The Mildew Resistance Locus O 4 Interacts with CaM/CML and Is Involved in Root Gravity Response

The plant-specific mildew resistance locus O (MLO) proteins, which contain seven transmembrane domains and a conserved calmodulin-binding domain, play important roles in many plant developmental processes. However, their mechanisms that regulate plant development remain unclear. Here, we report the functional characterization of the MLO4 protein in Arabidopsis roots. The MLO4 was identified as interacting with CML12 in a screening for the interaction between the proteins from Arabidopsis MLO and calmodulin/calmodulin-like (CaM/CML) families using yeast two hybrid (Y2H) assays. Then, the interaction between MLO4 and CML12 was further verified by Luciferase Complementation Imaging (LCI) and Bimolecular Fluorescence Complementation (BiFC) assays. Genetic analysis showed that the mlo4, cml12, and mlo4 cml12 mutants displayed similar defects in root gravity response. These results imply that the MLO4 might play an important role in root gravity response through interaction with CML12. Moreover, our results also demonstrated that the interaction between the MLO and CaM/CML families might be conservative.


Introduction
Mildew resistance locus O (MLO) was first identified as being resistant against powdery mildew (PM) infection in barley [1]. Then, the MLO proteins were found in many plant species. Arabidopsis has at least 15 MLO proteins that diverged into several clades and were involved in different physiological processes, such as sexual reproduction, PM infection, and root thigmomorphogenesis [2][3][4][5]. MLO7/NORTIA (NTA), a downstream component of the receptor-like kinase FERONIA (FER), is involved in pollen tube reception and PM infection [6]. Study showed that its function is dependent on NTA homooligomerization and its carboxyl-terminal tail identity [7]. MLO5/9/15 played roles in pollen tube responses to ovular signals, and MLO5/9 selectively recruit Ca 2+ channel CNGC18-containing vesicles to the plasma membrane [8]. The MLO1, MLO2, MLO3, MLO6, MLO8, MLO9, MLO10, and MLO13 are expressed in discrete domains during reproductive development [4]. The MLO2, MLO4, and MLO6 modulate defense responses against PM fungi and a number of other phytopathogens [9]. Furthermore, the MLO2 also function as a negative regulator in plant ROS responses involving biotic and abiotic stress [10]. Mutations in MLO4 and MLO11 exhibit abnormal root thigmomorphogenesis and gravity sensitivity [3,11]. Genetical complementation with MLO4 domains indicate that the C-terminal cytoplasmic domain of MLO4 is necessary for regulation of asymmetrical root growth [11].
All MLO proteins are predicted to share a conserved calmodulin-binding domain (CaMBD) in C-terminal cytoplasmic tail [12,13]. The gel overlay assays demonstrated in vitro Ca 2+ -dependent binding of CaM to the CaMBD of the MLOs from barley [14]. However, little is known about the interaction of individual MLO with the specific calmodulins (CaMs) or calmodulin-like proteins (CML) in Arabidopsis or other plants.
MLO4 and CML12 were expressed in many tissues, including roots. Although the highest expression levels of both MLO4 and CML12 did not present in roots, they were highly coexpressed in roots, implying both of them may function in roots.

LCI and BiFC Assays Further Demonstrated That MLO4 Interacted with CML12
At first, MLO4 and CML12 were chosen to verify the results of Y2H screening ( Figure  S1). Then, the LCI results showed that CML12 interacted with the full length of MLO4 in tobacco cells (Figure 2a). BiFC results demonstrated that the CML12 interacted with the cytoplasmic domain of MLO4, in concert to Y2H result. Furthermore, when the full length of MLO4 was used for the assays, the interactive signal was detected only on the plasma membrane, which was co-localized with the staining signals of the membrane dye FM 4-64 ( Figure 2b). This result was consistent to the ones obtained from the membrane localization of MLO4 with a signal peptide and seven transmembrane domains [11].

