2... .7. 5.... 1. . t . fans... :26...» “i 3 a. 1.2:... ‘31.. . .5 I... .z o. i: i 5., 1V: 2:“, Jr. ‘J- n- 5... S. i. :9 A1 I. 2...... k g . . I ., 2 13...: A €1.64 . 1:, v. .‘ui..f: v6.11 92:... ‘ “Whirfigfiihwméfi. v . . {2.9 .ra... 4...: {1.2.2.3}... a .ii.‘¢.f...l~.:i{2 in we... 8“,...“ zu . uh... .. um . gift... 3 Ilgiuzn «unyifz . 3L1“ .... 3.536. fix 2. .. livl‘z: \__’—————m ‘ LIBRARY fifiéehigan 5'; 1 Univerfisfiiy | , ~v—. ~.——-—- — This is to certify that the dissertation entitled ACETYL-COENZYME A CARBOXYLASE AND THE REGULATION OF FATTY ACID SYNTHESIS presented by Sarah Catherine Hunter has been accepted towards fulfillment of the requirements for Ph-D- degree in Botary and Plant Pathology { Major professor Date ‘4 QLKTLE? AMSU 1.; un Affirmulnv Ar Hun 'Equul Oppnrlumn‘ Irulilunun 0-1277] PLACE IN RETURN Box to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6’01 c:/CIRCIDateDuo_p65-p.15 ACETYL-COENZYME A CARBOXYLASE AND THE REGULATION OF FATTY ACID SYNTHESIS By Sarah Catherine Hunter A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Botany and Plant Pathology 1999 ABSTRACT ACETYL-COENZYME A CARBOXYLASE AND THE REGULATION OF FATTY ACID SYNTHESIS By Sarah Catherine Hunter While seeds often contain the bulk of a plant’s store of fatty acids in the form of triacylglycerols, fatty acids are found throughout plants and fatty acid synthesis (FAS) is one of the primary metabolic pathways. Some plants make “unusual” fatty acids, stored almost exclusively in seeds, which have desirable industrial properties. Factors limiting the use of these oils for industry are 1) plants may not be suitable for crop use and 2) seed may have neither enough total oil, nor a high enough proportion of the unusual fatty acid in the oil, to be cost-effective for industrial use. Current efforts to manipulate both quality and quantity of oil have had mixed results. This is due, in part, to our incomplete knowledge about the components and regulation of FAS in plants. One goal of this thesis has been to extend our understanding of how plants regulate FAS. In plants de novo FAS occurs primarily in plastids and the rate of FAS is regulated in part by acetyl- CoA carboxylase (ACCase). The dicot ACCase investigated in this thesis contains four subunits, one of which, Ach, is encoded on the plastidial genome. This work presents new data about the regulation of ACCase. In lysates of light- incubated chloroplasts, ACCase activity is at least two-fold higher than from dark-incubated ones. Additionally, acetyl-CoA activates ACCase in a chloroplast lysate. This activation occurs at 10 uM acetyl-CoA, which is similar to the acetyl-CoA concentration found in chloroplasts. When ACCase is assayed in vitra under conditions which simulate the in viva concentrations of metabolites, ACCase activity is five- to ten-fold lower than that required to sustain known in viva rates of FAS. Thus, we still have an incomplete understanding of factors which regulate this enzyme. This thesis also presents a model of the major initial reactions of FAS. This model was built using Stella 11 software, pool sizes of FAS metabolites, kinetic characterizations of FAS enzymes, and common sense. The simulation allows enzyme expression to be manipulated, either singly or jointly, between 0- and 25-fold. The model’s predictions match several results reported from over- and underexpression experiments - even when those results have been non-intuitive. The model predicts that the set of FAS enzymes which limit or co-limit flux vary depending an illumination and products (18:1A9 or 16:0). The model also predicts a method involving concurrent overexpression of three enzymes by which total P A exported from plastids can be increased. Finally, this work documents unsuccessful attempts to relocate Ach to the nuclear genome, reports the sequence of Oenothera haakeri Ach, and maps Ach on the plastidial genome of 0. haakeri. Copyright by Sarah Catherine Hunter 1999 to the primary producers of the planet ACKNOWLEDGMENTS Many people have contributed to this research and my graduate education. John Ohlrogge, my advisor, accepted me into his lab, suggested a fascinating (albeit frustrating) research problem, and gave me guidance over the years. He has the ability to be patient at the appropriate time and push at the appropriate time. This work would not have been completed had he not had both those abilities. My committee members, Ken Keegstra, Ken Nadler, and Jon Walton also provided assistance, ideas, and feedback. The work on Oenathera was started in Barb Sears’ lab when I needed a stable place to work my first summer. Without the support and co-operation of the above people, this work would not have been possible. The majority of my funding came from the Michigan State University Biotechnology Training program, however the Michigan Agriculture Experiment Station, the College of Natural Science, and the Graduate School also supported me at times throughout this process. The Cell and Molecular Biology Program, the Biotechnology Training Program, Paul Taylor funds from the Botany Department, the National Plant Lipid Cooperative, and the Conference Organizers at Iowa State all contributed travel funds for meetings. Additional research and travel support came from John and Mary Krimmel. Members of the Ohlrogge lab taught me many things in my time here. Thanks to Sergei Mekhedov and David Pan (for perspective and wonderful bad (Russian) jokes), Ed Cahoon (for numerous references and email support), Nicki Engeseth (who supplied written protocols which worked), Linda Savage (who assured me vi there would always be spinach and usually be peas, and was right), other members of the lab, Oskar Martinez de Ilarduya, Dave Shintani, Vicki Eccleston, Keith Roesler, David Schultz, Xaoming Bao, Mi-Chung Suh, Jim Todd, Mike Pollard, Thomas Girke, Ghislaine Tissot, Ray Pacovski, and Basil Shorrosh all provided techniques, lessons, ideas, and conversation. Visitors Manfred Focke, Abi and Arif, and Hajime provided new perspectives. The newest members of the lab, Jay and Gustavo, have mercifully left me alone the last few weeks. This is no small feat and a great help. I hope I have not forgotten anyone. I also received invaluable support from members of the MSU community. Thanks to Richard Allison for listening and ideas about biotechnology, Peter Walk for attempting to keep my thinking precise, and Bob Bandurski for numerous conversations ranging from methods to history to social concerns. Lara Stoike Steben provided ideas, information, and inspiration, and Georgina Vidal-Gavilan showed me the extremes and benefits of focusing hard at work then hard at play. She also reminded me to use both sides of a piece of paper (at least for drafts). Jim Miller challenged and inspired me to reevaluate my concepts about and methods for practicing science. These scientists are people who expressed faith in me when I lacked it in myself. My partner, Dave Hunter, deserves more than honorable mention for moving to Michigan, entertaining himself for over five years, and listening to countless hours of talk about plant biology for which he had little interest. He also vii "‘ deserves a few months where he doesn’t have to shop, do dishes, and cook. Ed Clayton not only kept me (literally) running throughout much of my time here, but on long runs often vetted my statistics and logic as only a high-energy physicist can. Ed also proofread and caught many mistakes in my “final” copy. Spiritual support from René Breuer and Pat Murphy, both from Portland, helped me choose and kept me on this path. Mary Krimmel (aka Mom) deserves credit for my mathematical approach to problems. The rest of my family, John A., John E., Sharon, Elizabeth, Billy, Bob, Berdie, Stephanie, Craig, Bronwen, Max, Bonnie, Ruth, Bill, Margot, Mark, Jordan, Krista, David, Cindy, and Erica provided challenges, perspectives, cool vacations (such as the San Diego desert when comet Hale-Bopp was visible), and graphic consultations. There are numerous others who gave me ideas or just said “hello” in the halls. It helped. Thank you. viii TABLE OF CONTENTS LIST OF TABLES .......................................... LIST OF FIGURES ......................................... LIST OF ABBREVIATIONS .................................. COMMON NAMES OF FATTY ACIDS .......................... CHAPTER 1 INTRODUCTION ........................................... Synthesis of Fatty Acids ................................ Regulation of FAS .................................... Proteins of fatty acid synthesis ........................ ACCase is a rate-determining enzyme of FAS ............. ACCase structure .................................. ACCase genes .................................... Oils are an important commodity ......................... Literature Cited ......................................... CHAPTER 2 REGULATION OF SPINACH CHLOROPLAST ACETYL-COA CARBOXYLASE ........................................... Abstract ............................................... Introduction ............................................ Materials and Methods .................................... ACCase activity assay A ................................ ACCase activity assay B ................................ Removal of acetyl-CoA from activated chloroplast lysates ...... ACCase assays in different metabolic environments ........... Results ................................................ Analysis of Products of ACCase assay by TLC ............... ACCase activity is two-fold higher in lysates of a light-incubated chloroplast suspension than lysates of a dark-incubated chloroplast suspension .......................................... ACCase activity increases afier pre-incubation of chloroplast lysates with CoA, acetyl-CoA, and short-chain acyl-CoA’s ...... ACCase activity of chloroplast lysates in the presence of DTT and thioredoxin .......................................... ACCase activity in different biochemical environments ......... Discussion ............................................. Light activation (or dark deactivation) of ACCase in a lysed chloroplast suspension ................................. ACCase activity after pre-incubation of chlorOplast lysates with substrates ........................................... ix xii xiii XV xvi MHWMQO‘MNH ## 17 l7 19 23 23 24 25 26 26 28 31 32 38 4o 41 41 42 Activation of ACCase by reductants ....................... ACCase activity in different biochemical environments ......... Acknowledgments ........................................ Literature Cited ......................................... CHAPTER 3 COMPUTER SIMULATIONS OF THE INITIAL REACTIONS OF FAS . . Abstract ............................................... Introduction ............................................ Materials and Methods .................................... Materials ........................................... Computer Program - Stella II ......................... Kinetic Data and Metabolite Pool Sizes .................. Numerical Methods ................................ Methods ............................................ Scope of the Model ................................. Implementation ................................... Assumptions ...................................... Model Validation .................................. Comparison of model predictions to transgenic plant results . . Results and Predictions .................................... ACCase, KAS I, and KAS III limit or co-limit flux through FAS in the light ............................................. ACCase, KAS I, and KAS II limit or co-limit flux through FAS in the dark ............................................ Increase of Fatty Acid Synthesis by Enzyme Overexpression ..... Overexpression of the enzymes at the end of the pathway does not increase FAS ........................................ Discussion ............................................. Model validation relies heavily on in viva KAS III overexpression experiments ......................................... Limiting and co-limiting enzymes in the light ................ Literature Cited ......................................... CHAPTER 4 ENGINEERING B-CT AND THE STRUCTURE OF Ach ............ Abstract ............................................... Introduction ............................................ Materials and Methods .................................... Materials ........................................... Surface Sterilization of Seeds ............................ Construction of B-CT transformation vectors ................. Plant Transformation .................................. Seed Germination Assays ............................... Neomycin phosphotransferase assays ...................... Western Blotting ..................................... 45 45 47 48 52 52 54 61 61 61 62 62 67 67 68 68 78 79 84 84 88 91 97 100 101 102 106 111 111 113 116 116 116 116 119 121 121 122 RNA Extraction 1 ..................................... RNA Extraction 2 ..................................... Northern Probes Construction ............................ Engineered B-CT .................................. Small Subunit Riboprobe ............................ 35S/B-CT probe ................................... Kanamycin probe .................................. Sequencing and Sequence Analysis ........................ Mapping of Ach on the 0. haakeri plastome ................ Results ................................................ Transformation of N. tabacum with B-CT constructs ........... Immunoblots of transformed plants ........................ RNA blots .......................................... Sequence of Oenathera haakeri Ach ...................... Mapping of Oenothera haakeri Ach on the plastidial genome . . . Discussion ............................................. Plant transformation and extra-plastidial expression of Ach . . . . Oenothera Ach location on the plastidial genome and sequence . . Acknowledgements ....................................... Literature Cited ......................................... CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTIONS .................... Future Directions ........................................ How does a plant control how much and what types of fatty acid are synthesized? ...................................... Is FAS Channeled? .................................... Light/dark regulation of FAS ............................ Will biotechnology be able and allowed to actually help? ....... Literature Cited ......................................... APPENDIX A . NUMERICAL METHODS .................................... Literature Cited ......................................... xi 122 123 123 123 124 124 125 125 126 126 126 127 128 131 136 140 140 144 147 148 150 153 153 155 157 160 162 167 172 Table 2.1 Table 2.2 Table 3.1 Table 3.2 Table 3.3 Table 3.4 Table 3.5A Table 3.58 Table 3.6 Table 3.7 Table 4.1 LIST OF TABLES Composition of light and dark stromal environments and resultant ACCase activity .......................... Effect of ethyl adenylate, CoA, malonyl-CoA, and short- chain acyl-CoA’s on ACCase activation and activity ...... Pool sizes of FAS intermediate metabolites from spinach and pea chloroplasts .............................. Kinetic data for FAS enzymes used in the Stella Model . . . . Additional criteria used to build the model ............. Components used in the Stella model of F AS ............ Kinetic data used to implement the Stella Model ......... Initial pool size of metabolites used in the Stella model of FAS and values of model Km’s ...................... Expression experiments and model predictions .......... Summary of limiting and co-limiting enzymes in the light and dark ....................................... Gene order in plastidial genomes ..................... xii 27 37 63 65 66 74 75 76 81 92 145 Figure 1.1 Figure 1.2A Figure 1.2B Figure 2.1 Figure 2.2A Figure 2.2B Figure 2.3 Figure 2.4 Figure 3.1 Figure 3.2A Figure 3.2B Figure 3.2C Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Figure 3.7 Figure 3.8 LIST OF FIGURES Major plastidial reactions of F AS ..................... ACCase structure ................................. ACCase localization ................................ Representative autoradiogram of acid-stable products separated on silica gel ............................. Specific activity of ACCase vs. total assay time .......... ACCase assay of chloroplast suspension lysed for various times .......................................... Activity of ACCase after removal of acetyl-CoA .......... Effect of DTT on ACCase activity ..................... Major plastidial reactions of FAS ..................... Stella diagram of the model of the initial reactions of FAS . . Stella diagrams of total CoA and ACP pools .............. Stella diagram of the simulation layer of the model ......... Theoretical consequences of altering enzyme activity and its effects on pathway flux ............................. Flux through the FAS pathway vs. enzyme activity under states of high flux ...................................... Flux through the FAS pathway vs. enzyme activity under states of low flux ....................................... The effect of altered expression levels of single enzymes in the model ........................................... Model predictions of the effects of double enzyme expression . Predictions of the effect of the altered expression of three enzymes ......................................... xiii 3O 33 34 36 39 56 69 71 72 85 86 89 94 95 96 Figure 3.9 Figure 3.10 Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 Figure 4.7 Figure 4.8 Figure A] Concurrent altered expression of ACCase, KAS I, KAS II, and KAS III ........................................ 98 Overexpression of the enzymes at the end of FAS pathway . . . 99 Construction of plasmids p0921-3.3 and p0921-3.5 ......... 117 Pea B-CT“ construct ................................ 120 RNA blots from plants transformed with plasmids p0921-3.3 and p0921-3.5 .................................... 129 RNA blots from plants transformed with plasmid B-CT* ..... 132 Preliminary 0. haakeri Ach sequence .................. 134 Protein alignment and consensus of selected .4ch genes ..... 137 '0. haakeri plastome maps and the location of the Ach gene . . 139 Areas where the Ach transcript may be improperly processed 142 Graphical representation of Euler’s method ............... 169 xiv LIST OF ABBREVIATIONS ACCase ................................... acetyl-CoA carboxylase ACP .......................................... acyl carrier protein ACS ....................................... acetyl-CoA synthetase ADP ....................................... adenosine diphosphate ATP ....................................... adenosine triphosphate BAP .............................................. benzyladenine BC ............................................ biotin carboxylase BCCP ................................. biotin carboxyl carrier protein chl ................................................. chlorophyll CoA .............................................. . Coenzyme A CT ............................. carboxyltransferase (transcarboxylase) DTT .............................................. dithiothreitol EDTA ................................ ethylenediaminetetraacetic acid FA .................................................. fatty acid FAS .......................................... fatty acid synthesis Hepes ............. N-(Hydroxyethyl) piperazine-N’-2-ethanesulfonic acid) KAS .......................... 3-ketoacyl-acyl carrier protein synthase MCT ................................. malonyl CoAzACP transacylase MCTE ................................... medium chain thioesterase MF .............................................. multifunctional min ................................................... minutes MS ............................................... multi-subunit NAA ......................................... naphaleneacetic acid NAD+ ................. nicotinamide adenine dinucleotide (oxidized form) NADH .................. nicotinamide adenine dinucleotide (reduced form) NADP+ ........ nicotinamide adenine dinucleotide phosphate (oxidized form) NADPH ........ nicotinamide adenine dinucleotide phosphate (reduced form) nd .................................................. not detected ORF .......................................... open reading frame PAGE .............................. polyacrylimide gel electrophoresis PFK .......................................... Phosphofructokinase RuBisCO ................. ribulose-1,5-bisphophate carboxylase/oxygenase SD ............................................ standard deviation SDS ........................................ sodium dodecyl sulfate SE ....................................... standard error of the mean TAG .............................................. triacylglycerol TLC ..................................... thin layer chromatography XV COMMON NAMES OF FATTY ACIDS laurate .................................................. 12:0 myristic ................................................. 14:0 palmitate ................................................ 16:0 petroselinic ............................................ 16: 1A4 stearate .................................................. 18:0 oleate ............................................ 18:1A9, 18:1 ricinolenic ........................................ 18:1A9, 120H cis—vaccenic ............................................ 18: IA] 1 crepenynic ..................................... C18:1A9, 12-C-'=C linoleic .................................................. 18:2 linolenic ................................................. 18:3 erucic ................................................ 22:1A13 xvi Chapter 1 INTRODUCTION Oilseeds produce around 80 million metric tons of oil per year and the total value of the vegetable oils produced is approximately 70 billion US dollars/year. About two-thirds of the oil produced is used for human consumption, while the rest is used by industry (Chrispeels and Sadava, 1994). While seeds often contain the bulk of a plant’s store of fatty acids in the form of triacylglycerols (oils), fatty acids are found throughout plants: as major components of cell membranes, as signaling molecules, as membrane anchors for proteins, and as components of waxes in the cuticle. Thus, fatty acid synthesis is one of the primary metabolic pathways. Recent reviews of fatty acid synthesis can be found in Ohlrogge and Browse (1995), Harwood (1996), and Ohlrogge and Jaworski (1997). Fatty acids consist of a hydrocarbon chain and a carboxyl group. Thus, they are extremely hydrophobic at one end and hydrophilic at the carboxy end. Free fatty acids are generally toxic to cells, and therefore are bound to other components: glycerol backbones in the case of triacylglycerols (TAG), or phospho-, sulfo-, or galacto-glycerols in the case of membranes. There are also acyl carrier proteins (ACP’s) used during fatty acid synthesis. The common fatty acids in plants have 16 and 18 carbon acyl groups, with zero to three double bonds. These double bonds are almost always in the cis configuration. The double bonds are usually in the A9, A12, and A15 positions. While the common names are occasionally used, more often fatty acids are designated by their chain length and the number of double bonds. Thus, 18:1A9, often just 18:1 (oleate), is a fatty acid with 18 carbons and one double bond. 16:0 (palmitate) has sixteen carbons and no double bonds. The other major fatty and “common” fatty acids found in all plants are 18:2 (linoleic) and 18:3 (linolenic). Plants can also manufacture “unusual” fatty acids (reviewed in van de Loo et al., 1993). These fatty acids are generally found in seed oils and may be shorter or longer than the common fatty acids, may have double bonds in non-standard positions, or may have additional functional groups: hydroxy groups in the case of ricinoleic acids or acetylenic groups in the case of crepenynic acid. These unusual fatty acids often have desirable industrial properties. However, the use of fatty acids in industry is limited by practical considerations: the plant which makes a specific desirable fatty acid may not be adaptable to large-scale agriculture or may not make and store the unusual fatty acid in a high enough quantity to be cost-effectively purified. Synthesis of Fatty Acids In plants, almost all de nova fatty acid synthesis takes place in plastids. In contrast, in other eucaryotic systems FAS occurs in the cytosol. Furthermore, plants also have a different type of fatty acid synthetase than other eucaryotes. Animals and yeast have a type I fatty acid synthetase. In this case, the numerous functions required to increase an acyl chain by two carbon units are encoded on the same gene and reside on the same protein for animals or on two genes/proteins for yeast. This is not true of the plant FAS system. The plant FAS system consists of a type II enzyme, which is very similar to bacterial fatty acid synthetase. Type II fatty acid synthetase consists of at least six separately encoded proteins that are easily separable but may be in a loose association. Thus, rather than being structurally related to other eucaryotic fatty acid synthetases, the plant fatty acid synthetase appears to be most closely related to bacterial fatty acid synthetase. How is a fatty acid made in a plant? Figure 1.1 shows the major initial reactions of FAS. Malonyl-ACP is the building block of FAS. An initial condensation between malonyl-ACP and acetyl-CoA is catalyzed by 3-ketoacyl acyl carrier protein synthase III (KAS III). This yields a four carbon 3-ketoacy1-ACP, 3-ketobutyryl-ACP, which is then reduced, dehydrated, and reduced a second time to yield a saturated four carbon acyl chain attached to ACP via a thioester bond. Further elongation of the acyl chain occurs when a second, third, fourth, etc., malonyl-ACP is added to the reactive thioester end of the chain. Normally, this process continues until the acyl chain is 16 or 18 carbons long. Sixteen carbon acyl chains can be hydrolyzed via thioesters and exported from the plastid, or they can be retained in the plastid by the action of acyltransferases which incorporate them into membrane lipids. Almost all eighteen carbon fatty acids are desaturated at the A9 position before being exported or incorporated .