Current Biology Magazine

Correspondence A B C No evidence for a Magnetic Magnetic magnetite-based otoconia otoconia Main Main magnetoreceptor otoconia otoconia in the lagena of Hair pigeons cells

E. Pascal Malkemper1, Lagena Daniel Kagerbauer2 , Lyubov Ushakova1, D Fe E Fe Ca Simon Nimpf1, Paul Pichler3 , K Christoph D. Treiber4 , Martin de Jonge5 , Jeremy Shaw6 , and David A. Keays1,* TV

It is well established that an array of avian species sense the Earth’s magnetic fi eld and use this information for orientation and navigation. While the existence F Fe G Fe of a magnetic sense can no longer be Cr disputed, the underlying cellular and biophysical basis remains unknown. It HCs has been proposed that pigeons exploit a magnetoreceptor based on magnetite crystals (Fe3O4) that are located within the lagena [1], a sensory epithelium of Figure 1. X-ray fl uorescence microscopy of serial sections reveals the absence of magnetic the inner ear. It has been hypothesised otoconia in the pigeon lagena. that these magnetic crystals form a (A) Diagram of the pigeon inner ear showing the lagena at the tip of the elongated cochlear bed of otoconia that stimulate hair cells duct. The otoconia are shown in black. (B) Diagram illustrating the magnetic otoconia hypothesis. Shown is a schematic cross-section through the lagena at the level indicated by the dashed line in transducing magnetic information into (A). Hypothetical magnetic otoconia could be linked to a specifi c population of hair cells that are a neuronal impulse. We performed a not associated with calcium-carbonate main otoconia. (C) Alternative to the hypothesis presented systematic high-sensitivity screen for iron in (B), magnetic structures could be interspersed within the primary array of calcium carbonate in the pigeon lagena using synchrotron otoconia in the pigeon lagena (based on [3]). (D) Cross-section of the pigeon lagena showing the X-ray fl uorescence microscopy coupled elemental maps for iron (red), calcium (green), and potassium (blue). The hair cell layer (HCs) and with the analysis of serial sections by tegmentum vasculosum (TV) are rich in iron. The scale bar shows 50 µm. (E) Iron map of the same lagenar cross-section shown in (D). The scale bar shows 50 µm. (F) Enlargement of boxed region transmission electron microscopy. shown in (E), showing the iron-rich hair cell layer (hair cells indicated) and an extracellular iron- We fi nd no evidence for extracellular rich structure. The scale bar shows 10 µm. (G) Enlargement of the extracellular iron-rich structure magnetic otoconia or intracellular boxed in (F). It contains chromium (yellow), indicative of laboratory contamination. The scale bar magnetite crystals, suggesting that shows 5 µm. if an inner ear magnetic sensor does exist it relies on a different biophysical it has been reported that iron is present magnetic otoconia. There could be either mechanism. in the lagenar otoconia of birds [3]. We a discrete subclass of magnetic otoconia Utilising pigeons as a model system, have further reported the presence of that are associated with a specifi c Dickman and Wu [1,2] have reported an iron-rich spherical organelle, the population of hair cells, or magnetic that various regions of the brain, cuticulosome, which is found apically in structures that are interspersed amongst including the vestibular nuclei and the cuticular plate of hair cells of birds the primary bed of calcium carbonate hippocampus, are activated when [4,5]. Cuticulosomes are predominantly otoconia (Figure 1A–C). It is challenging exposed to magnetic stimuli. Removal composed of ferrihydrite nanocrystals to test these ideas as magnetic otoconia of the lagena demonstrated that this and are therefore unlikely to act as cannot be visualized with classical iron- activation is dependent on the structural torque-based magnetoreceptors staining methods (for example Prussian integrity of the inner ear, leading the [6]; however, they closely resemble Blue), which dissolve the structures of authors to suggest that ferrimagnetic structures that are abundant in the cusp interest. Additionally, iron contamination particles stimulate lagenar receptors epithelia of chitons, which are involved from non-biological sources is a that form the basis of a magnetic in the biomineralization of magnetite persistent problem, and high sensitivity sensory apparatus [1]. This hypothesis is teeth [7]. is required to identify magnetite crystals attractive, as the lagena (like the utricle It is therefore conceivable that within a comparatively large anatomical and saccule) contains sensory hair cells cuticulosomes act as an intermediate structure. coupled to dense crystals of calcium iron store, supplying precursor material For this reason, we performed a carbonate, called otoconia. Moreover, for the formation of extracellular high-resolution elemental analysis of

