SUMO E3 Ligase AtMMS21 Regulates Drought Tolerance in Arabidopsis thaliana


  • Shengchun Zhang,

    1. Guangdong Key Lab of Biotechnology for Plant Development, College of Life Science, South China Normal University, Guangzhou 510631, China
    Search for more papers by this author
  • Yanli Qi,

    1. Guangdong Key Lab of Biotechnology for Plant Development, College of Life Science, South China Normal University, Guangzhou 510631, China
    Search for more papers by this author
  • Ming Liu,

    1. Guangdong Key Lab of Biotechnology for Plant Development, College of Life Science, South China Normal University, Guangzhou 510631, China
    Search for more papers by this author
  • Chengwei Yang

    Corresponding author
    1. Guangdong Key Lab of Biotechnology for Plant Development, College of Life Science, South China Normal University, Guangzhou 510631, China
    Search for more papers by this author

Tel/Fax: +86 20 8521 0855; E-mail:


Post-translational modifications of proteins by small ubiquitin-like modifiers (SUMOs) play crucial roles in plant growth and development, and in stress responses. The MMS21 is a newly-identified Arabidopsis thaliana L. SUMO E3 ligase gene aside from the SIZ1, and its function requires further elucidation. Here, we show that MMS21 deficient plants display improved drought tolerance, and constitutive expression of MMS21 reduces drought tolerance. The expression of MMS21 was reduced by abscisic acid (ABA), polyethylene glycol (PEG) or drought stress. Under drought conditions, mms21 mutants showed the highest survival rate and the slowest water loss, and accumulated a higher level of free proline compared to wild-type (WT) and MMS21 over-expression plants. Stomatal aperture, seed germination and cotyledon greening analysis indicated that mms21 was hypersensitive to ABA. Molecular genetic analysis revealed that MMS21 deficiency led to elevated expression of a series of ABA-mediated stress-responsive genes, including COR15A, RD22, and P5CS1 The ABA and drought-induced stress-responsive genes, including RAB18, RD29A and RD29B, were inhibited by constitutive expression of MMS21. Moreover, ABA-induced accumulation of SUMO-protein conjugates was blocked in the mms21 mutant. We thus conclude that MMS21 plays a role in the drought stress response, likely through regulation of gene expression in an ABA-dependent pathway.


All organisms use a variety of chemical modifiers for post-translational control of proteins that affect development and growth. Ubiquitin (Ub) is one such polypeptide that was first described to attach covalently to other proteins upon completion of their synthesis. SUMO (small ubiquitin-like modifier) is a type of protein termed as ubiquitin-like modifiers, that function in a manner analogous to ubiquitin. SUMO conjugation to protein substrates (sumoylation) is a reversible, post-translational modification that is regulated by environmental stimuli in animals and yeast (Johnson 2004). Plant sumoylation is involved in controlling cell growth and development (Miura and Hasegawa 2010), embryogenesis (Colby et al. 2006; Saracco et al. 2007), regulation of flowering time (Jin et al. 2008) and root growth and meristem maintenance (Huang et al. 2009; Ishida et al. 2009). In addition, sumoylation is involved in actions of both biotic and abiotic stresses (Kurepa et al. 2003), including salicylic acid-dependent pathogen defense (Lee et al. 2007), phosphate starvation responses (Miura et al. 2005), cold tolerance (Agarwal et al. 2006), drought response (Catala et al. 2007), basal thermotolerance (Yoo et al. 2006), excess copper tolerance (Chen et al. 2011), abscisic acid (ABA) signaling (Miura et al. 2009; Zheng et al. 2012) and nitrogen assimilation (Park et al. 2011). Similar to ubiquitination, sumoylation of substrates is catalyzed by a cascade of enzymes: the E1 SUMO-activating enzyme (AOS1-UBA2), the E2 conjugating enzyme (UBC9), and the E3 SUMO ligases (Hay 2005). However, to date, only two SUMO E3-ligases, SIZ1 and MMS21/HPY2, have been identified in Arabidopsis, and these proteins contain the SP-RING structural domain known to confer SUMO E3-ligase activity. The Arabidopsis SIZ1 is a prototype mammalian protein inhibitor of activated signal transducers and activators of transcription (PIAS), and is similar to the yeast Siz (scaffold attachment factors A/B/acinus/PIAS (SAP) and MIZ) family of SUMO E3-ligases (Miura et al. 2005). In contrast, MMS21/HPY2 constitutes a MMS21-type enzyme, an imperfect PIAS protein, which contains the SP-RING domain alone (Huang et al. 2009; Ishida et al. 2009).

Drought stress is one of the most severe environmental factors that greatly restricts plant distribution and crop production (Zhu 2002). To reduce the adverse effects of drought stress, plants have evolved multifaceted strategies, including morphological, physiological, and biochemical adaptations (Bohnert et al. 2006). Some of these strategies aim to avoid dehydration stress by increasing water uptake or reducing water loss, whereas other strategies seek to protect plant cells from damage when water is depleted and tissue dehydration becomes inevitable (Verslues et al. 2006). Moreover, the adaptive response to drought must be coordinated at the molecular, cellular, and whole-plant levels (Yu et al. 2008).