- I siliusl 4l NADP+ 3-hydroxyacyl-ACP NADPH H20 3-hydroxyacyl-ACP 3-ketoacyl-ACP + H” dehydrase reductase trans-A2 acyl enoyl-ACP 3-ketoacyl-ACP NAD(P)H + H“ C02 Enoyl-ACP reductase NAD(P)+ acyl-ACP (2 C longer) malonyl-ACP KAS III 16:0 ACP 13:0 ACP MCT CoA Acyl-ACP acetyl-CoA th' t noes erase ATP _ ACCase ACP ACP 16:0 (palmitate) malonyl-CoA ADP, Stearoyl-ACP Desaturase 18:1 ACP ACP Oleoyl-ACP thioesterase 18:1(oleate) Figure 1.1. Major plastidial reactions of FAS. Enzymatic steps are in boldface type. See text for further explanation. Figure after Ohlrogge and Browse (1995). into membrane lipids in the plastid. Further desaturation occurs outside the plastid. Malonyl-ACP, the building block, is formed by the carboxylation of acetyl-CoA, an ubiquitous metabolite, and the subsequent transfer of the malonyl moiety from CoA to ACP. Acetyl-CoA carboxylase (ACCase, EC. 6.4.1.2) catalyzes the first reaction. Malonyl-CoAzACP transacylase catalyzes the second. Regulation of FAS Although FAS is a primary pathway, it is still unclear how a plant cell regulates the amount of fatty acids it produces (reviewed in Ohlrogge and Jaworski, 1997). There are numerous demands for fatty acids that vary according to cell type and developmental state. For example, plant cells undergoing rapid expansion require fatty acids for membrane synthesis, developing seeds require fatty acids in the form of TAG during oil deposition, and epidermal cells require fatty acids for wax and cuticular lipids. To make the regulation more complex, there is also a compartmentalization issue: most, if not all, of the de nova FAS occurs in plastids. Membrane phospholipids, triacylglycerols, waxes, etc. are formed in the cytosol. Therefore, there must be some mechanism of communication between the two compartments. Light affects the rate of fatty acid synthesis both in vitra and in viva. Under optimal conditions, which includes 200 ,uE * m'2 *sec", in vitra rates of fatty acid synthesis are 1-2 nmoles 2 C units (acetate) "' hour'l * mg chl'1 in intact n- - ._——=—- chloroplasts (Roughan, 1987). This value agrees with recent unpublished data on the in viva rate of FAS (Bao et al., in preparation). Although isolated chloroplasts were unable to synthesize fatty acids without light, in leaf discs which were not exposed to light, the rate of FAS dropped six to ten-fold compared to leaf disks exposed to light (Browse et al., 1981). In the in viva studies, essentially no FAS was seen without illumination (Bao et al., in preparation). A number of studies have attempted to determine the mechanism of light activation of FAS. Light dependence has been attributed to changes in the stromal pH (Sauer and Heise, 1983). Many, but not all, of the FAS enzymes have pH optima between 8 and 8.5. Differences in ATP/ADP ratio, Mg“, and reductant concentrations have also been proposed as regulatory factors (Sauer and Heise, 1983). A number of these explanations have been questioned: ATP and reducing power are not sufficient to substitute for light activation (e. g. Stumpf et al., 1963; Nakamura and Yamada, 1975). In Chlamydomonas, electron flow through photosystem I is required for FAS, but the relationship is not linear (Picaud et al., 1991). Proteins of fatty acid synthesis The plant fatty acid synthesis system is made up of at least 11 enzymes as shown in Figure 1.1. The first committed step, carboxylation of acetyl-CoA, is performed by acetyl-CoA carboxylase (ACCase). The transfer of the malonyl moiety to ACP is catalyzed by malonyl-CoA2ACP transacylase (MCT). There are three enzymes which perform the condensing reactions, 3-ketoacyl-acyl carrier protein synthase 1, II, and III (KAS I, KAS II, and KAS III). The two reduction steps are performed by 3-ketoacyl-ACP reductase and enoyl-ACP reductase. The dehydration step is catalyzed by 3-hydroxyacyl-ACP dehydrase. There are two common thioesterases, oleoyl-ACP thioesterase and a less specific acyl-ACP thioesterase, and finally, a stearoyl-ACP desaturase. All of the above reactions, except ACCase, require ACP, a small (~9 kDa) protein, which is attached to the acyl chain during processing. AC Case is a rate-determining enzyme afFAS ACCase plays a major role in determining the rate of FAS. ACCase shares many in vitra properties with regulatory enzymes. Most importantly, however, in viva data also suggest a regulatory role for ACCase. If flux through a pathway can be increased or decreased, concomitant changes in the pool sizes of intermediates in the pathway can indicate which enzyme(s) is (are) regulatory. When flux is decreased, metabolites prior to the regulatory step will increase, and those after the regulatory step will decrease. The converse is true if flux is increased: metabolite pools prior to a regulatory step will decrease and metabolite pools after the regulatory step will increase (Rolleston, 1972). Flux through FAS can be decreased by removing light and for isolated chloroplasts, it can be increased by Triton X-100. Through analysis of pool size data under states of high flux and low flux, Post-Beittenmiller et al. (1991, 1992) determined that ACCase is regulatory for dicot systems. Analysis of control coefficients of barley and maize showed that ACCase is regulatory in gramineaceous systems (Page et al., 1994). Of course, other enzymes also play a role in rate-determination: KAS III, the enzyme that performs the initial condensation reaction, is likely also a player (Verwoert et al., 1995). AC Case structure ACCase is a biotin-containing enzyme with three functional subunits: a biotin carrier protein (BCCP), a biotin carboxylase (BC), and a carboxyltransferase (CT). Knowledge of the overall structure of ACCase has recently been revised (reviewed in Sasaki et al., 1995). It varies according to both cellular localization and plant species as shown in Figure 1.2. There is both a type I (also called homomeric or multifunctional) ACCase and a type II (also called heteromeric or multisubunit) ACCase. Dicots and most monocots have a type II ACCase in their plastids, but they also have a type I ACCase, presumably in the cytosol. Gramineaceous plants have only type I ACCases. However, there have been reports of homomeric ACCase in Brassica napus plastids (Schulte et al., 1997), so our picture of ACCase localization and structures may still be incomplete. ACCase genes The genes that encode three of the four subunits of ACCase are located on the nuclear genome. However, the gene that codes for the B-CT subunit, Ach, is retained in the plastidial genome in dicots and presumably most non-grass Fligur DiCO:S “5-1 :lIOSC ”99s, Figure 1.2A. ACCase structure. The functional units of ACCase can be on the same polypeptide or on different polypeptides. Multifunctional (MF) ACCase has the three functional subunits, a biotin carboxylase (BC), a biotin carrier protein (BCCP), and a carboxyltransferase (CT) on one peptide. Multisubunit (MS) ACCase has at least four separable subunits: BC, BCCP, a-CT and B-CT. Both a-CT and B-CT are necessary for the carboxyltransferase function. Figure after Ohlrogge and Browse (1995) Figure 1.2B. ACCase localization. Dicots and most monocots have both MS-ACCase and MF-ACCase. The MS-ACCase is located in the plastid and MF-ACCase is presumably in the cytosol. Grasses have only MF-ACCase. Figure after Ohlrogge and Browse (1995) homomeric multifunctional BC BCCP CT type I heteromeric multisubunit BC BCCP j a-CT/B-CT | type 11 Figure 1.2A. ACCase structure Dicots G Most Monocots rasses / MF-ACCase MF-ACCase \ (plastid) MS-ACCase \ J Figure 1.2B. ACCase localization. 10 C plastid) \ MF-ACCase J 30“; fr J _r monocots. In rice the plastidial genome has an ORF one-fifth the size of tobacco Ach with strong homology to the gene. However, it is clearly truncated and not a functional gene. Oils Are an Important Commodity Aside from curiosity, there are economic reasons to investigate the regulation of FAS in plants. Plants currently provide valuable seed oils, most of which are used for human consumption. While traditional breeding methods have been able to alter the amount of oil in seeds (e. g. high and low oil corn) and have been able to alter fatty acid composition (e. g. low- and high-erucic rapeseed lines), this process has been time-consuming. Plants make a wide variety of unusual fatty acids which could become important industrial feedstocks if the quantity and quality of the fatty acid in the seed oil of crop plants could be increased (reviewed in Ohlrogge, 1994; Topfer et al., 1995). Thus, many efforts have focused on genetically engineering plants to alter and increase their oils (e. g., Voelker et al., 1992; Broun and Somerville, 1997). These efforts have met with limited success (e. g. Roesler et al., 1997; and reviewed in Ohlrogge and Jaworski, 1997), especially when the goal has been an increase in total oil content. What controls how much and what types of fatty acids a plant makes? It is vital to answer these questions if plant oils are to increase their agricultural value. Thus, understanding the regulation of de nova plant fatty acid synthesis is 11 useful. Furthermore, ACCase regulation, as a major determinant of the rate of FAS, needs to be understood. The work described below adds to, confirms, and furthers the body of knowledge known about the regulation of ACCase and FAS. Chapter 2 presents data on the metabolic regulation of ACCase. A more precise, albeit time-consuming, ACCase assay was developed to investigate the differences between activity in light-incubated and dark-incubated lysed chloroplasts. Even when assayed in the same environment (e. g. pH, and ATP, acetyl-CoA, and Mg2+ concentrations), light-incubated chloroplasts had at least two-fold higher activity than dark- incubated ones. This difference decreased over time, until ACCase activity in dark-incubated lysed chloroplasts rose to the activity of light-incubated ones. Chapter 2 also presents the discovery of an additional activator for ACCase in chloroplast lysates: its substrate, acetyl-CoA. This activation occurs at low concentrations of acetyl-CoA that are likely to occur in chlorOplasts. Chapter 2 also introduces a theme discussed in Chapter 3. Which of the observations made in vitra can be applied in viva? Knowledge of the concentrations of chloroplastic metabolites are reasonably precise, yet ACCase activity in a simulated chloroplast environment is five- to ten-fold lower than that required to sustain known in viva rates of FAS. Additionally, reductants have been shown to activate ACCase in vitra, but evidence in Chapter 2 questions whether or not the activation is physiologically relevant. 12 Chapter 3 takes a different approach. Although Chapter 2 clearly argues against indiscriminate application of in vitra knowledge to in viva systems, the approach taken in Chapter 3 is a step further removed from in viva systems than in vitra data. Chapter 3 presents a computer simulation of the initial reactions of FAS. This computer simulation was built using metabolite pool sizes in the light, characterizations of F AS enzymes, and common sense. As detailed above, FAS is a complex pathway and many of the details of its regulation may be non- intuitive. The computer simulation allows enzymes to be expressed, either singly or jointly, between 0 and 25-fold. The model’s predictions match many results from transgenic over- and underexpression experiments, even when the results have been non-intuitive. KAS III overexpression is a case where the non- intuitive results from plants have been matched by the model. The model was then used to investigate limiting and co—limiting enzymes in the light and dark for production of both 18:1 and 16:0. The model predicts that different enzymes take over the major regulatory role(s) depending on both products and environment. Although there is precedent for this in carbohydrate metabolism, this is not something that has been seriously considered for FAS in the past. The model was also used to predict a method by which total FA exported from the plastids could be increased. Chapter 4 illustrates that testing the ideas put forth in Chapter 3 will not be straightforward. Two attempts to relocate the plastidial gene of ACCase, Ach, 13 to the nucleus are documented. Initially, the goal was to alter Ach expression to test coordinate regulation of subunits and, hopefully, alter the rate of FAS. Both these attempts failed. Chapter 4 details the attempts, and provides ideas as to why the relocation failed and how it might be made to succeed. In addition, Chapter 4 presents the sequence of Oenathera haokeri Ach, and maps its location on the plastidial genome. 14 Literature Cited Bao, X., Focke, M., Pollard, M., and Ohlrogge, J. B. Understanding in Viva carbon precursor supply for fatty acid synthesis in leaf tissue. In preparation. Broun, P., and Somerville, C. (1997). Accumulation of ricinoleic, lesquerolic, and densilipolic acids in seeds of transgenic arabidopsis plants that express a fatty acid hydroxylase cDNA from castor bean. Plant Physiol. 113, 933-942. Browse, J., Roughan, P. G., and Slack, C. R. (1981). Light control of fatty acid synthesis and diurnal fluctuations of fatty acid composition in leaves. Biochem. J. 196, 347-354. Chrispeels, M.J., and Sadava, D.E. (1994). Plants, Genes, and Agriculture. (Boston: Jones and Bartlett). Harwood, J. L. (1996). Recent advances in the biosynthesis of plant fatty acids. Biochim. Biophys. Acta 1301, 7-56. Nakamura, Y., and Yamada, M. (1975). Fatty acid synthesis by spinach chloroplasts 1. Property of fatty acid synthesis from acetate. Plant Cell Physiol. 16, 139-149. Ohlrogge, J. B. (1994). Design of new plant products: engineering of fatty acid metabolism. Plant Physiol. 104, 821-826. Ohlrogge, J., and Browse, J. (1995). Lipid biosynthesis. Plant Cell 7, 957-970. Ohlrogge, J. B., and Jaworski, J. G. (1997). Regulation of fatty acid synthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 48, 109-136. Page, R. A., Okada, S., and Harwood, J. L. (1994). Acetyl-CoA carboxylase exerts strong flux control over lipid synthesis in plants. Biochim. Biophys. Acta 1210, 369-372. Picaud, A., Creach, A., and Trémoliéres, A. (1991). Studies on the stimulation by light of fatty acid synthesis in Chlamydomonas reinhardii whole cells. Plant Physiol. Biochem. 29, 441-448. Post-Beittenmiller, D., Jaworski, J. G., and Ohlrogge, J. B. (1991). In viva pools of free and acylated acyl carrier proteins in spinach. J. Biol. Chem. 266, 1858-1865. 15 Post-Beittenmiller, D., Roughan, G., and Ohlrogge, J. B. (1992). Regulation of plant fatty acid biosynthesis: analysis of acyl-Coenzyme A and acyl-acyl carrier protein substrate pools in spinach and pea chloroplasts. Plant Physiol. 100, 923-930. Roesler, K., Shintani, D., Savage, L., Boddupalli, S., and Ohlrogge, J. (1997). Targeting of the arabidopsis homomeric acetyl-Coenzyme A carboxylase to plastids of rapeseeds. Plant Physiol. 113, 75-81. Rolleston, F. S. (1972). A theoretical background to the use of measured concentrations of intermediates in study of the control of intermediary metabolism. Curr. Top. Cellu. Reg. 5, 47-75. Roughan, G. (1987). Long-chain fatty acid synthesis and utilization by isolated chloroplasts. Meth. Enzymol. 148, 327-337. Sasaki, Y., Konishi, T., and Nagano, Y. (1995). Update on biochemistry: the compartmentation of acetyl-Coenzyme A carboxylase in plants. Plant Physiol. 108, 445-449. Sauer, A., and Heise, K. (1983). On the light dependence of fatty acid synthesis in spinach chloroplasts. Plant Physiol. 73, 11-15. Schulte, W., Tiipfer, R., Stracke, R., Schell, J., and Martini, N. (1997). Multi-functional acetyl-CoA carboxylase from Brassica napus is encoded by a multi-gene family: Indication from plastidic localization of at least one isoform. Proc. Natl. Acad. Sci. USA 94, 3465-3470. Stumpf, P. K., Bové, J. M., and Goffeau, A. (1963). Fat metabolism in higher plants. XX: Relation of fatty acid synthesis and photophosphorylation in lettuce chloroplasts. Biochim. Biophys. Acta 70, 260-270. Tiipfer, R., Martini, N., and Schell, J. (1995). Modification of plant lipid synthesis. Science 268, 681-686. van de Loo, F.J., Fox, B.G., and Somerville, C. (1993). Unusual fatty acids. In Lipid Metabolism in Plants. T. S. Moore, (Boca Raton: CRC Press). pp. 91-126. Verwoert, I. I. G. S., van der Linden, K. B., Walsh, M. C., Nijkamp, H. J. J., and Stuitje, A. R. (1995). Modification of Brassica napus by expression of the Escherichia cali fabH gene, encoding 3-ketoacyl-acyl carrier protein synthase III. Plant Mol. Biol. 27, 875-886. Voelker, T. A., Worrell, A. C., Anderson, L., Bleibaum, J., Fan, C., Hawkins, D. J., Radke, S. E., and Davies, J. M. (1992). Fatty acid biosynthesis redirected to medium chains in transgenic oilseed plants. Science 257, 72-74. 16 Chapter 2 REGULATION OF SPINACH CHLOROPLAST ACETYL-COA CARBOXYLASE‘ Abstract We have investigated several factors which influence acetyl-CoA carboxylase (ACCase) activity in lysed spinach chloroplasts. 1) When assayed after rapid lysis of light-incubated chloroplasts, ACCase activity was two-fold higher than activity from dark-incubated chloroplasts. Within five minutes after lysis, activity from dark incubated chloroplasts increased, suggesting a transient inactivation or inhibition of ACCase in the dark. 2) When lysed chloroplast suspensions were incubated with 30 to 100 uM acetyl-CoA before starting assays, activity was four-fold higher than if suspensions were not pre-incubated with acetyl-CoA. CoA, malonyl-CoA, propionyl-CoA and butyryl-CoA also activated ACCase. Full acetyl-CoA activation required MgATP and was essentially complete afier eight minutes. ACCase activity decreased upon removal of acetyl-CoA by gel-filtration and was partially restored by readdition of acetyl-CoA. Thus, ACCase activation by acetyl-CoA was reversible. 3) Dithiothreitol and thioredoxin stimulated ACCase activity, but only in 1 This chapter was previously published in Archives of Biochemistry and Biophysics 359, 170-178. COpyright O by Academic Press. Permission not required for use in author’s dissertation. Minor changes have been made. 17 am 3'3 preparations where ACCase activity was low. 4) ACCase was assayed in concentrations of ATP, ADP, NADPH, NADP+, Mg“, and C02/HCO3' which are estimated to occur in the stroma of chloroplasts under illumination or darkness. ACCase activity from lysed chloroplast suspensions was ten-fold higher when illuminated conditions were used. However, this activity was still five-fold to ten-fold lower than the rates required to sustain known in viva rates of fatty acid synthesis and in vitro rates achieved under optimum assay conditions with saturating substrates. 18 Introduction In spite of the central role of fatty acid synthesis (F AS) in plants and numerous studies of FAS regulation, it is still unclear how a plant cell regulates the amount of fatty acids it produces (reviewed in Ohlrogge and Jaworski, 1997). FAS is six- to ten-fold higher in illuminated leaves than non-illuminated ones (Browse et al., 1981), and a number of studies have attempted to determine the factors involved in this light control. Some of the light-dependence of FAS can be attributed to changes in the stromal pH, Mg” concentration, ATP/ADP ratio, and reductants (Sauer and Heise, 1983). However, not all of the changes in F AS can be accounted for by these factors. Early studies showed ATP and reducing power are not able to substitute for light and thus imply the role of light in fatty acid synthesis is more than simply to provide those substrates (Sauer and Heise, 1983; Stumpf et al., 1963; Nakamura and Yamada, 1975). Studies in Chlamydomonas mutants lacking photosystem I show light activation of fatty acid synthesis is photosystem I dependent, but the relationship between photosystem I electron flow and FAS light activation is not linear (Picaud et al., 1991). Thus, light does more than provide metabolites for FAS. Furthermore, while it is clear light does play a regulatory role, the mechanisms by which this happens are not clear. The carboxylation of acetyl-CoA to produce malonyl-CoA by acetyl-CoA carboxylase (ACCase, EC 6.4.1.2) is the first committed step of FAS and ACCase is the enzyme considered most responsible for the primary regulation of 19 FAS. Many of the properties of ACCase are consistent with it being a control enzyme. ACCase is the first committed step of the pathway. Metabolite pool size analyses under states of high flux (light-incubated leaves or chloroplasts) and low-flux (dark-incubated leaves or chloroplasts) are consistent with a regulatory role (Post-Beittenmiller et al., 1991 and 1992b). Flux control coefficients for ACCase from barley and maize leaves have been estimated to be 0.5 to 0.6 which also indicate that ACCase plays a major role in FAS regulation (Page et al., 1994). Furthermore, in animals and other non-plant systems, ACCase has been determined as a primary regulatory step (Salati and Goodridge, 1996). Understanding of plant ACCase structure has been substantially redefined in the last five years. In most plants, such as spinach or pea, plastidial ACCase is an easily dissociated multisubunit enzyme (Kannangara and Stumpf, 1972; Sasaki et al., 1995; Alban et al., 1994). However, grasses have a large homodimeric multifunctional plastidial enzyme (Sasaki et al., 1995) and there is evidence that a multifunctional ACCase exists in Brassica napus chloroplasts (Schulte et al., 1997). In most cases, the multisubunit and multifunctional enzymes appear to be regulated in a similar manner. Both have pH optima near 8.0 to 8.5 (Alban et al., 1994; Nikolau and Hawke, 1984; Bettey et al., 1992; Sauer and Heise, 1984; Sasaki et al., 1997), both are competitively inhibited by ADP and to a lesser extent AMP (Nikolau and Hawke, 1984; Bettey et al., 1992; Sauer and Heise, 1984; Eastwell and Stumpf, 1983). Citrate and long chain acyl-CoA are considered key regulators of animal ACCase (Salati and Goodridge 1996; Allred 20 and Reilly, 1997). Long chain acyl-CoA inhibits plant ACCase (Shintani and Ohlrogge, 1995), but reports on citrate regulation of plant ACCase have been inconsistent (Alban et al., 1994; Nikolau and Hawke, 1984; Mohan and Kekwick, 1980) The biochemical environment of the chloroplast stroma undergoes numerous changes upon illumination. Stromal H+ concentration decreases ten-fold, ATP concentration increases almost three-fold, the ATP/ADP ratio goes up four-fold, the NADPH concentration doubles, and Mg2+ concentration increases from 1 mM to 3 mM (Heldt 1979; Heineke et al., 1991). It has been suggested that these differences can account for the stimulation of ACCase in the light, and therefore FAS (Nikolau and Hawke, 1984). However, other studies suggest that changing metabolite concentrations may not be the sole method for light modulation of ACCase activity (Sauer and Heise, 1984; Sasaki et al., 1997). In 1984 Sauer and Heise reported results on the activity of ACCase before, during and after intact spinach chloroplasts were illuminated. They observed that even when assayed at the same pH, Mg2+ and ATP concentrations, lysates of light-incubated chloroplasts had approximately three-fold higher ACCase activity than lysates of dark-incubated chloroplasts. More recently, Sasaki et al. reported that reduced thioredoxin activates ACCase and is the link between light activation and FAS (1997) 21 While much research has focused on the role of light and its direct or indirect effects, it would be surprising if there were not aspects of plant ACCase regulation independent of light. Fatty acids are a required precursor of membrane phospholipids, signaling pathways, carbon and energy storage. These precursors are necessary even in the absence of light. Additionally, in bacterial systems, fatty acid synthesis is tightly coupled to phospholipid synthesis (Cronan and Rock, 1996). Thus, non-light factors such as feedback regulation and source or sink control may also be involved in FAS regulation (Ohlrogge and Jaworski, 1997) Although over the past several years increasingly strong evidence has accumulated that ACCase plays a major role in the rate determination of leaf FAS, the factors which determine ACCase activity are still unclear. There is now a more complete model of ACCase structure and more detailed knowledge of pool sizes for FAS metabolites, more detailed knowledge of the metabolite concentrations of the chloroplast stroma and more detailed knowledge of the changes which occur upon illumination. Because additional information on ACCase regulation should provide better understanding of how plant cells regulate their fatty acid production, the objectives of this study were to re- evaluate, verify and further study factors affecting spinach chloroplast ACCase activity. 22 Materials and Methods Intact chloroplasts were isolated using a Percoll cushion according to the method of Roughan (1987) and suspended in incubation buffer (330 mM sorbitol, 25 mM Hepes/KOH pH 8.0, 2.0 mM NasEDTA, 1.0 mM MgC12, 1.0 mM MnClz, 0.5 mM K2HPO4). Percoll gradient-purified chloroplasts for control experiments were isolated according to Perry et al. (1991) using homogenization buffer from Roughan (1987) instead of grinding buffer and incubation buffer for washes and suspension instead of import buffer. All suspensions were adjusted to 0.4 mg chlorophyll/ml before use. Chlorophyll (chl) was assayed in methanol (Wintermans and De Mots, 1965). ”Light“ refers '2 * s'1 for 5 to 10 minutes. "Dark" to an incubation with at least 150 uE "' m incubations were done in foil-covered tubes. “Non-illuminated” chloroplasts were kept in room light, which does not stimulate FAS. All enzyme assays were done in room light. ACCase activity assay A ACCase activity was measured as the acetyl-CoA dependent fixation of H14C03' into acid-stable, heat-stable products (Sauer and Heise, 1984). The final concentration of assay components was 100 mM Tricine/KOH (pH 8.5), 0.5 mM acetyl-CoA, 1 mM ATP, 2 mM MgClz, 30 mM KCl, 10 mM NaHCO; and 0.2% Triton X—100. Chl content, specific activity of H14C03' and assay volume varied 23 and are described in figure legends. Assays were started by adding chloroplast suspension to assay components and stopped by bringing the reaction to 1N HCl. Samples were dried at 65°C, and the acid-stable radioactivity was counted by liquid scintillation. Control samples without acetyl-CoA were included in each series of assays and enzyme activity was calculated after subtracting these control results. The Triton X-100 and low osmoticum in the assay buffer caused immediate lysis of chloroplasts after their addition to the assay mix. In some experiments chloroplasts were lysed separately from ACCase assay by dilution into lysis buffer (125 mM Tricine/KOH, pH 8.5, 0.33% Triton X-100). Usually additions to the lysis buffer were made. These additions are indicated in figure legends. Assays were then started by the addition of lysed chloroplast suspension to assay mix which contained the remaining components to bring the final assay to the standard assay conditions detailed above. AC Case activity assay B In a different set of experiments the products of H”C03' fixation were separated via TLC. Final assay concentrations were the same as in Assay A. Reactions were started by the addition of chloroplast suspension and stopped by the addition of 1/6 volume of 15% trichloroacetic acid. Thioesters were hydrolyzed with the addition of 1/ 10 volume of 2N NaOH. Samples were spotted on a TLC plate (Whatman, K6 or K6F Silica Gel, 60 A), then developed in 95% ethanol: 24 25% ammonium hydroxide: water (78: 12.5: 9.5, v/v/v) for 2 to 2.5 hours (Copius-Peereboom, 1969) in a saturated chamber. TLC plates were imaged and radioactivity determined by a Instant Imager (Packard Instrument Company, Meriden, CT). Relative ACCase activity was determined using the net counts of bands which co-migrated with a malonate standard. Removal of acetyI-CoA from activated chloroplast lysates Micro Bio-Spin® P-6 (Bio-Rad) columns were equilibrated with column buffer (125 mM Tricine/KOH, pH 8.5, 47 mM KCl, 1.6 mM ATP, 3 mM MgClz, 0.33% Triton X-100). A maximum of 75 11.1 of activated lysate was added to the column, columns were spun at approximately 1500 x g, eluant recovered and, if necessary, pooled. In a control experiment, 0.5 mM radiolabeled acetyl-CoA (1 mCi/mmol) was combined with chloroplast lysate, added to the column, and the column spun as above. The column was washed thrice with 250 111 column buffer. Eluants and washes were counted for radioactivity via scintillation counting. Less than five percent of the radiolabeled acetyl-CoA was recovered in the eluant from the first spin. Thus, this method was able to remove almost all acetyl-CoA from chloroplast lysates to perform deactivation experiments. 25 AC Case assays in different metabolic environments Non-illuminated chloroplasts were assayed using assay A for 10 minutes using the assay conditions detailed above, or substrate concentrations and pH were adjusted to mimic the light or dark stromal environment (Winter et al., 1994). ADP, NADPH, and NADP” were also added. Final concentrations of altered assay components are given in Table 2.1. Results We chose to investigate regulation of ACCase in chloroplast lysates rather than a partially purified ACCase preparation for several reasons. First, we and others (Shorrosh et al, 1996; Sasaki et al. 1993) have observed membrane association of ACCase which may influence regulation of the enzyme and would be lost during enzyme preparation. Second, some changes in ACCase activity occur rapidly (e.g. Sauer and Heise, 1984) and the time required to prepare partially purified ACCase preparations may obscure these changes. Third, if unknown regulatory factors or other enzymes are required for interactions with ACCase, partial purification may eliminate them. Thus, a crude chloroplast extract may more accurately mimic an in vivo system than a partially purified system, particularly for studying rapid changes in ACCase regulation. 