Differential Roles of Arabidopsis Dynamin-Related Proteins DRP3A, DRP3B, and DRP5B in Organelle Division


Tel: +1 517 432 4620; Fax: +1 517 353 9168; E-mail: huji@msu.edu


Dynamin-related proteins (DRPs) are key components of the organelle division machineries, functioning as molecular scissors during the fission process. In Arabidopsis, DRP3A and DRP3B are shared by peroxisomal and mitochondrial division, whereas the structurally-distinct DRP5B (ARC5) protein is involved in the division of chloroplasts and peroxisomes. Here, we further investigated the roles of DRP3A, DRP3B, and DRP5B in organelle division and plant development. Despite DRP5B's lack of stable association with mitochondria, drp5B mutants show defects in mitochondrial division. The drp3A-2 drp3B-2 drp5B-2 triple mutant exhibits enhanced mitochondrial division phenotypes over drp3A-2 drp3B-2, but its peroxisomal morphology and plant growth phenotypes resemble those of the double mutant. We further demonstrated that DRP3A and DRP3B form a supercomplex in vivo, in which DRP3A is the major component, yet DRP5B is not a constituent of this complex. We thus conclude that DRP5B participates in the division of three types of organelles in Arabidopsis, acting independently of the DRP3 complex. Our findings will help elucidate the precise composition of the DRP3 complex at organelle division sites, and will be instrumental to studies aimed at understanding how the same protein mediates the morphogenesis of distinct organelles that are linked by metabolism.


Eukaryotes are defined by the presence of membrane-bounded subcellular organelles, which house specific biochemical reactions. In addition to performing distinct roles, various organelles can also function together in some intracellular metabolic pathways. In plants for example, peroxisomes, mitochondria, and chloroplasts act in concert in photorespiration (Peterhansel et al. 2010), lipid bodies, peroxisomes, mitochondria, and the cytosol together mediate the consecutive biochemical steps in fatty acid metabolism (Baker et al. 2006; Penfield et al. 2006), and jasmonic acid biosynthesis spans across chloroplasts and peroxisomes (Kazan and Manners 2008). Given the crucial roles of organelles in maintaining basic cellular functions, organelle population and distribution need to be tightly controlled. In plants, peroxisomes, mitochondria, and chloroplasts can divide by binary fission from pre-existing organelles, and one of the major division factors is the evolutionarily conserved dynamin-related protein, DRP (Yang et al. 2008; Kaur and Hu 2009; Logan 2010).

DRP belongs to the dynamin superfamily, which plays critical roles in diverse cellular processes such as vesicle scission, organelle fusion and fission, and cytokinesis (Praefcke and McMahon 2004). Classic dynamin proteins contain five conserved domains. The GTPase domain hydrolyzes guanosine triphosphate (GTP), the middle domain (MD) is responsible for self-interaction of dynamins through its coiled-coil region, the GTPase effector domain (GED) is involved in protein complex formation and activation of GTPase activity, the pleckstrin-homology domain (PHD) binds to negatively charged lipids, and the prolin-rich domain (PRD) provides the binding site for dynamin-binding proteins (Hinshaw 2000; Heymann and Hinshaw 2009). DRPs are defined by having at least the first three characteristic domains (Heymann and Hinshaw 2009); however, there are a few exceptions, in which the DRPs only contain the GTPase domain (Hong et al. 2003). Similar to dynamins, Dnm1p, a yeast DRP involved in peroxisomal and mitochondrial division, has GTPase activity and can self-assemble to form a spiral-like structure whose size matches the mitochondrial constriction sites (Ingerman et al. 2005). The assembled DRP polymers have been suggested to act as mechanochemical enzymes or signaling GTPases in a GTP hydrolysis-dependent manner (Praefcke and McMahon 2004). Recently, the three-dimensional (3D) structure of Dnm1p was solved using cryo-electron microscopy. In the presence of GTP, the Dnm1p spiral-like structure undergoes a large conformational change that presumably initiates constriction and fission of the mitochondrial membrane (Mears et al. 2011). The high-resolution crystal structure of human Dynamin1 (Drp1), a shared protein in mitochondrial and peroxisomal division, was also solved in two independent studies, which demonstrated that the higher-order formation of Drp1 is achieved through criss-cross assembly of the stalks (Faelber et al. 2011; Ford et al. 2011).

