SEARCH

SEARCH BY CITATION

Keywords:

  • abscisic acid (ABA);
  • Arabidopsis;
  • BOTRYTIS SENSITIVE1 ;
  • cell death;
  • lesion mimic;
  • MYB108 ;
  • reactive oxygen species (ROS);
  • wounding

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Wounding results in the controlled cell death of a few rows of cells adjacent to disrupted cells resulting in physical wound closure, which combined with phenolic compound deposition, prevents water loss and pathogen entry. The control of these processes remains uncharacterized.
  • Cell death in a mutant of Arabidopsis thaliana lacking BOTRYTIS SENSITIVE1/MYB108 (BOS1/MYB108) function was characterized utilizing physiological, cell biological and genetic methods.
  • The bos1 mutant has a wound induced runaway cell death that includes enhanced reactive oxygen species (ROS) production that followed the extent of enhanced cell death. Exogenous abscisic acid (ABA) enhanced wound induced cell death in Col-0 plants and was sufficient to trigger cell death in bos1. Uncontrolled cell death was dependent of the production and perception of ABA. Furthermore, bos1 had altered sensitivity to and accumulation of ABA.
  • Arabidopsis possesses a genetic program controlling the extent of wound inducible cell death. BOS1 acts as a negative regulator of ABA induced cell death, which functions in the control of this wound sealing program. This program is distinct from other known cell death programs in that it is ABA dependent, but independent of salicylate biosynthesis, ethylene, jasmonate, metacaspases and ROS derived from RBOHD and RBOHF.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Wounding is a mechanical injury involving cellular rupture, loss of compartmentalization and disruption of outer protective layers, which requires wound closure to prevent water loss and pathogen entry. Indeed some wound and pathogen responses are overlapping. For example, the stress hormones abscisic acid (ABA), jasmonic acid (JA) and ethylene are common regulators of both pathogen and wound responses (Bari & Jones, 2009). The anatomy of wound closure sites is well characterized (Bostock & Stermer, 1989); in most dicot leaves, cell death occurs in the cells immediately adjacent to the wound and these cells are sealed with lignin and other phenolic compounds, but no meristematic activity is present under the wound, as is the case in other plant wound morphology types. The control of this wound induced cell death response remains uncharacterized. Previous knowledge of mechanisms controlling the propagation and limitation of cell death between dying cells and adjacent cells is substantially derived from the study of spontaneous cell death mutants like lesions simulating disease resistance1 (lsd1), accelerated cell death1 (acd1), acd11, defense no death1, among others (Lorrain et al., 2003). However, the relevance of these known cell death regulatory circuits to the wound response also remains unknown. Here we report that a mutant deficient in the transcription factor MYB108 exhibited mis-regulated cell death spreading from wound sites, giving insight to the genetic program controlling wound induced cell death.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Biological materials and growth conditions

Arabidopsis (Arabidopsis thaliana (L.) Heynh) seeds were sown in a mixture of soil and vermiculite (1 : 1), stratified 2 d at 4°C and cultivated in a growth room under 12 h : 12 h light  :  dark regime, 23 : 18°C and relative humidity 70 : 90%, with a light intensity of 160–220 μmol m−2 s−1. Three to six-week old plants, grown four per 8-cm pot, were used for all studies. Seedings for in vitro assays were grown in a controlled environment chamber (Sanyo, Sakata, Japan) under conditions of a 12 h : 12 h light : dark regime with constant 21°C and light intensity of 220 μmol m−2 s−1.

