Populus salicinoids: a thriving subfield in the omics era

Abstract

Members of the salicaceous genus Populus are primarily used by plant biologists as a model system for understanding the genetic underpinnings of woody plant growth and development. Beyond their importance to those conducting developmental research, Populus spp. are key members of ecosystems in the Northern Hemisphere and show promise as a vital renewable source of biomass for sustainable biofuel production. This genus also produces a class of signature herbivore-deterring and medicinally significant phenolic glycosides, commonly referred to as salicinoids. Although salicinoids in Populus are primarily associated with defense against biotic disturbances, they have also been implicated in structuring the chemotaxonomy of Populus and Salicaceae, shaping endophytic microbiomes, directing abiotic stress responses and participating in primary metabolism. Despite advancements in understanding these interactions through functional genomics and biotechnological techniques such as CRISPR/Cas9, much about their function and biosynthesis still remains obfuscated. Here, we summarize a global view of progress made in Populus salicinoid research, focusing particularly on studies conducted through a biotechnological lens, to elucidate the distribution, ecological significance, and biosynthesis of these compounds.

Introduction

The vast amount of structural diversity and specificity observed in plant secondary metabolites across and within taxa, and the ubiquity of these compounds in higher plants has led to the suggestion that production of these compounds served adaptive purposes necessary for the progenitors of land plants to make their transition out of the water c. 500 million years ago, and continue to serve adaptive purposes in modern higher plants (Close and McArthur 2002, Wink 2003, Boeckler et al. 2011, Maeda and Fernie 2021). Phenylpropanoids, or molecules originating from the reactions catalyzed by phenylalanine ammonia-lyase (PAL), are the most abundant class of plant secondary metabolites, and are key contributors to shoot structural integrity, antioxidant activity, molecular signaling, herbivory defense, pollinator attraction, and protection from abiotic stressors such as UV radiation (Harborne and Nash 1984, Smiley et al. 1985, Li et al. 1993, Jones et al. 2001, Vlot et al. 2009). For example, BAHD acyltransferases and UDP-dependent glycosyltransferases (UGTs) are some of the primary families of enzymes responsible for the synthesis of diverse secondary metabolites and have evolved to become highly neofunctionalized over the hundreds of millions of years that tracheophytes have spent living, surviving and diversifying in terrestrial habitats (Yonekura-Sakakibara and Hanada 2011, Moghe et al. 2023). Biotechnological innovations in plants have allowed for the elucidation of these enzyme families, and their subsequent use in the synthesis of pharmaceuticals, cosmetics, and other high-value chemicals and molecules (Schmidt et al. 2015, Kwon et al. 2023, Bibik et al. 2024, Kruse et al. 2024). Understanding how these valuable compounds are produced biologically, and equally important, could be produced at higher concentrations can be industrially valuable, and may even be used in tandem in scenarios where feedstock are designed to produce sustainable and less recalcitrant biofuels to mitigate anthropogenic climate change (Mottiar et al. 2016, Mahon and Mansfield 2019). Poplar trees hold much value as a potential biofuel feedstock, a platform to produce valuable plant metabolites, and a keystone species in delicate riparian ecosystems (Braatne et al. 1996, Sannigrahi et al. 2010, Bryant et al. 2020). A greater understanding of Populus’ unique metabolism and its medicinally and ecologically significant metabolites will undoubtedly prove valuable. This review summarizes current knowledge on Populus salicinoids and their metabolism, as well as what has been discovered about their ecological significance and the genomic architecture and mechanisms underlying their production, with the aspiration of providing a road map and guidance for future biotechnological studies in this ecologically and economically important tree species.

