Temporal physiological response of pine to Fusarium circinatum infection is dependent on host susceptibility level: the role of ABA catabolism

Abstract

Pine pitch canker (PPC), caused by Fusarium circinatum Nirenberg and O’Donnell, represents an important threat to conifer forests worldwide, being associated with significant economic losses. Although essential to develop disease mitigation strategies, little research focused on host susceptibility/resistance mechanisms has been conducted. We aimed to explore the response of a highly susceptible (Pinus radiata D. Don) and a relatively resistant (Pinus pinea L.) species to F. circinatum infection at different stages of infection. Morpho-physiological, hormonal and oxidative stress-related changes were assessed for each pine species and sampling point. Most of the changes found occurred in symptomatic P. radiata, for which an increased susceptibility to photoinhibition was detected together with decreased superoxide dismutase activity. Abscisic acid catabolism was activated by F. circinatum inoculation in both pine species, leading to the accumulation of the inactive dihydrophaseic acid in P. radiata and of the less-active phaseic acid in P. pinea. Hormone confocal analysis revealed that this strategy may be of particular importance at 6 d.p.i. in P. pinea, which together with photosynthesis maintenance to fuel defense mechanism, could in part explain the species resistance to PPC. These results are of great interest for the development of hormone-based breeding strategies or for the use of hormone application as inducers of resistance to F. circinatum infection.

Introduction

Conifer forests’ sustainability and productivity have been threatened by the emergence of several pathogens. Fusarium circinatum is one of the most important pathogens in at least 60 species of Pinus and Pseudotsuga menziesii, causing pine pitch canker (PPC) globally (Martín-García et al. 2019). The disease affects all stages of tree development, leading to needle wilting and chlorosis, canopy dieback and mortality in mature trees, and to seed contamination and seedling damping-off in nurseries (Gordon et al. 2015), which results in considerable environmental and economic costs.

Although some fungicides showed promising results to control PPC in vitro (e.g.,Berbegal et al. 2015, Mullett et al. 2017), its use in European forests is highly restricted in order to reduce its risks and impacts on human health and the environment (Directive 2009/128/EC), making the search for environmentally friendly control methods imperative. These methods comprise biological control, e.g., Trichoderma treatment (Amaral et al. 2019b), or the application of compounds such as phosphite (Cerqueira et al. 2017). Furthermore, the exploitation of genetic resistance has been arising as one of the most promising environmentally friendly and cost-effective strategies to control PPC, which would be especially easy at the species level, with resistant species replacing the ones currently used in large-scale forestry (Martín-García et al. 2019).

Several studies have focused on testing species susceptibility to F. circinatum infection (e.g., Iturritxa et al. 2013, Martínez-Alvarez et al. 2014). However, only recently Amaral et al. (2019a) aimed to unveil the mechanisms explaining the differential levels of susceptibility of Pinus species to F. circinatum by exploring physiological and hormonal changes and primary metabolite profiles of symptomatic plants. The highly susceptible Pinus radiata and Pinus pinaster, showing an intermediate response to F. circinatum inoculation, presented similar response profiles, contrasting with the relatively resistant Pinus pinea. Studies on the mechanisms behind the outcome of pathogen infection in different hosts are crucial to understand the disease and further select resistant genotypes.

The activation of early defense responses after pathogen recognition is known to be crucial to determine plants disease resistance, and mainly includes changes in protein phosphorylation, ion fluxes, reactive oxygen species (ROS) and other signaling transduction molecules such as hormones (Shen et al. 2017). Studies addressing early changes in Pinus elliottii var. elliottii (Davis et al. 2002) and P. radiata (Donoso et al. 2015, Carrasco et al. 2017) after F. circinatum inoculation focus on exploring changes in gene expression. Recently, transcriptomic studies considering early stages of infection with F. circinatum were also performed in P. pinaster (Hernandez-Escribano et al. 2020) and in the highly susceptible Pinus patula vs the resistant Pinus tecunumanii (Visser et al. 2019). These highlight important changes in genes associated with hormone regulation during disease progression. However, physiological studies (including hormone analysis) at early stages of PPC progression are not available.

Excess excitation energy may be generated in any stressful condition limiting photosynthesis (Bechtold et al. 2005), as is the case for F. circinatum inoculation in symptomatic P. radiata (Cerqueira et al. 2017, Amaral et al. 2019a, 2019b). Its dissipation may occur through thermal dissipation and/or by non-assimilatory photochemistry, mainly photorespiration and the water–water cycle (reviewed by Vitale (2015)). These are efficient photoprotection mechanisms against photoinhibition, i.e., decreased photosynthetic efficiency and/or maximum rate due to photo-oxidative damages resulting from ROS and other oxidative by-products derived from photosynthesis (reviewed by Niyogi (1999)). Other important antioxidant mechanisms are fundamental to avoid ROS-induced damage, namely the ascorbate-glutathione (GSH) cycle (Foyer and Noctor 2011). Its involvement in P. pinea relative resistance against PPC has been previously hypothesized (Amaral et al. 2019a). Moreover, Vivas et al. (2014) reported changes in P. pinaster antioxidant activity 4 weeks after F. circinatum infection. The study of susceptibility to photoinhibition and of antioxidant enzymes is therefore relevant to understand Pinus–F. circinatum interaction.

The study of symptomatic plants is of great interest to understand the Pinus–Fusarium interaction. However, it is also crucial to explore the changes occurring before visual symptoms take place in order to better understand the disease, support the development of innovative approaches for early disease detection and to search for early defense mechanisms responsible for plant survival under F. circinatum attack in resistant species such as P. pinea. Therefore, the aim of this study was to explore changes occurring in plant processes usually targeted under stressful conditions (photosynthesis, water relations, hormonal dynamics and antioxidant system) in a susceptible (P. radiata) and a relatively resistant (P. pinea) species at different stages of PPC progression. Besides the temporal analysis of the response of pine species with contrasting levels of susceptibility to PPC, this study provided further insights into the hormonal dynamics and antioxidative machinery in symptomatic P. radiata seedlings that complement previous works (Cerqueira et al. 2017, Amaral et al. 2019a, 2019b).

