Little overlap between transcripts differentially expressed on the growing sides of stems during phototropism and autostraightening

Heliotropic plants bend from east to west during the day then from west to east during the night. We previously demonstrated a role for the circadian clock in this process [22], but the involvement of other signaling pathways has not been extensively investigated. We wished to explore the possibility that daytime movements rely upon the phototropin signaling pathway and that nighttime reorientation involves an autostraightening response. Autostraightening, also known as autotropism, has been most widely studied in plant organs undergoing a gravitropic response [17,28]. To investigate whether sunflower stems responding to a light cue also show this behavior, we exposed 2-week-old sunflower plants grown in light/dark cycles to unidirectional blue light and monitored stem angle over many hours (Fig 1A). We found that while plants initially bend rapidly towards the light source, after around 10 hours of continuous illumination, they indeed bend away from the light in an autostraightening response (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Phototropic and autostraightening response in sunflower stems.

(A) Schematic of experimental design. (B) Changes in mean stem angles over time for 2-week old sunflower plants stimulated with unidirectional blue light at 3 hours after dawn (n = 3). Arrows indicate times of tissue collection for phototropism RNA-seq. (C) Change in mean stem angles over time for 2-week old sunflower plants stimulated with unidirectional blue light starting at different times of day (n = 6–9). White and gray areas represent subjective day and subjective night. Ribbons represent SEM. (D) Numbers of genes significantly more highly expressed on the lit (blue) or shaded (red) sides of stems are indicated. Significance determined using FDR < 0.05, based on pairwise comparisons of gene expression on the shaded and lit sides of the stem at indicated times. (E, F) Normalized CPM of BZR1-like (Ha412HOChr14g0662481) and IAA14-like (Ha412HOChr03g0143141) over time on shaded and lit sides of stems (mean +/− SEM, n = 6). (G) GO enrichment analysis of genes differentially expressed genes during phototropism at ZT4 and ZT15. All terms for processes involved in light signaling, hormone regulation, and growth are shown for lit and shaded sides of stems. The underlying raw data may be found in S1 Data for Fig 1D; in S2 Data for Fig 1G; in S13 Data for Fig 1B and 1C; and at NCBI GEO, accession GSE229654, for Fig 1E and 1F. CPM, counts per million; FDR, false discovery rate; GO, gene ontology; SEM, standard error of the mean; ZT, Zeitgeber Time.

https://doi.org/10.1371/journal.pbio.3002344.g001

We previously showed that sunflower phototropism is gated by the circadian clock, such that the degree of bending towards a unidirectional blue light over a 4-hour period is dependent on the time of day the stimulus is given [22]. We therefore next tested whether the autostraightening response is also time-of-day dependent: We exposed 2-week-old chamber-grown sunflowers to unidirectional blue light at 6 different time points, ranging from the normal time of lights on (Zeitgeber Time 0, or ZT0) to lights off (ZT16) (Fig 1C). Sunflowers exposed to unidirectional blue light during the first part of the day (ZT0, ZT3, and ZT6) displayed a similar autostraightening response, in all cases beginning at around ZT14. However, plants exposed to unilateral light in the later part of the day (ZT9, ZT12, and ZT16) did not undergo autostraightening at ZT14, nor did they autostraighten when they reached a maximum bending angle many hours later. This suggests autostraightening is gated by the circadian clock, such that the timing of the onset of phototropic bending influences the ability of the plant stem to autostraighten.

We next wished to identify genes with expression patterns correlated with sunflower autostraightening and phototropism. Many signaling components involved in phototropism have been characterized; however, most of this work has been done using the hypocotyls or coleoptiles of etiolated plants [7]. Moreover, little is known about the molecular mechanisms involved in autostraightening [17]. To identify candidate genes and pathways involved in sunflower stems undergoing phototropism and subsequent autostraightening, we performed RNA sequencing. Two-week-old light-grown plants were placed in front of unidirectional blue light at ZT3, a time permissive for robust phototropism and subsequent autostraightening. Stem peels, consisting of epidermal, cortex, and vascular tissue (S1 Fig), were collected from the lit and shaded sides of bending stems at 9 time points over an 18-hour period (Fig 1B). RNA was extracted, libraries constructed, and whole transcriptome sequencing was performed using the Illumina HiSeq platform. Reads were mapped to the Ha412HOv2 genome using the Eugene annotation [29]. Genes differentially expressed across the stem were determined by pairwise comparisons of gene expression between the lit and shaded sides at each time point. We found thousands of genes are differentially expressed across the 2 sides of stems, starting at 1 hour after exposure to the blue light and persisting until the end of our experiment (Fig 1D and S1 Data).

