Introduction
Our previous results concerning the antifeedant activity of xanthohumol (XN) and supercritical CO2 spent hops extract demonstrated that the extract has a moderately stronger deterrent activity than pure XN (Jackowski et al., Reference Jackowski, Hurej, Rój, Popłoński, Kośny and Huszcza2015). Therefore we decided to test other compounds that were previously identified in hops and spent hops by Stevens et al. (Reference Stevens, Ivanic, Hsu and Deinzer1997) or Chadwick et al. (Reference Chadwick, Nikolic, Burdette, Overk, Bolton, Van Breemen, Frolich, Fong, Farnsworth and Pauli2004, Reference Chadwick, Pauli and Farnsworth2006), except for xanthohumol flavone (FXN) and 1″,2″-dihydroxanthohumol K (DHXN K), which, along with naringenin (NAR), were used in our study as model derivatives. Additionally, the tested compounds were chosen based on the minute structure modifications of XN, resembling those that may occur during the storage and production of spent hops extracts (e.g., isomerisation, cyclization, oxidation, reduction) and may also provide some structure-activity relationship data. Little is known about the deterrent activity of flavonoids against grain products pests, thus different classes of flavonoids were used in the bioassay: chalcones: (XN, XN C, DHXN C, DHXN K, α,β-dihydrochalcone (DHXN), flavanones (IXN, NAR) and flavone (FXN).
Materials and methods
Racemic naringenin ((±)-NAR) was purchased from Sigma Aldrich.
XN was isolated from supercritical carbon dioxide extracted hops (‘Marynka’, harvest year 2012), obtained from Production of Hop Extracts (New Chemical Syntheses Institute, Puławy, Poland), following the method described by Bartmańska et al. (Reference Bartmańska, Huszcza and Tronina2009).
α,β-DHXN was obtained by the catalytic hydrogenation of XN, described by Popłoński et al. (Reference Popłoński, Sordon, Tronina and Huszcza2014).
Racemic isoxanthohumol ((±)-isoxanthohumol, IXN) was obtained by the alkaline isomerization of XN, described by Bartmańska et al. (Reference Bartmańska, Huszcza and Tronina2009).
XN C and 1″,2″-DHXN C were synthesized from XN applying methods, described by Vogel & Heilmann (Reference Vogel and Heilmann2008). The spectroscopic data of the obtained products are in agreement with literature values.
The spectroscopic data of the isolated XN and synthesized DHXN, IXN, XNC and DHXN C are in good agreement with the data given in the literature (Verzele et al., Reference Verzele, Stockx, Fontijn and Anteunis1957; Etteldorf et al., Reference Etteldorf, Etteldorf and Becker1999; Stevens et al., Reference Stevens, Taylor, Nickerson, Ivancic, Henning, Haunold and Deinzer2000; Chadwick et al., Reference Chadwick, Nikolic, Burdette, Overk, Bolton, Van Breemen, Frolich, Fong, Farnsworth and Pauli2004).
FXN was obtained in the reaction of IXN with iodine-pyridine complex described by Ndoile & van Heerden (Reference Ndoile and Van Heerden2013). The spectroscopic data of this new compound are included in the online supplementary materials of the manuscript.
1″,2″-DHXN K was obtained in the reaction of XN with trifluoracetic acid. The procedure of the synthesis of this compound is included in online supplementary materials of the manuscript.
Test insects and tests of feeding deterrent activity
The tests of feeding deterrent activity were carried out using three species of stored product pests: Granary Weevil (Sitophilus granarius L.) (Coleoptera, Curculionidae), Confused Flour Beetle (Tribolium confusum Duv.) (Coleoptera, Tenebrionidae) and Khapra Beetle (Trogoderma granarium Everts) (Coleoptera, Dermestidae), which had been selected originally by Nawrot et al. (Reference Nawrot, Harmatha, Kostova and Novotny1984) for their stored product pest status, and are still considered as model organisms for screening the deterrent activity of chemical compounds (Nawrot et al., Reference Nawrot, Bloszyk, Harmatha, Novotny and Drozdz1986, Reference Nawrot, Harmatha, Kostova and Ognyanov1989, Reference Nawrot, Dams and Wawrzeńczyk2009; Nawrot & Harmatha, Reference Nawrot and Harmatha2002; Jackowski et al., Reference Jackowski, Hurej, Rój, Popłoński, Kośny and Huszcza2015). Granary Weevil and Confused Flour Beetle are cosmopolitan, synanthropic species, which, in a temperate climate, occur only indoor. S. granarius feeds on stored grain of wheat, on barley, rye, oats, millet, maize and buckwheat. T. confusum feeds also on flour, cereal bran and on dried bread. The reason for including T. granarium is its tropical origin, as this fact broadens the possible scope of information about the activity of tested compounds and may render the results of the experiments meaningful for the regions to which this species is native.