Materials and methods

Plant material and fungal culture

Five-month-old P. radiata D. Don and P. pinea L. plants were obtained from Melo & Cancela Lda. (Anadia, Portugal). Plants were placed in 400 ml pots with a 3:2 (w/w) peat:perlite mixture and kept in a climate chamber (Fitoclima D1200, Aralab, Sintra, Portugal) under the following day/night conditions: 16/8 h photoperiod, 25/20 °C temperature, 60/65 % relative humidity and 500 μmol m² s⁻¹ photon flux density. Plants were acclimatized for seven weeks, and were regularly watered and fertilized.

The F. circinatum Nirenberg and O’Donnell FcCa6 isolate described by Martínez-Alvarez et al. (2012) was obtained from the Forest Entomology and Pathology Lab at the University of Valladolid (Spain). It was grown on potato dextrose agar (PDA; VWR Chemicals, Leuven, Belgium) at 25 °C until the mycelium covered 90 % of the Petri dish. Three pieces of mycelium (5 mm diameter) were grown under agitation on potato dextrose broth (PDB; CONDA, Madrid, Spain) for 24 h. Spores were counted using a haemocytometer.

Plant inoculation, evaluation of visual symptoms and sampling

After acclimatization, seedlings of each pine species were separated in two groups of 40 plants each: mock-inoculated controls (C) and inoculated with F. circinatum (F). Stem surfaces were wounded with a sterile scalpel prior to the application of 10⁴ FcCa6 spores and wounds were sealed using Parafilm®. Controls were similarly wounded and received a corresponding 10 μl of PDB. Plants were kept under the previously described environmental conditions.

Tip dieback, needle wilting and chlorosis, and resin formation were evaluated daily. Several sampling points were considered after inoculation with F. circinatum: 2 h.p.i. (hours post-inoculation; immediate response), 1 d.p.i. (days post-inoculation; early response), 6 d.p.i. (before the development of the first disease symptoms in inoculated P. radiata seedlings) and 10 d.p.i. (all inoculated P. radiata seedlings displayed tip dieback—see Figure S1 available as Supplementary Data at Tree Physiology Online). At each sampling time, seedlings from both species and groups were harvested for further analysis. The plant material collected was frozen in liquid nitrogen, stored at −80 °C for hormone quantification and antioxidant enzyme activity analysis. Immediate physiological measurements were also performed at each sampling point: needle gas exchange-related parameters, chlorophyll a fluorescence parameters, relative water content (RWC), midday shoot water potential and relative internal necrosis. For each sampling point, relative internal stem lesion of six biological replicates per treatment was determined in longitudinal stem cuts as the relation of internal lesion length to the total stem length. To confirm Koch’s postulate, stem cuts were plated onto PDA and incubated at 25 °C.

Plant water relations

Midday shoot water potential (ᴪ_md, MPa) was measured using a Scholander-type pressure chamber (PMS Instrument Co., Albany, OR, USA). The RWC from five needles per seedling was determined. Fresh weight (FW) was recorded and needles were immersed in distilled water overnight in the dark at 4 °C. The excess of water was removed from needle surface and turgid weight (TW) was recorded. Needles were oven-dried at 60 °C until reaching a constant dry weight (DW). RWC was calculated as |$\mathrm{RWC}\ \big(\%\big)=\frac{\big(\mathrm{FW}-\mathrm{DW}\big)}{\big(\mathrm{TW}-\mathrm{DW}\big)}\times 100$|⁠. Five biological replicates were measured for each group and species at each sampling point.

Needle gas exchange-related parameters

Net CO₂ assimilation rate (A, μmol CO₂ m⁻² s⁻¹), stomatal conductance (g_s, mol H₂O m⁻² s⁻¹), transpiration rate (E, mmol H₂O m⁻² s⁻¹) and sub-stomatal CO₂ concentration (C_i, v.p.m.) were measured using a gas exchange system (LCpro-SD, ADC BioScientific Ltd, Hoddesdon, UK) coupled to a conifer-type chamber. Controlled conditions were maintained inside the chamber: ambient CO₂ concentration (405.2 ± 0.6 v.p.m.), air flux (200 μmol s⁻¹), block temperature (25 °C). Light response curves of CO₂ assimilation/photosynthetic photon flux density (A/PPFD) were performed with the following PPFD: 2000, 1500, 1000, 750, 500, 250, 100, 50 and 0 μmol m⁻² s⁻¹. After A/PPFD analysis, punctual measurements were performed at saturation light intensity (1500 μmol m⁻² s⁻¹). Data were recorded when parameters remained stable. Five biological replicates were measured for each group and species at each sampling point.

Chlorophyll a fluorescence-related parameters

Chlorophyll a fluorescence-related parameters were measured at room temperature using an imaging chlorophyll fluorometer (Open FluorCam 800-O/101, Photon System Instruments Ltd, Brno, Czech Republic). Minimum fluorescence (F_o), maximum fluorescence (F_m) and maximum quantum yield of photosystem II (PSII) (F_v/F_m = (F_m-F_o)/F_m) were measured by applying a modulated measuring light (<0.1 μmol photons m⁻² s⁻¹) and saturation pulses (>47,500 μmol photons m⁻² s⁻¹), provided by red LED panels (emission peak at 612 nm, 40 nm bandwidth). Samples were dark-adapted for 30 min before the first measurement. Afterwards, samples were exposed to 1200 μmol photons m⁻² s⁻¹ for 90 min, after which measurements were carried out every 2 min during 16 min to allow the relaxation of reversible non-photochemical quenching. The susceptibility to photoinhibition was quantified by calculating the percent variation of F_v/F_m (%F_v/F_m) from the ratio of pre-illumination and post-illumination values (Serôdio et al. 2017). Triplicates for each of the three biological replicates per treatment and species were performed at each sampling point. The images obtained (512 × 512 pixels) were processed using FluorCam7 v.1.2.5.3 (Photon System Instruments Ltd) by manually defining the areas of interest (AOI), and parameters were calculated by averaging all pixel values in each AOI. Light stress-recovery experiments allowed to distinguish whether reversible photochemical down-regulation of photosynthesis or photoinhibition occurred at the different sampling points for each species.

