Most of us are familiar with the important role that clouds play in the Earth’s hydrological cycle (Fig. 1). What is less obvious is that clouds play a significant role in regulating the Earth’s climate through their interaction with incoming solar radiation and outgoing heat. Clouds can help cool the Earth by reflecting solar radiation back into space (albedo effect) and can also trap heat in the atmosphere. Consequently, understanding the distribution, lifetime, and properties of clouds is essential for understanding the Earth’s radiative budget and climate. However, there are many knowledge gaps in our understanding of the processes determining cloud properties, and clouds remain a major source of uncertainty in our models of climate and climate change.

Clouds form from tiny water droplets, ice crystals, or a mixture of water and ice. Even in the tropics, ice is a major component of clouds. Pure water droplets in the atmosphere do not freeze until the temperature falls to -38 °C. However, aerosol particles known as ice nucleating particles (INPs) catalyze the freezing of water at much warmer temperatures, thereby increasing the range of conditions where ice-containing clouds form. The role of biological particles in cloud formation is poorly characterized.

The objective of our research was to select specific and common biomolecules and determine whether they could act as effective INP. Previous work has shown that organic matter from a variety of sources can act as INP. Much of this research has been based on complex samples collected in the field or experiments with microorganisms in the laboratory. We took a reductionist approach by measuring whether specific chemical compounds or classes of chemical compound can act as effective INP. This approach avoided the complexity of mixed biological samples and we hoped that we would be able to identify the types of chemical compound and properties that contribute to ice nucleation in the atmosphere.

In the laboratory, we measured whether organic compounds acted as ice nucleating particles using an ‘ice microscope’ (Fig. 2). We placed a small droplet of pure water containing our sample on a hydrophobic slide, which was placed inside a chamber in which temperature and humidity are carefully controlled. The sample was cooled at a defined rate and a change in droplet opacity indicated when it froze.  This change is clearly visible in images taken using a digital camera. After many measurements on many samples, we found that a range of amino acids, proteins, and nucleic acids were effective immersion ice nuclei, catalyzing the freezing of water at temperatures in the range of -19 to -26 °C, significantly warmer than -38 °C. This shows that classes of compound found in all living organisms are potential INP.

There was one protein, RuBisCO, that stood out. Water droplets containing RuBisCO had an average freezing temperature of -7.9 ± 0.3 °C, making it one of the most effective known biological INPs. This was an exciting result as RuBisCO is one of the most abundant proteins on Earth, with a global distribution on land and in the ocean. RuBisCO’s full name is Ribulose-1,5-carboxylase/oxygenase and it is the enzyme that fixes carbon dioxide into organic compounds in plants and phytoplankton (Fig. 3).

We know that RuBisCO is all around us in plants, but is it in the atmosphere? To address this question, we placed a high-volume aerosol sampler on the roof of the tall building where our laboratories are located (Fig. 4). The collected aerosol were analyzed using a sensitive chemical assay specific for RuBisCO. We repeated this process and found that each sample contained RuBisCO. We don’t know how RuBisCO gets into the atmosphere, but it seems unlikely that leaves are the major source. Many microorganisms also contain RuBisCO, so it may come from microalgae blown into the atmosphere from soils or the surface of the ocean.

There are many unanswered questions, but a priority is to determine why RuBisCO is such an effective INP – is it its large size? or something to do with its shape? Or a specific sequence of amino acids? Understanding what makes an effective INP would allow us to make predictions about other sources of biological INP and develop mechanistic models of biological ice nuclei in the atmosphere. We also need data on the distribution of atmospheric RuBisCO to enable us to constrain its role in ice nucleation. This study illustrates that there are fascinating connections and potential feedbacks between the biosphere and the Earth’s climate system. It is an exciting time to be working on biological aerosol.