The insects were reared in permanent darkness in climatic chambers. S. granarius was reared at 24 ± 1°C, whereas T. confusum and T. granarium – at 30 ± 1°C. The relative humidity was maintained at 70 ± 5% for S. granarius and at 60 ± 5% for the two other species. S. granarius L. was offered wheat grain of cv. NATULA as food and oviposition substrate. T. confusum and T. granarium were reared on commercially available oat flakes mixed in equal proportion with oat flour. The tests were carried out in the same rearing chambers.
Wheat wafer test
In order to test the feeding deterrent activity of the considered compounds, the wheat wafer disk bioassay was used. It was run as described by Nawrot et al. (Reference Nawrot, Bloszyk, Harmatha, Novotny and Drozdz1986, Reference Nawrot, Dams and Wawrzeńczyk2009) and identically as used more recently by Jackowski et al. (Reference Jackowski, Hurej, Rój, Popłoński, Kośny and Huszcza2015). All the compounds were tested as 1.0% ethanol solutions, except for the FXN, for which anhydrous THF (tetrahydrofuran) was applied as a solvent. In order to check whether the THF itself would not affect the test insects feeding preferences, a separate wafer test was launched in ten replicates using ethanol-soaked, THF-soaked and dry wheat wafers. The test demonstrated no statistically significant difference in the insects acceptance of the three substrates.
The wheat wafer disks of 1 cm diameter were dipped in the solvent solution of the tested compound or in the solvent alone (reference) using pincers, and placed on large glass Petri dishes (of 210 mm diameter), treatment and reference on separate dishes, in order to air-dry. After 30 min of drying, the disks were weighed and placed on polystyrene Petri dishes of 90 mm diameter. After the weighing of the disks had been completed in all the treatments, the test insects, adults or larvae, were placed on the Petri dishes, which were then covered with lids and placed in climate chamber conditions appropriate for the test species used. The tests were carried out in eight replicates. In a single replicate, three adults of S. granarius were used, or 20 adults or ten larvae of T. confusum, or ten larvae of T. granarium, following the description by Nawrot et al. (Reference Nawrot, Dams and Wawrzeńczyk2009) and by Jackowski et al. (Reference Jackowski, Hurej, Rój, Popłoński, Kośny and Huszcza2015). Unsexed, 7–10 day old adults or 5–30 day old larvae were used for each test. After 120 h the insects were removed from the Petri dishes and the remnants of the wheat wafers were weighed again.
The tests were conducted in a three part run simultaneously: reference (two wheat wafers treated with solvent alone), choice test (one solvent-treated wafer and one compound solution-treated wafer) and no-choice test (two compound solution-treated wafers). Based on the depletion of the wafers' weight during the test time, three indices of the compound activity were calculated, following the formulae used by Nawrot et al. (Reference Nawrot, Bloszyk, Harmatha, Novotny and Drozdz1986): relative (R), absolute (A) and total (T) deterrency coefficient, where:
$$R{\rm} = \left( {(C - E)/(C + E)} \right) \times 100\;({\rm choice}\;{\rm test}),$$
$$A = \left( {(CC - EE)/(CC + EE)} \right) \times 100\;({\rm no - choice}\;{\rm test}),$$
$$T{\rm} = {\rm} A + R{\rm}, $$
and C, CC stand for the consumed amount of the reference disks, whereas E, EE stand for the consumed amount of the compound-treated disks. C and E markings refer to choice test (one wafer of each kind in a single replicate), whereas CC and EE markings refer to reference and to the no-choice test, respectively (two wafers of the same kind in a single Petri dish).
