Introduction

The genome sequence of the fruit fly Drosophila melanogaster was published in 2000. One year later, a study screening the fly genome for disease-related genes followed: 548 genes implicated in 714 human diseases were found. The first author of the study, Laurence Reiter, commented “this came as a bit of a surprise as most people don’t think to study hearing or cancer in Drosophila”. In terms of hearing, this is entirely true: we have known that Drosophila can hear since the 1960s; however, for a long time, only few scientists investigated hearing in flies. The findings of these studies were astonishing and hearing in fruit flies has thus become a very competitive field of research over the past years. Not only was the availability of new genetic methods to analyze and manipulate the function of sensory cells in Drosophila essential, but the identification of many functional parallels between fly and vertebrate hearing was also crucial. Another important aspect was the ability to analyze molecular hearing mechanisms noninvasively, due to the direct coupling between molecular processes and the systematic performance of the fly ear.

The Drosophila ear

Drosophila males use courtship songs to attract females. When females are in the proximity, males extend one of their wings and vibrate it in a regular pattern, resulting in a courtship song dominated by frequencies of around 200 Hz. These songs enhance the receptivity of the females to copulation and animate other males to sing along.

Males as well as females detect the courtship songs with their antennae. These are located at the front of the fly head (Fig. 1). Every antenna is composed of three main compartments—the proximal scapus, the pedicellus and the distal funiculus. The club-shaped funiculus and its feathery arista are the sound receivers, functionally corresponding to our tympanic membrane. As the arista is stiffly coupled to the funiculus, both compartments swing around the longitudinal axis of the latter upon exposure to sound. The funiculus rotation is directly detected by the mechanosensory cells of the Johnston’s organ—the fly equivalent of the organ of Corti in the human ear. The Johnston’s organ is located in the pedicellus of the antenna and consists of about 200 chordotonal stretch receptors. Each of these receptors contains two or three sensory cells and several supporting cells, which form via mitotic division of a precursor cell. This precursor cell is specified via the transcription factor atonal, which also regulates the development of hair cells in the vertebrate ear. Despite this similarity, fly auditory sensory cells and vertebrate hair cells are anatomically different: in flies these cells are bipolar mechanosensory neurons with a proximal axon and a distal ciliated dendrite that is directly attached to the funiculus via an extracellular cap. The vibrations of the funiculus are converted into electrical signals, which are then encoded as action potentials by the neurons and transmitted via their axons into the antennal mechanosensory motor center in the fly brain.

Regarding their axonal projection target, the approximately 500 mechanosensory neurons can be subdivided into five classes, of which only the first two are apparently true “hearing cells”: Calcium imaging showed that the 250 neurons of classes A and B are very sensitive to vibrations of the funiculus, whereas the 200 neurons of classes C and E preferentially respond to sustained mechanical deflection of the funiculus and detect gravity and wind. The remaining 50 class D neurons are activated by prolonged deflection and vibration of the funiculus, although large deflection and vibration amplitudes are required.

Auditory mechanisms: active amplification and transduction

The Drosophila ear does not contain a middle ear and sensory cells are directly coupled to the sound receiver. Therefore, the neurons directly influence the mechanistic behavior of the sound receiver, which can be measured noninvasively via laser Doppler vibrometry. Analyses of the vibration characteristics of the antennal receiver have shown that it exhibits all features of the cochlear receiver in vertebrate ears, i.e. it is an active amplifier based on the active movement of the hair cells. Genetic manipulations showed that corresponding active deflection of the mechanosensory neurons in the Johnston’s organ cause active amplification in the fly ear and that only class A and B auditory neurons are necessary: without these neurons, amplification is disrupted, whereas the loss of the other neuron classes does not affect this process.

In addition to active amplification characteristics, mechanical signatures of transduction can be identified by analyzing the performance of the fly antennal receiver and these are similar to the mechanisms detectable in the mechanosensory hair bundles of hair cells (Fig. 2). Quantitatively, these transduction signatures in the fly ear can be well explained by the “gating spring model” of hair cell transduction. This model assumes a serial arrangement of mechanically activated ion channels, “gating springs” and motors. The gating springs are elastic elements that transfer force to the ion channel, whereby the deformation of the spring determines the opening probability of the channel. Upon opening of the channel, the gating springs are relieved, which reduces the stiffness of the antennal receiver. After prolonged deflection, channels are closed by motors and the original stiffness is restored.

