Are arbuscular-mycorrhizal Alnus incana seedlings more resistant to drought than ectomycorrhizal and non-mycorrhizal ones?

(1)

UEF//eRepository

DSpace https://erepo.uef.fi

Rinnakkaistallenteet Luonnontieteiden ja metsätieteiden tiedekunta

2020

Are arbuscular-mycorrhizal Alnus incana seedlings more resistant to drought than ectomycorrhizal and non-mycorrhizal ones?

Kilpeläinen, Jouni

Oxford University Press (OUP)

Tieteelliset aikakauslehtiartikkelit

http://dx.doi.org/10.1093/treephys/tpaa035

https://erepo.uef.fi/handle/123456789/8149

Downloaded from University of Eastern Finland's eRepository

(2)

Are arbuscular-mycorrhizal Alnus incana seedlings more resistant to drought than ectomycorrhizal and non-mycorrhizal ones?

Jouni Kilpeläinen^1,2, Pedro J. Aphalo³, Aitor Barbero-López¹, Bartosz Adamczyk^4,5, Sammi Alam Nipu¹, Tarja Lehto^1*

1University of Eastern Finland, School of Forest Sciences, P.O. Box 111, 80101 Joensuu, Finland

2Present address: Natural Resources Institute Finland (Luke), Joensuu, Finland

3University of Helsinki, Department of Biosciences, Helsinki, Finland

4University of Helsinki, Department of Agriculture and Institute for Atmospheric and Earth System Research (INAR), Finland

5Present address: Natural Resources Institute Finland, Helsinki, Finland

Running head: Arbuscular- and ectomycorrhizal Alnus in drought

*Corresponding author, email tarja.lehto@uef.fi, tel. +358-50-4422818

Downloaded from https://academic.oup.com/treephys/advance-article-abstract/doi/10.1093/treephys/tpaa035/5809513 by Joensuun Yliopisto Kirjasto user on 08 May 2020

(3)

2 Abstract

Arbuscular mycorrhizas (AM) prevail in warm and dry climates and ectomycorrhizas (EM) in cold and humid climates. We suggest that the fungal symbionts benefit their host plants especially in the corresponding conditions. The hypothesis tested was that AM plants are more drought resistant than EM or non-mycorrhizal (NM) plants.

Grey alder (Alnus incana) seedlings were inoculated with two species of either AM or EM fungi or none. In one controlled-environment experiment, there was a watering and a drought treatment. Another set of seedlings were not watered until permanent wilting.

The AM plants were somewhat smaller than EM and NM, and at the early stage of the drought treatment the soil-moisture content was slightly higher in the AM pots. Shoot water potential was highest in the AM treatment during severe drought, while stomatal conductance and photosynthesis did not show a mycorrhizal effect. In the lethal-drought set the AM maintained their leaves longer than EM and NM plants, and the AM seedlings survived longer than NM seedlings. Foliar phosphorus and sulfur concentrations remained higher in AM plants than EM or NM but potassium, copper and iron increased in EM during drought.

The root tannin concentration was lower in AM than EM and drought doubled it.

Although the difference to EM plants was not large, the hypothesis was supported by the better performance of AM plants during a severe short-termed drought. Sustained phosphorus nutrition during drought in AM plants was a possible reason for this. Moreover, the higher foliar sulfur and lower metal-nutrient concentrations in AM may reflect differences in nutrient uptake or (re)translocation during drought, which merit further research. The much larger tannin concentrations in EM root systems than AM did not appear to protect the EM plants from drought. The differential tannin accumulation in AM and EM plants needs further attention.

Key words

arbuscular mycorrhiza, grey alder, ectomycorrhiza, nutrient, secondary metabolite, tannin, water deficit, water potential, water stress

(4)

3 Introduction

The majority of land plants live in symbiosis with soil fungi forming mycorrhizas. The most common type both in the plant kingdom and geographically is arbuscular mycorrhiza (AM).

Ectomycorrhizas (EM) are limited to a small number of host plants, but these are often dominant trees. Most ectomycorrhizal trees prevail in cool and humid regions (although not exclusively), while arbuscular mycorrhizal plants are widespread in different biomes,

including habitats that are too dry for trees. We have presented a general hypothesis that one of the reasons for the geographic distribution of these mycorrhiza types is a better drought resistance of arbuscular mycorrhizal fungi (AMF) than ectomycorrhizal fungi (EMF) while EMF are more resistant to cold (Lehto and Zwiazek 2011, Kilpeläinen et al. 2016, 2017). We suggest that these mycorrhiza types benefit their host plants most in the corresponding

environmental conditions. Here we explore the possibility that AMF benefit their host plants in dry conditions more than EMF.

The effect of AM and EM on the water relations and drought resistance of the host plants has been the topic of numerous studies, yet it is difficult to draw general conclusions because results have varied from one experiment to another (Augé 2001, Smith et al. 2010, Lehto and Zwiazek 2011, Augé et al. 2015). Nevertheless, AM have either increased the drought

resistance of the host plants or not shown any effect (Augé 2001). Also, for EM there are results showing both no effect and increased drought resistance. However, several studies have shown a negative effect of EM on the host water relations compared to non-mycorrhizal ones (NM), and reduced root hydraulic conductance is a likely reason for this (Lehto and Zwiazek 2011, Korhonen et al. 2019). Moreover, drought resulted in shorter fine-root lifespan in four temperate EM species compared to four AM species (Liese et al. 2019), and the loss of plant photosynthates by root exudation was much more increased in EM species than AM during drought (Liese et al. 2018).

In many cases an improved drought resistance by mycorrhizal colonization can be attributed to improved mineral nutrition, as adequate levels of N, P and other nutrients are essential for stress tolerance (Lehto and Zwiazek 2011). In P deficient plants, root hydraulic conductivity is impaired (Andersen et al. 1989, Coleman et al. 1990). Therefore, the nutrient status of the experimental plants in different mycorrhiza treatments should preferably be the same at the onset of the drought stress treatment, or at least it should be measured to allow assessing the

(5)

4 roles of nutritional and non-nutritional effects of mycorrhizas (Lehto 1992a). Moreover, nutrient uptake is reduced by drought both because of changes in soil-nutrient availability and the nutrient uptake capacity of roots and mycorrhizas. Only a few of the studies on the effects of mycorrhizas on plant performance have explored the nutrient uptake during drought (Lehto 1992b), even though nutrient uptake is arguably the most important function of mycorrhizas from the point of view of the host.

Comparisons between AM and EM physiology can be done using several species of each type (e.g. Liese et al. 2018, 2019), or alternatively, host species which form both types.

Previously we showed that grey alder (Alnus incana (L.) Moench) seedlings formed both EM and AM in experimental conditions more reliably than seven other tree species tested

(Kilpeläinen et al. 2019). In non-stress conditions, we showed that there was no difference in CO2 assimilation per unit leaf area, stomatal conductance or shoot water potential of AM, EM or NM grey alder seedlings.

Tannin accumulation in roots is one of the well-known plant reactions to EM colonization (Pfabel et al. 2012), but less attention has been paid to tannins in AM symbiosis. Woody plants tend to accumulate tannins also as a response to different stresses, and in cold climates, their concentrations increase strongly in the autumn in conifers (Rummukainen et al. 2007).

