Abstract
Background and Aims
Cleistogamy is considered to be an adaptive strategy resulting in plasticity in chasmogamous (CH) and cleistogamous (CL) flower production depending on environmental conditions and plant size. The aim of this study was to investigate whether CH and CL flower production in Portulaca oleracea is genetically differentiated among populations in association with climatic conditions.
Methods
First, we conducted growth experiments with P. oleracea seedlings from 16 populations under two temperature conditions. Secondly, we sowed seeds originating from the parents in the first experiment and grew the resulting plants to investigate whether flower production is heritable and whether plants in the same population show the same pattern of flower production.
Key Results
Two types of plants that produced only CH or CL flowers (referred to as CH and CL plants, respectively) were mainly observed, and the growing temperature conditions did not affect flower production. The frequency of CL plants increased with a decrease in the mean temperature in the original population. The CL plants tended to begin reproduction earlier than the CH plants, and the probability that a CH plant would flower decreased under the low growing temperature condition. Thus, CL plants may have some advantages in unfavourable environments in which early reproduction is necessary due to a short growing season and/or when CH flowers cannot open due to low temperatures. The progeny originating from CH and CL plants also produced only CH and CL flowers, respectively, suggesting that there is a genetic basis for the dimorphism in flower production in P. oleracea, represented by CH and CL plants.
Conclusions
In contrast to the previous hypothesis that the production of both CH and CL flowers would be plastic, the genotypes producing either CH or CL flowers occurred at different frequencies under varying climatic conditions.
Keywords: Cleistogamy, Portulaca oleracea, climatic effect, intraspecific variation, reproductive trait
INTRODUCTION
Cleistogamy is a mixed mating system that has been reported to occur in 693 angiosperm species that are members of 28 genera and 50 families (Culley and Klooster, 2007). Cleistogamous species produce both chasmogamous (CH) flowers, which are capable of outcrossing, and cleistogamous (CL) flowers, which are always closed and in which only self-fertilization is possible, within individual plants. CL flowers are characterized by a reduction in corolla and stamen size and/or stamen number relative to those of CH flowers, and CL flowers are less costly to produce than CH flowers (Culley and Klooster, 2007).
Schoen and Lloyd (1984) proposed a complex habitat model describing the process of selection for cleistogamy. This model suggests that heterogeneity in the environment may lead to the selection of phenotypes that reproduce with both CH and CL flowers, with different frequencies of these flowers depending on environmental conditions. Because of their low cost and reliability in terms of reproduction, the relative investment in the production of CL over CH flowers tends to be higher in environments with scarce resources, such as low light levels (Schemske, 1978; Masuda and Yahara, 1994; Culley, 2002) and low nutrients (Le Corff, 1993), and with pollen limitation (Redbo-Torstensson and Berg, 1995; Berg and Redbo-Torstensson, 1998; Ansaldi et al., 2018). Plant size also affects investment in CH vs. CL flowers, but such size dependence is variable among species. In some species, such as Oxalis montana and Viola praemorsa, individual plants produce a certain number of CH flowers independent of their size and increase CL flower production as plant size increases (Jasieniuk and Lechowicz, 1987; Berg and Redbo-Torstensson, 1998; Forrest and Thomson, 2008), possibly because an increase in the number of CH flowers causes geitonogamy and herbivory, and results in diminishing benefits (Forrest and Thomson 2008). On the other hand, in many cleistogamous annual species, which usually produce CL flowers prior to CH flowers, the relative investment in CH flowers increases as plant size also increases (Waller, 1980; Diaz and Macnair, 1998; Lu, 2002). This may occur because CH flower production is costly and because larger plants can invest more resources into CH flower production (Diaz and Macnair, 1998). Thus, cleistogamy is considered to be an adaptive strategy that shows plasticity in CH and CL flower production that is dependent on environmental conditions and plant size.
Variation in CH and CL flower production might be driven by factors other than plasticity; if climatic conditions greatly differ among populations, investment in CH and CL flower production might become genetically differentiated among these populations. Moreover, if either CH or CL flowers are advantageous under certain habitat conditions, genotypes producing only CH or CL flowers might also arise. For example, if there is little chance of outcrossing, it may be advantageous to produce only CL flowers to conserve resources, whereas, if pollinators and resources are always abundant, it may be advantageous to produce only CH flowers. Indeed, CH and CL flower production has been found to be differentiated among populations in some species (Sun 1999; Wang et al. 2017). For example, in Scutellaria indica, there are two types of plants, namely those that produce both CH and CL flowers and those that produce only CL flowers, and such flower production is presumed to be genetically determined (Sun 1999). However, plants producing only CL flowers were found in only a single population of this species and, hence, it is unknown whether this genotype commonly and stably exists. Additionally, it is unclear why this genotype exists in this population. If genotypes producing only CH or CL flowers have become differentiated, certain environmental factors should have enhanced such differentiation. For example, in an environment with abundant resources, the genotype producing only CH flowers might become dominant, whereas, in an environment with scarce resources, the genotype producing only CL flowers might become dominant. These factors should be clarified for a better understanding of the differentiation of CH and CL flower production.
