Arabinogalactan Proteins Accumulate in the Cell Walls of Searching Hyphae of the Stem Parasitic Plants, Cuscuta campestris and Cuscuta japonica

Abstract

Stem parasitic plants (Cuscuta spp.) develop a specialized organ called a haustorium to penetrate their hosts’ stem tissues. To reach the vascular tissues of the host plant, the haustorium needs to overcome the physical barrier of the cell wall, and the parasite–host interaction via the cell wall is a critical process. However, the cell wall components responsible for the establishment of parasitic connections have not yet been identified. In this study, we investigated the spatial distribution patterns of cell wall components at a parasitic interface using parasite–host complexes of Cuscuta campestris–Arabidopsis thaliana and Cuscuta japonica–Glycine max. We focused on arabinogalactan proteins (AGPs), because AGPs accumulate in the cell walls of searching hyphae of both C. campestris and C. japonica. We found more AGPs in elongated haustoria than in pre haustoria, indicating that AGP accumulation is developmentally regulated. Using in situ hybridization, we identified five genes in C. campestris that encode hyphal-expressed AGPs that belong to the fasciclin-like AGP (FLA) family, which were named CcFLA genes. Three of the five CcFLA genes were expressed in the holdfast, which develops on the Cuscuta stem epidermis at the attachment site for the host’s stem epidermis. Our results suggest that AGPs are involved in hyphal elongation and adhesion to host cells, and in the adhesion between the epidermal tissues of Cuscuta and its host.

Introduction

Parasitic plants have developed a specialized root-like organ called a haustorium to absorb water and nutrients from the host plant. The stem tissues of stem parasitic plants (Cuscuta spp.) increase cell division after touching the host stem (Lee 2007). This area of increased cell division is called the pre haustorium, from which the elongation of the haustorium originates. In the tips of haustoria, long cells are formed, which eventually become searching hyphae that elongate toward the host’s vascular tissues (Alakonya et al. 2012).

To establish a parasitic connection, Cuscuta spp. need to overcome the physical barriers of the host plant, particularly the cuticle and cell wall. During the penetration stage, the expression levels of genes that encode cell wall-degrading and -modifying enzymes increase at the parasitic interface (Ikeue et al. 2015). An enzyme responsible for the de-esterification of pectin (pectin methylesterase) has been detected in the haustoria of the root parasite Orobanche cumana, and in the adjacent host apoplast (Losner-Goshen et al. 1998). Pectin methylesterase increases cell wall degradation, because de-esterified pectin is a good substrate for pectolytic enzymes, such as polygalacturonase and pectate lyase (Wakabayashi et al. 2003). Similarly, high concentrations of de-esterified homogalacturonans in the cell wall adjacent to the haustorium, and high expression levels of pectate lyase, have been observed in a parasite–host complex of Cuscuta reflexa and Pelargonium zonale (Johnsen et al. 2015). This suggests that parasitic plants directly contribute to cell wall modification or remodeling.

Proteins and polysaccharides are the main structural components of plant cell walls. A previous study revealed the complexity of cell wall proteomes (Somerville et al. 2004), and cell wall proteins are key molecules in signaling and defense regulation (Ellis et al. 2010). Arabinogalactan proteins (AGPs) are cell wall proteins that have arabinogalactan side chains attached to the core protein, and are in many cases glycosylphosphatidylinositol (GPI) anchored. AGPs are involved in various biological processes, including cell division, pattern formation, pollen tube guidance and plant–microbe interactions (Seifert and Roberts 2007). Recently, AGP accumulation was mapped in and around the haustoria of C. reflexa invading P. zonale, indicating that extensins tend to increase in infected areas (Striberny and Kraus 2015). It has also been reported that contact between C. reflexa pre haustoria and the stem surfaces of tomato plants induces the expression of a tomato gene that encodes the AGP gene, attAGP (Albert et al. 2006). Suppression of attAGP expression in tomato by RNA interference (RNAi) significantly reduces the capability for attachment of C. reflexa, suggesting that C. reflexa infection induces a signal in the host that results in the expression of attAGP, which increases the parasite’s adherence (Albert et al. 2006). The results of these previous studies suggest that AGPs are involved in parasitic processes, but their precise role has not been ascertained.

In this study, we investigated the spatial distribution patterns of cell wall components at a parasitic interface using parasite–host complexes of C. campestris–Arabidopsis thaliana and C. japonica–Glycine max. We focused on AGPs because AGP accumulation has been commonly observed in the cell walls of searching hyphae of both C. campestris and C. japonica. Using in situ hybridization, we identified five C. campestris genes that encode hyphal-expressed AGPs that belong to the fasciclin-like AGP (FLA) family, which we named CcFLA genes. Three of the five CcFLA genes were also expressed in the holdfast, which develops on the Cuscuta stem epidermis at the attachment site for the host’s stem epidermis. Our results suggest that AGPs are involved in hyphal elongation and adhesion to host cells.

