Plant Biochemistry - sdarticle (10), Manuais, Projetos, Pesquisas de Química

Baixe Plant Biochemistry - sdarticle (10) e outras Manuais, Projetos, Pesquisas em PDF para Química, somente na Docsity! 8 Nucleic Acids and Proteins Eric Lam 8.1 Introduction 315 8.2 DNA 318 8.3 RNA 324 8.4 Proteins 338 8.5 Summary 349 References 350 8.1 INTRODUCTION Nucleic acids and proteins are two major con- stituents of all living cells with critical roles. Nucleic acids contain the information necessary for the synthesis of proteins with specific struc- tures and functions. On the other hand, proteins are needed for the metabolism of nucleic acids in addition to performing practically all the enzymic reactions in a living cell. With the advent of molecular techniques in every discipline of biol- ogy, our understanding of the chemical make-up and reactions within a living cell has undergone a major revolution in the last ten years. Although most of the fundamental work in the study of nucleic acids and proteins has been carried out in bacterial and animal model systems, many of the general principles appear to function in plants as well. In this chapter, our current state of knowl- edge on the properties of nucleic acids and proteins will be summarized, with special empha- sis on some of the plant-specific aspects. For more detailed descriptions of some of the general properties of nucleic acids and proteins, the reader should consult Stumpf 8c Conn (1981, 1989), Grierson 6c Covey (1988), and Alberts et al. (1989). 8.1.1 Nucleic acids There are two types of nucleic acid in living organisms: deoxyribonucleic acid (DNA) and ribonucleic acid (RNA). DNA and RNA are long polymers of subunits called nucleotides that are composed of a 5-carbon sugar backbone with one PLANT BIOCHEMISTRY ISBN 0-12-214674-3 or more phosphate groups attached at the C-5 position. In addition, nitrogen-containing bases called purines and pyrimidines are also attached to the C-1 position of the sugar. Nucleotides for DNA contain a hydrogen atom at the C-2 position whereas those for RNA contain a hydroxyl group instead. Polymers of deoxyribose or ribose nucleo- tides are linked by a phosphodiester bond between the C-3 hydroxyl group of one nucleotide and the phosphate group attached to the C-5 position of the adjacent nucleotide. Due to the stereospecifi- city of this arrangement, nucleic acids are di- rectional and are commonly diagrammatically displayed left to right from the 5'-phosphate con- taining end to the 3'-hydroxyl containing end (Fig. 8.1). For DNA, two different purines (adenine (A), guanine (G)) and two different pyrimidines (cytosine (C), thymine (T)) are found naturally. For RNA, the same bases are used with the exception that thymine is replaced by another pyrimidine uracil (U). The specific order of these bases on the sugar-phosphate backbone of DNA and RNA is the basis for their role as information storage devices. From these specific sequences, proteins with distinct structure and function are encoded and synthesized. In addition, regulatory information such as the precise expression pattern of the encoded protein is contained within the specific arrangement of these different nucleotides of DNA and RNA. For simplicity, a gene can be defined as a region of the DNA that contains the necessary information for the synthesis of one or more RNA transcripts. An important property of nucleic acids is their ability to interact with each other through hydrogen bonding between the nitrogen and Copyright © 1997 Academic Press Ltd All rights of reproduction in any form reserved 316 ERIC LAM 0 = P - 0 - C H 2 Q 0 " ^ Base 5' P-c A G HO- >-0H T C =i-P Figure 8.1 General structure of nucleic acids, (a) Chemical structure of the sugar-phosphate backbone. For simplicity, hydrogen atoms of the sugar moiety are not shown. The positions of the 5' and 3' carbon atoms are indicated. Four possible nitrogen bases can be attached to the Ci carbon atom, (b) Some essential features of a double-stranded nucleic acid are depicted. The anti-parallel arrangement is emphasized. Base- pairing between the two strands is suggested by the hydrogen bonds (vertical lines) shown between two pairs of nucleotides. oxygen atoms that are present in the keto and amino groups of the bases. Due to the structure of the different bases, interaction between them is highly specific. Thus, guanine can only interact with cytosine whereas adenine can interact with either thymine or uracil. These interactions, called base-pairing, are largely responsible for the secondary structures of DNA and RNA. Since three hydrogen bonds are formed between G and C as compared to two hydrogen bonds between A and T or U, G : C pairing is more stable. In nature, DNA is synthesized and maintained in a double- stranded form called the double-helix. This con- sists of two strands of DNA held together in an antiparallel fashion through base-pairing (see Fig. 8.1). In addition to hydrogen bonding between the bases, the double-helix structure is also stabilized by hydrophobic interactions between the planar aromatic surfaces of the bases that became stacked on each other in the center of the helix. Since interaction between bases is highly specific, the sequence of nucleotides from one strand of DNA will allow us to deduce the sequence of its complementary partner with great certainty. Due to the interaction between the bases, double- stranded nucleic acids have lower absorbance in ultraviolet light