An endoplasmic reticulum magnesium transporter is essential for pollen development in Arabidopsis

Plant Science 231 (2015) 212–220

Contents lists available at ScienceDirect

Plant Science journal homepage: www.elsevier.com/locate/plantsci

An endoplasmic reticulum magnesium transporter is essential for pollen development in Arabidopsis Jian Li a , Yuan Huang a , Hong Tan a , Xiao Yang a , Lianfu Tian a , Sheng Luan b,c , Liangbi Chen a,∗ , Dongping Li a,∗∗ a

College of Life Sciences, Hunan Normal University, Changsha 410081, PR China NJU-NJFU Joint Institute for Plant Molecular Biology, School of Life Sciences, Nanjing University, Nanjing 210093, PR China c Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720, USA b

a r t i c l e

i n f o

Article history: Received 5 September 2014 Received in revised form 11 November 2014 Accepted 8 December 2014 Available online 15 December 2014 Keywords: Endoplasmic reticulum Magnesium transporter Pollen abortion

a b s t r a c t Magnesium is one of the essential macro-elements for plant growth and development, participated in photosynthesis and various metabolic processes. The Mg-transport abilities of the AtMGT (Magnesium Transporter) genes were identified in bacteria or yeast mutant system. In our previous studies, both the AtMGT5 and AtMGT9 were found essential for pollen development in Arabidopsis. Here we report another AtMGT member, AtMGT4, which was localized to the endoplasmic reticulum, was essential for pollen development as well. AtMGT4 expressed notably in pollen grains from bicellular pollen stage to mature pollen stage. A T-DNA insertional mutant of the gene, named mgt4-1, showed pollen abortive phenotype, thus we could not get any homozygous mutant from progenies of self-crossed +/mgt4-1 plants. Meanwhile, nearly half of pollens in AtMGT4-RNAi transgenic lines were sterile, consistent with the phenotype of +/mgt4-1 mutant. Transgenic plants expressing AtMGT4 in the mgt4-1 background could recover the pollen fertility to the wild type. Together, our findings demonstrated that the disruption of AtMGT4 in Arabidopsis could cause a defect of pollen development. The visible pollen abortion appeared at bicellular pollen stage in +/mgt4-1. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Magnesium is one of the macro-elements essential for plant growth and development, involved in photosynthesis and various metabolic pathways [1]. The CorA proteins were first found in bacteria, conferring resistance to cobalt, and function as magnesium transporters [2–4]. CorA proteins have two closely spaced C-terminal transmembrane (TM) domains. The first TM domain ends with a GMN (Gly-Met-Asn) tripeptide motif and appears to be a conservative part of the entry pore for the Mg2+ [5]. The crystal structure of the Thermotoga maritima CorA protein has been identified as a pentamer with the pore lined by the first of the TM

Abbreviations: EGFP, enhanced green fluorescent protein; ER, endoplasmic reticulum; Mg, magnesium; PI, propidium iodide; RFP, red fluorescent protein; RNAi, RNA interference; SEM, scanning electron microscope; UDP, uridine 5 -diphosphate; UV, ultraviolet. ∗ Corresponding author at: No. 36, Lushan Road, Yuelu District, Changsha City, Hunan Province, PR China. Fax: +86 731 8887 2617. ∗∗ Corresponding author at: No. 36, Lushan Road, Yuelu District, Changsha City, Hunan Province, PR China. Fax: +86 731 8887 2724. E-mail addresses: [email protected] (L. Chen), [email protected] (D. Li). http://dx.doi.org/10.1016/j.plantsci.2014.12.008 0168-9452/© 2014 Elsevier Ireland Ltd. All rights reserved.

domains of each monomer [6,7]. Homologous proteins of CorA type are identified in a range of different species in archaea, eubacteria, and eukaryotes [8–11]. The best studied in the eukaryotic CorA homologs are Mg2+ transporters in Saccharomyces cerevisiae. The first yeast Mg2+ transporters identified were named as ALR1 and ALR2, conferring to aluminum resistance of the yeast [12]. Thereafter, a mitochondrial Mg2+ transporter was identified, named as MRS2 for its ability to correct an RNA splicing defect [9]. Other CorA homologs in yeast were identified one after another, such as LPE10 (a second mitochondrial Mg transporter) [13], MNR2 (necessary for cells to access Mg stored in the vacuole) [14]. In Arabidopsis, there are about 11 CorA-like genes including two pseudogenes, named AtMRS2 [15] or AtMGT [16]. AtMGTs can heterologously complement bacteria or yeast Mg2+ transport defective mutants [15,16]. The individual of the AtMGT members could reestablish Mg2+ uptake in mitochodria of yeast mrs2 mutant to varying degrees [17]. In recent years, Mg2+ transport properties of each AtMGT members are analyzed preliminarily in Arabidopsis. The physiological functions of AtMGT proteins are elaborated by studying gene knockout mutants. A strong magnesium-dependent phenotype of growth retardation was first found in atmgt7/atmrs27 mutant [17], then in AtMGT6/AtMRS2-4 RNAi transgenic plants

