Root architecture plasticity in response to endoparasitic cyst nematodes is mediated by damage signaling

Summary Plant root architecture plasticity in response to biotic stresses has not been thoroughly investigated. Infection by endoparasitic cyst nematodes induces root architectural changes that involve the formation of secondary roots at infection sites. However, the molecular mechanisms regulating secondary root formation in response to cyst nematode infection remain largely unknown. We first assessed whether secondary roots form in a nematode density‐dependent manner by challenging wild‐type Arabidopsis plants with increasing numbers of cyst nematodes (Heterodera schachtii). Next, using jasmonate‐related reporter lines and knockout mutants, we tested whether tissue damage by nematodes triggers jasmonate‐dependent secondary root formation. Finally, we verified whether damage‐induced secondary root formation depends on local auxin biosynthesis at nematode infection sites. Intracellular host invasion by H. schachtii triggers a transient local increase in jasmonates, which activates the expression of ERF109 in a COI1‐dependent manner. Knockout mutations in COI1 and ERF109 disrupt the nematode density‐dependent increase in secondary roots observed in wild‐type plants. Furthermore, ERF109 regulates secondary root formation upon H. schachtii infection via local auxin biosynthesis. Host invasion by H. schachtii triggers secondary root formation via the damage‐induced jasmonate‐dependent ERF109 pathway. This points at a novel mechanism underlying plant root plasticity in response to biotic stress.


Introduction
Plants utilize root plasticity as a key strategy to survive in a changing soil environment. Remodeling of root systems allows plants to cope with nutrient deficiencies, drought, salinity, and other abiotic stresses (Koevoets et al., 2016). However, little is known about root architecture plasticity in response to soil-borne biotic stresses. Infections by cyst nematodes are known to induce elaborate root architectural changes in host plants. Secondary roots form locally at cyst nematode infection sites (Grymaszewska & Golinowski, 1991;Goverse et al., 2000;Lee et al., 2011). Furthermore, the ability to form secondary roots in response to nematode infection can result in better maintenance of shoot growth in some potato and soybean cultivars (Trudgill & Cotes, 1983;Miltner et al., 1991). Nevertheless, the molecular mechanisms regulating secondary root formation in response to belowground herbivory are not well-understood.
Cyst nematodes are microscopic root endoparasites that cause large agricultural losses world-wide. These nematodes can persist in the soil in a dormant state for many years (Jones et al., 2013). Exudates from host roots trigger hatching of dormant secondstage juveniles (J2s) and guide their migration to the root surface. Here, the J2s penetrate the root epidermis of the differentiation or mature root zone by piercing plant cell walls with their needlelike oral stylet and by secreting plant cell wall degrading enzymes (Bohlmann & Sobczak, 2014). Subsequently, juveniles migrate intracellularly within the cortex, leaving behind a trail of destruction (Wyss & Zunke, 1986;Grundler et al., 1994). Plant cell wall fragments released during nematode migration can act as damage-associated molecular patterns triggering defense signaling in the host (Shah et al., 2017). Nematode migration also activates biosynthesis and signaling of the defense hormone jasmonate (JA) (Kammerhofer et al., 2015). Upon successful arrival at the vascular cylinder, cyst nematodes utilize stylet-secreted effectors to manipulate plant developmental pathways to transform host cells into permanent feeding sites (Gheysen & Mitchum, 2011). Together with permanent feeding site development, multiple de novo formed secondary roots emerge in clusters at nematode infection sites (Grymaszewska & Golinowski, 1991;Goverse et al., 2000;Lee et al., 2011).
Nematode feeding sites are characterized by the local accumulation of the plant hormone auxin (Karczmarek et al., 2004;Grunewald et al., 2009). Auxin transport and auxin-insensitive Arabidopsis mutants infected by cyst nematodes show smaller females and smaller feeding sites, respectively (Goverse et al., 2000;Grunewald et al., 2009). Additionally, auxin is an important regulator of secondary root formation. Oscillations of auxin maxima at the root tip determine the formation of lateral roots in a regularly spaced pattern along the primary root (Fukaki & Tasaka, 2009). However, these oscillations are not required for the de novo formation of secondary roots. Ectopic induction of local auxin biosynthesis in pericycle cells via an inducible promoter is sufficient to trigger de novo secondary root formation (Dubrovsky et al., 2008). Auxin accumulation in multiple neighboring pericycle cells can lead to the formation of secondary root clusters (Dubrovsky et al., 2008). The spatial co-occurrence of nematode feeding sites and secondary root clusters often corresponds to overlapping regions of auxin accumulation (Karczmarek et al., 2004;Absmanner et al., 2013). This suggests that secondary roots could be induced as the sole consequence of the auxin that accumulates during nematode feeding site development (Goverse et al., 2000). Alternatively, damage caused by nematode infection might also lead to local auxin accumulation and secondary root formation.
Tissue damage triggers auxin accumulation and de novo root formation via the JA-dependent ERF109 transcription factor in leaf explants Chen et al., 2016;Hu & Xu, 2016;Zhang et al., 2019). Herein, JA accumulates at the site of wounding within a few hours of leaf detachment and triggers expression of the transcription factor ERF109 via the JA receptor COI1. ERF109 binds to the promoter of the auxin biosynthesis gene ASA1, which induces root formation in a process referred to as de novo root organogenesis. Direct interaction of JAZ proteins inhibits ERF109 expression in a negative feedback loop to avoid wound hypersensitivity . Sterile mechanical injury in primary roots of Arabidopsis can trigger auxin accumulation at the wounding site and subsequent secondary root formation (Sheng et al., 2017). However, whether this occurs via the same damage signaling pathway as de novo root organogenesis from leaf explants is unknown. Furthermore, mechanical injury is an artificial condition, and therefore it remains unclear whether the JA-dependent ERF109 pathway is involved in the regulation of secondary root formation also upon naturally occurring damage by herbivory or pathogen penetration.
Previously, we showed that components of the JA-dependent ERF109 pathway are induced by root-knot nematode (Meloidogyne spp.) infection . Differently from cyst nematodes, root-knot nematodes penetrate roots at the elongation zone and migrate toward the root apical meristem by moving in between cells. Although this type of migration creates minimal tissue damage, root-knot nematode invasion of the root apical meristem induces expression of the ERF109 transcription factor. This eventually promotes tissue regeneration and reduces the inhibitory effect of nematode infection on primary root growth . Thus, wound signaling can mediate primary root growth compensation in response to damage by stealthily migrating root-knot nematodes. However, further research is needed to understand whether JA-dependent wound signaling regulates root architectural changes to compensate for tissue destruction by the more damaging cyst nematodes in the differentiation and mature root zones.
In this study, we hypothesized that local tissue damage by cyst nematode host invasion causes secondary root formation at infection sites via the JA-dependent ERF109 pathway. By challenging Arabidopsis seedlings with increasing numbers of J2s of the beet cyst nematode Heterodera schachtii, we found that secondary root formation is induced at infection sites in a nematode densitydependent manner. With time course confocal microscopy of JA biosensors and ERF109 reporter lines in Arabidopsis, we provide evidence that secondary root formation is preceded by the transient and local JA-dependent expression of ERF109. Moreover, the nematode density-dependent increase in secondary roots is abolished in coi1-2 and erf109 knockout mutants. By selectively applying the auxin biosynthesis chemical inhibitor L-kynurenine (L-kyn) to shoots and roots, we further found that the ERF109mediated formation of secondary roots is dependent on local auxin biosynthesis. We therefore conclude that tissue damage by host-invading cyst nematodes induces secondary root formation by altering local auxin biosynthesis via the JA-dependent ERF109 pathway. Altogether, our results show that damage signaling via the JA-dependent ERF109 pathway regulates root architectural plasticity in response to cyst nematode infection.