than the single-stranded forms. This behavior, called hyperchromicity, allows quantitative and kinetic measurement of the equilibrium between the two forms of nucleic acid. When a double-stranded nucleic acid in aqueous solution is heated to 100°C or when the solution pFi is increased to 13 or above, hydrogen bonds between the bases are disrupted and the two strands separate from each other. This process is called denaturation. Once denatured, the single- stranded nucleic acids can re-establish hydrogen bonds with other complementary sequences pro- vided that the right conditions, such as tempera- ture and pFi, are available. This process, which essentially involves the formation of a double- stranded nucleic acid from the single stran- ded forms (e.g. D N A : DNA, D N A : RNA, or RNA:RNA) is called hybridization and is a simple but powerful tool in molecular biology. The specificity of base-pairing allows the identifi- cation and quantitation of a small amount of specific RNA or DNA sequences within highly heterogeneous mixtures. Techniques based on sequence-specific hybridization of nucleic acids, such as genome mapping and polymerase chain reaction (PCR), offer exciting possibilities of understanding and manipulating complex genetic materials. In the first case, the relative locations of different regions of the chromosomes in an organism may be determined ^Little, 1992). In the latter, minute quantities of nucleic acids can be amplified millions of times through the action of an enzyme from a thermophilic bacteria called Taq DNA polymerase (Saiki et al,, 1988). 8.1.2 Proteins Enzymatic reactions within living cells are typi- cally mediated by proteins. The building blocks for proteins are amino acids that contain an identical backbone structure (Fig. 8.2) with one out of 20 possible side chains (R group) attached to it. Since their backbone structure is identical, the R group of amino acids distinguishes one amino acid from another. Each amino acid in a protein is connected to its neighbors via an amide linkage called a peptide bond, which involves the amino group of one residue and the carboxyl group of another. The primary structure of a protein is usually displayed left to right as the sequence of is constituent amino acids from the amino to the carboxyl termini. Like the purine and pyrimidine bases of nucleic acids, which are involved in specificity of base-pairing and other essential properties of DNA and RNA, the nature of the R-group in each of the 20 amino acids contributes to the functional and structural properties of proteins. Flowever, instead of only four different building blocks as in the case of nucleic acids, there are 20 different possible side NUCLEIC ACIDS AND PROTEINS 319 fast-renaturing portion is composed of highly repetitive sequences of 300-1000 base pair (bp) in size that are represented many times in the genome. Unique sequences that represent func- tional genes are in the slow-renaturing portion of the DNA. One type of repeat sequences that has been well-characterized in different plants is called satellite DNA. These DNA sequences are isolated based on their different apparent density from the rest of the nuclear DNA during centrifugation. They are usually found as a minor band of DNA of slightly lower density than the bulk of the genomic DNA. Their amount in different plants is variable and their repeat unit sizes range from a few to several hundred base pairs. The origin of satellite DNA is still unclear. In general, plants that have larger genomes tend to have also a larger portion of repetitive sequences in their DNA. For example, in pea, which has a large genome with about 4.8 X 10^ bp per haploid nucleus, about 85% of its DNA is repeated sequences (Murray et al., 1978). In contrast, there is only about 10% fast- renaturing DNA in the genome of Arabidopsis thaliana. In this plant, a 180 bp sequence has been identified to repeat between 4000 and 6000 times within the genome (Martinez-Zapater et al., 1986; Simoens et al., 1988). In addition, another 500 bp sequence is also found to be represented about 500 times per haploid nucleus. Like all eukaryotes, the telomeres at the ends of each of the five chromo- somes in Arabidopsis are composed of a 7bp sequence (CCCTAAA) repeated about 350 times. Together, these three types of characterized repeats account for about 20% of the repetitive sequences, or 2% of the total genomic DNA, that are present in this small plant. Aside from the telomeric repeats, which are likely to be involved in the maintenance of chromosomes, the function of the majority of the repetitive sequences present in the nuclear genome is largely unknown. In plants, however, species with a smaller size and a shorter life cycle tends to have smaller genome size and less repetitive sequences in their genome. In addition to nuclear DNA, plants have at least two other centers of DNA metabolism which reside in the mitochondria and the plastids. In each of these compartments of the plant cell, DNA replication and protein synthesis are carried out by organelle-encoded as well as nuclear-encoded gene products. Although many of the basic properties of organellar DNAs and their metabolism are likely to be similar, if not identical to their counterpart in the nucleus, important differences also exist. The percentage of the plastid genome in the total DNA of a given cell is variable among different organs of a plant since the number of plastids is dependent on