Salicinoids, also commonly referred to as phenolic glycosides or salicinoid phenolic glycosides (Boeckler et al. 2011), are a class of secondary metabolites generally associated with herbivory defense in the Salicaceae. Salicin (Figure 1 [I]), a salicyl alcohol ether-linked to a 1-D-glucopyranose moiety, was first characterized in 1828 by Johann Andreas Buchner, a German pharmacologist, following its isolation from willow bark. Following the initial identification, early work continued to refine extraction techniques and elucidate basic chemical information (Buchner 1836, Pelouze and Gay-Lussac 1930, Partington 1964, Fischer and Ganellin 2010). In tandem, a significant body of research since the first isolation of salicin explored its medical uses on rheumatism, influenza and other ailments (Maclagan 1876, Pernet 1926, Craig 1927). Since the initial discovery of salicin, other compounds that contain salicin as a base chemical skeleton, such as populin [III], have been identified (Braconnot 1831). In recent years, with the advent of nuclear magnetic resonance (NMR) spectroscopy and advanced chromatography mass spectroscopy, several salicin-based compounds have emerged among Salicaceae, including the identification of complex salicinoid compounds found in poplar bark and leaves, such as tremuloidin [IV], trichocarposide [VII], and deltoidin [V] (Pearl and Darling 1959, 1971c, Estes and Pearl 1967). At the same time, reports disclosed more ubiquitous salicinoids such as salicortin [VI] and tremulacin [IX] (Thieme 1964, Thieme and Richter 1966). These seminal studies used hot-water extraction techniques to establish the foundational knowledge in the field; however, later inquiries found that higher-order salicinoids were highly susceptible to degradation under these conditions, and consequently, it was shown that the higher-order, complex salicinoids could be hydrolyzed, and therefore, compound abundance is heavily dependent on the mode of tissue preparation (Lindroth and Pajutee 1987). Due to this revelation many years after, these foundational studies, although highly informative, should not be viewed as comprehensive, nor should they be used to suggest any distribution of these metabolites in their respective species, as some easily hydrolyzable molecules present in high quantities, such as HCH-salicortin [XI] in Populus deltoides (Rubert-Nason et al. 2018), was not detected. Salicinoids are often described as being exclusively produced in the Saliceae tribe of the Salicoidiae subfamily of Salicaceae, however, other salicaceous species, such as Idesia polycarpa, Homalium cochinchinensis, and Azara serrata have also been reported to produce salicinoid and salicinoid-like phenolic glycosides (Ishikawa et al. 2004, Huang et al. 2019, Hopfstock et al. 2024). These findings suggest that further chemical characterization of the Salicaceae, and potentially beyond, will be required to understand the evolution and distribution of these compounds. The development of biotechnological methods in Populus over the last two decades has been paramount in elucidating the function, molecular structure, and potential uses of this diverse group of naturally abundant compounds.

Figure 1.

Selection of salicinoid and salicinoid-like compounds, in order of molecular weight. Roman numerals are included for reference throughout the text.

Biosynthesis and general biology

The complete metabolic pathway responsible for salicinoid biosynthesis remains elusive (Figure 2), but prior to Populus being widely used as a model woody plant, isotopic labeling was employed as a highly informative tool, in both Populus and Salix, to explore the origin of said compounds. For example, an investigation feeding multiple putative C¹⁴ labeled salicinoid precursors in Salix purpurea found that reduced salicylic acid (SA) or free salicyl alcohol were not probable candidates as salicin precursors, while the phenylpropanoid products benzoic acid, benzyl alcohol, o-coumaric acid, and cinnamic acid were found to be highly incorporated into salicin (Zenk 1967). Furthermore, it was shown that labeled helicin [XII] accumulated after feeding S. purpurea C¹⁴ benzoic acid, which led to the suggestion that salicyloyl-CoA was reduced to salicylaldehyde and subsequently glycosylated to helicin, and therefore postulated as the salicin precursor (Zenk 1967). Contrasting these findings, Babst et al. (2010) fed ¹³C₆ labeled salicyl alcohol to Populus nigra and found it to be readily incorporated into salicin. However, these contrasting findings could potentially be due to alternate pathways of salicin biosynthesis between Populus and Salix—an area that warrants investigation. Despite this discrepancy, the findings of Babst et al. (2010) concur with the findings purported by Zenk (1967) that there is likely a phenylpropanoid origin for the salicinoids, as ¹³C₆ labeled cinnamic acid, benzoic acid, and benzaldehyde were found to be incorporated into the higher order salicinoid, salicortin. It was also found that salicylaldehyde and salicyl alcohol were not incorporated into salicortin, suggesting the absence of a shared biosynthetic pathway for all salicinoids. A more recent study applied D₇ labeled cinnamic acid to Populus trichocarpa and found that it was also incorporated into salicin. The same experiment dually confirmed that the hydroxy-6-oxo-cyclohex-2-en-1-carboxylic acid (HCC) moiety on salicortinoids and related non-salicinoid phenolic glycosides (Figure 1) also originates from cinnamic acid catabolism (Lackus et al. 2021).

Figure 2.

Schematic of what is known and hypothesized about salicinoid biosynthesis. Schematic represents a proposed salicinoid biosynthetic pathway, based on in vitro, in vivo and isotope labeling studies. Black arrows indicate in vivo findings; red arrows indicate C¹³ labelling findings; blue arrows indicate in vitro findings; multiple arrowheads indicate a multistep reaction; dotted arrows indicate hypothesized reactions. (References for the work being depicted are represented by numbers on the schematic, where: 1. Zenk 1967; 2. Babst et al. 2010; 3. Chedgy et al. 2015; 4. Lackus et al. 2020; 5. Fellenberg et al. 2020; 6. Lackus et al. 2021; 7. Gordon et al. 2022).

