The trans-Golgi-localized protein BICAT3 regulates manganese allocation and matrix polysaccharide biosynthesis

Abstract

Manganese (Mn²⁺) is essential for a diversity of processes, including photosynthetic water splitting and the transfer of glycosyl moieties. Various Golgi-localized glycosyltransferases that mediate cell wall matrix polysaccharide biosynthesis are Mn²⁺ dependent, but the supply of these enzymes with Mn²⁺ is not well understood. Here, we show that the BIVALENT CATION TRANSPORTER 3 (BICAT3) localizes specifically to trans-cisternae of the Golgi. In agreement with a role in Mn²⁺ and Ca²⁺ homeostasis, BICAT3 rescued yeast (Saccharomyces cerevisiae) mutants defective in their translocation. Arabidopsis (Arabidopsis thaliana) knockout mutants of BICAT3 were sensitive to low Mn²⁺ and high Ca²⁺ availability and showed altered accumulation of these cations. Despite reduced cell expansion and leaf size in Mn²⁺-deficient bicat3 mutants, their photosynthesis was improved, accompanied by an increased Mn content of chloroplasts. Growth defects of bicat3 corresponded with an impaired glycosidic composition of matrix polysaccharides synthesized in the trans-Golgi. In addition to the vegetative growth defects, pollen tube growth of bicat3 was heterogeneously aberrant. This was associated with a severely reduced and similarly heterogeneous pectin deposition and caused diminished seed set and silique length. Double mutant analyses demonstrated that the physiological relevance of BICAT3 is distinct from that of ER-TYPE CA²⁺-ATPASE 3, a Golgi-localized Mn²⁺/Ca²⁺-ATPase. Collectively, BICAT3 is a principal Mn²⁺ transporter in the trans-Golgi whose activity is critical for specific glycosylation reactions in this organelle and for the allocation of Mn²⁺ between Golgi apparatus and chloroplasts.

Activity of BIVALENT CATION TRANSPORTER 3 in the trans-Golgi is critical for glycosylation reactions, plant growth under Mn deficiency, pollen tube elongation, and intracellular distribution of Mn.

Introduction

Manganese (Mn²⁺) is an essential trace metal for plants, which functions as a cofactor of enzymes or a catalyst in metal clusters (Andresen et al., 2018; Alejandro et al., 2020). One of the most prominent roles of Mn²⁺ is to constitute the Mn₄O₅Ca catalytic cluster in the oxygen-evolving complex (OEC) of photosystem II (PSII) for water oxidation, and consequently the supply of electrons and protons for the photosynthetic electron chain and ATP production. Furthermore, THYLAKOID-ASSOCIATED PHOSPHATASE 38, which is critical for state transition between PSI and PSII, requires a binuclear Mn or Mg center (Wei et al., 2015). Besides its roles in photosynthetic activities, Mn²⁺ is also involved in ROS scavenging as a component of mitochondria- and peroxisome-localized manganese superoxide dismutase (He et al., 2021). Moreover, Mn²⁺ is essential for the catalytic activity of many Golgi-localized glycosyltransferases (GTs), which are critical for the glycosylation of proteins and lipids, and for the synthesis of complex cell wall matrix polysaccharides (He et al., 2021). Matrix polymers (hemicelluloses and pectins) account for two-thirds of the Arabidopsis (Arabidopsis thaliana) primary cell wall. They contribute substantially to the mechanical properties of the wall and are thus critical for plant growth and development (Cosgrove, 2016, 2018; Rui and Dinneny, 2020; Molina et al., 2021). Matrix polysaccharides influence the organization of cellulose microfibrils, the main load-bearing components of the cell wall, to modulate wall extensibility and organ formation (Park and Cosgrove, 2012; Rui and Anderson, 2016; Zhao et al., 2019). Golgi-localized xyloglucan xylosyltransferases (XXT1 and 2), which are involved in xyloglucan biosynthesis, are Mn²⁺-dependent (Culbertson et al., 2016, 2018). The xxt1xxt2 double mutant is smaller than the wild-type in shoot size, with a diminished amount and abnormal composition of xyloglucans. Biosynthesis of cell wall sugars may be mutually affected, as xxt1xxt2 not only displays a defect in xyloglucan production, but also in cellulose production and pectin methylation at the shoot apical meristem (Zhao et al., 2019). Another class of matrix sugars, pectins, plays pivotal roles in a plethora of biological processes, including cell adhesion and morphogenesis (Bouton et al., 2002; Altartouri et al., 2019; Haas et al., 2020), organ formation (Peaucelle et al., 2011), plant growth (Kim et al., 2015), pollen tube growth (Kim et al., 2015; Lund et al., 2020), fertilization (Duan et al., 2020), stomatal functions (Amsbury et al., 2016; Rui et al., 2019), and also confers resistance to biotic and tolerance to abiotic stress (Bacete et al., 2018; Wu et al., 2018; De Lorenzo et al., 2019).

In addition to matrix sugars, the N- and O-glycosylation of proteins and lipids are essential for plant development and stress responses (Kobayashi, 2016; Nagashima et al., 2018; Huby et al., 2020; Mortimer and Scheller, 2020; Seifert, 2020; Silva et al., 2020). Arabinogalactan proteins (AGPs) have been proposed to represent Ca²⁺ capacitors that reversibly bind and release apoplastic Ca²⁺ in a pH-dependent manner by their β-linked glucuronic acid (GlcA) residues and which thereby determine numerous biological processes, such as pollen tube growth and guidance, seedling growth, and Ca²⁺ oscillations in roots (Lamport et al., 2018; Lopez-Hernandez et al., 2020). Accordingly, mutants of Mn²⁺-dependent hydroxyproline galactosyltransferases (GALT2-6) that are involved in AGP biosynthesis, show defects in root, leaf, and pollen tube growth (Basu et al., 2015a, 2015b). Mn²⁺-dependent transfer of GlcA residues to glycosyl inositol phosphoryl ceramide sphingolipids by inositol phosphoryl ceramide glucuronosyltransferase 1 (IPUT1/PGSIP6/MOCA1) is also crucial for generative and vegetative development, and the glycosylated sphingolipids are believed to act as Na⁺ sensors (He et al., 2021). Altogether, an adequate Mn²⁺ supply for Golgi-localized glycosyl transfer processes appears to be crucial for most aspects of plant performance. However, as this assumption is derived from in vitro analyses of individual enzyme activities, the mechanism and bottlenecks for Mn²⁺ supply in planta remain poorly understood.

In Arabidopsis, Mn²⁺ is acquired by the high-affinity transporter NRAMP1 and may also enter the root in high amounts via the Fe uptake transporter IRT1. The pathway of xylem loading is dubious in dicots (Alejandro et al., 2020). Intracellularly, Mn²⁺ may be sequestrated in vacuoles by CATION DIFFUSION FACILITATOR/METAL TOLERANCE PROTEIN (CDF/MTP) and CATION/H⁺ EXCHANGER (CAX) transporters, which is an important mechanism of its detoxification and its storage in seeds (Eroglu et al., 2016, 2017; He et al., 2021; Höller et al., 2022). Transfer of Mn²⁺ into the endoplasmic reticulum by the P_2A-type ATPase ER-TYPE CA²⁺-ATPASE 1 (ECA1) also confers tolerance to the metal (Liang et al., 1997). To fulfill its function in the OEC, Mn²⁺ needs to cross the chloroplast envelope and the thylakoid membrane. Both steps are mediated by members of the BIVALENT CATION TRANSPORTER (BICAT) family, with BICAT1/PAM71/CCHA1 supplying the thylakoid lumen and BICAT2/CMT1 operating in the inner envelope (Schneider et al., 2016; Eisenhut et al., 2018; Zhang et al., 2018). Mn²⁺ supply of the chloroplast is also dependent on the NRAMP2 transporter in the trans-Golgi network (Alejandro et al., 2017). This transporter is functionally epistatic to the vacuolar Mn²⁺ exporters NRAMP3 and NRAMP4, pointing to a crucial function of vesicular compartments for intra-organellar Mn²⁺ distribution (He et al., 2021). This notion is also supported by the severely aberrant Mn²⁺ handling in mutants of MTP11, that are very hypersensitive to Mn²⁺ excess and hypertolerant to Mn²⁺ limitation. MTP11 has been localized to the prevacuolar compartment and the Golgi, and, based on lower Mn²⁺ levels in the mutants, suggested to promote Mn²⁺ detoxification by vesicular secretion (Delhaize et al., 2007; Peiter et al., 2007). Another Mn²⁺ translocator, the P_2A-type ATPase ECA3, has similarly been localized to the Golgi, and eca3 mutants are equally sensitive to Mn²⁺ toxicity (Mills et al., 2008).

In yeast (Saccharomyces cerevisiae) and humans (Homo sapiens), Mn²⁺ loading of Golgi vesicles is pursued by SPCA and SERCA-type ATPases and transporters related to the above-mentioned BICAT proteins operating side-by-side (He et al., 2021). Those transporters, GDT1 in yeast and TMEM165 in humans, permeate Ca²⁺ besides Mn²⁺ (Demaegd et al., 2013; Thines et al., 2018; Stribny et al., 2020), as was also shown for BICAT1 and BICAT2 (Frank et al., 2019). The absence of TMEM165 inflicts defects in protein glycosylation by limiting the activity of Mn²⁺-dependent β-1,4-galactosyltransferase 1 (B4GALT1), which co-localizes with TMEM165 in the trans-Golgi (Foulquier et al., 2012; Foulquier and Legrand, 2020).

In plants, not only protein glycosylation demands Mn²⁺ as a co-factor, but also the biosynthesis of matrix polysaccharides, mediated by a plethora of glycosyl transferases (He et al., 2021). It has been known for long that those reactions are spatially organized to form an efficient assembly line in the secretory pathway. Specific Mn²⁺ transporters may thus be required for discrete glycosylation steps, as it is the case in animals. In plants, the specific role of Golgi-localized Mn²⁺ transporters and the functional interaction of their activity with Mn²⁺-requiring processes in other organelles are poorly understood despite some recent progress (Yang et al., 2021; Zhang et al., 2021). Here, we demonstrate that a protein of the BICAT family that has been suggested to function as Golgi-localized Mn²⁺ transporter is required for the formation of specific glycosyl linkages, rendering it essential under numerous vegetative and generative circumstances. Importantly, BICAT3 affects Mn²⁺ accumulation as well as its subcellular distribution between Golgi and chloroplast. Double mutant analyses show that its physiological relevance is distinct from that of Ca²⁺/Mn²⁺-ATPase ECA3.

