Highly active and stable fungal ice nuclei are widespread among Fusarium species

Some biological particles and macromolecules are particularly efficient ice nuclei (IN), triggering ice formation at temperatures close to 0 ◦C. The impact of biological particles on cloud glaciation and the formation of precipitation is still poorly understood and constitutes a large gap in the scientific understanding of the interactions and co-evolution of life and climate. To investigate the frequency and distribution of IN activity within the fungal genus Fusarium, more than 100 strains from 65 different Fusarium species were screened. In total, ∼11 % of all tested species included ice nucleation-active (IN5 active) strains, and ∼16 % of all tested strains showed IN activity above -14 ◦C. Besides Fusarium species with known IN activity, F. armeniacum, F. begoniae, F. concentricum, and F. langsethiae were newly identified as IN-active. The cumulative number of IN per gram of mycelium for all tested Fusarium species was comparable to other biological IN like Sarocladium implicatum, Mortierella alpina, and Snomax®. Filtration experiments suggest that the single cell-free Fusarium IN is smaller than 100 kDa, and that aggregates can be formed in solution. Long-term storage and freeze-thaw cycle experiments revealed 10 that the Fusarium IN remain active in solution for several months and after repeated freezing and thawing. Oxidation and nitration reactions, as occurring during atmospheric aging, did not affect the activity of the Fusarium IN. The high frequency of Fusarium and the wide distribution of IN activity within the genus, combined with the high stability of the IN, suggest a significant impact of fungal IN on the Earth’s water cycle and climate.


Introduction
Ice particles in the atmosphere are formed either by homogeneous nucleation at temperatures below -38 • C or by heterogeneous nucleation catalyzed by particles or macromolecules serving as ice nuclei (IN) at warmer temperatures (Pruppacher and Klett, 1997).Biological particles in particular are expected to play an important role as IN in the temperature range from -15 • C to 0 • C, but the impact of biological particles on cloud glaciation and the formation of precipitation is still poorly understood (Coluzza et al., 2017).Several studies suggest a triggering effect of biological IN for cloud glaciation and formation of precipitation (Creamean et al., 2013;DeMott and Prenni, 2010;Pratt et al., 2009), and former studies have shown that 1 https://doi.org/10.5194/bg-2019-276Preprint.Discussion started: 31 July 2019 c Author(s) 2019.CC BY 4.0 License.biological particles are more efficient than mineral IN (DeMott and Prenni, 2010;Després et al., 2012;Hill et al., 2014;Hoose and Möhler, 2012;Huffman et al., 2013;Möhler et al., 2007;Morris et al., 2014;Murray et al., 2012;Pratt et al., 2009).
Other species can produce secondary metabolites known as mycotoxins that can cause a variety of acute and chronic health effects in humans and animals (e.g., Bush et al., 2004;Ichinoe et al., 1983).
Whereas the positive selective pressure for IN activity in Fusarium and other fungi has not been directly identified, an ecological advantage of initiating ice formation is easily conceivable.Indeed, most IN-active bacteria and fungi are isolated from regions with seasonal temperatures below 0 • C (Diehl et al., 2002;Schnell and Vali, 1972).Ice nucleation activity at temperatures close to 0 • C could be beneficial for pathogens or might provide an ecological advantage for saprophytic Fusarium species by facilitating in the acquisition of nutrients liberated during cell rupture of the host (Lindow et al., 1982).Furthermore, IN on the surface of the mycelium could avoid physical damage of the fungus by protective extracellular freezing (Fröhlich-Nowoisky et al., 2015;Zachariassen and Kristiansen, 2000) or to bind moisture as ice in cold and dry seasons (Pouleur et al., 1992).With increasing temperatures, the retained water can be of advantage in early vegetative periods and for bacterial movement on the mycelial water film known as fungal highway (Kohlmeier et al., 2005;Warmink et al., 2011).Moreover, IN activity might be beneficial for airborne Fusarium and for their return to the Earth's surface under advantageous conditions in a feedback cycle known as bioprecipitation (Després et al., 2012;Morris et al., 2013Morris et al., , 2014;;Sands et al., 1982).In addition, once the IN are released into the environment, they can adsorb to clay and might also be available in the atmosphere associated with soil dust particles (Conen et al., 2011;Fröhlich-Nowoisky et al., 2015, 2016;Hill et al., 2016;O'Sullivan et al., 2014O'Sullivan et al., , 2015O'Sullivan et al., , 2016;;Sing and Sing, 2010).