Although many salicinoids have been characterized as being constitutive in nature, select environmental, ontogenetic, and genotypic factors have also been shown to influence the concentration and distribution of salicinoids in Populus and Salix trees. For example, in Populus tremuloides, it has been found that although variation in salicinoid content can largely be attributed to genotype or clonal origin, younger leaf tissue has been found to contain higher concentrations of these compounds, particularly salicin (Massad et al. 2014), presumably due to the increased risk of herbivory in younger trees (Osier and Lindroth 2001, Donaldson et al. 2006, Bose et al. 2024). Drought and warming have also been found to affect salicinoid composition (Lastra et al. 2017, Tschaplinski et al. 2019, Dong et al. 2021). Most species of Populus are dioecious, and in Populus cathayana, sex specific induction of salicortin has also been observed, adding an additional dimension to the complexity of factors that may affect the production of these important compounds, in nearly every member of the genus (Dong et al. 2021). A sampling of the multitude of other environmental and ecological factors that affect the production of salicinoids in Populus will be expanded upon further later in this review.

Enzymes involved in salicinoid biosynthesis

Beyond their origination from cinnamic acid and the phenylpropanoid pathway, the majority of the mechanisms underpinning the presence of salicinoids in the Salicaceae have largely remained a mystery. That being said, the relatively recent expansion of molecular platforms in several poplar genotypes has allowed for the development of a nascent understanding of their synthesis (Figure 2). Chedgy et al. (2015) leveraged expression analysis of transgenic poplar overexpressing a transcription factor associated with condensed tannin biosynthesis, MYB134 (Mellway et al. 2009), to identify two clade VI BAHD acyltransferase enzymes that were downregulated by the overexpression of the transcription factor. Benzoyl-CoA:benzyl alcohol O-benzoyltransferase and benzoyl-CoA:salicyl alcohol O-benzoyltransferase (PtBABT and PtSABT) were identified and subsequently functionally tested in vitro. PtSABT was found to primarily use salicyl alcohol and benzoyl-CoA to catalyze the formation of salicyl benzoate, while PtBEBT was found to primarily employ benzyl alcohol and benzyl-CoA to form benzyl benzoate, although both enzymes demonstrated substrate promiscuity, which is characteristic of the BAHD acyltransferase enzyme family (Moghe et al. 2023). Proposing salicyl benzoate as a potential precursor for salicortin, Fellenberg et al. (2020) followed up on this finding by investigating two UGTs that were co-expressed with PtSABT and PtBEBT, and therefore, potential glucosylators of salicyl benzoate. PtUGT71L1 and PtUGT78M1 were subsequently shown to favor salicyl benzoate as a substrate in vitro. The two enzymes shared highly similar optimal conditions and substrate affinities, with both being able to glucosylate 2-hydroxycinnamic acid and salicylaldehyde in addition to salicyl benzoate. However, subsequently generated CRISPR/Cas9-mediated knockouts in Populus alba × tremula (INRA-717) hairy root culture paint a contrasting picture (Fellenberg et al. 2020). A significant portion of PtUGT71L1 biallelic knockouts demonstrated a complete loss of salicortin, tremulacin, and tremuloidin production, and a reduced amount of salicin production, while PtUGT78M1 knockouts conversely show no observable changes in salicinoid contents. This finding implicates glucosylated salicyl benzoate as a central precursor to salicortin and other salicinoids, and suggests that PtUGT71L1 is the main catalyst in the glucosylation of salicyl benzoate. A deeper examination of P. trichocarpa UGT71Ls found that the expression of UGT71L1 is highly correlated with PtBEBT and closely related to homologs in S. purpurea (Kulasekaran et al. 2021). Identifying two nearly identical homologs that were found to share substrate affinities and functionality, Kulasekaran et al. (2021) showed that SpUGT71L2 also favored salicyl benzoate, hinting at pathway elements that are shared between Populus and Salix. RNAi-mediated suppression of PtUGT71L1 in poplar trees showed reduce the accumulation of salicinoid-like compounds and distally related phenolics, such as chlorogenic acid. Although RNAi suppression is known to incompletely halt expression, the presence of low levels of salicinoids raised questions about the participation of other UGTs in the pathway (Kulasekaran et al. 2021). The authors also observed an unexpected lack of substrate buildup in RNAi trees, and this, in combination with a decrease in other phenolics, suggests that critically low levels of salicinoids can initiate feedback inhibition of phenylpropanoid or benzoate metabolism.

