|Home | About | Journals | Submit | Contact Us | Français|
The giant fiber system (GFS) of Drosophila is a well-characterized neuronal circuit that mediates the escape response in the fly. It is one of the few adult neural circuits from which electrophysiological recordings can be made routinely. This article describes a simple procedure for stimulating the giant fiber neurons directly in the brain of the adult fly and obtaining recordings from the output muscles of the giant fiber system.
The giant fiber system (GFS) mediates a fast escape behavior in adult flies (Allen et al. 2006). Behaviorally, it is characterized by an initial extension of the mesothoracic leg, to propel the flies off the substrate, followed by a wing downbeat to initiate flight. The efferent (output) pathways of the GFS have been well defined (Figure 1) for the most part by work from Wyman and others in the 1980’s using a combination of dye-injection, EM and electrophysiological techniques (Ikeda et al. 1980; King and Wyman 1980; Koto et al. 1981). The two largest interneurons in the fly, the aptly named giant fibers (GFs), relay the signal from the brain to the mesothoracic neuromere where each makes two identified synapses. The first is to a large motorneuron (TTMn) that drives the tergotrochanteral “jump” muscle (TTM), which is also referred to in the literature as the tergal depressor of trochanter or TDT. This GF-TTMn synapse is the largest central synapse in the fly and is a mixed synapse with the electrical gap-junction component encoded by the shaking-B (shakB) gene and the chemical component using acetylcholine as its neurotransmitter (Blagburn et al. 1999; Allen and Murphey 2007; Phelan et al. 2008). The second identified synapse of the GF is to another interneuron, the peripherally synapsing interneuron (PSI), which exits the ganglion via the posterior dorsal medial nerve (PDMN) and synapses with dorsal longitudinal motorneurons (DLMns) within the PDMN. The DLMns drive the large indirect flight muscles (DLMs). Electrophysiological recordings can be made from the GFS in a simple non-invasive manner to determine the function of the central synapses within the circuit. Using combinations of adult viable mutants and/or GAL4 lines that express in its neurons, the GFS has provided a useful model circuit to investigate the role of several molecules in the formation of central synapses including Glued, Rac1, Robo, Semaphorin1a and Neuroglian (Allen et al. 1999; Allen et al. 2000; Godenschwege et al. 2002a; Godenschwege et al. 2002b; Godenschwege et al. 2006). The GFS has also been used to investigate the effects of aging, sensitivity to anesthetics, the effects of neurodegeneration and the molecular basis of habituation (Engel and Wu 1996; Lin and Nash 1996; Engel and Wu 1998; Martinez et al. 2007; Watson et al. 2008).
The GFs can be activated directly with brain stimulation and the two output pathways can be monitored by recording simultaneously from the TTM and DLMs. The original rationale was that by placing the stimulating electrodes into the brain and slowly increasing the stimulation voltage, a point would be reached where only the GF interneurons would propagate an action potential since their large size would mean they have the least resistance and thus the lowest threshold. While this may theoretically be true, in practice accurate positioning of the electrodes is hard to achieve so the stimulation voltage given is much above threshold. This ensures that the GFs are activated directly and not by upstream neurons (unless that is desired, see below). Though many neurons in the brain may be activated, the only route to the TTMs and DLMs from the brain activated by this procedure seems to be via the GFs. This is suggested by the findings that genetic ablation of the GFs, or abrogation of the electro-chemical synapses between the GF and the TTMn and PSI, results in total loss of TTM and DLM responses upon brain stimulation (Allen et al. 2000; Allen and Murphey 2007). However, both TTMn and the DLMns have other unidentified inputs, one of which is triggered by looming stimuli (Fotowat et al. 2009). Once direct activation of the GFs is achieved, recordings from TTM monitor the function of the GF-TTMn central synapse along with the neuromuscular junction (NMJ) and recordings from DLM monitor the function of the GF-PSI and PSI-DLMns synapses as well as the NMJ.
The most commonly used tests for the GFS are the response latency, the refractory period and the ability to follow to high frequency stimulation. These will be described in turn.