Portulaca oleracea is a summer annual that is distributed worldwide, including in Japan. This species has a cleistogamous reproductive system. Our preliminary observations suggest that populations consisting of plants producing both CH and CL flowers and those producing only CL flowers occur in central to north-eastern Japan. We suppose that temperature, the most differentiated climatic factor among populations, might plastically or genetically affect flower production. Each individual produces either CH or CL flowers sequentially during the flowering period. Furthermore, in both plant types, the plants begin flowering in July and continue flowering until September, with a flowering peak in August.
The aim of this study was to investigate whether the production of CH and CL flowers in P. oleracea is differentiated according to climatic conditions. In Experiment 1, we sampled P. oleracea seedlings from 16 populations along a gradient of climatic conditions and conducted experiments evaluating the growth of these seedlings under controlled conditions. In Experiment 2, to investigate whether the production of CH and CL flowers is an inherited characteristic, we sowed seeds originating from the plants in Experiment 1 as parents and observed flower production among the progeny.
MATERIALS AND METHODS
Study species
Portulaca oleracea (Portulacaceae) is a summer annual that grows in sunny fields. It has a worldwide distribution in temperate and tropical zones, including in Japan. The main stems creep along the ground, and their lengths are approx. 10–30 cm. The CH flowers are yellow, and they have a diameter of approx. 5–10 mm when open. The CH flowers open for a few hours in the morning and never open again once they close, and both outcrossing and self-fertilization are possible. Hoverflies often visit CH flowers. However, CH flowers often fail to open on cloudy or rainy days; petals appear but do not expand. However, because the anthers dehisce before petal expansion, the ovules are probably successfully fertilized by self-pollen even if the petals do not open (T. Furukawa, pers. obs.). In contrast, CL flowers have petals but never open, and only self-fertilization is possible.
The mean bud sizes of CH and CL flowers were found to be 5.02 ± 0.54 and 4.12 ± 0.34 mm (mean ± s.d., n = 25), respectively (Supplementary data Fig. S1); their petal lengths were 4.05 ± 0.69 and 2.24 ± 0.40 (Supplementary data Tables S1, S2); and the mean numbers of stamens were 11.08 ± 1.91 and 7.02 ± 1.67 (mean ± s.d.) (Supplementary data Tables S1, S3). In addition, the petals of CH flowers appear out of the calyxes even if flowers fail to open, whereas the petals of CL flowers are completely or almost completely enclosed by the calyxes (Supplementary data Fig. S1).
Experiment 1: factors affecting flower production
To investigate the geographical variation in CH and CL flower production and the factors affecting it, we conducted a growth experiment using seedlings collected from 16 populations (Table 1; Fig. 1; Supplementary data Table S4) under controlled conditions.
Table 1.
Flower production in the plants used in Experiments 1 and 2
| Population | Experiment 1 | Experiment 2 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Only CH flowers (referred to as CH plants) | Only CL flowers (referred to as CL plants) | CH and CL flowers | Parent plant type | F1 progeny plant type | F2 progeny plant type | ||||
| CH plant | CL plant | CH plant | CL plant | ||||||
| Ninohe | 0 | 18 | 0 | CL | 0 | 9 | – | – | |
| Morioka | 0 | 3 | 0 | – | – | – | – | – | |
| Shiwa | 0 | 22 | 0 | – | – | – | – | – | |
| Kitakami | 0 | 8 | 0 | CL | 0 | 7 | – | – | |
| Oosaki | 0 | 16 | 0 | CL | 0 | 8 | – | – | |
| Sendai | 0 | 20 | 0 | CL | 0 | 9 | – | – | |
| Fukushima | 0 | 17 | 0 | CL | 0 | 9 | – | – | |
| Kouriyama | 0 | 13 | 0 | – | – | – | – | – | |
| Iwaki | 0 | 18 | 2 | CL | 0 | 8 | – | – | |
| Hitachi | 2 | 14 | 1 | CH | 6 | 0 | 3 | 0 | |
| CL | 0 | 9 | 0 | 6 | |||||
| Tsukuba | 0 | 12 | 0 | CL | 0 | 7 | – | – | |
| Matsudo | 3 | 14 | 0 | CH | 8 | 0 | – | – | |
| CL | 0 | 8 | – | – | |||||
| Ichihara | 2 | 11 | 0 | CH | 5 | 0 | 2 | 0 | |
| CL | 0 | 9 | 0 | 6 | |||||
| Fuji | 3 | 3 | 0 | – | – | – | – | – | |
| Minamiboso | 17 | 5 | 0 | CH | 9 | 0 | 9 | 0 | |
| CL | 0 | 8 | 0 | 3 | |||||
| Shizuoka | 0 | 13 | 1 | CL | 0 | 8 | – | – | |
| Total | 27 | 207 | 4 | 28 | 99 | 14 | 15 | ||
Numbers of individuals are shown for each flower production type.