The T coefficient is the only parameter finally used in the assessment of a compound activity. T values may range between −200 and +200. T values within the intervals 151–200 and 101–150 indicate very good and good deterrent activity, respectively. Compounds with values of T within the range 51–100 show medium deterrent activity and those showing T lower than 50 are weak deterrents. The negative value of the T coefficient indicates increased feeding. R and A values should not be interpreted alone as proving deterrency or increased feeding.
Statistical analysis
The total deterrence coefficient (T) served as the index of the biological activity of the tested materials. The data sets for individual compounds and test insect species and instars were checked for normality and homogeneity of their variances, based on Kolmogorov–Smirnov along with Lilliefors tests, and Levene test, respectively. Consequently, Kruskal–Wallis analysis of variance (ANOVA) of ranks was used to compare the T values obtained in wafer tests for particular compounds and insect species. The data were analyzed using Statistica 12 package (StatSoft, Inc. 1984–2014).
Results
In 19 out of 32 experiments, the Kolmogorov–Smirnov test demonstrated that the distribution of the T coefficient value was different from normal at P < 0.05, and in 31 out of 32 experiments the Lilliefors test had confirmed distribution different from normal at P < 0.01. The Levene test of the homogeneity of variances, run for the data categorized in the same way, revealed significant differences between the variances of the calculated T values (F = 3.14; P = 0.000001). The obtained data were therefore analyzed using Kruskal–Wallis ANOVA of ranks, categorized both by the tested compounds and by the test organisms. The resultant statistics are presented in table 1, and in the online Fig. S1 and Fig. S2, in the section A of the online Supplementary Materials. For the reader's convenience, the mean T-values and standard deviations (SD) are given in table 1, but the medians, 25–75 percentiles and minimum–maximum range, are shown in online Fig. S1 and Fig. S2, as they are considered more appropriate since the nonparametric statistics were used. The table 1 presents the test results in two different ways. In columns, the response of individual test organisms and their instars can be seen across the whole spectrum of the investigated chemicals. The rows, on the other hand, show the deterrent activity of each particular compound across the species and instars of the set of test organisms.
Deterrent activity (mean T-value ± SD and verbal description) of the tested compounds across the test insects and their instars.
1 The mean T-values followed by the same lowercase letter do not differ significantly within each single species category (columns, Kruskal–Wallis H and P in the bottom rows). The mean T-values followed by the same uppercase letter do not differ significantly within each single compound category (rows, Kruskal–Wallis H and P in the right-hand columns); ns and NS respectively – non-significant.
2 The negative value of T coefficient denotes, on the average, increased feeding rather than deterrent action.
Compound codes: XN- xanthohumol; IXN – (±)-isoxanthohumol; NAR – (±)-naringenin; DHXN – α,β -dihydroxanthohumol; FXN – xanthohumol flavone; XN C – xanthohumol C; DHXN C – 1″,2″-dihydroxanthohumol C; DHXN K – 1″,2″-dihydroxanthohumol K.
Out of the two species used in the adult stage, it was the Grain Weevil, S. granarius, which responded to three of the tested compounds with considerable reduction of its feeding activity (table 1: IXN, FXN, DHXN K; T > 100). The deterrency of IXN, FXN and DHXN K was not statistically different from that of the other substances that had shown only medium deterrency towards S. granarius (T < 100). It was only significantly higher compared with the weakly deterrent NAR (T < 50, table 1).
The adults of the Confused Flour Beetle, T. confusum, were less sensitive, with none of the tested compounds showing more than the medium deterrent effect on it, and with two of them showing weak deterrency only (table 1: XN, IXN; T < 50). Moreover, IXN, while showing good deterrency in S. granarius, worked as only a weak deterrent for T. confusum. Similar was true for XN, in which the T values indicated medium deterrency for S. granarius and weak deterrency for T. confusum (table 1, online Fig. S1). The activity of IXN in S. granarius was significantly higher compared with that of NAR, whereas the activity of both XN and IXN in T. confusum was not (table 1, online Fig. S1).