J. Li et al. / Plant Science 231 (2015) 212–220

[18]. AtMGT7/AtMRS2-7 was located to the endomembrane system [17], and AtMGT6/AtMRS2-4 in plasma membrane, participating in cellular Mg homeostasis and root Mg-uptake respectively [18]. AtMGT2/AtMRS2-1 and AtMGT3/AtMRS2-5 are targeted to the tonoplast and corresponding T-DNA insertion lines have perturbed mesophyll-specific vacuolar magnesium accumulation under serpentine conditions [19,20]. Some other AtMGT members have not yet been linked their physiological functions with magnesium transport [21,22]. However, gene knockout mutants of these members showed distinguishing phenotype of pollen development defect. AtMGT5/AtMRS2-6, locating in mitochondria in Arabidopsis, is essential for pollen development [23]. Another magnesium transporter AtMGT9/AtMRS2-2 is also identified to be essential for pollen development [24]. Here, we report another AtMGT member, AtMGT4/AtMRS2-3, which is located to endoplasmic reticulum (ER), is essential for pollen development. 2. Materials and methods 2.1. Plant material and growth conditions The mgt4-1 mutant (Ler. ecotype, CS170677), was ordered from ABRC. Seeds were sterilized and stratified for 2 days at 4 ◦ C. Then, seeds were sown on 1/2 MS medium containing 50 ␮g/ml kanamycin [25]. Heterozygotes of mgt4 mutants were selected on 1/2 MS medium based on their kanamycin resistance. Seven days after mgt4-1 mutants’ germination, seedlings were transferred to vermiculite with a photoperiod of 16 h light and 8 h darkness and a consistent temperature of 23 ◦ C. +/mgt41 mutants were also identified by PCR, primers used for the PCR were as follows: 4F (5 -ATGAGAGGAGCTAGACCCGATGAAT-3 ); 4R (5 -TCATTCAAGAAGGCGCTTGTACTTG-3 ); Ds3-1 (5 -ACCCGACCGGATCGTATCGGT-3 ). The insertion site of the mgt4-1 mutant was confirmed by sequencing. 2.2. Pollen vitality analysis Dehiscent anthers detached from +/mgt4-1 and wild type plants were placed on glass slides. By using a dissecting needle, pollen grains were released from these anthers with the extrusion of anthers. Pollens were suspended in PBS buffer (pH 7.2) with 50 ␮g/ml Hoechst 33342 dye and kept away from light. After incubation for 15 min at 25 ◦ C, glass slides coated with pollens were placed under Olympus fluorescence microscope BX51 to observe emitting light of pollens’ nuclei. Pollen grains from dehiscent anthers were analyzed under scanning electron microscopy as described previously [26]. 2.3. AtMGT4 expression pattern analysis To determine AtMGT4 expression pattern, a 294 bp DNA fragment between AtMGT4 and its upstream contiguous gene together with 149 bp from the start of AtMGT4 1st exon was amplified by PCR using the primers P4-F (5 -GCTCTAGACGACTCTTCCTTGGGAAT-3 ) and P4-R (5 -GCTCTAGAAGTACCAACCACGTCCT-3 ). The amplified AtMGT4 promoter (AtMGT4pro ) was digested with Xba I, and inserted into the multiple cloning site of the plant binary vector pBI101.2 to drive the expression of GUS reporter. The method of delivering the recombinant plasmid into Arabidopsis was described previously. Transgenic plants carrying GUS reporter driven by AtMGT4pro were screened on 1/2 MS mediums containing 50 ␮g/ml kanamycin. AtMGT4pro activity was visualized in different tissues of plants via histochemical staining of GUS. AtMGT4pro ::GUS transgenic

213

plants were immersed in staining buffer containing 0.1 M phosphate buffered sodium (pH 7.0), 10 mM EDTA, 2 mM potassium ferrocyanide, 0.1% (v/v) Triton X-100, 20% (v/v) methanol, 0.5 mM X-gluc (5-bromo-4-chloro-3-indolyl glucuronide), and incubated at 37 ◦ C overnight. After staining, tissues were fixed in Carnoy’s Fluid for further analysis. 2.4. Subcellular localization of AtMGT4 To identify the subcellular localization of AtMGT4, recombinant vector expressing AtMGT4-EGFP fusion protein was constructed. AtMGT4 CDS was amplified using primers MGT4-GFP-F (5 AACTCGAGATGAGAGGAGCTAGACCCGA-3 ) and MGT4-GFP-R (5 CGCAATTGCTTCAAGAAGGCGCTTG-3 ). Amplified fragments were digested with Xho I and Mfe I, and inserted into transient expression vector pEZS-NL. The analysis of transient gene expression in Arabidopsis mesophyll protoplasts was described previously. AtMGT4-EGFP and Bip-RFP (an endoplasmic reticulum fluorescent marker) were co-transformed into Arabidopsis mesophyll protoplasts. Fluorescence emitted by AtMGT4-EGFP and Bip-RFP were detected and analyzed by LEICA TCS SP2 laser scanning confocal microscope. EGFP protein is excited by laser at 488 nm, and the emitted fluorescence is collected with a band-pass filter at 490 nm–540 nm. RFP is excited by laser at 543 nm, and the emitted fluorescence is collected with a band-pass filter at 580 nm–650 nm. 2.5. Paraffin sections of anthers Inflorescences of mgt4-1 and wild type plants were vacuum infiltrated and fixed in Carnoy’s Fluid overnight. After that, flowers were gradually dehydrated in progressively increasing ethanol concentrations up to 100%. Samples were hyalo-cleared in 50% xylene and 50% ethanol for 30 min, the hyalo-clearing was continued in 100% xylene twice, 20 min each time. Next, the wax immersion of tissues was performed in 50% xylene and 50% paraffin firstly at 42 ◦ C for 1 h. Subsequently, tissues were transferred into 100% paraffin for 1 h at 60 ◦ C, which was repeated three times. At last, samples were embedded in paraffin, sections of 5 ␮m thickness were cut with knives in a Leica RM2235 rotary microtome, fixed to glass slides, and stained in 0.05% (w/v) hematine. For inflorescence of AtMGT4pro ::GUS transgenic plants, sections were stained in 0.05% (w/v) PI for a short time. 2.6. AtMGT4-RNAi and +/mgt4-1 complementary transgenic plants For preparing AtMGT4-RNAi construct, a 733 bp fragment in the AtMGT4 cDNA sequence, showing low nucleotide sequence identity to other AtMGT genes, was amplified respectively using two pairs of primers Anti-4F (5 -CGTCTAGACCTCAGCTGCTCGATCT-3 ) and Anti-4R (5 -CCCAAGCTTGTCCAGGGCTGGATGA-3 ), Sense-4F (5 CCGCTCGAGTCTAGACCTCAGCTGCTCGATCT-3 ) and Sense-4R (5 CGCAATTGAAGCTTGTCCAGGGCTGGA-3 ). They were subcloned into the pKANNIBAL vector as described previously to assemble the RNAi cassette. After that, AtMGT4 promoter, RNAi cassette and OCS terminator were transferred into the multiple cloning sites of pCAMBIA1300 one after another. Transgenic plants carrying AtMGT4pro ::AtMGT4-RNAi construct were screened on 1/2 MS mediums containing 50 ␮g/ml hygromycin B. For genetic complementation, AtMGT4 promoter, CDS and terminator were spliced with overlap-extension PCR method. The recombinant fragment was then subcloned into pCAMBIA1300 vector. Complementary plants in the +/mgt4-1 background were selected on 1/2 MS mediums containing 50 ␮g/ml hygromycin B and 50 ␮g/ml kanamycin.

