We now know how insects and bacteria control ice

Proteins help organisms form or inhibit ice crystals

University of Utah

Contrary to what you may have been taught, water doesn’t always freeze to ice at 32 degrees F (zero degrees C). Knowing, or controlling, at what temperature water will freeze (starting with a process called nucleation) is critically important to answering questions such as whether or not there will be enough snow on the ski slopes or whether or not it will rain tomorrow.

Nature has come up with ways to control the formation of ice, though, and in a paper published today in the Journal of the American Chemical Society University of Utah professor Valeria Molinero and her colleagues show how key proteins produced in bacteria and insects can either promote or inhibit the formation of ice, based on their length and their ability to team up to form large ice-binding surfaces. The results have wide application, particularly in understanding precipitation in clouds.

“We’re now able to predict the temperature at which the bacterium is going to nucleate ice depending on how many ice-nucleating proteins it has,” Molinero says, “and we’re able to predict the temperature at which the antifreeze proteins, which are very small and typically don’t work at very low temperatures, can nucleate ice.”

What is ice nucleation?

It’s long been known that life likes to mess with ice. Insects, fish and plants all produce various forms of antifreeze proteins to help them survive in below-freezing conditions. And plant pathogens, particularly the bacterium Pseudomonas syringae, employ proteins that promote the formation of ice to induce damage in their hosts. Before we can talk about how these proteins work, though, we need a quick refresher on how ice freezes.

Pure water, with no impurities, won’t freeze until it reaches -35 degrees C (-31 degrees F). That’s the temperature at which the water molecules will spontaneously arrange into a crystal lattice and start to recruit other molecules to join in. To start the freezing process at warmer temperatures, however, water molecules need something to hold on to, like a speck of dust, soot or other impurity, on which it can start building its crystal lattice. This is the process called nucleation.

Ice-nucleating proteins, such as those in Ps. syringae, bind to nascent ice crystallites in such a way as to reduce the energy cost of additional freezing. They can also aggregate together to further enhance their nucleating power. “It is a lot of group work!” Molinero says.

A snowmaking cannon

These proteins can be so efficient that they can nucleate ice at temperatures as warm as -2 degrees C (29 degrees F). Ice-nucleating proteins are already being put to use at ski resorts, with Colorado-based Snomax International marketing an additive containing Ps. syringae that gives snowmaking machines a boost.

Antifreeze proteins, however, also bind to ice, but force it to develop a curved surface that discourages additional freezing and requires much colder temperatures for ice to grow. Also, antifreeze proteins don’t aggregate together. “They have evolved to be loners, as their job is to find ice and stick to it,” Molinero says.

All of this was previously known, including the fact that antifreeze proteins were relatively small and ice-nucleating proteins were relatively large. What wasn’t known, though, was how the sizes and aggregating behaviors of the proteins affected the temperature of ice nucleation. That’s the question Molinero and her team set out to answer.

A “single bullet”

Molinero and graduate students Yuqing Qiu and in Arpa Hudait conducted molecular simulations of protein interactions with water molecules to see how they affected the temperature of ice nucleation. Antifreeze and ice-nucleating proteins, Molinero says, bind to ice with nearly equal strength.

“Nature is using a single bullet in terms of interactions to address two completely different problems,” she says. “And the way it has resolved between antifreeze or ice nucleation is by changing the size of the proteins and their ability to team up to form larger ice-binding surfaces.”

Antifreeze proteins, they found, nucleated at just above -35 degrees C, which matched experimental data. Lengthening the simulated proteins increased the nucleation temperature, which plateaued after a certain length. The simulations predicted that further assembling around 35 bacterial proteins into larger domains was key to reach the ice-nucleating performance of Ps. syringae, with a nucleation temperature of -2 degrees C (29 degrees F).

“Now we can design new proteins or synthetic materials that nucleate ice at a specific temperature,” Molinero says.

Why it matters

The implications of such a finding extend all the way to the future of water on Earth.

Precipitation begins as ice, which nucleates and grows until it’s heavy enough to precipitate. At high altitudes where it’s colder, soot and dust can do the job of triggering nucleation. But at lower altitudes, it’s not dust that triggers nucleation–it’s bacteria.

