Species-Specific Impact of Fusarium Infection on the Root and Shoot Characteristics of Asparagus

Soil-borne pathogens can have considerable detrimental effects on asparagus (Asparagus officinalis) growth and production, notably caused by the Fusarium species F. oxysporum f.sp. asparagi, F. proliferatum and F. redolens. In this study, their species-specific impact regarding disease severity and root morphological traits was analysed. Additionally, various isolates were characterised based on in vitro physiological activities and on protein extracts using matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS). The response of two asparagus cultivars to the different Fusarium species was evaluated by inoculating experiments. Differences in aggressiveness were observed between Fusarium species and their isolates on roots, while no clear disease symptoms became visible in ferns eight weeks after inoculation. F. redolens isolates Fred1 and Fred2 were the most aggressive strains followed by the moderate aggressive F. proliferatum and the less and almost non-aggressive F. oxysporum isolates, based on the severity of disease symptoms. Fungal DNA in stem bases and a significant induction of pathogenesis-related gene expression was detectable in both asparagus cultivars. A significant negative impact of the pathogens on the root characteristics total root length, volume, and surface area was detected for each isolate tested, with Fred1 causing the strongest effects. No significant differences between the tested asparagus cultivars were observed.


Introduction
Asparagus officinalis L. is a perennial horticultural crop grown over a wide range of soils and under various climatic conditions. It can remain profitable for about 10 years in temperate regions that have long growing seasons. However, a gradual decline in productivity and the growth of asparagus plants can be observed within the growing period due to the 'asparagus decline' syndrome [1] mainly associated with the elevated soil densities of pathogenic Fusarium spp. [2]. Additionally, 'early decline' and 'asparagus replant disease' can occur when previous asparagus fields are replanted. Fusarium crown and root rot disease affects almost all plant organs. Fusarium oxysporum (Schltdl.) f.sp. asparagi and Fusarium proliferatum (Matsush.) are considered the main pathogenic species causing asparagus rots worldwide [3][4][5][6][7] and therefore have been studied the most. Depending on the continent and country, there are further Fusarium species involved, such as Fusarium redolens (Wollenw.) and Fusarium solani (Mart.) in Europe [8]. Here, F. redolens is of interest because this species is closely similar to F. oxysporum [9]. F. oxysporum enters its host through the root and induces either root rots or tracheomycosis when invading the vascular system [10]. Symptoms on asparagus caused by different Fusarium species are similar and include brown lesions on roots, and stems as well as rots of roots, were assumed. Thus, the objectives of this study were to (i) characterise isolates of F. oxysporum, F. proliferatum and F. redolens at physiological, pathogenic and molecular levels; (ii) identify the defence response of two asparagus cultivars to isolates of the three Fusarium species by verifying fungal DNA presence in infected plants and analysing the expression of defence-related genes by RT-qPCR; and (iii) study root and shoot characteristics and identify root morphology characteristics of two asparagus cultivars in response to infection by the three Fusarium species using an automated analysis of root images.

Identification and Physiological Characterisation of Fusarium Spp.
In order to analyse whether the efficacy in degrading plant-derived carbon sources is correlated with disease severity caused by the Fusarium spp. isolates, we tested their ability to utilise different plant-derived polysaccharides. Fungal growth on complex polysaccharides and cell wall constituents, cellulose and pectin was slower as compared to the other tested substances (Figure 1). No species-specific growth response was found for cellulose as carbon source. However, F. oxysporum isolates showed better growth on pectin as compared to the other Fusarium species. In addition, extracellular enzymatic activities were analysed in plate tests as another factor influencing the usability of these polysaccharides as carbon source for fungal growth. By visualisation on chromogenic medium, no extracellular enzyme activity was detected when Fusarium spp. were cultivated on the two substrates cellulose and pectin. In contrast, extracellular enzymatic activity was detected on plates containing starch, trehalose, and cellobiose in all tested isolates, with the latter substance being degraded most efficiently ( Figure S1). Substrate utilisation of trehalose and starch was significantly higher in F. oxysporum isolates as compared to F. proliferatum and F. redolens (Figure 1). Pathogens 2020, 9, x FOR PEER REVIEW 3 of 20 oxysporum, F. proliferatum and F. redolens at physiological, pathogenic and molecular levels; (ii) identify the defence response of two asparagus cultivars to isolates of the three Fusarium species by verifying fungal DNA presence in infected plants and analysing the expression of defence-related genes by RT-qPCR; and (iii) study root and shoot characteristics and identify root morphology characteristics of two asparagus cultivars in response to infection by the three Fusarium species using an automated analysis of root images.