Results

BICAT3 primarily localizes to the trans-Golgi and is ubiquitously expressed

The genome of A. thaliana harbors five members of the functionally conserved UPF0016 family, BICAT 1–5. Previous phylogenetic analyses revealed that BICAT3 is the closest Arabidopsis homolog to human TMEM165 and yeast GDT1 (Demaegd et al., 2014). Both proteins are Ca²⁺/Mn²⁺ transporters localized in the Golgi and critical for Mn²⁺ supply for protein glycosylation (Foulquier et al., 2012; Colinet et al., 2016; Potelle et al., 2016; Dulary et al., 2018; Thines et al., 2019). To investigate the subcellular localization of BICAT3, we co-expressed BICAT3 with organelle markers fused with different fluorescent proteins in Arabidopsis mesophyll protoplasts. Complementation of the bicat3-1 mutant with BICAT3-Venus driven by the native BICAT3 promoter confirmed the functionality of the fusion construct (Supplemental Figure S1). In confocal laser scanning microscopy (CLSM) analyses, BICAT3 co-localized with the trans-Golgi marker sialyl transferase (ST; Wee et al., 1998) (Figure 1A), but not with markers targeted to mitochondria or peroxisomes (Supplemental Figure S2). Since conventional CLSM cannot reliably discriminate Golgi subcompartments, we recorded images at high speed and resolution with a STELLARIS 8 microscope (Leica, Wetzlar, Germany). This enabled a clear discrimination of fluorescent proteins targeted to the cis/medial-Golgi [α-mannosidaseI (ManI); Donohoe et al. (2013)] and the trans-Golgi (ST) (Figure 1B). While BICAT3 overlapped perfectly with the trans-Golgi marker, a shift was noted to that for the cis-Golgi, indicating that BICAT3 is not a cis-Golgi-specific protein, as previously claimed (Yang et al., 2021). Supporting the live-cell analyses, immunogold labeling with Venus antibodies in the complementation line primarily detected BICAT3 in the medial- and trans-Golgi (Figure 1C).

Figure 1.

BICAT3 localizes to the trans-Golgi and is ubiquitously expressed in Arabidopsis. A, Subcellular localization of BICAT3 in Arabidopsis mesophyll protoplasts. BICAT3 was co-expressed with a trans-Golgi marker (ST). Bar represents 5 µm. B, High-resolution co-localization of BICAT3, a cis-Golgi marker (ManI), and a trans-Golgi marker (ST) in N. benthamiana mesophyll cells. Magenta letters indicate proteins tagged with mCherry, green letters indicate proteins tagged with GFP. Bars represent 500 nm. C, Ultrastructural localization of BICAT3-Venus by immunogold staining. a, Wild-type Golgi apparatus as a negative control. b and c, Localization of BICAT3-Venus in the Golgi apparatus of bicat3-1 leaf cells expressing BICAT3-Venus under control of the native BICAT3 promoter. Scale bars represent 200 nm. D, Expression of BICAT3 at different growth stages as determined by GUS staining of a ProBICAT3-GUS line. a, Germinated seed. b, 10-day-old seedling. c, Inflorescence and siliques. d, Mature rosette leaf. e, Anther. Scale bars represent 300 µm (a), 200 µm (e) or 2 mm (b, c, and d).

To reveal the expression pattern of BICAT3 in different tissues, its promoter activity was visualized by using the β-glucuronidase (GUS) reporter. Strong GUS expression was detected in most of the tissues (embryo, root, leaves, stem, and flowers) throughout different growth stages (Figure 1D). The expression of BICAT3 was not affected by imbalanced Ca²⁺ or Mn²⁺ supply (Supplemental Figure S3).

BICAT3 acts as Ca²⁺ and Mn²⁺ transporter in yeast

Putative BICAT3 orthologs in yeast and humans mediate evolutionarily conserved Ca²⁺ and Mn²⁺ transport to feed compartments of the Golgi apparatus for essential physiological processes, such as glycosylation. To test if Golgi-localized BICAT3 has similar functions in cation allocation, BICAT3 was heterologously expressed in a pmr1Δgdt1Δ mutant, which displays a defect in Ca²⁺ transport into the Golgi due to the lack of the Ca²⁺ pump PMR1 and of GDT1, as well as in a pmr1Δ single mutant. The pmr1Δgdt1Δ mutant showed strong growth retardation, while pmr1Δ was only slightly affected by Ca²⁺ toxicity (Figure 2A) as previously reported (Demaegd et al., 2013). The Ca²⁺ sensitivity of pmr1Δgdt1Δ was complemented by the expression of BICAT3. The growth of pmr1Δgdt1ΔBICAT3 was comparable to the pmr1Δ single mutant under 600-mM CaCl₂, implying a similar ability of BICAT3 and GDT1 to transport Ca²⁺ in yeast. In addition to transport Ca²⁺, PMR1 translocates excess Mn²⁺ into the Golgi for detoxification. Heterologous expression of BICAT3 partially complemented the Mn²⁺ sensitivity of pmr1Δ (Figure 2B). In contrast, BICAT3 did not complement the Zn²⁺-, Co²⁺-, Cu²⁺-, and Fe³⁺-dependent growth defects in zrc1Δ, cot1Δ, cup2Δ, and ccc1Δ yeast mutants, respectively (Supplemental Figure S4). In conclusion, the results indicate that BICAT3 acts as a Ca²⁺ and Mn²⁺ transporter in yeast.

Figure 2.

BICAT3 complements Ca²⁺- and Mn²⁺-sensitive yeast strains. A, Growth of wild-type, pmr1Δ, and pmr1Δgdt1Δ with and without BICAT3 under Ca²⁺-toxic conditions. B, Growth of wild-type and pmr1Δ with and without BICAT3 under Mn²⁺-toxic conditions. Liquid cultures of the strains were serially diluted and dropped onto media as indicated.

Growth of bicat3 mutants is differentially affected by high and low supply of Mn²⁺ and Ca²⁺

To characterize the functions of BICAT3 in Arabidopsis, two T-DNA insertional mutants were obtained. bicat3-1 harbors a T-DNA in the sixth intron, and bicat3-2 carries a T-DNA insertion in the three prime untranslated region (3′-UTR) (Supplemental Figure S5). While bicat3-1 is a knockout mutant, bicat3-2 shows 6% of BICAT3 transcript level compared to the wild-type. Both mutants show no obvious vegetative growth differences compared to the wild-type at mature stage under nonstressed conditions.

To assess the relevance of BICAT3 in Ca²⁺ and Mn²⁺ dependence of Arabidopsis, we cultured bicat3-1 and wild-type seedlings on plates with different Mn²⁺ and Ca²⁺ supply. Intriguingly, young seedlings of bicat3-1 grew more vigorously than the wild-type under control conditions (Supplemental Figures S6–S8), but were overly sensitive to Mn²⁺ deficiency (0 mM; Supplemental Figure S6) and Ca²⁺ toxicity (50 mM; Supplemental Figure S7). Conversely, bicat3-1 seedlings were more tolerant of Mn²⁺ toxicity (1 mM; Supplemental Figure S8).

To examine the role of BICAT3 in older plants and a more controlled medium composition, bicat3-1 and wild-type were cultured in hydroponics with different Mn²⁺ and Ca²⁺ supply (Supplemental Figure S9). With surplus Mn²⁺ (350 µM) or deficient Ca²⁺ (0.05 mM) supply, shoot and root growth of bicat3-1 were enhanced compared to the wild-type. In contrast, shoot growth of bicat3-1 was more sensitive than that of the wild-type to elevated Ca²⁺ (25 mM). The visually most striking difference was found under Mn²⁺ deficiency (0.05 µM). Under this condition, the size of the bicat3-1 shoot was much smaller than that of the wild-type, and leaves of the mutant were curled (Supplemental Figure S9A). Despite this decrease in shoot size, shoot dry mass of bicat3-1 was comparable to that of the Mn²⁺-deficient wild-type (Supplemental Figure S9B).

To determine the role of BICAT3 in Mn²⁺ and Ca²⁺ homeostasis, we measured their concentrations in bicat3-1 and wild-type. Under adequate supply, bicat3-1 did not differ in Mn²⁺ concentration compared to the wild-type, but had a higher Ca²⁺ concentration in shoots (Supplemental Figure S9, C and D). However, under Mn²⁺ deficiency and Ca²⁺ toxicity, the Mn²⁺ concentration in bicat3 shoots was elevated in comparison to the wild-type, while bicat3-1 shoots accumulated less Mn²⁺ than the wild-type under Mn²⁺ toxicity (Supplemental Figure S9C). Altogether, BICAT3 plays a central role in the plant’s resilience to Mn²⁺ deficiency and contributes to Ca²⁺ and Mn²⁺ homeostasis in Arabidopsis. The bicat3-2 mutant showed the same Mn²⁺ deficiency-hypersensitive growth phenotype and comparable aberrances in Mn²⁺ and Ca²⁺ concentrations in shoots and roots as bicat3-1 (Figure 3). Mg²⁺ concentrations were also slightly but significantly lower in Mn²⁺-deficient mutant shoots, whereas Fe³⁺ and Zn²⁺ concentrations were unchanged (Supplemental Figure S10). Interestingly, relative water content was decreased in the mutants, particularly under Mn²⁺ deficiency (Supplemental Figure S11).

Figure 3.

Phenotypes of Col-0 wild-type, bicat3-1, and bicat3-2 under Mn²⁺ deficiency conditions. A, Growth phenotypes of 5-week-old plants cultivated in hydroponics with 3.5 µM Mn²⁺ (control) or 0.05 µM Mn²⁺ (-Mn). Images were digitally extracted for comparison. Scale bar represents 1 cm and applies to all images. B, Shoot and root DW of plants grown as in (A). C and D, Mn²⁺ and Ca²⁺ concentrations in shoots and roots of plants grown as in (A). Data indicate means + se of five biological replicates. Data were analyzed by two-tailed Student’s t test to identify significant differences between wild-type and mutant (^*P < 0.05; ^**P < 0.01; ^***P < 0.001). The experiment was repeated twice with similar results.

bicat3 mutants accumulate more manganese in chloroplasts and have improved photosynthesis under Mn²⁺ deficiency

Since a decrease in photosynthetic activity is the most prominent consequence of Mn²⁺ deficiency (Alejandro et al., 2020), we analyzed the chlorophyll fluorescence of the nutrient-starved plants. Intriguingly, despite the retarded growth of bicat3 mutants (Figure 3), both lines showed a substantially higher maximum quantum yield of PSII (Fv/Fm) and effective quantum yield of PSII [Y (II)] (Figure 4, A and B). Whole-plant analysis had shown that bicat3 mutants accumulate more Mn²⁺ in shoot compared to the wild-type under Mn²⁺ deficiency (Figure 3C;Supplemental Figure S9C). Since we hypothesized that photosynthetic improvement was caused by an enhanced supply of PSII with Mn²⁺, we determined the ion's concentration in isolated chloroplasts by inductively coupled plasma mass spectrometry (ICP–MS). Interestingly, bicat3-1 and bicat3-2 chloroplasts contained a higher amount of Mn²⁺ than the wild-type under Mn²⁺ deficiency (Figure 4C), demonstrating that BICAT3 determines cellular Mn²⁺ distribution, including Mn²⁺ supply to chloroplasts, which may subsequently have an effect on photosynthetic activity.