Proline acts as an osmoprotectant and cryoprotectant in organisms as diverse as bacteria, plants, and insects (Szabados and Savour 2010). Many plants accumulate proline in response to low water potential and dehydration caused by drought or freezing. Transpirational water loss through the stomata is a key determinant of drought tolerance (Xiong et al. 2002). The closing and opening of the stomata are mediated by a turgor-driven change in volume of the two surrounding guard cells (Yu et al. 2008). Guard cell turgor change is influenced by many factors, such as light, phytohormones, potassium ions, calcium ions, malate, NO, and H2O2 (Shimazaki et al. 2007). The guard cells sense and integrate environmental signals to modulate stomatal aperture in response to drought stress.

ABA is an important phytohormone that regulates many essential processes, including inhibition of germination, maintenance of seed dormancy, and regulation of stomatal behavior, and is an essential mediator in triggering plant responses to most of the common abiotic stresses, including drought, salinity, high temperature, oxidative stress, and cold (Finkelstein et al. 2002). Drought causes increased biosynthesis and accumulation of ABA, which can be rapidly catabolized following the relief of stress (Cutler and Krochko 1999). Nevertheless, high levels of ABA inhibit plant growth by affecting cell division and elongation (Finkelstein et al. 2002). The molecular mechanisms underlying ABA-mediated plant tolerance to drought stress are still not fully understood because of the complex nature of the plant response to ABA signaling and drought stress, but the study of stress-responsive transcription factors has been one of the foci in studies on drought stress tolerance (Fujita et al. 2011).

Based on genetic and biochemical analyses, stress-responsive SUMO conjugation in Arabidopsis is mediated mainly by the SIZ1 SUMO E3-ligase, which has been reported to participate in a series of development and stress-responsive processes (Miura and Hasegawa 2010). In particular, SIZ1 has been reported to play important roles in ABA signaling and plant drought tolerance (Catala et al. 2007; Miura et al. 2009; Zheng et al. 2012). Recently, it was reported that SIZ1 negatively affects stomatal closure and drought tolerance through the accumulation of salicylic acid-induced reactive oxygen species (ROS) (Miura et al. 2012). However, the Arabidopsis SUMO E3 ligase AtMMS21, a homologue of NSE2/MMS21, has only recently been identified to participate in root development (Huang et al. 2009; Ishida et al. 2009), and whether it plays roles in the process of stress-response is unclear. In this report, detailed characterization of the knockout mutant mms21 and MMS21 over-expression plants reveals that the E3 ligase MMS21 is involved in the responses of plants to ABA and drought stresses. We also performed a comparison between the expression profiles of mms21 with wild-type (WT) plants grown under control conditions. Our results reveal that MMS21 regulates the expression of an important set of genes in response to drought stress. These results indicate that MMS21 regulates the response to water deficit by changes in gene expression.


The AtMMS21 gene regulates drought tolerance negatively

It has been reported that plant SUMO E3-ligase SIZ1-mediated sumoylation plays broad and important roles in stress response. As a newly-identified SUMO E3-ligase, MMS21 plays crucial roles in root growth. However, its function in the stress responses is unclear. This study focuses on the function of MMS21 on the ABA-mediated plant drought tolerance.

Because mms21 mutant plants are always alive when facing drought stress, it is believed that mms21 mutants may have an increased tolerance to water deficit. To test this hypothesis, the mms21 mutant, AtMMS21 over-expression (MMS21-OE) and WT seeds (all of them are the same lines which were used by Huang et al. 2009) were germinated simultaneously and then planted in soil. Three weeks after germination, the plants were treated with natural drought (water was withheld). Thereafter, plants were challenged with drought by withholding water for 17 d. The plants were then re-watered and photographed after two d (Figure 1A), and the surviving plants were measured. The AtMMS21 over-expression plants showed wilting symptoms 3 d in advance compared with the WT plants. After 2 w without watering, the mms21 plants did not show drought stress symptoms, and the WT plants showed only weak drought stress symptoms, whereas the MMS21-OE plants exhibited severe drought symptoms. After 3 d of continued drought stress, the plants only showed only weak drought stress symptoms. When plants were re-watered, only 10% of the MMS21-OE plants survived, but almost all of the mms21 mutant plants were still alive (Figure 1B). These results suggested that mms21 plants were comparatively more resistant to drought stress than WT and MMS21-OE plants, and that AtMMS21 regulates plant drought stress negatively.

Figure 1.