Nematode sterilization
Heterodera schachtii (Woensdrecht population from IRS, the Netherlands) cysts were extracted from sand of Brassica oleracea infected plants as previously described (Baum et al., 2000) and incubated for 7 d in a solution containing 1.5 mg ml À1 gentamycin sulfate, 0.05 mg ml À1 nystatin, and 3 mM ZnCl 2 . Hatched J2s were purified by centrifugation on a 35% sucrose gradient and surface sterilized for 15 min in a solution containing 0.16 mM HgCl 2 , 0.49 mM NaN 3 , and 0.002% Triton X-100. After washing three times with sterile tap water, H. schachtii J2s were resuspended in a sterile 0.7% Gelrite (Duchefa Biochemie, Haarlem, the Netherlands) solution. A similar concentration of Gelrite solution was used as mock treatment.

Inoculation density-response curve
Individual Arabidopsis seeds were sown in square Petri dishes. Nine-day-old seedlings were inoculated with 0 (mock), 50, 100, 200, 350, or 500 H. schachtii J2s. Specifically, two 5 ll drops of solution (with J2s or mock) were pipetted at opposite sides of each seedling while keeping the Petri dishes vertical. This allowed for a homogeneous smear of J2s along the whole length of the root. At 7 d post-inoculation (dpi), scans were made of whole seedlings using an Epson Perfection V800 photo scanner (Epson, Nagano, Japan). The root architecture (total root length, primary root length, and total secondary root length) was measured using the WINRHIZO package for Arabidopsis (Regent Instruments Inc., Qu ebec, Canada). For the coi1-2 mutant, primary root length was measured manually because of the convoluted root system. The number of root tips was counted manually based on the scans. Furthermore, nematodes within the roots were stained with acid fuchsin and counted as previously described (Warmerdam et al., 2018). For comparisons between genotypes, the background effect of the mutation on the root architecture was corrected by normalizing each measured component in infected seedlings to the average respective component in mock-inoculated roots. Additionally, the presence of clusters and the number of secondary roots per cluster were scored using a binocular.

Histology and microscopy
Four-day-old Arabidopsis seedlings were inoculated with either 15 H. schachtii J2s or mock solution. The choice of using younger seedlings as previously done by Zhou et al. (2019) was made to reduce the damage inflicted to the seedling during sample preparation for microscopy. Root architecture was inspected using an Olympus SZX10 binocular with a 91.5 objective and 92.5 magnification (Olympus, Tokyo, Japan). Pictures were taken with an AxioCam MRc5 camera (Zeiss). For confocal and brightfield microscopy, single-nematode infection sites were selected for observation. For histochemical staining of b-glucuronidase (GUS) activity, seedlings were incubated in a GUS staining solution as previously described  for 4 h. Stained seedlings were mounted in a chloral hydrate clearing solution (12 M chloral hydrate; 25% glycerol) and inspected with an Axio Imager.M2 light microscope (Zeiss) via a 920 objective. Differential interference contrast images were taken with an AxioCam MRc5 camera (Zeiss). b-Glucuronidase saturation was quantified as previously described (Beziat et al., 2017) using FIJI software (Schindelin et al., 2012). For confocal laser scanning microscopy, seedlings were mounted either in water or in 10 lg ml À1 propidium iodide and imaged using a Zeiss LSM 710 system via 910 and 940 objectives. The wavelengths used were as follows: 600-640 nm for PI, 500-540 nm for GFP, 520-560 nm for YFP, and 590-680 nm for RFP. For pAOS::YFP N and JAS9-VENUS reporters, the fluorescent signal was imaged at the focal plane displaying the xylem vessels, where the nematode head is found. For the pERF109::GFP reporter, Z-stacks of six 13 lm slices were made of the entire root depth. Images were taken using ZEN 2009 software (Zeiss) and processed using the FIJI software. To make the fluorescence more visible, the brightness was enhanced for all the representative pictures in the same way using Adobe PHOTOSHOP 2021. Fluorescence intensity was quantified using the FIJI software. Specifically, the region of interest was selected using a set threshold, and then the integrated density was measured. Z-stacks were projected using the maximum intensity method.