Tannins play a well-known role in protection from pests and pathogens (Hagerman 2012) while their function in protection from abiotic stress remains less well understood. Tannins affect the carbon and nutrient economy of plants and their cycling, as they represent a substantial proportion of the biomass of trees, and they can retard litter decomposition and precipitate proteins in soils (Smolander et al. 2012).

The aim of this study was to compare the effect of AM and EM on drought resistance in grey alder seedlings. In one experiment we grew grey alder seedlings inoculated with either EMF, AMF or none in controlled conditions, subjected half of each group to a water-withholding treatment, and recorded their growth, water relations, photosynthesis and nutrient

accumulation. In another experiment, we followed the development of lethal drought in a different set of plants from the same cohort. We hypothesized that AM are more drought resistant than either NM or EM plants and accumulate more nutrients during drought.

Furthermore, the EM and NM treatments were expected not to differ from each other in this respect.

(6)

5 Materials and methods

Experimental plants and fungi

Grey alder was selected as a model species for comparisons of AM and EM in cool-climate studies after testing the formation of mycorrhizas in eight broadleaf species (Kilpeläinen et al.

2019). Grey alder is a pioneer species which regenerates and grows in a wide range of habitats, including both dry sandy sites and shores. The seeds originated from Loppi, southern Finland. In the simplified model system used, the N2-fixing symbiont Frankia was not included, and nutrients were supplied in soluble mineral forms. Two AMF and two EMF species were used. The AMF were Glomus hoi (Luke, Laukaa, Finland) and Rhizophagus intraradices (formerly sold as a commercial product Myko-Ymppi, Luke, Laukaa, Finland).

The AM fungi were originally isolated and identified using spore morphology by Mauritz Vestberg in Finland. One EMF was Paxillus involutus isolated from a fruiting body in Joensuu, Finland. The other one was an ascomycete fungus (EMF-359-2-1) previously isolated from mycorrhizas of A. incana grown in forest soil in experimental conditions (Kilpeläinen et al. 2016), and sequenced and identified as an ascomycete (Kilpeläinen et al.

2019). Recently it was identified as Scytalidium album, courtesy of Dr Thorunn Helgason. In preliminary experiments both grey alder and black alder (Alnus glutinosa) formed

mycorrhizas with both these EM fungi. Independently of us, Schweiger (2016) also used grey alder as a test species to study N isotope fractionation in AM and EM seedlings, and achieved extensive colonization with Rhizophagus intraradices (R. irregularis) and Paxillus involutus in different seedlings. Sometimes P. involutus is argued not to form mycorrhizas with alders.

However, this concept is not based on experimental work that would show failure of

inoculation, but a brief mention of not finding this fungus with alders (Laiho 1970). Laiho did not exclude the possibility, as his survey did not focus on alder.

Cultivation of plants and fungi

Alnus incana seedlings were grown under controlled conditions in a walk-in growth chamber (Conviron GR77, Controlled Environments, Winnipeg, MB, Canada). Seeds were sterilized in 30% H2O2 using tea infusers (Kilpeläinen et al. 2016) and sown in 165 cm³, 21 cm deep

(7)

6 plastic pots (SC10 U, Ray Leach „Cone-tainers‟, Stuewe and Sons, Tangent, OR, USA) sterilized in 70% ethanol overnight and filled with acid washed perlite in February, 2015. On day 1, eight seeds were sown in each pot, which were covered with Clingfilm and kept in the growth chamber. Plants were thinned to one per pot by cutting the culled ones on day 48 after sowing. The initial conditions, chosen to favour germination, were 90% air relative humidity and 22°C air temperature, illumination by fluorescent tubes (VHO 215 W, Sylvania Cool White, Sylvania, USA), 20 h light and 4 h dark, 80 μmol m^-2 s^-1 PAR. On day 12 PAR irradiance was increased to 200 μmol m^-2 s^-1, now also including incandescent bulbs (60 W, Oy Airam, Finland). On day 18 irradiance was increased again to 350 μmol m^-2 s^-1 PAR from both fluorescent tubes and incandescent bulbs. From day 18 day/night air temperature and relative humidity were set to 22°C/17°C and 70%/80% respectively.

On day 21 the seedlings were inoculated by placing inoculum near lateral roots. Perlite was gently shaken from the upper part of the pot to expose the upper roots. One-third of the seedlings were inoculated with EMFs, another third with AMFs and one-third were not inoculated (NM seedlings). The EM inoculum was three 5mm x 5mm pieces of agar with mycelium from the growing edges of colonies grown on Hagem agar for three weeks. The AM fungi were cultured with Trifolium repens as host in a growth room, and the soil with the roots was subsequently dried at room temperature and homogenized with scissors and by mixing. The homogenized soils containing the two AMF species were mixed in a volume ratio 1:1 and 2 ml of the mix were added to the corresponding pots. To avoid differences in the supply of nutrients all pots received Hagem agar either sterile or with mycelium and AM soil inoculum either autoclaved or not depending on the treatments.

During germination and early growth the pots were watered with tap water, and from day 39 onwards with liquid fertilizer, initially with 20 mg l^-1 of nitrogen (N) and other nutrients in proportion (below) to the pots five days per week, and from day 60 onwards 30 mg l^-1 N, from day 68 onwards 40 mg l^-1 N, and from day 81 until day 93, the start of the water- withholding experiments, 50 mg l^-1 N. An excess of the solution was applied each time to ensure that the solution in the pots was largely exchanged. The nutrients were initially from Kekkilä Garden irrigation fertilizer (Kekkilä Oy, Vantaa, Finland) (16.6% N, 4.0% P, 25.3%

K, 0.03% B, 0.014% Cu, 0.18% Fe, 0.10% Mn, 0.001% Mo, 0.023% Zn). From day 66 to day 73, 12 mg of Mg as Mg(NO3)2·6H2O was added to the mix of water and fertilizer. On day 68 the fertilizer was substituted with a complete nutrient solution (Riddoch et al. 1991).

(8)

7 Starting 63 days since sowing, seedling height was measured once a week to the nearest millimetre with a ruler from the soil surface to the stem apex. Starting day 84, the diameter of the stem base (root collar) was measured with a digital calliper together with height. On day 81, the ten tallest and shortest seedlings from each mycorrhiza treatment were culled. Part of the retained seedlings were selected at random for another study (Kilpeläinen et al. 2019).

On day 93, the seedlings were fertilized for the last time, and water-withholding treatments started (dry day 0) simultaneously in two drought experiments that were run in parallel:

Experiment A designed to compare watered and non-watered plants and Experiment B to study the time course of drought until the death of the seedlings. Plants from the same cohort were assigned to experiments A and B at random. For Exp. A, groups (blocks) of six

seedlings, two from each mycorrhiza treatment, were formed based on the size of the plants:

two tallest from each mycorrhiza treatment in block 1, two second tallest in block 2, etc. In Exp. B there was no block arrangement because all plants had the same treatment (see below).

