It has long been considered that in dioecious plant species, the growth rates and chemical defenses may differ between female and male individuals (Ågren et al., 1999; Cornelissen and Stiling, 2005; Avila-Sakar and Romanow, 2012). The gender-specific differences in plant carbon allocation likely result from a higher resource investment to reproduction in females, which may lead to a lower growth rate and higher concentrations of plant secondary metabolites (PSM) in females compared with males. Gender dimorphism is also suggested to influence plant responses to the loss of biomass in response to herbivory (Cornelissen and Stiling, 2005). In slow-growing woody plants, the loss of biomass often induces an increase in PSMs, which constitutes an important part of the plant resistance against invertebrate herbivory (Haukioja and Koricheva, 2000) and mammalian browsing (Bryant et al., 1991, 2014). Whether this is a defense reaction or an indirect result of environmental constraints that regulate the trade-off between plant growth and synthesis of secondary phenolic compounds has remained unclear (Tuomi et al., 1990; Jones and Hartley, 1999; Mattson et al., 2005). The capacity of plants to compensate for the lost biomass through increased growth rates (i.e., plant compensatory growth) constitutes another important means of plant tolerance to herbivory (Lehtilä, 2000; Tiffin, 2000; Peinetti et al., 2001; Cromsigt and Kuijper, 2011). Owing to gender dimorphism, males may more commonly respond to the loss of plant biomass by compensatory growth, whereas females respond by increasing plant secondary compounds (Cornelissen and Stiling, 2005).

The secondary compounds in plants exhibit a diverse spectrum of biological functions, and several factors regulate their synthesis (Koricheva et al., 1998; Stamp, 2003; Theis and Lerdau, 2003). Plant secondary metabolites can roughly be gathered in three classes of chemical compounds, namely alkaloids, phenolic compounds and terpenes, with 1000s of compounds in each class (Chomel et al., 2016). PSMs have important roles in plant development, plant–plant and plant–microbe/insect/herbivore interactions; they also govern the mechanisms of allelopathy, influencing intra- and interspecific competition between plants. Despite well-established importance of gender dimorphism, the majority of studies investigating the effects of environmental stresses on PSMs has focused on plant phenolics (e.g., Cornelissen and Stiling, 2005) while less is known about gender dimorphism of terpenoids. For understanding the ecological consequences of gender-specific differences on plant growth and survival, more studies are needed on the role of sexual dimorphism that would consider the multiple functions of the different classes of PSMs.

Common juniper (Juniperus communis) of the Cupressaceae family is an evergreen dioecious gymnosperm shrub that has one of the widest global distributions of any gymnosperm (Thomas et al., 2007). In Europe, the distribution of juniper ranges from the Mediterranean to the Arctic and within each area, is found in a very wide range of habitats, such as old pastures, forests, and peatlands. Juniper foliage is rich in secondary substances, particularly terpenoids and phenolics (Martz et al., 2009). Previous studies have demonstrated strong gender dimorphism in juniper. For example, juniper populations may be biased toward males, because females have higher mortality rates in resource-limited conditions (Ortiz et al., 2002). The males of Juniperus thurifera start flowering younger (Gauquelin et al., 2002). Different juniper genders also exhibit differing physiological responses to shading from neighboring plants (Verdú et al., 2004). Some studies on junipers, however, have not found gender-specific differences in growth or survival (Marion and Houle, 1996; Ortiz et al., 2002), or these differences have been highly site-specific (DeSoto et al., 2016). Northern boreal juniper populations provide an excellent opportunity to investigate gender dimorphism in plants, because these populations grow in stressful environments and have high concentrations of both phenolics and terpenes in their biomass (Martz et al., 2009).

Here, we analyzed gender dimorphism in juniper by analysing the secondary compounds and responses to shoot cutting in northern boreal forests in Finland. In northern boreal forest, junipers commonly form polycormic tall shrubs instead of trees (Figure 1). We assumed that if there are differences in the reproductive effort between the genders, male and female individuals should differ in growth rates and the concentrations of PSM as well as in their responses to a loss of shoot biomass. We predicted that (1) concentrations of secondary metabolites should be higher whereas growth rates lower in females than males. As woody plants respond to biomass loss both by compensatory growth and production of secondary metabolites, we further predicted that (2) shoot cutting should induce a stronger increase in the concentrations of PSMs in females than males, whereas the compensatory growth after biomass loss should be higher in males than females.