Auxin biosynthesis inhibition
For split plate assays, we used the method described by Matosevich et al. (2020). For the L-kyn split plate assay, the four treatment combinations prepared were as follows: MM (modified Knop medium and 0.02% DMSO), KK (modified Knop medium, 10 lM L-kynurenine (Sigma-Aldrich), and 0.02% DMSO), MK (L-kyn only in the root), and KM (L-kyn only in the shoot). The Yucasin split plate assay is described in Supporting Information Fig. S1. Four-day-old Arabidopsis seedlings were inoculated with 15 H. schachtii J2s. Sixteen-hours post-inoculation (hpi), when J2s are still migrating through the root, seedlings were transferred to the treatment plates, so that the shoot and the hypocotyl were in contact with the medium in the upper half of the plate and the nematode-infected root was on the medium in the lower half of the plate. For microscopy, seedlings were collected at 3 dpi, and GUS staining was performed. For root architecture inspection, scans were made of whole seedlings at 7 dpi using an Epson Perfection V800 photo scanner. The total number of secondary roots per plant was counted based on the scans. Additionally, the presence of clusters and the number of secondary roots per cluster were scored using an Olympus SZX10 binocular.

Reverse transcription-quantitative real-time PCR
For reverse transcription-quantitative real-time PCR (RT-qPCR) analysis, several hundred root segments (c. 0.2 cm) harboring nematode infection sites or similar root segments of mockinoculated 12-d-old seedlings of Arabidopsis were collected at 12 hpi. Attention was paid not to include root tips and secondary root primordia. Subsequently, RNA extraction and qPCR were performed as previously described (Chopra et al., 2021;Hasan et al., 2022). ERF109 was amplified using the primers CTTAT GATCGAGCCGCGATT and TCCTCCGTTCCATTGCTC  (Cai et al., 2014;Zhou et al., 2019). Three independent biological replicates of the experiment were performed, with three technical replicates per each biological replicate. Relative expression of ERF109 was calculated based on the endogenous control 18 S rRNA (Pfaffl, 2001). The average ERF109 expression in the mock-inoculated wild-type roots of the first biological replicate was used as a reference to normalize the average expression in the other samples (Hasan et al., 2022).

Statistical analyses
Statistical analyses were performed using the R software v.3.6.3. The correlation between variables was calculated using Spearman rank-order correlation coefficient. Significance of the differences between samples was calculated as indicated in the figure legends. The confidence interval of the inoculum density-response curves was calculated by loess regression (as per default in geom_smooth) in R.

Heterodera schachtii infection induces local formation of secondary roots in a nematode density-dependent manner
To test whether tissue damage by invading nematodes in roots triggers the formation of secondary roots, we analyzed root branching upon penetration by increasing numbers of nematodes. We inoculated seedlings with 0, 50, 100, 200, 350, or 500 J2s of H. schachtii and counted both the number of nematodes that penetrated the roots and the total number of secondary roots at 7 dpi ( Fig. 1). Here, the total number of secondary roots in infected seedlings was normalized to the average respective number in uninfected roots. We found that the number of nematodes that penetrated the roots increased by inoculum density for up to 350 J2s per plant, whereafter it remained the same (Fig. 1a). Furthermore, we observed that the number of penetrated nematodes correlated positively with the total number of secondary roots per plant (Fig. 1b). Next, we investigated whether the clustering of secondary roots around nematode infections sites also correlates with the inoculum density ( Fig. 1c,d). For this, we challenged Arabidopsis with four inoculum densities (0, 50, 100, and 350) to establish an incremental increase in the number of nematode infection sites per plant. Infection sites were identified by the local discoloration of root tissue due to cell necrosis along the migratory tract of the nematode (Grundler et al., 1994). Roots were counted as clusters when more than one secondary root emerged in the proximity of an infection site. Also, we counted the number of secondary roots per cluster. We found that uninfected seedlings showed a typical pattern of lateral roots regularly distributed along the primary root (Fig. 1c). However, in infected seedlings, clusters of secondary roots emerged close to nematode infection sites in an inoculum density-dependent manner ( Fig. 1c,d). Interestingly, the number of secondary roots per cluster also significantly increased at inoculum density 350 compared with 50 ( Fig. 1c,e). Moreover, higher inoculation densities caused more extensive discoloration at the infection sites indicating higher levels of tissue damage. Altogether, these observations showed that infection by H. schachtii triggers local density-dependent formation of secondary roots.