Figure 4.

Absence of BICAT3 causes improved photosynthesis and higher Mn²⁺ contents of chloroplasts compared to the wild-type under Mn²⁺ deficiency conditions. A, Fv/Fm false color images of 5-week-old Col-0, bicat3-1, and bicat3-2 grown in hydroponics with 3.5-µM Mn²⁺ (control) or 0.05-µM Mn (-Mn). Images were digitally extracted for comparison. Scale bar represents 1 cm and applies to all images. B, Efficiency of PS II [Y (II)] of plants grown as in A. Data indicate means ± se of five biological replicates. C, Mn²⁺ content of chloroplasts isolated from Col-0, bicat3-1, and bicat3-2 plants grown as in (A). Data indicate means + se of four replicates. Data in (B) and (C) were analyzed by two-tailed Student’s t test to identify significant differences between wild-type and mutant (^*P < 0.05; ^**P < 0.01). The experiment was repeated twice with similar results.

To analyze the dependency of growth defect and improved photosynthetic activity on Mn²⁺ supply level and its reversibility, plants were grown under mildly and severely Mn²⁺-deficient conditions. Both conditions led to an improved Fv/Fm in bicat3-1 compared to the wild-type (Supplemental Figure S12B). This difference increased with increasing degree of Mn²⁺ deficiency, and it could be reversed by a resupply with Mn²⁺ for 1 week. Pronounced differences in Y(II) between bicat3-1 and wild-type were only obtained under severe Mn²⁺ deficiency, and were also reversible by Mn²⁺ resupply (Supplemental Figure S12C). In contrast to photosynthetic parameters, the abnormal leaf morphology of bicat3-1 under severe Mn²⁺ deficiency did not change after Mn²⁺ resupply (Supplemental Figure S12A).

To determine if the prominent shoot phenotypes of bicat3 mutants are controlled systemically by the roots or locally in the shoot, reciprocal grafting of bicat3-1 and wild-type seedlings was performed. Plants consisting of bicat3-1 shoot and wild-type root (bC) showed identical shoot phenotypes to bicat3-1 self-grafted plants (bb) (Supplemental Figure S13). On the other hand, plants consisting of wild-type shoot and bicat3-1 root (Cb) showed a shoot phenotype similar to wild-type self-grafted plants (CC). Regardless of the amount of Mn²⁺ supplied, grafted plants comprising bicat3-1 roots (Cb and bb) accumulated higher shoot Ca²⁺ concentrations. Taken together, BICAT3 contributes differentially to Ca²⁺ and Mn²⁺ utilization in roots and shoots. Higher Ca²⁺ concentrations in the bicat3-1 shoots are caused by the absence of BICAT3 in roots, while improved photosynthetic activities, higher shoot Mn²⁺ concentrations, and the curly-leaf phenotype of bicat3-1 under Mn²⁺ deficiency are due to its function in the shoot.

BICAT3 is essential for cell expansion and matrix polysaccharide biosynthesis under low Mn²⁺ availability

One of the most notable changes in the morphology of the bicat3 mutants is the curly-leaf symptom under Mn²⁺ deficiency (Figure 3A;Supplemental Figure S9A). Cross-sections of bicat3-1 leaf blades grown under Mn²⁺-deficient conditions showed an extremely compacted mesophyll with diminished intercellular space (Supplemental Figure S14B), accompanied by smaller palisade parenchyma, spongy parenchyma, and lower epidermis cells compared to the wild-type (Supplemental Figure S14, C and D). In addition, sections of spongy parenchyma cells were more circular in the mutant (Supplemental Figure S14E). The more compact tissue likely explains the observed decreased relative water content (Supplemental Figure S11).

Regulation of plant cell size and shape is a highly complex process, encompassing cell growth and cell division. Complex polysaccharides and, to a lesser extent, structural proteins form walls that confine the expansion of plant cells. The Golgi apparatus is a critical site for the posttranslational modification of proteins and matrix polysaccharide biosynthesis. Since Mn²⁺ is an indispensable co-factor for many glycosylation reactions, we analyzed the cell wall composition of bicat3 mutants, which showed striking differences in growth compared to the wild-type under Mn²⁺ deficiency.

Under control conditions, bicat3 and wild-type shoot cell walls had similar monosaccharide and glycosidic linkage abundances (Figure 5). However, under Mn²⁺ deficiency, numerous alterations were observed: fucose (Fuc), galactose (Gal), xylose (Xyl), and galacturonic acid (GalA) levels significantly decreased in both bicat3 mutants as compared to the Mn²⁺-deficient wild-type (Figure 5A). Conversely, the abundance of glucose (Glc), rhamnose (Rha), and mannose (Man) in shoot cell walls of Mn²⁺-deficient bicat3 was significantly higher compared to those of wild-type plants. Glycosidic linkages, characteristic of different polysaccharides, were massively altered in Mn²⁺-deficient bicat3 (Figure 5B;Supplemental Table S1), with significant increases of 2-Rha, t-Man, 3-Glc, and 4-Man linkages, accompanied by reductions in 2-Gal, 4-Gal, and 4-Glc. The highly elevated abundance of 3-Glc indicated an accumulation of callose in bicat3 shoots under Mn²⁺ deficiency, which was confirmed by aniline blue and immunogold staining of callose (Supplemental Figure S15). Callose accumulation is often triggered by an accumulation of reactive oxygen species. 3,3'-diaminobenzidine (DAB) staining indeed showed an accumulation of H₂O₂ under Mn²⁺ deficiency in bicat3-1, but not in the wild-type (Supplemental Figure S16).

Figure 5.

Cell wall matrix sugar components are altered in bicat3 shoots compared to the wild-type under Mn²⁺ deficiency. A, Monosaccharide composition of shoot cell wall AIR. Col-0 and bicat3 plants were cultured for 5 weeks in hydroponics with 3.5-µM Mn²⁺ (C) or 0.05 µM Mn (-Mn). B, Glycosidic linkages of Col-0 and bicat3 shoot cell wall AIR. Values represent molar percentage of total carbohydrates detected. Data in (A) and (B) indicate means + sd of four and three independent biological replicates, respectively. The complete data set can be found in Supplemental Table S1. Data were analyzed by two-tailed Student’s t test to identify significant differences between wild-type and mutants (^*P < 0.05; ^**P < 0.01; ^***P < 0.001).

In contrast to the shoots, the monosaccharide and linkage composition of bicat3 roots was relatively similar to the wild-type (Supplemental Figure S17), indicating organ-specific differences in Mn²⁺ handling and requirements that have yet to be explored.

Mutation of BICAT3 hampers seed set and pollen tube growth

Besides the vegetative growth aberrance under Mn²⁺ limitation, the bicat3 lines displayed a severe generative growth defect: both mutants produced short siliques with an incomplete seed set (Figure 6, A and B; Supplemental Figure S18). Additionally, bicat3 mutant seeds were significantly larger, heavier, and germinated faster than wild-type seeds (Supplemental Figure S19), which is the likely cause of the faster seedling development of the mutants (Supplemental Figure S6–S8). To assess the biological basis of the short-silique phenomenon, reciprocal crosses between bicat3-1 and wild-type were performed. Siliques obtained from crosses of bicat3-1 pollen with maternal bicat3-1 or wild-type showed a reduced length compared to those from the wild-type self-cross, whereas the siliques from bicat3-1 flowers fertilized with wild-type pollen were comparable to those from wild-type self-cross (Figure 6C). This result indicates a male gametophyte defect to be responsible for the short-silique phenotype.

Figure 6.

bicat3 mutants show male gametophyte defects that lead to shorter siliques compared to the wild-type. A, Siliques of Col-0, bicat3-1, and bicat3-2. Scale bar represents 1 cm and applies to all images. B, Silique length of Col-0, bicat3-1, and bicat3-2. Data indicate means + se of 30 siliques. Data were analyzed by two-tailed Student’s t test to identify significant differences between wild-type and mutant (^***P < 0.001). C, Reciprocal crossing of Col-0 and bicat3-1. Scale bar represents 1 cm and applies to all images. The experiment was repeated twice with similar results.

This observation provoked the question whether BICAT3 is active in pollen. The expression of BICAT3 in pollen grains and tubes was confirmed in both ProBICAT3-GUS and ProBICAT3-BICAT3-Venus reporter lines (Figure 7, A and B), suggesting a direct role of BICAT3 in pollen. Aniline blue staining of pollen tubes grown for 24 h in vivo showed that only a minor proportion of bicat3-1 pollen tubes grew normally and reached the basal end of the transmitting tract (Supplemental Figure S20). To substantiate the defect in mutant pollen tube elongation, pistils were pollinated and the top part of pollinated pistils, including stigma and style, excised and subsequently cultured for 12 h (Figure 7; Supplemental Figure S21). Under those semi in vivo conditions, only few bicat3-1 and bicat3-2 pollen tubes grew out of the excised wild-type and mutant pistils after they were cultured on medium for 12 h (Figure 7C). Contrastingly, numerous wild-type pollen tubes penetrated the stigma and style of wild-type and mutant pistils and continued to grow under the same conditions. The abnormal growth of pollen tubes can be complemented by the expression of BICAT3 driven by its native promoter (Supplemental Figure S21). Together, these results indicated that BICAT3 is critical for pollen tube elongation and consequently seed production.

Figure 7.