214

J. Li et al. / Plant Science 231 (2015) 212–220

Fig. 1. Morphological observation of pollen grains from the wild type and +/mgt4-1 mutant plants. (A) Identification of the +/mgt4-1 mutants based on PCR method. (B) T-DNA insertion site in the +/mgt4-1 mutants. The ATG start codon and TGA stop codon were indicated. The black straight lines showed introns of AtMGT4 and the black boxes showed exons of AtMGT4 in genome. The T-DNA insertion was shown as the triangle above the gene diagram. (C) Hoechst 33342 staining of mature pollen grains from the Ler. ecotype and +/mgt4-1 plants respectively. Fluorescent signaling of the pollens (left panels) excited by UV were observed under a fluorescence microscope. The morphological characteristics of the pollens (right panels) were observed under bright field of the fluorescence microscope. Arrows showed abnormal pollen grains. Bars = 20 ␮m.

2.7. Semi-quantitative RT-PCR Total RNAs from inflorescences were extracted using the Trizol reagent (Sangon) following the manufacturer’s instructions. Samples were treated with DNase (Thermo Scientific) to eliminate genomic DNA contamination. One microgram of DNA-free RNA was used for reverse transcription by M-MLV Reverse Transcriptase (Invitrogen) with anchored oligo(dT18 ). For semi-RT-PCR, 1 ␮l cDNA was used with a total reaction volume of 15 ␮l. AtMGT4 was amplified using primers 4F (5 -ATGAGAGGAGCTAGACCCGATGAAT-3 ) and RT-4-R (5 -AATTCGGGAGGAAGCCTGTC-3 ) followed by 30 PCR reaction cycles. Arabidopsis AteIF4A1 gene was used as a quantification control with primers RT-eIF4A1-F (5 -CAAGGTTCACGCCTGTGTTG-3 ) and RT-eIF4A1-R (5 -ACGTTCGAAGGCAGAGCATC-3 ). 3. Results 3.1. Disruption of the AtMGT4 gene causes a defect of pollen development in Arabidopsis An AtMGT4 T-DNA insertion mutant line was identified from ABRC stock CS170677 and named mgt4-1 (Fig. 1A). The insertional site of T-DNA was mapped between the 19th and 20th nucleotide

of the 4th exon in AtMGT4 genomic sequence (Fig. 1B). We identified several heterozygous plants from the seed pool, however, no homozygous mutant was obtained. Subsequently, we tried to identify homozygous mutant from progenies of successively selfcrossed +/mgt4-1 plants, but failed, suggesting a defect in pollen and/or embryo development when AtMGT4 was disrupted. We then detected vitality of pollen grains from +/mgt4-1 mutant by staining with DNA fluorescence dye. Hoechst33342, a widely used dye for fluorescent staining of DNA and nuclei, was chosen to stain dispersed mature pollen grains from wild type and +/mgt4-1 plants. The normal mature pollen grains from wild type showed three nuclei, including two generative nuclei and one vegetative nucleus (Fig. 1C). The generative nuclei emitted bright fluorescence because of their intensive chromatin, while vegetative nucleus gave out dull fluorescence due to its disperse chromatin. However, many of the pollen grains from +/mgt4-1 lines could not detect nuclear fluorescence signal or had disperse fluorescence signal under fluorescent microscopy. Furthermore, these pollen grains presented abnormal, shriveled shape under light microscopy, displaying abortive phenotype (Fig. 1C). We made a count on the number of the abnormal and normal pollen grains in +/mgt4-1 and the wild type respectively. In wild-type plants, only 3.4% of the pollen grains were abortive. However, nearly half of the pollen grains were abortive in +/mgt4-1 mutant (Table 1).

J. Li et al. / Plant Science 231 (2015) 212–220

215

Table 1 Mean ratio of abortive pollens in +/mgt4-1. Line

Mean number of abortive pollensa

Mean number of fertile pollensa

Mean ratio of abortive pollens (%)

+/mgt4-1 self-fertilized WT self-fertilized

113 8

114 225

49.8 ± 1.6b 3.4 ± 0.4b

a b

The data represented the mean number of pollens from three independent plants. The data represented the mean ± SD.

