Specialized fungal parasites reduce fitness of their lichen hosts

Abstract

Background and aims

Understanding to what extent parasites affect host fitness is a focus of research on ecological interactions. Fungal parasites usually affect the functions of vascular plants. However, parasitic interactions comprising effects of fungal parasites on the fitness of lichen hosts are less well known. This study assesses the effects of the abundance of two highly specialized gall-forming fungi on growth of their two respective lichen hosts and tests whether these fungal parasites reduce lichen fitness.

Methods

The relative biomass and thallus area growth rates, and change in specific thallus mass of Lobaria pulmonaria and L. scrobiculata were compared between lichens with and without galls of the lichenicolous fungi Plectocarpon lichenum and P. scrobiculatae, cultivated in a growth chamber for 14 d. By estimating the thallus area occupied by the galls, it was also assessed whether growth rates varied with effective photosynthetic lichen surface area.

Key results

Plectocarpon galls significantly reduced relative growth rates of the lichen hosts. Growth rates decreased with increasing cover of parasitic galls. The presence of Plectocarpon-galls per se, not the reduced photosynthetic thallus surface due to gall induction, reduced relative growth rates in infected hosts. Specific thallus mass in the hosts changed in species-specific ways.

Conclusions

This study shows that specialized fungal parasites can reduce lichen fitness by reducing their growth rates. Higher parasite fitness correlated with lower host fitness, supporting the view that these associations are antagonistic. By reducing hosts’ growth rates, these parasites in their symptomatic life stage may affect important lichen functions. This fungal parasite–lichen study widens the knowledge on the ecological effects of parasitism on autotrophic hosts and expands our understanding of parasitic interactions across overlooked taxonomic groups.

Fungal parasites, gall-forming fungi, host fitness, lichenicolous fungi, Lobaria, parasitic interactions, Plectocarpon, relative growth rate, specific thallus mass

INTRODUCTION

Parasitic symbiosis is a major ecological interaction affecting individual fitness, driving population and community dynamics (Gilbert, 2002; Thomas et al., 2005; Preston and Johnson, 2010). Host–parasite relationships are often stories of winners and losers: the parasite benefits by diverting host resources for its own needs, reducing the host fitness and/or lifespan (Price, 1980; Ewald, 1994). The degree of virulence and specialization of the parasite for its host often determines the outcome of the relationship (Ewald, 1994; Leggett et al., 2013). Specialized and low-virulence parasites usually keep their hosts alive to secure their resource supply, often resulting in a reduced host fitness (Price, 1980; Legget et al., 2013, and references therein). However, symbiosis in nature is often a dynamic parasitic–mutualistic continuum that varies in time and depends on external factors and the life cycle phase of the organism (e.g. Sapp, 2004; Vasiliauskas et al., 2007; Martos et al., 2009; Zook, 2015), meaning that symbionts also sometimes can provide their host with significant services (see e.g. Spribille et al., 2016). Thus, to understand how parasites affect host fitness has been a focus of research in ecological interactions.

Many host–parasite interactions in vascular plants are ecologically well studied. Fungi are among the main parasites in plants (Burdon, 1987; Isaac, 1992; Gilbert, 2002) and can adversely affect various host functions and thus influence plant population dynamics. For instance, host-specific, low-virulence fungal parasites may reduce plant fecundity by altering the host reproductive systems (Clay, 1991) and plant growth by damaging leaves and roots (Isaac, 1992; Agrios, 2005). Interestingly, the outcome of symbiotic plant–fungal interactions may change across space and time; mycorrhizal fungi can act as parasite or a mutualistic partner depending on the resource availability for the plant (Sapp, 2004; Johnson, 2010). Whether such effects on host fitness occur in other fungal–autotrophic host interactions remain little explored [see reviews by Lawrey and Diederich (2003), and Davey and Currah (2006) for fungi–lichen and fungi–moss interactions, respectively]. Studies of overlooked autotrophic hosts will widen our current knowledge on the ecological effects of parasitism.