Yes, the same proteins in Ps. syringae that aid snowmaking at ski resorts also aid ice formation at warmer temperatures, allowing low-altitude clouds to precipitate. In a warming climate, Molinero’s findings can help climate modelers better understand the conditions of cloud formation and precipitation and forecast how warming will affect the amount of ice nucleation and precipitation in the future.

“The ability to predict whether the clouds are going to freeze or not is super important in climate models, because ice formation determines precipitation and also the ratio of solar energy absorbed and reflected by our atmosphere,” Molinero says. “The challenge to predict whether ice is going to nucleate or not in clouds is a major limitation the predictive ability of weather and climate models.”

At a much smaller scale, however, the antifreeze and ice-nucleating proteins can be employed together in a fine-tuned ice dance: Some insects use antifreeze proteins to protect themselves down to around -8 degrees C (18 degrees F), but then employ ice-nucleating proteins at lower temperatures to contain ice growth before it gets out of hand.

“The big picture is that we now understand how proteins use their size and aggregation to modulate how much they can nucleate ice,” Molinero says. “I think that this is quite powerful.”


Find this release here.

Find the full study here.

From EurekAlert!


43 thoughts on “We now know how insects and bacteria control ice

  1. Does this mean that we can make clouds freeze, thus to cause rain ?

    Instead of seeding a cloud with dry ice, ca we instead drip a spray of the correct bacteria ?


    • The hexagonal lattice of silver iodide, similar to the structure of snow/ice, makes it the preferred choice for cloud seeding. Dry ice was mainly used historically.

    • Ummm, EurekAlert!, people?

      In the real world, surprisingly enough, despite their computer simulations, we do not find any absolutely pure water that supercools and fails to freeze significantly below 0C. It’s not a big question of the ages, will water freeze? Applying what happens biologically inside an insect to the realm of clouds (how many orders of magnitude larger?) is practical in what way?

      It’s models all the way down.

      • Mechanical vibration which is always present to some degree (noise, ground vibration, etc.) also triggers the crystallization to ice (due to compression waves in the water). If the water were pure and in 0 g with no vibration, the lower temperatures would be approached.

  2. ” In a warming climate, Molinero’s findings can help climate modelers better understand the conditions of cloud formation and precipitation and forecast how warming will affect the amount of ice nucleation and precipitation in the future.” <== Mandatory Climate Change speculation required to obtain funding. Despite the fact that current climate models to not take cloud formation into account.

    • Molinero says. “The challenge to predict whether ice is going to nucleate or not in clouds is a major limitation the predictive ability of weather and climate models.”

      Don’t be so critical Edward. He is clearly stating that until climate models can model the “basic physics” of nucleation they are worthless for predicting how the climate system will respond to future warming.

      This is THE key question in climate models since it determines cloud feedbacks and ultimately climate sensitivity. It has taken nearly 40 years to get to this point of admitting that climate models are based of guesswork not “basic physics” for the key processes which would make them a useful model of climate response.

      • “He is clearly stating that until climate models can model the “basic physics” of nucleation they are worthless for predicting how the climate system will respond to future warming. ”

        In that he would be wrong. The intractable flaw of modeling is not basic physics. It is their sensitivity to initial conditions. I.e. lack of complete knowledge of the conditions in each cell.

        If a butterfly flapping its wings over Beijing can trigger a thunderstorm in Brazil two weeks later, there will never be sufficient knowledge of initial conditions to allow for meaningful long term forecasts.

        At this point, the modellers stick a rabbit, into their hats, pull it out, and say: “Behold, a rabbit”. The rabbit is the “forcing” they attribute to the presence of CO2 in the atmosphere. The rest of their act is pure flummery.

          • There is no real difference between weather models and climate models. They are both built on the same types of differential equations and bot suffer from the same problems.

    • “Mandatory Climate Change speculation ”

      Climate will change, it always has, who is expecting otherwise?

      He does use the cliche “in a warming climate” but he does not make any reference to CO2, CAGW or similar. What he is doing is relevant to one of the key failings of climate models. One area which certainly merits more funding. Don’t be unnecessarily critical.