To implicate distinct biological processes in these bending responses, gene ontology (GO) enrichment analysis was performed on genes differentially expressed at each time point. As expected, genes rapidly up-regulated on the lit sides of stems were enriched for homologs of genes involved in blue light and UV responses (Fig 1G). As also expected based on previous studies of phototropism [12,13,30], homologs of auxin-responsive and auxin signaling genes were significantly enriched among transcripts quickly up-regulated on the shaded sides of stems (Figs 1G and S2 and S2 Data; note that sunflower gene IDs for every enriched GO term are listed in these files). The same was true for genes predicted to act in the signaling pathways of other growth-promoting hormones such as brassinosteroid and gibberellic acid (Fig 1E and 1G). Consistent with the onset of bending occurring within 1 hour after light stimulation (Fig 1B), homologs of genes involved in cell growth and cell wall processes were also enriched among those rapidly up-regulated on the shaded sides of stems. Overall, these data are consistent with previous studies suggesting that phototropic bending towards blue light is mediated by the redistribution of auxin towards the shaded sides of stems [12,31].

Interestingly, we also observed a highly significant enrichment of genes implicated in translational and ribosomal processes among those up-regulated on the shaded sides of stems between ZT7 and ZT15 (S3 Fig). The lag between the enrichment of these terms and both the onset of growth (Fig 1B) and the enrichment of hormonal and cell wall–related terms (S3 Fig) suggests that induction of the translation machinery occurs in response to growth rather than as its driver.

We next examined differential gene expression across stems during the autostraightening response, 12 hours after exposure to unidirectional blue light (ZT15) (Fig 1B). Since autostraightening is driven by the faster elongation of cells on the lit sides of stems and phototropism by the faster elongation of cells on the shaded sides of stems [16,32], we speculated that similar growth pathways might be involved in both processes. Indeed, we found homologs of auxin-responsive genes enriched among transcripts up-regulated on the lit sides of stems at ZT15, similar to the enrichment of these types of genes on the shaded sides of stems at ZT4 (Fig 1F and 1G). However, we did not find enrichment for homologs of genes involved in auxin signaling, or brassinosteroid or gibberellic acid signaling or response, among autostraightening genes. Moreover, a direct comparison of genes differentially expressed at ZT4 and ZT15 revealed that few genes were up-regulated both on the shaded sides of stems at ZT4 and on the lit sides of stems at ZT15 (S4 Fig and S3 Data). The limited overlap between enriched functional categories (Figs 1G and S2) and specific genes (S4 Fig) at the phototropic and autostraightening time points suggests that different growth pathways may control these different bending responses.

The majority of genes expressed in heliotropic stems have diel rhythms with peak expression occurring either before dusk or dawn

Having identified genes implicated in phototropism and autostraightening, we next wanted to investigate patterns of gene expression in plants undergoing heliotropism in natural growth conditions (Fig 2A). We collected peels every 4 hours over 2 days from the east and west sides of 2-week-old sunflowers exhibiting robust solar tracking in the field (Fig 2B). RNA expression analysis was carried out as described for phototropic stems. To visualize the difference in expression profiles between samples, we performed principal component analysis (PCA) [33]. Both PC1 (22.2%) and PC2 (19.8%) separate the samples by time of day, suggesting that time of day rather than side of stem or date of collection is the most influential factor on gene expression in this experiment (Fig 2C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Transcriptome analysis of solar tracking sunflower stems.

(A) Schematic showing orientation of heliotropic plants at dawn and dusk. (B) Stem angles of 2-week-old sunflowers growing in the field (mean plotted with ribbon representing SEM, n = 6). White and gray areas represent day and night. (C) PCA analysis of gene expression during heliotropism. (D) Number of genes with statistically significant diel rhythms on east or west sides of stems (FDR < 0.05). (E, F) Normalized expression over time for LHY-like (Ha412HOChr15g0736201) and PRR5-like (Ha412HOChr08g0350221) transcripts during heliotropism (mean +/− SEM, n = 3). (G) Acrophase distribution of genes that are significantly rhythmic on both sides of sunflower stems. (H) Acrophases of genes on the east and west sides of stems. Genes with an acrophase difference greater than 3 hours are highlighted in red. (I) Amplitudes of significantly rhythmic genes on the east side and west sides of stems. Genes with amplitude differences greater than 2-fold across the stem are highlighted in red. (J) Numbers of genes significantly more highly expressed on the east or west sides are indicated (FDR < 0.05). Pairwise comparisons of gene expression on the east and west sides of stems at indicated times. (K, L) Normalized expression of XTH23-like and EXPL1-like transcripts during heliotropism (mean +/− SEM, n = 3). (M) GO enrichment analysis of genes differentially expressed during heliotropism. All terms shown in Fig 1G are included here for direct comparison. (E, F, K, L) Mean values are plotted +/− SEM, n = 3. The underlying raw data may be found in S4 Data for Fig 2D and 2G; in S6 Data for Fig 2H; in S7 Data for Fig 2I; in S8 Data for Fig 2J; in S9 Data for Fig 2M; in S13 Data for Fig 2B; and at NCBI GEO, accession GSE229654, for Fig 2C, 2E, 2F, 2K, and 2L. FDR, false discovery rate; GO, gene ontology; PCA, principal component analysis; SEM, standard error of the mean.

