Evolution of sensory signaling


How we sense the world has, ultimately, a cellular and molecular basis. We have these big brains that do amazingly sophisticated processing to interpret the flood of sensory information pouring in through our eyes, our skin, our ears, our noses…but when it gets right down to it, the proximate cause is the arrival of some chemical or mechanical or energetic stimulus at a cell, which then transforms the impact of the external world into ionic and electrical and chemical changes. This is a process called sensory signaling, or sensory signal transduction.

While we have multiple sensory modalities, with thousands of different specificities, many of them have a common core. We detect both light and odor (and our cells also sense neurotransmitters) with similar proteins: they use a family of G-protein-linked receptors. What that means is that the sensory stimulus is received by a receptor molecule specific for that stimulus, which then actives a G-protein on the intracellular side of the cell membrane, which in turn activates an effector enzyme that modifies the concentration of second messenger molecules in the cell. Receptors vary—you have a different receptor for each molecule you can smell. The effector enzymes vary—it can be adenylate cyclase, which changes the levels of cyclic AMP, or it can be phospholipase C, which generates other signalling molecules, DAG and IP3. The G-protein that links receptor and effector is the common element that unites a whole battery of senses. The evolutionary roots of our ability to see light and taste sugar are all tied together.

There’s another class of senses that seem to function in a different way, and are distinct from the G-protein mediated senses like sight and smell. The G-protein senses are characterized by specific receptors tuned to recognize discrete elements—photons, chemicals, transmitters—in the environment. Sensing thirst, touch, vibration, texture, pressure, though, are different. There the stimulus is physical distortion, not a specific chemical agent. These senses don’t use G-proteins, and are less well understood…but it’s beginning to look like there are commonalities here, too, and we can trace our ability to hear and touch to a bacterium’s ability to react to changes in salt concentrations.

What Kung proposes is that there are two very broad classes of primitive sensory signalling: one that detects solutes, molecules dissolved in the environment, which has diversified over evolutionary history to handle vision, smell, and taste; and another class that detects solvents, which has evolved to be used in our senses of hearing and equilibrium and touch. One way to think of it is that a bacterium’s ability to sense when it is raining is the precursor to our ability to listen to music.

Here, for example, is a rod-shaped E. coli bacterium in an environment with some concentration of salts dissolved in it; it’s in equilibrium. If it rains, though, turning our salty red world into a more dilute pink one, water flows into the bacterium, causing it to swell distressingly. The dangerous turgor pressure is detected by sensors that respond to the distortion of the membrane, opening pores large enough for internal solutes to flow out, restoring equilibrium and allowing the bacterium to relax back into its rod shape.

An E. coli cell in a normal environment (left) and in the rain (or upon dilution in the laboratory, right). A bacterium (shown as a rod), having adjusted its cytoplasm to the relatively high osmolarity of the surrounding milieu (shown in dark red, the red dots being solutes, not water), is confronted with a sudden dilution of its environment upon the onset of rain (light red). Entry of water (not shown) through the lipid bilayer swells the bacterium (now oval-shaped) and stretches open the MS channels to jettison solutes (red puffs), enabling it to reach a new equilibrium and escaping osmolysis (and returns to being rod-shaped).

So how do proteins detect the swelling of the cell? The answer is surprisingly direct: they have channels that respond to the tension in the cell membrane. Here, for example, is the MS (mechanosensitive) channel in E. coli. It contains a ring of helically organized rods that allow the channel to dilate open like the iris of a camera as the lipids in the membrane around it push and pull on its structure.

Helical segments (S1, S2, S3) and transmembrane helices (M1, M2) in one MscL subunit, as deduced from sequence and other analyses (left). Side (upper centre) and top (lower centre) views of the closed channel backbone structure of the E. coli MscL protein, by analogy to the crystal structure of the M. tuberculosis MscL homologue. The open structure deduced from both modelling and experimentation (right). Unlike MthK, the prokaryotic K+ channel that is equipped with a second constriction (the K+ filter), MscL is like the acetylcholine receptor/channel, in which the open gate doubles as the filter. Here the opening is huge (30 Å in diameter): befitting its ability to release solutes indiscriminately. The work to increase the area under tension constitutes the free energy difference that partitions the open and closed states.

