Whiteflies [Bemisia tabaci (Gennadius)] and aphids [Aphis gossypii Glover and Myzus persicae (Sulzer)] are serious threats to watermelon by direct feeding and by transmitting viruses of important virus diseases. The desert watermelon Citrullus colocynthis (L.) has been shown to exhibit resistance to these insect pests and could be a useful source for breeding resistance into watermelon [Citrullus lanatus var. lanatus (Thunbs) Matsum & Nakai]. Using high-performance liquid chromatography (HPLC), we found differences among the chemical profiles of two U.S. PIs of C. colocynthis, one PI of C. lanatus var. citroides, and two heirloom watermelon (C. lanatus var. lanatus) cultivars (‘Charleston Gray’ and ‘Mickey Lee’). Flavonoid and caffeic acid derivatives were identified in the leaf extracts by a combination of ultraviolet (UV) and mass spectrometry (MS) spectral analyses. Four phenolic derivatives of caffeic and/or ferulic acid were found to be essentially unique to C. colocynthis. Total flavonoid content was found to be approximately four to 18 times higher in C. colocynthis accessions and seven to nine times higher in C. lanatus var. citroides as compared with watermelon cultivars. Caffeoyl-glucose was also identified in the leaves of watermelon cultivars for the first time. Leaf sugar concentrations (198 to 211 mg·dL−1), read from a glucometer, were statistically the same among the various germplasm entries. These results will help in the development of pest-resistant watermelon.
Watermelon [C. lanatus var. lanatus (Thunbs) Matsum & Nakai] is an important crop globally. The origin of Citrullus spp. is in central or southern Africa (Jarret et al., 1997; Mujaju et al., 2010). On that continent, a wide variation of watermelon populations exists in diverse geographical regions and the fruit is considered a vital source of water and food for the native people and animals. As a result of many years of cultivation and selection for desirable qualities, a large number of the American heirloom watermelon cultivars shares a narrow genetic base and is susceptible to diseases and pests (Levi et al., 2001a; Simmons and Levi, 2002). On the other hand, the Citrullus spp. germplasm collected mainly in central and southern Africa shows a wide phenotypic and genetic diversity (Levi et al., 2001b).
The genus Citrullus includes several species or subspecies. Among them is the bitter watermelon C. colocynthis (L.) Schrad that thrives in the deserts of North Africa, the Middle East, and Asia. It has distinct morphological and biochemical features such as thick leaves, fairly small fruits, and a bitter odor that repels insects (Simmons and Levi, 2002). Additional species, found in southern Africa, are C. ecirrhosus Cogn. and C. rehmii De Winter (Robinson and Decker-Walters, 1997). C. lanatus var. lanatus is considered the progenitor of cultivated watermelon. The C. lanatus also includes the citron watermelon, C. lanatus (Thunbs) Matsum & Nakai var. citroides (L.H. Bailey), which thrives in the deserts of southern Africa. It is known as the “Citron Watermelon,” “Cow Watermelon,” or “Tzama” (Jarret et al., 1997; Mujaju et al., 2010) and is considered a valuable germplasm source because different accessions of this subspecies contain resistance to diseases or pests (Levi et al., 2001b; Thies and Levi., 2007). The Citrullus germplasm collection maintained by the USDA-ARS Plant Genetic Resources and Conservation Unit, Griffin, GA (http://www.ars-grin.gov) includes over 1800 U.S. PIs. These PIs have been useful sources of germplasm for identifying disease or pest resistance that through intensive breeding programs could be incorporated into elite watermelon cultivars.
Whiteflies [Bemisia tabaci (Gennadius)] and aphids [Aphis gossypii Glover and Myzus persicae (Sulzer)] are major pests that feed on and transmit viruses to watermelon plants (Simmons et al., 2010; Simmons and Levi, 2002). However, several C. colocynthis PIs possess resistance to the sweetpotato whitefly, B. tabaci (Simmons and Levi, 2002), the two-spotted spider mite, Tetranychus urtichae Koch (Lopez et al., 2005), and aphids (Simmons, unpublished data). These sources of germplasm should be useful for incorporating pest resistance into watermelon cultivars.
Very little work has been reported on the flavonoids and phenolics of watermelon, especially in the leaves. Most previous investigations of flavonoid or phenolic content of watermelon have been limited to determination by colorimetric methods (Asyaz et al., 2010; Chopra et al., 1974; Ibrahim et al., 2010; Tlili et al., 2011; Venkataramaiah and Narayana, 1983). Others have used acid hydrolysis before analysis to measure individual aglycone flavonoids (Harsh and Nag, 1988; Lugasi and Hovari, 2002; Meena and Patni, 2008) or phenolic acids (Das et al., 1967; Venkataramaiah and Narayana, 1983). Delazar et al. (2006) determined a number of flavonone-C-glycosides in Citrullus colocynths fruits. To our knowledge, only the report of Maatooq et al. (1997) reports specific and novel hydroxybenzyl-flavonoids in the leaves of watermelon (Citrullus colocynthis) and only that of Chopra et al. (1974) relates phenolic content to resistance and susceptibility to disease (Alternaria cucumerina). Furthermore, there is no sufficient information on the chemical profile and compounds that may lure or repel insect pests and affect their feeding habits and reproduction on plants of watermelon cultivars vs. C. colocynthis PIs.
