The Use of High Throughput Transcriptomics Data to Infer Signalling Networks Activated During Stress

To achieve understanding of complex biological systems it is necessary to integrate high-throughput biological studies. In plant research, systems biology is still in its dawn and very much at the stage of accumulating vast quantities of data, especially from high-throughput transcriptional profiling. An important problem is that of reconstruction of signalling networks from gene expression data.

Gene expression data convey information about pathways, as it is reasonable to presume that highly co-expressed genes work in the same pathway (Eisen et al. 1998; Marcotte et al. 1999). The standard way to select genes that belong to the same pathway is therefore to group them according to the correlation of their expression profiles over different conditions—this is sometimes referred to as "guilt by association" (Walker et al. 1999). Clearly, although helpful in identifying the components of a certain pathway, this methodology is still largely deficient in providing information on the hierarchical connection between these components.

Once some candidate genes have been identified using these in-silico methods, a necessary step is then to verify these predictions through reverse genetics and mutant analysis. In recent times, a number of excellent studies has been performed in the direction of constructing signalling networks activated during plant responses to stress and during defence. Plants respond to stimuli such as pathogen and pest attack, wounding, changes in light, temperature and availability of nutrients (see also chapters from Doerner P Signals and mechanisms in the control of plant growth; Durgardeyn J and Van Der Straeten D, Ethylene: inhibitor and stimulator of plant growth). A transi tion during development is also perceived by the plant as a change generating a signal. Activated responses generally lead to extensive transcriptional reprogramming of gene expression. In this review we will consider mainly responses at transcriptional level. Pathways historically involved in the perception and transduction of stress are those regulated by jasmonates (JAs), salicylic acid (SA), ethylene (ET) and abscisic acid (ABA). These molecules are involved in the local and/or systemic response of the plant (for relevant recent reviews see Fujita et al. (2006), Gfeller et al. (2006), Grant and Lamb (2006), Halim et al. (2006), Dreher and Callis (2007), Devoto and Turner (2005), Lorenzo and Solano (2005).

Particularly, in Arabidopsis, the availability of tools such as mutants and pathosystems as well as high-throughput technology such as transcription profiling by microarrays has greatly facilitated information gathering on the existence of components of signalling pathways as well as of crucial nodes facilitating communication within signalling networks. This approach has been used for example to identify regulatory nodes in the transcriptional network of systemic acquired resistance (SAR) in Arabidopsis. SAR is an in-ducible plant defence response involving a cascade of transcriptional events caused by SA through the transcription cofactor NPR1 (non-expressor of PR; (Kinkema et al. 2000). To identify novel regulatory nodes in the SAR network (Wang et al. 2006), performed microarray analysis in a stepwise approach on Arabidopsis plants expressing the NPR1-GR (Glucocorticoid Receptor) fusion protein. Since nuclear translocation of NPR1-GR requires dexametha-sone (Wang et al. 2005), the authors were able to control NPR1-dependent transcription and to identify direct transcriptional targets of NPR1 acting as crucial regulatory nodes during SAR. Disrupting these regulatory nodes compromised various functions assigned to NPR1. Specifically it was found that NPR1 directly upregulates the expression of five WRKY transcription factor genes that had never been placed before in the SAR network. Among these WRKY factors, both positive (WRKY18 and 53) and negative regulators (WRKY58) of SAR were found. In addition, fine tuning of SAR occurs when SA levels are high: signalling through positive WRKY factors were found to overcome the negative effect of WRKY58 to activate downstream gene transcription and the action of WRKY70 and WRKY54 prevent excessive SA accumulation.

The nature of the mobile signal as well as of the remotely activated networks responsible for establishing SAR, remains unclear (Grant and Lamb 2006). In a recent study (Truman et al. 2007) have shown that in Arabidopsis, despite the absence of pathogen-associated molecular pattern (PAMPS, Nurnberger et al. 2004) contact, systemically responding leaves rapidly activate a SAR transcriptional signature with strong similarity to local basal defence responses to herbivory and wounding. The signature shares secondary metabolism components with late basal defence responses. The RPM1 (Resistance to P. syringae pv. Maculicola 1) pathosystem (Grant et al. 1995)

has been used here to dissect both timing and nature of early transcriptional events in tissues associated with the establishment of systemic immunity after RPM1 recognition. A role as initiating signal for SAR has been attributed here to JAs which are suggested to act ahead of SA-dependent responses in systemic leaves. JAs, including the JA precursor oxophyto-dienoic acid (OPDA) and conjugated derivatives, such as methyl-JA (MeJA) or isoleucine-JA have been previously demonstrated to possess roles in defence signalling (Stintzi et al. 2001; Staswick and Tiryaki 2004; Sasaki-Sekimoto et al. 2005). These conclusions were reached by carrying out an extensive comparison of in-house microarray analysis with experiments representing host responses to biotic and abiotic stresses or hormone treatments from the ArrayExpress ( and NASCAr-rays ( database repositories. The novel work from M. Grant's laboratory has shown that SAR can be mimicked by foliar JA application and was found to be abrogated in mutants impaired in JA synthesis or response. De novo JA biosynthesis was found to be associated with the induction of jasmonate-responsive genes in systemic tissues. Therefore, although JAs have been generally regarded as antagonizing SA-dependent responses, the plant can benefit from the advantages of both signalling pathways when their activation is separated during time or space.

