Open preprint reviews by Leslie Vosshall

The wiring diagram of a glomerular olfactory system

Matthew E. Berck, Avinash Khandelwal, Lindsey Claus, Luis Hernandez-Nunez, Guangwei Si, Christopher J. Tabone, Feng Li, James W. Truman, Richard D. Fetter, Matthieu Louis, Aravinthan D. T. Samuel, Albert Cardona

I presented this paper to the Vosshall Lab Olfaction and Behavior Journal Club on 2/3/2016. Here is my synthesis of our discussion.

This is an important, foundational paper that provides for the first time a global view of all 'players' in the compact olfactory system of the Drosophila melanogaster larva. Because this animal has an early olfactory processing system that is superficially similar to that of vertebrates, lessons learned here will be broadly applicable to studies in other animals. This is a connectomics study that uses comprehensive transmission electron microscopy analysis to develop a parts list and a wiring diagram of the entire larval antennal lobe. The paper builds on earlier work that described the cell types and functional connectivity of the circuit based on light microscopy, but that missed many cell types. The strength here is that through the work of multiple labs, we have a complete inventory of incoming olfactory sensory neurons (OSNs), the odorant receptors they express, and the chemical ligands that activate these receptors and OSNs. The major surprises are the fascinating connectivity of multi-glomerular projection neurons, cells that have mostly been ignored by the field; and the beautiful wiring of local interneurons (LNs). It’s clear that much of the sophisticated processing of this system is achieved by LNs that reciprocally inhibit each other, and talk to incoming sensory afferents as well as projection neurons. It would not have been possible to access these cells without the kind of painstaking reconstructions afforded by connectomics. The work sets the field on a course to understand how this connectivity develops, and how the circuit computes odor information to make decisions according to the salience of the stimulus received.

We had the following comments and queries, in no particular order of priority:

1. It would be extremely helpful to have 3-dimensional models of the reconstructed antennal lobe so that the relationship of cell types and synapses can be seen more clearly. Obviously in a PDF only a 2-D flattened z-stack is possible. Even depth coding would help the reader grasp the relationships here. Perhaps you will host this on a server so that readers can “play” with adding and removing cells to see how it all fits together?

2. The non-expert will not understand how you mapped odorant receptor-specific OSNs onto the antennal lobe. Having Supplementary Figure 1, which was based on the Masuda-Nakagawa work of 2009 integrated earlier in the presentation would help the reader understand the logic of assembling the map. Also you misspelled her name in the text.

3. The discovery of the Keystone LN is important. This was our favorite neuron—it will be exciting to see in future functional work how it acts as a hub for routing information by inhibition. It would be helpful to have your thoughts about Keystone and possible homologues in other systems. How do you think it relates to the hub neuron that Cori Bargmann has described in C. elegans? Are there circuit motifs like this in mammalian CNS?

4. The distinction between “picky” and “choosy” LNs was lost on us. Maybe that can be explained more clearly, or names that are less interchangeable can be devised?

5. Neuromodulation is clearly important here, and we wondered why dopamine was not included in the analysis. Perhaps there are no dopamine neurons in the vicinity, or markers were not suitable for mapping onto EM stacks.

6. We wondered how feasible it would be to develop a theoretical neural circuit model based on the connectomics to predict the behavior of the larva given a known odor activating a known OSN.

7. Supplementary Figure 8 is interesting, but perhaps an orthogonal non sequitur in this connectomics paper.

show less

The Ionotropic Receptors IR21a and IR25a mediate cool sensing in Drosophila

Lina Ni, Mason Klein, Kathryn Svec, Gonzalo Budelli, Elaine C Chang, Richard Benton, Aravi DT Samuel, Paul A Garrity

Nice work!-very cool to expand the role of IRs in biology.

Here are my comments on the paper, ordered by appearance of line number

in the manuscript not by urgency/priority

L139 what is the evidence that IR21a is endogenously expressed in
DOCCs? Anything beyond the Gal4 line? Antibody/in situ? My former PhD student Kenta
Asahina was the king of larval in situs, but that was done before the IRs were

L166 did you do a rescue of Ir25a just in the Ir21a DOCCs? Ir25a is
*everywhere* which is a problem for claiming specificity

L192 I don’t understand how brv1 is showing such a clear phenotype, and
not necessary for DOCC responses. Must be some other cells elsewhere that
require this?

L195 I thought the conclusion here was overly strong given the strength
of the evidence offered

L226 This is a great experiment—very elegant

L216 wondered why you did pan-neuronal Ir21a expression rather than
just go with the much more selective HC>Ir21a. You could consider showing
the HC result in the main figure, and putting the pan-neuronal as data not
shown. Always makes me nervous to put a protein like that into *all neurons*

L246-253 could equally be a cell biological problem with trafficking
unrelated to any functionally relevant co-factor. I would not be so forthright
here (unless you have the answer in the form of the co-factor in hand already)

L257-259 I agree with this conclusion. I think Ir25a is receptor for
heat just as much as orco is a sensor for ethyl acetate. It’s the wrong way to
look at this. Of course a ‘co-receptor’ will have a selective phenotype, but
it’s wrong to conclude that it is the subunit responsible for the specific
sensory cue.

Figure 1 I don’t understand why you are doing huge temperature swings
of 14oc vs 20oc. You say these neurons are extremely sensitive to small changes
in temperature, why not image under those conditions. Also you have the chance
to analyze the kinetics of the response to extract party of the answer. Since
your temperature ramps are slow, you could calculate the onset of calcium
signal and the rise time, etc.

Figure 2b I found the cartoon very busy and confusing. All I cared
about was what the temperatures at the extremities were and that was not

Figure 2c what is navigational bias, not defined?

Figure 2c are the Ir21 mutants actually PREFERRING the cold?

Figure 2 general: it looks like you are not doing single animal
tracking here. Can you revise the data presentation to extract additional
information on speed, turning, path tortuosity, etc etc rather than just a
single index number.

Figure 3 very pretty! But could integrate into another figure because
it’s making a single elegant point, while the other figures are pretty crowded.

Figures 4-5 my lab had the hardest time with this experiment. We
understand what you are trying to do here, but it’s not easy to explain or
understand. Isn’t IR21a already expressed in these cells? What happens if you
overexpress Ir25a? or Ir8a? Or some other random protein? I worry that the
small bumps in Figure 5d are some nonspecific problem with the neuron. Can’t
exclude that with the current data? It looks like there is still some low
amplitude cycling in Ir21a mutants? (Figure 4d, f). Do you think that’s real?

show less