Fig. 1.
Frequencies of plants producing only CH flowers, only CL flowers and both CH and CL flowers in the sampled populations. The numbers in the figure show the numbers of individuals producing buds.
We selected 16 populations along a latitudinal gradient in eastern Japan ranging from 34.96° to 40.28°N, where climatic conditions are variable (Table 1; Fig. 1; Supplementary data Table S4). These sites occur on frequently disturbed bare ground (e.g. crop fields). From 26 June to 2 August 2018, we collected seedlings that had up to four true leaves (the greatest leaf length was <13 mm) and had not yet started to produce flower buds. These seedlings should not yet have been affected by the climatic conditions of their populations. We collected 17–26 seedlings from an area of approx. 18 000 m2 for each population, and seedlings growing at least 1.5 m away from each other were sampled. Each sampling area was homogeneous in terms of environmental conditions.
We planted the sampled seedlings in pots (4 cm in diameter and 5 cm in depth) filled with river sand (one individual per pot). We fertilized the seedlings once a week (Hyponex 6-10-5; Hyponex Japan, Osaka, Japan) and watered them at an interval of 1–2 d.
The experimental plants were grown under two temperature conditions because growing temperature conditions may affect the production and openness of CH flowers. We grew the seedlings at 30 °C/25 °C and 26 °C/21 °C (day/night temperature; 12 h daylengths) using two chambers (LH-410S, Nippon Medical & Chemical Instruments Co., Ltd., Japan; and MLR-351H, Sanyo, Japan). These temperatures were used because the means of the daily highest temperatures in July were 26.3 °C and 29.5 °C (30 year means for 1981–2010) at the northmost site (Ninohe) and the southmost site (Shizuoka), respectively. Photosynthetically active radiation (PAR) was approx. 104 and 92 μmol m–2 s–1 in the high- and low-temperature chambers, respectively (the two chambers used were different models, but we set the light intensities to be as similar as possible). We divided the individuals of each population between the two chambers. For each population, 10–16 and 6–10 individuals were subjected to the high- and low-temperature conditions, respectively.
We observed flower production for 45–51 and 65 d after planting under the high- and low-temperature conditions, respectively, during which almost all healthy individuals successfully produced flower buds (except for the Shiwa population, which produced buds at 25 and 49 d under the high- and low-temperature conditions, respectively). For each individual, we measured the stem length (plant size) on the day when the first flower bud was produced (when the bud length became >1 mm). We counted the number of CH flowers that were open (flowers whose width between the tops of opposite petals was ≥1 mm) and not open (width <1 mm) each day. To compare reproductive success between CH and CL plants, we also counted the number of mature fruits at intervals of 1–3 d during the observation period and collected 3–8 fruits from each of 156 individuals. For each fruit, the number of seeds and their total weight were measured, and then the mean weight per seed was calculated. We also estimated the total number of seeds produced by each individual (mean number of seeds per fruit × number of fruits). Most individuals produced only CH or CL flowers (see the Results), and we hence classified the individuals as CH or CL plants, respectively. Individuals that survived and produced at least one bud were included in the further analysis (Supplementary data Table S5).
We also obtained climatic data for each population from the nearest meteorological station; we examined the mean temperature, precipitation and sunshine hours in July to September (30 year averages from 1981 to 2010) (Supplementary data Table S4).
Experiment 2: whether flower production is heritable
To examine whether flower production (production of CH and CL flowers) is genetically determined, we investigated the production of flowers by the F1 and F2 generations of CH and CL plants. In addition, to determine whether there were differences in reproductive traits other than flower production between the CH and CL plants, we investigated the initiation of flower bud production.
We collected seeds produced by CH and CL plants that were grown in Experiment 1 originating from four and 12 populations, respectively (Table 1; Supplementary data Tables S5 and S6). All seeds were produced via self-fertilization because the parent plants were grown in chambers. We collected seeds from 1–4 fruits from each of 2–3 families (parent individuals) for each population and sowed them in Petri dishes with wet filter paper. A total of 19–66 seeds from each family were placed in a single dish and grown in a chamber at a 30 °C/25 °C day/night temperature and under approx. 144 μmol m–2 s–1 PAR (LH-410S, Nippon Medical & Chemical Instruments Co., Ltd.). The light level was set higher than that in Experiment 1 to enhance the possible production of CH flowers by CL plants. The seeds started to germinate 2 d after sowing. After a set of cotyledons expanded, we randomly selected nine individuals from each family and transplanted them into three pots (4 cm in diameter and 5 cm in depth; three individuals per pot) filled with river sand. We grew these individuals continually under the same temperature and light conditions (30 °C/25 °C day/night temperature and approx. 144 μmol m–2 s–1 PAR), fertilized them once a week (Hyponex 6-10-5; Hyponex Japan), and watered them at an interval of 1–2 d. After 20 d, the individual with the longest stem in each pot was identified, and the other two individuals were removed. If all individuals in a pot were dead within 20 d after transplanting, additional individuals were transplanted in the same way. Thus, three individuals per family were used for the experiment; in total, 4 populations × 2–3 families × 3 individuals for the progeny of CH plants and 12 populations × 3 families × 3 individuals for the progeny of CL flowers were used.