This is the time taken for the output muscle to respond to a single stimulus activating the GFs. In the TTM of wild type flies this is ~0.8 ms after GF activation and is via the monosynaptic pathway through the large electro-chemical GF-TTMn synapse. The response in a DLM, through the disynaptic pathway, is seen ~1.2 ms after GF activation. These latencies correspond to the escape behavior where the jump always occurs before the wing downbeat. This robust short-latency (SL) response is a good indicator of synaptic function and any abnormalities in the synapses of the GFS will result in an increase in the latency or a loss of the response. For example, loss of gap junctions or structural malformations of the synapse that alter its shape or size (Thomas and Wyman 1984; Oh et al. 1994; Allen et al. 1999; Allen et al. 2000; Godenschwege et al. 2002a; Godenschwege et al. 2002b; Godenschwege et al. 2006; Allen and Murphey 2007; Uthaman et al. 2008).
In addition to SL responses, intermediate-latency (IL) responses (TTM ~1.8 ms, DLM ~2.2 ms), and long-latency (LL) responses (TTM ~3.9 ms, DLM ~4.3 ms) can be elicited by simply reducing the voltage during brain stimulation, or providing a light-off stimulus to a tethered fly. All these responses are still conducted through the GF; note the delay between the TTM and DLM response is always ~0.4 ms, indicating the disynaptic pathway from GF to DLM via the PSI and DLMn. The longer IL and LL responses, during low-voltage electrical stimulation or a light-off stimulus, are attributed to indirect activation of the GF by the afferent neurons in the brain. These neurons still remain unidentified but have interesting properties as they show both sensitivity to anesthetics and habituation to repeated stimuli (Engel and Wu 1996; Lin and Nash 1996; Engel and Wu 1998).
In this test twin stimuli are given, initially 10 ms apart, and the responses from both TTM and DLM recorded. The interval between the two stimuli is then gradually reduced until the second stimulus fails to elicit a response. The shortest time between two stimuli that still produces two responses is defined as the refractory period. For TTM this is ~3 ms and DLM is ~5 ms due to the greater time needed for the PSI-DLMn chemical synapses to replenish their synaptic vesicles. This test is less common than the other two as similar information can be gleaned if you observe the responses to the first two stimuli in the “following at high frequencies” test (see below).
In this test a train of ten stimuli are given to the preparation at high frequency and the number of responses is recorded. These trains of stimuli are usually given at 100, 200 & 250 or 300 Hz. At 100 Hz (stimuli 10 ms apart) both TTM and DLM should respond 1:1 and give ten responses. At the higher frequencies e.g. 250 Hz (stimuli 4 ms apart), TTM will still respond 1:1 due to the robust GF-TTMn electro-chemical synapse, however, DLM recordings will start to show failures as the time between stimuli is less than the refractory period of the PSI-DLMns synapses. An alternative way of performing the test is to gradually increase the frequency of the stimuli until the response rates fall below 50% (5 out of 10). This is described as the Following Frequency50 (FF50) (Gorczyca and Hall 1984). This test will often reveal an abnormality in synaptic function that does not cause an abnormal response latency (Allen et al. 1999), although it usually confirms an aberrant response latency.
This protocol is a standard method for recording from the giant fiber system of Drosophila. It is a relatively non-invasive method that allows the investigator to stimulate the giant fibers in the brain and assay the function of several central synapses within this neural circuit by recording from the thoracic musculature.
General Note: Successful recording from the GFS relies on being able to arrange the 5 micromanipulators so that the electrodes can be placed within several mm of each other. It is worth spending some time moving and adjusting these before a preparation is introduced so that minimal adjustment is required when recordings are needed.
The recording electrodes are sliding on the cuticle without being able to pierce it on the appropriate spot in order to impale the correct muscle.
The more perpendicular the electrode is to the cuticle, the easier it is for the electrode to get through the cuticle. Move the electrode to a slightly different area within the target area, change the angle of the micromanipulator itself, try for the muscle on the contra lateral side or re-mount the fly in a differently angled position.
The recording electrodes are indenting the cuticle or the recording electrodes are bending without piercing the cuticle.
Make sure your electrode is not broken and has the appropriate shape. The tip of your electrode should have the approximate shape and size similar to the posterior Supra-Alars setae (Figure 3B). In case the electrode is not broken and has the appropriate shape, try gently tapping on the back of the forward moving knob of the micromanipulator (once there is slight indentation) to encourage penetration through the cuticle.
No stimulation artifact and no response.
Check whether all equipment is turned on. Double-check whether the fly is responding upon stimulation (step 4). If it doesn’t there is something wrong with your stimulation (check stimulation electrodes, ground and stimulator settings etc.). If the fly does respond then there is something wrong with your recording (check recording electrodes and amplifier settings etc.).