Hormone quantification and immunolocalization

Abscisic acid (ABA), phaseic acid (PA), dihydrophaseic acid (DPA), salicylic acid (SA), jasmonic acid (JA), JA-isoleucine (JA-Ile) and indole acetic acid (IAA) were quantified based on Durgbanshi et al. (2005). Freeze-dried needle tissue (50 mg) was extracted in 2 ml of ultrapure water after spiking with [²H₆]-ABA, [²H₃]-PA, dehydrojasmonic acid and [¹³C]-SA in a ball mill (MillMix20, Domel, Zelezniki, Slovenija). Supernatants were recovered after centrifugation (4000g, 10 min, 4 °C) and pH was adjusted to 3 using acetic acid. The extract was partitioned twice against diethyl ether. The upper layer was recovered and evaporated in a centrifugal concentrator (Speed Vac, Jouan, Saint HerblainCedex, France). The dry residue was resuspended in 10% MeOH by gentle sonication, passed through 0.22 μm regenerated cellulose membrane syringe filters (Albet SA, Barcelona, Spain), and directly injected into an UPLC system (Acquity SDS, Waters Corp., Milford, MA, USA). Analytes were separated using a reversed-phase C18 column (gravity, 1.8 μm, 50 × 2.1 mm, Macherey-Nagel, Düren, Germany) using a 300 μl min⁻¹ linear gradient of ultrapure H₂O (A) and MeOH (B) (both supplemented with 0.01% acetic acid). The gradient used was: (0–2 min) 90:10 (A:B), (2–6 min) 10:90 (A:B) and (6–7 min) 90:10 (A:B). Quantification was performed in a Quattro LC triple quadrupole mass spectrometer (Micromass, Manchester, UK) connected online to the output of the column through an orthogonal Z-spray electrospray ion source. The quantitation of hormones was achieved based on a standard curve. Five biological replicates per group and species were analyzed for each sampling time.

For the immunolocalization of ABA, JA, IAA and 1-aminocyclopropane-1-carboxylic acid (ACC), pools of five biological replicates were considered per group and species for each sampling time. Two needles per plant were collected, cut into 1 cm transversal segments and fixed for 24 h at 4 °C in 3% paraformaldehyde [w/v; in phosphate-buffered saline(PBS)] containing 0.1% (v/v) Triton X-100 (Fisher Scientific, Geel, Belgium) and 4% (w/v) 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (ThermoFisher Scientific, Rockford, IL, USA) to immobilize IAA, ABA, JA and ACC by covalent binding to proteins. Needle segments were then washed three times for 10 min in PBS and stored in 0.1% paraformaldehyde (w/v; in PBS) at 4 °C. Afterwards, samples were included in a cryostat medium (Tissue-Tek, Killik; Sakura Finetek USA, Inc., Torrance, CA, USA), 50 μm sections were cut using a sliding cryotome CM1510S (2002 Leica Microsystems, Wetzlar, Germany) and section slides were kept at −20 °C. Sections were immunolocalized according to Escandón et al. (2016) with the following modifications: primary antibodies (for ABA—ref. AS09 446, for IAA—ref. AS09 421, for ACC—ref. AS11 18,000, for JA—ref. AS11 1799, AGRISERA, Vännäs, Sweden) were used at 0.05 μg μl⁻¹ in 1% BSA (w/v; bovine serum albumine) and samples were incubated overnight. For the immunochemical detection of all hormones negative controls were used replacing the primary antibody by PBS. Fluorescence was visualized using a confocal microscope (Leica TCS-SP2-AOBS) connected to a workstation and the images were processed with Fiji Software (Schindelin et al. 2012).

Antioxidant enzyme activity

Specific activities of glutathione reductase (GR, EC 1.8.1.7), monodehydroascorbate reductase (MDHAR, EC 1.6.5.4), dehydroascorbate reductase (DHAR, EC 1.8.5.1), catalase (CAT, EC 1.11.1.6) and superoxide dismutase (SOD, EC 1.15.1.1) were determined using a microplate reader based on Murshed et al. (2008). All enzymes were extracted from 50 mg or 30 mg of needle material (for P. radiata and P. pinea, respectively; FW) by homogenizing in a ball mill (Mixer Mill MM400, Retsch, Haan, Germany) with 1 ml of extraction buffer, which consisted of 0.2 mM Tris-HCl pH 8.0, 10% (v/v) glycerol, 0.25% (v/v) Triton X-100, 5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride and polyvinylpolypyrrolidone [3% (w/v) for P. radiata and 6% (w/v) for P. pinea]. After centrifugation (17,000g for 30 min at 4 °C), the supernatants were added to the reaction buffer and substrate specific for each assay (see Table S1 available as Supplementary Data at Tree Physiology Online) up to a total volume of 200 μl per well. Measurements were performed at the corresponding wavelength (Table S1 available as Supplementary Data at Tree Physiology Online) at 25 °C over 5 min, except for SOD for which measurements were done before and after exposure to a 15 W fluorescent lamp for 15 min. Plates were always shaken before measurements. Measurements were done in triplicate for the four biological replicates per group and species for each sampling point. Enzymes specific activities (in μmol min⁻¹ mg protein⁻¹) were calculated considering the molar extinction coefficient of each substrate or product (Table S1 available as Supplementary Data at Tree Physiology Online) and the total protein on the extract, which was determined by using the Quick Start Bradford 1 × Dye Reagent (Bio-Rad, Hercules, CA, USA) following the manufacturer's instructions.

Statistical analysis

Two-way analyses of variance (ANOVAs) were performed individually for each pine species on all quantitative variables in order to evaluate the effects of F. circinatum inoculation and time post-inoculation independently, as well as their interaction. For the variables showing significant interaction between inoculation and time, Tukey’s HSD tests were applied to explore the differences between non-inoculated control plants (C) and plants inoculated with F. circinatum (F) for each sampling time. When data did not follow ANOVA’s assumptions (Shapiro–Wilk normality test and Bartlett test of homogeneity of variances, P ≤ 0.05), robust statistical methods were applied (Wilcox 2017). In particular, heteroscedastic two-way ANOVAs were carried out using the generalized Welch procedure and a 0.1 trimmed mean transformation. Similarly, when significant interaction between the two factors studied was found, differences between the groups of interest were performed testing linear contrasts involving trimmed means. In order to reduce the dimensionality and complexity of the quantitative variables, principal components analyses (PCA) were conducted individually for each pine species. Statistical procedures were conducted using R v3.5.0 core functions (R Development Core Team 2018) and the Wilcox’s Robust Statistics (WRS2) package (Mair and Wilcox 2020). Figures were built using the ggplot2 package (Wickham 2016). Data are presented as mean ± standard error (SE).