      • No, that area does not merit more funding. That’s because the models don’t merit more funding. That’s because they have too coarse a resolution ever to work. No matter how much you learn about nucleation, say, it won’t make the models effective, because their resolution will never be high enough to make proper use of it.

        But it gets worse. No matter how high a resolution the models get, and no matter how many improvements are made to them, they will never ever be able to simulate climate because they will never ever be able to get initial conditions accurate enough. Bear in mind that model runs have shown that variations of one trillionth of a degree (yes really just one trillionth) in initial conditions can give final result changes of 5 degrees or more. If an error in initial conditions can be multiplied by trillions in a model run without any rules being changed then no matter how much you fix up the rules you are never ever ever going to get a meaningful prediction.

        No funding increase is merited. Period.

    • Possibly , Deloss. But if some of you haven’t read Vonnegut’s Cat’s Cradle, please try to. I have returned to it a couple of times over the years, one of the most amusing illustrations of human stupidity I have ever come across.

  3. While on Cloud 9 for Cause & Effect mangulation, what about this:

    The ‘anchor point’ for this craic is the Wiki page for Greenhouse Effect describing Planet Earth’s energy balance.
    From geometry we calculate the incoming energy from Sol
    Then we apply an ‘adjustmnent’ for Earth’s albedo.

    Here comes the fun, Earth’s albedo is taken to be (measured as) ‘about’ 0.3
    First question- why is it not the same as Moon’s albedo = 0.12?
    Are they not made of the same stuff and generally taken that one is the calf of the other?

    No matter, here on Earth, we can go out and measure albedo of different places.
    Its been done and the figures are all ‘out there’
    Water has very low, snow is very high, dry desert sand is something else, conifers different again etc etc
    But we know how much of each there is:
    70% water
    10% sandy desert
    10& snow/ice
    and of the remaining 10%, do we say 2% conifer, 2% deciduous, 2% green grassland etc
    Make up your own mix and ‘area weight it’ and generate the sum

    I got an Earth albedo of 0.19 by doing that
    The kicker now being, still much less than the measured value of 0.3

    Why. Surely as easy as pie – the increased albedo is caused by clouds
    Thus clouds are reducing the incoming solar energy and thence, if you like, are Cooling The Earth.

    But what are clouds made of or you might ask what are they composed of?
    Water or water vapour.
    Coz if there was no water vapour, there’d be no clouds

    Major point of semantics will arise but it is (again) a point concerning Cause & Effect
    We will be told that clouds are composed of water droplets or tiny ice crystal and are no longer = Water Vapour.
    BUT and taking the extreme case, can you fly an airplane through a block of ice or a lake of water?
    Is it arguable that clouds are much closer to being a gas (vapour) than they are a solid or liquid?

    Entirely irrelevant as if there were no water vapour there’d be No Clouds. Period

    Now I’ve Pooped The Party, Burst The Bubble and Called Time ain’t I just?
    Water Vapour is NOT a Green House Gas that heats the place – it works to cool the planet

    = a conclusion derivable from the words of the Warmists themselves – hence my reference to the Wiki

    And no appeals to authority or consensus, no super computers, no fiddly little graphs and no bad language.

    Diehards will assert that somehow water vapour still DOES have a heat trapping effect in the troposphere despite that that instantly invokes a powerful positive feedback – patently not the case.

    Anyone who reads the junk I post into here will know where that came from:
    Esp: Why does a green growing wheat field have such a high albedo when it is such a dark shade of green and why does its albedo increase as it is fed upon nitrogen fertiliser = something that any farmer will tell you, causes plants to become darker shades of green than they previously were.
    Also from reading this very recently arrived:

    Seemingly the potato plants did better under the shade of the solar panels.
    Crazy innit, but there again, there’s a myriad of things going on in there and as we know, the solar panel enthusiasts will Cherry Pick the one they like.
    Still, a very interesting experiment as I (ymmv) would argue that the solar panels effectively increased the albedo of the plants underneath – just as how they respond to added fertiliser

    And back to the original topic here:
    What are the bacteria doing when they create ice and or clouds if not trying to protect themselves from heat?
    Bacteria survive inside the home freezer – they do not survive inside the cooker

    *Now* you see the Cause & Effect error in the research here presented?
    Puts an epic new slant on Global Greening too……..