The sources, abundance, and identity of biological IN are not well characterized (Coluzza et al., 2017), and it has been proposed that systematic surveys will likely increase the number of IN-active fungal species discovered (Fröhlich-Nowoisky et al., 2015).Fusarium is the best-known IN-active fungus, but the frequency and distribution of IN activity within Fusarium is not well known.In this study, more than 100 strains from 65 different Fusarium species were tested for IN activity in three  S1).
The USDA-ARS/Michigan State University strains were cultivated on dextrose peptone yeast extract agar, containing 10 g L -1 dextrose (VWR, Radnor, PA, USA), 3 g L -1 peptone (Difco Proteose Peptone No. 3, Becton, Dickinson and Company, Franklin Lakes, NY, USA), and 0.3 g L -1 yeast extract (Merck, Kenilworth, NJ, USA), filtered through a 0.2 µm pore diameter filter (PES disposable filter units, Life Science Products, Frederick, CO, USA).After filtration, 12 g L -1 agarose (Certified Molecular Biology Agarose, Bio-Rad, Hercules, CA, USA) was added, and the medium was sterilized by autoclaving at 121 • C for 20 min.The colonies were grown at 22 • C to 24 • C for 7 to 19 days.The strains from the Schmale Laboratory at Virginia Tech and the Kansas State University Teaching Collection were maintained in cryogenic storage at -80 • C and were grown on quarter-strength potato dextrose agar (Difco Laboratories, Detroit, USA) on 100 mm Petri plates at ambient room temperature for four days prior to ice nucleation assays.
For quantitative analysis, exposure experiments, freeze-thaw cycles, as well as short-and long-term storage tests a selection of IN-active tested strains was grown on full-strength potato dextrose agar (VWR International GmbH, Darmstadt, Germany) first at room temperature for four to six days and then at 6 • C for about four weeks.For filtration experiments, the fungal cultures were grown at 6 • C for up to six months.

Preparation and treatments of aqueous extracts
For LED-based Ice Nucleation Detection Apparatus (LINDA) (Stopelli et al., 2014) experiments (see Sect. 2.3), 4 mL of sterile 0.9 % NaCl was added to each of eight petri plates, and the fungal cultures were scraped with the flat end of a sterile bamboo skewer.The resulting suspension of mycelium and spores was filtered through a 100 µm filter (Corning Life Sciences, Reims, France).
For Twin-plate Ice Nucleation Assay (TINA) (Kunert et al., 2018) experiments (see Sect. 2.3) the fungal mycelium was scraped off the agar plate and transferred into a 15 mL tube (Greiner Bio One, Kremsmünster, Austria).The fresh weight of the mycelium was determined gravimetrically.Pure water was prepared as described in Kunert et al. (2018).Aliquots of 10 mL pure water were added before vortexing three times at 2 700 rpm for 30 s (Vortex-Genie 2, Scientific Industries, Inc., Bohemia, For filtration experiments, the 0.1 µm filtrate was further filtered successively through 300 000 MWCO and 100 000 MWCO PES ultrafiltration units (Vivaspin ® , Satorius AG, Göttingen, Germany).After each filtration step, the IN concentration was determined using TINA.
For exposure experiments, aqueous extracts of F. acuminatum 3-68 and F. avenaceum 2-106 were exposed to high concentrations of O 3 and NO 2 as described in Liu et al. (2017).Briefly, a mixture of 1 ppm O 3 and 1 ppm NO 2 was bubbled through 1 mL aliquots of aqueous extract for 4 h, which represents an exposure to an atmospherically relevant amount of approximately 200 ppb of each gas for about 20 h.Afterwards, the IN concentration was determined using TINA.
For freeze-thaw cycles, the IN activity of F. acuminatum 3-68 was determined shortly after preparation of the aqueous extract and after storage at 6 • C for 24 h using TINA.Then, the aqueous extract was stored at -20 • C for 24 h and thawed again.The IN activity was tested before storage at -20 • C for an additional 24 h.After thawing, the IN activity was determined again.
For long-term storage experiments, the aqueous extract of various Fusarium species was stored at 6 • C for about four months or at -20 • C for about eight months, and the IN activity was determined using TINA.

Ice nucleation assays
Two independent droplet freezing assays conducted in two laboratories were used to investigate the distribution of IN activity within Fusarium in an initial screening.