26 Table 2.1. Composition of light and dark stromal environments and resultant ACCase activity. light-like dark-like light pH / dark pH / environment environment other dark other light conditions conditions pH 8.0 7.0 8.0 7 .0 ATP 0.68 mM 0.24 mM 0.24 mM 0.68 mM ADP 0.29 mM 0.41 mM 0.41 mM 0.29 mM NaHCO; 0.74 mM 0.67 mM 0.74 mM 0.67 mM (C02) (10 uM) (84 uM) (10 11M) (84 11M) NADPH 0.11 mM 0.05 mM 0.05 mM 0.11 mM NADP+ 0.2 mM 0.2 mM 0.2 mM 0.2 mM MgC12 3 mM 1 mM 1 mM 3 mM activity 0.20 t 4% 0.02 :t 16% 0.20 i 12% 0.08 :t 7% Twenty ul of non-illuminated chloroplast suspension was assayed with the concentrations of Hi“, ATP, HC03', and Mg2+ altered from the standard assay. ADP, NADPH and NADP+ were also added. Final assay volume was 100 pl, specific activity of HMCOg' was 20 mCi/mmol. The concentration of C02 and HCO3’ was calculated based on pH and a total of 0.75 mM HI‘C03° in the assay mix; the actual substrate for ACCase is HC03’ (Cronan and Rock, 1996). Activity was determined by the average of eight separate assays using two different chloroplast suspensions. ACCase activity in standard assay conditions was 0.8 :l: 10% nmole CO2 fixed "‘ hour'l * ug chl'l. 27 We used cushion-purified chloroplasts in order to minimize extraction time and chloroplast handling. Although unlikely, multifunctional ACCase could be present as a cytosolic contaminant or a plastidial isoform (Schulte et al., 1997). Multifunctional ACCase has the ability to carboxylate propionyl-CoA to create methyl malonyl-CoA, whereas the multisubunit ACCase does not (Dehaye et al., 1994). To check the purity of cushion-purified chloroplasts and confirm that any multifunctional ACCase present did not contribute to our results, we used ACCase assay B, substituting propionyl-CoA for acetyl-CoA to check for propionyl-CoA carboxylase activity. In this experiment, all products of the assay remained at the origin of the TLC plate. This strongly suggests that all ACCase activity described in the experiments below is due to the activity of the multisubunit ACCase. Analysis of products of A C Case assay by TLC The assay system most frequently used for ACCase measures the acetyl-CoA dependent fixation of 14C02 into acid stable materials (e.g. Alban et al., 1994; Sauer and Heise, 1984; Sasaki et al., 1997). For assays with crude chloroplast extracts lasting several minutes or with purified enzyme, fixation of H‘4C03' and MC02 by controls in the absence of acetyl-CoA was minor (less than 10-30%) compared to the fixation in the presence of acetyl-CoA. However, in order to study ACCase light activation, previous studies (Sauer and Heise, 1984) and those described here require assays of less than one minute duration. Under these short assay times, fixation of H”C03' into products other than malonyl-CoA was 28 sflbfiantial, typically 50 to 90% of the total counts fixed. This background presumably reflects fixation of 1“CO2 by ribulose 1,5-bisphosphate carboxylase using endogenous pools of substrates. To establish that the standard bicarbonate fixation assay provides a true measure of ACCase activity, we analyzed the acid stable products by silica gel TLC. After stopping the reaction, acyl-CoA’s were hydrolyzed with NaOH, yielding radiolabeled malonate from the malonyl-CoA produced by ACCase. When the assay products were separated by silica gel TLC, three acid stable radiolabeled products were seen (Figure 2.1). A band with Rf = 0.34 co-migrated with a malonate standard. Its formation was acetyl-CoA, ATP, and time dependent. Occasionally, a minor upper band was seen which co-migrated with acetate. The remaining radioactive products remained at the origin. Their formation was acetyl-CoA independent but ATP dependent. The TLC assay separated the malonate produced by the ACCase assay from other acid-stable H’4C03 fixation products and thereby ensured that products which did not co-migrate with the malonate band were not included in the assay results. Assay A, on the other hand, uses controls without acetyl-CoA to differentiate the products fixed by ACCase from those fixed by other enzymes and for short assay times background can be substantial as discussed above. Our results using assay B which are described below confirmed assay A results from (Sauer and Heise, 1984). Thus the results provided by the standard assay system, assay A, were 29 malonate origin light dark light dark Figure 2.1. Representative autoradiogram of acid-stable products separated on silica gel. After hydrolysis of the CoA's from the products by altering the pH to 12, samples were spotted and run in an ethanol/ammonium hydroxide solvent system. Two replications are shown on this plate, light-incubated (lanes 1, 3) or dark-incubated (lanes 2, 4) lysed chloroplasts were assayed immediately after lysis. Assay B was used. Final assay volume was 60 1.11 with 3.2 ug chl, specific activity of H”C03' was 20 mCi/mmol. 30 tenable; the vast majority of product fixed in the presence of acetyl-CoA was malonyl-GOA and the high background could be adequately assessed by including samples without acetyl-CoA. Because analysis of the products by TLC and imaging the plates was slow, we preferentially used assay A for most experiments. AC Case activity is two-fold higher in lysates of a light-incubated chloroplast suspension than lysates of a dark-incubated chloroplast suspension ACCase activity was assayed by rapidly diluting an intact chloroplast suspension into Triton X-100 lysis/assay buffer. The TLC assay B was used in these experiments to distinguish products formed by ACCase from the high background l“C fixation. When assayed for 30 seconds immediately after lysis, the ACCase activity in lysates was two-fold higher from light-incubated than dark-incubated chloroplasts. The average light/dark ratio was 1.9 (SE of the ratios = 0.1, SD of the ratios = 1.1, n = 32). Light- or dark-incubated chloroplasts which had been lysed for five minutes before measuring activity had no significant difference in activity (average light/dark ratio = 0.9, SE = 0.1, SD = 0.2, n = 8). Relative ACCase activity in lysed, dark-incubated chloroplasts increased over the five minute lysis period, eliminating the light/dark difference seen in lysed and immediately assayed chloroplasts (data not shown). Therefore, the influence on ACCase activity appears to be transient and easily reversible. These results confirm and extend those of Sauer and Heise (Sauer and Heise, 31 1 (9'1)... . 1984). where a three-fold difference between light and dark activity of ACCase was reported. ACCase activity increases after pre-incubation of chloroplast lysates with CoA, acetyl-CoA and short-chain acyl-CoA ’s In one minute assays of rapidly lysed chloroplasts ACCase activity was 0.4 nmole CO2 fixed * hour'1 *ug chl'l (n=31). However, when assayed for 10 minutes under the same conditions, ACCase activity was 1.5 nmole CO2 fixed * hour'l *ug chl'l (n=22), or approximately four-fold higher. This result suggested a time dependent activation of ACCase may occur in the presence of assay components. Indeed, as shown in Figure 2.2A, ACCase activity of spinach chloroplast lysates increased over time, perhaps in response to components in the assay. In order to examine this activation, chloroplasts were lysed and pre- incubated in the presence of selected substrates. Figure 2.2B graphs the effect of adding acetyl-CoA, MgATP, or HCOa' to the lysis buffer before assaying ACCase activity for one minute using assay A. When chloroplasts were lysed in the presence of 30 to 833 11M acetyl-CoA before starting the ACCase assay, activity was increased by four-fold within eight minutes. Maximal acetyl-CoA activation required MgATP. Other substrates of the reaction, MgATP and HCOg', did not significantly activate spinach ACCase upon pre-incubation with the enzyme. Acetyl-CoA activation occurred in suspensions regardless if they were kept in room light, foil-wrapped, or illuminated. There is considerable variability in these assays. Hence, we believe that higher concentrations of acetyl-CoA 32 1.2 1— 1.0 ‘~ 0.6 ~~ ACCase activity 0.4 ~- i l l r 0.0 t 0.0 2.0 4.0 6.0 8.0 10.0 assay time (minutes) ~49— —1>— .4 Figure 2.2A. Specific activity of ACCase vs. total assay time. ACCase was assayed for varying times by adding 275 pl non-illuminated chloroplast suspension to 1100 [.11 assay buffer, then stopping 250 111 aliquots after 0.5, 1, 2, 4, and 8 minutes with 1 vol 2N HCl. Data are the average of three separate assays using the same chloroplast suspension, activity is given in nmole CO2 fixed * hour'l * ug chl". Specific activity of Hl‘CO3’ was 0.1 mCi/mmol. 33 .—l A 1 fl he N u—n O .o 00 .L .0 Ox 1 ACCase activity .9 .6. _o N l .0 o , minutes after lysis Figure 2.2B. ACCase assay of chloroplast suspension lysed for various times. Chloroplast suspension was lysed by five-fold dilution, then lysates were incubated for various times before ACCase assay. After the specified time, a 112.5 ul aliquot of lysed suspension containing 9 ug chl was added to assay mix. All assays were one minute. Lysis buffer (43-) had additions of 3.33 mM MgC12, 1.67 mM ATP, and 50 mM KCl ("B-- ); 10 mM NaHCO; and 50 mM KCl ("A"); 100 11M acetyl-CoA and 50 mM KCl {-B—); 833 uM acetyl-CoA, 3.33 mM MgCl2, 1.67 mM ATP and 50 mM KCl (+); 100 11M acetyl-CoA, 3.33 mM MgCl2, 1.67 mM ATP and 50 mM KCl (-I—); 30 11M acetyl-CoA, 3.33 mM MgCl2, 1.67 mM ATP and 50 mM KCl (+); 10 1.1M acetyl-CoA, 3.33 mM MgCl2, 1.67 mM ATP and 50 mM KCl (+). Data are averages of at least 3 replications. Final volume was 150 pl, specific activity of H‘4C03’ was 10-15 mCi/mmol. Activity is given in nmoles (:0, fixed * hour" * 11g chl". (e. g., 833 uM) most likely do not give significantly higher levels of activation. To determine if acetyl-CoA activation was reversible, acetyl-CoA was removed from the activated lysate by spinning through a gel filtration column. At least 95% of the acetyl-CoA was removed by this procedure. Immediately after the spin-column, ACCase activity was similar, or perhaps slightly higher, than the ACCase activity just prior to the spin column. However, 16 minutes after removal of acetyl-CoA, activity was reduced to about half of the maximal activation. Readdition of 200 to 400 uM acetyl-CoA to the same lysate caused ACCase activity to increase to about two-thirds of the maximal activation (Figure 2.3). Thus, the influence of acetyl-CoA is at least partially reversible. The activation described above was not specific to acetyl-CoA; other short chain acyl-CoA's and CoA also activated ACCase. Table 2.2 shows that over a period of 8 minutes, CoA, malonyl-CoA, propionyl-CoA and butyryl-CoA also activated ACCase three- to five-fold above initial activity. These activation experiments were done in a lysed chloroplast system, thus stromal enzymes were present. Since acetyl-CoA synthetase (ACS) is highly active in chloroplast lysates (Roughan and Ohlrogge, 1994; 6012 and Lichtenthaler, 1992), CoA added to the lysate could be converted to acetyl-CoA by ACS. To test if the activation of ACCase in presence of CoA was occurring due to CoA conversion to acetyl- CoA, ethyl adenylate, an ACS inhibitor (Golz and Lichtenthaler, 1992), was added to the lysis buffer. Table 2.2 shows that both acetyl-CoA and CoA 35 100% ‘1 90% '4 .233 80% ~ § 70% - '3 60% . 4 i: g 50% 1 a: 40% - 3 30°/ - Spin 2 0 column Readdition of 20% ‘ acetyl-CoA 10% - 0% r I t I r 1 l I 1 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 minutes after lysis Figure 2.3. Activity of ACCase after removal of acetyl-CoA. Non-illuminated chloroplast suspension was lysed by five-fold dilution. Lysis buffer additions were 100 uM acetyl-CoA, 50 mM KCl, 3.33 mM MgCl2, and 1.67 mM ATP. 40 ul aliquots of lysate were assayed for 1 minute at the times indicated. After 8 minutes acetyl-CoA was removed via a spin column prepared as described in Materials and Methods. After spinning, aliquots of elutant were assayed for 1 minute at the times indicated. After 25 minutes, acetyl-CoA was added to the elutant and aliquots were assayed. ACCase activity was normalized using the activity of a 10 minute assay done concurrently as 100% activity. N = 3 for times 0.5, 8, 13, 19, 25, and 34.5 minutes. N=1 for times 6.5 and 42.5 minutes. Assay A was used. Final assay volume was 50 111 (with 3.2 1.1g chl), specific activity of HI‘CO3' was 10 mCi/mmol. 36 Table 2.2. Effect of ethyl adenylate, CoA, malonly-CoA, and short-chain acyl- CoA’s on ACCase activation and activity. Additions to Lysis One minute assay Standard Buffer assay 0.5 minute 8 minute fold lysis lysis activation 100 11M acetyl-CoA 0.10 i 0.02 0.42 :1: 0.11 4 1.26 :t 0.04 100 uM acetyl-CoA, 50 uM ethyl 0.17 a; 0.15 0.33 i 0.22 2 1.17 :1: 0.05 adenylate not 100 11M COA 0.07 :t 0.02 0.38 :t 0.11 5 determined 100 1.1M CoA, 50 uM ethyl 0.06 :t 0.02 0.26 a: 0.15 4 1.14 :1: 0.04 adenylate 100 11M malonyl-CoA 0.05 :t 0.04 0.15 x 0.03 3 0.61 :1 0.07 100 11M propionyl-CoA 0.06 1 0.05 0.22 i 0.03 4 0.96 a: 0.05 100 uM butyryl-CoA 0.09 :t 0.03 0.36 :1: 0.02 4 1.15 a; 0.07 One minute assays: Non-illuminated chloroplast suspension was lysed by five- fold dilution and incubated for 30 seconds or 8 minutes before assay. Lysis buffer contained 1.67 mM ATP, 3.33 mM MgCl2, and 50 mM KCl as well as additions listed above. After the indicated time, 112.5 I.“ of lysed chloroplast suspension (equivalent to 9 ug chl) was added to 37.5 ul assay mix. Specific activity of H‘4C03' was 10 mCi/mmol. Assays were analyzed as in assay A. Standard assays: 22.5 111 of non-illuminated chloroplast suspension was added to 127.5 ul assay mix and assayed for 8 or 16 minutes. Specific activity of H14C03' was 2 mCi/mmol, assays were analyzed as in assay A. Activity is given in nmole CO2 fixed * hour'l *ug chl". 37 1% activated ACCase in the presence of ethyl adenylate, thus any effect of CoA was not due to its conversion to acetyl-CoA by ACS from the chloroplast lysate. Table 2.2 also shows the effects of short-chain acyl-CoA’s on 8 and 16 minute ACCase activity. The presence of 100 uM malonyl-CoA depressed ACCase activity approximately 50%. 100 1.1M propionyl-CoA depressed activity about 25%. Ethyl adenylate with acetyl-CoA, ethyl-adenylate with CoA, or butyryl- CoA did not significantly affect 8 and 16 minute ACCase activity. ACCase activity of chloroplast lysates in the presence of DTT and thioredoxin There have been reports that reducing agents such as DTT and thioredoxin stimulate ACCase activity both in spinach chloroplasts (Sauer and Heise, 1984) and partially purified pea ACCase (Sasaki et al., 1997). Sauer and Heise observed a 50% increase in ACCase activity when chloroplasts were pre- incubated with 5 mM DTT before chloroplast lysis and ACCase assay. Additionally, in a partially purified ACCase preparation from pea, Sasaki et al. (1997) report almost a six-fold increase in ACCase activity when the preparation includes 2 mM DTT or is pre-incubated with a combination of 50 11M DTT and 16 11M E. coli thioredoxin. In agreement with these results, we have observed stimulation by DTT and DTT plus E. coli thioredoxin on a clarified pea lysate as well as a resuspended and desalted ammonium sulfate precipitate of the pea lysate (data not shown). However, with lysed spinach chloroplasts, DTT stimulated ACCase only under conditions when activity was low (Figure 2.4). 38 .—s 00 J .—a O\ y—a h -__.___.1.._._--1._~.' .__J..-_ 1.0“ 0.8 * ACCase Activity 0.6- 0.4~ 0.22 i i 00‘ - A B c D E 1.3 3.4 2.4 1.6 1.5 1.2 Treatment Fold - Activation Figure 2.4. Effect of DTT on ACCase activity (nmole CO2 fixed "‘ hour'l * 11g chl“ ). Treatments A through E: Chloroplast suspension was lysed by five-fold dilution into lysis buffer (with additions of 1.67 mM ATP, 3.33 mM MgCl2, 50 mM KCl) containing no (I) or 2.67 - 5 mM DTT (El). After 1 (A), 2.5 (B), 4 (C), 8 (D) or 16 (E) minute(s) ACCase activity was assayed for 1 minute. Assay A was used, the final volume was 100 to 150 ul (5 - 9 ug chl), specific activity of HI‘CO3' was 10 mCi/mmol. Data are average of n=6 trials. Treatment F: Chloroplast suspension was assayed using standard assay conditions in the absence (I) or presence (1:!) of 0.05 or 1.0 mM DTT. Assay A was used, final assay volume was 100 to 150 u] (5 - 9 ug chl), specific activity of Hl‘COg' was 0.5 mCi/mmol. Data are average of n=4 trials. 39 With 1 minute assays immediately after lysis using assay A, we saw a stimulation by DTT of about two-fold, however, with 15 minute assays, where ACCase activity is higher, DTT had no significant effect. DTT at 50 11M and reduced Spirulina thioredoxin (50 to 151 uM) had no significant effect on the ACCase activity of lysed spinach or pea chloroplasts (data not shown). AC Case activity in different biochemical environments The effects of pH, Mg”, ATP and ADP on the activity of chloroplast ACCase have been described (Sauer and Heise, 1984; Sasaki et al., 1997; Eastwell and Stumpf, 1983). ADP is a competitive inhibitor of ATP (Eastwell and Stumpf, 1983) and stromal ADP levels rise in the dark (Heineke et al., 1991). Although the effects of pH and Mg2+ together have been detailed (Sauer and Heise, 1984; Sasaki et al., 1997), ACCase activity in the expected environment of the chloroplast, either in the light or dark, with respect to ATP, ADP, pH, NADPH, NADP”, Mg” and CO2/I-IC03' simultaneously, has not been described. Table 2.1 shows the effects that a light-like or a dark-like environment had upon ACCase activity. When assayed under the conditions described, ACCase activity was about ten-fold higher in a light-like environment (0.2 nmole CO2 fixed * hour’l I“1.1g chl'l) than a dark-like one (0.02 nmole CO2 fixed * hour‘l “ 11g chl"). When ACCase activity was assayed at pH 7 with concentrations of metabolites expected in the light, activity was four-fold greater than activity in a wholly dark-like environment. But, when assayed at the pH of illuminated chloroplast stroma (pH 8), ACCase activity was the same, 0.2 nmole CO2 fixed * hour'l *ug 40 chl", regardless of other metabolite concentrations. Thus, most of the influence of light-like and dark-like environments can be attributed to pH. Discussion Light activation (or dark deactivation) of ACCase in a lysed chloroplast suspension When lysed and assayed in the same buffer, the ACCase activity of a chloroplast extract is two-fold higher if chloroplasts were pre-incubated in light. Based on a value of 15 ug chl/u] stroma (Winter et al., 1994) and the final concentration of chl in the assay of 0.08 ug chl/111, lysis into assay buffer dilutes the stromal contents more than 100-fold, therefore, any metabolite differences of the illuminated vs. dark chloroplast suspensions would be overshadowed by the conditions in the assay buffer. Therefore, this difference in ACCase activity is likely to be either an enzyme activation connected to light, or a deactivation of the enzyme in the absence of light. Interestingly, when the chloroplasts are lysed for five minutes prior to assay, the ACCase activity of the dark-incubated chloroplasts increases relative to that of a short lysis period. Thus the regulation is most likely a reversible dark inactivation rather than a light activation. These data indicate that light regulation of ACCase is not solely a response to changes in ATP, pH, Mg”, etc., but that other factors are likely involved in establishing a lower activity of ACCase in dark chloroplasts. Because fatty acid synthesis must reSpond to a wide variety of cellular demands, we expect that metabolic feedback 41 controls on the enzyme are crucial to coordinating the production of fatty acids in the plastid with their use at the endoplasmic reticulum or other locations. Indeed, we have observed that exogenous fatty acids feedback on the rate of fatty acid synthesis in tobacco suspension cultures and that the site of this control is ACCase (Shintani and Ohlrogge, 1995). Thus, one possible explanation for the lower rate of ACCase in dark chloroplasts is that ACCase is inhibited by some feedback mechanism related to the lower demand for fatty acids in the dark. AC Case activity after pre-incubation of chloroplast lysates with substrates We observed an increasing activity of ACCase over assay time (Figure 2.2A) which could be attributed to the activating presence of acetyl-CoA in the assay buffer. As shown in Figure 2.28, pre-incubation with 30 to 833 11M acetyl-CoA increases ACCase activity by about four-fold. At 10 11M acetyl-CoA, two-fold activation is seen. The other substrates of the reaction, MgATP and HC03' do not appear to activate the spinach chloroplast enzyme. Laing and Roughan (1982) reported activation of ACCase by CoA, but did not investigate acetyl-CoA activation. We saw no significant difference between CoA, acetyl-CoA and other short chain acyl-CoA's when used for activation, and therefore these activators most likely use the same mechanism. The addition of ethyl adenylate at 12.5-fold its k, effectively blocked ACS activity, hence, any activation by CoA is due to the effect of CoA, not the effect of acetyl-CoA. 42 Removal of acetyl-CoA by gel filtration decreased activity, and furthermore, ACCase was partially reactivated by the readdition of acetyl-CoA. Since the column buffer contained all ACCase assay components except acetyl-CoA and HCOs', only acetyl-CoA, HCOg' and small molecules from the diluted stroma were removed by the spin column. If there were small molecules from the stroma which were removed from the lysate by the spin column, and these molecules were essential for acetyl-CoA activation of ACCase, readdition of acetyl-CoA would not have reactivated ACCase. Since readdition of acetyl-CoA at least partially reactivated ACCase, any small stromal molecules which may have remained on the column were not required for reactivation. Therefore, acetyl- CoA activation appears to be easily reversible, although full activation may require additional components. As shown in Figure 2.2B, ACCase is activated at low concentrations of acetyl- CoA, but does this occur in vivo? The concentration of total CoA's in chloroplasts is 10-20 11M, and the vast majority occurs as acetyl-CoA (Post- Beittenmiller et al., 1992b; Roughan 1997). Therefore, at the concentrations where we observe activation, it is likely that acetyl-CoA, rather than CoASH, or other acyl-CoA, is the form of CoA available to activate ACCase in vivo. Concentrations where we observe activation in vitro are similar to those that occur in vivo and thus acetyl-CoA activation of ACCase may be physiologically important. In order for this activation to be used in vivo, acetyl-CoA concentrations must change, at least locally to ACCase. Chloroplast acetyl-CoA 43 concentrations are very similar in light or dark leaves or chloroplasts. However, a transient decrease in acetyl-CoA occurs for two to three minutes after transfer of spinach chloroplasts from light to dark, but then stabilizes to light levels (Post-Beittenmiller et al., 1992b). The effect of acetyl-CoA upon ACCase is rapid: changes in the activity are apparent 30 seconds after lysis, which is in line with the rapid light-dependent increase in FAS (Browse et al., 1981). The calculated concentration of stromal acetyl-CoA (10-20 uM) is well below the K.,. of ACCase for acetyl-CoA (60-100 11M, Sauer and Heise, 1984; Mohan and Kekwick, 1980). Furthermore, at this concentration of acetyl-CoA and other stromal metabolites, calculated ACCase activity is several-fold lower than that needed to sustain in vivo rates of FAS (Table 2.1, Roughan 1997). The K.,. of spinach plastid ACCase for acetyl-CoA is about five times the calculated stromal concentration of acetyl-CoA. Based on this and other information, Roughan (1997) proposed that compartmentalization of FAS enzymes and substrates, including acetyl-CoA, may be occurring. Thus, the acetyl-CoA concentration actually presented to the ACCase active sites may be substantially higher than 10-20 uM. Although acetyl-CoA levels in seed plastids are not known, acetyl- ACP and malonyl-ACP levels are higher in seeds than in leaves (Post- Beittenmiller et al., 1991, Post-Beittenmiller et al., 1992a). If acetyl-CoA levels are likewise higher in seeds than leaves, this may suggest that acetyl-CoA activation of seed ACCase is associated with high rates of FAS during triacylglycerol accumulation in developing seeds. 44 Activation of AC Case by reductants Our results show that when ACCase activity is low, DTT has a stimulatory effect, increasing activity approximately two-fold. However, in cases where ACCase activity is higher, DTT by itself or in combination with Spirulina thioredoxin has no significant stimulatory effect either in spinach or pea chloroplast lysates. Although it has been reported that partially purified pea ACCase and spinach lysates are stimulated several fold by DTT and E. coli thioredoxin (Sauer and Heise, 1984; Sasaki et al., 1997) the activity of ACCase in those preparations was one-third to one-half (for pea, Sasaki et al., 1997) and one-tenth (for spinach, Sauer and Heise, 1984) the amount required to support observed in viva rates of FAS (Browse et al., 1981; Roughan, 1987). From our results, this is in the range where added DTT does have an effect. Possibly the preparation method of the partially purified pea ACCase altered its state to make the enzyme more sensitive to reductants than we observe in chloroplast lysates. Activation by reducing agents could be concurrent with acetyl-CoA activation - both operate when ACCase activity is relatively low. AC Case activity in different biochemical environments To a first approximation, the biochemical environment of the stroma is an important determinant of ACCase activity from lysed spinach chloroplasts, contributing a ten-fold difference in in vitro activity (Table 2.1). Furthermore, most of this effect appears to be due to the difference between light and dark stromal pH. But even in this model environment, in vitro ACCase activity is at 45 least five-fold lower than that required to sustain known rates of acetate incorporation by intact chloroplasts (1-2 umole * hour" "' mg chl", Roughan, 1987) or the in vitro rate of ACCase under optimized assay conditions. Thus, in the standard ACCase assay, where substrates are present well above their Km's and the reaction is optimized for ACCase, in vitro activity is similar to in vivo. This is not true for the model stromal environments. The expected concentration of HCOg' in the stroma is approximately one quarter of the spinach chloroplast K.,, for HCOg' (3.0 mM, Mohan and Kekwick, 1980). Thus, if ACCase activity in optimized assays is at vum and ACCase follows Michaelis-Menten kinetics, this lack of HCOg' would reduce ACCase activity in the light-like assay to about one fifth of van, activity. The difference in pH between standard assay conditions and illuminated stromal conditions further reduces ACCase activity by one half (Sauer and Heise, 1984). These changes alone can explain the difference in activity between a standard (optimized) ACCase assay and the model stromal environment. However, the low activity of ACCase under estimated stromal conditions suggests either the model environment is inaccurate or/and the ACCase enzyme in vitro is less active than in vivo. Roughan (1997) argues that the enzymes, substrates and cofactors for fatty acid synthesis must be compartmentalized to account for the types of differences discussed above. Our data support this conclusion, although at this time there is no clear picture of the nature of a FAS metabolon, nor can we rule out that ACCase must be in some “native” environment (membrane association?) to achieve maximum activity. Finally, in non-photosynthetic plastids those factors which are best understood to 46 influence FAS, such as light-dependence and pH changes, are unlikely to control ACCase, and other controls, such as substrate activation or feedback inhibition may play a dominant role in the regulation of this enzyme. Acknowledgments Linda Savage for assistance with the preparation of partially purified pea ACCase. Manfred Focke for the gift of ethyl adenylate. SCH was supported by the Michigan State University Biotechnology Training Program. This work was supported by a grant from the National Science Foundation (MCB 94-06466). We also acknowledge the Michigan Agricultural Experiment Station for its support of this research. 47 - M's-.- —.#*’-'ra-"" do”: ‘ . A" Literature Cited Alban, C., Baldet, P., and Douce, R. (1994). Localization and characterization of two structurally different forms of acetyl-CoA carboxylase in young pea leaves, of which one is sensitive to aryloxyphenoxypropionate herbicides. Biochem. J. 300, 557-565. Allred, J. B., and Reilly, K. E. (1997). Short-Term Regulation of acetyl-CoA Carboxylase in Tissues of Higher Animals. Prog. Lipid Res. 35, 371-385. Bettey, M., Ireland, R. J., and Smith, A. M. (1992). Purification and Characterization of Acetyl-CoA Carboxylase from Developing Pea Embryos. J. Plant Physiol. 140, 513-520. Browse, J., Roughan, P. G., and Slack, C. R. (1981). Light control of fatty acid synthesis and diurnal fluctuations of fatty acid composition in leaves. Biochem. J. 196, 347-354. Copius-Peereboom, J.W. (1969). Foodstuffs and Their Additives. In Thin-Layer Chromatography, a Laboratory Handbook. E. Stahl, ed. (New York: Springer-Verlag). pp. 630-636. Cronan, J.E. Jr., and Rock, C.O. Biosynthesis of Membrane Lipids. In E. Coli and Samonella: Cellular and Molecular Biology. F. C. Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger, Eds. (Washington, D. C.: ASM Press). pp. 612-636. Dehaye, L., Alban, C., Job, C., Douce, R., and Job, D. (1994). Kinetics of the two forms of acetyl-CoA carboxylase from Pisium sativum: correlation of the substate specificity of the enzymes and sensitivity towards aryloxyphenoxyproprinate herbicides. Eur J Biochem. 225, 1113-1123. Eastwell, K. C., and Stumpf, P. K. (1983). Regulation of Plant Acetyl-CoA Carboxylase by Adenylate Nucleotides. Plant Physiol. 72, 50-55. Golz, A., and Lichtenthaler, H. K. (1992). Inhibition of Plant Acetyl-CoA Synthetase by Alkyl-adenlyates. Z. Naturforsch 47c, 845-850. Heinke, D., Riens, B., Grosse, B., Bol’eriehter, P., Peter, U., Fliiffe, U., and Heldt, H. W. (1991). Redox Transfer across the Inner Chloroplast Envelope Membrane. Plant Physiol. 95, 1131-1137. 48 Heldt, H.W. (1979). Chapter 16: Light-Dependent Changes of Stromal H+ and Mg2+ Concentrations Controlling C02 Fixation. In Encyclopedia of Plant Physiology, New Series. M. Gibbs, and Latzko E., Eds. (Berlin: Springer-Verlag). pp. 202-207. Kannangara, C. G., and Stumpf, P. K. (1972). Fat Metabolism in Higher Plants: LIV. A Procaryotic Type Acetyl CoA Carboxylase in Spinach Chloroplasts. Arch. Biochem. Biophys. 