Identification and Physiological Characterisation of Fusarium Spp.
In order to analyse whether the efficacy in degrading plant-derived carbon sources is correlated with disease severity caused by the Fusarium spp. isolates, we tested their ability to utilise different plant-derived polysaccharides. Fungal growth on complex polysaccharides and cell wall constituents, cellulose and pectin was slower as compared to the other tested substances (Figure 1). No species-specific growth response was found for cellulose as carbon source. However, F. oxysporum isolates showed better growth on pectin as compared to the other Fusarium species. In addition, extracellular enzymatic activities were analysed in plate tests as another factor influencing the usability of these polysaccharides as carbon source for fungal growth. By visualisation on chromogenic medium, no extracellular enzyme activity was detected when Fusarium spp. were cultivated on the two substrates cellulose and pectin. In contrast, extracellular enzymatic activity was detected on plates containing starch, trehalose, and cellobiose in all tested isolates, with the latter substance being degraded most efficiently ( Figure S1). Substrate utilisation of trehalose and starch was significantly higher in F. oxysporum isolates as compared to F. proliferatum and F. redolens ( Figure  1).   (Fpro1, Fpro2, Fpro3), and F. redolens (Fred1, Fred2); cell wall components (cellobiose, pectin, cellulose) are shown on the left, storage carbohydrates (starch, trehalose) are shown on the right. Mycelial diameters were measured at four days (cellobiose, trehalose, starch) and seven days (cellulose, pectin) after plate inoculation. The values are the means ± standard errors (n = 8). Significant differences are indicated by different letters (p < 0.05; Tukey test).
Pathogens 2020, 9, 509 4 of 20 To test the potential of matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) for the species characterisation of Fusarium isolates, mass spectra on the basis of protein fingerprints of mycelium from overnight fungal cultures were obtained. A comparison of spectra from the different Fusarium spp. by hierarchical clustering showed a clear separation between F. proliferatum isolates and the remaining isolates, while no separation between F. oxysporum and F. redolens isolates was found ( Figure 2). Pathogens 2020, 9,  To test the potential of matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) for the species characterisation of Fusarium isolates, mass spectra on the basis of protein fingerprints of mycelium from overnight fungal cultures were obtained. A comparison of spectra from the different Fusarium spp. by hierarchical clustering showed a clear separation between F. proliferatum isolates and the remaining isolates, while no separation between F. oxysporum and F. redolens isolates was found ( Figure 2).

Aggressiveness of Fusarium spp. Isolates on Asparagus Plants
Possible differences in the aggressiveness of Fusarium isolates and the susceptibility of the two asparagus cultivars, 'Backlim' and 'Rapsody', were initially tested by monitoring disease symptoms and plant growth. The typical symptoms on roots were brown to reddish-brown necrotic spots and lesions of variable length on both storage and fibrous roots. The tissue within the rhizome was strongly affected so that the inner tissue of storage roots was degraded and roots consisted only of root rhizodermis and the central cylinder. This was predominantly observed by inoculation with F. proliferatum and F. redolens isolates ( Figure S2). Control plants did not show any symptoms. Two-way ANOVA of DS of roots showed no interaction between cultivars and isolates. Therefore, the influence of isolates on DS is statistically analysed across both cultivars. However, with regard to the biological difference of 'Backlim' and 'Rapsody' in plant performance, we show the effect of isolates on the DS of roots separately for each cultivar, being aware that statistical significance between isolates and control counts for both cultivars likewise. This type of presentation allows the comparison between treatments differing both in cultivar and isolate level [35]. Disease severity (DS) ratings of roots inoculated with Fusarium spp. isolates differed significantly from the control plants (Figures 3 and  S3). DS ratings showed a significant species-specific aggressiveness on the roots of both cultivars: F. redolens was the most aggressive species followed by F. proliferatum and thereafter F. oxysporum. Between the cultivars 'Backlim' and 'Rapsody' no significant difference in susceptibility of roots was