AtMMS21 mutation improves Arabidopsis thaliana drought tolerance
(A) Drought stress was imposed on 3-w-old seedlings for 14 d or 17 d, and then re-watering was done for 2 d. Drought experiments were repeated four times, and at least 36 plants for each individual line were used in each repeated experiment and one representative picture was shown. (B) The survival rate of different AtMMS21 genotype plants after drought for 17 d and re-watering for 2 d. (C) Rate of water loss by detached leaves from wild-type (WT), mms21 and MMS21 over-expression plants. Values are the mean ±SE (n = 6 plants). (D) Leaf stomata photographs of different AtMMS21 genotype plants. Bar = 100 μm. (E) Leaf stomatal density of different AtMMS21 genotype plants.

Transpirational water loss is one of the most important factors related to drought tolerance. To assess the water loss rate of mms21, MMS21-OE and WT plants, rosettes were detached and their fresh weight changes were measured over a 300-min period. The mms21 leaves showed a slower rate of water loss than WT leaves (Figure 1C). The reduced rate of water loss is attributable to the increased drought tolerance in mms21 mutant plants. It is speculated that the slower transpirational water loss may result from the decrease of leaf stomata density, so the leaf stomata density of three different MMS21 genotype plants was measured. The results showed that the leaf stomata density of mms21 is significantly larger than that of the WT and MMS21-OE plants (Figure 1D).

AtMMS21 gene mutation increases proline content

In order to test whether or not mms21 only has better drought tolerance under natural drought conditions, the 10-d-old seedlings of different AtMMS21 phenotype plants were cultured in medium supplemented with 10% polyethylene glycol (PEG) 6000, a stress treatment commonly used to mimic drought tolerance in the laboratory. Under the PEG stress condition, a part of the cotyledon showed a yellow phenotype. Obviously, the cotyledon yellow of the MMS21-OE plants was more severe that that of the mms21 and WT plants (Figure 2A). After 7 d of stress treatment, the yellow rate of the MMS21-OE plants reached to 80%, whereas that of the mms21 mutant and the WT plants was only about 20% and 40%, respectively (Figure 2C). Drought stress can cause increased accumulation of proline, which plays important roles in the drought response. To determine whether proline levels were affected, we quantified the proline content in different MMS21 phenotype plants. Under normal conditions, the proline content of mms21 is one-fold higher than in MMS21-OE plants. When exposed to 10% PEG 6000, mms21 accumulated one-fold higher proline content than the WT, and 3-fold higher than that of MMS21-OE plants (Figure 2B). In Arabidopsis, transcriptional up-regulation of P5CS1 is essential for proline accumulation. To determine whether the P5CS1 expression patterns were affected, we detected its transcript levels using quantitative RT-PCR. Under normal conditions, the P5CS1 level in the mms21 mutant was almost 1.3-fold higher than in the WT plants, and 3-fold higher than in the MMS21-OE plants. After PEG treatment, the expression levels of P5CS1 in the mms21 mutant plants and WT plants was induced significantly; however, it was only slightly increased in the MMS21-OE plants (Figure 2D). These results suggest that the elevated proline levels may contribute to drought tolerance in mms21, and the AtMMS21 gene may be a negative regulator of proline synthesis.

Figure 2.

AtMMS21 mutation increases proline content
(A) The mms21 plants are insensitive to 10% PEG stress. (B) Proline contents of different AtMMS21 genotype plants before or after polyetylene glycol (PEG) treatment. (C) Leaf yellow rate of different AtMMS21 genotype plants after exposure to 10% PEG for 1 w. (D) The expression level of the P5CS1 gene in different AtMMS21 genotype plants before or after different time PEG treatments.

AtMMS21 participates in the ABA pathway

In response to drought stress, stomata often close to limit water loss by transpiration. During this process, ABA plays a role in stomatal closure. Given that endogenous ABA was up-regulated in mms21, we further investigated whether mms21 affects the sensitivity of guard cells to ABA. Epidermal peels of WT, mms21 and MMS21-OE plants were incubated in a buffer solution under strong light conditions for 12 h to fully open the stomata. Next, the peels were treated with different concentrations of ABA for 2 h (Pei et al. 1997). The ratio of stomatal width to length indicated the degree of stomatal closure. Three different AtMMS21 genotype plants showed the same stomatal width-to-length ratio of fully-opened stomata without ABA treatment, but mms21 plants showed a lower stomatal width-to-length ratio than WT plants after treatment with ABA (Figure 3A). These results suggest that stomatal closure in mms21 is more sensitive to ABA than in the WT and MMS21-OE plants, which may be critical for mms21 mutants to adapt to drought stress, and also indicates that AtMMS21 may participate in ABA pathway.

Figure 3.

Exogenous stress affects the expression of AtMMS21 and stomatal closure
(A) Stomatal closure in response to abscisic acid (ABA) treatment. (B) The expression of AtMMS21 is inhibited by dehydration. (C) The expression of AtMMS21 is inhibited by exogenous ABA treatment. (D) The expression of AtMMS21 is inhibited by exogenous polyetylene glycol (PEG) treatment.

