Floral ultraviolet absorbance area responds plastically to ultraviolet irradiance in Brassica rapa

Abstract Solar ultraviolet (UV) radiation is known to have significant effects on the development and performance of plants, including flowers. In multiple species, UV‐absorbing floral patterns are associated with environmental conditions such as the solar UV exposure they typically receive. However, it is not known whether plants can increase the UV‐absorbing areas found on petals plastically when in a high‐UV environment. We grew Brassica rapa at three different UV radiation intensities (control, low, and high) and under two exposure duration regimes. We removed petals from flowers periodically during the flowering period and measured the proportion of the petal that absorbed UV. UV‐absorbing areas increased when plants were exposed to longer periods of UV radiation, and at high UV radiation intensities. UV‐absorbing area of petals of the UV intensity treatments decreased over time in long exposure plants. This study demonstrates that flowers can potentially acclimate to different UV radiation intensities and duration of exposure through an increase in UV‐absorbing areas even after a relatively short exposure time to UV. Such a rapid plastic response may be especially beneficial for dynamically changing UV conditions and in response to climate change.


| INTRODUC TI ON
For many years, we have known that ultraviolet (UV) stress is induced by exposure to relatively high intensity of UV radiation and/ or low intensity of photosynthetically active radiation (PAR; Ballaré et al., 2011;Jansen et al., 1998). This can translate into a reduction in biomass accumulation, reduction in photosynthesis, damaging large macromolecules such as DNA and other proteins, and inducing significant oxidative stress within the vegetative tissues of plants (Frohnmeyer, 2003;Jansen et al., 1998;Rozema et al., 1997).
UV-B is a portion of solar radiation that is biologically active, and can induce profound changes in gene expression, physiology, and morphology of plants (Jansen, 2002). Most of the solar UV-B radiation that reaches the Earth's surface is absorbed by the stratospheric ozone layer, and atmospheric ambient UV-B radiation varies depending on factors such as season, latitude, altitude, and time of day (McKenzie et al., 2007). More importantly, it is becoming well established that ozone depletion coupled with climate change is increasingly modifying the timing and duration of UV-B radiation exposure for plants (Bornman et al., 2019). Recent research indicates concurrent exposure to UV-A and PAR can minimize the detrimental effects of high UV-B (Verdaguer et al., 2017). It is now becoming clear that broad spectrum UV radiation may influence plant metabolism and other important processes (Jansen & Bornman, 2012;Jenkins, 2009), and accumulated evidence suggests that, under realistic exposures scenarios, the negative effects of UV-B on plant performance are indeed relatively minor or non-existent (Fiscus & Booker, 1995).
In most instances, low UV intensity can be characterized as regulatory or "eustress" to plants (i.e., good stress, Hideg et al., 2013), inducing plastic responses in both vegetative and reproductive plant tissues (Jansen, 2002;Llorens et al., 2015). In some cases, plastic responses to elevated UV such as rapid increase in flower production and earlier anthesis (Torabinejad et al., 1998;Ziska et al., 1992) are suggestive of a strategic stress response, increasing reproductive output before yielding to the stressor for example (Wada & Takeno, 2010). A good example of these plastic responses includes increased accumulation of UV-absorbing compounds (such as flavonoids) in the pollen walls of plants grown in UV-B treatments that are significantly higher than typical ambient levels (Demchik & Day, 1996), or increased accumulation of anthocyanins and flavonoids in flowers (Ben-Tal & King, 1997;Hennayake et al., 2006), overall producing petals with more intense color than petals of flowers grown in UV-absent conditions (Ben-Tal & King, 1997).