152, 83-91. Laing, W., and Roughan P. (1982). Activation of Spinach Chloroplast acetyl-Coenzyme A carboxylase by coenzyme A. FEBS Lett. 144, 341-344. Mohan, S. B., and Kekwick, R. G. O. (1980). Acetyl-Coenzyme A Carboxylase from Avocado (Persea americana) Plastids and Spinach (Spinacia oleracea) Chloroplasts. Biochem. J. 187, 667-676. Nakamura, Y., and Yamada, M. (1975). Fatty acid synthesis by spinach chloroplasts 1. Property of fatty acid synthesis from acetate. Plant Cell Physiol. 16, 139-149. Nikolau, B. J., and Hawke, J. C. (1984). Purification and Characterization of Maize Leaf Acetyl-coenzyme A Carboxylase. Arch. Biochem. Biophys. 228, 86-96. Ohlrogge, J. B., and Jaworski, J. G. (1997). Regulation of Fatty Acid Synthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 48, 109-136. Page, R. A., Okada, S., and Harwood, J. L. (1994). Acetyl-CoA carboxylase exerts strong flux control over lipid synthesis in plants. Biochim. Biophys. Acta 1210, 369-372. Perry, S. E., Li, B., and Keegstra, K. (1991). In Vitro Reconstitution of Protein Transport into Chloroplasts. Methods Cell Biol. 34, 327-344. Picaud, A., Creach, A., and Trémolieres, A. (1991). Studies on the stimulation by light of fatty acid synthesis in Chlamydomonas reinhardtii whole cells. Plant Physiol. Biochem. 29, 441-448. Post-Beittenmiller, D., Jaworski, J. G., and Ohlrogge, J. B. (1991). In vivo Pools of Free and Acylated Acyl Carrier Proteins in Spinach. J. Biol. Chem. 266, 1858-1865. Post-Beittenmiller, D., Jaworski, J.G., and Ohlrogge, J.B. (1992a). Regulation of Lipid Synthesis in Castor Seeds: Analysis of the In Vivo Acyl-Acyl Carrier Protein Pools. In Seed Oils for the Future. S. L. MacKenzie, and D. C. Taylor, eds. (Champaign, Illinois: AOCS Press). pp. 44-51. 49 Post-Beittenmiller, D., Roughan, G., and Ohlrogge, J. B. (1992b). Regulation of Plant Fatty Acid Biosynthesis: Analysis of Acyl-Coenzyme A and Acyl-Acyl Carrier Protein Substrate Pools in Spinach and Pea Chloroplasts. Plant Physiol. 100, 923-930. Roughan, G. (1987). Long-Chain Fatty Acid Synthesis and Utilization by Isolated Chloroplasts. Methods Enzymol. 148, 327-337. Roughan, P. G. (1997). Stromal concentrations of coenzyme A and its esters are insufficient to account for rates of chloroplast fatty acid synthesis: evidence for substrate channelling within the chloroplast fatty acid synthase. Biochem. J. 327, 267-273. Roughan, P. G., and Ohlrogge, J. B. (1994). On the Assay of Acetyl-CoA Synthetase Activity in Chloroplasts and Leaf Extracts. Anal. Biochem. 216, 77-82. Salati, L.M., and Goodridge, A.G. (1996). Fatty acid synthesis in eukaryotes. In Biochemistry of Lipids, Lipoproteins and Membranes. D. E. Vance, and J. E. Vance, eds. (Amsterdam: Elsevier). pp. 101-127. Sasaki, Y., Hadamada, K., Suama, Y., Nagano, Y., Furusawa, I., and Matsnne, R. (1993). Chloroplast-encoded Protein as a Subunit of Acetyl-CoA Carboxylase in Pea Plant. J. Biol. Chem. 268, 25118-25123. Sasaki, Y., Konishi, T., and Nagano, Y. (1995). Update on Biochemistry: The compartmentation of Acetyl-Coenzyme A Carboxylase in Plants. Plant Physiol. 108, 445-449. Sasaki, Y., Kozaki, A., and Hatano, M. (1997). Link between light and fatty acid synthesis: thioredoxin-linked reductive activation of plastidic acetyl-Coenzyme A carboxylase. Proc. Natl. Acad. Sci. USA 94, 11096-11101. Sauer, A., and Heise, K. (1983). On the Light Dependence of Fatty Acid Synthesis in Spinach Chloroplasts. Plant Physiol. 73, 11-15. Sauer, A., and Heise, K. (1984). Regulation of Acetyl-Coenzyme A Carboxylase and Acetyl-Coenzyme A synthetase in Spinach Chloroplasts. Z. Naturforsch 39c, 268-275. Schulte, W., Tiipfer, R., Stracke, R., Schell, J., and Martini, N. (1997). Multi-functional acetyl-CoA carboxylase from Brassica napus is encoded by a multi-gene family: Indication from plastidic localization of at least one isoform. Proc. Natl. Acad. Sci. USA 94, 3465-3470. 50 Shintani, D. K., and Ohlrogge, J. B. (1995). Feedback inhibition of fatty acid synthesis in tobacco suspension cells. Plant J. 7, 577-587. Shorrosh, B. S., Savage, L. J., 8011, J., and Ohlrogge, J. B. (1996). The Pea Chloroplast Membrane-Associated Protein, IEP96, is a Subunit of Acetyl-CoA Carboxylase. Plant J. 10, 261-268. Stumpf, P. K., Bové, J. M., and Goffeau, A. (1963). Fat Metabolism in Higher Plants. XX: Relation of Fatty Acid Synthesis and Photophosphorylation in Lettuce Chloroplasts. Biochim. Biophys. Acta 70, 260-270. Winter, B., Robinson, D. G., and Heldt, H. W. (1994). Subcellular volumes and metabolite concentrations in spinach leaves. Planta 193, 530-535. Wintermans, J. F. G. M., and De Mots, A. (1965). Spectrophotometric Characteristics of Chlorophylls a and b and their pheophytins in Ethanol. Biochim. Biophys. Acta 109, 448-453. 51 ‘Pt Chapter 3 COMPUTER SIMULATIONS OF THE INITIAL REACTIONS OF FAS Abstract Biosynthesis of the major fatty acids produced by plastids, palmitic and oleic (16:0 and 1821159), requires 43 and 50 enzymatic steps respectively; thus de novo fatty acid synthesis (FAS) is a complex process. Through genetic engineering alterations of lipid composition have been achieved (e.g. increased laurate in rapeseed, addition of epoxy fatty acids in Arabidopsis) and oil quantity has been increased in Arabidopsis and Brassica. However, these accomplishments are balanced by numerous failures to achieve expected results and thus techniques for metabolically engineering a plant to produce a predetermined quantity and/or quality of fatty acids is still largely “hit-or-miss.” In addition to the conventional biochemical and genetic methods which are currently applied in metabolic engineering this chapter presents an additional approach: the use of modeling and computer simulations to control conceptual complexity and focus experiments in promising areas. A modeling program, Stella II, was used to create a computer simulation of the initial reactions of FAS. This simulation 1) integrates and combines the large amount of biochemical data available from studies of both enzyme kinetics and 52 metabolite pool sizes; 2) allows enzyme activities to be easily altered before or during a simulation; 3) mimics the behavior of metabolite pools under both high and low rates of FAS; 4) mimics the behavior of several transgenic over- and underexpression studies; and 5) qualitatively predicts most of the changes in metabolic pools seen in a light/dark transition, which consists of a six- to ten-fold drop in the rate of FAS. While the model is still primitive, it provides a proof-of-concept that models of the complex FAS pathway can be built, can mimic (or predict) changes in pool size and can mimic (or predict) the consequences of over- and underexpression of genes. This model predicts a series of results, a number of which correspond to known properties. In addition, it predicts that the set of enzymes which limit and co-limit the production of fatty acids are different depending upon 1) whether 18:1A9 or 16:0 fatty acid is considered and 2) if the system is in a state of high flux or low flux. The model also predicts that in order to substantially increase the amount of FA produced, ACCase, KAS III, and KASI will all have to be concurrently overexpressed. Any one enzyme or pair alone will not suffice. The model also predicts that increasing expression of genes in the latter part of the pathway will have very little effect on the overall rate of F AS. 53 Introduction How can the amount and type of fatty acids synthesized by a plant be altered? Almost forty years after the first papers on plant fatty acid synthesis were published, key aspects of plant fatty acid synthesis regulation are not understood. The majority of fatty acids are used in membranes and as storage products, although small amounts of specific fatty acids such as jasmonate and myristate play important cellular roles in signaling and anchoring proteins to membranes. Eighty to 85% of the FA in a non-seed plant cell is derived from 18:1A9, whereas about 15-20% is derived from 16:0 or its derivatives. Even in plants such as castor (Ricinus communis) or California Bay (Umbellularia californica) which are known for synthesizing large amounts of unusual fatty acids such as ricinolenic (18:1A9, 12-OH) or laurate (12:0) in seeds, 16:0 and 18:1 fatty acids are still the major components of membranes. Ideally, we would like to be able to genetically engineer specialty crop plants to create more or less oil and/or change the composition of that oil. To engineer fatty acid synthesis we need a clear understanding of the limits and regulators of fatty acid and triacylglycerol production. De nova FAS occurs in plastids and many factors affect the rate. In leaves one major regulator is light; FAS is six- to ten-fold higher in light-incubated chloroplasts than in dark- incubated chloroplasts (Browse et al., 1981). Light affects FAS in numerous ways: the pH of the stroma increases one unit, thus enzyme activity is affected, 54 and the concentrations of substrates and cofactors change. Light also indirectly generates the NAD(P)H used in the two reduction steps in each cycle of F AS. In chloroplasts the first committed and a rate-determining step of FAS is performed by acetyl-CoA carboxylase (ACCase). ACCase activity is at least two-fold higher in extracts of light-incubated chloroplasts than in extracts of dark-incubated chloroplasts, even when the assay conditions are identical (Chapter 2). Light also indirectly generates ATP which is required for the initial carboxylation of acetyl- CoA. The concentration of a cofactor, Mg2+ also increases over two-fold in the light from 1 mM to 3 mM (Heldt, 1979). Reduced thioredoxin can also stimulate ACCase activity, and Sasaki et al. have proposed this as the connection between light and FAS regulation (1997). The “on/off” mechanism of FAS is still not completely understood, although ACCase has been identified as a probable control point and some factors which influence its activity are known. Figure 3.1 shows the metabolic pathway of the plastidial reactions of FAS. Briefly, acetyl-CoA is carboxylated. Then the three carbon moiety of the resultant malonyl-CoA is transferred to acyl carrier protein (ACP) and this malonyl-ACP is used as a building block for fatty acid synthesis. The initial elongation reaction, catalyzed by KAS III, condenses malonyl-ACP and acetyl-CoA to form 3-ketobutyryl-ACP. The 3-ketobutyryl-ACP is reduced to 3-hydroxybutyryl-ACP, dehydrated to trans-AZ-butenoyl-ACP, then reduced again to butyryl-ACP. This butyryl-ACP is then elongated by another 2 carbons by condensation with another malonyl-ACP, and the cycle of reduce, dehydrate, reduce repeats. This cycle is 55 NADP+ 3-hydroxyacyl-ACP NADPH H20 3-hydroxyacyl-ACP 3-ketoacyl-ACP + H+ dehydrase reductase trans-A2 acyl enoyl-ACP 3-ketoacyl-ACP NAD(II)H C02 + H Enoyl-ACP reductase NAD(P)+ acyl-ACP (2 C longer) malonyl-ACP KAS III 16:0 ACP 18:0 ACP MCT CoA Acyl-ACP acetyl-CoA thioesterase ATP ACCase ACP ACP 1620 (palmitate) malonyl-CoA ADP, Pi Stearoyl-ACP Desaturase 18:1 ACP ACP Oleoyl-ACP thioesterase 18:1(oleate) Figure 3.1. Major plastidial reactions of FAS. Enzymatic steps are in boldface type. See text for further explanation. Figure afier Ohlrogge and Browse (1995). 56 repeated six times until the chain is 16 carbons in length. These elongations are performed by KAS I. The final steps of FAS are either hydrolysis of the 16:0 moiety from ACP to yield 16:0 fatty acids or a further elongation by KAS II to 18:0-ACP, then desaturation to 18:1A9-ACP followed by hydrolysis to yield 18:1A9 fatty acids. Acyl transferases can also move 16:0 and 18:1A9 from ACP onto glycerol for use in the chloroplast glycerolipids (Roughan and Slack, 1982; Somerville and Browse, 199]). The biochemistry of plant FAS is anything but simple. Of the 15 or so proteins involved in common fatty acid production, 10 have been partially purified or purified in some plants. Sixteen of the 18 or so genes have been cloned; of these, 10 have been used for under- or overexpression studies. However, the pathway is actually more complex than that shown in Figure 3.1. Enzymes specific to unusual fatty acid synthesis are still being discovered. A few examples: desaturase for petroselinic acid production (Cahoon et al., 1992), thioesterase for 12:0 production (Pollard et al., 1991), and KAS IV in Cuphea endosperm (Dehesh et al., 1998). Furthermore, there are specific co-factors which interact with these enzymes: ACP isoforms in Thunbergia alata and coriander which yield higher activity when used in in vitro desaturation assays (Suh et al., 1999) and likely a specific ferredoxin for the T. alata desaturase (Schultz and Ohlrogge, 1998). Where do the myriad of factors which regulate FAS act? Through analysis of metabolite pools in spinach leaf chloroplasts, Post-Beittenmiller et al. have argued 57 that ACCase must be rate limiting (1991, 1992). This work relies on the theory that it is possible to compare the pool sizes of metabolites in a pathway under states of high or normal flux and low flux. Under a state of low flux, the rate-limiting enzyme will be saturated and a backlog of some metabolite will be waiting to be processed, increasing the pool size just before the rate-limiting enzyme. Pool sizes downstream from the rate-limiting enzyme will be smaller than in the normal or high flux state (Rolleston, 1972). The studies of Post- Beittenmiller et al. used the difference in rates of FAS between the light and dark to alter flux (1991, 1992). Since Triton X-100 stimulates FAS in isolated chloroplasts, this method was also used to alter the rate of flux (Post- Beittenmiller et al., 1992). The studies showed that 1) in light-incubated chloroplasts, malonyl-CoA and malonyl-ACP were detectable, but were not detected in dark-incubated chloroplasts, 2) although the concentration of acetyl- CoA, the substrate for ACCase, was constant, the concentration of acetyl-ACP increased five-fold in dark-incubated chloroplasts. Likewise, when FAS is stimulated by Triton X-100, the acetyl-CoA pool size decreases slightly, but the malonyl-CoA pool size increases several-fold (Post-Beittenmiller et al., 1992). Taken together, these data point to ACCase as a major component of FAS rate- determination. Other enzymes in the pathway may also be regulatory. From results of an overexpression experiment of the E. cali KAS III homolog, fabH, in Brassica napus and E. cali, Verwoert et al. (1995) suggested that regulation of FAS is shared by a number of enzymes in the pathway. 58 In any case, a logical next step to alter oil yield in plants was to overexpress ACCase. Roesler et al. (1997) targeted the homomeric cytosolic ACCase to the plastid using a napin promoter. They saw a two-fold increase in ACCase activity in the seed, but only a 5% oil increase in plants during a growth chamber experiment (Ohlrogge, personal communication; Roesler et al., 1997). This pattern has held in field trials (Ohlrogge, personal communication). Why did total oil yield not increase more than 5%? Although there may be many reasons, one possible explanation is that some regulatory and/or controlling components remain unidentified. Antisense experiments in other primary pathways have also yielded unexpected results. Phosphofructokinase (PFK) is often described as a rate-limiting enzyme in glycolysis. PFK was overexpressed with no final change in respiration and only a slight increase in the intermediate metabolite, pyruvate (Burrell et al., 1994). Thus, if more than one enzyme is operating close to its maximal activity in a pathway, increasing just that one enzyme may not significantly increase flux and the final product. In tryptophan synthesis in yeast, genes were singly overexpressed ten- to fifty-fold (Niederberger et al., 1992). In most cases, overexpression had little effect on tryptophan accumulation or flux. The largest single-gene increase in flux was 30%, which was achieved when the enzyme involved had twenty-fold higher activity. Overexpression of four genes increased flux up to 2.5-fold. An eight-fold increase in flux was finally obtained: it took fifteen- to twenty-five-fold overexpression of the five genes in the pathway. 59 Non-intuitive regulation is also apparent in cyclic pathways. Reactions that are reversible in viva are generally not candidates for pathway control points (Fell, 1997). However, aldolase, an enzyme in the Calvin Cycle that catalyzes a reversible reaction, inhibits photosynthesis when antisensed (Haake et al., 1999). The activity of fructose-1,6-bisphosphatase is decreased in these plants under both high and low irradiance growth conditions (Haake et al., 1999). Underexpression of enzymes that are postulated to play a controlling role, e.g., RuBisCO, shows the “rate-limiting” enzymes to have 40-60% excess capacity under non-stress conditions. (Haake et al., 1999; Stitt, 1999) The above examples from both linear and cyclic pathways, as well as experience specific to fatty acid synthesis, argue that altering final product levels may require manipulation of many genes. In a short pathway, it may be possible to test single and multi-gene overexpression in a few experiments. This is not as feasible in a long pathway. With the advent of powerful and relatively easy-to-use computer simulation programs there is an additional approach that may be used. In this chapter, a computer simulation is described which incorporates the pool sizes of FAS metabolites, enzyme kinetic data (Km’s) and, when available, the relative activity of enzymes. This model was tested to see if it behaved in a “plant-like” manner under states of high and low flux. It can easily run multiple over- and underexpression studies. Ideally, the model will suggest a limited number of bench experiments for increasing oil production in plants. The computer simulation allows control of the conceptual complexity of FAS, and enforces consistent logic to scenarios tested. This chapter 60 describes 1) the simulation, 2) the test cases used to validate the model, 3) the model’s predictions, and 4) the implications of those predictions for both increasing oil production in plants and conceptualizing the regulation of plant fatty acid synthesis. Materials and Methods Materials Computer Program - Stella II Stella 11 (version 5.1, High Performance Systems, Inc., Hanover, NH) is a general-purpose modeling program available for both Windows and Macintosh platforms. Models in Stella consist of “stocks,” “flows,” “reservoirs,” “connectors” and “converters.” For the FAS model, “stocks” represent metabolite pools and are represented by rectangular boxes. Stocks move from one pool to another by means of a “flow.” The enzymes of FAS are the flows, and are represented by uni- or bi-directional pipes with faucets. Flows have user-defined rate equations which are based on Michaelis-Menten equations for this model . A “reservoir” is an infinite stock; these represent metabolite pools that are in constant, infinite supply in the model. The only true reservoir used in this model of FAS is the source of carbon, assumed to be acetate. A “converter” modifies the flow rate; these are gene expression factors in this implementation. A “connector” is a link between two other objects, either stocks or flows. Connectors may be 61 used to define the value of the flow in a model. Connections from stocks (metabolite pools) to flows (enzymes) allow pool size data to be used in the rate equations. Kinetic Data and Metabolite Pool Sizes The simulation was built using metabolite pool sizes in the light and kinetic data from enzymes. The pool sizes of FAS metabolites are shown in Table 3.1. This table also contains changes that occur during a light to dark transition when that data was available. The pool sizes have been determined in a variety of ways and general agreement of the values has been achieved over the last few years. Table 3.2 contains kinetic values for the enzymes of F AS gleaned from the literature. While kinetic data and pool size data were used as a base, they alone were not enough to create a realistic model of FAS. Therefore, other common-sense criteria were applied and are listed in Table 3.3. These criteria were used to tweak the system in order to get the system to behave in a biologically reasonable manner. Numerical Methods Stella II generates a system of differential equations, one for each flow into and out of a stock. This system of equations can be solved by standard mathematical methods which provide excellent approximations. Stella II provides three numerical methods for solving the system of rate equations generated: Runge- 62 Table 3.1. Pool sizes of FAS intermediate metabolites from spinach and pea chloroplasts. Values from references which were given in nmol per mg chl were adjusted according to Winter et al., 1994. Additional calculations have been performed on values in italics. Details of those calculations are given in the notes below the table. Nd = not detected. Metabolite 1.1M References and Notes Light Dark Acetyl-CoA 10-20 same as Post-Beittenmiller et al., 1992 1' 2 light CoA nd, nd Post-Beittenmiller et al., 1992 2' 3 < 0.8 Acetyl-ACP 0.5 — 1.0 2. 7 Post-Beittenmiller et al., 1991, 0.6 1992 “ 8 same as Post-Beittenmiller et al., 1991, Total ACP light 1992; 10-20 Kuo and Ohlrogge, 1984 free ACP 1.7 - 1.9 decreases Post—Beitenmiller et al., 1991, 4.6 1992 ‘ Malonyl-CoA 0.2 — 1.2 nd Post-Beittenmiller et al., 1992 nd, decreases Post-Beittenmiller et al., 1991, Malonyl-ACP 0.5—2.8 stays same 1992 4' 5 0.4 Williamson and Wakil, 1966 6 4:0-ACP 0.3 Post-Beittenmiller et al., 1991 ‘ 6:0-ACP 0.5 Post-Beittenmiller et al., 1991 ‘ 0.17 — 0.13 Roughan, 1998 8:0-ACP 0.2 Post-Beittenmiller et al., 1991 ‘ 0.33 - 0.40 Roughan, 1998 10:0-ACP 0.2 Post-Beittenmiller et al., 1991 4 0.24 — 0.31 Roughan, 1998 12:0-ACP 0.2 Post-Beittenmiller et al., 1991 1 0.26 - 0.26 Roughan, 1998 14:0-ACP 0.2 Post-Beittenmiller et al., 1991 4 0.24 - 0.23 Roughan, 1998 16:0-ACP 0.2 Post-Beittenmiller et al., 1991 " 1.23 - 0.73 Roughan, 1998 18:0-ACP 0.5 Post-Beittenmiller et al., 1991 T 0.64 — 0.54 Roughan, 1998 18:1-ACP 0.2 Post-Beittenmiller et al., 1991 ‘ 1.06 - 1.11 Roughan, 1998 Total esterified 5 Kopka, et al., 1995 ACP 63 1"- -4:'i~‘—* ”LI" “— Table 3.1, continued. 3-ketoacyl- nd Roughan, 1998 ACP < 0.005 trans A2 enoyl- nd Roughan, 1998 ACP < 0.005 3-hydroxyacyl- 0.023 - Roughan, 1998 ACP, 8 C 0.016 3-hydroxyacyl- 0.049 - Roughan, 1998 ACP, 10 C 0.040 3-hydroxyacyl- 0.070 — Roughan, 1998 ACP, 12 C 0.081 3-hydroxyacyl- 0.103 - Roughan, 1998 ACP, 14 C 0.137 3-hydroxyacyl- 0.089 — Roughan, 1998 ACP, 16 C 0.109 3-hydroxyacyl- 0.094 - Roughan, 1998 ACP, 18 C 0.091 , Notes: 1) Acetyl-CoA shows a transient decrease in the dark over 2-3 minutes after light/dark transition, but it returns to its original value after 6-8 minutes. 2) Essentially all of the CoA pool is found in acetyl-CoA. The pool size is slightly lower in light-incubated chloroplasts under optimal conditions for F AS than in freshly isolated chloroplasts 3) The detection limit was 0.8 uM 4) The reported values include the total acyl-ACP concentration and the proportion of specific acyl-ACP’s of this total determined by western blot analysis. These data were combined to yield the values in this table. 5) Detection methods differed. 6) Value calculated from the keq for the reaction and malonyl-CoA concentration. 64 Table 3.2. Kinetic data for FAS enzymes used in the Stella Model. Enzyme K.,, (uM) References ACCase acetyl-CoA 60-300 Bettey et al., 1992; Alban et al., 1994; Mohan and Kekwick, 1980, Sauer and Heise, 1984 ACT acetyl-CoA 8 Shimakata and Stumpf, 1983 MCT ACP 34 Post-Beitenmiller et al., 1989; malonyl-CoA 9-16 Shimakata and Stumpf, 1983; Guerra and Ohlrogge, 1986 KAS III acetyl-CoA 5-45 Gulliver and Slabas, 1994; Verwoert et malonyl-ACP 14 al., 1995; Clough et al., 1992 KAS I 6:0-ACP 7 Shimakata and Stumpf, 1982 10:0-ACP 5-10 14:0-ACP 9 KAS II 10:0-ACP 13 Shimakata and Stumpf, 1982 14:0-ACP 14 16:0-ACP 3 0TB 1610-ACP 4 D6rmann, et al., 1993 18:0-ACP 7 18:1A9-ACP 1 Desaturase 18:0-ACP 0.38 McKeon and Stumpf, 1982 Fatty Acid ACP 5 Post-Beittenmiller et al., 1989 Synthase 65 Table 3.3. Additional criteria used to build the model. Metabolite Pools and Enzyme Kinetics The model must reach a steady state in the light. The model must reach a steady state in the dark. Total CoA pool must remain constant. Total ACP pool must remain constant. Under wild type simulations, 15 to 20% of the FA produced should be 16:0. Steady state metabolite concentrations achieved in the model under wild- type (light) conditions should be approximately the same as measured in viva concentrations. Steady state metabolite concentrations achieved in the model under decreased ACCase activity (dark conditions) should be approximately the same as measured in viva concentrations. About 90% of the CoA should be found in free CoA under both light and dark conditions. About 60% of the ACP should be non-esterified in the light. The MCT reaction should be close to equilibrium. Light to Dark Transition The rate of FAS in the dark should be 1/10th to 1/6“I of the rate in the light. The acetyl-ACP concentration should climb four- to five-fold in the dark. Non-esterified ACP should decrease after transition and most of the ACP in the dark should be in acetyl-ACP. Malonyl-ACP levels should decrease in the dark. Malonyl-CoA levels should increase in the dark. Short and medium chain (C4 to C14) aCyl-ACP’s should remain constant. Palmitoyl-ACP, stearoyl-ACP, and oleoyl-ACP concentrations should be lower after the light to dark transition. 66 Kutta 2, Runge-Kutta 4, and Euler's Method. In theory, the simulation should predict the same behavior, independent of the numerical method used to solve the problem. Due to computational issues, this is not always the case. Detailed explanations of the numerical methods are in appendix A. Stella II provides both tabular and graphical views of stocks and flow rates as a function of time. Tabular data may be easily exported to other programs, e.g. Excel 97 (Microsoft Corporation, Redmond, WA) for data analysis. Methods Scope of the Model . An all-inclusive model of FAS would be extremely difficult to build. Furthermore, FAS is complex. ACP isoforms, probable differences in enzyme presence and specificity for unusual fatty acids would need to be incorporated. Given the limited knowledge of the enzyme kinetics and metabolic pools involved, it is likely that too many assumptions would be required to build a meaningful all-inclusive model at this time. Furthermore, a model that is too simplistic would likely not predict anything useful. This model is limited to the reactions of FAS that occur in the plastid and also to the production of 16:0 and 18:1A9 fatty acids. These two fatty acids or their derivatives comprise about 95% of the fatty acids found in plants, of which 15 to 20 % of the fatty acids produced are 16:0, the remainder is 18:1A9. 67 Implementation The biochemical detail of the model is shown in Figures 3.2A and 32B. In this model, all stocks (metabolite pools) are non-negative. In addition, altered gene expression and thus activity are represented by converters that can be manipulated by sliders, this model detail is shown in Figure 3.2C. All Stella objects are required to have unique names beginning with a letter, thus when an enzyme uses more than one metabolite pool, numbers and letters have been appended to form a unique name. For example, KAS III normally uses acetyl-CoA and malonyl-ACP to form 3-ketobutyryl-ACP, however, acetyl-ACP can also be used in place of acetyl-CoA with lower activity. Thus, the activities of KAS III are represented by KAS III a and KAS III b for the malonyl-ACP and acetyl-CoA condensation. Additionally, KAS III c and KAS III (1 represent the malonyl-ACP and acetyl-ACP condensation. Table 3.4 shows the components of the model and separates them by Stella object type. Table 3.5 shows the kinetic information and pool size data actually used in the model. Assumptions In order to create a workable model, simplifying assumptions almost always must be made. In the best possible case, these assumptions are not only correct, but have no bearing on the output and predictions of the model. In fact, often the most difficult part of creating a model is deciding which simplifications are valid (Bender, 1978). Modeling is an iterative process. As data become available and the model matures additional complexity is added. Many of the assumptions made 68 Figure 3.2A. Stella diagram of the model of the initial reactions of FAS. The reactions of F AS have been separated by function into four large classes: 1) reactions involving acetyl-CoA, 2) reactions involving the building block malonyl-CoA, 3) the elongation reactions which create acyl-ACP’s, and 4) the final reactions of the pathway. The details of flow into and out of CoA and ACP pools are shown in Figure 3.2B. Flow of the metabolites (square boxes) through the pathway is shown by double-lined arrows. Single lined arrows denote connections between components of the model, thus flow through ACCase is modulated by the concentration of acetyl-CoA, its substrate, and malonyl-CoA, its product. Flow rates are user-specifiedand in this model are all based on Michaelis-Menten kinetics. Expression factors (EF’s) exist for all enzymes and are used to modulate wild-type activity before or during simulations. Some icons appear in multiple places, such as Malonyl-ACP or KAS III EF. These are the exact same Stella component, but have been duplicated to increase diagram simplicity. More details of the icons, their use in the model, and their meaning are presented in Table 3.4. 