Roots of potato (Solanum tuberosum cv. Desiree), soybean (Glycine max cv. Toliman) and three spring wheat cultivars (Triticum aestivum cv. Bobwhite, Cadenza, and Fielder) were infected with freshly hatched J2s of G. pallida, H. glycines, and CCNs (H. avenae and H. filipjevi) respectively.

Individual potato chits were removed from sprouting tubers and transferred to growth pouches (Mega International, St Louis Park, MN, United States). Soybean and wheat seedlings were similarly transplanted into pouches after germination on moist filter paper as described previously (Urwin et al., 2002). Cultivation was then carried out at 20°C (potato and wheat) or 25°C (soybean) under a 16 h/8 h photoperiod in a growth chamber (Sanyo MLR). Wheat seedlings were supplied with full-strength Hoagland nutrient solution (MP Biomedicals, Europe) once the third leaf was about to emerge. Water was added daily to replace evaporation losses and the nutrient solution was replaced after the first 2 weeks then subsequently once per week.

Second-stage juveniles of G. pallida and H. glycines were hatched from cysts in host root exudate at 20 or 25°C respectively and collected as described (Urwin et al., 1995). H. avenae and H. filipjevi cysts were rinsed in sterile tap water and stored in a 1.5 mL tube at 4°C for 1 month before hatching. Cysts were then placed into a sterile hatching jar and incubated at 10 or 4°C in the dark. Newly hatched J2s were collected every 2–3 days and could be stored at 4°C until required. At approximately 7 days after transfer to growth pouches, selected root tips of all plants were each inoculated with 20 J2s of the compatible nematode species and covered by a piece of GF/A paper (Sigma). Filter papers were removed 24 h post-inoculation. Infected and comparable non-infected regions of roots were excised at different time points post-infection.

Rat monoclonal antibodies used in this study were: LM10 and LM11 which bind to heteroxylan (McCartney et al., 2005), LM12 to feruloylated heteroxylan (Pedersen et al., 2012), LM28 to glucuronosyl-containing heteroxylan (Cornuault et al., 2015), LM15 to XXXG motif of xyloglucan (Marcus et al., 2008) and LM25 to XXXG/galactosylated xyloglucan (Pedersen et al., 2012), LM19 directed against low/non methyl-esterified HG (Verhertbruggen et al., 2009) and LM20 to highly methyl-esterified-HG (Verhertbruggen et al., 2009), LM5 to (1–4)-β-D-galactan (Jones et al., 1997), LM6 to (1–5)-α-L-arabinan (Willats et al., 1998), JIM20 to extensins (Smallwood et al., 1994; Knox et al., 1995) and MLG (Meikle et al., 1994).

Lengths of root harboring an established parasitic nematode and its associated syncytial feeding site were excised. Fixation and embedding together with subsequent sectioning and in situ analysis, were carried out using a described method (Davies et al., 2012). Serial transverse sections were collected from the mid-point of each syncytium to obtain the optimal size for downstream labeling and analysis. Any set of root sections that clearly harbored more than one syncytium was discarded. A minimum of five independent syncytia were sectioned and analyzed for each experiment with at least 24 technical replicate sections observed for each antibody. In order to eliminate background autofluorescence in wheat sections, a 5 min incubation using 0.1% Toluidine Blue O (pH 5.5, 0.2 M sodium phosphate buffer) was carried out after the immunolabeling steps. For bright field optical images, equivalent sections were stained with Toluidine Blue O solution (1% Toluidine Blue O dissolved in 1% sodium borate aqueous solution and filtered) for 5–10 min at room temperature then excess dye was washed off.

For unmasking of the LM25 xyloglucan epitope, pectic HG was enzymatically degraded with pectate lyase. Sections were first treated with 0.1 M sodium carbonate (pH 11.4) for 2 h at room temperature; washed with deionised water, then incubated in pectate lyase (Aspergillus sp.; Megazyme International, Ireland) at 25 μg/ml in 50 mM CAPS buffer with 1 mM CaCl₂ (pH 10) for 2 h. Following three washes in deionised water immunolabeling was performed, as described in Davies et al. (2012).

