Edited by Mark Estelle, University of California, La Jolla, CA, and approved April 1, 2010 (received for review January 23, 2010)
With global warming, plant high temperature injury is becoming an increasingly serious problem. In wheat, barley, and various other commercially important crops, the early phase of anther development is especially susceptible to high temperatures. Activation of auxin biosynthesis with increased temperatures has been reported in certain plant tissues. In contrast, we here found that under high temperature conditions, endogenous auxin levels specifically decreased in the developing anthers of barley and Arabidopsis. In addition, expression of the YUCCA auxin biosynthesis genes was repressed by increasing temperatures. Application of auxin completely reversed male sterility in both plant species. These findings suggest that tissue-specific auxin reduction is the primary cause of high temperature injury, which leads to the abortion of pollen development. Thus, the application of auxin may help sustain steady yields of crops despite future climate change.
Plant reproduction processes are threatened by high temperature (HT) injury caused by recent global warming (1). Lobell and Field (2) have reported that, at least in the cases of wheat, maize, and barley, there is clearly a negative correlation between worldwide crop yields and increased temperatures, and for these crops, recent warming has resulted in an annual combined loss of ≈40 megatons or $5 billion. Previously, we used double-rowed barley to show that increasing temperatures principally influence the early phase of anther development, causing premature progression through meiosis of pollen mother cells and proliferation arrest and premature degradation of anther wall cells (3⇓–5). Complete male sterility can result from elevated temperatures for 4 days or longer when they occur during the early phase of anther development, because pollen grains abort (3, 4). Several morphological abnormalities can arise during HT injury, including mitochondrial swelling and vacuolization, and comprehensive alterations to transcription in anther wall cells (5). Male sterility caused by abortion of pollen development can be observed widely among other temperature-stressed plant species, such as wheat, tomato, cowpea, and Arabidopsis (6⇓⇓⇓–10). However, the molecular and physiological mechanism(s) underlying HT injury and method(s) for reversing such damage have not been fully identified.
Auxin is a phytohormone that orchestrates many physiological and developmental processes (11). It is also known that HT promotes auxin-mediated hypocotyls elongation in Arabidopsis (12). A tryptophan aminotransferase-encoding gene TAA1/TIR2 involved in one of several auxin biosynthetic pathways (13, 14) is required for the elongation (15). This gene expression is positively regulated with increased temperatures in hypocotyls, cotyledons, and root (15). Two Arabidopsis cytochrome P450s, CYP79B2 and CYP79B3, implicated in other pathway of tryptophan-dependent auxin biosynthesis via an intermediate indole-3-acetaldoxime (IAOx), also play a role in the hypocotyls elongation (16). In addition, expressions of YUCCA flavin monooxygenases implicated in IAOx biosynthesis with different pathway are temporally and spatially controlled in developing anthers (17⇓–19). In Arabidopsis, double or triple mutants that include yuc2 and yuc6, completely lose male fertility and form short stamens lacking pollen grains (17). Interestingly, these mutant phenotypes are quite similar to HT injury to male reproductive development, whereas HT generally increases auxin levels in certain other tissues. Here, we used barley and Arabidopsis to examine the effects of increasing temperatures on expression of endogenous auxin and YUCCA genes, as well as the application of exogenous auxin for reversing male sterility caused by HT.
To define the relationship between auxin and HT injury, we measured endogenous auxin levels and auxin signal transduction activity. In barley, we observed that the endogenous auxin levels of developing anthers are reduced in response to HT (Fig. 1). In untreated controls (5-mm-long panicles), auxin accumulated abundantly throughout the developing anther cells, i.e., parietal, epidermal, and sporogenous cells, and in rachis cells around vascular bundles (Fig. 1 B and D). This stage of development is just before development of tapetum, middle layer, and endothecial cells from the parietal cells (4, 5). In plants exposed to HT for <3 days, no morphological abnormalities were observed, and after a temperature downshift, male fertility was recovered (3, 4). On the other hand, when 5-mm-long panicles were examined after HT treatment for 3 days, anther parietal and epidermal cells significantly decreased auxin levels, showing 52.9% intensity of fluorescent signals per area in each anther locule compared with levels of control plant (Fig. 1 H and J). However, HT increased the signal intensity in the rachis cells around vascular bundles by >148.7% (Fig. 1 H and J). When 10-mm-long panicles were examined after exposure to HT for 5 days, auxin levels had decreased even further (39.4% intensity) in the pollen mother and tapetum cells at center of the anthers (Fig. 1 F and L).
