Conceived and designed the experiments: CH CL. Performed the experiments: CH FK. Analyzed the data: CH JL CL. Contributed reagents/materials/analysis tools: CH CL. Wrote the paper: CH JL KF.
The authors have declared that no competing interests exist.
Honeybees (
The waggle dances of honeybees communicate the information of direction and distance from the hive to a new food source was first hypothesized by von Frisch
Magnetoreception of honeybees has been proposed on the basis of numerous behavioral evidences. They can be summarized as follows: The behavioral changes in comb building, and homing orientation when extra magnetic field is added
To explain the behavior of magnetoreception and to search for biomagnetites, researches on magnetic remanence and iron biomineralization in honeybees have been carried out for a long time. The results of these works can be summarized as follows: The magnetic remanence was detected in the abdomen of honeybees by superconducting quantum-interference device magnetometers (SQUID)
All of magnetic remanence from SQUID, the typical X-band resonance spectra from EPR, and hysteresis curve from SQUID were obtained from the whole body or abdomen of honeybees. In addition, SM was observed in four IGs of trophocyte. Whether the results of magnetic measurement are derived from the SM of IGs remains an open question. Although SM was observed in the IGs of trophocyte under HRTEM, the study was carried out on a very small sample (four IGs)
In this study, we establish a size-density purification procedure to recover quantitative amount of IGs from trophocytes, needed for characterization and demonstrating the presence of SM in IGs. We determined the magnetic properties of the purified IGs by SQUID, atomic force microscope (AFM), magnetic force microscope (MFM), EPR, and electron spectroscopy for chemical analysis (ESCA). We also applied additional magnetic field (1 Gauss) to examine whether the size of IGs in trophocytes can be induced to change along with release of calcium ions.
In this study, we established a size-density procedure new for purification of IGs from trophocytes (
A flow diagram showing the purification processes of the size-density purification technique.
IGs purified from the trophocytes of honeybees. (A) A large amount of IGs appear in precipitates purified from the trophocytes. Scale bar, 1 µm. (B) Purified IGs enclosed with lipid-bilayer membranes (arrowhead). Scale bar, 100 nm. (C) EDX spectrum is obtained from an IG in the purified precipitates. Iron (Fe), calcium (Ca) and phosphorus (P) are present in an IG. STEM mode of 100-kV accelerating voltage was used; counts were made for 100 seconds. (D) EDX spectrum of spurr's resin is devoid of IGs in the purified precipitates. The experimental conditions are the same as in C. (E) Size distribution of purified IGs calculated from the TEM pictures (N = 274). N, total number of granules used in this calculation.
The morphology of electron-dense granule vesicles in the purified precipitates is similar to that of IDVs observed in the trophocytes of adult workers
The density of the purified IGs was measured. This was done by first precipitating the granules by centrifugation against 6.0 molality (m) sucrose solutions at 23,000×g, with the precipitates re-suspended on 6.25 m sucrose solutions. The density of 6.0 m and 6.25 m sucrose solution is 1.2506 g/cm3 and 1.2539 g/cm3, respectively. Therefore, the density of IGs should be at the interval of 1.2506–1.2539 g/cm3.
Various magnetic properties of purified IGs were determined.
By the use of SQUID magnetometers, an averaged SQUID magnetization curve showed that IGs possess a residual magnetization of 2.5×10−5 emu, a saturation magnetization of 2.83×10−4 emu, and an intrinsic coercivity of 100–150 Gauss (
The magnetic properties of IGs determined by magnetic-determining techniques. (A) An averaged magnetization curve of SQUID from the purified IGs at 300°K for six runs. (B) An image of AFM from one of the purified IGs. (C) An image of MFM from one of the purified IGs. (D) A spectrum of ESCA from the purified IGs. The left thin line shows the Fe(2p1/2) of FeOOH. The middle thin line shows the Fe(2p3/2) of Fe2O3. The right thin line shows the Fe(2p1/2) of FeO. (E) A photoelectron spectrum of elemental composition of ESCA from the purified IGs. (F) Typical X-band resonance spectra from the purified IGs at different temperatures. Right inset shows a magnified image from the arrow. Left inset shows a magnified image from the arrowhead.