Immunolabeled sections were observed under a Leitz DMRB Fluorescence Microscope [Leica Microsystems (UK) Ltd] and images were taken by QImaging QICAM digital camera (QImaging, Canada) using Q-capture pro software (QImaging, Canada). Further necessary image editing and composition was carried out using CorelDRAW X7 (Corel Corporation, Canada) and PaintShop Pro X7 (Corel Corporation, Canada). Fluorescence intensity measurements at 30–70 locations per cell type per antibody were carried out in ImageJ for the walls of the major host root cell types, including xylem, phloem elements, cortex and epidermis, as well as syncytial walls. For each antibody, the mean normalized value for each cell type was compared by one-way ANOVA and significant differences between syncytial walls and those of other cells were determined following Tukey’s multiple comparisons test.

Cell wall architectures of nematode-induced syncytia in dicotyledonous plants were analyzed using potato roots infected with G. pallida (Figure 1 and Supplementary Figure S1) and soybean roots infected with H. glycines (Figure 2 and Supplementary Figure S2), both at 14 days post-inoculation (14 dpi). Structural changes within the root associated with nematode parasitism were revealed by toluidine blue O staining of sectioned, fixed roots, and syncytial regions were outlined in bright-field images of each figure to aid interpretation. The xylan antibody LM11, which in dicots binds specifically to secondary cell walls, was used to visualize xylem vessels within vascular cylinders in both potato (Figure 1B and Supplementary Figure S1B) and soybean (Figure 2B and Supplementary Figure S2B) roots. This, together with the bright-field staining, allowed orientation of key root structures in spite of the gross changes in morphology that occurred during syncytium formation. The lower magnification images provided in Supplementary Figures S1, S2 allow the features of the syncytium to be viewed in context with the surrounding root cells.

The distribution of xyloglucan, the major non-cellulosic polysaccharide in dicots, was revealed mainly by the anti-xyloglucan probe LM25 in equivalent transverse sections through the syncytia of infected potato (Figures 1C,D and Supplementary Figures S1C,D) and soybean (Figure 2E and Supplementary Figure S2E) roots. The cell walls of syncytia induced by both G. pallida and H. glycines contained xyloglucans and the removal of pectic HG using pectate lyase was required to fully unmask the LM25 epitope in syncytial walls of potato (Figure 1D) similar to previous reports (Marcus et al., 2008; Davies et al., 2012). A further two probes which recognize different epitopes in xyloglucans (LM15 and LM24) also bound (data for soybean shown in Figures 2C,D, data for potato shown in Supplementary Figure S3) although the strongest binding was observed using LM25.

The presence of pectic HG, the major pectic polymer, was visualized using mAbs LM19 and LM20 (Figures 1E–G, 2F–H and Supplementary Figures S1E–G, S2H,I), directed toward low/non methyl-esterified HG and highly methyl-esterified HG respectively. Differential binding of these two antibodies together with the effect of Na₂CO₃ pre-treatment to remove methyl-esters suggested that the syncytial cell wall of both potato and soybean possessed abundant pectic HG that was heavily methyl-esterified. LM20 bound strongly to syncytial cell walls in untreated sections (Figures 1G, 2H and Supplementary Figures S1G, S2I) whereas LM19 only bound effectively following incubation with Na₂CO₃ (Figures 1F, 2G and Supplementary Figure S1F). Strikingly different binding patterns of LM6 and LM5, recognizing 1,5-α-arabinan and 1,4-β-galactan side chains of pectic RG-I polysaccharides respectively, were observed in the two plant species. The LM6 arabinan epitopes were abundant within syncytial cell walls in both plants (Figures 1I, 2J and Supplementary Figures S1I, S2G) while the LM5 galactan epitopes were barely detectable in walls of syncytia formed in potato (Figure 1H and Supplementary Figure S1H) and absent from those in soybean (Figure 2I and Supplementary Figure S2F).