The absence of BICAT3 hampers pollen tube growth. A, Activity of BICAT3 promoter in pollen grains and pollen tubes as detected by GUS staining of ProBICAT3-GUS line. Scale bar represents 50 µm. B, Activity of BICAT3 promoter in pollen grains and pollen tubes of ProBICAT3-BICAT3-Venus line. Scale bars represent 100 µm. C, Semi in vivo pollen tube growth assay of Col-0, bicat3-1, and bicat3-2 germinated on Col-0, bicat3-1, or bicat3-2 pistils. Pictures were taken 12 h after pollination. Images taken from Supplemental Figure S21. Scale bar represents 1 mm and applies to all images.

bicat3 pollen tubes accumulate less low methyl-esterified homogalacturonan compared to wild-type pollen tubes

Cell wall polysaccharides such as pectin are crucial for pollen tube extensibility and growth (Hepler et al., 2013; Mollet et al., 2013; Dehors et al., 2019), suggesting that bicat3 pollen tubes may have altered cell walls. In addition to the retarded pollen tube growth of bicat3 mutants observed in vivo and in semi in vivo growth assays (Figure 7C;Supplemental Figure S20), bicat3 pollen tubes showed unusual swelling and branching in vitro (Supplemental Figure S22). Ruthenium red primarily stained pectin at the tip of wild-type and complementation line pollen tubes, which was reduced in both bicat3 mutants and accompanied by elevated staining at the branch points.

Pectin deposition was further analyzed by immunostaining of in vitro-grown pollen tubes with the JIM5 monoclonal antibody (Dardelle et al., 2010), recognizing partially methylesterified homogalacturonan (HG). Wild-type pollen tubes accumulated partially methylesterified HG in their apical region except for the very tip (Figure 8, A and B). Staining of partially methylesterified HG was severely reduced in both bicat3 mutants. In typical bicat3 tubes, the low, wild-type like intensity at the apex did not increase, but rather decreased, in the subapical region (Figure 8B), causing an overall lower staining intensity (Figure 8C). However, staining patterns of individual pollen tubes varied considerably. Partially methylesterified HG showed a heterogeneous distribution in swollen and branched bicat3 pollen tubes with increased accumulation at branch points or tips and a much lower abundance in the sub-apical region, in particular in swollen sections (Figure 8D). However, as indicated by the brightness adjustment, staining intensity in these accumulation hotspots was much weaker than that in areas of high fluorescence in wild-type tubes. In some notable exceptions, higher intensity that extended to the very tip was observed (Figure 8, C and D). The aberrant partially methylesterified HG distribution was abolished by complementation of the bicat3-1 mutant with the wild-type BICAT3, confirming that the absence of BICAT3 was causal for this phenotype (Figure 8, A–C). In summary, the data demonstrate the requirement of BICAT3 for pectin deposition in pollen tubes, whereby the heterogeneity in growth defects is mirrored by a heterogeneity in pectin deposition aberrances.

Figure 8.

bicat3 mutant pollen tubes contain less partially methylesterified HG with more variable deposition compared to the wild-type. A, Immunofluorescence labeling of in vitro-grown pollen tubes with JIM5 monoclonal antibody. Images were taken with identical settings. Scale bar represents 20 µm and applied to all images. B, bicat3 mutant pollen tubes contain a similar amount of JIM5-labeled partially methylesterified HG as the wild-type at the tip, but less in sub-apical regions. Fluorescence intensity of pollen tubes (10 µm from tip) of Col-0, bicat3-1 complementation line, and bicat3 mutants. Data indicate means ± se of at least six biological replicates. C, Fluorescence intensity of JIM5-labeled pollen tube apex (0–10 µm from tip) of Col-0, bicat3 complementation line, and bicat3 mutants. The box plot shows the range of the values (upper and lower bar), median (black line), means (dotted white line), and the lower and upper quartiles. Single data points are indicated by gray circles. Data were analyzed by Kruskal–Wallis one-way analysis of variance on ranks with Dunn’s method to identify significant differences between wild-type and mutants or Comp and mutants (^*P < 0.05; ^**P < 0.01). D, Heterogeneity of JIM5 labeling of in vitro-grown bicat3 pollen tubes. a–d, Pollen tubes show low fluorescence in the sub-apical region and a, higher fluorescence at the base of a branch; b, higher fluorescence at the tip and very low fluorescence at the swollen part; c, lower fluorescence at the swollen shank and higher fluorescence at a constricted section; d, homogenous but low fluorescence at the apical and sub-apical region. e, Exceptional tubes show high fluorescence at the apex. Scale bars represent 10 µm. The experiment was repeated twice with similar results.

The physiological relevance of BICAT3 is distinct from that of ECA3

The growth defect of bicat3 mutants was only apparent at low Mn²⁺ supply, and their male-sterility was incomplete. Both phenotypes are related to alterations in matrix polysaccharide composition of the cell wall, most likely inflicted by malfunction of Mn²⁺-dependent glycosyl transferases in the Golgi. Since Mn²⁺ is essential for these reactions, mechanisms other than BICAT3 have to exist to supply the Golgi with the metal. The CDF-type Mn²⁺ transporter MTP11 has been localized to the Golgi apparatus (Peiter et al., 2007). However, mtp11 mutants have a phenotype opposite to that of bicat3 mutants, showing hypertolerance to Mn²⁺ limitation. In contrast, mutants of the Mn²⁺-transporting P-type ATPase ECA3, which has also been localized to the Golgi, are hypersensitive to low Mn²⁺ supply (Mills et al., 2008). We therefore hypothesized that ECA3 may operate alongside BICAT3, and that double mutants would hence show exacerbated phenotypes. However, in contrast to bicat3 mutants, silique length was not reduced in an eca3 mutant as compared to the wild-type (Figure 9A). Together with the very low expression of ECA3 in pollen (Qin et al., 2009), this indicates that this pump does not play a notable role in Mn²⁺ supply of the pollen Golgi.

Figure 9.

BICAT3 and ECA3 are both involved in Mn²⁺ efficiency, but with distinct roles. A, Siliques of Col-0, bicat3-1, bicat3-2, and eca3-2. Scale bar represents 1 cm and applies to all images. B, Growth phenotypes of 6-week-old Col-0, bicat3-1, eca3-2, and dko (bicat3-1 eca3-2) plants in grown in hydroponics with 3.5 µM Mn²⁺ (control) or 0.05 µM (-Mn). Scale bar represents 3 cm and applies to all images. C, Fv/Fm images of Col-0, bicat3-1, eca3-2, and dko plants grown under control and Mn²⁺-deficient conditions. Images were digitally extracted for comparison. Scale bar represents 3 cm and applies to all images. The experiments were repeated twice with similar results.

The situation was different in vegetative tissues. As apparent in Figure 9B, the rosette size of 6-week-old plants grown under Mn²⁺ deficiency was diminished in eca3-2 as well as in bicat3-1, as observed before (Figure 3A), and the double mutant displayed an exacerbation of this phenotype, which was accompanied by a likewise decrease of root and shoot dry weight (DW) (Supplemental Figure S23). The effect of both mutations on photosynthesis was unequal. Whereby the decrease in Fv/Fm and Y(II) under Mn²⁺ deficiency was strongly alleviated in bicat3-1 and the double mutant, this was less pronounced in eca3-2 (Figure 9B;Supplemental Figure S23, D and E). The previously observed increase in Mn²⁺ concentration in Mn²⁺-deficient bicat3 was more moderate in eca3-2 and more pronounced in the double mutant (Supplemental Figure S23B), whereas for Ca²⁺, an additive effect of both mutations was not noticeable (Supplemental Figure S23C). The mutant phenotypes indicate that BICAT3 and ECA3 do not function redundantly, but partly cooperate for efficient plant performance under limited Mn²⁺ availability.

Discussion

Manganese is an indispensable metal for a diversity of processes in plants. One of the essential roles of Mn²⁺ is as a cofactor of enzymes that mediate protein glycosylation and polysaccharide biosynthesis in the Golgi apparatus (He et al., 2021). Despite the crucial role of Mn²⁺ in the plant’s Golgi, little is known about the transport proteins responsible for the Mn²⁺ supply of Golgi functions. In this work, we found that the ubiquitously expressed trans-Golgi-localized BICAT3 protein transports Mn²⁺ and Ca²⁺ in a heterologous expression system. In the absence of BICAT3, plants showed numerous aberrances under Mn²⁺ deficiency, such as growth retardation, changes in rosette morphology, and defects in glycan synthesis, albeit a strikingly better maintenance of photosynthetic activity was observed. Besides, bicat3 mutants have a defective pollen tube growth with aberrant pectin accumulation and, ensuing from this, produce shorter siliques with fewer seeds. Thus, BICAT3 plays a vital role in maintaining Golgi functions for both vegetative and reproductive growth of the plant. The biochemical phenotypes of BICAT3 revealed in this work point to a direct involvement in Mn²⁺ homeostasis, whereby additional roles for Ca²⁺-dependent processes cannot be excluded and may also underlie some of the macroscopic phenotypes.

BICAT3 determines responses to varying Mn²⁺ and Ca²⁺ supply

Heterologous expression of BICAT3 in yeast mutants that are sensitive to different metals showed that BICAT3 restored their Ca²⁺ and Mn²⁺ tolerance (Figure 2). BICAT3-complemented pmr1Δgdt1Δ yeast grew comparably to the pmr1Δ single mutant under 600-mM CaCl₂, indicating the functional conservation of BICAT3 and GDT1. The complementation experiments suggest that BICAT3 is specifically involved in Ca²⁺ and Mn²⁺ transport when heterologously expressed in yeast. Such a poor selectivity between Ca²⁺ and Mn²⁺ is characteristic of transport proteins belonging to numerous families. In plants, these include ER-type Ca²⁺-ATPases (e.g. ECA1 Wu et al., 2002) and ECA3 (Li et al., 2008; Mills et al., 2008), Calcium Exchangers (e.g. CAX2; Shigaki et al., 2003), as well as BICAT proteins, for which transport of Ca²⁺ and Mn²⁺ have been directly demonstrated in BICAT1 and 2 (Schneider et al., 2016; Wang et al., 2016; Frank et al., 2019). The physiological relevance of the dual functionality is largely unresolved, and it is unclear how homeostasis of both elements is regulated independently by nondiscriminating transporters (He et al., 2021).