These pores don’t require any accessory proteins to do their job: they can be inserted into artificial lipid membranes, and they still function, responding to to distortion of the membrane by changing their permeability, which can be measured as a flow of current. Another interesting property is that they are sensitive to the lipid content of their surrounding membrane, which changes the forces exerted on them. Many anesthetics are readily dissolved in membrane lipids, and this may explain their mode of action—the E. coli MS channels can respond to exposure to procaine and tetracaine. That shot of painkiller you get at the dentist may work by making your nerve cell membranes more fluid and slippery, changing the way they can exert force on pain receptor channels.

We animals have other kinds of mechanoreceptors. Hair cells, indispensable for hearing and balance, rely on elaborate cilia coupled to TRP (transient receptor potential) channels in the membrane with tethers—tug on a hair, and it pulls or pushes on a structure imbedded in the membrane. The principle is the same, though, with tension between a protein and the lipids around it inducing a change in channel properties.

TRP channels have been located in complex auditory sensory cells, even though the mechanism by which ciliary vibrations (arrow pairs) lead to the iris-like opening of the channels on the side of the cilia is not clear. a, The antennal chordotonal organ of Drosophila. CM, cap-cell matrix; DC, dendritic cap; CD, ciliary dilation. Red marks the location of NAN (a TRPV-type channel subunit encoded by the Nanchung gene). b, A vertebrate hair cell. St, stereocilia; K, kinocilium; PZ, pericuticular zone. Red marks the location of TRPA1. c, Models of the vertebrate hair-cell transduction channel. Molecular identifications have transformed the biophysical trapdoor model (left) to one with a TRPA channel and a stiff cadherin-containing tip link (right). The elastic element of transduction is now assigned to the ankyrin repeats in the four (presumably) TRPA subunits (shown as coils), which are presumed to be attached to cytoskeleton and/or myosin (not shown). This current model is compatible with one in which the displacement of the channel protein, with respect to the lipid bilayer, ultimately triggers the channel conformation change, right. However, none of these models should be taken literally since we do not yet know the true composition of the transduction channel(s) and how the various channel components contact each other and the lipids.

I’m feeling the keys under my fingers and seeing the screen in front of me and hearing the music on my headphones using a suite of tools derived from some primeval microbe’s acquisition of sensors for dissolved nutrients and osmotic pressure. We’ve elaborated and refined and added new layers of complexity, but deep down we can still see echoes of our microscopic ancestors.

a, A diagram of an imaginary early cell equipped with two types of receptors that are required to sense solutes and solvents—the two ingredients of life’s chemistry. The dots in the grey background represent water molecules (the solvent) and the red circles represent solutes (molecules dissolved in water). When a cell accumulates solutes, the internal water concentration is reduced and the tendency of water to enter the cell results in a turgor. Both the lock-and-key type of receptors (red) for different solutes (ligands), as well as the turgor sensors (blue) for water (the solvent), are needed for even an early cell to survive. b, A hypothetical diagram (not to be mistaken for phylogenetic trees) on the grouping of various senses that emphasizes the discrete separations of the lock-and-key type of sensing of the solutes (red) from the force-from-bilayer type of sensing of the solvent (blue).

Some people seem to think the linkage to our history demeaning, and take offense at being found similar to an ape. Personally, I find it uplifting and wonderful to see our unity with bacteria, fungi, worms, jellyfish, herps, and fish. What I love about biology is the way it binds us closer and closer to the world around us, and shows us over and over again that we’re part of this place.

Kung C (2005) A possible unifying principle for mechanosensation. Nature 436:647-654.