Cell death measurements

Cell death was measured using chlorophyll retention assays, ion leakage and lesion diameter. For chlorophyll retention, four fully expanded leaves per plant were cut with a scalpel blade at 3 mm intervals. Six plants for each treatment were harvested at the indicated time in days post wounding (dpw), pigments extracted into 80% acetone and absorbance of the extract measured at 645 and 663 nm using a spectrophotometer (Agilent 8453; Agilent Technologies, Waldbronn, Germany). The total chlorophyll concentration was calculated using Arnon's (1949) equation: total chlorophyll (g l−1) = 0.0202 A645 + 0.00802 A663 with leaf chlorophyll content expressed in μg mg l−1 FW. Ion leakage was measured from rosette leaves from six plants for each treatment, wounded by crushing with forceps at 3 mm intervals and immersed in 5 ml of MQ water for 4 h. Conductivity (in μS cm−2) was measured with a Mettler conductivity meter (Mettler Toledo GmbH, Greifensee, Switzerland). Lesion diameter was determined from leaves punctured with a needle, photographed at the indicated time in dpw with a size standard and measured with ImageJ (http://rsbweb.nih.gov/ij/).

Reactive oxygen species (ROS) and phenolic stains

Rosette leaves were wounded and then stained with 3,3′-diaminobenzidine (DAB) and nitroblue tetrazolium (NBT) to visualize reactive oxygen species (ROS) production, and with trypan blue to visualize plant cell death according to Jabs et al. (1996) and Torres et al. (2002). The callose and phenolic compound were stained according to Adam & Somerville (1996). Tissues were examined with a microscope (Leica DML; Leica Microsystems, Wetzlar, Germany) under bright light or 365 nm ultraviolet (UV) light.

Phytohormone treatment

Leaves were detached, punctured and immediately put in ¼ MS liquid solution supplemented with indicated concentration hormones or solvent control in plates covered to prevent drying under standard growth conditions described earlier. ABA treatments of attached leaves were prepared with six 5 μl droplets of 200 μM ABA or solvent control placed on the surfaces of punctured or intact control growing leaves in covered trays to prevent drying. Photographs were taken at the indicated time for further measurements. Seeding growth assays were performed with sterilized seeds on 0.5× MS plates with 1% agar, 1% sucrose and supplemented with 0.2 μM ABA or solvent control. After 3 d stratification, the plates were placed in chambers and seedlings photographed after 1 wk of growth. Since germinating seeds on ABA plates alters both germination and root growth, a second root growth only assay was performed. Four-day old plants were transferred from normal MS plate to 50 μM ABA or control plate and photographed 5 d later. Root lengths were measured with ImageJ.

ABA concentration measurement

Four fully expanded leaves of 5-wk old plants were cut with scissors at 3 mm intervals. Five biological replicates of each treatment were harvested at 24 h post-wounding, and ground to a fine powder in liquid nitrogen. Extractions were done according to Müller & Munné-Bosch (2011). Extracted plant samples were analyzed by Waters Acquity UPLC – Synapt G2 HDMS mass spectrometer (Waters, Milford MA, USA). The hormones were separated on an Acquity UPLC BEH C18 column (1.7 μm, 50 × 2.1 mm; Waters, Wexford, Ireland) at 40°C and analyzed in electrospray negative ion mode. Mass range was set to 100–600. The mobile phase consisted of H2O and acetonitrile (Chromasolv grade; Sigma-Aldrich, Steinheim, Germany) both containing 0.1% HCOOH (Sigma-Aldrich). A linear gradient was used starting from 95% of A to 30% in 6 min, then back to 95% in 6.1 min and left to equilibrate for 1 min. The injection volume was 2 μl and flow-rate of the mobile phase was at 0.6 ml min−1.

Double mutant construction

All mutants were in Col-0 background (Supporting Information Table S1 including the methods for genotyping). All studies were performed with F2 or F3 generation plants with double mutants confirmed at every stage of crossing by phenotyping or PCR genotyping.

