RAB5 is a small GTPase that acts in endosomal trafficking. In addition to canonical RAB5 members that are homologous to animal RAB5, land plants harbor a plant-specific RAB5, the ARA6 group, which regulates trafficking events distinct from canonical RAB5 GTPases. Here, we report that plant RAB5, both canonical and plant-specific members, accumulate at the interface between host plants and biotrophic fungal and oomycete pathogens. Biotrophic fungi and oomycetes colonize living plant tissues by establishing specialized infection hyphae, the haustorium, within host plant cells. We found that Arabidopsis thaliana ARA6/RABF1, a plant-specific RAB5, is localized to the specialized membrane that surrounds the haustorium, the extrahaustorial membrane (EHM), formed by the A. thaliana -adapted powdery mildew fungus Golovinomyces orontii . Whereas the conventional RAB5 ARA7/RABF2b was also localized to the EHM, endosomal SNARE (soluble N -ethylmaleimide-sensitive factor attachment protein receptor) and RAB5-activating proteins were not, which suggests that the EHM has modified endosomal characteristic. The recruitment of host RAB5 to the EHM was a property shared by the barley-adapted powdery mildew fungus Blumeria graminis f.sp. hordei and the oomycete Hyaloperonospora arabidopsidis , but the extrahyphal membrane surrounding the hypha of the hemibiotrophic fungus Colletotrichum higginsianum at the biotrophic stage was devoid of RAB5. The localization of RAB5 to the EHM appears to correlate with the functionality of the haustorium. Our discovery sheds light on a novel relationship between plant RAB5 and obligate biotrophic pathogens.
Filamentous pathogens such as fungi causing powdery mildew and rust, and oomycetes causing downy mildew establish parasitic relationships with living host plant cells through the formation of an elaborated feeding organ, the haustorium. The presumed function of the haustorium is delivering effector proteins into host cells to suppress the immunity of the plant and taking up host cellular nutrients through the extrahaustorial matrix (EHMx) and the extrahaustorial membrane (EHM) ( Yi and Valent 2013 ). In electron micrographs, the EHMx is composed of fine granular materials and often contains membrane-bounded vesicles ( Micali et al. 2011 ). The structural singularity of the EHM has been well documented for the haustoria formed by powdery mildew fungi, whose EHM exhibits a thick and convoluted appearance, which is distinct from the thin, smooth host plasma membrane (PM) ( Gil and Gay 1977 , Celio et al. 2004 , Micali et al. 2011 ). An electron-dense haustorial neckband is observed at the junction of the host PM and the EHM, probably sealing the intermembrane space to retain the EHMx in the haustorial complex ( Bushnell 1972 , Gil and Gay 1977 ). The properties and molecular composition of the EHM are distinct from those of the PM. The EHM withstands treatments with chemicals such as detergents and acids, which normally dissolve the PM ( Gil and Gay 1977 ), and all examined PM proteins are absent from the EHM ( Koh et al. 2005 , Micali et al. 2011 ). Despite the importance of the EHM as a pathogenic interface and a potential target of effective fungicides, its molecular constituents remain obscure.
The physiological significance of membrane trafficking in plant–microbe interactions has recently been recognized ( Inada and Ueda 2014 ). A single round of membrane trafficking consists of several sequential steps, including cargo sorting at, and membrane vesicle budding from, the donor organelle, transport of the membrane vesicle to its destination and tethering of the membrane vesicle to the target membrane, followed by the fusion of these membranes ( Inada and Ueda 2014 ). RAB GTPase is one of the key factors that regulates the final step, the tethering and fusion of the transport vesicles. Individual trafficking pathways involve specific RAB GTPases, and RAB5 is a key regulator of a wide spectrum of endocytic events in eukaryotic organisms, including plants ( Ebine and Ueda 2009 ). In addition to the canonical RAB5, which is anchored to the membrane via a geranylgeranyl moiety attached to the C-terminal cysteine motif, plants harbor another type of RAB5 termed the ARA6 group, the members of which are anchored to membranes through dual fatty acylations at their N-termini ( Ueda et al. 2001 ). A clear homolog of ARA6 has not yet been identified in non-plant systems, suggesting that plants acquired this specific subtype of RAB5 during evolution. The model plant Arabidopsis thaliana (hereafter referred to as Arabidopsis) harbors three RAB5-related proteins, two canonical RAB5s (RHA1/RABF2a, At5g45130; and ARA7/RABF2b, At4g19640) and one plant-unique RAB5 (ARA6/RABF1, At3g54840), which are localized to distinct populations of multivesicular endosomes (MVEs) with considerable overlap ( Ueda et al. 2004 , Haas et al. 2007 , Ebine et al. 2011 ). RHA1 and ARA7 redundantly regulate the endocytic/vacuolar trafficking pathways ( Sohn et al. 2003 , Kotzer et al. 2004 , Ebine et al. 2011 ), whereas ARA6 acts in the trafficking pathway from the endosomes to the PM ( Ebine et al. 2011 ) and may play a role in vacuolar trafficking ( Bottanelli et al. 2011 ). ARA6 function is required for normal abiotic stress responses ( Ebine et al. 2011 ) and starch homeostasis ( Tsutsui et al. 2015 ), suggesting that the ARA6-mediated trafficking pathway was exploited to maximize plant fitness during evolution.