Imaging by AFM and MFM of an IG showed that single domain inherent in each IG reflects superparamagnetic behavior (
EPR spectra also show that SM is present in IGs (
The ESCA spectra of IGs showed that FeO and Fe2O3 coexist in IGs (
These results demonstrate unequivocally that SM is present in the purified IGs, and IGs are magnetic granules (MGs).
To study the role of MGs as magnetoreceptors, we examined under confocal microscope size fluctuation of MGs in trophocytea when additional magnetic field (1 Gauss) was applied. At low magnification, MGs appear as tiny black particles in both live (
The images of MGs in the trophocytes obtained by confocal microscope. (A) MGs (arrowhead) appear as tiny black particles under low magnification in the live trophocytes. Arrow indicates oil body. Scale bar, 8 µm. (B) MGs (arrowhead) appear as black granules under high magnification in the live trophocytes. The images of MGs are obtained without the application of 1 Gauss magnetic field. Arrow indicates oil body. Inset shows a magnified MG. Scale bar, 1 µm. (C) The same images obtained by the application of 1 Gauss magnetic field in the live trophocytes. White arrow indicates the direction of magnetic field. Arrowhead indicates MGs. Arrow indicates oil body. Inset shows a magnified MG. Scale bar, 1 µm. (D) MGs (arrowhead) appear as tiny black particles under low magnification in the dead trophocytes. Scale bar, 8 µm. (E) MGs (arrowhead) appear as black granules under high magnification in the dead trophocytes. The images of MGs are obtained without the application of 1 Gauss magnetic field. Inset shows a magnified MG. Scale bar, 1 µm. (F) The same images obtained by the application of 1 Gauss magnetic field in the dead trophocytes. Arrowhead indicates MGs. White arrow indicates the direction of magnetic field. Inset shows a magnified MG. Scale bar, 1 µm.
To study signal transduction elicited by MGs for magnetoreception, the variation of calcium ion in trophocytes and fat cells labeled with flou 4 were measured under confocal microscope. Of these two types of cells, trophocytes are larger (about 75 µm in diameter) and are dark green in color, while fat cells smaller (about 30 µm in diameter) and bright green (
The images of increase of [Ca+2]i in the live trophocytes obtained by confocal microscope. (A) The image of trophocytes and fat cells under confocal microscope. Trophocytes (arrowhead). Fat cells (arrow). Scale bar, 40 µm. (B) The image of trophocytes and fat cells labeled fluo 4 under confocal microscope. Trophocytes (arrowhead). Fat cells (arrow). Scale bar, 40 µm. (C) The merge of Figure A and Figure B. Scale bar, 40 µm. (D) The fluorescence intensity of trophocytes (○), fat cells (Δ), and trophocytes and fat cells together (□) with 1 Gauss magnetic field. The fluorescence intensity of trophocytes (•), fat cells (▴) and trophocytes and fat cells together (▪) without 1 Gauss magnetic field as control. (E) The fluorescence intensity of trophocytes (○), fat cells (Δ), and trophocytes and fat cells together (□), which is inhibited by colchicine and with 1 Gauss magnetic field. The fluorescence intensity of trophocytes (•), fat cells (▴) and trophocytes and fat cells together (▪), which is inhibited by colchicine without 1 Gauss magnetic field as control. (F) The fluorescence intensity of trophocytes (○), fat cells (Δ), and trophocytes and fat cells together (□), which is inhibited by latrunculin B and with 1 Gauss magnetic field. The fluorescence intensity of trophocytes (•), fat cells (▴) and trophocytes and fat cells together (▪), which is inhibited by latrunculin B without 1 Gauss magnetic field as control.
The effect of magnetic field on [Ca+2]i could be inhibited by colchicine (
Taken together, these results show that the increase of [Ca+2]i in trophocytes may be caused by the fluctuation of MGs in size and is mediated by microtubules and microfilaments.