MLO4 and CML12 Displayed a Similar Expression Pattern
To understand the relationship between MLO4 and CML12, we further investigated whether MLO4 and CML12 have a co-expression pattern in vivo. First, GUS staining of pCML12::CML12-GUS transgenic lines showed that CML12 was expressed in most of the vegetable tissues, such as seedlings, roots, root hairs, leaves, trichomes, stems, siliques, calyxes, and chalazal end of ovules but not in pollen (Figure 3a-j). In addition, pCML12::CML12-GFP transgenic lines showed that the CML12-GFP fusion protein was detected in the root cap ( Figure 3k). Real-time PCR results showed that MLO4 and CML12 shared a similar expression pattern. Both genes were expressed constitutively in many tissues, including seedlings, roots, stems, leaves, inflorescence, and siliques. Expression levels of both genes were closely similar in seedlings. MLO4 was expressed with twice the CML12 expression in roots (Figure 3l). A previous study revealed that the pMLO4::MLO4-GFP was also expressed in the roots, and MLO4-GFP fusion protein was detected in the root epidermal cell [11]. Therefore, both CML12 and MLO4 proteins were expressed in roots, suggesting that they may function coordinately.

MLO4 and CML12 Displayed a Similar Expression Pattern
To understand the relationship between MLO4 and CML12, we further investigated whether MLO4 and CML12 have a co-expression pattern in vivo. First, GUS staining of pCML12::CML12-GUS transgenic lines showed that CML12 was expressed in most of the vegetable tissues, such as seedlings, roots, root hairs, leaves, trichomes, stems, siliques, calyxes, and chalazal end of ovules but not in pollen (Figure 3a-j). In addition, pCML12::CML12-GFP transgenic lines showed that the CML12-GFP fusion protein was detected in the root cap ( Figure 3k). Real-time PCR results showed that MLO4 and CML12 shared a similar expression pattern. Both genes were expressed constitutively in many tissues, including seedlings, roots, stems, leaves, inflorescence, and siliques. Expression levels of both genes were closely similar in seedlings. MLO4 was expressed with twice the CML12 expression in roots (Figure 3l). A previous study revealed that the pMLO4::MLO4-GFP was also expressed in the roots, and MLO4-GFP fusion protein was detected in the root epidermal cell [11]. Therefore, both CML12 and MLO4 proteins were expressed in roots, suggesting that they may function coordinately.