    Just think about it, that’s all I ask.

    • That Ma Nature is a bitch, she got me there too.
      Why do the bacteria make ice up in the top of the sky?

      Because it would protect them from the high levels of UV radiation there is up there.
      It really is a Magical Place, this Earth, innit?

    • Love your comments Peta. Yes water definitely cools the planet. I have arrived at that conclusion, however, by a different route, namely:
      At water phase change to vapor absorbed energy is converted to Latent Heat at constant temperature. Thus in the Planck equation the sensitivity coefficient is zero. (ie: no resulting temperature increase). This vapor is buoyant wrt dry air and thus rises towards space, taking with it some 680 Watthrs/Kg. for dissipation.
      A cloud comprises both vapor and liquid/solid. We can only see the latter, not the vapor, which is tending to rise while the rest tends to fall; but with less energy content.
      Models do not include this process in their calculations thus they arrive at too high a figure for Climate Sensitivity, now becoming very evident in the actually observed figures.

      Incidentally: 1): The water/atmosphere contact area is far greater than the usually used figure of 70% of the earth’s surface. Consider the leaves of a tree and extrapolate up for all the flora on earth! See what I mean?
      2): Water is a greenhouse gas; but, as described above it does NOT result in an increase in temperature. Thus it has as you rightly say a strong negative feedback to the CO2 influence. Contrary to the IPCC claim.


    • Peta
      You said, “… Water has very low, [albedo]” That isn’t true. The albedo of water is predominantly controlled by diffuse reflectance from particles in suspension, such as silt and plankton. It can range from zero to double digits, depending on the concentration of the suspended reflectors and their inherent reflectivity. However, water is a specular reflector where the directed reflectivity (not albedo) ranges from about 2% with the sun at solar noon and the measuring platform directly overhead as well, to 100% with the sunlight having a glancing angle of incidence and the measuring platform looking towards the sun.


  4. A moderately interesting paper.

    But the climate-change waffle is on very shaky ground. There is zero evidence that bacterial proteins are important for cloud nucleation. But it is like the irrelevant Lenin quote in soviet-era science papers. It signals that you are loyal to the party line.

  5. I just hopeone thinks about using the psuedomonas a bit more,
    if its a plant killer then spraying it IN water used for icemaking is going toleave it in the area to infect plants
    thats not so bright.
    but then I guess its all about the almighty tourist $

  6. How does this effect the conclusions of the Marsh-Svensmark experiments at the CERN cloud chambers?
    I believe the conclusion was that Cosmic Rays (increasing as the sun’s magnetic field weakens) create nuclei too small to nucleate cloud droplets?
    A stretch–but this could be an additional amplification factor!

  7. “Pure water, with no impurities, won’t freeze until it reaches -35 degrees C (-31 degrees F)”

    Something is not right here.

    • Also with this: “These proteins can be so efficient that they can nucleate ice at temperatures as warm as -2 degrees C (29 degrees F).” When a paper makes such basic errors, it detracts from the overall quality. While there is some interesting stuff here, their attempt to broaden its importance to climate further detracts from it. I give it about a C -.

  8. I’m disappointed at the inaccurate statement “water doesn’t always freeze at 32f”! If it doesn’t then by definition it isn’t “water”!!! And this coming from people at my alma mater! Sheesh!!

    • With all due respect, it does change with changing pressure as does the boiling point. Water is a wonderful substance with a wide range of properties over temperature and pressure. A state chart would illustrate some of this.

      • For those interested in the miraculous liquid we all take so much for granted, it is virtually impossible to melt ice at any temperature we can achieve at pressures higher than 70,000 bar, but it is not ice as we know it under those conditions, having a specific gravity higher than that of water. Rocky probably had less extreme conditions in mind and a more “normal” kind of ice!

          • While insects are mentioned a couple times, the “domain” of the article is not insects. Fish and clouds are also mentioned. The “domain” of the article is how proteins affect the freezing point of water.

      • …as explained at length in the 798 pages of one of my favorite college text books: Geochemistry of Hydrothermal Ore Deposits, Second Edition, edited by Hubert Lloyd Barnes, 1979.

