGs in the trophocytes of worker bees after ferritin2 RNAi. (A) The morphology of IGs on 12 (a), 14 (b), and 20 (c) days after ferritin2 dsRNA injection, 12 (d), 14 (e), and 20 (f) days after GFP dsRNA injection, and 12 (g), 14 (h), and 20 (i) days after nuclease free water injection. Scale bar in images a–i, 0.3 μm. (B) Size changes of IGs in different days after ferritin2 dsRNA injection. Bar represents the means ± SEMs (n = 50). ◊, control; □, GFP dsRNA; Δ, ferritin2 RNAi. Asterisk indicates a statistically significant difference (*** < 0.001; one-way ANOVA).

Full size image

Ferritin1 RNAi inhibits the mRNA production of ferritin1

To evaluate the effect of ferritin1 RNAi, we assayed the ferritin1 mRNA levels in the trophocytes and oenocytes of worker bees at 6 days after ferritin1 or GFP dsRNA injection. We found that worker bees treated ferritin1 RNAi had lower ferritin1 mRNA levels than the water control and the GFP RNAi control. Statistical analyses revealed that the mRNA levels of ferritin1 were significantly different the water control and the GFP RNAi control (n = 10, P < 0.05; Fig. 3) indicating that ferritin1 RNAi suppressed ferritin1 mRNA production.

Figure 3
figure 3

The production of ferritin1 mRNA in worker bees at 6 days after ferritin1 RNAi. The mRNA production of ferritin1 in trophocytes and oenocytes was measured via qPCR. Actin served as the loading control. The results were normalized to those of the water control (control) and are shown as fold changes, representing the mean ± SEMs (n = 10). Fer1, ferritin1. Asterisk indicates a statistically significant difference (***P < 0.001; one-way ANOVA).

Full size image

Ferritin1 RNAi inhibits the formation of IGs

To evaluate the effect of ferritin1 knockdown on IGs formation, we examined the morphology of IGs at the 6, 8, 10, 12, 14, and 20 days after dsRNA injection. We found that the IGs of worker bees with ferritin1 knockdown had decreased iron mineralization at 12, 14, and 20 days after ferritin1 RNAi (Fig. 4A,a–c), as compared to the GFP RNAi control (Fig. 4A,d–f) and the water control (Fig. 4A,g–i). Statistical analyses revealed that iron mineralization in the worker bees with ferritin1 knockdown was significantly different from the GFP RNAi control and the water control (n = 30, P < 0.05; Fig. 4B) indicating that ferritin1 RNAi inhibited the formation of IGs. The GFP RNAi control is also significantly different from the water control at 12 and 14 days.

Figure 4
figure 4

The morphology of IGs in trophocytes after ferritin1 RNAi. (A) The morphology of IGs on 12 (a), 14 (b), and 20 (c) days after ferritin1 dsRNA injection, 12 (d), 14 (e), and 20 (f) days after GFP dsRNA injection, and 12 (g), 14 (h), and 20 (i) days after nuclease free water injection. Scale bar in images a–i, 0.2 μm. (B) Size changes of IGs in different days after ferritin1 dsRNA injection. Bar represents the means ± SEMs (n = 50). ◊, control; □, GFP dsRNA; Δ, ferritin1 RNAi. Asterisk indicates a statistically significant difference (** < 0.01; *** < 0.001; one-way ANOVA).

Full size image

The different deposition morphology of 7.5-nm diameter iron particles from ferritin2 and ferritin1RNAi

The deposition morphology of 7.5-nm diameter iron particles via ferritin2 RNAi was different from that of ferritin1 RNAi at 6 days after dsRNA injection, indicating that ferritin2 and ferritin1 may play different roles in the formation of IGs (Fig. 5).

Figure 5
figure 5

The deposition morphology of 7.5-nm diameter iron particles in the trophocytes of worker bees at 6 days after by ferritins dsRNA injection. (A) An IDV at 6 days after ferritin2 RNAi. Scale bar, 0.2 μm. Arrow, 7.5 nm particles. (B) An IDV at 6 days after ferritin1 RNAi. Scale bar, 0.2 μm. Arrow, 7.5 nm particles.

