Impacts of previous crops on inoculum of Fusarium culmorum in soil, and development of foot and root rot of durum wheat in Tunisia

Copyright: © 2020 E. Khemir, S. Chekali, A. Moretti, M.S. Gharbi, M.B. Allagui, S. Gargouri. This is an open access, peer-reviewed article published by Firenze University Press (http:// www.fupress.com/pm) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


INTRODUCTION
In Tunisia, cereals and their by-products are the main sources of dietary calories, and are the common base of all diets and the historical base of the Mediterranean diet. The average per capita consumption of cereals in Tunisia is 184 kg per year (Slama et al., 2005;ONAGRI, 2016). Cereals are grown on approx. 1 million hectares of the agricultural land, with the number of the farmers involved estimated to be approx. 250,000. About 63% of farmers are smallholders, each with land areas of less than 10 ha (Bachta, 2011;Anonymous 2, 2018). Fusarium foot and root rot (FFRR) has been recognized as one of the most important diseases of cereals in Tunisia since the 1970s (Ghodbane et al., 1974), and this disease is responsible for significant economic losses, especially in arid and semi-arid regions (Van Wyk et al., 1987;Bateman, 1993;Hollaway et al., 2013). Yield losses of up to 26% in durum wheat and 18% in barley have been recorded (Chekali et al., 2013).
FFRR is difficult to manage since F. culmorum can survive as hyphae in stubble residues of cereals (Cook, 1968;Bateman and Murray, 2001;Burgess, 2011;Khemir et al., 2018) and other grasses (Wallwork et al., 2004), and has the ability to persist as chlamydospores in soil (Sitton and Cook, 1981). This survival ability has major implications for designing strategies for effective disease management. Control of FFRR relies on agronomic management strategies, since in-crop fungicides are largely inefficient, and host resistance is limited (Burgess et al., 2001;Pereyra and Dill-Macky, 2004;Wisniewska and Kowalczyk, 2005).
The market situation in Tunisia has influenced the diversity of crops that are grown, where continuous wheat cropping and short crop rotations have become popular during the last 30 years. These cropping systems promote the increase of soil-borne diseases and resulting yield losses. In addition, durum wheat, which is highly susceptible to FFRR (Burgess et al., 2001), is the dominant crop, representing approx. 60 % of the cereal growing area (Slama et al., 2005;Gharbi and Felah, 2013). In the context of climate change, this disease could become more important, especially as the Mediterranean region has been qualified as a "hot spot for climate change" (Giorgi, 2006;Vicente-Serrano, 2006;Anonymous 1, 2019). It is important, therefore, to consider the best cultural practices for FFRR control. Crop rotations with non-hosts such as legumes has been reported to efficiently control the disease in many countries including the United Kingdom (Bateman and Kwasana, 1999;Bateman and Murray, 2001), Australia (Felton et al., 1998;Kirkegaard et al., 2004;Evans et al., 2010) and South Africa (Lamprecht et al., 2006).
Very little research has been conducted in the Mediterranean basin, including Tunisia (Chekali et al., 2016), to evaluate the impacts of break crops on development of FFRR. Fenugreek, an annual legume that grows well in Mediterranean climates (Duke et al., 1981), and which has demonstrated strong insecticidal, nematicidal and antifungal activity (Pemonge et al., 1997;Zia et al., 2001;Evidente et al., 2007;Haouala et al., 2008;Omezzine et al., 2014), has not been evaluated for reducing F. culmorum levels in soil, and its impacts on the disease. Very few studies have focused on monitoring F. culmorum soil inoculum under different cropping systems (McKenzie and Taylor, 1983;Evans et al., 2010).
The present study aimed to assess: i) the potential benefits of previous crops for reducing F. culmorum populations in the soil; ii) the impacts of previous crops on the incidence and severity of FFRR on stems and roots of durum wheat; and iii) the relationships between preplanting inoculum of F. culmorum and FFRR disease expression and grain yields.

Site characteristics and cultural practices
To understand the effect of previous crops on populations of F. culmorum in the soil and incidence of FFRR in durum wheat, a field trial and three demonstration trials were established during the cropping seasons 2012/13, 2013/14 and 2014/15, in Northwest Tunisia.