HT causes reduction of endogenous auxin levels in barley anther cells. Effects of HT treatment on IAA distribution in anther and rachis cells. Plants were treated for 3 days (5-mm panicles at the early phase of anther development; A–D, control; G–J, HT) or 5 days (10-mm panicles; E and F, control; K and L, HT), starting at the five-leaf stage (2-mm panicles). Anti-IAA antibody (25) and Alexa 488-conjugated goat anti-mouse IgG antibody were used. Visualization was performed under constant excitation light and the same exposure period for fluorescence (B, D, F, H, J, and L) and DIC microscopy (A, C, E, G, I, and K). An, anthers; Pi, pistils; Ra, rachis cells.
To study whether the phenomenon of auxin reduction could be observed in the anther cells of other plant species, we tested a recombinant Arabidopsis line that expressed β-glucuronidase (GUS), with the gene under control of a synthetic auxin response element (DR5-GUS) and the natural auxin response gene (ARF19-GUS) (20, 21). The disappearance of DR5-GUS signals has been reported in anther cells of auxin perception defective tir1 and afb multiple mutants (19). In addition, the significant reduction of DR5-GUS signals was observed in young leaves of triple mutants of YUCCA genes yuc1, -2, and -6 or yuc1, -4, and -6 (17). HT injury was observed in recombinant Arabidopsis grown at 33 °C for >7 days, with plants forming short stamens and rarely producing any pollen because of premature abortion of microsporogenesis (Fig. 2 A, B, G, and I). In the case of recombinants exposed to 31 °C for 7 days, microspores could be observed in anthers at stage 10; however, pollen maturation and filament elongation were ultimately repressed (Fig. 2 C and M). The strongest GUS activity in DR5-GUS line appeared in pollen mother and tapetum cells at stage 10 of floral development (ref. 19; Fig. 2 D and G and Fig. S1). In recombinants exposed to HT (31 °C or 33 °C) for >1 day, the DR5-GUS signals in anther cells decreased significantly at stage 10 (Fig. 2). In developing anther cells, GUS activity in ARF19-GUS line were lower than that in DR5-GUS line, and these weak signals disappeared completely after 1 day of 33 °C HT treatment (Fig. 2 N and O).
Increasing temperatures cause repression of GUS activity in developing anthers of Arabidopsis DR5-GUS and ARF19-GUS lines. The strongest GUS signal appeared in developing anthers at stage 10 (ref. 19 and Fig. S1). After 7 days at 31 °C and 33 °C, complete male sterility was observed in wild-type A. thaliana (ecotype Columbia) and its derivatives expressing DR5-GUS and ARF19-GUS. Representative SEM images of mature flowers (stage 14) grown normally for 7 days at 23 °C (A; control) or with HT at 33 °C (B) and 31 °C (C). Position of anthers is indicated by an arrow. Representative images of GUS signals in the flower of DR5-GUS line at stage 10 under the following conditions: control (D and G); exposure to HT at 33 °C for 1 day (E and H) or 3 days (F and I); and exposure to HT at 31 °C for 1 (J), 3 (K), or 7 days (L and M). Weak GUS signals in ARF19-GUS were observed in the anthers at stage 10 (N; control), but not detected after treatment at 33 °C for 1 day (O), indicated by an arrow.
In contrast to the GUS staining in anther cells, the signals in the top of gynoecium and around vascular cells of petals were amplified with increased temperatures not only at stage 10 (Fig. 2) but also at earlier stages (Fig. 3 A–C). In mature flowers at later stage, the HT-caused induction around vascular cells of carpel was still observed (Fig. 3 D–F). However, the signals in anther cells at earlier and later stages were silent under HT condition (Fig. 3). In seedlings, HT treatment increased DR5-GUS signals in petiol of cotyledon, hypocotyl, and central cylinder of root (Fig. 3 G–I). These results indicate clearly that HT treatment represses the auxin signaling in an anther cell-specific manner, leading to similar abortion of pollen development and filament elongation in Arabidopsis as observed in barley.