Physical simulations have revealed that the interplay between opening of the ion channels and the resulting motor movements is sufficient to explain active amplification in the fly ear. This implies that active amplification and transduction are coupled. The same transduction-driven amplification mechanism explains the active motility of the sensory hair bundles of vertebrate hair cells, which, alongside Prestin-driven somatic motility, contributes to active amplification in the vertebrate ear.

Channels, gating springs and motors

Several ion channels of the transient receptor potential (TRP) family are involved in transduction and amplification in the fly ear, including the TRPN1 channel “No mechanoreceptor potential C” (NOMPC), and the TRPV channels Nanchung (Nan) and Inactive (Iav).

NOMPC is a mechanically activated mechanotransduction channel that localizes to the distal ciliary region of the mechanosensory neurons. Loss of the channel leads to a complete loss of active amplification and sensitive hearing response. Without NOMPC, only loud sounds can evoke an electric response in the fly ear. The latter responses likely originate from the gravity and wind-sensitive class C and E neurons, since they persist when class A and B neurons are ablated and correlate with the calcium signals in class C and E neurons. Signatures of the channel opening mechanics in the antennal receiver provide evidence that NOMPC could be the transduction channel of the class A and B auditory neurons: analyzing the mechanical responses of the sound receiver over a broad range of stimulus amplitudes, the opening signatures of at least two ion channel types are visible. One channel type is very sensitive and mainly associates with class A and B auditory neurons. The second channel type is rather insensitive and mainly affiliated with the gravity- and wind-sensitive class C and E neurons. Class C and E neurons function in a NOMPC-independent manner; however, mechanical opening of the sensitive channels requires NOMPC. Without NOMPC, these channels are mechanistically decoupled from the antennal receiver, suggesting that NOMPC forms the sensitive channels or transfers the vibrations of the sound receiver to the channels.

Both TRPV channels Nan and Iav are located in the proximal ciliary region of the mechanosensory neurons, possibly forming heteromeric Nan–Iav channels. Loss of Nan–Iav leads to complete loss of electric responses in the neurons; however, active amplification persists and becomes excessive. The complete abolishment of electrical responses indicates that Nan–Iav could form the transduction channel. NOMPC would then act as preamplifier and, together with the motors, generate the amplification, whereas Nan–Iav would generate the gross transduction current. Alternatively, Nan–Iav could propagate the electric signals of NOMPC and the as yet unidentified less sensitive transduction channels. However, as long as it remains unclear whether Nan–Iav channels are mechanically activated mechanotransduction channels, we cannot discriminate between these two possibilities.

Gating springs, the elastic components that transmit mechanical force to the mechanotransduction channels, have not yet been identified on a molecular level. NOMPC itself seems to be a good candidate, as its N-terminus contains a molecular spring composed of 29 ankyrin residues. Recent findings suggest that this spring couples the NOMPC channel to microtubules. Manipulation of the number of ankyrin residues will be needed to show if these act as a gating spring. It is also possible that the cell membrane embedding these channels acts as the gating spring—a possibility that is currently discussed for hair cells.

Similarly, the motors that apparently drive active amplification in the fly ear together with NOMPC have also not been molecularly identified, whereby the cilia of the mechanosensory neurons in the fly ear show dynein-like arms and axonemal dyneins are therefore excellent candidates. A genetic screen recently identified several axonemal dynein genes expressed in the fly ear. Initial results suggest that these genes are required for active oscillation amplification. If dynein motors do drive active amplification in the fly ear, the amplification and transduction mechanisms used by the sensory cells in the fly would be similar to vertebrate hair cells, but the molecules involved would be different. While TRP channels function as auditory transduction channels in the fly, transmembrane channel-like (TMC) proteins are discussed as auditory transduction channels in hair cells. In addition, hair cells use prestin and myosin motors for active amplification. Although prestin is present in the fly ear, it remains unclear whether it plays a role in the amplification process.

Ciliary genes and chemo- and photoreceptor proteins

The auditory sensory cells of flies and vertebrates are evolutionarily related, although they evidently use different proteins for amplification and transduction. This is mirrored by the use of a similar genetic toolkit: of 274 recently identified genes expressed in the fly ear, every fifth has a counterpart in humans that is involved in hearing defects. Many of these genes encode motors and channels. Moreover, many conserved ciliary genes are involved in fly hearing and mutations in several of these genes cause primary ciliary dyskinesia and Kartagener syndrome in humans. These inheritable diseases are caused by defects in primary cilia, the same type of cilia that play a role in fly hearing. New genes causing these diseases were identified using the fly ear, thus proving it to be an interesting model system and not just for studying hearing.