Field trial
The field trial was carried out in a farm field located at Bou Salem (N36 53.012 E9 31.251). The area has a typical Mediterranean climate (Köppen, 1936), with hot summers and cold, wet winters. The monthly rainfall during the three cropping seasons at Bou Salem is summarized in Table 1, where the 10-yr annual average rainfall is approx. 500 mm.
The trial was established in 2012/13 as a randomized complete block design (RCBD) with three crop-ping sequences and four replications. An area of 7,500 m 2 was divided into four blocks. Each block was subdivided into three plots (each of 12 × 50 m) which were sown with durum wheat (Triticum durum) var. 'Karim', faba bean (Vicia faba) var. 'Bachar' or fenugreek (Trigonella foenum-graecum) var 'Rihana'. The impacts of previous crops were assessed during the two following cropping seasons of 2013/14 and 2014/15. Durum wheat was sown at the recommended rate of 150 kg ha -1 ; faba bean was sown at 140 kg ha -1 and fenugreek at 35 kg ha -1 . All these crops were seeded in early to mid-November of each year. Diammonium phosphate (P) (150 kg ha -1 ) was applied with seed (Zadoks growth stage (ZGS) 1; Zadoks et al. 1974), and ammonium nitrate fertilizer (N) (150 kg ha -1 ) was applied at tillering (ZGS 21) and at stem elongation (ZGS 32). For uniform suppression of grass weeds in the faba bean and fenugreek plots, clethodim (Select super®) was applied at 1 L ha -1 post-emergence of grass weeds from the 3-leaf stage of the crop plants.

Demonstration trials
Demonstration trials were established in three farmer's fields (36°37'29.28'N, 8°42'04.56'E; 36°36'47.56'N, 8°34'39.42'E and 36°37'01.10'N. 8°40'35.90'E) at Fernana. This region is also characterized by hot summers and cold, wet winter, but with greater annual rainfall than at the field trial site. The monthly rainfall during the three cropping seasons is summarized in Table 2, and the 15 y average rainfall was approx.780 mm at this location.
In 2012/13, an area of 3,000 m 2 at each farm was divided into three plots, which were each sown with faba bean (Vicia faba) var. 'Bachar', durum wheat (Triticum durum) var. 'Karim' or a local variety of vetch (Vicia sativa). The plots were 10 m wide and 100 m long. The effects of previous crops were evaluated during the two subsequent cropping seasons (2013/14 and 2014/15). These trials were direct seeded (without tillage) to retain uniform and high inoculum levels of F. culmorum at the beginning of the study. The three previous crops were applied according to common farmer practices. Durum wheat was seeded at 150 kg ha -1 , faba bean at 130 kg ha -1 and vetch at 100 kg ha -1 . Diammonium phosphate (P) (150 kg ha -1 ) was applied with seed (Zadoks growth stage (ZGS) 1; Zadoks et al. 1974), and ammonium nitrate fertilizer (N) (150 kg ha -1 ) was applied at tillering (ZGS 21) and at stem elongation (ZGS 32).

Fusarium culmorum inoculum density
Soil sampling Soil samples were collected in July 2013, following the crop sequence treatments from the 12 plots of Source: National Institute of Meteorology. the field trial at Bou Salem, and the three plots from each of the demonstration trials at Fernana. One hundred soil cores, including any plant residus, were taken within each plot to a depth of 100 mm, using a 15-mm diam. Accucore corer (Spurr Soil Probes), and were then bulked to provide a single sample for each plot. The samples were air-dried for several days. Dry soil (500 g) for each plot was sent to the South Australian Research and Development Institute (SARDI), Adelaide, Australia, for assays of F. culmorum DNA ('Predicta® B' Soil Testing Service). The remaining soil from each plot was used for soil dilution plating (below).

