An ice-binding protein from an Arctic population of American dunegrass, Leymus mollis

Several cold-hardy grasses have been shown to have ice-binding proteins (IBPs) that protect against freeze-thaw injury. Here, we looked for IBP activity in an Alaskan coastal grass, Leymus mollis (Pooidae), that had not previously been examined. Rhizome tissue had strong ice-structuring and ice recrystallization inhibiting (IRI) activities, indicating the probable presence of IBPs. The gene sequence of an IBP was obtained. The sequence encoded a 118-amino acid IRI domain composed of eight repeats and that was 80% identical to the IRI domain of the IBP of perennial ryegrass Lolium perenne. The predicted 3D structure of the IRI domain had eight beta-roll coils like those in L. perenne IBP


Introduction
Within the grass family (Poaceae), the subfamily Pooideae includes many cold-adapted grasses including wheat, barley and forage grasses. These grasses have developed a large family of ice-binding proteins (IBPs) that protect the plants from freezing damage (Sandve et al., 2008). The IBPs, which have negligible effect on the freezing point, are characterized by a C-terminal domain consisting of several repeating units that is associated with strong ice recrystallization inhibition (IRI) activity (Sandve et al., 2008). This region is thus called the IRI domain. Ice recrystallization, in which larger ice grains grow at the expense of smaller grains, occurs mostly at warmer sub-zero temperatures and is thought to cause damage to cell walls. Two of the most studied of these IBPs are from the grasses Lolium perenne (Sidebottom et al., 2000) and Deschampsia antarctica (John et al., 2009). Their amino acid sequences are very similar in both the N-terminal and IRI domains. The 118-a.a. IRI domain of the L. perenne IBP has been crystallized and its structure determined by X-ray crystallography (Middleton et al., 2012). The IRI domain has a beta-roll fold with eight similar coils, one side of which is the predicted ice-binding site. The spacing of the coils is very close to the repeat distance along the a-axis of ice. Interestingly, heterologous expression of the Lolium IRI domain in tomato was also shown to increase chilling (4°C) tolerance (Balamurugan et al., 2018), although the mechanism remains unclear. Here, we describe a related IBP from a pooid grass from Utqiaġvik (Barrow), Alaska, Leymus mollis. This site, at 71.3° N, is the highest latitude at which IBPs have been examined in a grass It is characterized by a short growing season from mid-June to late August with an average air temperature of 4°C (Tieszen, 1972), and frequent subzero air temperatures (Table S1, Extended data).
L. mollis, also known as American dunegrass, is found in coastal habitats, especially sand dunes across North America and Greenland. It is typically subjected to many stresses, such as low nutrient levels, salt spray, little freshwater, inundation during storms, wind abrasion, and ice storms (Gagné & Houle, 2002). Here we characterize the IBP of L. mollis and compare it with the IBP of L. perenne, a plant that is not as well adapted to such northern regions (Helgadóttir et al., 2018).

Grass
A grass sample was collected from a gravel beach at Utqiaġvik (Barrow), Alaska (71.3°N) on 6 October 2019, and stored at -80°C. The grass was shipped frozen to University of Nevada Las Vegas for analysis, but accidentally rose to ~10°C for about one day during shipment.

Activity measurements
To obtain a sample for measuring ice structuring and IRI activities, the brown outer layers of a rhizome from just below the soil surface were peeled off, revealing green tissue underneath. About 70 mg of the green tissue was ground in 1 ml water in a mortar and pestle. The suspension was centrifuged at 14,000 rpm for 5 minutes to yield a slightly cloudy supernatant. Stem tissue from lawn grass (Festuca sp.) at the University of Nevada Las Vegas (52 mg) was similarly homogenized and used as a control. Ice-structuring activity was observed by examining the growth of an ice seed crystal with well-defined ice c-and a-axes submerged in the grass extract supernatant, as described previously (Raymond & Fritsen, 2001). Briefly, the sample was placed in a rectangular tube and the tube was submerged in a controlled temperature bath with front and rear windows. The growth of the crystal was observed at a temperature slightly below the freezing point (~-0.2°C) with a horizontally mounted dissecting microscope. Ice-structuring activity was defined as the appearance of sharply defined facets on the crystal surface. IRI activity was observed as described previously (Raymond & Fritsen, 2001). Briefly, 3 µl drops of supernatant were placed on a slide cover glass. A liquid nitrogen-cooled slide glued to a handle was pressed against the drops to form highly polycrystalline ice between the slide and cover glass. The samples were stored at -3°C in hexane and changes in recrystallization were monitored in the temperature bath described above over 21 hours. Photographs were taken through crossed polarizers.