ABA plays an important role in regulating plant responses to different stresses. To understand the relationship between MMS21 and ABA, real-time RT-PCR analysis was conducted to examine whether MMS21 is responsive to dehydration, ABA and PEG stresses. After exposing the 10-d-old WT seedlings to dehydration, 10 μM ABA or 20% PEG conditions for 3 h or 6 h, MMS21 expression was down-regulated significantly in response to dehydration, ABA or PEG treatments, and the inhibition degree was positively correlated with the treatment time (Figure 3B–D). These results indicate that the expression of MMS21 was inhibited by the exogenous or stress-induced ABA, and imply that MMS21 plays a role in abiotic stress.

Our results revealed that AtMMS21 expression was inhibited by abiotic stresses, such as ABA and PEG (Figure 3B). Given the function of ABA in seed dormancy, inhibitory experiments of seed germination have provided useful insights into components of ABA signaling, so we explored the functions of AtMMS21 in seed germination. We tested the germination frequency of WT, mms21, and MMS21-OE plant seeds on Murashige and Skoog (MS) medium supplemented with 0 μM, 0.2 μM, 0.5 μM or 1.0 μM ABA, and compared for differences in germination and post-germinative growth. In the absence of ABA, there was no significant difference among WT, mms21 and MMS21-OE lines. In the presence of ABA, the ABA-sensitive response of mms21 occurred at concentrations as low as 0.2 μM ABA, and mms21 germinated later than WT plants. The MMS21-OE seeds displayed a higher germination ratio than WT seeds, whereas the seed germination frequency of mms21 was significantly lower than that of the WT in different ABA concentrations. After germination for 24 h, the germination rate of mms21 seeds was only 30%, whereas that of the MMS21-OE seeds reached 90% on basal MS supplementation with 1.0 μM ABA (Figure 4B). Moreover, early seedling growth of mms21 was also more sensitive to ABA than the WT and MMS21-OE plants (Figure 4A). When seedlings were grown on basal MS medium supplemented with 1.0 μM ABA for 7 d, no mms21 mutant seedlings exhibited greening, whereas the cotyledon greening rate of the MMS21-OE seedlings reached 90% (Figure 4C), and the growth of mms21 mutant seedlings was severely arrested. After germination for 2 d, the seedlings were transferred to the ABA medium. The results showed that the mms21 mutant root is more sensitive to exogenous ABA than the WT and MMS21-OE plants (Figure 4D, E). It is thus suggested that AtMMS21 not only affects seed germination, but also post-germination early growth.

Figure 4.

The mms21 mutant is more sensitive to abscisic acid (ABA)
(A) Different AtMMS21 genotype plant seeds germinate and grow in Murashige and Skoog (MS) medium supplemented with different ABA concentrations for 1 w. (B) Quantification of seed germination. Seed germination percentage of three genotypes grown on different concentrations of ABA was recorded at 1 d after stratification. (C) Quantification of cotyledon greening. Cotyledon-greening percentage of three genotypes grown on medium containing different concentrations of ABA was recorded at 7 d after stratification. (D, E) The mms21 mutant root is more sensitive to exogenous ABA than the wild-type (WT) and MMS21-OE plants.

AtMMS21 regulates the expression of stress-responsive genes

The enhanced tolerance of mms21 plants to drought, along with the MMS21 involved in ABA signaling during germination and post-germination growth, and the inhibitory effect of ABA on AtMMS21 gene expression in WT plants, prompted us to evaluate whether the expression of ABA-mediated stress-responsive genes in mms21 mutant plants was affected. Plants have evolved some mechanisms to adapt to stress environments such as drought, salinity, and cold. Some stress-responsive genes can be induced to help plants to survive under drought conditions. To explore the molecular mechanisms of AtMMS21 function in drought tolerance, we determined whether AtMMS21 regulates the expression of stress-responsive genes. In plants exposed to drought, ABA is increased in the tissues, inducing stomatal closure and promoting the expression of many stress-related genes such as RAB18, RD29A and RD29B, that enhance water stress tolerance or avoidance (Yamaguchi-Shinozaki and Shinozaki 1993), and RD29A is a stress-responsive marker used as a control for stress treatments. ABF3 is induced by dehydration, high salinity, or ABA treatment in vegetative tissues, and its gain-of-function mutants show enhanced drought stress tolerance (Fujita et al. 2005). As expected, these four genes were up-regulated in mms21 plants under ABA treatment (Figure 5). In contrast, their expression decreased in MMS21-OE compared with that in WT plants under ABA-induced conditions. In order to find out whether the expression of other stress-responsive genes had changed in the mms21 mutants, a microarray experiment was conducted, comparing the mms21 mutant with WT plants under normal conditions. The results showed that a series of ABA- and stress-responsive genes were up-expressed in the mms21 mutants under normal growth conditions (Table 1), including COR15A, COR15B, P5CS1, RAB18, ERD11 and RD22 and so on, which are related to dehydration stress. These results indicate that MMS21 negatively regulates these stress-responsive genes, which may account for the enhanced drought tolerance in the mms21 mutant.