Soil dilution plating
This was carried out to estimate populations of F. culmorum (Cook, 1980), following the procedure of Snyder and Nash (1968), with modifications as described by Bateman and Coskun (1995), Bateman et al., (1998) and Bateman and Murray (2001). Briefly, 10 g of soil from each plot were added to 100 mL of 0.1 % water agar and mixed thoroughly for 30 sec. Further dilution series were made using 1 mL aliquots from the starting suspension. An aliquot of 1 mL of the final dilution (1/100, based on premilary tests) was transferred to each of five Petri dishes containing Peptone-PCNB agar medium. This medium contained 15 g Difco peptone, 20g agar, 1 g KH 2 PO 4 , 0.5g MgSO 4 7H 2 O, 1 g pentachloronitrobenzene (PCNB), 100 mg L -1 streptomycin sulfate and 80 mg L -1 neomycin. Aliquots were dispersed by circular agitation of the Petri dishes (Steinkellner and Langer, 2004). The Petri dishes were incubated at 20°C for 5-7 d with a 16 h light/8 h dark regime. Fusarium culmorum was identified morphologically according to Leslie and Summerell (2006).
To confirm fungal identifications, some colonies were transferred to ¼ strength potato dextrose agar (PDA) and then to carnation leaf agar (CLA). Amounts of F. culmorum in the soil samples were expressed as the numbers of colony forming units per gram of air-dried soil (CFU g -1 soil).
CFU g -1 was calculated using the formula: Number of colonies × dilution factor CFU g -1 = Volume of culture plate With dilution factor = 10 2 , and volume of culture plate = 1 mL.

Quantitative real-time polymerase chain reaction (qPCR)
The amount of F. culmorum DNA present in soil from each plot was evaluated, after applying the crop sequence treatments (July 2013). Soil samples were submitted to the SARDI Root Disease Testing Service. This includes a suite of tests for the quantification of F. culmorum DNA in soil. QPCR assays using rDNA (TaqMan) probe sequences specific to F. culmorum were made for the total pathogen DNA extracted from the soil. Levels of F. culmorum DNA in soil can be related to disease risk in the following crop year (Hogg et al., 2007). Extraction protocols and analyses of DNA from soil were done according to Ophel-Keller et al. (2008). Inoculum concentrations of DNA per gram of soil were expressed in pictograms.

Pathogen isolation and disease incidence
During the 2013/14 and 2014/15 seasons, 50 plants of durum wheat were collected randomly along a W-shaped transcript, from each of the 12 plots of the field trial, and from the nine plots of the three demonstration trials, at the durum wheat anthesis stage (ZGS 85). The incidence of FFRR was calculated as the frequency of isolation of F. culmorum, separately from roots or stem bases.
Stem base sections and root sections (2-3 mm length) were surface-sterilized with 70% ethanol for 15 s, followed by 3% sodium hypochlorite for 1min, and then rinsed three times with distilled water, and were then air-dried on sterile filter paper. Five sections from each sample were then plated onto ¼ strength PDA containing 20 mL L -1 streptomycin sulfate (100 mg L -1 ) and 12 mL L -1 neomycin sulfate (80 mg L -1 ) in Petri dishes. The plates were incubated up to 7 days at 25°C in a 12 h light/12 h dark regime. Hyphae of Fusarium spp. were subsequently transferred onto CLA to promote conidium production. Fusarium culmorum was identified using morphological criteria, according to Leslie and Summerell (2006).

Disease severity
FFRR disease severity was assessed only during the 2014/15 cropping season in the three farm demonstration trials, at the durum wheat anthesis stage (ZGS 85). Severity of F. culmorum was evaluated by measuring surface of stem bases showing browning. Fifty durum wheat plants were randomly removed from each plot and scored using a scale 0 to 3, where: 0 = no browning (no visible symptoms); 1 = 0.1 to 1.9 cm extent of browning; 2 = 2 to 3.9 cm of browning; and 3 = > 4 cm browning.
Ratings were converted to severity indices using the following formula: Where: N = number of evaluated plants; n 0 = number of plants in class 0; n 1 = number of plants in class 1; n 2 = number of plants in class 2; and n 3 = number of plants in class 3.

Yield parameters
Grain yields of durum wheat, subjected to different crop rotations, were assessed from the three demonstration trials in the 2013/14 and 2014/15 cropping seasons. The yields (kg ha -1 ) were based on harvesting all of the grain from each plot.

Field trial
The trial was arranged in a randomized complete block design (RCBD) with four replicates. All data were subjected to analysis of variance (ANOVA) using 'SPSS Statistics version 20' software published by 'IBM Crop 2011'. The year and previous crop were considered as main factors. Means comparisons were performed using the Student's LSD test (at P = 0.05 or P = 0.01).
Regression analyses used Pearson's correlation test, to identify possible associations between the incidence of F. culmorum isolated from stem bases and roots (2013/14 cropping season data) and the amounts of F. culmorum in the soil (DNA and CFU) left after each treatment, measured at plant maturity in July 2013.