In this study, we established a new procedure based on size-density for the purification of large quantities of IGs from honeybee trophocytes. With this procedure, we verified that SM was present in IGs. We determined the density of IG to be 1.25 g/cm3. We showed that additional magnetic field (1 Gauss) induces size fluctuation in MGs, which triggers the increase of [Ca+2]i via cytoskeletons. A cellular mechanism which allows honeybees to have such a sensitive system for orientation and positioning is thus formulated.
It has been reported that IGs occupy approximately 7% of the volume of the trophocytes in honeybees
Phosphorus is present in IGs of bacteria and honeybees
It has been suggested that the IGs found in honeybees may be related to the exclusion of biological waste or the reduction of biological toxicity of metals. However, the diet of honeybees is honey and pollen. Pollens are the main source of minerals, which include mostly a metal mineral (Ca) and a little trace elements (Al, Cr, and Fe)
Previous histological studies show that trophocytes are the only cell containing IGs in the abdomen including both the dorsal and ventral regions
The saturation magnetization of purified IGs is about 2.83×10−4 emu from the abdomen of 2000 honeybees. Thus, the saturation magnetization of purified IGs per individual is about 1.4×10−7 emu, which is lower than the value of 2.4×10−7 emu as previously reported
Earlier studies analyzed the whole body and abdomen of honeybee by SQUID and showed that intrinsic coercivity is at the interval of 83–103 Oe and 44 Oe, respectively, as derived from hysteresis curves
AFM and MFM have been generally used to study SM particles
EPR has been used to study the magnetic material in the whole body of migratory ants
The temperature dependencies of the HF linewidth of purified IGs spectra are shown in
The resonance linewidth and anisotropy fields of the broad resonance lines (HF) of purified IGs. (A) Temperature dependence of the resonance linewidth of the HF. The solid line is the best fit of the HF data according to Eq. 1, with Δ
The EPR spectra of de-oxygen horse spleen ferritin show that g = 9 signal can be shifted to g = 4.3 with the increment of temperature (80K to 290K) and the intensity of g = 4.3 increases with the increment of temperature (142K to 290K). g = 2.066 signal can be shifted to g = 2.014 with the increment of temperature (19K to 290K)
The EDX spectra of a large amount of purified IGs of ESCA (in this article) are the same as the EDX spectra of an iron granule of STEM
We show in this study the release of calcium ion occurs principally in the trophocytes, but there is a small effect observed in the fat cells. This effect is due to the magnetic field acting on fat cells. Previous studies show that exposure to a static magnetic field resulted in an increased Ca+2 level in numerous cells, including macrophages
Additional magnetic field can induce the fluctuation of MGs in size, which corresponds to the property of SM
Based on the results of cellular studies, we postulate a magnetic map hypothesis for the magnetoreception as follows: the magnetic field induces the size of MGs to shrink at the paralleled direction to the magnetic field and to enlarge at the vertical direction in the horizontal plane (
The magnetic map hypothesis of magnetoreception for orientation and positioning during foraging and homing. (A) When honeybees fly in the direction paralleled to external magnetic field, MG expands due to the repellent force of SM. When honeybees fly in the perpendicular direction to external magnetic field, MG contracts in the same direction due to the attractive force of SM. (B) The direction of magnetic field of SM is parallel to the direction of Earth's magnetic field at flower 1. SM would be in a side-by-side position, creating repulsion and subsequent expansion of the MGs. At the flower 3, the direction of magnetic field of SM is vertical to the direction of Earth's magnetic field. SM would be positioned end-to-end, creating attraction and subsequent contraction of the MGs. This contraction induces the tensing of cytoskeletons and triggers the increase of [Ca+2]i in trophocytes. At flower 2, the relaxing and tensing of cytoskeletons is a mixture of flower 1 and flower 3. The direction of magnetic field of SM at the flower 1 is the same as flower 5. Flower 2 is the same as flower 6, and so on. The distinction in both cases is dependent on the direction of waggle dance.