Attempting to elucidate the true effect of PtUGT71L1 loss on salicinoid biosynthesis and phenylpropanoid metabolic flux, Gordon et al. (2022) generated stable PtUGT71L1 knockouts and conducted greenhouse growth trials. Corroborating the findings of Fellenberg et al. (2020), salicortin and tremulacin were found to be greatly diminished by biallelic knockouts, but supporting the findings of Kulasekaran et al. (2021), they were detectable, albeit at low concentrations. The presence of basal levels of salicinoids in the PtUGT71L1 knockouts implies multiple routes of synthesis. These plants also displayed a reduced growth phenotype, which was associated with high levels of the defense hormones, SA and jasmonic acid (JA) (Groszmann et al. 2015). Upon further investigation and discovery that SA and its glucoside had increased (~15-fold), and JA and its biologically active conjugates had increased ~ 17- to 27-fold, it was postulated that the biallelic knockout of PtUGT71L1 forced the release of SA from salicyl salicylate that would otherwise be glucosylated and continuously metabolized into downstream salicinoids (Gordon et al. 2022). Curiously, the PtUGT71L1 knockout plants also displayed a reduction in another non-salicinoid phenolic glycoside specific to Populus, grandidentatin [XVI], which inherently implies it is metabolized from the same pool of phenolic moieties as the salicinoids. A combination of D₇-labeled cinnamic acid and RNAi-suppression of P. trichocarpa peroxisome-localized benzenoid biosynthesis enzymes has revealed that these phenolic moieties may derive from a yet-to-be-characterized cytosolic pathway of non-volatile benzenoid biosynthesis (Lackus et al. 2021). These experiments have propelled knowledge of salicinoid biosynthesis much farther in the last few of years, but the prospect of alternative biosynthetic pathways and the wide variety of salicinoids indicate that more work is needed to fully understand the biosynthesis of salicinoids in Populus.

The number of identified naturally occurring salicinoids in Populus has grown in parallel with the elucidation of the biosynthetic pathway. Recent work, employing liquid chromatography (LC) coupled with tandem mass spectrometry (LC–MS–MS) and NMR, has highlighted the presence of a sulfate side group present on salicin, producing the previously undescribed salicin-7-sulfate [II] (Noleto-Dias et al. 2018). Further research, using similar methods, found another close structurally related sulfated salicinoid-like compound, salirepin-7-sulfate [XIV] (Lackus et al. 2020). Additional efforts investigating the formation of salicinoids through sequence homology, as well as in vitro enzyme assays, disclosed the identity and location of the sulfotransferase (SOT) responsible for the addition of the sulfate groups (Lackus et al. 2020). Interestingly, sulfated salicinoids were found in several other members of the Salicaceae and Populus, with the exception of P. nigra, which was found to contain a frameshift mutation in the SOT1 locus (Lackus et al. 2020). Contrary to expectations, there was no signal that this compound affected herbivory preferences in Lymantria dispar when RNAi SOT1 INRA-717 were deployed in feeding trials (Lackus et al. 2020). To date, the leading theory for the role of these sulfated metabolites is that they contribute to sulfur storage, however, there is also speculation that they could be potentially used in defense against specialist herbivores (Lackus et al. 2020). With the increased ability to further identify and quantify modified salicinoids and their unmodified counterparts, a potential role for salicinoid compounds in primary metabolic processes may be found. Through their investigations of UGT71s, Kulasekaran et al. (2021) also identified several novel salicinoids, such as 6′-HCH-salicin and 6′-HCH-tremulodin, raising further questions about the timing and mechanism of the addition of the HCH-moiety. Another recent effort identified 2′-(Z)-cinnamoylsalicortin [X], which was isolated from Populus tremula, and has been shown to be significant in shaping the chemotypic diversity in Swedish populations of the species (Keefover-Ring et al. 2014a, 2014b).

Genomics

While agrobacterium-mediated transformation in Populus has primarily driven the work to unravel the tangled web of salicinoid biosynthesis thus far, exploration into the genomic and phenotypic structure of multiple sects of Populus in tandem has illuminated multiple aspects of the evolution, distribution, and genetic basis of these compounds. Several genomic regions associated with salicinoids, specifically salicortin and HCH-salicortin, have been identified independently, as suggestive QTLs in P. alba, Populus fremontii × P. angustifolia, P. alba × P. tremula (Populus × canescens), and P. tremula (Caseys et al. 2015, Woolbright et al. 2018). Caseys et al. (2015) used the naturally hybridized populations of European Populus × canescens to explore the genetic structure of secondary metabolic traits. Relative to flavonoids, salicinoid chemotypes were found to exhibit species-specific patterns and fall under strong phenotypic correlations, but similar to a previous investigation, salicinoids failed to produce a strong geographic signal (Keefover-Ring et al. 2014a). Two detected QTLs for salicortin and HCH-salicortin were also found to correspond with those found in North American poplar (Woolbright et al. 2018). Woolbright et al. (2018) took advantage of the naturally hybridizing HCH-salicortin producing P. fremontii and P. angustifolia, which produce little to no HCH-salicortin, to examine the genetic basis of salicinoids in Populus. A significant and still yet to be described QTL on Chr. 12 was putatively identified as an N-hydroxycinnamoyl/benzoyltransferase. Another BAHD acyltransferase identified as having QTL-associated regions was hydroxycinnamoyl transferase 6 (HCT6) (Woolbright et al. 2018); although demonstration of HCT6’s ability to affect the salicinoid biosynthetic pathway is not present in the literature. Additionally, generated hybrids were found to lack a significant association between leaf salicortin concentration and HCH-salicortin concentration (Woolbright et al. 2018), adding support to the hypothesis that the salicinoid biosynthetic pathway is complex and difficult to intuit. QTL regions were also identified for salicortin, although as both species contained significant amounts of salicortin, the results were less dramatic than what was observed for HCH-salicortin.