The pollen grains were observed further by using scanning electron microscopy (SEM). Dehiscence anther and released mature pollen grains from the wild type appeared normal (Fig. 2A and C), whereas a number of abnormal pollen grains with irregular and shriveled shape were apparent in the anther from +/mgt4-1 mutant (Fig. 2B and D–F). The percentage of abnormal pollen grains was consistent with the abortive ratio determined by the nuclear fluorescent signals. To further identify whether ovules vitality was interfered, we performed reciprocal cross test between +/mgt4-1 and wild type plants. The genotypes of the F1 and F2 progenies were identified by PCR (Table 2). When +/mgt4-1 line was used as the male parent and the wild type (Ler.) as the female parent, all progenies were wild type genotype. Nevertheless, when the wild type was used as the male parent and +/mgt4-1 line as the female parent, the segregation ratio between wild type and +/mgt4-1 genotypes closed to 1:1. These indicated it is the mutated male gametes, but not mutated female gametes that cannot transmit to the progenies.

Furthermore, the cross of mgt4-1 × Ler. produced full seat siliques, and nearly all the embryos in the silique developed normally observed by the whole clearing technique (Fig. S1). These results showed that the vitality of the ovules in +/mgt4-1 was natural. It was in the male gametophytes, but not in the female gametophytes, that the disruption of AtMGT4 led to the failure of producing homozygous mutants from +/mgt4-1 line. Supplementary Fig. S1 related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.plantsci.2014.12.008. To verify that the abortive pollens were caused by the loss-of-function of AtMGT4 gene, we carried out a genetic complementation experiment by introducing AtMGT4pro ::AtMGT4 recombinant construct into +/mgt4-1 mutant. In T2 population, obtained from self-fertilized T1 plants, several independent lines were homozygous for the mgt4 allele (assessed by PCR, Fig. 3A). In all tested T3 progenies of the homozygous lines, nearly all the mature pollen grains grew normal, and each pollen grain had two generative nuclei and one vegetative nucleus when stained

Fig. 2. SEM observation of the +/mgt4-1 anthers and pollen grains. (A) and (B) Whole dehiscence anthers and pollens of the wild type and +/mgt4-1 respectively under scanning electron microscope (SEM). Bars = 50 ␮m. (C) and (D) The close-up of the pollen grains of the wild type anther, as shown in (A), and of the +/mgt4-1 anther, as shown in (B). Bars = 10 ␮m. (E) Pollen grains isolated from the +/mgt4-1 mutants. Bar = 10 ␮m. (F) The close-up of the pollen grains as shown in (E). Bar = 10 ␮m.

216

J. Li et al. / Plant Science 231 (2015) 212–220

Fig. 3. Analysis of pollen fertility in the AtMGT4-RNAi and +/mgt4-1 complementary transgenic plants. (A) Identification of homozygous allele of mgt4 in +/mgt4-1 complementary lines based on PCR method. (B) Transcriptional level of AtMGT4 in the AtMGT4-RNAi and +/mgt4-1 complementary transgenic plants respectively. (C) Hoechst 33342 staining of mature pollen grains of Col. ecotype and the AtMGT4-RNAi transgenic plants respectively. Bar = 20 ␮m. (D) Hoechst 33342 staining of mature pollen grains of Ler. ecotype and +/mgt4-1 complementary transgenic plants. Bar = 20 ␮m. (E) Bar chart of the ratio of abortive pollens in the AtMGT4-RNAi and +/mgt4-1 complementary transgenic lines. Error bars showed the standard deviations from three independent plants.

with Hoechst 33342 (Fig. 3D), and the ratio of abortive pollen in the complementary lines lowered significantly to the level in the wild type (Fig. 3E). The transcriptional level of AtMGT4 in complementary lines recovered to that in wild type, determined by RT-PCR (Fig. 3B). All these data displayed that AtMGT4 transgene fully complemented the male sterile phenotype of +/mgt4-1 plants. Since there was only one T-DNA insertion mutant available, we used the RNA interference approach to make AtMGT4 gene knock-down transgenic plants. The construct of AtMGT4 hairpin was driven by AtMGT4 promoter itself. Several transgenic plants harboring the RNAi construct were obtained. Taking AteIF4A1 gene as a control, we analyzed the transcriptional levels of AtMGT4 in the floral organs of AtMGT4pro ::AtMGT4-RNAi transgenic lines by semi-quantitative RT-PCR method. The result showed that AtMGT4 mRNA level decreased in the AtMGT4pro ::AtMGT4-RNAi lines (numbered #1 and #2 respectively) as low as in the mgt4-1, and AtMGT4 mRNA level in RNAi line #5 was the same as in wild type (Col. ecotype) (Fig. 3B). Then we stained the pollen grains from the AtMGT4-RNAi lines with Hoechst 33342, both of AtMGT4-RNAi lines #1 and #2 had defects in pollen development and nearly 50% mature pollens were abnormal, similar to those of +/mgt4-1 lines. While RNAi line #5, similar to wild type, had a low ratio of abortive pollens (Fig. 3C and E).

These results above demonstrated that loss-of-function of AtMGT4 would cause the defect of pollen development in Arabidopsis. 3.2. AtMGT4 displays distinct expression patterns in Arabidopsis To analyze the expression pattern of AtMGT4 gene, we made a construct with the ␤-glucuronidase (GUS) reporter gene driven by AtMGT4 promoter and transformed it into wild type plants. Several AtMGT4pro ::GUS transgenic lines were selected randomly to detect GUS staining signals. They all showed similar expression patterns. GUS activity was detected mainly in anthers (Fig. 4A-a), ovules (Fig. 4A-b), shoot apical meristems (Fig. 4B-a), young true leaves (Fig. 4B-a), vascular tissues of the mature leaves and hypocotyls (Fig. 4B-a, b, and c), the sites of lateral root elongation (Fig. 4B-c), and root tips (Fig. 4B-d). Because disruption of AtMGT4 gene led to pollen abortion, we focused on the expression pattern of AtMGT4 in the anthers and pollens. We first isolated the pollen grains at various developmental stages from the AtMGT4pro ::GUS transgenic plants for GUS staining. The GUS signals in the pollens were obtained by X-gluc staining, and the pollen developmental stages were determined by Hoechst 33342 fluorescent staining. The GUS signals appeared distinctly from bicellular pollen to mature pollen stages,