Results

Progression of PPC disease

No visual symptoms of disease or internal necrosis were registered for non-inoculated control plants. Likewise, P. pinea plants did not develop disease symptoms throughout the entire experiment and changes on stem relative internal necrosis were not statistically significant (Figure 1, Table 1). On the other hand, the first disease symptoms in P. radiata appeared 7 d.p.i., when 18.75% of the P. radiata seedlings inoculated with F. circinatum displayed tip dieback (Figure 1). The percentage of P. radiata plants exhibiting symptoms continued to increase progressively (37.5% at 8 d.p.i. and 93.75% at 9 d.p.i.) until all plants inoculated with the pathogen were symptomatic (10 d.p.i.) (Figure 1). At this stage of the disease, relative internal stem necrosis increased significantly in P. radiata seedlings inoculated with F. circinatum in comparison with its respective non-inoculated control (Figure 1). Besides the significant effect of both inoculation and time and its interaction on relative internal stem necrosis in P. radiata (Table 1), no significant changes were found between non-inoculated and inoculated plants for the remaining sampling points (Figure 1).

Figure 1.

Time course of symptoms development and relative internal stem necrosis length in P. radiata and P. pinea seedlings inoculated with F. circinatum. Lines represent the percentage of plants showing symptoms. Bars show relative internal stem necrosis length of inoculated plants at each sampling point (red points). Necrosis data are presented as mean ± SE. For significant interaction, asterisks indicate significantly different means in comparison with mock-inoculated controls for each species and sampling point (post hoc test, α = 0.05). No symptoms or necrosis were observed in the controls.

Table 1

Two-way ANOVA summary table for the quantitative variables at the different sampling points (time factor) in P. radiata and P. pinea control plants and in plants inoculated with F. circinatum (inoculation factor). The significance of the interaction of both factors was also tested.

		P. radiata				P. pinea
Variable	Source of variation	Df	F	Significance		Df	F	Significance
RIN	Inoculation	1	303.2	<0.001	***	1	2.343	0.179	NS
	Time	3	287.2	<0.001	***	3	6.588	0.212	NS
	Interaction	3	287.2	<0.001	***	3	6.588	0.212	NS
gs	Inoculation	1	73.467	<0.001	***	1	17.293	<0.001	***
	Time	3	152.302	<0.001	***	3	351.249	<0.001	***
	Interaction	3	446.517	<0.001	***	3	8.002	0.109	NS
E	Inoculation	1	367.818	<0.001	***	1	8.662	0.008	**
	Time	3	2310.636	<0.001	***	3	611.459	<0.001	***
	Interaction	3	2134.821	<0.001	***	3	1.665	0.682	NS
A	Inoculation	1	174.882	<0.001	***	1	1.068	0.312	NS
	Time	3	1371.446	<0.001	***	3	13.84	0.02	*
	Interaction	3	1579.315	<0.001	***	3	19.576	0.005	*
Ci	Inoculation	1	269.442	<0.001	***	1	1.206	0.292	NS
	Time	3	727.061	<0.001	***	3	75.836	<0.001	***
	Interaction	3	573.205	<0.001	***	3	27.735	<0.001	***
%Fv/Fm	Inoculation	1	20.391	<0.001	***	1	7.692	0.012	*
	Time	3	15.656	<0.001	***	3	11.978	0.043	*
	Interaction	3	5.389	0.004	**	3	12.371	0.039	*
RWC	Inoculation	1	7.144	0.012	*	1	3.023	0.102	NS
	Time	3	30.981	<0.001	***	3	6.91	0.153	NS
	Interaction	3	8.635	<0.001	***	3	2.831	0.495	NS
Ψmd	Inoculation	1	0.159	0.696	NS	1	0.014	0.909	NS
	Time	3	270.51	<0.001	***	3	120.398	<0.001	***
	Interaction	3	4.601	0.278	NS	3	10.027	0.072	NS
ABA	Inoculation	1	154.18	<0.001	***	1	1.76	0.202	NS
	Time	3	179.323	<0.001	***	3	3.614	0.381	NS
	Interaction	3	175.359	<0.001	***	3	0.321	0.961	NS
PA	Inoculation	1	2.614	0.133	NS	1	11.114	0.002	**
	Time	3	3.486	0.405	NS	3	2.622	0.067	NS
	Interaction	3	2.761	0.502	NS	3	0.205	0.892	NS
DPA	Inoculation	1	7.586	0.02	*	1	1.061	0.322	NS
	Time	3	5.485	0.243	NS	3	7.317	0.122	NS
	Interaction	3	2.483	0.558	NS	3	1.159	0.788	NS
JA	Inoculation	1	3.201	0.114	NS	1	0.918	0.355	NS
	Time	3	11.555	0.041	*	3	15.305	0.017	*
	Interaction	3	4.401	0.304	NS	3	10.121	0.06	NS
JA-Ile	Inoculation	1	2.503	0.154	NS	1	0.383	0.55	NS
	Time	3	14.318	0.02	*	3	4.011	0.34	NS
	Interaction	3	8.425	0.093	NS	3	5.188	0.238	NS
SA	Inoculation	1	0.371	0.551	NS	1	0.326	0.588	NS
	Time	3	19.688	0.006	*	3	3.718	0.382	NS
	Interaction	3	0.891	0.848	NS	3	2.173	0.599	NS
IAA	Inoculation	1	3.337	0.118	NS	1	2.441	0.16	NS
	Time	3	4.493	0.315	NS	3	9.465	0.074	NS
	Interaction	3	4.97	0.276	NS	3	6.311	0.177	NS
SOD	Inoculation	1	1.537	0.224	NS	1	1.537	0.224	NS
	Time	3	16.895	<0.001	***	3	16.895	<0.001	***
	Interaction	3	6.647	0.001	**	3	6.647	0.001	**
CAT	Inoculation	1	0	0.998	NS	1	1.688	0.23	NS
	Time	3	11.823	0.035	*	3	77.036	<0.001	***
	Interaction	3	3.169	0.434	NS	3	1.863	0.647	NS
MDHAR	Inoculation	1	0.019	0.894	NS	1	0.28	0.613	NS
	Time	3	60.552	<0.001	***	3	127.97	<0.001	***
	Interaction	3	10.416	0.058	NS	3	1.36	0.753	NS
DHAR	Inoculation	1	1.928	0.197	NS	1	0.087	0.777	NS
	Time	3	25.637	0.003	**	3	143.944	<0.001	***
	Interaction	3	4.362	0.319	NS	3	0.812	0.869	NS
GR	Inoculation	1	0.002	0.967	NS	1	1.264	0.272	NS
	Time	3	5.751	0.203	NS	3	54.678	<0.001	***
	Interaction	3	11.929	0.037	*	3	0.052	0.998	NS