The Arabidopsis DRP family contains 16 members that can be divided into six subgroups (DRP1–6) based on their structural and sequence similarities and molecular activities (Hong et al. 2003). Among them, DRP1A and DRP1C have been suggested to function in clathrin-mediated membrane endocytosis (Collings et al. 2008; Konopka and Bednarek 2008; Fujimoto et al. 2010). DRP1C and DRP1E have also been implicated in mitochondrial morphogenesis (Jin et al. 2003). Like the DRP1s, DRP2A and DRP2B play a role in clathrin-mediated endocytosis (Bednarek and Backues 2010; Fujimoto et al. 2010; Taylor 2011). The DRP3 subfamily has two members, DRP3A and DRP3B, which are grouped in the same phylogenetic subclade as Drp1 and Dmn1p, peroxisomal and mitochondrial division factors from mammals and yeast (Saccharomyces cerevisiae) (Arimura and Tsutsumi 2002; Hong et al. 2003; Miyagishima et al. 2008). Consistent with this phylogenetic analysis, DRP3A and DRP3B are dual-localized and mediate the division of both peroxisomes and mitochondria (Arimura et al. 2004; Mano et al. 2004; Aung and Hu 2009; Fujimoto et al. 2009; Zhang and Hu 2009). Phosphorylation of DRP3A and DRP3B can also promote mitochondrial fission during mitosis (Wang et al. 2012). The DRP5 subfamily is composed of DRP5A and DRP5B. DRP5A is involved in cytokinesis (Miyagishima et al. 2008). DRP5B (ARC5), however, associates with chloroplasts and peroxisomes and mediates their division. Lack of a functional DRP5B (ARC5) leads to enlarged and dumbbell-shaped chloroplasts and highly aggregated peroxisomes that are impaired in fission (Gao et al. 2003; Zhang and Hu 2010).

DRP3A and DRP3B are partially redundant in the division of peroxisomes and mitochondria. However, despite having a small stature and mitochondria and peroxisomes that are grossly distorted morphologically, the drp3A drp3B double mutant is still fertile and contains both organelles in all cell types (Fujimoto et al. 2009; Zhang and Hu 2009). These data suggest that additional DRPs (such as DRP5B) or a DRP-independent machinery may function to maintain a low level of peroxisomal and mitochondrial division in the drp3 double mutant. In addition, although DRP3A, DRP3B, and DRP5B are all involved in peroxisomal division, it is unclear whether DRP5B's role is partially redundant to that of the DRP3s in this process, and whether DRP5B can form complexes with DRP3s in vivo. To address these questions, we generated a drp3A drp3B drp5B triple mutant, analyzed the impact of DRP5B on mitochondrial division, and tested the interaction among the three DRP proteins. Surprisingly, DRP5B affects mitochondrial division/morphogenesis in spite of lacking a stable physical association with this organelle. Furthermore, a DRP3-containing protein complex was detected in vivo in which DRP3A appears to be a major component. DRP5B does not directly interact with DRP3 in vitro, and is not a major component of the DRP3 protein complex. Our data have revealed the unexpected role of DRP5B in mitochondrial morphogenesis/division and DRP5B's DRP3-independent mode of action in organelle division.


DRP5B is involved in mitochondrial division despite its lack of stable association with mitochondria

DRP3A and DRP3B belong to a subclade of DRPs that are conserved across eukaryotic species (Miyagishima et al. 2008). They each contain three functional domains: the GTPase domain, the MD, and the GED (Figure 1A). DRP5B, on the other hand, is a plant and algal-specific protein that is composed of four functional domains: the GTPase domain, the MD, the GED, and the PHD (Figure 1A) (Hong et al. 2003). To test the hypothesis that DRP5B is another division factor in the same division machinery as DRP3A and DRP3B, we generated the drp3A-2 drp3B-2 drp5B-2 triple knockout mutant (Figure 1B) by crossing drp3A-2 drp3B-2 (Zhang and Hu 2009) with drp5B-2 (Zhang and Hu 2010), both of which are in the Col-0 background. Triple mutants were identified by PCR of genomic DNA in the F2 generation (Figure 2A). Consistent with previous reports (Fujimoto et al. 2009; Zhang and Hu 2009, 2010), the drp3A-2 drp3B-2 double mutant displayed severe dwarfness, whereas the drp5B-2 single mutant was slightly smaller than the wild type (Figure 2B, C). The drp3A-2 drp3B-2 drp5B-2 triple mutant resembled the drp3A drp3B double mutant in terms of growth (Figure 2B, C).

