MUR1‐mediated cell‐wall fucosylation is required for freezing tolerance in Arabidopsis thaliana

Summary Forward genetic screens play a key role in the identification of genes contributing to plant stress tolerance. Using a screen for freezing sensitivity, we have identified a novel freezing tolerance gene, SENSITIVE‐TO‐FREEZING8, in Arabidopsis thaliana. We identified SFR8 using recombination‐based mapping and whole‐genome sequencing. As SFR8 was predicted to have an effect on cell wall composition, we used GC‐MS and polyacrylamide gel electrophoresis to measure cell‐wall fucose and boron (B)‐dependent dimerization of the cell‐wall pectic domain rhamnogalacturonan II (RGII) in planta. After treatments to promote borate‐bridging of RGII, we assessed freeze‐induced damage in wild‐type and sfr8 plants by measuring electrolyte leakage from freeze‐thawed leaf discs. We mapped the sfr8 mutation to MUR1, a gene encoding the fucose biosynthetic enzyme GDP‐d‐mannose‐4,6‐dehydratase. sfr8 cell walls exhibited low cell‐wall fucose levels and reduced RGII bridging. Freezing sensitivity of sfr8 mutants was ameliorated by B supplementation, which can restore RGII dimerization. B transport mutants with reduced RGII dimerization were also freezing‐sensitive. Our research identifies a role for the structure and composition of the plant primary cell wall in determining basal plant freezing tolerance and highlights the specific importance of fucosylation, most likely through its effect on the ability of RGII pectin to dimerize.

. Expression of CBF1-3 and the CBF target genes KIN2 and GOLS3 are all expressed to normal wild type levels in sfr8.           S11. sfr8 is more sensitive to freezing than WT even without cold acclimation. Table S1. Candidate SNPs identified using Galaxy.   S1. Expression of CBF1-3 and the CBF target genes KIN2 and GOLS3 are all expressed to normal wild type levels in sfr8.
Relative Quantitation (RQ) of transcripts using qRT-PCR in 8-day-old seedlings exposed to cold (5°C) or ambient (20°C) temperature. (a) CBF1-3 transcripts after 2 h of exposure, (b) KIN2 and (c) GOLS3 after 6 h of exposure. Expression is shown after normalisation to PEX4. Values were calculated using the ΔΔCT method and the error bars represent RQMIN and RQMAX and constitute the acceptable error level for a 95% confidence level according to Student's t-test. Error bars indicate the level of variation between technical replicates within one biological replicate experiment. Data are representative of 2 biological replicate experiments. Primers with the following sequences were used to detect transcripts as (a) Cold-acclimated Col-0 wild type and two insertional mutants one week after a 24 h freezing treatment at -8.5°C.
(b) Both mutants showed reduced levels of At3g50910 transcript compared to wild type. qRT-PCR data showing Relative Quantitation (RQ) of At3g50910 transcripts in the two mutants compared to Col-0 wild type. Error bars present RQMIN and RQMAX for three technical replicate measurements as detailed above in the legend for Figure S1.  . S4. sfr8 fails to convert GDP-mannose to GDP-fucose.

Protein 1 M A S E N N G S R S D S E S I T A P K
The conversion of radiolabelled GDP-mannose to GDP-fucose was assayed using the TLC-based method described by Bonin et. al. 1997. Arrows mark the positions of GDP-fucose and GDP-mannose standards. Extracts from wild type (WT) plants and two other freezing sensitive mutants (sfr4 and sfr5) catalysed the conversion of GDP-mannose substrate to GDP-fucose after 70 minutes. Extract from sfr8 did not.

Mobile
Phase GDP-Mannose GDP -Fucose (a) Freeze-induced damage is greater in sfr8 than wild type but is restored in two additional complemented lines. Percentage electrolyte leakage from leaf discs of Col-0 wild type (WT) plants, sfr8 and complemented lines 1 (L1) and 8 (L8) after freezing at -6, -8 or -10 °C. Complemented lines 1 and 8, like complemented line 14 (referred to as sfr8-C throughout the main text), show restored levels of freezing tolerance. Results from a single experiment are presented; each data point corresponds to the mean of 6 pseudo-biological replicate samples per genotype comprising 3 leaf discs per pseudoreplicate. Error bars represent +/-1 SEM calculated from arcsine-transformed data. Percentage electrolyte leakage values from leaf discs of Col-0 wild type (WT), sfr8 (a) and mur1-1 (b) plants after freezing at -7, -9.5 or -12°C. Plants were grown and cold acclimated as described in the main text and plants sprayed once a week with 60 ml of 10 mM fucose or water during both the growth and cold acclimation periods. Results from a single experiment are presented; each data point corresponds to the mean of 6 pseudo-biological replicate samples per genotype/treatment comprising 3 leaf discs per pseudo-replicate. Error bars represent +/-1 SEM calculated from arcsine-transformed data. Percentage electrolyte leakage in Col-0 wild type (WT) and mur2 plants after freezing to -7, -9.5 or -12°C. Plants were grown and cold acclimated as described in the main text. Each data point represents the average of three separate biological replicate experiments. Each experiment used six pseudo-biological replicate samples per genotype comprising three leaf discs per pseudo-replicate. Error bars represent +/-1 SEM calculated from arcsine-transformed data. Arcsine-transformed percentage leakage data were analysed by a least-squares means comparison at each temperature point (*, P < 0.05). A small increase in electrolyte leakage was observed at -7°C only (P = 0.047). Col-0 wild type (WT) and mur1-1 plants at five weeks old after watering with deionised water (-BA) or 20 mg/l boric acid (+BA). Boric acid (BA)-supplemented mur1-1 plants show restoration of wild type leaf shape from cup-shaped to spatulate and restoration of petiole length.

Fig. S10. MUR1 is not inducible by cold.
(a) Relative quantitation (RQ) by qRT-PCR of MUR1 and KIN2 (a known cold-inducible CBF-target gene) transcripts after 1, 3 ,6, 12, 24, 96 or 168 h at 20°C (white bars) or 5°C (grey bars). Expression is shown after normalisation to PEX4. Values were calculated using the ΔΔCT method and the error bars represent RQMIN and RQMAX and constitute the acceptable error level for a 95% confidence level according to Student's t-test. Error bars indicate the level of variation between technical replicates. KIN2 and PEX4 primer sequences are shown in Fig. S1; MUR1 primer sequences were: 5'-ACACCCAGCGAATCAACCAT-3' (forward) and 5'-CGATCCAACGACGGAGAGAG-3' (reverse).  Table S3 and show no significant differential expression between contrast groups.
(a) (b) Fig. S11. sfr8 is more sensitive to freezing than WT even without cold acclimation.
Freeze-induced damage is greater in sfr8 than wild type even in non-acclimated plants. Percentage electrolyte leakage from leaf discs of Col-0 wild type (WT) plants, sfr8 after freezing at -2, or -4°C. Plants were grown as described in the main text. Each data point represents the average of two separate biological replicate experiments. Each experiment used six pseudo-biological replicate samples per genotype comprising three leaf discs per pseudo-replicate. Error bars represent +/-1 SEM calculated from arcsine-transformed data. Arcsine-transformed percentage leakage data were analysed by a leastsquares means comparison at each temperature point (***, P < 0.001).