Integrated bioprocessing of native and engineered components from tobacco (Nicotiana tabacum L.)

by 1978- Whitfield, Matthew Bruce

WHITFIELD, MATTHEW BRUCE. Integrated Bioprocessing of Native and Engineered Components from Tobacco (Nicotiana tabacum L.). (Under the direction of David A. Danehower and R. C. Long) The bioprocessing of tobacco (Nicotiana tabacum L.) is attracting attention as a means for the production of transgenic proteins, as well as native proteins and secondary metabolites, and is doing so at a time when recent reductions in demand for conventionally produced tobacco have created a need for alternative crops. Of the potentially valuable natural products produced by tobacco, many are sequestered on the leaf surface as cuticular wax or trichome exudate. Because of the potential high value of many of these leaf surface components as pharmaceuticals, fragrance elements, and pesticides, as well as the known ease of their removal from the leaf via a simple solvent wash, procedures have been developed to extract these compounds in a manner that could be readily integrated into a tobacco bioprocessing operation. To this end, well-established methods for the analytical extraction of the leaf surface chemistry have been modified to improve their suitability for such integration. A battery of solvents including methanol, ethanol, isopropanol, n-propanol, and acetone were selected for their theorized ability to extract the components of interest as well as the improvements in washed tissue condition and process cost and safety they would likely confer relative to methylene chloride, the analytical solvent of choice. The amounts of some major components (cis-abienol, ?- & ?-cembratrienediols, docosanol, and sucrose esters) extracted from tissue by sequential, timed washes with each solvent were quantified by gas chromatography and compared to amounts extracted by methylene chloride. Most of the solvents tested extracted high levels of the components analyzed with varying degrees of efficiency. Methanol, ethanol, and n-propanol were found to quickly extract levels of components generally comparable to those extracted by methylene chloride. Regardless of any other considerations, methylene chloride and similar solvents are generally not suitable for extraction of components from the surface of the tobacco leaf prior to extraction of proteins because they cause denaturation and precipitation of proteins subsequently extracted from tissue, as well as tissue wilting due to disruption of cellular membranes. All the solvents studied were found to have less deleterious effects on the integrity of the leaf than methylene chloride. In addition, solvent effects on the extraction and crystallization of ribulose bisphosphate carboxylase/oxygenase (Rubisco; “Fraction 1 protein”) from washed tissue were studied. Fortuitously, methanol and ethanol washed tissue were found to result in higher levels of Rubisco crystallization than unwashed tissue. All other solvents, however, including n-propanol, were found to reduce crystallizable Rubisco. Notably, crystallizable Rubisco levels in all cases were increased relative to the amounts of other soluble (“Fraction 2”) proteins. To generally demonstrate the effect of surface washing on transgenic proteins produced in tobacco, tissue from plants modified to produce bovine stomach lysozyme (BSL) was washed with methylene chloride and the alternative solvents. Of the solvents tested, only methanol and isopropanol resulted in significant decreases in lysozyme activity in a total protein extract of washed tissue. Moreover, immunoblotting of extracts from washed tissue indicated that the observed activity reductions may have been caused by inactivation of the enzyme, rather than decrease in extractability. On the bases of surface extraction efficiency, Rubisco crystallization, and lysozyme activity, ethanol would appear to be the best choice for leaf surface extraction in an integrated bioprocess. However, solvent selection would also be dependent upon the relative importance of the products affected; if Rubisco is of much lower value than the other products, then it may be that n-propanol would be a better choice. Furthermore, it is readily apparent that solvent wash effects would need to be studied on a case by case basis for any other native or engineered proteins of interest.