Aggressiveness of Fusarium spp. Isolates on Asparagus Plants
Possible differences in the aggressiveness of Fusarium isolates and the susceptibility of the two asparagus cultivars, 'Backlim' and 'Rapsody', were initially tested by monitoring disease symptoms and plant growth. The typical symptoms on roots were brown to reddish-brown necrotic spots and lesions of variable length on both storage and fibrous roots. The tissue within the rhizome was strongly affected so that the inner tissue of storage roots was degraded and roots consisted only of root rhizodermis and the central cylinder. This was predominantly observed by inoculation with F. proliferatum and F. redolens isolates ( Figure S2). Control plants did not show any symptoms. Two-way ANOVA of DS of roots showed no interaction between cultivars and isolates. Therefore, the influence of isolates on DS is statistically analysed across both cultivars. However, with regard to the biological difference of 'Backlim' and 'Rapsody' in plant performance, we show the effect of isolates on the DS of roots separately for each cultivar, being aware that statistical significance between isolates and control counts for both cultivars likewise. This type of presentation allows the comparison between treatments differing both in cultivar and isolate level [35]. Disease severity (DS) ratings of roots inoculated with Fusarium spp. isolates differed significantly from the control plants ( Figure 3 and Figure S3). DS ratings showed a significant species-specific aggressiveness on the roots of both cultivars: F. redolens was the most aggressive species followed by F. proliferatum and thereafter F. oxysporum. Between the cultivars 'Backlim' and 'Rapsody' no significant difference in susceptibility of roots was observed. DS values were significantly discriminable between inoculation with the isolates of the three Fusarium species on 'Backlim' and 'Rapsody'; except for isolate Fpro3 that had similar impact like Foa1 and Foa4. Among the F. oxysporum f.sp. asparagi isolates Foa1 and Foa4 were likewise aggressive and more Pathogens 2020, 9, 509 5 of 20 aggressive than Foa2 and Foa3 but they still did not cause DS scores higher than 2 and fell below the normalised rank mean of 0.5 ( Figure 3). No difference in aggressiveness was observed among the F. proliferatum isolates Fpro1 and Fpro2, with both causing higher DS than Fpro3 ( Figure 3). The highest DS ratings on both cultivars were caused by the two F. redolens isolates Fred1 and Fred2, which differed significantly from each other. In contrast to root, the DS ratings of ferns were very low, with highest value of 1.9, and almost no differences among the Fusarium spp. (Figure S4). The DS of ferns did not differ significantly between the Fusarium species, and the correlation between the DS of roots and of stems was not significant (r 2 = 0.347).
Pathogens 2020, 9, x FOR PEER REVIEW 5 of 20 observed. DS values were significantly discriminable between inoculation with the isolates of the three Fusarium species on 'Backlim' and 'Rapsody'; except for isolate Fpro3 that had similar impact like Foa1 and Foa4. Among the F. oxysporum f.sp. asparagi isolates Foa1 and Foa4 were likewise aggressive and more aggressive than Foa2 and Foa3 but they still did not cause DS scores higher than 2 and fell below the normalised rank mean of 0.5 ( Figure 3). No difference in aggressiveness was observed among the F. proliferatum isolates Fpro1 and Fpro2, with both causing higher DS than Fpro3 ( Figure 3). The highest DS ratings on both cultivars were caused by the two F. redolens isolates Fred1 and Fred2, which differed significantly from each other. In contrast to root, the DS ratings of ferns were very low, with highest value of 1.9, and almost no differences among the Fusarium spp. ( Figure  S4). The DS of ferns did not differ significantly between the Fusarium species, and the correlation between the DS of roots and of stems was not significant (r 2 = 0.347). Two-way ANOVA of root fresh weight (FW) revealed no interaction between cultivars and isolates. We found a significant main effect of cultivar and isolate on root FW accumulation; root FW between the cultivars was significantly different. Thus, mean values of cultivars are depicted separately and statistical significance between isolates and control counts for both cultivars likewise. Root FW after inoculation with all isolates was significantly lower than the control. Among all isolates, Foa2, Foa3, and Foa4 had significantly lowest impact on FW reduction compared to Fpro3, Fred1, and Fred2. The highest impact was observed on plants inoculated with F. redolens isolate Fred1 ( Figure 4A).  Two-way ANOVA of root fresh weight (FW) revealed no interaction between cultivars and isolates. We found a significant main effect of cultivar and isolate on root FW accumulation; root FW between the cultivars was significantly different. Thus, mean values of cultivars are depicted separately and statistical significance between isolates and control counts for both cultivars likewise. Root FW after inoculation with all isolates was significantly lower than the control. Among all isolates, Foa2, Foa3, and Foa4 had significantly lowest impact on FW reduction compared to Fpro3, Fred1, and Fred2. The highest impact was observed on plants inoculated with F. redolens isolate Fred1 ( Figure 4A).
Two-way ANOVA of fern FW revealed no interaction between cultivars and isolates. We found significant main effect of cultivar and isolate on fern FW accumulation; fern FW between the cultivars differed significantly ( Figure 4B). Thus, mean values of cultivars are depicted separately and statistical significance between isolates and control counts for both cultivars likewise. Compared to the control, a significant reduction in fern FW was only observed when plants were inoculated with F. proliferatum isolate Fpro3 and F. redolens isolates Fred1 and Fred2 ( Figure 4B). A medium but significant correlation was observed between the DS of roots and root FW (r 2 = 0.539), and between root DS and fern FW (r 2 = 0.533). The correlation between the FW of root and fern was highly significant (r 2 = 0.835).  Two-way ANOVA of fern FW revealed no interaction between cultivars and isolates. We found significant main effect of cultivar and isolate on fern FW accumulation; fern FW between the cultivars differed significantly ( Figure 4B). Thus, mean values of cultivars are depicted separately and statistical significance between isolates and control counts for both cultivars likewise. Compared to the control, a significant reduction in fern FW was only observed when plants were inoculated with F. proliferatum isolate Fpro3 and F. redolens isolates Fred1 and Fred2 ( Figure 4B). A medium but significant correlation was observed between the DS of roots and root FW (r 2 = 0.539), and between root DS and fern FW (r 2 = 0.533). The correlation between the FW of root and fern was highly significant (r 2 = 0.835).

Molecular Detection of Fusarium spp. in Stem Tissue
The specific detection of Fusarium spp. was carried out in the stem bases of harvested plants to provide evidence of fungal spread in inoculated plants. Species-specific primers were based on sequence variations, previously analysed and standard curves were calculated in the presence of plant DNA as described. Although proof of presence of the specific fungal DNA was clear evidence of infection, the amount of fungal DNA did not necessarily correlate with the severity of disease symptoms. However, a certain relationship between fungal DNA in the stem bases of asparagus and root FW after inoculation with Fusarium spp. was detectable in both cultivars 'Backlim' (r 2 = 0.402) and 'Rapsody' (r 2 = 0.690) ( Figure S5).

Expression of Defence-Related Genes
Quantitative real-time reverse transcription PCR was performed to analyse the RNA accumulation of defence genes coding for pathogenesis-related proteins. Differences in the relative expression levels of pathogenesis-related protein 1 (PR1-2), pathogenesis-related protein 1-like (PR1-4), cationic peroxidase 1-like (POX), and phenylalanine ammonia-lyase (PAL) in the stem bases of harvested plants are shown in Figure 5. While the steady-state levels of mRNA coding for PAL were unchanged in both cultivars 8 weeks after inoculating roots with Fusarium spp. isolates, a significant reduction in the steady-state levels of mRNA of POX was detectable in 'Rapsody' after inoculation with F. oxysporum and F. redolens isolates. An induction of PR-genes was found in both cultivars, with the strongest increase in relative gene expression after inoculation with F. redolens. The expression increase of PR1-2 was significant and higher in 'Backlim' than in 'Rapsody'. sequence variations, previously analysed and standard curves were calculated in the presence of plant DNA as described. Although proof of presence of the specific fungal DNA was clear evidence of infection, the amount of fungal DNA did not necessarily correlate with the severity of disease symptoms. However, a certain relationship between fungal DNA in the stem bases of asparagus and root FW after inoculation with Fusarium spp. was detectable in both cultivars 'Backlim' (r 2 = 0.402) and 'Rapsody' (r 2 = 0.690) ( Figure S5).