The involvement of host MVEs at the early stages of the plant–fungus interaction has been previously suggested. MVEs and ARA6-positive compartments accumulate near cell wall appositions (papillae) that form at the attempted invasion site of a non-adapted powdery mildew fungus ( An et al. 2006 , Nielsen et al. 2012 ). The papillae-targeted secretion of PEN1/SYP121, which is a soluble N -ethylmaleimide-sensitive factor attachment protein receptor (SNARE) functioning in membrane fusion and resistance against non-adapted fungal challenge, is also mediated by MVEs ( Collins et al. 2003 , Meyer et al. 2009 , Nielsen et al. 2012 ). However, its role in haustorium formation and the establishment of parasitic interactions has been less well studied.
In the present study, we report that RAB5 GTPases in Arabidopsis, both plant-specific ARA6 and conventional ARA7, are recruited to the EHM of obligate biotrophic powdery mildew fungi. The identity of the RAB5-labeled EHM differs from the RAB5-positive membrane in the host cytoplasm, suggesting a modulation of the RAB5 function upon pathogen infection. Furthermore, we determined that the modulation of the ARA6-mediated pathway could be shared by the phylogenetically distinct obligate biotrophic oomycete, but not the hemibiotrophic fungus. These results suggest that RAB5 is involved in distinct plant–pathogen interactions between obligate biotrophic and hemibiotrophic pathogens.
ARA6 forms a membrane around the haustorium of the adapted powdery mildew fungus Golovinomyces orontii
To obtain insights into the relationship between plant-specific ARA6 and the EHM, Arabidopsis plants expressing native promoter-driven ARA6 as a fusion with green fluorescent protein (GFP) were infected with Arabidopsis-adapted Golovinomyces orontii ( Go ) ( Micali et al. 2008 ). Go invades Arabidopsis epidermal cells and forms the haustorium by 1 day post-inoculation (dpi). We found a uniform membrane-like localization of ARA6–GFP enclosing round structures in Go -infected epidermal cells at 20–24 hours post-inoculation (hpi) by confocal microscopic observation ( Fig. 1 A). Propidium iodide (PI) staining to visualize fungi ( Koh et al. 2005 ) revealed that ARA6–GFP localized to the membranous structure surrounding haustoria ( Fig. 1 B) in addition to the punctate endosomes (indicated by an arrow in Fig. 1 B). Three-dimensional (3D) reconstruction of the distribution of ARA6 indicated that ARA6–GFP completely enclosed the haustorium ( Fig. 1 C; Supplementary Movie S1 ), the membranous localization of which appeared to start from the haustorial neckband and not to extend to the host PM ( Fig. 1 B).
ARA6–GFP surrounded 100% of haustoria when ARA6–GFP leaves were observed from 12 hpi to 3 dpi (172 haustoria were observed in 28 independent experiments), except for those encased with the callosic encasement (described below). Previously it was reported that the endoplasmic reticulum, Golgi apparatus and peroxisomes localized close to the haustorium ( Koh et al. 2005 ); however, the localization pattern of those organelles was patchy and discontinuous, which was clearly different from the localization pattern of ARA6–GFP entirely surrounding the haustorium.
