Brassinosteroids, de-etiolation and the re-emerging art of plant hormone quantification

Abstract

An increase in the use of molecular techniques has provided a significant insight into the function of genes, and how they are regulated and interact. However, in the field of plant hormone physiology, the increased use of these techniques has been accompanied by a reduction in the direct measurement of plant hormone levels by physiochemical methods. Instead, the transcript (mRNA) levels of genes involved in hormone metabolism are often used to predict endogenous hormone levels. The validity of this approach was recently tested by comparing the expression of a range of genes involved in BR synthesis, catabolism and perception, with the actual endogenous BR levels in pea seedlings grown under different light conditions.^1,2 Based on this comparison, we now argue that gene expression analysis alone is not always a reliable indicator of endogenous hormone levels.

Addendum to: Symons GM, Smith JJ, Nomura T, Davies NW, Yokota T, Reid JB. The hormonal regulation of de-etiolation. Planta 2008; 227:1115-25.

The Re-emerging Art of Plant-Hormone Analysis

Since the early 1990s, the field of plant biology and biological sciences generally, has witnessed a dramatic increase in the use of molecular techniques to investigate gene sequence, function, regulation and interaction. For researchers working on plant hormones this revolution has provided an unprecedented insight into many of the genes involved in hormone synthesis, catabolism, perception and signal transduction, across a wide range of species. Coupled with continual development and refinement of molecular techniques, this new information is now being utilised to successfully address many fundamental questions regarding plant hormone biology.3

However, one unfortunate casualty of this revolution in research methodology has been the direct measurement of endogenous plant hormone levels. Instead, the transcript (mRNA) levels of genes involved in hormone synthesis, catabolism and perception/signalling are often used to presuppose endogenous hormone levels, without confirmation by direct measurement (for example, see refs. 4–9). However, the validity of using gene expression analyses alone to “determine” the endogenous hormone levels, and the accuracy of such predictions have not been widely debated in the literature.

In response to this situation, we presented evidence in a recent paper, published in Planta,1 which clearly demonstrates that the transcript levels of hormone synthesis/catabolism genes are not always an accurate indication of the endogenous hormone levels. As part of this study we investigated the relationship between endogenous BR levels and the expression of genes involved in BR synthesis, catabolism and perception, by examining transcript levels of a number of such genes in pea seedlings subjected to different light treatments.1 It has been clearly established that in pea, BR levels are significantly higher in light-grown seedlings than in dark-grown seedlings (a result which runs counter to a previous conventional wisdom on this issue).2,10 However, our quantitative, real time RT-PCR results show that there is no consistent pattern of expression of these genes under the different light conditions that could be said to accurately reflect the endogenous BR levels (Table 1). Indeed, the expression patterns of the individual genes varied dramatically, and are either increased, decreased or unchanged in light-grown seedlings relative to those grown in the dark (see Table 1).

In some instances there was no disconnect between gene expression and hormone levels. For example, the increase in the BR synthesis gene, PsDWF4, and the decrease in BR metabolism gene, PsBAS1, in light-grown plants (relative to the dark-grown controls) were consistent with the higher BR levels measured under these conditions.2 This is consistent with previous studies conducted in Arabidopsis and tomato, where expression levels of selected BR-synthesis genes were also closely correlated with the actual endogenous BR levels.11–14 However, this situation did not hold true when we examined the full complement of genes involved in the BR synthesis pathway. Indeed, there were just as many examples where the expression of individual genes did not reflect the endogenous BR levels. A clear example can be seen in the transcript levels of the two largely redundant PsBR6ox genes, which catalyse the conversion of 6-DeoxoCS to the bioactive compound, CS.15 In this case, PsBR6ox1 transcript levels were slightly elevated in the light relative to the dark (Table 1), which is consistent with the higher endogenous BR levels under these conditions.2 However, in direct contrast, PsBR6ox6 transcript levels were clearly reduced in the light (Table 1), possibly as a result of the feedback regulation of this step in response to the high BR levels.15

Table 1 The change in transcript levels, of genes involved in brassinosteroid (BR) synthesis (S), catabolism (C) or perception (P), in light grown pea (Pisum sativum) seedlings

Gene	Gene function	Change in transcript level relative to dark grown controls
LK	S	-
LKA	P	--
LKB	S	--
PsCPD1	S	+++
PsCPD2	S	0
PsDWF4	S	+
PsBR6OX1	S	+++
PsBR6OX6	S	--
PsBAS1	C	--

Whilst endogenous bioactive BR levels were significantly increased in the light-grown seedlings, relative to those grown in the dark, gene transcript levels were either increased (+), decreased (-) or unchanged (0). The magnitude of changes is indicated as follows: 0 (< 10%); -/+ (10–30% change); --/++ (31–50% change); ---/+++ (51–100% change).