The mlo4 and cml12 Mutants Shared Similar Phenotypes
The MLO4 has been characterized as regulating root thigmomorphogenesis, including gravitropism in roots [11]. Previous characterization of cml12 revealed that the mutations in CML12 exhibited a similar phenotype [30]. To study the genetic relationship between MLO4 and CML12, we further characterized their roles in root growth.
The cml12-3 mutant (SALK_122731) was obtained from Dr. Sherryl R. Bisgrove (Department of Biological Sciences, Simon Fraser University, Burnaby, BC, Canada). Real-time PCR showed that cml12-3 was a knock-out mutant, as reported by Gleeson et al. [31] ( Figure S2a,d). To test its gravity response, cml12-3 seedlings were vertically grown on the agar minimal medium (0.25% sucrose) for 7 d (16 h light/8 h dark cycle); then, the plates were rotated by 90 • in the clockwise direction. The roots bent down in response to gravity change until root tips became parallel with new gravity vector ( Figure 4a). Compared to the time (12.60 ± 0.66 h) for the wild-type roots to form a bend, the cml12-3 roots took less time (9.33 ± 0.75 h) to form a bend (Figure 5a), suggesting that cml12-3 roots were more sensitive to gravity response.
To verify the phenotype of cml12-3 in gravity response, a complementation assay was performed using the construct pCML12::CML12-GFP in PCAMBIA 1300 vector containing the native promoter (1500 bp upstream ATG) and coding sequence of CML12 (975 bp). Two pCML12::CML12-GFP transgenic cml12-3 mutant lines were chosen to further characterize their root phenotype. The average times for the roots from each pCML12::CML12-GFP (cml12-3/-) transgenic lines to form a bend were 12.10 ± 1.42 h and 12.90 ± 0.96 h, respectively, similar to that of wild-type roots (12.60 ± 0.66 h) (Figure 4a). The real-time PCR results showed that CML12 expression was higher in transgenic roots than that in cml12-3 mutant roots (Figure 5a). Therefore, CML12 was able to restore the defect of cml12-3 mutants in gravity response, further demonstrating that CML12 is involved in gravity response.
The previous study showed that mlo4-1, mlo4-3, and mlo4-4 mutants exhibited different growth patterns on the reclined medium than wild-type [11]. In this study, we identified another Salk allele mlo4-5 in which the T-DNA was inserted in the last exon of MLO4 ( Figure S2b), different from the previously-reported alleles ( Figure S2c). Real-time PCR showed that mlo4-5 was a knock-out mutant and expressed truncated protein ( Figure S2e). At first, an allelic analysis was performed. The F1 progeny of mlo4-5 from the crosses with mlo4-4 exhibited a resembled phenotype in growth pattern ( Figure S3). The loops in each root of mlo4-4/+; mlo4-5/+ plant were denser with shorter wavelengths, and its phenotype was similar to that of single mutant-mlo4-4/-and mlo4-5/-. The result demonstrated that mlo4-5 also was an effective mutant of MLO4 gene. Therefore, the mlo4-5 was used for further characterization.
Starch grains in root tips are sensors of gravity response [30]. Therefore, the starch stain assays were further performed to compare the starch contents in the mutants and wild-type root tips using KI-I 2 staining. Contents of the starch grains in the single mutant cml12-3 and mlo4-5, as well as the double mutant cml12-3 mlo4-5, were obviously higher than that in wild-type root tips (Figure 4c).

Mutation in MLO4 Affected the Expression of CML12
CaM/CMLs bound calcium and were supposed to regulate target proteins downstream [16]. To further study the possible regulatory relationship between CML12 and MLO4, their relative expression levels in each other mutants were analyzed. The real-time PCR assay showed the expression of CML12 was significantly elevated in mlo4-5 mutant, far more than MLO4 expression changing in cml12-3 mutant (Figure 5d). Thus, MLO4 likely modulates the expression of CML12.

MLO4 and CML12 Function Coordinately in Root Thigmomorphogenesis
In this study, the cml12-3 roots were more sensitive to gravity, as demonstrated by taking less time for them to form a bend in gravity response assays. Moreover, the CML12 gene could restore the cml12-3 phenotype. These results suggest that CML12 regulates root gravity response. The transgenic plants of pCML12::CML12-GUS and pCML12::CML12-GFP showed that CML12 protein was expressed in root, especially in root cap. Previous studies showed that gravity sensing happens in the root cap and is related to the starch sediment in the root cap [32][33][34][35][36]. Starch-filled amyloplasts sediment settled in the bottom of statocytes under gravity [35]. In starch-deficient mutants, amyloplasts did not precipitate and gravity response was deficient [37]. Characterization of both starch-less and starch-more mutants suggested that starch amount was directly associated with gravity sensing [30]. In cml12-3 and mlo4-5 mutants, starch grains detected by I 2 -KI stain were more than those in wild-type. Both mutants exhibited more sensitive gravity sensing than wild-type. Therefore, MLO4 and CML12 may be involved in gravity sensing through regulating starch content in root caps.
In the mlo4-5 mutant, CML12 expression was dramatically upregulated, implying that MLO4 may negatively modulate the expression of CML12. The mlo4-5 and cml12-3mutants shared a similar growth pattern that exhibited a more sensitive response to gravity on the reclined medium plates in comparison with that of wild-type plant (Figures 4a and 5a). Moreover, MLO4 and CML12 could interact with others. These results indicate that MLO4 and CML12 may function coordinately to regulate root thigmomorphogenesis. However, it is unclear how loss of MLO4 could promote expression of CML12. CML12, also called TOUCH3/TCH3, is a touch-induced gene whose mRNA increases massively within seconds in response to various environmental stimuli, including touch, wind, and darkness [38]. A newly changed gravity vector may be a stimulus upregulating CML12 expression, or there may be also other unknown regulators involved in regulation of CML12 expression. Nevertheless, more studies are required to explain the regulatory mechanism of MLO4 and CML12.

The Conservative Amino Acids of CaMBDs and C-terminal Sequence of MLO Protein May Determine Their Specific Interacting CaM/CMLs
MLO proteins are encoded by medium-sized gene families present in all investigated land plant species and they share a conserved CaMBD in cytoplasmic C-termini, which consists of 20 amino acids [39,40].
As reported previously, the barley MLO C-terminus interacted with soy bean CaM1. When the conservative amino acids (L420R, W423R) in CaMBD were mutated, the interaction of mutated HvMLO with SCaM1 was weak and even became undetectable [14]. These results suggest that the MLO-CML interaction is likely to depend on the specific CaMBD in C-terminus of MLO proteins. The C-terminal sequences of MLO proteins are obviously diversified. Studies showed that C-terminal regions of the MLO proteins could determine specific binding to the individual CaM/CML proteins to fulfill their distinct biofunction [13,41,42]. Domain-swap experiments testified that the identity of the C-terminal intracellular tail contributes to MLO function [43].
As confirmed in our study, the Y2H assays also showed that, for each MLO protein, only the C-terminus region was sufficient for binding the CaM/CML proteins (Table 1). Each MLO might interact with more than one CaM/CML proteins. The specific interaction between MLO and CaM/CML proteins are due to CaMBD of the MLO C-terminus. First, Y2H screen results revealed that C-terminus region of NTA/MLO7 interacted with CML8-12 ( Figure S5). Then, Y2H assays was conducted with site mutation in CaMBD of Cterminus. NTA variants with modification of the conservative sites (L452R, W455R) could not interact with CML12, while NTA variants with changes in the non-conservative sites (K457N, L464R) remained able to interact with CML12. All NTA variants (L452R, W455R, K457N, L464R) failed to interact with CML8 or CML9 but still bound CML10 and CML11 ( Figure S6). Therefore, the invariant and conservative amino acids may enable CaMBD to bind CaM/CMLs. Meanwhile, the variant amino acids and C-terminal sequence may determine their specific interaction with CaM/CMLs, with the former directly affecting binding ability and the latter indirectly influencing MLO-CaM/CML spatial distance via conformation.