Statistics

All the experiments were repeated at least three times. Student's t-test was used in all the data analysis. All error bars represent one standard error (SE) of the means. The asterisk or letters above the bars indicate significant difference (< 0.01).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Here we report that a mutant deficient in the MYB108 transcription factor, botrytis sensitive1 (bos1; Mengiste et al., 2003), but not Col-0 wild type, exhibited mis-regulated cell death spreading from wound sites made by cutting with a blade, crushing with forceps or needle puncture (Fig. 1). Visible chlorosis adjacent to wounds of bos1 leaves (Fig. 1a) was manifest 3 dpw and spread by 5 dpw, as quantified by chlorophyll loss (Fig. 1b). Symptoms advanced and cell death occurred, as measured by ion leakage (Fig. 1c), resulting in dead dry tissue by 5 dpw. Lesion propagation was unaltered in the dark and high humidity (Fig. 1d), suggesting it is not light or desiccation dependent. Beyond 5 dpw, cell death advanced slowly but continuously, consuming the entire leaf and spreading further to other unwounded leaves (Fig. 1e) via the petioles (Fig. 1f) through the rosette, ultimately resulting in the death of the plant within several weeks (Fig. 1e). In Col-0, cell death was limited to the cells adjacent to wounds (Fig. 1a). We conclude that BOS1 acts as a negative cell death regulator limiting spreading cell death to cells adjacent to wounds.

image

Figure 1. Cell death spread from wound sites in the Arabidopsis thaliana mutant bos1. Cell death spreading from a scalpel blade inflicted wound, measured at the indicated time in days post wounding (dpw) measured by: (a) photographs taken at 1, 3 and 5 dpw; (b) chlorophyll retention assay; (c) cell death quantified as ion leakage. Letters above error bars (± SE of means, = 6) represent statistical significance (Student's t-test, < 0.01). (d) Cell death spread from forceps pinching inflicted wound in bos1 leaves also occurred when leaves were placed in the dark or under high humidity conditions immediately after wounding and remained there continuously until photographed. Representative photographs at 5 dpw are shown. (e) Cell death spread from the wounding sites from the excision of four leaves advances through rosette core into unwounded leaves. This process takes 3–5 wk post-wounding (wpw), photograph taken at 4 wpw. (f) Photograph depicting cell death spread into petioles 7 d after leaf excision: Col-0 petioles remain healthy (left; with six leaves removed) while bos1 exhibited cell death from leaf excisions wounds had spread into the petioles (right; with five leaves removed). Bars, 1 cm.

Download figure to PowerPoint

In order to address if bos1 was deficient in wound sealing defensive compounds, such as callose or lignin, their wound induced deposition was observed. Initially, both callose and autofluorescent phenolic compound deposition was similar in bos1 and Col-0 at 1 dpw (Fig. 2a,c). Subsequently, accumulation advanced in a circle with spreading cell death around the puncture wound in bos1 but was restricted to a few cells adjacent to the wound in Col-0 at 5 dpw (Fig. 2b,d). We conclude that bos1 plants executed a full uncompromised but at the same time unrestricted wound sealing response.

image

Figure 2. Reactive oxygen species (ROS) accumulation and deposition of phenolics in Arabidopsis thaliana leaves. Bright field (left) and UV (right) micrographs showing accumulation of callose (a, b) and autofluorescent phenolic compounds (c, d) around a scalpel blade cut (a) or needle puncture (b–d) at 1 and 5 d post wounding (dpw). (e, f) ROS visualized in forceps pinch or blade cut induced wounds: H2O2 by 3, 3′-diaminobenzidine (DAB; left) and superoxide anion by nitroblue tetrazolium (NBT; right) staining assays at 1 min post wounding (e) and 5 dpw (f). (g) Cell death visualized by trypan blue staining (TB). Bars, 500 μm.

Download figure to PowerPoint

To further define the enhanced ROS accumulation seen in bos1 (Mengiste et al., 2003; Kraepiel et al., 2011), wound induced H2O2 and superoxide distribution was monitored. At early time points there was no difference in ROS staining between bos1 and Col-0 (Fig. 2e), indicating BOS1 does not regulate initial wound-induced ROS production. At 5 or more dpw, ROS accumulation remained restricted to the wounding sites in Col-0, while in bos1 enhanced ROS accumulation advanced in a circle (Fig. 2f) together with progressing cell death, visualized with trypan blue staining (Fig. 2g). Taken together with our results earlier this suggests that prolonged ROS production may be associated with wound sealing rather than a cause of cell death.