Results

Localization of cell wall components at the parasitic interface between C. campestris and A. thaliana

We investigated the distribution patterns of cell wall components in a non-targeted manner by immunohistochemistry, at the interface between C. campestris and A. thaliana at 120 hours after attachment (haa) (Table 1; Fig. 1). Using the CCRC-M1 antibody (that binds to fucosylated xyloglucan), no significant signal was detected in the haustorium (Fig. 1A, B), whereas in the host’s stem, the label was detected in the xylem. Using the LM15 antibody (that binds to xyloglucan with a XXXG-motif), a weak signal was detected in the central part of the haustorium and a strong signal was detected in the host’s xylem adjacent to the tip of the haustorium (Fig. 1C, D). Using the LM19 antibody (that binds to low-esterified homogalacturonan), a significant signal was detected in the lateral part of the haustorium (Fig. 1E, F). In contrast, using the LM20 antibody (that binds to high-esterified homogalacturonan), a strong signal was detected in the central part of the haustorium (Fig. 1G, H) and, using the LM6 antibody [that binds to (1–5)-α-l-arabinan], a strong signal was detected that was localized to the cell walls of searching hyphae (Fig. 1I, J).

Table 1

Antibodies used for immunostaining

Antibody	Epitope detected
LM2	β-linked glucuronic acid of an arabinogalactan protein
LM6	(1–5)-α-l-arabinan
LM15	XXXG-motif of xyloglucan
LM19	Homogalacturonan with a low degree of esterification
LM20	Homogalacturonan with a high degree of esterification
CCRC-M1	α-Fuc-(1,2)-β-Gal glycan group of fucosylated xyloglucans

Fig. 1

Localization of cell wall components at the parasitic interface of a Cuscuta campestris–Arabidopsis thaliana complex at 120 hours after attachment (haa). (A, C, E, G, I, K) Transmitted light images. (B) CCRC-M1 antibody, (D) LM15 antibody, (F) LM19 antibody, (H) LM20 antibody, (J) LM6 antibody and (L) control serum. Cc, C. campestris; At, A. thaliana; x, xylem; ha, haustorium. Asterisks indicate searching hyphae. Scale bar = 50 µm. The dotted line shows the outline of C. campestris.

LM6 epitopes were localized to the cell walls of searching hyphae of C. japonica

To test whether the distribution patterns of cell wall components are conserved in different parasite–host complexes, we immunostained the interface tissue between C. japonica and G. max. The side surfaces of the haustoria were labeled with control antiserum (Fig. 2L), which was due to non-specific antibody binding. Labeling with CCRC-M1 resulted in a strong signal in the metaphloem of G. max adjacent to C. japonica haustoria, indicating an accumulation of fucosylated xyloglucan in the metaphloem of the host (Fig. 2A, B). Immunostaining with LM15 was scarcely detected in the haustoria of C. japonica (Fig. 2C, D), and LM19 (Fig. 2E, F) and LM20 (Fig. 2G, H) signals were similar in the haustoria and adjacent host tissues. The differential localization of LM19 and LM20 epitopes that was observed in C. campestris (Fig. 1F, H) was not observed in C. japonica. The boundary between the host’s and parasite’s tissues at the tip of the haustorium was clearly observed using the LM6 antibody (Fig. 2I, J).

Fig. 2

Localization of cell wall components at the parasitic interface of a Cuscuta japonica–Glycine max complex at 120 hours after attachment (haa). (A, C, E, G, I, K) Transmitted light images. (B) CCRC-M1 antibody, (D) LM15 antibody, (F) LM19 antibody, (H) LM20 antibody, (J) LM6 antibody and (L) control serum. White triangles in (A) indicate metaphloem tissue of G. max in which the CCRC-M1 signal density increased with the progression of the parasitic process. Cj, C. japonica; Gm, G. max; x, xylem; mp, metaphloem; ha, haustorium. Asterisks indicate searching hyphae. Scale bar = 50 µm. The dotted line shows outline of C. japonica.

The distribution patterns of cell wall components in the C. japonica–G. max complex were slightly different from those in the C. campestris–A. thaliana complex. However, an accumulation of LM6 epitopes in the tips of the haustoria was observed in both complexes, and was investigated further.