Surprisingly, some of the 274 genes identified in the fly hearing organ encode chemo- and photoreceptor proteins. One such example is the new family of ionotropic receptors, chemoreceptor proteins present in certain chemosensory cells of the fly. Additionally, nearly all components of the phototransduction cascade from the fly eye, including visual rhodopsin and the TRP channels forming the phototransduction channels, were found in the fly ear. Functional analyses have shown that mutations in most of these genes lead to hearing defects in flies. Even the visual rhodopsins, known to collect photons in the fly eye, have proven to be essential in fly hearing. Several visual rhodopsins that differ in their sensitivity to spectral light are expressed in the auditory sensory cells in the fly ear. The rhodopsins localize to the sensory cilia of the cells and increase the sensitivity of the transduction channels, possibly by modulating the stiffness of the cell membrane. Similar rhodopsin-dependent stiffness modifications were recently described in fly photoreceptors, where these modifications are likely to activate the phototransduction channels. In contrast, the function of rhodopsins in the fly ear seems to be independent of light, raising the question of which stimulus activates the rhodopsins in the fly ear.

Evolution of sensory signaling cascades

The function of chemo- and photoreceptor proteins in the auditory sensory cells of the fly is interesting in terms of their evolution: fly chemo- and photoreceptors detecting chemical or light stimuli via ionotropic receptors or visual rhodopsins, respectively, as well as auditory sensory cells and vertebrate hair cells are specified by the transcription factor atonal. This suggests that these cells are evolutionarily related and originate from a common ancestral cell. These ancestral or protosensory cells were presumably present in every segment of the fly. This serial arrangement is still conserved in the chordotonal stretch receptors used in fly hearing, indicating that these receptors are very similar to these protosensory cells. During the course of segment specialization mediated by Hox genes, these protosensory cells evolved into chemo- and photoreceptor cells in certain body parts; whereas cells in other body regions retained their function as chordotonal stretch receptors. The fact that these stretch receptors use chemo- and phototransduction proteins suggests that these proteins did not arise with the evolution of chemo- and photoreceptor cells, but were already present and played a sensory role. Analyzing the function of these proteins in the fly ear promises to provide insights into the evolution of sensory signaling cascades and possibly the original sensory function of chemo- and photoreceptor proteins.

The fly ear as multifunctional organ

As mentioned above, in addition to hearing, the hearing organ of the fly serves the detection of gravity and wind. New findings indicate that the auditory sensory cells are also important for sensing temperature. Temperature sets the circadian clock and an important gene involved in this process is expressed specifically in chordotonal stretch receptors, including those in the fly ear. Similarly, a connection between temperature sensation and chordotonal stretch receptors was found in fly larva. Additionally, TRP channels, which play a key role in temperature sensation, are expressed in a subset of sensory cells in the flies Johnston’s organ. It is as yet unclear how the fly ear detects temperature and how it encodes mechanical as well as thermal stimuli. Considering the functional diversity and the multiple key components yet to be discovered, research on the fly ear remains fascinating.

Fig. 1
figure 1

Fly hearing organ a lateral view of the fly head. The antenna comprises the scapus (s), the pedicellus (p) and a funiculus (f) with a laterally attached arista (a). b Schemata of the antennal ear. The arista (a) and the funiculus (f) together form the sound receiver and rotate around the longitudinal axis of the latter when stimulated acoustically (dotted line and black arrows). Vibration of the funiculus is directly detected by the mechanosensory neurons of the Johnston’s organ in the pedicellus (p), which thus act as gravity and wind receptors or auditory sound receptors. c Schemata of a single mechanosensory neuron of the Johnston’s organ with its proximal dendritic cilium, soma and axon. The Nan-Iav TRP ion channels localize to the proximal part of the cilium, the NOMPC TRP channels to the distal ciliary tip

Full size image
Fig. 2
figure 2

a Active amplification and transduction. Sound-induced vibrations of the antennal receiver (1) are transferred to the neurons of the Johnston’s organ (2), which then convert the vibration into an electric signal (3). The transmission is realized by mechanically activated ion channels. The coupling of the mechanical stimulus to the ion channels occurs via “gating springs”. Together with motors (4), opening of the ion channels causes an active movement of the neurons, which amplify the sound-induced vibrations of the sound receiver. b Structure of the NOMPC protein. Besides a channel pore region, NOMPC contains 29 ankyrin residues that possibly act as a “gating spring”. Adapted from Göpfert [4] (a) and Zanini und Göpfert [10] (b)

Full size image