ARA6 localizes to the extrahaustorial membrane
The membrane-like fluorescence enclosing the haustoria could result from ARA6–GFP on/in (i) the cytosol; (ii) the tonoplast; (iii) the EHMx or (iv) the EHM. The localization of ARA6–GFP was distinct from the fluorescent pattern observed in cytosolic GFP; cytosolic free GFP (cytGFP) yielded vague fluorescence with variable thicknesses, in some cases with very faint fluorescence ( Supplementary Fig. S1A–C ). Conversely, ARA6–GFP exhibited a continuous localization with uniform thickness around the haustoria in all confocal slices we observed ( Supplementary Fig. S1D–F ). Although weak fluorescence was observed in the cytosol in cells expressing ARA6–GFP, its intensity was much lower than that due to ARA6–GFP localized on the endosomes and around the haustorium. The immunoblot analysis using the anti-GFP antibody further confirmed the marginal amount of free GFP in Go -infected leaves ( Supplementary Fig. S2 ). Those results indicate that the ARA6–GFP surrounding the haustorium did not result from a cytosolic population. The possibility of vacuolar localization of ARA6–GFP was also ruled out, as described below. Furthermore, ARA6–GFP around the mature, lobed haustoria was uniformly membranous ( Supplementary Fig. S1E , F ), in contrast to the expected localization to the EHMx, which embeds the haustorial lobes.
These results strongly suggest localization of ARA6–GFP to the EHM. We conclusively demonstrated this by immunoelectron microscopy using the same anti-GFP antibody used in the immunoblot analysis ( Supplementary Fig. S2A ), which specifically recognized GFP. Gold particle labeling was observed only at the outermost region of the haustorial complex ( Fig. 1 D, E ), and no specific labeling was detected in the sections that were processed without the primary antibody ( Fig. 1 F). Thus, ARA6–GFP was localized to the EHM of the Go haustorium.
Localization of ARA6 to the EHM is correlated with the function of the haustorium
Only a few plant proteins have been identified as localized to the powdery mildew fungal EHM. The EHM localization of those proteins, pea glycoproteins ( Roberts et al. 1993 ) and the Arabidopsis resistance protein RPW8.2 ( Wang et al. 2009 , Micali et al. 2011 ), is limited to specific stages of haustorium development ( Roberts et al. 1993 , Wang et al. 2009 , Micali et al. 2011 ).
To investigate if the localization of ARA6 to the EHM is also limited to specific stages of haustorium development, we observed various maturation stages of haustoria in leaves at 3 dpi. At this stage, Go has formed many haustoria at different developmental stages along the hyphae. Young round haustoria just after the invasion are observed around the tip of elongating hypha, whereas mature haustoria with more elongated shape and many lobes are observed near the conidium ( Inada et al. 2016 ). The localization of ARA6–GFP to the EHM was found with all haustoria at any developing stages ( Fig. 2 ). We also conducted time-lapse imaging at the Go invasion stage (12–14 hpi) and found that ARA6–GFP accumulated at the site of fungal invasion, after which its accumulation at the cell surface disappeared and the young haustorium surrounded by ARA6–GFP emerged ( Supplementary Fig. S3 ).
In contrast, ARA6–GFP was not present on the EHM of haustoria at the senescing stage covered with a callosic encasement. The callosic encasement begins to form at the basal part of the haustorium and grows apically ( Meyer et al. 2009 ). The EHM has been shown to locate inside the encasement ( Wang et al. 2009 ). ARA6–GFP was located only outside the callosic encasement, but was absent on the part of EHM surrounded by the encasement (indicated by arrows in Fig. 3 A), whereas the uncovered apical part was still decorated by ARA6–GFP ( Fig. 3 B, C ; similar patterns of ARA6–GFP were observed for nine haustoria that were partially surrounded by the encasement). These results suggest that the EHM that is surrounded by the encasement is qualitatively different from uncovered EHM.
Arabidopsis is a non-host of barley-adapted Blumeria graminis f.sp. hordei ( Bgh ), and Bgh occasionally develops a haustorium for a short time before the infected cell undergoes cell death ( Collins et al. 2003 ). Only punctate or sporadic distributions of ARA6–GFP were observed around the Bgh haustorium ( Fig. 3 D). This difference in the distribution patterns of ARA6–GFP around the haustoria between the adapted and non-adapted powdery mildew fungi was not due to the cell death process, as the punctate localization of ARA6–GFP on the endosome was still visible in cells undergoing cell death ( Supplementary Fig. S4 ).
The EHM exhibits modified endosomal characteristics
The localization of ARA6–GFP to the Go EHM suggests that this EHM has, at least in part, the identity of the host endosomal membrane. To investigate these endosomal characteristics further, we examined the EHM localizations of other GFP-tagged endosomal proteins, specifically ARA7, VPS9a and VAMP727, whose expression was driven by their native regulatory elements.