Research into UV floral pigmentation has identified UV-B as a potential selective agent of the UV-absorbing areas (bullseye) floral pattern (Koski & Ashman, 2015). The bullseye is a pattern formed by the arrangement of UV reflecting petal apices and UV-absorbing petal bases (Guldberg & Atsatt, 1975;Harborne & Nash, 1984). UV bullseyes, and their ecological development, have been traditionally understood as the result of pollinator-driven interactions. Much of the floral diversity we see today, and in particular in UV bullseye pattern, indicates that UV bullseyes function to aid pollinator perception, orientation, visitation, foraging efficiency, and ultimately plant fitness (Koski & Ashman, 2013;Rae & Vamosi, 2013;Sheehan et al., 2016).
However, the recent findings that Argentina anserina (Rosaceae), a globally widespread plant, produce flowers with a higher proportion of UV-absorbing area (UV proportion ) with decreasing latitude (UV-B irradiance is greater at low latitudes), and that the UV proportion of 177 species within the Potentillae correlated with broad geographic variation in UV-B irradiance (Koski & Ashman, 2016) suggests that UV-B could be an important overlooked abiotic driver of phenotypic variation in plants (Koski & Ashman, 2015), even though this has not been started to be experimentally tested until recently (Finnell & Koski, 2021).
Flowers producing UV bullseyes with greater UV proportion are thought to be the object of selection because they significantly improve their male fitness in elevated UV-B environments (Koski & Ashman, 2015). Flowers with more UV reflectance are hypothesized to experience decreased pollen viability via diffuse reflection of UV onto pollen-bearing anthers; conversely, flowers with larger bullseye protect pollen by absorbing UV radiation (Koski & Ashman, 2015).
The extent of protection is such that the germination rate of pollen from flowers in relatively high UV conditions, but with large bullseyes, is comparable with that of flowers grown in UV-absent conditions (Koski & Ashman, 2015;Peach et al., 2020), even though this may not be ubiquitous in plants. It should be also noted that these previous studies could not elucidate whether UV-B irradiance was the ultimate factor mediating this response in the wild. Many other environmental variables also change with latitude (i.e., temperature, PAR, seasonality) and there are strong links between UV-B radiation and climate change (Bornman et al., 2019). Therefore, in the present study, we tested whether UV proportion variation directly reflects changing UV-B radiation intensity and duration of exposure during plant growth using Brassica rapa flowers, a species known to show UV proportion variation among genotypes and cultivars (Yoshioka et al., 2005). Because previous studies have indicated that UV bullseyes are the target of selection, functioning in part to protect pollen from harmful levels of UV-B damage (Koski & Ashman, 2015;Peach et al., 2020), it now seems pertinent that we examine whether plants have the capacity to increase the UV-absorbing areas of their petals in response to increased exposure and intensity of UV-B radiation. This capacity may be especially important if the pace of climate change exceeds the speed that wild plants can evolve these features. We predict that UV-absorbing areas are phenotypically plastic, and expect that UV proportion will be greater in the petals of plants exposed to high UV-B radiation intensities. B. rapa is insectpollinated and represents a valuable agro-economic resource and improved knowledge of its floral response(s) to UV-B may inform about the appropriate use of genotypes and cultivars on a regional to global scale, where UV radiation varies according to both latitude and season (Herman, 2010).

| Study species
Field Mustard B. rapa L. (Brassicaceae) is an annual crucifer native to Europe. It has been extensively cultivated, and is now distributed globally, and found on waste ground, roadside, and both river and stream bankside habitats. To the human eye, the flowers have four bright yellow petals which are arranged symmetrically, with six stamens in the center. In addition, B. rapa has a clear UV-bullseye (Yoshioka et al., 2005). To test whether UV is a driver of phenotypic plasticity in a specific floral trait (UV bullseye), we used the rapidcycling B. rapa, developed by Williams and Hill (1986) via several cycles of selection. Compared to their wild relatives, rapid-cycling B. rapa plants flower earlier (~18 days from planting), producing up to 20-25 flowers per plant.