Full size image

Discussion

Actin, myosin, ferritin2, and ATP synthase are involved in the actin-myosin-ferritin transporter system which participates in the formation of IGs in the IDVs of trophocytes12. In this study, we demonstrate that the mRNA and/or protein production of ferritin2 or ferritin1 and the formation of IGs in trophocytes can be inhibited by ferritin2 or ferritin1 RNAi. This technique may be used to further explore the magnetoreception of honey bees.

Ferritins RNAi inhibits the mRNA and/or protein production of ferritins

Ferritin packaged with iron ions is a 7.5-nm spherical iron particle in IDVs. Ferritin2 RNAi inhibits the mRNA and protein production of ferritin2. Likewise, ferritin1 RNAi inhibits the mRNA production of ferritin1. These phenomena are consistent with previous studies showing that the mRNA production of honey bees can be inhibited by RNAi19,20,21,22.

Ferritin2 and ferritin1 RNAi inhibit the formation of IGs

7.5-nm diameter iron particles are formed in the cloudy area beneath the inner membrane of IDVs and move to the center of IDVs for aggregation which forms IGs11. The IGs formation is proposed and summarized as follows: Fe2+ from the cytoplasm is transported into the acidic space (pH < 7) between the outer and inner IDV membranes via a transporter protein. An H+/Fe2+ antiporter on the inner IDV membrane then simultaneously transports one molecule of Fe2+ into the acidic space and two molecules of H+ out of the acidic space to maintain a pH < 7 in the acidic space and a pH > 7 in the alkaline lumen of IDVs. Fe2+ then becomes partially oxidized to Fe3+, and Fe2+/Fe3+ is integrated into apoferritin in the cloudy layer of IDVs to form 7.5-nm spherical iron particles (ferritin)11,12,30. Then, ferritin attached to myosin is transported along an actin chain to the center of IDVs in a manner that is dependent on Ca2+ and ATP12. This transporter system is called the actin-myosin-ferritin transporter system. Actin, myosin, and ferritin2 are involved in actin-myosin-ferritin transporter system12. Ferritin2 RNAi inhibited the formation of IGs demonstrated that ferritin2 has a relationship with the formation of 7.5-nm diameter iron particles and ferritin2 is involved in the formation of IGs. This finding is consistent with a previous study12. In addition, ferritin1 RNAi inhibited the formation of IGs demonstrated that ferritin1 has a relationship with the formation of 7.5-nm diameter iron particles and ferritin1 is involved in the formation of IGs. The technique developed here can be used to further explore magnetoreception in honey bees. The GFP RNAi control is also significantly different from the water control at 12 and 14 days. The most likely reason is that GFP dsRNA interferes the synthesis of ferritins at 12 and 14 days, a period of mass synthesis of ferritins for iron deposition.

Ferritin2 and ferritin1 play different roles in the formation of IGs

The image of IGs after ferritin2 knockdown showed that the accumulation of 7.5-nm diameter iron particles in the center of IDVs of ferritin2 RNAi bees is similar to that of the water control11. However, this accumulation in ferritin2 RNAi bees is slower than that in the water control due to 7.5-nm diameter iron particles deposited slowly11. This phenomenon shows that ferritin2 may be involved in the formation of 7.5-nm diameter iron particles. The deposition morphology of 7.5-nm diameter iron particles after ferritin1 knockdown showed that most of 7.5-nm diameter iron particles accumulated on the periphery of the center of IDVs and are different from that of ferritin2 RNAi. This phenomenon shows that ferritin1 may have a relationship with the transportation of 7.5-nm diameter iron particles. Ferritin1 may be associated with the tail of myosin12.

These phenomena were only observed at the early stage of ferritin RNAi. The most likely reason is that the dsRNA dosage of one injection did not inhibit gene expression and the formation of IGs throughout the entire observation period.