Table 2 Gene segregation analysis of +/mgt4-1 by reciprocal cross. Female × male

Genotype of plants WT

+/mgt4-1

Expected ratio

2

df

p

−/−

F1

+/mgt4-1 × WT WT × +/mgt4-1

268 504

254 0

0 0

1:1 –

0.3754 –

1 –

0.50–0.95 –

F2

+/mgt4-1 × WT WT × +/mgt4-1

279 452

268 0

0 0

1:1 –

0.2212 –

1 –

0.50–0.95 –

J. Li et al. / Plant Science 231 (2015) 212–220

217

Fig. 4. Expression patterns of AtMGT4 by GUS activity analysis. AtMGT4pro ::GUS transgenic plants are stained with X-Gluc. (A) GUS activities detected in generative organs (a and b, bars = 1 mm) and pollen grains at different stages of development (c). The pollen grains were also stained by fluorescent dye Hoechst 33342, and the fluorescent signaling was detected under a fluorescence microscope (c, bar = 20 ␮m). (B) GUS activities detected in vegetative organs, including the whole seedling (a, bar = 1 mm), hypocotyl-root transition zone (b, bar = 0.2 mm), the site of lateral root elongation (c, bar = 40 ␮m) and root tips (d, bar = 0.2 mm). (C) GUS activities detected in anthers at different stages of development with transverse paraffin sections. A short-term staining with PI (propidium iodide) gave a red color to the cell walls. The images from (a) to (e) indicated stage 7, 8, 9, 11 and 13 of anthers development respectively. Bars = 20 ␮m.

but weakly at microspore stage (Fig. 4A-c). Then, the X-gluc stained anthers at various developmental stages were sectioned by transverse paraffin section method and the GUS signals were observed under microscopy. Phase of Arabidopsis thaliana anther development had been described previously in detail [27]. No GUS signal was detected at stage 7 (Fig. 4C-a), while a weak GUS activity initiated in microspores of stage 8 anther (Fig. 4C-b). At stages 9 to 11, GUS was detected obviously in tapetum tissue, but weaker in pollen grains (Fig. 4C-c and d). GUS was strongly expressed in mature pollen grains at stage 13 anther (Fig. 4C-e). 3.3. AtMGT4 is localized to endoplasmic reticulum in Arabidopsis protoplasts Subcellular localization can also provide valuable cues for gene function dissection. Earlier studies showed that AtMGT5, an essential member for pollen development, is located to the mitochondria in flower tissues [23]. To determine the subcellular

localization of AtMGT4 protein, we constructed a recombinant plasmid 35S::AtMGT4-EGFP to express AtMGT4-EGFP fusion protein. We transfected Arabidopsis mesophyll protoplasts with this plasmid transiently, then observed the GFP fluorescence by confocal microscopy (Fig. 5). AtMGT4-EGFP was detected in intracellular organelles with varying and threadlike shape. Based on the shape of GFP fluorescence, we speculated this organelle to be endoplasmic reticulum (ER). Then we used Bip-RFP, an ER fluorescence marker [28], to co-localize AtMGT4-EGFP fluorescence signal. We found that EGFP signal overlapped well with the RFP signal, indicating that AtMGT4 was localized to the ER in Arabidopsis cells. 3.4. Pollen abortion initiates at bicellular stage in +/mgt4-1 mutant pollens To determine at which stage of anther development the pollens emerge visibly defect, we first observed the pollen vitality by nuclear status and nuclear number at different pollen

218

J. Li et al. / Plant Science 231 (2015) 212–220

Fig. 5. Subcellular localization of AtMGT4 in protoplasts from Arabidopsis mesophyll cells. The recombinant plasmid 35S::AtMGT4-EGFP and 35S::BiP-RFP (ER marker) were used to cotransfect Arabidopsis mesophyll protoplasts transiently. AtMGT4-EGFP fusion protein was excited by laser at 488 nm, and the emitted fluorescence signaling was collected with a band-pass filter at 490 nm–540 nm (GFP panel). Bip-RFP marker was excited by laser at 543 nm, and the emitted fluorescence was collected with a band-pass filter at 580 nm–650 nm (RFP panel). The GFP and RFP fluorescence signaling could mostly overlapped (Overlay panel). DIC panel showed the image of the protoplast under bright field of the fluorescence microscope. Bar = 20 ␮m.

developmental stages (Fig. 6). Pollens were stained with Hoechst 33342 and observed under fluorescence microscopy. At microspore stage, pollens in +/mgt4-1 mutants showed normal single nucleus fluorescence as in the wild type (Fig. 6A). The ratio of abnormal

microspores was approximately 1.28% in +/mgt4-1 mutants and 1.11% in the wild type (Fig. 6D). However, at the stage of bicellular pollen, nearly half of the pollens in +/mgt4-1 mutants had disperse fluorescence and did not show unambiguous nuclei, while

Fig. 6. Pollen development in the +/mgt4-1 lines with Hoechst33342 staining. (A)–(C) Hoechst 33342 staining of pollens at different stages of development in the Ler. ecotype and +/mgt4-1 plants respectively. Fluorescence excited by UV was observed under a fluorescence microscope (left panels). Pollen morphology was also observed under bright field of the fluorescence microscope (right panels). Arrows showed abnormal pollen grains. Bar = 20 ␮m. (D) Bar chart of the ratio of abortive pollens at different stages of development in the wild type and +/mgt4-1 lines. Black bars represented abortive pollens ratio in wild type, while gray bars represented the ratio in +/mgt4-1 lines. Error bars showed the standard deviations from three independent plants.