| Experimental design: plants
Individual B. rapa plants were exposed to one of two periods of UV exposure prior to anthesis: (i) long-term UV exposure (~24 days) and (ii) short-term (~6 days) UV exposure ( Figure 1).
For the long-term UV exposure, 54 plants were grown in March 2018 from seed within two SANYO MLR-351 H, Panasonic plant chambers (18 plants per UV intensity treatment; Figure 2). The total duration of the experiment was set at 75 days. The seeds were potted in pots (7 cm × 7 cm × 7 cm) containing compost and vermiculite (both Verve), and were randomly assigned to a UV intensity treatment (control, low, high, see Section 2.3).
For the short-term exposure, B. rapa seeds were potted and grown (soil and pots used as above) in a glass greenhouse for 18 days prior to placing them in the chambers for 57 days (Figure 2), so the experiment was terminated when plants from both long-term and short-term exposure were 75 days old to allow meaningful comparisons. The short duration of UV exposure (~6 days) was defined as 25% of UV radiation (in days) received by the long exposure plants prior to flowering. Mean day of first flowering in the long-term UV exposure experiment was day 24; therefore, we estimated that the plants should receive approximately 6 days of UV radiation in the growth chambers before flowering would begin. On day 18 of growth, 51 plants growing in the greenhouse were randomly assigned to UV treatment groups and transferred to the growth chambers (Figures 1 and 2). Daily Solarmeter® (Model 6.2 Sensitive UV-B Meter) measurements indicated the greenhouse UV-B irradiance was 0 W/m 2 for the duration of growth prior to transfer to the plant chambers. The growth conditions were set as the above long-term UV exposure experiment, and the plants were grown for a further 57 days.

| Experimental design: UV-B intensity treatments
Three UV radiation intensities were administered across two plant chambers (SANYO MLR-351 H, Panasonic; Figure 2). Treatments were (i) control (no UV exposure), (ii) low UV exposure, and (iii) high UV exposure. Low UV intensity was achieved in one chamber using 15 × 36 W 6% UV-B bulbs (ArcadiaReptile) and for the high UV intensity, a second chamber was fitted with 15 × 36 W 12% UV-B bulbs (ArcadiaReptile). In both treatments, lamps provided in addition to 6% or 12% UV-B the full light spectrum +30% UV-A as specified by the supplier (ArcadiaReptile). These percentages represent the total spectral power output that falls within the ranges of UV-B  (Baker et al., 1997). UV treatments were significantly different (F 2,30 = 377.8, p < .001), and Tukey multiple pairwise comparisons indicated that all three UV treatments significantly differed from one another (all p < .001).
All plants were exposed to 12 h light/12 h dark cycle, 23°C light/20°C dark cycle, and 60% relative humidity within their chambers, and were watered every second day for the duration of the experiment. The control plants and boxes were rotated between chambers fortnightly and plants in the low and high UV treatments F I G U R E 1 Schematic diagram showing the duration (black boxes) of the long-(24 days) and short-term (6 days) UV exposure prior anthesis in the experimental plants. Long-term UV-exposed plants were grown exclusively in the growth chambers (yellow box) for the duration of the study. Short-term exposure plants were grown for 18 days in a glass greenhouse (blue box) before moving them in the growth chambers (yellow box). UV, ultraviolet.
were rotated and assorted within and across shelves within their

| Flower collection and petal UV photography
In each long-term and short-term UV exposure plants, all flowers were removed upon anthesis, noting plant age (i.e., days since planting). The natural curvature of B. rapa petals made it difficult to accurately measure the UV-absorbing part of the petals, so we removed the petals from each flower and pressed them in a plant-press in the dark for at least 7 days, ensuring that all petals were adaxially presented and evenly flattened. Koski and Ashman (2013) indicated that this method of drying and pressing specimens to measure the size of UV-absorbing area does not influence measurements, which we corroborated for our species too.
A total of 726 pressed flowers were photographed in UV-A light.
All photographs were taken with a Canon EOS 400D camera, fitted with a Baader UBVRI 1¼′′ Photometric U-filter. The filter permits the transmission of UV-A light between 320 and 390 nm, peaking with 85% transmission at 350 nm, while removing infrared and daylight wavelengths. The availability of naturally occurring UV light in the daylight spectrum can be limited indoors; the image field was therefore illuminated with a UV-A torch (Convoy S2+ Nichia UV waterproof LED flashlight) with a peak UV emission at 365 nm. The camera was mounted on a tripod and arm to achieve an overhead view of the pressed specimen.
All photographs included a standard scale (i.e., ruler, Figure 3), and the UV torch was also positioned approximately 15 cm above the specimen. For all UV photographs, the exposure was 10 s, and the aperture f/5.6. Alongside each image, we include a piece of paper covered in sunscreen which appeared dark in images as it absorbed UV.
From the 726 pressed flowers, the total UV-absorbing area (UV area , mm 2 ) and total petal area (mm 2 ) could be accurately mea-