69 mm = w<¥ a = 93. Alllll‘ .__ 2.. Q/ H 92 BYOB) _ m5. > cone 1 . a a an _ m5. ., no _ m .8; mm 0.2 m» .32 9: mm _ ms. ao< 5 _ 9:. m. m _ m5. MU \— os, z ao< um top up n _ m 200,000 1 i ' H '\ afl’ r; ,. 180,000 -‘ ’3" 4. - 1; m 160,000 ‘ Pd 1 ,5 ‘ ,l' 5 a 140,000 . z / r a; - I I ‘ .9, 120,000 - I / 5 .... o / A g 100,000 — . I x a a / I .8 80,000 « - I / a t: I I .8 60,000 7 g r, a . / 0 40,000 4 '4 I g A 20,000 J, - 0 A T ' Y T 1 0.0 0.2 0.4 0.6 0.8 1.0 B 10,... 8 '5 8,000 d a ‘5‘ 6,000 a 53 1,8 4,000 -' O E ; Z : 2,000 - o 0.0 C 4,000 -. g 3,500 1: °r v: .1 , "' '_ _ +537." "I" 2 3,000 ,r 2 ‘ _. g I x - g , , ,3; 2,500 2 - 5 , . z ‘9‘ f r. ’5’ ’ 9 2,0001i I , 2 l ‘ 7‘ I l I g 1,500 2, - - a l ’ 2 1,000 l; 1”, l [A g 500 —f ,’ _ 4’ 0 AF 1 . . - 0.0 0.2 0.4 0.6 0.8 1.0 87 expression levels. ACS is also co-limiting to some extent but this may be artifactual since the in viva source of acetyl-CoA for FAS is unknown. The remainder of the enzymes, ACT, MCT, KAS II, the desaturase, and the thioesterases are not limiting. Counter-intuitively, the limiting enzymes are different when the product of F AS is 16:0 free fatty acid rather than 18: 1A9. In this case, TE 16:0, KAS I and KAS III underexpression have the most effect on limiting flux toward 16:0. ACCase and ACS are co-limiting. ACT, MCT, the desaturase, and TE 18:1 are non-limiting. KAS II underexpression increases flux through the 16:0 free fatty acid pathway. ACCase, KAS I, and KAS II limit or co-limitflux through FAS in the dark FAS is six to ten-fold higher in light-incubated leaf discs than in dark-incubated leaf-discs (Browse et al., 1981), in addition ACCase activity is at least two-fold higher in lysed, light-incubated chloroplasts than in dark-incubated ones (Chapter 2). Incubation of chloroplasts in the dark is mimicked by lowering ACCase activity in the model to one-eighth its wild-type value. That is, we set the expression factor, ACCase EF, to 0.125; in essence, this turns off the lights in the model. The same set of simulations as those shown in Figure 3.4 were run. The results, where the lights are “off,” are shown in Figure 3.5. Figure 3.5 shows that when total carbon in fatty acid or 18:1A9 fatty acids are considered, the limiting enzyme is ACCase, and KAS I is co-limiting. However, 88 Figure 3.5 Flux through the FAS pathway vs. enzyme activity under states of low flux. Computer simulations were run with ACCase EF set to 0.125. Then each expression factor of each enzyme was set to 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8 0.9 and 1.0. The final pool sizes of FAS products, 16:0 and 1821159 are shown in theoretical nmoles. A) Total carbon theoretically incorporated into FAS products. KAS II, TE 16:0, ACS, TE 18:1, ACT, MCT, and the Desaturase have essentially no effect on carbon incorporated into total fatty acids. B) Flux into 18:1A9 products only. KAS 1, ACS, TE 18:1, ACT, MCT, and the Desaturase have essentially no effect on carbon incorporated into 18: 1A9 fatty acids. C) Flux into 16:0 products. ACS, ACT, MCT, and the Desaturase have essentially no effect on carbon incorporated into 16:0 fatty acids. Legend + 0.125X ACCase - 'A- ' KAS I —e— KAS II - 'l- " KAS III + TE 1620 -—A—- ACS - -O- - TE 18:1 - '8- ' ACT - “‘- ' MCT g + Desaturase l 89 Carbon in total fatty acids 3’ Nmoles 18:1 fatty acids w C Nmoles 16:0 fatty acids 60,000 50,000 40,000 30,000 20,000 10,000 2,500 - 2,000 j 1,500 ~ 1,000 a 500 a 1.-- --__1_ 0.0 0.2 0.4 0.6 0.8 1.0 2,000 - 1,800 4; 1,600 —3 1,400 4 1,200 - 1,000 — 800 ~ 600 - 400 4, 200 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 90 unlike the predictions in the light, KAS III becomes detrimental to total fatty acid production, whereas in the light KAS III is co-limiting. These different roles are upheld when total fatty acid synthesis is broken down into 18:1.A9 and 16:0 components. In the dark, ACCase and KAS II are limiting, and KAS III and the 16:0 thioesterase become detrimental. In the case of 16:0 production in the dark, ACCase and KAS I are limiting, the 16:0 thioesterase is co-limiting. KAS II is detrimental and KAS III borders on detrimental. These relationships are summarized in Table 3.7. Similar to FAS simulation in the light, a different set of enzymes limit and co-limit 18:1A9 and 16:0 production. However, these sets are different in the light and dark. That is, KAS II and, to some extent, ACCase, limit 18:1A9 production in the dark, but in the light, 18:1A9 production is limited by ACCase alone while KAS II is co-limiting. So, not only does the model question the assumption that 18: 1A9 and 16:0 production are regulated similarly but it also predicts that regulation may be controlled by different sets of enzymes under states of high flux (light) or low flux (dark). There is precedent for this: the regulatory role of RuBisCO varies with irradiance (reviewed in Stitt and Schultze, 1994). These possible complexities in regulation will need to be explored further if we want to engineer plants to produce fatty acids on demand. The model may not be correct, but the predictions are worth exploring further. Increase of Fatty Acid Synthesis by Enzyme Overexpression One major purpose in creating this computer simulation was to attempt to predict which enzymes, when overexpressed, would allow the most effective increase in 91 Table 3.7. Summary of limiting and co-limiting enzymes in the light and dark. Fatty Acid Enzyme Role Light Dark Limiting ACCase ACCase Total Co-limiting KAS I KAS I KAS III LACS) Detrimental none KAS III Limiting ACCase KAS II ACCase Co-limiting KAS I none 18:1A9 KAS II KAS III (ACS) Detrimental TE 16:0 TE 16:0 KAS III Limiting TE 16:0 ACCase KAS III KAS I Co-limiting KAS I TE 16:0 16:0 ACCase (ACS) Detrimental KAS II KAS II (KAS III) 92 «a. fatty acid synthesis. Since overexpression of medium chain thioesterases and desaturases have already been well-explored and shown to change the composition of seed oils (Knutzon et al., 1992; Voelker et al., 1992; and reviewed in Kinney, 1994), increased production of 18:1A9 was the focus of this study. The simulation’s predictions of overexpression of any one of the four enzymes which limit or co-limit 18:1A9 production are shown in Figure 3.6. It is clear from the graphs that the model does not predict a continual increase in yield when any single enzyme is overexpressed. After two-fold overexpression of ACCase, 18:1A9 production is only predicted to increase 50%. Flux does not increase substantially further when ACCase is overexpressed up to five-fold. Predictions of double-enzyme overexpression experiments are shown in Figure 3.7. Once again the model makes some sobering predictions: 18:1A9 can only be increased two-fold, despite several-fold simultaneous overexpression of two enzymes. The model predicts the highest 18:1A9 yield when KAS II/ACCase dual expression is used. KAS I/ACCase overexpression is only slightly lower. An increase in 16:0 is perhaps not as limited. KAS I/KAS III overexpression is predicted to yield a 2.5- fold increase in 16:0, the largest in the simulation. Since single and double overexpression experiments do not appear to significantly increase FAS, will triple-enzyme overexpression make a difference? The results of simulated overexpression of three of the four major limiting enzymes are shown in Figure 3.8. In this case, The largest increase in pathway flux to 18:1A9 is predicted if ACCase, KAS I and either KAS II or KAS III are overexpressed 93 2‘57 ..................................................................................................... A i < 2.0.. ..................................................................................................... u, . a ._ ................... . ............................. n ....... ,., ................. h .................. 8 lji n U U U £1[JACCase 5 . A AKASI .0~~ s - - u go 1 W oKASII '5 0.5 ............................................................................................ IKASIII 0.0.. .................................................................................................... B 25' """"""""""""""""""""""""""""""""""""""""""" 2.0: ooooooooooooo i oooooooooooooooooooooooooooooooooooooooooooooooooooo 3 1,5-» ------------------------- C: ----- 3 ----------- 3 ---------- QUACCase .E 3 C C floKASII glfl‘ ------ as ----- t. -------- a. fiAKASI '5 0_5.l.. . ........................................................ IKASIII 0.0- ------------------------------------------------------------------ C 3.0- ........................................................................................................ C 2 l .................................................................................. 5'. ~15, AKASI g) ...... \i: DACCase 6 ............................................. .KASIII *3 oKASII 0 l 2 3 4 5 Expression Factor Figure 3.6. The effect of altered expression levels of single enzymes in the model. The simulation was run with each enzyme expressed 0.05 to five-fold. The ratio of altered-expression production/wild-type expression production vs. EF is plotted. A) total carbon in fatty acids, B) carbon in 18:1A9, C) carbon in 16:0. 94 ‘-—-< A 20. ............. 1 ...................................................... < E 1.5 ............. , ............ ' c1 ACCase/KAS I g * ACCase/KAS 11 E10. .. IKAS I/KASIII '0 ' o KAS I/KAS 11 g” as o ACCase/KAS 111 .c: 3' U . A KAS II/KAS III 00* B 25 a .............................................................................................................. 2.0., ...................... y..............................................................;x. ................... .x* ACCase/KAS II 2': 5 S 8 fit: ACCase/KAS I E 1.54 ................................................... ,5 4130 KAS I/KAS II 0 OD 1.0- :;, ................................. g 1’57 5 [/2 - KAS I/KAS III 0.5+ ..... ( , ........ . ............................................................. p ................... . ACCase/KAS III s/ ‘ A KAS II/KAS 111 0.0 «L ............................................................................................................. C25 - ........................................................................................................... - KAS I/KAS III 20.. ..................... , ................................................................................ E 1.5. .................................................................................................... .5 104 :S”\ - 1-1 B 8 r3 3 0 ACCflSC/KAS III 9% . H... ‘ on :, WNHHHN ............................................................... "‘ D ACCase/KAS I g “ Ra‘ . U 0.5,... ......... , .................... \~n~;k A .................... o KAS I/KAS II - «\l‘ * ACCase/KAS II 0.0 i . . A KAS II/KAS 0 l 2 3 4 5 Expression Factor Figure 3.7 .Model predictions of the effect of double enzyme expression. The simulation was run with each pair of enzymes expressed by 0.05 to five- fold. The ratio of altered-expression production/wild-type production vs. EF is plotted. A) total carbon in fatty acids, B) carbon in 18:1A9, C) carbon in 16:0. 95 OOOOOOOOOO .......... --------------------------------------------------- . ACCase, KAS I, _KAS III e ACCase, KAS I, KAS 11 ...O KAS I, KAS II, KAS III 0.5 a» ---------------------------------------------- 1:1 ACCase, KAS II, 0.0 . ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, KAS III B 3.0 _, ........................................................................................................... . ACCase, KAS I, 25 a, ........................................................................................................... KAS III '.-: .................................................... ,- —-—-—0 ACCase. KAS 1. a 2.0 1 '2 KAS II go o KAS I, KAS II, .5 KAS III 05 1 ........................................................................................... D ACCase, KAS 119 0.0 -l .................... 1 ...................................................................................... KAS 111 C 3.0. ................................................................... 25 . ........................................................... l ACCase, KAS I, KAS III 20. ............................................................... 9 1.51 ................................................................ :9. o KAS I, KAS II, .5 1.021 ----- 0 KAS III D E” 0.5 + ............................ 1:1 ACCase, KAS II, D KAS III 0.0 i 1 T * eACCase, KAS I, 0 1 2 3 4 5 KAS II Expression Factor Figure 3.8. Predictions of the effect of the altered expression of three enzymes. The simulation was run with each triplet of enzymes expressed by the same amount, from 0.05- to five-fold. The ratio of altered-expression production/wild- type production vs. EF is plotted. A) total carbon in fatty acids, B) carbon in 18:1A9 , C) carbon in 16:0. 96 simultaneously. Figure 3.9 shows the effect of overexpressing all four of the enzymes simultaneously. In this case, the graph is almost linear indicating these enzymes in consort are limiting according to the scheme in Figure 3.3. Additionally, concurrent five-fold overexpression of ACCase, KAS I, KAS II, and KAS III yields almost a three-fold increase in flux to 18: 1A9. If this were actually true in viva, it is likely that transport or downstream processes would start to limit TAG production. Overexpression of the enzymes at the end of the pathway does not increase FAS One universal method proposed to increase metabolic methods is to overexpress all enzymes of the pathway (Kacser and Acerzena, 1993). Clearly at this time, this is not feasible for FAS as it would require at least 14 enzymes to be overexpressed. However, this highly theoretical treatment also predicts that enzymes near the end of a pathway have more control over the output of the pathway than those at the start. This is due to end products relieving feedback inhibition at the start of the pathway (Kacser and Acerzena, 1993). Therefore, enzymes at the end of the pathway were overexpressed in the Stella FAS model. Equal and concurrent overexpression of the two thioesterases, the desaturase, and KAS II are shown in Figure 3.10. It is clear that overexpression of these terminal enzymes does not significantly alter total carbon in fatty acid or the 18:1A9/16:0 ratio. 97 crow-":9;rt:l..s.ltln‘!\rcl\! r Change in total FA ea Changein 18:1 C 3.0 T ----------------------------------------------------------- 2.5 -* -------------------- — ....................................... .o. 2.0 *1- ----------------------------------------------------------- .‘2 .5 1,5 1 ----------------------------------------------------------- 8’11 3 1.0 . ................................................. U 0.5 er ----------- r ——————————————————————————————————————————————— 0.0 T a, f , 0 l 2 3 4 5 Expression Factor Figure 3.9. Concurrent altered expression of ACCase, KAS I, KAS II, and KAS III. The simulation was run with the enzymes expressed the same amount, from 0.05- to five-fold. The ratio of altered-expression production/wild-type expression production vs. EF is plotted. A) total carbon in fatty acids, B) 18:1A9 , C) 16:0. 98 1.11 _° 18:1 3’ total fatty acids 3: 1.0 ~ -------------------------------------- g 16:0 9. ‘7 0,9 1 . . . . 0 1 2 3 4 5 Fold Expression Figure 3.10. Overexpression of the enzymes at the end of FAS pathway. TE 16:0, TE 18:1, KAS II, and desaturase were all simultaneously overexpressed one- to five-fold. The amount of total, 16:0, and 18:1 fatty acids produced during overexpression were compared to the amount of total 16:0, and 18:1 fatty acids produced during wild-type expression. The ratio of overexpression production/wild-type production vs. level of expression is plotted. Only a slight increase in total fatty acids is seen. 99 Discussion In many modeling situations it is easy to forget that the model is not the in viva system. One purpose of any model is to allow us to understand the system under study more completely and to point out areas where further work is needed. This model of FAS and the computer simulations are no exception. The model presented above is a vast simplification of a living plant system, and therefore the model’s predictions may not hold in viva. However, the model raises some interesting questions. These questions are 1) is the accumulation of 16:0 regulated differently than the accumulation of 18:1A9? and 2) is FAS regulated in the same manner under states of high flux (light) and low flux (dark)? While these questions may not be novel questions, investigating them becomes more important because the model raises them. A short diversion into science philosophy will explain this. Progress in science depends on making worthwhile observations that either help confirm or (even better) invalidate conceptual models (Popper, 1953; Platt, 1964). So, when testing a conceptual model there are criteria which make a test, usually a bench experiment, worth doing. One criterion of a “good” test is that any prediction made from the conceptual model must be logically deducible (Giere, 1984). This is sometimes difficult to ensure with a simple model; however, with a complex model system, such as the FAS model presented here, it would be essentially impossible to integrate and check every assumption and connection without some sort of external control. The computer simulation provides this 100 consistent and complete check on deducibility. Therefore, the issues raised in the computer simulations are logical and deducible consequences of the model and not simply well-reasoned, though perhaps humanly flawed, opinion. Model validation relies heavily on in viva KAS III overexpression experiments The KAS III overexpression data are critical for model validation. This is due to the non-intuitive nature of the results - both in the model and in viva. Certainly the model’s predictions do not quantitatively match reality, however they d_ qualitatively. The E. coli KAS III homolog, fabH, was under and overexpressed in E. coli by Tsay et al. (1992) as well as Verwoert et al. (1995). Membranes were analyzed for 14:0, 16:0, 18:0 and 18:1 (cis-vaccenic) fatty acids. In the case of KAS III underexpression, 16:0 in E. cali membranes decreased by slightly over 12% while 18:1 (cis-vaccenic) increased by 18%. With overexpression of KAS III, membrane 14:0 increased lO-fold, 16:0 increased 24%; conversely 18:0 decreased 60% and 18:1 (cis-vaccenic) decreased 86% (Tsay et al., 1992). The authors did not comment on growth rate. Verwoert et al. (1995) observed the same pattern and also reported significantly slower growth of E. cali overexpressing fabH. They proposed the accumulation of 14:0 may be a stress response of E. coli and not a direct consequence of KAS III overexpression (Verwoert et al., 1995). However, if KAS III overexpression decreases fatty acid availability for cell membranes, signaling, and growth, then this would probably be stressful in viva. 101 The KAS III overexpression simulations used to validate the model suggest some additional experiments for existing overexpressing plants. In a wild-type plant and in the normal model simulation over half the ACP is non-esterified (ACP-SH). KAS III overexpression decreases the amount of ACP-SH to very low levels in the Stella model. KAS III overexpression in tobacco plants results in a two-fold increase in total ACP (Jaworski and Hinneburg-Wolf, 1998). Is the in viva increase enough to maintain wild-type ratios of acyl-ACP pools? Or is ACP expression at its maximal level? Given the model predictions of very low ACP-SH, it is more probable that ACP expression is as high as possible. Limiting and ca-limiting enzymes in the light Most of the predictions the model makes in Figures 3.4 and 3.5 are intuitive. Post-Beittenmiller et al. (1991, 1992) determined that ACCase is a limiting factor in total fatty acid synthesis in the light, and the simulation reflects this as well. The model predicts that decreasing the activity of the 16:0 thioesterase (TE) will increase the amount of 18:1A9 produced. This makes sense, since 16:0-ACP is a substrate for both the 16:0 thioesterase and KAS II reactions. Decreasing the 16:0 thioesterase activity increases the amount of 16:0-ACP available for KAS II elongation. Likewise, the predictions the simulation makes for the limiting and co-limiting steps of 16:0 production are logical. The final step catalyzed by the 16:0 TB is limiting. In the production of 16:0, the initial condensation, performed by KAS III, is also limiting. Here the bottleneck seems to be the initiation of fatty acid synthesis rather than the co-limiting reactions of malonyl-CoA synthesis and 102 intermediate elongations catalyzed by ACCase and KAS I respectively. In a manner similar to the production of 18:1A9, decreasing the activity of the competing KAS II reaction in the final branch point of 16:0 production increases the production of 16:0. One surprising prediction the model makes is apparent when the simulation’s predictions for the regulation of 18:1A9 and 16:0 production are compared. The production of 18:1A9 is limited by ACCase, whereas the production of 16:0 is limited by the thioesterase and KAS 111. Since ACCase becomes co-limiting when the production of 16:0 is considered and KAS III becomes co-limiting when the production of 18:1A9 is considered, it appears that ACCase and KAS III may switch regulatory roles when different products are examined. KAS I is co-limiting in the production of both 18:1A9 and 16:0. Thus, FAS may be regulated differently depending on whether the production of 16:0 or 18:1A9 is considered. Until now, 18:1A9 and 16:0 fatty acids have been combined when considering the regulation of FAS. The pathway of fatty acid synthesis is taught and studied as one pathway, in spite of different exit points. This assumption needs to be more fully examined. In many cases, oil composition has been altered by changing the expression level of the thioesterases and desaturases (reviewed in Kinney, 1994). For example, Knutzon et al. (1992) antisensed the 18:0 desaturase and increased the 18:0 level in seeds. Overexpression of an acyl-ACP thioesterase increased the amount of 103 16:0 and 18:0 (Kinney, 1994). In many of the cases where enzyme overexpression studies have been done, results for both 16:0 and 18:1A9 have been reported. Often, the model predicts only part of the results correctly (Table 3.6). For example, when KAS II is antisensed, 16:0 increases approximately three-fold and there is essentially no significant change in the amount of 18:1A9 (Kinney and Hitz, 1995). The model predicts the increase in 16:0, but predicts a 50% decrease in 18:1A9. Likewise, thioesterase overexpression causes a two-fold increase in 16:0 and no significant change in 18:1A9 (Kinney and Hitz, 1995). The model correctly predicts the increase in 16:0, but predicts a 25% decrease in 18:1A9. In both cases, the model predicts reduced levels of 18:1A9 which are not seen in viva. This may be due to strict requirements for 18:1A9 which control its production in viva. It is clear from the above discussion that there may be a very fine balance between FAS enzyme activities, not only for ACCase, KAS I, KAS II, and KAS III, but also for the set of terminal enzymes: the thioesterases and the desaturase. In addition, slight differences in enzyme substrate affinities may change the composition of the fatty acids produced and exported by the plastids. Alternatively, the enzymes of fatty acid synthesis may be flexible and subject to fine regulation. Both scenarios are speculative at this point in time. If varying substrate affinities are important for regulation, there are both positive and negative consequences for manipulation of seed oils. On the positive side, if small changes are needed, then perhaps the enzymes involved can be engineered for 104 n = .,5I. )‘.A-_,'.- 97'“ H. specificity. For example, Shankin and co-workers (Cahoon et al., 1997) have changed a A6 desaturase to a A9 desaturase and vice versa with five or fewer amino acid changes. On the negative side, a consequence of fine and precise control is that precise biochemical data for the enzymes will be required to model FAS on a quantitative level. Furthermore, the enzyme data must be valid in viva. Given this requirement, it is perhaps surprising that the model mimics reality in any way. Modeling, as a technique to understand systems, is widely practiced. Any abstraction or simplification of a problem is a model. Formalized mathematical models are also widely used (for examples, see Bender, 1978). The model of fatty acid synthesis presented in this chapter synthesizes data from three sources: biochemical data, successful and not-so-successful transgenic expression experiments, and a computer simulation package. The data presented by the model highlight areas where further work is indicated in order to engineer fatty acids, namely, more precise biochemistry. The model points out one area where reconceptualization of the FAS pathway may be in order: are there separate signals and mechanisms for 18: 1A9 and 16:0 production? Finally, the model yields a sobering prediction if increased total flux is the goal: overexpress at least three enzymes. While the predictions of the model may be incorrect, modeling is an iterative process. This model, with its predictions and questions, is a starting point for further investigations. 105 Literature Cited Alban, C., Baldet, P., and Dance, R. (1994). Localization and characterization of two structurally different forms of acetyl-CoA carboxylase in young pea leaves, of which one is sensitive to aryloxyphenoxypropionate herbicides. Biochem. J. 300, 557-565. Bender, E. A. (1978). An Introduction to Mathematical Modeling. (New York: John Wiley & Sons). Bettey, M., Ireland, R. J., and Smith, A. M. (1992). Purification and characterization of acetyl-CoA carboxylase from developing pea embryos. J. Plant Physiol. 140, 513-520. Browse, J., Roughan, P. G., and Slack, C. R. (1981). Light control of fatty acid synthesis and diurnal fluctuations of fatty acid composition in leaves. Biochem. J. 196, 347-354. Burrell, M. M., Mooney, P. J., Blundy, M., Carter, D., Wilson, F., Green, J., Blundy, K. S., and ap Rees, T. (1994). Genetic manipulation of 6-phosphofructokinase in potato tubers. Planta 194, 95-101. Cahoon, E. B., Lindqvist, Y., Schneider, G., and Shanklin, J. (1997). Redesign of soluble fatty acid desaturases from plants for altered substrate specificity and double bond position. Proc. Natl. Acad. Sci. USA 94, 4872-4877. Cahoon, E. B., Shanklin, J., and Ohlrogge, J. B. (1992). Expression of a coriander desaturase results in petroselinic acid production in transgenic tobacco. Proc. Natl. Acad. Sci. USA 89, 11184-11188. Clough, R. C., Matthis, A. L., Barnum, S. R., and Jaworski, J. J. (1992). Purification and characterization of 3-ketoacyl-acyl carrier protein synthase III from spinach. J. Biol. Chem. 267, 20992-20998. Dehesh, K., Schneider, G., Lindquvist, Y., and Byrne, J. (1998). KAS enzymes: crystal structure and identification of a new, medium chain-specific member of the family, KAS IV. Presented at l3"I Int. Symp. Plant Lipids, Sevilla. Diirmann, P., Spener, F., and Ohlrogge, J. B. (1993). Characterization of two acyl-acyl carrier protein thioesterases from developing Cuphea seeds specific for medium-chain- and oleoyl-acyl carrier protein. Planta 189, 425-432. Fell, D. (1997). Understanding the Control of Metabolism. (London: Portland Press). 106 Furbank, R. T and Taylor, W. C. (1995). Regulation of photosynthesis in C3 and C4 plants: A molecular approach. Plant Cell, 7, 797-807. Giere, R. N. (1984). Understanding Scientific Reasoning. (New York: Holt, Reinhart). Guerra, D. J., and Ohlrogge, J. B. (1986). Partial purification and characterization of two forms of malonyl-Coenzyme Azacyl carrier protein transacylase from soybean leaf tissue. Arch. Biochem. Biophys. 