Experiment A: Comparison of drought-treated and watered plants

InExperiment A, there were 84 plants per mycorrhizal treatment, 252 altogether. Within each block, one seedling of each mycorrhiza treatment was allocated at random to the watered treatment and the other to the water-withholding treatment. The plants in the watered treatment received 20 ml deionized water each day to replace the water lost through evapotranspiration but avoiding leaching of nutrients through percolation.

Seven blocks (each including one seedling per mycorrhiza and watering treatment combination) were measured at each harvest. On days 2, 3, 4, 6 and 9 of drought, gas exchange, leaf chlorophyll and shoot water potential (Ψ) were measured. The dry mass of leaves, stems and roots were measured on days 3 and 4 and leaves additionally on day 2.

After this, some leaves were shed, and roots became difficult to separate from the soil.

Mycorrhizas were assessed on plants harvested on dry day 2. Other measurements were done on all harvests and seedlings except for gas-exchange in some seedlings on dry day 9 that had only brittle leaves left. Shoot water potential was additionally measured on dry day 10.

(9)

8 Exp A: Soil water, shoot water potential, gas exchange and chlorophyll

CO2 and H2O exchange of the youngest fully expanded leaf was measured with a portable instrument (LI-6400, LI-COR, Lincoln, NE, USA) using the LI-COR flat-leaf chamber and a red-blue LED light source. The area of the chamber opening was 6 cm², and equal to the measured area except for 11 plants with small leaves for which the enclosed leaf area was estimated visually. Air flow rate was set at 500 μmol s^-1, ambient air temperature at 24 °C, PAR at 300 µmol m^-2 s^-1, CO₂ concentration in the incoming air (C_a) at 450 µmol mol^-1, water vapour pressure at 2.2 kPa corresponding to relative humidity close to 70%, giving a leaf to air vapour pressure difference of 0.9 to 1 kPa depending on leaf temperature. After the leaf had been enclosed in the chamber for 1 min, and readings were stable, three

measurements were saved. Stomatal conductance (gs), net photosynthetic rate (A) and intercellular CO2 concentration (Ci) were calculated by the LI-6400. From these three data records, the one with the median value of stomatal conductance was retained for further data analysis.

Immediately after gas exchange measurement, the seedlings were excised at the base of the stem and Ψ was measured with a Scholander-type pressure chamber (Model 1000, PMS Instruments, Albany, OR, USA). If sap did not show on the cut surface at chamber pressure of 2.6 MPa the measurement was terminated, and the measurement recorded as -2.7 MPa.

Each day gas exchange and water potential were measured in all harvested seedlings within two hours after midday.

Chlorophyll was assessed with an SPAD-502 Plus meter (Konica Minolta Inc., Japan) after measuring Ψ. Three leaves were measured per sapling and the values averaged. The raw estimates were converted to chlorophyll per unit leaf area using the equation Chl [µmol m^-2]

= 10(M^0.26721)

where M is the instrument reading and the constant 0.26721 derived from calibration data for our SPAD unit (Randriamanana et al. 2012), based on the equation by Markwell et al. (1995).

Exp A: Mycorrhizal colonization and root tips

(10)

9 The pots from dry day 2 were stored at -18°C. They were thawed in batches consisting in one pot from each treatment combination. Roots were separated from perlite in water using forceps.

Two subsamples from each root system were taken for microscopy, one from depth of 0.5 cm to 3.0 cm and the other from depth of 4.0 cm to 6.5 cm. Each subsample had at least 150 root tips and 100 intersections. If subsamples contained fewer roots, the whole root system was divided into two. If it was still too small, the whole root system below the depth 0.5 cm was measured. The samples from EM seedlings were stained using Ponceau stain (Daughtridge et al. 1986). The samples from AM and NM seedlings were incubated in 10% K₂O overnight, followed by 20min in alkaline H₂O₂ solution, then 1% HCl for 2h at room temperature, then 1.5h in a lactic acid, glycerol and methyl blue mix preheated at 80ºC (Grace and Stribley 1991).

Each stained subsample was spread on a grid with a line spacing of 1.27 cm drawn below a 90 mm petri dish and observed under a stereo microscope. Root length was estimated from the number of root intersections with the gridlines. For AM analysis, intersecting roots were classified into mycorrhizal and non-mycorrhizal, and spores counted (Giovannetti and Mosse 1980). In EMF-inoculated seedlings, the EM tips with a mantle and tips with no mantle were counted.

Exp A: Plant growth, foliar nutrients and root tannins

For foliar nutrient analyses, leaves were pooled to have enough material. For dry days 2, 3, 4, 6, 9 and 10, two replicates per treatment and day were analyzed; one with pooled leaves of three seedlings and the second with four seedlings. Nitrogen concentration was measured by Kjeldahl‟s method (Halonen et al. 1983). For B, Ca, Cu, K, Mg, Mn, P, S and Zn, a MARS5 microwave wet digestion in HNO3 and H2O2 in Teflon containers (method based on Epa 3051) was done and then ICP-OES (Iris Intrepid II XSP, Thermo Elemental, Franklin, MA, USA) was used.

The tannin concentration of roots harvested on dry day 3 was measured with the protein- precipitating method according to Hagerman and Butler (1978), see Adamczyk et al. (2008).

This method measures both condensed and hydrolysable tannins precipitated by protein.

(11)

10 Following reaction with protein, the tannins are determined spectrophotometrically based on oxidation of OH group with FeCl3 and formation of iron-phenolate complex. Three laboratory replicates were measured for each root sample, and their means are reported.

Tannic acid (Sigma Chemicals) was used as a standard (Adamczyk et al. 2012), and the results are presented as mg of tannic acid equivalents per 1 g of dry root.

Experiment B: lethal drought

For Experiment B, 50 plants per mycorrhizal treatment, 150 altogether, were all subjected to the water-withholding treatment. The temperature of one young fully expanded leaf in each seedling was measured on dry days 2-7 between 14:00 and 15:00 using a portable infrared thermometer with a laser sight and macro-optics (Optris LS, Optris GmbH, Berlin, Germany).

The pots were weighed, and all leaves were counted and categorized as either alive or dead daily (except on dry days 8 and 14). When all the leaves had fallen, the plant was considered dead. Monitoring ended on day 20 of drought, when all the plants were dead. This was confirmed by re-watering the seedlings and inspecting them in the growth room during seven days; no new growth was observed.

Statistical analysis for Exp A and Exp B

The mycorrhiza colonization percent, root tip number per root length (dry day 2) and the root tannin concentrations (dry day 3) were measured in individual seedlings (replicates). These results were subjected to ANOVA by fitting linear-mixed-effects models with factorial combinations of mycorrhiza (M) and watering (W) treatments and block as a random factor in IBM SPSS Statistics 25 software. Holm‟s correction (1979) was applied for the post-hoc tests of mycorrhiza treatment.