MLO-CaM/CML Interaction May Relate to Ca 2+ Signaling and Mediate Vesicle Transport
The Ca 2+ dynamics were detected in various cells upon responding to extracellular signals, for instance, the root development and root hair growth, the synergid cells sensing the arriving pollen tube, and vice versa, the pollen tube tips sensing ovules during sexual reproduction [8,[44][45][46][47]. Meanwhile, more and more evidence has pointed out that vesicle transport was revealed in the corresponding root tips, synergid cells, and pollen tube tips [43,46,48].
The CaM/CML proteins distinguish themselves with EF-hands that can bind Ca 2+ , a second messenger generated by various stimulations [49]. Previous study has demonstrated that CML12 could bind to Ca 2+ in vitro [38]. Additionally, nuclear Ca 2+ signatures are detected in root cells and associated with root development [46,47]. MLO4 protein was located in vesicle membrane of root cells, and its subcellular localization responded to vesicle traffic inhibitor BFA. The expression of auxin efflux carrier fusion, PIN1-GFP, was altered at root tip of mlo4 mutant seedlings, implying that MLO4 protein participated in auxin transport [11]. Now that MLO4 interacts with CML12 (Figure 2), this case provides the first scattered clues where Ca 2+ signaling and vesicle transport may be related by MLO4-CML12 interaction.
The second case is Ca 2+ oscillations displayed in the synergid cells and the FER-NTA pathway is required for Ca 2+ responses [44,45,48]. NTA/MLO7 protein is located in vesicular-membrane and evenly distributed in synergid cells before arrival of the pollen tube. Upon pollen tube arrival, it is then polarly transferred to the micropylar end. In this way, MLO might deliver out certain signal through vesicle traffic by responding to the approaching pollen tube [6]. The interaction between NTA/MLO7 and CML8-12 was also verified by Y2H, BiFC, and LCI assays ( Figures S5-S8), although later genetic analysis showed that the quadruple mutant cml8 cml9 cml11 cml12 has normal fertility, not similar to that of the nta/mlo7 mutant, indicating the possible functional redundancy of the CML proteins (CML10/26/30/37/38/40/44) (Table 1). Therefore, NTA/MLO7 protein may mediate Ca 2+ signaling and vesicle transport by interacting with CMLs during pollen tube reception.
The third case is that Ca 2+ formed a tip-focused gradient in the pollen tube and underwent oscillation in the tip region during pollen tube growth [8,43]. The pollenspecific MLO5/9 was distributed on vesicle membrane in the tip region of the pollen tube and directly bound Ca 2+ channel CNGC18, which recruited CNGC18-containing vesicles to the plasma membrane through the R-SNARE proteins VAMP721 and VAMP722 in trans mode, in response to ovular signals [8]. In short, MLO5/9 mediated the exocytosis of a Ca 2+ channel to reply to ovular guidance. In addition to that, our Y2H screen disclosed that MLO5 interacted with CML18/26/40/44, MLO9 interacted with CML10/42/44 (Table 1), which, of course, needs further verifying. Therefore, it is rational to deduce that MLO5/9-CML interaction may relate to Ca 2+ signaling and mediate vesicle transport.
In conclusion, responding to extracellular signals, MLO proteins seem to modulate SNARE-dependent vesicle transport-associated processes at plasma membrane, in close relation to Ca 2+ signaling, possibly through interacting with CaM/CMLs.
The MLO proteins are unique to plants [40]. Although some mlo mutants have been characterized, the mechanisms of the modulation of CaM by MLO proteins remain poorly understood. The biochemical function of MLO in modulating Ca 2+ -loaded CaM/CML also remains elusive.

Plant Materials and Growth Conditions
Arabidopsis thaliana plants in Columbia (Col) background were used in this study. The cml12-3 (SALK_122731) was a gift from Dr. Sherryl R. Bisgrove [31]. The primer pairs used for the mutant characterization were LBa1 and SK122731-1 (Table S3). The mlo4-4 (SAIL_266_C02) and mlo4-5 (SAIL_563_E06) were purchased from the Arabidopsis Biological Resource Center (ABRC, www.arabidopsis.org, accessed on 19 May 2021). The primer pairs for identifying mlo4-4 and mlo4-5 were SAIL-LB3/SAIL266C02-1 and SAIL-LB3/SAIL563E06-1, respectively (Table S3). The seeds were surface-sterilized in the solution containing NaClO:ddH 2 O (1:8) for 5 minutes and washed with sterilized ddH 2 O three times. Then, the seeds were sowed on 1/2 MS culture medium and stood in 4 • C for 2-4 days. The vernalized seeds on 1/2 MS were cultured under a 16 h/8 h day/night cycle at 22 • C for 7-10 days for further growth. The resulting seedlings were then planted into soil under a 16 h/8 h day/night cycle at 22 • C as previously described [50].

Root Phenotype Analysis
For gravity response, the seeds were planted on the agar medium (1/2 MS, 0.8% Agar, 0.05% MES, 1% Sucrose, pH 5.8) and grown inside the agar on vertical plates for 7 d. The root bend formation times were counted from the moment when the plates were rotated by 90 • in the clockwise direction to the time when root growth became parallel with new gravity vector [31]. For root growth pattern, the seedlings were grown for 2 d on agar-solidified minimal medium (0.25% sucrose) that was vertically set. Then, the plates were slanted at a 45 • angle for 4 days [11].

