Photoreceptor-mediated kin recognition in plants
Summary
- Although cooperative interactions among kin have been established in a variety of biological systems, their occurrence in plants remains controversial.
- Plants of Arabidopsis thaliana were grown in rows of either a single or multiple accessions.
- Plants recognized kin neighbours and horizontally reoriented leaf growth, a response not observed when plants were grown with nonkin. Plant kin recognition involved the perception of the vertical red/far-red light and blue light profiles. Disruption of the light profiles, mutations at the PHYTOCHROME B, CRYPTOCHROME 1 or 2, or PHOTOTROPIN 1 or 2 photoreceptor genes or mutations at the TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS1 gene required for auxin (growth hormone) synthesis impaired the response. The leaf-position response increases plant self-shading, decreases mutual shading between neighbours and increases fitness.
- Light signals from neighbours are known to shape a more competitive plant body. Here we show that photosensory receptors mediate cooperative rather than competitive interactions among kin neighbours by reducing the competition for local pools of resources.
Introduction
Preferential helping of relatives has been observed for a wide range of taxa. For instance, in vertebrate (bird, mammal) species, helpers preferentially aid closer relatives during breeding (Griffin & West, 2003). In the social amoebae Dictyostelium discoideum, cells cooperate preferentially with relatives and aggregate to form multicellular fruiting bodies (Hirose et al., 2011). In humans, as the cost of helping increases, the share of help given to kin increases, whereas that given to nonkin decreases (Stewart-Williams, 2007). These differential responses require mechanisms to discriminate between kin and nonkin, but these mechanisms are often difficult to establish (Mehlis et al., 2008) and some have only recently been elucidated. In Dictyostelium discoideum, kin recognition depends on the presence of matching pair of alleles of two genes involved in cell–cell adhesion (Hirose et al., 2011). In humans, sibling detection involves monitoring subsystems specialised for registering maternal perinatal association with the individual's biological mother, and duration of sibling coresidence (Lieberman et al., 2007).
The occurrence of preferential cooperation among plant kin has become controversial, particularly after the pioneer studies of Dudley and collaborators. The reduced root growth of plants of Cakile edentula grown with sibling compared to strangers (Dudley & File, 2007) has been interpreted as the result either of chemical signals leading to cooperative interactions or of enhanced competition (Dudley & File, 2008; Klemens, 2008; Murphy & Dudley, 2009; Biedrzycki & Bais, 2010; Biedrzycki et al., 2010; Masclaux et al., 2010; Willis et al., 2010; Bhatt et al., 2011; Fang et al., 2013). It is difficult to investigate the occurrence of cooperative interactions among kin by analysing quantitative changes in growth, because kinship can also reinforce the competition for local pools of resources available for growth (Milla et al., 2009, 2012; Masclaux et al., 2010; Willis et al., 2010; File et al., 2012b; Lepik et al., 2012; Dudley et al., 2013). Artemisia tridentata plants exposed to volatile cues from clipped close relatives experience less leaf herbivory over the growing season than those that received cues from clipped neighbours that were more distantly related (Karban et al., 2013). Mycorrhizal partners become more abundant when Ambrosia artemisiifolia is grown with siblings compared to strangers (File et al., 2012a).
Due to selective light absorption by photosynthetic pigments, plant canopy shade-light exhibits low red to far-red ratios (R : FR) and low blue-light irradiances (Smith, 2000; Casal, 2013). In sparse canopies, the R : FR of the radiation propagating horizontally can be reduced by selective far-red light reflection on neighbour foliage even before mutual shading among plants is established (Ballaré et al., 1987; Smith et al., 1990). As the canopy closes, horizontally propagating blue light is also reduced before the horizontally placed leaves are severely shaded (Ballaré et al., 1991). The responses to low R : FR and low blue-light signals include increased growth of stems and/or petioles, and a redirection of leaf growth vertically and/or horizontally (Casal, 2013). These responses, called ‘shade-avoidance reactions’ increase the competitive ability of plants (Schmitt, 1997; Schmitt et al., 1999, 2003; Smith, 2000; Weinig, 2000), which place their leaves at better light strata within the canopy at the expense of overtopped neighbours that become shaded. The changes in R : FR and blue light caused by neighbours are (respectively) perceived mainly by phytochrome B (phyB) and cryptochrome 1 (cry1) (Sellaro et al., 2010; Keller et al., 2011). Blue light can also be perceived by phototropin 1 (phot1) and phot2 (Christie, 2007) but their role in shade-avoidance responses is not unequivocally documented.