246, 274-285. Gulliver, B. S., and Slabas, A. R. (1994). Acetoacyl-acyl carrier protein synthase from avocado: its purification, characterization and clear resolution from acetyl CoA: ACP transacylase. Plant Mol. Biol. 25, 179-191. Haake, V., Geiger, M., Walch-Liu, P., Engels, C., Zrenner, R., and Stitt, M. (1999). Changes in aldolase activity in wild-type potato plants are important for acclimation to growth irradiance and carbon dioxide concentration, because plastid aldolase exerts control over the ambient rate of photosynthesis across a range of growth conditions. Plant J. 17, 479-489. Heath, R. J., and Rock, C. O. (1995). Enoyl-acyl carrier protein reductase (fab!) plays a determinant role in completing cycles of fatty acid elongation in Escherichia coli. J. Biol. Chem. 270, 26538- 26542. Heldt, H. W. (1979). Chapter 16: Light-dependent changes of stromal H+ and Mg2+ concentrations controlling CO2 fixation. In Encyclopedia of Plant Physiology, New Series. (Berlin: Springer-Verlag) pp 202-207. Jaworski, J. G., Clough, R. C., and Barnum, S. R. (1989). A cerulenin insensitive short chain 3-ketoacyl-acyl carrier protein synthase in Spinacia oleracea leaves. Plant Physiol. 90, 41-44. Jaworski, J., and Hinneburg-Wolf, B. (1998). Analyzing partial reactions of mutagenized spinach 3-ketoacyl-acyl carrier protein synthase III (KAS III). Presented at 13’“ Int. Symp. Plant Lipids, Sevilla. Kacser, H., and Acerenza, L. (1993). A universal method for achieving increases in metabolic production. Eur. J. Biochem. 216, 361-367. Kinney, A. J., and Hitz, W. D. (1995). Improved soybean oils by genetic engineering. Presented at Biochem. Mol. Biol. Plant Fatty Acids Glycerolipids Symp., South Lake Tahoe, CA. 107 I" '.__C ‘..- Kinney, A. J. (1994). Genetic modification of the storage lipids of plants. Curr. Op. Cell Biol. 5, 144-151. Knutzon, D. S., Thompson, G. A., Radke, S. E., Johnson, W. B., Knauf, V. C., and Kridl, J. C. (1992). Modification of Brassica seed oil by antisense expression of a stearoyl-acyl carrier protein desaturase gene. Proc. Natl. Acad. Sci. USA 89, 2624-2628. Kopka, J., Ohlrogge, J. B., and Jaworski, J. G. (1995). Analysis of in Vivo levels of acyl-thioesters with gas chromatography/mass spectrometry of the butylamide derivative. Anal. Biochem. 224, 51-60. Kuo, T. M., and Ohlrogge, J. B. (1984). The primary structure of spinach acyl carrier protein. Arch. Biochem. Biophys. 234, 290-296. McKeon, T. A., and Stumpf, P. K. (1982). Purification and characterization of the stearoyl-acyl carrier protein desaturase and the acyl-acyl carrier protein thioesterase from maturing seeds of safflower. J. Biol. Chem. 257, 12141-12147. Mohan, S. B., and Kekwick, R. G. O. (1980). Acetyl-Coenzyme A carboxylase from avocado (Persea americana) plastids and spinach (Spinacia oleracea) chloroplasts. Biochem. J. 187, 667-676. Niederberger, P., Prasad, R., Miozzari, G., and Kacser, H. (1992). A strategy for increasing an in viva flux by genetic manipulations. The tryptophan system of yeast. Biochem J. 287, 473-479. Ohlrogge, J., and Browse, J. (1995). Lipid biosynthesis. Plant Cell 7, 957-970. Ohlrogge, J. B., and Jaworski, J. G. (1997). Regulation of fatty acid synthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 48, 109-136. Platt, J. R. (1964). Strong inference. Science 146, 347-353. Pollard, M. R., Anderson, L., Fan, C., Hawkins, D. J., and Davies, H. M. (1991). A specific acyl-ACP thioesterase implicated in medium chain fatty acid production in immature cotyledons of Umbellularia caltfarnica. Arch. Biochem. Biophys. 284, 306-312. Popper, K. R. (1953). Science: Conjectures and refutations. In Philosophy of Science and the Occult, P. Grim, ed. (1982). (Albany, NY: SUNY Press). pp 87- 93. 108 Post-Beittenmiller, D., Jaworski, J. G., and Ohlrogge, J. B. (1991). In vivo pools of free and acylated acyl carrier proteins in spinach. J. Biol. Chem. 266, 1858-1865. Post-Beittenmiller, D., Roughan, G., and Ohlrogge, J. B. (1992). Regulation of plant fatty acid biosynthesis: analysis of acyl-Coenzyme A and acyl-acyl carrier protein substrate pools in spinach and pea chloroplasts. Plant Physiol. 100, 923-930. Post-Beittenmiller, M. A., Schmid, K. M., and Ohlrogge, J. B. (1989). Expression of halo and apo forms of spinach acyl carrier protein I in leaves of transgenic tobacco plants. Plant Cell 1, 889-899. Roesler, K., Shintani, D., Savage, L., Boddupalli, S., and Ohlrogge, J. (1997). Targeting of the arabidopsis homomeric acetyl-Coenzyme A carboxylase to plastids of rapeseeds. Plant Physiol. 113, 75-81. Rolleston, F. S. (1972). A theoretical background to the use of measured concentrations of intermediates in study of the control of intermediary metabolism. Curr. Top. Cellu. Reg. 5, 47-75. Roughan, P. G. (1998). Substrate channelling within the chloroplast fatty acid synthase: concentrations of acyl-acyl carrier proteins are very low during fatty acid synthesis. In preparation. Roughan, P. G., Holland, R., and Slack, C. R. (1979). On the control of long-chain-fatty acid synthesis in isolated spinach (Spinacia oleracea) chloroplasts. Biochem. J. 184, 193-202. Roughan, P. G., and Slack, C. R. (1982). Cellular organization of glycerolipid metabolism. Annu. Rev. Plant Physiol. 33, 97-132. Sasaki, Y., Kozaki, A., and Hatano, M. (1997). Link between light and fatty acid synthesis: thioredoxin-linked reductive activation of plastidic acetyl-Coenzyme A carboxylase. Proc. Natl. Acad. Sci. USA 94, 11096-11101. Sauer, A., and Heise, K. (1984). Regulation of acetyl-Coenzyme A carboxylase and acetyl-Coenzyme A synthetase in spinach chloroplasts. Z. Naturforsch 39c, 268-275. Schultz, D. J., and Ohlrogge, J. B. (1998). Ferredoxin may be a key component to production of unusual monoenoic fatty acids in transgenic oilseeds. Presented at l3“I Int. Symp. Plant Lipids, Sevilla. 109 Shimakata, T., and Stumpf, P. K. (1982). Isolation and function of spinach leaf B-ketoacyl-[acyl-carrier-protein] synthesis. Proc. Natl. Acad. Sci. USA 79, 5808-5812. Shimakata, T., and Stumpf, P. K. (1983). The purification and function of acetyl Coenzyme Azacyl carrier protein transacylase. J. Biol. Chem. 258, 3592-3598. Shintani, D. K., and Ohlrogge, J. B. (1995). Feedback inhibition of fatty acid synthesis in tobacco suspension cells. Plant J. 7, 577-587. Somerville, C., and Browse, J. (1991). Plant lipids: metabolism, mutants, and membranes. Science 252, 80-87. Stitt, M., and Schulze, D. (1994). Does Rubisco control the rate of photosynthesis and plant growth? An exercise in molecular ecophysiology. Plant Cell & Environ. 17, 465-486. Stitt, M. (1999). Coordinate regulation in central metabolic pathways. Presented at Plant Biol. & Mol. Biol. Symposium: Metabolic Networking in Plants, Ames, IA. Suh, M. C., Schultz, D. J., and Ohlrogge, J. B. (1999). Isoforms of acyl carrier protein involved in seed-specific fatty acid synthesis. Plant J. 17, 679-688. Tsay, J., Oh, W., Larson, T. J., and Jackowski, S., Rock, C. O. (1992). Isolation and characterization of the B-ketoacyl-acyl carrier protein synthase III gene (fabH) from Escherichia cali K-12. J. Biol. Chem. 267, 6807-6814. Verwoert, I. I. G. S., van der Linden, K. H., Walsh, M. C., N ijkamp, H. J. J., and Stuitje, A. R. (1995). Modification of Brassica napus by expression of the Escherichia cali fabH gene, encoding 3-ketoacyl-acyl carrier protein synthase 111. Plant Mol. Biol. 27, 875-886. Voelker, T. A., Worrell, A. C., Anderson, L., Bleibaum, J., Fan, C., Hawkins, D. J., Radke, S. E., and Davies, J. M. (1992). Fatty acid biosynthesis redirected to medium chains in transgenic oilseed plants. Science 257, 72-74. Williamson, 1. P., and Wakil, S. J. (1966). Studies on the mechanism of fatty acid synthesis: XVII. Preparation and general properties of acetyl Coenzyme A and malonyl Coenzyme A-acyl carrier protein transacylases. J. Biol. Chem. 241, 2326-2332. Winter, H., Robinson, D. G., and Heldt, H. W. (1994). Subcellular volumes and metabolite concentrations in spinach leaves. Planta 193, 530-535. 110 Chapter 4 EXPRESSION OF B-CT AND THE STRUCTURE OF ACCD Abstract Plastidial acetyl-CoA carboxylase (ACCase), one of the rate determining enzymes of fatty acid synthesis (FAS), is encoded by four different genes in dicot species. Although three of these genes are nuclear-encoded, the fourth, Ach, is encoded on the plastid genome. Originally the goal of the work presented in this chapter was to create transgenic tobacco plants with variable levels of the Ach protein, B-carboxylase transferase (B-CT). These plants would have allowed investigation of l) coordinate regulation of ACCase subunits, 2) mutations in Ach structure, and 3) the effects of low rates of FAS. This chapter presents two unsuccessful attempts to relocate Ach to the nuclear genome. The first attempt used tobacco Ach behind the promoter and transit peptide from the small subunit of RuBisCO, while the second attempt used the pea Ach behind a 358 promoter and the transit peptide from the small subunit of RuBisCO. In neither set of transformed plants was any expressed B-CT protein detected, nor was message from the constructs detected. Sequence analysis shows the pea and tobacco Ach sequences to be AT-rich with non-plant codon usage. In addition, many motifs which may cause non-desired transcript processing such as polyadenylation or degradation are found in the 111 coding regions. While none of these motifs is known to be sufficient for processing, the abundance of possible motifs may be the cause of non-expression of the protein. In addition, this chapter presents preliminary sequence data for the Ach gene of Oenothera hookeri. The 5 ’ end of the Oenothera sequence has essentially no conservation with the 5’ end of the other Ach genes and contains many repeated motifs from 9 to 45 bp in length. The Ach gene was also mapped on the Oenothera plastidial genome. Rather than being between the rbcL and petA genes as in the tobacco, pea, pine, and liverwort plastidial genomes, it is located between the petA and atpA genes at the opposite end of the large single copy region of the plastome. 112 Introduction Acetyl-CoA carboxylase (ACCase, EC 6.4.1.2) is one of the rate limiting enzymes of fatty acid synthesis (FAS). ACCase is a biotin-containing enzyme that catalyzes the carboxylation of acetyl-CoA to form malonyl-CoA, the building block of FAS. ACCase contains three functional subunits: a biotin carrier protein (BCCP), a biotin carboxylase (BC), and a carboxyl-transferase (CT). There are two different types of acetyl-CoA carboxylase in plants: a multi-subunit form and a multi-functional form. In dicots and most monocots both forms exist: the multi-subunit form is located in the plastids, while the multi-functional form is presumably located in the cytosol. In grasses, only the multi-functional form of ACCase exists. In this case, there are at least two isoforms of this greater-than 250 kDa protein, one of which is imported into plastids and used in de nova FAS and which therefore replaces the multi-subunit form found in most plant plastids. ACCase structure and localization have been recently reviewed by Sasaki et al. (1995). The BCCP and BC subunits of the multi-subunit ACCase are each encoded by nuclear genes. In the multi-subunit ACCase, the carboxyl-transferase function is encoded by two genes: AccA encodes the a-CT subunit and Ach, the subject of this chapter, encodes the B-CT subunit. While a-CT is a nuclear- encoded gene, the B-CT gene is located on the plastidial genome. It is the only gene of plant lipid metabolism (of more than 100) which is plastome-encoded 113 and perhaps the only plastidial gene of known function which is not involved in photosynthesis or protein synthesis. Of the four genes which encode the multi- subunit form, only one gene has been identified for each of the two CT and the BC subunits, while at least two BCCP-encoding genes have been identified (Mekhedov et al., in preparation). Proteins with multiple subunits may have coordinate expression of the genes encoding the subunits. For example, the gene which encodes the small subunit of RuBisCO is a nuclear gene, whereas the gene which encodes the large subunit is a plastidial gene. When the large subunit of RuBisCO is not available, the small subunit is imported into the chloroplast but rapidly degraded (Schmidt and Mishkind, 1983). Thus, RuBisCO subunits are kept in strict stoichiometric amounts. While the genes which encode the ACCase subunits are transcribed at the same time and thus expressed coordinately (Nikolau and Wurtele, 1998), underexpression of BC does not affect BCCP accumulation until very little BC is present (Shintani et al., 1995). However, since the AccA and Ach genes code for different subunits of the same functional component, over- or underexpression of these genes may still affect gene expression or protein stability of the other component. Numerous genes from the ancestral plastidial genome have been lost or transferred to the nuclear genome during the evolution of plants (for review see Gray, 1992). Generally, the transfer of a single gene to the nucleus has occurred 114 in parallel in independent lineages (Martin et al., 1998). That is, even though there have been independent endosymbioses of chloroplasts, the same genes tend to be lost. Although Ach is present in dicot and most monocot plastidial genomes, it has been removed multiple times from the genomes of Cyanophora, Odontella, Euglena, and Porphyra (Martin et al., 1998). Other genes which have been removed with the same frequency are the ndh genes and the cysA and cysB genes (Martin et al., 1998). In the monocots, both Zea mays and Orzya sativa have vestiges of the Ach gene in their plastid genomes. However, in these species and other grasses, the present-day functional plastidial ACCase is a multi-functional, not a multi-subunit, enzyme. The original purpose of the work of described in this chapter was two-fold 1) determine if B-CT and a-CT subunits accumulate in strict stoichiometric amounts similar to the large and small subunits of RuBisCO, and 2) alter the expression of B-CT using a light-regulated promoter, remove the Ach gene from the plastidial genome, and use those plants to study the effects of altered B-CT and presumably ACCase availability. This chapter describes unsuccessful attempts to move the B-CT gene from the plastidial genome to the nuclear genome. In addition, the chapter 1) presents evidence that the pea and tobacco Ach genes must be re-engineered to be functional in a nuclear background, 2) presents the sequence of the Oenothera haokeri .4ch gene, 3) analyzes the tobacco and pea Ach sequences for possible transcript instability motifs and 115 pO\yadenylation signals, and 4) maps the Ach gene on the 0. hoakeri plastidial genome. Materials and Methods Materials Restriction enzymes used in these studies were from Boheringer Mannheim. E. coli strains used were DHSa and DH50t’. A. tumefaciens strains used were LBA4404 and ABI (Monsanto). The N. tabacum variety was SR-l, and the pea variety was Little Marvel. Accession numbers for the genes involved are N. tabacum: 200044, bases 59793 to 61331; P. sativum, X54750; P. thunbergii D17510; B. napus, 250868; and M. polymorpha (liverwort) X04465. Surface Sterilization of Seeds Seeds were agitated in 10% Clorox bleach with 0.001% Tween-20 for 10-20 min, washed three times for 5-10 min each with sterile H2O. Construction of fi-C T transformation vectors The construction of plasmids p0921-3.3 and p0921-3.5 is shown in Figure 4.1. Standard protocols were used (Sambrook et al., 1989; Magic Minipreps, Promega; GeneClean II, BiolOl, La Jolla, CA). One ng of genomic N. tabacum DNA (gift of Keith Roesler) was amplified with Taq polymerase (Boheringer Mannheim) using primers JO360 (5’- CAU CAU CAU CAU ATG CAT GCA 116 E coil] /NotI EcaRI / Nsil F... Hindi] ”mm" Ecdil P1490 l p0921-3.5 (11.6) p0921-3.3 0 219 17 = 3003 OCS Ach RGLIIPWVQM small subunit promoter/transit HindIII Sal] Nsil HindIII 0 0921’3-5 1219 2787 3003 small subunit promoter/transit MQVWPIILGRAch Perch? Hindi]! Nsil Sal! Hind]! Figure 4.1. Construction of plasmids p0921-3.3 and p0921-3.5. 117 GGT GTG GCC AAT AAT TTT AGA GCG AAT GAC TA) and JO329 (CUA CUA CUA CUA GTC GAC TAC TTG ATT TTG CTT GA) for 30 cycles. Primer JO360 contains the sequence encoding the first 10 amino acids of the mature small subunit of RuBisCO in tobacco. Both primers were from the Michigan State University Macromolecular Structure, Sequencing and Synthesis Facility (East Lansing, MI) and contained deonyMP (U), used in conjunction with uracil DNA glycosylase to insert the PCR product into the pAMPl vector (Gibco/BRL). Each PCR cycle consisted of a denaturation step (94°C, 1 min), a primer annealing step (45°C, 2 min) and an elongation step (72°C, 1.5 min). The thermocycling was preceded by an initial denaturation step of 94°C for 3 min and followed by a final elongation of 10 min at 72°C. The PCR product was inserted into the pAMPl vector (Gibco/BRL) according to the manufacturers protocol. This plasmid was designated p0714. The EcoRI/Notl fragment from this vector was inserted into pBluescript II KS+ (Stratagene) The EcaRI/NsiI fragment from p1490 (Post-Beittenmiller et al., 1989) was added to the pBluescript II KS+ plasmid with the B-CT gene. This EcaRI/NsiI fragment contains the RuBisCO small subunit promoter and transit peptide. The termination signal was provided by insertion of the 3-kb EcoRI/Sall fragment from the pBluescript II KS+ construct into pUCA7-TX (provided by C. Gatz, University of Gottingen, Germany). The HindIII fragment containing the promoter, transit peptide, B-CT gene, and OCS terminator was inserted into the plant transformation vector pPZP122 (Hajdukiewicz et al., 1994). The plasmid with the B-CT gene construct coding in the same direction of the AacCI coding region was denoted p0931-3.3, 118 the plasmid with the B-CT construct coding in the opposite direction was denoted p0931-3.5. AacCI codes for gentamycin resistance. The constructs were sequenced by the Michigan State University DNA Sequencing Facility (East Lansing, M1) to verify that the junction between the transit peptide, additional amino acids, and the start of the Ach gene remained in frame. A second construct was designed and built by Oskar Martinez de la Ilarduya (University of California, Riverside) and is depicted in Figure 4.2. Pea (Little Marvel) genomic DNA was amplified using custom primers from GibCO/BRL, Ach.DO (5’-NN GAA TTC TC ATG AAA AGA AAG TGG) and Ach.UP (5’-CTT CAT TTA TCA TTT GTC GTT T). The sense primer was designed to. amplify 130 bases prior to the published start site of the pea genomic sequence. The EcoRI and SacI sites added during PCR were used to insert the PCR product, pea B-CT*, into pUC19. The BspHI/Sacl fragment which contained the ATG start codon was transferred to pMON25661 to add the small subunit transit peptide and E93’ termination signal. The BglII/NatI fragment from this vector was then ligated into pMON10098 to provide the 358 promoter and plant transformation sequences. The construct was sequenced to verify junctions. Plant Transformation Standard protocols were used for plant transformation. Surface sterilized leaf discs or strips were incubated in A. tumefaciens culture containing the desired plasmid, briefly blotted, then transferred to M8104 plates (4.31 g/l MS Salts 119 I 1 l_ l I / 7 7 — p35S SSUAA B-CT“ E93’ p35S KAN NOS 3’ Figure 4.2. Pea B-CT“ construct. This construct has a 35S promoter with an E93’ termination signal. In addition to the published coding region of the pea Ach gene, this construct carries 130 bases 5’ of the coding region. The construct is in the pMON10009 vector (Monsanto) which contains sequence for plant transformation and kanamycin resistance, also behind a 358 promoter. 120 WWW??? (Gibco/BRL); 30 g/l sucrose; 100 mg/l myo-inositol; 10 mg/l thiamine-HCl; 1 mg/l nicotinic acid; 1 mg/l pyridoxine-HCl; 8 g/l phytoagar (Gibco/BRL)) with 1.0 ug/ml benzyladenine (BAP) and 0.1 ug/ml naphaleneacetic acid (NAA). Afier two to four days, leaf pieces were transferred to M8104 plates with antibiotics (100 mg/l carbenicillin and either 150 mg/l kanamycin or 100 mg/l gentamycin depending on vector), and the hormones BAP and NAA. Shoots from these leaf pieces were transferred to root-induction media (MS104 with 100 mg/l carbenicillin and either 150 mg/l kanamycin or 100 mg/l gentamycin depending on vector), allowed to root, then transferred to soil. Seed Germination Assays Surface sterilized seeds from p0921-3.3- and p0921-3.5-transformed plants were germinated on seed germination medium (Fisher and Guiltinan, 1995) in the presence of gentamycin. Surface sterilized seed from B-CT"‘-transformed plants were germinated on filter paper wetted with 0.1X MS-Salts in the presence of 100 mg/l kanamycin (Knutzon et al., 1992). Seeds were scored resistant if they were green after two weeks. Neomycin phosphastranferase assays Neomycin phosphotransferase (NPT) assays were performed using a kit from Agdia, Inc. (Elkhart, IN) according to manufacturer’s protocol. 121 .' 2" Western Blotting Protein was extracted from approximately 20 mg (fresh weight) tissue from transformed plants by freezing, grinding in 1X SDS sample buffer (0.25 M Tris, pH 6.8, 8% SDS, 20% glycerol, 0.04% bromphenol blue, 5 mM DTT), then boiling for 10 min. Clarified extract was separated on 10% or 7.5% SDS-PAGE gels and transferred to membranes and detected according to standard protocols (Hoefer Scientific Instruments, 1994). All antibodies were used at 1:1000 and were all gifts from Linda Savage. The MSU-l and UMC-l antibodies were prepared from E. cali-expressed tobacco B-CT protein. The MSU6(E) antibody was prepared from E. -coli-expressed pea B-CT protein and could detect as little as 10 ng protein. Neither tobacco nor pea antibodies cross-reacted with each other. RNA Extraction 1 One hundred to 250 mg (fresh weight) tissue from young leaves was frozen in liquid N2 and ground to a powder. Total RNA was extracted similarly to the method described in Shintani (1996). One ml of 80°C homogenization buffer (200 mM sodium borate, 30 mM EGTA, 1% SDS, 1% deoxycholate, 2% PVP 40,000, 10 mM DTT, pH 8.0) was added to the ground powder, followed by 1 ml of phenol: chloroform (1:1). The mixture was vortexed for several seconds, then centrifuged at 10,000 x g for 15 min. The aqueous layer was removed and saved. The organic phase was back-extracted with 1 ml H2O, spun for 15 min at 10,000 x g, and the aqueous phase combined with the previously-saved aqueous phase. 122 The Combined aqueous phases were extracted with one vol of chloroform and spun for 5 min at 2700 rpm to separate phases. The aqueous layer was adjusted to 2M LiCl and incubated overnight on ice. The precipitate was collected by centrifugation at 10,000 x g at 4°C. RNA pellets were dried under vacuum. Pellets were resuspended and reprecipitated by the addition of 1/10 vol of 2M potassium acetate and two vol of ice cold ethanol. The precipitate was recovered by centrifugation at 10,000 x g for 20 min, the pellet dried under vacuum, and resuspended in 200 111 H2O. Total RNA was stored at -20°C for less than one week. RNA Extraction 2 Fifty to 100 mg tissue from young leaves was frozen in liquid N2 and ground to a powder. Ground powder was kept in liquid N2 until all samples were ground. One ml TRIZOL® reagent (Gibco/BRL) was added and tubes were incubated 5 min, then spun 10 min at 12,000 x g, 4°C to pellet debris. The supernatant was transferred to a fresh tube and total RNA isolated using the manufacturer’s protocol. Northern Probes C anstructian Engineered Q-C T: Plasmid p0714 (from Figure 4.1) was digested with NsiI/SaII. This 1.2-kb fragment contains the first 10 codons from the mature small subunit of RuBisCO and the tobacco Ach coding region. This fragment was isolated using a Geneclean II kit (Bio 101 Inc., La Jolla, CA), then labeled with random 123 hexfimers using 32P-CTP. Unincorporated nucleotides were removed with a G-50 spin column (Sambrook et al., 1989). Small Subunit Riboprobe: A linearized plasmid containing the tobacco RuBisCO small subunit coding region (gift of David Shintani) was transcribed in the antisense direction using anP-UTP and a MAXIscriptTM In Vitro Transcription kit according to the manufacturer’s protocol (Ambion). This yielded a 250-bp transcript, of which 180 bp matched the small subunit coding region; the remainder matched the multiple cloning site from pBluescript II KS+. Unincorporated nucleotides were removed with a G-50 spin column (Sambrook etal,1989) 35SAB-CT probe: The 3SS/B-CT" plasmid from Figure 4.2 was digested with BglII/SacI. This fragment contains the SSU transit peptide and the pea B-CT coding region. The fragment was inserted into pBluescript II KS+ cut with BamHI/SacI. This plasmid was transformed into E. coli to amplify it, then digested with SacI/SmaI. The resultant 2.1-kb fragment was isolated using a JETSORB kit (Genomed, Research Triangle Park, NC) and labeled with [a-32P]dATP using Ambion’s Strip-EZTM DNA kit (Ambion, Austin, TX). Unincorporated nucleotides were removed with a Micro Bio-Spin® P-6 column (Bio-Rad Laboratories, Hercules, CA) using the manufacturer’s protocol. 124 W: Vector DNA from pMON28860 (Monsanto) was digested with Kim]. This yielded a 1.4-kb fragment containing the kanamycin-resistance coding region and the 3’ nopaline synthase terminator region (NOS). This fragment was inserted into pBluescript II KS+ cut with XhoI. The plasmid was then transformed into E. cali to amplify it. Plasmid DNA was digested with XhoI. The 1.4-kb fragment was isolated with a JETSORB kit (Genomed, Research Triangle Park, NC), and the resultant fragment labeled with [a-32P]dATP using Ambion’s Strip-EZTM kit. Unincorporated nucleotides were removed with a Micro Bio-Spin® P-6 column (Bio-Rad Laboratories, Hercules, CA) using the manufacturer’s protocol. Sequencing and Sequence Analysis Plasmid pOjSalE (from B. Sears, Michigan State University) was used for sequencing the 0. hookeri Ach gene. This plasmid contains 7 kb of the plastidial genome in pRL498 (Elhai and Wolk, 1988). Sequencing was performed by the Michigan State University DNA Sequencing Facility (East Lansing, MI) using primers made by the Michigan State University Macromolecular Structure, Sequencing and Synthesis Facility (East Lansing, MI). The gene was sequenced in one direction, from one to five times, depending on sequence data. Oenothera sequence data were assembled using a combination of tools: BLAST (Altschul et al., 1990; http://www.ncbi.nlm.nih.gov), GCG Sequence Analysis Software Package 7.3-Unix (Genetics Computer Group, Madison, WI), and Squeb version 1.1 (Genetics Computer Group, Madison, WI). Unpublished Oenothera 125 sequence for alignment and comparison was also provided by R. G. Herrmann (Ludwig Maximilians University) in collaboration with Barbara Sears (Michigan State University). Mapping of Ach on the 0. hookeri plastome A DNA blot was prepared from a clone library from Oenothera haakeri using standard protocols (Sambrook et al., 1998). Plasmid pETllaNtTC (gift from Linda Savage), a pETlla (Novagen) expression vector containing the complete tobacco Ach gene, was labeled with 32P and hybridized to the DNA on the membrane. The blot was exposed to film for 4 h at -70°C. Results Transformation of N. tabacum with B-C T constructs Leaf discs were incubated in Agrabacterium culture containing one of the three B-CT constructs detailed above and in Figures 4.1 and 4.2. Each construct contained an Ach gene fused to a chloroplast transit peptide from the small subunit of RuBisCO. Two constructs, p0921-3.3 and p0921-3.5 used a light- regulated promoter and conferred resistance to gentamycin and they differed in the orientation of the Ach gene in the plasmid. These constructs contained the Ach gene from tobacco. The third construct, B-CT“, used a 35S promoter and contained the Ach gene from pea. 126 Ten plants which went through the plant transformation protocol using plasmid p0921-3.5 and which grew roots on root-induction plates containing 100 ug/ml gentamycin were transferred to soil. Of these ten plants, four produced seed which germinated in the presence of 100 ug/ml gentamycin. Of the six plants transformed with p0921-3.3 two survived and produced seed capable of germinating on gentamycin. None of the plants transformed with the pPZP122 vector alone produced seed which germinated on gentamycin. Thirteen plants transformed with the 3SS/B-CT“ construct were transferred to soil. Nine of these expressed the neomycin phosphotransferase (NPT) gene at various levels. These NPT expressing plants were the only ones able to germinate on media containing 150 ug/ml kanamycin. Six plants transformed with pMON10009 were transferred to soil. Of these, four expressed NPT and were able to germinate on media containing 150 ug/ml kanamycin. Most of the antibiotic-resistant lines which produced germinating seed had germinationznon- germination ratios of 7 5%, although in some lines all seeds tested were resistant. Immunoblots of transformed plants Most of the plants transformed with the pPZP-derived vectors containing the tobacco Ach gene were checked for expression of the B-CT protein. Plants putatively transformed with p0931-3.3 and p0931-3.5 were probed with two antibodies raised against tobacco B-CT (gifts of Linda Savage). Neither antibody detected any B-CT protein in transgenic plants. Plants transformed with the 127 pMON-derived vector containing the pea Ach gene were also checked for expression of the B-CT protein. Although the MSU-6E antibody could easily detect 10 ng protein, no pea B-CT protein was detected in the transgenic plants. RNA blots As all attempts to detect B-CT protein in transgenic plants led to negative results, an indication was sought that the message for the B-CT protein was present. Figure 4.3 shows that mRNA corresponding to the plasmids from the plants transformed with p0921-3.3 and p0921-3.5 cannot be detected. The blot contains total RNA from nine lines, five of which are transformed according to seed germination tests on kanamycin. Figure 4.3A shows the blot after hybridization with the engineered B—CT probe. This probe contains 30 bases of the coding region from the small subunit of RuBisCO and 1536 bases which encode the Ach gene. It is clear that there is no correlation between transformation and hybridization. In order to confirm that the small subunit promoter was functional and the blot had detectable mRNA, a second hybridization was done to the same membrane. Figure 4.3B, shows the same blot probed with an RNA probe containing the antisense message to the small subunit transit peptide coding region. Since Figure 4.3B shows that small subunit message RNA was present, the blot did contain detectable mRNA, and the light-regulated small subunit promoter functional, at least for the native gene. 128 ‘ ‘£‘* Figure 4.3. RNA blots from plants transformed with plasmids p0921-3.3 and p0921-3.5. RNA was extracted from young leaves using extraction method 1. Twenty ug of total RNA was fractionated on a 1.2% agarose RNA gel (1X MOPS, 2% formaldehyde, 10 ug/ml ethidium bromide). RNA was transferred to Zeta-Probe® membrane (Bio-Rad Laboratories, Hercules, CA) using the manufacturer’s Northern Blotting protocol. Membranes were prehybridized in 5X SSC, 10X Denhardt’s, 0.1% SDS, 100 mM potassium phosphate, pH 6.8, 100 ug/ml salmon sperm DNA at 42°C overnight using a hybridization oven (Hybaid, Hybaid Instruments, Holbrook, NY) at the oven speed 6 setting. Lanes left to right: 1003- (pPZP122), 1302, 2301+, 2302+, 1503-, 2504+, 2505+, 2506+, 2507-. Lines which are denoted with + germinated on gentamycin containing media. A) Engineered B-CT probe (5 x 10‘5 CPM/ml buffer) was added to hybridization buffer (5X SSC, 10X Denhardt’s, 250 mM potassium phosphate, pH 6.8, 30% formamide, 100 ug/ml salmon sperm DNA) and the membrane was hybridized overnight at 42°C at the oven setting of speed 6. The membrane was washed four times. The first wash was at 50°C in 2X SSC, 0.5% SDS for 5 min; the second at 50°C in 2X SSC, 0.5% SDS for 20 min; the third at 65°C in 2X SSC, 0.5% SDS for 20 min; and the last at 65°C in 0.2X SSC, 0.5% SDS for 30 min. The membrane was phosphoimaged overnight. B) Afier imaging the blot was stripped by washing twice in 0.1X SSC, 0.5% SDS at 80°C for 20 min, the blot was prehybridized as above, then labeled small subunit riboprobe (4.9 x 10’ CPM/ml buffer) added and hybridized overnight. The membrane was washed as follows. 