Functional Genomics Project

Drought, high salinity and low temperature are stress factors through which the physical environment affects plant development, metabolism, performance and yield. One commonality between these three factors is the problem of maintaining water supply (homeostasis) because water is either scarce (drought) or hard to get to (high salinity) or difficult to transport (low temperature). Irrespective of commonality, plants sense these stresses as distinct and respond distinctly – with some illuminating overlaps in the response programs.

In collaboration with several associated labs, our lab wanted to know how many and which genes might be involved in abiotic stress responses, including the identification of sensors, signaling intermediates, and the metabolic and physiological end products (“downstream reactions”). Also, we wanted to find such genes in an evolutionary context, interested in seeing whether Saccharomyces cerevisiae, Aspergillus nidulans, Synechocystis sp. PCC6803, Dunaliella salina, Mesembryanthemum crystallinum, Arabidopsis thaliana and Oryza sativa possessed and expressed common gene networks and response patterns and, if so, what the constituents of these networks were and how they functioned. See www.stress-genomics.org for access to a grant proposal that has been funded by the National Science Foundation (www.nsf.gov; DBI 98-13360 & DBI 02-23905).

The stress-genomics group uses several approaches to address this question:

• Generating cDNA libraries from “stressed” organisms, establishing one-pass sequenced ESTs, annotating and determining expression profiles.
• Using these ESTs in large genome-wide microarray printing and smaller designer arrays for selected pathways. Using hybridizations for determining transcript amounts and changes under stress conditions.
• Generating reporter lines for Arabidopsis expressing luciferase under the control of a stress-inducible promoter, and then inserting T-DNA into those lines. If stress-relevant genes are targeted, an altered response to stress will easily be monitored.

Here are some statements:

• Following these strategies, we see, based on microarray analyses, stress-related changes in transcript abundance amount to approximately 5-10% of all genes in an organism, with the numbers very similar in, for example, yeast and rice.
• When comparing transcript profiles it is not only clear that diagnostic sets of genes can be established for different organs, tissues, or cells, it is also possible to recognize types of abiotic stresses, and possibly biotic stresses as well from the composition of these profiles.
• Stress by shocking the organism (e.g. 0 to 1M NaCl in yeast; or 0 to 150 mM NaCl in rice) elicits similar responses. A time course (on a scale of minutes for yeast, and hours for rice) of specific responses is indicated starting with changes in transcript amounts for (some) signaling components as an IMMEDIATE RESPONSE, followed by components that indicate RESTRUCTURING OF THE PROTEOME. In this category are found ubiquitin-dependent processes, many proteases, and transcripts for other degradative enzymes, and, significantly, transcripts that indicate increased synthesis of the translation machinery, including ribosomes. This phase is followed by the induction of – seemingly species-specific – responses, which might be termed METABOLIC ADJUSTMENT, that lead to altered biochemical pathways, altered cell wall synthesis, changes in ion uptake and transport systems, water transport, or hormone synthesis. A last phase in the adaptation process is recognizable after most transcripts have returned to the pre-stress abundance level. It is characterized by high transcript amounts for transcripts encoding defense functions, such as PR proteins, or glutathione S-transferases, ROS-scavenging and chaperone activities.
• Salinity stress responses are subject to the severity of the stress. Life-threatening stress, for example in yeast mutants incapable of osmotic adjustment, brought about genome stress, the induction of transcripts for functions associated with transposons and transposition activities (Yale and Bohnert, 2001). Rice Pokkali, moderately salt-tolerant, and rice IR29, sensitive, the most distinguishing characters were a lack of early responses, a less pronounced induction of the proteome restructuring activity but metabolic changes were much the same and of the same magnitude in both lines. However, Pokkali, 24 h after salt shock, had regained pre-stress levels for most transcripts while IR29 was dead at this time. Permanently up-regulated in rice Pokkali remained only few transcripts (Kawasaki et al., 2001).
• When stress is imposed gradually, responses are much more difficult to monitor. To use microarray analysis for such experiments that are close to a natural progression of stress in the environment, it will be important that most or all genes in an organism are represented on the array, that gene-specific array elements are present (i.e., 3’ end targets), that alternatively spliced forms of genes are included and that complete (several weeks) time courses are analyzed.
• Repeat hybridizations are essential components of the experiments. We generally carry out two hybridizations with the RNAs from one experiment and average the data. A third hybridization is done with RNAs from a complete and separate experiment. If significantly matched regulation is observed, no further repeats are considered necessary. Based on our data, a log10 ratio of ~+/-0.15 (expt/control) is sufficient to call up- or down-regulation. This is equivalent to an approximately 1.6-fold difference in regulation. In practically all experiments, a variable number of DNA elements provide widely varying intensities and must be flagged and excluded from the analysis.
• Problems with targets for microarray hybridizations are the following;
• At present, our microarrays include only between ~5% (rice) to ~30% (arabidopsis) of all genes or putative genes in these organisms. Ideally one would wish to have whole-genome microarrays for exploratory work and designer arrays for specific tasks.
• The fact that the elements that are printed are ESTs, i.e., not completely sequenced means that we are not able to distinguish isologs and often not even paralogs of gene families. Ideally, one would have targets on each array slide that report transcripts for an entire family, for a specific sub-family and for each individual gene in a sub-family.
• Only 10-20% of our clones are full-length, very much depending on a specific library. The future lies in full-length cDNAs, i.e., including the 5’ UTR.
• For additional discussions on microarray experiments and problems, see: http://www.stress-genomics.org/stress.fls/expression/array_tech/tech_aspects_index.htm (Laboratory of DW Galbraith, U. Arizona)
• A collection of hybridization data from different organisms, predominantly salt stressed, can be found at: http://www.OSMID.org. This database is searchable for transcripts by name or accession number, or individual experiments may be browsed.
• The generation of mutants in Arabidopsis thaliana is based on work pioneered by the lab of Jian-Kang Zhu, Arizona (Ishitani et al. 1997; Xiong and Zhu, 2002). Mutants are generated by EMS treatment or by T-DNA insertion. Two primary screens are used.
• In one screen, M2 seeds are plated individually on MS-agar plates that are stored vertically which leads to directional root growth. Germinated seedlings are then placed on plates that contain various concentrations of NaCl such that the roots pointed up. Plants with a defect in genes providing the marginal salinity stress protection to which Arabidopsis is capable will continue to grow and their roots will bend. Plants affect in some aspect of tolerance will cease growth or show aberrant growth.
• More than 100 mutants in this class have been identified after screening ~70,000 seeds.
• A second screen uses a different population of M2 seeds from Arabidopsis plants into which a gene has been inserted before mutagenesis. This gene consists of the promoter for a stress-inducible gene driving luciferase expression. Most lines express luciferase under the control of RD29A promoter but recently other promoters with different induction characteristics have been used. Into such a line, T-DNA is inserted randomly by Agrobacterium tumefaciens infection. Mutants are detedcted by altered luciferase expression using fluorescence imaging cameras (Roper Scientific, Inc.) located at UA, Purdue, and UIUC.
• More than 100 mutants in this class of camera mutants have been identified after screening ~50,000 plantlets.
• Generation of different luciferase lines, T-DNA insertion mutagenesis, screening and analysis of the luciferase-mutants are done at Purdue University because of its exceptional greenhouse facilities and excellent management (Ray A Bressan, P. Michael Hasegawa).