In addition to polysaccharides, plant cell walls also possess various structural proteins. The JIM20 antibody-targeted epitope of extensin, a type of hydroxyproline-rich glycoprotein (HRGP), was detected at a low level in syncytial cell walls of soybean, especially the internal walls (Figure 2K and Supplementary Figure S2J) but was absent from walls of syncytia induced in potato (Figure 1K and Supplementary Figure S1K). Analysis of cell surface arabinogalactan-proteins (AGPs) using JIM13 showed that epitopes for this antibody were present within syncytial elements of potato (Figure 1J and Supplementary Figure S1J) but absent from those of soybean (data not shown).

At 14 dpi, the syncytia induced by female cyst nematodes have reached their maximum size (Urwin et al., 1997). The characteristics of cell walls during the extensive cell expansion and incorporation phase of syncytial formation may differ from those of a fully developed syncytium. Therefore we also analyzed sections of potato roots infested with juvenile G. pallida at 7 dpi when the syncytia were clearly less developed and still undergoing expansion (Supplementary Figure S4). The binding patterns of the monoclonal antibodies (Supplementary Figure S4) were very similar to those observed at 14 dpi (Figure 1 and Supplementary Figure S1). This suggests that the syncytial cell walls maintain a consistent structural composition during the course of syncytium formation and nematode development.

The overall binding patterns of the various mAbs to syncytial cell walls within the context of the entire root section were also analyzed and shown in Supplementary Figure S1 for potato whole root and Supplementary Figure S2 for the soybean vascular cylinder. In comparison to the other root cells of the vascular cylinder, the cyst nematode-induced syncytia were seen to have a distinctive cell wall chemical composition. For example, syncytial cell walls lack xylan compared to host xylem elements (Supplementary Figures S1B, S2B), and contain more abundant xyloglucan (Supplementary Figures S1C,D, S2C–E) but lack galactan (Supplementary Figures S1H, S2F) compared to the phloem elements. The differential abundance of xyloglucan, galactan, arabinan, and methyl-esterified HG was investigated in more detail for the walls of syncytia, xylem, phloem, cortical, and epidermal cells in potato roots (Figure 3). A fluorescence quantification method was applied over the G. pallida-infested potato root sections, and the results show the clearly distinct nature of syncytial walls from walls of other major root cell types (Figure 3). Cell walls of syncytia contained significantly more xyloglucan detected by LM25 (Figure 3A) and arabinan detected by LM6 (Figure 3C) than those of any other cell type analyzed. The galactan content of the syncytium cell walls was significantly lower than for cortex or phloem (Figure 3B) but significantly higher than for either xylem vessels or epidermal cells. Of the four components analyzed, the widely present methyl-esterified HG was most similar between walls of the syncytium and other cell types (Figure 3D), however it was still significantly more abundant in the syncytium than either phloem or cortical cells. No other cell type in the potato roots displayed an identical wall composition to syncytia, confirming the unique nature of the syncytium.

The globally important wheat crop can be severely affected by CCNs (Nicol et al., 2011). We carried out a comprehensive analysis of cell wall structure encompassing syncytia induced by the two most economically important species H. avenae and H. filipjevi in three spring wheat cultivars (Bobwhite, Cadenza, and Fielder) at different stages of the infection process. Roots were infected with J2s of H. avenae and H. filipjevi, and samples were collected when similar stages of adult females of both nematode species were observed on the root surface (21 dpi for H. avenae and 28 dpi for H. filipjevi).

Heteroxylans are the major cross-linking glycans in PCWs of grasses. Three monoclonal antibodies (LM10, LM11 and LM12) that recognize unsubstituted xylan, unsubstituted xylan/arabinoxylan and feruloylated xylan respectively were used to reveal both the presence and the substitutions of xylans in syncytial cell walls. Antibody binding patterns revealed the absence of heteroxylan recognized by LM10 (Figures 4B,L and Supplementary Figures S5B,L, S6B,L) and LM11 (Figures 4C,M and Supplementary Figures S5C,M, S6C,M) within syncytial cell walls. Strong binding was observed for LM12 (Figures 4D,N and Supplementary Figures S5D,N, S6D,N), suggesting that syncytial cell wall xylans are extensively substituted with ferulic acid. The xyloglucan epitopes recognized by LM15 and LM25 were present within syncytial walls induced by both nematode species, with that of LM25 relatively more abundant (Figures 4F,G,P,Q and Supplementary Figures S5F,G,P,Q, S6F,G,P,Q). MLG is mainly restricted to the grass family and was evident in a range of cell types in the root sections. However, whilst it was present in the cell walls of syncytia induced by H. filipjevi (Figure 4O and Supplementary Figures S6E,O), it was almost entirely absent from H. avenae-induced syncytia (Figure 4E and Supplementary Figures S5E,O).