J. Li et al. / Plant Science 231 (2015) 212–220

219

Fig. 7. Observation of anther development of the +/mgt4-1 lines. The anthers at different developmental stages from wild type and +/mgt4-1 lines were sliced with transverse paraffin sections. Tissues were stained with hematine. Arrows showed abnormal pollen grains. Bars = 20 ␮m.

98.50% of the pollens in the wild type showed two bright nucleus fluorescences (generative nuclei) and one dull nucleus fluorescence (vegetative nucleus) (Fig. 6B and D), suggesting pollen abortion initiated in the mutant pollens at bicellular stage, which responds to the 11th stage of anther development. At the stage of mature pollen, approximately 50% pollens in +/mgt4-1 mutants, but only 3.23% of pollens in the wild type, displayed irregular and shriveled shape and no nucleus fluorescence signals appeared (Fig. 6C and D). We then observed the anthers and the enclosed pollens at different developmental stages [27] from the wild type and +/mgt4-1 with transverse paraffin sections method (Fig. 7). From stage 5 to stage 9 anthers, the pollens in +/mgt4-1 mutant anthers appeared normal as in the wild type. At stage 11 anther, a ring of vacuole appeared in some of the pollens of +/mgt4-1 mutant, which showed light color when stained with hematine (arrows indicated). Subsequently a number of abnormal pollen grains with irregular and shriveled shapes were obvious at stage 13 of anther development (arrows indicated). These results showed that visible changes in +/mgt4-1 mutant pollens emerged at stage 11 of anther development. 4. Discussion 4.1. Mg2+ homeostasis inside ER tends to be important for the pollen development As is known, ER is not only the sites for protein synthesis, protein folding, glycosylation, and secretion [29], but also functions majorly in lipid biosynthesis [30], and calcium homeostasis maintaining [31]. In Arabidopsis, studies revealed that normal physiological functions of ER were crucial to pollen development. Disruption of certain genes in ER could cause pollen abortion. It was reported that endoplasmic reticulum- and golgi-localized phospholipase A2 (PLA2 ) impacted deformation and trafficking of the endomembrane system and played critical roles in Arabidopsis pollen development [32]. Phospholipase A2 (PLA2 ) hydrolyzed the phospholipid molecule to produce lysophospholipid and a free fatty acid, which was important for the deformation and trafficking of endomembrane system in mammalian cells [33,34]. ER of pollen grains of PLA2 -RNAi plants had an irregular shape and a fragmented membrane. The nucleotide sugar transporters AtUTr1 and AtUTr3, involved in the transport of UDP-glucose into the ER lumen, were essential for the proper folding of glycoproteins in the ER lumen, and disruption of the two genes resulted in pollen abortion in Arabidopsis thaliana [35,36]. In addition, adding tunicamycin to induce

the accumulation of unfolded proteins, leaded to an increase in the uptake of UDP-glucose at the ER. Thus, import of UDP-glucose into the ER lumen appeared to be a critical step during protein quality control in ER, which was also responsible for the pollen abortion of the atutr mutant. Immunoglobulin-binding proteins (BiPs) are molecular chaperones in the ER, functioning in protein translocation, protein folding and quality control in the ER [37,38]. There are three Bip genes (Bip1, Bip2 and Bip3) in Arabidopsis, and disruption of these genes (bip1 bip2 bip3 triple mutant) led to microspore lethality during pollen development [39]. Kim et al. found that endoplasmic reticulum was distinguishable in microspores at the highly vacuolated stage and then stacked ER was extensively distributed in the vegetative cytoplasm during sperm cell formation [32,40]. The pollen abortion in the +/mgt4-1 emerged at bicellular stage, which was largely consistent with the period of formation of ER in the vegetative cytoplasm of the developing pollen, suggesting that the pollen abortion in the +/mgt4-1 was relevant to the abnormal differentiation or function of ER. It was reported that AtMGT4 could re-establish Mg2+ uptake in yeast mrs2 mutant heterologously [17]. As an ER-localized Mg2+ transporter, AtMGT4 probably plays a role to maintain magnesium homeostasis in ER, and loss-of-function of AtMGT4 in pollens might disturb the magnesium homeostasis in ER. The Mg2+ imbalance might change the inner environment of ER and result in the occurrence of ER stress, subsequently trigger apoptosis procedure [41], and lead to the pollen abortion eventually. However, the underline mechanism on how the Mg2+ in ER impacts pollen development needs to be clarified further. 4.2. Potential interaction between AtMGT4 and other AtMGT genes In the yeast mating-based split-ubiquitin system (mbSUS), AtMGT4 showed proneness to non-selective, strong interactions with the other AtMGT proteins [42]. In consideration of ER localization of AtMGT4 and the wide distribution of ER through the whole cell, otherwise, various organelles in the cell were enclosed by ER [43], it is reasonable that AtMGT4 possessed interactions with other AtMGT proteins in Arabidopsis. Our previous work showed that AtMGT9, another member of Mg2+ transporters, was essential for pollen development in Arabidopsis [24]. The expression pattern of AtMGT9 was similar to AtMGT4, both of them expressed not only in anthers but also vascular tissues and meristems. Interestingly, AtMGT9 was also located to ER as AtMGT4 (data not shown). The identical localization and