developmental as well as environ- mental cues. It can range from about 10 to 20% of the total DNA in a leaf cell whereas in roots, it constitutes only about 1% of the DNA. In each plastid, 10-200 copies of a double-stranded DNA exist as circular molecules with sizes ranging from 70000 to about 180000 base-pairs in different species. Plastid DNAs of several plant species have been completely sequenced. Potential and known protein-coding regions in these plastid genomes have been identified (Shimada & Sugiura, 1991). The plastid genome has thus been shown to encode 120-150 genes. These include many of the proteins involved in photosynthesis and house- keeping functions such as protein synthesis. These and other studies have also shown that plastid DNA of most plant species studied to date contains an inverted repeat region in which a significant portion of the genome is duplicated on the opposite sides of the circular DNA. An interesting exception is that of the pea plastid DNA where no evidence for an inverted repeat region is found. The size of this inverted repeat region constitutes about 10-20% of the total plastid genome and the number and identity of the genes located within it is somewhat variable between different species. Many of the ribosomal RNA and tRNA genes are located within this inverted repeat region. An interesting mechanism, called copy correction, apparently exists in the plastids to insure that the sequence on the inverted repeats are identical. Thus, if advantageous mutations and insertions are introduced to one of the repeats during plastid transformation procedures, the alterations will be duplicated onto the other repeat unit (Svab et al., 1990). This process may provide a buffer against muta- tions in critical genes for plastid biogenesis. At present, the mechanism for this interesting phe- nomenon is unknown. Mitochondrial DNA of higher plants has a rather complex mode of maintenance. The number as well as the size of its genome is variable in a given species and is in striking contrast to that found in yeast and mammalian cells where a well-defined circular DNA molecule is found. In the case of human mitochondria, the nucleotide sequence of all 16569bp has been determined (Anderson et al., 1981). Typically, mitochondrial DNA of different plant species contains about 200000 to 2 million base pairs. In maize, seven different circular forms from 570000 to 47000bp in size have been described. An important genetic trait, called cytoplasmic male sterility, has been shown to be caused by mutations in genes encoded in the mitochondrial genome of some plant species 320 ERIC LAM (Levings, 1983). Unlike higher plants, the mito- chondrial genome of the liverwort Marchantia polymorpha is contained within a single circular molecule of 186 608 bp. The complete DNA sequence of this genome has been determined and genes encoding ribosomal RNAs, tRNAs and proteins involved in oxidative phosphorylation were identified (Oda et al., 1992). 8.2.2 Cellular dynamics of DNA structure and organization The nucleus of a eukaryotic cell is a dense struc- ture in which the genomic DNA is packaged in a complex nucleoprotein structure called chroma- tin. The structure of chromatin is not static and can be dramatically different in the life of a cell. In recent years, biochemical and molecular stud- ies have clearly shown that the organization of chromatin as well as DNA structure can have profound effects on gene expression (Grunstein, 1992). Thus, the understanding of DNA archi- tecture and chromatin organization is important from a functional as well as a structural point of view. The double-helix of DNA can be envisioned as two tightly linked, antiparallel ribbons that are wound around an imaginary rod. One important consequence of this structure is that it is stereo- specific. Thus, one can wind these linked ribbons in two different directions and produce molecules that are mirror images of each other. Due to the geometry of the sugar-phosphate backbone, most double helical DNA in nature exists in the 'right- handed' form called B-DNA. There are about 10.5 base-pairs in each turn of the double helix in this conformation. Short stretches of DNA that con- tain alternating purine/pyrimidine sequences have been found to adopt the 'left-handed' form of the double helix called Z-DNA. Evidence for the presence of Z-DNA in vivo have been obtained in animal and plant systems (Wittig et ai, 1991; Ferl & Paul, 1992). However, the function of these Z- DNA structures remains obscure. In addition to double helical structures, new structures of DNA such as DNA triplex and quadruplex have been shown to be possible in vitro (Griffin & Dervan, 1989; Kang et al., 1992). Recently, the structure of the telomeric repeats from the ciliated protozoan Oxytricha have been determined by X-ray crystal- lography (Kang et al., 1992) and N M R spectro- scopy (Smith & Feigon, 1992). Both studies show that these G-rich repeats exist as a DNA quadru- plex. These examples serve to illustrate the diverse possibilities that DNA structure can assume. The eukaryotic nucleus is a complex organelle that has a dynamic structure. Typically, it contains an envelope of two lipid bilayers with numerous pore structures involved in macromolecular trans- port in and out of the nucleus. Adjacent to the inside of this nuclear envelope is a matrix of proteins called lamins. Nuclear DNA packaged with basic proteins called histones are thought to be