Fusarium graminearum‐induced shoot elongation and root reduction in maize seedlings correlate with later seedling blight severity

Abstract Fusarium graminearum seedling blight is a common disease of maize (Zea mays). Development of genetic resistance to seedling blight in maize germplasm requires efficient and accurate quantitative assessment of disease severity. Through artificial inoculation experiments under controlled growth conditions, we determined that host genotype, pathogen genotype, and infection dose influence the extent to which F. graminearum induces shoot elongation and inhibits root growth in maize seedlings. A comparison of 15 maize inbred lines showed independent variation of these two fungus‐induced effects on seedling growth. In a broader survey with nine commercial maize hybrids and three field‐collected fungal isolates, there was significant correlation between these seedling growth responses, as well as with later seedling blight severity. Analysis of variance suggested that this variation and the observed correlative relationships were primarily driven by differing pathogenicity of the three fungal isolates. Together, our results indicate that F. graminearum‐induced shoot elongation and root reduction in maize seedlings have distinct underlying physiological mechanisms, and that early observations of seedling growth responses can serve as a proxy for investigating natural variation in host resistance and pathogen aggressiveness at later growth stages.

fungicide treatment is not always sufficient and may promote the development of resistance in the fungal population. Genetic sources of seedling blight resistance have yet to be established in commercial maize germplasm.
In addition to causing direct yield loss through tissue rotting, F. graminearum can also cause crop contamination with mycotoxins such as deoxynivalenol and zearalenone, leaving the grain and stover unsafe for human and livestock consumption, and imposing further financial burdens on the growers (Arunachalam & Doohan, 2013;Maresca, 2013). Furthermore, F. graminearum can also cause destructive scab and head blight diseases in barley and wheat (McMullen et al., 2012;Trail, 2009). Considering the overwinter persistence and airborne infectious ascospores of this pathogen, unchecked F. graminearum diseases in maize could have a far-reaching negative impact on other crop species in the vicinity.
Breeders and researchers interested in improving resistance against F. graminearum in maize have been doing artificial inoculation experiments in both field and controlled environments, primarily focused on infected maize ears and stalks. Genetic mapping of Gibberella ear rot resistance based on symptom severity has identified a large number of quantitative trait loci (QTL), each with small effect size. Moreover, these QTL tend to be sensitive to experimental methods and environmental conditions, making them difficult to validate across different studies (Ali et al., 2005;Brauner et al., 2017;Kebede, Woldemariam, Reid, & Harris, 2016). In contrast, major genetic loci contributing to Gibberella stalk rot resistance have been identified with bi-parental mapping populations (Chen et al., 2017;Ma et al., 2017;Yang et al., 2010;Zhang et al., 2012). Ye et al. (2013) observed that maize near-isogenic lines resistant against Gibberella stalk rot also showed less severe seedling root browning and shrinkage after F. graminearum inoculation. This suggests that resistance against Gibberella stalk rot and F. graminearum seedling blight, both diseases in vegetative tissues, are likely controlled by similar genetic and physiological mechanisms, and are distinct from those affecting Gibberella ear rot resistance. This is consistent with the known influence of tissue type and developmental stage on the interactions between maize and F. graminearum (Zhang et al., 2016).
Genetic improvement of F. graminearum resistance depends on accurate quantitative assessment of resistant and susceptible phenotypes. Here, we show that F. graminearum inoculation of maize seedlings not only inhibits root growth, but can also promote shoot elongation in a dosage-and genotype-dependent manner. Shoot elongation in response to F. graminearum root infection is significantly correlated with seedling survival, suggesting that these traits could be used to assess seedling resistance, as well as serving as a quantitative measure of the aggressiveness of different F. graminearum isolates.  Table S1). The maize hybrids were selected to cover different maturation time and provider-assessed seedling vigor levels. The accession numbers of the tested hybrids are masked per the provider's request. Three field-collected F. graminearum isolates (Supporting Information Table S2) that have shown different levels of aggressiveness in a previous laboratory test (Kuhnem, Del Ponte, Dong, & Bergstrom, 2015) were used for inoculations (Supporting Information Table S2). Fungal chemotype, namely the type of major mycotoxin accumulated, was not considered during isolate selection, since it was not significantly associated with aggressiveness in the previous test.

