Isolation and identification of Desmodium root exudates from drought tolerant species used as intercrops against Striga hermonthica

http://dx.doi.org/10.1016/j.phytochem.2015.06.026 0031-9422/ 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). ⇑ Corresponding author. E-mail addresses: tony.hooper@rothamsted.ac.uk (A.M. Hooper), john.caulfield@ rothamsted.ac.uk (J.C. Caulfield), bing.hao@rothamsted.ac.uk (B. Hao), john. pickett@rothamsted.ac.uk (J.A. Pickett), cmidega@icipe.org (C.A.O. Midega), zkhan@ icipe.org (Z.R. Khan). A.M. Hooper a,⇑, J.C. Caulfield , B. Hao , J.A. Pickett , C.A.O. Midega , Z.R. Khan b


Introduction
In subsistence farming regions of sub-Saharan Africa (SSA), where parasitic weeds of the Striga genus devastate subsistence cereal crops, a successful intervention called push-pull technology effectively inhibits Striga damage and has been adopted by more than 96,000 farmers (http://www.push-pull.net/, Fischler, 2010). The technology utilises forage legumes in the Desmodium genus (Fabaceae) as an intercrop which provides the key chemical components for inhibiting development of Striga in the field. The mechanism by which this occurs has been shown to be caused by allelopathic root exudates from Desmodium that both stimulate the germination of Striga, (Khan et al., 2002) but more importantly inhibit the subsequent development of the germinated seed, (Tsanuo et al., 2003;Hooper et al., 2010) so that in the field, almost no parasitism is observed Pickett et al., 2010). The technology blends in with the cultural practise of mixed cropping and has been taken up by farmers introduced to it through their peers in farmer-led groups (Khan et al., 2008(Khan et al., , 2014Murage et al., 2011). It provides additional benefits to insect pest (cereal stemborer) and weed control by providing fodder for zero-grazed livestock while improving soil quality through organic carbon and fixed N (Khan et al., 2008a;Midega et al., 2013). In order to expand the technology into areas that are either more arid or under threat of increased drought through climate change, new species of Desmodium must be identified that are resilient to drought and which can also provide the chemistry that confers inhibition of Striga. Once identified, the intercrop should protect drought tolerant cereal crops such as sorghum or new drought tolerant maize varieties against Striga, thereby increasing food security for small-holder farmers. In addition, the further characterisation of active root exudate chemistry provides the basis for identifying the enzymic and genetic basis for allelochemical production in planta (Hooper et al., 2009;Khan et al., 2008b).
In our screening for drought tolerance, three Desmodium species were originally identified as showing potential as they survived extended periods of drought when grown in African screen-house conditions. They are Desmodium intortum (Mill.) Urb., Desmodium incanum (G. Mey.) DC. and Desmodium ramosissimum G. Don. We hypothesise that taxonomically related Desmodium species may produce similar root exudate chemistry to the Desmodium species known to be effective in the field (Desmodium uncinatum (Jacq.) DC.), and so be potential intercrops for inhibition of Striga parasitism of subsistence cereals in more arid agronomic environments,. In order to investigate their potential for Striga inhibition, the root exudates of these plants were collected and analysed by HPLC and liquid chromatography-mass spectrometry (LCMS) to determine whether the same Striga inhibiting chemistry previously identified from D. uncinatum  was present and which could be attributed to inhibition of Striga. In addition, by performing screen-house pot experiments, the drought tolerant Desmodium plant root exudates were tested against Striga hermonthica on maize, to demonstrate Striga inhibiting biological activity. The drought tolerant Desmodium species were then tested as Striga inhibiting intercrops for sorghum in rain fed plot experiments under field conditions in Western Kenya.