High temperatures increase DR5-GUS expression in several tissues except for anthers of Arabidopsis. Representative images of GUS signals in the flowers of DR5-GUS line at earlier stages (A–C; prior stage 10) and later stage (D–F; mature flower) under the following conditions: control (A and D), exposure to HT at 31 °C for 1 (B and E) or 3 days (C and F). Representative images of DR5-GUS signals in the seedlings under the following conditions: control (G), exposure to HT at 31 °C for 2 days (H), and at 33 °C for 2 days (I). (Scale bars: 200 μm.)
To study the effect of HT on expression of certain auxin-related genes, we used stereomicroscopy to dissect Arabidopsis anthers at stage 9. Real-time RT-PCR of the YUCCA genes YUC2 and -6 showed significant repression 1 day after temperature upshift to 33 °C (Fig. 4A). After 3 days of 33 °C, YUC6 expression was even more severely reduced. TAA1/TIR2 expression was reduced after 3 days (Fig. 4A). In control plants, expression of the auxin-induced gene IAA1 (22) is weak at stage 9 (19). However, IAA1 expression decreased by ≈50% under HT conditions (Fig. 4A). Although treated with HT for 3 days, we did not detect any change in expression of the auxin receptor gene TIR1 (23), or ARF6 and 8, which are involved in jasmonic acid production and flower maturation (24), in anthers at stage 9 (Fig. 4A). Under normal conditions, the barley YUCCA genes unigene Nos. 31993 and 5729 in HarvEST: Barley Version 1.68 Assembly No. 35, show increased expression during the early development of panicles, but the up-regulation of No. 31993 expression could not be detected and the No. 5729 expression was severely repressed to ≈60% normal levels by increased temperatures (Fig. 4 B and C).
Effect of increasing temperatures on expression of YUCCA genes in Arabidopsis and barley. (A) Expression profiles of the Arabidopsis genes YUC2, YUC6, TAA1, IAA1, TIR1, ARF6, and ARF8 in anthers at stage 9. Total mRNA was isolated from ≈50 stamens of plants treated with or without 33 °C HT for 1 or 3 days. The relative ratio of gene expression was normalized by using expression of the ACT2 gene. (B and C) The expression levels of barley YUCCA genes, Unigene Nos. 31993 and 5729 (Contig11792_at) in HarvEST: Barley Version 1.68 Assembly No. 35, were determined from panicles of ≈2 mm (time = 0, n = 20), 3 mm (1 day after HT treatment at 30 °C/25 °C, n = 10) and 5 mm (3 day after HT treatment, n = 10), starting at the five-leaf stage. The relative ratios were normalized by using expression of the EF-1α gene (Unigene No. 13677: HB22P12r_x_at). All real-time RT-PCR experiments were performed in biological triplicate. Vertical bars represent standard error.
Recent Arabidopsis studies have reported that auxin affects pollen maturation, filament growth, and anther dehiscence (17, 19). Auxin is most likely biosynthesized in developing anther cells at floral stages 8–11, with the auxin response beginning to appear after stage 9 (19). The auxin biosynthesis genes YUC2 and -6 are strongly expressed in anther cells; their inactivation leads to short stamens and rarely produced any pollen (17). Thus, in the developing anther cells of Arabidopsis and barley plants, expression of auxin biosynthesis genes is susceptible to increased temperatures. Reduced expression of these genes probably causes anther-specific auxin depletion, and reduction of the auxin response.
Our previous result also shows that HT induces expression of auxin response genes in barley seedlings but not in panicles (5). In addition, Figs. 1–3 show opposite effects of HT on both endogenous auxin level and auxin signaling between developing anther cells and other tissues, especially around vascular cells; namely, they decrease in the former and increase in the latter. Thus, HT effects on auxin level appear to differ among plant tissues. Furthermore, auxin transport and perception mutants show a reduction in filament length but only exert a moderate effect on pollen maturation and anther dehiscence in Arabidopsis (19). Likewise, application of an auxin transport inhibitor does not affect the DR5-GUS staining in anthers (19). These results suggest that biosynthesis of endogenous auxin in developing anthers is a major factor responsible for the HT-caused abortion of pollen and auxin reduction. Accordingly, HT-tolerant plants might be obtained by controlling anther-specific auxin biosynthesis genes.