| Plant growth and fungal inoculation methods
Maize seeds were germinated in moisturized rolls of germination paper after surface sterilization with 10% household bleach for 30 min. Germinated seedlings were transplanted to 7.5 cm × 7.5 cm plastic pots with Turface ® MVP ® calcined clay (Profile Products LLC, Buffalo Grove, IL) when their primary roots reached approximately 9 cm, which would take 3-5 days depending on the specific maize genotype. For inoculation experiments, seedling roots were immersed in F. graminearum spore suspension or mock solution (0.03% Phytagar) for 1 hr prior to transplanting. Fungal spore sus- For the concentration gradient experiment, 24 seedlings of each maize hybrid were grown, and six were inoculated with each tested fungal spore concentration and mock inoculum. For each pair of maize hybrid and fungal isolate included in the diversity screening, at least 15 seedlings were inoculated with fungal spores, and 5 were mock-inoculated. Transplanted seedlings were kept at 26°C (Day)/ 23°C (Night), 60% humidity, and a 16 hr long-day light cycle. Seedlings inoculated with different F. graminearum isolates or mock inoculum were kept in separated growth chambers set to identical growth conditions.

| Phenotypic data collection and analysis
The heights of fungus-and mock-inoculated seedlings were measured 9 days post-inoculation. On the same day, seedlings were removed from pots to measure their total root length using an image analysis algorithm implemented in the RootReader 2D software (Clark et al., 2013). After measurement, seedlings were repotted in the same Turface particles. At 14 days post-inoculation, the number of surviving seedlings within each fungus-inoculated maize hybrid population was counted as a measurement of F. graminearum seedling blight severity. Seedlings were considered dead when they broke at the soil line, with browning and rot at the breakpoint.
For seedling height and total root length, an average ratio of mock-and fungus-inoculated seedlings for each maize hybrid-fungus isolate combination was calculated to shown variation in these fungus-induced responses. Measurements from fungus-and mockinoculated seedlings were compared using Student's t tests. Average ratios were rounded to one when measurements from the two groups were not significantly different from one another. For twoway ANOVA, maturation times of maize hybrids were grouped into four bins, as indicated in parentheses in Supporting Information Table S1. The average ratios of seedling height and total root length between fungus-and mock-treated seedlings were used as quantitative measurements to calculate linear correlation with seedling survival rate measured at 14 days after infection.

| Fusarium graminearum inoculation at low concentration induces shoot elongation and root reduction in maize seedlings
Different concentrations of spore suspensions have been used to study natural variation in maize seedling responses to F. graminearum inoculation (Kuhnem et al., 2015;Ye et al., 2013). To determine the optimal concentration of F. graminearum spores that leads to the largest change in plant morphology, two concentration gradient experiments were performed with Gz014, a field-collected F. graminearum strain (Supporting Information Table S2), and two commercial maize hybrids with differing seedling vigor levels (Hybrids IV and V in Supporting Information Table S1, as assessed by the provider, Blue River Organic Seed Company). At 1.2 × 10 6 spores/ml, a concentration that is comparable to a previous study, the seedling stunting phenotype that had been reported was not observed in either of the tested maize hybrids (Kuhnem et al., 2015). Instead, the hybrid with high seedling vigor showed no significant change in seedling height, whereas the moderately vigorous hybrid seedlings grew significantly taller when inoculated with F. graminearum spores. At lower concentrations, fungus-induced shoot elongation became obvious in both highly and moderately vigorous hybrid seedlings ( Figure 1a). Underground, F. graminearum infection significantly reduced root growth in the moderately vigorous hybrid, with no evidence of a dosage response.
This fungus-induced change was not observed in the highly vigorous hybrid, which showed no significant change in total root length at all tested doses (Figure 1b).
For both fungus-induced growth responses, inoculation with the lowest fungal spore concentration (3 × 10 5 count/ml) resulted in a response that was stronger than or comparable to that observed with the two higher inoculum concentrations. Therefore, the 3 × 10 5 spores/ml F. graminearum dose was used for all subsequent experiments.

| Fusarium graminearum-induced shoot elongation and root reduction vary independently across maize inbred lines
To confirm the F. graminearum-induced shoot elongation and root reduction observed in the concentration gradient experiment, the artificial inoculation experiment was repeated with the reference maize genetic inbred line B73 with the same fungal isolate following the same protocol. Consistently, the same fungus-induced growth responses were observed (Figure 2). To investigate how F. graminearum-induced seedling shoot elongation and root reduction may be related, seedlings of 15 maize inbred lines were inoculated with Gz014 or mock inoculum and their shoot and root growth were quantified 9 days after inoculation. Although F. graminearum inoculation reduced total root length in most inbred lines by approximately 50%, no significant difference was observed between mock-and F. graminearum-inoculated NC350, NC358, and Mo17 seedlings, (Figure 3a).
In the same experiment, F. graminearum-induced shoot elongation also varied, ranging from 0 to over 25% increase ( Figure 3b).
Noticeably, the two fungus-induced growth responses do not correlate with one another across the 15 tested inbred lines tested (Pearson's R 2 = 0.0079, p > 0.05). Two inbred lines showing no significant F. graminearum-induced root reduction, NC350 and NC358, grew 16% and 24% taller, respectively, after F. graminearum inoculation.
Conversely, Tx303 only showed obvious root reduction (40%) but no significant shoot elongation (Figure 3). Inbred line Mo17 showed neither of the tested fungus-induced growth responses.