Striga inhibition by drought tolerant Desmodium species root exudates
In screen house experiments, the parasitism of maize by S. hermonthica was inhibited strongly by irrigation with the root exudates of all the tested Desmodium species compared with irrigation through soil alone (Fig. 1). The mean number of Striga plants emerged on root exudate treated maize differed among treatments with distilled water control ( Fig. 1; F 5,107 = 98.23, P < 0.001) showing the root exudates for all tested Desmodium species were effective in inhibiting Striga parasitism.

Striga inhibition by drought tolerant Desmodium species in demonstration plots
The four species of Desmodium for which exudates were studied, were grown as intercrops with drought tolerant sorghum (the early maturing commercial hybrid Gadam Hamam) in rain fed plots on the icipe Thomas Odhiambo Campus on the shores of Lake Victoria. The performance of the plots after the short and rainy seasons of 2014 were assessed (Fig. 2). All the drought tolerant Desmodium species tested showed significant Striga inhibiting properties in the field, as they did in the pot experiments, with no significant differences to the performance of D. uncinatum the intercrop now used widely in farmer fields.

Exudate analysis
The root exudates of D. uncinatum, D. intortum and D. incanum were collected one month after the transfer of plants into hydroponic solution and analysed by HPLC (Fig. 3). The seedlings were transferred to the hydroponic solution at two weeks after germination from seeds. The structure of the major flavonoid peaks in these initial exudates were identified by HPLC co-elution with known standards, isolated from extracting leaf or root tissue of conspecific plants which were fully characterised by nuclear magnetic resonance (NMR) spectroscopic and electrospray ionisation mass spectrometry (ESIMS) analysis (Fig. 3). The major components seen in the C-glycosylflavonoid (CGF) region of D. uncinatum are isoschaftoside (5), vitexin (8) and a broad signal for 2 00 -O-glucosylvitexin (7). In the case of D. intortum, the major components are vicenin-2 (2) and isoschaftoside (5). The component isoschaftoside (5) has been reported previously as being able to interfere with Striga development post-germination . However, analysis of the exudate from D. incanum revealed that it comprised a number of CGFs structures, of which isoschaftoside was only a minor component. All three species were found to produce isoschaftoside (5) in their exudate in varying quantity but there are very clear differences in the compounds that comprise the exudate at this stage in plant development.
The same three species D. uncinatum, D. intortum, D. incanum along with another species, D. ramosissimum, which was added to the experiment subsequently because it was found to regenerate after an extended period of drought (>2 months), were grown in hydroponics for four months and the root exudates collected again for HPLC and LCMS analysis. At this time of plant growth, although some differences still existed in the chemical make-up of the exudates, there was now a high degree of similarity as demonstrated in Fig. 4. The many differences observed in young plant root exudates were now reduced and the root exudates matched closely. After HPLC analysis and tentative identification of the flavonoid components through retention time, LCMS analysis generated their molecular weight and their fragmentation was used to verify the nature of attachment of hexose or pentose moieties to the flavonoid aglycone. LCMS experiments using higher cone voltages showed fragmentation of the metabolites which generated losses of 120 and 90 Da, typical of C-linked hexoses, and losses of 90 and 60 Da typical for C-linked pentoses. This identified the hexose or pentose moieties seen in the molecules to be both C-linked. The structures were then identified by co-elution with standards obtained and fully characterised from conspecific plant tissues. Specific chemicals identified and present in each of the four Desmodium species plant root exudates are labelled and shown in Fig. 4. The presence of di-C-glycosylflavones coupled to phenolic acids was not detectable in the early stages of exudate production and were also not detected in the tissues of the plants extracted for natural product standards. They are presumably exuded shortly after they are biosynthesised, implying an important role in rhizosphere chemical ecology yet to be elucidated. The absence in Desmodium plant tissue means no chemical standard could be fully structurally elucidated and the position of the ester and the position and type of hexose and pentose remain targets for characterisation.

