Doctor of Philosophy / Department of Biochemistry / Gerald R. Reeck / Prolamins are the alcohol soluble storage proteins found in the endosperm of seeds from cereals and related grasses. The physical and biochemical properties of prolamins vary between species; and due to their relative abundance can greatly affect the properties and healthfulness of foods from those sources. In this work I investigate peptides from the high molecular weight glutenin of wheat, which is linked to dough elasticity and finished product quality. Using 2D NMR I determined the three-dimensional structure for the repeat peptide Ac- GQQPGQG-Am, which makes up ~50% of the 700 residue central domain. The structure was found to be a flexible β-hairpin with a type II β-turn across residues QPGQ. The NMR structure was later compared to 33 proteins with known three-dimensional structure carrying the exact sequence (backbone RMSD=0.802Å). This finding provides useful insight into the structure of high molecular weight glutenin and the molecular nature gluten elasticity. Alternatively, I studied the kafirin storage prolamins from sorghum, which do not have important physical properties, but are poorly digestible by humans and livestock. Improving digestibility of sorghum could significantly impact human health and nutrition in countries where sorghum is a dietary staple. In this work I devised a unique protocol to isolate kafirins under both non-reducing and reducing conditions. I studied kafirin extracts using SDS-PAGE, HPLC and MALDI-TOF MS, then purified β-kafirin, for the first ever characterization of this single protein. Past studies implicate β-kafirin as a source of poor digestibility due to extensive intermolecular disulfide cross-linking. Contrary to this claim I found more than 50% of β-kafirin was extractable without reducing agents. I used chymotrypsin to digest pure β-kafirin and map 10 cysteine residues to 5 intra-molecular disulfide bonds. Precise pairings have yet to be determined although the protein is largely intact after 12 hours of digestion. This work challenges us to think about sorghum protein body formation and the mechanism that leads to disulfide cross-linking during seed desiccation at maturity.

Sorghum is a valuable grain crop and a principle source of food of particular importance in human and animal nutrition in the semi-arid regions of Africa and Asia. Despite its value, sorghum grain quality is a major limitation to its productivity and profitability. Sorghum grain is usually discounted as feed grain when compared to wheat and barley, predominantly because of its poor digestibility. The sorghum endosperm is composed of a complex starch protein matrix, whereby the starch is physically bound within the storage proteins, the kafirins. The kafirins are synthesised on the membrane bound polysomes and have a signal peptide which targets them to the lumen of the endoplasmic reticulum. Growth of protein bodies occur as - kafirins fill in the interior with  and γ kafirins occupy the periphery. Despite the -kafirins being more digestible and composing of 60- 80% of the kafirins, they are not easily accessible due to  and γ kafirins which have a high content of intermolecular disulphide bonds (S-S), rendering them highly resistant to proteases. Alteration of the structure of the protein bodies and change of the location of the-kafirins could result in a higher digestibility of sorghum proteins. This could be achieved by upregulating or downregulating the -kafirins. The improvement of grain quality, both in increased protein and starch digestibility would substantially enhance the digestibility of sorghum as animal feed as well as for human consumption. Various techniques have been utilised to classify the kafirins according to their mobility on SDS PAGE electrophoresis, Reverse Phase High Performance Liquid Chromatography (RP-HPLC), Free Zone Capillary Electrophoresis (FZCE) and Lab on Chip. Until recently the characterisation and classification of the kafirins generally have relied on the characterisation of zeins from maize. Zeins have about 70% homology to the kafirins both at the nucleotide and amino acid level. Based on the high similarity of the -kafirins to the -zeins, the - kafirins have been classified as 19 and 22 kDa. Despite their 70% homology the migration of the - kafirins on SDS PAGE is quite different to that of the zeins. Hence, I propose a new classification of the -kafirins as 23 kDa and 25 kDa based on their mobility on SDS PAGE Characterisation and cloning of the 23 and 25 kDa genes was performed using QL41 the Queensland inbred line of sorghum. Ten positive clones were isolated from a cDNA library for the 25 kDa and two clones for the 23 kDa -kafirins. The isolated clones of the 25 kDa -kafirins showed 98-99% homology with each other and also with the GenBank sequences. The major finding was the characterisation of the 23 kDa -kafirins. The two clones obtained showed 100% homology to each other as well as to the published sequences on the GenBank, and were full-length sequences. Also a partial sequence was obtained that lacked the signal peptide and was different to the other two clones. Whilst characterising the 23 kDa a second group of the 25 kDa -kafirins was identified from the genomic DNA, of all the three genotypes (QL41, 296B and QL12), which was unique from the previously isolated clones. This group of -kafirins was not among the cluster but was 5’ upstream of the cluster. This group had a higher content of the glutamine compared with the other 25 kDa group. The expression level was studied to show how each gene family contributed to the level of - kafirins. QL41 and 296B were used for this study. From the studies it was shown that the 23 kDa - kafirins genes were 20% more expressed than the 25 kDa. An attempt to identify suitable markers for the -kafirins was investigated using RFLP and SSR analysis. Thirty-two different genotypes were utilised for this study. The observed variation indicated by cluster analysis (4-38%) clearly showed variation of the -kafirins in genotype and within the kafirin genes as elucidated by the sequences in Chapter 4. Markers able to identify this variation could help in the selection of highly digestible mutants. Hence, there is potential for sorghum grain improvement using marker-assisted breeding. The need to identify a tissue specific promoter was essential, especially for a strong promoter that could drive expression in the endosperm of the monocots. A vector construct consisting of the - kafirin promoter driving the GUS reporter genes was used for transient expression from QL41. This was assessed in the sorghum and barley calli, sorghum endosperm and leaves and corn endosperm. Tissue specific expression as well as higher levels of transient expression were seen using the - kafirin promoter, compared with the ubiquitin promoter. Preliminary experiments have illustrated the potential use of a gene silencing mechanism that could enhance the digestibility of sorghum grain. The 25 kDa -kafirin gene was used as the target for gene silencing using the mechanism of iRNA. Transformation constructs were developed using the throughput vector pSTARGATE in an effort to silence the 25 kDa -kafirins. The characterisation of the -kafirins has provided valuable information for future sorghum improvement research.

Although work on the prolamins of pearl millet (Pennisetum glaucum) has revealed partial amino acid sequences for several alcohol-soluble storage proteins (Marcellino et al. 2002) the genes encoding them have not yet been isolated. We constructed a cDNA library from developing P. glaucum seed tissue and screened it using maize zein gene probes to isolate several α-prolamin-like gene sequences. The proteins encoded by these genes generally fall into two size classes: 20.6kD and 27.1kD, which we call the 21kD and 27kD pennisetins. Both proteins are similar in composition and sequence to α-prolamins from maize, sorghum and Coix. Protein bodies that appear as occlusions within the rough ER of P. glaucum endosperm cells are also very similar in size and shape to maize and sorghum protein bodies. The SDS-PAGE gel of the alcohol soluble protein fraction shows two distinct bands in the region corresponding to the 19kD and 22kD of maize α-zein. Both classes of pennisetins appear to be more similar to the 19kD α-zein of maize than to the 22-kD α-zein judging from sequence homology and maize antibody binding. Phylogenetic reconstruction suggests that P. glaucum may have branched from maize prior to the gene duplication which created the 19kD and 22kD α-zein families.