Conversely to adults, the larvae of T. confusum seemed to be more sensitive to FXN and to XN, as well as to IXN, and less sensitive to DHXN K (table 1, online Fig. S1). FXN can be classified as a good deterrent for the larvae, whereas XN, IXN and DHXN K – as medium or weak deterrents (table 1). When offered to the larvae of T. confusum, all the tested compounds have shown the deterrent activity that makes them assigned, by the Kruskal–Wallis test used, to one homogenous group (table 1, online Fig. S1).
The response of the larvae of the Khapra Beetle, T. granarium, to the assemblage of the tested compounds was most diverse compared with one of the other test species, as it varied from good deterrency to faintly increased feeding, spanning four classes on the applied assessment scale. The responses to chemicals classified into the neighbouring classes of ‘weak deterrency’ (0 < T < 50) and ‘increased feeding’ (T < 0) were not significantly different in statistical terms (table 1, Trogoderma granarium). The response of T. granarium to DHXN was more apparent compared with that observed in the larvae of T. confusum (table 1, medium vs. weak deterrency), and it differed significantly in T. granarium from the response to all the compounds showing the ‘increased feeding’ activity (table 1, T. granarium: XN C, DHXN C, DHXN K, T = −34.4 through −5.0). T. granarium responded to FXN at the level similar to that shown by the larvae of T. confusum and the adults of S. granarius, and the response was statistically different from the response to other compounds, except for DHXN, which made a homogenous group with FXN (table 1, online Fig. S1).
When the responses of different test insects were grouped for each of the tested compounds, the most common deterrent activity was apparently observed for FXN, which was a good deterrent for three out of the four test organisms used, and a medium deterrent for the adults of T. confusum: (table 1, online Fig. S2). Among the other compounds, DHXN K and IXN had shown good deterrent activity in S. granarius, being significantly higher than the deterrency observed in the two remaining test organisms (table 1, online Fig. S2).
All the other substances investigated (XN, NAR, DHXN, XN C, DHXN C) performed in all the test organisms at medium or weak deterrency level only (table 1, 50 < T < 100, T < 50, respectively), and three of the tested compounds, including DHXN K, mentioned before as good deterrent in S. granarius, had induced an increased feeding in one of the test species: T. granarium (table 1, online Fig. S2: XN C, DHXN C, DHXN K).
Discussion
Insect response to the tested compounds
In the majority of the results obtained in the experiment, the compounds classified into different deterrency classes, lie within one homogenous group determined by the statistical test. It is particularly true for the neighbouring deterrency classes, as may be seen in the adults of S. granarius and T. confusum, but can also be noticed in the larvae of T. confusum, in which all of the substances tested, although spanning three classes of deterrency, make up one homogenous group in statistical terms (table 1). The situation may point to an inadequate precision of the bioassay used, but it may equally well mirror the intrinsic differences in the acceptance of the treated food by the individual insects. The T value observed in the test replicates exceeded the value of 150 a number of times, indicating very good deterrency in individual insects. The examples are FXN, DHXN C and IXN for the larvae of T. confusum: T = 190.2; T = 170.9; T = 184.6, respectively; as well as FXN and DHXN for the larvae of T. granarium: T = 160.8; T = 158.1, respectively. These numbers, being replicate values, are not shown in the table 1, their occurrence may be nevertheless inferred from percentiles represented in online Fig. S1 and Fig. S2. The fact that the mean T values merely exceeded or were <100 or even <50 for these same substances (table 1: see IXN, DHXN K and DHXN C in S. granarius, DHXN in the larvae of T. confusum and FXN in all species) is likely a manifestation of the broad physiological variation in the laboratory colonies of the test insects.