ARA7 is a conventional RAB5 ( Ueda et al. 2001 ) that localizes to a distinct population of endosomes that shows considerable overlap with those occupied by ARA6 ( Ueda et al. 2004 , Ebine et al. 2011 ). GFP–ARA7 also localized to the Go EHM ( Fig. 4 A) in a pattern that perfectly overlapped with that of ARA6–monomeric red fluorescent protein (mRFP) ( Fig. 4 B–D). We also observed ARA7 localization to the EHM of all haustoria we observed during 1–3 dpi (27 haustoria were observed in 12 independent experiments). ARA7 was also absent from the part of the EHM enclosed by the haustorial encasement ( Supplementary Fig. S5 ).
RAB GTPase functions as a molecular switch by cycling between activated GTP-bound and inactivated GDP-bound states ( Inada and Ueda 2014 ). This cycling is catalyzed by guanine nucleotide exchange factor (GEF). In vegetative tissues of Arabidopsis, VPS9a (At3g19770) acts as a sole activating factor for all RAB5 members ( Goh et al. 2007 ). Although VPS9a co-localized effectively with ARA6 and ARA7 on punctate endosomes ( Fig. 4 E, F ), it was not associated with the membranous structure surrounding the haustorium ( Fig. 4 G).
VAMP727 (At3g54300), a SNARE protein that is responsible for the fusion of endosomes with the vacuoles and the PM and that was demonstrated to co-localize with RAB5 members on punctate endosomes in Arabidopsis cells ( Ebine et al. 2008 , Ebine et al. 2011 ), was also not associated with the membranous structure around the haustorium ( Fig. 4 H). These results indicate that the EHM has a unique specificity for endosomes, at which RAB5 members were located but RAB5 GEF and endosomal SNAREs were not.
The EHM of powdery mildew fungus does not acquire RAB7, a marker of late endosomes and tonoplast
In the case of symbiotic Rhizobium bacteria, the symbiosomal membrane does not contain host legume RAB5 but instead harbors RAB7 ( Limpens et al. 2009 ), which localizes to the tonoplast in Arabidopsis ( Saito et al. 2002 ). RAB7 members in Arabidopsis and Nicotiana benthamiana [RABG3c (At3g16100) and NbRabG3c-a, respectively] are also shown to be recruited to the EHM of the obligate biotrophic oomycete Phytophthora infestans when transiently expressed in N. benthamiana epidermal cells ( Bozkurt et al. 2015 ). We thus investigated the localization of RAB7 at the EHM formed by Go .
We observed transgenic plants expressing Arabidopsis RAB7 (RABG3f, At3g18820) conjugated with GFP (GFP–RAB7) under regulation of the native promoter inoculated with Go . However, as the distance between the tonoplast and the EHM is frequently <200 nm, as shown in Fig. 1 D, which is below the resolution limit of the light microscope, it was difficult to distinguish the fluorescence from the tonoplast from the signal on the EHM by confocal microscopy. To overcome this problem, we observed haustoria located near the host cell periphery, where the tonoplast should only partially surround the haustorium, whereas the EHM should completely surround the haustorium ( Fig. 5 A; Supplementary Fig. S6 ). The GFP–RAB7-positive membrane showed a discontinuity near the cell periphery, whereas ARA6–mRFP exhibited a continuous distribution around the haustorium ( Fig. 5 B–D; Supplementary Fig. S7 ). In addition, when the haustorium was closely associated with the host nucleus ( Fig. 5 E), GFP–RAB7 surrounded both the haustorium and the nucleus, whereas ARA6–GFP surrounded only the haustorium ( Fig. 5 F–K). GFP–SYP22/VAM3 (At5g46860), another marker of the tonoplast ( Sato et al. 1997 , Uemura et al. 2010 ), exhibited a localization pattern similar to that of GFP–RAB7 ( Supplementary Fig. S8 ).
These results indicate that RAB7 was not localized to the EHM formed by Go , and that the RAB5 localization was distinctive from that of the tonoplast.
Localization of RAB5 to the EHM is conserved among haustorium-forming obligate biotrophic pathogens
Finally, we evaluated whether the recruitment of RAB5 to the EHM is conserved among other haustorium-forming obligate biotrophic pathogens.