Both young (2-week-old) and older (5-week-old) plants of bicat3 loss-of-function mutants showed altered responses to different Ca²⁺ and Mn²⁺ supply levels compared to the wild-type. In conditions of Mn²⁺ deficiency and Ca²⁺ toxicity, bicat3 mutants performed worse than the wild-type, while they responded superiorly to Mn²⁺ toxicity and Ca²⁺ deficiency (Supplemental Figure S6–S9). The most striking growth difference between bicat3 mutants and the wild-type was found under Mn²⁺ starvation. Collectively, these results indicate a role of BICAT3 in Mn²⁺ and Ca²⁺ homeostasis of Arabidopsis, in particular during low Mn²⁺ availability. It remains to be clarified if the altered response of bicat3 mutants to extreme Ca²⁺ supply is due to a Ca²⁺ transport activity of the transporter, or due to the Mn²⁺-versus-Ca²⁺ selectivity of the processes that back up BICAT3. In mammalian cells, Mn²⁺-mediated rescue of glycosylation defects inflicted by deletion of the BICAT3 homolog, TMEM165, is likely dependent on ER-resident SERCA-type ATPases, indicating the ER potentially contributes to the supply of Mn²⁺ to the Golgi (Houdou et al., 2019). In contrast, the Golgi-localized SPCA1 Ca²⁺/Mn²⁺-ATPase did apparently not contribute to the rescue of Mn²⁺-dependent glycosylation in tmem165 mutants (Houdou et al., 2019). It is unresolved if the normal vegetative growth and glycosylation pattern of bicat3 mutants under control conditions are due to other Golgi-localized Mn²⁺ transport proteins or a supply via vesicular trafficking from the ER. The comparative analysis of bicat3, eca3, and bicat3 eca3 mutants (Figure 9) indicated that the Golgi-localized Mn²⁺ pump ECA3 does most likely not operate alongside BICAT3, but both proteins may function in series, either subcellularly in different cisternae, or in different cell types.

The bicat3 mutants were more tolerant to Mn²⁺ toxicity in both plate and hydroponic cultures under long-day and short-day conditions, represented by a higher shoot and root biomass than that of the wild-type (Supplemental Figures S8 and S9). Interestingly, bicat3 shoots accumulated less Mn²⁺ compared to those of the wild-type under elevated Mn²⁺ conditions, which likely explains the better growth of bicat3 under Mn²⁺ toxicity. In contrast to our results, a previous study reported a higher sensitivity of bicat3 (pml3) mutants to Mn²⁺ toxicity compared to the wild-type (Yang et al., 2021). These findings are difficult to reconcile with our findings and might be due to different media (nutrient composition and concentration), growth conditions (light cycle and light intensity), or plant age.

BICAT3 determines matrix polysaccharide biosynthesis

BICAT3 operates as a Golgi-localized Mn²⁺ transporter, and the biosynthesis of the cell wall matrix polysaccharides, pectin, and hemicellulose, is mediated by Golgi-localized glycosyl transferases, many of which require Mn²⁺ (He et al., 2021). We, therefore, hypothesized that the prominent aberrations in leaf shape, and cell size and morphology of Mn²⁺-deficient bicat3 mutants are related to an altered cell wall matrix. This notion was strongly supported by a combined analysis of monosaccharide and glycosidic linkage composition, which showed substantial changes in pectin structure (Figure 5). A reduction in the abundance of GalA indicated that the overall pectin content was lower in shoot cell walls of Mn²⁺-deficient bicat3 mutants as compared to those of the wild-type. Additionally, the abundance of Rha and the 2-Rha linkage was increased, while that of Gal and the 4-Gal linkage was severely decreased. These aberrances indicate that the β-1,4-galactan side chain substitution of RG I was affected in bicat3 shoot cell walls (He et al., 2021). In Arabidopsis, this addition of galactose from UDP-α-D-Gal to growing β-1,4-galactan chains is catalyzed by highly Mn²⁺-dependent galactan synthases (GALS1 to 3) (Ebert et al., 2018; Laursen et al., 2018), which were probably not sufficiently supplied with their metal cofactor in the bicat3 mutants.

Galactan side chain substitution of RG I occurs in the trans cisternae of the Golgi (Zhang and Staehelin, 1992), which corresponds to the primary localization of BICAT3, as identified by high-resolution CLSM and immunogold staining (Figure 1). In this respect, a previously claimed preferable localization of BICAT3 (PML3) in the cis-Golgi (Yang et al., 2021), which was concluded from conventional CLSM data, cannot be confirmed.

Along with xyloglucans, galactan domains of RG I interact with cellulose microfibrils, hence maintaining cell wall structure (Zykwinska et al., 2007). A truncation of galactan side chains also interferes with porosity and thus water-binding capability of the plant cell walls (Klaassen and Trindade, 2020). However, a gals1 gals2 gals3 triple mutant did not display a macroscopic phenotype (Ebert et al., 2018), suggesting that the severe reduction of 4-Gal does not account for the impaired growth of Mn²⁺-deficient bicat3 mutants.

By association with receptor-like proteins (RLP) or wall-associated kinases (WAK), pectins not only modulate the extensibility, flexibility, and rigidity of the cell wall, but are also involved in feedback regulation to control wall homeostasis during cell expansion (Franck et al., 2018). Thereby WAKs are essential for cell expansion and pathogen resistance (Kohorn and Kohorn, 2012). WAK2 binds to pectin in the cell wall, and WAK2 antisense plants show small and curly-bent leaf phenotypes with smaller cell size (Wagner and Kohorn, 2001). The curled-leaf phenotype of bicat3 might thus be related to signaling pathways induced by the defects in pectin biosynthesis under Mn²⁺ deficiency.

As compared to the shoot, aberrances in monosaccharide composition and linkage in the root were much less pronounced (Supplemental Figure S17). This is in disagreement with a previous analysis, which describes a morphological disintegration of the root apex (Yang et al., 2021). This was not observed in any of our experiments, with exemplary images shown in Supplemental Figure S24. Again, the reason for this disagreement is unclear and may lie in the growth conditions.

The abundance of callose was drastically increased in cell walls of bicat3 shoots (Supplemental Figure S15), which coincided with a higher abundance of Glc and 3-Glc linkage (Figure 5). Callose is not synthesized in the Golgi, but by callose synthases at the plasma membrane. Hence, callose deposition in bicat3 mutants likely is an indirect response to glycan defects inflicted by the absence of BICAT3. Callose deposits play multiple roles in plants, in particular in the regulation of plasmodesmata and in responses to biotic and abiotic stresses (De Storme and Geelen, 2014; Rissel et al., 2017; Hunter et al., 2019; Wang et al., 2021). Recently, callose synthesis was reported to be induced by Ca²⁺ deficiency, preventing cell death (Shikanai et al., 2020). Due to its function in cross-linking pectins, Ca²⁺ deficiency is believed to induce cell wall damage that is counteracted by callose depositions. In support of this, expression of a cell wall damage-induced transcription factor, OsWRKY42, in Arabidopsis leads to intensified callose accumulation (Pillai et al., 2018). We therefore hypothesize that the abnormal accumulation of callose in bicat3 mutants is a result of cell wall defect-induced signaling pathways under Mn²⁺ deficiency. This is supported by the increased accumulation of the ROS H₂O₂ (Supplemental Figure S16), which is a common component of those pathways.

A previous study also found alterations in shoot cell wall composition of bicat3 mutants, which, however, did not fully correspond to our observations (Yang et al., 2021). While similar Alcohol-insoluble residue (AIR) preparation methods were used in the two studies, different analytical methods (high-performance anion-exchange chromatography coupled with pulsed amperometric detection [HPAEC-PAD] here and gas chromatography–mass spectrometry [GC–MS] in Yang et al. (2021)) were used for the final quantification. While the GC–MS method only detected neutral sugars after derivatization, HPEAC-PAD allowed us to also quantify the abundance of GalA, the main building block of pectin. Nevertheless, the major changes in the Mn²⁺-deficient mutant shoots differed between the two studies: less Man, Ara, Xyl, and Glc compared to the wild-type in Yang et al. (2021); less Gal, but more Glc and Man relative to the wild-type in this work (Figure 5A). These differences should not be the result of the different analytical techniques used, but instead suggest that the precise cultivation conditions likely have a profound effect on modulating the bicat3 mutant chemotypes.

Taken together, under limited Mn²⁺ supply, the loss of BICAT3 causes multiple direct and indirect changes in cell wall polysaccharide biosynthesis. It remains to be determined which of the observed macroscopic phenotypes are a direct result of which of those alterations.

BICAT3 determines pollen tube growth and seed set

Apart from their drastic vegetative phenotype under Mn²⁺ deficiency, bicat3 mutants were severely affected in reproduction, producing fewer but larger seeds in shorter siliques (Figure 6; Supplemental Figure S19). As evident from reciprocal crossing, this reduced fertility was caused by a defect in the male gametophyte (Figures 6, C and 7, C). Pollen tubes of bicat3 mutants showed highly heterogeneous morphologies and were often branched or swollen (Figure 8; Supplemental Figure S22).

We hypothesized that the defective pollen tube growth was caused by a Mn²⁺-dependent function of the Golgi. Elongation of the pollen tube requires the deposition of cell wall material at the tip. The most abundant pectin, HG, is synthesized in the Golgi by HG synthetases of the GAUT family (Atmodjo et al., 2011), which contain a Mn²⁺-binding DXD motif. For the GAUT1:GAUT7 enzyme complex an absolute dependence on Mn²⁺ has been experimentally shown (Amos et al., 2018). Numerous GAUT genes are expressed in pollen, most prominently GAUT13 and 14. Pollen tubes of the gaut13 gaut14 mutant are swollen and defective in tube growth (Wang et al., 2013), similar, but more severely, to what we observed in bicat3 mutants. Also mutants of the functionally redundant GAUT5, 6, and 7, that are more weakly expressed in pollen, were defective in tube elongation (Lund et al., 2020). In those GAUT mutants, HGs are unconventionally distributed, with weakly esterified HG, detected by the JIM5 antibody, being severely reduced in the gaut13 gaut14 mutant (Wang et al., 2013). This defect closely resembles our observations on bicat3 pollen. The abundance of weakly esterified HG in subapical regions was generally severely diminished in bicat3, which most likely led to a loss in rigidity and hence in a swelling of the otherwise perfectly cylindrical tube.

The pattern of pectin deposition and the tube morphology showed substantial variability in the bicat3 mutant, which may be related to varying Mn²⁺ resources in the pollen population, affecting Mn²⁺-dependent processes to different degrees (Figure 8). This was also evident in vivo, where only some pollen tubes extended sufficiently to reach an ovule. The abnormal swelling was always associated with a decreased HG abundance, and abnormal branching was accompanied by a localized HG deposition. In agreement with previous analyses (Chebli et al., 2012; Zhang et al., 2021), the deposition of weakly esterified HG was weaker at the tip of the wild-type tube, and this tip-low pattern was not apparent in the bicat3 mutants (Figure 8B). However, the distribution of weakly esterified HG in bicat3 differed somewhat to that found in a previous report, which described a consistently increased deposition at the tip of bicat3, with no absolute difference to the wild-type in subapical regions (Zhang et al., 2021). Those different patterns may have been caused by different intensities of Mn²⁺ limitation.