To delineate signaling pathways involved in bos1 spreading cell death, wounds were treated with exogenous stress hormones and double mutants of several hormone or defense signaling mutants in the bos1 background were used to address the genetic requirements for spreading cell death. The stress hormones, salicylic acid (SA), ethylene and JA had no effect on the propagation of cell death in both exogenous feeding and genetic double mutant assays (Fig. 3a–c). Hormones traditionally studied in the context of development also play roles in plant defenses (Robert-Seilaniantz et al., 2011). However, both auxin and 24-epibrassinolide failed to restrict the wound induced death spread in bos1 (Fig. 3d). ROS produced via the NADPH oxidases, RESPIRATORY BURST OXIDASE HOMOLOGS (RBOH), are regulators of HR-like spreading cell death (Torres et al., 2005) prompting us to construct bos1 rbohD, bos1 rbohF and bos1 RbohD-overexpression double mutants. No significant difference in the lesion development was observed in these double mutants (Fig. 3c). Metacaspases are executors of programmed cell death in plants (Bozhkov et al., 2010). The hypothesis that BOS1 regulates metacaspase (mc) dependent cell death was tested with double mutants of bos1 to mc1, mc4, mc5, mc6, mc7, mc8 and mc9; none of which could block the bos1 spreading cell death. The lesion sizes of two representative metacaspase double mutants, bos1 mc1 and bos1 mc8, were quantified and they exhibited no difference with the bos1 single mutant (Fig. 3c; mc4, mc 5, mc 6, mc7, mc8, mc9, data not shown). These results demonstrate RBOHD and RBOHF and metacaspases were not required for spreading cell death in the bos1 mutant.

image

Figure 3. Exogenous hormone feeding and genetic analysis. (a, b) Stress hormone in vitro treatments. Needle punctured detached leaves of Arabidopsis thaliana were supplied with stress hormones as indicated and phenotypes monitored in (a) the chlorophyll retention assay at 3 d post wounding (dpw) or (b) visually in photographs at 2 dpw. (c) Genetic requirements of bos1 spreading cell death: Double mutants of hormone and defense mutants in the bos1 background were wounded by needle puncture, photographed and lesions quantified at 5 dpw; (top) photographs of representative lesion phenotypes, (bottom) quantification of lesion size (in mm). (d) Needle punctured detached leaves supplied with 20 μM auxin or 10 μM 24-epibrassinolide. Lesions photographed and measured at 3 dpw demonstrated these growth hormones had no effects on the spreading cell death in bos1 leaves. (e, f) Exogenous abscisic acid (ABA) treatment of attached leaves: Six droplets of 5 μl 200 μM ABA were put on the leaves and at 5 dpw lesion sizes were quantified of both Col-0 and bos1 with wounding (e) or without wounding (f). Error bars represent standard error (SE) of means (= 24). Asterisks and letters above the bars denote < 0.01 (Student's t-test). Bars, 1 cm. Abbreviations: Col-0, Columbia 0 ecotype; bos1, botrytis sensitive1; aba2, abscisic acid deficient2; aba3, abscisic acid deficient3; abi1, aba insensitive1; abi4, aba insensitive4; sid2, salicylic acid induction deficient2; ein2, ethylene insensitive2; eto1, ethylene overproducer1; coi1, coronotine insensitive1; jar1, jasmonate resistant1; RBD-oe, RESPIRATORY BURST OXIDASE HOMOLOG D – over expression line; rbohD, respiratory burst oxidase homologD; rbohF, respiratory burst oxidase homologF; mc1, metacaspase1; mc8, metacaspase8.

Download figure to PowerPoint

Only exogenous ABA significantly enhanced cell death (Fig. 3a,b,e,f). In bos1 both chlorophyll loss (Fig. 3a) and visible chlorotic cell death symptoms were enhanced by ABA (Fig. 3b). Enhanced cell death occurred in circles surrounding the wound with 10 μM ABA, and consumed nearly the whole leaf at 50 μM ABA (Fig. 3b). ABA enhanced wound associated cell death in both bos1 and Col-0 wild type attached leaves (Fig. 3e). Since ABA enhanced wound-induced cell death, the sufficiency of ABA alone to trigger cell death was tested revealing that 200 μM ABA droplets triggered cell death in bos1 and to a much lesser extent also in Col-0 (Fig. 3f). The ABA deficient loci ABA2, ABA3 and ABA insensitive ABI1, but not the ABA responsive transcription factor ABI4, were required for the full bos1 spreading lesion phenotype (Fig. 3c). Thus, ABA was necessary and sufficient for spreading cell death in bos1.