We observed flower production for 50 d in the same way as in Experiment 1. All plants also produced only CH or CL flowers, and we hence classified the individuals as CH or CL plants, respectively.
For each individual, we recorded the day when the first flower bud was produced (when the bud size became >1 mm) and measured the stem length on that day. Individuals that survived and produced at least one bud were included in the further analysis (Supplementary data Table S6).
Finally, to investigate heritability more strictly and to exclude maternal effects, we observed flower production by the F2 generation (next generation of Experiment 2). We collected seeds produced in the above experiment and sowed and grew them in the same way as in the above experiment. We grew 30 individuals of the F2 generation, and 29 individuals survived and produced at least one bud (Supplementary data Table S7). We observed flower production until the first flower of each individual flowered.
Statistical analysis
The following data analyses were conducted using R 3.4.0. (R Core Team, 2018). A generalized linear mixed model (GLMM) with a binomial error distribution and logit link function was applied to the data obtained from Experiment 1 to examine the effects of climatic factors associated with the sampled populations and the growing temperature conditions on the occurrence of CH and CL plants in the populations. The response variable was the reproductive system of each plant (assigning 0 and 1 to CH vs. CL plants, respectively). The fixed effects were the three climatic factors (mean temperature, total precipitation and total sunshine hours), which were standardized to a mean of zero and a variance of one, and the growing temperature condition (high/low), and the random effect was the sampled population. The analysis was carried out using the glmmML function in the glmmML package in R.
To examine the effect of growing temperature conditions on the opening of CH flowers, a bias-reduced generalized linear model (brGLM) with a binomial error distribution and a logit link function was applied to the data obtained from Experiment 1. The response variable was the ratio of the number of open flowers/number of flowers produced by each CH plant, and the explanatory variables were the same as those in the above analysis. This analysis was carried out using the brglm function in the brglm package in R (Kosmidis and Friar, 2007).
To compare the reproductive success of CH and CL plants in relation to plant size and growing temperature conditions, we conducted three GLMM analyses for the data obtained from Experiment 1. In the two analyses, the response variable was either the mean seed weight or the estimated total number of seeds produced by an individual, and the fixed effects were plant type (CH/CL plant), the three standardized climatic factors and the growing temperature condition. We used a gamma and a negative binomial error distribution with a log link function for the mean seed weight and the estimated total number of seeds produced by an individual, respectively. For the analysis of mean seed weight, individual nested within population was treated as a random effect. For the analysis of the estimated total number of seeds, population was treated as a random effect. In the rest of the analyses, the response variable was the total number of fruits produced by an individual. The fixed effects were the plant size when flower bud production began and the growing temperature condition, and the random effect was the sampled population. The analysis was carried out using the glmer function in the lme4 package in R.
To examine the probability of the development of CH and CL plants from seeds produced by CH and CL parents, a brGLM with a binomial error distribution and logit link function was applied to the data obtained from Experiment 2. The response variable was the occurrence of CL plants, and the explanatory variable was the plant type of the parents. The analysis was carried out using the brglm function in the brglm package in R (Kosmidis and Friar, 2007).
To evaluate the initiation of flower bud production in CH and CL plants, a GLMM with a gamma error distribution and log link function was applied to the data obtained from Experiment 2. The response variable was either the day or plant size at which flower bud production began. The fixed effects were the plant type and the three standardized climatic factors, and the random effect was family nested within population. The analysis was carried out using the glmer function in the lme4 package in R.
RESULTS
Flower and fruit production
Among the 238 plants producing flower buds in Experiment 1, a total of 27 plants produced only CH flowers, 207 plants produced only CL flowers and four plants produced both CH and CL flowers (Table 1; Supplementary data Table S5). Each of the four plants in the last group produced >10 CL flowers and only a single small CH flower (open width was <3 mm, and the petals were small) and may therefore be categorized as CL plants (but we did not include them in the analysis). The CH and CL plants (excluding the four plants producing both flowers) produced 2.6 ± 4.0 and 6.8 ± 7.0 mature fruits (mean ± s.d.) during the observation period, respectively.
In Experiment 2, the seeds originating from CH and CL parents in Experiment 1 became CH and CL plants (Table 1; Supplementary data Table S6 and S8), respectively. Each of these CH and CL plants produced 4.8 ± 5.5 and 18.0 ± 13.0 fruits (data from four populations, namely Hitachi, Ichihara, Matsudo and Chikura), respectively. The individuals of the F2 generation also produced the same type of flowers as their F1 parents (Table 1; Supplementary data Tables S7 and S9). Thus, the dimorphism in flower production is suggested to be a genetically determined characteristic.