Despite the close match of aberrant HG deposition and growth defects, a contribution of other Mn²⁺-dependent processes to the phenotype cannot be excluded. Naturally, Golgi-localized enzymes involved in the synthesis of further cell wall polyssacharides and glycoproteins may also be affected. For example, MGP4 involved in RG-II biosynthesis (Liu et al., 2011), GALT2-6 and HPGT1-3 involved in AGP biosynthesis (Ogawa-Ohnishi and Matsubayashi, 2015; Basu et al., 2015a, 2015b), and HPAT1-3 involved in extensin biosynthesis (Ogawa-Ohnishi et al., 2013), all show pollen tube growth defects, and any contribution of these proteins to the bicat3 phenotypes remains to be demonstrated.

BICAT3 restricts Mn²⁺ bioavailability for chloroplasts

Intriguingly, although growth of bicat3 mutants suffered substantially under Mn²⁺ deficiency, photochemical efficiency of the mutants was superior to that of the wild-type (Figure 4). This coincided with moderately higher Mn²⁺ concentrations in shoots of the mutants (Figure 3). Grafting experiments revealed that improved photosynthesis and the morphological phenotype of bicat3-1 were both caused by the loss of BICAT3 in the shoot (Supplemental Figure S13). Since the photosynthetic light reaction is highly sensitive to Mn²⁺ deficiency (Alejandro et al., 2020), we assumed that the intracellular allocation of Mn²⁺ to chloroplasts may be altered in bicat3 mutants. Indeed, under Mn²⁺ deficiency bicat3-1 and bicat3-2 chloroplasts accumulated more Mn²⁺ than those of the wild-type, rendering the Mn²⁺ limitation less severe (Figure 4). This indicates that BICAT3 is able to govern the competition of different cellular compartments for Mn²⁺. The effect of the bicat3 mutations on photosynthesis was more pronounced than that on chloroplast Mn²⁺ content. However, we do not expect both parameters to be linearly associated but rather in a threshold-like manner. Below a certain concentration of free Mn²⁺ in the chloroplast, synthesis, and stability of the water-splitting complex will be severely affected. A small increase in chloroplast Mn²⁺ concentration may therefore cause a marked improvement of photosynthesis at low Mn²⁺ supply. It currently cannot be discerned whether Mn²⁺ supply of other organelles, such as mitochondria, is also improved in the absence of BICAT3.

In this respect, the TGN-localized transporter NRAMP2, in conjunction with the vacuolar transporters NRAMP3 and 4, has previously been proposed a role in inter-organellar Mn²⁺ distribution toward chloroplasts under Mn²⁺-limiting conditions. This appears to be supported by decreased or increased chloroplast Mn²⁺ concentrations in NRAMP2 knockdown or overexpression plants, respectively (Alejandro et al., 2017). However, since expression of NRAMP2, 3, and 4 in leaves is strictly confined to the vascular system (Thomine et al., 2003; Alejandro et al., 2017), the operation of an NRAMP2-driven Mn²⁺ supply chain of chloroplasts and a possible interference of BICAT3 with this process demands further scrutiny.

Taken together, BICAT3 occupies a central position in Mn²⁺ use efficiency of Arabidopsis, both directly by supplying Golgi cisternae with Mn²⁺, and indirectly by determining intracellular Mn²⁺ distribution. Aberrant Mn²⁺ homeostasis in the Golgi apparatus causes specific defects in glycosyl transfer reactions, affecting cell size, leaf shape, and the tip growth of pollen tubes. The fact that this Mn²⁺ transporter and Mn²⁺-dependent processes depending on it are located in the same cisternae supports the model of a subcompartmental arrangement of the polysacharide assembly line. Numerous Mn²⁺ transporters have by now been identified in the plant’s secretory pathway. These do not function merely redundantly, but have distinct and sometimes even opposite physiological functions. The differential relevance of BICAT3 and ECA3 is exemplary for this not yet fully understood division of labor. Elucidation of the functional interaction of those Mn²⁺ transporters on subcompartmental, subcellular, and tissue level is a consequential next step to be pursued.

Materials and methods

Plant materials and growth conditions

Arabidopsis (A. thaliana) T-DNA insertion mutant lines were supplied by the Nottingham Arabidopsis Stock Centre. bicat3-1 (GABI_027F07) and bicat3-2 (SALK_97998) mutants with a Columbia-0 (Col-0) background were obtained from the GABI collection and the SALK collection, respectively (Ülker et al., 2008). Homozygous plants were identified by using gene-specific primers as listed in Supplemental Table S2. The insertions were identified with GABI_LB and SALK_LB primers for the lines of the GABI collection and the SALK collection, respectively. For identification of gene knockout, gene-specific primers were used in reverse transcription–PCR (RT–PCR). Amplification of Beta-6 Tubulin (TUB6; At5g12250) was used as an internal control. To quantify the expression of BICAT3 in the knock-down mutant, RT–qPCR was performed. Actin2 (ACT2; At3g18780) was used as housekeeping gene.

To generate a bicat3-1 complementation line, a genomic DNA fragment containing the 5′-UTR and promoter region (2,017-bp upstream of the ATG), the coding sequence, and the 3′-UTR (493-bp downstream of the TAG) of BICAT3 was cloned into pGreenII, yielding pGreenII-gBICAT3. bicat3-1 plants were stably transformed with the pGreenII-gBICAT3 construct by floral dip (Clough and Bent, 1998). Seedlings were selected by spraying with Basta. This line was used in the semi in vivo pollen assay and in immunostaining of pollen tubes.

To assemble the pCAMBIA2300u-PrBICAT3-BICAT3-Venus-tNos vector, the region 1,811-bp upstream of the BICAT3 start codon, BICAT3-Venus, and a Nos terminator were amplified from Col-0 genomic DNA, pART7-BICAT3-Venus, and pILTAB381, respectively, using USER cloning-compatible primers (Nour-Eldin et al., 2006). The pART7-BICAT3-Venus vector was generated by inserting the BICAT3 CDS into pART7-Venus using XmaI-containing primers. bicat3-1 plants were stably transformed with the pCAMBIA2300u-PrBICAT3-BICAT3-Venus-tNos construct by floral dip. Seedlings were selected on half-strength Murashige and Skoog (1/2 MS) plates containing kanamycin (50 μg mL⁻¹). This line was used in pollen tube and immunogold localization experiments.

To cultivate plants on 1/2 MS agar plates, seeds of wild-type and transgenic Arabidopsis were sterilized in 70% (v/v) ethanol for 1 min, 30% (v/v) bleach for 5 min and rinsed 5 times with sterile water. Seeds were stratified for 3 days at 4°C in darkness.

For testing the growth of plants under Mn²⁺ deficiency on plates, stratified seeds were sown on self-made 1/2 MS agar with or without Mn²⁺ (Supplemental Table S3), containing 8 g L⁻¹ agar (No. 2,266.1, Carl Roth, Karlsruhe, Germany) and 2.5 g L⁻¹ sucrose. Plants were cultured for 2 weeks in a growth cabinet (AR75, Percival Scientific, Perry, IA, USA) set to 21°C day and 19°C night temperature and 16-h light period with a light intensity of 150-µmol m⁻² s⁻¹ photons. For root length measurements, plates were scanned every other day from the 4th until the 14th day. Primary root length of seedlings was measured by the NeuronJ plugin of Fiji.

For testing the growth of plants under Ca²⁺ and Mn²⁺ toxicity on plates, stratified seeds were precultured for 7 days on 1/2 MS agar including B5 vitamins (M0231, Duchefa Biochemie, Haarlem, the Netherlands), containing 8-g L⁻¹ agar (Phyto Agar P1003, Duchefa Biochemie) and 2.5-g L⁻¹ sucrose, adjusted to pH 5.8 with KOH. Plants were cultured in a growth cabinet as above. Seven-day-old seedlings were transferred to new 1/2 MS media with or without additional 50-mM CaCl₂ or 1-mM MnSO₄, and cultured for another 8 days. Plates were scanned every other day after transferring the seedlings for root length measurement. Primary root length of seedlings was measured as above.

For cultivation of Arabidopsis in hydroponics, stratified seeds were sown on 1/2 MS plates as above. After 12 days of growth, seedlings were transferred to seedling holders created by removing the tip of a 1.5-mL centrifuge tube and a cylindrical polyethylene foam piece (BKF A. Fleuren, Friesoythe, Germany). A nutrient solution based on Peiter et al. (2007) was used as hydroponics basal medium (Supplemental Table S3). After growing for 10 days in 1-L containers with basal medium, roots and seedling holders were washed 3 times with ddH₂O. Seedlings were transferred to 4-L containers containing control or treatment media, and cultivated further for 2 weeks. In the case of recovery experiments, plants were cultivated in basal media for one more week after 2 weeks of Mn²⁺ deficiency treatment. Plants on plates and in hydroponics were cultured in a growth cabinet (ATC26, Conviron, Winnipeg, Canada) set to 21°C day and 19°C night temperatures and 10-h light period with a light intensity of 150-µmol m⁻² s⁻¹ photons. The nutrient solution was changed twice a week.

Subcellular localization

The full-length BICAT3 coding sequence (CDS) was used to construct pART7-BICAT3-Venus and pART7-BICAT3-mCherry. Organellar markers for co-localization experiments were obtained from Nelson et al. (2007). Protoplasts of the Arabidopsis Col-0 wild-type and an ST-GFP trans-Golgi marker line (Hawes and Satiat-Jeunemaitre, 2005) were prepared and transiently transformed with the constructs as previously described (Peiter et al., 2007). Twenty four hours after transformation, fluorescence was observed by confocal laser microscopy using a LSM 880META containing a PMT detector and with a planapochromatic lens (63×/1.4 Oil) (Carl Zeiss, Jena, Germany). Images were acquired in lambda mode (Figure 1) or channel mode (Supplemental Figure S2) with a resolution of 1,024 × 1,024 pixels and 512 × 512 pixels, respectively. Excitation light source was a 488-nm multiline ion laser and a 561-nm argon laser. Images were acquired by using the Zeiss ZEN blue software; the display range was optimized by applying the Best Fit function in the software. In channel mode, pictures were acquired by using the multitrack function. Venus and mCherry were detected by using a 446–553 and 588–651-nm band pass filter, respectively. Images in lambda mode were acquired at a collection bandwidth of 410–651 nm in plane scan mode for single lines with four repetitions. GFP and mCherry signals were separated by linear unmixing using previously generated spectra obtained from protoplasts transiently transformed with a pART7 vector containing the respective fluorophore.