This scenario poses a significant problem for anyone wishing to predict the endogenous hormone content based on gene expression alone. Faced with this data set, and in the absence of direct hormone measurements, which of the two genes, PsBR6ox1 and PsBR6ox6, would we use to predict plant hormone levels? Similarly, if we had examined the expression of only one, or a select few genes involved in BR biosynthesis/catabolism (as occurs in many studies), could we have any confidence that the expression profile of those genes would accurately reflect the expression profiles of the other relevant genes, or the actual BR levels? Based on our results we strongly believe that the answer to this question is no.

In contrast, we might expect that examining the expression of genes that reflect the transduction of the hormone signal may be a more accurate indication of endogenous hormone levels. One commonly used example is DR5, a synthetic auxin-responsive promoter element, linked to the GUS (β-glucronidase) coding sequence. The DR5-GUS reporter system is inserted into the Arabidopsis genome by Agrobacterium-mediated transformation, its transcription is upregulated by elevated auxin levels, and this can be visualised in plants by detection of the GUS gene. As a consequence, this system is often used as a marker to study endogenous auxin distribution and levels, and in some circumstances it has been shown to closely reflect actual IAA distribution as determined by direct measurement of IAA levels.16,17 However, even the use of this technique to ‘measure’ IAA levels is controversial because BRs18 and other auxins19 can also activate DR5-GUS expression independently of IAA.

Case Study: The Role of Brassinosteroids in De-Etiolation

A classic case, which highlights the need for direct measurement of plant hormone levels, is the long-standing view that BRs play a negative regulatory role in de-etiolation (the transition from growth in the dark to growth in the light). This suggestion first arose in the 1990's when it was reported that many brassinosteroid deficient Arabidopsis mutants exhibited signs of de-etiolation when grown in the dark.20,21 The logical conclusion drawn from these, and other data,22 was that a reduction in endogenous levels may be a normal part of the de-etiolation response and that this reduction, at least in part, may facilitate the development of a de-etiolated (light-grown) phenotype.23–25

In subsequent years this theory was tested and disproved in species such as pea and rice, using direct measurement of BR levels.2,10,26 However, it remained possible that these results reflected species-specific differences in the hormonal regulation of de-etiolation, and were not applicable to the situation in Arabidopsis. The critical, and admittedly more technically difficult, experiments which would examine BR levels during de-etiolation in Arabidopsis, were not forthcoming. In the absence of direct evidence, expression levels of BR biosynthesis genes have often been used to support the widely-accepted, but incorrect, assumption that BR levels are lower in light-grown than in dark-grown Arabidopsis seedlings.5,6,27 The record has finally been corrected in our recent paper published in Planta,1 which demonstrates, through direct measurement, that BR levels are in fact higher (not lower) in light-grown than in dark-grown Arabidopsis seedlings. This example demonstrates that even accurate gene expression data may be misleading if used as an indicator of plant hormone levels, and highlights the need for direct measurements.

Conclusion

Hormone biosynthesis and catabolism are undoubtedly complex processes, often involving multiple parallel pathways leading to the production of the bioactive compound/s, and many reactions (often controlled by multi-gene families), that are differentially regulated by environmental and developmental inputs, including self-regulation (feedback and feedforward) in response to bioactive hormone levels. Given this level of complexity it is overly simplistic to assume that, when a single gene that is involved in hormone synthesis is highly expressed, each and every gene involved in the synthesis of that hormone will also be highly expressed, and that the bioactive levels of that hormone will be elevated. This often-used logic is not supported by the results presented in Planta.1 This study clearly demonstrates that expression analysis of hormone biosynthesis genes is not a panacea for determining endogenous plant hormone levels. Indeed, gene transcript levels can be misleading and result in predicted hormone levels that are, quite simply, incorrect. It should be clear from the recent results published in Planta1 that there is no substitute for the direct measurement of that hormone via modern physiochemical methods. The widespread acceptance of this reality should see the continued re-emergence of GC/LC-MS-based plant-hormone quantification as an important tool for use in a multidisciplinary approach to plant hormone research.

Acknowledgements

We thank John Ross for his comments on this manuscript and the ARC for financial support.

Addendum to: