Broomrapes (Phelipanche spp. and Orobanche spp.) are holoparasitic plants that parasitize many dicotyledonous crops, causing tremendous losses in crop yield and quality worldwide (Joel et al., 2007). The parasite produces tiny seeds (0.2–0.3 mm) that remain viable in the soil for many years and germinate in response to substances—mostly are strigolactones—secreted from plant roots (Xie et al., 2010). The germinating seed produces a radicle that grows and reaches the root surface and produces a haustorium (the attachment stage). Intrusive haustorium cells penetrate the host root via enzymatic activity and physical force and form connections between the vascular systems of the host and parasite. The parasite develops into a tubercle that grows underground on the host root surface, sucking all of its nutrient needs from the host. Toward the end of its life cycle, the parasite develops flowering shoots that emerge aboveground, producing hundreds of thousands of seeds per plant (Parker and Riches, 1993).

Several methods were suggested for broomrape management that may be classified to resistant varieties, cultural, physical, biological, and chemical control by herbicides (Dhanapal et al., 1996; Ali, 2007). Resistant varieties and chemical control are probably the most commonly used commercially. Effective chemical control of broomrape is extremely difficult to achieve because these plants are anatomically and physiologically connected to the host. In many cases, when a broomrape sensitive crop is grown for the first time, the farmer is unaware of the parasite’s presence or infestation level in the field until broomrape shoots appear aboveground, at which point it is too late because most of the damage has already been done. In addition, control of aboveground shoots is much more difficult than at early developmental stages (Eizenberg et al., 2013). For effective control with systemic herbicides, the host should be highly selective to the herbicide without reducing the latter’s phytotoxicity as it passes through the host conductive tissues toward the roots (Joel et al., 2007). Broomrapes are chlorophyll-lacking parasites. Therefore photosynthesis-inhibiting herbicides cannot be used because there is no target site. To date, only amino acid biosynthesis-inhibiting herbicides have been used successfully to control broomrapes at a commercial level. These include imidazolinones, sulfonylureas, and glyphosate (Eizenberg et al., 2013). Imidazolinones and sulfonylureas inhibits acetolactate synthase (ALS), also known as acetohydroxy acid synthase (EC 4.1.3.18), a key enzyme in the biosynthesis of the branched-chain amino acids leucine, isoleucine and valine – HRAC group B (Herbicide Resistance Action Committee [HRAC], 2010). Glyphosate inhibits the enzyme 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS; EC 2.5.1.19), which is essential for the synthesis of the aromatic amino acids phenylalanine, tryptophan and tyrosine – HRAC group G (Schönbrunn et al., 2001; Duggleby et al., 2008; Herbicide Resistance Action Committee [HRAC], 2010).

Phelipanche aegyptiaca is a devastating parasite on tomato (Solanum lycopersicum L., synonyms Lycopersicon esculentum Mill.), that endangers the existence of the processing tomato industry in Israel, among other countries (Hershenhorn et al., 2009). Sulfosulfuron (SS) has been registered in Israel for P. aegyptiaca control in tomato. Its usage involves three split applications with precise timing, each application followed by upper irrigation (Hershenhorn et al., 2009). The herbicide is not active when applied on tomato foliage, probably due to detoxification by a P450-type oxidase enzyme (Buker et al., 2004). To be effective, the herbicide should be washed into the soil so that it can establish direct contact with the germinating P. aegyptiaca seeds (Eizenberg et al., 2004). This method of control has several drawbacks. Any deviation from the prescribed application timing may drastically reduce control efficacy (Hershenhorn et al., 2009). Upper irrigation systems are no longer in use for tomato growth as all fields are drip-irrigated, so the farmer must invest in expensive irrigation equipment such as moving pivots. In addition, SS has a long residual effect in the soil, jeopardizing proper crop rotation. The upper irrigation, which enhances disease development, the expensive equipment needed, the precise timing and the long residual effect of SS in the soil considerably limit the number of fields in which this method is implemented in Israel.