According to this hypothesis, (1) The flight paths should be straight from the hive to the food source, as has been determined in the foraging and homing of honeybees
‘How do MGs found in the abdomen function as magnetoreceptors’ is an enigma yet to be resolved. Suffice to note that peripheral neurons of insects may play a role independent of the brain, such that a male cockroach can continue with mating, with its head bitten off by his female partner. Certainly, a magnetoreception system for positioning and orientation exists in honeybees, and this simple, primitive, and highly accurate sensing mechanism may be present in all other magnetotactic organisms.
Honeybees,
Two thousand worker bees were freshly collected from a hive, anesthetized with ice-distilled water, washed with ice-distilled water three times, and dissected with stainless tools. Ventral abdomens were obtained, pollen washed out through stainless filter with honeybee saline (156.4 mM NaCl, 2.7 mM KCl, 1.8 mM CaCl2, 22.2 mM glucose, pH 7.3) three times, and centrifuged against a stainless network filter in a centrifugal tube at 3000×g for 2 minutes three times. The abdomens were pressed by stainless bar, dissected into small pieces, and centrifuged against a stainless network filter in a centrifugal tube at 15000×g for 5 minutes three times in honeybee saline containing 0.5 M sucrose and 0.5 M NaCl. Trophocytes were collected into 10 ml honeybee saline containing 0.5 M sucrose, 0.5 M NaCl and 50 µl phenylmethylsulfonyl fluoride, and homogenized by polytron at max speed 30,000 rpm for 3 minutes. The homogenized solution was filtered through a series of filters containing 350 meshes stainless network filter, 10 µm plastic network filter, 20 µm membrane filter, plastic network filters with 5 µm and 1 µm, as well as membrane filters with 10 µm, 8 µm, 5 µm, 3 µm, 1.2 µm, 0.8 µm, and 0.65 µm. The filtered solution was centrifuged against 6.0 m sucrose solutions at 23000×g for 1 hour. The precipitate was collected, resuspended in honeybee saline, and centrifuged against 6.0 m sucrose solutions at 23000×g for 1 hour again. The precipitate was collected, sucrose washed out with honeybee saline, and centrifuged against a 0.22 µm membrane filters in a centrifugal tube at 23000×g for 10 minutes. The precipitate on 0.22 µm membrane filters was collected and analyzed.
The purified precipitates were fixed in 2.5% glutaraldehyde in a 0.1 M phosphate buffer with 0.35 M sucrose at pH 7.4 for 30 minutes at 25°C and were postfixed in 1% osmium tetroxide in a 0.1 M phosphate buffer with 0.35 M sucrose at pH 7.4 for 1hour at 25°C. The purified precipitates 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 JEOL JEM-2000EXII TEM, operating at an accelerated voltage of 100 kV. The elemental composition and crystal structure in selected areas of thin sections were analyzed by energy-dispersive X-ray microanalysis and by selected-area electron diffraction in a JEOL JEM-2000FX STEM operating at an accelerated voltage of 100 kV
The purified IGs from 2 thousand bees were lyophilized with a lyophilizer for six runs. The resulting powder was transferred to tubes and sealed under nitrogen flux to prevent oxygen contributions to the SQUID signal at low temperatures. The magnetization and hysteresis of each sample was measured at a temperature of 300 K and magnetic field intensity from−5000 to 5000 Gause under a SQUID magnetometer (Quantum Design MPMS2).
The purified IGs from 8 thousand bees were prepared according to TEM techniques. Sections (100 nm in thickness) were cut with a diamond knife, and then transferred to silicon chips. A commercially available AFM/MFM (SPA300, Seiko Scientific Instrument, Tokyo, Japan) was used to generate the images. The microscope was operating in Tapping/Lift mode to obtain images of both topography and magnetic force over a region of the samples. High-resolution AFM images of the sample surface topography and magnetic interaction was provided simultaneously.