R14 Current Biology 29 , R1–R15, January 7, 2019 Crown Copyright © 2018 Published by Elsevier Ltd. Current Biology Magazine

the pigeon lagena with synchrotron- crystals. Employing this magnifi cation, Austrian Science Fund (Y726 to D.A.K.) and based X-ray fl uorescence microscopy we screened the complete hair cell the Australian Synchrotron Access Program (XFM) coupled with an assessment of layer at regular 30 µm intervals (n = 3 (AS171/XFM/11722 and AS181/XFM/12934). serial sections by transmission electron birds, 161 sections, approximately S.N. is supported by a DOC-fellowship from microscopy. XFM relies on the analysis 17,500 hair cells in total) for crystalline the Austrian Academy of Sciences. We wish of X-rays emitted from a sample of structures from the apical to basal edges to thank Boehringer Ingelheim who fund interest and generates elemental maps of the lagena (Figure S2). We did not basic scientifi c research at the Institute of with exquisite sensitivity. We exploited a identify any structures consistent with Molecular Pathology (IMP). We are indebted to the excellent support facilities at the IMP beamline at the Australian Synchrotron intracellular magnetite crystals. and VBCF, including histology, electron for this purpose, which was optimised In summary, our data do not support microscopy, and bio-optics. to detect iron above a concentration the existence of a magnetite-based 2 of 174 ng iron/cm . Critically, this magnetoreceptor in the lagena of REFERENCES minimal detection limit would permit the pigeons. We fi nd no evidence for identifi cation of a chain of single domain extracellular magnetic otoconia or 1. Wu, L.Q., and Dickman, J.D. (2011). magnetite crystals that could act as a intracellular single-domain crystalline Magnetoreception in an avian brain in part mediated by inner ear lagena. Curr. Biol. 21, torque-based magnetoreceptor (297 ng magnetite in the lagena. We acknowledge 418–423. iron/cm2; Figure S1 and supplemental that our TEM screen may have failed 2. Wu, L.Q., and Dickman, J.D. (2012). Neural calculations in the Supplemental to identify single superparamagnetic correlates of a magnetic sense. Science 336 , 1054–1057. Information). magnetite crystals smaller than 10 3. Harada, Y., Taniguchi, M., Namatame, H., and We prepared 100 µm thick sections nm in size; however, such crystals Iida, A. (2001). Magnetic materials in otoliths of bird and fi sh lagena and their function. Acta. from the apical to basal edges of could not serve as a torque-based Otolaryngol. 121, 590–595. the lagena (n = 3 birds), which were magnetoreceptor that would provide 4. Lauwers, M., Pichler, P., Edelman, N.B., screened using a beam of 10 keV information on the polarity of the Resch, G.P., Ushakova, L., Salzer, Marion, C., Heyers, D., Saunders, M., Shaw, J., and X-rays. A total of 14 elements were magnetic fi eld [8]. Moreover, our fi ndings Keays, D.A. (2013). An iron-rich organelle in the analysed by XFM and X-ray maps were are consistent with that of Zhao et al. cuticular plate of avian hair cells. Curr. Biol. 23, 1–6. generated for iron, calcium, chromium [9], who performed mass-spectrometric 5. Nimpf, S., Malkemper, E.P., Lauwers, M., and potassium (Figure 1D–G and analysis on the lagenar otoconia of Ushakova, L., Nordmann, G., Figure S1). The calcium carbonate pigeons and show a near absence Wenninger-Weinzierl, A., Burkard, T.R., Jacob, S., Heuser, T., Resch, G.P., et al. otoconia were readily identifi able, as of iron. How can these fi ndings be (2017). Subcellular analysis of pigeon hair cells well as the hair cell layer and tegmentum reconciled with the data that magnetic implicates vesicular traffi cking in cuticulosome formation and maintenance. eLife 6 , e29959. vasculosum, both of