        Packed with enough phase diagrams to choke a horse, the book deals with about every combination of temperature, pressure, and dissolved mineral component found in nature! A challenging read and one that defines several types of water (surface, connate, metamorphic, and magmatic) but not once did it try to trick the reader with the silly hook line that water freezes at other than 32f because when it does, the astute student knows that it really isn’t “water”! And maybe biologists don’t know that!

        Standard temperature and pressure may not be taught at university anymore; it has been a while since I taught a grad course in geochemistry!

      • …none of which applies to insects! The sligtest perturbation in a supercooled liquid disrupts the metastable state! Hence, flash solidification!

  9. Hey, fellows! Freezing of water is a complex subject. Nucleation is even worse. Water does not freeze until the energy balance says that the crystalline water has a lower entropy than the liquid water. any contaminants, solutes (see “Molar Depression of freezing point”) or, apparently proteins affect the freezing point.

    Still wrong, but probably better to say pure water does not start to melt until it rises to 273o Kelvin (o degrees C) at STP.

  10. From the abstract
    “ …and, after validating the theory against Thet of the proteins in molecular simulations, we use it to predict…..

    Of course, this is hypothesis stacking, but the significant point is the role of what have been long known as colligative properties of water. In the 1989 killing freezes in the Gulf of Mexico, sea water of lower salinity, as in closer to river mouths, froze quicker and thicker. Unfortunately, there is not a lot of data, most holed up behind the devices I write from. That’s not sarc.

    Simple question, water can be full of organic matter, sea water with lots of free proteins around, collecting in still incompletely understood ways. There is a type of “rain” in the ocean, something to do with gravity, but never heard of proteins raising the freezing point. Wish they would do an actual experiment which used to be the rule, but there is nothing in the abstract to show this.

  11. Something to consider: Since the invention of antibiotics, they have become ubiquitous. They are not only used in human disease treatment, but in prodigious quantities in animal husbandry. Antibiotics can be found throughout the environment because animals excrete the compounds. What if the anomalous quantities of antibiotics are making their way into the atmosphere and affecting the floating microbes that in turn affect nucleation of water droplets in clouds? The amount of precipitation would be impacted by declining populations of nucleating microbes, and the energy balance would be shifted because the heat of crystallization released would be less frequent, particularly at lower altitudes. Might it be more than coincidence, or a spurious correlation, that it has been warming since the invention of antibiotics? Or, might antibiotics be playing an unidentified role in the very complex interplay of the elements of weather and climate?

    • I’d like to give a Jordan Peterson response to that one: “No.”

      I seriously doubt that microbes are a significant source of nucleation sites in clouds compared to natural and manmade pollution aerosols, soot, dust, pollen, or ionization trails from cosmic rays for that matter. Even if they were dead bacteria, they would still be nucleation sites and they only get into the atmosphere in the first place presumably through convection. It is not as if there is a cloud of bacteria reproducing in mid-air.

      • Rich Davis
        It sounds as though you didn’t bother to read the article that I provided a link for, before responding.

  12. “These proteins can be so efficient that they can nucleate ice at temperatures as warm as -2 degrees C (29 degrees F).”

    Minus 2 degrees C is 28.4 degrees F. If the authors are going to be so condescending as to assume that we can’t easily do the conversion, they should at least be able to round off correctly.

  13. Nature has come up with ways to control the formation of ice

    And other varieties of the same concept in the article.

    Nature doesn’t ‘do’ anything. It is not an active thing. It just is. Basically what we observe.

    This is a common theme. People like to anthropomorphise natural things. Often people pretend that evolution itself has a purpose.

    Nature does not have a purpose. Evolution does not have a purpose. The Earth, solar system, galaxy and universe do not have a purpose. Life has no purpose. It all just is!

    For supposed scientists to claim, or even imply, otherwise just proves that they are not scientists at all.

  14. “Molinero and graduate students Yuqing Qiu and in Arpa Hudait conducted molecular simulations ”
    Note that this whole study was done with computer model simulations. Take it with a grain of salt and pinch lightly.

Leave a Reply

Your email address will not be published. Required fields are marked *