Distribution and chemotaxonomy

Several investigations have sought evidence of signals of local adaptation and intraspecific geographic structure in salicinoid abundances. Sampling of natural populations of P. tremuloides in the Great Lakes region and the Rocky Mountains in North America suggests a geographic signal in both salicortin and tremulacin variation when comparing major populations across the continent (Lindroth et al. 2023). There are many possibilities for this apparent discretion in very closely related species, including levels of gene flow between the sampled populations, the range of the respective studies, G×E interactions, and more. Although in all aforementioned publications, genotype was still found to be the greatest driver of variation in chemical phenotypes among populations, chemotypic plasticity has been observed in natural accessions of Populus as well. A common garden experiment consisting of representatives from 16 populations of P. fremontii, replicated across a mean annual temperature and elevation gradient, found differences in potential for chemical plasticity when comparing genotypes from provenances with differing climates (Eisenring et al. 2022). Individuals from cooler provenances and higher elevations displayed greater plasticity than individuals from hotter provenances and lower elevations (Eisenring et al. 2022). Additionally, it was found that individuals from low elevation provenances growing at low elevations contained a lower abundance of constitutively produced salicinoids than individuals from high elevation provenances growing at high elevations (Eisenring et al. 2022). This may mean that high elevation P. fremontii, or members of the greater population who experience infrequent herbivory and high levels of climatic variation, have adapted to prioritize the direction of carbon flux toward growth until higher levels of defense compounds are needed (Eisenring et al. 2022). Although this type of plasticity in chemical phenotype has not been explored in other members of Populus, this dynamic will be important to understand while undertaking future attempts to elucidate the distribution of chemotypes across populations, within populations and in different environments within a species. In contrast, although subject to tremendous gene flow, P. tremula is composed of distinct populations separated across a latitudinal gradient (Bernhardsson et al. 2013), yet the distribution of chemotypes appears to be ‘remarkably consistent’ across latitudinal clines in Sweden (Keefover-Ring et al. 2014a), which has been corroborated by other explorations into P. tremula populations and chemotypes across the European subcontinent (Caseys et al. 2015). A greater magnitude of exploration into the distribution of salicinoid chemotypes in natural populations of Populus spp. will be required to expand on the role of these compounds in local adaptation, and their potential for chemotaxonomic identification of phenotypically distinct populations of Populus.

Although the presence of a distinct chemotaxonomic pattern for the various intergenic sects of Populus has yet to be detected, the salicinoid profiles of species are distinct enough for identification via LC–MS (Caseys et al. 2015). As salicinoid content and presence/absence of specific compounds do not appear to strictly correspond to phylogenetic relationships (Figure 3; Sanderson et al. 2023), it may be possible that these traits are derived from interspecific variation in regulation of gene expression and substrate availability in certain tissues. Many Populus species and sections have overlapping ranges (Figure 3), and understanding the additive effect of hybridization on salicinoid composition as well as chemotaxonomic distribution may allow for determination of genotype/pedigree for many of these naturally occurring hybrids, as well as a greater understanding of the complex relationships within Populus.

Figure 3.

Approximate geographic ranges (according to Little 1971, 1976, Thompson et al. 1999) and a selection of known salicinoid present, conferring chemotaxonomy within monophyletic groupings of North American Populus species (Sanderson et al. 2023) belonging to the traditionally assigned sections of the genus, excluding section Abaso. An asterisk (^*) denotes very minimal concentrations. Note: higher order salicinoids are highly susceptible to degradation, and it has been shown that the more derived salicinoids can be hydrolyzed during extraction (Lindroth and Pajutee 1987). Therefore, compound presences or lack thereof reported from older experiments and literature may be due to the method of tissue preparation used. (References are represented by numbers in the figure, where: 1. Clausen et al. 1989; 2. Estes and Pearl 1967; 3. Pearl and Darling 1969; 4. Pearl and Darling 1971a; 5. Pearl and Darling 1971b; 6. Pearl and Darling 1977; 7. Pobłocka-Olech et al. 2021; 8. Rehill et al. 2005; 9. Rubert-Nason et al. 2014; 10. Snyder et al. 2015; 11. Tschaplinski et al. 2019; 12. Woolbright et al. 2018).