  1. Will says

    You write:

    The evolutionary roots of our ability to see light and taste sugar are all tied together.

    There’s another class of senses that seem to function in a different way, and are distinct from the G-protein mediated senses like sight and smell. The G-protein senses are characterized by specific receptors tuned to recognize discrete elements–photons, chemicals, transmitters–in the environment. Sensing thirst, touch, vibration, texture, pressure, though, are different.

    Photons are light, so I’m not sure were sight falls into this duality? Is it a G-protein mediated sense or not? In your final paragraph you again refer to “seeing the screen” as if it were part of the non-G-protein mechanism.



  2. Will says

    Never mind. I see that you were linking the two to a comon mechanism in the last paragraph. Sorry. Will

  3. TheBlackCat says

    Very good post. I do have one minor quibble. You cannot really lump taste together like that. There are taste receptors that are G protein-coupled receptors (GPCRs), for instance sweet (carbohydrates) and some bitter (poison) receptors. However, others like salt (ions) and sour (acids) could be considered more like the “solvent” channels. They are not GPCRs, they are simply regular permanately open ion channels that let specific ions flow through, causing a change in membrane potential and thus activity in the receptors. They may decect solutes but physically they are more akin to what you call solvent receptors.

    What makes this deal with using GPCRs in both taste, smell, and vision odd is that taste receptors are not neurons. They are epithelial cells. They connect to neurons and have several of the features of neurons but are not actually neurons themselves. Epithelial cells are similarly involved in some touch receptors which, as you pointed out, do not use GPCRs at all (touch receptors are just as variable as taste receptors in terms of what they detect and how they detect it, but variable in different ways).

    And the similarities between senses are not all at the molecular level. There is something in senses called “lateral inhibitition”. If you look at say a digital camera photograph, what it is looking at is the raw pattern of energy falling on it. That is ultimately what it records (perhaps with some compression later on). Our senses do not operate like that. Instead, each receptor (meaning a cell in this case) tries to inhibit all the nearby receptors, with receptors farther away being less inhibited than those closer. The strength of the inhibitory signal sent out is proportional to the strength of the signal recieved by that particular receptor. This means all the receptors are competing with each other, and the one that is receiving the strongest input shuts down all the receptors surrounding it. If all the receptors are recieving the same level of stimulus, they all shut down. This serves to sharpen the shape detection in the image by enhancing borders and attenuating uniform areas. This is easy enough to think about in vision, where everything but borders is basically deleted. But it also occurs in hearing relative to frequency, touch relative to location on the skin, and even smell regarding particular similar chemical features. It is not known to happen in taste as far as I am aware, but that does not mean it is not happening only that we may not have found it yet. It is one of the things that is uniform across practically all of our senses. And there is no reason for it to be so. Smell evolved along completely different lines than all the other senses. It follows compltely different pathways to the brain and does completely different things once there.

    This is related to another feature common to all senses (except perhaps taste): maps. All senses are mapped relative to one or more features. For instance in vision, the visual cortex in the brain is actually laid out as a distorted veresion of the area you are looking at. Areas that are close to each other in the visual scene are close to each other in the cortex, and spatial relationships are also largely preserved (with some distortion). This organization is not only at the first level of the cortex, but is preserved for several additional levels. It is similar with touch, where there is a distorted map of the body in the early stages of sensory processing in the cortex. A similar thing occurs in terms of frequency of a sound in at least the earlier stages of processing of hearing in the brainstem, and there may even be spatial maps regarding the location of sounds in space (they are pretty sure this is the case with barn owls, it is not as clear with humans). Smell has a similar map except it is mapped relative to certain chemical features. The exact organization of the map is not clear but it is clearly uniform between individual within the same species, at least for rats.

