عالية الدقة قياس سلوك رائحة يحركها يرقات ذبابة الفاكهة في

Hi, my name is Matt Lui. I'm a postdoctoral fellow in the laboratory of Leslie Sal at the Rockefeller University. And today I'd like to tell you about the experiments we've been doing to study larval chemotaxis.

The problem we are interested in, it's fairly simple. It is actually represented here in this setup. I've just transferred a group of Joseph Larvae on this surface of agros.

This environment is fairly hostile and they're close to starving. Somehow these animals will need to find their way to the closest source of food and here are some banana emitting smell which will attract the larvae. This environment is not ideal to study chemotaxis behavior, so the first improvement will be to replace the banana by a chemical.

This chemical is iso acetate and it is one of the flavor which is predominant in banana. So this is our artificial flavor and by using this we have a much better control of the olfactory stimulus. The next improvement is to use a closed environment and try to prevent in such a way the fluctuation in the other distribution.

So here is the assay we've been using in the lab for quite a long time. This is a Petri dish with a surface of agarro at the bottom. Agarro is ideal for larvae to chemo attacks because it keeps them moisturized.

On the two sides of the plate, you find an other cap that is made of Teflon and what we going to do now is introduce or isam acetate on one side and no order on the other side. Once this is done, the idea is to close the petro dish, wait for a little time for theodor to diffuse, and then afterwards open the dish. Introduce an animal here, and we let the animal move for a certain period of time.

This assay rely on the assumption that by introducing theodor on one side of the dish, we have a gradient that is fairly quickly established and will remain stable for the period of the test. Now the main problem is that no one has ever demonstrated that the gradient exists in the plate. Could very well be that very quickly after the addition of theodor, you have homogeneous distribution of theodor in the plate in such a way that the gradient would be very shallow or it could be that several minutes are required for the gradient to be established.

This will of course depend mainly on the vapor pressure of theodor that is being used. The second issue is related to what is quantified during the experiment. Imagine that an animal starts from here, the X and after three minutes is found at this end plate, the red X.Now there are two possible approaches.

One thing you can do is only record the final position of the animal after a certain period of time, three or five minutes. Then the only information you would capture is the position of the start and, and we believe that it is important to have a good resolution about the behavior of the animal over time. For instance, an animal could have hardly moved from the start position to the end position such as here, or it could have very well visited the region surrounding theodor for a significant portion of time and return it before the end of the trial.

The conclusion we would draw from these two situations is very different. In one case, the animal may not have responded to the stimulus while in the other case it certainly has. The two improvements we try to achieve is first, control the shape and stability of the other gradient in the arena.

Second, track the behavior of the animal in real time and keep all this information. In the next sequence, Sylvia is going to present you how we prepare order dilution. My name is S ti.

I'm a research assistant in the VAs laboratory working for material lui. Today I'm going to show you how to prepare the materials for the chamber system that we use to track animals. The primary issue with our chamber system is that we're using odor in dispersed in air.

This means that for example, if the agros plates that we use are not flat, you might get fluctuations and concentration between between experiments, which will introduce noise into the system. So I will show you and give you small tips to try to reduce the noise in the system. So the materials for preparing your odor dilution are a glass jar, preferably darkened glass in case you're using light reactive odors with an aliquot of the odor that you're interested in.

Glass pipettes to prepare this aliquot and an adequate mechanical pipette aid. Then paraffin oil pipette tips, your standard pipet man and small vials for the dilution. Once again, these need to be glass and preferably darkened and then a weighing scale to make accurate measurements and reduce noise between experiments.

Since odors are highly volatile and some of them are toxic, they're done within a hood. The first step to repairing an odor dilution is to aliquot the odor that you're going to be diluting. To prepare accurate dilutions and minimize the variation between experiments, it is necessary to weigh everything.

So start off by labeling, gonna make a one molar concentration of ethyl butyrate. First, you prepare the paraffin oil. Paraffin oil is very viscous, so it's important to try to dip the pipette tip as little as possible into into the paraffin oil, just enough to be able to suck up the paraffin oil without generating any bubbles.

Then it's good to wipe off the extra paren oil on the side of the bottle. Once again, because paren oil is viscous, it is very important to let it out slowly into the vial. Once you have pipetted the para oil into your into your bottle, do not throw away the tip immediately because more will collect at the bottom of the tip.

It's good to prepare a sheet of calculations before you start diluting where you have already measured out the volume of paren oil that you're going to be adding and also the weight, and then do the same also for your odor with also the weight. That way you can check on the scale that you have, all the paraffin oil that you need and all the and all the odor as well. If necessary, you pour in more paren oil and remeasure until you have the right weight.

Once you have the correct paren oil weight, all you need to do is add odorant using the same principles. All right. Here I'm going to show how to prepare larvae for tracking.

We usually use larvae which are six days old. We do this because they're large enough that they're easily picked up by our software and they're also still very active to tell whether your larvae are too old for use in tracking, you just have to check whether there are larvae climbing up the walls. If there are larvae climbing up the wall, then they're too old and they will not be particularly active.