The prevalence of pectic polysaccharides in the syncytial cell walls of wheat was analyzed using LM19, LM20, LM5, and LM6. In contrast to the syncytia formed in potato and soybean, only trace levels of methyl-esterified HG were observed, indicated by the slight binding of LM20 (Figures 4H,R and Supplementary Figures S5H,R, S6H,R). However, this was clearly due to a lack of HG, rather than an altered methyl-esterification status, as the LM19 epitope was entirely absent (data not shown). Low levels of galactan were detected with LM5 (Figures 4I,S and Supplementary Figures S5I,S, S6I,S) whilst arabinan epitopes detected by LM6 were much more abundant (Figures 4J,T and Supplementary Figures S5J,T, S6J,T).

Importantly, no differences in antibody binding patterns (for instance, abundant vs. absent) were observed between cell walls of syncytia induced by the same nematode species in the three wheat cultivars. Therefore, for investigation of temporal changes in cell wall composition during nematode and syncytium development, sections of syncytia induced by H. avenae and H. filipjevi in only the cultivar Bobwhite were analyzed. The antibody binding was generally stable in syncytia induced by the same nematode species among different stages (Supplementary Figure S7 for H. avenae and Supplementary Figure S8 for H. filipjevi) and syncytia induced by both nematode species shared high similarities in cell wall chemical composition, apart from LM5 and MLG (Figure 5). The feruloylated heteroxylan in syncytial walls also contained glucuronic acid decorations (indicated by LM28, data for H. avenae shown in Supplementary Figure S7), indicating the abundance of xylan substitutions in walls of syncytia induced by both nematode species. The LM5 galactan epitope was largely absent from H. filipjevi-induced syncytial walls among all the stages analyzed (Figures 5Q–T) whilst shown to be present in H. avenae-induced syncytia at certain stages (Figures 5E–H). The previously observed absence of MLG in syncytia of H. avenae and presence in those of H. filipjevi MLG was most likely not related to differences in syncytial development. MLG was barely detectable in syncytia of H. avenae at a range of developmental stages (Figures 5I–L) whilst it remained present in all syncytial stages of H. filipjevi (Figures 5U–X).

The cell walls of syncytia induced by both H. avenae and H. filipjevi in wheat roots were rich in highly substituted heteroxylan and arabinan and also contained xyloglucan. Syncytia induced by the same nematode species had generally conserved cell wall chemical compositions, both among different host cultivars and among various nematode developmental stages within the same cultivar. Apart from the different binding of pectic galactan and MLG, syncytia of both nematode species in wheat possess similar cell wall microstructures. As analyzed using the same host at a series of developmental stages, those differences of galactan and MLG are more likely related to the nematode species although the precise reasons are yet to be elucidated.

Xyloglucan has been found in almost every land plant species analyzed and is the most abundant hemicellulose in primary walls of spermatophytes except for grasses (Scheller and Ulvskov, 2010; Pauly et al., 2013). It was detected in the cell walls of all syncytia analyzed, irrespective of host or nematode species although, consistent with its distribution in the plant kingdom, it was more abundant in syncytia within potato and soybean roots than those formed in wheat roots. The function of xyloglucan in the plant cell wall is described by the load-bearing xyloglucan/cellulose framework model: increasing xyloglucan tethers between the cellulose microfibrils causes increased rigidity of the cell wall, while degradation of these tethers causes the walls to loosen (Pauly et al., 2013). The dynamic changes of xyloglucan cross-linking, catalyzed by enzymes such as xyloglucan endo-transglycosylase/hydrolase (XTH), could mediate the cell wall restructuring that is required during syncytium formation. For instance, the expression of Arabidopsis XTH9 and an endo-xyloglucan transferase XTR6 were both up-regulated in syncytia of H. schachtii, whilst XTR9 was down-regulated (Szakasits et al., 2009), reflecting the reconstruction of host cell walls during syncytium formation and functioning. Walls of syncytia in soybean also possess abundant xyloglucan, however, in this case a homolog of Arabidopsis XTR6 was found to be slightly downregulated during the interaction with H. glycines (Ithal et al., 2007a). Several xyloglucan endo-transglycosylase (XET) genes were highly upregulated at the early stage of syncytium induction by H. glycines, with transcripts of one gene localized in 5-day old syncytia (Ithal et al., 2007b). XET activity can also be involved in deposition of new wall material (Nishitani and Tominaga, 1992), therefore the dynamic change of this network, breaking down or generating new cross-links, may be regulated as the nematode develops.

