Optical mapping of ground reaction force dynamics in freely behaving Drosophila melanogaster larvae

Reviewer #1 (Public Review):

This work introduces a new method of imaging the reaction forces generated by small crawling organisms and applies this method to understanding locomotion of Drosophila larva, an important model organism. The force and displacement data generated by this method are a qualitative improvement on what was previously available for studying the larva, improving simultaneously the spatial, temporal, and force resolution, in many cases by an order of magnitude. The resulting images and movies are quite impressive.

As it shows the novel application of recent technological innovations, the work would benefit from more detail in the explanation of the new technologies, of the rationales underlying the choice of technology and certain idiosyncratic experimental details, and of the limitations of the various techniques. In the methods, the authors need to be sure to provide sufficient detail that the work can be understood and replicated. The description of the results and the theory of motion developed here focus only on forces generated when the larva pushes against the substrate and ignores the equally strong adhesive forces pulling the larva onto the substrate.

The substrate applies upward, downward, and horizontal forces on the larva, but only upward and downward forces are measured, and only upward forces are considered in the discussions of "Ground Reactive Forces." An apparent weakness of the WARP technique for the study of locomotion is that it only measures forces perpendicular to the substrate surface ("vertical forces" in Meek et al.), while locomotion requires the generation of forces parallel to the substrate ("horizontal forces"). It should be clarified that only vertical forces are studied and that no direct information is provided about the forces that actually move the larva forward (or about the forces which impede this motion and are also generated by the substrate). Along with this clarification, it would be helpful to include a discussion of other techniques, especially micropillar arrays and traction force microscopy, that directly measure horizontal forces and of why these techniques are inappropriate for the motions studied here.

The larvae studied are about 1 mm long and 0.1 mm in cross-section. Their volumes are therefore on order 0.01 microliter, their masses about 0.01 mg, and their weights in the range of 0.1 micronewton. This contrasts with the force reported for a single protpodium of 1 - 7 micronewtons. This is not to say that the force measurements are incorrect. Larvae crawl easily on an inverted surface, showing gravitational forces are smaller than other forces binding the larva to the substrate. The forces measured in this work are also of the same magnitude as the horizontal forces reported by Khare et al. (ref 32) using micropillar arrays.

I suspect that the forces adhering the larva to the substrate are due to the surface tension of a water layer. This would be consistent with the ring of upward stress around the perimeter of the larva visible in S4D, E and in video SV3. The authors remark that upward deflection of the substrate may be due to the Poisson's ratio of the elastomer, but the calibration figure S5 shows that these upward deflections and forces are much smaller than the applied downward force. In any case, there must be a downward force on the larva to balance the measured upward forces and this force must be due to interaction with the substrate. It should be verified that the sum of downward minus upward forces on the gel equals the larva's weight (given the weight is neglible compared to the forces involved, this implies that the upward and downward forces should sum to 0)

Much of the discussion and the model imply that the sites where the larva exerts downward force on the gel are the sites where horizontal propulsion is generated. This assumption should be justified. Can the authors rule out that the larva 'pulls' itself forward using surface tension instead of 'pushing' itself forward using protopodia?

More detail should be provided about the methods, their limitations, and the rationale behind certain experimental choices.

Three techniques are introduced here to study how a crawling larva interacts with the substrate: standard brightfield microscopy of a larva crawling in an agarose capillary, ERISM imaging of an immobilized larva, and WARP imaging of a crawling larva. The authors should make clear why each technique was chosen for a particular study - e.g. could the measurements using brightfield microscopy also be accomplished using WARP? They should also clarify how these techniques relate to and possibly improve on existing techniques for measuring forces organisms exert on a substrate, particularly micropillar arrays and Traction Force Microscopy.

As written, "(ERISM) (19) and a variant, Wavelength Alternating Resonance Pressure microscopy (WARP) (20) enable optical mapping of GRFs in the nanonewton range with micrometre and millisecond precision..." (lines 53-55) may generate confusion. ERISM as described in this work has a much lower temporal resolution (requires the animal to be still for 5 seconds - lines 474-5); In this work, WARP does not appear to have nanonewton precision (judging by noise on calibration figures) and it is not clear that it has millisecond precision (the camera used and its frame rate should be specified in the methods).

It would be helpful to have a discussion of the limits of the techniques presented and tradeoffs that might be involved in overcoming them. For instance, what is the field of view of the WARP microscope, and could it be increased by choosing a lower power objective? What would be required to allow WARP microscopy to measure horizontal forces? Can a crawling larva be imaged over many strides by recentering it in the field of view, or are there only particular regions of the elastomer where a measurement may be made?

Reviewer #2 (Public Review):

With a much higher spatiotemporal resolution of ground dynamics than any previous study, the authors uncover new "rules" of locomotory motor sequences during peristalsis and turning behaviors. These new motor sequences will interest the broad neuroscience community that is interested in the mechanisms of locomotion in this highly tractable model. The authors uncover new and intricate patterns of denticle movements and planting that seem to solve the problem of net motion under conditions of force-balance. Simply put, the denticulated "feet" or tail of the Drosophila larva are able to form transient and dynamic anchors that allow other movements to occur.

The biology and dynamics are well-described. The physics is elementary and becomes distracting when occasionally overblown. For example, one doesn't need to invoke Newton's third law, per se, to understand why anchors are needed so that peristalsis can generate forward displacements. This is intuitively obvious. Another distracting allusion to "physics" is correlating deformation areas with displaced volume, finding that "volume is a consequence of mass in a 2nd order polynomial relationship". I have no idea what this "physics" means or what relevance this relationship has to the biology of locomotion.

The ERISM and WARP methods are state-of-the-art, but aside from generally estimating force magnitudes, the detailed force maps are not used. The most important new information is the highly accurate and detailed maps of displacement itself, not their estimates of applied force using finite element calculations. In fact, comparing displacements to stress maps, they are pretty similar (e.g., Fig 4), suggesting that all experiments are performed in a largely linear regime. It should also be noted that the stress maps are assumed to be normal stresses (perpendicular to the plane), not the horizontal stresses that are the ones that actually balance forces in the plane of animal locomotion.

But none of this matters. The real achievements are the new locomotory dynamics uncovered with these amazing displacement measurements. I'm only asking the authors to be precise and down-to-earth about the nature of their measurements.

It would be good to highlight the strength of the paper -- the discovery of new locomotion dynamics with high-resolution microscopy -- by describing it in simple qualitative language. One key discovery is the broad but shallow anchoring of the posterior body when the anterior body undertakes a "head sweep". Another discovery is the tripod indentation at the tail at the beginning of peristalsis cycles.

As far as I know, these anchoring behaviors are new. It is intuitively obvious that anchoring has to occur, but this paper describes the detailed dynamics of anchoring for the first time. Anchoring behavior now has to be included in the motor sequence for Drosophila larva locomotion in any comprehensive biomechanical or neural model.