220

J. Li et al. / Plant Science 231 (2015) 212–220

similar expression patterns offered the possibility of interaction between AtMGT4 and AtMGT9. It is highly plausible that AtMGT4 and AtMGT9 form functional hetero-oligomer Mg2+ transporter to mediate intracellular Mg2+ homeostasis in pollens, or AtMGT4 and AtMGT9 function as influx and efflux Mg2+ transporter respectively. Hence knockout one of which would disturb the Mg2+ homeostasis and lead to pollen abortion. As no homozygous mutant could be obtained, it is hard to dissect the function of AtMGT4 in organs other than the pollen. It might be a practicable approach to develop a chimeric mutant by expressing AtMGT4 with a pollen-specific promoter in +/mgt4-1 plant to screen AtMGT4 pollen-expressed but other organ-knockout mutant plants. Then the function in vascular tissues and meristems where AtMGT4 is expressed could be elucidated. Acknowledgements We are grateful to Inhwan Hwang (Pohang University of Science and Technology, Korea) for providing Bip-RFP marker. This work was supported by National Science Foundation of China grants (NSFC-31170229; 31371244), Key Project of Hunan Provincial Education Department (12A096), Hunan Provincial Natural Science Foundation (12JJ6021), Hunan Provincial Construct Program of the Key Discipline in Ecology (0713), Hunan Provincial Cooperative Innovation Center of Engineering and New Products for Developmental Biology (20134486). References [1] C. Hermans, N. Verbruggen, Physiological characterization of Mg deficiency in Arabidopsis thaliana, J. Exp. Bot. 56 (2005) 2153–2161. [2] J.E. Lusk, E.P. Kennedy, Magnesium transport in Escherichia coli, J. Biol. Chem. 244 (1969) 1653–1655. [3] D.L. Nelson, E.P. Kennedy, Magnesium transport in Escherichia coli inhibition by cobaltous ion, J. Biol. Chem. 246 (1971) 3042–3049. [4] M.H. Park, B.B. Wong, J.E. Lusk, Mutants in three genes affecting transport of magnesium in Escherichia coli: genetics and physiology, J. Bacteriol. 126 (1976) 1096–1103. [5] R. Smith, J. Banks, M. Snavely, M. Maguire, Sequence and topology of the CorA magnesium transport systems of Salmonella typhimurium and Escherichia coli identification of a new class of transport protein, J. Biol. Chem. 268 (1993) 14071–14080. [6] S. Eshaghi, D. Niegowski, A. Kohl, D.M. Molina, S.A. Lesley, P. Nordlund, Crystal structure of a divalent metal ion transporter CorA at 2.9 angstrom resolution, Science 313 (2006) 354–357. [7] V.V. Lunin, E. Dobrovetsky, G. Khutoreskaya, R. Zhang, A. Joachimiak, D.A. Doyle, A. Bochkarev, M.E. Maguire, A.M. Edwards, C.M. Koth, Crystal structure of the CorA Mg2+ transporter, Nature 440 (2006) 833–837. [8] D.G. Kehres, C.H. Lawyer, M.E. Maguire, The CorA magnesium transporter gene family, Microb. Comp. Genomics 3 (1998) 151–169. [9] D.M. Bui, J. Gregan, E. Jarosch, A. Ragnini, R.J. Schweyen, The bacterial magnesium transporter CorA can functionally substitute for its putative homologue Mrs2p in the yeast inner mitochondrial membrane, J. Biol. Chem. 274 (1999) 20438–20443. [10] R.L. Smith, E. Gottlieb, L.M. Kucharski, M.E. Maguire, Functional similarity between archaeal and bacterial CorA magnesium transporters, J. Bacteriol. 180 (1998) 2788–2791. [11] M.B.C. Moncrief, M.E. Maguire, Magnesium transport in prokaryotes, J. Biol. Inorg. Chem. 4 (1999) 523–527. [12] C.W. MacDiarmid, R.C. Gardner, Overexpression of the Saccharomyces cerevisiae magnesium transport system confers resistance to aluminum ion, J. Biol. Chem. 273 (1998) 1727–1732. [13] J. Gregan, D.M. Bui, R. Pillich, M. Fink, G. Zsurka, R.J. Schweyen, The mitochondrial inner membrane protein Lpe10p, a homologue of Mrs2p, is essential for magnesium homeostasis and group II intron splicing in yeast, Mol. Gen. Genet. 264 (2001) 773–781. [14] N.P. Pisat, A. Pandey, C.W. Macdiarmid, MNR2 regulates intracellular magnesium storage in Saccharomyces cerevisiae, Genetics 183 (2009) 873–884. [15] I. Schock, J. Gregan, S. Steinhauser, R. Schweyen, A. Brennicke, V. Knoop, A member of a novel Arabidopsis thaliana gene family of candidate Mg2+ ion transporters complements a yeast mitochondrial group II intron-splicing mutant, Plant J. 24 (2000) 489–501. [16] L. Li, A.F. Tutone, R.S. Drummond, R.C. Gardner, S. Luan, A novel family of magnesium transport genes in Arabidopsis, Plant Cell 13 (2001) 2761–2775.