Experiments earlier indicated that exogenous ABA enhanced wound associated cell death in both bos1 and Col-0 (Fig. 3). This suggests that BOS1 may regulate ABA accumulation or sensitivity, the latter of which can be assessed via seed germination and root growth assays on plates with ABA. Germination of bos1 on ABA was delayed resulting in a shorter root than that of Col-0 (Fig. 4a; left). Root growth assays after transfer to ABA plates also demonstrated enhanced growth inhibition by ABA in bos1 (Fig. 4a; right). In addition to reduced root growth, bos1 leaves were bleached in seedlings transferred to 50 μM ABA plates (Fig. 4b; bottom). We conclude that bos1 is hypersensitive to growth inhibition by ABA. As ABA production increases after wounding (Hildmann et al., 1992), wound induced ABA content in bos1 was measured. ABA concentration significantly increased in Col-0 upon wounding. However, in bos1 ABA was already elevated under control conditions to levels seen in Col-0 after wounding but did not further increase upon wounding (Fig. 4c). This suggests that BOS1 negatively regulates production of ABA under control but not wound induced conditions.

image

Figure 4. Altered abscisic acid (ABA) response in the Arabidopsis thaliana mutant bos1. (a, b) ABA sensitivity assay: germination on 0.2 μM ABA plates and root growth after transfer from control to 50 μM ABA plates, both photographed (b) and quantified as root length in mm (a). The bos1 mutant exhibited leaf bleaching after being transferred to 50 μM ABA plates (b; bottom). Bars, 1 cm. (c) ABA concentration assay: control and wound-induced ABA concentration determined at 1 d post wounding (dpw). Error bars represent standard error (SE) of means. Asterisks and letters above the bars denote < 0.01 (Student's t-test).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The analysis of wound-induced cell death using the bos1 mutant provides evidence of a genetically regulated developmental cell death program that determines the extent of cell death following wounding. BOS1 acts in the control of this program, likely as a negative regulator. Accumulation of polymerized phenolic compounds, such as callose, suberin or lignin, is considered the first physical and antimicrobial chemical barrier at the wounding and pathogen infection sites in plants (Bostock & Stermer, 1989; Jacobs et al., 2003). The accumulation of these wound sealing compounds in cells undergoing spreading cell death in the bos1 mutant indicates that cell death and wound sealing are coupled in this mis-regulated wound response. The prolonged accumulation of ROS while wound sealing is occurring suggests this ROS may be related to the production of polymerized phenolic compounds, such as lignin, rather than a regulator of cell death. However, further studies are required to confirm this.

Lesion mimic mutants ectopically activate pathogen induced cell death and defense programs in the absence of pathogens and have contributed significantly to our understanding of cell death control. These mutants can be classified into initiation mutants, which ectopically initiate cell death that is contained in defined lesions, and propagation mutants, which result in uncontained ‘runaway’ cell death. The bos1 mutant represents a novel type of ectopic cell death mutant with some similarities, but also key differences, to propagation class lesion mimics. The stress hormones, SA, ethylene and JA had no effect on cell death spreading in both exogenous feeding and genetic double mutant assays (Fig. 3a,c). Therefore, the bos1 spreading cell death type was distinct from the SA-dependent cell death, such as in lsd1 and acd1 (Lorrain et al., 2003). However, since only the sid2 mutant, deficient in the main SA biosynthesis enzyme isochorismate synthase, was used, we cannot exclude that other components in downstream SA signaling may be involved in bos1 regulated cell death. RBOHD and RBOHF and metacaspases were not required for spreading cell death in bos1, further distinguishing its cell death regulation from that of other well characterized lesion mimics, such as lsd1 (Torres et al., 2005; Coll et al., 2010).