Five populations included both CH and CL plants, and 11 populations included only CL plants, but no population was composed of only CH plants (Fig. 1; Supplementary data Table S5). Among the three climatic factors associated with the original populations, only mean temperature affected the occurrence of CL plants; the probability of occurrence of CL plants increased with a decrease in mean temperature in the population of origin (Table 2; Fig. 2). However, the growing temperature conditions did not affect the frequency of CL plants.
Table 2.
GLMM analysis of the effects of climatic factors associated with the original populations and growing temperature conditions on the occurrence of CL plants
| Estimate | s.e. | z-value | P-value | |
|---|---|---|---|---|
| (Intercept) | 5.3432 | 1.5613 | 3.422 | 0.000621 |
| Mean temperature | –3.6178 | 1.7767 | –2.036 | 0.041723 |
| Total precipitation | 0.9897 | 0.7165 | 1.381 | 0.167229 |
| Total sunshine hours | –0.3101 | 0.7886 | –0.393 | 0.694128 |
| Low-temperature conditions | –0.5355 | 0.6136 | –0.873 | 0.382815 |
Estimates are standardized regression coefficients.
Fig. 2.
Effects of climatic factors associated with the original populations and growing temperature conditions on the probability of the occurrence of CL plants in a population.
Initiation of flower bud production and seed production
The day on which flower bud production began was earlier and the plant size was smaller at the initiation of flower bud production in the CL plants than in the CH plants (Table 3; Fig. 3). Among the three climatic factors associated with the original populations, only total sunshine hours affected the day on which flower bud production began; the day on which flower bud production began was later among the plants from the populations for which the total sunshine hours was longer (Table 3). No climatic factor affected the plant size at which flower bud production was initiated (Table 3).
Table 3.
GLMM analysis of the effects of plant type (CH or CL plant) and climatic factors associated with the original populations on the day and size at which flower bud production began
| Estimate | s.e. | t-value | P-value | |
|---|---|---|---|---|
| Day when flower bud production began | ||||
| (Intercept) | 3.4001 | 0.0612 | 55.52 | <0.001 |
| CL plant | –0.2340 | 0.0661 | –3.54 | <0.001 |
| Mean temperature | –0.0108 | 0.0453 | –0.24 | 0.812 |
| Total precipitation | –0.0187 | 0.0352 | –0.53 | 0.596 |
| Total sunshine hours | 0.0909 | 0.0417 | 2.18 | 0.029 |
| Plant size when flower bud production began | ||||
| (Intercept) | 1.1051 | 0.0857 | 12.90 | <0.001 |
| CL plant | –0.6119 | 0.0985 | –6.22 | <0.001 |
| Mean temperature | –0.0633 | 0.0564 | –1.12 | 0.261 |
| Total precipitation | 0.0313 | 0.0442 | 0.71 | 0.478 |
| Total sunshine hours | 0.0596 | 0.0490 | 1.22 | 0.224 |
Estimates are standardized regression coefficients.
Fig. 3.
Effect of plant type (CH/CL plant) on the day and plant size at which flower bud production began. The boxes indicate the interquartile (25–75 %) ranges.
The mean seed weight was significantly higher in the CL plants than in the CH plants, but was not affected by the climatic factors associated with the original populations or the growing temperature conditions (Table 4; Fig. 4). The estimated total number of seeds produced by an individual did not significantly differ between the CH and CL plants (Supplementary data Table S10; Fig. S2) but increased in the plants grown under the low-temperature conditions (Supplementary data Table S10; Fig. S2). No climatic factor associated with the original populations affected the estimated total number of seeds (Supplementary data Table S10). The plant size at which flower bud production began enhanced the total number of fruits produced by the plant under the high-temperature conditions, but this effect diminished under the low-temperature conditions (Supplementary data Table S11; Fig. S3).
Table 4.
GLMM analysis of the effects of plant type (CH or CL plant) and climatic factors associated with the original populations on mean seed weight
| Estimate | s.e. | t-value | P-value | |
|---|---|---|---|---|
| (Intercept) | –2.51760 | 0.07248 | –34.74 | <0.001 |
| CL plant | 0.31776 | 0.07588 | 4.19 | <0.001 |
| Mean temperature | –0.01942 | 0.03935 | –0.49 | 0.622 |
| Total precipitation | 0.00701 | 0.03018 | 0.23 | 0.816 |
| Total sunshine hours | 0.02056 | 0.03917 | 0.52 | 0.600 |
| Low-temperature conditions | 0.05837 | 0.04229 | 1.38 | 0.168 |
Estimates are standardized regression coefficients.
Fig. 4.
Mean seed weight in the CH and CL plants. The boxes indicate the interquartile (25–75 %) ranges.
Effect of temperature on CH flower opening
All CH flowers opened under the high-temperature conditions, but the probability that CH flowers would open was lower under the low-temperature conditions (Table 5). All three climatic factors associated with the original populations affected the probability that the CH flowers would open; the likelihood of flower opening increased with an increase in mean temperature and with decreases in total precipitation and total sunshine hours (Table 5).