The objective of this study was to determine if differences exist in the chemical profiles of leaves of C. colocynthis PIs that showed whitefly resistance (Simmons and Levi, 2002) vs. those of susceptible watermelon cultivars (C. lanatus var. lanatus) and in a representative PI of the citron watermelon (C. lanatus var. citroides).
Materials and Methods
The plants in this study included the watermelon cultivars Charleston Gray and Mickey Lee (C. lanatus var. lanatus), the C. var. citroides PI 500354, and the C. colocynthis PI 386015 and PI 432337. The plants were grown in pots in the greenhouse using a standard watering regime.
Methanolic extract of fresh leaves of each Citrullus accession was made by clipping five healthy leaves from 7-week-old plants of each accession and cutting them into pieces with a pair of scissors. Three-gram portions of the leaf samples of each watermelon accession was then placed in separate 14 × 7-cm, 118.5-mL glass bottles (a Teflon-lined cap was used) and 100 mL of methanol was added to each bottle. Chrysin (Sigma-Aldrich, Milwaukee, WI; recrystallized from amyl alcohol) was used as an internal standard. Three milliliters of a methanolic solution (4.8 mg chrysin/3 mL) was added to each leaf extract and the solution mixed. The submerged leaves were subsequently further cut into smaller pieces with scissors and ground for ≈1 min with a polytron (Kinematic-PCU-2; Brinkmann Instruments, Inc., Westbury, NY) equipped with a 6-mm diameter sawtooth grinder type of tissue cutter. The solutions were filtered through 0.45-μm nylon-66 filters in preparation for HPLC analysis.
High-performance liquid chromatography analysis.
Extracts were analyzed once by reversed-phase HPLC using a H2O/MeOH linear gradient from 10% to 100% MeOH in 35 min, a flow rate of 1 mL·min−1, and detection at 340 nm. Each solvent contained 0.1% H3PO4. Analyses were performed with a Beckman Ultrasphere C18, 5 micron (4.6 × 250 mm; Beckman Instruments, Norcross, GA) column using a Hewlett-Packard 1050 diode array HPLC (Palo Alto, CA). Quantitation was performed by using chrysin's response factor. In addition to chrysin, other standards used were apigenin, caffeic acid, chlorogenic acid for tuning the HPLC-Mass Spectrophotometer (Sigma-Aldrich, St. Louis, MO), and isoorientin (Indofine Chemical Co., Belle Mead, NJ).
Preliminary identifications of compounds were by ultraviolet spectra and retention time correlations with standards. Also, mass spectra were obtained with a Thermo-Finnigan LCQ HPLC/MS. For HPLC/MS, chlorogenic acid was used for tuning. Spectra were obtained in the negative ion mode. Gas chromatographic (GC) analyses were performed with a Hewlett-Packard 5890 GC fitted with a DB-5 capillary column (30 m × 0.25 mm i.d., 1-mL·min−1 flow rate); injector, 250 °C; flame ionization detector, 350 °C; linear temperature program, 100–320 °C at 8 °C·min−1; splitless mode. Compounds were analyzed as their silylated derivatives.
Peaks A and B were isolated from the methanolic leaf extract of PI-500354 after concentration by rotary evaporation to remove methanol. The solution was then placed on a preparative liquid chromatography column packed with PrePAK-500 C18 packing (Waters Millipore Corp., Milford, MA; washed with MeOH and recycled to water, 20 psi nitrogen pressure-aided flow). The column was eluted with MeOH/water solutions and Peak A was eluted with 10%, whereas Peak B was eluted with 30% MeOH/water.
The HPLC peaks were labeled alphabetically based on the retention times (the retention times were based on 340 nm milliabsorption units) at which the respective chemical compounds were eluted and detected (Figs. 1 and 2). The internal standard (chrysin) eluted consistently at 28 min. The mathematical area occupied by each peak correlated with the amount of the compound and was either directly provided by the software or was mathematically derived from similar (or neighboring) peaks (Table 1).
High-performance liquid chromatography analysis showing the amount (μg·g−1) of each chemical compound (lettered A to M) detected in the various Citrullus accessions.
Leaf sugar concentration.