First, we examined the barley– Bgh pathosystem. We isolated the cDNA for barley ARA6 (HvARA6), which shares structural characteristics, including a consensus sequence for dual fatty acylation at the N-terminus, with Arabidopsis ARA6 ( Supplementary Fig. S9 ). When transiently expressed in barley leaf epidermal cells, HvARA6–GFP generated motile, punctate fluorescent patterns in the cytosol ( Fig. 6 A; Supplementary Movie S2 ), similar to those observed for the endosomes labeled with ARA6–GFP in Arabidopsis cells. Following Bgh infection, HvARA6–GFP exhibited a continuous membranous structure that surrounded the haustorium with finger-like protrusions ( Fig. 6 B), indicating that HvARA6 was localized to the EHM (all of the 15 haustoria observed in four independent experiments exhibited a similar localization of HvARA6–GFP) .
We then examined another obligate biotrophic pathogen, the Arabidopsis-adapted oomycete Hyaloperonospora arabidopsidis ( Hpa ) ( Coates and Beynon 2010 ). ARA6–GFP localized to the continuous membranous structures, exhibiting bright fluorescent dots around the Hpa haustoria that formed in mesophyll cells at 4 dpi ( Fig. 6 C; Supplementary Movie S3 ), indicating its localization to the EHM. ARA6 has been reported to accumulate in the haustorial encasement ( Lu et al. 2012 ). However, we noted that the Hpa encasement at 6 dpi emitted strong autofluorescence, which was detected as the signal from GFP or RFP using conventional filter-based spectrometry ( Supplementary Fig. S10 ). Efficient separation of the autofluorescence of the haustorial encasement from the GFP fluorescence by spectral imaging microscopy revealed that the EHM localization of ARA6–GFP was absent from the haustoria with encasements ( Fig. 6 D). All haustoria without the encasement we observed were surrounded by ARA6–GFP (53 haustoria were observed in three independent experiments). Thus, ARA6 localized to the EHM of the Hpa haustoria without the encasement.
We then extended our investigation to the hemibiotrophic pathosystem. The Arabidopsis-adapted hemibiotrophic fungus Colletotrichum higginsianum ( Ch ) initially grows as biotrophic hyphae in living host cells and then switches to a destructive necrotrophic phase ( O’Connell et al. 2004 ). The plant-derived membrane surrounding the biotrophic primary hyphae is continuous with the plant PM, and is labeled with the Arabidopsis PM marker GFP–SYP121 (At3g11820) at the early infection stage ( Shimada et al. 2006 ). We also found that GFP–SYP132 (At5g08080), another PM marker ( Uemura et al. 2004 , Enami et al. 2009 ), entirely surrounded the biotrophic hyphae at the early infection stage as in the case of SYP121 ( Fig. 6 E–H; Supplementary Fig. S11 ). Although ARA6–mRFP localized close to the primary biotrophic hyphae, the fluorescence intensity around the hyphae was weaker than that due to ARA6–mRFP on endosomes, and the localization patterns were slightly different between GFP–SYP132 and ARA6–mRFP ( Fig. 6 E–H). The plot of fluorescence intensity also indicated that the peak of ARA6–mRFP located outside that of GFP–SYP132 ( Fig. 6 I), confirming that ARA6–mRFP did not target to the membrane surrounding Ch biotrophic hyphae.
Thus, our results indicated that the nature of the plant-derived membrane surrounding the biotrophic hyphae of the hemibiotroph Ch differs from the nature of the EHM of obligate biotrophic pathogens.
In this study, we demonstrated that the host plant RAB5, both plant-specific ARA6 and conventional ARA7, localized to the host-derived membrane surrounding the haustorium formed by the powdery mildew-causing fungus, Go . This membrane did not acquire the RAB5 activator and endosomal/vacuolar SNARE molecules, suggesting its modified endosomal characteristics. Recruitment of ARA6 to the EHM was also observed for the other obligate biotrophic fungus and oomycete. In contrast, the hemibiotrophic fungus at the biotrophic stage did not recruit ARA6 to the membrane surrounding the intracellular hypha. Therefore, ARA6 localization at the plant–pathogen interface would be specific to the haustorium-forming obligate biotrophs.