Figure 5.

Expression levels of stress-responsive genes in the mms21 mutant
Relative expression levels of ABF3, RAB18, RD29A and RD29B. The 2-w-old seedlings were used for abscisic acid (ABA) treatment.

Table 1.  Abscisic acid (ABA) and stress-responsive genes up-expressed in mms21 plants under normal growth conditions
Gene nameAGI CodeRatio*Description
  1. *Expression fold increase in mms21 relative to wild-type (WT) plants.

COR15A At2g4254014.26 Cold-regulated 15A
COR15B At2g4253010.36 Cold-responsive protein/cold-regulated protein (cor15b)
GEA6 At2g401704.39Late embryogenesis abundant 6
LTP3 At5g593204.16Lipid binding
KIN2 At5g159703.47Cold-responsive 6.6
ADH1 At1g771203.35Alcohol dehydrogenase
VSP2 At5g247702.87Acid phosphatase
P5CS1 At2g398002.85Proline biosynthetic process, response to abscisic acid stimulus
RAB18 At5g664002.80Responsive to ABA 18
ERD11 At1g029302.78Early responsive to dehydration 11
RD22 At5g256102.66Responsive to desiccation 22
ATHB-7 At2g466802.53Arabidopsis thaliana homwobox-7
COR314-TM2 At1g293902.41Cold regulated 314 thylakoid membrane 2

AtMMS21-mediated SUMO-Protein conjugates in response to ABA

Abiotic stresses, such as heat shock, low temperatures, ethanol, and H2O2 and phosphorus deficiency, have been reported to trigger a significant increase in SUMO-protein conjugate levels (Kurepa et al. 2003; Murtas et al. 2003; Yoo et al. 2006; Miura et al. 2007). It was reported that SUMO-protein conjugate levels were also elevated by drought treatment, but the induction appeared to be ABA-independent, because no significant difference was observed between the WT and the ABA-deficient mutant aba2 (Catala et al. 2007). To explore the possible role of MMS21 in this process, we compared changes in SUMO-protein conjugate levels in WT, mms21, and MMS21-OE plants. ABA-induced accumulation of SUMO-protein conjugates was obviously increased in the WT and MMS21-OE plants, but was reduced significantly in mms21 mutants (Figure 6) after different concentrations ABA treatment, suggesting that MMS21 mediates, in part, the increase of SUMO-protein conjugate levels in response to ABA.

Figure 6.

AtMMS21 mutation inhibits the formation of abscisic acid (ABA)-induced SUMO-protein conjugates
In planta sumoylation profiles of wild-type (WT), mms21 and AtMMS21 over-expression seedlings that were grown for 7 d on full-strength Murashige and Skoog (MS) medium, and then treated with sterile water or ABA for 3 h. Ponceau staining of total protein was used as the loading control. Protein extracts were analyzed by protein gel blots using anti-SUMO1polyclonal antibodies to detect SUMO-protein conjugates. The large subunit of Rubisco (55 kD) was used as a loading control. Each lane contained 10 μg of protein.

These results also imply that AtMMS21-mediated sumoylation plays important roles in response to ABA-mediated stress.


Sumoylation plays important roles in floral initiation, root growth, ABA signaling, and plant response to various stresses like heat and cold shock, phosphate deficiency, and pathogen resistance (Kurepa et al. 2003; Lois et al. 2003; Murtas et al. 2003; Miura et al. 2005; Yoo et al. 2006; Miura et al. 2007; Huang et al. 2009). In this report, we show that the null mms21 mutant displays increased drought tolerance and hypersensitive responses to exogenous ABA. Our physiological, phenotypic, and molecular characterisation of mutants demonstrates that AtMMS21 plays a crucial role in drought tolerance and ABA signaling.

ABA is an essential mediator in triggering plant responses to most of the common abiotic stresses (Finkelstein et al. 2002). To evaluate the specific effects of MMS21 in ABA signaling, we analyzed whether MMS21 affects responses to exogenous ABA application. Our data suggest that the exogenous application of ABA resulted in decreased levels of MMS21 transcripts (Figure 3), and increased sensitivity of mms21 plants to ABA (Figure 4). On the other hand, mms21 mutant plants showed an increased tolerance to drought (Figure 1A). Measurements of water loss in detached leaves from irrigated plants showed a slower water loss in the mms21 mutant (Figure 1C). The ability of guard cells to respond to environmental changes is one of the major mechanisms that governs water loss in plants (Schroeder et al. 2001; Sirichandra et al. 2009). Since ABA regulates stomatal activity (Hetherington 2001), the stomatal response to ABA was also examined in three genotypes. Interestingly, though the stomatal density of mms21 is the largest, the stomatal aperture of mms21 was the smallest in ABA-induced stomatal closure (Figure 3A). These results indicate that MMS21 may play a crucial role in the drought tolerance response by regulating stomatal opening and closing by an ABA-dependent pathway.