AGP accumulation in the cell walls of searching hyphae

The LM6 antibody can recognize (1–5)-α-l-arabinan in pectic polymer and AGPs. To clarify whether AGPs were present in the hyphal cell walls, we stained the interface tissue with Yariv reagent, which specifically binds to a glycan moiety of AGPs. In C. campestris, the cell walls of searching hyphae were stained a distinct brown color with β-glucosyl Yariv reagent (βGlcY), but not with α-galactosyl Yariv reagent (αGalY) (Fig. 3A, C). Similarly, staining specific to βGlcY was detected at the boundary between C. japonica hyphae and the host’s xylem cells (Fig. 3B, D). We further confirmed the presence of AGP in hyphal cell wall of C. campestris by using LM2 antibody which recognizes β-linked glucuronic acid of AGP (Fig. 3E). These results demonstrate that AGPs were present in the cell walls of searching hyphae. In C. campestris, βGlcY staining was not only detected in cell walls in contact with Arabidopsis cells, but also in hyphal cell walls not in contact with Arabidopsis cells (Fig. 3A), indicating that AGPs are produced in hyphal cells.

Fig. 3

Arabinogalactan protein (AGP) localization in hyphal cell walls by staining with Yariv reagent and LM2 antibody. Haustorium tips of Cuscuta campestris (Cc; A and C) and Cuscuta japonica (Cj; B and D) harvested at 120 hours after attachment (haa) and stained with β-glucosyl Yariv (βGlcY) (A and B) or α-galactosyl Yariv (αGalY) (C and D). βGlcY specifically binds to a glycan moiety of AGPs, whereas αGalY does not, so αGalY was used as a negative control. (E) Haustorium tip of C. campestris at 240 haa stained with LM2 antibody detected by using Alexa Flour^® 488-conjugated secondary antibody (green). Asterisks indicate searching hyphae. Scale bar = 50 µm.

AGP accumulation in hyphal cell walls depended on the parasitic stage

Next, we investigated AGPs at different stages of the parasitic process. At 48 haa, when pre haustoria emerged in the stem cortex of C. campestris, βGlcY labeling in the cell walls was not detected in the tips of the pre haustoria (Fig. 4A), whereas the LM6 antibody clearly labeled the walls of the tip cells (Fig. 4C). LM2 antibody signal was mainly detected in the inner part of the cells, not in the cell wall (Fig. 4D). At 72 haa, when the haustoria penetrated the Arabidopsis stem tissues, significant staining of the hyphal cell walls with βGlcY reagent, LM6 and LM2 was detected (Fig. 4E, G, H). Therefore, pectin was present in the hyphal cell walls at the pre haustorium stage, but AGPs were not. However, at the penetration stage, AGPs and pectin accumulated in the hyphal cell walls (Fig. 4E, G, H). These results suggest that AGP accumulation in hyphal cell walls is developmentally regulated. In addition to the hyphal cell walls, significant βGlcY labeling was also observed on the epidermal surfaces, or holdfast, of C. campestris stems attached to the host’s stem surface (Fig. 4A, E).

Fig. 4

Stage-dependent accumulation of arabinogalactan proteins in hyphal cell walls. Cuscuta campestris haustorium tips at 48 hours after attachment (haa) (A–D) and 72 haa (E–H) stained with β-glucosyl Yariv (βGlcY) (A and E), α-galactosyl Yariv (αGalY) (B and F), LM6 antibody (C and G) or LM2 antibody (D and H). Cc, C. campestris; At, Arabidopsis thaliana; pha, pre haustorium; ha, haustorium. Asterisks indicate searching hyphae. Arrows indicate cells in the holdfast. Scale bar = 50 µm. LM6 and LM2 antibodies were detected using Alexa Flour^® 488-conjugated secondary antibody (green), and the section was counterstained with calcofluor white (blue).

Identification of C. campestris AGP genes expressed in the hyphae

To identify C. campestris genes that encode hyphal-expressed AGPs, we searched for tentative C. campestris consensus sequences that exhibited similarity with 11 AGP contigs of C. japonica, which are up-regulated during the parasitic process (Ikeue et al. 2015). We selected 19 AGP contigs of C. pentagona, which is phylogenetically closer to C. campestris than C. japonica, that has similarity to the C. japonica contigs (Supplementary Table S1; Ranjan et al. 2014), and cloned 17 C. campestris AGP cDNAs based on contig sequences of C. pentagona (CcFLA1–CcFLA17; Supplementary Data S1). Next, we performed in situ hybridization using parasitic interface tissue of the C. campestris–A. thaliana complex. Of the 17 candidate genes, five were localized to hyphal cells (CcFLA5, CcFLA7, CcFLA12, CcFLA16 and CcFLA17; Fig. 5B–F). It is noteworthy that the expression of three of these five genes, CcFLA7, CcFLA16 and CcFLA17, was also localized to the epidermal cells of the holdfast, which is an adhering organ that surrounds the basal end of the haustorium (Fig. 5C, D, E). This is consistent with the AGP accumulation observed on the surfaces of the holdfasts (Fig. 4A, E). Of 12 candidate genes which were not localized to hyphae, CcFLA3, CcFLA6, CcFLA8 and CcFLA9 were detected in the cells in the lateral boundary between the haustorium and the host’s stem, and CcFLA10 showed significant signal in developing xylem tissue of the main stem of C. campestris (Supplementary Fig. S1).