Xylans, a diverse group of glycans, are the dominant non-cellulosic polysaccharide in the secondary cell walls of dicotyledonous plants. They are commonly substituted with glucuronosyl residues to form glucuronoxylans and previous research has shown that syncytia formed in Arabidopsis roots possess no secondary cell walls and correspondingly contain no xylans (Davies et al., 2012; Wieczorek et al., 2014). Similarly, no xylan epitopes were detected within syncytial cell walls in potato and soybean roots. Conversely, xylans are abundant in primary walls of commelinid monocots such as Poales including wheat, rice and maize, and usually contain many arabinose residues attached to the backbone, forming arabinoxylans (AXs) and glucuronoarabinoxylans (GAXs) (Buanafina, 2009; Scheller and Ulvskov, 2010). The AXs serve an important role by cross-linking cellulose microfibrils as well as oxidatively linking with each other. The walls of growing cells predominantly contain highly substituted AXs rather than the less branched xylans (Buanafina, 2009). Syncytial cell walls in wheat contained no un-/low-substituted xylans but a small amount of AX/GAX. The strong binding of LM28, targeting glucuronosyl-containing heteroxylans (Cornuault et al., 2015) also suggested high substitution of heteroxylans in walls of syncytia formed in wheat. Meanwhile LM12, derived to recognize feruloyl residues attached to a range of sugars (Pedersen et al., 2012), bound strongly within wheat syncytial cell walls, most likely to feruloylated xylans. Together, this indicates the importance of xylan substitution in both formation and function of syncytia in wheat. Ferulate esters can be oxidatively cross-linked in a variety of ways, potentially causing cell wall stiffening and reduced growth and expansion. Feruloylation is also correlated with cell wall degradability (Buanafina, 2009). Cell wall integrity was compromised in transgenic plants expressing a fungal feruloyl esterase AnFAE that caused a reduction of ferulic acids and the plants were more susceptible to fungal pathogens (Reem et al., 2016). A similar study has also shown the link between the expression of fungal ferulic acid esterase and cell wall digestibility (Badhan et al., 2014). The abundance of xylan substitutions in syncytial cell walls of wheat may indicate that various cross-links are formed to maintain syncytial wall integrity and provide mechanical strength. The reported near absence of heteroxylans in the walls of syncytia formed by H. avenae in barley, a host also with heteroxylan-rich primary walls (Aditya et al., 2015) may result from the use of only anti-xylan LM11 in that study. Further analysis with a wider range of probes that recognize substituted forms of xylan, would likely be informative.

Pectic polysaccharides, of which HG is the major group, are abundant in plant cell walls, comprising as much as 30% of dicot, gymnosperm, and non-Poales monocot cell walls, but considerably less in cell walls of grasses (Caffall and Mohnen, 2009). Homogalacturonan was abundant and predominantly methyl-esterified in the cell walls of syncytia induced by G. pallida and H. glycines. Interestingly, methyl-HG was detectable at a low level within several syncytial wall samples of wheat, despite that host generally lacking abundant pectin in PCWs. Thus, the methyl-esterification status of HG in cell walls of syncytia formed in potato, soybean and even in wheat is similar to that previously described for mature syncytia in Arabidopsis (Davies et al., 2012; Wieczorek et al., 2014) and barley (Aditya et al., 2015), although HG was reported to be largely unmethylesterified at very early stages of syncytium formation (Wieczorek et al., 2014).