Maleic hydrazide (MH; 1,2-dihydro-3,6-pyridazinedione) has been extensively used as a commercial systemic plant growth regulator and as herbicide since its introduction in 1949 (Schoene and Hoffman, 1949). After application to foliage, MH is freely translocated in plants to meristematic tissues, with mobility in both phloem and xylem (Meyer et al., 1987). Its mode of action in plants is not clear, although several hypotheses have been proposed and investigated, such as inhibition of cell division by mitotic disruption. Others have suggested that MH acts as an anti-auxin, anti-gibberellin or regulator of auxin metabolism and other plant growth regulators (Hoffman and Parups, 1964). Some carcinogenic effects of MH in mice, rats and cultured human lymphocytes raised a concern of its risks to man (Swietlińska and Žuk, 1978; De Marco et al., 1995; Ribas et al., 1996). MH is registered in the United States to control sprouting of potatoes and onions, suckers in tobacco, and growth of weeds, grasses, and trees in and along lawns, turf, ornamental plants, non-bearing citrus, utility and highway rights-of-way, airports and industrial land (Environmental Protection Agency [EPA], 1994). MH is also registered in Europe (except Austria, Denmark, Finland), Canada, Brazil, Argentina, China, Dominican Republic, Israel, Malaysia, South Africa, and other countries as well for sprout suppression on onion, shallot, garlic, carrots, and sprout suppression and control of volunteers on potatoes (Boyd, 2006; Coresta, 2014; European Food Safety Authority [EFSA], 2016). Previous attempts to use MH for broomrape control were reported in watermelon for P. aegyptiaca (Prokudina, 1976) and P. aegyptiaca and P. ramosa in tobacco (Evtushenko et al., 1973a; Danko, 1980; Lolas, 1986, 1994; Zhelev, 1988; Covarelli, 2002; Brault-Hernandez and Mornet, 2007) with moderate to high control efficacy. MH was also considered for dodder (Cuscuta spp.) control on sugar beet (Evtushenko et al., 1973b).

The limited number of effective herbicides for broomrape control and their extremely narrow range of modes of action, make it hard to match an herbicide to a crop plant for successful broomrape control. Glyphosate and ALS inhibitors are among the most widely used herbicides in the world, resulting in the appearance of glyphosate- and ALS-herbicide-resistant weed populations (Heap, 2014). This phenomenon greatly increases the chances of developing sulfonylurea-, imidazolinone-, and glyphosate-resistant broomrape races, because they are plants with high level of genetic flexibility, especially in areas where the herbicides are heavily used (Hershenhorn et al., 2009). Therefore, there is an urgent need to find new agents that effectively control broomrape by combining new modes of action with ease of use. MH’s mode of action differs from ALS and EPSPS inhibition and it has no residual soil effect.

The objectives of this work were: to study the effect of MH on early developmental stages of P. aegyptiaca and its translocation within tomato roots; and to optimize its ability to control of this parasitic weed in tomato cultivation under greenhouse and field conditions.

Maleic hydrazide has never been used for broomrape control in tomato, and its ability to control the parasite has thus never been compared to other relevant herbicides such as SS and IM (Eizenberg et al., 2013). The Petri-dish experiments were used to investigate MH’s effect on P. aegyptiaca seeds at early developmental stages, i.e., seed germination and germ tube development and elongation, before reaching the host root. MH demonstrated limited capacity to prevent seed germination. Only the two highest MH concentrations, 11.2 and 22.4 mM, were able to reduce the seed-germination percentage by 25 and 45%, respectively, as compared to the untreated controls. Lattice seed germination was reduced by 93% in response to 15 mM MH, (Haber and White, 1960), and Grant and Harney (1960) did not find any inhibitory effect on tomato seed germination at 10 mM MH, while Albizia lebbeck L. seeds responded in a positive germination rate to increase of MH concentration (Tomar, 2008). These results demonstrate the variation in seed sensitivity of different plant species to MH. Using dormant lettuce and y-irradiated wheat seeds, Haber and White (1960) concluded that MH inhibits cell division but has no effect on cell expansion. Similar experiments were conducted by Hershenhorn et al. (1998) with sulfonylurea herbicides. Chlorsulfuron, bensulfuron, primisulfuron, and thifensulfuron under similar conditions reduced seed germination rate by 17, 18, 25, and 19% at a concentration of 10 μM, which is 1000–2000 times lower than the MH concentrations used in our experiments. The same concentration, 10 μM, of seven phytotoxins—fusarenon X, nivalenol, deoxynivalenol, T-2 toxin, HT-2 toxin, diacetoxyscirpenol, and neosolaniol produced by Fusarium species caused 100% inhibition of Phelipanche ramosa seed germination (Zonno and Vurro, 2002).

