Allelopathic effects of volatile organic compounds released from Pinus halepensis needles and roots

Abstract The Mediterranean region is recognized as a global biodiversity hotspot. However, over the last decades, the cessation of traditional farming in the north part of the Mediterranean basin has given way to strong afforestation leading to occurrence of abandoned agricultural lands colonized by pioneer expansionist species like Pinus halepensis. This pine species is known to synthesize a wide range of secondary metabolites, and previous studies have demonstrated strong allelopathic potentialities of its needle and root leachates. Pinus halepensis is also recognized to release significant amounts of volatile organic compounds (VOC) with potential allelopathic effects that have never been investigated. In this context, the objectives of the present study were to improve our knowledge about the VOC released from P. halepensis needles and roots, determine if these VOC affect the seed germination and root growth of two herbaceous target species (Lactuca sativa and Linum strictum), and evaluate if soil microorganisms modulate the potential allelopathic effects of these VOC. Thirty terpenes were detected from both, needle and root emissions with β‐caryophyllene as the major volatile. Numerous terpenes, such as β‐caryophyllene, δ‐terpinene, or α‐pinene, showed higher headspace concentrations according to the gradient green needles < senescent needles < needle litter. Seed germination and root growth of the two target species were mainly reduced in presence of P. halepensis VOC. In strong contrast with the trend reported with needle leachates in literature, we observed an increasing inhibitory effect of P. halepensis VOC with the progress of needle physiological stages (i.e., green needle < senescent needle < needle litter). Surprisingly, several inhibitory effects observed on filter paper were also found or even amplified when natural soil was used as a substrate, highlighting that soil microorganisms do not necessarily limit the negative effects of VOC released by P. halepensis on herbaceous target species.


| INTRODUC TI ON
Plant community organization and dynamics are under the control of biotic processes, particularly plant-plant interactions such as resource competition, facilitation, and allelopathy (Callaway & Walker, 1997). A strong attention has been paid during the last decades to allelopathy, demonstrating the key implication of plantplant chemical interaction as a driver of plant community structure and ecosystem functioning (Inderjit, Wardle, Karban, & Callaway, 2011;Meiners, Kong, Ladwig, Pisula, & Lang, 2012;Wardle, Nilsson, Gallet, & Zackrisson, 1998). Seed germination and seedling performance are the main life stages usually affected by allelochemicals, and frequent negative allelopathic effects are inhibition of seed germination (Fernandez et al., 2013;Herranz, Ferrandis, Copete, Duro, & Zalacain, 2006), delay of seed germination (Fernandez et al., 2013;Hashoum et al., 2017), and inhibition of seedling growth (Gavinet et al., 2019;Santonja, Le Rouzic, & Thiebaut, 2018) by altering physiological processes (e.g., photosynthesis, nutrient uptake, cell division, or elongation; Inderjit & Duke, 2003). However, the persistence, availability, and biological impacts of the allelochemicals could be modulated by soil microbial communities (Cipollini, Rigsby, & Barto, 2012;Inderjit, 2005). Indeed, by using different substrates such as filter paper, natural or sterilized soils, several studies highlighted the key role played by soil microorganisms which suppressed the potential negative allelopathic effects (Fernandez et al., 2013;Inderjit, 2005;Kaur, Kaur, Kaur, Baldwin, & Inderjit, 2009). In addition to be highly variable among species, the diversity and quantity of allelochemicals produced by a given species and their influence on a target species are strongly dependent on its phenological stage Hashoum et al., 2017;Santonja, Le Rouzic, et al., 2018). Surprisingly, most of the published allelopathy studies were performed by using only green leaves (or needles) and thus neglected the allelopathic potentialities of chemicals contained in senescent leaves or leaf litter. Hashoum et al. (2017) reported that the germination velocity of two target herbaceous species (Festuca ovina L. and Linum perenne L.) was inhibited by aqueous extracts of senescent leaves of woody species (Acer monspessulanum L., Cotinus coggygria Scop., and Quercus pubescens Willd.) while their seedling growth was affected by aqueous extracts of green leaves.
In addition, terpene emissions from plant species are predicted to increase substantially due to a warmer climate and dense vegetation communities (Peñuelas & Staudt, 2010) indicating the need for further research on the role played by these VOC in ecosystem functioning. In this context, it is of prime interest to improve our knowledge about the allelopathic potentialities of P. halepensis VOC.
To fill this gap, we performed a laboratory experiment in order to (a) identify the VOC released from P. halepensis needles and roots (green needles, senescent needles, needle litter, and roots), (b) determine if these VOC affect seed germination and root growth of two herbaceous target species (Lactuca sativa L. and Linum strictum L.), (c) test if these VOC and their effects vary according to the organs (needles vs. roots) and the needle physiological stages, and finally (d) evaluate if soil microorganisms modulate the potential allelopathic effects of these VOC by using filter paper and natural soil as substrate.