First, a thermal cycler (PTC200, MJ Research, Hercules, CA, USA) was used as described in Fröhlich-Nowoisky et al. (2015) to screen 30 Fusarium strains from seven species from USDA-ARS/Michigan State University in the temperature range from -2 • C to -9 • C. Mycelium was picked with sterile pipette tips (Eppendorf, Westbury, NY, USA) into 80 µL aliquots of 0.2 µm pore diameter filtered dextrose peptone yeast extract broth in sterile 96-well polypropylene PCR plates (VWR International, LLC, Radnor, PA, USA).
Second, the LED-based Ice Nucleation Detection Apparatus (LINDA) was used as described by Stopelli et al. (2014) to screen 13 strains from the Schmale Laboratory at Virginia Tech and 69 strains from the Kansas State University Teaching Collection.Aliquots of 200 µL of each aqueous extract were transferred to three separate 500 µL tubes and placed on ice for 1 h prior to the LINDA experiments.LINDA was run from -1 • C to -14 • C, and images of the samples were recorded every six seconds.As positive control, aqueous suspensions of Pseudomonas syringae CC94 from the collection of INRA (Avignon, France) (Berge et al., 2014) (with a final OD 580 of 0.5 to 0.7, i.e. ∼10 9 bacteria mL -1 ) were used for each experiment.The bacteria were grown on King's medium B (King et al., 1954) at 22 • C to 25 • C for 48 h, and aqueous suspensions were incubated at 4 • C for 1 h to 4 h before LINDA experiments.The aqueous extract was prepared in 0.9 % NaCl solution, which reduced the freezing temperatures about 0.5 • C based on theoretical calculations.Ice nuclei of selected Fusarium species were further analyzed using the high-throughput Twin-plate Ice Nucleation Assay (TINA) (Kunert et al., 2018).The aqueous extract was serially diluted 10-fold with pure water by a liquid handling station with the associated error using the Vali formula and the Gaussian error propagation (Kunert et al., 2018;Vali, 1971).For each experiment, the cumulative number of IN was averaged over all dilutions.If the experiment was repeated, the cumulative number of IN was averaged over all experiments, and the standard error was calculated.Three independent experiments with aqueous extract from three individual fungal culture plates of the same strain showed similar results with only slight variation.
An example of results is presented for F. armeniacum 20970 (Fig. S1).

IN-active Fusarium species
Although several IN-active Fusarium species are known, the frequency and distribution of IN activity within the fungal genus Fusarium is still not well studied (Hasegawa et al., 1994;Humphreys et al., 2001;Pouleur et al., 1992;Richard et al., 1996;Tsumuki and Konno, 1994).Two initial screenings in the temperature range from -1 • C to -14 • C were performed to better evaluate the frequency of IN activity within Fusarium.
In total, ∼ 16 % (18/112) of the tested strains showed IN activity with initial freezing temperatures of -3.5 • C to -11.2 • C (Table 1) in the typical range known for Fusarium (-1 • C and -9 • C) (Hasegawa et al., 1994;Humphreys et al., 2001;Pouleur et al., 1992;Richard et al., 1996;Tsumuki et al., 1992;Tsumuki and Konno, 1994).Most formerly reported initial freezing temperatures were obtained with different Fusarium strains, growth conditions, and freezing assays, which might explain differences compared to our results.The high proportion of IN-active strains within F. acuminatum is consistent with previous reports (Pouleur et al., 1992;Tsumuki et al., 1995).Overall, ∼ 11 % (7/65) of the tested species included IN-active strains.In addition to strains from Fusarium species with known IN activity, four Fusarium species were newly identified as IN-active: F. armeniacum, F. begoniae, F. concentricum, and F. langsethiae.In further experiments, the IN activity of F. begoniae and F. concentricum could not be verified.
