A Neuron Pair Controls a Fruit Fly's Reaction to Black Coffee

I had written this article to communicate the science behind this study a year ago. A friend was going to publish it, but it hasn't happened yet. So I figured I might as well put this out here.

At a very basic level, our brains compute our reactions to sounds, smells, tastes and problems in communication/mathematics and decision making of all kinds at the very highest levels. What makes this computer work? What are the circuits in our brain that make any of this even possible?

Some sensations make our bodies react immediately. Most of us cringe at the sound of nails on a chalkboard, pull away our fingers from hot pans and jump if we step on a sharp pin. Smoke makes us cough. We’ve all squinted, wrinkled our faces and puckered at the first taste of lemon juice as babies. We don’t think about our reactions. They just happen.

But, if you do think about it, these reflexes aren’t that simple. Messages carrying information about a screeching sound, hot pan and sour juice need to be generated and conveyed through nerves to somewhere. That information then gets processed and a command reaches different parts of our body to move in very particular ways. Much like a computer, there is input, processing and output of information.

Nerves work a little like cables in a computer. Tiny wires inside a cable fan out when the covering is peeled. Much like that, animal nerves are bundles of the long wiry arms of neurons (nerve cells). These arms are called axons. The part of the neuron that holds the nucleus is called the cell body. The cell body has smaller arms that are a fraction of the size and length of axons. These smaller arms are called dendrites. The cell body is sometimes very far away from the other end of the axon. Obviously, neurons have to communicate: exchange information to make input-output type decisions. This kind of information exchange happens at the ends of axons and dendrites. The ends of some neurons almost look like they are shaking hands with other neurons. This is the site of information exchange between the neurons. These handshakes are called synapses. One neuron can either “shake hands”, more technically “Synapse”, with one or more other neurons. It is strange to think about, but sometimes neurons “shake hands” with themselves[1]. 

Nerves and neurons have very specific functions. One group of nerves takes information from the skin, nose, tongue, eyes and ears to the animals processing unit: the brain and spinal cord. These are called sensory neurons. The other takes processed information to whichever organ it needs to go to. The neurons that tell say, muscles what to do e.g. move limbs, are called motor neurons. There are other nerve cells that actually process incoming information from the senses. Then directly or through other nerve cells, the output is passed to motor neurons. These nerve cells are called interneurons. When many neurons (which could be of different types) interconnect, the network is called a neural circuit.

So, we just discussed how different neurons have different functions, shape and positions. One has to wonder. How does one neuron get its function? What makes an interneuron in the brain be different from another interneuron from the brain or from a sensory neuron from the right little toe or a motor neuron from the same toe? When a neuron, or any cell, takes on a specific identity and function, it is said to be specified.

As we grow in the womb, our cells sense signals from neighboring cells. These signals can be chemical in nature or even physical (like stretching). These signals trigger changes inside foetal cells. Different stretches of DNA get activated in response to different signals in different cells. This activation of different stretches of DNA is what makes tissues, like skin vs brain, different as we become fully formed babies. This is a crucial step in the process of specification of cells. It is also how an interneuron in the brain be different from another interneuron from the brain or from a sensory neuron from the right little toe or a motor neuron from the same toe.

It is important to understand what we mean by DNA activation. A stretch of DNA is identified by its sequence of deoxynucleotides. Depending on signals from outside a cell, certain sequences of DNA are needed to influence the transcription of specific genes. This means, specific sequences of DNA in response to external signals or internal cell events, activate genes that encode proteins. The specific DNA sequences are called promoters and enhancers. Enhancers are not necessarily close to the gene on the chromosome. Yet, they play an important role in gene activation. Different tissues have different active enhancers. Enhancers can be active in cells at different times. In our context, different kinds of neurons have different active enhancers.

I hope by now it is clear that a “pin-prick = move away” type movement, is brought about by a non-trivial process. Throughout development, different DNA sequences are active in different cell types. A whole collection of neurons kicks into to action to get an output for one input. Imagine the number of combinations of sensory inputs we receive and how many neural circuits work together for even the simplest output. We’re not even talking about solving Rubik’s cubes while simultaneously juggling them (look for youtube videos)! No wonder we have billions of neurons!

Therefore to understand the neural circuit that controls say, lifting one arm, we need a model system with few neurons but repeatable and reliable neural control. As an analogy: without blueprints and circuit diagrams, how does one understand the workings of a supercomputer? It might be easier to start with a very basic old computer, figure out the basics by changing cables, adding and removing parts and then, think about how a supercomputer works.

