The kin selection hypothesis (KSH) has been invoked in several studies to explain the observation of decreased herbivory in plants exposed to VOCs emitted from damaged plants that are more genetically related compared to those that are less related (Karban et al. 2013, 2014a; Moreira et al. 2016; Hussain et al. 2019). Many of these studies neglected to consider a simpler alternative hypothesis of volatile-mediated associational resistance (VMAR), where repellent or toxic VOCs are adsorbed on neighbouring plants rendering these intact plants more resistant to herbivores (Himanen et al. 2010, 2015; Kessler and Kalske 2018). Our results demonstrated that sagebrush plants actively responded to VOC cues from damaged plants. Moreover, stronger volatile-mediated induced resistance (VMIR) was elicited when emitter and receiver plants shared the same heritable chemotype but only for two of the five chemotypes tested. We detected the ability for sagebrush plants to adsorb and subsequently release repellent VOCs. Consequently, both VMIR and VMAR may be responsible for observed decreases in herbivory after exposure to damaged plants depending on the chemotypes of receiver and emitter plants involved.
In alignment with numerous studies from several researchers that have repeatedly demonstrated the ability of sagebrush plants to respond to wounding cues from damaged neighbors [e.g., (Karban et al. 2006; Pezzola et al. 2017; McMunn 2017; Grof-Tisza et al. 2020)], we found that receiver plants exposed to DIPVs experienced less herbivory by a generalist herbivore compared to control plants. More notably, we showed for the first time that this reduction in herbivory was associated with the up-regulation of defense-related genes. Exposure to DIPVs can prime plants without transcriptional changes (Engelberth et al. 2004; Heil 2014). Consequently, not detecting a response at the gene-level is not sufficient evidence to conclude that plants did not perceive DIPVs. Conversely, finding an increase in gene expression is a strong indicator that plants actively perceived and responded to VOC cues. Our ability to detect direct induction through the up-regulation of genes enabled us to distinguish between active and potentially passive responses. These results validated our approach to investigate volatile-mediated interactions under laboratory conditions in a system historically studied using manipulative field studies.
While it is well established that plants can respond to DIPVs emitted by nearby plants (Karban et al. 2006; Kost and Heil 2006; Li and Blande 2017), how the traits of emitter and receiver plants and their interactive effects influence the ability of receiver plants to respond to these cues is less understood. Several investigations reported that the degree to which a plant responded depended on its relatedness to the damaged emitter (Karban et al. 2013; Moreira et al. 2016; Hussain et al. 2019). A few of these studies specifically assessed the role of heritable chemotypes in volatile-mediated signalling (Karban et al. 2014a; Hussain et al. 2019). For example, Karban et al. (2014) concluded that communication was more effective between sagebrush plants sharing the same chemotype relative to when the chemotypes were different. Based on this finding, the authors posited that kin selection could explain the selection for ‘private-channels of communication’ between related individuals. These authors, like many of those listed above, assumed plants actively responded to DIPVs, but never confirmed this assumption by quantifying physiological or transcriptional changes. This additional verification is crucial to distinguish between active and passive responses as these same observations are possible via VMAR. Here, we obtained the equivalent result as first described by Karban and colleagues (2014) in our second exposure experiment with two chemotypes, α-thujone and artemiseole; exposure of a receiver plant to DIPVs from an emitter plant of the same chemotype resulted in less herbivory compared to that from a different chemotype. Unlike the original study, we obtained evidence indicative of an active response. No evidence was detected to suggest that the decrease in herbivory was a function of adsorbed VOCs. Taken together, these results confirm the conclusions of previous work and lend additional support for the KSH.
As mentioned above, a previous field study with sagebrush investigated volatile-mediated interactions between two chemotypes, a-thujone and camphor (Karban et al. 2014b). Due to a low sample size resulting from lack of herbivory in our feeding assays, we could not rigorously assess this combination of chemotypes in our first exposure experiment with four chemotypes which included the two originally tested. The pairing of chemotypes in our second exposure experiment involving α-thujone and artemiseole was chemically similar to that of α-thujone and camphor. The volatile blend of camphor resembles that of artemiseole (Fig. 6) (Grof-Tisza et al. 2021). One hypothesis that may explain this parallel pattern of induction between studies involving plants of similar chemotypes is that the chemical dissimilarity between α-thujone and the camphor dominant chemotypes of camphor and artemiseole prevent the cross recognition of volatile cues. In ordination space, α-thujone clusters quite distinctly from artemiseole and camphor (Fig. 6). This hypothesis is supported by one model explaining the evolution of plant-to-plant communication, which is thought to have evolved as a by-product of within-plant signalling (Heil and Karban 2010). Selection may reinforce the specificity of chemotype-specific alarm cues as the ability to respond to specific cues and minimize eavesdropping confers a competitive advantage. Numerous studies have demonstrated the high level of specificity of volatile cues involved in plant signalling (Erb et al. 2015; Moreira et al. 2018; Ninkovic et al. 2021).