Sequencing
Green tissue was obtained and homogenized as described above. DNA was extracted with a NucleoSpin Plant II kit (Machery Nagel; catalog number 740770) according to the manufacturer's instructions. To make IBP primers, the nucleotide sequences of IBPs from L. perenne, D. antarctica and Triticum aestivum (GenBank accession numbers EU680848, FJ663044 and KU204387, respectively) were aligned. Regions with high identities at the 5' and 3' ends were selected for primer sequences to obtain an amplicon that covered as much of the gene as possible. Primers for 18S ribosomal RNA were selected from conserved regions in the 18S sequences from Dupontia fisheri, L. perenne and D. antarctica (GenBank accession numbers KP794861, KJ598999 and MH628292, respectively) that flanked a region of high variability. At that time, we

Amendments from Version 1
Many revisions were made in response to reviewer comments. Leymus mollis is not strictly an Arctic grass so the title was changed to "an Arctic population of American dunegrass, Leymus mollis". The percent identity to L. perenne was added to the Abstract. Details about the growing season on the Arctic coast were added to the Introduction. We clarified the significance of facets on ice crystals in the Results and discussion section. We added another control (Festuca) to Figure 2 to more clearly show the higher IRI activity of L. mollis. A sentence was added to clarify how IBPs are thought to bind to ice to the Results and discussion section. We revised the final paragraph to provide some comparative aspects of cold tolerance and adaptation in plants.
We have added a supplementary table (Table S1) as Extended data, which gives the frequency of subzero temperatures during the growing season.

REVISED
suspected the Utqiaġvik grass sample was D. fisheri and did not use the L. mollis 18S sequence. PCR was carried out with an Eppendorf Mastercycler Personal thermal cycler, with 3 minutes initial denaturation at 95°C, followed by 35 cycles of 30 seconds denaturation at 95°C, 30 seconds annealing at 59°C and 40 seconds extension at 72°C, followed by 3 minutes final extension at 72°C using Promega GoTaq polymerase (catalog number M300B). PCR products were electrophoresed on a 2% agarose gel, stained with ethidium bromide and observed and photographed with a UVP transilluminator. IBP and 18S bands of the expected sizes were obtained (see Underlying data (Raymond, 2020c)), cleaned up with a Nucleospin gel and PCR clean-up kit (Machery Nagel, catalog number 740609) and sequenced in both directions at the UNLV Genomics Core with an Applied Biosystems 3130 sequencer. The primers that were used for sequencing are shown in Table 1. the Leymus IRI domain based on the template. We selected the one with the highest sequence identity (76.5%), which also was the only one that had eight orderly coils like those in the L. perenne template. The free energy of this structure was then minimized (from -42.56 to -52.58 MJ mol -1 ) with the YASARA Energy Minimization Server (Krieger et al., 2009) and displayed with YASARA View. Stereoviews were obtained by rotating the molecule around its vertical axis by 3°. The average distance between coils in the IRI domain was measured with the YASARA distance function as the distance between the alpha carbons of Ser15 and Ser116 divided by seven.

Species identification
The obtained 18S rRNA sequence (GenBank accession no. MT506010) most closely matched the sequence of Leymus mollis (GenBank accession no. EF581964) (98.3% identity). Among the Pooidae, L. mollis is a member of the Triticodae, which includes wheat (Triticum) and barley (Hordeum), while Lolium and Deschampsia are members of the Poodae. The grass was confirmed as L. mollis from photos of the remnant spikes and leaves by Matthew Carlson (University of Alaska Anchorage) and Carolyn Parker, University of Alaska Fairbanks (personal communications to T.S.).

Ice-structuring and recrystallization inhibition activities
An extract from green rhizome tissue strongly affected the growth of an ice seed crystal, causing it to develop sharp facets ( Figure 1A). The facets are an indication of the presence of ice-binding proteins. The facets themselves do not have any role in freezing tolerance, nor are they likely to form in frozen tissue; they are only a demonstration that molecules are binding to the ice surface, which is necessary for IRI. In constrast, ice grown in the presence of extract from lawn grass stem tissue (control) is smooth and rounded with no evidence of facet development ( Figure 1B). The Leymus extract also had strong IRI activity. Polycrystalline ice formed from water and the lawn grass extract recrystallized significantly after 21 hours at -3°C, while the supernatant of the Leymus extract showed no recrystallization ( Figure 2).