Because photosensory receptors are able to perceive subtle signals generated by neighbours (Ballaré et al., 1987; Smith et al., 1990) and reorient growth of aerial organs (Maddonni et al., 2003; Christie, 2007), we investigated whether they are involved in kin recognition and cooperative interactions. Single plants of Arabidopsis thaliana were grown in pots (to minimise belowground interference) forming rows of kin or nonkin neighbours.
Materials and Methods
Growth conditions
Plants of Arabidopsis thaliana (L.) Heynh were grown in single rows of seven cylindrical pots (10-mm diameter, 74-mm height), containing one seedling each and secured by means of Styrofoam supports, in a glasshouse (Supporting Information Fig. S1).
Experiments and plant material
In a first experiment (Fig. 1), leaf-position ratio was measured in seedlings of the accessions Columbia, Landsberg erecta, Nossen, Cvi, SHA, BAY and RLD grown in rows of a single genotype (pure rows of kin) or rows of mixed genotypes (where different accessions were included at random without repetitions within each row).
In a second experiment (Fig. 2), leaf-position ratio was measured in seedlings of the accession Columbia at different times after their arrangement in rows. Leaf-position ratio was simultaneously measured in controls grown in similar pots but isolated from neighbours.
In a third experiment (Fig. 3) seedlings of the accessions Columbia, Landsberg erecta, Nossen, Cvi, SHA, BAY and RLD were grown in rows of a single genotype (pure rows of kin) and used for the measurements of light profiles.
In a fourth experiment (Fig. 4), leaf-position ratio was measured in seedlings of the Columbia wild-type and of the phyB-9 (Reed et al., 1993), cry1 (hy4-B104) (Bruggeman et al., 1996), cry2-1 (Guo et al., 1998), phot1-5 (Liscum & Briggs, 1995; Huala et al., 1997), phot2-1 (Kagawa et al., 2001), cry1 (hy4-b104) cry2-1 (Buchovsky et al., 2008), phot1-5 phot2-1 (Liscum & Briggs, 1995; Kagawa et al., 2001), ft-10 or ft-10 phyB-9 (Strasser et al., 2010) mutants (all in the Columbia background) grown in pure rows of kin. Leaf-position ratio was also measured in seedlings of the Columbia wild-type grown in pure rows placed under either green (Lee filters number 089) or yellow (Lee filters number 101) plastic filters with lateral ventilation to distort either the normal R : FR signals (sunlight, 1.09; under filter, 0.24) and/or the blue-light signals.
In a fifth experiment (Fig. 5), seedlings of the Columbia wild-type were grown isolated (i.e. without nearby neighbours) and a green plastic filter (1.5-cm tall, 2-cm wide) was placed vertically, North–South-oriented, 1 cm from the centre of the plant either towards the East or the West side of the plant, to simulate the signals of the horizontally propagating light generated by a neighbour. A clear plastic filter was used as control. The number of leaves growing towards the East, West, North or South was recorded 10 d later.
In a sixth experiment (Fig. 6), leaf-position ratio was measured in seedlings of the Landsberg erecta wild-type grown in rows formed either by coetaneous or noncoetaneous (age difference of 7 d) seedlings for 3 wk. The youngest plants had two pairs of leaves, the same as the plants used for the second and third experiments.
In a seventh experiment (Fig. 7), leaf-position ratio, the degree of mutual and self-shading, and seed yield were measured in seedlings of the Columbia wild-type and of the sav3 mutant (Tao et al., 2008) grown in pure rows of kin. Isolated plants were also included to evaluate whether the mutation affected fitness independently of the presence of controls. In a variant of this experiment (conducted separately), leaf-position ratio and seed yield were measured in seedlings grown in alternate rows of sav3 and Columbia seedlings (where each genotype was grown at alternate positions within the same row).
In an eighth experiment (Fig. 8), fitness was evaluated in seedlings of the accessions Columbia, Landsberg erecta, Nossen, Cvi, SHA, and RLD grown either in pure rows of kin or in alternate rows of two accessions (all possible combinations).