attached to this nuclear matrix. Evidence has been obtained that shows that at least one kind of lamin proteins, called lamin B, can interact specifically with certain DNA sequences that are thought to be involved in chromatin organization (Luderus et ai, 1992). There are also structural proteins such as tubulins and intermediate fila- ments within the nucleus. Their precise architec- ture and dynamics are complex and variable during the cell's life cycle. Another structure within the nucleus is the nucleolus. This is the site of most of the ribosomal RNA synthesis within the cell. Nuclear DNA is normally packaged in protein complexes with highly basic proteins called histones. There are five distinct groups of histone proteins called H I , H2A, H2B, H3 and H4. The structure of these proteins is very well conserved among eukaryotes, especially in the case of histones H3 and H4 where their primary sequences differ very little between cows and peas. The sequence of the other three histones are more variable. Nucleosomes consist of a core of two molecules each of histones H2A, H2B, H3 and H4. About 150 base-pairs of double-stranded DNA are wound twice around this octameric protein and one molecule of histone H I is bound to this nucleoprotein complex. This stage of DNA packaging produces fibers of about 10 nm in diameter and shortens the apparent length of the DNA about seven times. The next step of chromatin organization is the formation of a solenoid structure from the nucleosome arrays. This step is favored under conditions of higher ionic strengths in vitro, typically with Na"*" and Mg^"^, and involves the formation of a helical structure with six nucleosomes per turn. This process further condenses the DNA into fibers of about 30 nm in width. These 30 nm fibers of DNA can be organized into higher order structures that are less well characterized. Histone H I phos- phorylation is likely to modulate at least part of this condensation process. The Packaged DNA are also tightly associated with an ill-defined network of proteinaceous material called the nuclear scaffold. Specific sequences in nuclear DNA, called scaffold attachment regions (SARs), are found to be tightly associated with this structure NUCLEIC ACIDS AND PROTEINS 321 even after exhaustive extractions with salts to remove histones. SARs have been characterized from various eukaryotes including fruit flies and tobacco (Gasser & Laemmli, 1987; Hall et al., 1991) and are rather A/T rich in their sequences. These DNA elements of about 300-1000 bp in length are thought to anchor different regions of the chromatin onto the nuclear scaffold with the intervening DNA organized into nucleosomes by interaction with histones. These intervening DNAs can be visualized by electron microscopy as loops of DNA that emanate from the nuclear scaffold and are commonly described as a halo. It is believed that the association of the SARs with the nuclear scaffold may have important functional consequences on gene expression (Gasser &c LaemmH, 1987). The fact that DNA molecules in nature are predominantly double-stranded and are either circular, as in the case of organellar DNAs, or anchored to nuclear proteins create topological problems for the process of replication and transcription. During these critical processes, multimeric protein complexes are thought to unravel the two strands while sliding along the double helix structure. This action compresses the DNA in front of the enzyme complexes and introduces torsional stress known as positive supercoiling while negative supercoiling is intro- duced to the DNA behind. Another topological problem arises during DNA replication when the double-stranded daughter molecules are required to separate from each other. An important class of proteins called topoisomerases are known to be involved in the relief of torsional stress in nuclear and organellar DNA. These enzymes catalyze transient breakage in the DNA to allow the two strands to rotate freely with respect to each other before rejoining them. They are critical to the survival of eukaryotic and prokaryotic cells, as demonstrated by mutants with conditional defects in genes that encode these enzymes (Goto &c Wang, 1985; Wang, 1991). Topoisomerases have been char- acterized in plants and they share similarities with their counterparts in other systems (Lam & Chua, 1987; Kieber et al., 1992). Since alterations in DNA topology have been shown to affect critical processes such as transcription and replication in vitro (Wang, 1991), topoisomerases may play a regulatory role in modulating these functions. The phosphorylation of yeast topoi- somerase II has recently been shown to increase its enzyme activity in vitro and is correlated with the onset of mitosis in the cell cycle (Cardenas ^^^/. ,1992). Nuclear DNA of various organisms have also been shown to be methylated by enzymes called methylases present in eukaryotes. As much as 30% of the nuclear DNA in pea is methylated. The most common form of methylation is the addition of a methyl group at the 5-carbon position of the cytosine base. Usually, this takes place in the cytosine of a CpG site where p stands for the location of the phosphodiester bond. Because of the symmetry of this dinucleotide pair, the cytosine base of the opposite strand is also modified. It is important to note that not all CpG sites of the genome are methylated. In fact, methylated DNA is usually thought to be asso- ciated with regions of the