Lichens are symbiotic organisms comprising one species-specific heterotrophic fungus, one or more autotrophic partners (photobionts) and often a cortical Basidiomycete yeast (Spribille et al., 2016). Lichens serve as essential fodder for specialized herbivores, pathogenic micro-organisms and lichenicolous fungi (reviewed by Asplund and Wardle, 2017). The more than 1800 known lichenicolous fungi are usually considered as parasites. They induce species-specific symptoms on the lichens depending on their host specialization and virulence (Hawksworth, 1982; Lawrey and Diederich, 2003, 2016). Many gall-forming lichenicolous fungi are highly host-specific and low-virulence parasites (Hawksworth, 1982; Lawrey and Diederich, 2003). Yet, there is a need to explore how host-specific parasites influence the fitness of their lichen hosts. In plants, fungal parasites often reduce growth rates of their hosts, not only by acquiring energy and other resources (Ayres, 1981; Isaac, 1992; Agrios, 2005), but also by, for example, affecting the host’s water economy (Ayres, 1978; Agrios, 2005). In lichens, there is also a need to understand how their fungal parasites interfere with host growth. Lichen growth comprises (1) areal extension requiring full hydration to expand fungal hyphae, and (2) biomass gain per thallus area depending on C-fixation rates (Palmqvist, 2000; Larsson et al., 2012). The balance between these two processes shapes the specific thallus mass (STM = biomass/area), an important driver of water-holding capacity (WHC) (Gauslaa and Coxson, 2011; Merinero et al., 2014; Esseen et al., 2015).

Here, we studied the host–parasite system comprising the epiphytic lichens Lobaria pulmonaria and L. scrobiculata (Lobariaceae, Ascomycota), colonized by the highly specialized gall-forming lichenicolous fungi Plectocarpon lichenum and P. scrobiculatae (Roccellaceae, Ascomycota), respectively. Plectocarpon mainly reproduces sexually in basally constricted galls on the lichen surface (Ertz et al., 2005). To what extent Plectocarpon occurs without inducing visible galls (i.e. asymptomatically) in their hosts is unknown. However, recent molecular studies have revealed that its presence is undetectable in several thalli without galls and at >1 cm from the galls in infected specimens, suggesting that Plectocarpon is localized to the gall and nearby host tissue (S. Merinero, L. Plaza and M. Prieto, unpubl. data). According to Ertz et al. (2005), Plectocarpon spp. are considered commensalistic lichenicolous fungi because they visibly modify only local parts of lichen thalli by gall induction (Ertz et al., 2005). The life cycle and nutritional strategy of Plectocarpon is unknown, but Grube and de los Ríos (2001) suggested that P. lichenum is a mycoparasite because it only causes visible necrosis in the mycobiont. Additionally, the presence of Plectocarpon galls has been linked to reduced carbon-based secondary compound concentration (Merinero et al., 2015a; Asplund et al., 2016) deterring grazers and/or screening excess solar radiation (Solhaug and Gauslaa, 2012), suggesting that Plectocarpon in its symptomatic life stage affects lichen fitness. Thus, these host–parasite systems serve as a model to test the effects of highly host-specific fungal parasites on host fitness. The main objective of the present study was to compare the relative growth rates in thallus area and biomass and change in STM of L. pulmonaria and L. scrobiculata with and without Plectocarpon galls. Consistent with a trade-off in host–parasite fitness, we hypothesized that the presence and abundance of Plectocarpon galls reduce the lichen fitness by reducing area and/or biomass growth rates in both lichen hosts.