Degrees of freedom (Df) and F values are presented for each source of variation. Significance: ^*P ≤ 0.05, ^**P ≤ 0.01, ^***P ≤ 0.001. NS, non-significant. Variables: RIN: stem relative internal necrosis length; gs: stomatal conductance; E: transpiration rate; A: net CO₂ assimilation rate; Ci: sub-stomatal CO₂ concentration; %Fv/Fm: percent variation of maximum quantum yield of PSII; RWC: relative water content; ψ_md: midday shoot water potential; ABA: abscisic acid; PA: phaseic acid; DPA: dihydrophaseic acid; JA: jasmonic acid; JA-Ile: JA-isoleucine; SA: salicylic acid; IAA: indole acetic acid; SOD: superoxide dismutase; CAT: catalase; MDHAR: monodehydroascorbate reductase; DHAR: dehydroascorbate reductase; GR: glutathione reductase.

Plant water status, needle gas exchange and susceptibility to photoinhibition

In contrast with P. pinea, P. radiata RWC was significantly affected by both inoculation and time, as well as by its interaction (Table 1). Post hoc analysis revealed significant differences on RWC between non-inoculated controls and inoculated P. radiata plants at the last sampling point (10 d.p.i., Table 2). Although a significant effect of time on plant water potential was found for both P. radiata and P. pinea plants, neither F. circinatum inoculation nor its interaction with time significantly influenced this variable (Table 1).

Table 2

Water status of P. radiata and P. pinea seedlings inoculated with F. circinatum (F) and respective controls (C) overtime: RWC and midday shoot water potential (ψ_md).

		RWC (%)	Ψmd (MPa)
P. radiata
2 h.p.i.	C	73.81 ± 2.39	−0.45 ± 0.01
	F	75.68 ± 1.09	−0.51 ± 0.01
1 d.p.i.	C	79.88 ± 3.14	−0.48 ± 0.03
	F	81.73 ± 2.86	−0.48 ± 0.03
6 d.p.i.	C	82.03 ± 2.64	−0.22 ± 0.01
	F	79.23 ± 2.56	−0.22 ± 0.02
10 d.p.i.	C	70.53 ± 1.79	−0.26 ± 0.06
	F	51.77 ± 1.24*	−0.23 ± 0.02
P. pinea
2 h.p.i.	C	80.62 ± 2.3	−0.58 ± 0.07
	F	79.98 ± 0.9	−0.5 ± 0.06
1 d.p.i.	C	86.12 ± 2.8	−0.32 ± 0.01
	F	83.16 ± 3.42	−0.37 ± 0.02
6 d.p.i.	C	86.71 ± 2.19	−0.23 ± 0.01
	F	77.01 ± 4.26	−0.2 ± 0
10 d.p.i.	C	89.64 ± 6.37	−0.2 ± 0
	F	86.47 ± 0.51	−0.28 ± 0.06

Data are presented as mean ± SE. For significant interaction, the asterisks and bold style on the F rows indicate significantly different means in comparison with the respective control for each species and sampling time (post hoc test, α = 0.05).

Inoculation with F. circinatum and time post-inoculation was shown to affect P. radiata leaf gas exchange-related parameters, and a significant interaction was also found between the two factors (Table 1). The inoculation of P. radiata plants with F. circinatum resulted on a significant decrease of gs, E, and A 10 d.p.i. in comparison with its respective non-inoculated controls, together with a significant increase of Ci (Figure 2). Moreover, at 2 h.p.i. P. radiata plants inoculated with F. circinatum showed significantly higher gs and E values than the controls (Figure 2A and B).

Figure 2.

Needle gas exchange-related parameters of P. radiata and P. pinea seedlings inoculated with F. circinatum (F) and respective controls (C) overtime. (A) Stomatal conductance (gs); (B) transpiration rate (E); (C) net CO₂ assimilation rate (A); (D) sub-stomatal CO₂ concentration (Ci). Data are presented as mean ± SE. For significant interaction, asterisks on F columns indicate significantly different means in comparison with the respective control for each species and sampling time (post hoc test, α = 0.05). The bracket common to F columns represents a significant effect of F. circinatum inoculation respective to controls (ANOVA, P ≤ 0.05).

On the other hand, for P. pinea only time had a significant effect on all leaf gas exchange-related parameters, although no interaction with F. circinatum inoculation was found (Table 1). However, the inoculation of P. pinea plants with F. circinatum significantly influenced E and gs, but with no significant interaction between the two factors considered (Table 1, Figure 2A and B). In addition, a significant interaction between inoculation with Fusarium and time post-inoculation was found for A and Ci in P. pinea (Table 1), with inoculated plants showing higher Ci values than controls 1 d.p.i. (Figure 2D). However, post hoc tests did not detect significant changes between inoculated and control P. pinea plants regarding A (Figure 2C).

For both P. radiata and P. pinea F. circinatum inoculation, time post-inoculation and its interaction significantly affected the percent variation of F_v/F_m, although at higher significance levels for P. radiata (Table 1). However, significant changes on %F_v/F_m were only found for P. radiata 10 d.p.i., with inoculated plants presenting lower percentages than its respective non-inoculated controls (Figure 3). Initial pre-illuminated Fv/Fm was consistent among species, treatments and sampling times (0.75 ± 0.01).

Figure 3.

Chlorophyll a-related parameters of P. radiata and P. pinea seedlings inoculated with F. circinatum (F) and respective controls (C) over time. (A) Percent variation of maximum quantum yield of PSII (%F_v/F_m). Data are presented as mean ± SE. For significant interaction, asterisks on F columns indicate significantly different means in comparison with the respective control for each species and sampling time (post hoc test, α = 0.05). (B) Exemplary images of the data obtained using imaging chlorophyll fluorometry. Maximum quantum yield of PSII (Fv/Fm) values of each of the three replicates per species and conditions (C and F) are represented by the color scale. Fv/Fm pre-illumination results and after light stress-recovery experiments are shown for P. radiata and P. pinea 1 d.p.i..