genome that are inactive in gene expression. Many proteins involved in binding to DNA are known to be sensitive to methylation of their recognition sequences. Thus, differential modification of DNA can have pro- found effects on the expression of the information carried by the particular DNA sequence. In fact, treatment of cells with the cytosine analogue 5- azacytosine, which cannot be methylated, can cause dramatic phenotypic changes. At present, the regulation of DNA methylase activities remains to be elucidated. Although the exact DNA sequence and the relative position of genes in a particular organism remains fairly stable, an important mechanism called recombination is used to maintain certain levels of genetic variation. This process insures that there is sufficient diversity in the genome to increase the chance of species survival in a changing environment. Two types of recombina- tion events are commonly known: general genetic recombination and site-specific recombination. In general recombination, homologous DNA strands from two copies of the same chromosome can be exchanged. This process can be initiated by a broken phosphodiester bond on one of the strands of the two sister chromosomes and usually involves base-pairing between the strands of the two neighboring DNA. An important example of this process is the exchange of sections of chromosomes during the course of meiosis. This event, called crossing-over, involves the exchange of genetic information between different parts of the closely apposed chromosomes during the early development of gametes and allows for greater diversity of the resultant progenies. In contrast to general recombination, site-specific recombination involves the movement of genetic material in the genome without the requirement of extensive base- pairing. This event is catalyzed by a class of enzyme called recombinases and requires only a relatively small, but specific DNA sequence. 324 ERIC LAM Okazaki fragments RNA primer Figure 8.3 DNA synthesis at a replication fork. The discontinuous synthesis of DNA on the lagging strand is contrasted with that of the leading strand. The Okazaki fragments and the RNA primer synthesized by DNA primase are indicated with arrows. the Okazaki fragments requires the removal of the RNA primers by a 5' to 3' exonuclease activity that is present in DNA polymerase a, follow^ed with the repair of the resultant gap using a specialized DNA polymerase, and finally ligation of the fragments by the enzyme DNA ligase. The polymerization of DNA on RNA primers is likely to involve the soluble forms of DNA polymerases a and 7 v^hereas DNA gap repair is likely to involve the chromatin-bound DNA polymerase /?. One impor- tant activity of DNA polymerases is their ability to work as a 3' to 5' exonuclease. When an error is made during DNA polymerization and a wrong base is incorporated, DNA synthesis will be inhibited. This inhibition is due to the fact that DNA polymerase absolutely requires the 3' hydroxyl end of a base-paired primer in order to add any new nucleotides and a mismatch at the 3' end of the primer will not be tolerated. In this case, the 3' to 5' exonuclease activity intrinsic to DNA polymerase will remove the unpaired base and thus allow the polymerase activity to proceed. This 'proof-reading' capability of DNA poly- merases is critical to their faithful duplication of large amounts of genetic information. In plants, DNA replication has been studied by radioiabeling plants with radioactive thymidine followed by autoradiography of the isolated DNA (Van't Hof et al., 1978). Each molecule of chromosomal DNA is found to contain many points of replication and they all proceed in both directions from each origin of replication. In total, there are tens of thousands of active replication forks in a plant cell nucleus during S phase. Apparently, different sets of these replication origins in the chromosome can be active at different times. The control of DNA replication and its site of initiation in plants, and in eukaryotes in general, remain to be elucidated. The origins of DNA replication have been identified in E. colt and in yeast. It is believed that they act as binding sites for sequence-specific DNA binding proteins. The characterization of these origin-binding proteins is at present under active study. In a number of plants, the activity of DNA polymerase a has been found to increase significantly during seed germination, in meriste- matic cells at the shoot and root apices, and upon treatment with auxins and cytokinins (reviewed in Dunham & Bryant, 1985). These conditions are all known to be associated with high rates of cell division in plants and are in keeping with their requirement for higher levels of DNA poly- merase activity. However, the regulatory mechan- isms underlying these observations remain to be elucidated. 