In contrast to the previously reported proteins that localize to the EHM formed by powdery mildew fungi, the localization of ARA6 to the EHM was observed throughout the biogenesis and maturation of the haustorium. This result suggests that the study of ARA6 could lead to an understanding of the novel mechanism that is responsible for the biogenesis and development of the haustorium that is formed by powdery mildew fungi. Furthermore, we found that ARA6 was absent from the EHM of senescent haustoria enclosed by a callosic encasement, which are thought to form as a result of the plant immune response ( Meyer et al. 2009 ). This absence of ARA6 from the EHM of encased haustoria may suggest that (i) the recruitment of ARA6 to the EHM is abrogated when the host defense prevails; or (ii) the recruitment of ARA6 to the EHM occurs at the stage of haustorium development that is later than the formation of the encasement. However, as we showed in Fig. 2 and Supplementary Fig. S3 , ARA6 localization to the EHM was seen from the initiation of haustorium formation and throughout the maturation; thus, the second possibility is unlikely. Therefore, we conclude that ARA6 is actively excluded from the EHM when it becomes surrounded by the encasement. This absence of ARA6 on the EHM surrounded by callosic encasement and the fact that Bgh haustoria do not acquire ARA6 throughout the EHM membrane could suggest a correlation between ARA6 localization to the EHM and the integrity and/or functionality of the haustorium. Thus, ARA6 is also expected to be a useful marker of the functional EHM of powdery mildew fungi.
With respect to the EHM formed by oomycetes, several factors have been identified as constituents of the EHM by two recent studies: a semi-comprehensive analysis of membrane trafficking-related factors around the haustoria formed by Hpa and Phytophthora infestans ( Lu et al. 2012 ) and a localization analysis of transiently expressed REMORIN1.3 (REM1.3) and RAB7 (NbRabG3) in N. benthamiana infected by P. infestans ( Bozkurt et al. 2014 ). It has also been reported that effector molecules secreted by oomycetes, such as Hpa HaRxL17 ( Caillaud et al. 2012a , Caillaud et al. 2012b ) and P. infestans AVRblb2 ( Bozkurt et al. 2011 ), are found on the EHM. Among those EHM-localizing factors, HaRxL17 is shown to localize to the EHM of all haustoria, including those encased by the haustoria ( Caillaud et al. 2012a ). Conversely, REM1.3 and RAB7 target the EHM of approximately half of the haustoria formed by P. infestans ( Bozkurt et al. 2014 , Bozkurt et al. 2015 ). Our result shows that ARA6 localized to the EHM of all Hpa haustoria without encasement but was excluded from encased haustoria and could thus indicate that ARA6 is involved in a layer of regulation of EHM function different from the previously identified oomycete EHM proteins.
The acquisition of RAB5 GTPases, plant-unique ARA6 and conventional ARA7, and the absence of their activating factor and endosomal/vacuolar SNAREs from the EHM strongly suggest the modification of host endocytic components by the powdery mildew fungal pathogen. Recently, rerouting of host vacuolar trafficking to the EHM was proposed by a study of N. benthamiana infected by P. infestans , which reported that several late endosomal/vacuolar proteins including RAB7 (NbRabG3) were localized to the EHM ( Bozkurt et al. 2015 ). However, our result demonstrated that RAB7 was absent from the EHM formed by Go in Arabidopsis. These results suggest the distinct modulation of endosomal/vacuolar trafficking pathways between biotrophic pathogens, whereas RAB5 is commonly recruited to the EHM of haustoria formed by biotrophic fungi and oomycetes.
As ARA6 and ARA7 regulate distinct endosomal trafficking events ( Ebine et al. 2011 ), it will be a future study of interest to investigate whether these RAB5 members contribute to the plant–pathogen interaction in the same or a distinct manner. Moreover, further studies of the precise functioning and regulatory mechanisms of RAB5 recruitment to the EHM in each pathosystem will provide a comprehensive understanding of the relationship between membrane trafficking and plant–pathogen interaction, and thus will contribute to establishing new strategies to overcome the threat of plant diseases.