All the other variables in Exp. A were measured on several days during the drought treatment but on different seedlings at each date. A linear mixed effects model was fitted to these data, with treatments M and W as explanatory factors, time since the last watering (days) as a covariate and blocks as the grouping factor for a random effect on the intercept. On dry day 9 gas exchange could not be measured in several of the seedlings in the dry treatment,

consequently, data for this day were not included in the analysis. No data transformations were applied as transformations alter the hypotheses tested in factorial designs (see Quinn

(12)

11 and Keough 2002, section 4.3.3). When variance is not homogeneous, the systematic change in variance is usually dependent on an explanatory variable. When this variable is known, this relationship can be incorporated into the model fitted so that parameter and probability

estimates remain valid. In the present case, when heterogeneity of variance was detected, we included the fitted values as the explanatory variable for the error variance, using a power of variance function to describe the shape of this dependency (Pinheiro and Bates 2000). Models were fitted using the nlme package 3.1-140 (Pinheiro et al. 2019) and plots drawn with

packages ggplot2 3.2.0 (Wickham 2016), ggpmisc 0.3.2 (Aphalo 2019) and patchwork (Pedersen 2017) under R 3.6.0 (R Core Team 2019). Contrasts were used to test differences among mycorrhiza treatments when the main effect or interactions involving this factor had P<0.10. Contrasts were fitted with package gmodels 2.18.1 (Warnes et al. 2018) and P-values adjusted using the correction method of Holm (1979).

For Exp. B, the percentages of living seedlings out of the total number of seedlings (50 per mycorrhiza treatment) were calculated for each measurement day. These data are binary, dead or alive, and were analyzed with a statistical procedure specific for survival data (Harrington and Fleming 1982) as implemented in package survival 2.44-1.1 (Therneau 2015).

The percentages of living leaves per plant at each measurement day were calculated with reference to the total number of leaves in the seedling on the day of last watering (day 0).

These as well as height, diameter and leaf temperature measurements were repeated through the experiment on the same seedlings. We fitted linear mixed-effects (LME) models to these data using the individual seedling as grouping factor for a random effect on the intercept. As for experiment A, no data transformations were applied. In those cases when variance was not homogeneous, the dependency of this heterogeneity on the fitted values was accounted for by the model that was fitted, as described above for Exp. A. Contrasts among levels of factor M were fitted as described above. The packages mentioned above were used under R 3.6.0. Additionally, survival time since the last watering was tested for M with one-way ANOVA in IBM SPSS Statistics 25 and the post-hoc test significances were adjusted with the method of Holm (1979). P-values from ANOVA and other a-priori tests, adjusted P-values from fitted contrasts, as well as comparisons among fitted models are given numerically when P<0.1.

(13)

12 Results

Exp A: Soil water, shoot water potential, gas exchange and chlorophyll

In the following, the probabilities for significance of the main effects and their interactions are denoted with subscripts for factors: T for time, W for watering and M for mycorrhiza.

The soil water content (% of dry mass) decreased rapidly in the dry treatment after the

watering ceased (P_T, P_W and P_TxW <0.001, Fig. 1a, Fig. S2). The mycorrhizal treatments dried in a different way in the watering treatments (P_M<0.001, P_WxM=0.001), and therefore the ANOVA was done separately for the two watering treatments. The AM pots were slightly drier in the watered treatment (PM=0.023, contrast EM to AM, Padj=0.035; AM to NM, Padj=0.039) and less dry in the dry treatment (PM=0.024, contrast EM to AM, Padj=0.033; AM to NM, Padj=0.028).

The Ψ was similar in all the watered plants, about 0.4 MPa (Fig. 1b) while it decreased with time in the dry treatment (PT, PW and PTxW <0.001). The decrease was different among the mycorrhiza treatments (PWxM=0.024, PTxWxM=0.001). Within the dry treatment Ψ decreased more slowly in the AM plants than others (PM=0.053, PTxM =0.001). The contrasts for difference between the mycorrhiza treatments against time showed Padj=0.005 for both AM vs EM and AM vs NM, and 0.717 for EM vs NM.

The gs decreased with time (PT and PW <0.001, Fig. 1c) especially in the dry treatment (P_TxW=0.079). As there was some indication for an interaction (P_WxM=0.077), the watering treatments were analyzed separately. Then the P_M=0.053 in the watered treatment as g_s was highest in AM plants; however, the contrasts had P_adj>0.116. In the dry treatment there was more variability, and no mycorrhiza effect was detected. The A declined with drought (data not shown) and the ANOVA result was similar as for gs except that MxW had a higher P=0.126.

The Ci/Ca ratio was lower in the dry than watered treatment on day 2 of drought, and higher on day 4 and after that (PTxW <0.001, Fig. 1d). However, the variability in Ci/Ca was high.

The relation between A and gs showed that during drought, the decline in A was steeper than gs and no difference among mycorrhiza treatments was detected (Fig. 2).

(14)

13 Chlorophyll concentration was about 320 µmol m^-2 in all treatments in the beginning of the drought. In the droughted plants it was reduced to about 290 and in the watered plants to about 260 µmol m^-2 by the end of the experiment (P_W=0.084, P_T<0.001 and P_WxT=0.036).

There was no indication of a mycorrhizal effect (data not shown in detail).

Exp A: Mycorrhizal colonization and root tips

The EM colonization rates were 65%-90% on dry day 2 (Table 1). The AM colonization rates were ≤65%; however, all seedlings except one had some arbuscules and spores. Three

seedlings had vesicles although the percentage of them in roots was below 0.6%. No

contamination of AMF and EMF was detected in NM seedlings or cross-contamination in the mycorrhizal treatments. No Frankia nodules were observed either. There was no watering effect and the treatments were pooled. The total number of root tips per root length was higher in EM plants compared to AM and NM, which did not differ from each other (Table 1).

Exp A: Plant growth, foliar nutrients and root tannins

Plant growth was affected by both M and W but there was no significant interaction. The total seedling mass was lower in the dry treatment than watered (PW<0.001), and lower in the AM treatment than EM and NM (PM<0.001, contrast to AM, Padj<0.001 for both EM and NM, which did not differ, Fig. 3a). The dry mass ratio of leaves was larger in the watered treatment (P_W<0.001), and larger in AM (P_M<0.001, contrasts AM to EM and NM P_adj=0.005 and 0.001, respectively, no difference was detected between EM and NM). The root mass ratio was larger in the dry treatment (P_W<0.001). It was largest in EM, NM was intermediate, and AM lowest (P_M<0.001, all contrasts P_adj<0.001) (Fig. 3b). Stem mass ratio was slightly lower in the dry treatment (PW<0.001).

The foliar nitrogen (N), phosphorus (P), potassium (K) and sulfur (S) concentrations decreased during the experiment (PT<0.001, Fig. 4). They were lower and reduced faster in the watered treatment and remained about the same in the dry treatment (PW<0.001,

PTxW≤0.012). The AM plants had highest concentrations of these nutrients (PM=0.002 for N, PM<0.001 for P, K and S). However, for N, the contrasts between the mycorrhiza treatments

(15)

14 did not show difference (Padj>0.10). For P, contrasts indicated a difference between AM and NM (Padj=0.005) and between EM and AM (Padj=0.023), but not between EM and NM. For K, the contrast between AM and NM had Padj=0.052 and AM and EM, Padj=0.078, while EM and NM did not differ. For K, the slope for EM plants in the drought treatment showed an increase with time unlike the other treatments (P_TxM =0.005). By contrast, the difference in S concentrations in favour of AM was larger in the dry treatment than watered (PWxM=0.044).