To assess how an exogenous application of auxin would affect HT injury to anther early development, we applied 10−6, 10−5, or 10−4 M indole-3-acetic acid (IAA; natural auxin), or the synthetic auxins 1-naphthaleneacetic acid (NAA) or 2,4-dichlorophenoxyacetic acid (2,4-D). Control or auxin-containing solutions contained 0.1% DMSO and 0.1% (vol/vol) Tween 20. Solutions were applied to barley plants in the mornings of days 18, 19, 21, and 23 (Fig. 5A). At the heading stage, control anthers grew to 2.99 ± 0.05 mm in length, whereas anthers exposed to high temperatures were only 1.48 ± 0.04 mm in length and contained no pollen grains (Fig. 5 B and C). Irrespective of auxinic compound used, auxin application restored anther length (1.8–2.5 mm), mature pollen grains (Fig. 5), and seed setting rate (Fig. 6) in a dose-dependent manner. Although anthers developed normally with application of 10−4 M 2,4-D, negative effects of auxin were detected, including premature blighting of leaves and loss of mature seeds (Fig. 6). Two applications of auxin (at days 19 and 21) were sufficient to restore anther development under HT conditions (Fig. S2). We have observed transcriptional repression of certain replication related genes including DNA replication licensing factor MCM5, and cell-proliferation arrest in anther wall cells and sporogenous cells, under HT conditions (5). The application of exogenous auxin could completely restore the MCM5 gene expression (Fig. S3), and it probably conducts normal proliferation and development of anther cells.
Effects of exogenous auxin on HT injury of pollen development in barley. (A) Schematic illustration of experiment for HT injury and auxin application. Applications occurred four times (days 18, 19, 21, and 23) starting at the five-leaf stage. (B) Representative structures (pistil, anthers, and pollen) are shown for each treatment. Application of auxin (IAA, NAA, or 2,4-D) reversed the abortion of pollen development under HT conditions. Mature pollen grains stained dark brown with iodine solution at the heading stage. (C) All auxins restored anther length at the heading stage in a dose-dependent manner. Vertical bars indicate SE. *, Statistical significance at P < 0.01 (Student's t test: n = 81 anthers used in each treatment). (D) A population of mature pollen grains was recovered from anthers after all auxin applications tested. More than 6 plants were tested independently under each set of conditions.
Exogenous auxin reversed male sterility and restored seed setting rate in barley. (A) Representative structures of mature ears after each treatment. In addition to pollen maturation, fertile seeds also developed normally following auxin application. (Scale bars: 10 mm.) (B) Seed setting rate was restored by all types of auxin, in a dose-dependent manner, with the exception of the 10−4 M 2,4-D application. All auxins were applied four times. The 10−4 M 2,4-D application restored pollen development (Fig. 5), but the plants withered before seed maturation, indicating that high concentrations of auxin cause blighting. Vertical bars indicate standard error. *, statistical significance at P < 0.01 (Student's t test: n = 9 plants used in each treatment).
The apex of each Arabidopsis inflorescence was sprayed once with 10−7 or 10−6 M IAA or NAA solution containing 0.1% DMSO just before increasing the temperatures to 31 °C. Seven days after the temperature upshift, HT injury resulted in significantly reduced stamen length, whereas in auxin-treated plants, this reduction could be suppressed in a dose-dependent manner (Fig. 7 A and B). In the stamens >1.0 mm length, mature pollen stained normally with an iodine solution. These results clearly suggest that a reduction in male tissue-specific auxin is the primary cause of HT injury. Furthermore, the resulting abortion of pollen development and male sterility can be reversed by the application of exogenous auxin. These phenomena are highly conserved in not only monocots but also dicots. To date, auxins have been used widely as potent and selective herbicides. Our results show that auxin is useful for the promotion of plant fertility and maintenance of crop yields under the global warming conditions.