Materials and Methods
Transgenic plants expressing GFP–VAMP727, GFP–SYP22/VAM3, GFP–RAB7, GFP–SYP132, GFP–SYP121, VPS9a–GFP or VPS9a–Venus were generated as described previously ( Goh et al. 2007 , Ebine et al. 2008 , Enami et al. 2009 , Ebine et al. 2011 , Ebine et al. 2014 , Sunada et al. 2016 ). For visualization of ARA6, the cDNA for GFP or mRFP was inserted in front of the stop codon in the 5.7 kb genomic fragment of ARA6 (At3g54840), including the 2.7 kb promoter, exons, introns and the 3'-flanking region. The chimeric genes were subcloned into the binary vector pGWB1 ( Nakagawa et al. 2007 ), and the resulting plasmids were used for transforming Col-0 and ara6-1 . For visualization of ARA7, the cDNA for GFP or mRFP was inserted in front of the start codon or behind the stop codon, respectively, in the 4.5 kb genomic fragment of ARA7 (At4g19640), including the 1.5 kb promoter, exons, introns and the 3'-flanking region. The chimeric genes were subcloned into the binary vector pGWB1, and the resulting plasmids were used for transforming Col-0. Transgenic plants expressing GFP–ARA7 and ARA6–mRFP, GFP–RAB7 and ARA6–mRFP, or GFP–SYP132 and ARA6–mRFP were generated by crossing Col-0 plants expressing GFP–ARA7, GFP–RAB7 or GFP–SYP132 and the Col-0 plant expressing ARA6–mRFP. The transgenic plant expressing VPS9a–GFP and mRFP–ARA7 was generated by transforming vps9a-1 plants expressing VPS9a–GFP with the binary vector pBGW ( Karimi et al. 2002 ) containing mRFP– ARA7 . The transgenic plant expressing VPS9a-Venus and ARA6-mRFP was generated by crossing the vps9a-1 mutant expressing VPS9a-Venus and the ara6 mutant expressing ARA6-mRFP.
Pathogen inoculation and disease resistance scores
For the Go susceptibility assay, 4-week-old Arabidopsis plants were inoculated with Go conidiospores. Conidiophore counting was performed as previously described ( Inada and Savory 2011 ). For the Hpa experiments, 1-week-old Arabidopsis seedlings were sprayed with a suspension of 4.0×10 4 spores ml –1 of the Hpa isolate Noco2. For the Ch experiments, one drop (2 μl) of conidial suspension (1–2×10 5 conidia ml –1 ) was spotted on each cotyledon of 1-week old Arabidopsis seedlings. The inoculated plants were placed at 100% relative humidity for 2–3 d.
Go -infected leaves were stained with PI as described previously ( Koh et al. 2005 ) and then observed using a Leica confocal system equipped with ×20 N.A. 0.7 and ×63 N.A. 1.4 objective lenses (TCS-SP5, Leica Microsystems). GFP fluorescence was excited at 488 nm and monitored with a band-pass filter at 500–550 nm. The PI was excited at 561 nm and monitored with a band-pass filter from 580–650 nm. The images were reconstructed and analyzed using ImageJ64 version 1.46f ( http://rsbweb.nih.gov/ij/ ). Hpa -infected Arabidopsis leaves were observed with a Zeiss confocal microscope LSM710 equipped with a × 60 N.A. 1.4 objective lens (Carl Zeiss Microscopy GmbH). Samples were excited at 488 nm using a laser, and images were processed using ZEN software (Carl Zeiss). Ch -infected Arabidopsis leaves were observed using a confocal microscope LSM780 (Carl Zeiss Microscopy GmbH). For the analysis of Ch primary infection hyphae, 1-week-old Arabidopsis seedlings were inoculated with either wild-type Ch or Ch expressing mRFP ( Hiruma et al. 2010 ). The leaves were then observed using a confocal microscope LSM780 (Carl Zeiss) equipped with a ×63 N.A. 1.40 objective lens. GFP was excited at 488 nm with a DPSS laser and the emission from 493 nm to 553 nm was observed. ARA6–mRFP was excited with a 561 nm DPSS laser and the emission from 561 nm to 641 nm was observed. Ch expressing mRFP was excited with a 561 nm DPSS laser and the emission from 561 nm to 695 nm was observed.