All the nutrient contents (total amount of N, P, K and S in the foliage of a seedling on dry days 2, 3 and 4) were lower in the dry treatment (P_W=0.015 for N and <0.001 for P, K and S), indicating that the dry mass of droughted plants decreased more than the concentrations decreased in the watered plants by this time (insets in Fig. 4). The N content was highest in NM (PM=0.057, contrast Padj=0.069 to both AM and EM) while AM and EM did not differ from each other. For P, there was no M effect. The K content was higher in AM than EM (PM=0.007, contrast AM vs EM Padj=0.043) while NM remained intermediate. The S content was lowest in EM (PM<0.001, contrast to NM Padj=0.029) and AM (Padj=0.001) while AM and NM did not differ.

Copper (Cu) and iron (Fe) concentrations showed a distinctive pattern (Fig. 5). In average the Cu concentrations decreased in time and were higher in the dry treatment, like most other nutrients (PT<0.001, PW<0.001 and PTxW <0.001). The Cu concentrations had a trend of decrease of AM concentrations and increase in EM with time in the dry treatment, although with low probabilities (PM =0.070, PTxWxM=0.094). The iron concentrations also varied with time (PT<0.001), and the drying affected the mycorrhiza treatments in a different way (P_M=0.004, P_WxM< 0.001 and P_TxWxM=0.092). Within the watered treatment, the AM plants had highest Fe concentrations (P_M<0.001, P_adj<0.001 for contrasts both AM vs EM and AM vs NM). Within the dry treatment the result was the opposite, as AM had the lowest

concentrations (P_M<0.001, contrasts P_adj<0.001 for both AM vs EM and AM vs NM). In EM the concentration increased in time within the dry treatment (PTxM=0.048 and the contrast for slope in time AM vs EM Padj=0.053).

The foliar contents of Cu and Fe during the first days were reduced by the dry treatment (PW<0.001, Fig. 5 insets). Copper content did not show a difference between mycorrhiza treatments. Iron contents in the watered treatment did not differ but in the dry treatment the

(16)

15 AM were lowest (PWxM<0.001, within dry, PM=0.008, and contrasts AM vs EM and AM vs NM Padj=0.013).

The B, Ca, Mg, Mn and Zn concentrations were higher in the dry treatment (P_W<0.001 except 0.028 for Zn, Fig. S1). The concentrations varied in time in different ways (P_T<0.001 except 0.049 for Zn). The Ca concentration remained the same with time in the watered treatment but increased in the dry treatment (PTxW=0.009). By contrast, Mg decreased in time in the watered treatment but remained about the same in the dry treatment (P_TxW<0.001). For Zn, the increase in time was steeper in the dry treatment (P_TxW<0.001). Calcium and Mn concentrations were highest in AM (P_M<0.001).

The dry treatment about doubled the tannin concentrations and the mycorrhiza effect was significant, but there was no interaction (PW<0.001, PM=0.018, Fig. 6). The largest tannin concentrations were in the EM roots and lowest in the AM roots (pairwise Padj=0.015). The NM treatment remained in the middle and did not show probability for difference from either.

Exp B: Height and diameter

The height and diameter were lowest in the AM plants during the whole experiment

(PM≤0.001). The difference in height was smallest in the beginning of the measurements, five weeks before the drought experiments PTxM≤0.001; contrast Padj<0.001 for the change in time AM vs NM and AM vs EM) (Fig. 7). The difference between AM and EM plus NM was about 7% in height and 4% in diameter during drought. The plants did not grow after the third day of drought.

Exp B: Leaf temperature and leaf loss

Substrate moisture decreased during Exp. B in the same way as in Exp. A, and it was similarly higher in the AM plants between days 4 to 8 of drought (P_M<0.001, Fig. S1). Leaf temperature increased nearly 2°C after two days without water, indicating partial stomatal closure (Fig. 8). It increased further to about 26.3°C on dry day 3. After this there was no change and the last measurement was on dry day 5.

(17)

16 The percent of surviving leaves out of the original number of leaves in each plant decreased in the same way in all mycorrhiza treatments in the beginning, but from day 9 of drought, the AM plants retained live leaves for longer than EM and NM (PM=0.007) (Fig. 9a).

Whole seedlings started dying on day 7, first in EM and NM, and after day 9 there was a steep decrease in remaining live plants, quickest in NM and slowest in AM. The survival analysis gave P=0.010 for the overall effect of mycorrhiza, and contrasts detected a

difference between AM and NM (P_adj=0.012) but not between EM and AM or NM and EM (Fig. 9b). The mean±SE lifetime in drought was 11.9±0.32 d in NM plants, 12.5±0.31 d in EM plants and 13.2±0.37 d in AM plants and when subjected to ANOVA and pairwise comparisons the only significant difference was the longer survival time in AM than NM (PM=0.015, pairwise Padj=0.014). The last living plants (assessed by retention of non-brittle leaves) were recorded 19 days after the last watering, hence on day 20, all plants were dead, which was confirmed by the observation of no growth in a week after rewatering.

Discussion

The arbuscular-mycorrhizal (AM) plants maintained higher shoot water potentials (Ψ) than ectomycorrhizal (EM) and non-mycorrhizal (NM) grey alder seedlings in the severe drought treatment (Experiment A). Together with the longer survival in the lethal-drought experiment (Exp. B) this indicates better drought resistance by AM plants. The stomatal conductance (gs) of AM plants was slightly higher than EM and NM when the soil was not dry; however, at the onset of the drought treatment, the AM plants closed their stomata at the same rate as the EM and NM plants. These responses suggest better adaptation of AM plants to varying soil water conditions, as the stomatal opening allows CO₂exchange when water is available, and during drought, stomatal closure can save the plant from long-term damage or death.

The AM plants were somewhat smaller than EM and NM, and this may have contributed to their better performance, as smaller plants may not dry their limited pot volume out as soon as larger ones. The largest difference in plant water potential on day 10 was observed when the soil moisture was already similarly low in all mycorrhizal treatments. Moreover, the differences in leaf and seedling survival were detected after this phase. Yet, it cannot be ruled out that the smaller size and slightly more remaining water may have been a factor in the

(18)

17 better performance of AM. Additionally, the spatial distribution of the remaining water and the roots and hyphae within the pots remains unknown.

The present experiments were short termed with a fast imposition of drought, and the growth ceased in all mycorrhiza treatments in a few days. It was not possible to get reliable results on growth and allocation at the later stages of the experiment; however, on days 3 and 4 of water withholding, we found more allocation to leaves in the AM plants, and more to roots in the EM plants in both watering treatments. This contrasted to the findings of earlier harvests from the same plant population in favourable conditions, where no difference in the growth allocation was detected (Kilpeläinen et al. 2019).

