4.2. Investment of C in isoprenoid emissions shifts to heavier compounds when photosynthesis is reduced in the dry season
From the wet to dry season, we found a general trend of decreasing photosynthesis rates, coupled with relatively constant total isoprenoid emission rates (Fig. 4). This combination led to a three-fold average increase in C allocation to isoprenoid emissions relative to photosynthesis (though this was variable and non-significant, Fig. 4). Even so, the C sink to isoprenoid emissions was less than 2% of photosynthesis in most plants, except in strongly sesquiterpene-emittingP. hebetatum , whose emissions averaged 6.1% of photosynthesis (occasionally > 10%) in the upland in the dry season. Associated with this increasing investment in emissions, plants also allocated more C toward heavier isoprenoid compounds (Fig. 4). Increasing ‘isoprenoid mass investment’ was a result of reduced isoprene emissions and increased monoterpene and sesquiterpene emissions in the dry season (Figs. S2).
Previous studies indicated that seasonality in isoprenoid emissions from Amazonia is controlled by changes in solar radiation, temperature and leaf demography and phenology (Kuhn et al. 2004b; Alves et al. 2016, 2018). It is well known that the central Amazon dry season reaches maxima in solar radiation, temperature and leaf flushing, and the latter translates into a higher proportion of canopy young leaves during this season (Lopes et al. 2016; Wu et al. 2016; Gonçalves et al. 2020). In terms of changes in chemical composition of isoprenoid emission capacities, leaf age is an important factor because the synthesis of these compounds changes with leaf ontogeny. Isoprene and monoterpenes are produced from dimethylallyl diphosphate (DMADP), which is catalyzed by isoprene synthase and monoterpene synthases, in the methylerythritol 4-phosphate (MEP) pathway in chloroplasts. The DMADP pool size varies with leaf ontogeny, meaning that in early developmental stages, leaves have low DMADP – due to high demand for leaf growth and greening, and it increases with leaf maturation. A low DMADP pool in young leaves may be responsible for a down-regulation in the isoprene synthesis and up-regulation in the monoterpene synthesis; this might be in part explained by the fact that the enzymes involved in the monoterpene synthesis pathway have higher affinity for DMADP, implying that these enzymes achieve maximum catalytic efficiency already at low DMADP concentration, which favors the synthesis of monoterpenes rather than isoprene in young leaves (Kuhnet al. 2004b). As constitutive emission of monoterpenes and isoprene has been suggested to have similar functional roles to protect plants against high temperatures and oxidants (Loreto et al.1996), when these stresses - typical of the Amazon dry season - are imposed to young leaves and isoprene synthesis is constrained by enzyme deficits or low DMADP concentration (Kuhn et al. 2004b), the production of monoterpenes might be preferable and then confer the high demand for leaf protection.
Additionally, the plasticity in emission capacities and chemical compositions may not exclusively be a response to abiotic stress, but they may signal up-regulation of defenses against herbivores and pathogens, to which the plant is more susceptible under abiotic stress (Monson, Weraduwage, Rosenkranz, Schnitzler & Sharkey 2021). Monoterpenes and sesquiterpenes have long been recognized as important compounds for plant defense against herbivores and pathogens (Gershenzon & Dudareva 2007). Young leaves have more soft tissues and less physical and structural defenses that make them more palatable and attractive to predators than mature leaves. In this sense, the increase in emitted monoterpenes and sesquiterpenes (in P. hebetatum ) during the dry season might also be interpreted as an increase in the demand for defense by new leaf tissues of the young leaves.
A higher proportion of young leaves during the dry season is clearly seen in the study site presented here (Lopes et al. 2016; Gonçalves et al. 2020). However, it was also observed that approximately 30% of trees do not have a massive leaf flushing in any time of the year, meaning that they are possibly flushing leaves continuously and therefore do not have a distinct seasonal leaf age distribution (Lopes et al. 2016). As the pattern of leaf phenology is plant species-specific and we do not have information on the leaf age for the individuals that we performed our measurements with, we cannot exclude other reasons for the shift in composition of isoprenoid emission capacities rather than leaf age. If significant changes in leaf age is unaccountable (i.e. specifically for these tree species and in the moments that our measurements were performed), then what might have been playing a role is the leaf physiological response to the environment, which is also affected by the environmental conditions previous to measurements, and not only to the standard conditions imposed to leaves during measurements.
The dry and wet season environment in central Amazonia is different for the leaf - the wet season imposes more light-limitation and the dry season imposes more vapor pressure deficit and higher temperature limitation to photosynthesis (Wu et al. 2017). The limitations imposed during the dry season are more stressful for plants, which might require more or different defense mechanisms. Recent work on the integration of isoprene, monoterpenes, sesquiterpenes and other secondary metabolites in signaling networks that mediate growth-defense tradeoffs (Monson, Weraduwage, Rosenkranz, Schnitzler & Sharkey 2021), suggests that plasticity and changes in emission composition (as can be observed seasonally and/or across habitats) could enhance plant chemical defense and therefore confer to the physiological system more resistance against different types and levels of stress - e.g., reducing isoprene and increasing monoterpene and sesquiterpene in the Amazon dry season might represent a switch in plant gas composition in order to cope with stresses such as heat and drought, and to prime the plant to associated biotic stresses.