Reconstruction and computer visualization of the 3D distribution of ARA6–GFP
The serial confocal images were recorded at an image size of 1,024×1,024×78 voxels, with a voxel dimension of 0.05×0.05×0.21 μm using TCS-SP5 equipped with a ×63 objective lens (N.A. 1.4) and processed using ImageJ version 1.46j with KBI ImageJ plugins ( http://hasezawa.ib.k.u-tokyo.ac.jp/zp/Kbi/ImageJKbiPlugins ). For the original image stacks, we applied a 2D Gaussian filter with a standard deviation of 1 μm to reduce noise and eliminate minor aberrations of the point-spread function using the ‘bandPassOps’ mode of the Kbi_Filter2d plugin. To improve the response of the volume rendering, as described below, the smoothed images were down-sampled by a factor of four both horizontally and vertically using a bi-cubic interpolation algorithm in the ‘scaleXy’ mode of the Kbi_Filter2d plugin. The down-sampled images were converted to Visualization Tool Kit (VTK, http://www.vtk.org/ ) format and saved as VTK files in the ‘saveVol’ mode of the Kbi_Vtk plugin. Once the VTK files were saved, the 3D distribution of ARA6–GFP was displayed by volume rendering using the ‘volume’ module of the visualization software Amira version 1.5 ( http://www.amira.com/ ). We captured the image displayed in the rendering window of Amira and encoded it to a movie file.
For structural observation, ARA6–GFP leaves at 5 dpi with Go were excised as 2 mm×2 mm sections and fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0.05 M cacodylate buffer (pH 7.4) at 4°C overnight. After this fixation, the samples were washed three times with 0.05 M cacodylate buffer for 30 min each, and were post-fixed with 2% osmium tetroxide in 0.05 M cacodylate buffer at 4 °C for 3 h. The samples were dehydrated in graded ethanol solution, and then were continuously dehydrated in 100% ethanol at room temperature overnight. After dehydration, samples were infiltrated with propylene oxide, and with Quetol-651 resin (Nisshin EM Co.). Ultrathin sections were collected on copper grids and stained with 2% uranyl acetate and lead. For immunoelectron microscopy, ARA6–GFP leaves at 5 dpi with Go were excised as 2 mm×2 mm sections and were fixed with 0.1% glutaraldehyde and 0.5% tannic acid in 0.05 M sodium cacodylate buffer (pH 7.4) at 4 °C for 90 min, and then washed three times in 0.1 M cacodylate buffer for 20 min each. Samples were dehydrated in a graded ethanol series and then infiltrated with LR white resin (London Resin Co. Ltd.). Ultrathin sections were collected on nickel grids and subjected to immunolabeling using 1 : 20 diluted anti-GFP antibody (rabbit, ab6556, Abcam) and 1 : 50 diluted 10 nm gold particle-conjugated anti-rabbit antibodies (Goat, Funakoshi) followed by 2% uranyl acetate staining. The sections were observed using transmission electron microscopy (JEM-1400Plus; JEOL Ltd.).
Observation of HvARA6
Total RNA was extracted from 1-week-old barley (Kobinkatagi) leaf blades using TRIzol reagent (Life Technologies Japan) according to the manufacturer’s protocol. cDNA synthesized from 1 μg of RNA using a 20 nucleotide oligo(dT) primer and ReveTra Ace reverse transcriptase (TOYOBO Life Science) were used to clone HvARA6–GFP. HvARA6-specific primers were designed based on a sequence deposited in GenBank (AK248897_1). The 1.5 kb HvARA6–GFP fragment was subcloned into pUbq-GW ( Taoka et al. 2011 ), and the resultant plasmid was delivered to the epidermal cells of the adaxial side of detached primary leaves from barley (Kobinkatagi) using a Hepta adaptor and a Biolistic ® PDS-1000/He particle delivery system (Bio-Rad Laboratories) according to the manufacturer’s instructions. The bombarded specimens were subjected to Bgh race1 inoculation as previously described ( Zhou et al. 2001 ). Primers used for cloning of HvARA6–GFP are listed in Supplementary Table S1 .
The results are expressed as the mean ± SE from an appropriate number of experiments as indicated in the figure legends. Student’s t -test was used to analyze the statistical significance.
Supplementary data are available at PCP online.
This study was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan [Grants-in-Aid for Scientific Research to N.I. and T.U.]; the Japan Science and Technology Agency (JST) [PRESTO (to T.U.)]; Nara Institute of Science and Technology (NAIST) [Grants-in-Aid for Scientific Research].
We thank Ms. S. Nakagawa (NAIST) and Dr. T. Inoue (University of Tokyo) for technical assistance. The Arabidopsis line expressing GFP–SYP121 and GFP-SYP132 was provided by Dr. K. Enami and Professor M.H. Sato (Kyoto Prefectural University).
The authors have no conflicts of interest to declare.
Blumeria graminis f.sp. hordei
guanine nucleotide exchange factor
green fluorescent protein
monomeric red fluorescent protein
soluble N -ethylmaleimide-sensitive factor attachment protein receptor