To establish the size and organization of the rice alpha-amylase multigene family, we have isolated 30 alpha-amylase clones from three independent genomic libraries. Partial characterization of these clones indicates that they fall into 5 hybridization groups containing a total of 10 genes. Two clones belonging to the Group 3 hybridization class have more than one gene per cloned fragment. The nucleotide sequence of one clone from Group 1, lambda OSg2, was determined and compared to other known cereal alpha-amylase sequences revealing that lambda OSg2 is the genomic analog of the rice cDNA clone, pOS103. The rice alpha-amylase genes in Group 1 are analogous to the alpha-Amy1 genes in barley and wheat. lambda OSg2 contains sequence motifs common to most actively transcribed genes in plants. Two consensus sequences, TAACAAGA and TATCCAT, were found in the 5' flanking regions of alpha-amylase genes of rice, barley and wheat. The former sequence may be specific to alpha-amylase gene while the latter sequence may be related to a 'CATC' box found in many plant genes. Another sequence called the pyrimidine box (TCCTTTTTC) was found in the alpha-amylase genes as well as other genes regulated by gibberellic acid (GA). Comparisons based on amino acid sequence alignment revealed that the multigene families in rice, barley and wheat shared a common ancestor which contained three introns. Some of the descendants of the progenitor alpha-amylase gene appear to have lost the middle intron while others maintain all three introns.

Human alpha-1-antitrypsin (AAT), the most abundant protease inhibitor found in the blood, was expressed in rice embryonic tissue suspension cell culture. This was accomplished by cloning the codon-optimized AAT gene into a vector containing the rice RAmy3D promoter and its signal sequence. The synthetic gene incorporates codons synonymous with those found in highly expressed rice genes. Approximately 1000 stable transformed calli were produced by particle bombardment mediated transformation and were screened for high AAT expression using a porcine elastase inhibitory activity assay. The band shift assay also confirmed that rice-derived AAT is functional regarding its binding capability to the elastase substrate. Time course studies were conducted to determine the optimum, postinduction expression levels from cell culture. AAT expression equivalent to 20% of the total secreted proteins was achieved, and a purification scheme was developed that yielded active AAT with purity greater than 95%. The potential applications of purified plant-derived AAT for treatments of various AAT-deficient diseases are discussed.

Two cDNA clones, pOS103 and pOS137, were isolated which code for distinct alpha-amylase isozymes in germinating rice seeds. Sequence analysis indicated that the clones encode polypeptides of approximately 48 kDa, both of which possess a signal peptide involved in directing secretion of the protein. Comparison of the two rice alpha-amylase amino acid sequence showed that they are 76% similar to each other, while showing 85% to 90% similarity with other cereal alpha-amylases. A comparison of eleven cereal alpha-amylases also revealed three new conserved regions (I', II', and IV') not previously identified in the animal, bacterial, and fungal alpha-amylases. Regions I' and IV' are sites for intron splicing while region II' is probably involved in calcium binding. One of the rice alpha-amylase cDNAs, pOS103, encodes a protein that has two potential N-glycosylation sites, one in the signal peptide and the other in the mature portion of the protein. The cDNA clone, pOS137, encodes an alpha-amylase with a single glycosylation site in the signal peptide, suggesting that the mature OS137 isozyme is not glycosylated. Analysis of the expression of these genes in germinating rice seeds indicated that mRNA corresponding to pOS103 and pOS137 could be detected throughout a 48 h period of seed imbibition. RNA levels, however, were dramatically stimulated by treatment of embryoless half-seeds with exogenous GA3. Our results demonstrate that at least two forms of alpha-amylase are expressed in germinating rice seeds and that the expression of these genes is regulated by the phytohormone GA3.

The promoters of a majority of cereal a-amylase genes contain three highly conserved sequences (gibberellin response element, box I, and pyrimidine box). Recent studies have demonstrated the functional importance of four regions that either coincide with orare immediately proximal to these three conserved elements as well as an upstream Opaque-2 binding sequence. In this study, we describe the characterization of nuclear protein factors from barley aleurone layers whose binding activity toward gibberellin response complex sequences from the barley low-pl a-amylase gene (Amy32b) promoter is stimulated by gibberellin A3 (a3) treatment. Barley proteins isolated from crude nuclear extracts prepared from aleurone layers incubated with or without GA3 were fractionated by anion exchange fast protein liquid chromatography and studied using band shift assays, sequence-specific competitions, and DNase I footprinting. A GA3-dependent binding activity eluting at 210 mM KCI was shown to bind specifically to the gibberellin response element and the closely associated box 1. DNase I footprinting with the proteins in this fraction indicated interactions with sequences in the gibberellin response element and box 1. A second DNA binding activity eluting at 310 mM KCI was present constitutively in extracts prepared from tissues incubated both in the absence and in the presence of hormone. Proteins in this fraction were able to bind to many DNA sequences and, in general, were largely nonspecific. DNase I footprinting with the proteins in this fraction indicated a large ama of protection with a single unoccupied region located at the 3' end of box 1. The possible function of such an activity in hormone regulation of the a-amylase genes is discussed.

Rice genomic clones containing eight different alpha-amylase genes have been previously classified into five groups based on DNA hybridization studies and restriction site mapping. This report describes the clustering of three Group 3 genes (RAmy3A, RAmy3B and RAmy3C) within 28 kb of genomic DNA. The genes are separated from each other by about 5 kb and transcribed in the same direction. At the protein level, RAmy3B and RAmy3C are 95% homologous while each is 78% homologous to RAmy3A. All three genes have relatively small introns in the first and third positions. RAmy3A; however, has an additional 409 bp intron in the second intron insertion site. Nucleotide sequence comparisons of the coding and 3' flanking regions suggest that clustering of the RAmy3 genes occurred by gene duplication resulting from unequal crossing-over at repetitive sequences. A comparison of the 5' flanking regions revealed several sequences that may be involved in transcription. Expression of RAmy3B/C first appears in the germinating seed after two days and at a higher level after four days. Quantitative primer extension analysis indicates that RAmy3B and RAmy3C contribute 25% and 75%, respectively, of the transcripts from this cluster at four days of germination. No primer extension band specific to RAmy3A transcripts could be detected at this time point. However, RAmy3A PCR products could be amplified from RNA isolated from embryo-derived callus tissue.