Ecology

Herbivory is one of the most significant biotic disturbances that salicaceous species face, usually by insects or mammalian herbivores, such as the spongy moth or the mule deer, both prolific herbivores of Populus. Inquiries into the salicinoid chemical interactions between herbivores and poplar trees often highlight the role of salicinoids as being toxic or a feeding deterrent, and the differential production of salicinoids due to continued defoliation. Ungulate herbivores have been shown to pose significant threats to P. tremuloides regeneration, leading to reduced recruitment (Holeski et al. 2016). The feeding habits of ungulates have been observed to be negatively affected when salicinoid concentrations are above ~16–17% dry weight (Wooley et al. 2008). In investigations comparing ‘high’ to ‘low’ salicinoid-containing populations, high intrapopulation variation in salicinoid content has led to an array of findings, as interpopulation differences are muted. These studies have either high similarity in salicinoid content or have concentrations over the ~16–17% threshold across their populations. In these studies, non-salicinoid factors, such as condensed tannins, protein and nitrogen content of vegetative tissues, have been found to be relevant to ungulate herbivory (Wooley et al. 2008, Villalba et al. 2014, Holeski et al. 2016, Heroy et al. 2018). Some studies, which inherently had a wider range of salicinoid content between groups, have observed that P. tremuloides clones that are higher in salicinoid content have lower rates of defoliation (Lastra et al. 2017), increased chances of successful regeneration (Rhodes et al. 2018), and are more often able to deter herbivores (Wooley et al. 2008). Interestingly, studies have noted that repeated continuous feeding by ungulates, especially in the apical meristem, leads to an ontogenetically stunted ramet that maintains high salicinoid concentration within its leaves (Rhodes et al. 2017). Furthermore, ungulates have been shown to naturally select for P. tremuloides with increased salicinoid content within branches, according to a study by Bailey et al. (2007), where a significantly lower stand-level salicortin and tremulacin content was observed in a gated ungulate exclosure, compared with measurements taken after exclosure removal, with tremulacin being the significant predictor of aspen mortality.

Adding to the complexity of these ecosystem studies, it was found that nutritional factors (Holeski et al. 2016, Heroy et al. 2018) and herbivore density (Rhodes et al. 2017) are also significant predictors of herbivory susceptibility in Populus, which may trump the deterrence of salicinoids. Future probes should look into the usage of highly predictable poplar foliage sources, such as transformants or characterized clones, to provide an effective control on nutritive factors, such as nitrogen content, protein content, and fiber content. Another consideration would be to measure a wider range of salicinoids, as most papers solely measure tremulacin and salicortin, two HCC-containing, higher-order salicinoids. Furthermore, the denotation of different sites in feeding experiments as high and low salicinoid-containing sites should be avoided, as the high interpopulation differences of salicinoids could be a confounding factor, unless all sampled trees are measured.

A historical examination of the role of herbivory and insect metabolism on salicinoids is comprehensively covered by Boeckler et al. (2011); however, since then, several investigations conducted have revealed additional layers of complexity and expounded on the proposed mechanisms. In MYB134 overexpressing P. tremula × tremuloides, which produces high levels of condensed tannins, another phenylpropanoid associated with anti-herbivory, it was found that although condensed tannin content was high, salicinoids were produced at low levels, which led to better survivorship and higher weight gain in the generalists Malacosoma disstria and L. dispar (Boeckler et al. 2014). A more recent study found that a near-complete defoliation of P. tremuloides by the invasive spongy moth (L. dispar) manifested in an average 8.4-fold increase in the salicinoids in reflushed leaves (Lindroth et al. 2024). This delayed systemic induction is corroborated by previous research, which also found that salicinoids were not immediately induced, but later observations saw an increase throughout the plant (Stevens and Lindroth 2005). The 8.4-fold increase in salicinoids in the P. tremuloides leaves led to significant increases in mortality in lab-reared native polyphemus moth larvae (Anthereae polyphemus; Lindroth et al. 2024). Further supporting this evidence, Rubert-Nason et al. (2015) found localized, leaf-level increases in salicinoids of around 22% after insect herbivory, as well as an average increase of 30% in unattacked leaves, indicating systemic induction. In contrast, in P. nigra, after defoliation by L. dispar, the salicinoid content of the partially consumed leaves decreased over the span of 4 weeks (Boeckler et al. 2013). Although contradictory, the leaves measured were old leaves, in comparison with those examined by Lindroth et al. (2024) and Stevens and Lindroth (2005), who sampled newly flushed leaves. Further contrasting data also find no systemic induction in total salicinoids after paired below ground and above ground herbivory in P. nigra, although a significant increase in the salicinoid homaloside D [XVII] was noted (Lackner et al. 2019). Another P. nigra greenhouse study employing clones of natural populations found an induction in salicin, but not salicortin and homaloside D, throughout herbivory trials by one specialist and two generalist insect herbivores. This result could implicate salicin induction as a general herbivory response, at least for P. nigra (Fabisch et al. 2019). The induced salicinoid responses of both P. tremuloides and P. nigra seem to differ, however, the ontogenetic differences between the experiments, as well as tissue types, and time lag before measuring inductive effects, make it difficult to compare and/or generate a comment on global response in poplar species. Additional research using similarly aged trees and the same tissue types, i.e., same-aged re-flushed leaves, would further help resolve such disparities.