Demonstration trials
Regression analyses were carried out (as above), to identify possible correlations between F. culmorum DNA content in soil (determined by qPCR), the number of CFU of F. culmorum per gram of soil (determined by soil dilution plating) and the incidence of F. culmorum isolations from stem bases and roots of durum wheat (2013/14 cropping season data). In addition, regression analysis was used to identify possible association between FFRR expression and grain yields from subsequent durum wheat crops.

Fusarium culmorum inoculum in the soil
Soil dilution plating and morphological identification revealed several Fusarium species from the soil samples. These included F. acuminatum, F. avenaceum, F. culmorum, F. compactum, F. equiseti, F. oxysporum and F. solani. However, F. culmorum was by far the dominant pathogenic species in the samples. The DNA analyses conducted for detection of the two recognized pathogenic species, F. culmorum and F. pseudograminearum, showed the presence of only F. culmorum in the samples.

Field trial
Soil dilution plating revealed the presence of F. culmorum in all 12 plots of the field trial at Bou Salem. ANOVA of data from the trial revealed that the number of CFU of F. culmorum in the upper 10 cm soil layer was significantly (P = 0.002) affected by the different break crops. Faba bean and fenugreek used as previous crops, vs. durum wheat monoculture, reduced the number of F. culmorum propagules in soil by 83% for faba bean and 85% for fenugreek (Figure 1). The number of propagules ranged from 133 to 1,733 CFU g-1 of soil, with a mean of 1100 for durum wheat, 183 for faba bean and 166 for fenugreek. DNA analyses confirmed the presence of F. culmorum in soil from the 12 plots of the field trial. The amounts of DNA of the pathogen were significantly (P = 0.004) different among previous crop treatments, ranging from 27 to 212 pg DNA g -1 of soil. Overall, the amounts of F. culmorum DNA in the soil were decreased by 58% following faba bean and by 65 %, following fenugreek. The greatest amounts of F. culmorum DNA occurred in the treatment where durum wheat was cultivated in monoculture, and here was no statistically significant difference in DNA amounts between the two legume treatments (Figure 2). Soil dilution plating and qPCR both showed greater amounts of F. culmorum in soil following a previous durum wheat crop, compared to previous faba bean or fenugreek crops. Pearson's correlation test showed that F. culmorum DNA concentration (pg DNA g -1 soil) was correlated (r = 0.92; P < 0.001) with the number of propagules of this fungus (CFU g -1 soil) detected in the soil 1 year after treatment.

Demonstration trials
Based on soil dilution plating, propagules of F. culmorum in the soil were more abundant after wheat (average of 866 propagules g -1 of soil) compared to vetch (358 propagules g -1 of soil) or faba bean (289 propagules g -1 of soil) (Table 3). Thus, faba bean and vetch, used as previous crops, reduced the number of F. culmorum propagules in soil by 66 % after faba bean and 58 % after vetch, compared to durum wheat monoculture. In addition, the results indicated that F. culmorum propagules were less abundant after faba bean as a previous crop compared to vetch.
The DNA analyses showed greater amounts of F. culmorum DNA in soil from Farm 1 and 2 compared to Farm 3 (Table 3). Based on real-time qPCR, the amount of DNA of F. culmorum was less following a legume compared to wheat at Farms 1 and 2. However, at Farm 3, where the DNA amounts were low in all cases, there were no differences in F. culmorum DNA between the different previous crops. Based on the SARDI risk assessment protocol, DNA amounts for all plots at Farm 3 were classified as leading to low risk of subsequent crop damage.
Overall, for the 12 plots of the field trial and the six plots of the demonstration trials at Farms 1 and 2, legumes used as previous crops reduced the inoculum of F. culmorum, as detected by qPCR and by soil dilution plating.

Disease assessments
A total of 4,200 stems and roots were examined to determine the impacts of crop sequence on soil-borne diseases of cereals in the field and demonstration trials. Visual assessments revealed evidence of take-all, eyespot and FFRR. Only the results for previous crop effects on FFRR development are presented here.