Exogenous auxin rescues HT injury in Arabidopsis anther development. (A) Representative structures of Arabidopsis flowers at anthesis after HT treatment at 31 °C for 7 days with or without 10−6 M IAA application. The position of anthers is indicated by an arrow. (B) In plants grown for 7–11 days under HT treatment, the auxins IAA and NAA restored stamen length (including filament) of mature flowers in a dose-dependent manner. Thirty to 50 stamens were measured for each treatment and three independent experiments were scored.
Under normal conditions, double-rowed barley plants (H. vulgare L. cv “Haruna-nijyo”) were maintained for the entire growth period in a growth cabinet (LH350S, NK system) at 20 °C during the day and 15 °C at night, with a 16-h photoperiod. As described, HT treatment was started at the five-leaf stage (2-mm-long panicle), when the tip of the fifth leaf had emerged. Plants were grown at 30 °C during the day and 25 °C at night for 5 days (3⇓–5). During this period, pollen mother and tapetum cells developed from anther sporogenous and parietal cells, respectively (4, 5). Arabidopsis thaliana Columbia (wild-type) and its derivatives expressing the GUS reporter, were grown at 23 °C (day and night) in a growth chamber with a 16-h photoperiod. Light conditions were 60–80 μE/m2 per s. To study the effects of elevated temperature on reproductive development, we monitored the primary inflorescence of plants transferred to growth chambers at 31 °C or 33 °C under a 16-h photoperiod.
We assessed the effects of exogenously applied auxin on HT injury of anthers during early development. The entire barley shoot was sprayed with 6 mL of control or 10−7, 10−6, 10−5, or 10−4 M auxin-containing solution [0.1% DMSO and 0.1% (vol/vol) Tween 20]. The natural auxin IAA, or the synthetic auxins NAA or 2,4-D, were applied in the morning of days 18, 19, 21, and 23 (Fig. 4). Tween 20 was eliminated from auxin solutions for Arabidopsis treatments. A single application of ≈0.1 mL of auxin solution was sprayed on the inflorescence of each Arabidopsis plant in the morning of the day that the temperature upshift occurred. Stamen length and pollen morphology were observed in the mature flowers 7–11 days after the temperature upshift. Ten to 12 plants were grown into a single pot, and the three pots were used in a series of experiments for each treatment. Each series of experiments was carried out independently in triplicate.
Barley anthers and Arabidopsis stamens were measured at the heading stage and anthesis, respectively, after dissection with a stereo microscope and CCD camera (SZX12 and DP20; Olympus). Pollen grains in the anthers were stained with an iodine solution (Lugol solution; MERCK). Dark brown pollen grains were scored as mature pollen. To observe mature Arabidopsis flowers with the scanning electron microscope (SEM: JEOL JSM-5800LV at 5 kV), excised flowers were fixed in FAA [3.7% formaldehyde, 50% ethanol, and 5% glacial acetic acid (vol/vol)] for 3 h at room temperature, dehydrated with ethanol, replaced with isoamyl acetate, and dried with a critical point drier (JEOL JCPD-5).