Figure 1.

Domain organization and gene structure of the dynamin-related proteins in this study. 

(A) Domain organization of DRP3A, DRP3B, and DRP5B. The GTPase, middle (MD), pleckstrin-homology (PHD), and the GTPase effector (GED) domains are indicated. α-DRP3A and α-DRP3B show positions of the peptides used to raise antibodies in this study. 
(B) Schematic diagrams of the DRP3A, DRP3B, and DRP5B genes. Black boxes are exons within the coding region, 5′ and 3′ UTRs are represented by gray boxes, black triangles indicate positions of the T-DNA insertions, and the asterisk indicates the point mutation in drp5B-1 (arc5).

Figure 2.

Growth phenotypes of the drp mutants. 

(A) Genotyping of the drp mutants by PCR of genomic DNA. The upper band (G) was amplified using gene-specific primers, and the lower band (T) is a T-DNA insertion-specific product amplified using gene-specific primers together with a T-DNA-specific primer (see Materials and Methods). 
(B) Images of four-w-old plants taken under the same magnification. 
(C) Eight-w-old plants.

To compare organelle morphologies in the mutants, a Col-0 line co-expressing the peroxisomal marker Cyan Fluorescent Protein–Peroxisome Targeting Signal type 1 (CFP-PTS1/SKL) and the mitochondrial marker (Saccharomyces cerevisiae Cytochrome C Oxidase IV-Yellow Fluorescent Protein, or ScCOX4-YFP) (Nelson et al. 2007), which was utilized in our previous study (Aung and Hu 2011), was crossed with the drp single, double, and triple mutants. The respective homozygous mutants co-expressing the two organelle markers were identified in the F2 and F3 generations by PCR genotyping (for homozygosity of the mutations) and confocal fluorescent microscopy (for the presence of the organelle markers). Confocal imaging of the mutants showed various degrees of impaired mitochondrial and peroxisomal division in the drp3 single and double mutants, phenotypes that are consistent with previous reports (Figure 3A, B, C and E) (Mano et al. 2004; Aung and Hu 2009; Fujimoto et al. 2009; Zhang and Hu 2009). Surprisingly, besides having enlarged and dumbbell-shaped chloroplasts and clustered peroxisomes, as shown in previous studies (Gao et al. 2003; Zhang and Hu 2010), mitochondria in drp5B-2 also exhibited extended tubular structures and membrane extensions (Figure 3D) instead of the rod-shaped morphology seen in the wild type (Figure 3A). This mitochondrial phenotype in drp5B is to some extent similar to that of drp3A-2 and drp3B-2, indicating that like the DRP3s, DRP5B also plays a role in mitochondrial division, possibly in the final fission step. In the drp3A-2 drp3B-2 drp5B-2 triple mutant, chloroplast division was impaired just like in the drp5B single mutant. The peroxisomal phenotype also resembled that of drp3A-2 drp3B-2, whereas mitochondria displayed enhanced elongation and aggregation compared with the double mutant (Figure 3F). These results reveal an unexpected role of DRP5B in mitochondrial division and morphogenesis, and this function appears to be partially redundant with that of the DRP3 proteins. However, the role of DRP3s in peroxisome division seems to be predominant compared with that of DRP5B.

Figure 3.

Organelle morphologies of the drp mutants. 

Confocal images were taken from leaf epidermal (for mitochondria and peroxisomes) or mesophyll (for chloroplasts) cells of four-w-old plants in various genetic backgrounds as indicated. Mitochondria were labeled by ScCOX4-YFP, chloroplast signals were generated by autofluorescence of the chlorophyll, and peroxisomes were marked by CFP-PTS1. Scale bars = 10 μm.