Real-Time PCR Assay
The RNA samples were extracted from different tissues of Arabidopsis plants using a total RNA extraction kit (TIANGEN, Beijing, China). The cDNA templates were prepared using SuperScript II kit (Invitrogen, http://www.invitrogen.com/, accessed 2 July 2019). Using ABI 7500 real-time system (Applied Bioystems, http://www.appliedbiosystems. com/, accessed 2 August 2019) and the Power SYBR Green PCR Master Mix, all the realtime-PCR assays were performed with the primers listed in Table S3. The RNA input was normalized with the ACTIN2/8 (At3g18780) RNA levels. The mRNA levels were calculated with comparative Ct (threshold cycle) method as previously described [50].

GUS Assay for Expression Pattern
Because the CML12 coding sequence (CDS) had three regions that shared high similarity [38], it was difficult to clone the full sequence of CML12 CDS. For the CDS, 18 bp upstream the start codon ATG was included and the promoter sequence was from 1500 bp upstream start codon to 18 bp upstream start codon. The 1482 bp promoter fragment and 993 bp coding sequence of CML12 were amplified using the primer pairs HindIII-PCML12/PCML12-XbaI and XbaI-TCH3/TCH3-BamHI (Table S3) from Col genomic DNA and cDNA, respectively. The resulting fragments were then inserted into the modified pCAMBIA1300 vector (CAMBIA, http://www.cambia.org/, accessed on 13 June 2020) to generate the construct pCML12::CML12-GUS. After being verified by sequencing, the resulting construct was introduced into Col plants using the Agrobacterium-mediated infiltration method [51]. GUS staining was performed as described previously [52].

Subcellular Localization of CML12 Protein
The 993 bp coding sequence DNA and 1482 bp promoter of CML12 were amplified by PCR using the primer pairs HindIII-PCML12/PCML12-PstI and PstI-TCH3/TCH3-BamHI listed in Table S3. These two fragments were then inserted into the Ti-derived pCAMBIA1300 vector (CAMBIA, http://www.cambia.org/, accessed on 13 June 2020) to generate the construct pCML12::CML12-GFP. The construct was then introduced into cml12-3 mutant plants. The GFP signals in the root epidermal cells were detected using the Zeiss confocal microscope (LSM710 Meta; Zeiss, http://www.zeiss.com, accessed on 21 March 2020).

Yeasts Two Hybrid (Y2H) Assay
The DNA fragments of CML12 and cytoplasmic domain of MLO4 (MLO4C) were amplified by PCR using the primer pairs NdeI-CML12/CML12-BamHI and NdeI-MLO4C/MLO4C-ECoRI listed in Table S3. The resulting fragments were inserted into the assay vector pairs pGADT7 (AD) and pGBKT7(BD) (Clontech, www.clontech.com/, accessed on 21 October 2020), respectively, as described previously [53]. The yeast strain AH109 was co-transformed with generated construct pairs. SD/-Leu-Trp plates were used to culture the transformed yeast cells for 3 days at 30 • C. Then, SD/-Trp-Leu-His-Ade plates were used to select the colonies with interacting protein pairs for 3 to 7 days.

Luciferase Complementation Imaging (LCI) Assay
The coding sequences of CML12 and MLO4 were amplified by PCR using the primers KpnI-CML12/CML12-SalI and KpnI-MLO4/MLO4-SalI (Table S3). These fragments were then subcloned into the vectors pCAMBIA1300-nLUC and pCAMBIA1300-cLUC, respectively. The positive control pair was SGT1-nLUC and RAR1-cLUC, which was a kind gift from Dr. Shuhua Yang (China Agricultural University, Beijing, China). Applying electroporation, all the constructs were introduced into A. tumefaciens strain GV3101, which was then injected into 3-week-old tobacco leaves to transiently transform the tobacco cells. Under the detection of the low-light cooled charge-coupled device camera (iKon-L936; Andor Technology, http://www.andor.com/, accessed on 21 April 2020), the fluorescence signal was detected 3-6 days after infiltration, as previously described by Chen et al. [54] and Hua et al. [55].