Leaf-position ratio
The number of leaves (> 3-mm length) within or outside the rows was recorded in 3-wk-old seedlings. Plants within the rows were observed from above (view perpendicular to the soil) through a cross drawn on a clear acetate sheet and centred on each seedling to define the sectors within or outside the rows (Figs 1, S1b).
Light profiles
The vertical profile of the R : FR and blue light was measured with a SpectroSense2 meter (Skye Instruments Ltd, Powys, UK) for six plants per accession. After removing the cosine corrector, the terminals of the red, far-red or blue channel sensors were placed facing the plant to be analysed (15 mm apart), at different heights. For Columbia, horizontal radiation coming from each cardinal point was also characterised with the sensor facing East, West, North or South.
Self-shading and mutual plant shading
The plants were photographed from above with a digital camera (A590 IS, 8MP, ×4; Canon, Tokyo, Japan) first within the row and then isolated from neighbours. Finally, to measure plant leaf lamina area, all the leaves were separated from the plant and photographed. The petioles were erased by using Photoshop (Adobe Systems Inc., San José, CA, USA) software before measuring lamina area (Fig. S1d,e). The degree of self-shading was measured by erasing from the total leaf lamina area of a plant the area exposed by this plant photographed isolated from neighbours before dissection (Fig. S1d). The degree of shading by neighbours was measured by erasing from the total leaf lamina area of a plant the area exposed by this plant photographed within the row (i.e. the area not shaded by neighbours) (Fig. S1e). Leaf shading data are expressed as a proportion of the total leaf lamina area. Total area was unaffected by the sav3 mutation (Columbia = 368 ± 30 mm2, sav3 = 416 ± 18 mm2). A plant was considered heavily shaded by neighbours when the latter covered > 20% of its lamina area because the distribution of frequency of mutual shading intensities was biphasic, with peaks separated by a minimum frequency at 20% mutual shading.
Fitness
In order to evaluate fitness, seeds were harvested separately per plant and stored 7 d at room temperature before weighing.
Results
Kin recognition
In order to investigate separately the occurrence of kin recognition and its fitness consequences, we searched for outputs of kin recognition involving no immediate reductions in growth (which can be interpreted as the result of exacerbated competition for local pools of resources). We cultivated seedlings of Arabidopsis thaliana forming a single-row canopy and recorded leaf position. Plants grown with kin showed a high ratio between the number of leaves positioned away from the neighbours and the number of leaves placed towards the neighbours in the row (Fig. 1a,b). In plants grown with nonkin neighbours (random mixture of accessions within the row) leaf position did not significantly deviate from random (leaf-position ratio = 1, Fig. 1a).
The leaf-position ratio changed rapidly in response to kin neighbours (Fig. 2a). To investigate the nature of the response we recorded the position of target leaves at different time points. The leaf-position ratio of the cohort present at time 0 changed from the original value of 1.1 (random disposition) to 1.4 (SE = 0.1, P < 0.01) 5 d later. Therefore, the change in leaf position was caused by a reorientation of leaf growth towards the empty space out of the row of plants (Fig. 2b) and not by changes in phylotaxis. During leaf transition to its new position, the petiole often showed bending and/or torsion leading to a vertical position of the lamina (Fig. 2c).
Mechanism of kin recognition
Because one seedling was grown in each pot, in current experiments kin recognition was likely based on features of the aerial environment. The presence of neighbours within the row generated a radially asymmetric light environment, where both R : FR and blue light were reduced towards the neighbours compared to open places (Fig. 3a). The vertical profile of R : FR and of blue light was recorded for the radiation propagating horizontally within the rows formed by different accessions. The minimum value of R : FR and of blue light correlated with the height of more dense foliage of the neighbour and therefore differed between individuals with different body shape (Fig. 3b).
In order to test the role of R : FR signals in kin recognition, we used the phyB mutant because phyB is the main photoreceptor of R : FR. phyB failed to rearrange its leaves in response to kin neighbours (Fig. 4a). Because the phyB mutant flowers early and this could affect the leaf position ratio, we also analysed the effect of the phyB mutation in the late-flowering ft background, which overrides the flowering phenotype of phyB. The results confirmed the failure of the phyB mutant to accommodate its leaves in response to kin neighbours (Fig. 4b). This response was also impaired when wild-type kin were grown under selective filters that distort the R : FR (Fig. 4a).