Hormone concentration and immunolocalization

From the hormones quantified, only ABA’s concentration in P. radiata was significantly affected by both F. circinatum inoculation, time post-inoculation and the interaction of both factors (Table 1). In fact, inoculation of P. radiata plants with F. circinatum led to a significant increase of ABA concentration 10 d.p.i. in comparison with the respective controls (Table 3). A significant effect of the inoculation with F. circinatum was also detected on DPA for P. radiata (Table 1, Table 3). Although the time post-inoculation was shown to significantly influence JA, JA-Ile and SA in P. radiata, no effect of F. circinatum inoculation or of its interaction with time was found (Table 1). No significant effects were detected for PA and IAA in P. radiata (Table 1).

Table 3

Hormone concentration in needles of P. radiata and P. pinea inoculated with F. circinatum (F) and its respective controls (C) overtime: abscisic acid (ABA); phaseic acid (PA); dihydrophaseic acid (DPA); jasmonic acid (JA); JA-isoleucine (JA-Ile); salicylic acid (SA); indole acetic acid (IAA).

		Hormone concentration (ng g DW−1)
		ABA	PA	DPA	JA	JA-Ile	SA	IAA
P. radiata
2 h.p.i.	C	144.671 ± 21.918	24.759 ± 7.302	0.098 ± 0.001	54.554 ± 17.321	48.051 ± 14.156	106.196 ± 25.989	31.22 ± 11.708
	F	149.275 ± 19.402	34.035 ± 12.884	5.288 ± 2.855*	19.019 ± 7.971	16.013 ± 6.109	84.068 ± 7.965	23.055 ± 8.49
1 d.p.i.	C	112.605 ± 23.759	35.675 ± 15.049	2.087 ± 0.784	6.519 ± 2.432	6.724 ± 0.977	93.122 ± 17.214	31.732 ± 17.049
	F	150.577 ± 33.337	36.025 ± 12.076	5.585 ± 1.553*	4.402 ± 1.768	6.193 ± 1.263	69.31 ± 10.312	85.636 ± 44.431
6 d.p.i.	C	178.489 ± 30.23	24.803 ± 7.026	1.069 ± 0.854	4.37 ± 1.214	5.788 ± 1.495	116.946 ± 23.016	14.278 ± 5.335
	F	168.422 ± 40.424	26.088 ± 13.883	4.991 ± 4.264*	5.835 ± 2.709	4.891 ± 2.234	118.71 ± 5.879	31.069 ± 9.664
10 d.p.i.	C	150.862 ± 11.088	28.54 ± 4.487	5.305 ± 4.01	3.821 ± 1.03	5.87 ± 2.922	194.111 ± 30.381	22.586 ± 10.955
	F	2276.532 ± 158.132*	91.454 ± 35.087	24.671 ± 9.413*	4.827 ± 2.057	13.669 ± 2.665	198.073 ± 41.54	273.968 ± 163.741
P. pinea
2 h.p.i.	C	66.07 ± 17.729	86.234 ± 11.566	11.637 ± 1.854	7.563 ± 2.253	9.767 ± 1.14	102.055 ± 29.534	39.484 ± 13.265
	F	76.184 ± 13.827	112.29 ± 12.249*	14.281 ± 4.037	4.021 ± 1.033	5.682 ± 2.151	137.812 ± 14.333	40.821 ± 7.287
1 d.p.i.	C	56.558 ± 7.217	107.55 ± 7.227	7.579 ± 1.295	4.381 ± 1.543	9.734 ± 5.951	118.697 ± 20.21	36.399 ± 11.261
	F	61.862 ± 6.162	131.448 ± 17.376*	9.708 ± 1.148	3.592 ± 0.931	8.892 ± 2.233	133.186 ± 18.388	26.091 ± 8.781
6 d.p.i.	C	47.764 ± 4.235	73.775 ± 2.961	6.864 ± 1.321	0.792 ± 0.292	3.986 ± 1.483	229.729 ± 107.874	21.234 ± 13.203
	F	56.24 ± 7.128	109.848 ± 14.911*	8.162 ± 1.506	2.628 ± 0.458	6.589 ± 1.241	100.113 ± 20.503	91.623 ± 60.167
10 d.p.i.	C	53.726 ± 3.36	83.758 ± 6.651	7.062 ± 0.961	2.305 ± 0.733	8.895 ± 2.278	95.759 ± 9.769	45.689 ± 10.513
	F	67.878 ± 12.109	102.862 ± 9.003*	6.647 ± 1.572	1.708 ± 0.438	6.503 ± 1.687	107.168 ± 13.897	90.464 ± 16.762

Data are presented as mean ± SE. For significant interaction, the asterisks and bold style on the F rows indicate significantly different means in comparison with the respective control for each species and sampling time (post hoc test, α = 0.05). This together with the underlining represents a significant effect of F. circinatum inoculation (ANOVA, P ≤ 0.05).

Few statistically significant changes were observed in P. pinea regarding hormone concentrations, with none of these variables being affected by the interaction of F. circinatum inoculation and time post-inoculation (Table 1). Time post-inoculation was shown to significantly affect JA concentration, but no significant interaction with F. circinatum inoculation was found (Table 1). However, an interesting effect of F. circinatum inoculation on PA concentration was found for P. pinea plants (Table 1, Table 3).

Needle hormone immunodetection complemented quantification data. Pinus radiata presented a high ABA signal in the control (Figure 4A b) and 2 h.p.i. and 6 d.p.i. in inoculated plants (Figure 4A c and e). This signal was reduced 10 d.p.i. (Figure 4A f). For inoculated P. pinea, ABA accumulated progressively reaching a maximum 6 d.p.i., followed by an accentuated decrease 10 d.p.i. (Figure 4B). For both species, JA immunolocalization revealed low levels in controls (Figure 4). While in P. pinea JA signal increased rapidly in mesophyll cells after inoculation with F. circinatum and was maintained at all sampling points (Figure 4B), in P. radiata JA signal reached a maximum 6 d.p.i. and was decreased afterwards (Figure 4A). In P. radiata, although IAA seems to be absent in controls and in inoculated plants 2 h.p.i. and 6 d.p.i., it was distributed in mesophyll cells and endodermis 1 d.p.i. and also in vascular tissue 10 d.p.i. (Figure 4A). Evident changes on IAA signal were not found for the different species or sampling times studied. In general, a complete distribution of the ethylene acid precursor ACC was registered for both pine species (Figure 4), although with lower intensity 2 h.p.i. and 6 d.p.i. for P. pinea (Figure 4B c and e).

Figure 4.