Fig. 5

Identification of CcFLA genes expressed in the hyphae (A–F) and their stage-dependent expression profiles (G–K). Paraffin sections were prepared from tissues at 72 hours after attachment (haa). (A) Negative control with a sense probe against CcFLA16, (B) antisense probe against CcFLA12, (C) antisense probe against CcFLA16, (D) antisense probe against CcFLA17, (E) antisense probe against CcFLA5 and (F) antisense probe against CcFLA7. ha, haustorium; hf, holdfast. Triangles indicate signals detected in the hyphae, and arrows indicate signals detected in the holdfast. Scale bars = 50 µm. Expression levels of (G) CcFLA12, (H) CcFLA16, (I) CcFLA17, (J) CcFLA5 and (K) CcFLA7 at 0, 72 and 120 haa were quantified with qRT-PCR by using parasitic interface tissues. The expression level was normalized by that of the C. campestris ribosome protein S18 gene (CcRPS18). The mean and SE of three biological replicates are indicated. PCR primers are listed in Supplementary Data S3.

Expression of CcFLA12 increased in the early stage from 0 to 72 haa, and decreased in the later stage from 72 to 120 haa (Fig. 5G). On the other hand, CcFLA16, CcFLA17, CcFLA5 and CcFLA7 decreased in the early stage, and then increased in the later stage (Fig. 5H–K).

Structure of CcFLA proteins expressed in the hyphae

All five CcFLA genes that were expressed in the hyphae contained a fasciclin-like domain (FAS1, IPR000782), so were classified as members of the FLA family. We performed a phylogenetic analysis of the FLA proteins in C. campestris, C. japonica and A. thaliana (Fig. 6), and a domain structure analysis of the five FLA genes expressed in the hyphae (Fig. 7; Supplementary Data S2). According to the structural classification conducted by Johnson et al. (2003), CcFLA16 and CcFLA17 belong to Group B, which is characterized by two FAS1 domains flanking an AGP region that contains at least two non-contiguous proline residues separated by alanine or serine. This short motif serves as a hydroxyproline-O-glycosylation site (Tan et al. 2003). N-Glycosylation motifs were found in the H1 region of FAS1 domains close to the N-termini of both proteins. CcFLA7 was also classified into Group B, and contained one FAS1 domain and five AGP regions (Fig. 7A). The CcFLA12 amino acid sequence was similar to that of Group A proteins, but lacked the AGP region (Fig. 7A). The domain structure of CcFLA5 was similar to that of Group C proteins, but lacked a C-terminal GPI anchor signal (Fig. 7A). The phylogenetic analysis revealed that the FLA proteins expressed in the hyphae do not cluster into a specific subclade, suggesting that several FLAs with different domain structures are involved in hyphal functioning. Alignment of the FAS1 domains of the five hyphal-expressed CcFLA proteins revealed that the H1 and H2 regions were conserved in all of the FAS1 domains (Fig. 7B). N-Glycosylation motifs were found in CcFLA16_FAS1-1 (Fig. 7B) and CcFLA17_FAS1-1 (Fig. 7B). The YH-motif, which interacts with integrins and mediates human fibroblast adhesion (Kim et al. 2002), was found in CcFLA17_FAS1-2 (Fig. 7B). Although the FAS1 domains of CcFLA5, CcFLA7 and CcFLA16_FAS1-2 had conserved tyrosine–histidine, they lacked two hydrophobic amino acids flanking the C-terminal side of the conserved tyrosine–histidine (Fig. 7B). The canonical DI-motif, which interacts with integrins and mediates human corneal cell adhesion (Kim et al. 2000), was not found in any of the five CcFLA proteins.

Fig. 6

Phylogenetic analysis of fasciclin-like arabinogalactan (FLA) proteins in Cuscuta campestris (red), Cuscuta japonica (blue) and Arabidopsis thaliana (black) based on deduced amino acid sequences using MEGA7 (Kumar et al. 2016). Boxed genes are CcFLA genes detected in the hyphae by in situ hybridization. Percentages of replicate trees in which associated taxa clustered together in the bootstrap test (500 replicates) are shown next to the branches (Felsenstein 1985). A, B, C and D indicate FLA groups according to the classification conducted by Johnson et al. (2003).