Previous fluorescence microscopic studies showed that GFP-DRP5B localizes discontinuously to a ring structure at the division sites of bilobed chloroplasts, and evenly on peroxisomes (Gao et al. 2003; Zhang and Hu 2010). Because our genetic analysis suggested an additional role for DRP5B in mitochondrial division, we tested whether DRP5B is physically associated with mitochondria. To this end, we tagged YFP to the N-terminal end of DRP5B to generate YFP-DRP5B, which was then stably expressed under the control of the 35S constitutive promoter in Col-0 plants that are already expressing the mitochondrial marker ScCOX4-CFP. T2 transgenic plants co-expressing the CFP and YFP fusion proteins were selected and subjected to confocal imaging for subcellular localization analysis. YFP-DRP5B signals were seen at the constriction sites of chloroplasts, but none were found on mitochondria (Figure 4) in any of the transgenic plants analyzed, suggesting that, unlike DRP3A and DRP3B, DRP5B is not stably associated with mitochondria.

Figure 4.

Subcellular localization of YFP-DRP5B in Arabidopsis thaliana. 

Confocal images were taken from leaf mesophyll cells of transgenic Col-0 plants co-expressing 35Spro:YFP-DRP5B and the mitochondrial marker ScCOX4-CFP. YFP signals are in green, chloroplast signals are in red, and CFP signals are in magenta. Scale bars = 5 μm.

The lack of a physical association between DRP5B and mitochondria led us to speculate that the mitochondrial phenotype in the drp5B mutants might be caused by the enlarged chloroplasts, which may drastically disrupt the intracellular architecture, including mitochondrial morphology. To investigate this possibility, we introduced the mitochondrial marker ScCOX4-YFP into another null drp5B allele, drp5B-1 or arc5 (Figure 1B) (Gao et al. 2003) and wild type Ler, from which arc5 was generated, by floral dipping. ScCOX4-YFP was also introduced by genetic cross into two other plastid division mutants, arc6 (Robertson et al. 1995; Vitha et al. 2003) and the ftsZ2 double knockout (Schmitz et al. 2009), both of which are in the Col-0 background. Confocal micrographs taken from the mutants expressing the mitochondrial marker confirmed the mitochondrial elongation phenotype in drp5B-1 (Figure 5A, B). Quantification analysis showed approximately 35% (drp5B-1 vs. Ler) and 38% (drp5B-2 vs. Col-0) reduction of the total number of mitochondria from the respective wild type parents (Figure S1). Although arc6 and the ftsZ double mutants each contain a few enormous chloroplasts in their cells, they possess wild type-looking mitochondria (Figure 5C, D, E), suggesting that the massive changes in chloroplast morphology do not have an obvious influence on mitochondrial division. Therefore, the mitochondrial phenotype in drp5B resulted from the lack of a functional DRP5B to control mitochondrial morphology, and is not a secondary effect of impaired chloroplast division.

Figure 5.

Mitochondrial and chloroplast morphologies in chloroplast division mutants. 

Confocal images were taken from leaf epidermal (for mitochondria) or mesophyll (for chloroplasts) cells of the indicated plants expressing the mitochondrial marker ScCOX4-YFP. Chloroplast signals were generated by autofluorescence of the chlorophyll. Scale bars = 10 μm.

Complex formation between DRP3A, DRP3B, and DRP5B

During organelle fission, DRP forms a collar-like structure around the mitochondrial division site through the assembly of DRP multimeric complexes (Ingerman et al. 2005). Consistent with this notion, DRP3A and DRP3B have been shown to form homo- or heterodimers in vitro or in vivo (Arimura et al. 2008; Fujimoto et al. 2009; Zhang and Hu 2010), suggesting that the two proteins might function cooperatively in the division of peroxisomes and mitochondria. To test whether DRP5B functions together with DRP3A and DRP3B, we employed the yeast two-hybrid (Y2H) approach to examine the physical interaction between the DRPs. Consistent with previous Y2H studies (Arimura et al. 2008; Fujimoto et al. 2009), DRP3A and DRP3B can form homo- and hetero-dimers; however, neither self-interaction of DRP5B nor interaction between DRP5B and DRP3s were detected in these assays (Figure 6), suggesting that DRP5B may not function in the same complex as the DRP3 proteins, and may not even form homooligomers.

Figure 6.

Yeast two-hybrid analysis to test interaction between the DRP5B and DRP3 proteins. 

SD/Glucose (-UTH) selects for transformants, and SD/Galactose (-UTHL) plus X-gal select for protein-protein interactions. Empty, the pB42AD-GW vector; BD, GAL4 DNA-binding domain; AD, GAL4 activation domain.