For localization of fusion proteins in Golgi cisternae, Agrobacterium tumefaciens GV3101 containing the desired constructs were grown to an optical density (600 nm) of 0.6, harvested, and resupended in Infiltration Buffer (10-mM MgCl₂, 5-mM MES [KOH pH 5.3], 150-µM acetosyringone). The desired combinations were mixed, and leaves of 5-week-old Nicotiana benthamiana plants were infiltrated with a 1-mL syringe. After 48 h, infiltrated leaf pieces were cut out and water-soaked using a syringe. The leaf disc was placed on a high-precision coverslip (170 ± 5 µm, Marienfeld, Lauda-Königshofen, Germany) with water as mounting medium and covered with a second coverslip. Images were taken with a STELLARIS 8 microscope (Leica, Wetzlar, Germany) equipped with a HC PL APO CS2 63×/1.30 glycerol objective. A White Light Laser was used as excitation light source, and GFP and mCherry were excited at 489 and 587 nm, respectively. Emitted light was detected with the Hybrid-Detector (HyD) with the respective ranges of 494–572 and 592–649 nm. Images were edited using the LAS X software (Leica).

For ultrastructural localization of BICAT3-Venus, leaf segments were cryo-fixed and freeze-substituted as described (Tabassum et al., 2020). For immunolabeling of ultrathin sections, we used a polyclonal anti-Venus antibody (St John’s Laboratory, London, UK; diluted 1:300) detected by a rabbit anti-goat secondary antibody conjugated with 10-nm gold (G5527, Sigma; diluted 1:100). Sections were poststained with uranyl acetate and lead citrate using an EMSTAIN instrument (Leica, Wetzlar, Germany) and observed with a Libra 120 transmission electron microscope (Carl Zeiss) operating at 120 kV. Images were taken by using a dual-speed on-axis SSCCD camera (BM-2k-120; TRS, Moorenweis, Germany).

For observation of BICAT3-Venus in pollen and pollen tubes, pollen of bicat3-1 stably expressing BICAT3-Venus was harvested and cultured in vitro for 3 h as described below and subsequently observed by fluorescence microscopy using a AxioCam MRm Rev. 3 camera mounted on an Axio Observer.Z1 (Carl Zeiss) microscope equipped with filter set FS46HE (excitation 488–512 nm, emission 520–550 nm) and a Plan Apochromatic lens (40×/1.3 oil). Excitation light source was a HXP-120 lamp. Images were acquired by using the Zeiss ZEN blue software; the display range was optimized by applying the Best Fit function of the software.

Histochemical GUS staining

For promoter-GUS studies, the BICAT3 promoter region 1,811-bp upstream of the start codon was amplified from Col-0 genomic DNA and cloned into pBl101 using Xma I, upstream of the uidA gene (Peiter et al., 2007). Arabidopsis Col-0 plants were stably transformed with this construct by the floral-dip method (Clough and Bent, 1998). Transformants were selected using kanamycin. Plants were cultivated either on sterile 1/2 MS agar plates or in soil. Plant materials were transferred to GUS staining solution containing 100-mM sodium phosphate buffer (pH 7.0), 10-mM EDTA, 3-mM K₄[Fe(CN)₆], 0.5-mM K₃[Fe(CN)₆], 0.1% (v/v) Triton X-100, and 2-mM 5-Bromo-4-chloro-3-indolyl-beta-d-glucuronic acid (X-Gluc, X-Gluc Direct, Malaga, Spain) in DMSO. Samples were vacuum-infiltrated with staining solution 2 times for 5 min and subsequently stained for 3 h at 37°C. Chlorophyll was removed by 80% (v/v) ethanol. Of 15 independent promoter-GUS lines, 13 showed the same staining pattern. Two homozygous single-insertion lines were chosen for in-depth analysis. The stained tissues were documented with an AxioCam HRc digital camera (Carl Zeiss) mounted on a SteREO Discovery V.20 stereomicroscope (Carl Zeiss).

For GUS staining of pollen and pollen tubes, pollen was harvested and cultured in vitro for 3 h as described below and subsequently stained with GUS staining solution for 3 h at 37°C.

Yeast complementation

The pRS416-BICAT3 plasmid was prepared by inserting the BICAT3 full-length sequence in pRS416 using BamHI and SalI. Yeast (S. cerevisiae) wild-type (BY4741, Y00000, Euroscarf, Oberursel, Germany) and mutants BY4741-pmr1Δ and BY4741-gdt1Δpmr1Δ were transformed with pRS416 or pRS416-BICAT3 as described (Frank et al., 2019). Transformants were selected on synthetic complete (SC) media lacking uracil. Yeast drop assays were performed on SC-Ura (pH 5.5) and AP-Ura (pH 5.5) media to assess Mn²⁺ toxicity and Ca²⁺ toxicity, respectively, as described previously (Peiter et al., 2007; Frank et al., 2019).

Metal concentration measurements

For determination of metal concentrations, shoots were excised and roots were sequentially washed in ice-cold washing buffer I (1-mM MES-KOH, pH 5.8) for 10 min, in washing buffer II (10-mM EDTA and 1-mM MES-KOH, pH 5.8) for 10 min, and in MilliQ-H₂O for several times. Shoots and roots were dried for 3 days at 65°C in a drying cabinet (Linn High Therm, Eschenfelden, Germany). Dried shoots and roots were weighed into PFA vessels (CEM, Matthews, NC, USA) and digested in 65% (v/v) HNO₃ for 20 min at 190°C in a MARS 5 Xpress (CEM) microwave oven with the temperature ramped to 190°C in 15 min. Elements were analyzed by microwave plasma—atomic emission spectrometry (MP-AES; 4210, Agilent Technologies, Palo Alto, CA, USA).

For the determination of metal concentrations in chloroplasts, plants were cultivated and treated in a hydroponic system as described above. After removing midrib veins, leaf samples (around 2-g fresh weight) were homogenized in 45-mL homogenization buffer (0.4-M sorbitol, 20-mM Tricine-NaOH [pH 8.4], 10-mM EDTA, 1-g L⁻¹ BSA, 5-mM NaHCO₃, and 1 -mM MgCl₂) by using a blender (Waring, Stamford, CT, USA) at low speed for 5 s. The mixture was filtered through 75-µm mesh (prewetted with homogenization buffer) into 50-mL tubes. Subsequently, the extracts were squeezed through two layers of prewetted 10-µm nylon mesh and concentrated at 1,500 g for 5 min. The pellet was resuspended in at least 2-mL washing buffer (80-mM sorbitol, 4-mM Tricine-NaOH [pH 8.4], 0.5-mM EDTA, and 1-mM MgCl₂) and then loaded gently onto a discontinuous 40/85% (w/v) Percoll gradient in washing buffer followed by centrifugation (4,000 g, 15 min, 4°C) in a swinging bucket rotor. Around 1-mL intact chloroplasts were recovered from the 40/85% (w/v) Percoll interphase. Three volumes of washing buffer (around 3 mL) were added, gently mixed, and centrifuged (10 min at 200 g followed by 1 min at 1,700 g; at 4°C). Chloroplasts were washed 3 times with 2-mL cold washing buffer. The integrity of the chloroplasts was verified by microscopy, and the number of chloroplasts was determined by counting with a Neubauer chamber. An aliquot of 500-µL chloroplast suspension together with 100-µL 65% (v/v) HNO₃ were pipetted into a 15-mL Falcon tube and digested overnight at 70°C. The ion concentrations were determined by ICP-MS using an iCAP-RQ ICP-MS instrument (Thermo Fisher Scientific, Bremen, Germany) fitted with a MicroFlow PFA-200 nebulizer and a Cetac ASX-560 (Teledyne, Cetac Technologies, Omaha, NE, USA) autosampler.

Photosynthesis measurements

Plants were dark-adapted for 30 min, and all fluorescence measurements were performed on intact plants at room temperature. Images were captured, and maximal photochemical efficiency of PSII (Fv/Fm) and photochemical efficiency of PSII (Y (II)) changes with increasing light intensities were determined by IMAGING-PAM MAXI Version (IMAGING-PAM M-Series, Walz, Effeltrich, Germany). Saturating flashes (902 μmol m⁻² s⁻¹) were used to measure the maximum quantum yield of PSII. The photochemical efficiency of PSII was measured under a series of flashes with increasing intensity (1, 7, 21, 42, 69, 101, 140, 185, 237, 361, 512, 681, and 902 μmol m⁻² s⁻¹).

Grafting

For reciprocal grafting of bicat3-1 and Col-0 wild-type, stratified seeds were sown onto 1/2 MS plates, which were placed vertically in a growth cabinet set to 21°C day and 19°C night temperatures and 10-h light period with a light intensity of 150-µmol photons m⁻² s⁻¹ and grown for 5–6 days. Grafting was performed by using the Two Segment Shoot-Root Graft method according to Melnyk (2017) with some modifications. Sterile water was added to an empty petri dish containing two pieces of sterile Whatman 3MM Chr cellulose chromatography paper (WHA3030917, Whatman, Maidstone, UK) and one sterile Amersham Hybond-N⁺ membrane (RPN303B, Cytiva, Marlborough, MA, USA) strip on the top. One cotyledon of the seedling was cut off and discarded, and a transverse butt-end cut through the hypocotyl close to the shoot was made. Scions and rootstocks were grafted on the Hybond membrane strip, and the redundant water in the petri dish was removed by sterile Whatman paper strips. Three days after grafting, 1-mL sterile water was added to plates. One week after grafting, successfully grafted plants which were well attached and without adventitious roots were transferred to 1/2 MS plates containing 0.8% (w/v) agar and 0.25% (w/v) sucrose. After recovering for 1 week, well-developed grafted plants were transferred to hydroponic solution. Grafted plants grew in complete hydroponic basal media for 12 days and were cultured for two more weeks after transfer to treatment media. Plants were cultured as described above for hydroponic cultures.

Leaf sectioning and cell size measurements

To visualize leaf flatness, leaves were embedded in 30-g L⁻¹ low-melting agarose (Biozym Scientific, Hessisch Oldendorf, Germany); 40-μm sections were prepared using a vibrating microtome (Hyrax V 50, Carl Zeiss). To quantify leaf cell size and shape, leaf segments were fixed and embedded as described (Eschen-Lippold et al., 2012); semi-thin sections (1 µm) were transferred to glass slides and stained with 1-g L⁻¹ toluidine blue. Images were taken by using an Axioskop 20 microscope (Carl Zeiss) equipped with an AxioCam MRc camera (Carl Zeiss). Morphometric measurements were performed with the iTEM software (Olympus SIS, Münster, Germany).