While the dry-season increase in ‘isoprenoid mass investment’ was most significant for P. hebetatum , it was mirrored by non-significant trends within most species and habitats (Fig. 3). We devised the ‘isoprenoid mass investment’ metric to assess whether changing environmental conditions are associated with shifts in C allocation to emissions along the spectrum from lighter to heavier isoprenoids. Many light dependent isoprenoids are evidenced to play related metabolic-support roles as components of cellular signalling networks and antioxidant defenses (Vickers, Gershenzon, Lerdau & Loreto 2009; Riedlmeier et al . 2017; Frank et al. 2021; Zuo et al. 2019; Monson, Weraduwage, Rosenkranz, Schnitzler & Sharkey 202; Harrison et al. 2013). While the diversity of specific compounds is high, there are general trends along the mass gradient from 5C isoprene through 10C monoterpenes to 15C sesquiterpenes. Notably, along the low-to-high mass gradient, volatile isoprenoid production has a decreasing dependency on photosynthesis. Compared to monoterpene synthase enzymes, isoprene synthase has a low reaction affinity for the primary isoprenoid substrate DMADP (Harrison et al. 2013; Kuhnet al. 2004b). Isoprene therefore requires a larger substrate pool in the chloroplast that is derived from photosynthesis (Sharkey & Monson 2014). This variation in emission capacity with variation in the supply of photosynthetically derived substrates is likely an explanation for why species either emit monoterpenes or isoprene at high rates, but not both (Harrison et al. 2013). Sesquiterpene biosynthesis is at least partially localized to the cytosol and use Isopentenyl Pyrophosphate (IPP) and DMADP from the Mevalonate pathway (MVA) (Sallaud et al. 2009), which further reduces their dependency on photosynthesis and DMADP pools in the chloroplast, relative to the smaller isoprenoids. This implies that seasonal changes in photosynthesis may affect isoprene and higher isoprenoid production, since the amount of DMADP available for isoprene synthesis is not only dependent on upstream flux through the MEP pathway, but also its downstream use for the production of higher compounds, such as monoterpenes and sesquiterpenes (Monson, Weraduwage, Rosenkranz, Schnitzler & Sharkey 2021), that can be demanded after distinct biotic and abiotic stresses over seasons (Harrison et al. 2013).  In addition, the heavier compounds tend to be much more reactive, as evidenced for example by the typical atmospheric half-lives of isoprene (hours), monoterpenes (minutes), and sesquiterpenes (minutes) (Atkinson & Arey 2003). Higher reactivity during the short residence time of volatiles within the cell may allow for a stronger effect as signaling molecules or even as direct antioxidants at lower production rates.
We hypothesize that as photosynthesis declines, an investment shift toward emissions of heavier isoprenoid compounds may allow sustained metabolic support due to their increasing reactivity and decreasing dependency on photosynthesis for production. Consistent with this hypothesis, our results show that when photosynthesis was reduced under dry season conditions, there was a tendency among all species to shift emission compositions toward heavier compounds. Interestingly, the only species to emit detectable sesquiterpenes, P. hebetatum , emitted them only in the dry season, and showed significantly increasing emission rates along the soil hydrologic gradient from shallow (wetter) to deep (drier) water table depths (Fig. 2), suggesting an association between investment in sesquiterpenes and the seasonal drought conditions experienced by individual plants. We suggest that our hypothesis is consistent with our empirical data and general theory, but that further experimental physiology work is required to demonstrate emission rate tradeoffs, photosynthetic dependencies, and the relationship between emission rates and metabolic responses for specific isoprenoid compounds.
Although there is variability within and among the tree species studied here and it is not possible to generalize results from three species to a whole plant community, here we presented results of species that are widespread in the central Amazon and have meaningful contribution to plant biomass and productivity (Fauset et al. 2015). Furthermore, the trend of reduction in isoprene emission with proportional increase in monoterpene and sesquiterpene emission during the dry season has implications for atmospheric chemical and physical processes. As for aerosol formation, laboratory-determined secondary organic aerosol (SOA) yield from isoprene has been reported as < 6% (Kroll, Ng, Murphy, Flagan & Seinfeld 2005; Xu, Kollman, Song, Shilling & Ng 2014) or higher over forested regions when isoprene is emitted in much more quantities relative to other compounds (Carlton, Wiedinmyer & Kroll 2009). Nonetheless, if other isoprenoids are more emitted, the contribution to the SOA formation could be also higher, as shown by SOA yields of ~5-10% for monoterpenes (Griffin, Cocker, Flagan & Seinfeld 1999a; Griffin, Cocker, Seinfeld & Dabdub 1999b) and ~20-70%  for sesquiterpenes (Hoffmann et al.1997; Griffin et al. 1999b; Lee et al. 2006a b; Chen, Li, McKinney, Kuwata & Martin 2012; Jaoui, Kleindienst, Docherty, Lewandowski & Offenberg 2013).
In this light, it is important for future studies to consider a wider range of volatile organic compounds with their synergetic importance in plant ecophysiological processes and the subsequent impact on the atmosphere. Our results suggest that emissions of monoterpenes and sesquiterpenes might be higher than anticipated and indicate a seasonal change in the composition of the emitted isoprenoids. In fact, seasonal shifts in monoterpene composition have been already reported in ambient air (Jardine et al. 2015; Yáñez-Serrano et al. 2018); but sesquiterpenes might have been underestimated given their high reactivity with ozone and OH and thereby the difficulty to detect in ambient air (Yee et al. 2018), indicating that only leaf level measurements are likely to give us a true measure of forest sesquiterpene emissions. Even though our observations do not offer results to make a clear statement of how isoprene, monoterpenes and sesquiterpenes work together in the face of abiotic stress such as heat and drought during the dry season, they supports the recent findings that isoprene does not confer plant tolerance by itself, but rather it is likely part of an intricate network that involves other plant metabolic processes and therefore other compounds (Harvey, Li, Tjellström, Blanchard & Sharkey 2015; Monson, Weraduwage, Rosenkranz, Schnitzler & Sharkey 2021). This opens more questions in order to understand what are the processes regulating isoprenoid emission capacities of Amazonian trees, and the need to consider seasonal shifts in isoprenoid composition in emission modeling.