Ice-binding protein
Primers based on the IBPs of other pooid grasses succeeded in amplifying a sequence that encoded the C-terminal part of the IBP gene of L. mollis, including the entire IRI domain. The sequence (GenBank accession no. MT506011) encoded 260 a.a., which corresponds to all but the first 25 a.a. of the L. perenne IBP sequence. The sequence included the stop codon and contained no introns. Although the sequence was close to the sequence of L. perenne (65% in the region of overlap), it most closely matched an IBP-like protein from Triticum aestivum (QBE94480) (86% identity, 91% similarity), in agreement with L. mollis's classification as a member of the Triticodae. When only the IRI domains were compared, the Leymus IRI domain was also closer to the Triticum IRI domain than it was to the Lolium IRI domain. The ice-binding activities of Tritium spp. IBPs have not yet been reported.
The IRI domain of L. mollis, like that of L. perenne, has eight repeats, each consisting of 14 or 15 residues ( Figure 3A). The consensus sequences of the two IBPs are virtually the same ( Figure 3A).

Model of structure
The predicted structure of the L. mollis IBP IRI domain, obtained from using the L. perenne IBP IRI domain structure as template, resembles the Lolium IRI domain in several respects: it has eight repeats corresponding to eight coils in a beta-roll fold ( Figure 3B); a flat side (the a side) that is populated with two rows that are rich in Thr and Ser residues ( Figure 3B); an irregular side (the b side) that has a bulge in the first three coils ( Figure 3C); and an interior that is dominated by Asn/His  Ice seed crystals were placed in the extracts at slightly below the freezing point. Scale bars, 1 mm.

Figure 2. Comparison of ice recrystallization in 3 µl drops of water and extracts from lawn grass (Festuca) and L. mollis (Leymus).
Scale bar, 1 mm. Figure 3A). The a side has been identified as the ice-binding site of L. perenne IBP (Middleton et al., 2012). In Lolium, the average spacing between the alpha carbons of the row of amino acids on the ice-binding site is 4.5 Å, which is very similar to the spacing along the a-axis of ice (4.51 Å). In Leymus, the average spacing was calculated as 4.77 Å. The regular array of hydrophilic residues is thought to order water molecules that act as a glue between the IBP and the ice (Middleton et al., 2012).

ladders (red residues in
In summary, we describe an ice-binding protein from the grass Leymus mollis with ice-structuring and ice recrystallization inhibition (IRI) activities. The collection site of the grass, a gravelly beach on the Chukchi Sea, is an extreme habitat subject to numerous environmental stresses. L. mollis appears to be better adapted to high latitudes than L. perenne, which does not grow well north of 60°N (Helgadóttir et al., 2018). It seems unlikely that the greater cold hardiness of Leymus can be attributed to its IBP as the IBPs of the two grasses are very similar. However, many proteins contribute to the cold hardiness of forage grasses, including transcription factors, cell membrane proteins, and proteins involved in cell signalling, cellular transport and photosynthesis (Sandve et al., 2011). It is thus likely that many of these factors are responsible for L. mollis's survival in the high Arctic. One approach to identifying them would be a comparison of transcriptomes of the Arctic population with a California population that never experiences freezing conditions.  The authors have observed ice structuring, recrystallisation inhibition activity in aqueous homogenates of the rhizome tissue. They have sequenced DNA from similar homogenates using primers based upon the conserved sequences from other IBP containing grasses. The obtained DNA sequences were translated into the amino acid sequences presumably involved in the ice recrystallisation inhibition and ice structuring. The amino acid sequence of the IRI domain of L. mollis was closely matched to the Triticum aestivum IRI domain.

○
The 3D structure of the ice binding domain was predicted using the obtained sequence and the structure of the IRI domain in Lolium perenne. It appears that the tree dimensional structure of the ice binding domain of L. mollis IBP quite resembles that of L. perenne.

○
The paper is an interesting comparative account of the ice related properties of a grass found at high latitudes -again indicating that ice structuring and recrystallisation inhibition is very important in plants adapted to cold climates. ○ I miss an extension of the discussion of the comparative aspects of cold tolerance and adaptive aspects of the findings in the paper.
○ Figure 1 could be explained a bit better as some readers may not understand the differences between the A and the B panel very well.
○ Figure 2 is not very clear and also this figure would benefit from a bit more explanation in the text.