The first step in preparing your larvae is to immerse them in 15%sucrose. The reason we use 15%sucrose is that the larvae will be Boyd in this solution and the food instead will either partly dissolve in it or drop down to the bottom. You pour enough sucrose so that it's about, I know a little less than an inch from the top of your, of your bottle or vial.

Then you take a spatula. You can already see the larvae floating up and you gently dig into the food to dislodge any animals that are digging into the food, swirl it around a bit to separate them from the food and then just sit it down for a couple of minutes. Once the food and the larvae have separated out adequately, take a pre-labeled Petri dish and pour them into the beeper part.

Keep the top of the petri dish as you will fill this with water and use it as a receptacle for tracked larvae so that you can separate which ones you have tracked and which ones you have not. Now that you have poured the larvae into the petri dish, you have to wait 30 minutes for them to adapt to the sucrose and reach a state of starvation. This means that they will behave better in the tracking setup during the 30 minutes that the larva are adapting to.

The sucrose is a good moment to prepare the plates for tracking. So these are tops of 96 well plates. You can buy them separately and this is a 3%aros solution that has already been heated up.

We use 3%because it prevents larvae from digging, but it's also dense enough for them to be able to walk on it and remain humidified on each plate top. You want to pipette 25 mils of agros. It is important to try to avoid forming large bubbles in the in the gel.

It's also important to place the plates on a flat surface, otherwise you may aros gels that are on a slope and may affect the behavior of the larvae. After the aros has set, we stack the tops and wrap them up in soran wrap so as to prevent them drying out. This should keep them in a good state for tracking for several days.

In this sequence, I'm going to present you our behavioral assay, which differs slightly from the petrol dish assay, which I told you about earlier. What you will need are leads of 96 well plates as well as those containing the agro layer. The assays composed of three lids, which are piled up.

The first lip contains nothing is empty and is only used to isolate the system from the surface on which it lies. The second layer contains Diego agro surface. Here the animals will be placed on the surface and now as I will show you next, we introduce theodor in the plate.

It is important to use a fresh and new lid every time you load theodor. This is to prevent any contamination from one experiment to another one. You will also need other dilution.

We normally don't use a dye to load theodor, but here I will make use of it for the sake of illustration. This is high concentration of theodor. While this one is very diluted, my goal will be to create a gradient along a line.

Say this line and we want to have concentration increasing from the left to the right. To achieve this, I will use several droplets of order placed in the condensation ring with increasing concentration. I start with the highest concentration, which will be placed at the extremity.

It is important to disperse it in the ring. Now I proceed with the next concentration, which is diluted twice compared to the first one, and I place it to this ring. So there No, I have six droplets.

The first one contains a concentration of say one. The second one concentration of 0.5, twice as less as the first one. Third one is concentration, 0.25, and so on.

So we have a geometric series along the line. The next step consists in inverting this lid on the AAL surface and now by surface tension, all the droplets remain suspended from the top of the lid. We devise a technique based on infrared spectroscopy to measure all the concentration in the plate.

By applying this technique, we were able to show that if you have a geometrical series along deodorant line here, you end up with a concentration of theodor along the length that increases exponentially as shown here and along the width of the plate. We have a profile that peaked right under the odorant line. Here is how the actual geometry of the gradient looks like.

If you consider the concentration of theodor in the air, the good thing is that we were able to show that the gradient its geometry is stable for at least 15 minutes. Upon the introduction of the top lid, we wait for 30 seconds to allow the gradient to be established. Then we introduce the animal in the arena.

The animal is introduced under the other end line between position three and four, we use a small pin to transfer the larvae from the sucrose solution to the agro surface. The transfer is made by fishing a larva around the head of the nail. We quickly open the lid, the larva is deposited on the surface of agarro.

Then we close the lid. It is important to minimize the time that the lid remains open. Next, you record the animal motion for three or five minutes.

The actual recording is made. With this tracking setup, we have a camera that is connected to a computer. The image tracking software that we using is Aton vision from Nordis to collect the information about the animal motion and store it as a digital file.

Now we are about to start the behavioral experiment. We have the the lid with the aro layer. We have the dilutions.

Diego is stuck on the first lid. Before loading theodor, it is important to vortex each vial. We use our dilution with the smallest one on the left and the highest one here on the right.

All the six droplets will be loaded on the plate at the same time. To avoid odor loss, we are using here a multichannel Pipeter six dip, 10 microliter of the odor is taken from each dilution sample, and theodor is loaded at the same time. Upon loading, the lid is now inverted on the surface of agros.

The gradient has been created in the plate over the past 30 seconds as I showed you before. Now I'm going to introduce an animal on the Diego surface. As soon as the animal has been introduced, I start the recording.

The animal is here and its motion is now being tracked. As you can see, the animal is following the odorant line and is heading towards the highest concentration point of the gradient. Remember that the concentration increases from left to right with an exponential shape.

As you can see, the direction of motion is constantly corrected. With respect to the stimulus, the animal has now reached the highest concentration point of the gradient, and we remove it from the plate. The animal is discarded on another plate to make sure that we not tracking the same individual twice directly.