Maleic hydrazide was much more efficient at preventing germ tube elongation. The two highest MH concentrations reduced germ tube length of all treated germinating seeds to less than 3 mm. Hershenhorn et al. (1998) reported that the sulfonylurea herbicides chlorsulfuron, bensulfuron, nicosulfuron, triasulfuron and thifensulfuron reduced germ tube length by ∼50% at a concentration of 10 μM. The fact that MH and the sulfonylurea herbicides were less effective at preventing seed germination and much more efficient at inhibiting seed germ tube elongation is not surprising, because most herbicides are active at the post-seed-germination stages rather than during initial seed germination phase (Cathey, 1964; Parka and Soper, 1977).

A PEB system was used to study the ability of MH, as compared to SS and IM, to prevent or inhibit the initial parasitism stages of P. aegyptiaca, namely, penetration and establishment in the host root tissues. In the first set of experiments, MH, SS, and IM were applied 2 days after the roots were exposed to GR24. The number of tubercles that formed on the control roots 7 DAA was high whereas in the treated plants, almost no tubercles formed, with no differences between the treatments. These results indicated very high control efficacy when the herbicides were applied shortly after the P. aegyptiaca seeds were triggered to germinate. These and the Petri-dish results suggest that the main MH inhibition mechanism is located at the penetration and establishment stages, because MH’s effect on the germination process in the Petri-dish experiments was rather limited. The observation made 14 DAA revealed no change in the treatments’ efficacy; except for 75 μM MH which, although somewhat less effective, still reduced tubercle number by 72% as compared to the control plants. To investigate the effect of the herbicides on a later stage of P. aegyptiaca parasitism, namely after tubercle formation, they were applied 2 WAA of GR24, when tubercles were already formed and were present on the roots. There were no significant differences between the treatments and the control plants in the number of healthy tubercles or their size at the time of herbicide application. However, all treatments significantly reduced the number of healthy tubercles and their size 2 WAA of herbicides. All treatments were equally effective, even the lower MH rate of 75 μM. This set of experiments demonstrated that MH can control P. aegyptiaca at the tubercle stage and is as effective as SS and IM, albeit at higher rates.

Translocation and movement of MH in the root system could enhance control efficacy if the compound flows, without being detoxified, into the P. aegyptiaca attachments, which serve as a strong sink (Fernández-Aparicio et al., 2016). The MH would then accumulate in the attachments until a lethal dose is achieved. To investigate MH translocation and flow in the roots, we used the double-PEB system. All attachments in the treated PEB were dead 2 WAA of herbicides regardless of the herbicide used—MH, SS, or IM. As expected, these results were in complete agreement with those obtained when using a PEB containing the whole plant root system. In the adjacent non-treated PEBs containing the other portion of the plant’s root system, 93, 77, and 75% of the attachments were dead in the MH, IM, and SS treatments, respectively. These results suggest that MH moves through the root system without being degraded to non-toxic metabolites. Uptake from solution of MH by barley roots and its translocation to shoots is pH dependent. Uptake and translocation was higher with the decrease of the solution pH probably by passive diffusion of the undissociated form of the acid. Translocation to shoots was approximately proportional to the chemical concentrations in the roots (Briggs et al., 1987). Radioactive labeled MH applied to leaves of Zebrina pendula, Tradescantia fluminensis and barley seedlings moved relatively free in both phloem and xylem (Crafts and Yamaguchi, 1958). MH absorption from Ricinus communis petioles was rather slow but was readily phloem-translocated from the mature leaves allowing appreciable quantities to reach the roots and the nutrient solution (Rigitano et al., 1987). MH absorption by foliar application, translocation to roots and between roots without being detoxified suggests an increase chances for successful broomrape control under greenhouse and field conditions.

Greenhouse experiments in pots revealed that MH up to 600 g a.i. ha^-1 applied twice, at 200 and 400 GDD, has no negative effect on tomato foliage or root biomass, demonstrating that MH is very safe for use on tomato plants. Further applications were not conducted and fruit number and weight were not determined as the pot volume became the main limiting factor for plant growth. The total number of P. aegyptiaca counted on the roots was significantly lower at application rates of 75 g a.i. ha^-1 or more as compared to the non-treated plants, and P. aegyptiaca biomass was reduced significantly at 25 g a.i. ha^-1, the lowest rate tested. Those results prompted us to test MH control of P. aegyptiaca under field conditions.