8.3 RNA 8.3.1 Property and diversity of plant RNAs RNA is a molecule that is at once simpler and more complex than DNA. It is simpler than DNA because it consists of only a single-stranded polymer of nucleic acids rather than being double-stranded. This fact makes its synthesis and metabolism a more straightforward problem in comparison to that of DNA. However, the secondary structure of RNA is more difficult to predict since the possible locations of intramole- cular base-pairing within an RNA molecule can be numerous. RNA is also more labile than DNA. In vitro, RNA is very sensitive to high pH since low levels of hydroxide ion will result in the formation of 2',3'-cyclic phosphates from RNA. Nucleic acids without the 2'-hydroxyl group, as in the case of DNA, are much more stable under alkaline conditions. RNases, enzymes that degrade RNA avidly, are usually also very stable proteins. This fact makes it harder to isolate intact RNA compared to DNA. Cells contain three types of RNAs that can be distinguished by their structure, mode of synthesis and function. These are called messenger RNA (mRNA), ribosomal RNA (rRNA) and transfer RNA (tRNA). In a typical plant cell, rRNA accounts for about 85% of the total RNA. Of this pool of rRNA, the contribution from the plastids is highly variable. In photosyn- thetic tissues, where the copy number of plastids per cell is high, chloroplast rRNA can make up NUCLEIC ACIDS AND PROTEINS 325 25-40% of the total RNA. All nuclear-encoded genes are first transcribed within the nucleus and subsequently transported to the cytosol, where they are usually involved in protein synthesis. Thus, the process of translation is spatially and temporally uncoupled from that of transcription for nuclear genes. In contrast, organellar and bacterial transcription is thought to be closely linked to the process of translation. (a) Messenger RNAs mRNAs provide the templates on which proteins of specific amino acid sequences are synthesized. Most nuclear mRNAs carry information, via a triplet code called codons, for the production of a specific polypeptide sequence. This is referred to as monocistronic. In contrast, bacterial and plastid genes are often transcribed as polycistronic mRNAs that encode multiple polypeptides. In a typical cell, mRNA usually represents about 5-10% of the total RNA by weight. The number of distinct mRNA species is likely to be in the range of about 10000 in a particular cell and the relative amounts of these different RNAs can differ by more than 100-fold at steady state. Thus, regulatory signals must be present to affect the differential accumulation of mRNA transcripts in the cytosol. A mature mRNA is typically between 400 and 4000 nucleotides in length. In higher eukaryotes such as plants, nuclear mRNA are usually synthesized in a precursor form that contains introns ranging in size from about 100 to a few thousand nucleotides. The introns are usually spliced out before the mature RNA is transported to the cytosol. Transcription of mRNA is carried out by the enzyme RNA polymerase II (Pol II) and proceeds from the S' end of the encoded transcript. After polymeriza- tion of about 20-30 nucleotides, the S' triphos- phate group of the first ribonucleotide is covalently linked to a guanosine by a nuclear enzyme. In plants as well as in animals, this guanosine is then methylated at the N-7 nitrogen of the guanine base. This process is called capping and appears to be universal for nuclear-encoded mRNAs. The presence of a cap structure at the S' end of the mRNA is thought to promote translational initia- tion as well as to protect the RNA from 5'- exoribonucleases. However, it is not an essential for the translation reaction since uncapped mRNA can be translated in vitro. The 3' end of most nuclear mRNA is also modified by the addition of multiple adenylate residues through the action of the enzyme poly(A) polymerase. A well-known exception is the transcripts encoding histones. which do not contain polyadenylated 3'-ends. Transcription of mRNA normally proceeds about 50 to several thousand bases beyond the 3'-end residue that will eventually be polyadenyl- ated. A conserved sequence AAUAAA that is about 30 nucleotides upstream from the poly (A) site is required to generate the proper 3' end of the RNA with the co-operation of other sequence elements that are downstream from the adenyla- tion site. Since RNAs with or without polyadenyl- ation can be translated efficiently in vitro., it is unlikely that the poly (A) tails are required for translation per se. However, studies in animal systems have suggested that the length of the poly(A) tails is directly correlated with the stability of the mRNA in the cytosol (Bernstein 6c Ross, 1989). In plant protoplasts, it has been shown that the presence of 15 adenylate residues at the 3'-end of the mRNA can enhance the half-life of a transcript by about 3-fold (Gallie et al., 1989). In addition, the presence of the poly(A) tail can also stimulate translation of the mRNA. Another covalent modification of mRNA that takes place in the nucleus is the methylation of some adenylate residues. Up to two methylated adenine have been found per mRNA of 1600 nucleotides. At present, the significance of adenine methylation in mRNA is not clear since yeast mRNA apparently lacks this modification. Thus, it may not be an essential process that is conserved among eukaryotes. The mRNAs of mitochondria and plastids differ from those of the cytosol in several respects. (1) they do not contain any apparent 5' methyl-guanosine cap; (2) they do not contain poly(A) tails at the 3' end and (3) in plastids, polycistronic mRNAs are observed for many of the genes found in this organelle. Pro- cessing of these mRNA species in turn generates a complex pattern of transcripts (Gruissem, 1989). (b) Ribosomal RNAs The cytosol of eukaryotes contains 805 ribosomes as the protein complex responsible for the translation of mRNA into proteins. These ribo- somes are either free in the cytoplasm or are membrane associated. Under appropriate condi- tions, the 80S