Heterodera schachtii host invasion induces JA biosynthesis and signaling
Artificially induced tissue damage can trigger the formation of roots via JA-dependent signaling pathways. For instance, wounding induces JA-dependent de novo root organogenesis in leaf explants . Infective juveniles of H. schachtii invade the host by destructive thrusts of the oral stylet and release of plant cell wall degrading enzymes causing extensive cell damage during host invasion (Grundler et al., 1994;Tytgat et al., 2002;Vanholme et al., 2007). We hypothesized that secondary root formation in the proximity of H. schachtii infection sites might be regulated by JA, in response to tissue damage associated with nematode host invasion. To test our hypothesis, we investigated whether JA biosynthesis and signaling were activated during H. schachtii infection using the JA biosynthesis reporter line pAOS:: YFP N (Poncini et al., 2017) and the JA signaling biosensor JAS9-VENUS (p35S::JAS9-VENUS/p35S::H2B-RFP) (Larrieu et al., 2015) (Fig. 2). We chose three time points that reflect the early parasitic stages of intracellular host invasion (12 hpi), permanent feeding site initiation (24 hpi), and permanent feeding site expansion (168 hpi) (Tytgat et al., 2002;Hewezi et al., 2014;Kammerhofer et al., 2015;Marhavy et al., 2019). Importantly, to avoid interference of signals due to the presence of multiple nematodes at one infection site, we selected single-nematode infection sites for our observations. We found that infection with H. schachtii significantly induces transient expression of pAOS::YFP N , with the highest level of expression at 12 hpi ( Fig. 2a-d). Likewise, JAS9-VENUS showed a strong JA signaling activity (i.e. low VENUS : RFP ratio) in infected roots at 12 hpi, which decreased over time to the level of uninfected root tissue at 168 hpi (Fig. 2eh). These observations demonstrated that both JA biosynthesis and JA signaling are strongly induced during and shortly after H. schachtii host invasion close to the nematode infection site. We therefore concluded that tissue damage caused by H. schachtii during intracellular host invasion triggers local JA biosynthesis and signaling in Arabidopsis.

COI1-mediated JA signaling regulates ERF109 expression upon H. schachtii infection
Root tip resection or wounding in leaf explants induces ERF109 expression in a COI1-dependent manner Zhou et al., 2019). To determine whether H. schachtii-induced JA signaling also triggers ERF109 expression, we monitored pERF109::GFP expression within single-nematode infection sites in the coi1-2 mutant and wild-type Arabidopsis Col-0 plants during the early stages of infection by H. schachtii (Fig. 3). Similar to that observed for JA biosynthesis and signaling, ERF109 expression was induced at early time points (12 and 24 hpi) of H. schachtii infection around the migratory track of the nematodes in wildtype Col-0 ( Fig. 3a-

Research
New Phytologist pERF109::GFP fluorescence was significantly reduced compared with wild-type Arabidopsis (Fig. 3d). Nevertheless, we observed a slight increase in fluorescence in coi1-2 mutant over time, which reached the fluorescence levels detected in Col-0 at 168 hpi. The fluorescence detected in wild-type Col-0 and coi1-2 mutant at 168 hpi might be caused by tissue autofluorescence from the cell walls of the permanent feeding sites (Hoth et al., 2005). Since pERF109::GFP has a nuclear-cytoplasmic localization  due to GFP diffusion into the nucleus (Hanson & Kohler, 2001), autofluorescence from cell walls in syncytia cannot be easily distinguished from the cytoplasmic part of the pERF109:: GFP signal. However, after quantifying only nuclear-localized pERF109::GFP, the initially observed fluorescence in the nematode-infected wild-type Col-0 and coi1-2 mutant plants at 168 hpi was not detected anymore, pointing at autofluorescence as the most plausible cause (Fig. S2). To independently verify a COI1-dependent regulation of ERF109 expression, we also performed a RT-qPCR on ERF109 transcripts in root segments containing nematode infection sites collected from coi1-2 and wildtype Col-0 at 12 hpi. Consistent with the observed regulation of Nine-day-old Arabidopsis Col-0 seedlings were inoculated with increasing numbers of H. schachtii second-stage juveniles (J2s), ranging from 0 (mock) to 500 J2s per seedling. At 7 d postinoculation (dpi), scans were made of the root systems, and the total number of secondary roots per plant was counted. Fuchsin staining was performed to count the number of J2s that had penetrated the roots. Additionally, the presence of clusters and the number of secondary roots per cluster were scored. (a) Number of nematodes that successfully penetrated the roots per inoculum. (b) Number of secondary roots formed per number of nematodes inside the roots. The total number of secondary roots in infected seedlings was normalized to the average respective component in uninfected roots and correlated with the number of nematodes inside the roots. Data from three independent biological repeats of the experiment were combined. Correlation (R) between two variables was calculated using Spearman's rank-order correlation coefficient (n = 30; P < 0.0001 pERF109::GFP fluorescence, we found significantly fewer transcripts of ERF109 in coi1-2 than in wild-type Col-0 (Fig. 3e). We therefore concluded that H. schachtii induces ERF109 expression during host invasion in a JA-dependent manner.
COI1 and ERF109 regulate secondary root formation upon H. schachtii infection Next, we asked whether the activation of JA-dependent expression of ERF109 is required for the formation of secondary roots during H. schachtii infections. If this holds true, the nematode density-dependent increase in secondary roots observed for wildtype Col-0 should be altered in both coi1-2 and erf109 mutants.
To test this, we performed the same density-response experiment as shown in Fig. 1(a,b). At 7 dpi, the number of nematodes that had successfully penetrated the roots did not differ significantly between wild-type Col-0 and the erf109 mutant (Fig. 4). By contrast, the number of nematodes was significantly higher in roots of the coi1-2 mutant than wild-type Arabidopsis plants, indicating a role of COI1 in plant susceptibility to penetration by H. schachtii (Fig. 4b). However, it must be noted that the uninfected coi1-2 mutant had a much larger root system than wild-type Arabidopsis Col-0 (Fig. S3), which also may influence the number of nematode penetrations. Nevertheless, while nematode infections in wild-type Arabidopsis induced the formation of secondary roots, no such increase was observed for erf109 and coi1-2 mutants (Fig. 4c). In conclusion, both COI1 and ERF109 regulate the density-dependent induction of secondary root formation by H. schachtii. This induction of secondary root formation is independent from plant susceptibility to nematode penetration.   Values represent log 2 of the fluorescence ratio between JAS9-VENUS and H2B-RFP raw integrated densities. Data from three independent biological repeats of the experiment were combined. Significance of differences between fluorescent intensities in nematode-infected and noninfected seedlings over the different time points was calculated by analysis of variance followed by Tukey's HSD test for multiple comparisons (n = 30; P < 0.0001). For boxplots, the horizontal line represents the median, the whiskers indicate the maximum/minimum range, and the black dots represent the outliers. Different letters indicate statistically different groups. White arrowheads indicate the nematode head; white dotted lines outline the nematode body. TL, transmission light. Bar, 100 lm. (2023)   (e) Fig. 3 ERF109 expression upon Heterodera schachtii host invasion is dependent on COI1-mediated jasmonate signaling. (a-d) Four-day-old Arabidopsis seedlings were either inoculated with 15 H. schachtii second-stage juveniles (J2s) or mock-inoculated. At 12, 24, and 168 h post-inoculation (hpi), seedlings were mounted in 10 lg ml À1 propidium iodide (PI) and then imaged using a fluorescent confocal microscope. Single-nematode infection sites were selected for observation. (a-c) Representative pictures of infected and mock-inoculated seedlings expressing the pERF109::GFP construct in either wild-type Col-0 or mutant coi1-2 background at 12 hpi (a), 24 hpi (b), and 168 hpi (c). To make the fluorescence more visible, the brightness was enhanced for all the representative pictures in the same way. (d) Quantification of pERF109::GFP fluorescent intensity induced by infection of Col-0 and coi1-2 roots. Values represent log 2 of the fluorescence ratio between the GFP integrated density of infected and noninfected roots. Data from two independent biological repeats of the experiment were combined. Significance of differences between fluorescent intensities in Col-0 and coi1-2 roots over the different timepoints was calculated by analysis of variance followed by Tukey's HSD test for multiple comparisons (n = 20; P < 0.05). For boxplots, the horizontal line represents the median and the whiskers indicate the maximum/minimum range. Different letters indicate statistically different groups. White and black arrowheads indicate the nematode head; white dotted lines outline the nematode body. TL, transmission light. Bar, 200 lm. (e) Twelve-day-old Col-0 and coi1-2 Arabidopsis plants were inoculated with H. schachtii. At 12 hpi, RNA was extracted from root segments of c. 0.2 cm harboring nematode infection sites or similar root segments of mock-inoculated seedlings. Data represent three independent biological replicates with three technical replicates per biological replicate. Relative expression of ERF109 was first calculated based on the endogenous control 18 s rRNA and then normalized to the mock-inoculated wildtype samples in the first biological replicate. Significance of differences between ERF109 relative expression in Col-0 and coi1-2 infected roots was calculated by ANOVA followed by Tukey's HSD test for multiple comparisons (n = 3 biological replicates; P < 0.01). Different letters indicate statistically different groups. Error bars represent SE of the mean.