The leaves started falling after days 6-7 when the shoot water potential had decreased to about -1.8 MPa. Loss of leaves may be beneficial in extreme drought conditions, as it reduces the loss of water by cuticular transpiration, even if the stomata are closed. We could not determine root longevity, but it is likely to be an important aspect of drought resistance and recovery. Liese et al. (2019) found lower fine-root life span in seedlings of AM tree species than EM. Some plant species shed fine roots readily in dry soil; the advantage of this is that photosynthates are not needed for keeping them up (Espeleta et al. 1999). This is consistent with the lower root/soil respiration rates in AM than NM during drought in Citrus (Espeleta et al. 1999). However, if a plant can avoid premature leaf and root loss by saving water by earlier stomatal closure, it will be in better shape after the drought; here, the leaf shedding was delayed in AM plants compared to EM and NM plants.

The ratio between the intercellular and ambient CO₂ (C_i/C_a ratio) was slightly lower in the droughted plants in the first measurements on dry day 2, indicating that the stomata had started closing to conserve water, which resulted in depletion of CO₂ in the stomatal cavity.

Afterwards, on day 4, C_i/C_a increased in all droughted plants. The high C_i/C_a shows that there was no shortage of CO2 for assimilation during the latter dry days despite the stomatal

closure. This indicates declining function of the photosynthetic “machinery” for assimilating carbon (Lawlor and Tezara 2009).

The net photosynthetic rate (A) and stomatal conductance did not show differences among the mycorrhiza treatments before or during drought, similarly as in our previous work (Lehto 1992a, Kilpeläinen et al. 2019). AM colonization has been shown to increase the

(19)

18 belowground sink strength enough to trigger increased photosynthesis (Gavito et al. 2019), but this may be context dependent. Moreover, the relationship between photosynthesis and stomatal conductance remained similar in all mycorrhiza treatments through the drought experiment, and neither C_i/C_a nor the leaf chlorophyll concentrations differed among the mycorrhizal treatments. Hence, there was no evidence for a mycorrhizal effect on photosynthetic functions.

The foliar concentrations of several nutrients declined in both watering treatments, as nutrients were withheld also from the watered plants. As the droughted plants ceased their growth after the first days, their nutrient concentrations remained higher than in the watered plants. Drought generally reduces nutrient uptake because of reduced mobility of nutrients in drying soils, the reduced ability of roots or mycorrhizas to take them up, and on the longer term, the reduction of absorbing surface because of death of fine roots and mycorrhizas (Kreuzwieser and Gessler 2010). The magnitude of the effect varies among nutrients, and N accumulation tends to be less affected than P (Peuke and Rennenberg 2004, He and Dijkstra 2014).

AM plants had higher P concentrations in the beginning of the dry treatment, and they remained higher in both watering treatments. Also, K, S, Ca and Mn levels were highest in AM plants. This was not completely caused by the smaller size of AM plants, as the foliar contents (mg nutrient in the foliage of a seedling) of P, K and S were either higher or similar in the AM treatment compared to EM and NM. Sustained better P uptake from dry soil by AM roots than NM was shown by Neumann and George (2004). This could be explained by the mycorrhizal hyphae accessing P outside the depletion zone near the roots, in smaller soil pores and by continued contact with soil particles by hyphae (Smith et al. 2010). However, this same explanation applies also to EM roots and external mycelium, and sustained foliar Ca, K, Mg, Mn, P and S concentrations were shown in EM poplar while the concentrations decreased in NM during drought (Danielsen and Polle 2014). It is possible that P uptake from organic sources is more efficient by EM (e.g. Dickie et al. 2014), but in a tracer study, no EM effect was found on phosphate uptake in dry conditions (Reid and Bowen 1979). Here, nutrients were supplied in soluble forms, and the substrate did not contain organic matter. In similar conditions, there was little difference in the N uptake of AM and EM grey alder, although the EM decreased the δ¹⁵N in the plants more consistently (Schweiger 2016). The contribution of N2 fixing by Frankia was excluded for simplicity, however, in further studies

(20)

19 on alders the role of Frankia and its interactions with mycorrhizas needs to be included.

Increased sulfur uptake both by AM and EM compared to NM has been shown (Rennenberg 1999, Allen and Shachar-Hill 2009), but here the S concentrations in AM were higher than NM and EM in both watering treatments.

By contrast to the results on N, P and S, the K, Fe and Cu concentrations increased in the EM plants in the dry treatment towards the end of the experiment. Additionally, Fe was highest in AM plants in the watered treatment and lowest in the dry treatment. These results suggest that while the AM sustained the non-metal nutrient uptake from mineral forms during drought, the EM maintained the metal uptake. On the other hand, EMF are known to accumulate metals (Berthelsen et al. 1995), and it is also possible that the severe drought caused a release of some metals from the fungal partner, although any mechanism for such release and transport remains unknown. More research is needed on nutrient uptake from different sources in dry soils, especially as global change is likely to bring more frequent and severe drought episodes in regions which are already affected by P deficiency (IPCC 2014).

Here, the foliar nutrient concentrations did not show deficiency (comparing to data on birch in Reinikainen et al. 1998), but with the decreases during the experiment, they were

approaching suboptimal levels at least for N, P and S. Therefore, the differences between the mycorrhiza treatments were likely to affect the performance of the plants. The sustained P uptake probably helped the AM plants maintain their hydraulic conductance, which is reduced by low P supply (Radin and Eidenbock 1984, Andersen et al. 1989, Coleman et al.

1990) and contributed to the higher water potential and slower loss of leaves in AM plants.

Potassium functions in osmotic adjustment and stomatal control (Nieves-Cordones et al.

2019). Also, a good N status is needed for maintaining favourable water relations (Radin and Boyer 1982). The roles of other nutrients in drought resistance are less well known.

The AM root systems had considerably lower tannin concentrations than EM in both watered and droughted plants. Earlier studies on tannins in plants have been done almost exclusively in foliage, but in Lotus japonicus, NM roots had more condensed tannins than AM, while in shoots the result was the reverse (Solaiman and Senoo 2018). Here, the tannin accumulation did not appear to increase drought resistance, on the contrary, the AM plants with least root tannins were somewhat more drought resistant than EM and NM. Tannins have been suggested to have a functional role in drought resistance, as they can act as antioxidants in

(21)

20 trees (Gourlay and Constabel 2019). In some studies there has been drought-induced tannin accumulation in tree leaves (Zhang et al. 2012, McKiernan et al. 2016, Top et al. 2017), but not always (McKiernan et al. 2017), and the observed drought effects on tannins have not been as strong as here in the alder roots. Generally, herbaceous plants tend to have low tannin concentrations compared to trees (Gourlay and Constabel 2019), and they are also typically AM. Tannin accumulation was one of the clearest differences between AM and EM, and the consequences of this in the stress resistance of mycorrhizal fungi and their hosts need more attention.

To conclude, the AM plants had somewhat better control over their water use than EM and NM as shown by higher shoot water potential and slower wilting during severe drought.