1) and 2) 50°C for 20 min in 2X SSC, 0.5% SDS at speed 10; 3) 50°C for 20 min in 0.2X SSC, 0.5% SDS at oven speed setting 10; 4) 50°C for 20 min in 0.2X SSC, 0.5% SDS at speed 5; 6) 60°C for 30 min in 0.1X SSC, 0.5% SDS at oven speed setting 5. 7) The final washes were at 65°C and 70°C, each for 30 min in 0.1X SSC, 0.5% SDS at the oven’s setting of speed 5. 129 -nomm +oomm momm +VOmm -mom_ +~om~ +~omm Nom_ -moo~ It". 130 Figure 4.4 shows the Northern data from a second set of transformed plants. This set of plants was transformed with the plasmid 3SS/B-CT*. Figures 4.4A and 4.43 show the blots probed with 3SS/B-CT. Once again, no mRNA hybridizing to the pea B-CT gene was detected. Figures 4.4C and 4.4D show the same blots hybridized with a kanamycin gene probe. These panels confirm that the 35S promoter is expressed and the plants are transformed. Sequence of Oenothera haakeri Ach At the time this work was initiated, Ach sequences of higher dicot plants were available only for tobacco and pea. These two sequences diverged greatly at the N-terminal region of the protein. Therefore, to help obtain a more complete picture of the primary structure of this protein in plants and to take advantage of available materials, the genomic sequence of 0. haokeri Ach was determined. The sequence was only determined in one direction, therefore, only preliminary sequence data are available. This data is shown in Figure 4.5. Possible start codons are at 319 and 556. There are five different direct repeat motifs in the 5’ coding region of the gene, three of which are 20-21 bp in length. First, a 21-bp repeat with low C content is located at 644 and 689. Two imperfect repeats of this region are at 583 and 614. Second, adjacent to these repeats is another shorter 9-bp repeat. This repeat is located after the first three repeats, but before the fourth repeat. A third 45-bp repeat is located at 416 and 488. A fourth repeat of 21-bp with low TC content is located at 394 and 466 with imperfect copies at 728 and 761. The fifth repeat is a 20 bp repeat with no T’s and one C. These 131 Figure 4.4. RNA blots from plants transformed with plasmid B-CT*. RNA was extracted from pooled apical meristems and small leaves of transformed lines using extraction method 2. Unless otherwise noted, 30 ug of total RNA was fractionated on 1.2% agarose RNA gels. RNA was transferred to Zeta-Probe membrane as in Figure 4.3. Membranes were prehybridized at least 2 hours. A) and B) B-CT probe (2.8 x 106 CPM/ml buffer) was added to hybridization buffer and membranes were hybridized overnight at 42°C at oven speed setting 6. Membranes were washed four times at 42°C. The first wash was in 2X SSC, 0.5% SDS for 10 minutes at oven speed setting 10; the second in 2X SSC, 0.5% SDS for 15 minutes at oven speed setting 7. The third and fourth washes both used 1X SSC, 0.5% SDS at oven speed setting 7, and each lasted 20 minutes. Membranes were phosphoimaged overnight. Blots were stripped using Ambion’s Strip-EZ kit according to the manufacturer’s protocol. C) and D) The same blots were prehybridized as above, then labeled kanamycin probe (2.1 x 106 DPM/ml buffer) was added to hybridization buffer and membranes were hybridized overnight and washed as in A) and B). A) and C) Lanes left to right: 0.5 ng unlabeled B-CT probe, 0.5 ng unlabeled kanamycin probe, nothing, V-3+, V-13+, B-l+, B-2+, B—8, B-11+, 15 ug B-12, nothing, 2.5 ng unlabeled kanamycin probe, 0.1 ng unlabeled B-CT probe. B) and D) Lanes left to right: 2.5 ng unlabeled B-CT probe, 0.1 ng unlabeled kanamycin probe, nothing, B-13, B- 10+, B-9+, nothing, V-15+, V-11+, V-1+, wild type, 0.5 ng unlabeled kanamycin probe, 0.5 ng unlabeled B-CT probe. Lines which are denoted with + expressed NPT and germinated on kanamycin containing media. 132 133 Figure 4.5. Preliminary 0. haakeri Ach sequence. Since the sequence data is preliminary, upper case letters indicate a consensus between the different sequencing runs. Lower case letters indicate that different runs produced different results, with the most likely base given. Codes used when data did not yield a consensus are Y = C or T; M = A or C; and N = any base. Potential start codons and the stop codon are in boldface. Boldfaced numbers on the left side indicate the sequence of the region where the protein is conserved (see Figure 4.6). The promoter used to transcribe Ach is a nuclear-encoded RNA polymerase. The YRT motifs required for this type of promoter are italicized. The only Ach promoter sequence published is for N. tabacum (-129 ATATAAAGGGGGTTCCAACATATTAATATATAGTGTAAGT-) (Leire and Maliga, 1999). Repeats described in the text are indicated as below. [21 by, low C] 1:161), AG-rich] 245 hp inherent 134 l 61 121 181 241 301 361 421 481 541 601 661 721 781 841 901 961 1121 1081 1141 1201 1261 1321 1381 1441 1501 1561 1621 1681 1741 1801 1861 1921 1981 . . AGCCTACC 'TGAGGATAAG AAAAAGCCTC AAAAGGAGTA CCTACCAAGG TTAGATCCTT tctaacgaga {AECQTACCBA a-‘i—tae AGAaGGAtAA AAGCCTCAAA GATGCCAAGA TTGAGGATAA aGAAAAAGCy CTAAAAGGAG TAGATCCTaT agcggaagag IGTCTAACGAC. a’AagccTAAc gaAAgatgGG GGAGTAAGAA.AAAGCCTCAA AGGAGTAAGA AICAAGCCTA CCAAGGCTAA GATCAAGATC CTATCAAGCC TCAAGATCCT GGATGGCCAA TCTTGAGGAT TAAGAAAAAG GAAGCCTaCC cttgaagggg anemone heatereaee ATttTCGtAc AAGGATGGCC AAGaTCAAGA CCTCTTGAGG ATAAGGATGG AAGCCTGAAg GGCCTAAAAA AAAAGCCTCT AAGGAGTAAG TACCAAGGCT TCCTZTCAAG CCAAGathT GCCtaacaag atabtaaggd bgaccttgaa EQCTTthfigaGGATHATAAG gngCf.fi.f GAAGACCth AGGAAQCGGA .AéésCflGee G¢§§§T§AGC tCaAT Tamg chfi’g’éC 25.; if ' 258683 6155825 oooooooooooooo ooooooooooooooooo Cue—1.849593?IfieCCCECZ-‘ECIETC—areas an mAgGAagaCl .AacACCtTaA aGGGGATaAT _aTGAge—ae 9 new liatAataagc CCTAAGCATT GAGCGCGAGT tCTCTAaAAT TATAaGAAAA ATGCATAGTT GAAGGCCTCC CGGGTTGCTT TACCTAAAAC ATgGGATCCG TTACCTCTTA TTGATGCAAA TTCTATGTAG TgGGGGATAT TTGAyCAAaC ATAAGGGCTT TATTTAaTtT TGTaTAGyIA CAATTGTAGA GAACTCCCIA ctcAtaAGga eéI-fFHT-ffiifilié’ktimtikfi‘i; a GGGgCTGTCC TthtCCCGC CGGGCTATGC ttaTGAAGTC CAGATCGAAT TTTCTCTAGA CTTATAAAAG GTATTCACCT TAGTCGGCGA TTCTAGTGTG TGGCTAAAAT CTCTCCTTAC CAcACaAGAA ATTGAATATT ATTCGATCAA yCACGGyTTt GytTATCTgA AAGAATCGTG GGA agaccttgaa gctgaaggac CCaCCGTAAG TCAGTCTACC GCATTTATGG CaaaaaCAAT CGACCTTATG TCCTATTGAA AAAGACAGGA AGGAATTGGA GAGAATCACC TGCTTCCGGA ATCTCCTGCT ATCTCCCACT GAACCTAATG GAAGTACCTG ATTGTACCAC GTTaCTtTGA aTCAAAGTAA CGGACAATTG 135 gAGTATGAgg aATAAgCCTc ggggaTa tgtcCCTGgC CATCGGGATC aaGCGGGATa GtTCaCTcCA GTTTGTGaaC CTTGATCCAC TTTCATTCGG TTATCGGAAG CTTATGGATT CGTCTGGACG GGGGCACGTA TCATCTGATT ACGGGTGGTG yCTACATcGC .AGGGTTCACA GCAATCCttT mTTCTCAAGT AAAAAAATGT nTTTATATTG '{b -'-.' I. e .‘l o~o-o-n..- 04.6...‘O~.I.-.-.l-.-.o l-O-OeO'O-O-DD-O-otO-u.- {Cf 33.499. 51.859. .6. .. .9 Ct. Esflea 95?. '_ . . - _ jam: artefiartniariét. .ACTGGTTTCa gCAAgtCGGT CCGAtccha cACTTTGtTC aATGTGGATC AGACTTGGGC AGAAAGACCC CTATTCACAC TTCAGTTTAT AGTATGCTAC TGCAAGAAGG ATCAATTCAA TGACGGCTAG ATTCGCAGGT AACGGCTGAA AAAGGGTTCT aATTaaTATt ATTTTTGGTG ATATTCCTGA an CCaGATTTAC TCCCGCTAAG tTCCGAGgCG TGGATTCAAT TCATTTGGAA TCCTATGCAT TTATAAAGAT GGGCAGAGGT GGGGGGTAGT CAATCGAGTT GAGTGGGAGC TAGAACGGTA CTTTGggaTG AAAAGAGTAA TATTTATtCG tTGAGTGAGT yATAATTATT AAATAAGATT TTACTATCAG repeats are located at 818 and 839. The significance, if any of these repeats is unknown. The promoter sequence for the tobacco Ach transcript has been described (Leire and Maliga, 1999). This sequence has conserved motifs which consist of a YRT core in an AT-rich region. (Y = C or T). Due to the high GC content of the repeated motifs, there are few truly AT-rich regions in this area, however, YRT motifs are also shown in Figure 4.5. When the putative 0. hookeri protein is aligned with other B-CT proteins, the protein is more conserved at the carboxy end of the protein. Furthermore, there is no significant protein alignment of the amino-terminal end of the protein until the coding region is out of the highly repeated area of the gene (Figure 4.6). Mapping of Oenothera hookeri Ach on the plastidial genome Early attempts at obtaining the Oenothera Ach gene via PCR were unsuccessful. Due to the unconserved 5’ regions of the tobacco and pea genes, these attempts used primers based on the highly-conserved 3’ end of the gene that precedes it in both the tobacco and pea plastidial genomes - the large subunit of RuBisCO. Since these attempts were unsuccessful and the Oenothera plastidial genome has some unusual qualities, the location of Ach on the plastome was investigated. The map of the clones in the library and results of the hybridization are shown in Figure 4.7. An expression vector containing the tobacco Ach gene hybridized with the pOjSalE, Bam3a and Bam3b clones. This places the Ach gene between the petA and atpA genes in 0. hookeri. 136 tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} l MTIHLLYFHA 51 TNEDPNRKVK TTISKDRFIY LLSDKIFSIR LTRKRKSLKG -T-DK-R-I- 101 DTFLVSDRNG DTFFVRDSNK TNITKYKWTN AMSMRKTLEY DTF-V-D-N- 151 SNYRNNGFRG SSYLNNRSKG NDTNDISDTN PHKEDLEGDN S-Y--N---- 201 IISGSENYGD FLSDSNNSNE TNDTNDIYG. PWHWFHQIYP --S-S-NYG- 251 LEVTQKYRHL FDITKNYNQL PNCMEKLARL ASLKSGYAHL ---T-KYAHL NRGQENSMER ~~~~~~~MEK NEDPSSLTDM ~~MKPTKPEG N ------ ME- NIHS...W.. DMDKNFYG.. DSNSN..... IIRKTLKGIK DI--N--c—- DSYSIYFDIE NSYSIYFDIE NINR.CIESY DDEYEEDLE. -SYSIYFDIE EDPYYN...S DNDLHYD..P DISDTNDTND KPHKEDLEGD D----ND--D SYIYRAVCGG SYIYNFICS. .IYDPDDIYE KHWGCPHRKH SYIY---C-- WVQCENCYGL WIQCDNCYGL WVQCETCYGL WVHSTLCSGF WVQCENCYGL WWFNSMLFKK SWFNLMFSKG DNNIDSWKNN PKKPNKSNEK -WFN-M--KK ....RNRDNS I I O .WGERSS SLTS.LTRKR ------- R-S NHIFE.IDND KKKFE.INND LRSQICEDID .YDDEYEEDL ---FE--DND YMYDTQYSWN YIKDTKYNCN TNAIYDPFDI NKPHKEDLEG Y--DT-Y--N ESRNSSENEG ESGSGKIRES TNIKDICERY RDRKSVPAKE ESRKS---E- NYKKFLKSKM KYK...KVEM NFKQFFRPKM NYKKIMKSKN NYKKF-KSKM EFERRCGLNK ELEYRGELSK SENSSYSHAD RKSLKGIIRK E-E-RG-L-K SCSNVDYLFG YYNNVDLLVN ...IY KSLKGISRRI --—NVD-L-- HSFLSELESS ...LSDLEIF FNSDICDKVQ EYDDEEYDDE ---~SELE-- NHINSCIDSY NHINSCIDSY SDTNDTNEIY DKQKEEEEKQ NHINSCI-SY SSRRTRTKGS KNDKIRTN.S SEIYPRNREK RELVP..AQS SE--PRT--S NICEQCGYHL NVCEECGHYL NICEHCGEHL NVCEQCGSHL N-CEQCG-HL 50 SMGSLGPIEN AMDSFAPIEK SLADVSNIDN TLKGLKSLTS S--SL-PIEN 100 VKDIRNFISD SKDIRNFISD DIYYAYDTND RKMGFSYSIM -KDIRNFISD 150 FYSY....RN FYSY....CS RTIIILIRST YEEDLEGDNK FYSY ------ 200 LQSQICIDTS FRSHICINSH DPFFILDIND KEEEEKLKDC --S-ICI--- 250 DLTIRESSND NRNNLMSSKD STFVPIDYSD TKRDTDPDSE ------ SSSD 300 KMSSSDRIEL KMTSSERIEL KMSSSDRIDL EMHSSDRIDL KMSSSDRI-L Figure 4.6. Protein alignment and consensus of selected Ach genes. The Oenothera Ach genomic sequence was translated using GCG and aligned with predicted proteins from tobacco, B. napus, and pea. The protein alignment was based on one done by L. Savage as well as the GCG program, pileup. The repeated region of Oenothera corresponds to amino acids 1 to 181. 137 tobacco} B. napus} Pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} tobacco} B. napus} pea} Oenothera} Consensus} 301 LIDPGTWDPM SIDPGSWNGM SIDRDTWNPM MLDPQTWAPM SIDPGTWNPM 351 LDSEEESSKD PLFEDIMNQL 451 GIGQLNGIPV GTGQLNGIPV GTGQLNGIPL FMGGSMGSVV GTGQLNGIPV 501 ASGGARMQEG SSGGARMQEG ASGGARMQEG KISPASSDYQ ASGGARMQEG 551 FGMLGDIIIA LGMLGDIIIA FGMLGDIIIA QVKE*~~~~~ FGMLGDIIIA 601 IVPRNLLKSV IVPRNPLKGV VVPRHLLKEF ~~~~~~~~~~ IVPRNLLK-V Figure 4.6, continued. DEDMVSLDPI DEDMVSADPI DEDMVSVDPI HEGLLSLDPI DEDMVSLDPI DLYLQTAKNR ..YKDRVASY __Y ....... AIGVMDFQFM ALGVMDFQFM ALAVMDSEFI GERITRLDEY ALGVMDFQFM SLSLMQMAKI SLSLMQMAKI SLSLMQMAKI FNRTVFYVAL SLSLMQMAKI EPNAYIAFAG EPYAYIAFAG EPNATIAFAG ~~~~~~~~~~ EPNAYIAFAG LSELFKLHAF VSELFQLHAF LTELFQFHGF ~~~~~~~~~~ LSELFQLHAF EFHSEEEP.. KFHSREEP.. KFDSIKELGS EFHSEKDP.. -FHSE-EP-- SYIDRLDSYQ VDSEEEKDQS KRKTGLSEAI GGSMGSVVGE GGSMGSVVGD AGSMGCVVGB ATNRVLPLIL -GSMGSVVGE SSALYDYQLN SSVLCDYQSS SSALYNYQIN LTSPTTGGVT SSALYDYQ-N KRVIEQTLNK KRVIEQTLKK KRVIEQLLNK ~~~~~~~~~~ KRVIEQTLNK 138 EKTGLPETVQ YKDRIDSYQR YKKRIASAQK YIDRLDSYQE HTGRGYLKRI YKDRIDSYQ- KITRLIEYAA KITRLIEYAT KITRLIEYAT VCASGGARMQ KITRLIEYAT KKLFYVSILT KKLFYISILT QKLFYVAILT ASFGMWGISH KKLFYVSILT TVPEGSQAAE AVPEGSQAAE EVPEGSQSAD ~~~~~~~~~~ -VPEGSQAAE ..PLNQKSSK ..PLNKNEIK ..PLT*~~~~ ~~~~~~~~~~ --PLN----K 350 TGTDQREEIH 450 KTGLTEAVQT KTGLTDAIQT KTGLPEAVQT HLGIGLMDFQ KTGLTEAVQT 500 NQILPLIIVC NQCLPLILVC NLLLPLIIVC EGSGSLMQMA NQ-LPLIIVC 550 SPTTGGVTAS SPTTGGVTAS SPTTGGVTAS KKNLMXTSHS SPTTGGVTAS 600 YLFQKGLFDL SLLRKGLLDA LLFDRGLLDA ~~~~~~~~~~ -LF-KGLLDA 643 IK* *~~ PstI clones on outer ring Bam3B rbcL, SaII clones on psb A inner ring 1 Bam2 Bam3B(A54) SalE SalI . PstG .' .. _ ‘ up i. ‘ 1W6; zo‘.‘ 2'13.- u.,.,..—r ..*¢.~‘«- . . v.4 . 's,‘hr~m haul-mu: Ll.» 11>» Figure 4.7 0. hookeri plastome maps and the location of the Ach gene. Clones from the plastidial genome of 0. hookeri were obtained from B. Sears, DNA prepared and blotted to membrane. The membrane was probed with an expression vector (from Linda Savage) containing the tobacco Ach coding 139 Discussion Plant transformation and extra-plastidial expression of Ach It is clear from the immunoblots and northern blots that moving the plastidial B-CT gene to the nuclear genome is not a straight-forward project. Although the efficiency of transformation was disappointing, especially with the pPZP-derived vectors, it is clear from the seed germination and NPT assays that the plants were able to incorporate single and multiple copies of foreign genes into their nuclear genomes. Thus, the plants were transformed. Both plant transformation vectors had the drug resistance genes adjacent to the left border which is transferred to the plant after the right border, therefore, the constructed Ach genes were transferred to the plant genome. In addition, the use of two different constructs, one based on the tobacco B-CT gene, the other based on the pea B-CT gene, decreases the probability that some non-conserved region is responsible for the lack of expression. Why did this project not yield detectable expression of Ach when it was located in the nuclear genome? The tobacco Ach gene has an AT content of 63% while the pea Ach AT content is 65%. The average dicot AT content is 55% (Sinibaldi 1992, reviewed in Diehn et al., 1996). Thus the Ach transcript may contain instability motifs and polyadenylation signals which are AT-rich and which prevent its successful translation in a nuclear background. Problems of this sort have been well documented for the Bacillus thuringiensis toxin genes (for review 140 see Diehn et al., 1998). Figure 4.8 highlights potential problem areas in the .4ch sequence. They are numerous in both tobacco and pea. While none of the regions indicated are known to be sufficient to cause incorrect processing and translation by themselves, the sheer number of problem areas suggests that relocation of .4ch to the nuclear genome may require substantial modification of the gene. In addition to these instability motifs and polyadenylation signals, the codon usage of the pea and tobacco Ach genes are not plant-like. Plant genes have a G or a C in the third position about 40% of the time (Wada, 1992), while the pea Ach gene has an G or a C in the third position 28% of the time and tobacco 32% of the time. Although the most likely explanation of the lack of B-CT protein is due to incorrect transcript processing or stability, there is also a small possibility that the B-CT DNA from the vectors was not stably incorporated into the plant genome. The techniques used to confirm this are generally Southern blotting or PCR (Register, 1997). This technique was not used on the tobacco plants because the gene is still present in numerous copies in the plastidial genome. While it is possible for occasional gene rearrangements or incomplete transfer of vector DNA to the plant genome, it is highly unlikely that every construct with both vectors chanced upon this problem, so if B-CT was rearranged out, it is likely to be due to a conserved region of the sequence. 141 Figure 4.8. Areas where the Ach transcript may be improperly processed. The promoter (Leire and Maliga, 1999) is italicized. Start and Stop codons are in boldface. Sequence prior to the ATG start codon was not included in the construct, thus no potential problem areas in that region are marked. ATTTA instability motifs which may cause premature degradation (Ohme-Takagi et al., 1993) are in boldface and underlined. Near upstream elements (NUE’s) are used as polyadenylation signals 40% of the time they are present. One mismatch with the consensus sequence has little effect on polyadenlyation efficiency (Rothnie, 1996). Although NUE’s are difficult to find via 5 uence inspection, potential NUE’s and NUE’s with one mismatch are shaded and lightgray respectively. The large box indicated by the dashed line indicates the region where the protein is conserved. 142 CA TA TTAA TA TA TAG TGAAG TAC T CACAAT AG T C T GATAG C C T TAT TAG T GAAATAAGAT TATTCATCTA TTGTATTTTC AAAGATGGTG TGTGGGCT ‘ Jigsaw IIIIGATCCA, ACAATTCTAG GGAATTTCAT AGTTATTCCA CGATCATTCT ACTCGAAT TA TCTTACATGT CAT TGATAGT GTGGTAGTGA GGTGAAAGTC CACGAAGGGC GTTTAATTCG AATCGAAAAG TTGCAGTAAT CTCTGATGAC TCTATTTTGA TTTCTGAGTG TCGGAATAAT ATGATACTCA TATCTTCAGT GAATTACGGT GAAATAGTAG AGTGATTTAA TGTTCCCCCA TAATAATCCT ATTCAAATAA ATGCAAATAG ATGTTGTTTA GGGCAGTCTT TGAAAAACAT GATTCAGAAC AGTGATTGGA ATAATTTTAT GGGGchQAA AGAAGGAGTT GGTCCTATTG” TCATAGTTGG GTTGAT TATT ACTTTTTTAG TATTGAAAAT AACTAGAAAG (GGATTTAGGG ATATAGTTGG CTCAAATCTG GACAGT TACA TGAAAACGAG TATTCGGCGT TTAGTGATAG CATATTTTTG TTCTTTTTAT GCGAAGATCC nagaagCACA TATAGATACT TTTATAGGGC GGTTCCAGTA TTTTTTTCAA TTTTATGCTT AGCGAAJT’CAC _AACTCTATGG CGAACGCAGG AAAATACC'I AGGAATCGTG TAAGACATTC GAATGGAGAC AGATTGACAA AGTTATCGAA CTACTATAAT TTAATAGTTG TCCATTATAA CGTTTGTGGT GACGAACTCG CTATAAGAGA TAACTCAAAA TTAAATTATA TGGATATCAT ATCCGGGTAC AITGAATTTC TCAAAGAAAG TAAACGGCAT GGTAGTATGG GCGCATGCAG CTGCTTTATA CTTACATCTC GGATATCATT GAGTAATTGA GCTGAATACT TCTTTTAAAA CGCTGCCAAT. CTTTGAATCA. ATACAGGCAT .AGAAATTTTT TTGAAAATGA TTGGGATCCT ATTCGGAGGA ACAGGATTAA TCCCGTAGCA GATCCGTAGT CAAATTTTAC GAAGGAAGTT TGATTATCAA' CGACAACTGG ATTGCCGAAC ACAAACAT T G ' TATTCCAGAA AGCGTTCTGA AAAGTCAAGC TTGTGGGTTC AAGTTCTAAT AATGCGAAAA GAAATCWTATTT GTAGTTCAGA CATGQATGAAG GGAGCCTTAT CCGAGGCTGT ATTGGGGTTA CGGAGAGAAA CCCTTATTAT TGAGCTTGAT Mm T G GAG T GACA TAGAATTGAA ACATGGTCTC ,AAAGATCGTA TCAAACAGGC TGGATTTTCA ATCACCCGTT AGTGTGTGCT GCAAATGGCT AGTTATTTTA GCTAGTTTTG CCAACGCCTA CATTGCATTT IIEIIIIACAG GGGTTTATTC TACCCGAAGG GACCTAATTG TAAGCTCCAC GTGAGTTATT AAAATCAAGT GAICTCGAGG TTGTTATGGA GTGAACAATG CTTTTGATCG TCTAGATCCC TTGATTCTTA ATAGGCCAAC GTTTATGGGG TGATTGAATA TTGGGGGGGC AAAATATCGT TGTATCAATC GTATGTTGGG GCAGGTAAAA TTCACAAGCA TACCACGTAA GCCTTTTTTC AG 143 Oenothera Ach location on the plastidial genome and sequence In the plastid genomes which were sequenced at the time this work was done (tobacco, rice, black pine, liverwort), Ach is close to rbcL, usually separated by one tRNA gene or ORF, and the Ach gene precedes the petA gene. However, in the 0. hookeri plastome, rbcL and petA are at opposite ends of the large single copy region. Ach is still next to petA, but near atpA and far from rbcL. The gene order is given in Table 4.1. The region around Ach appears to have been an area where the chloroplast genome could recombine and rearrange. While the 0. hookeri line used in these studies was a not a line with an active plastome mutator (PM), in the evolution of the four different plastidial genomes found in Oenothera repeat amplifications are common (B. Sears, personal communication; Sears et al., 1996). As seen in Figure 4.5, there are many direct repeats in the 5’ region of Ach. While the significance, if any, of these repeats is not known, the appear to not be tolerated in the 3’ region of the gene. Perhaps ancestral rearrangements which occurred in this region were not viable. The DNA sequence and protein alignments do little to clear up confusion - in fact, they raise more questions. Although the carboxy end of the protein is more conserved than the amino end, the putative Oenothera protein does not contain some of the conserved motifs. For example, the first putative phosphorylation site is missing, and the carboxy biotin binding site is poorly conserved. Is the 144 Table 4.1. Gene order in plastidial genomes. N. tabacum P. thunbergii 0. sativa M. polymorpha (tobacco) (black pine) (rice) (liverwort) rb cL rb cL rb cL rb cL trnR 0RF133 trnR Ach Ach ORF] 06 ORF 31 6 (A ch) (Ach) 0RF184 0RF119 0RF36 0RF36 ORF 228 pm] 0RF185 0RF184 petA 0RF184 0RF230 ORF 434 0RF261 petA petA petA 145 Oenothera protein longer than most? Could the protein be truncated? Although the earliest start site is at base 319, the sequence presented here differs from unpublished sequence generated elsewhere from the same clone (R. G. Herrmann, Ludwig Maximilians University). This difference is due to a difference in the repeated region.0ne of the imperfect repeats presented in Figure 4.5 is a perfect repeat in the Herrmann et al. sequence. This difference causes a frameshift. Thus, Herrmann et al. assign the start codon to base 556 (B. Sears, personal communication). This difference may be due to sequencing CI’I'OI'S . Analysis of the regions 5’ to the open reading frame in search of a promoter region does not shed more light on the start codon or length of the Oenothera transcript. There are two different plastidial RNA polymerases, one nuclear encoded (NEP), the other encoded by the plastidial genome (PEP) (Allison et al., 1996; reviewed in Weihe and Berner, 1999). The PEP uses a eubacterial-type promoter and primarily transcribes photosynthetic genes. The NEP usually transcribes Ach, and the promoter from tobacco has been described (Leire and Maliga, 1999). The tobacco Ach promoter is a class Ia NEP promoter. Due to the high GC content of the repeated motifs, there are few truly AT-rich regions in this area. The tobacco Ach promoter has an AT content of 85%. In the Oenothera sequence there are many YRT motifs (Y = C or T, R = A or G) in the region 5’ to the potential start codons. The AT content of the surrounding bases ranges from 65% to 30%, so no region strongly suggests itself as a promoter. 146 ..-1-. L... Based on the discussion above, some of the traits of Ach which may interfere with relocation of the Ach gene do not appear to be confined to any one species. All of the genes are AT-rich and there is little conservation in the 5’ region. All the species discussed in this chapter have retained a functional Ach gene in their plastidial genomes. However, neither the tobacco nor the pea gene was successfully relocated, and the Oenothera gene has a unique 5’ end and a possibly unique location on the plastidial genome. The true significance of these differences and gene sequences are unknown; however, they may have a significant bearing to the ability of Ach to be easily engineered. Acknowledgements Linda Savage for antibodies and assistance, Sue Stoltzfus for sequencing advice, Dave Shintani for plasmids and technical assistance, Lara Stoike Steben and Gabi Schewe for technical assistance, Mark Johnson for RNA stability discussions, Barbara Sears for guidance, clones, and sequence data. 147 Literature Cited Allison, L. A., Simon, L. D., and Maliga, P. (1996). Deletion of rpoB reveals a second distinct transcription system in plastids of higher plants. EMBO J. 15, 2802-2809. ' Diehn, S. H., E. J. De Rocher, and, Green, P J. (1996). Problems that can limit the expression of foreign genes in plants: lessons to be learned from B. t. toxin genes. Genetic Engineering. 83-99. Elhai, J., and Wolk, C. P. (1988). A versatile class of positive-selection vectors based on the nonviability of palindrome-containing plasmids that allows cloning into long polylinkers. Gene 68, 119-138. Fisher, D. K., and Guiltinan, M. J. (1995). Rapid, efficient production of homozygous transgenic tobacco plants with Agrobacterium tumefaciens: a seed-to-seed protocol. Plant Mol. Biol. Rep. 13, 278-289. Hajdukiewicz, P., Svab, Z., and Maliga, P. (1994). The small versatile pPZP family of Agrobacterium binary vectors for plant transformation. Plant Mol. Biol. 25, 989-994. Knutzon, D. S., Thompson, G. A., Radke, S. E., Johnson, W. B., Knauf, V. C., and Kridl, J. C. (1992). Modification of Brassica seed oil by antisense expression of a stearoyl-acyl carrier protein desaturase gene. Proc. Natl. Acad. Sci. USA 89, 2624-2628. Kutzelnigg, H., and Stubbe, W. (1974). Investigations on plastome mutants in Oenothera. 1: general considerations. Sub-Cell. Biochem. 3, 73-89. Liere, K., and Maliga, P. (1999). In vitro characterization of the tobacco rpoB promoter reveals a core sequence motif conserved between phage-type plastid and plant mitochondrial promoters. EMBO J. 18, 249-257. Martin, W., Stoebe, B., Goremykin, V., Hansmann, S., Hasegawa, M., and Kowallik, K. V. (1998). Gene transfer to the nucleus and the evolution of chloroplasts. Nature 393, 162-165. Mekhedov, S., Martinez de Ilarduya, 0., and Ohlrogge, J. (1999). Towards the functional catalogue of plant genes: a survey of genes for glycerolipid biosynthesis. In preparation. Nikolau B., J., and Wurtele, E. S. (1998). In situ and transgenic studies of the regulation of acetyl-CoA carboxylase gene expression. In Advances in Lipid Research, J. Sénchez, E. Cerda-Olmedo, and E. Martinez-Force, eds. (Sevilla: Secretariado de Publicaciones), pp 50-53. 148 Post-Beittenmiller, M. A., Schmid, K. M., and Ohlrogge, J. B. (1989). Expression of holo and apo forms of spinach acyl carrier protein I in leaves of transgenic tobacco plants. Plant Cell 1, 889-899. Register, J. C. III. (1997). Approaches to evaluating the transgenic status of transformed plants. Trends Biotechnol. 15, 141-146. Sambrook, Fritsch, and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual (New York: Cold Spring Harbor Press). Sasaki, Y., Konishi, T., and Nagano, Y. (1995). Update on biochemistry: the compartmentation of acetyl-Coenzyme A carboxylase in plants. Plant Physiol. 