ERF109-mediated induction of secondary root formation compensates for primary root growth inhibition by H. schachtii
The induction of secondary root formation by cyst nematodes might compensate for a possible inhibition of root growth by nematode invasion. To test this hypothesis, we investigated whether the total length of the entire root system, the primary root length, and the total length of the secondary roots were altered in the infected erf109 mutant compared with wild-type Col-0 (Fig. 5). To eliminate the background effect of the mutation on the root architecture, we normalized each measured component in infected seedlings to the average respective component in uninfected roots. We found that the total length of the root system of wild-type Col-0 at increasing numbers of nematodes remains similar to that of uninfected plants (i.e. close to 1 in Fig. 5a). By contrast, the total length of the entire root system in the erf109 mutant decreased by nematode density as compared to uninfected plants. As the total length of the root system is the sum of the lengths of the primary roots and the secondary roots, we also analyzed these components separately. The primary root length of both wild-type Col-0 and the erf109 mutant declined by nematode density (Fig. 5b). This decline was slightly but significantly exacerbated by the erf109 mutation. However, we found a more striking difference in the total length of the secondary roots between wild-type Col-0 and the erf109 mutant (Fig. 5c). In wild-type Col-0, we observed a significant increase in the total length of the secondary roots by nematode density, sufficient to compensate for the loss in primary root length. However, we observed no significant increase in the total length of the secondary roots by nematode density in the erf109 mutant, which explains why the total length of the root system by nematode density remained stable for wild-type Col-0, but not for the erf109 mutant. Based on our data, we conclude that ERF109mediated formation of secondary roots compensates for primary root growth inhibition by H. schachtii.

ERF109 regulates local auxin biosynthesis at the nematode infection site
ERF109 mediates JA-induced secondary root formation by directly binding to the promoter of auxin biosynthesis genes ERF109 and COI1 regulate the nematode density-dependent secondary root formation that is triggered by Heterodera schachtii infections. Nine-dayold wild-type Col-0, erf109, and coi1-2 seedlings were inoculated with increasing numbers of H. schachtii second-stage juveniles (J2s), ranging from 0 (mock) to 500 J2s per seedling. At 7 d post-inoculation (dpi), scans were made of the root systems, and the number of secondary roots per plant was counted. Fuchsin staining was performed to count the number of J2s that had penetrated the roots. (a) Representative pictures of wild-type Col-0, erf109, and coi1-2 infected seedlings at 7 dpi. (b) Number of nematodes that successfully penetrated the roots per inoculum. (c) Secondary roots formed per number of nematodes inside the roots. The total number of secondary roots in infected seedlings was normalized to the average respective component in mock-treated roots. Data from three independent biological repeats of the experiment were combined. Significance of differences between genotypes was calculated by analysis of variance followed by Tukey's HSD test for multiple comparisons (n = 30; P < 0.0001). Gray area indicates 95% confidence interval. Bar, 2 cm.