| Statistical analysis
We first tested whether the total number of flowers and duration of flowering of each plant varied with UV-intensity treatment (control, low, and high) and exposure length (short and long), using a two-way Kruskal-Wallis test.
We then tested variation in UV proportion and petal area, which were modeled using a linear mixed-effects model (LMM), with individual plant included as a random effect, with flower as a random effect nested within individual plant. We first tested the relationship between radiation intensity treatment (control, low, and high), exposure length (short and long), and their interaction on UV proportion and petal area. UV proportion was logit-transformed using the car package (Fox & Weisberg, 2019), to improve the normality of model residuals. Following this analysis, we also included plant age (days since planting) as a quadratic covariate to investigate further how this may affect petal responses to increase UV-B. We included the quadratic relationship based on a reviewer's suggestion, and based on likelihood ratio tests, the models performed significantly better than plant age as a linear covariate. We again tested whether the inclusion of interactions between flowering date, treatment and exposure increased model fit, and the likelihood ratio tests supported the inclusion of the two-way interactions over other model types.
LMM modeling was carried out using the statistical package "lme4" (Bates et al., 2015) run in R version 3.6.3 (R Development Core Team, 2020). Post hoc testing between factors was carried out using the package "emmeans" (Lenth, 2018).

| Effects of UV intensity and duration of exposure on flower production and duration
There was a significant effect of both duration of exposure (H = 7.13,  UV light (b, c). The UV proportion (%) of petals is the measure of (c) UV area (mm 2 ) relative to (b) total petal area (mm 2 ). UV, ultraviolet.

| Changes in petal UV proportion and petal area in relation to UV intensity and duration of exposure
There was a significant effect of UV intensity treatment and exposure on the UV proportion of petals, but no significant effect of the interaction (Table 1). For long exposure plants, those grown in the high intensity treatments had significantly greater UV proportion than both control (t-ratio = −3.28, p = .005) and low intensity (t-ratio = −3.68, p = .002) plants, with no significant difference between control and low intensity plants (t-ratio = 0.13, p = .991; Figure 6). Among short exposure plants, there were no differences between intensity treatments detected (all p > .9). Overall, the UV proportion between them was greater in long exposure plants compared to short exposure plants ( Figure 6).
For petal areas, there was a significant effect of exposure, but not UV intensity treatment (Table 1); the petals of short exposure plants were approximately 13% larger than in long exposure plants (10.55 ± 0.35 vs. 9.35 ± 0.38 mm 2 ).

| Changes in UV proportion and petal area in relation to plant age
For UV proportion , there were significant two-way interactions between plant age and exposure and plant age and UV intensity treatment (Table 2). Petals had significantly smaller UV proportion as plants became older in long exposure plants, but not short exposure plants ( Figure 7a). UV proportion also became smaller with plant age in high intensity UV treatments, with a decrease also in low or control plants that was however not statistically significant ( Figure S1).
For petal area, there were similar significant two-way interactions between plant age and exposure, and plant age and UV intensity (Table 2). Petal area became smaller with plant age in short exposure plants, but not in long exposure plants (Figure 7b). Petal area also declined in high UV intensity treatments, but did not change in control or low UV intensity plants ( Figure S2).