MATERIAL AND METHODS

Lichen material

Two widespread N-fixing foliose lichens were selected: Lobaria pulmonaria (L.) Hoffm. and L. scrobiculata (Scop.) P. Gaertn. The cephalolichen L. pulmonaria contains the green alga Dictyochloropsis reticulata as the main photobiont and the cyanobacterium Nostoc sp. localized in scattered internal cephalodia, while the cyanolichen L. scrobiculata has Nostoc sp. as its only photobiont. The lichenicolous fungus Plectocarpon lichenum (Sommerf.) D.Hawksw. inhabits L. pulmonaria and two Pseudocyphellaria species, while P. scrobiculatae Diederich and Etayo exclusively occurs on L. scrobiculata (Ertz et al., 2005). Plectocarpon induces the development of basally constricted galls mainly on the lichen upper surface (Ertz et al., 2005). Galls consist mostly of mycobiont and parasite hyphae (Grube and de los Ríos, 2001;,Ertz et al., 2005). The parasitic reproducing structures (ascomata) develop on top of each gall and form sexual (ascospores) and asexual (conidia) propagules (Ertz et al., 2005). The nomenclature of the species follows the MycoBank (www.mycobank.org; Crous et al., 2004).

The lichens were collected in two Mediterranean oak (Quercus pyrenaica Willd.) forests in Central Spain in March 2015. Lobaria pulmonaria with and without P. lichenum galls was collected in Montejo de la Sierra, Madrid (41°06′44″N, 03°29′33″W; 1263 m a.s.l.; mean annual precipitation: 818 mm; mean temperature 9.5 °C). Lobaria scrobiculata with and without P. scrobiculatae galls was collected in a drier and warmer southern location in Hontanar, Toledo (39°29′27″N, 04°36′33″W; 797 m a.s.l.; mean annual precipitation: 638 mm; mean annual temperature: 13.6 °C; climate data extracted from CLIMOEST; Sánchez-Palomares et al., 1999). For each thallus with Plectocarpon galls, one without galls was sampled from the same tree. Lichen specimens with similar size and none or a few reproductive structures were selected. We gathered 26 and 25 pairs of L. pulmonaria and L. scrobiculata thalli with and without Plectocarpon galls, respectively, cleaned them from debris and kept them air-dried for 48 h before the experiment.

Before and after cultivation, each thallus was saturated by spraying de-ionized water, their upper surface was scanned with a digital scanner (Epson Perfection 4870) and the total lichen surface area (A) and the thallus area occupied by Plectocarpon galls (A_galls) were estimated with Adobe Photoshop 6.0 software (Adobe Systems, San Jose, CA, USA). Thalli were then air-dried for 48 h and air-dry mass was recorded (±0.1 mg). Ten extra control thalli of each species and parasitic-galls category (yes/no) were weighed; then oven-dried 72 h at 65 °C before recording dry mass (DM). The DM of cultivated thalli was computed using the mass reduction factor of the sacrificed control thalli. A_start (mean ± 1 s.e.) of L. pulmonaria and L. scrobiculata thalli without Plectocarpon galls was 9.94 ± 0.44 and 7.07 ± 0.75 cm², respectively, and 9.58 ± 0.52 and 7.19 ± 0.88 cm² for L. pulmonaria and L. scrobiculata with Plectocarpon galls, respectively. The effective photosynthetic thallus surface area for all thalli, A_effective = A_start − A_galls, was calculated, hereafter referred to as effective photosynthetic thallus surface.

Experimental design

Lichens were cultivated in a SELECTA Hot-Cold GL 2101507 growth chamber (Grupo Selecta, Barcelona, Spain) for 14 d by slightly modifying the protocol of Bidussi et al. (2013). We used the temperature regime providing the fastest growth rates in both species (21/16 °C day/night; 14-h daily photoperiod at 120 μmol photons m⁻² s⁻¹ provided by Osram HotCold 60 W F48T12/0W/HO fluorescent lamps. Thalli were cultivated in 26 open square Petri dishes (four thalli; one of each parasite category and species in each dish) on top of ten layers of filter paper. During cultivation, lichens and filter paper were kept moist by spraying with sterilized and de-ionized water twice (in the morning and at night). The amount of added water was adjusted for each species to keep them equally well hydrated until the end of each photoperiod. At the end of each day, there was a short period (3–4 h) with at least partial drying of all thalli before the next spraying at night. During this period, thalli were transferred to dry filter papers to ensure daily drying (Bidussi et al., 2013). Within each species, thalli with and without Plectocarpon galls received the same amount of water. Preliminary experiments were run to adjust the amount of added water enabling all specimens to remain hydrated until the end of each photoperiod. Because light varied slightly across and between shelves, Petri dishes were daily rotated and interchanged among the chamber shelves.