Table 5.
GLM analysis of the effects of climatic factors associated with the original populations and growing temperature conditions on the probability of CH flowers opening (number of open flowers/number of flowers produced)
| Estimate | s.e. | z-value | P-value | |
|---|---|---|---|---|
| (Intercept) | 0.8012 | 1.7240 | 0.465 | 0.6421 |
| Mean temperature | 26.9892 | 7.8476 | 3.439 | <0.001 |
| Total precipitation | –4.3265 | 1.3772 | –3.142 | 0.0017 |
| Total sunshine hours | –1.6524 | 0.5312 | –3.111 | 0.0019 |
| Low-temperature conditions | –29.1989 | 7.4685 | –3.910 | <0.001 |
Estimates are standardized regression coefficients.
DISCUSSION
Dimorphism in flower production in P. oleracea
In Experiment 1, two types of plants, each producing only CH or CL flowers, were mainly observed, although there were four exceptions that produced many CL flowers and a single CH flower (Table 1; Supplementary data Table S5). Furthermore, flower production was not affected by the growing temperature conditions (Table 2). In Experiment 2, the plants originating from CH and CL parents produced only CH and CL flowers, respectively (Table 1; Supplementary data Tables S6 and S8). Moreover, the F2 plants originating from CH and CL F1 parents also showed the same pattern (Table 1; Supplementary data Tables S7 and S9). These results suggest that there is genetic dimorphism in terms of flower production, represented by CH and CL plants, in P. oleracea, although we cannot rule out that CH and CL plants produce CL and CH flowers, respectively, because we did not observe the whole flowering period (we made observations at 41–50/65 d under the high-/low-temperature conditions, but the whole flowering period was approx. 90 d). However, there were many individuals producing only CL flowers throughout their flowering periods in naturally growing populations, regardless of plant size (T. Furukawa, pers. obs.). Thus, it is likely that CL plants, which rarely show plasticity in flower production, are genetically differentiated in P. oleracea. This contrasts with the previous hypothesis that, in cleistogamous species, single plants produce both CH and CL flowers and show plasticity in such flower production that is dependent on environmental conditions and plant size (Culley and Klooster, 2007). Only a few studies have shown that the percentage of CL flowers among the flowers produced is largely heritable (Clay, 1982) or that the numbers of CH and CL flowers are not dependent on environmental conditions (Munguía-Rosas et al., 2012). However, no previous studies have reported that two types of plants producing only CH or CL flowers commonly exist within a species.
It should be noted that a certain degree of plasticity might be expressed if plants are grown under other conditions. In fact, four plants producing both CH and CL flowers were observed under the high-temperature conditions (Supplementary data Table S5). In this study, we examined the effects of temperature conditions on flower production. However, the frequency of cleistogamy is also affected by other growing conditions, such as light intensity (Schemske, 1978; Masuda and Yahara, 1994; Culley, 2002), nutrient availability (Le Corff, 1993) and photoperiod (Lord, 1982; Barnett et al., 2018); hence, the effects of other factors on flower production should be examined.
Effects of climatic factors on the frequency of CH and CL plants
What conditions favour CL plants? The frequency of CL plants increased with a decrease in mean temperature experience by the original population (Table 2; Fig. 2). The CL plants began reproduction earlier and at smaller sizes than CH plants (Table 3; Fig. 3), probably because CL flowers are less costly to produce than CH flowers and because the time needed to attain flower maturity in CL flowers is shorter than that for CH flowers (Mayers and Lord, 1983). Thus, CL plants may be advantageous in terms of rapid reproduction in low-temperature environments where the growing season is short. Other studies have also suggested that rapid flowering and reproduction are selected for in locations where the growing season is short (Weber and Schmid, 1998; Montague et al., 2008; Prendeville et al., 2013). In addition, the mean seed weight was higher in CL plants than in CH plants (Table 4; Fig. 4), although the total number of seeds produced by an individual did not significantly differ between these two plant types (Supplementary data Table S10; Fig. S2). Heavier seeds tend to perform better in terms of germination and growth (Leishman and Westoby, 1994; Milberg and Lamont, 1997; Easton and Kleindorfer, 2008), and thus CL plants may have an advantage over CH plants in terms of initial growth. Portulaca oleracea, which is mainly distributed in tropical and temperate regions (Ocampo and Columbus, 2012; Ohwi and Kitagawa, 1983), might have evolved cleistogamy in the parts of its distribution where the temperature is relatively low.
On the other hand, CH flowers sometimes fail to open under low-temperature conditions (Table 5). A previous study on P. oleracea also mentioned that its flowers become CL under conditions of reduced light and temperature (Kim and Carr, 1990), and such CL flowers were likely to be considered as unopened CH flowers in our study. The failure of CH flowers to open reduces the opportunity for outcrossing, and CH plants may thus have a disadvantage in low-temperature environments.