Studies on dioecious plants have documented gender-related differences in growth rates and concentrations of secondary metabolites with higher growth but lower PSM concentrations in males compared with females (e.g., Ågren et al., 1999; Cornelissen and Stiling, 2005). Consequently, in dioecious plant species, the intensity of herbivory may be biased toward males (Cornelissen and Stiling, 2005). Studies have found exceptions to these generalizations (Avila-Sakar and Romanow, 2012), and among the species belonging to the dioecious Juniperus genus, previous evidence on gender dimorphism has also been mixed. Contrasting the generalization, Massei et al. (2006) found higher concentrations of terpenes and phenolics in male than female individuals in Mediterranean J. oxycedrus macrocarpa, and McGowan et al. (2004) that the females of prostrate juniper J. communis ssp. nana in northern Scotland were more frequently subjected to herbivory than males. No gender-specific difference in juniper growth was observed along an altitudinal gradient in Sierra Nevada despite decreasing reproductive success along this gradient (Ortiz et al., 2002). Here, except for a few individual compounds, we found no effects of gender in phenolic and terpenoid concentrations in J. communis in northern boreal forests. Although, as predicted, shoot cutting significantly increased phenolic concentrations, the level of PSMs increased, similarly, in both males and females. Further, theories on gender dimorphism state that males should be more growth-orientated than females (e.g., Ågren et al., 1999; Cornelissen and Stiling, 2005), but we found marginally (P < 0.10) higher growth after shoot cutting in females compared with males.

The lack of clear gender dimorphism in PSMs and growth in northern juniper populations compared with e.g., boreal deciduous trees (Nissinen et al., 2018; Ruuhola et al., 2018; Zhang et al., 2018) is noteworthy, because junipers in these systems seem to have a particularly high level of chemical defense. Along a large geographical gradient from the southern to northern boreal forests, Martz et al. (2009) found that secondary compound concentrations in the common juniper significantly increased with latitude. Interestingly, these trends were similar in both terpenoids and phenolics, although these compound groups exert largely differing ecological functions. Terpenoids may have a key role in the defense against mammalian herbivory (Mutikainen et al., 2000), and the low palatability of juniper to herbivores is commonly derived from oils found in the needles, cones and wood, dominated by monoterpenes (Thomas et al., 2007). Despite the low palatability, junipers are commonly subjected to herbivory in different habitats (Livingston, 1972; Miller et al., 1982; Fuhlendorf et al., 1997), and in the boreal forests, herbivory on juniper is dominated by shoot consumption by moose (Alces alces) during the winter. Contrasting with terpenoids, phenolics may have a higher importance in the plant protection against invertebrate herbivory (Koricheva et al., 1998), and photooxidative stress (Close and McArthur, 2002). The need for antioxidative compounds in relation to light intensity may increase in conditions of both nutrient deficiency and low temperature (Close and McArthur, 2002), which could explain why their concentrations increase with latitude (Stark et al., 2008; Martz et al., 2009, 2010). Traditional theories also suggested that greater concentrations of phenolics may result from allocation of extra photosynthesized carbon to secondary metabolites as a result of nutrient deficiency (Bryant et al., 1983; Mattson et al., 2005).

Although both phenolic and terpenoid concentrations in juniper seem to show a similar gradient with increasing latitude (Martz et al., 2009), here, we found that the responses of PSM concentrations to shoot cutting varied greatly among the different compounds and compound groups. More specifically, cutting lead to increased concentrations of soluble phenolics and a higher abundance of flavonoids (specific flavones and flavonol glycosides). As we analyzed samples 1 year after the shoot cutting, our analyses depict delayed inducible reactions that take place after a time lag from the loss of biomass (Tuomi et al., 1990). Although previous studies have indicated that delayed inducible reactions in conifers is relatively uncommon when compared with the deciduous trees (Nykänen and Koricheva, 2004), increasing phenolic concentrations in response to shoot cutting agree with previous findings on e.g., Pinus species (Honkanen et al., 1999; Roitto et al., 2008). Flavonols in juniper are mainly quercetin derivatives that exert strong antioxidant activity due to their chemical features (Rice-Evans et al., 1997); thus, shoot cutting led to higher abundances of compounds with a high antioxidant capacities that are more efficient in relation to the carbon cost to their synthesis (sensu Close and McArthur, 2002).