| Material collection
The sampling site was located in the Luberon Natural Regional Park, SE France. This site is a secondary succession following abandonment agricultural lands, including different stages of P. halepensis colonization from fallow land with few young pines (<5 years old) to old pine forests (>60 years old).
The soil used as substrate for the bioassays was collected outside the zone of influence of P. halepensis (i.e., in a fallow without pine), sieved to a mesh size of 2 mm, and kept at room temperature until the start of the experiment. Green needles, senescent needles, needle litter, and roots of P. halepensis were collected in a young P. halepensis forest (about 10 years old) at the beginning of summer. While green needles, senescent needles, and roots were directly collected on the trees, needle litter of the current year was collected on the ground under the canopy of the corresponding trees. Material from 10 individuals was collected and pooled every 2 days since fresh pine material was renewed every 2 days in order to perform the laboratory experiments, for a total of 110 individuals sampled during the study.
Two herbaceous species were selected as target of P. halepensis VOC. Firstly, L. strictum as this herbaceous species is naturally present in the first secondary succession stages following abandonment of agricultural lands in the studied area and had been reported to be highly sensitive to green needle leachates (Fernandez et al., 2006(Fernandez et al., , 2013. Secondly, L. sativa since this species is known for its sensitivity to allelopathic substances and is frequently used for bioassays (e.g., Bousquet-Mélou et al., 2005;Fernandez et al., 2006). Seeds of L. strictum were collected from wild populations on the study site outside the zone of influence of P. halepensis and then stored in a cold chamber at 4°C until the start of the experiment. Seeds of L. sativa were purchased in a garden shop (Truffaut, www.truff aut.com).

| Allelopathic bioassay with Pinus halepensis VOC
This bioassay was conducted with two doses of VOC by suspending 2.5 or 10.0 g (equivalent dry mass, DM) of plant material in a 1 L microcosm ( Figure 1). Fresh material was renewed every 2 days.
Petri dishes were filled with 50.0 g DM of soil or with two layers of filter paper deposited at the bottom of the microcosm closed with nalophane to prevent VOC from escaping ( Figure 1). We compared results from natural soil and filter paper as substrate in order to assess the role of natural soil microbial communities in shaping allelopathic effects (Fernandez et al., 2013;Inderjit, 2005;Kaur et al., 2009). Each Petri dish was sown with 25 seeds of one of the two target species that were watered every 2 days with 2 ml of deionized water for filter paper substrate or 5 ml for soil substrate (Figure 1).
Four replicates were performed for each treatment (target species × Pinus VOC source × dose × substrate) for a total of 96 microcosms. Bioassays were conducted under natural photoperiod (15 hr: 9 hr day: night regime) and controlled temperature (21 ± 1°C).