New Phytologist
ASA1 and YUC2 (Cai et al., 2014). We hypothesized that ERF109 regulates secondary root formation by inducing local auxin biosynthesis at the nematode infection site. Thus, we used a split plate assay containing growth media with and without Lkyn to chemically inhibit auxin biosynthesis in the shoots and/or the roots of infected wild-type and erf109 plants (Fig. 6). The local accumulation of auxin was monitored using the DR5::GUS reporter (Fig. 6a). When seedlings were grown on regular medium or when auxin biosynthesis was inhibited by L-kyn only in the shoots, DR5::GUS was expressed at nematode infection sites in wild-type Col-0 seedlings. However, when auxin biosynthesis was inhibited in both shoots and roots or only in the roots by treatment with L-kyn, no DR5::GUS expression was observed (Fig. 6b,c). This suggested that auxin accumulation at nematode infection sites was dependent on local auxin biosynthesis in the roots. Importantly, we observed that the auxin accumulation at nematode infection sites via root-localized auxin biosynthesis was disrupted in the erf109 mutant. Indeed, DR5::GUS expression was significantly lower at the nematode infection sites in erf109 seedlings than wild-type Col-0 when auxin biosynthesis was permitted in the root (Fig. 6b,c). To determine whether the differences in DR5::GUS between the two Arabidopsis genotypes were only local at the nematode infection site or systemic throughout the root system, we also looked at DR5::GUS expression in root tips (Figs 6d, S4). In contrast to nematode infection sites, when auxin biosynthesis was inhibited only in the shoots, we observed no difference between erf109 and wild-type Col-0 in DR5::GUS expression in the root tip (Figs 6d, S4). Since L-kyn has been shown to also inhibit ethylene-induced auxin biosynthesis (He et al., 2011), we also performed the experiment using the auxin biosynthesis inhibitor Yucasin (Yuc). Due to the higher concentration of DMSO used to dissolve Yuc, an overall lower frequency of DR5:GUS staining was observed. Nevertheless, the Yuc split plate assay showed the same trend as the L-kyn experiment (Fig. S1). From these results, we concluded that ERF109 regulates local auxin biosynthesis at infection sites of H. schachtii.

ERF109-induced secondary root formation upon H. schachtii infection is dependent on local auxin biosynthesis
We found that ERF109 regulates local auxin biosynthesis at H. schachtii infection sites. This raised the question of whether the ERF109-mediated secondary root formation upon H. schachtii infection is dependent on this local biosynthesis of auxin. To test this, we inoculated 4-d-old wild-type Col-0 and erf109 seedlings with either 15 H. schachtii J2s or a mock solution. At 16 hpi, seedlings were transferred to the four previously described split plates containing medium with and without 10 lM L-kyn (Fig. 6a). At 7 dpi, the total number of secondary roots was scored. As expected, the different treatment combinations with and without L-kyn in the shoots and/or roots led to a different number of lateral roots in the uninfected roots (Fig. S5). Therefore, to calculate the number of additional secondary roots induced by nematode infection, the number of secondary roots in infected roots was normalized to the average respective component in uninfected roots. Additionally, we scored how often a cluster of roots occurs in the proximity of an infection site and the number of secondary roots per cluster (Fig. 7). When auxin biosynthesis was inhibited in both shoots and roots or only in the roots, no additional secondary roots formed in infected Col-0 wild-type seedlings (Fig. 7a,b). Consistently, no clusters of secondary roots were found at nematode infection sites (Fig. 7c-e). However, inhibition of auxin biosynthesis in the shoots alone led to a significant reduction in the total number of secondary roots in infected seedlings (Fig. 7a,b) as well as in the number of clusters and the number of secondary roots per cluster compared with when auxin biosynthesis was permitted in both shoots and roots (Fig. 7c-e; treatment MM vs KM). Thus, secondary root formation upon H. schachtii infection is dependent on local auxin biosynthesis, although polar auxin transport from the shoots might still play a role. Furthermore, the mutation in erf109 strongly affected secondary root formation when auxin biosynthesis was permitted in the roots. Indeed, a significant decrease in the number of additional secondary roots, the number of clusters of secondary roots, and the number of secondary roots per cluster was observed for erf109 than wild-type Col-0 (Fig. 7). Altogether, we concluded that ERF109-dependent secondary root formation upon H. schachtii infection relies at least partially on local auxin biosynthesis.

Discussion
Root architecture plasticity in response to stress by soil-borne pathogens and pests is a largely unexplored field of research. Root parasitism by cyst nematodes is often associated with the formation of secondary roots in the proximity of infection sites (Grymaszewska & Golinowski, 1991;Goverse et al., 2000;Lee et al., 2011). However, the molecular mechanisms regulating secondary root formation in response to cyst nematode infection have thus far remained unclear. Here, we provide evidence for a model wherein formation of secondary roots near H. schachtii infection sites is triggered by tissue damage caused by nematode invasion. This response is regulated by the JA-dependent ERF109-activated local biosynthesis of auxin.

