A sensory feedback circuit coordinates muscle activity in Drosophila

Abstract

Drosophila larval crawling is a simple behavior that allows us to dissect the functions of specific neurons in the intact animal and explore the roles of genes in the specification of those neurons. By inhibiting subsets of neurons in the PNS, we have found that two classes of multidendritic neurons play a major role in larval crawling. The bipolar dendrites and class I mds send a feedback signal to the CNS that keeps the contraction wave progressing quickly, allowing smooth forward movement. Genetic manipulation of the sensory neurons suggests that this feedback depends on proper dendritic morphology and axon pathfinding to appropriate synaptic target areas in the CNS. Our data suggest that coordination of muscle activity in larval crawling requires feedback from neurons acting as proprioceptors, sending a "mission accomplished" signal in response to segment contraction, and resulting in rapid relaxation of the segment and propagation of the wave.

Figure 1. Larval crawling and its three interacting systems

A) Wildtype larva with Mhc-GFP labeled muscles, crawling from left to right. In 5 seconds, 11.5 peristaltic waves repeat from tail to head in a stereotyped pattern (red stripes). A single wave (box) is enlarged in B). Images are false-colored: areas of highest muscle contraction and highest GFP-intensity are red, while the most relaxed areas are blue. C) The muscle pattern of a whole larva is shown with Mhc-GFP. D) The CNS and PNS are revealed by scratch-Gal4 driving UAS-cd8-GFP. Anterior is to the right for all panels.

Figure 2. Inhibiting the PNS disrupts larval crawling

A, B) 5-40-Gal4 driving UAS-cd8-GFP and UAS-nls-RedStinger reveals the sensory neuron nuclei (yellow) and cell membranes with elaborate arborizations (green). Expression includes md neurons (blue labels), es neurons (green labels) and cho neurons (red labels). The diagram shows Gal4 driver expression patterns, with blue, green, and red boxes corresponding to the five classes of md neurons, the es neurons as a whole, and the four sets of cho neurons. Colored boxes represent expression, gray boxes represent lack of expression. C) The negative control larva (w/UAS-shi^ts; Mhc-GFP/+ ; UAS-shi^ts/+) shows a typical wildtype crawling pattern of about 2 waves/sec (5 sec shown here and in all moviestrips). D) Inhibiting the PNS with 5-40-Gal4 driving UAS-shi^ts causes a major disruption to the crawling pattern, here showing 5 sec with only a single very slow, tight forward wave. E) Another 5-40-Gal4 inhibition experiment shows two slow, tight backwards waves and head curls. F, G) NP5092-Gal4 driving UAS-shi^ts shows similar defects. H-K) Enlarged views of 12 frame segments marked with white lines in D-G. L) The sensory neurons of a single abdominal hemisegment are represented with mds (blue), chos (red) and es neurons (green), shown with anterior to the left. M) Expression patterns of the Gal4 drivers in third instar larvae are diagrammed.

Figure 3. Multidendritic neurons provide the major component of sensory feedback

A) 109(2)80-Gal4 driving UAS-cd8-GFP and UAS-nls-RedStinger reveals the nuclei (yellow) and arborizations (green) of the multidendritic neurons (blue labels). B) Inhibiting the md and es neurons with clh201-Gal4 driving UAS-shi^ts disrupts the crawling pattern, here showing only 1.5 waves in 5 sec. C) Inhibiting only the mds with 109(2)80-Gal4 driving UAS-shi^ts produces a similar defect. D, E) In contrast, inhibiting the chordotonal neurons with 9-20-Gal4 or 8-73-Gal4 driving UAS-shi^ts does not disrupt crawling. F-I) Enlarged views of 12 frame segments marked with white lines in B-E.

Figure 4. Only a subset of the md neurons is necessary for proper crawling

A) Inhibiting just the bipolar dendrite and class I mds with NP2225-Gal4 driving UAS-shi^ts produces defective crawling with slow, tight waves. B) Inhibiting the class II and III neurons with 1003.3-Gal4 driving UAS-shi^ts does not interfere with normal crawling. C, D) Inhibiting the class IV mds with clh24-Gal4 or 7-33-Gal4 does not interfere with normal crawling. E-H) Enlarged views of 12 frame segments marked with white lines in A-D.

Figure 5. The bipolar dendrites and class I mds are partially redundant

A) clh8-Gal4 driving UAS-cd8-GFP and UAS-nls-RedStinger shows expression in the three bipolar dendrite neurons (dbd, lbd, and vbd) as well as dmd1. B) 2-21-Gal4 driving UAS-cd8-GFP shows the three class I neurons (ddaD, ddaE, and vpda) plus dbd. C) Combining the two Gal4 drivers gives expression in all bipolar dendrites and class I mds, plus dmd1. D, E) Inhibiting the bipolar dendrites with clh8-Gal4 or 8-113-Gal4 driving UAS-shi^ts causes mild defects in crawling, seen here as slower, tighter waves. F) Inhibiting the class I mds with 2-21-Gal4 also causes mild crawling defects. G, H) Inhibiting all bipolar dendrite and class I mds simultaneously with combinations of the Gal4 drivers causes severe disruptions in the crawling pattern. I-L) Enlarged views of 12 frame segments marked with white lines in E-H.