108, 445-449. Schmidt, G. W., and Mishkind, M. L. (1983). Rapid degradation of unassembled ribulose 1,5-bisphosphate carboxylase small subunits in chloroplasts. Proc. Natl. Acad. Sci. USA 80, 2632-2636. Sears, B. B., Stoike, L. L., and Chiu, W. (1996). Proliferation of direct repeats near the Oenothera chloroplast DNA origin of replication. Mol. Biol. Evol. 13, 850-863. Shintani, David K. (1996). How plants manage their fatty assets: a study into the organization and regulation of the plant fatty acid biosynthetic pathways. Dissertation, Michigan State University. Shintani, D., Shorrosh, B., Roesler, K., Savage, L., Kolattukudy, P. E., and Ohlrogge, J. (1996). Alterations of tobacco leaf fatty acid metabolism using antisense-expression and reverse genetic engineering approaches. Presented at 12‘h Int. Symp. Plant Lipids. Toronto. Wada, K., Wada, Y., Ishibashi, F., Gojobori, T., and Ikemura, T. (1992). Codon usage tabulated from the GenBank genetic sequence data. Nucleic Acids Res. 20, Supplement, 2111-2118. Weihe, A., and Biirner, T. (1999). Transcription and the architecture of promoters in chloroplasts. Trends Plant Sci. 4, 169-170. 149 , . :fi'ms-ar-cssmflfl; . Chapter 5 CONCLUSIONS AND FUTURE DIRECTIONS The work described herein used a variety of methods and approaches in an attempt to understand the regulation of FAS and to develop strategies to manipulate FAS. Even as the amount and types of fatty acids made and stored in a plant are being genetically engineered and grown as crops (e.g., Facciotti et al., 1999), the work presented in Chapter 2 demonstrates that physiologically important in vitro biochemical properties of individual enzymes can still be discovered. For example, in Chapter 2, new information about the metabolic regulation of ACCase was presented: in a lysed chloroplast system ACCase is activated by its substrate, acetyl-CoA. This activation occurs between 10-100 uM acetyl-CoA. Since the concentration of acetyl-CoA in the chloroplast is approximately 10-20 uM, this activation may be physiologically relevant. Less is known about the in vivo behavior of individual enzymes, let alone enzyme systems or complexes. Chapter 2 also provides more fodder for the sometimes-forgotten truism that laboratory experiments done in test tubes with artificial systems do not automatically transfer to the living organisms under study. While previously reported work (Sauer and Heise, 1984; Sasaki et al., 1997) clearly present results that DTT and thioredoxin stimulate ACCase, these results could not be 150 duplicated in a lysed chloroplast system when ACCase was activated (Chapter 2). It is likely that both thioredoxin and acetyl-CoA activate ACCase, but the effect is not additive. Given the data of Sasaki et al. (1997), it appears that both thioredoxin and acetyl-CoA can activate ACCase to a similar extent. It is not surprising that there are multiple methods of activating ACCase. It is, after all, a major determinant of the rate of F AS (Post-Beittenmiller et al., 1992). In view of the results presented in Chapter 2 (Table 2.1); namely that in vitro results do not accurately reflect in vivo conditions, the computer simulation approach presented in Chapter 3 may initially appear overly simplistic. This is true: a modeling approach is simplistic if built without in vivo data or followed up by in vivo experiments. In Chapter 3, a model of the major plastidial reactions of FAS and its predictions were presented. This model was validated using the non-intuitive results of KAS III overexpression. In vivo experiments in spinach show that KAS III overexpression alters the 16 carbon to 18 carbon fatty acid ratio (Jaworski and Hinneburg-Wolf, 1998). Although the quantitative values of in vivo results and model predictions differ, the model qualitatively predicts the in vivo result. In the artificial and simpler system of the model, the behavior is easy to dissect - the results are due to increased competition for malonyl-ACP by KAS III, KAS I and KAS 11. Further validation of the model was provided by comparison of model predictions with in vivo pool size changes during a light/dark transition and other antisense/overexpression experiments. 151 Data presented in Chapter 3 also questions some assumptions that workers have made about the regulation of F AS. Many workers have reasonably assumed that the regulation of 16:0 and 18:1 production is essentially similar and they have based the interpretation of their results on this assumption, either consciously or unconsciously. The model does not make this assumption, and in fact, predicts that the production of the two fatty acids has different co-limiting steps. Thus, the model has brought out an assumption we have been using. Which is correct: the assumption or the model? We don’t know yet. The model presented in Chapter 3 also predicts that the most efficient way to increase FAS would be to overexpress three enzymes in conjunction with each other - ACCase, KAS III, and KAS I. KAS II, the enzyme which does the final elongation from 16:0 to 18:0, is apparently not required. Although this is a logical result, to our knowledge it has not (yet) been attempted. Finally, Chapter 4 presents two unsuccessful attempts to accomplish a seemingly simple genetic engineering project - relocation of the plastidial Ach gene to the nucleus. This is not novel; the plastidial gene which codes for the large subunit of RuBisCO has been moved to the nuclear genome (Kanaveski and Maliga, 1994). But evidence presented in Chapter 4 indicates that the Ach transcript is not functional in a nuclear background. Thus, while the data presented in Chapter 3 encourages efforts to overexpress ACCase, this may be technologically difficult for the heteromeric ACCase. The plastidial ACCase is comprised of four subunits, and the homomeric ACCase by itself increases oil yield 5% (Roesler et 152 .000 ‘ \ W— al., 1997). Chapter 4 also presents the sequence of Oenothera hookeri Ach. This region is a hotspot for recombination in 0. hookeri (B. Sears, personal communication) and Ach has been deleted from the plastidial genome in many lineages (Martin et al., 1998). Furthermore, mRNA instability motifs and polyadenylation signals may exist in the coding region. These issues make relocating the Ach gene as outlined in chapter 4 risky proposition. Future Directions The questions that started these research projects, “what regulates FAS in a plant?” and “how can we alter F A composition, quality, and yield in a plant?” are still unanswered, although progress has been made. Answering these questions is imperative if we are to realize the vision of biological plants as large scale special purpose chemical factories (Ohlrogge, 1994; Brown, 1996; Ohlrogge 1999; Kinney, 1998). Therefore, future research should focus in this direction. One of the questions involves very basic research and the other question very applied research. The knowledge gained from probing either question suggests solutions for (or more often, invalidates approaches being tried by) the other approach. How does a plant control how much and what types of fatty acid are synthesized? What regulates plant FAS? Some pertinent questions are: how many different FAS pathways exist, and are they regulated in the same manner? Specific ACP’s 153 (Suh et al., 1999), ferredoxins (Schultz and Ohlrogge 1998), desaturases (Cahoon et al., 1992) and thioesterases (Pollard et al., 1991) exist for the synthesis of unusual fatty acids. This is ample evidence that unusual fatty acids are synthesized by a different set of enzymes, but it is unknown how much overlap exists between the set which produces normal fatty acids and the set specific to each different unusual fatty acid. Unusual fatty acids are excluded from membranes by some unknown mechanism and found almost completely in seed oils. While it is possible that plants which make unusual fatty acids have specific pathways m for the unusual fatty acid and another pathway for 18:1 and 16:0 production, another scenario is that fatty acids destined for membranes are synthesized by a separate pathway or system than fatty acids destined for storage triacylglycerols in all plants. Another alternative is that the separation occurs in the cytosol, outside the chloroplast, and fatty acids are funneled into either membranes or oil bodies. In any case, there must be some manner of sensing and signaling demand for, or supply of, fatty acids. When MCTE is overexpressed, laurate is synthesized and degraded, but due to presumed lack of 18C fatty acids, overall fatty acid production is induced (Eccleston and Ohlrogge, 1998). How will this knowledge help to manipulate the products of FAS? If enzymes are specific to membrane, oil biosynthesis, or housekeeping functions, they are essentially in different pathways. This is important information when attempting overexpression experiments because overexpressing an enzyme outside of the pathway targeted for manipulation is unlikely to yield productive results. For 154 example, if the role of homomeric ACCase is to provide malonyl-CoA for further elongation of fatty acids, phytoalexins, and fiavonoids (Sasaki et al., 1995), this enzyme may not respond to signaled demand for fatty acids destined for membranes or TAG. On the other hand, if enzymes are not specific to storage or membrane pathways there are fewer to find and target. This question will not be easy to answer at this point in time. If a set of enzymes which make unusual fatty acids can be localized in the chloroplast and that localization is different than the set of enzymes which make common fatty acids, this is a good indication that the pathways are separated. This would require specific antibodies or other probes to many enzymes, some of which may be undiscovered. These probes would have to be labeled differently for common and unusual FAS. However, insight to the question may be received by further investigation into one of the lipid field‘s hotly debated topics: channeling. Is FAS channeled? Currently, of the three intermediates in the reduce/dehydrate/reduce cycle, only hydroxyacyl-ACP’s have been detected (Roughan, 1998, personal communication). This has led Roughan and others to argue that FAS is channeled. Evidence for this hypothesis consists of l) chloroplasts permeabilized with low concentrations of Triton X-100 lose RuBisCO activity, but retain FAS competence (Roughan and Ohlrogge, 1996); 2) in vivo concentrations of CoA and its esters are too low to support in vivo rates of FAS (Roughan, 1997); 3) the 3-ketoacyl-ACP and 2-enoyl-ACP intermediates have yet to be detected 155 (Roughan, 1998, in preparation); 4) both ACP (Slabas and Smith, 1998) and ACCase are membrane associated (Sasaki et al, 1993; Shorrosh et al., 1995); 5) estimates of the concentration of the stearoyl-ACP desaturase enzyme exceed the known concentrations of stearoyl-ACP (Mehkedov et al., in preparation); finally, 6) data in Chapter 2 clearly demonstrate that an in vitro ACCase assay with in viva-like conditions will not support the known in vivo rate of FAS. While none of the lines of evidence are definitive, they all suggest that FAS may, in fact act as a complex in vivo. Channeling can occur in both stable and dynamic forms (e.g. Cornish-Bowden, 1997). A stable complex is, of course, easier to find than a dynamic one. In dynamic channeling, an enzyme-substrate (El-S) complex diffuses to meet a second enzyme (E2) faster than the free intermediate product diffuses to meet E2. This can occur if the concentration of El-S is much higher than the free intermediate product (Cornish-Bowden, 1997). If enzymes remain membrane- associated, this is not far-fetched. Two membrane-associated proteins would diffuse in two-dimensions, over the m of a membrane, while a soluble substrate, or protein-substrate complex, such as an acyl-ACP, would diffuse in three dimensions, in the mug; of the stroma. Furthermore, some FAS enzymes, KAS III and perhaps a-CT, require detergent for maximal recovery during extraction (Gulliver and Slabas, 1994; Shorrosh et al., 1996). If channeling exists, it could be the mechanism by which fatty acids destined for storage products and for membranes are segregated. 156 Light/dark regulation of FAS The mechanism of light/dark regulation and fatty acid synthesis in the light and dark needs to be further examined. ACCase activity is lower in the dark than the light (Chapter 2). Are other enzymes of FAS also influenced by light? If only ACCase is affected, it would be wise to look into the specific mechanism and target the specific mechanism. If other enzymes also have higher activity in the light than in the dark, then the light-regulation is likely to be a global mechanism or signal. Recent evidence where l3C02 was used to label fatty acids (Bao et al., in preparation) indicates that the rate of fatty acid synthesis in the dark is essentially zero. However, while the demand for fatty acids may decrease in the dark, it, does not disappear - plant cells still expand. Are new lipids for membrane biogenesis needed for expansion and if so are they stored? Another approach for manipulating metabolism is to find and alter transcription factors. This approach has worked in maize cell cultures where three transcription factors were overexpressed (Grotewold et al., 1998). Over- expression of two of the transcription factors led to an increase of one set of fiavonoid products, overexpressing the other led to an increase of a different flavonoid product. In another recent example, Thomashow and coworkers overexpressed a transcription factor for cold acclimation which led to an increase in the expression of genes involved in cold-acclimation (Jaglo-Ottensen et al., 1998). There are two tasks to accomplish before this method can be applied to 157 fatty acid synthesis. First, the appropriate transcription factor is necessary, and none have yet been identified. More importantly, neither of the other pathways were primary metabolic pathways. Secondary pathways likely have fewer interconnections to other pathways and may be more amenable to metabolic engineering. There are numerous connections between carbon and nitrogen metabolic pathways. Genes for nitrogen metabolism are directly and indirectly controlled by sugars (Coruzzi, 1999) and carbon and nitrogen metabolism are extensively linked (reviewed in Huppe and Turpin, 1994). Although largely unexplored, interconnections between FAS and other metabolic pathways may exist and may need to be controlled. In spite of our metabolic goals when engineering an organism, the organism will generally attempt to optimize its survival and reproduction, regardless of our goals (Varner and Ramkrishna, 1998). Thus, in order to successfully manipulate F A metabolism, discovering the nature of interconnections between FAS and other pathways and deciphering the resultant interconnections and signals may be necessary. A more complete computer simulation model could help to address a number of the above issues and increase progress in the manipulation of FAS products and quantity. The model presented in Chapter 3 is immature. It lacks one control point we know exists: feedback on ACCase (Shintani and Ohlrogge, 1995). It does not incorporate cofactor limitations of NADPH and NADH. The concentration of NADH is not significantly different in illuminated and non- illuminated chloroplasts, but the concentration of NADPH in a non-illuminated 158 chloroplast is half the concentration of NADPH in an illuminated chloroplast (Winter et al., 1994). These factors, if incorporated into the model, could affect the model’s predictions. Furthermore, if connections to other metabolic pathways are made, competition for cofactors could substantially alter model predictions. Any model, at best, can only be as good as the data of which it is comprised. The model is currently imprecise and only qualitatively matches reality. Precision can be increased by improving the biochemical inputs to the model: specifically, re-checking Km and V“...x values, and confirming that in vitro data has realistic physiological meaning. To increase precision, work on detecting changes in pool sizes should also continue. Current methods to determine concentration are based on lengthy TLC and HPLC methods as well as evaluation of product partitioning via immunoblotting (e. g., Roughan 1998, in preparation; Post-Beittenmiller et al., 1991, 1992). A method has recently been developed to rapidly and concurrently measure metabolites in crude rabbit extracts using capillary electrophoresis (Dillon and Sears, 1998). If the precision of measuring metabolite pools can be increased by either increased throughput or a new method, the quality of the model inputs and, therefore, outputs can be improved. Models should be insensitive to unavoidable unknowns. The model in Chapter 3 has not been tested for sensitivity to the accuracy of kinetic values. This can probably be easily remedied. GEPASI is another software package for modeling (Mendes, 1993). It was designed for use in biochemical networks - precisely the 159 application which was described in Chapter 3. GEPASI was not used to develop the Chapter 3 model because it requires the biological system to be in a steady state before simulation. Furthermore, it is difficult to build a model in incremental stages with GEPASI. Due to those limitations, Stella II was used to build the model described in Chapter 3. But now that the model is built it could be moved to GEPASI. GEPASI has the ability to alter parameters, such as K.,l and vmx in a sequential series of simulations. This feature can be used to test whether or not the model’s predictions vary with changes in these unknown parameters. To be an efficient prediction tool, the model needs to incorporate as much experimental data as possible. The changes above should be incorporated. The model can also be rebuilt in several ways: alternative control pathways can be incorporated, and model predictions which are different between alternatives should be tested. The genetic expression experiments proposed by the model should also be investigated. Will biotechnology be able and allowed to actually help? A final area where further work needs to be done is in the acceptance and public perception of and knowledge about biotechnology. The direction and discussion of the work presented here and its usefulness presume that transgenic plant products will be accepted by the public and their governments. This is not a given. There have been unintended consequences from some applications of 160 biotechnology. Transgenic canola comprises 70% of the crops grown in Canada (Ohlrogge, 1999), but the Roundup Ready® gene has escaped to wild mustard, another member of the Brasssicaceae family (R. Allison, personal communication). Pollen from corn expressing the B. thuringiensis toxin is lethal to Monarch butterflies (Losey et al., 1999; Milius, 1999). Although tomato paste from transgenic tomatoes was marketable in the United Kingdom, public acceptance of transgenic products in that country is low (Gavaghan, 1999). There is great concern and debate about the escape of genes used to make crops resistant to herbicides and pests (e.g. Roundup Ready® soybeans and/or B. thuringiensis toxin producing plants), However, technology such as the “terminator” gene, which not only makes the plant sterile but also makes gene escape highly improbable, is also publicly unacceptable. (Service, 1999) In addition, there is debate about whether or not biotechnology will increase or decrease sustainability (e.g. Crouch, 1995; Zechendorf 1999). These issues of societal (mis)understanding of biology, what it can and can not accomplish, and the converse issue of scientists not understanding or accepting societal concerns are important future considerations. 161 Literature Cited Bao, X., Focke, M., Pollard, M., and Ohlrogge, J. B. Understanding in vivo carbon precursor supply for fatty acid synthesis in leaf tissue. In preparation. Brown, K. S. (1996). Life on the molecular farm. BioScience 46, 80-83. Cahoon, E. B., Shanklin, J., and Ohlrogge, J. B. (1992). Expression of a coriander desaturase results in petroselinic acid production in transgenic tobacco. Proc. Natl. Acad. Sci. USA 89, 11184-11188. Cornish-Bowden, A. (1995). Fundamentals of Enzyme Kinetics. (London: Portland Press). Coruzzi, G. (1999). Amino acid signaling in central metabolic pathways. Presented at Plant Biol. and Mol. Biol. Symp.: Metabolic Networking in Plants, Ames, IA. Crouch, M. L. (1995). Biotechnology is not compatible with sustainable agriculture. J. Agric. and Environ. Ethics 8, 98-111. Dillon, P. F., and Sears, P. R. (1998). Capillary electrophoretic measurement of tissue metabolites. Am. J. Physiol. - Cell Physiol. 43, C840-C845. Eccleston, V. S., and Ohlrogge, J. B. (1998). Expression of lauroyl-acyl carrier protein thioesterase in Brassica napus seeds induces pathways for both fatty acid oxidation and biosynthesis and implies a set point for triacylglycerol accumulation. Plant Cell 10, 613-621. Facciotti, M. T., Bertain, P. B., and Yuan, L. (1999). Improved stearate phenotype in transgenic canola expressing a modified acyl-acyl carrier protein thioesterase. Nature Biotech. 17, 593. Gavaghan, H. (1999). Britain struggles to turn anti-GM tide. Science 248, 1442-1444. Grotewold, E., Chamberlain, M., Snook, M., Siame, B., Butler, L., Swenson, J., Maddock, S., St. Clair, G., and Bowen, B. (1998). Engineering secondary metabolism in maize cells by ectopic expression of transcription factors. Plant Cell 10, 721-740. Gulliver, B. S., and Slabas, A. R. (1994). Acetoacyl-acyl carrier protein synthase from avocado: its purification, characterization and clear resolution from acetyl CoA2ACP transacylase. Plant Mol. Biol. 25, 179-191. 162 Huppe, H. C., and Turpin, D. H. (1994). Integration of carbon and nitrogen metabolism in plant and algal cells. Annu. Rev. Plant Physiol. Plant Mol. Biol. 45, 577-607. Jaglo-Ottosen, K. R., Gilmour, S. J., Zarka, D. G., Schabenberger, 0., and Thomashow, M. F. (1998). Arabidopsis CBFl overexpression induces COR genes and enhances freezing tolerance. Science 280, 104-106. Jaworski, J., and Hinneburg-Wolf, B. (1998). Analyzing partial reactions of mutagenized spinach 3-ketoacyl-acyl carrier protein synthase III (KAS III). Presented at 13th Int. Symp. Plant Lipids, Sevilla. Kaneveski, I., and Maliga, P. (1994). Relocation of the plastid rbcL gene to the ' nucleus yields functional ribulose-l,S-bisphosphate carboxylase in tobacco chloroplasts. Proc. Natl. Acad. Sci. USA 91, 1969-1973. Kinney, A. J. (1998). Plants as industrial chemical factories - new oils from genetically engineered soybeans. FETT/Lipid 100, 173-176. Losey, J. E., Rayer, L. S., and Carter, M. E. (1999). Transgenic pollen harms monarch larvae. Nature 399, 214. Martin, W., Stoebe, B., Goremykin, V., Hansmann, S., Hasegawa, M., and Kowallik, K. V. (1998). Gene transfer to the nucleus and the evolution of chloroplasts. Nature 393, 162-165. Mekhedov, S., Martinez de Ilarduya, 0., and Ohlrogge, J. (1999). Towards the functional catalogue of plant genes: a survey of genes for glycerolipid biosynthesis. In preparation. Mendes, P. (1993). GEPASI: a software package for modelling the dynamics, steady states and control of biochemical and other systems. CABIOS 9, 563-571. Milius, S. (1999). Bt-corn pollen can kill monarchs. Science News 155, 324. Ohlrogge, J. B. (1994). Design of new plant products: engineering of fatty acid metabolism. Plant Physiol. 104, 821-826. Ohlrogge, J. (1999). Plant metabolic engineering: are we ready for phase two? Curr. Op. Plant Biol. 2, 121-122. Pollard, M. R., Anderson, L., Fan, C., Hawkins, D. J., and Davies, H. M. (1991). A specific acyl-ACP thioesterase implicated in medium chain fatty acid production in immature cotyledons of Umbellularia californica. Arch. Biochem. Biophys. 284, 306-312. 163 .4 '1‘ t'Yv—‘wn, _. Post-Beittenmiller, D., Jaworski, J. G., and Ohlrogge, J. B. (1991). In vivo pools of free and acylated acyl carrier proteins in spinach. J. Biol. Chem. 266, 1858-1865. Post-Beittenmiller, D., Roughan, G., and Ohlrogge, J. B. (1992). Regulation of plant fatty acid biosynthesis: analysis of acyl-Coenzyme A and acyl-acyl carrier protein substrate pools in spinach and pea chloroplasts. Plant Physiol. 100, 923-930. Roesler, K., Shintani, D., Savage, L., Boddupalli, S., and Ohlrogge John. (1997). Targeting of the Arabidopsis homomeric acetyl-Coenzyme A carboxylase to plastids of rapeseeds. Plant Physiol. 113, 75-81. Roughan, P. G. (1997). Stromal concentrations of coenzyme A and its esters are insufficient to account for rates of chloroplast fatty acid synthesis: evidence for substrate channelling within the chloroplast fatty acid synthase. Biochem. J. 327, 267-273. Roughan, P. G. (1998). Substrate channelling within the chloroplast fatty acid synthase: concentrations of acyl-acyl carrier proteins are very low during fatty acid synthesis. in preparation. Roughan, P. G., and Ohlrogge John B. (1996). Evidence that isolated chloroplasts contain an integrated lipid-synthesizing assembly that channels acetate into long-chain fatty acids. Plant Physiol. 110, 1239-1247. Sasaki, Y., Hadamada, K. Suama, Y., Nagano, Y., Furusawa, 1., and Matsnne, R. (1993). Chloroplast- encoded protein as a subunit of acetyl- -CoA carboxylase 1n pea Plant. J. Biol. Chem. 268, 25118- 25123 Sasaki, Y., Konishi, T., and Nagano, Y. (1995). Update on biochemistry: the compartmentation of acetyl-Coenzyme A carboxylase in plants. Plant Physiol. 108, 445-449. Sasaki, Y., Kozaki, A., and Hatano, M. (1997). Link between light and fatty acid synthesis: thioredoxin-linked reductive activation of plastidic acetyl-Coenzyme A carboxylase. Proc. Natl. Acad. Sci. USA 94, 11096-11101. Sauer, A., and Heise, K. (1984). Regulation of acetyl-Coenzyme A carboxylase and acetyl-Coenzyme A synthetase in spinach chloroplasts. Z. Naturforsch 39c, 268-275. Schultz, D. J., and Ohlrogge, J. B. (1998). Ferredoxin may be a key component to production of unusual monoenoic fatty acids 1n transgenic oilseeds. Presented at 13th Int. Symp. Plant Lipids, Sevilla. 164 Service, R. F. (1998). Seed-sterilizing 'Terminator Technology' sows discord. Science 282, 850-851. Shintani, D. K., and Ohlrogge, J. B. (1995). Feedback inhibition of fatty acid synthesis in tobacco suspension cells. Plant J 7, 577-587. Shorrosh, B. S., Roesler, K. R., Shintani, D., van de Loo, F. J., and Ohlrogge, J. B. (1995). Structural analysis, plastid localization, and expression of the biotin carboxylase subunit of acetyl-Coenzyme A carboxylase from tobacco. Plant Physiol. 108, 805-812. Shorrosh, B. S., Savage, L. J., Soll, J., and Ohlrogge, J. B. (1996). The Pea Chloroplast Membrane-Associated Protein, IEP96, is a Subunit of Acetyl-CoA Carboxylase. Plant J. 10, 261-268. Slabas, A. R., and Smith, C. G. (1998). Immunogold localization of acyl carrier protein in plants and Escherichia coli: Evidence for membrane association in plants. Planta 175, 145-152. Suh, M. C., Schultz, D. J., and Ohlrogge, J. B. (1999). Isoforms of acyl carrier protein involved in seed-specific fatty acid synthesis. Plant J. 17, 679-688. Varner, J., and Ramkrishna, D. (1998). Application of cybernetic models to metabolic engineering: investigation of storage pathways. Biotech. Bioeng. 58, 282-291. Winter, H., Robinson, D. G., and Heldt, H. W. (1994). Subcellular volumes and metabolite concentrations in spinach leaves. Planta 193, 530-535. Zechendorf, B. (1999). Sustainable development: how can biotechnology contribute? Trends Biotech. 17, 219-225. 165 APPENDIX A 166 Appendix A NUMERICAL METHODS This section describes and discusses the numerical methods to solve the equations generated by Stella II. The point of all these methods is to solve a set of differential equations, balancing computational efficiency against error. A mathematical function is “well-behaved” if it is both differentiable and continuous. Since the system is a biological one, it is quite probably described by a system of well-behaved functions. The model generates a system of rate equations that are a set of simultaneous differential equations, and numerous methods exist to solve these equations numerically. The system of equations is too complex to be solved analytically (which would give an exact solution). All the methods described below are iterative, single-step methods. That is, they evaluate the function at some specific value, in this case, time. The resultant value is used to evaluate the function at some further point in time. The method needs a function, an initial condition, and the first derivative of the function. In general, the closer the time intervals are, the more closely the value obtained from the numerical integration matches the value that would be obtained if the equations were solved analytically. These methods are based on series expansions of functions. A series expansion allows evaluation of a function, f(x) at some value, i.e. x, given an initial value 167 (x0) and the function. Euler’s and Runge-Kutta numerical methods are based on the Taylor series expansion. This expansion is shown below (Swokowski, 1978). fix)=f+f