| Allelopathic bioassay with β-caryophyllene
In addition to the use of P. halepensis material, we tested the effects of β-caryophyllene, the main VOC released from both, needles and roots (Appendix 1; Figure 2), on the germination and growth of L. sativa. β-caryophyllene (99% high purity standard) was obtained from Sigma-Aldrich. The experiment was performed with filter paper as substrate ( Figure 1). Each Petri dish was sown with 25 seeds that were watered every 2 days with 2 ml of deionized water. The pure compound was dissolved in ethanol (Reigosa & Pazos-Malvido, 2007) and five different dilutions (10, 50, 100, 500, and 1,000 µM) were prepared. Every 2 days, a piece of filter paper (2 × 1 cm) was impregnated with 1 ml of each solution (or only ethanol for the control treatment), kept few seconds outside the microcosm to evaporate ethanol and then suspended in the 1 L microcosm. Four replicates were performed for each concentration for a total of 24 microcosms. Bioassays were conducted under the same conditions than before (natural photoperiod and controlled temperature: 21 ± 1°C).
F I G U R E 1 Schematic drawing (a) and picture (b) of the experimental design used to test the effects of the volatile organic compounds released from Pinus halepensis needles and roots on seed germination rate and root growth of Lactuca sativa and Linum strictum. Filter paper or natural soil were used as substrate in the glass Petri dish, and the 1,000 ml glass jar was closed with nalophane to prevent volatile from escaping ………. 25 seeds of target species Petri dish with substrate 1,000 ml glass jar Nalophane Pine needle or root Germination rate, root growth, and corresponding RAE values were obtained as previously described.

| Chemical analysis
Headspace Solid Phase Micro Extraction (SPME) was performed to collect and characterize the chemical composition of volatiles released from P. halepensis needles and roots (Jassbi, Zamanizadehnajari, & Baldwin, 2010). A SPME fiber coated with Polydimethylsiloxane/Divinylbenzene (PDMS/DVB, fiber diameter 65 μm, needle size 24 ga, StableFlexTM) was exposed for 4 hr to 10.0 g DM of suspended plant material in each microcosm 1 hr after the pine material was put into the microcosm ( Figure 1).
The SPME fibers were analyzed on a Hewlett-Packard GC6890

| Statistical analyses
Statistical analyses were performed with the R software (version 3.3.1).
Significance was evaluated in all cases at p < 0.05. Normality and homoscedasticity of the residuals of the models were visually checked.
Firstly, differences of seed germination rate and root growth according to target species (L. sativa vs. L. strictum), substrate type (filter paper vs. natural soil), and their interactions in absence of VOC (i.e., in the control treatments) were assessed using two-way ANOVAs, followed by Tukey HSD tests for post hoc pairwise comparisons.
Secondly, three-way ANOVAs, followed by Tukey HSD tests for post hoc pairwise comparisons, were used to test the effects of P. halepensis VOC source (green needle, senescent needle, needle litter, and root), dose (low and high), substrate type (filter paper and natural soil), and their interactions on the RAE on seed germination and root growth of the two target species.
Thirdly, Kruskal Wallis tests, followed by post hoc multiple range tests (Fisher's Least Significance Difference), were used to test the effects of β-caryophyllene concentration on the seed germination and root growth of L. sativa.

| Detected volatiles
Solid Phase Micro Extraction fibers were used to trap the emitted VOC from Pinus needles and roots (Appendix 1; Figure 2    from both, needle and root emissions, while 20 terpenes were detected only from needles (Appendix 1). In addition, 12 terpenes were detected in emissions from senescent needles and needle litter but not from green needles. Higher headspace concentrations of terpenes were observed in microcosms containing needles compared to roots. Sesquiterpenes from needles were emitted twice more than monoterpenes (58% vs. 32%), whereas the ratio was 72% versus 22% for roots. β-caryophyllene was the major emitted volatile from both, needles and roots. The other major emitted volatiles from needles were myrcene, δ-terpinene, and α-pinene, while α-pinene, α-muurolene, and copaene were the other major emitted volatiles from roots. Finally, numerous terpenes, such as β-caryophyllene, δ-terpinene, and α-pinene, showed an increasing concentration according to the gradient green needles < senescent needles < needle litter.