The newly identified IN-active species are cosmopolitan.Fusarium armeniacum is a toxigenic saprophyte (Burgess et al., 1993) causing seed and root rot on soybeans (Ellis et al., 2012).The geographical distribution has been reported as tropical and subtropical (Leslie and Summerell, 2006), but it was also found in Minnesota, USA (Kommedahl et al., 1979) and Australia (Burgess et al., 1993).Fusarium begoniae is a plant pathogen of Begonia found in Germany with a potential wider distribution (Nirenberg and O'Donnell, 1998).Fusarium concentricum is a plant pathogen, frequently found in Central America and isolated from bananas (Aoki et al., 2001;Leslie and Summerell, 2006), and F. langsethiae is a broadly distributed cereal pathogen (Torp and Nirenberg, 2004).Some strains of these newly identified IN-active species are known to produce mycotoxins, which can threaten the health of humans and animals (Fotso et al., 2012;Kokkonen et al., 2012;Wing et al., 1993a, b).least to contain a proteinaceous compound (Hasegawa et al., 1994;Pouleur et al., 1992).Their production requires energy, and we might assume that this trait would not be expressed or maintained unless there was an ecological advantage.It is known that Fusarium can regulate the gene expression for IN production depending on environmental conditions such as nutrient availability (Richard et al., 1996), and some Fusarium species reduce or lose their IN activity after several subcultures (Pummer et al., 2013;Tsumuki et al., 1995).Thus, we cannot exclude that all Fusarium strains have the ability to produce IN.From the phylogenetic distribution of IN activity across the genus Fusarium, we can speculate that IN activity is a very old trait, but either the gene expression requires a trigger, which is not yet identified, or the trait might be in the process of being lost.It is unlikely, however, that the age of the genetic determinants of fungal IN activity is older than that in bacteria, since fungi diverged well after the age that has been attributed to the bacterial IN gene (Morris et al., 2014), and the genetic determinants are not the same as those in bacteria.

Quantification and size determination of IN from selected Fusarium species
A selection of IN-active Fusarium species was further investigated by extensive droplet freezing assay analysis using TINA.All tested Fusarium strains initiated ice nucleation between -3 • C and -4 • C (Fig. 1).Differences in the initial freezing temperature between the initial screening and the quantitative analysis can be due to different growth conditions and freezing assays.The cumulative number of IN (N m ) per gram of mycelium was in the range between 10 8 g -1 and 10 13 g -1 .Fusarium acuminatum  Kunert et al., 2018), and the bacterial IN-active substance Snomax ® containing Pseudomonas syringae (10 12 g -1 , Budke and Koop, 2015;Kunert et al., 2018).
The size of the Fusarium IN was investigated by filtration experiments.Filtration through a 5 µm and a 0.1 µm filter did not affect the IN activity (Fig. 2), revealing that Fusarium IN are cell-free, easily removed from the fungus, and stay active in solution.This is consistent with previous studies (O 'Sullivan et al., 2015;Pouleur et al., 1992;Tsumuki and Konno, 1994).
Moreover, the IN are smaller than 100 nm for all tested Fusarium strains.Filtration through a 300 000 MWCO filter unit decreased the cumulative number of IN per gram of mycelium about 50 % to 75 % depending on the Fusarium species, but a tremendous number of IN (10 10 -10 13 g -1 ) still passed through the filter.The initial freezing temperature was slightly shifted towards lower temperatures.Further filtration through a 100 000 MWCO filter unit reduced the IN number to 10 8 -10 10 g -1 , which is less than 1 % of the initial IN concentration.Additionally, the initial freezing temperature was shifted about one degree towards lower temperatures.
As IN activity was found in all filtrates, the aqueous extract of Fusarium consists of a mixture of IN-active proteins with different sizes.We hypothesize that Fusarium IN are single proteins smaller than 100 kDa, which agglomerate to large protein complexes in solution.Some of these complexes fall apart upon filtration, so that the single IN proteins can pass through the filter.The small shift in the initial freezing temperature suggests that these proteins reassemble again to aggregates after filtration, as larger IN nucleate at warmer temperatures (Govindarajan and Lindow, 1988;Pummer et al., 2015).Assuming the protein as smooth spherical particle, the minimum diameter of the single IN protein would be smaller than 6.1 nm according https://doi.org/10.5194/bg-2019-276Preprint.Discussion started: 31 July 2019 c Author(s) 2019.CC BY 4.0 License.
to Erickson (2009).Our results are in accordance with Lagzian et al. (2014), who cloned and expressed a 49 kDa IN-active protein from F. acuminatum.
As Fusarium IN are cell-free and can easily be washed off the fungal surface, they can be released in high numbers into the environment.If they are not degraded by microorganisms before, the IN can adsorb to soil dust and be aerosolized attached to these particles (Conen et al., 2011;Fröhlich-Nowoisky et al., 2015;Hill et al., 2016;O'Sullivan et al., 2014O'Sullivan et al., , 2015O'Sullivan et al., , 2016;;Sing and Sing, 2010).This is in good agreement with Pruppacher and Klett (1997), who found a positive correlation between IN number concentration and particles in the coarse mode.Other releasing processes cannot be excluded, however, it is unlikely that the single proteins are present in the atmosphere as individual aerosol particles.

Stability of Fusarium IN
In the atmosphere, IN can interact with other aerosol particles or gases.They can be exposed to chemically modifying agents like ozone and nitrogen dioxide, and physical stressors like low temperatures and quickly changing temperatures.To investigate the stability of Fusarium IN, we performed exposure experiments, freeze-thaw cycles, and long-term storage tests.