Schoonhoven et al. (Reference Schoonhoven, Van Loon and Dicke2012), while discussing ‘experience-induced changes in host-plant preference’ point at the phenomenon, which they call habituation to deterrents. The authors discuss natural situations, in which a herbivore is confronted to an assemblage of plant secondary compounds simultaneously, and one or more of them show deterrent action towards the invader. In such a case, habituation may occur on the condition that, as the authors put it, ‘desensitizing of the gustatory systems’ would not kill the organism as a result of the deterrent compound being a toxin, which the insect cannot detoxify. Contrastingly, the present experiment makes the test insects face just one chemical at a time, offered on the inert substrate. The more, some degree of habituation seems possible, particularly in the no-choice part of the wafer test, and this may also lead to the considerable intra-group variation, demonstrated in many cases by the non-significant statistics. As the variation expected in genuine, local populations of the stored product pests can be higher compared with the laboratory colonies, drawing applicative conclusions at this stage of the research would be unwise, for it might well result in misjudging the compounds that could potentially prove effective under different, non-laboratory, circumstances.
Structure – activity relationship
Our results indicate some deterrent activity differences related to the structure of the tested compounds (table 1). The most significant increase in antifeedant activity in all the test insects was observed in FXN compared with IXN, i.e., after the introduction of additional C-2, C-3 double bond in flavanone. The difference between the IXN and FXN was significant only in adults of T. confusum and in larvae of T. granarium, although in the latter species the distance was wider on the deterrency class (table 1: T. granarium, weak vs. good, IXN vs. FXN, respectively). Similar observations were made in the study by Stompor et al. (Reference Stompor, Dancewicz, Gabryś and Anioł2015), using peach potato aphid (Myzus persicae Sulz.) (Hemiptera), where flavone was more active than flavanone. On the other hand, studies of the activity of different flavanones and flavones containing additional C-3 hydroxyl group on termites (Coptotermes formosanus Shiraki) (Blattodea (Isoptera)) (Ohmura et al., Reference Ohmura, Doi, Aoyama and Ohara2000) indicate that such statement is imprecise. Yet, as the compounds investigated in our study, except for XN and IXN, had not been tested before on stored product pests, comparing the results with those obtained in other insect orders is nothing but a necessity. Otherwise, one may presume that the more directed and coherent an activity across a number of taxonomic groups used in tests, the more likely there is a genuine biological phenomenon behind the observed results.
Comparing chalcone vs. flavanone classes of flavonoids, thus considering heterocyclic dihydropyran ring formation as a factor of deterrence, it appears from our studies that IXN (flavanone) exhibits higher deterrent activity in adults of S. granarius than XN (chalcone) (table 1). It is worth mentioning, despite the use of different test insects that Simmonds et al. (Reference Simmonds, Blaney, Delle Monache and Marini Bettolo1990) had observed a significant increase in the deterrent activity of flavanones compared with the corresponding chalcones in Spodoptera exempta (Walker) and Spodoptera littoralis (Boisduval) (Lepidoptera) (also reported by Simmonds (Reference Simmonds2001)), and Stompor et al. (Reference Stompor, Dancewicz, Gabryś and Anioł2015) had recorded a similar relationship in Myzus persitae (Homoptera). In other work, although not including flavanones, Morimoto et al. (Reference Morimoto, Kumeda and Komai2000), while analyzing different chalcones and flavones, had noted the importance of the pyran ring as a factor positively affecting the antifeedant activity of flavonoids against the common cutworm (Spodoptera litura F.). Scarce literature data and uncertain results comparing chalcone and dihydrochalcone, as the most similar flavonoid classes, represented in the present study by XN and DHXN, make any general conclusions difficult. Furthermore, a dihydrochalcone – phloretin, tested in the studies of Ohmura et al. (Reference Ohmura, Doi, Aoyama and Ohara2000), was less active than the related NAR, but still more active than other tested flavans and flavonols. Therefore it may be provisionally concluded that the rigid structure of flavones, containing both heterocyclic ring and the C-2, C-3 double bond, is most probably the appropriate factor making the difference between the deterrent activity of these flavonoid classes.
Another structure–activity relationship may be inferred from the comparison of structures between flavonoid classes. NAR, used here as a reference natural deterrent and at the same time a model flavanone, works as a weak deterrent for all the test organisms (table 1). This may suggest the importance of the prenyl moiety at C-8 position that is absent in NAR, for the deterrent activity of flavanones (table 1: compare the activity of XN and IXN in S. granarius and in the larvae of T. confusum, and that of FXN in all the test organisms). The observation of the weak deterrency of NAR may be associated with the evolutionary adaptation of the species of test insects to the presence of NAR in their food. NAR occurs in grain pericarp and is even overproduced as a result of Fusarium infection, as was demonstrated by Buśko et al. (Reference Buśko, Góral, Ostrowska, Matysiak, Walentyn-Góral and Perkowski2014). It therefore seems likely that grain feeders face it on a regular basis in natural situations. Schoonhoven et al. (Reference Schoonhoven, Van Loon and Dicke2012) have described the phenomenon of herbivore habituation to feeding deterrents, which may partly explain the weak response of all the test organisms to NAR in the present study.