In order to test the role of blue-light signals in kin recognition, we used cry1, cry2, cry1 cry2, phot1, phot2, and phot1 phot2 mutants. All of these mutants showed reduced leaf-position responses to kin neighbours, with the strongest phenotype corresponding to cry1 cry2 (Fig. 4a,c). This response was also impaired when wild-type kin were grown under selective filters that distort the blue-light environment (Fig. 4a).
The aforementioned results suggest that the leaf-position response is caused by neighbour signals of the horizontally propagating radiation, and signals of insufficient magnitude reach the leaves when the neighbour is not a kin. To test this possibility, plants were grown isolated from nearby neighbours with a green plastic filter placed vertically either towards the East or towards the West side of the plant, simulating the presence of a neighbour. This filter produced strong neighbour signals by reducing the R : FR (from 0.94 ± 0.03 in the clear control filter to 0.40 ± 0.01) and blue irradiance (54 ± 9% of the clear filter control) of the horizontally propagating irradiation (measured at 4 mm height) coming from the side where it was placed. Leaves grew away from the plastic filter that simulated the presence of a neighbour (Fig. 5). A clear plastic filter, used as control, did not affect the direction of leaf growth (Fig. 5).
We compared the leaf position responses in rows formed by coetaneous or noncoetaneous neighbours of the same genotype. The rows formed by alternate plants with an age difference of 7 d failed to show the leaf-position response (Fig. 6). Neither the eldest nor the youngest plants rearranged their leaves away from the row of neighbours (Fig. 6). Because plants of different age have different stature, these results reinforce the idea that kin recognition requires matching body shapes. Furthermore, they underscore the importance of horizontally propagating radiation because the smaller (younger) plants were shaded by their taller neighbours from above and yet did not rearrange their leaves.
The promotion of petiole growth and the vertical reorientation of leaf growth in response to low R : FR require the growth hormone auxin (Tao et al., 2008; Kozuka et al., 2010). The horizontal leaf-growth response reported here could also require auxin. To test this idea we used the sav3 mutant, affected in the TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS1 (TAA1) gene involved in the synthesis of auxin in response to low R : FR (Tao et al., 2008). sav3 failed to reorient leaf growth in the presence of kin neighbours (pure sav3 rows, Fig. 7a,b).
Fitness consequences: kin recognition provides mutual benefit
We investigated whether recognizing a kin provides any advantage when the plant grows with kin neighbours. We took advantage of the fact that sav3 mutant seedlings are morphologically indistinguishable from their wild-type, with the exception of its impaired leaf-position response. We compared pure rows of either wild-type plants (that respond to their kin) or sav3 plants (that do not respond to their kin). The occurrence of leaf-position responses towards kin neighbours showed fitness consequences that can be disaggregated into positive and negative components. First, the re-orientation of leaf-growth direction increased the degree of leaf self-shading (Fig. 7c). The leaf illustrated in Fig. 2(b) departed from its original position facing the neighbour but ended up partially shaded by a leaf of the same plant. Therefore, the leaf-position response involves a cost (Pignocchi et al., 2006; West et al., 2007a,b). Second, the leaf-position response also involves a benefit because it reduced the degree of mutual shading among plants (Fig. 7c). Actually, the leaf-position response strongly reduced the probability of becoming severely shaded by neighbours (proportion of plants with > 20% of the leaf area under shade: Wild-type = 0.05, sav3 = 0.55, P < 0.01). In the present system, cost (increased self-shading) and benefit (reduced mutual shading) can be measured in the same currency. The decrease in mutual shading outweighed the increase in self-shading (Fig. 7c) and therefore the leaf-position response involves a net benefit for the actor within its lifespan.
Shading can compromise fitness by reducing the radiation available for photosynthesis and to test this possibility, plants were allowed to grow to maturity and produce seeds. When grown with kin neighbours, seed yield was significantly higher in the wild-type than in the sav3 mutant (Fig. 7d). This provides unequivocal evidence of adaptive kin discrimination (Biernaskie, 2011). Plants of the sav3 mutant showed wild-type seed yield when grown isolated from neighbours (Fig. 7d), indicating that the fitness consequences of the sav3 mutation are expressed only when grown with kin. Therefore, the cooperation among Arabidopsis kin plants is a case of mutual benefit (Pignocchi et al., 2006; West et al., 2007a,b).