https://doi.org/10.1371/journal.pbio.3002344.g002

To examine daily expression patterns, the Cosinor method available in the R package DiscoRhythm was used to estimate the acrophase, amplitude, and rhythmicity of each expressed gene [34,35]. Of the 34,071 expressed genes, 21,325 (62.6%) were assessed as having a diel rhythm (qvalue < 0.05) on both the east and west sides of the stem (Fig 2D and S4 Data). Although some genes appeared to be uniquely rhythmic on one side or the other, most of those are removed if more stringent statistical cutoffs for rhythmicity are applied, suggesting these may represent transcripts with marginal rhythmicity. Therefore, only the 21,325 genes scored as significantly rhythmic on both the east and west sides of stems were used for acrophase and amplitude analysis. These high-confidence rhythmic genes include homologs of core clock genes, LHY-like and PRR5-like [36], which displayed strong daily rhythms with similar peak phases to their Arabidopsis homologs [37,38] (Fig 2E and sF). Importantly, there was no significant difference in expression across the 2 sides of the stem either for these or other homologs of core clock genes (Figs 2E, 2F and S5). These data indicate that the circadian clock is robustly rhythmic and entrained to the same phase on the opposite sides of solar tracking stems.

We next examined the overall patterns of expression for genes with diel patterns of expression on the east and west sides of stems in field-grown sunflowers. Of these high-confidence rhythmic genes, the majority had peak expression either before dusk or before dawn (Fig 2G). To identify functional categories enriched among genes with peak expression at distinct times of day, we placed genes into 4-hour bins based on their times of peak expression and performed GO analysis (S5 Data). As expected, genes with peak expression in the early morning (ZT0 to ZT4) were enriched for terms involved in the response to different wavelengths of light and in photoprotection. Genes with peak expression in the first part of the day (ZT4 to ZT8) were enriched for heat response and protein folding, suggesting up-regulation of pathways involved in abiotic stress responses at this time. Genes with peak expression in the afternoon (ZT8 to ZT12) and in the evening (ZT12 to ZT16) were enriched for functions in DNA repair, DNA replication, and the cell cycle, suggesting cell division might be preferentially phased towards the end of the day. Intriguingly, we also found enrichment of genes implicated in RNA modification, especially in chloroplasts and mitochondria, in the afternoon and evening. RNA editing is essential for organelle biogenesis [39], suggesting that replication of endosymbiotic organelles may be coordinated with that of the nuclear genome. Homologs of brassinosteroid signaling components were enriched during the evening and early night (ZT12 to ZT16 and ZT16 to ZT20). This steroid hormone has been shown to play epidermal- and vasculature-specific roles in local and systemic signaling [40–43]; it is possible that one or more of these processes is controlled in a time-of-day–dependent manner. Plant defense-related terms were enriched in the early night (ZT16 to ZT20), consistent with reports of reduced plant susceptibility to fungi and bacterial pathogens in the late night and at dawn compared to near dusk [44–46]. We also found enrichment for autophagy and cell growth terms in the early night (ZT16 to ZT20) and for auxin-related terms in the late night (ZT20 to ZT24). Overall, these data reveal that the majority of the stem transcriptome is rhythmically expressed in field-grown plants and suggest ways in which plant physiology may be temporally partitioned in natural conditions.

Transcriptome analysis implicates auxin signaling in heliotropic growth at night but not during the day

We next wished to identify genes with different expression patterns on the east and west sides of heliotropic stems. For most genes, the diel expression patterns across the stem are quite similar, consistent with the near-identical patterns of expression of clock genes on the east and west sides of stems (Figs 2E, 2F and S5). However, when comparing the times of peak expression of transcripts significantly rhythmic on the 2 sides, we found that 1,369 genes have acrophase differences of greater than 3 hours on opposite sides of the stem (Fig 2H and S6 Data). Many of these, for example, XTH23-like, appear to have different waveforms on the 2 sides as well (Fig 2K). We next compared the rhythmic amplitude of transcripts on the east and west sides of stems (Fig 2I and S7 Data). As was true for our phase analysis, most genes were very similar, with only 811 having a greater than 2-fold difference in amplitude across the stem. Some of these transcripts, such as EXPL1-like, have much higher overall expression levels on one side of the stem compared to the other (Fig 2L).

We next compared expression across the stem at individual time points to identify genes and pathways associated with heliotropic movement. This analysis revealed relatively few differentially expressed genes early in the day, at ZT0 and ZT4, but a larger number of transcripts differentially expressed in the afternoon and early evening, ZT8 and ZT12. Surprisingly, we found the largest numbers of differentially expressed genes at night, ZT16 and ZT20 (Fig 2J and S8 Data), times when there cannot be unequal illumination across the stem. These data demonstrate there is significant differential expression across the stems of solar tracking plants, but not phased to the times of day we had expected.