In the case of another invasive species, the Asian longhorned beetle, a known xylophagous generalist, feeding and oogenesis were both significantly negatively affected by the exogenous addition of salicinoids, and according to frass analysis, the salicinoids were not fully metabolized. (Mason et al. 2021). Boeckler et al. (2016) also fed artificial diets supplemented with salicortin, tremulacin, and salicin to L. dispar larvae to elucidate the effect of each individual salicinoid on larval development and survival. In contrast to a previous investigation (Lindroth et al. 1988), it was shown that salicin has an inhibitory impact on herbivore growth when supplemented at a concentration accounting for the breakdown of higher-order salicinoids rather than at the average concentration that salicin is found inherently in Populus foliage (Boeckler et al. 2016). This also suggests that there could be an additive effect from consumption of the HCC moiety lacking salicinoids, such as trichocarposide and populoside [VIII], which may implicate them as anti-herbivory compounds. Boeckler et al. (2016) also found that the ester-cleaved moieties of higher-order salicinoids, such as tremulacin, have a synergistic negative effect on larval growth and survival that is not seen when the compounds are supplemented separately. This catabolism of higher-order salicinoids into salicin during insect generalist herbivore digestion is further reflected in other studies. Mason et al. (2021) investigated fecal samples of larvae feeding on P. tremula × tremuloides, and showed that the breakdown of HCC resulted in the formation of catechol, which had an inhibitory effect on growth similar to salicin (Boeckler et al. 2016). Surprisingly, it was observed that L. dispar detoxified ingested salicinoids via phosphorylation, consuming about half of the phosphorus they ingested through poplar herbivory, although glycine conjugation seemed to be used to detoxify ingested benzoic acid, resulting in the formation of the well-studied hippuric acid (Boeckler et al. 2016). Finally, the hypothesis that the toxicity of catechol is due to oxidation of the molecule to form quinones was suggested to be a strong possibility, as larvae fed pure catechol had darker feces, characteristic of a high concentration of quinones. Although outside of the scope of this review, many advances in understanding the highly specialized metabolic pathways and other specializations of monophagous and certain oligophagous insects have also been investigated, such as the cases of Cerula vinula (Feistel et al. 2018, Schnurrer and Paetz 2023, Schnurrer et al. 2024) and Chrysomela populi (Pentzold et al. 2019).

Although it remains under-explored, there is significant potential for salicinoids to play a role in fungal pathogen resistance. In P. nigra × deltoides hybrids, it was found that mature leaves displayed higher intensity of rust infection, greater uredinia size and higher spore production rates, and that leaves with higher salicinoid content displayed less intense rust infection, smaller uredinia size and lower spore production rates (Maupetit et al. 2018). Examinations of these dynamics in Salix suggest that specifically, salicortin may have an inhibitory effect on rust severity (Julkunen-Tiitto et al. 1994), and that Salix members may induce salicortin protection when infected (Hakulinen 1998). Due to their structural similarities, the catechol degradation product of salicortinoids and other 6-HCH-containing salicinoids could potentially have similar inhibitory effects through oxidant quenching, as catechin, a flavan-3-ol, appears to play a role in Populus fungal pathogen resistance (Ullah et al. 2017).