Growth rates, STM and chlorophyll fluorescence measurements

Relative thallus area growth rate, RT_AGR = (ln(A_end/A_start)) × 100/Δt (mm² cm⁻² d⁻¹) and relative biomass growth rate, RGR = ((ln(DM_end/DM_start)) × 1000/Δt (mg g⁻¹ d⁻¹), were quantified following Evans (1972). Δt was 14 d. STM (= DM/A) was calculated at the start and the end of the experiment. Percentage changes in STM were calculated as ΔSTM = 100 × (STM_end − STM_start)/STM_start. After the experiment, P. lichenum galls were carefully removed from hydrated L. pulmonaria thalli by cutting at the base of the galls with a razor and they were then weighed after being oven-dried for 72 h at 65 °C. New DM_end values were then recorded for infected L. pulmonaria thalli with galls removed. Assuming that these galls had not changed their biomass during the experiment, DM_start of the host only was computed by subtracting the gall mass. The contribution of the Plectocarpon gall mass to RGR, STM and change in STM in L. pulmonaria could thus be estimated. This was not done for L. scrobiculata thalli because these were used afterwards in another experiment (Asplund et al., 2016). We removed from the dataset two L. scrobiculata thalli with P. scrobiculatae galls that for unknown reasons had strongly reduced mass due to fragmentation and deterioration.

Chlorophyll a fluorescence was recorded before and after the experiment in all thalli after 1 h of hydration at low light intensity at the end of the last night period (10 h). The maximum photochemical efficiency of photosystem II (F_V/F_M) was measured with a portable Handy PEA chlorophyll fluorometer (Hansatech Instruments Ltd, Kings Lynn, UK) after 15 min of dark adaptation.

Statistical analyses

Two-way analyses of variance (ANOVA) were used to analyse differences in growth rates between thalli with and without Plectocarpon galls, including the binary factors lichen species, presence of Plectocarpon galls and the interaction between the two. To identify significant differences in growth rates between parasite and lichen species, a post hoc Tukey-type multiple comparison test was used (package lsmeans; Lenth, 2013). Analysis of covariance (ANCOVA) was used to disentangle the effects of the presence of Plectocarpon galls from the possibly reduced effective thallus surface in infected specimens. Here, growth rate was used as a dependent variable, parasite (presence or absence of Plectocarpon-galls) as a factor, effective thallus surface as a covariate, and a parasite × effective thallus surface interaction for each growth rate and lichen species. A significant interaction term indicates that the presence of the parasite modifies the slope of the relationship between lichen growth rates and effective thallus surface. Linear models were used to test the relationship between growth rates of infected thalli and the percentage cover of Plectocarpon galls (based on A_galls), which was log-transformed to improve model fitting. ANOVAs were run to detect the effect of gall mass on growth rates, STM and change in STM in L. pulmonaria. In all models, residuals were checked for normality and homocedasticity. Analyses were run in R 3.1.0. (R Foundation for Statistical Computing, Vienna, Austria). Means ± 1s.e. are given in the text and the significance value set at P < 0.05.