Sustained P uptake from mineral sources is a possible explanation to the better performance of the AM plants. Although the study was done on a limited number of host-fungus

combinations, and the AM plants were smaller, this is the first study comparing the drought resistance of EM, AM and NM plants. Our results support the general hypothesis on a better adaptation of AM to warm and dry climates and concomitant advantage for the host plants (Lehto and Zwiazek 2011). The results complement the theory of “Mycorrhizas in

ecosystems” (Read 1991, Smith and Read 2008) indicating that P availability has a major role in the predominance of AM. Now we showed that this applies during severe drought

conditions, which are less typical of the biomes where EM dominate. At the same time, there were novel results on enhanced uptake of also S by AM, contrasting with enhanced Cu, Fe and K uptake by EM during drought, which deserve further attention. If the tannin

accumulation is a more general stress reaction in EM roots than AM roots, it may be asked, if tannins are more useful in cold resistance than drought resistance (Rummukainen et al. 2007).

No doubt the results may vary depending on the species and genotype of the study plants and fungi, as always in inoculation experiments, and therefore the generalization of the results requires further studies on other species combinations. Study questions for future include differential performance, particularly nutrient uptake by AM and EM plants in different soil conditions and under different environmental stresses, including slower induction of drought and cyclic drought, waterlogging and adverse temperature regimes.

Data and materials availability

(22)

21 The data will be deposited in Zenodo.

Supplementary data

Supplementary figures S1 and S2.

Conflict of Interest

The authors declare no conflict of interest.

Funding

Funding was provided by the Academy of Finland, decision numbers 268279, 311455 and 324648.

Acknowledgements

We thank Dr Thorunn Helgason for identifying the EM ascomycete, and Leena Kuusisto and Maini Mononen for skilful assistance in running the experiment and for nutrient analyses.

Author contribution

J.K., P.J.A. and T.L. planned and supervised the study. A.B-L. and S.A.N. implemented most of experiments A and B, respectively. B.A. analyzed and interpreted the tannin results. P.J.A.

was responsible for the statistics. All authors took part in writing the manuscript.

References

Adamczyk B, Kitunen V, Smolander, A (2008) Protein precipitation by tannins in soil organic horizon and vegetation in relation to tree species. Biol Fert Soils 45:55–64.

(23)

22 Adamczyk B, Salminen J-P, Smolander A, Kitunen V (2012) Precipitation of proteins by tannins: effects of concentration, protein/tannin ratio and pH. Int J Food Sci Tech 47:875–878.

Allen JW, Shachar-Hill Y (2009) Sulfur transfer through an arbuscular mycorrhiza. Plant Physiol 149:549–560.

Andersen C, Sucoff E, Dixon R, Markhart A (1989) Effects of phosphorus deficiency on root hydraulic conductivity in Fraxinus pennsylvanica. Can J Bot 67:472–476.

Aphalo PJ (2019) ggpmisc: Miscellaneous extensions to 'ggplot2'. https://cran.r- project.org/package=ggpmisc (21 August 2019, date last accessed).

Augé RM (2001) Water relations, drought and vesicular-arbuscular mycorrhizal symbiosis.

Mycorrhiza 11:3–42.

Augé RM, Toler HD, Saxton AM (2015) Arbuscular mycorrhizal symbiosis alters stomatal conductance of host plants more under drought than under amply watered conditions: a meta-analysis. Mycorrhiza 25:13–24.

Berthelsen BO, Olsen RA, Steinnes E (1995) Ectomycorrhizal heavy metal accumulation as a contributing factor to heavy metal levels in organic surface soils. Science Total Environ 170:141–149.

Coleman M, Bledsoe C, Smit B (1990) Root hydraulic conductivity and xylem sap levels of zeatin riboside and abscisic-acid in ectomycorrhizal Douglas fir seedlings. New Phytol 115:275–284.

Danielsen L, Polle A (2014) Poplar nutrition under drought as affected by ectomycorrhizal colonization. Environ Exp Botany 108:89–98.

Daughtridge A, Boese S, Pallardy S, Garrett H (1986) A rapid staining technique for assessment of ectomycorrhizal infection of oak roots. Can J Bot 64:1101–1103.

Dickie IA, Koele N, Blum JD, Gleason JD, McGlone MS (2014) Mycorrhizas in changing ecosystems. Botany 92:149–160.

Espeleta JF, Eissenstat DM, Graham JH (1999) Citrus root responses to localized drying soil:

A new approach to studying mycorrhizal effects on the roots of mature trees. Plant Soil 206:1–10.

Gavito ME, Jakobsen I, Mikkelsen TN, Mora F (2019). Direct evidence for modulation of photosynthesis by an arbuscular mycorrhiza-induced carbon sink strength. New Phytol 223:896–907.

Giovannetti M, Mosse B (1980) Evaluation of techniques for measuring vesicular arbuscular mycorrhizal infection in roots. New Phytol 84:489–500.

(24)

23 Gourlay G, Constabel CP (2019) Condensed tannins are inducible antioxidants and protect

hybrid poplar against oxidative stress. Tree Physiol 39:345–355.

Grace C, Stribley DP (1991) A safer procedure for routine staining of vesicular-arbuscular mycorrhizal fungi. Mycol Res 95:1160–1162.

Hagerman A, Butler L (1978) Protein precipitation method for the quantitative determination of tannins. J Agr Food Chem 26:809–812.

Hagerman AE (2012) Fifty years of polyphenol–protein complexes. In: Cheynier V, Sarni‐

Manchado P, Quideau S (eds) Recent advances in polyphenol research, volume 3. John Wiley & Sons, Hoboken, NJ, USA, pp. 71–97.

Halonen O, Tulkki H, Derome J (1983) Nutrient analysis methods. Finnish Forest Research Institute Research Papers 121:1–28.

Harrington DP, Fleming TR 1982) A class of rank test procedures for censored survival data.

Biometrika 69:553–566.

He M, Dijkstra FA (2014) Drought effect on plant nitrogen and phosphorus: a meta-analysis.

New Phytol 204:924–931.

Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6:65–70.

IPCC (2014) Core Writing Team, Pachauri RK, Meyer LA, eds. Climate Change 2014:

Synthesis Report. Contribution of working groups I, II and III to the fifth assessment report of the Intergovernmental Panel on Climate Change. IPCC, Geneva, Switzerland, 155 pp. https://www.ipcc.ch/report/ar5/syr/ (21 August 2019, date last accessed).

Kilpeläinen J, Vestberg M, Repo T, Lehto T (2016) Arbuscular and ectomycorrhizal root colonisation and plant nutrition in soils exposed to freezing temperatures. Soil Biol Biochem 99:85–93.

Kilpeläinen J, Barbero-López A, Vestberg M, Heiskanen J, Lehto T (2017) Does severe soil drought have after-effects on arbuscular and ectomycorrhizal root colonisation and plant nutrition? Plant Soil 418:377–386.

Kilpeläinen J, Barbero-López A, Adamczyk B, AphaloPJ, Lehto T (2019) Morphological and ecophysiological root and leaf traits in ectomycorrhizal, arbuscular-mycorrhizal and non-mycorrhizal Alnus incana seedlings. Plant Soil 436:283–297.

Korhonen A, Lehto T, Heinonen J, Repo T (2019) Whole-plant frost hardiness of mycorrhizal (Hebeloma sp. or Suillus luteus) and non-mycorrhizal Scots pine seedlings. Tree Physiol 39:951–960.

Kreuzwieser J, Gessler A (2010) Global climate change and tree nutrition: influence of water availability. Tree Physiol 30:1221–1234.