Previous work has reported an increase in the accumulation of SUMO-protein conjugates in response to abiotic stresses, such as heat shock, cold, ethanol, or H2O2 (Kurepa et al. 2003; Murtas et al. 2003; Yoo et al. 2006; Miura et al. 2007). Here, we show that Arabidopsis plants exposed to ABA accumulate increased levels of sumoylated proteins by an ABA-dependent pathway. Moreover, this increase is also highly dependent on MMS21 activity, since the accumulation of SUMO-protein conjugates is significantly lower in mms21, and higher in MMS21-OE compared with WT plants (Figure 6). Refer to previous reports that siz1 mutants show a decrease in the accumulation of SUMO-protein conjugates under heat shock and cold (Miura et al. 2005, 2007; Yoo et al. 2006), our results suggest that MMS21 plays a general role in abiotic stress responses. We note that the accumulation of SUMO-protein conjugates in mms21, although significantly decreased compared with WT plants, remain inducible by ABA (Figure 6), indicating the existence and contribution of additional SUMO E3 ligases or other proteins with SUMO E3 ligase activity, except SIZ1, which accumulates increased levels compared to sumoylated proteins through an ABA-independent pathway (Catala et al. 2007).

Proline acts as an osmoprotectant and cryoprotectant in organisms as diverse as bacteria, plants, and insects (Szabados and Savour 2010). Many plants accumulate proline in response to low water potential and dehydration caused by drought or freezing. In Arabidopsis, transcriptional up-regulation of Δ1-pyrroline-5-carboxylate synthetase1 (P5CS1) is essential for low water potential-induced proline accumulation, and proline accumulation of p5cs1 mutants is only 15–20% of the WT level (Yoshiba et al. 1995; Szekely et al. 2008; Sharma and Verslues 2010). The basal expression level of P5CS1 was increased significantly in the mms21 mutant and reduced in the MMS21-OE plants compare to that of the WT plants (Figure 2D, Table 1). Consistent with the P5CS1 expression level, the proline content also was one-fold higher in mms21 than in the MMS21-OE plants. After PEG treatment, stress-inducible P5CS1 expression and proline accumulation were inhibited by MMS21 over-expression (Figure 2C, D). These results indicate that MMS21 is a crucial regulator of proline synthesis. It was reported that ABA induction of P5CS1 represents a primary transcriptional regulation of proline content (Strizhov et al. 1997). Salt induction of proline accumulation and P5CS1 gene expression are inhibited in the ABA-deficient aba1 and ABA-insensitive abi1 Arabidopsis mutants (Savoure et al. 1997; Strizhov et al. 1997; Nambara et al. 1998). Combined with the results that the expression of MMS21 was down-regulated by PEG or ABA treatment (Figure 3), we speculate that the ABA-inducible proline accumulation pathway is partially blocked by the loss of function of MMS21. Overall, basal and drought-inducible expression patterns of P5CS1 and increased proline accumulation levels in mms21 may contribute to the mutant's insensitivity to drought stress.

In plants exposed to drought, ABA increases in the tissues, inducing stomatal closure and promoting the expression of many stress-related genes such as RAB18, RD29A and RD29B that enhance water stress tolerance or avoidance (Acharya and Assmann 2009). The expression of RAB18, RD29A and RD29B, three typical genes induced by ABA, was strongly and quickly induced in the mms21 mutants compared with WT plants, and was obviously inhibited in MMS21 over-expression plants after ABA treatment (Figure 5), indicating that the MMS21-dependent pathways are also involved in the regulation of these genes in response to drought stress. The tolerance of Arabidopsis to dehydration is mediated mainly by three independent signaling pathways: the first one is dependent on ABA, the second one is regulated by the transcription factor DREB2A, and the last one regulates ERD1 gene expression (Shinozaki et al. 2003; Sakuma et al. 2006; Shinozaki and Yamaguchi-Shinozaki 2006). MMS21 is needed for the basal expression of some ABA-dependent genes, including COR15A, COR15B, KIN2 and more at the WT levels, because these genes increased significantly in the mms21 mutants under normal conditions (Table 1). The drought induction of other ABA-dependent genes (i.e., RAB18 and RD29B, Figure 5) is also affected in mms21 mutants. Moreover, the expression of ABI5 is also inhibited in the MMS21 over-expression plants (Figure 5). Therefore, our results suggest that MMS21 regulates part of the ABA-dependent signaling pathway by an ABA-dependent process, and MMS21 likely acts in a new signaling pathway in the response to dehydration stress.