RESULTS

The Plectocarpon infection significantly reduced growth rates in biomass (RGR) and area (RT_AGR) of their lichen hosts, although the magnitude of the effects differed between lichen species (Table 1; Fig. 1A, B). RGR in L. pulmonaria was twice as high in specimens without P. lichenum galls (16.68 ± 0.04 % biomass gain) as in infected thalli (8.48 ± 0.56 % biomass gain; Fig. 1A). This contrast was stronger in L. scrobiculata with six times higher RGR in thalli without P. scrobiculatae galls than in infected ones (14.95 ± 1.53 vs 2.4 ± 1.2 % biomass gain; Fig. 1A). Also, RT_AGR was lower in infected than in non-visibly infected thalli for both host species (Fig. 1B). Lobaria pulmonaria without galls had four times higher RT_AGR than thalli with galls (8.98 ± 0.61 vs 2.16 ± 0.44 % area gain, respectively; Fig. 1B), whereas L. scrobiculata without galls had two times higher RT_AGR than thalli with galls (6.97 ± 0.62 and 4.11 ± 0.48 % area gain, respectively; Fig. 1B). Thus, the difference in RT_AGR between thalli with and without galls was significantly larger for L. pulmonaria than for L. scrobiculata (Fig. 1B), and area gain across species was not strictly coupled to biomass gain.

Table 1.

ANOVA results for relative growth rates in biomass (RGR), area (RT_AGR), specific thallus mass at the start of the experiment (STM) and change in STM (ΔSTM) in the lichens Lobaria pulmonaria and Lobaria scrobiculata with and without galls of the fungal parasites Plectocarpon lichenum and P. scrobiculatae, respectively, cultivated for 14 d in a growth chamber

Source	d.f.	RGR		RTAGR		STMstart		ΔSTM
		F	P	F	P	F	P	F	P
Species (S)	1	16.71	<0.001	0.02	0.884	1.95	0.166	15.99	<0.001
Parasite (P)	1	119.1	<0.001	83.63	<0.001	57.88	<0.001	30.89	0.001
S × P	1	5.94	0.016	13.60	<0.001	8.50	0.004	22.61	0.002
Residuals	96
Total	99
r2adj		0.58		0.49		0.40		0.40

Values in bold are significantly different at P < 0.05.

Fig. 1.

Mean (± 1 s.e.) relative growth rates in biomass [RGR (A)], area [RT_AGR (B)], specific thallus mass at the start [STM (C)] and change in STM (D) in the lichens Lobaria pulmonaria and Lobaria scrobiculata with and without Plectocarpon galls cultivated for 14 d in a growth chamber. Different letters indicate significant differences at the P < 0.05 level, according to two-way ANOVA (Table 1) and post hoc Tukey test.

At the start, thalli with Plectocarpon galls had significantly higher STM, and thallus thickness was highest in infected L. pulmonaria (Fig. 1C). Lobaria pulmonaria with and without P. lichenum galls became thicker after cultivation (Fig. 1D). By contrast, the discrepancy in A and DM growth between L. scrobiculata with and without P. scrobiculatae galls resulted in strong differences in thickness (Table 1; Fig. 1D). While L. scrobiculata without galls became thicker, thalli with parasitic galls experienced a reduction in thickness (Fig. 1D).

The effective photosynthetic thallus surface area (A_effective) did not differ either between specimens with and without Plectocarpon galls in L. pulmonaria (ANOVA, F_1,50 = 3.32, P = 0.074), or in L. scrobiculata (ANOVA, F_{1, 46} = 0.43, P = 0.514). The ANCOVA data showed that the presence of Plectocarpon galls was associated with strongly lower growth rates in infected thalli (Table 2). Irrespective of the presence of galls, thalli of L. pulmonaria with larger A_effective increased their growth, whereas this variable did not significantly influence any growth rate in L. scrobiculata (Table 2). The parasite × A_effective interaction was only significant for RT_AGR in L. scrobiculata because thalli without galls decreased their growth with increasing A_effective (i.e. A_start) (linear regression, estimate = −0.02, F_1,23 = 4.61, P = 0.042, r²_adj = 0.13), but RT_AGR did not significantly vary with A_effective in infected thalli (linear regression, F_1,21 = 0.61, P = 0.445).

Table 2.

ANCOVA summary testing the effects of the factor presence of Plectocarpon galls (parasite) and the covariate photosynthetic effective thallus surface area at start (A_e) and their interaction on relative growth rates in biomass (RGR), area (RT_AGR) and change in specific thallus mass (ΔSTM) in the lichens Lobaria pulmonaria and Lobaria scrobiculata cultivated for 14 d in a growth chamber

Source	d.f.	RGR			RTAGR			ΔSTM
		Coeff (± s.e.)	F	P	Coeff (± s.e.)	F	P	Coeff (± s.e.)	F	P
Lobaria pulmonaria
Parasite (P)	1	−4.34 (2.51)	53.60	<0.001	−0.53 (0.21)	73.44	<0.001	1.41 (2.53)	0.13	0.72
Areaeffective (Ae)	1	0.55 (0.19)	38.76	<0.001	0.01 (0.02)	15.39	<0.001	0.63 (0.19)	18.85	<0.001
P × Ae	1	−0.02 (0.26)	0.005	0.942	0.01 (0.02)	0.007	0.638	−0.17 (0.26)	0.44	0.508
Residuals	48
Total	51
r2adj			0.63			0.63			0.24
Lobaria scrobiculata
Source
Parasite (P)	1	−8.02 (2.18)	60.41	<0.001	−0.39 (0.11)	15.25	<0.001	−6.04 (3.49)	29.23	<0.001
Areaeffective (Ae)	1	0.04 (0.19)	0.001	0.330	−0.02 (0.01)	0.69	0.409	0.38 (0.31)	1.66	0.205
P × Ae	1	−0.01 (0.28)	0.001	0.972	0.03 (0.01)	4.39	0.042	−0.44 (0.46)	0.94	0.337
Residuals	44
Total	47
r2adj			0.55			0.27			0.38

Values in bold are significantly different at P < 0.05.

The mean cover of Plectocarpon galls (a surrogate of parasite fitness) in was 8.76 ± 1.14 % in L. pulmonaria and 4.99 ± 0.94 % in L. scrobiculata. Increasing cover of Plectocarpon galls significantly reduced RGR in L. pulmonaria, but not in L. scrobiculata having significantly lower mean RGR (Fig. 2A). By contrast, increasing cover of Plectocarpon galls significantly decreased RT_AGR in both species (Fig. 2B). Thus, higher abundance of parasitic galls correlated with lower lichen growth rates. Increasing cover of Plectocarpon galls was significantly associated with increasing STM_start in both lichen species (Fig. 2C), although the increase in STM with increasing cover of galls was only significant in L. scrobiculata (Fig. 2D). No changes were detected in the cover of Plectocarpon galls or the occurrence of new Plectocarpon galls during the experiment (data not shown).

Fig. 2.

Regression plots of relationships between the cover of Plectocarpon galls (log-transformed) and relative growth rates in biomass [RGR (A)], area [RT_AGR (B)], specific thallus mass at the start of the experiment [STM (c)] and change in STM (D) of infected Lobaria pulmonaria and Lobaria scrobiculata. Significant regression equations: L. pulmonaria: RGR = 8.47 – 3.15 log (cover galls), r² = 0.25, P = 0.009; RT_AGR = 0.33 – 0.21 log (cover galls), r² = 0.16, P = 0.043; STM_start = 10.1 + 6.3 log (cover galls), r² = 0.50, P < 0.001. L. scrobiculata: RT_AGR = 0.4 – 0.22 log (cover galls), r² = 0.36, P = 0.002; STM_start = 12.43 + 2.71 log (cover galls), r² = 0.28, P = 0.009, STM_change = −2.77 + 2.24 log (cover galls), r² = 0.23, P = 0.021.