Over 30 field experiments were conducted during 2010–2014 to optimize the timing, number and rates for maximal P. aegyptiaca control and tomato safety, as well as to minimize the total amount of MH applied. MH is registered in Israel for the prevention of onion and potato sprouting at rates of 2.2 and 2.5 kg ha^-1, respectively. Our goal was to limit the total rate used for P. aegyptiaca control to this range, in order to facilitate MH registration for its new purpose—P. aegyptiaca control in tomato.

The first set of field experiments was aimed at detecting the most effective rate under a constant number of applications. We used five split applications following the application scheme established for P. aegyptiaca control in tomato with SS (Hershenhorn et al., 1998, 2009; Eizenberg et al., 2004, 2013; Ephrath et al., 2012), which dictates three split applications of sulfonylurea herbicides at 200 GDD intervals starting at 200 GDD, followed by two applications of IM, for a total of five split applications. A significant reduction of P. aegyptiaca shoots in the 2010 experiments at Ein Harod Ihud and Hulata was achieved with 270 g a.i. ha^-1 MH and a further reduction with 540 g a.i. ha^-1, a total of 1.35 and 2.7 kg a.i. ha^-1 of MH, respectively. Gradually increasing the rates along the applications had no influence on P. aegyptiaca shoot number in Ein Harod Ihud but significantly decreased their number in Hulata at the early observation of 65 DAP; this effect then diminished and disappeared by 76 DAP. A positive correlation was found between P. aegyptiaca control and yield in Ein Harod Ihud. The yield was significantly higher with the treatments that were also effective at reducing the number of P. aegyptiaca shoots. Based on the results of these experiments, we further examined the minimal number of applications needed with 540 g a.i. ha^-1, the most successful rate in the 2010 experiments.

We conducted three experiments aimed at testing the efficacy of an increased number of applications under a constant MH rate of 540 g a.i. ha^-1. The results (Table 3) indicated that even one or two MH applications significantly reduce the number of P. aegyptiaca shoots. This was true for the two Ein Harod Ihud experiments whose results were combined. At Mevo Hama, two applications were more efficient than one application, which also reduced the number of P. aegyptiaca shoots significantly as compared to the non-treated control plots. In all three experiments, three and four MH applications gave superior results, with no aboveground P. aegyptiaca shoots in Mevo Hama and one aboveground P. aegyptiaca shoot per m² in the Ein Harod Ihud experiments. The higher control efficacy achieved in Mevo Hama may be explained by the lower P. aegyptiaca infestation level in that field, with 18 shoot m^-2 in the non-treated plots as opposed to about 30 shoot m^-2 in the Ein Harod Ihud control plots. Although the infestation level in Mevo Hama was lower and the P. aegyptiaca control efficacy higher than at Ein Harod Ihud, the yield—irrespective of the control efficacy—was lower. This was probably a result of pest and disease outbreaks in the field toward the end of the growing season. All treatments in Ein Harod Ihud improved the yield regardless of the number of applications.