Field trial
Incidence of infection by F. culmorum in durum wheat was greater in durum wheat stem bases than in roots. Statistical analyses showed that there was a difference (P = 0.047) between the two growing seasons (2013/14 and 2014/15) for incidence of F. culmorum recovered from durum wheat stem bases. No significant difference was detected in the incidence of F. culmorum recovered from wheat roots during these two cropping seasons (Figure 3).
ANOVA of data from the field trial revealed that there was a highly significant (P < 0.001) effect of previous crop on the incidence of FFRR infection of stem bases and roots of durum wheat. Fenugreek or faba  bean used as the previous crop decreased the incidence of F. culmorum affected durum wheat roots more than 3× for fenugreek and 9× for faba bean, compared to durum wheat. Similarly, fenugreek and faba bean decreased the incidence of F. culmorum in durum wheat stem bases, by more than 3× for fenugreek and 5× for faba bean. Data analyses showed that the incidence of infection by F. culmorum on durum wheat stem bases and roots was least (P < 0.001) when faba bean was the previous crop ( Figure 4).

Demonstration trials
Data from the three demonstration trials showed that the presence of F. culmorum in durum wheat roots and stem bases when they were sampled dur-ing the 2013/14 and 2014/15 cropping seasons. Assessments of development of FFRR showed that stem bases were much more likely to be infected by F. culmorum than roots. Faba bean and vetch, used as previous crops, resulted in a less isolation incidence of F. culmorum than for durum wheat monoculture. Compared to durum wheat, the mean reduction in F. culmorum incidence after faba bean was 59% in stem bases and 68% in roots, and after vetch was 62% in stem bases and 87% in roots ( Figure 5).
FFRR severity was evaluated only in spring 2015 (2014/15 cropping season). Compared to durum wheat monocropping, vetch used as the previous crop reduced disease severity by 80% and faba bean as the previous crop reduced severity by 50 % (Figure 6).

Field trial
According to Pearson's correlation test, the amounts of F. culmorum DNA detected in soil samples 1 year after treatment were positively correlated with the incidence of F. culmorum, durum wheat roots (R 2 = 0.56, P = 0.005) and in stem bases (R 2 = 0.80, P = 0.001) for wheat grown in the following year (Figure 7).
The association between pre-planting inoculum concentrations, indicated by F. culmorum CFU g -1 of soil, and disease expression, indicated by incidence of F. culmorum in wheat roots, was significantly positive (R 2 = 0.58, P = 0.004) ( Figure 7A). This was also true for this relationship in stem bases (R 2 = 0.83, P = 0.000) ( Figure 7B).

Demonstration trials
Inoculum levels of F. culmorum detected in the soil after each treatment, as determined by soil dilution plating, were positively correlated with incidence of F. culmorum in wheat roots (R 2 = 0.66) and stem bases R 2 = 0.64 (Figure 8, A and B). In addition, a positive relationship was observed between the amounts of F. culmorum DNA and the incidence of F. culmorum in wheat roots (R 2 = 0.65) and stem bases (R 2 = 0.7) after each treatment ( Figure 8, A and B).