Extracts of 0.5 mL volume from leaves representing three groups of germplasm (Mickey Lee, PI 500354, and PI 386015) were obtained by leaf grinding and diluted with equal amounts (0.5 mL) of distilled water and centrifuged for 4 min at 1000 rpm. A drop of each of the diluents was applied to a clinitest strip, and the sugar content was read from a glucometer. Data on sugar content were analyzed for any statistically significant variance using SAS statistical software (SAS Institute, 2003). Significant differences were determined at P < 0.05.
Results and Discussion
Noticeable differences exist in the HPLC profiles between the C. colocynthis and C. lanatus var. lanatus or C. lanatus var. citroides accessions (Tables 1 and 2; Figs. 1 and 2). The flavonoid compounds (Peaks B and D; Table 1; Figs. 1 and 2) are low in the watermelon cultivars (‘Charleston Gray’ and ‘Mickey Lee’) but are significantly higher in the C. lanatus var. citroides (PI 500354) or the C. colocynthis accessions (PI 386015 and PI 432337). Peaks G and H showed similar results. Peaks C, E, K, and L are unique to the C. colocynthis PI 386015 and PI 432337 (Fig. 2). A clear peak M was observed only in PI 432337, whereas peaks A, J, and F occurred in sufficient amounts in all accessions analyzed (Table 1; Figs. 1 and 2). The consistent elution of chrysin at 28 min for all analyzed samples suggests an overall high accuracy of the results.
Total flavonoid content (μg·g−1 fresh leaf weight) of various Citrullus accessions.z
A higher number of compounds was seen in chromatograms with C. colocynthis accessions as compared with those of C. lanatus var. lanatus or C. lanatus var. citroides accessions (Table 1; Figs. 1 and 2). Ultraviolet spectra and chromatography elution points are only indicators of what may be in a peak. It is possible that one or several of these compounds could be associated with the relative resistance of C. colocynthis to fluid-feeding pests. This could be possible for phenolic compounds of peaks K, L, or M. Although chlorogenic acid compounds are found in certain plants (Harrison et al., 2008), chlorogenic acid is unique to C. colocynthis accessions in Citrullus (Wu, 2007). The spectral analysis indicated that Peak A was caffeoyl-glucose: MS: 341 (M-H); MS2: 179 (M-caffeoyl), 161 (M-glucose). Acid hydrolysis yielded caffeic acid (confirmed by HPLC and GC retention time and MS) and glucose (confirmed by GC retention time). Peak B represented Isovitexin-2″-O-glucoside: MS: 593 (M-H); MS2: 503 (M-H-90), 473 (M-H-120). Ultraviolet analysis indicated an apigenin or kaempferol aglycone. After isolation, this flavonoid (Peak B) appeared to lose a glucose moiety on acid hydrolysis and had a molecular mass of 594, suggesting that it might be either apigenin-diglucoside or kaempferol-rhamnosyl-glucoside. After hydrolysis, isovitexin (MS:431; M-H) was identified by HPLC/MS and liberated glucose by GC retention time. Mass fragmentation yielded ions of masses 503, 473, and 311. The abundance of the M-H-90 ion in the MS2 spectrum showed that the C-bound sugar was attached to the C-6-position, indicating an isovitexin structure. The data are in agreement with literature (Qimin et al., 1991) for 6-C-glucosyl-O-glycosyl-apigenin (glucosyl-isovitexin). Peak C represented Isoorientin: MS: 447 (M-H); MS2: 357 (M-H-90), 327 (M-H-120). Ultraviolet spectra indicated a luteolin aglycone base.
Peaks D to M had broad ultraviolet maximum near 310 nm, which indicated they were phenolic in nature, but HPLC/MS was inconclusive and requires further study.
Leaf sugar content was tested to investigate if sugar may play a direct role in the feeding preference on Citrullus accessions by pests such as whiteflies and aphids. Overall means for sugar concentrations were not significantly different among the genotypes (P < 0.05) (mean = 198 to 211 mg·dL−1). This suggests that leaf sugar content plays no role in insect preference of one Citrullus accession over another. Sugar specificity was not assayed. Hence, it is not known what types of sugars were involved.
The C. colocynthis thrives in the deserts of northern Africa, the Middle East, and Asia, and a relatively wide genetic diversity exists among accessions of this species collected in these locations (Levi et al., 2001b). A wide genetic distance exists between C. colocynthis and watermelon cultivars (C. lanatus var. lanatus) (Jarret et al., 1997; Levi et al., 2001b). Still, the C. colocynthis should be a valuable germplasm source to improve resistance of watermelon cultivars to insect pests. A recent study (Hadizadeh et al., 2009) identified C. colocynthis as having antifungal activities. The HPLC analysis in this study identified several compounds unique to C. colocynthis (Table 1; Figs. 1 and 2). However, further analysis is needed to determine if these differences exist in a large number of accessions representing the Citrullus species or subspecies and to determine what role these compounds may have in pest resistance. We have constructed genetic populations derived from crosses between C. colocynthis and C. lanatus var. lanatus or C. lanatus var. citroides PIs that will be further evaluated for the presence or absence of the compounds identified in this study and determine if any of these compounds are associated with resistance to whiteflies, aphids, or other pests.