(25)

24 Laiho, O. (1970). Paxillus involutus as a mycorrhizal symbiont of forest trees. Acta Forestalia

Fennica 106:1–73.

Lawlor DW, Tezara W (2009) Causes of decreased photosynthetic rate and metabolic

capacity in water-deficient leaf cells: a critical evaluation of mechanisms and integration of processes. Ann Bot-London 103:561–579.

Lehto T, Zwiazek JJ (2011) Ectomycorrhizas and water relations of trees: a review.

Mycorrhiza 21:71–90.

Lehto, T (1992a) Mycorrhizas and drought resistance of Picea sitchensis (Bong.) Carr. II. In conditions of adequate nutrition. New Phytol 122:669–673.

Lehto T (1992b) Effect of drought on Picea sitchensis seedlings inoculated with mycorrhizal fungi. Scand J For Res 7:177–182.

Liese R, Leuschner C, Meier, IC (2019) The effect of drought and season on root life span in temperate arbuscular mycorrhizal and ectomycorrhizal tree species. J Ecol 107:2226–

2239.

Liese R, Lübbe T, Albers NW, Meier IC (2018). The mycorrhizal type governs root exudation and nitrogen uptake of temperate tree species. Tree Physiol 38:83–95.

Markwell J, Osterman JC, Mitchell JL (1995) Calibration of the Minolta SPAD-502 leaf chlorophyll meter. Photosynth Res 46:467–472.

McKiernan AB, Potts BM, Brodribb TJ, Hovenden MJ, Davies NW, McAdam SAM, Ross JJ, Rodemann T, O'Reilly-Wapstra JM (2016) Responses to mild water deficit and rewatering differ among secondary metabolites but are similar among provenances within Eucalyptus species. Tree Physiol 36:133–147.

McKiernan AB, Potts BM, Hovenden MJ, Brodribb TJ, Davies NW, Rodemann T, McAdam SAM, O'Reilly-Wapstra JM (2017) A water availability gradient reveals the deficit level required to affect traits in potted juvenile Eucalyptus globulus. Ann Bot-London

119:1043–1052.

Neumann E, George E (2004) Colonisation with the arbuscular mycorrhizal fungus Glomus mosseae (Nicol. & Gerd.) enhanced phosphorus uptake from dry soil in Sorghum bicolor (L). Plant Soil 261:245–255.

Nieves-Cordones M, Ródenas R, Lara, A, Martinez M, Rubio F (2019) The combination of K+ deficiency with other environmental stresses: What is the outcome? Physiol

Plantarum 165:264–276.

Pedersen TL (2017) patchwork: The composer of ggplots. R package version 0.0.1.

https://github.com/thomasp85/patchwork (21 August 2019, date last accessed).

(26)

25 Peuke A, Rennenberg H (2004) Carbon, nitrogen, phosphorus, and sulphur concentration and

partitioning in beech ecotypes (Fagus sylvatica L.): phosphorus most affected by drought. Trees-Struct Funct 18:639–648.

Pfabel C, Eckhardt KU, Baum C, Struck C, Frey P (2012) Impact of ectomycorrhizal colonization and rust infection on the secondary metabolism of poplar (Populus trichocarpa x deltoides). Tree Physiol 32:1357–64.

Pinheiro JC, Bates DM (2000). Mixed-effects models in S and S-plus. Springer, New York, NY, USA.

Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team (2019) nlme: Linear and nonlinear mixed effects models. R package version 3.1-140. https://cran.r-

project.org/package=nlme (21 August 2019, date last accessed).

Quinn GP, Keough MJ (2002) Experimental design and data analysis for biologists.

Cambridge University Press, Cambridge, UK.

R Core Team (2019) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ (21 August 2019, date last accessed).

Radin J, Boyer J (1982) Control of leaf expansion by nitrogen nutrition in sunflower plants - role of hydraulic conductivity and turgor. Plant Physiol 69:771–775.

Radin J, Eidenbock M (1984) Hydraulic conductance as a factor limiting leaf expansion of phosphorus-deficient cotton plants. Plant Physiol 75:372–377.

Randriamanana T, Wang F, Lehto T, Aphalo PJ (2012) Water use strategies of three

Malagasy Adansonia species under drought. South African Journal of Botany 81:61–70.

Read DJ (1991) Mycorrhizas in ecosystems. Experientia 47:376–391.

Reid CPP, Bowen GD (1979) Effect of water stress on phosphorus uptake by mycorrhizas of Pinus radiata. New Phytol 83:103–107.

Reinikainen A, Veijalainen H, Nousiainen H (1998) Puiden ravinnepuutokset -

metsänkasvattajan ravinneopas. Finnish Forest Research Institute Research Papers 688:1–44.

Rennenberg H (1999) The significance of ectomycorrhizal fungi for sulfur nutrition of trees.

Plant Soil 215:115–122.

Riddoch I, Lehto T, Grace J (1991) Photosynthesis of tropical tree seedlings in relation to light and nutrient supply. New Phytol 119:137–147.

(27)

26 Rummukainen A, Julkunen-Tiitto R, Räisänen M, Lehto T (2007) Phenolic compounds in

Norway spruce as affected by boron nutrition at the end of the growing season. Plant Soil 292:13–23.

Schweiger PF (2016). Nitrogen isotope fractionation during N uptake via arbuscular mycorrhizal and ectomycorrhizal fungi into grey alder. J Plant Physiol 205:84–92.

Smith SE, Read DJ (2008) Mycorrhizal symbiosis. 3rd edition. Academic Press, Cambridge, MA, USA.

Smith SE, Facelli E, Pope S, Smith FA (2010) Plant performance in stressful environments:

interpreting new and established knowledge of the roles of arbuscular mycorrhizas. Plant Soil 326:3–20.

Smolander A, Kanerva S, Adamczyk B, Kitunen V (2012) Nitrogen transformations in boreal forest soils - does composition of plant secondary compounds give any explanations?

Plant Soil 350:1–26.

Solaiman ZM, Senoo K (2018) Arbuscular mycorrhizal fungus causes increased condensed tannins concentrations in shoots but decreased in roots of Lotus japonicus L.

Rhizosphere 5:32–37.

Therneau TM (2015) A package for survival analysis in S. version 2.38. https://cran.R- project.org/package=survival (21 August 2019, date last accessed).

Top SM, Preston CM, Dukes JS, Tharayil N (2017) Climate influences the content and chemical composition of foliar tannins in green and senesced tissues of Quercus rubra.

Front Plant Sci 8:article 423.

Warnes GR, Bolker B, Lumley T, Johnson RJ (2018) gmodels: Various R programming tools for model fitting. R package version 2.18.1. https://cran.r-project.org/package=gmodels (21 August 2019, date last accessed).

Wickham H (2016) ggplot2: Elegant graphics for data analysis. Springer-Verlag, New York, NY, USA.

Zhang LH, Shao HB, Ye GF, Lin YM (2012) Effects of fertilization and drought stress on tannin biosynthesis of Casuarina equisetifolia seedlings branchlets. Acta Physiol Plant 34:1639–1649.