ribosomes can be dissociated into two subunits one of which is twice the size of the other. The larger ribosomal subunit contains three RNA species of 25, 5.8 and 5S whereas the smaller subunit has one RNA component of 18S. Together, these RNAs make up about 50% in the total mass of the cytosolic ribosomes. The 70S ribosomes found in the plastids of higher plants have very similar characteristics to those of 326 ERIC LAM prokaryotes and they consist of 50S and SOS subunits. The 50S subunit of the plastid ribosome contains rRNAs with the size of 23, 5 and 4.5S. A single species of 16S rRNA is present in the smaller subunit. Plant mitochondrial ribosomes from various species sediment at 77-78S and contain three rRNAs. In the larger subunit of the ribosome, a 24-26S rRNA with sequences similar to those from the mitochondria of fungi and animals is found. In addition, a 5S rRNA apparently unique to this organelle in higher plants is also present. A single 18S rRNA is found in the smaller subunit of the plant mitochondria ribosomes. Although the sedimenta- tion coefficients of the subunits of plant mitochon- drial ribosomes are more similar to their counterparts in the cytosol, the sequence of their rRNA resembles more of those found in 70S ribosomes. The size of the three classes of rRNAs (i.e. 2 3 - 26S; 16-18S; 4.5-5.8S) found in ribosomes of various compartments of the plant cell are approximately 3200, 1800 and 120 kilobases (kb) in length. The organization of rRNA genes in the nucleus is remarkably conserved among eukar- yotes. They are typically arranged in large tandem arrays of a repeated transcription unit that consists of the 18S, 5.8S and 25S rRNAs with transcribed 'spacer' regions between these sequences. This organization is illustrated in Fig. 8.4. Intergenic regions between the repeats have also been found to contain internal repeating structures and may function to regulate expression of the rRNA genes. The rRNA repeat unit is present in a few hundred to several thousand copies in the nucleus and is organized into fibrillar centers in the nucleolus and in a condensed form in heterochromatin. In addition to the 18S-5.8S-25S rRNA gene clusters, the 5S rRNA genes also exist as long tandem arrays in various regions of the plant chromo- somes. The sequence and predicted secondary structure of 5S rRNAs from diverse sources in nature are highly conserved. Interest- ingly, the rate of 5S rRNA synthesis is apparently not correlated with that of the other rRNAs. Although the sequence of rRNAs are fairly well- conserved, the number of rRNA genes can be variable even within the same species. When active during interphase of the cell cycle, the condensed units in the nucleolus are uncoiled and transcrip- tion of many of the repeat units is initiated simultaneously. This can be visualized by electron microscopy as well as quantitated by radioisotope labeling of the newly synthesized rRNAs with [^H]uridine. From such studies, it has been shown that not only the transcription of multiple adjacent rRNA repeat units can be carried out simulta- neously, each repeat unit can also initiate new RNA synthesis before termination of the earlier ones. These properties, in addition to the numer- ous numbers of these genes, accounts for the high proportion of rRNA within the total RNA pool of the cell. Since there are several million ribosomes per cell, the maintenance of high levels of rRNA is critical for the cell survival and proliferation. The 18S-5.8S-25S repeat units are transcribed by RNA polymerase I to produce initial transcripts of about 45S. These 45S transcripts are then processed in the nucleolus to produce the three mature rRNAs before they are transported to the cytosol. However, methylation of highly conserved regions is known to take place even before the completion of 45S pre-rRNA synthesis. Although the exact role of rRNA methylation is not known for certain, it appears to be necessary for pre- rRNA processing. Ribosomal proteins are also known to be associated with the pre-rRNA during its synthesis, before its exit to the cytosol. After the transcription of pre-rRNA is complete, the 18S rRNA is cleaved from the precursor transcript. Subsequently, the 5.8S rRNA is hydrogen-bonded to complementary sequences at the 5' end of the 25S rRNA sequence. Cleavages at the spacer region between the 5.8S and 25S rRNAs then Intergenic Region 18S rRNA 5.8S rRNA 25S rRNA Figure 8.4 Structure of a rRNA gene cluster. The arrangement of the 18S, 5.8S and 25S rRNA-encoding genes are shown in an idealized form. An intergenic region between two rRNA repeats is indicated with an arrow. NUCLEIC ACIDS AND PROTEINS 329 flower and leaf morphogenesis have been found to encode putative transcription factors (Yanofsky et ai, 1990; Vollbrecht et al., 1991). Thus, as in the w^ell-documented case of segmentation genes of Drosophila, a cascade of transcriptional regulators is likely to underlie many of the pathw^ays involved in developmental control. The central enzyme in transcription is called RNA polymerase. Mech- anistically, the best-characterized RNA polymer- ase is that of the eubacteria w^hich served as the model system in which the process of RNA synthesis has been studied in detail. The first step of this reaction involves the recognition and binding of RNA polymerase to a specialized region of the gene called a promoter. This is usually an asymmetric sequence located in the 5' end of the transcribed portion of the