| Allelopathic bioassays with P. halepensis VOC
Germination rate of L. sativa seeds was three times higher than L. strictum in the control treatments (F = 174.2, p < 0.001,

| Seed germination
Pinus VOC source and dose had significant effects on seed germination (Table 2). Needle litter exhibited higher negative effects on seed germination than the three other VOC sources for both herbaceous target species (Figure 3a

| Root growth
Allelopathic effects on root growth varied across needle physiological stages, with a clear trend to increasing negative effects according to the following order: green needle < senescent needle < needle litter (Figure 5a,d). In addition, root growth of both target species decreased with increasing dose (Figures 5b,e) and these negative effects were higher with natural soil compared to filter paper ( Figure 5c,f). However, significant interactions between Pinus VOC source, dose, and substrate type were observed ( Table 2).
As reported for L. sativa seed germination, Pinus VOC source and substrate type interacted in their effects on root growth for both target species (Table 2; Figure 4). The inhibitory effects of Pinus roots and green needles (only for L. strictum) were enhanced with natural soil compared to filter paper (Figure 4b,c). By contrast, the inhibitory effects of both senescent needles and needle litter on L. sativa root growth were reduced with natural soil compared to filter paper (Figure 4b).
Regarding the significant dose × substrate type interaction on L. sativa root growth ( inhibitory effect at higher dose was remarkably higher on natural soil compared to filter paper ( Figure 6).
Finally, the significant Pinus VOC source × dose interaction (Table 2)

| D ISCUSS I ON
As inhibited seed germination rate and seedling root growth imply a decrease in recruitment and survival of individuals, our results evidenced a strong potential control of P. halepensis VOC on the dynamics of herbaceous species populations. Two previous studies already highlighted an allelopathic effect of green needle and root leachates on L. strictum seed germination and seedling growth during laboratory bioassays (Fernandez et al., 2006(Fernandez et al., , 2013. Floristic inventories performed by Fernandez et al. (2013)  Needle physiological stage was a key factor of the observed allelopathic effects in the present study as they explained a large part of RAE variance on both seed germination and root growth of the two target species (percentages of sums of squares, Table 2).
Interestingly, in contrast with the trend reported by Nektarios et al. (2005) with the use of P. halepensis needle leachates, we observed a clear increasing inhibitory effect related to VOC according to the evolution of needle physiological stage (green needle < senescent needle < needle litter). Since the allelochemicals released in leachates belong usually to phenolics Santonja, Le Rouzic, et al., 2018), the water solubility, and rapid leaching of these compounds, could explain the decreasing allelopathic potentialities along needle physiological stages (Chomel et al., 2014;Hashoum et al., 2017).  and Chomel et al. (2014) reported that 40% and 80% of phenolics disappeared after 2 and 6 months of P. halepensis needle litter decomposition, respectively, supporting the findings of Nektarios et al. (2005).
β-caryophyllene was the main volatile released from P. halepensis needles and roots ( Figure 2). Several previous studies showed that β-caryophyllene is constitutively present in P. halepensis branches and litter and is naturally released in a large variety of growing conditions Ormeño, Fernandez, Bousquet-Mélou, et al., 2007;Ormeño, Fernandez, & Mévy, 2007). Ormeño et al. (2009) showed that β-caryophyllene occurs within needle litter of P. halepensis in the field. Likewise, Ormeño, Fernandez, Bousquet-Mélou, et al. (2007) reported β-caryophyllene emissions from P. halepensis branches growing in six Mediterranean natural forest sites. Likewise, Ormeño, Fernandez, and Mévy (2007) reported that P. halepensis seedlings also synthesize and emit β-caryophyllene. In the present study, this sesquiterpene exhibited strong inhibitory effects on both, seed germination and root growth of L. sativa, a finding in line with previous studies that suggested that this compound may act as allelochemical to influence neighboring plant growth (Kong, Hu, & Xu, 1999;Sanchez-Muñoz, Aguilar, King-Díaz, Rivero, & Lotina-Hennsen, 2012;Wang, Pen, Zeng, Ding, & Xu, 2009). Wang et al. (2009) reported that β-caryophyllene inhibited both, seed germination and seedling growth of Brassica campestris L. and Raphanus sativus L., given thus potential support for the successful invasion of Mikania F I G U R E 7 Relative allelopathic effect (RAE) on root growth of Lactuca sativa (a) and Linum strictum (b) according to the Pinus halepensis VOC source × dose interaction (Table 2). Values are mean ± SE. Different letters denote significant differences between VOC sources with a > b > c (lower case = low dose; upper case = high dose). GN, green needle; NL, needle litter; SN, senescent