In a review article in the June 03 issue of Current Opinion in Plant Biology 2003, 6:288-296, Thomas J. Buckhout and Oliver Thimm at the Institute of Biology, Humboldt University, Berlin and the Max-Planck Institute of Molecular Plant Physiology, Golm, respectively, describe how DNA microarray technology is being used to analyze simultaneous expression of thousands of genes under a variety of experimental conditions. The authors draw attention to the fact that setting up a microarray facility may be costly but once it is in place, functional analyses of a multitude of genes is relatively straightforward. According to these authors, the difficulty in using this technology lies in interpreting the results rather than carrying out an array experiment. The authors lay stress on the importance of using the two new emerging sciences, namely, proteomics (analysis of protein structure and function) and metabolomics (analysis of biochemical concentration at a particular metabolic state in a sample) to strengthen microarrays studies.
The advent of DNA microarray tools has greatly facilitated on research of gene expression patterns. DNA microarrays may be constructed, using (a) cDNAs to express sequenced tags (ESTs), (b) gene sequences or (c) gene-specific oligonucleotides. Of the above three, the use of oligonucleotides is presently considered the best choice because 1) problems associated with false signals can be minimized, 2) it simplifies gene annotation and 3) it gives sequence reliability without clone contamination.
Gene expression during seed development in Arabidopsis: The authors cite the studies of Dr. John Ohlrogge (Michigan State University) who used DNA microarrays to determine how specific genes act to regulate primary lipid biosynthetic pathways at three distinct stages of seed development. In the initial stage, there is cell division, followed by tissue and organ development. In the second stage, marked by the seed maturation process, there is an accumulation of storage products. In the third and last stage, seed desiccation takes place, halting further seed development. Data were collected at six different stages of development between 5 and 13 days after flowering (DAF). Dr. John Ohlrogge’s group observed that the maximum gene activity in early seed development was confined to genes involved in the transport of glucose phosphate into the chloroplast and cytosol. The genes that acted in the intermediate phase of seed filling were those involved in plastid glycolysis and lipid biosynthesis. Finally, genes encoding storage proteins were found to be implicated in the final stage of seed development. In other words, there was correspondence between the stage of seed development and the genes that encoded the specific enzymatic proteins that were required to carry out those specific reactions. These microarray findings corroborated previous-technology-based results on gene expression related to the three stages of seed development. Of most importance, the patterns of induction and repression for genes of known function can now be compared with the pattern of genes of unknown function or whose function in seed development was not previously known. This knowledge can serve to formulate hypotheses, for example to describe carbon flow during seed development, that can be tested in a straight-forward manner.
Arabidopsis plants grown in Fe-deficient environment: Dr. Oliver Thimm (one of the authors) and his associates carried out a microarray analysis to study the pattern of gene expression in Arabidopsis plants grown in an Fe-deficient environment for one, three and seven days. They observed changes in the expression of genes that responded to the iron-deficiency conditions after three days. The genes that responded were the ones that encode metabolic enzymes such as hexokinase and those involved in phosphate translocation and sucrose export. Furthermore, there was an increase in the quantity of glucose-6-phosphate and glucose concentrations by approximately 20% and 45% in Fe-deficient conditions compared to those where there was no iron deficiency. The findings point to a significant change in primary metabolism in response to iron deficiency.
Gene expression during adaptation to changing N supply: The authors describe how the study of DNA microarrays has helped identify a subset of Arabidopsis genes that are activated in response to a fresh supply of nitrate after a 3-day period of nitrogen starvation. Among others, the genes encoding enzymes that are involved in processes of nitrate transport, reduction to ammonium and assimilation were rapidly induced following nitrate re-supply. Some nitrate-induced expression was initiated and completed in 2 hours. Another significant observation made in this context is the activation of several genes, not known to directly participate in NO3- assimilation. For example, Ca2+ transporter (CAX1), protein kinases and phosphatases are some of the enzymes that are considered essential in the transduction of the NO3- signal. Auxin is also believed to have a functional role in this process. Analyses of microarrays revealed that providing nitrate to nitrate-starved tomato plants triggers the function of regulatory protein genes and transcription factors. It was also observed that the genes responding to nitrate supply in tomato and Arabidopsis are not identical, indicating that, to date, only a subset of the regulatory genes have been identified.
Predicting metabolic activity is difficult on microarray data : The authors point out some of the difficulties in predicting metabolic activity for Arabidopsis and tomato on the basis of microarray data alone. Considering that carbon- and nitrogen metabolism represent highly connected pathways, the present microarray data do not explain why only a few genes encoding enzymes of carbon metabolism are induced when nitrate is re-supplied to Arabidopsis or tomato plants after they were grown in nitrate-deficient medium. Clearly, mechanisms other that gene regulation are functioning in parallel.
Comparison of microarray studies between nodules and normal roots: The authors discuss a recent comparative microarray analysis between nodule formation in Lotus japonicus and its normal roots. It was shown that the genes that encode enzymes for glycolysis and control other steps in sucrose metabolism are induced in the process of nodule formation. For instance, compared to roots, the expression of genes encoding malate dehydrogenase (MDH) and aspartate transaminase (ATT) was found enhanced in the nodules. These data can now serve as a basis to analyze carbon and nitrogen flow during symbiotic nitrogen fixation at the gene level
Two sets of genes: one set responding to a specific environment and the other induced under different environments: Microarray analyses have made it possible to analyze a large number of genes that are either induced or repressed in response to environmental cues. The authors point out that there are a large number of genes that are reportedly induced in different experiments, while there are some that are specific to a single response. In support of their statement, they have cited the results of microarray analyses in which many of the genes that were induced under oxygen deficiency (hypoxia) were also induced under Fe-deficiency. Similarly, they report that the some genes that are associated with nitrogen assimilation were found to be induced when changes in Fe supply take place.
Interconnection of cellular metabolism pathways: Different biochemical pathways carry out a multitude of functions in cellular metabolism, and some of the pathways are interconnected. In other words, some of these pathways share common primary signal-transduction pathways; they influence other metabolic processes as a secondary effect. For instance, the induction of fermentation genes takes place as a result of Fe-deficiency but is not likely the result of hypoxia. The authors remind investigators that it is important to confirm microarray data by means of metabolite concentrations in order to avoid arriving at an erroneous conclusion.
Conclusion: Although micro-array analytical tools have so far been applied for confirming current levels of knowledge of gene expression, the authors conclude that new applications such as manipulation of large datasets will be found in the near future. Analyses of these datasets, which are increasing at a phenomenal rate, will increase our understanding of the hitherto unknown gene functions