Conversely, Lane et al. (Reference Lane, Sutherland and Skipp1987) when testing nine isoflavones as feeding deterrents against 3rd instar larvae of Costelytra zealandica (White) and Heteronychus arator Fabr. (Coleoptera, Scarabaeidae), had observed the opposite activity of prenyl moiety: it reduced the antifeedant activity of the tested compounds. A similar conclusion may be found in the studies of Simonds et al. (Reference Simmonds, Blaney, Delle Monache and Marini Bettolo1990), which compared prenylated and non-prenylated chalcones and flavanones while using lepidopteran test species (Spodoptera exempta and S. littoralis). As the test species used in both the studies mentioned are either root feeders (the larvae of scarabaeid beetles in Lane et al., Reference Lane, Sutherland and Skipp1987), or leaf eaters (the lepidopteran species studied by Simmonds), it may be speculated that the deterrent value of the prenyl moiety in isoflavones occurs only in grain-feeders such as the coleopterans used in the present study, whereas the same structure reduces the deterrent effect in species foraging on fresh plant material.
Interesting structure–activity interactions may also be speculated from the results obtained for XN C, DHXN C and DHXN K, being chalcones with a cyclized prenyl group, containing saturated and unsaturated dimethylpyran moiety. In general, the cyclization of the prenyl group increases deterrent activity in the adults of S. granarius and in the larvae of T. confusum, although the relationship is inversed in the two organisms. In the case of the adults of S. granarius, DHXN K is a good deterrent, whereas DHXN C and XN C work at a medium deterrency level. In the larvae of T. Confusum, DHXN K works as a weak deterrent, whereas XN C and DHXN C show higher (medium) deterrent activity. In the adults of T. confusum all three of the discussed compounds fit into the ‘medium’ deterrency class and are there along with DHXN and FXN, which confirms that the insect remains indifferent in the face of the structural alterations presented to it (table 1).
Simmonds et al. (Reference Simmonds, Blaney, Delle Monache and Marini Bettolo1990) and Lane et al. (Reference Lane, Sutherland and Skipp1987) also observed that the cyclization of prenyl group is an important factor that positively affects the deterrent activity of the tested compounds. Additionally, the results obtained by Simmonds et al. (Reference Simmonds, Blaney, Delle Monache and Marini Bettolo1990) coincide with our conclusions regarding the direction of prenyl group cyclization. The comparison of XN C and DHXN C activity for all the tested pests indicates no significant effect of additional π-bond in the dimethylpyran ring on the deterrent activity, although the additional bond slightly reduces antifeedant potency (table 1: T XNC < T DHXNC). Surprisingly, all the tested compounds containing dimethylpyran moiety induce an increased feeding in the larvae of T. granarium.
Conclusions
We conclude that the introduction of additional structural fragments, such as prenyl or dimethylpyran moiety, to a flavonoid structure may significantly alter its deterrent activity. The prenyl moiety in flavonoids is likely the reason of higher deterrent activity of a compound in grain-feeder species, such as the coleopteran store-product pests used in the present study. On the other hand, it appears from other studies (Lane et al. Reference Lane, Sutherland and Skipp1987; Simonds et al. Reference Simmonds, Blaney, Delle Monache and Marini Bettolo1990), that the same structure reduces the deterrent effect of a compound in species foraging on fresh plant material.
For the future studies it can also be assumed that, as in most insect interactions with the flavonoids described in the literature, also in S. granarius and in T. confusum, the introduction of dimethylpyran moiety to the flavonoid structure may increase the deterrent activity of the tested compound, although it may be kept in mind that an increased feeding was observed in T. granarium in response to the same modification.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0007485317000050