Expression of Defence-Related Genes
Quantitative real-time reverse transcription PCR was performed to analyse the RNA accumulation of defence genes coding for pathogenesis-related proteins. Differences in the relative expression levels of pathogenesis-related protein 1 (PR1-2), pathogenesis-related protein 1-like (PR1-4), cationic peroxidase 1-like (POX), and phenylalanine ammonia-lyase (PAL) in the stem bases of harvested plants are shown in Figure 5. While the steady-state levels of mRNA coding for PAL were unchanged in both cultivars 8 weeks after inoculating roots with Fusarium spp. isolates, a significant reduction in the steady-state levels of mRNA of POX was detectable in 'Rapsody' after inoculation with F. oxysporum and F. redolens isolates. An induction of PR-genes was found in both cultivars, with the strongest increase in relative gene expression after inoculation with F. redolens. The expression increase of PR1-2 was significant and higher in 'Backlim' than in 'Rapsody'.

Changes in Root Morphology
Based on the significant disease symptoms in the roots of asparagus plants that were inoculated with Fusarium spp. isolates, a more quantitative investigation of root morphology was performed

Changes in Root Morphology
Based on the significant disease symptoms in the roots of asparagus plants that were inoculated with Fusarium spp. isolates, a more quantitative investigation of root morphology was performed using image analysis. Two-way ANOVA revealed no interaction between cultivars and isolates. Statistically significant difference between isolates and control counts for 'Backlim' and 'Rapsody' likewise. Because of biological cultivar specification, also, here, the mean values of cultivars are depicted separately. No difference between cultivars was observed by none of the traits except for fibrous root radius. The inoculation of asparagus plants with the nine Fusarium spp. isolates was associated with a decrease in total length of both, fibrous and storage roots, as measured separately ( Figure 6). Total root lengths of plants inoculated with F. oxysporum isolate Foa1, F. proliferatum isolates Fpro1 and Fpro3, and F. redolens isolates Fred1 and Fred2 were significantly lower than control plants ( Figure 6). Isolate Fred1, provoking the highest DS among the nine tested Fusarium spp. (Figure 3), caused a storage root Pathogens 2020, 9, 509 8 of 20 length reduction of 46% compared to control plants. A significant reduction of storage root length of 36 and 40% was also observed when plants were inoculated with Fpro1 and Fpro3 ( Figure 6). depicted separately. No difference between cultivars was observed by none of the traits except for fibrous root radius. The inoculation of asparagus plants with the nine Fusarium spp. isolates was associated with a decrease in total length of both, fibrous and storage roots, as measured separately ( Figure 6). Total root lengths of plants inoculated with F. oxysporum isolate Foa1, F. proliferatum isolates Fpro1 and Fpro3, and F. redolens isolates Fred1 and Fred2 were significantly lower than control plants ( Figure 6). Isolate Fred1, provoking the highest DS among the nine tested Fusarium spp. (Figure 3), caused a storage root length reduction of 46% compared to control plants. A significant reduction of storage root length of 36 and 40% was also observed when plants were inoculated with Fpro1 and Fpro3 ( Figure 6). F. redolens isolate Fred1 significantly reduced the fibrous root lengths of both cultivars and induced a decrease of approximately 77% ( Figure 6). In addition, Fred2 caused a reduction of fibrous loot length by 63%. Among the F. proliferatum isolates Fpro1 and Fpro3 caused significant fibrous root length reductions of ca. 50 and 60%. F. oxysporum isolate Foa1 had a significant effect on plants when compared to controls. DS of roots correlated significantly with fibrous root length (r 2 = 0.499), but not with the storage root length (r 2 = 0.406). High correlation values between root length and fern FW (r 2 = 0.759), and between root length and root FW (r 2 = 0.890) show a close relation between these two traits.
No impact of F. oxysporum was observed on the root radius, subdivided into fibrous and storage roots radius ( Figure S6). F. proliferatum isolate Fpro3 caused a significant reduction in storage and fibrous roots' radius As a consequence of decreased root lengths, the fibrous and storage roots' volumes of plants that were inoculated with the Fusarium spp. isolates Foa1, Fpro1, Fpro2, Fpro3, Fred1, and Fred2 were significantly reduced (Figure 7). F. redolens isolate Fred1 significantly reduced the fibrous root lengths of both cultivars and induced a decrease of approximately 77% ( Figure 6). In addition, Fred2 caused a reduction of fibrous loot length by 63%. Among the F. proliferatum isolates Fpro1 and Fpro3 caused significant fibrous root length reductions of ca. 50 and 60%. F. oxysporum isolate Foa1 had a significant effect on plants when compared to controls. DS of roots correlated significantly with fibrous root length (r 2 = 0.499), but not with the storage root length (r 2 = 0.406). High correlation values between root length and fern FW (r 2 = 0.759), and between root length and root FW (r 2 = 0.890) show a close relation between these two traits.
No impact of F. oxysporum was observed on the root radius, subdivided into fibrous and storage roots radius ( Figure S6). F. proliferatum isolate Fpro3 caused a significant reduction in storage and fibrous roots' radius As a consequence of decreased root lengths, the fibrous and storage roots' volumes of plants that were inoculated with the Fusarium spp. isolates Foa1, Fpro1, Fpro2, Fpro3, Fred1, and Fred2 were significantly reduced (Figure 7).
The root surface area of 'Backlim' and 'Rapsody' control plants was 850 and 875 cm 2 , respectively ( Figure 8). Inoculation with Fpro1, Fpro3, Fred1 and Fred2 reduced root surface area to nearly half of that of the control plants. The utmost reduction about more than 500 cm 2 was caused by inoculation with F. redolens isolate Fred1. F. oxysporum f.sp. asparagi isolate Foa1 and Fpro2 also had a significant impact on the surface area reduction of plants ( Figure 8). The correlation values between root surface area and root DS, root FW and stem FW (r 2 root DS = 0.583; r 2 root FW = 0.908; r 2 stem FW = 0.689) indicate a tense relation between these traits.