We previously showed that the east sides of solar tracking stems grow faster than the west sides during the day while the west sides grow faster than the east sides at night [22]. To identify biological processes correlated with these growth patterns, we next performed GO analysis on genes differentially expressed at each time point. Since previous studies, along with our own findings, have suggested involvement of light responses, hormone signaling, and growth-related processes in phototropism [9], we focused on these terms. We found enrichment for blue light and UV responses on the east sides of stems in the early morning (ZT0 and ZT4), as well on the west sides in the early afternoon (ZT8), consistent with the position of the sun relative to plant stems at these times (Fig 2A and 2M). We next considered enriched growth-related GO categories. As expected, we found enrichment for cell wall–related processes at ZT8 and ZT12 on the east side, and at ZT16 and ZT20 on the west side, coinciding with the times of antiphasic growth on the opposite sides of the stem (Fig 2B). Surprisingly, however, auxin signaling and response terms were not enriched among genes more highly expressed on the east side during the day. However, these terms were enriched among genes more highly expressed on the west side in the evening and early night (ZT12 and ZT16) (Fig 2M and S9 Data). These results suggest a potential role for the auxin signaling pathway during nighttime west-to-east reorientation, but not during daytime east-to-west growth.

Transcriptional responses to phototropism and heliotropism are quite distinct

To investigate whether phototropism towards blue light and heliotropism in the field rely on similar molecular pathways, we next systematically compared the patterns of gene expression occurring during these 2 processes. First, we asked if genes with different diel phases across the stem during heliotropism (Fig 2H) might also be differentially expressed across the stem during phototropism. We examined the expression of S15a-like, which encodes a protein with homology to Arabidopsis thaliana ribosomal protein S15a (AT1G07770). During heliotropism, this gene has a phase difference of 4.2 hours between the east and west sides, with a peak expression later in the day on the east side of the stem compared to the west (Fig 3A). This gene is also differentially expressed in the chamber during phototropism, with increased levels on the shaded side of the stem between 4 and 12 hours after the onset of directional illumination (Fig 3B). To determine if this pattern is consistent for other genes with different phases of peak expression across heliotropic stems, we subdivided all differentially phased genes into 3 subcategories based on times of peak expression (Figs 3C and S6). Similar to S15a-like, the normalized average expression of all genes with comparable peak phases of expression in the field is higher on the shaded side of the stem 4 to 12 hours after the onset of directional illumination during phototropism (Fig 3D). Gene expression patterns during phototropism were similar for the 2 other groups of transcripts with different phases of expression across heliotropic stems (S6 Fig). In all cases, up-regulation on the shaded sides of stems occurs many hours after the initial phototropic bending (Fig 1B). GO analysis of transcripts in all 3 phase groups revealed enrichment for terms involved in ribosomal biogenesis, translation, and DNA replication (S10 Data). This enrichment suggests roles for these genes in cellular growth induced in response to bending, rather than for roles in the initiation of this tropic response. Overall, transcripts expressed with different phases across solar tracking stems are correlated neither with the rapid onset of growth on the shaded sides of stems nor with autostraightening growth on the lit sides of stems.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Distinct patterns of gene expression during phototropism and heliotropism.

(A, B) Normalized transcript levels over time for S15a-like (Ha412HOChr05g0204731) during heliotropism and phototropism, respectively. (C) Genes with different acrophases on the east and west sides of stems; west side acrophases between ZT0 and ZT8. (D) Zscore-normalized mean expression of the differentially phased genes shown in (C) during phototropism. (E, F) Normalized transcript levels over time for LNG1-like (Ha412HOChr17g0822481) during heliotropism and phototropism, respectively. (G, H) Expression patterns of phototropic genes. Their Zscore-normalized mean expression during heliotropism and phototropism, respectively, are shown. (I, J) Normalized transcript levels over time for ARF19-like (Ha412HOChr03g0103791) during heliotropism and phototropism, respectively. (K, L) Expression patterns of autostraightening genes. Their Zscore-normalized mean expression during heliotropism and phototropism, respectively, are shown. (M) Numbers of genes more highly expressed on the growing sides of stems during phototropism (ZT4 shaded (ZT4_Sh) and ZT15 lit (ZT15_L)) and during heliotropism (ZT8 east (ZT8_E) and ZT16 west (ZT16_W)). (N) GO enrichment analysis of genes shared between any 2 of the growth categories shown in (M). Terms shown are those shared with Figs 1G and 2M. (A, E, I) Mean CPM +/− SEM, n = 3. (B, F, J) Mean CPM +/− SEM, n = 6. (G, K) Lines fitted with a generalized additive model. (D, H, L) Lines fitted with LOESS. Ribbons represent 95% confidence intervals. The underlying raw data may be found at NCBI GEO, accession GSE229654, for Fig 3A, 3B, and 3D-L; in S3 and S8 Data for Fig 3M; in S6 Data for Fig 3C; in S17 Data for Fig 3D, 3G, 3H, 3K, and 3L; and in S18 Data for Fig 3N. CPM, counts per million; GO, gene ontology; SEM, standard error of the mean; ZT, Zeitgeber Time.

https://doi.org/10.1371/journal.pbio.3002344.g003

Given that hundreds of genes have different amplitudes of expression on the opposite sides of heliotropic stems (Fig 2I), we investigated their expression patterns during phototropism. The normalized mean expression patterns of these transcripts in phototropic plants reveals higher expression on the shaded sides of stems 4 to 12 hours after the onset of directional illumination during phototropism (S7 Fig), with up-regulation occurring much slower than the onset of plant bending (Fig 1B). GO analysis of these transcripts did not reveal any significantly enriched functional categories. Although the pattern of expression of these genes during phototropism is similar to that observed for transcripts with different phases of expression in solar tracking plants (Figs 3D and S6), there is only limited overlap between genes expressed with different phases and amplitudes in the field (S7 Fig). As is true for the differentially phased genes, transcripts expressed with different amplitudes across solar tracking stems are not correlated with the rapid onset of growth on the shaded sides of stems or with autostraightening growth on the lit sides of stems.