The interaction between the DRP3 proteins suggests that DRP3A and DRP3B may form supercomplexes in the cell to mediate organelle division. To test this hypothesis and to analyze the expression of endogenous DRP3 proteins in the plant, we generated peptide antibodies against DRP3A and DRP3B (see Methods section). The α-DRP3A antibodies detected the endogenous DRP3A protein in Col-0, drp3B-2, and drp5B-2, but not in drp3A-2 (Figure 7A). Similarly, α-DRP3B detected endogenous DRP3B in Col-0, drp3A-2, and drp5B-2, whereas the corresponding band was missing in drp3B-2 (Figure 7B). These results confirm the specificity of the antibodies in recognizing the cognate endogenous proteins. In addition, the absence of one functional DRP3 protein does not seem to have an obvious effect on the steady-state level of the other DRP3 in Arabidopsis.

Figure 7.

Detection of endogenous DRP3A and DRP3B proteins and complexes in two–w-old Arabidopsis thaliana plants. 

(AB) SDS-PAGE analysis to detect the steady-state level of endogenous DRP3A and DRP3B proteins. The large subunit of Rubisco is shown in the lower panel as a loading control. Arrowheads mark the expected size for DRP3A (89.7 kDa) and DRP3B (89.9 kDa) proteins. 
(CD) Blue-native (BN) PAGE analysis to show the presence of DRP3A and DRP3B supercomplexes. The photosystem I and photosystem II (PSI/PSII) core dimer (green band) was used as the loading control. 
Numbers indicate molecular weight in kDa. The α-DRP3A or α-DRP3B peptide antibodies were used in both analyses.

Blue native polyacrylamide gel electrophoresis (BN-PAGE) was then employed to detect the presence of DRP3 protein complexes in vivo in wild type and drp3A-2, drp3B-2, and drp5B-2 single mutants. Both DRP3A and DRP3B antibodies detected protein complexes as smear signals around 1,000 kDa in the wild-type Col-0 (Figure 7C, D), suggesting that DRP3 can form a higher-order protein complex in vivo. The lack of a major band might be due to the formation of various degrees of DRP3 multimers. The DRP3A antibodies detected the same ∼1,000-kDa protein complex in drp3B-2 and drp5B-2, and a faint signal at the same size in drp3A-2, suggesting that DRP3A is an essential component of the DRP3A supercomplex (Figure 7C). Likewise, DRP3B is also required for the formation of the DRP3B protein complex, as only a weak signal of the DRP3B complex was detected in drp3B-2 (Figure 7D). Interestingly, the level of the DRP3B protein complex was also significantly decreased in drp3A-2, suggesting that DRP3A is a major component of the protein complex that presumably contains both DRP3A and DRP3B. No differences were observed in the level of the DRP3A or DRP3B complex in the drp5B-2 mutant (Figure 7C, D), suggesting that DRP5B is not a constituent of the DRP3 complex.


DRP5B has an additional role in mitochondrial division

DRP5B (ARC5) was initially identified as a chloroplast division protein, which is recruited to a ring-shaped structure at the chloroplast division site (Gao et al. 2003), and was later demonstrated to be targeted to peroxisomes and involved in peroxisome division as well (Zhang and Hu 2010). In this study, we uncover an additional role of DRP5B in mitochondrial division, as the absence of a functional DRP5B results in mitochondrial elongation (Figures 3, 5) and a reduction in mitochondrial abundance (Figure S1). This mitochondrial phenotype is specific to drp5B but not to other chloroplast division mutants such as arc6 and ftsZ2 (Figure 5). Thus, it is clear that DRP5B has a direct role in mitochondrial morphogenesis/division. However, stable association of YFP-DRP5B was not detected on the surface of mitochondria like it is on chloroplasts and peroxisomes, leading to speculation that DRP5B may be very transiently associated with mitochondria during the fission process. Alternatively, DRP5B's association with mitochondria may only occur at a specific stage of the cell cycle. Plant proteins involved in the morphogenesis of three different types of organelles had not been identified prior to this study. Thus, our findings will be instrumental to future studies to investigate mechanisms by which the same protein functions at multiple subcellular compartments.