DISCUSSION
In this study effects of previous crops on inoculum levels of F. culmorum in soil, development of FFRR, and yield parameters of durum wheat were evaluated in Tunisian farming conditions. Data were collected from an experimental trial, and from demonstration trials established in three farmer's fields.
Inoculum levels of F. culmorum in the soil were quantified using two methods, soil dilution plating and qPCR. Both methods showed the dominance of F. culmorum in the upper 10 cm soil layers of the trial sites. Fusarium pseudograminearum, which has been isolated in other studies from Tunisia (Gargouri et al., 2001(Gargouri et al., , 2007Chekali et al., 2016), was not detected in the present study. However, this is not surprising since it has been demonstrated that F. pseudograminearum is restricted to arid zones in Tunisia (Gargouri et al., 2001), whereas in more humid regions, such as Fernana and Bou Salem where this study was conducted, F. culmorum is the dominant species. These observations have also been confirmed in other regions of the world, including    the Pacific Northwest of United States of America, and in Australia (Backhouse et al., 2004;Poole et al., 2012).
The results presented here demonstrated that break crops, including faba bean, vetch and fenugreek, reduced the number of propagules of F. culmorum in the soil. These data are confirming previous results which showed that populations of F. culmorum in the soil are affected by the previous crops (Snyder and Nash, 1968;McKenzie and Taylor, 1983). Steinkellner and Langer (2004) also demonstrated that the number of Fusarium CFU in soil was affected by previous crops.
Overall, the results from soil dilution plating were confirmed by qPCR, except for those from Farm 3 in the demonstration trials. At this farm, the amounts of F. culmorum DNA detected were very low, and this may have affected the qPCR analyses. The correlations observed between the two techniques suggest that both are effective for quantifying inoculum levels of F. culmorum in field soil. To our knowledge, this is the first report examining correlations between soil dilution and qPCR data for F. culmorum and durum wheat. In addition, this study was the first to measure the levels of F. culmorum inoculum in Tunisian soils, and may have been the first to make these assessments in the Mediterranean region.
The amount of F. culmorum DNA in the soil was generally reduced when legumes were used as previous crops before durum wheat. This is consistent with the results of Evans et al. (2010) in south-east Australia, who demonstrated that DNA of F. culmorum in soil was much less after field pea (115 pg DNA g -1 ) and greater after durum wheat (974 pg DNA g -1 ) or barley (1,196 pg DNA g -1 ).
The numbers of propagules of F. culmorum detected in the present study (0 to 1,974 CFU g -1 of soil) was low compared to other studies. In Austria, Steinkllner and Langer, (2004) reported that inoculum of F. culmorum and F. pseudograminearum in the upper 10 cm layer, ranged between 0 to 8,750 CFU g -1 . Similarly, the amounts of F. culmorum DNA recorded in the present study were less than those found in Australia (Evans et al., 2010;Halloway et al., 2013).
Development of soil-borne diseases depends on the concentration of the inoculum in the soil (Cook, 1981). Cook (1968) demonstrated that 100 propagules of F. culmorum per gram of soil were sufficient to cause damage even when host plants are not under stress. In addition, McKay et al. (2008) suggested guidelines that related DNA amounts of F. culmorum in the soil before sowing to the risk of FFRR and yield losses in bread and durum wheat. They indicated that pre-planting DNA soil amounts less than 100 pg of fungal DNA g -1 of soil for bread wheat and less than 25 pg for durum wheat, would lead to minimal risk of yield losses due to crown rot caused by Fusarium spp.
The low amounts of F. culmorum inoculum in the soil may be explained by several factors. Differences in climatic conditions (Doohan et al., 2003;Xu et al., 2018), tillage methods (Steinkellner and Langer, 2004) and various cultural practices (Klem et al., 2007;Muller et al., 2010;Blandino et al., 2012;Chekali et al., 2016;Hemissi et al., 2018) are known to affect Fusarium populations in the soil. However, the low levels of inoculum observed in the present study were likely related to soil tillage. Lowor no-till are not common practices in Tunisia. Even in the demonstration trials reported here, direct seeding was applied for the first time. Steinkellner and Langer, (2004) found that the total number of CFU of Fusarium spp. was affected by tillage treatment. This agrees with previous studies by Toledo-Souza et al. (2008), who demonstrated that Fusarium inoculum was greater in a notillage system than a conventional one. Paulitz (2006) also showed that FFRR can cause significant damage when no-tillage practices and stubble retention are used. Recently, in Tunisia, Khemir et al. (2018) showed that stubble residue retention on soil surfaces increased F. culmorum inoculum levels in plant residues. The impact of no-tillage was also previously reported by Evans et al. (2010), who demonstrated that tillage reduced inoculum levels in soil by burying and promoting rapid degradation of infested plant residues. Furthermore, type and quantity of crop residues left on the soil surface, influence microorganisms and microbial processes in soil (Kandeler et al., 1999). Yi et al. (2002) showed that decomposition of crop residues was associated with a decline in CFU of Fusarium species.