Salicinoids have also been associated with abiotic stress responses in Populus. Tschaplinski et al. (2019) conducted a greenhouse trial comparing the metabolomic profiles of P. deltoides clones in cyclic and acute drought. The leaves of clones subject to acute drought exhibited an induction of high levels of salicinoid compounds and showed high levels of osmotic adjustment to reduce the effects of water loss. These results implicate the participation of salicinoid phenolic glycosides in drought response, and acute drought trials using transgenics with constitutively higher levels of salicinoids would help to support this hypothesis. In accordance with this hypothesis, it has also been observed in P. tremuloides that ramets growing in drier sites accumulate higher levels of salicortin and tremulacin (Lastra et al. 2017), although in field trials, there are too many variables to fully attribute this finding to soil moisture.

Another area of intense research has been elucidating how and if salicinoids are translocated throughout the plant. A C¹³ labeling study revealed that while carbon is rapidly converted into salicinoids in immature leaves, corroborating previous evidence and demonstrating that these tissues have the highest concentration of salicinoids, it takes around a week for the label to appear in the bark and roots (Massad et al. 2014). The leading hypothesis, therefore, is that salicinoids are produced in source leaves and translocated to the various sink tissues. Tracking the labeled isotope late into the growing season, Massad et al. (2014) also found that salicortin increased in the bark throughout the season, supporting the overarching premise that salicortin may serve as a mammalian feeding deterrent in woody tissue post-bud set. Trading isotope labeling for molecular methods and using the same UGT71L1-KO plants as generated in Gordon et al. (2022), Hillabrand et al. (2023) found that, unlike non-structural carbohydrates (such as starches), once salicinoid phenolic glycosides are synthesized and translocated from source to sink tissues, the glycosylated components that accompany them are unable to be remobilized back into the primary metabolic network.

Salicinoids have been shown to compose up to 20–28.5% of the dry weight of Populus spp. vegetative tissue (Donaldson et al. 2006, Lindroth et al. 2023), and as such may be a contributor to the allelopathic effects seen from cinnamic acid derivatives present in abscised leaves (Song et al. 2024). In addition to shaping the extended phenotype of Populus spp. communities through mediating their interactions with other macroorganisms, salicinoids have also been implicated in shaping their microbiomes. Two recent studies have examined how tissue specialization and secondary metabolite biosynthesis contributes to creating phyllosphere community organization in P. nigra. In one, it was found that the developing shoot apex contains the highest concentration of salicinoids, while being relatively absent in woody tissue (Bose et al. 2024). Although more research is required, these tissue-specific concentrations may play a role in shaping the distinct communities in the phyllosphere found between Populus organs. A second study focused on how the endophyte Cladosporium cladosporioides alters the leaf chemistry of inoculated individuals, and it was found that salicin and the salicinoid-like compound nigracin [XV] were upregulated via inoculation with a symbiotic fungus, although defensive hormones JA and SA were unaffected. Additionally, an interaction between endophyte infection and herbivory was found, and the combined effect of these two plant stressors was determined to significantly increase salicin, nigracin, and salicortin when compared with the stressors alone. It is clear that salicinoid metabolism is affected by this extended phenotype, and vice versa, and further investigations into the interplay between microbial communities and salicinoid richness and concentration are necessary to comprehensively elucidate the role of salicinoids in Populus biology (Walther et al. 2024). Intriguingly, while it has also been found that endophyte composition in P. tremula is tightly linked to genotype, just like salicinoid chemotype, this specificity is lost after herbivory by a salicinoid-attracted Chrysomela sp. leaf beetle (Albrectsen et al. 2018). Herbivory by the yeast-harboring specialist leads to higher uniform diversity and abundance in fungal endophyte communities within leaves (Albrectsen et al. 2018).

Conclusion: biotechnological potential and future perspectives

The omics era and development of molecular platforms in Populus have exponentially expanded our current understanding of salicinoid secondary metabolites in Salicaceae. However, there remains a plethora of experiments necessary to truly understand their distribution, their biological role, and the mechanism(s) of their effects. Elucidation of the complete salicinoid biosynthetic pathway using CRISPR and agrobacterium-mediated transformation, determination of the chemotaxonomy of Populus and Salicaceae via LC–MS and NMR, the usage of transgenics to determine the exact role salicinoids play in biotic/abiotic stress resistance, and population genetic studies across multiple species and more populations are all promising avenues for future research. Outbreaks of invasive and native insect herbivores exacerbated by climate change, recruitment decline of native Populus spp., challenges in managing biofuel feedstock forests, and loss of riparian ecosystems are still ever-present challenges in today’s poplar-rich ecosystems.

Author contributions

N.J.K. and J.Y. wrote the drafts of the manuscript. S.D.M. and N.J.K. conceptualized the review. S.D.M. and B.G.F. edited the manuscript.

Conflict of interest

None declared.

Funding

This work was supported by the US Department of Energy, Office of Science, Basic Energy Sciences under award #DE-SC0023013 to B.G.F. and S.D.M.

Data availability

Raw data can be provided upon request to the corresponding author.