Figure 6. The bipolar dendrites and class I mds are necessary for proper wave speed and tightness of contraction

Parameters of crawling behavior were quantified from five individuals for each experiment. Gal4 drivers that inhibit the bipolar dendrite and class I mds are represented by blue bars, drivers inhibiting either the bipolar dendrites or the class I mds have light blue bars, and drivers that do not inhibit the bipolar dendrites and class I mds have yellow bars. Negative controls A and B refer to experiments with the UAS-shi^ts responder alone and the 5-40-Gal4 driver alone, respectively, while wt larvae carry only the Mhc-GFP insert. A) Duration of the peristaltic wave is significantly longer when the bipolar dendrites and class I mds are inhibited. Wave duration is the time from the contraction at the tail to contraction at the head, and excludes the between-wave pause interval. Wildtype waves have a short duration of <500ms (yellow bars), but when the bd and class I md neurons are inhibited the wave duration increases up to 2000-3000ms (blue bars). When only the bds or class I mds are inhibited separately, the wave duration is near the wildtype range (light blue bars). Because 109(2)80-Gal4 × UAS-shi^ts crawling was dominated almost entirely by backwards waves, the wave duration was calculated from backwards waves. B) Wave frequency is decreased when the bipolar dendrites and class I mds are inhibited. Wave frequency is about 1.5-2.0 waves/sec for negative controls and for all the non-bd/class I experiments (yellow bars), but waves are much less frequent (<0.5 waves/sec) when the bd and class I mds are inhibited (blue bars). When only the bds or class I mds are inhibited alone, the wave frequency is within the normal range (∼1.5 waves/sec; light blue bars). C) Segments contract more tightly when the bd and class I mds are inhibited. Bars show the segment width for segment A4 at the maximum contraction, as a fraction of the relaxed segment width between contractions, calculated from 5 waves from 5 individuals per experiment. D) The fraction of backwards waves is increased strongly in the 109(2)80-Gal4 line, and weakly with some of the other sensory inhibited lines. Bars are slightly offset to clarify the several data points with zero backwards waves. Error bars represent standard deviation.

Figure 7. Altering sensory neuron morphology or axon projections disrupts sensory feedback

Misexpression of the identity gene cut is known to result in defective dendritic arborizations in md neurons (Grueber et al., 2003) and possibly projection defects as well (Merritt et al., 1993). Misexpression of unc5 is known to cause repulsion of neurons from the CNS midline (Keleman and Dickson 2001). Expression of either gene in sensory neurons interferes with their sensory feedback function and disrupts crawling behavior. A, B) Misexpression of cut in the md and es neurons with clh201-Gal4 driving UAS-cut results in defective crawling consisting of slow tight waves, similar to the result of inhibited the neurons. C, D) Enlarged views of frames marked with white lines in A,B). E) Normal pattern of sensory projections within the CNS neuropile in the negative control (w 5-40-Gal4/w; UAS-cd8-GFP Mhc-GFP/+). Top, whole CNS; bottom, boxed region enlarged. F,G) When Unc5 is misexpressed, the sensory projections fail to project normally within the CNS, and either stall in the lateral region (asterisk) or fail to enter the CNS at all (arrow; genotype is w 5-40-Gal4/w; UAS-cd8-GFP Mhc-GFP/GS-unc5; +/UAS-Unc5). The CNS shown in (F,G) belonged to the larva whose crawling is shown in (I). H) FasII staining (red) reveals the normal pattern of axon bundles (arrow), showing that the overall structure of the CNS is not disrupted despite misrouting of the sensory neurons (green). Anterior to left for (E-H). I, J) Misexpression of Unc5 in the PNS with 5-40-Gal4 and UAS-unc5 causes crawling defects similar to the inhibition experiments. K, L) Enlarged views of frames marked with white lines in G,H). For all Cut and Unc5 experiments, first instar larvae are shown because the crawling defects causes lethality at later stages.

Figure 8. A “mission accomplished” model for sensory feedback

Crawling requires the integration of motorneuron activity, muscle contractions, and sensory feedback. We suggest that the main function of the bd/class I md sensory feedback is to report a “mission accomplished” signal to the CNS after a successful contraction, and thereby induce muscle relaxation and forward propagation of the wave. Larval musculature is shown at top, with abdominal segments A3-A5 marked (blue box). An enlargement of the CNS shows the corresponding neuromeres A3-A5 (orange box). In A-C, approximately 100 msec are shown as the contraction wave passes through A5, A4, and A3 (blue boxes). The corresponding activity in the neuromeres are shown (orange boxes) with active motorneurons (orange cells) and the subsequent sensory input (blue arrows). Focusing on the contraction in A4 (B), we suggest a four part mechanism: 1) the motorneurons fire, 2) muscles contract, 3) the bd and class I mds respond to the change in tension, 4) sending a “successful contraction” message to the CNS which causes muscle relaxation in A4 and a propagation signal to A3.