Selection for the production of CH or CL flowers
The CL plants in P. oleracea may exhibit a novel phenomenon in which CL annuals have lost their plasticity in CH and CL flower production. Why do CL plants produce only CL flowers? This probably occurs because there is little chance of outcrossing in CH flowers; they have a small display (5–10 mm), the degree of herkogamy is low (the length between the stigma and anther is <1 mm; T. Furukawa, pers. obs.) and prior autogamy occurs before anthesis. Hence, it does not seem to be advantageous to produce CH flowers for outcrossing. Sun (1999) investigated the genetic variation in S. indica populations in which both CH and CL flowers or only CL flowers were observed, and found that genetic variation was very low in both types of populations. Thus, selfing rates may be as high in the populations in which CH flowers are observed as in the populations in which only CL flowers are observed, suggesting that CL flowers have been selected for.
If so, the existence of CH plants, which may be less advantageous in terms of rapid growth and reproduction, is unclear. Plants probably need less time to become large in high-temperature environments than in low-temperature environments. Thus, CH plants might produce more seeds than CL plants even though CH plants begin reproduction later than CL plants. In fact, the number of seeds produced was greatest in the plants that began flower bud production at a larger plant size under high-temperature conditions (Supplementary data Table S11; Fig. S3). On the other hand, the low frequency of CH plants suggests that CL plants have more advantages than CH plants overall and thus that CL plants will become dominant in the future. Further studies are necessary to examine whether, in terms of early reproduction in habitats with less chance of outcrossing, CL plants are more advantageous than CH plants and plants producing both CH and CL flowers in the environments where this species is currently distributed (including the environments where only CH plants are distributed).
SUPPLEMENTARY DATA
Supplementary data are available online at https://academic.oup.com/aob and consist of the following.
Figure S1: buds of CH and CL flowers.
Figure S2: estimated total number of seeds produced by a CH plant and by a CL plant during the observation period.
Figure S3: effects of plant size when flower bud production began and growing temperature conditions on the estimated total number of seeds produced by a plant during the observation period.
Table S1: comparison of petal length and stamen number between CH and CL flowers.
Table S2: GLM analysis of the effect of plant type on petal length.
Table S3: GLM analysis of the effect of plant type on stamen number.
Table S4: sampled populations.
Table S5: flowers produced by the plants in Experiment 1.
Table S6: flowers produced by the plants of the F1 generation in Experiment 2.
Table S7: flowers produced by the plants of the F2 generation in Experiment 2.
Table S8: GLM analysis of the occurrence of CL plants in the F1 generation that developed from seeds produced by CL parents.
Table S9: GLM analysis of the occurrence of CL plants in the F2 generation that developed from seeds produced by CL parents.
Table S10: GLM analysis of the effects of plant type and climatic factors associated with the original populations on mean seed weight.
Table S11: effects of plant size when flower bud production began and growing temperature conditions on the total number of fruits produced by a plant.
ACKNOWLEDGEMENTS
We wish to extend our thanks to M. Maki and M. Kondo for comments on an early version of this manuscript, to K. Hikosaka for technical advice on the growth experiments, to S. Matsuhashi for advice on the experimental design, to M. Oguro and M. Nomura for statistical advice, to S.-I. Morinaga and D. Kyogoku for valuable comments on the results, and to M. Asai for valuable advice regarding the flowering of P. oleracea. We are also grateful to all colleagues who assisted in fieldwork and laboratory experiments.
LITERATURE CITED
- Ansaldi BH, Weber JJ, Franks SJ. 2018. The role of phenotypic plasticity and pollination environment in the cleistogamous, mixed-mating breeding system of Triodanis perfoliata. Plant Biology 20: 1068–1074. [DOI] [PubMed] [Google Scholar]
- Barnett LL, Troth A, Willis JH. 2018. Plastic breeding system response to day length in the California wildflower Mimulus douglasii. American Journal of Botany 105: 779–787. [DOI] [PubMed] [Google Scholar]
- Berg H, Redbo-Torstensson P. 1998. Cleistogamy as a bet-hedging strategy in Oxalis acetosella, a perennial herb. Journal of Ecology 86: 491–500. [Google Scholar]
- Clay K. 1982. Environmental and genetic determinants of cleistogamy in a natural population of the grass Danthonia spicata. Evolution 36: 734–741. [DOI] [PubMed] [Google Scholar]
- Culley TM. 2002. Reproductive biology and delayed selfing in Viola pubescens (Violaceae), an understory herb with chasmogamous and cleistogamous flowers. International Journal of Plant Sciences 163: 113–122. [Google Scholar]