| Fusarium graminearum-induced growth responses and F. graminearum seedling blight severity are primarily determined by the fungal genotype
To further survey the natural variation in F. graminearum-induced seedling shoot elongation and root reduction, the artificial inoculation experiment was expanded to the nine commercial maize hybrids, including the two used in the concentration gradient experiment described above (Supporting Information Table S1), and three F. graminearum isolates (Supporting Information Table S2). Both F. graminearum-induced growth responses were highly variable across this panel (Figure 4). Two-way analysis of variance (ANOVA) performed on both traits showed that fungal isolate, rather than maize maturation time or provider-assessed seedling vigor, is the main source of assay variance (Supporting Information Table S3).
After quantifying F. graminearum-induced growth responses, the fungus-infected seedlings were kept under the same growth conditions, and F. graminearum seedling blight severity was quantified as the percentage of seedlings surviving in each maize hybrid population 14 days after inoculation. As in the case of the morphological data, ANOVA of the seedling survival rate data identified the F. graminearum isolate as the most important contributor. Whereas both Gz014 and Gz829 infection left less than 50% of inoculated seedlings surviving after 14 days, more than 85% of Gz835-ZHOU ET AL.  Information Table S3).

| Fusarium graminearum-induced shoot elongation is positively correlated with F. graminearum seedling blight severity
To investigate the relationship between F. graminearum-induced growth responses and F. graminearum seedling blight severity, linear regression analysis was performed on the extent of these changes and seedling survival. This showed that F. graminearum-induced seedling elongation has a significant negative correlation with seedling survival rate at 14 days after inoculation (Figure 6a; Pearson's R 2 = 0.244; p < 0.005). Thus, stronger fungus-induced shoot elongation may be predictive of more severe F. graminearum seedling blight. Unexpectedly, the extents of F. graminearum-induced shoot elongation showed strong positive correlation with root length (Figure 6b; Pearson's R 2 = 0.676; p < 0.005), and a significant negative correlation between root reduction and seedling survival rate F I G U R E 1 Fusarium graminearum inoculation induces shoot elongation and root reduction in maize seedlings in a dosage-and host genotype-dependent manner. Shoot height (a) and total root length (b) measurements on seedlings inoculated at each fungal spore concentration were compared to mock-treated seedlings with Dunnett's tests relative to mock-infected controls (N = 6 for each group; * p < 0.05). Group means are indicated by the black bars, and the interquartile range by the upper and lower edges of the boxes. Outliers extending more than 1.5× of the interquartile range are indicated by individual circles, and the whiskers show the range of each group excluding the outliers. Fungal strain Gz014 was used for this experiment. All measurements were taken 9 days post-inoculation. Fungal spore concentration is in counts/ml  Maize hybrid accession Shoot height ratio (fungus/mock inoculated)

Root length ratio (fungus/mock inoculated)
F I G U R E 4 Fusarium graminearuminduced growth responses in maize seedlings vary with both plant and fungal genotype. Nine maize hybrid lines were inoculated with three F. graminearum isolates (Gz014, Gz829, and Gz835). Total root length (a) and shoot height (b) of each maize hybrid-fungus isolate pair were compared to measurement from a corresponding mock treatment group with Student's t-test (*p < 0.05). All measurements were taken at 9 days postinoculation. Fungal spore concentration = 3 × 10 5 count/ml ( Figure 6c; Pearson's R 2 = 0.263; p < 0.005). This leads to the counter-intuitive inference that resistance to fungus-induced root growth inhibition co-occurs with stronger shoot elongation and more severe F. graminearum seedling blight.

| DISCUSSION
Under field conditions, maize seedlings infected by F. graminearum as well as other fungal and oomycete pathogens are often characterized by stunted growth and dehydration in the aboveground tissues before death. Similar seedling stunting symptoms have also been observed in a previous artificial inoculation test with F. graminearum spores under laboratory conditions (Kuhnem et al., 2015). However, in our concentration gradient experiment with the same fungal isolates, low dosage F. graminearum inoculation on maize seedling root induced shoot elongation (Figures 1a and 2a). Based on the natural variation in this F. graminearum-induced response across different maize inbred lines and hybrids reported in this study (Figures 3a and   4a), it is likely that the inconsistent observations between studies could arise from inoculum concentration, maize genotype, or unintended differences between experimental conditions. Root length ratio (fungus/mock inoculated)