The ABRE-binding (AREB) proteins or ABRE-binding factors (ABFs) were isolated by using ABRE sequences as bait in yeast one-hybrid screenings (Choi et al. 2000). AREB1/ABF2, AREB2/ABF4, and ABF3 are induced by dehydration, high salinity, or ABA treatment in vegetative tissues, and their gain-of-function mutants show enhanced drought stress tolerance (Fujita et al. 2005). The ABF3 basal expression is induced by MMS21 deficiency, and the ABA-mediated inducible-expression of ABF3 is inhibited by MMS21 over-expression (Figure 5), indicating that the MMS21 mutation affects the ABF3 normal expression pattern. However, the promoter sequence analysis of the MMS21 gene showed that three ABRE cis-elements exist in the 284 base pairs promoter region. Therefore, these results indicate that the function of MMS21 in the response of Arabidopsis to drought stress is complex, and it may regulate different signaling pathways at different levels.

In conclusion, our results show that MMS21 is an important component in the control of plant drought stress responses. Though the basal expression of some stress-responsive genes are obviously changed in the mms21 mutants under normal conditions, the comparison of drought- or ABA-induced gene expression patterns between mms21 and WT plants could give us more clues to understand the detailed mechanism behind MMS21-mediated drought tolerance. Also, the ABA content in MMS21 genotype plants needs to be determined to test whether or not ABA biosynthesis is affected by MMS21. Considering that MMS21 is a SUMO E3 ligase that most probably targets multiple substrates, the isolation and identification of sumoylated proteins related to MMS21 will help to clarify the biological functions of the sumoylation machinery in plant drought tolerance.

Materials and Methods

Plant materials and growth conditions

Seeds of Arabidopsis thaliana (Columbia 0 ecotype background, either wild-type (WT), mutant or over-expression transgenic plants) were sown in a Petri dish containing sterile solid medium consisting of 1× Murashige and Skoog (MS) salt, 1.5% sucrose, and 0.8% Phytagel (Sigma, St. Louis, MO, USA) at pH 5.7. Seeds were first surface sterilized and arranged on the surface of the solid medium, and were given a cold treatment at 4 °C for 72 h. Seeds were germinated and seedlings were grown on a light shelf under a 16 h/8 h light/dark cycle for 12 d, and the seedlings were then transferred into a well-watered potting mix (FAERDIGBLANDING SUBSTRATE, Pindstrup Inc., Danmark). Light was supplied by cool- and warm-white fluorescent bulbs, reaching an intensity approximately 100 μmol m−2 s−1 on the surface of the shelf. The mms21 and MMS21-OE plants are the mms21-1 and 35S-MMS21-GFP lines from the study by Huang et al. (2009).

ABA treatments and seed germination, cotyledon greening, root growth measurements

Plants were grown in the same conditions and seeds were collected at the same time. For each comparison, seeds were planted in the same plate containing MS medium (0.5×MS salts, 1% sucrose, 0.8% agar) without or with different concentrations of ABA. Plates were chilled at 4 °C in the dark for 2 d (stratified) and moved to 22 °C with a 16-h light/8-h dark cycle. The percentage of seed germination was scored at indicated times. Germination was defined as an obvious emergence of the radicle through the seed coat. Cotyledon greening is defined as obvious cotyledon expansion and greening. After germination for 2 d, the seedlings were transferred to the medium with or without 10 μM ABA and cultured vertically under light for 7 d, after which the relative root growth on the ABA medium was measured and compared to that on the MS medium. All experiments were performed with three biological replicates.

Drought treatment and measurement of transpiration rate

For the soil-grown plant drought tolerance test, one-w-old seedlings were transplanted to the soil for two w under standard growth conditions, and plants were then subjected to progressive drought by withholding water for specified times. To minimize experimental variations, the same numbers of plants were grown on the same tray. The entire test was repeated a minimum of three times. To measure the transpiration rate, detached fresh leaves were placed with the abaxial side up on open petri dishes and weighed at different time intervals at room temperature. Leaves of similar developmental stages (the third to the fifth rosette leaves) from 3-w-old soil-grown plants were used. All experiments were performed with three biological replicates.

Stomatal aperture measurements

Stomatal aperture was measured according to a previously-published procedure, with minor modifications (Zhang et al. 2004). Epidermal peels were stripped from fully expanded leaves of 2-w-old plants, and were floated in a solution of 30 mM KCl and 10 mM 2-(N-morpholine)-ethanesulphonic acid (MES-KOH; pH 6.15) in Petri dishes. After incubation for 3 h under cool white light (150–200 μM m−2 s−1) at 22 °C to induce stomatal opening, appropriate concentrations of ABA were added. Stomatal apertures were recorded under a light microscope (BX51; Olympus,, Japan). Measurements were performed using the free software DIGIMIZER (, and all experiments were performed with three biological replicates.