| DISCUSS ION
The reproductive structures of plants and especially pollen are sensitive to UV radiation (Feng et al., 2000). Several studies have shown that one of the most effective defensive mechanisms in flowers against UV-induced stress is the accumulation of a wide range of UV-absorbing phenolic compounds in the petals and in other structures such as pollen (Demchik & Day, 1996;Jenkins, 2009;Rozema et al., 1997). However, it is not clear whether plants can increase these UV-absorbing plastically when in a high UV environment.
In this study, we found that B. rapa plants decreased flower production when grown with a short UV exposure and especially under high UV intensity, which also decrease the time they were flowering F I G U R E 5 Mean ± SE number of flowers produced for control, low, and high UV intensity treatments Brassica rapa plants from short and long UV exposure treatments. SE, standard error; UV, ultraviolet.

TA B L E 1
Petal UV proportion and petal area (mm 2 ) in relation to UV exposure (short/long) and UV intensity treatment (control, low, and high). larger bullseyes in artificial flowers provide enhanced protection to UV sensitive pollen, as compared with flowers with smaller UV bullseyes (Koski & Ashman, 2015). Thus, our findings suggest that the UV proportion of B. rapa plants can respond plastically to increase UV absorption in response to UV radiation, demonstrating for the first time that petals are capable of plastic change in response to ambient UV radiation even after a relatively short exposure time.
We also found that longer exposure increased overall UV proportion in petals. It is possible that increased duration of UV exposure early in plant growth (i.e., from seedling stage) may be important for the production of flavonoids that are responsible for the UV-absorbing pigments in floral tissues. While UV exposure early in plant development may increase flavonoid accumulation in both vegetative tissues of B. rapa, and later in floral tissues, it is at present unclear by what mechanism exposure of vegetative tissues to UV affects later floral development (Musil & Wand, 1993, but see Llorens et al., 2015. Nevertheless, exposure to UV during seedling development may be important in altering floral tissue components (Yao et al., 2015).
Besides affecting UV proportion , UV intensity also affected petal size in our study. Overall petal area was smallest in long exposure treatments, while no significant difference was observed between UV intensity. In comparison, previous studies have reported a range of negative to positive effects or no changes in petal area in response to UV-B (see Llorens et al., 2015 and references therein). As well as study-specific differences (UV dose, exposure, and measurement), evidence suggests that how plants respond morphologically is also species specific and depends on the parameter measured.
This is probably because UV-mediated morphogenesis is the result of multiple compensatory molecular and physiological mechanisms that interact with other environmental variables (Barnes et al., 1990;Jansen et al., , 2019.  et al., 1996), the observed increase is likely a time lagged response between stress and effect on phenotype, as UV-exposed plants balance the damage, repair, acclimation, and adaptation responses at different molecular and physiological levels, which require different tempos (Jansen et al., 1998), translating into responses to UV modulated by organ/organism ontogeny.

| CON CLUS IONS
The plastic responses of floral UV-absorbing area demonstrated in this study indicate the potential for plants to persist when faced with environmental UV-B change by decreasing flower production but enhancing petal UV proportion in these flowers. We identified a significant positive correlation between the intensity and duration of UV exposure and UV proportion in B. rapa. Even though not tested specifically here, these responses may be important for the hypothesized reduction in exposure to UV radiation of the floral reproductive structures caused by a reduction in the proportion of reflected UV radiation that they receive as a consequence of the increased UV bullseye. Moreover, our study demonstrates that flowers can potentially acclimate to different UV radiation intensities and duration of exposure through an increase in UV-absorbing areas. Such a rapid plastic response may be especially beneficial in regard to climate change, even though field studies involving multiple species and environmental factors should be employed to fully understand the complexity of UV response in plants.

ACK N OWLED G M ENTS
The authors thank the Douglas Bomford Trust for funding this work, Natalie Ben Alaya and Marco Perez Sanchez for support with watering and plant maintenance, and four previous reviewers for insightful comments on a previous draft.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available at