AsyazS.HussainI.KhanF.MuranA.KhanI.U.2010Evaluation of chemical analysis profile of Citrullus coloynthis growing in southeastern area of Khyber Pukhtunkhwa PakistanWorld Appl. Sci. J.10402405
ChopraB.JhootyJ.S.BajajK.L.1974Biochemical differences between two varieties of watermelon resistant and susceptible to Alternaria cucunerinaPhytopathol. Zeitscrift794752
DelazarA.GibbonsS.KosariA.R.NazemiyehH.ModarresiM.NaharL.SarkerS.D.2006Flavone-C-glycosides and cucurbitacin glycosides from Citrullus colocynthisDaru J. Faculty Pharmacy Tehran Univ. Med. Sci.14109114
HadizadehI.PeivasteganB.KolahiM.2009Antifungal activity of nettle, colocynth, oleander and konar extracts on plant pathogenic fungiPak. J. Biol. Sci.125863
HarrisonH.F.JrMitchellT.R.PetersonJ.K.WechterW.P.MajetichG.R.SnookM.E.2008Contents of caffeoylquinic acid compounds in the storage roots of sixteen sweetpotato genotypes and their potential biological activityJ. Amer. Soc. Hort. Sci.133492500
IbrahimT.A.El-HefnawyH.M.El-HelaA.A.2010Antioxidant potential and phenolic acid content of certain cucurbitaceous plants cultivated in EgyptNat. Prod. Res.2415371545
JarretR.L.MerrickL.C.HolmsT.EvansJ.AradhyaM.K.1997Simple sequence repeats in watermelon [Citrullus lanatus (Thunb.) Matsum. & Nakai]Genome40433441
LeviA.ThomasC.E.KeinathA.P.WehnerT.C.2001aGenetic diversity among watermelon (Citrullus lanatus and Citrullus colocynthis) accessionsGenet. Resources Crop Evol.48559566
LeviA.ThomasC.E.WehnerT.C.ZhangX.2001bLow genetic diversity indicates the need to broaden the genetic base of cultivated watermelonHortScience3610961101
LopezR.LeviA.ShepardB.M.SimmonsA.M.JacksonD.M.2005Sources of resistance to two-spotted spider mite (Acari: Tetranychidae) in Citrullus sppHortScience4016611663
MeenaM.C.PatniV.2008Isolation and identification of flavonoid ‘quercetin’ from Citrullus coloynthis (Linn.)Schrad. Asian J. Exp. Sci.22137142
MujajuC.SehicJ.WerlemarkG.Garkava-GustavssonL.FatihM.NybomH.2010Genetic diversity in watermelon (Citrullus lanatus) landraces from Zimbabwe revealed by RAPD and SSR markersHereditas147142153
QiminL.Van Den HeuvelH.DelorenzoO.CorthoutJ.PietersL.A.C.VlietinckA.J.ClaeysM.1991Mass spectral characterizations of C-glycosidic flavonoids isolated from a medicinal plant (Passiflora incarnata)J. Chromatogr. B Biomed. Sci. Appl.562435446
RobinsonR.W.Decker-WaltersD.S.1997Cucurbits. CAB International Publishing Oxon UK
SAS Institute2003SAS/STAT user’s guide. SAS Institute Cary NC
SimmonsA.M.KousikC.S.LeviA.2010Combining reflective mulch and host plant resistance for sweetpotato whitefly (Hemiptera: Aleyrodidae) management in watermelonCrop Prot.29898902
SimmonsA.M.LeviA.2002Sources of whitefly (Homoptera:Aleyrodidae) resistance in Citrullus for the improvement of cultivated watermelonHortScience37581584
ThiesJ.A.LeviA.2007Characterization of watermelon (Citrullus lanatus var. citroides) germplasm for resistance to root-knot nematodesHortScience4215301533
TliliI.HdiderC.LenucciM.S.RiadhI.JebariH.DalessandroG.2011Bioactive compounds and antioxidant activities of different watermelon [Citullus lanatus (Thunb.) Mansfeld] cultivars as affected by fruit sampling areaJ. Food Compost. Anal.24307314
VenkataramaiahC.NarayanaR.K.1983Studies on indoyl-3-acetic acid oxidase and phenolic acid pattern in Cucurbitaceous fruitsIntl. J. Plant Physiol.111459463