Microarray analyses

Total RNA of the experimental (7-d-old mms21 mutant) and reference (WT) were extracted with the TRIzol reagent (Invitrogen, USA) and purified with a NucleoSpin® RNA clean-up Kit (MACHEREY-NAGEL, Germany) following the manufacturer's instructions. The Arabidopsis genome-wide long oligonucleotide microarray was constructed in-house at the CapitalBio Corporation (Beijing, China). Briefly, 5′-amino-modified 70-mer probes representing 26,173 Arabidopsis genes from the Arabidopsis Genome Oligo Set Version 3.0 (Operon), and internal and external controls, were printed on amino silane-coated glass slides using a SmartArray microarrayer (CapitalBio, China). Fluorescent-labeled DNA (Cy3 and Cy5-dCTP) was produced through Eberwine's linear RNA amplification method and subsequent enzymatic reaction. Briefly, double-stranded cDNA containing a T7 RNA polymerase promoter sequence was synthesized with 1 μg of total RNA using a Reverse Transcription System, RNase H, DNA polymerase I and T4 DNA polymerase, according to the manufacturer's recommended protocol (CapitalBio). The resulting labeled DNA (labeled control and test samples) was quantitatively adjusted based on the efficiency of Cy-dye incorporation, and mixed into 80-μL hybridization solution (3× SSC, 0.2% SDS, 25% formamide and 5× Denhart's solution). Individually labeled cRNAs were not pooled before hybridization. DNA in hybridization solution was denatured at 95 °C for 3 min prior to loading onto a microarray. Arrays were hybridized using a CapitalBio BioMixerTM II Hybridization Station overnight at a rotation speed of 8 rpm and a temperature of 42 °C, and were washed with two consecutive solutions (0.2% SDS, 2× SSC at 42 °C for 5 min, and 0.2× SSC for 5 min at room temperature). Arrays were scanned with a confocal LuxScan 10K-A scanner, and the images obtained were then analyzed using LuxScanTM 3.0 software (Both from CapitalBio). For individual channel data extract, faint spots were removed, for which intensities were below 400 units after the background was subtracted in both channels (Cy3 and Cy5). A space- and intensity-dependent normalization based on a LOWESS program was employed. To determine the significant differentially-expressed genes, Significance Analysis of Microarrays was performed using one class comparison in the Significant Analysis of Microarray software (SAM, version 3.02). Genes were determined to be significantly differentially-expressed with a selection threshold of false discovery rate, FDR < 5% and fold change > 2.0 in the SAM output result, and all these genes are listed in Table S1. Number of biological replicates = 3.

Plant treatments and proline content measurement

For the MMS21 expression level detection, 10-d-old WT Arabidopsis seedlings were treated with 10 μM ABA, 20% PEG or dehydration for 3 h or 6 h. For the phenotype of plant sensitivity to the PEG, the 10-d-old plants were transplanted to the medium with 10% PEG for 7 d, and the proline content was then measured. The free proline content was measured in seedlings by colorimetric assay according to the method described by Bates (1973). For the stress-responsive gene expression analysis, all the MMS21 genotype plants were treated with 100 μM ABA for 3 h or 6 h.

Gene expression analysis

Real-time PCR was performed using the ABI Prism 7300 Fast Real-time PCR system (Applied Biosystems Inc., USA) with SYBR Premix Ex Taq (Takara Bio, Inc., Japan). Total RNA was extracted as described above. cDNAs were synthesized from 0.5 μg of total RNA using a PrimeScriptTM RT reagent Kit (Perfect Real Time, Takara Bio, Inc.). Each PCR reaction contained 1× SYBR Premix Ex Taq, 0.2 μM of each primer, and 2 μL of a 1:10 dilution of the cDNA in a final volume of 20 μL. The following PCR program was used: initial denaturation, 95 °C, 15 s; 40 cycles of 95 °C for 4 s, 60 °C for 15 s and 72 °C for 31 s. In a melting curve analysis, PCR reactions were denatured at 95 °C, re-annealed at 60 °C, and then we monitored the release of intercalators from PCR products or primer dimers by an increase to 95 °C. cDNA quantities were calculated by ABI Prism 7300 SDS Software Ver.1.3 (Applied Biosystems Inc.), and transcript data were normalized using the UBQ10 gene as an internal control. Error bars were presented to indicate the standard error of the mean. All experiments were performed with three biological replicates. The primers used in this paper for Quantitative RT-PCR are listed in Table S2.

Analysis of sumo-conjugation proteins

Total protein of seedlings incubated at 24 °C or treated by sterile water with or without 10 μM and 100 μM ABA for 3 h were extracted and separated by SDS–PAGE. The gel blot was probed with the SUMO1 antibody (Bioworld,, USA) and detected using ECL plus (Amersham Pharmacia,, USA).

Accession numbers

Sequence data from this article can be found in The Arabidopsis Information Resource database (TAIR) under the following accession number: MMS21, At3g15150.

(Co-Editor: Giovanna Serino)


This work was supported by the National Natural Science Foundation of China (30900789, 31170269), the Ministry of Agriculture (National Key Program for Transgenic breeding 2009ZX08009-006B) and the Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme 2010. We gratefully acknowledge the Arabidopsis Biological Resource Center (ABRC) (Ohio State University) for the mutant lines.