Maleic hydrazide at a rate of 700 g a.i. ha^-1 applied on tobacco once at 40 DAP reduced the number of P. ramosa plants by 60–80% depending on the tobacco cultivar. The same rate applied twice at 40 and 60 DAP reduced the number of P. ramosa plants by 80–90% and led to a considerable yield increase (Lolas, 1986). Lower rate of 4500 + 4100 g a.i. ha^-1 applied at 40 and 60 or 50 and 70 DAP resulted only in a moderate control (Lolas, 1994). Those results are in agreement with the results obtained in the present study were increased MH rates improved the control efficacy. However, Lubanov (1973) reported complete broomrape control with one MH application of 450 g a.i. ha^-1 on tobacco at the bud formation stage. This complete control may be explained by the fact that the field was previously disinfected with metam-sodium. Similar results were also reported by Jelev (1988) were MH was applied once to tobacco 40 DAP. MH at 2000 g a.i. ha^-1 applied once on tobacco foliage 55 DAP prevented completely the appearance of P. ramosa plants above ground but the same rate applied 20 days later had no effect on the number of P. ramosa plants as compared to the control plots (Covarelli, 2002) emphasizing the importance of the application timing. The main parameter dictating broomrape parasitism process and as a consequence herbicide application is the accumulated heat expressed as GDD (Joel et al., 2007; Hershenhorn et al., 2009; Eizenberg et al., 2013). Comparison to our results cannot be conducted without available GDD data. However, it should be noted that MH early application had phytotoxic effect on tobacco plants. In our experiments one MH application of 540 g a.i. ha^-1, only 25% of the rate used by Covarelli (2002), 20 DAP reduced the number of P. aegyptiaca plants by 12% in the two experiments conducted at 2011 and 2012 at Ein Harod Ihud and by 34% in Mevo Hama without any phytotoxic effects. The host plant roots grow and penetrate in to deeper soil layers during the growing season and stimulate new broomrape seeds to germinate. Therefore repeated MH applications are needed to effectively control the parasite during the whole growing season. It may be hypothesized that complete control of the parasite by one MH application may occur if MH stays at an effective lethal concentration level in the tobacco roots during all the growing season. These contrasting results demonstrate the problematic of comparing experiments conducted in different crops under different conditions. Differences in physiology and morphology of the host plant species, soil type, environmental conditions such as temperature and rainfall, irrigation regime and equipment, growth conditions and as a result plant status and stress, broomrape species and sampling methods, spraying equipment and MH formulation and surfactants, infestation level and cultivation practices are only some of the factors which influence control efficacy.

In irrigated water melon a reduction of the infestation level by 67–96% was achieved with MH at 5000 g a.i. ha^-1 repeated twice at 10–20 days (Prokudina, 1976). Such high MH rates were not tested in our experiments. In our preliminary experiments one MH application of 1100 g a.i. ha^-1 caused a significant damage to tomato plants (curly and silvery leaves, flower drops and as a result extensive vegetative growth) and therefore were not continued to be tested (Sarkar et al., 2014).

In the last set of experiments conducted at Kibbutz Mesilot and Havat Eden in 2012 (Table 4), we tried two different options to reduce the total MH amount applied to the field during the growing season. Since we had found, in preliminary pot experiments, that moving the first application to 100 GDD has no negative effect on the tomato plant, we reduced the rate of the two first applications to 270 g a.i. ha^-1 and started the first application at 100 GDD. The second option was a constant rate of 400 g a.i. ha^-1, also starting at 100 GDD. As MH may be absorbed by roots and was proven in our double-PEB experiments to flow through the root system to P. aegyptiaca attachments, we applied a very high amount of MH through the drip-irrigation system. Five foliar applications, regardless of rate, had excellent control efficacy, almost completely preventing the appearance of P. aegyptiaca shoots aboveground. This was especially impressive at Kibbutz Mesilot as the infestation level was high. MH applied three times through the drip-irrigation system had no effect on the number of P. aegyptiaca shoots in Havat Eden, and three split applications of 2700, 5200, and 7900 g a.i. ha^-1 reduced shoot numbers significantly as compared to the non-treated plots but were not as effective as the foliar treatments. The effective control of the foliar applications resulted in a significant yield increase in both locations. The drip-irrigation applications had no effect on yield at Kibbutz Mesilot and a negative effect at Havat Eden. The inefficiency of MH control when applied through the soil, even at the very large amounts applied in our experiments, may be explained by the soil characteristics. Although MH is readily absorbed by roots, it rapidly disappear form soil (Levi and Crafts, 1952; Hoffman et al., 1962; Helweg, 1974). There are several soil components that enhance MH adsorption and inactivation, such as clay, specific surface area and pH, but not organic matter content (Hermosin et al., 1987) The end result is less herbicide available to the plant roots and therefore to P. aegyptiaca.

Based on these results and many other field and observation trials, MH was registered in March 2013 for P. aegyptiaca control in processing tomato fields in Israel under the two protocols presented in Table 4. Since its registration, MH has been widely used by Israeli tomato growers with excellent control results, crop safety and yield increases.

AV contributed to the conception of the work, conducted the experiments and analyzed the data, ED conducted the statistical analysis, contributed to data interpretation and edited the manuscript, GA conducted the experiments, DP conducted the experiments and contributed to the design of the work, ES conducted the experiments, JH planned the study, analyzed and interpreted the data, drafted the manuscript and ultimately approved the version to be published.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer BC and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.