We examined the IAA distribution in early developing barley panicles (5-mm-long, i.e., just before development of tapetum, middle layer, and endothecial cells from the parietal cells), after treatment with or without HT stress for 3 days. As described, IAA was detected by using an anti-IAA monoclonal antibody (25), with the modifications indicated below. To cross-link IAA, excised panicles were immediately prefixed for 2 h in 3% (wt/vol) EDAC (Sigma-Aldrich) at room temperature and then transferred to FAA for 24 h at 4 °C. Fixed panicles were dehydrated in a series of ethyl alcohol and tertiary butyl alcohol washes, and then embedded in paraffin (PARAPLAST Plus, Oxford Labware). Samples were serially sectioned (10 μm thick) with a microtome and then affixed onto MAS-coated slides (Matsunami Glass Industry). After overnight drying at 42 °C, sections were deparaffinized with xylene and hydrated by using an ethanol-water series. Specimens were incubated in 10 mM PBS (PBS; 2.68 mM KCl, 0.15 M Na2HPO4, and 0.086 M KH2PO4) containing 0.1% (vol/vol) Tween 20, 1.5% glycine, and 5% BSA for 45 min at 22 °C. Samples were then rinsed in a regular salt rinse solution [RSRS; 10 mM PBS, 0.88% NaCl, 0.1% (vol/vol) Tween 20 and 0.8% BSA] and washed briefly with 10 mM PBS containing 0.8% BSA solution (PBS+BSA) to remove the Tween 20. After the application of anti-IAA monoclonal antibody (No. A0855, Sigma-Aldrich; 400 μL of a 1:1,000 dilution from 2.1 mg/mL stock) to each slide, samples were incubated overnight in a humidity chamber at room temperature. After hybridization, samples were subjected to a series of vigorous washes, twice with a high-salt rinse solution [HSRS; 10 mM PBS, 2.9% NaCl, 0.1% (vol/vol) Tween 20 and 0.1% BSA] for 10 min, once with RSRS for 10 min, and briefly with PBS+BSA. The Alexa 488-conjugated goat anti-mouse IgG antibody (Invitrogen; 400 μL of a 1:300 dilution of 2 mg/mL stock) was then placed on each slide, and these were incubated for 4–6 h in a humidity chamber at room temperature. After washing with RSRS twice for 15 min, specimens were mounted with an anti-fade reagent (ProLong Gold; Molecular Probes), covered with a coverslip, and observed under a fluorescent microscope (BX51; Olympus). The intensity of fluorescent signal was measured by Image J software.
GUS staining was performed as follows: Arabidopsis inflorescences were fixed in acetone (90%) for 2 h at −20 °C, and then infiltrated for 15 min with 2 mM ferricyanide and 2 mM ferrocyanide in Na phosphate buffer (staining buffer). Samples were then incubated in 2 mM X-Gluc (in staining buffer) at 37 °C for 15 h (for inflorescences) or 0.5 h (for seedlings), and the whole tissues were cleared by using Hoyer's medium. Observations were performed by using DIC optics (BX51; Olympus). To examine the GUS expression pattern in tissues, stained samples were fixed in FAA for 30 min at room temperature. Then samples were dehydrated, embedded in paraffin, and serially sectioned at a thickness of 10 μm.
Isolation of total RNA from barley panicles and Arabidopsis stamens was carried out by using TRIzol Reagent (Invitrogen). Real-time quantitative RT-PCR was performed by using the SYBER ExScript RT-PCR Kit (TaKaRa), and the primer sets are listed in Table S1. Triplicate PCRs in each sample were carried out in a series of experiments, and each series was performed in biological triplicate.
Statistics were calculated with MS Excel. Statistical significance was assessed by an unpaired Student two-tailed t test. Values were considered statistically significant at P < 0.05.
We thank the following: Drs. T. J. Guilfoyle, University of Missouri (Columbia, MO), and J. W. Reed, University of North Carolina (Chapel Hill, NC), and the Arabidopsis Biological Resource Center at The Ohio State University (Columbus, OH) for kindly supplying the DR5, ARF6, -8, and -19 GUS recombinant lines and wild-type A. thaliana; Drs. J. S. Heslop-Harrison, F. Berger and Y. Hotta for critical reading of the manuscript; and C. Watanabe, N. Fujii, and T. Sato for helpful suggestions. This work was funded in part by Ministry of Education, Culture, Sports, Science and Technology Grants 21-COE, G-COE, 18075003, 1807512, 20678001, and 19043004; Ministry of Agriculture, Fisheries, and Food Grant IPG-0019; and a Ground-Based Research Announcement for Space Utilization promoted by the Japan Space Forum. This work based on a Patent PCT/JP2010/50101 by A.H., T.S., and M.W.
↵1T.S., T.O., S.M., and M.T. contributed equally to this work.
Author contributions: T.S., H.T., M.W., and A.H. designed research; T.S., T.O., S.M., M.T., Y.T., N.H., Y.M., and A.H. performed research; T.S., T.O., S.M., M.T., Y.T., N.H., and A.H. analyzed data; and T.S., Y.M., H.T., M.W., and A.H. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1000869107/-/DCSupplemental.