The influence of chemical processing on the Fusarium IN, in particular oxidation and nitration reactions as occurring during atmospheric aging, was investigated by exposing aqueous extracts from F. acuminatum 3-68 and F. avenaceum 2-106 to high concentrations of ozone and nitrogen dioxide in liquid phase.Figure 3 shows that for both species neither the initial freezing temperature nor the cumulative number of IN per gram of mycelium was affected by exposure.These results demonstrate a high stability of Fusarium IN under oxidizing and nitrating conditions.This is in contrast to bacterial IN (Snomax ® ), which were reduced upon exposure (Kunert et al., 2018).
To study the effects of short-term storage and freeze-thaw cycles on the IN activity of F. acuminatum 3-68, IN measurements of the same aqueous extract were performed at different time points (Fig. 4).The results of freshly prepared aqueous extract revealed that the highest activity of fungal IN was already developed during preparation of the filtrate and no time for equilibration was required.Storage of aqueous extract at 6 • C for 24 h did not affect the IN activity.Also, further storage at -20 • C for another 24 h, and repeated freeze-thaw cycles had no impact on the IN activity.This means, that, once in the atmosphere, the IN can undergo several freeze-thaw cycles without losing their activity and are still able to influence cloud glaciation and the formation of precipitation.This could be an explanation why not all fungi are always IN-active as their IN are highly stable and quasi recyclable.Ice nuclei might influence the availability of moisture over long times periods, and if enough moisture is available in the environment, the necessity of IN production would be omitted and the fungus could safe energy.
In addition, the stability of Fusarium IN was studied in long-term storage tests, where aqueous extract of various Fusarium species was stored at different temperatures for a long period of time.
https://doi.org/10.5194/bg-2019-276Preprint.Discussion started: 31 July 2019 c Author(s) 2019.CC BY 4.0 License.laboratories with different freezing methods.A high-throughput droplet freezing assay was used to quantify the IN of selected Fusarium species, and filtration experiments were performed to estimate the size of the Fusarium IN.Furthermore, the stability of Fusarium IN upon exposure to ozone and nitrogen dioxide, during freeze-thaw cycles, and after short-and long-term storage under various conditions was investigated.2 Materials and methods 2.1 Origin and growth conditions of fungal cultures Thirty Fusarium strains from USDA-ARS/Michigan State University (L.Hanson, East Lansing, MI, USA), 13 strains from the Schmale Laboratory at Virginia Tech (D. Schmale, Blacksburg, VA, USA), and 69 strains from the Kansas State University Teaching Collection (J.Leslie, Manhattan, KS, USA) were screened for IN activity (Table https://doi.org/10.5194/bg-2019-276Preprint.Discussion started: 31 July 2019 c Author(s) 2019.CC BY 4.0 License.NY, USA) and centrifugation at 4 500 g for 10 min (Heraeus Megafuge 40, Thermo Scientific, Braunschweig, Germany).For all experiments the aqueous extract was filtered successively through a 5 µm and a 0.1 µm PES syringe filter (Acrodisc ® , Sigma-Aldrich, Taufkirchen, Germany), and the aqueous extract contained IN from spores and mycelial surfaces.
https://doi.org/10.5194/bg-2019-276Preprint.Discussion started: 31 July 2019 c Author(s) 2019.CC BY 4.0 License.(epMotion ep5073, Eppendorf, Hamburg, Germany), and 96 droplets (3 µL) were tested per dilution with a continuous cooling rate of 1 • C min -1 from 0 • C to -20 • C. The temperature was measured with an accuracy of 0.2 K (Kunert et al., 2018).The obtained fraction of frozen droplets (f ice ) and the counting error were used to calculate the cumulative number of IN (N m ) Fusarium is more widespread than previously known.Not all Fusarium species include IN-active strains and not all strains within one species show IN activity.Fusarium IN are thought to be proteins or at https://doi.org/10.5194/bg-2019-276Preprint.Discussion started: 31 July 2019 c Author(s) 2019.CC BY 4.0 License.

Figure 1 .Figure 2 .Figure 3 .Figure 4 .Figure 5 .
Figure 1.Overview of IN activity for selected Fusarium species and strains: cumulative number of IN (Nm) per gram of mycelium plotted against the temperature (T); arithmetic mean values and standard error of three independent experiments with aqueous extracts from different fungal culture plates.