ISTITUTO SPERIMENTALE PER LA CEREALICOLTURA,

VIA STEZZANO 24, 24126 BERGAMO, ITALY

Transcriptome analysis of opaque2 and opaque7 mutants in maize endosperm

Pirona, R, Hartings, H, Rossi,V, Motto, M

In maize, the zein synthesizing system is particularly adapted for the study of the regulating mechanisms of plant genes because i) its expression is restricted to a specific tissues and stage during seed development and ii) for the availability of mutants useful in dissecting the regulatory processes taking place in the developing seed. (Pirona et al., Maydica 50:515-530, 2005). Studies on genetic mutations that affect the accumulation of different zeins have demonstrated the existence of several regulatory signals controlling the expression of specific members of the zein family which confer an opaque phenotype to the endosperm (Motto et al., pp: 479-522, In: Larkins, BA, Vasil, 1K (Eds), Cellular and Molecular Biology of Plant Seed Development, Kluver Acad. Publ., The Netherland, 1997). For example, the recessive mutation opaque2 (o2) and opaque7(o7) induces a specific decrease in accumulation of 22 and 19-kD alpha-zeins, respectively, while the opaque15 (o15) mutation exerts its effect primarily on the 27-kD gamma zeins. The recessive mutation opaque6 (o6) and the dominant or semi-dominant mutations Floury (Fl2), Defective endosperm *B30 (De*B30), and Mucronate (Mc) cause a more general reduction in accumulation of all zein classes.

The 02 mutation has been widely studied at molecular, genetic, and biochemicai levels (see Pirona et al., 2005). The product of the 02 gene is a basic leucine zipper (bZiP) transcriptionai regulator that is specifically expressed in the endosperm and activates the expression of 22 kDa alpha-zein and 15 kDa gamma-zein, together with the B-32 gene, encoding an endosperm specific ribosome nactivating protein. Other possible direct or indirect target genes of the 02 factor have been shown to belong to various metabolic pathways, suggesting that 02 may play an important role in the developing grain, as a coordinator of the expression of storage protein, and nitrogen and carbon metabolism genes.

In recent years, the development of extensive maize cDNA libraries, along with computer software to systematically characterize them, has made it possible to analyze gene expression in developing maize endosperm more thoroughly. Accordingly, we have used cDNA microarray technology to investigate the transcription profiles and differential gene expression of maize endosperm from two different opaque mutants (o2 and o7) and in double mutant combination (o2o7).

Microarrays were assembled using clones obtained from the EC ZeaStar project ( Edwards et al., unpublished results). Briefly, 20 part-normalized cDNA libraries were prepared from tissues covering 5 key stages in both endosperm and kernel development. Approximately 20,000 ESTs were sequenced, aligned, assembled into contigs using a similarity score of 80%, and annotated using BLASTA and TBLASTN software. Contigs and singleton cDNAs were used to construct a unigene set of 8,950 sequences. EST sequences were analyzed with the BLAST2GO software (http://www.blast2go.de). First, homology searches using public domain non-redundant databases were performed and identified significantly homologous sequences for 48.4% of the ESTs considered. These ESTs represented 3,090 single hit (71.3%) and 1,240 multiple hit sequences. Subsequently, an attempt was made to associate biological functions to each of the ESTs showing sequence homology using the gene ontology (http://www.geneontology.org) and KEGG databases (http://www.genome.jp/kegg). Approximately 85% of the ESTs analyzed could be associated with GO database entries. The results of this analysis permitted to divide the mentioned ESTs into 24 functional group total of 7,250 clones were spotted in duplicate.

Microarray slides containing the entire Zeastar unigene set were hybridized with probes derived from endosperm tissue harvested 15 days after pollination (DAP) and derived from the A69Ywt, A69Yo2, A69Yo7, and A69Yo2o7 isogenic lines. To reduce hybridization artefacts, all probes were labelled both with Cy3 and with Cy5 and used in dye-swapping experiments on series of three independent slides. The expression data obtained were assayed for consistency by performing T-tests at 95% confidence levels.

All microarray experiments were performed in triplicate using dye swapping, hence giving rise to 12 independent measurements for each EST, considering the presence of duplicate spots on each slide. Raw measurements of spot fluorescence intensities were collected from hybridized slides using a Genepix 4100A scanner and Genepix Pro4 software (Axon Instruments, Union, CA). Subsequently, the obtained spot values were corrected for background fluorescence and analyzed using the Vector Xpression3 software (Informax, Frederick, MD). The data were log2 transformed and normalized by equalizing the mean intensity of each channel to 1. To verify reproducibility between spots and between channels, T tests were performed applying a 95% confidence threshold and allowing to remove inconsistent hybridization results. Ratios between wild type and mutant expression levels were calculated and ESTs exhibiting ratios below 0.5 or over 2 were selected for further analysis.

Average signal values derived from the four probes used were graphed using a logarithmic scale. The graphical representations clearly showed the prevalence of genes showing distinct expression patterns in the A69Ywt and A69Yo2 genotypes. Conversely, the A69Ywt and A69Yo7 genotypes show less evident differences in expression levels. The A69Yo2o7 double mutant exhibits differences in expression patterns resembling those obtained for the A69Yo2 genotype. A plot of A69Yo2 vs. A69Yo7 expression levels showed the cumulative effect of both genotypes revealing a high number of genes with distinct expression patterns.

Consistently performing spots in T-tests were selected and used to calculate wtlmutant expression ratios. Among the ESTs considered, 17,1% exhibited a down-regulated expression profile. The o2 mutation could be associated with 649 down regulated ESTs, 508 down-regulated ESTs were identified in A69Yo7 background, whereas 759 ESTs showed a reduced expression pattern in ‘ A69Yo2o7. Up-regulated expression profiles were found for 3.23% of the ESTs considered. One hundred and thirteen up-regulated ESTs were identified in the A69Yo2, 26 in the A69Yo7, and 86 in an A69Yo2o7 backgrounds, respectively. Among the ESTs identified, 36.7% exhibited relevant homology with sequences deposited in public databases and could be univocally associated with known biological processes related to amino acid and carbohydrate metabolism, signal transduction, protein turnover, transport, and protein folding. In addition, three transcription factors different from O2 appear down-regulated. Collectively, the results may provide a framework for investigating a common mechanism that underlines the o2 and o7 kernel phenotypes.