We need a living being, where there is a reproducible and repeatable physical movement. We should also be able to activate and deactivate specific neurons to check if there is a change in the specific physical movement. Practically, we’d like to have these results as quickly as possible. After a century of research, few animals are as suited to this task as the fruit fly (Drosophila melanogaster). We know that Fruitflies extend their mouthparts (proboscis) in response to sweet substances like sugar water. We know they do not extend the proboscis in response to bitter substances[2]. Fruitfly researchers have developed specific tools that allow control of protein function in specific tissues at any time during the life of the fly. In this way, depending on the protein whose function is controlled, the cell itself can be switched on or off[3].

Ali Asgar Bohra from the lab of Prof K VijayRaghavan at NCBS wanted to find out exactly what neural circuit in Fruitflies controls proboscis extension, or the lack of extension, in response to bitter substances. The approach he used can identify single interneurons in this circuit. This method with variations, might be able to identify every single neuron in the bitter-tasting circuit in fruitflies.

We know that sensory neurons at tips of Drosophila legs and other places on the fly body can taste structures around them. It is weird, but yes, Drosophila can taste with their legs. Starved Drosophila reflexively extend their proboscis in response to sweet substances and do not extend their proboscis in response to bitter substances.

Other researchers trying to identify neural circuits have found that sensory neurons in the proboscis that respond to taste, extend their axons to specific brain region. It has also been shown that the set of neurons in this brain region that respond to sweet taste is different from the set of neurons that respond to bitter taste. Now, Ali decided to use enhancers that distinguish neurons within the bitter responding set, to identify individual neurons of this set. Just to remind you, different enhancers are active in different sets of cells. Using a set of enhancers, Ali activated and deactivated specific neurons in the fly brain. He found that by deactivating one specific cell group, fruitflies extend their probosces even if they taste a bitter substance. Therefore under normal circumstances, if the fly tastes something bitter, these neurons are responsible for keeping the proboscis unextended. Using genetic tricks Ali then showed that from this group of cells, a single interneuron is sufficient to control proboscis extension in response to bitter taste. Remarkably, he was able to record video of this neuron in action[4].

I think this is a good time to pause and reflect on this finding. One neuron, among millions of cells in a fruitfly, controls the movement of organs that crucial to fruitfly survival. Also, from a design perspective, one neuron is a critical link, a bottleneck step, in the “bitter taste input”-“don’t extend proboscis” output process. How many such bottleneck steps exist and how would they have been selected for are questions other neuroscientists can ponder.

Here I have not described the number of control experiments he did to be certain of these conclusions. You’ll have to read the paper for that and let us know if you have questions. Using the very same approach he is looking for other neurons in the bitter sensing neural circuit. He hopes to be able to cleanly identify as many neurons in this circuit as possible.

The hope is, if we understand basic reactions to everyday stimuli, we will be able to understand more complicated behaviors. If the principles on which circuits operate are found, they might someday explain human neural disorders. Physicists and engineers may probably draw up theoretical plans and models to understand these responses or design their own, but it crucial to know what actually exists inside living animals right now. Not only do we find out what really works, it may inspire more designs of information processing for decision making. As yet, a millimeter size self reproducing robot that senses its environment to feed, to adapt, to escape predators and pathogens and proliferate is far from reality.

1.     Deleuze C, Pazienti A, Bacci A. ScienceDirect Autaptic self-inhibition of cortical GABAergic neurons: Synaptic narcissism or useful introspection? Current Opinion in Neurobiology. Elsevier Ltd; 2014;26: 64–71. doi:10.1016/j.conb.2013.12.009

2.     Gordon MD, Scott K. Motor Control in a Drosophila Taste Circuit. Neuron. Elsevier Ltd; 2009;61: 373–384. doi:10.1016/j.neuron.2008.12.033

3.     Refined LexA transactivators and their use in combination with the Drosophila Gal4 system. 2010;: 1–6. doi:10.1073/pnas.1005957107/-/DCSupplemental/pnas.201005957SI.pdf

4.     Ali Asgar Bohra et al. Identification of a Single Pair of Interneurons for Bitter Taste Processing in the Drosophila Brain. Current Biology. Elsevier Ltd; 2018;28: 847–858.e3. doi:10.1016/j.cub.2018.01.084