Discussion
Asparagus plants can be infected by different Fusarium species and are colonised via the roots. The influence of Fusarium on the root growth characteristics of asparagus depending on the species has not been studied yet. In this study, the impact of four isolates of F. oxysporum, three of F. proliferatum and two of F. redolens was analysed on the asparagus cultivars 'Backlim' and 'Rapsody', with different growing parameters. Hence, we expected differences in response to Fusarium infection depending on the cultivars.
Since different Fusarium spp. can occur on asparagus, the availability of a rapid and safe species-specific diagnosis tool would be useful. Nowadays, the identification of microbial pathogens is routinely performed using matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS, Bruker Daltonik, Bremen, Germany ), especially for bacterial iso lates [36]. Additionally, mass-spectrometry-based approaches have also been applied successfully for the identification and differentiation of Fusarium species based on spore protein profiles [37,38]. Furthermore, MALDI-TOF MS peptide/protein fingerprints have been used to distinguish closely related species in the Fusarium fujikuroi species complex [39], and for in situ diagnosis in infected plant tissue [40]. Here, we have tested the potential of MALDI-TOF MS to distinguish the used Fusarium species based on protein fingerprints of mycelium. A reliable separation between F. proliferatum and the remaining F. oxysporum and F. redolens isolates was found. A comparable pattern of similarities among these Fusarium spp. was shown by Baayan et al., who performed AFLP fingerprinting followed by hierarchical clustering [9]. With an increasing database of fungal protein fingerprints, the future application of this rapid and economic tool for pathogen diagnosis should lead to the establishment of better and, also, in situ fungal pathogen detection.
The comparative analysis of translated fungal genomes revealed that the highest number of carbohydrate-active enzymes occurs in necrotrophic and hemibiotrophic plant pathogenic fungi, with Fusarium spp. ranking among the highest [41]. The synthesis of cell wall degrading enzymes, a subset of carbohydrate-active enzymes, is of high importance during plant infection with Fusarium spp. [42], as some of those are secreted to facilitate infection and to exploit the plant carbon source. In this study, we intended to answer the question if differences in the activity of carbohydrate-active enzymes between the tested Fusarium spp. are given. It is to assume that a relation between the ability to degrade plant-derived carbon sources and aggressiveness exist. A species-specific substrate utilisation was found for pectin, trehalose and starch, although the response was contrarious to what was found for isolates of F. culmorum [43,44]. The least aggressive F. oxysporum isolates showed the highest ability to degrade pectin and to utilise starch and trehalose, indicating that other factors, despite the tested carbohydrate-active enzyme activities account for species-specific aggressiveness to asparagus.
The impact of F. oxysporum, F. proliferatum and F. redolens isolates was investigated at an early growth stage of asparagus with particular emphasis on changes in the root morphology. Based on the present pathogenicity test, we found that our tested isolates differed significantly in their aggressiveness, and, to a certain extent, a classification of the Fusarium spp. in four pathogenicity groups was possible. F. redolens isolates were the most aggressive strains followed by the moderate aggressive F. proliferatum and the weakly aggressive F. oxysporum isolates. Based on the severity of disease symptoms, significant differences in aggressiveness were observed between the isolates of each species. Following this classification, the F. proliferatum isolate Fpro3 should rather be ranged in the group of the two pathogenic F. oxysporum isolates Foa1 and Foa4. Following the argumentation of Elmer and Stephens [45] who considered isolates causing DS higher than 20% as pathogenic on asparagus, the two isolates Foa2 and Foa3 that were not able to cause DS scores higher than 1 (that includes a diseased root area up to 20%) could be rated as non-pathogenic. However, it has to be noted that our testing was performed with rather young plants and long-term observations in fields may reveal stronger and increasing impact over time. It is worth mentioning that, in the past, many pathogenicity tests with Fusarium strains performed in different countries resulted in diverse outcomes. Borrego-Benjumea et al. [7] showed in their pathogenicity tests that F. oxysporum, F. solani and F. acuminatum were highly aggressive, whereas F. proliferatum and F. redolens were weakly aggressive. The most aggressive Fusarium spp. in Spain were F. proliferatum and F. solani [30], and, in the Netherlands, Blok and Bollen [11] found, in F. oxysporum isolates, the highest aggressiveness. In Japan and the United Kingdom, F. oxysporum and F. proliferatum were the most pathogenic Fusarium spp. responsible for asparagus decline [29,46]. In summary, this demonstrates that the assessment of pathogenicity depends on the analysed isolates, the cultivars, and the interaction of both with the environmental conditions.
The results in the present experiments show a comparable level of susceptibility in the two cultivars, while they were chosen because of differences in growth and robustness. There was no cultivar-specific significant difference in the DS of roots caused by the various Fusarium spp. Thus, both cultivars can likely be regarded as equally susceptible to at least the tested Fusarium spp. isolates.
The aggressiveness of the isolates of the three Fusarium species is largely reflected in the root traits. Compared to control plants, the root FWs of plants were reduced when they were inoculated with each isolate. A decrease in root FW was noticed with increasing DS of roots. These results are in accordance with the findings of Corpas-Hervias et al. [30], who found a general trend in the decrease in root dry weight with increasing severity of root rot symptoms. An impact of the isolates on fern FW was obvious when plants were inoculated with F. proliferatum isolate Fpro3 and F. redolens isolates Fred1 and Fred2, with Fred1 being the most aggressive isolate. The impact of the F. oxysporum f.sp. asparagi isolates was low, highlighting the weak aggressiveness of this Fusarium species, but also Fpro1 and Fpro2 had only weak impact on the FW of fern.