The present study showed that frequency of isolation of F. culmorum from wheat roots and stem bases was affected by the previous crop in both trials, and was reduced by up to 70%. However, the results showed that in general the incidence of the pathogen was greater in stem bases than in roots. Knudsen et al. (1995) reported that FFRR was a consequence of initial stem base infections, while Beccari et al. (2011) and Covarelli et al. (2012) found that the stem bases of durum wheat were more heavily infected by F. culmorum compared to the roots. Knight and Sutherland (2013;2017) reported that F. culmorum caused minor necrosis of the primary roots of cereals (durum wheat, bread wheat and barley), but caused serious cortical rot in leaf sheath tissues. These results emphasize the importance of crop residues left on the soil at the level of plant crowns for infection by the Fusarium species and the development of FFRR.
The use of break crops, including faba bean, fenugreek and vetch, decreased the incidence and sever-ity of FFRR and increased grain yield in the subsequent durum wheat crop. These results are consistent with previous studies (Felton et al., 1998;Montanari et al., 2006;Zhou and Everts, 2007;Evans et al., 2010;Chekali et al., 2016). Other studies also demonstrated that preceding crops are key factors increasing the risk of Fusarium diseases on cereals (Klem et al., 2007;Muller et al., 2010;Blandino et al., 2012;Chekali et al., 2016). Rotations with chickpea (Felton et al., 1998;Chekali et al., 2016), faba bean (Kirkegaard et al., 2004), pea Evans et al., 2010) or lentil (Chekali et al., 2016) have been reported to reduce the incidence of FFRR. Chekali et al. (2016), showed, in northwest of Tunisia, that faba bean and chickpea used as previous crops could contribute to decreased incidence of infection of roots and stem bases of durum wheat by F. culmorum, and increase grain yields compared to durum wheat as a previous crop. Rasmussen et al. (2002) hypothesized that this reduction in F. culmorum could be due to the non-host character of legumes, or to their high cellulose contents which increases microbial activity in the soil and leads to decrease of Fusarium spp. survival. In contrast, a high level of infection of wheat by F. culmorum has been reported when monoculture is practiced (Blecharczyk et al., 2006;Kurowsk et Adamiak, 2007;Kraska et Mielniczuk, 2012). In the present study, the Mediterranean forage legume fenugreek was reduced soil inoculum levels of F. culmorum by more than 60%, and decreased F. culmorum incidence on durum wheat stem bases, compared inoculum and disease incidence in durum wheat monoculture. These results confirm the antifungal activity of fenugreek. Omezzine et al. (2014) assessed in vitro antifungal activity of aqueous extracts from T. foenumgraecum, against F. oxysporum f. sp. radicis-lycopersici and F. oxysporum f. sp. lycopersici, and reported the antifungal and allelopathic potential of extracts from aerial parts of a Tunisian fenugreek cultivar. Other studies have also reported similar activities (Devasena and Menon, 2003;Oddepally and Guruprasad, 2015;Dharajiya et al., 2016;Sudan et al., 2020). Here, we have demonstrated, for the first time in Tunisia, the beneficial effects of fenugreek as a previous crop to decrease F. culmorum populations in soil and the development of FFRR in durum wheat.
Our results showed positive correlation between F. culmorum inoculum concentrations in the soil prior to sowing and subsequent FFRR expression. A negative correlation was also observed between FRRR expression and durum wheat grain yield. These results are similar to those of Smiley et al. (2005), Hollaway et al. (2013) and Chekali et al. (2016). Hollaway et al. (2013) suggest-ed that pre-planting amounts of F. culmorum inoculum and expression of FFRR were positively correlated. Given that control options for FFRR are limited, this information would be valuable for grain producers, to prevent the yield losses caused by these diseases. In Australia, a DNA-based soil testing service (the 'PreDicta® B' system) was developed in 1997, to improve assessment of risks from root diseases and assist grain producers in planning cropping programmes . Previous applications of this system have been mainly utilized in Australia and the United States (Paulitz et al., 2010;Bithell et al., 2012). This analysis tool is currently not available in Tunisia or in other Mediterranean countries. Our results encourage the use of systematic surveys across years and regions to document Fusarium inoculum levels in soil and FFRR occurrence and severity, and grain yield of durum wheat. This will assist development of a predictive model useful to farmers in Tunisia and the wider Mediterranean region.
In this study, the demonstration trials which involved farmers may impact on extension workers and other grain production stakeholders, to promote adoption of crop rotation as a tool for reducing soil-borne diseases of wheat in Tunisian cropping systems. Further studies would be useful to develop and implement a predictive model based on Fusarium inoculum levels in soil, to assess the risks and estimate the potential yield losses resulting from FFRR. This could be achieved by adjusting cropping sequences and other cultural practices to maximize yields. Our results could be useful for the Mediterranean basin, where FFRR is an endemic disease wherever wheat is cultivated.