Since we found little overlap between genes expressed with different rhythmic parameters on the opposite sides of heliotropic stems and genes differentially expressed during the onset of phototropic bending, we next examined how genes correlated with phototropic or autostraightening growth are expressed during heliotropism. One transcript rapidly up-regulated on the shaded sides of stems during phototropism is LNG1-like (Fig 3F), a homolog of the Arabidopsis gene LONGIFOLIA1 (AT5G15580), which promotes unidimensional cell growth and is induced rapidly in response to auxin [47–49]. During heliotropism, this gene has a similar waveform and phase on the east and west sides of stems but a higher amplitude on the west side (Fig 3E). This results in a modestly higher level of expression on the east side of the stem around midday (ZT4 to ZT8), and on the west side of the stem in the late day and early night (ZT12 to ZT16). To determine whether this pattern is consistent for all genes rapidly up-regulated on the shaded sides of stems during phototropism, we examined their mean normalized expression in field-grown plants undergoing heliotropism. We found the same pattern of modest amplitude differences across the stem observed for LNG1-like (Fig 3G). This pattern is quite distinct from the large differences in mean expression levels and kinetics observed for these genes across stems during phototropism (Fig 3H). Overall, genes rapidly up-regulated on the shaded sides of stems during phototropism show only limited differences in expression across the stems of solar tracking plants.

We next reasoned that genes implicated in autostraightening responses might be differentially expressed during solar tracking. We first examined ARF19-like, a transcript that encodes a protein with homology to auxin response factors (ARFs) and that is a likely activator of transcription [50]. This ARF-like transcript is up-regulated late in the phototropic time course on the lit sides of stems (Fig 3J), correlating with the timing of the autostraightening response (Fig 1B). ARF19-like has very similar expression patterns on the east and west sides of heliotropic stems (Fig 3I). However, inspection of the mean expression patterns of all transcripts correlated with autostraightening (Fig 3L) during heliotropism revealed a slightly advanced late-afternoon/early-night phase on the west side of solar tracking stems compared to the east side (Fig 3K). This pattern can also be observed for the ARF19-like transcript on the second day of the solar tracking time course (Fig 3I). Therefore, similar to our findings for genes correlated with phototropism, genes with expression patterns correlated with autostraightening have only subtle differences in expression across the stems of solar tracking plants.

Because phototropic and autostraightening genes displayed slight differences in amplitude and phase on the opposite sides of plant stems during heliotropism, we wanted to directly compare genes scored as differentially expressed during phototropism, autostraightening, or heliotropism at times of asymmetrical growth. We compared genes: up-regulated on the shaded sides of phototropic stems at ZT4 (phototropic genes), up-regulated on the lit sides of phototropic stems at ZT15 (autostraightening genes), up-regulated on the east sides of field-grown plants at ZT8 (daytime heliotropism), and up-regulated on the west sides of field-grown plants at ZT16 (nighttime heliotropism) (Fig 3M). Pairwise comparisons revealed small but statistically significant overlaps between all of these categories (S1 Table). To investigate possible biological functions of genes shared between 2 categories, we next performed GO analysis. Transcripts present in both the phototropic (ZT4, shaded sides of stems) and daytime heliotropic (ZT8, east sides of stems) categories are enriched for unidimensional cell growth, cell wall organization, and xyloglucan metabolic processes (Fig 3N). However, auxin-related terms are not enriched among genes differentially expressed in both conditions. In contrast, transcripts differentially expressed in both the phototropic (ZT4, shaded sides of stems) and nighttime heliotropic (ZT16, west sides of stems) conditions are enriched for auxin response and auxin signaling terms. These auxin terms are also enriched among the transcripts differentially expressed in both the autostraightening (ZT15, lit sides of stems) and nighttime heliotropic (ZT16, west sides of stems) conditions. These data suggest that although there is significant overlap between transcripts differentially expressed during heliotropism and phototropism, those shared between phototropic and daytime heliotropic growth are more likely involved in carrying out growth programs. In contrast, transcripts shared between phototropic, autostraightening, and nighttime heliotropic reorientation are more likely responsible for initiating growth.