Seedling
F I G U R E 5 Fusarium graminearum-induced growth responses and seedling blight severity vary significantly between fungal isolates. The extent of F. graminearum-induced root reduction and shoot elongation at 9 days post-inoculation (a & b) and survival rates of maize seedlings inoculated with different F. graminearum isolates at 14 days post-inoculation (c) were compared using two-way analysis of variance with Tukey's HSD. Group means are indicated by the black bars, and the interquartile range by the upper and lower edges of the boxes. Outliers extending more than 1.5× of the interquartile range were indicated by individual circles, and the whiskers showed the range of each group excluding the outliers. Significantly different groups are denoted with different letters. Fungal spore concentration = 3 × 10 5 count/ml. Gz829 Gz835 Shoot height ratio (fungus/mock infected)

Root length ration (fungus/mock infected)
F I G U R E 6 Fusarium graminearum-induced growth parameter changes in roots and shoots, and seedling blight severity, which are significantly correlated with each other. Each mark represents a maize hybrid-fungal isolate pair, with different fungal isolates represented by different colors. Root length and seedling height were measured at 9 days after infection. Survival was assessed at 14 days after infection Interestingly, the observed F. graminearum-induced shoot elongation phenotype is reminiscent of the foolish seedling disease in rice, which is caused by a related fungal pathogen Gibberella fujikuroi. Rice seedlings affected by foolish seedling disease demonstrate abnormal seedling elongation resulting from the production of gibberellic acid (GA), a growth-promoting phytohormone, G. fujikuroi (Yabuta & Sumiki, 1938). Fungus-derived GA has been hypothesized to be a virulence factor by negatively influencing the signaling transduction pathway of the immunity-regulating phytohormone, jasmonic acid (Navarro et al., 2008). However, unlike G. fujikuroi, F. graminearum is not known to produce GA, and its genome does not contain biosynthetic genes required for GA biosynthesis (Cuomo et al., 2007).
However, we cannot rule out the possibility that F. graminearum infection induces endogenous maize GA production.
Consistent with previous reports (Kuhnem et al., 2015;Ye et al., 2013), F. graminearum infection induced maize seedling root growth reduction in a host genotype-dependent manner (Figures 1b, 2b, 3b and 4b). Application of a computer image-processing tool, RootReader 2D (Clark et al., 2013), in our study enabled more accurate, quantitative, and efficient measurement of fungus-induced growth response, and its natural variation with host and pathogen genotype.
Fungus-induced root reduction and shoot elongation measurements from the same experiment on 15 maize inbred lines showed no correlation in the extent of these two traits, with some inbred lines showing only one or the other fungus-induced change (Figure 3a,b).
This observation suggests that these two responses may be governed by distinct physiological mechanisms.
As one of the most widespread fungal phytopathogens, F. graminearum is known to be highly diverse, and different strains are highly variable in their aggressiveness (Backhouse, 2014;Boutigny et al., 2011;Lee et al., 2010;Sampietro et al., 2011;Wang, Ndoye, Zhang, Li, & Liao, 2011). Our survey of nine commercial maize hybrids and three fungal isolates (Figures 4 and 5) is consistent with prior observations that the prevalence of F. graminearumrelated diseases is highly variable across multiple locations, depending on the aggressiveness of the predominant isolate and environmental permissiveness (Backhouse, 2014;Boutigny et al., 2011;Lee et al., 2010;Sampietro et al., 2011;Wang et al., 2011).
Unexpectedly, maize hybrid-fungal isolate pairs with stronger fungus-induced root growth inhibition tended to show a lower level of induced shoot elongation ( Figure 6c) and better survival (Figure 6b). This was surprising, because severe reduction in root growth was expected to significantly impair the seedlings' ability to acquire water and nutrients from the soil, which could lead to seedling death. Two-way ANOVA and a closer look at the data distribution suggest that these correlative relationships are primarily driven by differences between the fungus isolates. Specifically, the isolate that induced significantly stronger shoot elongation, Gz835, induced less root reduction and seedling death ( Figure 5). This suggests that these fungus-induced growth responses can be used for quantitative assessment of the aggressiveness of F. graminearum isolates, and that different F. graminearum isolates may have distinct pathogenic mechanisms when interacting with maize seedlings. Further research with maize seedlings growing in agricultural fields will be required to determine whether F. graminearum-induced shoot elongation can serve as a reliable predictor of fungal sensitivity in commercially grown maize.

ACKNOWLEDG MENTS
The authors would like to thank the Blue River Organic Seed Company for providing the maize hybrids used in this study, Jon Shaff and Eric Craft for training and assistance in using the RootReader