DRP5B may function independently of DRP3

DRP5B is distantly related to DRP3A and DRP3B, sharing only 15.2% and 15.3% amino acid sequence identity respectively with the two DRP3 proteins. In addition, DRP5B contains a PHD, which has a putative function in lipid binding (Hong et al. 2003). In this study, DRP5B does not interact with DRP3A or DRP3B in Y2H assays. This data contradicts our previous findings, which showed physical interaction between DRP5B and DRP3 in Bimolecular Fluorescence Complementation (BiFC) assays and the ability of these proteins to form complexes in co-immunoprecipitation experiments (Zhang and Hu 2010). This discrepancy may be explained in several ways. First, the association between DRP5B and DRP3 is transient or not strong enough to be detected by the Y2H system. Second, our previous BiFC assays may have detected random interactions between the two halves of the fluorescent protein instead of true interactions between DRP5B and DRP3 – a problem that occurs frequently in these types of assays. Third, other proteins are required to bridge DRP5B and DRP3. Lastly, DRP5B contains the putative lipid binding domain PHD, thus it may not target to the yeast nucleus, which is a requirement for testing protein-protein interaction using the Y2H system. However, DRP3s are localized to spots on the surface of peroxisomes (Mano et al. 2004; Zhang and Hu 2009), whereas DRP5B is evenly distributed on the peroxisome (Zhang and Hu 2010). These differential localization patterns of DRP3 and DRP5B on peroxisomes seem to support the conclusion that DRP5B does not work together with DRP3s on the organelle. Consistent with this conclusion, DRP5B does not have an apparent impact on the formation of the DRP3-containing protein complexes (Figure 7C, D), suggesting that DRP5B functions independently of the DRP3 complex.

DRP3A is a major component of the organelle division machinery

Based on studies of the yeast and mammalian DRP proteins, Arabidopsis DRP3 was suggested to form a higher order protein complex during organelle division (Hu 2009; Kaur and Hu 2009), yet biochemical evidence of this had not previously been shown. Here, we provide evidence for the existence of native DRP3A- and DRP3B-containing high-molecular-mass protein complexes in Arabidopsis. DRP3A and DRP3B are present in the same protein complex in which DRP3A is a major component (Figure 7C, D). In line with this finding, genetic screens for peroxisome division mutants repeatedly identified DRP3A, but not DRP3B or DRP5B (Mano et al. 2004; Aung and Hu 2009; Zhang and Hu 2009). Furthermore, the peroxisomal phenotype is much stronger in drp3A than in drp3B (Fujimoto et al. 2009; Zhang and Hu 2009), and DRP3A was able to rescue the peroxisomal phenotype in the drp3B mutant, yet DRP3B could not restore the peroxisomal division defects in drp3A (Fujimoto et al. 2009). To understand how exactly the DRP3 and DRP5B proteins mediate organelle division, it will be necessary to elucidate the precise configuration of the DRP3 supercomplex at organelle division sites.

Additional DRPs or DRP-independent machinery might be involved in the division of organelles

The drp3A drp3B drp5B triple knockout mutants generated in this study are vital and contain mitochondria, peroxisomes, and chloroplasts in all cell types, suggesting that the organelles still undergo division and are able to be inherited to the daughter cells. This implies that other DRP members may partially compensate for the loss of DRP5B and DRP3. Supporting this prediction is the finding that DRP1C and DRP1E are involved in the division or morphogenesis of mitochondria (Jin et al. 2003). The relationship between the DRP1, DRP3, and DRP5B proteins needs to be established. Comprehensive analyses of organelle morphologies in all the drp single and higher-order mutants, and protein-protein interaction between the DRP proteins will provide a more complete picture of the contribution of Arabidopsis DRP proteins to organelle fission. Alternatively, there might be a DRP-independent machinery that regulates organelle division. Consistent with this hypothesis, we recently reported the role for the Arabidopsis peroxisomal and mitochondrial division factor1 (PMD1) in mediating the division of peroxisomes and mitochondria in a DRP3-independent manner (Aung and Hu 2011).

During the last decade, we have witnessed the discovery of organelle division factors across diverse organisms. Among the most striking findings in this field is the sharing of components in the division machineries of distinct organelles such as mitochondria, peroxisomes and chloroplasts, although these organelles have very distinct origins and ultrastructures (Hu 2007, 2009; Kaur and Hu 2009; Hu et al. 2012). The use of shared division factors might be critical to maintaining cellular homeostasis, as organelles within eukaryotic cells are highly interactive and are often connected by metabolic pathways.