Monosaccharide and sugar linkage determination

Plants were cultivated in a hydroponic system as described above. AIR was prepared from 18 to 25 mg of lyophilized tissue per sample essentially as previously described (Polko et al., 2018). In brief, the tissue was homogenized with a ball mill, sequentially washed with 70% (v/v) ethanol, chloroform: methanol (1:1 v/v), and acetone, followed by enzymatic de-starching. De-starched AIR polysaccharides were hydrolyzed and analyzed for monosaccharide composition via HPAEC-PAD as previously described (Voiniciuc and Günl, 2016), with the exact instrument and eluent gradient described by Mielke et al. (2021). Partially methylated alditol acetates of glycosidic linkages in 1 mg of the AIR material were derivatized and analyzed by GC–MS exactly as described in a recent publication (Robert et al., 2021).

Callose detection

Callose depositions in leaves were analyzed by aniline blue staining according to Rissel et al. (2017) with slight modifications. Leaves were fixated and destained in 1:3 acetic acid/ethanol, washed in 150-mM K₂HPO₄ for 30 min, and subsequently incubated in Aniline blue solution (0.1 g L⁻¹ Aniline blue [CI 42780, Carl Roth], 150-mM K₂HPO₄) in the dark for 3 h. Stained leaves were embedded in 50% (v/v) glycerol. Callose depositions were documented using a AxioCam MRm Rev. 3 camera mounted on an Axio Observer.Z1 (Carl Zeiss) microscope equipped with filter set 49 (DAPI, Carl Zeiss).

For ultrastructural localization of callose, the material was prepared, treated, and observed as described (Eschen-Lippold et al., 2012).

Reciprocal crossing

Four- to five-week-old plants grown in the greenhouse were used for reciprocal crossing. Anthers were removed from the late unopened flowers, avoiding damage of the female reproductive part. Emasculated plants were cultivated overnight in the greenhouse for further maturation of stigmata. The stigmatic surfaces of emasculated flowers were pollinated with the desired pollen. Pollinated plants were further cultivated in the greenhouse until siliques were matured. Siliques were documented by SLR camera.

Seed germination assay

Sterilized and stratified seeds were sown on 10 layers of sterilized wet Blue Roll paper and cultured in darkness at 22°C. Germinated seeds with emerged radicles were counted every 8 h.

Pollen tube growth assays

For the in vivo assay, 4—to 5-week-old plants grown on soil in the greenhouse were used. Anthers were removed from late unopened flowers, avoiding damage of the female reproduction part. Emasculated plants were cultivated further overnight for maturation of the stigmata. The stigmatic surfaces of emasculated flowers were pollinated with desired pollen. Pistils were harvested 48 h after pollination and destained in 1:3 acetic acid/ethanol until they were transparent. Pistils were further incubated in 8-M NaOH overnight and subsequently stained with Aniline blue solution (0.1-g L⁻¹ Aniline blue [CI 42780, Carl Roth], 150-mM K₂HPO₄) for 5 h in the dark after washing 3 times with 150-mM K₂HPO₄. Pictures were acquired with an LSM 880META with a Plan-apochromatic lens (20x/0.8) (Carl Zeiss) in tile scanning mode.

The semi in vivo pollen tube growth assay was performed as described (Dickinson et al., 2018). One hour after hand-pollination, pollinated pistils were excised above the junction to the ovary and subsequently placed on a gel pad in a humid chamber. Pictures were acquired with a SteREO Discovery V.20 stereomicroscope (Carl Zeiss).

For detection of pectin in pollen tubes, pollen were cultured in vitro for 6 h as described (Ischebeck et al., 2008). Pollen tubes were subsequently stained by 0.1 g L⁻¹ Ruthenium Red (Sigma-Aldrich, St Louis, MO, USA) for 10 min. Pictures were acquired with an AxioCam MRc camera mounted on an Axioskop 20 microscope (Carl Zeiss).

Immunostaining of pollen tubes

Freshly opened flowers from 4- to 5-week-old Arabidopsis plants grown on soil in the greenhouse were harvested and incubated in a humid box at 22°C for 30 min. Pollen was collected into filter-sterilized liquid pollen germination media (5-mM KCl, 1-mM MgSO₄, 5-mM CaCl₂, 0.1-g L⁻¹ H₃BO₃, 100-g L⁻¹ sucrose, pH 7.5) by moving flowers up and down in the medium. Pollen grains were incubated at 30°C for 40 min before further incubation at 22°C for 3 h. For immunostaining, pollen tubes were fixed in freshly prepared 30-g L⁻¹ paraformaldehyde and 100-g L⁻¹ sucrose in PIPES buffer (50-mM PIPES [pH 6.9], 2-mM EGTA, 2-mM MgSO₄) at room temperature for 2 h. Pollen tubes were subsequently washed 2 times in PIPES buffer and 3 times in PBS (pH 7.2). For removing the residual paraformaldehyde, pollen tubes were incubated in 0.1-M NH₄Cl PBS buffer for 5 min and once in PBS. For detection of low-esterified HG, pollen tubes were further incubated overnight in the 1:10-diluted primary antibody solution (JIM5, Plant Probes, University of Leeds, UK) in PBS buffer with 50-g L⁻¹ BSA at 4°C after blocking in PBS buffer with 50-g L⁻¹ BSA for 0.5 h. Pollen tubes were washed 4 times in PBS buffer with 1-g L⁻¹ BSA, followed by 90-min incubation with 1:500 diluted secondary antibody solution (goat anti-rat IgG/Alexa Fluor 488 [A-11006, Thermo Fisher Scientific, Waltham, MA, USA]). Pollen tubes were subsequently washed 3 times with PBS buffer and transferred to microscopy slides. Fluorescence was observed by CLSM using a LSM 880META with a Plan-apochromatic lens (40×/0.95) (Carl Zeiss). Pictures of pollen tubes were taken in channel mode with the same gain and a resolution of 1,024 × 1,024. Excitation light source was a 488-nm multiline ion laser. Alexa Fluor 488 was detected by using a 493–598-nm band pass filter. Images were acquired by using the Zeiss ZEN blue software.

Statistical analyses

Statistical analyses were performed with SigmaPlot version 13.0 (Systat, San Jose, CA, USA). The comparison of two groups was done by Student’s t test. Significance values are defined as ^*P <0.05, ^**P<0.01, and ^***P<0.001. The number of replicates and the repetition of experiments is indicated in the figure legends.

Accession numbers

The Arabidopsis Genome Initiative locus numbers for the genes mentioned in this article are as follows: BICAT3 (AT5G36290); ECA3 (AT1G10130).

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Expression of BICAT3:Venus driven by the native BICAT3 promoter complements the growth defect of the bicat3-1 mutant under Mn²⁺ deficiency.

Supplemental Figure S2. BICAT3 does not co-localize with mitochondria and peroxisomes.

Supplemental Figure S3. GUS staining of ProBICAT3-GUS seedlings grown under different Mn²⁺ and Ca²⁺ supply levels.

Supplemental Figure S4. BICAT3 does not complement Fe³⁺-sensitive (ccc1Δ), Zn²⁺-sensitive (zrc1Δ), Cu²⁺-sensitive (cup2Δ), and Co²⁺-sensitive (cot1Δ) yeast strains.

Supplemental Figure S5. Characterization of bicat3 mutants.

Supplemental Figure S6. bicat3-1 shows retarded growth compared to the wild-type under Mn²⁺ deficiency.

Supplemental Figure S7. bicat3-1 shows retarded root growth compared to the wild-type under Ca²⁺ toxicity.

Supplemental Figure S8. bicat3-1 shows improved growth compared to the wild-type under Mn²⁺ toxicity.

Supplemental Figure S9. Growth of Col-0 and bicat3-1 under different Ca²⁺ and Mn²⁺ supply.

Supplemental Figure S10. Fe²⁺, Zn²⁺, and Mg²⁺ concentrations of 5-week-old plants cultivated in hydroponics with 3.5 µM Mn²⁺ (control) or 0.05 µM Mn²⁺ (-Mn).

Supplemental Figure S11. Relative water content of shoots of 5-week-old Col-0, bicat3-1, and bicat3-2 plants cultivated in hydroponics with 3.5 µM Mn²⁺ (control) or 0.05 µM Mn²⁺ (-Mn).

Supplemental Figure S12. Phenotypes of Col-0 and bicat3-1 under different Mn²⁺ supply levels.

Supplemental Figure S13. Phenotypes of reciprocally grafted Col-0 and bicat3-1 plants under control and Mn²⁺ deficiency (0.05 µM Mn²⁺) conditions.

Supplemental Figure S14. The cell size and shape of bicat3-1 leaves changes under Mn²⁺ deficiency (0.05 µM Mn²⁺) compared to the wild-type.

Supplemental Figure S15. bicat3-1 accumulates more callose in leaves compared to the wild-type under Mn²⁺ deficiency (0.05 µM Mn²⁺).

Supplemental Figure S16. bicat3-1 accumulates more H₂O₂ in leaves compared to the wild-type under Mn²⁺ deficiency (0.05 µM Mn²⁺).

Supplemental Figure S17. Cell wall matrix sugar components in bicat3 roots compared to the wild-type under Mn²⁺ deficiency (0.05 µM Mn²⁺) and control conditions.

Supplemental Figure S18. bicat3-1 and bicat3-2 mutants produce shorter siliques compared to the wild-type.

Supplemental Figure S19. The bicat3-1 mutant produces larger seeds and germinates faster than the wild-type.

Supplemental Figure S20. Aniline blue staining of Col-0 and bicat3-1 pollen tubes grown for 48 h in vivo.

Supplemental Figure S21. Semi in vivo pollen tube growth assay of Col-0, bicat3-1, bicat3-2, and bica3-1 complemented by expression of BICAT3 driven by its native promoter.

Supplemental Figure S22. bicat3 mutant pollen tubes grow aberrantly in vitro and show abnormal pectin distribution.

Supplemental Figure S23. BICAT3 and ECA3 distinctly determine growth, Mn²⁺ and Ca²⁺ accumulation, and photosynthesis under Mn²⁺ limitation.

Supplemental Figure S24. Roots of bicat3 mutants show no morphological defects under Mn²⁺ deficiency.

Supplemental Table S1. Glycosidic linkages of Col-0 and bicat3 shoot and root cell wall AIR.

Supplemental Table S2. Primers and constructs used in this study.

Supplemental Table S3. Media used in this study.

 

E.P. and J.H. designed the research. J.H., B.Y., G.H., N.R., and T.P.V. performed research. M.S. and C.V. contributed new analytic tools. J.H., B.Y., G.H., N.R., C.V., and E.P. analyzed data; J.H. and E.P. wrote the article with contributions of all authors.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Edgar Peiter (edgar.peiter@landw.uni-halle.de).