Removing the P. lichenum galls after the experiment showed that there were no differences in STM_start between infected and non-visibly infected L. pulmonaria specimens (Fig. 3A; ANOVA, F_1,50 = 3.96, P = 0.052, r² = 0.05). This after-experiment manipulation also resulted in a significantly higher RGR in infected thalli with galls removed than in thalli hosting P. lichenum galls, yet lower than in thalli without galls (Fig. 3B; ANOVA, F_2,79 = 28.58, P < 0.001, r² = 0.42), and a higher increase in STM in thalli with galls removed (Fig. 3C; ANOVA, F_2,79 = 7.88, P < 0.001, r² = 0.17). F_V/F_M between thalli with and without galls of each species did not differ at any stage of the experimental period (data not shown).

Fig. 3.

Differences in (A) specific thallus mass at the start of the experiment (STM), (B) relative biomass growth rate (RGR) and (C) change in STM in Lobaria pulmonaria without Plectocarpon lichenum galls (white), infected L. pulmonaria with P. lichenum galls (black) and infected L. pulmonaria with P. lichenum galls removed after the growth experiment (grey). Different letters indicate significant differences at the P < 0.05 level, according to one-way ANOVA and post hoc Tukey test.

DISCUSSION

This is the first study reporting direct, adverse effects of highly host-specific lichenicolous fungi on lichen growth rates and thus on the primary metabolism of their hosts. Consistent with our first hypothesis, the presence and abundance of parasitic Plectocarpon galls significantly reduced relative growth rates of their hosts L. pulmonaria and L. scrobiculata. Infected thalli of both lichens exhibited lower biomass and area gain than thalli without galls. These negative effects of lichenicolous fungi on their hosts are consistent with adverse effects of fungal parasites on vascular plants (Burdon, 1987; Isaac, 1992; Gilbert, 2002).

Fungal parasites destroying photosynthetic parts of host plants reduce their photosynthetic output, often resulting in declining plant growth rates (Agrios, 2005, chapter 3). There was no support for the second hypothesis predicting that a Plectocarpon gall-induced reduction of the photosynthetic lichen surface area should cause decreasing RGR in thalli with galls. Instead, the effective photosynthetic thallus surface area did not differ between gall-free and infected thalli in either of the two lichen species. Therefore, the effective thallus surface area could not explain the reduced growth rates in infected thalli. In fact, reduced RGRs were associated with the presence of Plectocarpon galls per se (Table 2). Only in L. scrobiculata, RT_AGR of thalli without galls decreased with increasing effective thallus surface (Table 2), but this is probably a size-dependent effect because small thalli grow faster than larger thalli (e.g. Gauslaa et al., 2009; Merinero et al., 2015b). Consistent with parasitic fungal–plant interactions (Isaac, 1992; Agrios, 2005), Plectocarpon presumably draws C and other nutrients from their hosts. Because P. lichenum is probably a mycoparasite (Grube and de los Ríos, 2001), we hypothesize that Plectocarpon benefits from C and/or nutrients taken from the mycobiont.

Consistent with a trade-off in host–parasite fitness (Price, 1980; Agrios, 2005), our results indicate a trade-off between lichen growth rates and fitness of the parasite. Increasing cover of parasitic galls led to decreasing area growth rates in both species, and decreasing biomass investment in L. pulmonaria (Fig. 2), consistent with recorded patterns in plant–parasitic fungi relationships (e.g. Ayres, 1981; Isaac, 1992; Agrios, 2005). Likewise, Asplund et al. (2016) documented a trade-off between defence compound investments in L. pulmonaria and Plectocarpon fitness, leading to increased grazing damage in thalli with galls.

No gall area expansion was detected during the cultivation experiment and it was assumed that gall mass had not changed during the experiment, so STM and RGR were re-calculated for infected L. pulmonaria with galls removed. However, the change in computed RGR by removal of the Plectocarpon galls in L. pulmonaria (Fig. 3B) suggests that the galls had increased in density or in thickness. The significant correlation between the cover of galls and STM at the start (Fig. 2C) suggests a contribution of the galls themselves to total mass per thallus area, as confirmed by the gall-removal manipulation (Fig. 3A).