which were noted stimuli cause neuronal activation in the 6. Jandacka, P., Burda, H., and Pistora, J. (2015). for their high iron content (Figure 1D). vestibular nuclei of pigeons? It is possible Magnetically induced behaviour of ferritin The latter observations were expected that inner ear magnetoreceptors rely on corpuscles in avian ears: can cuticulosomes function as magnetosomes? J. R. Soc. Interface. as it is known that pigeon hair cells electromagnetic induction, rather than 12, 20141087. express high levels of ferritin and a magnetite-based mechanism [10]. It 7. Shaw, J.A., Macey, D.J., Brooker, L.R., Stockdale, E.J., Saunders, M., and Clode, P.L. contain cuticulosomes [4,5], and the has been proposed that the semicircular (2009). Ultrastructure of the epithelial cells tegmentum vasculosum is a highly canals and conductive endolymph may associated with tooth biomineralization in vascularised tissue that is responsible serve as an anatomical circuit for this the chiton Acanthopleura hirtosa. Microsc. Microanal. 15 , 154–165. for producing endolymph. In none of purpose. As a pigeon moves through 8. Winklhofer, M., and Kirschvink, J.L. (2010). A the birds analysed did we observe a a static magnetic fi eld, a voltage could quantitative assessment of torque-transducer models for magnetoreception. J. R. Soc. discrete concentration of extracellular be induced that might be detected Interface 7 , 273–289. iron particles. In some sections we did by a highly sensitive electroreceptor. 9. Zhao, Y., Huang, Y.N., Shi, L., and Chen, L. observe iron-rich structures within the Alternatively, it is conceivable that the (2009). Analysis of magnetic elements in otoliths of the macula lagena in homing pigeons with main otoconia; however, analysis of the brainstem vestibular nuclei are involved inductively coupled plasma mass spectrometry. chromium spectra revealed that these in the multimodal integration of linear Neurosci. Bull. 25 , 101–108. 10. Jungerman, R.L., and Rosenblum, B. (1980). structures were contaminants, most acceleration, gravitational input, and Magnetic induction for the sensing of magnetic likely from the preparation of lagenar magnetic information that originates fi elds by animals--an analysis. J. Theor. Biol. 87 , sections (Figure 1F,G). from primary magnetoreceptors at an 25–32. Next, we considered the possibility unknown location [2]. that single domain magnetite crystals 1Research Institute of Molecular Pathology may reside intracellularly within the SUPPLEMENTAL INFORMATION (IMP), Vienna Biocenter (VBC), Campus- Vienna-Biocenter 1, 1030 Vienna, Austria. iron-rich hair cell layer in the lagena. To 2 Supplemental Information includes Atominstitut, TU Wien, Stadionallee 2, 1020 this end, we prepared serial ultrathin Vienna, Austria. 3School of Life Sciences, experimental procedures, calculations of sections encompassing the entire University of Sussex, Brighton BN1 9QG, UK. detection limits and two fi gures and can be sensory epithelium of the lagena for 4 Centre for Neural Circuits and Behaviour, The analysis by transmission electron found with this article online at https://doi. University of Oxford, Oxford, UK. 5Australian org/10.1016/j.cub.2018.11.032. microscopy (TEM). Bright-fi eld TEM Synchrotron, ANSTO, 800 Blackburn Road, 6 imaging of magnetotactic bacteria Clayton, Victoria 3168, Australia. Centre for ACKNOWLEDGEMENTS Microscopy, Characterisation and Analysis, revealed that a 7,100 x magnifi cation The University of Western Australia, Crawley, enables the swift identifi cation of This work was supported by the European WA 6009, Australia. electron dense 50 nm size magnetite Research Council (336725 to D.A.K.); the *E-mail: [email protected]

Current Biology 29 , R1–R15, January 7, 2019 R15