Fig. 7

(A) Schematic representation of hyphal-expressed CcFLA proteins deduced from cDNA sequences. Red, FAS1 domain (SM000554). Dark gray, Pfam:fasciclin domain (PF02469). Green, AGP region (hydroxyproline-O-glycosylation motif; Gaspar et al. 2001). Blue, N-glycosylation motif (Johnson et al. 2003). Light gray, glycosylphosphatidylinositol anchor signal. Numbers indicate positions. (B) Alignment of FAS1 domains. FAS1-1 and FAS1-2 represent FAS1 domains close to the N- and C-terminus, respectively. The H1 and H2 conserved regions that are characteristic of FAS1 domains are indicated below the alignment. Identical residues are shaded dark gray, and similar residues are shaded light gray. N-Glycosylation motifs in CcFLA16_FAS1-1 and CcFLA17_FAS1-1 are boxed. Conserved tyrosine–histidine residues of the YH-motif are in bold, and conserved hydrophobic residues are underlined.

Discussion

Spatially different distributions of pectin and xyloglucan were observed in the C. campestris–A. thaliana complex, but not in the C. japonica–G. max complex

Cell walls between parasitic plants and host plants play important roles in the host’s defense against parasitic plants and in establishing parasitic junctions, and would be expected to undergo biochemical changes during the penetration of haustoria through the host’s tissues, and once the tissues of both plants are connected. Johnsen et al. (2015) conducted a comprehensive analysis of cell wall composition in parasitic interface tissue of a stem parasitic plant, C. reflexa, and its host, P. zonale. They found that low-esterified homogalacturonan accumulated in infected cell walls of P. zonale, which coincided with high expression levels of pectate lyase genes in C. reflexa, suggesting that cell wall remodeling occurs during parasitic infection (Johnsen et al. 2015).

We used two parasite–host complexes, C. campestris–A. thaliana and C. japonica–G. max, to identify common changes in cell wall composition in two different complexes. Regarding pectin accumulation, we did not observe a large accumulation of low-esterified homogalacturonan in the area adjacent to the haustorium, which has been reported in the C. reflexa–P. zonale complex (Johnsen et al. 2015). In C. campestris, low- and high-esterified homogalacturonans accumulated in the side and central parts of the haustorium, respectively (Fig. 1F, H). However, this differential accumulation pattern was not observed in C. japonica (Fig. 2F, H).

Neither the CCRC-M1 antibody (Figs. 1B, 2B), which recognizes fucosylated xyloglucan, nor the LM15 antibody (Figs. 1D, 2D), which binds to xyloglucan with a XXXG-motif, strongly labeled cell walls in the haustoria of C. campestris or C. japonica. In contrast, fucosylated xyloglucan was highly accumulated in the metaphloem of G. max that was parasitized by C. japonica (Fig. 2B). The Arabidopsis mur1mur2 double mutant, which contains low levels of fucose in pectin, glycoprotein and xyloglucan, has a low cell wall tensile strength (Vanzin et al. 2002). Therefore, it can be hypothesized that increased levels of xyloglucan fucosylation strengthen cell walls in host phloem tissue, and increase defense against invasion by haustoria. However, an accumulation of fucosylated xyloglucan in the phloem was not observed in the C. campestris–A. thaliana complex (Fig. 1B). It has been reported that AGPs contain fucose, and fucosylated AGP is required for root cell elongation (van Hengel and Roberts 2002, Liang et al. 2013). In C. campestris–A. thaliana and C. japonica–G. max complexes, the CCRC-M1 epitope was not co-localized with AGP, suggesting that AGPs in the cell wall of hyphae were free of fucosylated xyloglucan. We did not, however, exclude the possibility that fucosylated AGP which can be detected by eel lectin (van Hengel and Roberts 2002, Liang et al. 2013) was present in hyphal cell wall, and involved in cell elongation.

AGP accumulation in the tips of haustoria was commonly observed in both parasitic complexes

We focused on AGP accumulation, which was observed in both parasitic complexes. Based on AGP detection at different parasitic stages (Figs. 1,4), we hypothesized that AGP accumulation is associated with haustorium development, which starts with the formation of a meristematic region called the pre haustorium in stem cortex tissue. Pre haustoria at 48 haa already have elongated cells, which is a characteristic of searching hyphae. AGPs, however, were not detected using βGlcY in the cell walls of these cells (Fig. 4A). At a later stage of haustorium development (72 haa), when searching hyphae penetrate the stem cortex tissue of the host, AGPs were clearly detected in hyphal cell walls (Fig. 4E). At 120 haa, when hyphal cells are still searching for their targets, even after the establishment of a vascular connection, AGP accumulation was observed in hyphal cell walls (Fig. 1J). These results suggest that AGP accumulation is developmentally regulated, or depends on contact between hyphae and host cells.