The two lichen species had similar RGRs in the absence of Plectocarpon galls, whereas the magnitude of the gall-induced depression of host RGR depended on lichen species. Galls strongly reduced area expansion in L. pulmonaria and biomass gain in L. scrobiculata (Fig. 1). This discrepancy in growth responses suggests parasite-specific effects on host lichens. This is consistent with Asplund et al. (2016) documenting species-specific effects of Plectocarpon galls on the stoichiometry and palatability on their two respective lichen hosts: In L. pulmonaria, infected thalli had lower C and C: N, whereas these stoichiometric patterns did not vary between non-visibly infected and infected L. scrobiculata having substantially higher N concentrations than L. pulmonaria. The lower C: N ratio in infected L. pulmonaria (Asplund et al., 2016) may have contributed to its lower RGR (Fig. 1A).

STM in L. pulmonaria increased similarly in thalli with and without Plectocarpon galls (Fig. 1D), meaning that both infected and non-visibly infected thalli increased in thickness. By contrast, the strong reduction in RGR in L. scrobiculata with Plectocarpon galls led to a decline in STM (Fig. 1D), which would imply a lower WHC according to Gauslaa and Coxson (2011) and Merinero et al. (2014). Reductions in STM and the strongly correlated WHC would lead to shorter active photosynthetic periods (see Gauslaa and Solhaug, 1998). Reduced water storage is detrimental for cyanolichens such as L. scrobiculata depending exclusively on liquid water to activate photosynthesis (Lange et al., 1986). They rely on a relatively high WHC to prolong their photosynthetic activity (Gauslaa et al., 2012) to compensate for their photosynthetic dependence on rainy periods occurring more infrequently than episodes of humid air that activate photosynthesis in lichens with green algae as main photobionts (Lange et al., 1986).

Recent research on arbuscular mycorrhiza dynamics indicates that resource availability shapes the parasitic or mutualistic character of the mycorrhiza (Sapp, 2004; Johnson, 2010). Our results indicate that Plectocarpon in its symptomatic life stage and under favourable conditions for lichen growth acts as a parasite that reduces lichen growth. Therefore, this association is hardly a commensalistic relationship (Ertz et al., 2005). However, further investigations replicating this study under different levels of resources will shed light on possible mutualistic–parasitic continuum dynamics of lichenicolous fungi. Plectocarpon species seem to be more frequent in humid and optimal habitats for Lobaria species (Ertz et al., 2005), in which the lichen species may afford losing some of their resources to their parasite. However, because fungal parasites have the potential to alter the dynamics of host population and local communities (Burdon, 1987; Gilbert, 2002; Thomas et al., 2005), Plectocarpon could also affect their hosts’ populations and thus the lichen community.

In conclusion, this study shows for the first time that highly host-specific fungal parasites can reduce the fitness of their lichen hosts. Lichen growth declined with the presence and with increasing cover of parasitic galls per se, and not because of reduced photosynthetic thallus surface due to gall induction. By reducing lichen growth rates, the parasite could also adversely affect important lichen size-dependent fitness components. The physiological mechanisms underlying these highly specialized parasitic interactions deserve further investigation. Our results are consistent with ecological effects of parasite–host interactions documented in vascular plants. By focusing on overlooked parasitic interactions, this study extends the knowledge on adverse effects of fungal parasitism on autotrophic organisms.

ACKNOWLEDGEMENTS

This work was supported by a grant from the Spanish Association of Terrestrial Ecology (AEET) to S.M. We thank J. L. Margalet for technical assistance with the growth chamber and the staff in the protected area ‘Hayedo de Montejo’ (owned by Ayuntamiento de Montejo de la Sierra and managed by Comunidad de Madrid) for sampling permissions. R. Belinchón, J. Ehrlén and two anonymous reviewers provided insightful comments that helped to improve the manuscript. S.M. held a postdoctoral fellowship from the Regional Government of Madrid (Remedinal 3-CM: S2013/MAE-2719).