Fitness consequences of cheating kin neighbours
We also investigated whether when mixed with wild-type seedlings, the sav3 mutant behaves as a ‘selfish cheater’ (West et al., 2007a,b). For this purpose we evaluated the behaviour of the wild-type and the sav3 mutant in mixed rows formed by alternate seedlings of each one of the two genotypes. The wild-type placed the leaves out of the position of sav3 neighbours, while sav3 did not show the leaf-position response (Leaf-position ratio: wild-type = 1.30 ± 0.08; sav3 = 1.09 ± 0.05, P < 0.05). By the end of the experiments with mixed rows, sav3 showed higher seed yield per plant than the wild-type (seed weight, mg, mean ± SE, wild-type = 20 ± 1, sav3 = 28 ± 2, P < 0.05).
Fitness consequences: Is it better to grow with a kin?
We also investigated whether kin recognition results in kin selection; that is, whether for a plant it is better to grow with a kin or a nonkin. We cultivated plants of different accessions either in pure rows (i.e. with kin neighbours) or as alternate mixtures of two accessions within the row (i.e. nonkin neighbours) in all possible pair combinations (Fig. 8a). We recorded seed yield per plant and to eliminate the intrinsic differences among accessions, for each accession we normalised seed yield to the values observed in pure rows (i.e. average normalised seed yield in kin rows was 1 for each accession) (Fig. 8a). Considering the average seed-yield ratio for all the accessions there was no significant difference between growing with kin or non kin (Fig. 8b). The latter average resulted from accessions with a ratio not significantly different from 1, and some accessions with a ratio significantly higher or lower than 1 (Fig. 8b). We do not regard these differences among accessions as general; rather, we consider that the hierarchy is valid only for the tested conditions.
Discussion
We propose a mechanism of kin recognition involving: (1) the generation of different vertical patterns of R : FR and blue light of the horizontally propagating radiation (Fig. 3); (2) the perception of these R : FR and blue-light patterns by the photosensory receptors phyB, cry1, cry2 and phot1, phot2 (Fig. 4); (3) the action of the auxin (growth hormone)-synthesis gene TAA1 downstream of the photoreceptors (Fig. 7a,b); and (4) the redirection of leaf growth away from the position of kin neighbours (Figs 1, 2). The evidence in favour of this mechanism is based on: (1) the selective reorientation of leaf growth away from the direction of neighbours observed in rows formed by kin and not in rows formed by nonkin (Fig. 1); (2), the measurement of selective light signals generated by kin neighbours (Fig. 3); (3), the impaired leaf response observed in mutants of photosensory receptor or auxin synthesis genes and when the light signals are distorted (Figs 4, 7); and (4), the reorientation of leaf growth in response to simulations of the light environment produced by kin neighbours (Fig. 5). This mechanism of kin recognition involves phenotype matching because the height where the light signal produced by the ‘sender’ is more intense (particularly the strongest reduction in R : FR) and coincides with the height where the foliage of the ‘receiver’ is denser when ‘sender’ and ‘receiver’ are kin neighbours (Fig. 3b). Nonkin produce comparatively weak light signals at the position where the ‘receiver’ has its leaves. Accordingly, plants did not show the leaf-position response when grown with noncoetaneous kin (Fig. 6).
The alternative hypothesis that in our experiments another sensory mechanism is involved in the perception of kinship and photosensory receptors are required only for the expression of the leaf-position response appears unlikely. The occurrence of belowground signals was minimised by the used of one pot per seedling. Volatile signals could be produced, but we observed comparable leaf-position responses to the presence of plastic filters that simulate the light signals of neighbours without producing volatiles (Fig. 5). To the best of our knowledge, neither belowground nor volatile signals have been shown to change the growth direction of leaves. Clearly, our data do not argue against the contribution of belowground signals to the control of root growth (Dudley & File, 2007; Biedrzycki et al., 2010; File et al., 2012a; Fang et al., 2013) and the contribution of volatile signals to control plant defences (Karban et al., 2013) under natural conditions.