The systematic comparison of the expression profiles between wild-type and mutants still represents one of the most exhaustive ways to elucidate critical steps in host responses, at least at the transcriptional level. A recent example of expression profiling and mutant analysis aimed at further dissecting Arabidopsis defence response to the necrotrophic pathogen Botrytis cinerea infection was provided by the laboratory of T. Mengiste (AbuQamar et al. 2006). In this study, mutants exhibiting enhanced susceptibility to Botrytis were used to correlate changes in mRNA profiles with impaired disease resistance responses of whole plants. Arabidopsis wild-type plants were compared to coi1 (coronatine insensitive 1) and ein2 (ethylene insensitive 1) mutants and to plants carrying the nahG (salicylate hydroxylase) gene. In wild-type plants, the expression of 621 genes representing approximately 0.48% of the Arabidopsis transcriptome was induced.

The expression of 181 Botrytis induced genes (BIGs) was dependent on a functional COI1 gene, a well-known component of JA signalling (Feys et al. 1994; Xie et al. 1998), whereas the expression of 63 and 80 BIGs were dependent on ET signalling or SA accumulation, respectively. Thirty BIGs encode putative DNA-binding proteins previously found to be regulating ET responses such as zinc-finger, MYB, WRKY, and HD-ZIP family transcription-factor proteins. Importantly, T-DNA insertion mutants in two BIGs, encoding putative DNA-binding proteins ZFAR1 (At2g40140) and WRKY70 (At3g56400), showed increased susceptibility to Botrytis infection. ZFAR1 is also required for germination on ABA, and encodes a putative transcription-factor protein containing zinc-finger and ankyrin-repeat do mains. The transcriptional activation of genes involved in plant hormone signalling and synthesis, removal of reactive oxygen species, and defence and abiotic-stress responses, coupled with the susceptibility of the wrky70 and zfar1 mutants, highlights the complex genetic network underlying defence responses to Botrytis in Arabidopsis. The above study by (AbuQamar et al. 2006) represents probably one of the latest examples about the previously demonstrated interconnection between signalling pathways such as JA, ET and SA (Devoto and Turner 2005; Devoto et al. 2005; Lorenzo and Solano 2005; Gfeller et al. 2006; Liechti et al. 2006).

Indeed, hormones rule every aspect of the biology of plants. Stress and development are regulated in comparable ways by multiple hormones and as also highlighted above by recent key studies, the existence of widespread crosstalk among different hormonal signalling pathways has been revealed (Finkelstein et al. 2002; Guo and Ecker 2004; Sun and Gubler 2004; Vert et al. 2005; Woodward and Bartel 2005). Crosstalk refers to the case that two inputs (in this review: stresses of biotic and abiotic origin) work through different signalling pathways but combine forces to regulate outputs and ultimately development. Intensive experimental work has revealed numerous potential paths for crosstalk. Despite the apparent integration of inputs from multiple hormones in regulating development, it has been recently shown that the level of convergence, defined by co-expression, on a common set of transcriptional targets is reduced only to a few genes (e.g., only seven genes were changed in the same direction by GA (gibberellic acid 3), IAA (auxin, indole acetic acid), and BL (brassinosteroids) treatments, none with known function) and that therefore there is not a core transcriptional growth-regulatory module in young Arabidopsis seedlings (Nemhauser et al. 2006). In this work, data produced by the AtGen- Express Consortium ( in which the effects of seven plant hormones at three time points were surveyed with Affymetrix ATH1 GeneChips representing nearly all protein-coding transcripts of Ara-bidopsis were compared. The compounds assayed included ABA, GA, IAA, 1-amino-cyclopropane-1-carboxylic acid (ACC; ethylene precursor), zeatin (CK; cytokinin), BL and MeJA. These studies revealed that a major part of early hormone response in plants is specific and independent of the effects of other hormones. It has to be highlighted however that despite the low numbers of shared transcriptional targets, an ample evidence of one hormone-regulating genes involved in the metabolism of another hormone was observed. It is possible that this could be due to a knock-on effect from one hormone re-setting many systems within the plant. Caution is therefore needed in drawing conclusions regarding the existence of crosstalk from a limited number of genes that appear to be similarly regulated by different hormones. Despite that such a comparison has not been carried out yet in a similar manner for plants "under attack", it is possible that similar conclusions might be applicable. Interestingly, the work carried out by (Wang et al. 2006) highlights that even different members of the same gene family of transcription factors may have very specific functions within the same pathway.

0 0

Post a comment