New Phytologist
Our data demonstrate that secondary root formation is most likely initiated by tissue damage brought about by cyst nematode infections. The number of secondary roots induced by H. schachtii correlated positively with the number of nematodes that penetrated the roots. This increase in the number of secondary roots may be simply due to an increase in the number of infection sites.
However, we also observed more nematodes within infection sites at higher inoculation densities, which correlated well with the number of secondary roots per infection site. This may mean that infection sites containing multiple nematodes developed a higher number of secondary roots per cluster than single-nematodeassociated infection sites. Moreover, we saw more extensive root secondary roots in infected vs noninfected roots of wild-type Col-0 and erf109 seedlings. Data from two independent biological repeats of the experiment were combined. Significance of differences in secondary roots between the different treatment combinations was calculated by analysis of variance followed by Tukey's HSD test for multiple comparisons (n = 43-45; P < 0.05). (c) Representative images of nematode infection sites in wild-type Col-0 and erf109 mutant. (d) Number of secondary root clusters that are associated with H. schachtii infection sites. Data from two independent biological repeats of the experiment were combined. Statistical significance was calculated by a pairwise Z-test (n = 31-33; ****, P < 0.0001). (e) Number of secondary roots per cluster. Data from two independent biological repeats of the experiment were combined. Significance of differences between secondary roots within a cluster was calculated by aligned rank transform for nonparametric factorial ANOVA followed by Tukey's HSD test for multiple comparisons (n = 31-33; P < 0.0001). For boxplots, the horizontal line represents the median, the whiskers indicate the maximum/minimum range, and the black dots represent the outliers. Difference in letters indicates statistically different groups. Black arrowheads indicate the infection site. Bar, 0.5 cm.
New Phytologist (2023) 237: 807-822 www.newphytologist.com tissue damage (i.e. root discoloring) at infection sites harboring multiple nematodes. We therefore consider tissue damage by infective juveniles inside roots as the likely cause of enhanced local secondary root formation. Tissue damage in Arabidopsis leaf explants triggers de novo root organogenesis in a JA-dependent manner . We found that intracellular host invasion by H. schachtii transiently induces JA biosynthesis and signaling and that the JA receptor mutant coi1-2 is defective in secondary root formation upon H. schachtii infection. Our results are in line with whole transcriptome analyses of root segments of Arabidopsis harboring migrating juveniles of H. schachtii at 10 hpi, which also showed that JA biosynthesis and signaling genes are upregulated during host invasion (Kammerhofer et al., 2015;Mendy et al., 2017). By contrast, recent reports indicate that host invasion by H. schachtii does not activate the JA signaling biosensor JAZ10::NLS-3xVENUS in Arabidopsis roots (Marhavy et al., 2019). The discrepancy between our observations with the JAS9-VENUS biosensor and the observations with the JAZ10:: NLS-3xVENUS biosensor might be due to differences in sensitivity of both sensor constructs. Compared to JAZ10::NLS-3xVENUS, the JAS9-VENUS biosensor is particularly sensitive to biologically active JA (JA-isoleucine), enabling the visualization of local JA signaling in response to stress in Arabidopsis roots at a high spatiotemporal resolution (Larrieu et al., 2015). Furthermore, JAS9-VENUS has been used to monitor the dynamics of JA signaling in response to single-cell ablation and intercellular migration of the less-damaging root-knot nematodes in Arabidopsis roots . Therefore, based on the activity of the JAS9-VENUS biosensor in our experiments, we conclude that the tissue damage associated with host invasion triggers a JA signal in cells close to the infection site of H. schachtii. Moreover, the transient nature of the JA signal suggests that the damage trigger decreases after nematode host invasion or that JA signaling is actively suppressed by H. schachtii when infective juveniles become sedentary.
Jasmonate signaling during H. schachtii migration also results in activation of plant defense responses (Kammerhofer et al., 2015). We observed that the coi1-2 mutant is more susceptible to penetration by H. schachtii, which is in line with previous findings showing a negative effect of exogenous JA on H. schachtii penetration rate (Kammerhofer et al., 2015). However, after nematode penetration, COI1 does not affect the rate at which J2s induce a permanent feeding site (Marhavy et al., 2019). Altogether, these findings suggest that JA signaling both negatively regulates host penetration rate by H. schachtii and mediates secondary root formation at H. schachtii infection sites.
The damage-induced formation of secondary roots by H. schachtii appears to be regulated by the JA-dependent expression of ERF109. We found that the expression of ERF109, which showed the same transient induction pattern as the JA biosynthesis reporter AOS and JAS9-VENUS biosensor, was abrogated in the coi1-2 mutant. Moreover, the erf109 mutant was as defective as the coi1-2 mutant in the density-dependent secondary root formation upon H. schachtii infection. Consistent with our data, ERF109 expression showed a COI1-dependent transient expression upon wounding in leaf explants . Furthermore, the erf109 mutation also disrupted the induction of secondary root formation by exogenous application of JA (Cai et al., 2014). Altogether, our findings show that tissue damage by invading nematodes triggers a JA signal, which induces the ERF109-dependent formation of secondary roots.
Next, our data provide evidence that damage-induced activation of ERF109 regulates the formation of secondary roots via local auxin biosynthesis. The local accumulation of auxin at nematode infection sites (i.e. expression of the auxin reporter DR5::GUS) was strongly reduced in the erf109 mutant than in wild-type plants. However, when auxin biosynthesis was blocked in whole seedlings or only in roots, the local accumulation of auxin at nematode infection sites was completely abolished in both the erf109 mutant and the wild-type Arabidopsis. Taken together, this demonstrates that auxin accumulation at nematode infection sites is at least partially dependent on ERF109regulated local auxin biosynthesis. Importantly, the patterns observed for local accumulation of auxin at nematode infection sites matched the patterns of secondary root formation in the absence or presence of the auxin biosynthesis inhibitor. The inhibition of auxin biosynthesis in the roots, but not in the shoots, abolished the formation of secondary roots upon nematode infection. Previously, ERF109 was shown to regulate secondary root formation by binding the promoter of auxin biosynthesis genes upon exogenous application of JA (Cai et al., 2014). Here, our data show that tissue damage by nematodes activates JA signaling and subsequently induces ERF109, which in turn regulates secondary root formation via local biosynthesis of auxin.
After blocking auxin biosynthesis in the shoots, we observed auxin accumulation and formation of secondary roots at nematode infection sites, which indicates that polar auxin transport from the shoots is not required for secondary root formation at nematode infection sites. Nevertheless, we noted a quantitative effect of the inhibition of auxin biosynthesis in shoots, leading to the formation of fewer secondary root clusters and fewer secondary roots per cluster as compared to untreated plants. This implies that polar auxin transport from the shoots may still play a complementary role in secondary root formation at nematode infection sites, albeit below the detection levels of the DR5::GUS reporter. Polar auxin transport from the shoots and further redistribution in root tissue results from the coordinated activities of auxin influx and efflux carrier proteins (Petrasek & Friml, 2009). Lee et al. (2011) showed that H. schachtii induces the formation of secondary roots in double aux1lax3 and quadruple aux1lax1lax2lax3 influx carrier mutants, which are otherwise unable to form secondary roots. This suggests that the accumulation of auxin and subsequent formation of secondary roots may be regulated independently of the activity of these influx carriers. There is ample evidence that auxin efflux carriers (i.e. PIN proteins) are important for the susceptibility of Arabidopsis to infections of H. schachtii (Grunewald et al., 2009). However, if and how they might contribute to the accumulation of auxin underlying the damage-induced formation of secondary roots needs further investigation.
Here, we demonstrate that ERF109-mediated local adaptations in root architecture compensate for primary root growth