Is the work clearly and accurately presented and does it cite the current literature?
Yes

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Author Response 30 Jul 2020 James Raymond, University of Nevada Las Vegas, Las Vegas, USA • I miss an extension of the discussion of the comparative aspects of cold tolerance and adaptive aspects of the findings in the paper.
We revised the last paragraph to give some perspective to the cold tolerance of L. mollis and how it probably differs from that of L. perenne.
• Figure 1 could be explained a bit better as some readers may not understand the differences between the A and the B panel very well.
We added some sentences to better explain the difference between Leymus and Festuca.
• Figure 2 is not very clear and also this figure would benefit from a bit more explanation in the text.
We added new photos to show RI of Festuca. It more clearly shows that Leymus has much stronger RI than Festuca.

Competing Interests:
The authors declare no competing interest.
Reviewer Report 23 July 2020 Yes Are the conclusions drawn adequately supported by the results? Yes to model the three-dimensional structure of the domain, which looks very much like that of Lollium perenne. Another set of primers was used to amplify a diagnostic section of the 18S ribosomal RNA gene whose sequence identified the species.
There is an element of comparative science here, and one would not need to see a repeat of this detailed analysis on other grasses/cereals. The presence of sequence homologs in transcriptomes as a way of seeing the distribution and variation of this IBP type should suffice from here on. Nevertheless, it is interesting to see this grass from a high latitude in Alaska (Beaufort Sea coast) extracted and analyzed. This report is also a good illustration of the use of molecular modeling to avoid solving a structure that would not differ much from the template. Indeed, the main features of the Lolium perenne IBP fold are retained in the L. mollis model, including the outward bulge of the beta-solenoid at one end on the non-ice-binding side.

Suggestions for improvement:
The Abstract should include the percent identity of the IRI domain between the two grass species. This type of information should also be provided in the Introduction where the sequences of the IBPs of L. perenne and D. antarctica are compared. It is not helpful to say their sequences "are very similar". In the last paragraph of the Introduction the L. mollis IBP sequence is again said to be "very similar" to those of the other two grasses. Please provide % identity and % similarity.
○ Figure 1A shows ice structuring. However, since panel B also shows ice crystals it is important to point out to the average reader that the distinction lies in the faceting seen in A where flat, angular surfaces are the result of IBP binding to ice. The ice in B is smoother and rounded.
○ Figure 2 is a rather poor presentation of ice recrystallization. One must look closely at the ice control images to see the grains are larger after 21 h. The methods say the ice recrystallization was done in water. From our experience the inclusion of 150 mM NaCl makes a huge difference to the visibility of the ice grains. The other suggestion is to take the starting picture at zero time with the ice held colder than -3 °C. We see substantial recrystallization happening in the first hour, even at -6 °C, and will typically hold the ice wafer at -10 °C until the starting picture is captured, and then raise the temperature for more rapid recrystallization. I recommend that this analysis be repeated and improved.

○
In Methods, under 'Structure prediction', it is stated that three structures were proposed by SWISS MODEL operating with L. perenne IBP as a template. Then, the one with the 'highest sequence identity' was selected. Surely the sequence identity is the same for all models because they are based on the same template. We would typically compare models by root mean square deviation (RMSD) -how the atoms deviate from the model on average. Is this what is meant here? ○ Although the sequence of the L. mollis open reading frame was incomplete, it would be useful to report the sequence upstream of the ice-binding domain. Presumably this is a series of leucine-rich repeats similar to those seen in other grass IBPs? ○ In reference to ice crystal axes the a-and c-should be italicized.
○ other grasses (for use as a forward primer) but we were unable to find such a sequence.

Competing Interests: the authors declare no competing interests
Author Response 28 Jul 2020 James Raymond, University of Nevada Las Vegas, Las Vegas, USA Percent identity was added to Abstract and the Results. Identities were removed from Introduction to make a revision requested by reviewer 2.  Swiss-Model: Swiss Model tries different regions of the sequence to model, e.g., residues 131-247 in one model and residues 144-228 in another. Each gives a slightly different model. We selected the one that had the highest a.a. sequence identity.
N-terminal region of IBP: For this, we would need a consensus sequence of the 5' UTR in other grasses (for use as a forward primer) but we were unable to find such a sequence.

Competing Interests: The authors declare no competing interests
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com