In our experiments, the DS of ferns was very low and we observed no significant impact of the fern DS on root FW (r 2 = 0.198) or on fern FW (r 2 = 0.347), and also no significant correlation existed between the DS of roots and DS of ferns (r 2 = 0.347). This is in contrast to the results of Corpas-Hervias et al. [30], who found a highly significant correlation between root rot and fern severity ratings. It is assumed that the rather short duration of our experiment with plant harvest eight weeks post inoculation may have minimised the impact of the pathogens on fern FWs, although the DNA of the pathogen was already detectable in the stem bases. The presence of fungal DNA clearly provides proof of plant infection; however, the amount of fungal DNA measured in the respective tissue does not necessarily correlate with DS. Since we have inoculated plant roots, the detection of fungal DNA in stem bases is an expression of fungal spread into aerial parts of the plant. This is the typical strategy of soil-borne fungal pathogens that cause vascular wilt in host plants. After expansion in the host they produce resting structures in plant tissues that get incorporated into soil; from there they again start interaction by first challenging the roots and subsequently spread inside the vascular tissue [47]. In addition, the first evidence of the ferns' response to the presence of the pathogen was detected using gene expression analysis. The expression of PR1 was significantly higher and the expression of another pathogenesis-related gene (PR9) POX was unchanged or even reduced in stem bases, thus resembling a typical PR-gene expression profile upon Fusarium infection [48]. The expression of a key biosynthetic enzyme, mainly involved in defence mechanisms, that catalyses the first step in the synthesis of a variety of polyphenol compounds PAL [49] remained unchanged. This is in contrast to other experiments, where the presence of F. oxysporum f.sp. lycopersici strains induced PAL expression as a clear evidence of induced systemic acquired resistance in tomato [50]. Symptoms on ferns and the reduction of fern FW seems to be the result of a longer period of interaction between plant and the soil-borne Fusarium spp., and it would be interesting to analyse the impact of the pathogens on aboveground parts of plants at this later stage of pathogenesis.
The morphology of inoculated asparagus roots classified in storage and fibrous roots was measured by means of a new automated processing pipeline. No significant difference between 'Rapsody' and 'Backlim' was observed regarding these traits, except for fibrous roots radius. A general tendency in the decrease in root length could be observed with increasing severity of root symptoms. The measured root volume values showed a general tendency like the results of the total root length, and no clear conclusion could be drawn regarding data of radius of fibrous and storage roots. We assume that most probably after attack with the pathogen, particularly fibrous roots die off and hence the remaining roots that were measured were almost not injured and stayed unaffected by the pathogen. Additionally, symptoms caused by Fusarium spp. on affected roots comprised lesions that were not yet so severe on root tissue at assessment time and therefore root radius was less influenced. The decrease in total root length resulted in a reduction of root surface area. Apart from F. proliferatum isolates Fpro1, Fpro2, and Fpro3, and both F. redolens isolates Fred1, and Fred2 also F. oxysporum f.sp. asparagi isolate Foa1 caused a clear reduction in the root surface area. Altogether, it can be concluded that the two F. redolens isolates had the highest impact followed by F. proliferatum isolates Fpro1 and Fpro3. F. oxysporum f.sp. asparagi isolates had the lowest effect on root traits and fresh weight. Thus, the variation in aggressiveness of the isolates was to some degree reflected in the effect of the isolates on fresh weight and root morphology.
These observations on root morphology are in accordance with other studies with Fusarium spp. and further soil-borne pathogens. For example, a decrease in root length, root weight, root surface area, and root volume was observed on tomato plants inoculated with F. oxysporum f.sp. lycopersici [34]. Crosby [51] stated that total root length, fine and small root length, and the number of root tips were greater for a melon cultivar that was tolerant to the soil-borne fungus Monosporascus cannonballus. On the other hand, there are also opposed statements on the correlation between root traits and the impact of soil-borne pathogens. For example, Cichy et al. [52] could not find significant interactions between attacks with F. solani f.sp. phaseoli and the root system traits of beans. They made the short experiment duration of 32 days responsible for the missing impact of the pathogen on the host. In a soil compaction-F. solani root rot interaction study with large-and medium-rooted pea lines, Kraft and Boge [53] asserted that larger-rooted pea lines had significantly more roots than did the smaller-rooted line. However, they observed that reduction in root growth was stronger for the larger-rooted lines than for the smaller-rooted pea line.
Insights into the root morphology of plants are fundamental for the comprehension of root functionality. Fibrous roots of asparagus are mainly responsible for water and nutrient uptake and the storage roots have the role as storage locations for reserve carbohydrates (mostly fructans). The considerable decreases in fresh weight, root length, root volume, and root surface area as shown in this study entail the limitation of water and nutrient uptake of diseased plants and result in weakness and stunted growth of plants. The ability of asparagus plants to store carbohydrates in these roots is limited when storage roots are reduced and this has a negative effect on the production of spears for the growing season and hence on yield. In this study, we could clearly differentiate the aggressiveness of Fusarium spp. isolates. Moreover, the measured root traits, including root fresh weight, total root length, root volume, and root surface area, turned out to be meaningful indicators for the general status of the plant. The correlations between the severity of root rot symptoms and root traits were mostly significant. Arias et al. [54] observed for some Fusarium spp. isolates significant linear relationships between soybean yield and root symptoms and concluded from their studies that the root morphological characteristics of soybean were more consistent indicators of yield loss than root rot severity caused by Fusarium spp. on soybean. The yield loss in asparagus that is related with the asparagus decline is also related to the presence of Fusarium spp. in soil, but following Elmer [55] apart from this soil-borne pathogen there are many other indicators that contribute to asparagus decline. Guo [56] specified that asparagus cultivars with high yield potential exhibit a big root-stem-ratio and underlined the relevance of root mass for yield. To promote vigorous root systems, proper balance of the nutrient elements N, P, and K is essential particularly to help plants to withstand biotic and abiotic stress. By means of management practices to overcome poor nutrition, abiotic stressors for the implication of asparagus decline, such as water insufficiency and malnutrition, can be resolved [55]. In this short-time experiment, we could not prove the consequences of limited water and nutrient uptake, but we show that the root system is heavily restricted when roots are attacked by the pathogenic Fusarium spp. and poor nutrition can also be initiated by diseased root systems that are restricted in water and nutrient uptake. This is why the prevention of lesions and injuries of the root system of asparagus is a main concern.
Most Fusarium species synthesise toxic secondary metabolites, known as mycotoxins, that are defined as evoking a toxic response when introduced in low concentrations to animals. It was shown for isolates of F. proliferatum from asparagus spears origin to produce moniliformin and fumonisin B1, and F. oxysporum to produce only moniliformin [57,58]. Moreover, the respective toxins can be found in basal parts of asparagus spears that were infected with the respective Fusarium species [59,60]. However, environmental factors affecting mycotoxin biosynthesis in the fungus and the roles of such mycotoxins in the infected plants are not yet fully understood [61,62]. The species-specific biosynthesis of such mycotoxins in the Fusarium asparagus interaction and its impact on root and shoot characteristics of asparagus plants should be the topic of a targeted future analysis.