In the parasitic complexes, AGP accumulation was also observed on the surfaces of stem epidermal cell walls of the parasite, or holdfast, where the stem epidermis of the parasitic plant made physical contact with the host plant (Fig. 4A, E). A previous study on the morphology of the epidermis of C. pentagona revealed that epidermal cells excrete an amorphous substrate, or ‘cement’, in the space between the cuticle and the outer layers of the cell wall, and this cement makes a tight seal between the parasite and the host’s epidermal layers (Vaughn 2002). The deposition of an amorphous substrate was observed on the attaching epidermal surfaces of C. campestris and C. japonica, and AGP accumulation appeared to be co-localized with the amorphous substrate (Fig. 4A, E). These results suggest that AGPs play a critical role in the adhesion of the parasite to the host’s surface.

Structure of hyphal-expressed CcFLA genes and their putative roles in parasitic processes

We identified five CcFLA genes that were expressed in hyphal cells (Fig. 5B–F). We need to mention that probe sequences for CcFLA16 and CcFLA17 shared 80% identity, and we could not exclude the possibility of the occurrence of cross-hybridization. Expression of CcFLA16, CcFLA17, CcFLA5 and CcFLA7 was higher at 0 haa, soon after stem attachment, and decreased in 0–72 haa (Fig. 5H–K). On the other hand, expression of CcFLA12 was at a low level at 0 haa, and increased in 0–72 haa (Fig. 5G). These expression patterns, combined with the holdfast expression of CcFLA16, CcFLA17 and CcFLA7, imply that CcFLA16, CcFLA17, CcFLA5 and CcFLA7 might have roles in stem surface adhesion, while CcFLA12 has a role in the adhesion of hyphal surface to the host’s tissue. All of the deduced amino acid sequences had one or two FAS1 domains, and so belonged to the FLA family. A phylogenetic analysis of FLA family members of C. campestris, C. japonica and A. thaliana revealed that the five hyphal-expressed genes were not clustered into a specific clade, but CcFLA genes from four different structural groups were expressed in the hyphae (Fig. 6).

A sequence comparison of FAS1 domains in the five hyphal-expressed CcFLA proteins showed that CcFLA7, CcFLA16 and CcFLA17 have putative hydroxyproline-O-glycosylation sites (Fig. 7B). An analysis of a galt2galt5 double mutant found that the hydroxyproline-O-glycosylation of FLA4/SOS5 in Arabidopsis is essential for normal root growth (Basu et al. 2015), and defects in the O-glycosylation of FLA4/SOS5 result in root tip swelling caused by abnormalities in cellulose synthesis (Basu et al. 2015). Similarly, the elongation of cotton fibers, which are trichomes that have differentiated from the outer integuments of the ovule, is inhibited by treatment with βGlcY reagent (Huang et al. 2013). These results suggest that the hydroxyproline-O-glycosylation of FLAs plays a role in cell elongation.

CcFLA16 and CcFLA17 have an N-glycosylation motif in one of the two FAS1 domains (Fig. 7B). The results of recent studies suggest that protein N-glycosylation plays a pivotal role in the intracellular trafficking of proteins that are in cell walls and plasma membranes (Yamamoto et al. 2014, von Schaewen et al. 2015). Although we did not investigate whether CcFLA16 and CcFLA17 proteins are N-glycosylated, N-glycosylation of these proteins may facilitate proper targeting to the hyphal surface.

The presence of a YH-motif, which is involved in human cell adhesion, in CcFLA17 suggests that the FLA is involved in the adhesion of hyphae to host cells. Interestingly, CcFLA17 expression was also detected in holdfast cells, which are C. campestris stem epidermal cells that contact host stem epidermal cells (Fig. 5), suggesting that the FLA may be involved in the adhesion of stem epidermal cells, and may keep the two plants in contact once haustorium intrusion has been initiated. In contrast to Cuscuta spp., a root obligate parasitic plant, Orobanche aegyptiaca, does not accumulate AGP and/or pectin in the haustorial tip or holdfast (Supplementary Fig. S2). Since soil particles counteract the growth of the intrusive organ, O. aegyptiaca may not need adhesive material that prevents pushing the parasite and host away during the intrusive phase (Heide-Jorgensen 2008). Alternatively, AGP accumulation in hyphae might be a phenomenon specific to Cuscuta. In addition to cell elongation and adhesion, another possible function of hyphal-expressed CcFLA genes is involvement in the perception of extracellular signals, as has been proposed for the FLA4/SOS5/FEI1–FEI2 pathway (Basu et al. 2016). The perception of extracellular signals from host cells may stimulate cell differentiation in the haustorium.

The involvement of AGPs in various biological processes in plants has frequently been documented (Ellis et al. 2010); however, their mode of action has not been elucidated. To rectify this, we are currently developing RNAi plants to investigate the mechanisms underlying hyphal-expressed CcFLA genes.