Pectin HGs are highly methyl-esterified when initially deposited into cell walls and can subsequently be de-esterified by the action of pectin methylesterases (PMEs) to facilitate the formation of HG-calcium complexes, the so called ‘egg-box’ model (Caffall and Mohnen, 2009; Wolf et al., 2009). The presence of the HG-calcium structure is postulated to induce pectic gel formation and thus cause cell wall stiffening, therefore the abundant methyl-HG might contribute to wall flexibility required during nematode feeding (Muller et al., 1981; Bohlmann and Sobczak, 2014). The rheological properties of the syncytium cell wall, including its porosity and extensibility, could be modulated by spatially distributed modifying enzymes such as PMEs and PMEIs (PME inhibitors) (Wolf et al., 2009). A secreted cellulose-binding protein (HsCBP) from the nematode H. schachtii specifically interacts with Arabidopsis PME3 to facilitate cyst nematode parasitism. Its expression peaks at the parasitic J3 stage, suggesting a role during the early phase of syncytium formation (Hewezi et al., 2008). However, both PMEs and PMEIs are large gene families with 67 PMEs (PF01095, EC 3.1.1.11) in Arabidopsis (Lombard et al., 2014) and 69 PMEIs (Saez-Aguayo et al., 2013). Their complex activities and interactions in the syncytia are likely to be responsible for regulating the abundant pectin methyl-HG in the syncytial walls, thus facilitating nematode parasitism. Further investigations should be made to elucidate the methyl-HG accumulation mechanism in syncytial walls as well as its impact on syncytium mechanical properties, especially considering the innate complexity of the role of possible pectin cross-links (Peaucelle et al., 2012).

The high abundance of RG-I pectic arabinan side chains was a striking common feature of all syncytial walls assessed, regardless of plant host or nematode species. This implies a conserved and important role, most likely to be in maintaining cell wall flexibility as reported for other specialized plant cell types. High arabinan content helps to maintain flexibility in guard cell walls (Jones et al., 2003), and is also implicated in the response of other plant cells facing water loss and therefore a change in cell volume (Moore et al., 2006, 2008). Due to the large volumes withdrawn by the feeding nematodes (Muller et al., 1981) and the high turgor pressure recorded inside syncytia (Böckenhoff and Grundler, 1994), in addition to the fluctuation in size during nematode development, such wall flexibility seems to be essential for general syncytial function.

In contrast to the abundant arabinan side chains, pectic galactan (indicated by LM5) was found to be absent from soybean root syncytia, as also shown previously for Arabidopsis (Davies et al., 2012), and was present at a very low level in walls of syncytia formed in potato and wheat. The higher natural abundance of galactan in potato (Oxenbøll Sørensen et al., 2000) may account for the difference among the dicot hosts. Pectic galactan may also be related to mechanical rigidity of cell walls (McCartney et al., 2000) although in this case, presence of pectic (1→4)-β-D-galactan correlates with firmer plant tissues. Therefore the relative absence of galactan residues is also likely to contribute to syncytial cell wall flexibility.

As a nutrient sink and the sole supply of food to the cyst nematode, the syncytium has a generally stable cell wall composition of complex polysaccharides during the nematode life cycle to maintain its function throughout parasitism. The substitutions and modifications of those cell wall polymers indicate that the syncytial wall is a dynamic, flexible structure, capable of fulfilling various requirements in plant–nematode interactions. Cell walls of syncytia are clearly distinct from those of surrounding root cells. Nevertheless, as a fusion of remodeled host plant cells, syncytial cell walls also possess key characteristics of their host plant cell walls: syncytia in potato, soybean and previously Arabidopsis hosts with dicot-type primary walls, contain abundant pectins and xyloglucans while syncytial cell walls in the commelinid monocot host wheat, possess large amounts of substituted heteroxylans, small amount of xyloglucans and very few pectic polysaccharides. This reflects the host-specific adaptations in syncytial formation, involving the use of existing cell wall synthesis mechanisms in host roots.