needle micrantha Kunth in China. In addition, Kong et al. (1999) reported that β-caryophyllene, by inhibiting seedling growth of several crop species (Solanum Lycopersicon L., Raphanus sativus L., and Vigna radiate (L.) R. Wilczek), could partly explain the strong allelopathic potentialities of the widespread weed Ageratum conyzoides L. in south China and Southeast Asia. Despite it is well known that allelopathic interactions are not due to a single compound but rather to a pool of several allelochemicals acting synergistically to inhibit or stimulate growth (Reigosa et al., 1999), the increasing release of β-caryophyllene along needle physiological stages (Figure 2) could partly explain the increasing allelopathic effects along needle physiological stages in the present study. Other terpenes known as allelochemicals such as α-pinene, 3-carene, or limonene (Abrahim, Braguini, Kelmer-Bracht, & Ishii-Iwamoto, 2000;De Martino et al., 2010;Singh, Batish, Kaur, Arora, & Kohli, 2006) showed the same trend of increasing release as β-caryophyllene (Appendix 1), giving additional support to increasing allelopathic effects according to needle physiological stages. However, we acknowledge that the volatile emission was measured only 1 hr after the pine material was put in the microcosm while this pine material remained in the microcosm for 2 days, suggesting that potentially other volatiles have been released by P. halepensis needles and roots during the experiment but not detected by the SPME fibers.
Laboratory bioassays using filter paper are frequently reported to overestimate the ability of allelochemicals to influence the germination and growth parameters of neighboring target plants (Fernandez et al., 2013). Indeed, the effects of allelochemicals are less inhibitory, disappear, or even become positive under natural soil (e.g., Fernandez et al., 2013;Hashoum et al., 2017) as microbial communities strongly influence the persistence, availability and biological activity of allelochemicals through volatile assimilation, degradation, and transformation (Blum & Shafer, 1988;Inderjit, 2005;Kaur et al., 2009). However, in the present study, numerous inhibitory effects observed with filter paper as a substrate were similar or amplified when using natural soil, highlighting that soil microorganisms are not necessarily able to limit the negative effect of VOC released from P. halepensis on herbaceous target species. This was particularly the case for the impact of VOC released from P. halepensis roots on seedling growth, suggesting that microbial degradation/transformation of these VOC could lead to degraded products with increased negative allelopathic effects. In addition, the impact of VOC released from green needles was enhanced, while those from senescent needles or needle litter were reduced with natural soil as substrate. We can speculate that a better ability of soil microorganisms to degrade the VOC released by senescent needles or needle litter leads to a reduction in their allelopathic effects as compared to those released by green needles. However, we acknowledge that we only used soil free from P. halepensis influence in the present study, that is, soil whose microbial community was not frequently in contact with the allelochemicals (phenolics and terpenes) released by P. halepensis.
The soil microbial community under the influence of P. halepensis could be completely different and, as a result, could have an altered effect on the outcome of plant-plant chemical interaction mediated by P. halepensis VOC. These hypotheses would need new laboratory experiments specifically designed to study such microbial-driven chemical transformations.

| CON CLUS ION
The present study confirms the strong allelopathic potentialities of P. halepensis as seed germination and seedling growth of the two target herbaceous species were mainly inhibited by VOC released from pine needles and roots. In addition, we demonstrated for the first time a clear increasing inhibitory effect of VOC according to the evolution of needle physiological stage (green needle < senescent needle < needle litter). Finally, our results pointed out that soil microorganisms are not necessarily able to limit the negative effect of VOC on herbaceous target species.

ACK N OWLED G M ENTS
This study was funded by the CNRS within the framework of the Zone Atelier "Arrière-pays Méditerranéen." We gratefully thank Sylvie Dupouyet, Elisabetta D'Addio and Flore Corallo for their technical assistance during both material collection and laboratory experiments.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
MS, ABM, and CF designed the experiment. MS and SG performed the experiment. MS analyzed the data and led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication. Index based on Adams (2007); Peak area, peak area of the compound in the chromatogram; Percentage, peak surface percentage in the chromatogram; Tr., Traces (peak surface percentage < 0.1%).

DATA ACCE SS I B I LIT Y
* Identification confirmed by commercial standard coinjection. APPENDIX 1 (Continued)