The occurrence of preferential cooperation among plant kin is controversial (Dudley & File, 2008; Klemens, 2008; Milla et al., 2009, 2012; Murphy & Dudley, 2009; Biedrzycki & Bais, 2010; Biedrzycki et al., 2010; Masclaux et al., 2010; Willis et al., 2010; Bhatt et al., 2011; File et al., 2012b; Lepik et al., 2012). The reason for this is that fitness consequences of the interaction with neighbours depend on the competitive ability of the neighbour, niche partitioning and cooperative interactions. Therefore, the comparison between kin and nonkin neighbours shows the balance between these components and not a clean evaluation of cooperative interactions. Conversely, by comparing kin with or without the ability to exhibit leaf-position responses to kin we were able to show that if a plant has kin neighbours a mutually beneficial rearrangement of the leaves alleviates the competition for local pools of light (mutual shading) (Fig. 7c) and increases fitness (Fig. 7d). This does not mean that for a plant it is necessarily better to grow with a kin than with a nonkin neighbour (Masclaux et al., 2010; Willis et al., 2010; Biernaskie, 2011; Milla et al., 2012). Actually, under our conditions, the average of multiple accessions showed similar fitness with kin and nonkin (Fig. 8b). Because the differences in competitive ability are expected to cancel out when multiple accessions are averaged, the similar average yield with kin or nonkin indicates that the other two components that define the consequences of the interaction with neighbours have similar magnitude; that is, the mutual benefit observed between kin neighbours approximately compensates the enhanced competition for local pools of resources between them. In other words, a fitness benefit from the leaf behaviour indicates kin selection, but the potential of niche partitioning and other processes (File et al., 2012ab) would preclude seeing higher fitness in kin groups.
Our genetic analysis has revealed several mutants that fail to rearrange their leaves in response to kin. When mixed with the wild-type, the sav3 mutant did not rearrange its leaves and this conferred a fitness advantage over the wild-type. The latter result does not indicate that the population frequency of sav3 mutants should be favoured over evolutionary time at the expense of the wild-type that shows ‘honest’ behaviour. First, ‘honest’ behaviour implies direct benefit to the wild-type actor when grown next to wild-type neighbours (Fig. 7d) and indirect benefits through wild-type relatives (West et al., 2007a). Second, the sav3 mutation also impairs competitive interactions with taller neighbours, which would reinforce the direct benefit of the wild-type gene to the actor (West et al., 2007a).
Shade-avoidance responses involve the perception of low R : FR mainly by phyB and of low blue light mainly by cry1 (Sellaro et al., 2010; Keller et al., 2011). Although phototropins provide adaptive value with respect to establishment and growth of plants, the physiological processes that generate this advantage have not been defined (Pedmale et al., 2010). Both phot1 and phot1 could be important for acclimation to shade (Casal, 2013). Whilst a role in vertical shade-avoidance (enhanced stem and petiole growth, leaf hyponasty) has not been documented, here we show that phot1 and phot2 are important for horizontal leaf shade-avoidance responses (Fig. 4a,c).
Because plants are sessile organisms, the genotype of the interacting neighbour is not a choice for them. Due to limited dispersal, plants often have coetaneous kin neighbours that tend to place the leaves at similar vertical strata, which would enhance mutual shading and competition for light, but this competition is alleviated by the horizontal rearrangement of the leaves towards empty areas with higher R : FR and blue-light signals. Classical shade-avoidance reactions increase plant stature placing the leaves at better light strata within the canopy at the expense of the light received by the overtopped neighbours; that is, shade signals were known to increase the competitive ability of plants (Schmitt, 1997; Schmitt et al., 1999, 2003; Smith, 2000; Weinig, 2000). Conversely, here we show that phyB-mediated perception of R : FR and cry1, cry2, phot1 and phot2-mediated perception of blue light generate cooperative rather than competitive interactions when the neighbour is a coetaneous kin. These cooperative interactions would be particularly important in agricultural crops involving genetically similar individuals grown at high densities with different distances among plants within and between rows.
Acknowledgements
We thank Victor Sadras (SARDI, Australia) for helpful comments on the manuscript. Supported by ANPCYT (PICT 1819) and UBA (20020100100437) grants to J.J.C.