(A) Inmunolocalization of hormones in transversal sections of needles of P. radiata inoculated with F. circinatum (c—2 h.p.i.; d—1 d.p.i.; e—6 d.p.i.; f—10 d.p.i.) and mock-inoculated controls (b; 2 h.p.i.) using confocal microscopy. Differential interference contrast is represented on (a). Blue signal corresponds to DAPI (4'6-diamidino-2-phenylindole) staining, whereas hormones signal is represented in green. (B) Inmunolocalization of hormones in transversal sections of needles of P. pinea inoculated with F. circinatum (c—2 h.p.i.; d—1 d.p.i.; e—6 d.p.i.; f—10 d.p.i.) and mock-inoculated controls (b; 2 h.p.i.) using confocal microscopy. Differential interference contrast is represented on (a). Blue signal corresponds to DAPI staining, whereas hormones signal is represented in green.

Activity of antioxidant enzymes

The interaction of F. circinatum inoculation with time post-inoculation had a significant effect on GR specific activity in P. radiata (Table 1). However, no significant changes between inoculated P. radiata plants and its respective non-inoculated controls at each sampling time were found regarding GR specific activity (Figure 5E). The specific activities of MDHAR, DHAR, CAT and SOD in P. radiata were shown to be affected by the time post-inoculation, but only SOD specific activity was significantly influenced by its interaction with F. circinatum inoculation (Table 1), with inoculated plants showing a significant decrease on SOD specific activity in comparison with its respective controls at 10 d.p.i. (Figure 5A).

Figure 5.

Antioxidant enzymes activity fold-change in needles of P. radiata and P. pinea seedlings inoculated with F. circinatum in respect to their respective controls overtime. (A) superoxide dismutase (SOD); (B) catalase (CAT); (C) monodehydroascorbate reductase (MDHAR); (D) dehydroascorbate reductase (DHAR); (E) gluthatione reductase (GR). For significant interaction, asterisks indicate significant changes for each species and sampling time (post hoc test, α = 0.05). APx: ascorbate peroxidase; Asc: ascorbate; MDHA: monodehydroascorbate; DHA: dehydroascorbate; GSH: reduced glutathione; GSSH: oxidized glutathione; NADPH: reduced nicotinamide adenine dinucleotide phosphate; NADP: oxidized nicotinamide adenine dinucleotide phosphate.

Although time post-inoculation was shown to significantly affect GR, MDHAR, DHAR, CAT and SOD specific activities in P. pinea, a significant interaction with F. circinatum inoculation was found only for SOD specific activity (Table 1). However, post hoc tests did not detect any significant differences on SOD specific activity between inoculated P. pinea plants and its respective controls at each sampling point (Figure 5A).

Integrated data analysis

The principal component analysis highlighted the separation of inoculated P. radiata samples from the remaining groups at the last sampling point (10 d.p.i.; Figure 6A). This separation is mainly based on PC1, which explains 48.7 % of the variation found and is highly influenced by relative internal stem necrosis, ABA and Ci. In contrast, for P. pinea the PCA was not able to find a clear pattern of response, either over time or comparing mock-inoculated controls with samples inoculated with F. circinatum (Figure 6B).

Figure 6.

Principal components analyses of (A) P. radiata and (B) P. pinea seedlings inoculated with F. circinatum (F) and respective controls (C) over time. The first two components considered for each pine species are plotted in the graphs. The proportion of variance explained by each component is represented on the axis labels. Analyses were performed considering all quantitative variables: RIN: stem relative internal necrosis length; gs: stomatal conductance; E: transpiration rate; A: net CO₂ assimilation rate; Ci: sub-stomatal CO₂ concentration; %Fv/Fm: percent variation of maximum quantum yield of PSII; RWC: relative water content; ψ_md: midday shoot water potential; ABA: abscisic acid; PA: phaseic acid; DPA: dihydrophaseic acid; JA: jasmonic acid; JA-Ile: JA-isoleucine; SA: salicylic acid; IAA: indole acetic acid; SOD: superoxide dismutase; CAT: catalase; MDHAR: monodehydroascorbate reductase; DHAR: dehydroascorbate reductase; GR: glutathione reductase.

Discussion

The analysis carried out allowed us to identify novel responses associated to F. circinatum infection in symptomatic P. radiata, as well as in P. radiata and P. pinea at earlier stages of disease progression. The greatest changes were found for symptomatic P. radiata seedlings. These are in accordance with previous studies (Cerqueira et al. 2017, Amaral et al. 2019a, 2019b) reporting photosynthesis impairment, ABA accumulation associated with stomatal closure and reduced RWC in P. radiata after inoculation with F. circinatum. In addition, although P. radiata maximum quantum yield of PSII remained stable in the presence of F. circinatum, an increased susceptibility to photoinhibition was detected 10 d.p.i.. This may indicate that D1 repair rate in the core of the PSII [see Theis and Schroda (2016) for further detail] may be insufficient to control the damage occurring at this stage and that the mechanisms responsible for the dissipation of the excess excitation energy present after complete CO₂ fixation shut-down seem to fail. In fact, previous studies showed that although the abundance of GOX (glycolate oxidase; one of the first enzymes in photorespiration) transcripts increased in symptomatic P. radiata, the role of photorespiration in the P. radiata–F. circinatum interaction is unclear (Amaral et al. 2019a).

Moreover, the decrease of SOD activity in symptomatic P. radiata suggests the failure of the water–water cycle to control the levels of ROS under F. circinatum attack at this point. SOD conversion of superoxide into H₂O₂ is the first line of defense against plant oxidative stress. It is known that once plants, antioxidant activity is insufficient to fight ROS effects, cell death and tissue necrosis occur (Gill and Tuteja 2010). In P. pinaster, showing a moderate susceptibility to F. circinatum, increased antioxidant activity was related to necrosis resulting from inoculation with the pathogen (Vivas et al. 2014). Although this could explain the development of PPC symptoms in the susceptible species, previous studies did not report the occurrence of membrane damage in symptomatic seedlings of P. radiata inoculated with F. circinatum (Cerqueira et al. 2017, Amaral et al. 2019b). An up-regulation of SOD genes was found for the highly susceptible P. patula 7 d.p.i. with F. circinatum, but not in the resistant P. tecunumanii (Visser et al. 2019).