The exclusion of ROS derived from RBOHD and RBOHF as regulators of wound induced cell death suggests other ROS sources function in this process. In tomato RBOHs were required for systemic wound signaling (Sagi et al., 2004), suggesting wound induced cell death and systemic signaling utilize distinct ROS sources. In addition to RBOHs, candidate wound induced ROS sources include peroxidases (Cheong et al., 2002) and polyamine oxidases (Angelini et al., 2008).

The bos1 seedlings exhibited enhanced growth inhibition by ABA, indicating ABA hypersensitivity in this mutant. Alternatively, activation of cell death programs also inhibits plant growth; thus loss of an inhibitor of ABA induced cell death would also enhance growth inhibition by ABA without changing ABA sensitivity, per se. ABA levels were also elevated in the control, but not wounded bos1, suggesting BOS1 negatively regulates ABA accumulation in unwounded leaf tissue. We conclude that BOS1 acts as a negative regulator of both ABA biosynthesis and ABA induced cell death. Expression of BOS1/MYB108 is induced by JA (Mengiste et al., 2003); this together with our results earlier suggest this transcription factor may act to integrate signals from several hormones. BOTRYTIS SUSCEPTIBLE1 INTERACTOR (BOI1) is a RING E3 ligase shown to interact with and target BOS1 for proteasome mediated degradation (Luo et al., 2010). BOI1 RNAi lines exhibited enhanced Botrytis susceptibility and altered gibberellin (GA) response. A mutually antagonistic hormone interaction between GA and ABA is well documented in multiple plant tissues (Weiss & Ori, 2007). Taken together with our ABA results, this suggests that BOS1 may be involved in an interaction between these two hormones. BOS1/MYB108 is also involved in the regulation of JA-mediated stamen maturation (Mandaokar & Browse, 2009). However, it remains to be seen what similarities, if any, this developmental process has with MYB108 wound and pathogen responses.

The studies presented here have implications for other phenotypes observed in the bos1 mutant. BOS1 was first characterized in a screen for fungal susceptibility (Mengiste et al., 2003). Necrotrophic pathogens such as Botrytis thrive on dead tissue, suggesting that in bos1, loss of cell death control and resulting spreading cell death, rather than deficient defense signaling, may be the mechanism responsible for its extreme susceptibility to necrotrophic pathogens. Accordingly, bos1 had enhanced symptom development, but unaltered resistance to biotrophic pathogens. The abiotic and ROS induced stresses previously tested would all promote ABA biosynthesis. As a consequence, altered cell death and growth responses seen in these stress responses may result from enhanced ABA-induced cell death. However, further work will be required to test these predictions.

In summary, we characterized a new type of spreading cell death triggered by wounding and negatively regulated by BOS1/MYB108. The cell death is phenotypically distinct from other spontaneous cell death mutants and dependent on ABA but not on other known wounding signals (JA and ethylene), SA biosynthesis, metacaspases or ROS produced via RBOHs. Elucidation of the signals responsible for cell death spreading in bos1 may significantly contribute to our understanding of cell death control and wound response mechanisms in plants.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The authors thank Leena Grönholm for excellent technical assistance and the following colleagues for the gift of mutant seeds: Tesfaye Mengiste for bos1, Miguel Torres for rboh mutants, Hannele Tuominen for metacaspase mutants, Julian Schroeder for abi1 in Col-0 and the NASC for remaining mutants used. F.C. and K.O. are associated with the Finnish Doctoral Program in Plant Biology (FDPPB). These studies were supported by the Chinese Scholarship Council (F.C., S.T.), the University of Helsinki research allocation (K.O.) and Academy of Finland Fellowship Program (decision no. 251397 and 256073 to K.O.).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

FilenameFormatSizeDescription
nph12456-sup-0001-TableS1.pdfapplication/PDF103KTable S1 Genotyping information for mutants used