Research
New Phytologist inhibition in response to nematode infection. In wild-type Arabidopsis, increasing densities of J2s led to a decline in the length of infected primary roots. However, this reduction in the length of infected primary roots did not result in a smaller root system, because of an increase in the total length of secondary roots. Our data show that these adaptations in root architecture depend on the transient and local activation of ERF109 by JA at nematode infection sites. Consistently, the JA signaling mutant coi1-2 showed a similar impairment as erf109 in compensating primary root length inhibition by an increase in total secondary root length (Fig. S6). Nevertheless, since COI1 also affects plant susceptibility to nematode penetration, more complex defense vs growth trade-offs may influence root growth in the coi1-2 mutant. Importantly, loss-of-function mutations in ERF109 do not alter the susceptibility of Arabidopsis to H. schachtii penetration but instead affect root architecture plasticity in response to nematode infection. Further research is needed to understand whether ERF109-mediated compensatory adaptations in root architecture could mediate tolerance of Arabidopsis to infections by H. schachtii.
It was previously shown that meristem damage caused by M. incognita root tip penetration triggers regeneration via JA-and ERF109-mediated damage signaling . Here, we show that H. schachtii penetration of the mature root zone causes damage-induced secondary root formation, which compensates for primary root growth inhibition. Therefore, we consider root tip regeneration and secondary root formation as two different outcomes of the same compensatory mechanism in response to tissue damage in different root zones.
Furthermore, we show the first case of a naturally occurring and biotic stress that triggers damage signaling-mediated secondary root formation. Primary roots can form two types of secondary roots (Sheng et al., 2017). One type, referred to as a lateral root, forms during the physiological postembryogenic development of plants and is regulated by ARF7 and ARF19 auxin response factors. The other type is induced by sterile mechanical injury of the mature root zone, soil penetration, or osmotic stress and is dependent on the transcription factor WOX11. Sterile mechanical injury causes a different type of root tissue damage compared with a biotic stress, such as cyst nematodes (Marhavy et al., 2019). Sterile mechanical injury damages many root cells at one time. Instead, cyst nematode host invasion causes the rupture of multiple single cells one after the other over the course of many hours (Wyss & Zunke, 1986). Thus, our results provide biological relevance for a mechanism so far only observed upon artificial conditions. As a natural trigger for damage signaling, H. schachtii can be used to further elucidate the pathway leading to secondary root formation. ERF109 was previously found to be responsive to reactive oxygen species (ROS) (Kong et al., 2018). It would be interesting to test whether ROS mediates ERF109-dependent secondary root formation upon H. schachtii infection. Furthermore, follow-up research could investigate whether damage receptors activated during H. schachtii migration (Shah et al., 2017) act upstream of ERF109. The root-knot nematode M. javanica triggers the expression of LBD16, a downstream target of both WOX11, and ARF7 and ARF19 (Cabrera et al., 2014;Olmo et al., 2017). Moreover, M. javanica infection of primary roots induces secondary root formation independently from ARF7 and ARF19 (Olmo et al., 2017). This suggests that nematode-induced secondary root formation could be regulated by WOX11. However, whether WOX11-mediated secondary root formation acts downstream of the ERF109-damage signaling pathway remains unknown.
In summary, we showed that H. schachtii triggers the formation of secondary roots via JA-and ERF109-mediated damage signaling (Fig. 8). Furthermore, ERF109-mediated secondary root formation compensates for primary root growth inhibition associated with H. schachtii infection. Thus, damage signalinginduced formation of secondary roots points at a novel mechanism underlying plant root architecture plasticity to biotic stress. Further research is needed to investigate whether damage- induced root architecture plasticity can contribute to plant tolerance to belowground herbivory.

Supporting Information
Additional Supporting Information may be found online in the Supporting Information section at the end of the article.

Research
New Phytologist