The muscle response has an unusual shape with multiple peaks.
This occurs when the microelectrode is not recording from a single muscle cell. This can occur in recordings from either muscle but is more common in TTM recordings since this muscle is composed of many small fibers and maintaining the position of the electrode after several muscle contractions is problematic. An unusual shaped or multi-peaked response trace does not affect the data since response latencies and followings will still be preserved.
There is a very large stimulation artifact obscuring the muscle repsonse and/or recordings of multiple stimuli are drifting on the recording monitor.
Double-check whether the ground electrode is properly in the fly and double-check the voltage and duration of the stimuli given. Also, when the hemolymph dries up around the ground wire it results in its loss of conductance. This can be prevented and recovered with a little drop of saline on the fly where the ground electrode enters the abdomen.
Obtaining long latencies or no responses in wild type flies.
Double-check whether you are in the correct target area for the appropriate muscle. Alternatively, your electrode might have gone in too far and you pierced through the correct muscle. Both muscles are just underneath cuticle. An approximate measure is that the cuticle is no thicker than 2–3x of the thickness of a posterior Supra-Alars setae at it thickest visible point (Figure 3B). Alternatively your stimulation is below threshold, therefore try increasing the voltage (duration). If you have used CO2 to anaesthetize the fly, either leave the fly to recover from CO2 longer before testing or anaesthetize flies using ice. Finally, your wild type fly may be a mutant.
Obtaining very short latencies for both, TTM (<0.7 ms) and DLM (< 1 ms).
This occurs if the ventral nerve cord, and thus the TTMn and DLMn motorneurons are being activated directly. Check the position of the stimulating electrodes and replace them in the brain if necessary.
In wild type flies average response latencies to a single stimulus are in the range 0.8 ms +/− 0.1 ms for the GF-TTM pathway and 1.4 ms +/− 0.3 ms for the GF-DLM pathway depending on genotype and genetic background. Similarly, with respect to following frequencies the GF-TTM path is able to follow 10 stimuli 1:1 up to 300 Hz and the GF-DLM pathway up to 100 Hz but variability between individual flies of different genotypes and genetic background have been observed. Hence, it is important to choose carefully the appropriate control flies when analyzing the electrophysiological phenotypes of mutants or targeted disruptions in the GFS. Two classical mutants that do affect the function of the GFS dramatically are shakB2 and bendless (Thomas and Wyman 1984; Blagburn et al. 1999; Allen and Murphey 2007; Phelan et al. 2008; Uthaman et al. 2008). In shakB2 flies the GF-TTMn synapse lacks the gap junctions but the chemical component is still present. The average response latency for the TTM in these flies is consistently increased to an average of 1.5 ms and it is not able to follow stimuli given at either 100, 200 or 300 Hz due the weak labile nature of resultant GF-TTMn synapse. In addition, no responses are obtained from the DLM when the GF is stimulated in the brain. Proof that the lack of responses are not due to a defect at the NMJ comes from the ability to record responses from the DLM muscle, when the motorneurons are stimulated directly by placing the stimulation electrodes in the thorax (Thomas and Wyman 1984). In contrast, in bendless flies the GF-DLM pathway remains unaffected when compared to wild type control flies. However, the GF-TTM connection is consistently increased to an average of more than 2 ms and is not able to following stimuli given at either 100, 200 or 300 Hz.
The reason that these indirect electrophysiological tests of these central synapses of the GFS are successful is that the NMJs at both TTM and the DLMs are large and extensive with many synaptic boutons. They rarely fail; the motorneurons can be stimulated directly at frequencies up to 500 Hz and the muscles will still show 1:1 responses to stimuli (MJA & TAG, unpublished data). Thus any effects seen on transmission through the pathways from the GF can be attributed to central synaptic defects. If defects are seen when testing it is always prudent to stimulate the motoneurons directly to confirm that the NMJs are functioning correctly in at least a few flies of the same genotype, because some mutants do affect the adult NMJ (Huang et al., 2006).
Work in the M.J.A. lab has been supported by the Wellcome Trust and the Leverhulme Trust. T.A.G. is supported by R01 HD050725. Thanks to Robin Konieczny for the artwork in Figure 1. We also owe much to R.K. Murphey for his enthusiasm and encouragement regarding the GFS.