The purified IGs from 40 thousand bees were lyophilized with a lyophilizer. The resulting powder was pressed into a disk and characterized by ESCA on a Perkin Elmer Phi 1600 ESCA system. Each sample was analyzed by survey and high-resolution spectra employing an Al Kα X-ray anode, energy 1486.6 eV. From the survey spectra, the qualitative and quantative elemental composition and state of the uppermost layer were computed. The binding state of elements Fe was obtained by acquisition of high-resolution spectra and subsequent curve-fitting.
The purified IGs from 80 thousand bees were lyophilized with a lyophilizer. The resulting powder was transferred to EPR quartz tubes and sealed under nitrogen flux to prevent oxygen contributions to the EPR signal at low temperatures. Measurements were performed in a commercial X-band (ν = 9.607GHz) EPR spectrometer (Bruker EMX-10) operation at a microwave power of 20 mW, with a 100 kHz modulation field of about 1.6 Oe (2 A m−1) in amplitude from 77 to 250 K and 4 to 300 K, respectively. Spectra double integration was obtained using WINEPR software (Bruker). The purified IGs spectra were fitted using Origin 7.5 with the HF. Gaussian and Lorentzian derivative-shaped lines were used to obtain the temperature dependence of the resonance linewidth (δ
A large layer of trophocytes and fat cells were freshly detached from the ventral abdomen in honeybee saline, mounted on the transfer slides with a little honeybee saline, covered with cover slides, and then the fluctuation of iron granules in size with or without 1 Gauss of magnetic field was measured under confocal laser scanning microscope (Leica TCS SP2 MP). The distribution of size was analyzed by photoshop software. The change rate (%) of IGs in size was calculated from the following equation:% change = the difference of size of IGs [(test-control)/control]×100%-the difference of size of white spot [(test-control)/control]×100%.
Qualitative changes in [Ca2+]i levels were determined using the cell-permeable fluorescent [Ca2+]i indicator fluo-4-AM. A large layer of trophocytes and fat cells were freshly detached from the ventral abdomen in honeybee saline, loaded for 1 h with 5 µM fluo 4-AM at room temperature, centrifuged at 10,000 xg to wash fluo 4 out for 1 min twice, mounted on the transfer slides with a little honeybee saline, covered with cover slides, and then the variation of fluorescence intensity was measured under confocal laser scanning microscope (Leica TCS SP2 MP). The fluorescence intensity of trophocytes and fat cells were independently measured at the beginning time and at about 12 sec, respectively. 1 Gauss of magnetic field at the center of two magnets was applied to trophocytes and fat cells as closely as possible at about 12 sec, persisted for about 34 sec, and fluorescence intensity was measure at about 46 sec. Same experiments were done on controls without 1 Gauss of magnetic field. Fluo 4 was excited with the 488 nm line of a HeNe laser. The emitted light was collected at 505 nm
Colchicine inhibits microtubule and latrunculin B inhibits microfilaments. A large layer of cells were freshly detached from the ventral abdomen in honeybee saline, treated with 1 µM colchicine and 62.5 nM latrunculin B for 30 min at room temperature, washed colchicine and latrunculin B out with honeybee saline for three times, loaded for 1 h with 5 µM fluo 4-AM at room temperature, centrifuged at 5,000 xg to wash fluo 4 out for 1 min twice, mounted on the transfer slides with a little honeybee saline, covered with cover slides, and the variation of fluorescence intensity was measured under confocal laser scanning microscope (Leica TCS SP2 MP). Fluorescence intensity of trophocytes and fat cells were independently measured at the beginning time and at about 12 sec, respectively. 1 Gauss of magnetic field at the center of two magnets was applied to trophocytes and fat cells as closely as possible at about 12 sec, persisted for about 34 sec, and fluorescence intensity was measured at about 46 sec. Same experiments were done on controls without 1 Gauss of magnetic field. Fluo 4 was excited with the 488 nm line of a HeNe laser. The emitted light was collected at 505 nm
We thank C.S. Ng for the measurement of AFM/MFM, C.T. Hsieh for the discussion of EPR, and P.C. Huang for discussions and critical reading of the manuscript.