Chemical characterisation
Compounds that were isolated from tissues are shown in Table 1. These were subsequently used to verify the structures of the chemicals in the plant exudates by HPLC co-elution and LCMS analysis. Compounds 1, 3, 4 and 6 were isolated from D. incanum root tissue, compound 2 from D. incanum leaf tissue and compounds 5, 7 and 8 from D. uncinatum root tissue. The C-linked sugar moieties show restricted rotation in NMR experiments and without recording NMR data at very high temperature, rotamers or broadened peaks were observed. In the case of 3, two major rotamers could not be coalesced at 350 K in methanol, but interconverted at 290 K and all rotamers were shown to be stable on the NMR timescale at 243 K. Both major rotamers were therefore characterised at 290 K.

Conclusions
Allelochemicals generated from the root exudates of plants in cropping systems possess the potential for delivering desirable biological activities in both low input and high-input agriculture as they may be delivered in the desired site for activity McNally et al., 2003;Macias et al., 2008Macias et al., , 2009). Our initial screens of desmodium species performed to test for tolerance of extended periods of drought resulted in three species being taken forward to examine whether they can act as potential intercrops for cereal production The root exudates of the drought tolerant intercrop species D. intortum, D. incanum and D. ramosissimum are similar to each other and to D. uncinatum after these plants have developed in hydroponic solution for four months. While all plants produce isoschaftoside, an allelochemical previously identified as inhibitory to post-germination Striga development, over time, a blend of di-C-glycosylflavones are exuded by all the species examined when maturing. This class of compounds has been reported to possess various ecological and biological activities such as anti-microbial activity (McNally et al., 2003), mycorrhizal colonisation stimulants of melon root (Akiyama et al., 2002), antifeedant activity (Chul-sa et al., 2009;Byrne et al., 1996), and activity antagonistic to Striga  as well as many other medicinal activities (Talhi and Silva, 2012). The root exudate blends characterised here demonstrated Striga inhibition in pot experiments and so provide the chemical basis for intervention against Striga inhibition under rain fed field conditions. Demonstrating Striga inhibition in research field plots by drought tolerant Desmodium in semi-arid rain fed agriculture shows that the field potency predicted by screen house pot experiments was achieved. The application of these plants as intercrops with drought tolerant cereals, such as sorghum, in participating farmer fields is now to be tested. Desmodium root exudate activity in arid conditions rather than hydroponics, where moisture in the rhizosphere may effect allelochemical production or movement may not be important as there must be sufficient moisture in the ground to grow cereal crops. In the research plots here under semi-arid conditions, the data show the exudates are effective. In addition, the Desmodium intercrop is perennial and produces exudates year round, so while leaching due to an instance of heavy rainfall may reduce soil concentrations, they can be replenished over time. The identification of root exudate metabolites also allows the creation of synthetic root exudates or blends containing one or more components, to probe further the chemical basis of Striga inhibition by Desmodium. The drought tolerant species have demonstrated resilience to the arid conditions where push-pull technology can make an impact in the yield of staple cereals for small-holder farmers. Climate change is also expected to increase drought stress in existing areas where the technology has been

Plant material
Seeds of D. uncinatum, D. intortum, D. incanum and D. ramosissimum were obtained from the seed bank at the International Livestock Research Institute (Addis Ababa, Ethiopia) and grown in seed multiplication plots at the field station of the International Centre for Insect Physiology and Ecology, Thomas Odhiambo Campus (icipe-TOC) on the shores of Lake Victoria in western Kenya.

Pot experiments
Screen house trials were conducted at icipe-TOC using methodologies adapted from Khan et al., 2002. The allelochemical activity of Desmodium intercrops against S. hermonthica parasitism on maize was tested by irrigating maize grown in Striga infested soil with an aqueous solution of chemical components, eluting from established Desmodium plants. Pots (20 cm) containing the drought tolerant desmodium species, D. intortum, D. repandum, D. incanum and D. ramosissimum, along with D. uncinatum as a positive control and soil alone as a negative control, were placed on shelves and received distilled water at a rate of 1.25 ml/min, thus allowing the flow of water by gravity through the Desmodium root mass and into pots containing maize situated below. These pots contained a Striga-susceptible maize variety (WH505) planted in soil which had been inoculated with approximately 3000 S. hermonthica seeds/pot. Autoclaved soil was used in all experiments, and no additional nitrogen was applied. Emergence of S. hermonthica parasites on maize were monitored in all treatments.