Contrasting with phenolics, we found no overall shoot cutting effect on terpenoids. In line with our investigation, the previous study by Honkanen et al. (1999) also found increasing phenolic concentrations in response to defoliation in boreal coniferous tree Pinus sylvestris L. but no effects on terpenoids, despite they are considered the main class of anti-herbivore defensive compounds. Contrasting with this idea, a meta-analysis by Nykänen and Koricheva (2004) concluded that the damage in woody plants quite commonly reduces the concentrations of terpenes. This may be true also in the case of juniper, as reduced yields of essential oils in junipers after a severe browsing damage have previously been found (Markó et al., 2008). Condensed tannins and flavonoids, among many other phenolic compounds are derived from phenylalanine via the phenylpropanoid pathway (Vogt, 2011) and terpenoids are synthesized via acetyl-CoA, pyruvate and glyceraldehyde-3-phosphate in parallel cytosolic of plastid metabolic pathways (Singh and Sharma, 2015). This shows that disturbance of the general carbon metabolism due to, for example biotic/abiotic stress or compensatory growth will have consequences on the PSM content.

Although the mechanisms underlying our findings remain uncertain, our study adds to previous evidence showing unclear or ambiguous effects of gender on growth, secondary compound concentrations and reproduction in Juniperus species (e.g., Marion and Houle, 1996; Ortiz et al., 2002; Verdú et al., 2004; Massei et al., 2006; DeSoto et al., 2016). For example, aged juniper populations are male-biased, but this bias does not seem to be easily explained by the gender-specific differences in the cost of reproduction (Gauquelin et al., 2002; Ortiz et al., 2002). Earlier reviews have already concluded that there is a need to revise theories predicting how gender dimorphism affects plant performance and responses to environmental stresses (Avila-Sakar and Romanow, 2012; Vega-Frutis et al., 2013). For woody plants, the tolerance of herbivory is a major component of plant resistance, because the probability of herbivory is high due to large size and long life span. However, the recovery potential of these species could be driven by the type of herbivory they commonly experience (Haukioja and Koricheva, 2000). Noteworthy, García et al. (2001) found that frugivory at Juniperus communis in the Mediterranean mountains depended largely on population characteristics rather than on individual attributes. If herbivory, such as browsing by moose, is commonly centered on locations with numerous and dense juniper populations to provide large food quantity, the chemical quality of the different individuals might not exert a primary role in the food selection of herbivores. Under these conditions, the capacity for regrowth could outweigh the importance of PSMs in herbivory tolerance. Further, the lack of clear gender dimorphism in northern boreal juniper populations could also result from the high need of PSMs at high latitude to protect from abiotic stresses (Martz et al., 2009). Although herbivory might even constitute one of the driving forces behind the evolution of dioecy in plants (Bawa, 1980; Cornelissen and Stiling, 2005; Avila-Sakar and Romanow, 2012; Vega-Frutis et al., 2013), studies on juniper populations have demonstrated that gender-related differences in growth and resource storage may be a consequence of local adaptation to environmental conditions (DeSoto et al., 2016). Protection against abiotic stresses through the PSMs could have such major significance for plant success under northern conditions that it might override any gender-specific differences in the carbon allocation for synthesizing PSMs.

Ecological factors that limit plant success in each specific conditions could also explain why we detected a marginally higher capacity for compensatory growth in females despite a presumably higher resource cost for reproduction (sensu Cornelissen and Stiling, 2005). Over a large geographical gradient from the Mediterranean to the sub-Arctic vegetation zone across Europe, García et al. (2000) concluded that the juniper population viability in the north may in fact be under less pressure compared with juniper populations at mid-latitudes, because these populations are free from seed predation. Experimental evidence on dioecious plants has suggested that increasing resources may weaken gender-specific differences within plant species, and consequently, the gender-specific differences in phenolic concentrations could be pronounced under high resource limitation (Palumbo et al., 2007). Further, woody plants seem to recover from the loss of biomass better under low than high resource availability, possibly because under these conditions, plants generally grow below their potential maximum growth rate (Hawkes and Sullivan, 2001). As northern conditions with low nutrient availability, low temperatures and high light likely require high defense through PSMs (Martz et al., 2009), the combination of high need for chemical defense and low pressure on reproduction (García et al., 2000) might direct gender dimorphism from gender-specific differences in defense toward the importance of compensation after the loss of biomass.

SS designed the experiments and performed field sampling and measurements. FM ran the laboratory analyses and conducted statistical tests. SS and FM jointly wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling Editor declared a shared affiliation, though no other collaboration, with one of the authors FM at the time of review.