Materials and Methods

Plant materials, growth conditions, and identification of T-DNA insertion plants

Arabidopsis thaliana plants were grown in growth chambers with 16 h light (∼70-80 μmol/m2 s2 white light), at 20 °C and 70% humidity. T-DNA insertion mutants drp3A-2 (SALK_147485), drp3B-2 (SALK_112233), and drp5B-2 (SAIL_71D_11) were characterized previously (Zhang and Hu 2009, 2010). drp5B-1 (arc5), arc6 and the ftsZ2 double mutant were provided by Kathy Osteryoung (MSU). drp5B-1 is in the Ler background, while all other T-DNA insertion mutants are in the Col-0 background. Presence of the T-DNA and the homozygosity of the mutants were verified by PCR using genomic DNA as templates. LBb1.3 (5′-ATTTTGCCGATTTCGGAAC-3′) together with gene-specific primers were used to genotype the SALK mutants. SALK_147485-LP (5′- AACCACAGGTTTACCTCCTGG-3′) and SALK_147485-RP (5′- ACGCCTCCTTCTTCTTCTACG-3′) were used for drp3A-2; and SALK_112233-LP (5′-TAAAATGGCCTTCAGGAAAGG-3′) and SALK_112233-RP (5′-TGAGGAGAGAAATAGCACCTTTG-3′) were used for drp3B-2. To genotype drp5B-2, the following primers were used: SAIL_71D_11-LP (5′- TGTGTTGGATGCCCTTAAGAC-3′), SAIL_71D_11-RP (5′-TGTCACCTGATGAAGGAAAGG-3′), and SAIL_LB3 (5′-GCATCTGAATTTCATAACCAATC-3′).

Plant transformation

The floral dipping method was used to introduce the mitochondrial marker ScCOX4-YFP into Ler and drp5B-1. The same method was also used to introduce 35Spro:YFP-DRP5B into Col-0, which already contained ScCOX4-CFP. To select for 35Spro:YFP-DRP5B, T1 seeds were grown on soil and sprayed with 0.1% (v/v) of Basta (Finale; Farnam Companies) and 0.025% (v/v) Silwet L-77 at 7 and 9 d after germination. The transgenic plants were further confirmed for the expression of Yellow Fluorescent Protein (YFP)-fusion proteins using epifluorescence microscopy.

Gene cloning

The full-length coding regions of DRP3A, DRP3B, and DRP5B were amplified from cDNA generated from total mRNA of Arabidopsis Col-0 seedlings using the following primers: DRP3A-attB1 (5′-ggggacaagtttgtacaaaaaagcaggcttcatgactattgaagaagtttccg-3′), DRP3A-attB2 (5′-ggggaccactttgtacaagaaagctgggtcttagaatccgtatccattttggtg-3′), DRP3B-attB1 (5′-ggggacaagtttgtacaaaaaagcaggcttcatgtccgtcgacgatctccc-3′), DRP3B-attB2 (5′-ggggaccactttgtacaagaaagctgggtcttacatatgaagccgtccgttc-3′), DRP5B-attB1 (5′-ggggacaagtttgtacaaaaaagcaggcttcatggcggaagtatcagc-3′) and DRP5B-attB2 (5′-ggggaccactttgtacaagaaagctgggtgtcaatgctgcaccgaagg-3′). The amplified product was cloned using the Gateway® system (Invitrogen, Carlsbad, CA, USA). pEarleyGate101 (CD3-683) (Earley et al. 2006) was used for constructing 35Spro:YFP-DRP5B, and pGilda-GW and pB42AD-GW were used to generate bait and pray fusion proteins for yeast two-hybrid assays using the Matchmaker system (Clontech, USA).

Microscopic analysis and mitochondrial quantification

Confocal images were taken with a confocal laser scanning microscope (Zeiss LSM 510 META), using the 63x oil immersion objective. Freshly excised Arabidopsis leaf discs were mounted with water for imaging. YFP and chlorophyll fluorescence were excited with a 514 nm argon laser line, and the signals were collected using BP530-600 nm and BP630-700 nm respectively. CFP signals were excited using a 458 nm argon laser line and detected by BP465-510 nm. For co-localization analysis, CFP, YFP, and chlorophyll fluorescence were excited and detected simultaneously using the same settings described above.