Gene expression analysis suggests the onset of heliotropism and blue light–dependent phototropism are regulated by distinct pathways

We next wondered if the limited similarity between genes differentially expressed during daytime heliotropism and in phototropism was because the latter process is an acute response to light while the former involves the circadian system [22]. While sunflowers grown in a chamber with overhead light do not bend [51], heliotropic-type movements can quickly be induced using directional lights sequentially turned on and off to mimic the movement of the sun [22]. In addition, rhythmic bending can be observed for several days after plants are moved from the field into constant environmental conditions. Thus, while directional light is necessary to initiate tracking, and later to reinforce it, it is not absolutely required for heliotropic growth patterns. We therefore asked whether the onset and continuation of solar tracking are characterized by different transcriptional programs and in particular whether the initial onset is transcriptionally similar to phototropism in a growth chamber. We therefore characterized the growth and gene expression patterns of plants at the onset of heliotropism. To this end, 2-week-old chamber-grown plants were transplanted into the field after dusk and imaged for 3 days. During the first day in the field, sunflowers bent east a small amount in the morning and bent west to a greater degree in the afternoon. To our surprise, these plants reoriented during the first night so that they were inclined to the east before dawn (Fig 4A). These data suggest that heliotropism begins the first day young plants are transferred to the field.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Initiation of heliotropism shows a distinct transcriptional profile and occurs in different light conditions.

(A) Stem angles of 2-week old chamber-grown sunflowers transplanted into the field after sunset (mean +/− SEM, n = 17). (B-H) Normalized expression of selected phototropically regulated genes during the first 3 days in the field (mean +/− SEM, n = 3, except ZT64 west, n = 2). Statistical significance calculated using Welch’s t test, * represents qvalue < 0.05; † represents qvalue = 0.063. (B) IAA4-like (Ha412HOChr10g0435441). (C) SAUR12-like (Ha412HOChr07g0303111). (D) IAA4-like (Ha412HOChr01g0004341). (E) PIN7-like (Ha412HOChr01g0035481). (F) PHYB-like (Ha412HOChr02g0079361). (G) HB2-like (Ha412HOChr13g0623611). (H) BIM1-like (Ha412HOChr02g0064141). (I) Wavelengths of light measured inside 3 filter boxes. (J) Zscore-normalized stem angles of sunflowers inside filter boxes after their transfer from growth chamber to field (mean +/− SEM, n = 5). Data from one representative experiment. (K) Percentage of plants with significant rhythmic solar tracking in each of the 3 filter boxes (pvalue < 0.001, n = 22). (L) Percent increase in length of sunflower stems in each filter box (J) from dawn on day 1 to dawn on day 5. The underlying raw data may be found in S13 Data for Fig 4J; in S14 Data for Fig 4A-H; and in S15 Data for Fig 4I, 4K, and 4L. BIM1, BES1-INTERACTING MYC-LIKE1; HB2, HOMEOBOX PROTEIN 2; IAA4, INDOLE-3-ACETIC ACID INDUCIBLE4; PHYB, PHYTOCHROME B; PIN7, PIN-FORMED7; SAUR12, SMALL AUXIN UP-REGULATED RNA12; SEM, standard error of the mean; ZT, Zeitgeber Time.

https://doi.org/10.1371/journal.pbio.3002344.g004

Next, we examined transcriptional responses across the stem during these first few days. Two-week-old chamber-grown sunflowers were transplanted into the field as described above, and peels from the east and west sides of the stem were collected every 2 hours during the first day and every 4 hours for 2 more days. RNA was extracted from the stem peels, and gene expression was analyzed by quantitative reverse transcription polymerase chain reaction (qRT-PCR). First, we examined expression of genes that were rapidly up-regulated and then quickly down-regulated on the shaded sides of stems during phototropism (Figs 3H and S8). These included homologs of the known Arabidopsis auxin-related genes INDOLE-3-ACETIC ACID INDUCIBLE4 (IAA4), SMALL AUXIN UP-REGULATED RNA12 (SAUR12), and PIN-FORMED7 (PIN7) [52]. Surprisingly, only one of these auxin-related transcripts (IAA4-like (Ha412HOChr10g0435441)) was differentially expressed across the stem at any time point during the first day in the field despite the obvious heliotropic bending during that time (Figs 4B–4E and S8A–S8D). However, at no time did we observe the rapid and one-sided up-regulation of expression these transcripts demonstrate during phototropism in the growth chamber (S8A–S8D Fig). During the second and third days, however, their expression patterns began to mirror those observed in the RNA-seq data generated from plants grown in the field for over a week (S8A–S8D Fig).

We next examined the expression of 2 transcripts that undergo rapid and sustained up-regulation on the shaded sides of stems during phototropism (S8E and S8F Fig). These encode homologs of Arabidopsis PHYTOCHROME B (PHYB) and HOMEOBOX PROTEIN 2 (HB2), genes known to be involved in shade avoidance [53,54]. These transcripts are rapidly up-regulated on the west sides of stems 2 hours after dawn; however, this is only true on the first day after transfer to the field (Fig 4F and 4G). A homolog of BES1-INTERACTING MYC-LIKE1 (BIM1), another gene involved in shade avoidance [55], also had increased expression on the west side at ZT2, albeit not at levels reaching statistical significance (Fig 4H). We next examined the expression patterns of these 3 genes in our RNA-seq data of plants undergoing established solar tracking. Only the homolog of HB2 is differentially expressed in this condition, being more highly expressed on the east sides of stems at ZT8 (S8 Data). In contrast to what we observed in plants initiating heliotropism, none of these 3 genes are similarly differentially expressed across the stems of established heliotropic sunflowers early in the day. Together, these data demonstrate that even though plants responding to unidirectional blue light in a chamber and those responding to the rising sun on their first day in the field exhibit similar bending, their underlying transcriptional responses are quite different. Our data also suggest the potential involvement of phytochrome-mediated signaling pathways in the initiation of solar tracking.