Plants can adsorb and reemit toxic or repellent VOCs that confer associational resistance to insects (Himanen et al. 2010, 2015; Mofikoya et al. 2020), pathogens (Camacho-Coronel et al. 2020) and can alter the cues used by natural enemies to locate their hosts (Bui et al. 2021). We detected artemisia ketone in the headspace of receiver plants generally not associated with this volatile compound suggesting it was adsorbed and reemitted. Additionally, we detected higher concentrations of this VOC in the headspace of receiver plants of the artemisia ketone chemotype after exposure to damaged emitter plants of the same chemotype. Because artemisia ketone is known to repel insects (Liu et al. 2021), it is plausible that receiver plants might have benefited from the increased protection provided by the adsorbed VOC in addition to its own emission. This could explain the observed preference by C. morosus for leaves other than those exposed to artemisia ketone emitter plants; if herbivores respond to adsorbed VOCs in a dose-dependent fashion, exposure to DIPVs of a plant sharing the same chemotype would result in less damage. This observation might lead to the erroneous conclusion that plant-to-plant signalling is more effective between like chemotypes when it is a result of VMAR. Here, we found that plants exposed to artemisia ketone were associated with increased transcription of all genes tested relative to control plants suggesting that the reduced herbivory was at least in part a function of VMIR. It is conceivable that both VMAR and VMIR contributed to decreased consumption of artemisia ketone exposed leaves. However, we are unable to determine the relative importance of each in this study. We did not detect adsorption of the other chemotype-dominant compounds tested, α-thujone, camphor, artemiseole and β-thujone. Studies investigating interspecific VMAR benefited from the presence of uniquely expressed VOCs, enabling researchers to easily track deposited VOCs (Bui et al. 2021). Contrastingly, the VOCs assessed here were not uniquely emitted by each chemotype, and their emission rates varied substantially even within the same chemotype. Considering this variation, differentiating between adsorption and primary emission is not easily accomplished. It is possible these VOCs contributed to VMAR to some extent.
Several studies have demonstrated threshold effects in plants in response to stress [reviewed in (Niinemets et al. 2014)]. For example, Karl et al. (2008) found that under moderate levels of thermal stress, volatile phytohormones and induced VOCs were absent or were detectable at low levels. Above a particular threshold, LOX products and methyl salicylate increased substantially. Though largely untested, theory predicts the selection for threshold-mediated responses to volatile alarm cues as a means to conserve resources for more substantial or immediate threats (Orrock et al. 2015). Indeed, a mechanistic explanation of plant priming is that threshold levels of stressors that trigger the activation of plant defense are decreased when plants are primed (Morrell and Kessler 2014). VMAR in conjunction with threshold-mediated induced resistance provides an alternative explanation to that of plants differentially responding to cues from kin and strangers. The accumulation of adsorbed VOCs on a plant emitted by damaged neighbours of a similar chemotype could trigger transcriptional changes upon reaching a critical threshold of a particular VOC cue. While the effect of exposure to a chemically similar individual is the same as VMIR, the selective driver of this effect may not involve kin recognition. To our knowledge, such a mechanism has not been described.
We detected minor differences in the emissions of several VOCs across multiple functional classes between exposed plants and filtered air. The emission of several of these compounds are known to be inducible and exhibit repellent properties, reduce damage by herbivores, or aid in indirect defenses. For example, β-caryophyllene emission was increased in intact maize after exposure to wounding signals (Engelberth et al. 2004); it was shown to repel a psyllid pest using Arabidopsis over-expression and knock-out lines (Alquézar et al. 2017) as well as attract natural enemies of Spodoptera caterpillars (Köllner et al. 2008). A few of these differentially emitted VOCs were used in chemotype assignment, including α-thujone, artemisia ketone, and artemisia triene, (Grof-Tisza et al. 2021) and are known to function as direct defenses (Obeng-Ofori et al. 1998; Mesbah et al. 2006; Tampe et al. 2015; Liu et al. 2021). Direct induced emissions are often less pronounced than the emissions seen following secondary damage after previous VOC exposure (i.e., priming) [reviewed in (Frost et al. 2008)]. Mechanically damaging receiver plants prior to the collection of headspace VOCs may have yielded larger effects although this was not tested here. It is possible that the observed increased emissions for at least some of these VOCs may stem from the adsorption and reemission and not direct induction. Given the growing evidence of VMAR, researchers should exercise caution when interpreting VOC emissions of plants exposed to DIPVs of strongly aromatic plants like sagebrush.