Fusarium spp. Cultivation
The Fusarium spp. isolates used in this study are listed in Table 1. The isolates were provided either on synthetic nutrient-poor agar [52] or on potato-dextrose agar (PDA; Potato Dextrose Agar, Carl Roth GmbH, Karlsruhe, Germany) plates. Single-spore isolates were made of each isolate and cultivated on PDA in the dark at 25 • C for 10 days for the studies.

Fungal Degradation of Polysaccharides and Production of Extracellular Hydrolytic Enzymes
The Fusarium spp. isolates were tested for their growth on plant-derived polysaccharides by growing them on cellobiose (Fluka Honeywell Specialty Chemicals, Seelze, Germany), trehalose (Carl Roth GmbH), Avicel (50 µm size cellulose, Sigma-Aldrich, Darmstadt, Germany), pectin (Sigma-Aldrich), or starch (neoLab, Heidelberg, Germany). The media contained 0.5% of respective polysaccharides, 0.1% yeast nitrogen base (Carl Roth GmbH), 0.05% congo red (Carl Roth GmbH) and 1.5% agar (VWR, Darmstadt, Germany). The plates (Ø 9 cm) were inoculated by placing, at the centre of the plate, an agar plug that was taken from the rim of 10-day-old actively growing colonies of each isolate on PDA (Table 1). Isolates were incubated at 25 • C for 4 (cellobiose, trehalose, and starch) or 7 days (cellulose, pectin) on the respective agar media. Seven days of incubation on cellulose and pectin was chosen because of the slower growth of these media. The mycelial diameter was measured as an indicator for carbon utilisation, as well as the diameter of congo red discoloration as an indicator of the production of extracellular enzymes. Eight replicate plates were tested per isolate and per carbon source.

Species-Discrimination by MALDI-TOF Mass Spectrometry
Overnight cultures of Fusarium spp. isolates (Table 1), grown in potato-dextrose broth at room temperature under shaking conditions, were pelleted by centrifugation. Proteins contained in the mycelia were extracted using formic acid-based method [63]. A total of 1 µL of protein extract was spotted onto a polished steel target (Bruker Daltonik, Bremen, Germany) and air-dried before the addition of a saturated α-cyano-4-hydroxycinnamic acid solution as matrix. The MALDI method was calibrated using a bacterial test standard (Bruker Daltonik).
An ultrafleXtreme MALDI-TOF mass spectrometer (Bruker Daltonik), working in linear positive mode, was used for acquiring mass spectra in the range of m/z 200-20,000. Measurements were performed by flexControl v3.4 software (Bruker Daltonik). The MALDI Biotyper v3.1 software (Bruker Daltonik) was used to process the raw spectra, perform a database search against the Bruker Filamentous Fungi Library v1.0 (364 entries) for isolate identification, and to generate a hierarchical clustering of m/z values. Species-discrimination was performed three times with independently grown fungal cultures.

Plant Inoculation and Pathogenicity Testing of Fusarium Spp. Isolates on Asparagus Plants
Asparagus plants ('Backlim' and 'Rapsody') were carefully removed from the substrate. Roots were rinsed with tap water and prepared for inoculation with spore suspensions. The spore suspension of each isolate (Table 1) was made by flooding single-spore isolate cultures on PDA plates with sterile distilled water and filtering it through four layers of sterile cheesecloth. The number of conidia was microscopically counted (Zeiss Axioskop 2, Carl Zeiss Microscopy GmbH, Jena, Germany) using a haemocytometer and adjusted to 6-8 × 10 6 conidia mL −1 by dilution with sterile distilled water. Plant inoculation was performed by dipping roots for 30 min in 10 mL of the spore suspension of the respective Fusarium isolate and adding 9 mL of the spore solution around the stem after the plants were transplanted singly in pots (10 × 10 × 11 cm, 0.69 L) containing sterile substrate (sand:Fruhstorfer Erde type P 1:1 (vol/vol)). The control plants were inoculated in the same manner but with sterile distilled water. After inoculation, plants were grown in growth chambers as described above until disease assessment. The experiment was performed three times. Each experiment included three replications for each isolate and cultivar. One replicate comprised six plants of the respective cultivar. Pots were placed in a randomised complete block design. Eight weeks after inoculation, Fusarium symptoms were assessed on roots and stems. Roots were removed from pots and thoroughly rinsed in tap water. Disease severity (DS) was rated based on the percentage of affected stems and root system on a 1-5 scale, where 1 = 1 to 20%, 2 = 21 to 40%, 3 = 41 to 60%, 4 = 61 to 80% and 5 = 81 to 100% of stems showing chlorosis, necrosis or wilt (wilted stems were counted as DS 5) and of roots showing lesions [30]. The means and standard error of DS of stems and roots were calculated over the 9 replicates in total, i.e., from the three performed experiments each including three replications for each isolate and cultivar. After disease assessment, the plants were cut directly above the hypocotyl for fresh and dry weight estimation. Stems and roots were dried at 80 • C for 48 h. The means and standard error of fresh and dry weight were calculated over the nine replicates of each treatment and cultivar.
In addition to the molecular detection of Fusarium spp. in plant tissue, one plant out of each replicate was used for the re-isolation of Fusarium spp. isolates on plates to fulfil Koch's postulate. Root pieces (2-3 cm) were washed twice with tap water and dipped for 2 min in ethanol (2 mL, 70%) and thereafter dipped for 8 min in NaOCl (1 mL, 5% + Tween20) and finally rinsed with sterile water. The treated root pieces were then placed on PDA amended with streptomycin sulphate (50 mg L −1 ), penicillin-G-potassium (100 mg L −1 ) and tetracycline hydrochloride (10 mg L −1 ) and incubated at 25 • C for 2 days in the dark. Fusarium was determined by a visual and microscopic analysis of conidia [65].