- Culley TM, Klooster MR. 2007. The cleistogamous breeding system: a review of its frequency, evolution,and ecology in angiosperms. The Botanical Review 73: 1–30. [Google Scholar]
- Diaz A, Macnair MR. 1998. The effect of plant size on the expression of cleistogamy in Mimulus nasutus. Functional Ecology 12: 92–98. [Google Scholar]
- Easton LC, Kleindorfer S. 2008. Effects of salinity levels and seed mass on germination in Australian species of Frankenia L. (Frankeniaceae). Environmental and Experimental Botany 65: 345–352. [Google Scholar]
- Forrest J, Thomson JD. 2008. Pollen limitation and cleistogamy in subalpine Viola praemorsa. Botany 86: 511–519. [Google Scholar]
- Jasieniuk M, Lechowicz MJ. 1987. Spatial and temporal variation in chasmogamy and cleistogamy in Oxalis montana (Oxalidaceae). American Journal of Botany 74: 1672. [Google Scholar]
- Kim I, Carr GD. 1990. Reproductive biology and uniform culture of Portulaca in Hawaii. Pacific Science 44: 123–129. [Google Scholar]
- Kosmidis I, Friar K. 2007. Bias reduction in exponential family nonlinear models. Biometrika 96: 793.– . [Google Scholar]
- Le Corff J. 1993. Effect of light and nutrient availability on chamogamy and cleistogamy in understory tropical herb Calathea micans (Marantaceae). American Journal of Botany 80(12):1392–1399 [Google Scholar]
- Leishman MR, Westoby M. 1994. The role of seed size in seedling establishment in dry soil conditions – experimental evidence from semi-arid species. Journal of Ecology 82: 249–258. [Google Scholar]
- Lord EM. 1982. Effect of daylength on open flower production in the cleistogamous species Lamium amplexicaule L. Annals of Botany 49: 261–263. [Google Scholar]
- Lu Y. 2002. Why is cleistogamy a selected reproductive strategy in Impatiens capensis (Balsaminaceae)? Biological Journal of the Linnean Society 75: 543–53. [Google Scholar]
- Mayers AM, Lord EM. 1983. Comparative flower development in the cleistogamous species Viola odorata. I. A growth rate study. American Journal of Botany 70: 1548–1555. [Google Scholar]
- Masuda M, Yahara T. 1994. Reproductive ecology of a cleistogamous annual, Impatiens noli-tangere L., occurring under different environmental conditions. Ecological Research 9: 571–580. [Google Scholar]
- Milberg P, Lamont BB. 1997. Seed/cotyledon size and nutrient content play a major role in early performance of species on nutrient-poor soils. New Phytologist 137: 665–672. [Google Scholar]
- Montague JL, Barrett SCH, Eckert CG. 2008. Re-establishment of clinal variation in flowering time among introduced populations of purple loosestrife (Lythrum salicaria, Lythraceae). Journal of Evolutionary Biology 21: 234–245. [DOI] [PubMed] [Google Scholar]
- Munguía-Rosas MA, Parra-Tabla V, Ollerton J, Cervera JC. 2012. Environmental control of reproductive phenology and the effect of pollen supplementation on resource allocation in the cleistogamous weed, Ruellia nudiflora (Acanthaceae). Annals of Botany 109: 343–350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ocampo G, Columbus JT. 2012. Molecular phylogenetics, historical biogeography, and chromosome number evolution of Portulaca (Portulacaceae). Molecular Phylogenetics and Evolution 63: 97–112. [DOI] [PubMed] [Google Scholar]
- Ohwi J, Kitagawa M. 1983. New Flora of Japan. Tokyo: Shibundo Publishers. [Google Scholar]
- Prendeville HR, Barnard-Kubow K, Dai C, Barringer BC, Galloway LF. 2013. Clinal variation for only some phenological traits across a species range. Oecologia 173: 421–430. [DOI] [PubMed] [Google Scholar]
- R Core Team 2018. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. [Google Scholar]
- Redbo-Torstensson P, Berg H. 1995. Seasonal cleistogamy: a conditional strategy provide reproductive assutrance. Acta Botanica Neerlandica 44: 247–256. [Google Scholar]
- Schemske DW. 1978. Evolution of reproductive characteristics in Impatiens (Balsaminaceae): the significance of cleistogamy and chasmogamy. Ecology 59: 596–613. [Google Scholar]
- Schoen DJ,Lloyd DG. 1984. The selection of cleistogamy and heteromorphic diaspores. Biological Journal of the Linnean Society 23: 303–322. [Google Scholar]
- Sun M. 1999. Cleistogamy in Scutellaria indica (Labiatae): effective mating system and population genetic structure. Molecular Ecology 8: 1285–1295. [DOI] [PubMed] [Google Scholar]
- Waller DM. 1980. Environmental determinants of outcrossing in Impatiens capensis (Balsaminaceae). Evolution 34: 747–761. [DOI] [PubMed] [Google Scholar]
- Wang C-H, Du W, Wang X-F. 2017. Reproductive investment in a cleistogamous morph of Polygonum jucundum (Polygonaceae). Plant Systematics and Evolution 303: 559–563. [Google Scholar]
- Weber E, Schmid B. 1998. Latitudinal population differentiation in two species of Solidago (Asteraceae) introduced into Europe. American Journal of Botany 85: 1110–1121. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.