Materials and Methods

Plant materials

The host plants, G. max and Nicotiana tabacum, were grown in soil (Sukoyaka-baido, Yammar) mixed with the same volume of vermiculite (GS30L, Nittai) under a 16/8 h light/dark cycle (Plant-lux, FL40S-BRN) at 25°C. Plants were grown with Hyponex (S8038-1, HYPONeX JAPAN Co. Ltd.) diluted 1,000 times. Arabidopsis thaliana (Col-0) was grown on vermiculite under a 16/8 h light/dark cycle (Plant-lux) at 22°C. Plants were grown with Hyponex diluted 1,000-fold.

Cuscuta japonica was grown and parasitized to G. max as described previously (Ikeue et al. 2015). The interface region containing both C. japonica and G. max was harvested at five stages: 24, 48, 72, 96 and 120 haa.

Dry C. campestris seeds were dipped in concentrated sulfuric acid for 30 min, washed three times with distilled water, plated on wet glass-fiber paper and grown in the dark at 25°C for 3 d. Three-day-old seedlings were transferred to vermiculite and grown under a 16/8 h light/dark cycle at 25°C for 1 d. Parasitism was induced by attaching the subapical region of C. campestris to the stem of N. tabacum or A. thaliana and illuminating the junction with blue light (BC-BML4 λ_p = 450 nm, Biomedical Science K.K.) for 1 h. The plants were then kept in darkness for 24 h before being grown under a 16/8 h light/dark cycle (Plant-lux) at 25°C. Approximately 5 cm of the lateral branch of C. campestris, which had already parasitized N. tabacum, was excised, and parasitism was induced by attaching the excised lateral branch to an inflorescence stem of Arabidopsis and illuminating the junction with blue light for 1 h. After illumination, the plants were kept in darkness for 24 h, before being grown under a 16/8 h light/dark cycle (Plant-lux) at 25°C. The interface region containing both C. campestris and Arabidopsis was harvested at 24, 48, 72, 96 and 120 haa.

Seeds of O. aegyptiaca were provided by Dr. Joseph Hershenhorn (Volcani Center, Israel) and imported with permission by the Ministry of Agriculture, Forestry and Fisheries, Japan. Seeds were germinated and then parasitized to tomato (‘Micro-Tom’) root as described by Aly et al. (2011). ‘Micro-Tom’ seeds (TOMJPF00001) were provided by National BioResource Project Tomato (Shikata et al. 2016).

Sectioning

Infection sites were sampled in 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) (Wako Pure Chemical Industries), degassed for 15 min, and then incubated at 4°C for 1 d. Fixed samples were embedded in paraffin (Paraplast, Leica Biosystems) according to the conventional protocol (Ruzin 1999). Paraffin blocks were cut with a microtome (PR-50, Yamato Kohki) into 8–16 µm thick sections. The resulting paraffin ribbons were extended, attached to slides and dewaxed, as previously described (Mochizuki and Ohki 2015). The paraffin sections were used for immunostaining, Yariv staining and in situ hybridization.

Immunostaining

For immunostaining, sections that had been completely de-paraffinized were rinsed with PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na₂HPO₄ and 1.47 mM KH₂PO₄) at room temperature for 10 min, and then blocked with a blocking reagent of 1% (w/v) bovine serum albumin (BSA; Nacalai Tesque) dissolved in PBS at room temperature for 1 h. The sections were then incubated with a primary antibody (LM2, LM6, LM15, LM19 or LM20 from Plant Probes, and CCRC-M1 from the University of Georgia) diluted 1 : 10 in blocking solution at room temperature for 1 h. The antibodies used and epitopes detected are presented in Table 1. The sections were washed three times with PBS and incubated in a 1 : 100 dilution of a blocking reagent of anti-rat IgG antibody (Alexa Flour^® 488 goat anti-rat IgG; Invitrogen) at room temperature for 1 h. The sections were then washed three times with PBS to remove the remnants of unbound antibodies and counterstained with 10 ng ml^–1 calcofluor white (Sigma-Aldrich). They were then observed using a laser scanning confocal microscope (Leica TCS SP8, Leica Biosystems). Normal rat IgG (Santa Cruz Biotechnology) was used as a control.

Yariv reagent staining

Yariv solution was prepared by dissolving 2 mg of βGlcY (Biosupplies Australia Pty) in 1 ml of 0.15 M NaCl. The βGlcY solution was applied to the sections for 1 h at room temperature. αGalY reagent (Biosupplies Australia Pty) was used as a control. The sections were then observed under a light microscope (BX53-33-PH, Olympus).