Mitochondrial numbers were quantified using ImageJ software (ImageJ 1997) as previously described (Desai and Hu 2008; Aung and Hu 2011). P values were calculated using the Student's two-tailed t-test against the wild type.

Yeast two-hybrid assays

The Matchmaker two-hybrid system (Clontech, USA) was used to test the homo- and heteromeric interactions between DRP5B and DRP3. A Frozen-EZ Yeast Transformation Kit (Zymo, Irvine, CA, USA) was used to simultaneously co-transform bait and pray plasmid DNAs into the yeast strain EGY48. The desired transformants were selected using standard synthetic dropout medium (SD/Glc-Ura-Trp-His). Physical interaction between the tested proteins was examined on SD/Gal-Ura-Trp-His-Leu agar plates containing X-Gal.

SDS-PAGE analysis

50 mg fresh weight of two-w-old Arabidopsis seedlings from each genotype was ground with a plastic pestle using liquid nitrogen with the addition of 500 μL of SDS-containing extraction buffer, which contains 60 mM Tris-HCL pH 8.8, 2% SDS, 2.5% glycerol, 0.13 mM EDTA pH 8.0, and 1X Complete Protease Inhibitor Cocktail (Roche, USA). The samples were vortexed for 30 s and heated at 70 °C for 10 min, followed by centrifugation at 13,000 g twice (5 min/each time) at room temperature. The supernatants (total proteins) were then transferred to new tubes. 5 μL of the total proteins were mixed with the same volume of 2x NuPAGE LDS sample buffer (Invitrogen, Carlsbad, CA, USA), separated on a 4%–12% NuPage gel (Invitrogen, Carlsbad, CA, USA), and blotted to a polyvinylidene difluoride (PVDF) membrane.

Blue-native polyacrylamide gel electrophoresis (BN-PAGE)

50 mg fresh weight of two-w-old Arabidopsis seedlings from each genotype was homogenized with a pestle and liquid nitrogen. Total native proteins were isolated using a NativePAGE™ Sample Prep Kit (Invitrogen, Carlsbad, CA, USA). 100 μL of 1x loading buffer containing 2% detergent (DDM) was added to homogenized samples, followed by incubation of the mixtures on ice for 30 min. The incubated samples were centrifuged at 17,000 g two times (15 min/each time) at 4 °C, and the supernatants were transferred to new tubes. 2 μL of 5% G-250 sample additive was added to 20 μL of the sample solution, which was then separated on 4%–12% NativePAGE™ using 1x NativePAGE™ Running Buffer and Cathode Buffer Additive. The gel was incubated with 2x NuPAGE™ transfer buffer for 10 min before being transferred to the PVDF membrane.

Antibody generation and immunoblot analysis

Antibodies against DRP3A (α-DRP3A: 489-RKRMDEVIGDFLREGLEP-506) and DRP3B (α-DRP3B: 535-HPVARPDTVEPER-548) were designed and synthesized by Open Biosystems, Inc. To clean up the antibodies, the serum was diluted to 1:400 in the blocking buffer and incubated with the PVDF membrane that contained total proteins transferred from drp3A-2 or drp3B-2 at 4 °C overnight. The treated antibodies were used in the immunoblot analysis.

The PVDF membrane from SDS-PAGE or BN-PAGE was blocked with 3% BSA in 1x TBST (50 mM Tris-base, 150 mM NaCl, 0.05% Tween 20, pH 8.0) overnight at 4 °C, before being probed with the antibodies prepared in the blocking buffer. The treated antibodies were hybridized to the PVDF membrane containing transferred proteins of interest for 1 hr at room temperature. The hybridized membrane was then rinsed three times (10 min/each time) with 1x TBST before being probed with a secondary antibody (1:20,000 goat anti-rabbit IgG) and HRP-conjugate (Millipore, USA) in 1x TBST. Signals were detected with 4x-diluted SuperSignal® West Dura Extended Duration Substrate (Pierce Biotechnology, Rockford, IL, USA). The membrane was exposed to a film to visualize the signals.

(Co-Editor: Chris Hawes)


We thank Kathy Osteryoung (Michigan State University) for seeds of arc5 (drp5B-1), arc6, and the ftsZ2 double mutant. This work was supported by grants from the National Science Foundation Arabidopsis 2010 program (MCB 0618335) and the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy (DE-FG02-91ER20021) to J.H.