Statistical analysis
The number of emerged Striga plants was analysed using oneway analysis of variance with blocking (batch) in GenStat 16. Fischer's least significant tests were used to compare means. There were six pots in three batches to give 18 total replicates.

Striga inhibition by drought tolerant Desmodium species in demonstration plots
Studies were conducted at Thomas Odhiambo Campus of the International Centre of Insect Physiology and Ecology (icipe), Mbita, located along the eastern shores of Lake Victoria, western Kenya (0°25 0 S, 34°12 0 E). Field trials were conducted during the long and short rainy season of 2014 to investigate effectiveness of the various Desmodium species in suppression of S. hermonthica. During this calendar year, rainfall was approximately 900 mm. Field plots measured 6 m by 6 m, and were arranged in a complete randomized design in four replications. The treatments comprise sorghum intercropped with each of the four Desmodium species and a monocrop. Sorghum was planted at a row-to-row distance of 60 cm and a plant-to-plant distance of 30 cm within a row. The sorghum variety used was the S. hermonthica susceptible, early maturing commercial hybrid Gadam Hamam, recommended for mid-altitude regions. Sorghum was planted at an inter-row spacing of 75 cm and an intra-row spacing of 30 cm, while Desmodium was planted through a drilling system in furrows between the rows of sorghum. The number of emerged Striga plants was counted from 45 plants per plot from within a radius of 15 cm around the base of each sorghum plant, and data expressed as the number of emerged Striga per plot. The observation was made at 10 weeks following sorghum emergence since Striga emergence often peaks at about 10 weeks of crop emergence in the region (Midega et al., 2013).

HPLC analysis of exudates
Desmodium seeds were germinated in a sandy soil and grown until approximately 4 cm high so that a significant root system had established. Subsequent removal of the plants was facilitated by the high sand content. The plants were washed and placed in a hydroponic solution (40% Long Ashton) in plastic tubs containing about 2 litres for approximately 50 plants. During 1 week the hydroponic solution was circulated at 5 mL/min using a peristalsis pump through a Pasteur pipette containing XAD-4 Amberlite polymer to trap out organic material. The trap was washed with MeOH (25 mL) and the solvent removed. The residue was dissolved in 1 mL MeOH for HPLC analysis on a Shimadzu VP series HPLC system using an ACE AQ C-18 column (250 mm x 4.6 mm, 5 lm).

LCMS analysis of exudates
The exudates isolated as described in Section 4.5, evaporated to dryness and re-dissolved in 1 mL methanol, were subject to LCMS analysis using the Micromass Quattro Ultima bench top triple quadrupole mass spectrometer attached to Waters Acquity UPLC system (Ultra Performance Liquid Chromatography). The mass spectrometer was operated in negative ion mode, with a capillary voltage of 2.7KV, cone voltage 50-180 eV, mass range 50-1000 m/z. Source temperature 130°C, desolvation temperature 350°C, desolvation gas flow 1000 L/h (nitrogen) and cone gas flow 60 L/h (nitrogen). Where required, selected ions were admitted to the collision cell for MSMS analysis with argon admitted at a pressure of 2.1 e À3 mbar, causing CID. Samples were injected via the Acquity sample manager, injecting 1 ll onto an Acquity UPLC BEC HSS C18 1.8 lm 2.1 Â 150 mm column. Run time was 63 min at a flow rate of 0.4 mL/min. Solvents used are defined A (water, 0.05% formic acid) and B (methanol). The mobile phase used a