Molecular Detection of Fusarium Spp. Isolates in Plant Tissue
A 1-cm piece of each stem was cut directly at the stem basis of each plant for pathogen DNA detection and gene expression analysis. Samples per plant were immediately frozen in liquid nitrogen and stored at −80 • C. Samples were taken from five plants per replicate of each treatment. Plant material was ground in liquid nitrogen in an orbital ball mill for 2 min at a frequency of 30 Hz s −1 (MM400 Retsch GmbH, Haan, Germany) with three balls of stainless steel with a 5-mm diameter.
Total DNA was extracted from 50 mg ground tissue using the DNeasy Plant Mini Kit (Qiagen GmbH, Hilden, Germany). DNA was quantified spectrophotometrically at 260 nm (Infinite M200PRO, Tecan Trading AG, Männedorf, Switzerland) using the NanoQuant plate, and quality was checked using absorption at 260 and 280 nm with a ratio >1.8 as acceptable. For the fungal DNA standard curves, mycelium was harvested from single-spore isolates grown for 5 days in liquid culture in PD medium in the dark at 25 • C. DNA from selected isolates per Fusarium spp. were extracted from 50 mg mycelium using the same methods. The standard curves were prepared using serial diluted samples of 2.

Analysis of the Root Morphology
The root systems of two randomly chosen plants of each replicate and cultivar of two experiments were used for imaging analysis. The data of the two experiments were pooled. To avoid overlapping, the washed roots were carefully separated on top of a DIN A4 sized bright frame and were imaged with a stage mounted digital camera (Canon EOS60 + EF 17-40 mm). Approximately 5 RAW images were obtained from each root tray with the following settings: aperture 10, ISO setting: 100 and focal length: 38 mm. The effective image resolution was about 450 DPI. A processing pipeline was realised by a MATLAB script (MATLAB. (2017) version 9.3 (R2017b). The MathWorks Inc., Natick, MA, USA) that provided cleaned and margin-free binary root images to IJ-Rhizo (Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, MD, USA, https://imagej.nih.gov/ij/, 1997-2018) [67] and performs additional pre-and post-processing steps. IJ-Rhizo was used to measure the root length of different diameter classes.
This automated processing pipeline consisted of the following steps: keyboard macrobased RAW-file conversion (Digital Photo Professional, 2011), median filtering over a tif-converted raw image file stack (n~5 images), margin detection and removal via hough transform, root segmentation and dirt detection and removal. Root segmentation and dirt detection was based on fitted Gaussian mixture models and corresponding posterior probabilities. Obtained root diameter versus length histograms from IJ-Rhizo were cumulated over all subsamples per plant and then discriminated into fibrous and (thick) storage roots using the Otsu segmentation method [68]. Using this diameter threshold and adopting a cylinder geometry for roots, length, volumes and diameters were computed for both root categories from diameter versus length frequency table. Estimated root volumes showed a good correspondence (r 2 = 0.74; p < 0.05, n = 30) to root fresh weight in a test set.

Statistical Analyses
The data of DS were analysed by rank-based ANOVA according to the ANOVA-type statistics [69] using SAS PROC RANK and PROC MIXED with the ANOVAF option (SAS Institute 2016, Cary, NC, USA). Comparisons of means of the replicates (n = 9) were performed using χ 2 test (p < 0.0001). Based on the multiple test procedure of comparisons of means p values were adjusted according to Holm [70]. Least-squares means were normalised to interval [0, 1] and used for correlation analyses with means of root traits (p < 0.05). The data of fresh and dry weight of roots and stems (n = 9), total root length, root volume, root surface area, and root radius (n = 6) were analysed by two-way ANOVA according to a randomized complete block factorial design. Comparisons of means of the replicates were performed using Tukey's test (p < 0.05).
For the statistical testing of the fungal utilisation of plant-derived polysaccharides, one-way ANOVA was applied. The identification of differences among the isolates within the substrate groups was performed with Tukey test (p < 0.05). Analyses of variance (ANOVA) were conducted using STATISTICA 13.5 software (StatSoft Inc., Tulsa, OK, USA).

Supplementary Materials:
The following are available online at http://www.mdpi.com/2076-0817/9/6/509/s1, Figure S1: Extracellular enzymatic activity of Fusarium spp. isolates on plates. Figure S2: Typical symptoms caused by Fusarium spp. isolates on roots. Figure S3: Disease severity (DS) of roots of Asparagus officinalis. Figure  S4: Disease severity of ferns of Asparagus officinalis. Figure S5: Relationship between concentration of fungal DNA in stem bases of Asparagus officinalis and root fresh weight after inoculation with Fusarium spp. Figure