Prioritization and cloning of C. campestris genes that encode AGPs

Sequences of 11 C. japonica AGP contigs, which are up-regulated during the parasitic process (Ikeue et al. 2015), were compared with nucleotide sequences of contigs of C. pentagona (Ranjan et al. 2014) which is closer to C. campestris than C. japonica using BLASTN (e-value < 1 e-50). Additionally, expression profiles were inspected to find C. pentagona AGP contigs whose expression levels were higher in the pre haustorium or haustorium than in the seedling or stem. By this search, 19 C. pentagona contigs were selected, and C. campestris cDNAs were cloned by using sequences of the C. pentagona contigs reported previously (Supplementary Table S1; Ranjan et al. 2014). For cDNA isolation, parasitic junction tissues that were frozen in liquid nitrogen were homogenized using a pestle and mortar and pre-cooled to –80°C. Total RNA was extracted using a Qiagen RNeasy Plant Mini Kit (Qiagen), according to the manufacturer’s protocol. cDNA was synthesized using a ReverTra Ace^® quantitative PCR (qPCR) reverse transcription kit (Toyobo) and an oligo(dT) primer (Toyobo). Using gene-specific primers that were designed based on the sequences of the 17 candidates (Supplementary Data S3), partial sequences of target genes were amplified, extracted from agarose gel using the Wizard^® SV Gel and PCR Clean-Up System (Promega) and cloned into a pCR^™-Blunt II-TOPO^® vector using a Zero Blunt^® TOPO^® PCR Cloning Kit (Thermo Fisher Scientific).

In situ hybridization

Partial CcFLA sequences in pCR^™-Blunt II-TOPO^® vectors were used to produce an RNA probe for in situ hybridization. Probe synthesis and hybridization were performed as described previously (Mochizuki and Ohki 2015). Primers used to produce probes are listed in Supplementary Data S3.

Analysis of the domain structures of CcFLA proteins

Amino acid sequences that had been deduced from DNA sequences of CcFLA genes in C. campestris (17 genes), C. japonica (11 genes) and A. thaliana (30 genes) were subjected to phylogenetic analysis using MEGA7 (Kumar et al. 2016). Phylogenetic relationships were inferred using the maximum likelihood method based on the JTT matrix-based model (Jones et al. 1992). A bootstrap consensus tree was inferred from 500 replicates. Amino acid sequences deduced from cDNA sequences of five CcFLA transcripts that were expressed in the hyphae were subjected to domain structure analysis using the sequence analysis module in SMART (http://smart.embl-heidelberg.de/; Schultz et al. 1998). The GPI anchor signal was searched for using big-PI Predictor (http://mendel.imp.ac.at/gpi/gpi_server.html; Eisenhaber et al. 2003), and amino acid sequence alignment of the FAS1 domains was performed using MEGA7 (Kumar et al. 2016).

Supplementary data

Supplementary data are available at PCP online.

Funding

This study was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT, Japan) [Scientific Research in Innovative Areas project ‘The Plant Cell Wall as an Information Processing System’ grant No. 15H01237 to K.A. and grant Nos. 24114001 and 24114005 to K.N.].

Abbreviations

Abbreviations
 
• AGP

  arabinogalactan protein

 
• FLA

  fasciclin-like arabinogalactan protein

 
• αGalY

  α-galactosyl Yariv reagent

 
• βGlcY

  β-glucosyl Yariv reagent

 
• GPI

  glycosylphosphatidylinositol

 
• haa

  hours after attachment

 
• PBS

  phosphate-buffered saline

 
• RNAi

  RNA interference

Footnotes

Footnotes
 
• The nucleotide sequences reported in this paper has been submitted to the DNA Data Bank of Japan under the following accession numbers: CcFLA1, LC269164; CcFLA2, LC269165; CcFLA3, LC269166; CcFLA4, LC269167; CcFLA5, LC269335; CcFLA6, LC270788; CcFLA7, LC269168; CcFLA8, LC269336; CcFLA9, LC269337; CcFLA10, LC269169; CcFLA11, LC269170; CcFLA12, LC269171; CcFLA13, LC269338; CcFLA14, LC269339; CcFLA15, LC270789; CcFLA16, LC270790; CcFLA17, LC269172

Acknowledgments

We thank Dr. Kyoji Yamada (Toyama University) and Dr. Kenji Matsui (Yamaguchi University) for providing us with the C. japonica seeds, Dr. Joseph Hershenhorn (Volcani Center, Israel) for providing us with O. aegyptiaca seeds, and National BioResource Project Tomato (AMED, Japan) for providing us with ‘Micro-Tom’ seeds. We also thank Dr. Junpei Takano and Dr. Akira Yoshinari (Osaka Prefecture University) for the confocal laser canning microscopy.

Disclosures

The authors have no conflicts of interest to declare.