Sunflower solar tracking is not affected by alterations in light quality

Our gene expression analysis suggested that early transcriptional events in heliotropism might have more in common with shade avoidance responses than with classical blue light–mediated phototropism. To investigate the possible involvement of multiple light signaling pathways in heliotropism, we tested the effects of changing light quality on the onset and maintenance of solar tracking. To do this, we built boxes using 3 types of filters: one clear, one to limit blue light, and one to limit red light (S9 Fig). Light measurements were taken from inside each box (Fig 4I) in the field. Light inside the clear box had a very similar spectrum to that of unfiltered light measured in the field and was used as a control. In the blue-depleted box, blue as well as violet and UV wavelengths were strongly reduced, while the red-depleted box removed virtually all red and a substantial portion of far-red light (Fig 4I). Soon after dusk, we transplanted 2-week-old chamber-grown sunflowers into one of these 3 boxes in the field and acquired time-lapse images of plant bending. Measurement of stem angles revealed that plants in all 3 types of boxes began tracking the first day, followed by reorientation on the first night (Fig 4J). Plants in all conditions maintained tracking movements for the following 3 days. However, plants in both the blue-depleted and red-depleted boxes exhibited additional movements at night (Fig 4J). The stems of these plants grew much longer than those in the clear box, likely due both to light quality differences and to overall lower light levels in these boxes compared to the clear box (Fig 4L). We suspect the noise in the nighttime reorientation of the plants grown in these 2 depletion boxes is caused by increased circumnutation accompanying the greater growth rates in these conditions compared to the clear box [56]. Despite these increased movements at night, estimations of rhythmicity of the bending movements revealed that over 75% of the plants in each condition, over 5 separate experiments, had significantly rhythmic tracking movements (Fig 4K). Of the 3 conditions, the blue-depleted plants had the lowest percentage of rhythmic trackers, perhaps due to the strong biphasic growth pattern, but overall diel rhythmicity of bending is still very robust (Fig 4J). These data show that both the onset and maintenance of heliotropism are resistant to changes in light quality and suggest that multiple light signaling pathways are involved in the regulation of solar tracking.

Plant growth conditions

Sunflower seeds (HA412-HO) were obtained from USDA North Central Regional Plant Introduction Station (USDA Germplasm Resources Information Network ID: PI 603993). Seeds were germinated in flats (filled with SunGro Sunshine mix #1) covered with a lid for 2 days. Lids were then removed and plants were grown in controlled environment chambers in 16 hours light:8 hours dark with 180 μmol m⁻² s⁻¹ fluorescent light at constant 25 °C. For phototropism experiments, 2-week old sunflowers were placed in front of unidirectional blue lights (Yescom 225 Blue LEDs Grow Light Ultrathin Panel, 280 μmol m⁻² s⁻¹) in a growth chamber at constant 25 °C. Sunflowers used for the onset of heliotropism experiments were grown in chambers with cycling temperature and photoperiod mimicking the field conditions for 2 weeks after germination and were then transplanted to the field after the sun had fully set. Plants used for established heliotropism field experiments were transplanted to the field 4 to 5 days after germination.

Tropism measurements

For phototropic and autostraightening bending analysis, time-lapse images were taken with a raspberry Pi 3 with infrared light and camera adapted from (https://maker.danforthcenter.org/tutorial/raspberry%20pi/led/raspberry%20pi%20camera/RPi-LED-Illumination-and-Imaging) [82]. For heliotropic bending analysis in the field, time-lapse images were taken using BirdCam 2.0 (Wingscapes) cameras. Stem angles from the base of the first internode to the apex were measured using ImageJ software [83].

RNA sequencing analysis

For all RNA-seq experiments, stem peels were collected from the first internodes of 2-week old plants using a vegetable peeler. Peels were collected from just below the stem apex to the bottom of the first internode. Consistent pressure was used for each peel to ensure similar thickness for each sample. The entire peel from each plant was frozen in liquid nitrogen and homogenized for use in extractions. For the field RNA-seq experiment, samples were collected at 4-hour intervals over 48 hours, starting at dawn on August 18, 2016. Day length was 13.5 hours; temperatures during this period are shown in S11A Fig. Tissue from the east and west sides of stems from 6 plants was collected at every time point. Two peels were pooled before RNA extraction to generate 3 biological replicates per time point. For the phototropism experiment, samples were collected from the lit and shaded sides of stems at ZT3, ZT5, ZT7, ZT9, ZT12, ZT15, ZT18, and ZT21. Stem peels from 2 different plants were pooled for each biological replicate, for a total of 6 biological replicates for each side at each time point. RNA was extracted and libraries were built using a previously described high-throughput RNA library preparation protocol [84]. Libraries from each experiment were separately multiplexed and sequenced at the California Institute for Quantitative Biosciences (QB3) using HI-seq4000 and 50-base pair single-end runs.