After we introduce a new larva and start a fresh recording, here are the rules we follow for the tracking. We stop the tracking as soon as the animal has reached the highest concentration row of the plate, which correspond to this, this one here. Also, if an animal hits the wall, the tracking is interrupted because we want to avoid any interference.

The first question we addressed was fairly could. What we wanted to know was whether or not an animal could smell theodor at a certain concentration. To make such a test, what we typically do is to have a single source of odor instead of having several.

This allows you to have one degree of freedom in changing the concentration in theodor end droplet. So here for instance, you have a high concentration droplet, and here you have a low concentration droplet. What we typically do is introduce the animal on Diego surface right under Theodor droplet and see whether or not the animal is able to detect its presence and stay in its neighborhood.

Now I'd like to make a more advanced use of our assay. Imagine you have one droplet of order and you record several trajectories which are super imposed here, and you have three of them. You have one individual which generate this green track.

You have this orange track that is less center, and then you have this pink trajectory which represent an animal apparently circle around the source. We believe there's much information to begin about the details about the trajectories. For this reason, we import the trajectory as series of possession into matlab, and we perform the analysis after the analysis with matlab.

Here is the distance to the source probably as a function of the time. For the three trajectories which were displayed previously, you have the green trajectory, which remains very close from the source. You have the orange trajectory, which displays significantly more noise.

And then you have this purple trajectory, which interestingly showed that the animal left the source to stay at a distance, which probably corresponded to a concentration Which was optimal. Here you see the illustration Of 10 trajectories, animals which are unable to smell. These are mutants or three B mutants, which are non ellers.

And as you can see, all these trajectory illustrate that animals failed to detect the source of odor and just left it. Now in star contrast, we recorded here 10 consecutive trajectories for white tap animals. These animals are of course able to smell, and as you can see, all of them accumulate right under deodorant source.

The 20 trajectories, which are just presented, were in fact acquired without knowing the identity of the animals. This demonstrate the efficiency and reliability of our assay. In determining whether or not some Mutants are able to smell here using an Exponential gradient, we generate several track and we want to have a sense of how well the animals are able to follow the other end stimulus.

So I'm going to show you now 15 trajectories recorded separately, and as you can see, the animals do excellent job in following the stimulus under the other end line. Now we'd like to proceed with a simple control. It may very well be that the animals in the exponential gradient followed the line because there was some stimulus irrespectively with its concentration, just the animals went straight.

So in order to show that they are in fact computing the concentration or something related to the concentration and making use of it to direct their motion, we have now this geometry, which is close to a rooftop. The larva is introduced here as usual, and what we expecting is that they will move along the line and then detect that the concentration is dropping after this point and return to it. So we performed the test with five animals.

We collected the trajectories, and now we are representing them superimposed on this slide. And as you can see, the animals started from here, moved along the line, then explore the region corresponding to the highest concentration and mainly stay there. This strongly suggests that they are able to compute other concentration in real time using the single source assay.

We were able to ask mainly one big question. Can an animal smell detect an odor at a certain concentration by using a multi-source device? We are, we have more flexibility in changing the shape of the gradient and determining whether or not an animal is able to make use of this information.

I'd like to illustrate this idea by performing a simple experiment. In the gradients we set up so far, we had an exponential shape along the lens. Now the question is, if we challenge the al more by having the the concentration increasing linearly along the lens, are they still able to find their way under the other end line?

This graph represent the two types of gradient. We tested the exponential one in purple and the linear gradient in green. This is the linear gradient, and I'm going to write the different concentrations of isam acetate diluting paren oil.

This is an arithmetic series. We have a linear gradient that will get established and as you can see, the slope is fairly shallow compared to the exponential gradient here where the concentration of the extremities also 0.5, but goes up to 0.03. The question is, are navi able to chemo attacks when they are faced with such a shallow gradient?

Here, the actual concentration of acetate used in the six droplets for the exponential gradient and for the linear gradient. As you can see, even though the highest concentration are the same in both case for the linear gradient, the lowest concentration here is fairly high. We perform the experiment with the linear gradient, and 15 trajectories are represented here of wild type animal, and we believe it's fairly clear that these animals were able to detect the linear gradient and follow this signal along deodorant line.

So I think our point is clear. Before finishing up here, I'd like to speak briefly about future directions for this tracking setup. Although this tracking setup allows you to do a lot of interesting things, for example, seeing complex behavior in larvae over a larger, a longer period of time, it does have some limitations.

For example, you are limited to the size of your plate, so that limits how long you can track for and the complex behaviors that you can look at. So it'd be nice if we could develop a system where either the plate is larger or you're not limited by the boundaries of a plate at all. Another problem is that with the software that we have right now, it is difficult to track more than one animal at the same time, so it is not possible to observe population behaviors in real time.

Another interesting aspect of the behavior that we cannot test through this setup is the actual three-dimensional movements of the larvae in space. As you track animals, you will see that, for example, when they reach an odor droplet, they might lift their head or they might sweep their head around trying to detect changes in concentration. Possibly this isn't possible with our particular setup right now, so it would be nice to develop something that would allow us to do this.