[17] M. Gebert, K. Meschenmoser, S. Svidova, J. Weghuber, R. Schweyen, K. Eifler, H. Lenz, K. Weyand, V. Knoop, A root-expressed magnesium transporter of the MRS2/MGT gene family in Arabidopsis thaliana allows for growth in low-Mg2+ environments, Plant Cell 21 (2009) 4018–4030. [18] D. Mao, J. Chen, L. Tian, Z. Liu, L. Yang, R. Tang, J. Li, C. Lu, Y. Yang, J. Shi, L. Chen, D. Li, S. Luan, Arabidopsis transporter MGT6 mediates magnesium uptake and is required for growth under magnesium limitation, Plant Cell 26 (2014) 2234–2248. [19] S.J. Conn, V. Conn, S.D. Tyerman, B.N. Kaiser, R.A. Leigh, M. Gilliham, Magnesium transporters, MGT2/MRS2-1 and MGT3/MRS2-5, are important for magnesium partitioning within Arabidopsis thaliana mesophyll vacuoles, New Phytol. 190 (2011) 583–594. [20] B.M. Waters, Moving magnesium in plant cells, New Phytol. 190 (2011) 510–513. [21] C. Hermans, S.J. Conn, J. Chen, Q. Xiao, N. Verbruggen, An update on magnesium homeostasis mechanisms in plants, Metallomics 5 (2013) 1170–1183. [22] H. Lenz, V. Dombinov, J. Dreistein, M.R. Reinhard, M. Gebert, V. Knoop, Magnesium deficiency phenotypes upon multiple knockout of Arabidopsis thaliana MRS2 clade B genes can be ameliorated by concomitantly reduced calcium supply, Plant Cell Physiol. 54 (2013) 1118–1131. [23] L.G. Li, L.N. Sokolov, Y.H. Yang, D.P. Li, J. Ting, G.K. Pandy, S. Luan, A mitochondrial magnesium transporter functions in Arabidopsis pollen development, Mol. Plant 1 (2008) 675–685. [24] J. Chen, L.G. Li, Z.H. Liu, Y.J. Yuan, L.L. Guo, D.D. Mao, L.F. Tian, L.B. Chen, S. Luan, D.P. Li, Magnesium transporter AtMGT9 is essential for pollen development in Arabidopsis, Cell Res. 19 (2009) 887–898. [25] T. Murashige, F. Skoog, A revised medium for rapid growth and bio assays with tobacco tissue cultures, Physiol. Plant 15 (1962) 473–497. [26] R. Gupta, J.T. Ting, L.N. Sokolov, S.A. Johnson, S. Luan, A tumor suppressor homolog, AtPTEN1, is essential for pollen development in Arabidopsis, Plant Cell 14 (2002) 2495–2507. [27] P.M. Sanders, A.Q. Bui, K. Weterings, K. McIntire, Y.-C. Hsu, P.Y. Lee, M.T. Truong, T. Beals, R. Goldberg, Anther developmental defects in Arabidopsis thaliana male-sterile mutants, Sex. Plant Reprod. 11 (1999) 297–322. [28] J. Lee, H. Lee, J. Kim, S. Lee, D.H. Kim, S. Kim, I. Hwang, Both the hydrophobicity and a positively charged region flanking the C-terminal region of the transmembrane domain of signal-anchored proteins play critical roles in determining their targeting specificity to the endoplasmic reticulum or endosymbiotic organelles in Arabidopsis cells, Plant Cell 23 (2011) 1588–1607. [29] Y. Shibata, T. Shemesh, W.A. Prinz, A.F. Palazzo, M.M. Kozlov, T.A. Rapoport, Mechanisms determining the morphology of the peripheral ER, Cell 143 (2010) 774–788. [30] J. Ohlrogge, J. Browse, Lipid biosynthesis, Plant Cell 7 (1995) 957–970. [31] D. Sanders, J. Pelloux, C. Brownlee, J.F. Harper, Calcium at the crossroads of signaling, Plant Cell 14 (2002) S401–S417. [32] H.J. Kim, S.H. Ok, S.C. Bahn, J. Jang, S.A. Oh, S.K. Park, D. Twell, S.B. Ryu, J.S. Shin, Endoplasmic reticulum- and golgi-localized phospholipase A2 plays critical roles in Arabidopsis pollen development and germination, Plant Cell 23 (2011) 94–110. [33] W.J. Brown, K. Chambers, A. Doody, Phospholipase A2 (PLA2 ) enzymes in membrane trafficking: mediators of membrane shape and function, Traffic 4 (2003) 214–221. [34] H.T. McMahon, J.L. Gallop, Membrane curvature and mechanisms of dynamic cell membrane remodelling, Nature 438 (2005) 590–596. [35] F. Reyes, L. Marchant, L. Norambuena, R. Nilo, H. Silva, A. Orellana, AtUTr1, a UDP-glucose/UDP-galactose transporter from Arabidopsis thaliana, is located in the endoplasmic reticulum and up-regulated by the unfolded protein response, J. Biol. Chem. 281 (2006) 9145–9151. [36] F. Reyes, G. León, M. Donoso, F. Brandizzí, A.P.M. Weber, A. Orellana, The nucleotide sugar transporters AtUTr1 and AtUTr3 are required for the incorporation of UDP-glucose into the endoplasmic reticulum, are essential for pollen development and are needed for embryo sac progress in Arabidopsis thaliana, Plant J. 61 (2010) 423–435. [37] S. Nishikawa, J.L. Brodsky, K. Nakatsukasa, Roles of molecular chaperones in endoplasmic reticulum (ER) quality control and ER-associated degradation (ERAD), J. Biochem. 137 (2005) 551–555. [38] A.L. Horwich, Molecular chaperones in cellular protein folding: the birth of a field, Cell 157 (2014) 285–288. [39] D. Maruyama, T. Sugiyama, T. Endo, S. Nishikawa, Multiple BiP genes of Arabidopsis thaliana are required for male gametogenesis and pollen competitiveness, Plant Cell Physiol. (2014). [40] A. Kuang, M.E. Musgrave, Dynamics of vegetative cytoplasm during generative cell formation and pollen maturation in Arabidopsis thaliana, Protoplasma 194 (1996) 81–90. [41] Y. Deng, R. Srivastava, S.H. Howell, Endoplasmic reticulum (ER) stress response and its physiological roles in plants, Int. J. Mol. Sci. 14 (2013) 8188–8212. [42] J. Schmitz, A. Tierbach, H. Lenz, K. Meschenmoser, V. Knoop, Membrane protein interactions between different Arabidopsis thaliana MRS2-type magnesium transporters are highly permissive, BBA-Biomembranes 1828 (2013) 2032–2040. [43] C. Doyle, The endoplasmic reticulum: a social network in plant cells, J. Integr. Plant Biol. 54 (2012) 840–850.