The sunflower transcriptome was built using the Rpackage cufflinks v2.2.1 gffRead using the Ha412HOv2.0 genome and the Eugene annotation (HAN412_Eugene_curated_v1) [29,85]. Reads were quality filtered using the fastx toolkit quality filter to have a minimum quality score of 20 in at least 95% of bases [86]. Reads counts were generated using salmon-1.4.0 to align and quantify the reads against the generated transcriptome [87]. The field data had an average of 16.8 million counts per sample, and the phototropism data had an average of 12.1 million counts per sample. Genes were removed if they did not have at least 10 counts in at least 3 samples for the field experiment, or in at least 6 samples in the phototropism experiment. Reads were normalized using the R package edgeR with the TMM method, and differentially expressed genes were determined by pairwise comparisons of sides of the stem (east versus west or shaded versus lit) at each time point, FDR < 0.05 [88]. For the differential expression analysis of the field data, the 2 days were combined into one, so each time point had 6 replicates per side. Rapidly induced genes during phototropism were identified by the overlap of the genes more highly expressed on the shaded side at ZT4 over ZT3, and ZT4 shaded over ZT4 lit during phototropism. Lit late induced genes during phototropism were identified by the overlap of those more highly expressed on the lit side at ZT15 over the shaded size at ZT15, and ZT18 lit over ZT9 lit during phototropism (FDR <0.05).

GO enrichment was performed using the R package clusterProfiler [89]. Arabidopsis homologs of each sunflower gene were found using blastx against the Arabidopsis proteome (TAIR10_pep_20101214), with an evalue cutoff of < 9.9 × 10⁻¹⁰ (S11 Data). GO terms were then assigned to each sunflower gene based on the top Arabidopsis hit using the terms found in ATH_GO_GOSLIM.txt.gz. Enriched GO terms were then found for each list of differentially expressed genes using all expressed genes as the background, FDR < 0.05.

Diel expression patterns of the field edgeR normalized counts were analyzed using the R package DiscoRhythm [35]. The Cosinor method was used to calculate acrophase, amplitude, and rhythmicity (FDR < 0.05) for each gene on both sides of the stem. Genes with an acrophase difference of >3 hours were considered differentially phased across the stem. Genes with >2-fold differences in amplitude across the stem were considered to have different amplitudes. PCA was performed on the heliotropism edgeR log transformed normalized counts using the R function prcomp [33].

qPCR analysis

Sunflowers at 2 weeks after germination were transplanted into the field after sunset. Stem tissue was collected from the east and west sides of plants, starting at dawn, every 2 hours for 1 day and every 4 hours for 2 more days. Tissue was collected from 6 plants for each time point, and 2 plants were pooled for each biological replicate for a total of 3 biological replicates per time point. RNA was extracted using Spectrum plant total RNA kit (Sigma), and cDNA was synthesized from 300 ng of total RNA using Superscript III reverse transcriptase (Invitrogen). Quantitative PCR was performed using BioRad CFX96 as previously described [90]. Expression was normalized to 2 house keeping genes, ELONGATION FACTOR 1-α (Ha412HOChr11g0495951) and UBIQUITIN-CONJUGATING ENZYME 21 (Ha412HOChr06g0274091), and analyzed using BioRad CFX Manager 3.1 software. Statistical significance was calculated using Welch’s t test, pvalue < 0.05. Primers used for qPCR are found in S12 Data.

Filter box assays

Experiments with filter boxes were conducted between August 12, 2021 and September 17, 2021, in Davis, California; day lengths during this period ranged between 13 hours 37 minutes and 12 hours 21 minutes. Mean, maximum, and minimum temperatures during these trials are shown in S11B Fig. Filter boxes were constructed with Rosco Roscolux light filters (Stage Lighting Store, Florida), #00 clear, #15 deep straw, and #375 cerulean blue. Light filters were wrapped and secured around garden support stakes to form a large box. A space of 1 to 2 inches was left at the top of the box between the walls and roof to provide air flow, and awnings of the appropriate filters were added to prevent nonfiltered light from entering through the openings. The boxes were placed into the field, held in place by garden stakes driven into the dirt. Light quality measurements were taken using a Black Comet spectrophotometer (StellarNet, Florida) from inside each of the boxes at the position and height of the plants. When measured at 11:30 AM, total PAR (400 nm to 700 nm) was 1,363 μmol m⁻² s⁻¹ within the clear box, 821 μmol m⁻² s⁻¹ within the deep straw box, and 410 μmol m⁻² s⁻¹ within the cerulean blue box. Approximately 4 to 6 plants were transplanted and monitored in each box for 3 to 4 days per trial. Mean and minimum and maximum daily temperatures during each of the 5 trials are shown in S11 Fig.

Rhythmicity of stem angle measurements was analyzed for each plant using the Biodare2 [91] eJTK test method with a pvalue < 0.001, n = 22. Percent stem growth was calculated by measuring the length from the bottom of the first internode to just below the apex, from dawn on the first day until dawn of day 5. Statistical significance was calculated using Welch’s t test, pvalue < 0.05.