Although few changes are reported regarding the first sampling points, an interesting adjustment of stomata occurs in inoculated P. radiata seedlings 2 h.p.i. with F. circinatum. However, this significantly higher stomatal conductance (accompanied by increased transpiration rate) is not maintained in the following sampling points and drastically decreases when plants display symptoms, indicating complete stomata closure. In contrast, inoculation of P. pinea with F. circinatum resulted on a generalized increase of both stomata opening and transpiration rate. The ABA-induced stomatal closure in a water stress-like scenario resulting from the inoculation of susceptible species with F. circinatum has been discussed in previous works, as well as the association of ABA accumulation with increased disease susceptibility (Cerqueira et al. 2017, Amaral et al. 2019a, 2019b).

Abscisic acid catabolism may have a key role in the outcome of Pinus–F. circinatum interaction. In this process, ABA 8′-hydroxylase is responsible for the conversion of ABA into PA, which may then be transformed into DPA. While F. circinatum inoculation significantly increased DPA levels in the susceptible species, an increase of PA levels occurred instead in P. pinea. Although Visser et al. (2019) found a down-regulation of an ABA hydroxylase 7 d.p.i. with F. circinatum in the resistant P. tecunumanii and an up-regulation in the susceptible P. patula, the authors discuss that further physiological studies are needed. In fact, it is known that changes at the transcriptomic level are not always transcribed into the responses expected due to post-translational regulatory mechanisms.

The accumulation of DPA in P. radiata may suggest an attempt of the host to control the pool of bio-active ABA and maintain ABA homeostasis through its inactivation. Hernandez-Escribano et al. (2020) reported the up-regulation of ABA 8′-hydroxylase in the moderately susceptible P. pinaster 5 and 10 d.p.i. with F. circinatum. The infection of a susceptible wheat cultivar with Fusarium graminearum resulted in a significant increase of ABA, PA and DPA 4 d.p.i., with DPA showing the highest increase with respect to the control (Qi et al. 2016). Low levels of ABA, PA and DPA were found by the same authors in mycelia of F. graminearum but not in its growth media, supporting that the changes observed were mostly a result of plants defense mechanism rather than a direct effect of fungal hormone production.

Unlike DPA, PA can bind to certain ABA receptors and have some ABA-like activity, and it has been proposed to be related to water stress tolerance (Weng et al. 2016). Therefore, the catabolism of ABA to a less active form may in part determine P. pinea fate under F. circinatum infection. However, Wang et al. (2018) reported that the most resistant wheat cultivar to fusarium head blight accumulated the lowest PA levels and high DPA levels in spikelets 4 d.p.i. with F. graminearum, but ABA accumulation was associated with susceptible cultivars only.

Furthermore, while a fluctuation of the (active and non-active) ABA signal was observed by confocal analysis for inoculated P. radiata, in P. pinea this signal seems to gradually increase reaching a maximum 6 d.p.i.. This sampling point also corresponds to the greater difference between control and inoculated plants PA levels in P. pinea. This may be related to the infection outcome, occurring just before the development of the first disease symptoms in the susceptible P. radiata 7 d.p.i.. Similarly, Carrasco et al. (2017) suggested that the resistance of selected genotypes of P. radiata to F. circinatum could be explained by higher gene expression 6 d.p.i., including the overexpression of ABA 8′-hydroxylase. Therefore, this could be considered a turning point, which together with the maintenance of energy production through photosynthesis [conceivably to fuel effective defense mechanisms; reviewed by Bolton (2009)] may be crucial to determine P. pinea resistance to PPC.

This study evidences the increased oxidative status of symptomatic P. radiata and its possible relation with necrosis development and with changes on the PSII machinery. Moreover, it discusses the role of ABA catabolism in Pinus defense response against F. circinatum infection, suggesting that the accumulation of DPA may be associated to plant susceptibility and PA accumulation to plant resistance, with 6 d.p.i. as a decisive turning point to define the outcome of Pinus–Fusarium interaction. Further studies are needed since the role of ABA catabolism in plant--pathogen interactions is not well documented, while representing a promising field of research. Once the complex regulation of hormone homeostasis during plant--pathogen interaction is unveiled, hormone-based breeding strategies could be applied to improve plant disease resistance (Denancé et al. 2013). Moreover, the application of ABA and/or its catabolites should be further explored as a means to induce Pinus resistance to F. circinatum. Previous experiments demonstrated that both SA (Fitza et al. 2013) and methyl jasmonate (Vivas et al. 2012) application was not effective to improve resistance to PPC.

Acknowledgments

We thank our colleagues Giorgia Bastianelli (University of Tuscia, Italy) and Pedro Monteiro (University of Aveiro, Portugal) for their assistance on plant inoculation and sampling. We also acknowledge Marta Pitarch’s (Universitat Jaume I, Spain) help during hormone quantification. Furthermore, we thank Laura Lamelas (University of Oviedo, Spain) and Jorge Martín-Garcia (University of Valladolid, Spain) for help regarding statistical analysis.

Conflict of interest

None declared.

Funding

This research was supported by FEDER through COMPETE (Programa Operacional Fatores de Competitividade) (POCI-01-015-FEDER-016785) and by National Funds through the Portuguese Foundation for Science and Technology (FCT) within the project URGENTpine (PTDC/AGR-FOR/2768/2014). Thanks are due to FCT/MCTES for financial support to CESAM (UID/50017/2020+UIDB/50017/2020) through national funds. FCT also supported J.A. (SFRH/BD/120967/2016). The Spanish Ministry of Economy and Competitiveness supported L.V. through the Ramón y Cajal Programme (RYC-2015-17871) and M.E. by the grant FJCI-2017-31613. COST Action FP1406 PINESTRENGTH (PPC—strategies for management of Gibberella circinata in greenhouses and forests), supported by COST (European Cooperation in Science and Technology), awarded B.C. with a STSM. The James Hutton Institute receives support from the Rural and Environment Science and Analytical Services Division of the Scottish Government.

Authors’ contributions

G.P., A.A. and J.A. designed and supervised the experiment. J.A., B.C. and C.J. performed the experiment. J.S. designed and supervised chlorophyll a fluorescence analysis, which were performed by C.J.. A.G.-C. supervised hormonal quantification, which was performed by B.C.. J.A., B.C. and M.E., and C.J. performed antioxidant enzyme activity analysis under supervision from R.D.H. and L.-T.D.. M.E. performed hormone immunolocalization. L.V. supervised statistical analysis, which was performed by J.A.. J.A. wrote the manuscript. All authors discussed the data and reviewed the manuscript.