    Another thing that is uniform across all senses is a power-law relationship. You do not detect the raw intensity of a stimulus. You detect the logarithm of the intensity of the stimulus. You know how we use decibels for sound. This is 10 times the log of the signal power divided by some arbitrary baseline. Using a logarithm scale for sound is not arbitrary, it reflects how we really hear. But that also occurs for all other senses as well, and logarithm scales are used for light intensity as well in sensory research (I am not as familiar with the conventions for measuring other senses).

    The really cool thing about the GPCRs used in vision is really how arbitrary they are. They are really nothing more than standard GPCR neurotransmiter receptors (usually called metabotropic receptors) that have been gerry-rigged to detect the presence of retinal (the molecule that responds to light) instead of more conventional neurotransmitters. When retinal is present, it is inactive. When retinal is forced out by interacting with light, it behaves just like any other GPCR. The other cool thing is that they work in reverse. Light reduces the activity of light receptors in vertebrates, while darkness increases the activity. Your eyes are working the hardest when you are asleep and in bright daylight become less active. The opposite is true in the octopus and squid, so there is no reason that has to be the case.

    The interesting thing about the hair cells in hearing is that they are not only used in hearing. Basically identical cells are used for sensing motion and head position in space. Instead of detecting motion of a membrane due to pressure waves they detect the flow of fluid in a tube (for angular acceleration) or what amounts to the direction a heavy weight is hanging for lateral acceleration and head orientation (which are equivalent as far as our senses can determine).

  4. Torbjörn Larsson says

    I’m not sure what you are trying to say on lateral inhibition and logarithmic sensitivity. Isn’t the former a primary reason we have a logarithmic sensitivity across senses? The later should enable us to distinguish informatory details in all sensory maps, whether they are spatial or nonspatial, amplitude or frequency resolved. The complexity of a near-neighbour confined relationship should be a simple method to give a power-law behaviour, which match what you claim.

    (I believe there are other reasons. IIRC at least rods have quite a photon-counting sensitivity at lowest illumination due to intracellular chemical cascades, and the later account for 2-3 orders of magnitude desensitivity. But I could be wrong. It is very old memories now, from when I looked at simple descriptions of eye constructions to compare with electronic microsensor capabilities.)

  5. TheBlackCat says

    All I am saying is that there are other interesting evolutionary parallels between the senses besides just the molecular-level signalling.

  6. miko says


    I was unaware of lateral inhibition in olfactory sensory neurons… of all the senses it is the most discrete, and sensory integration occurs partly at the bulb level but mainly in the cortex (I know OR mapping is maintained all the way from the epithelium to the cortex, though it is a “collapsed” map by the end). OSNs are not generally organized in the epithelium according to what class of odorants they bind (they are roughly organized by genomic cluster, I think, but for the most part scattered at random), so how could lateral inhibition among sensory neurons make sense?

    Also, I’m not sure that maps reflect shared underlying functional properties of different sensory systems… to use smell again, there is probably not much of a sensory basis for the spatial organization of odor maps, though it seems to sort things like basic vs. acidic amino acids. Retaining close mapping two chemically similar odorants doesn’t really make sense the way it does for continuous variable modalities like vision and hearing. Two chemicals can be very similar but “smell” completely different because they bind to different ORs or subsets of ORs. This is important because small differences in chemical structure can mean the difference between food and poison.

    It seems more likely that mapping is common among the senses not because it is needed for a shared functional property, but because it is a robust developmental mechanism for organizing axonal inputs and outputs across multiple tissues (it’s also been argued that it’s the most efficient in wiring costs). Topographic mapping based on gradients both reduces the number of distinct cues needed and allows target tissues to be very plastic in receiving more or less innervation and organizing it logically, i.e. so spatial relationships are maintained across multiple brain areas (thus contributing to the evolvability of functional connections in the brain). Maps are common in non-sensory brain areas as well.

    I’m not saying maps don’t have functional consequences for sensation, but they are more there as a developmental constraint than being selected for their functional characteristics. Maybe. Just throwing that our there.

