Formins are evolutionarily conserved multi-domain proteins participating in the control of both actin and microtubule dynamics. Angiosperm formins form two evolutionarily distinct families, Class I and Class II, with class-specific domain layouts. The model plant Arabidopsis thaliana has 21 formin-encoding loci, including 10 Class II members. In this study, we analyze the subcellular localization of two A. thaliana Class II formins exhibiting typical domain organization, the so far uncharacterized formin AtFH13 (At5g58160) and its distant homolog AtFH14 (At1g31810), previously reported to bind microtubules. Fluorescent protein-tagged full length formins and their individual domains were transiently expressed in Nicotiana benthamiana leaves under the control of a constitutive promoter and their subcellular localization (including co-localization with cytoskeletal structures and the endoplasmic reticulum) was examined using confocal microscopy. While the two formins exhibit distinct and only partially overlapping localization patterns, they both associate with microtubules via the conserved formin homology 2 (FH2) domain and with the periphery of the endoplasmic reticulum, at least in part via the N-terminal PTEN (Phosphatase and Tensin)-like domain. Surprisingly, FH2 domains of AtFH13 and AtFH14 can form heterodimers in the yeast two-hybrid assay—a first case of potentially biologically relevant formin heterodimerization mediated solely by the FH2 domain.

SH3P2 (At4g34660), an Arabidopsis thaliana SH3 and Bin/amphiphysin/Rvs (BAR) domain-containing protein, was reported to have a specific role in cell plate assembly, unlike its paralogs SH3P1 (At1g31440) and SH3P3 (At4g18060). SH3P family members were also predicted to interact with formins—evolutionarily conserved actin nucleators that participate in microtubule organization and in membrane–cytoskeleton interactions. To trace the origin of functional specialization of plant SH3Ps, we performed phylogenetic analysis of SH3P sequences from selected plant lineages. SH3Ps are present in charophytes, liverworts, mosses, lycophytes, gymnosperms, and angiosperms, but not in volvocal algae, suggesting association of these proteins with phragmoplast-, but not phycoplast-based cell division. Separation of three SH3P clades, represented by SH3P1, SH3P2, and SH3P3 of A. thaliana, appears to be a seed plant synapomorphy. In the yeast two hybrid system, Arabidopsis SH3P3, but not SH3P2, binds the FH1 and FH2 domains of the formin FH5 (At5g54650), known to participate in cytokinesis, while an opposite binding specificity was found for the dynamin homolog DRP1A (At5g42080), confirming earlier findings. This suggests that the cytokinetic role of SH3P2 is not due to its interaction with FH5. Possible determinants of interaction specificity of SH3P2 and SH3P3 were identified bioinformatically.

Formins are a large, evolutionarily conserved family of actin-nucleating proteins with additional roles in regulating microfilament, microtubule, and membrane dynamics. Angiosperm formins, expressed in both sporophytic and gametophytic tissues, can be divided into two subfamilies, Class I and Class II, each often exhibiting characteristic domain organization. Gametophytically expressed Class I formins have been documented to mediate plasma membrane-based actin assembly in pollen grains and pollen tubes, contributing to proper pollen germination and pollen tube tip growth, and a rice Class II formin, FH5/RMD, has been proposed to act as a positive regulator of pollen tube growth based on mutant phenotype and overexpression data. Here we report functional characterization of the Arabidopsis thaliana pollen-expressed typical Class II formin FH13 (At5g58160). Consistent with published transcriptome data, live-cell imaging in transgenic plants expressing fluorescent protein-tagged FH13 under the control of the FH13 promoter revealed expression in pollen and pollen tubes with non-homogeneous signal distribution in pollen tube cytoplasm, suggesting that this formin functions in the male gametophyte. Surprisingly, fh13 loss of function mutations do not affect plant fertility but result in stimulation of in vitro pollen tube growth, while tagged FH13 overexpression inhibits pollen tube elongation. Pollen tubes of mutants expressing a fluorescent actin marker exhibited possible minor alterations of actin organization. Our results thus indicate that FH13 controls or limits pollen tube growth, or, more generally, that typical Class II formins should be understood as modulators of pollen tube elongation rather than merely components of the molecular apparatus executing tip growth.

Mildew resistance locus o (MLO) proteins are heptahelical integral membrane proteins of which some isoforms act as susceptibility factors for the fungal powdery mildew pathogen. In many angiosperm plant species, loss-of-function mlo mutants confer durable broad-spectrum resistance against the powdery mildew disease. Barley Mlo is known to interact via a cytosolic carboxyl-terminal domain with the intracellular calcium sensor calmodulin (CAM) in a calcium-dependent manner. Site-directed mutagenesis has revealed key amino acid residues in the barley Mlo calcium-binding domain (CAMBD) that, when mutated, affect the MLO-CAM association. We here tested the respective interaction between Arabidopsis thaliana MLO2 and CAM2 using seven different types of in vitro and in vivo protein-protein interaction assays. In each assay, we deployed a wild-type version of either the MLO2 carboxyl terminus (MLO 2CT), harboring the CAMBD, or the MLO2 full-length protein and corresponding mutant variants in which two key residues within the CAMBD were substituted by non-functional amino acids. We focused in particular on the substitution of two hydrophobic amino acids (LW/RR mutant) and found in most protein-protein interaction experiments reduced binding of CAM2 to the corresponding MLO2/MLO2 CT LW/RR mutant variants in comparison to the respective wild-type versions. However, the Ura3-based yeast split-ubiquitin system and in planta bimolecular fluorescence complementation (BiFC) assays failed to indicate reduced CAM2 binding to the mutated CAMBD. Our data shed further light on the interaction of MLO and CAM proteins and provide a comprehensive comparative assessment of different types of protein-protein interaction assays with wild-type and mutant versions of an integral membrane protein.