  7. Torbjörn Larsson says

    Okay. And thank you, it was interesting facts.

    Wow. I remember that the conclusion was that electronic sensing, and the microelectronic construction methods we were developing then, had a long way to go if we should think of catching up on biology. Now it seems even longer. ;-)

  8. TheBlackCat says


    It may not make sense, but neither does plenty of things in the nervous system. The lateral inhibition occurs at the level of the olfactory bulb. There does not appear to be much organization at the sensory epithelium level, but the olfactory bulb is highly and very specifically organized in terms of what pattern of activity is triggered by a particular chemical. That is where the lateral inhibition occurs, with regions of the olfactory bulb inhibiting nearby regions. Paticular chemicals cause very specific patterns of activity in the olfactory bulb that are symmetrical between the bulbs. It was known that something akin to lateral inhibition must occur since certain chemicals will mask the sent of other chemicals. It was found that these chemical trigger activity in nearby regions of the bulb. It was also known that mitral cells in the olfactory bulb form inhibitory connections with adjacent mitral cells by means of granule cells after leaving the glomeruli. Also, more related chemicals cause more similar patterns of activity in the bulb, giving a good reason for lateral inhibition because it will sharpen discrimination between similar chemicals.

    You have to remember, ORNs do not bind chemicals explicitly. They bind particular functional groups and shapes within chemicals. The information on the functional groups and shapes present is then used to assemble a chemical signiture in the olfactory bulb. That is how we are able to smell chemicals that do not occur in nature. With that in mind lateral inhibition makes more sense. You have to be able to discriminate between chemicals with a very similar set of functional groups and shapes, and thus lateral inhibition whose whole purpose is to improve discrimination between similar objects is very useful.

    And I never said mapping was functionally essential. All I said was that it exists. It may be useful in some cases, but it is by no means essential in any of them as far as I know. I just said it was something most senses have in common for whatever reason. There is no reason it has to be that way, but the fact that it is that way is definitely a similarity between senses. It can’t be essential, at least the way we have it, since the degree or form of mapping varies between senses (rats lack humans’ “pinwheel” mapping in the visual cortex for instance, if I remember correctly).

  9. bern says

    Raivo Pommer


    Die Schweiz, Luxemburg und Österreich suchen einen Weg, wie sie einen Rest ihrer Bankgeheimnisse retten können

    Vier Wochen vor dem Treffen der 20 wichtigsten Wirtschaftsnationen der Welt (G20) kommt Bewegung in die Riege der europäischen Steueroasen. Die Schweiz, größter Finanzplatz für internationale Privatvermögen, hat sich am Freitag offiziell zu einer weiteren Aufweichung ihres Bankgeheimnisses bereit erklärt. Ähnliche Signale werden von einem Treffen der Finanzminister der Schweiz, Österreichs und Luxemburg am Sonntag in Luxemburg erwartet.

    “Wir wollen die internationale Zusammenarbeit bei Steuerdelikten verbessern”, sagt der Schweizer Bundespräsident und Finanzminister Hans-Rudolf Merz nach einer Kabinettssitzung in Bern. Die Drohungen von Seiten großer Industriestaaten seien ungerechtfertigt, aber ernst zu nehmen. Für die Schweiz gelte es zu verhindern, dass sie von der G20 oder einem ähnlichen internationalen Forum auf eine Schwarze Liste gesetzt werde. “Auf Verträge einzugehen, die unter Sanktionen entstanden sind, wäre ganz schlecht”, sagte Merz. Erwartet wird, dass Österreich und Luxemburg bei dem Treffen am Sonntag ein ähnliche Position einnehmen werden. Belgien und Luxemburg, so sagte der britische Botschafter in Bern jüngst, hätten ohnehin signalisiert, dass sie das Bankgeheimnis nach 2013 “nicht in dieser Form weiterführen werden”. Ein Sprecher des Luxemburger Finanzministers wies diese Darstellung am Freitag jedoch zurück.