gene. Subsequent to the formation of a polymerase/ promoter complex, local 'melting' of the transcrip- tion start site ensues to produce single strands of DNA within a short region. Base-pairing then occurs between free ribonucleoside triphosphate monomers and one of the two strands in this open complex before the adjacent nucleotides are joined by the formation of phosphodiester bonds. This creates a short region of D N A : RNA hybrid in the active site of the polymerase. The polymerase then moves along the DNA toward the 3'-end of the growing RNA molecule while unwinding small stretches of the transcribed region one at a time. As the transcription complex passed through, the DNA double helix is rewound behind the poly- merase and the RNA strand is displaced. When the polymerase encounters a special sequence in the gene called termination signal, the enzyme is released from the template DNA along with the completed RNA chain. To accomplish these tightly orchestrated reactions, some bacterial viruses encode a single subunit of RNA polymer- ase that can function without the aid of other proteins. The bacterial and eukaryotic enzymes, however, are large multi-subunit complexes of over 500 kilodaltons (kDa) in apparent molecular weight. It is probable that this difference reflects the need for tight regulation of transcription in most cells. One of the most well-characterized RNA polymerases is that of E. coli. The holoenzyme consists of five distinct subunits called a, /?, /?', a and LU. Two copies of the a subunit are present in each holoenzyme complex whereas the other subunits are present in single copies. The a subunit acts as an initiation factor and provides the sequence specificity to the polymerase holoenzyme to preferentially transcribe from promoter ele- ments of the bacterial genome. There are multiple (7 subunits present in bacteria and they are involved in gene activation under specific condi- tions. One of the best known of these are the a^^ factors of E. coli, that are responsible for the activation of heat-shock responsive genes, and the multiple a factors of B. subtilis that regulate genes involved in sporulation (Doi & Wang, 1986). Thus, there are likely to be distinct holoenzymes that have RNA polymerase activity in any given bacteria at a particular time. The predominant form of a factor in E. coli is called cf^ and this protein has been shown to contain two different DNA-binding domains that recognize different sequence motifs in the prokaryotic promoter (Dombroski et at., 1991). These sequences are the so-called '—10' and '—35' elements located at the specified regions upstream of the transcription start-site and they have the consensus sequence of TATAAT and TTGACA, respectively. Genes that are tightly regulated in bacteria usually have very different - 1 0 and - 3 5 elements from this con- sensus and it is likely that a different a factor will provide distinct sequence preference for the resultant RNA polymerase holoenzyme in order to transcribe them under a particular set of conditions. After transcription has been initiated in the open complex, a factor is released from the polymerase. To promote efficient elongation and correct termination, elongation factors are then bound to the polymerase. Although sequence elements that are responsible for the regulation of bacterial gene expression is commonly thought to reside around the transcription initiation site, more recent studies have demonstrated that repressors and activators can also exert their effects over kilobases away to influence the action of RNA polymerase (Gralla, 1989). These modulators of transcription are typically sequence- specific DNA-binding proteins and at least in the case of the NRi regulator of the genes involved in nitrogen assimilation, protein phosphorylation has been shown to mediate the activation process. In plant cells, at least three distinct compart- ments carry out RNA synthesis and each has their own special properties. These are the nucleus, plastids and mitochondria. In the nucleus, three different classes of RNA polymerase activities can be differentiated by their sensitivity toward the fungal toxin a-amanitin and the types of RNA that they produce. RNA polymerase I (Poll) is localized in the nucleolus and is responsible for the synthesis of large rRNAs. This enzyme activity is insensitive to a-amanitin. RNA polymerase II (Pol II) is localized in the nucleoplasm, very sensitive to a-amanitin and is responsible for the synthesis of most, if not all mRNAs as well as 330 ERIC LAM some snRNAs in the nucleus. Lastly, RNA polymerase III (Pol III) is responsible for the synthesis of tRNAs, the small rRNAs, and some of the snRNAs. This enzyme is inhibited by a-amanitin at relatively high concentrations of the toxin. Organellar RNA polymerases, like those of prokaryotes, are not sensitive to a-amanitin and do not appear to differentiate between different types of RNA. Frequently, polycistronic tran- scripts are made and multiple processing steps are required to generate the correct RNA pro- ducts. Biochemically, there is evidence that two distinct types of RNA polymerase may exist in chloroplasts. The soluble form of this enzyme behaves much like that of eubacteria whereas a chromosome-bound RNA polymerase activity, called transcriptionally-active chromosome (TAG), can be distinguished by its activity under high salt stress (100-200 mM NaCl) and insensitivity to heparin, an inhibitor of transcription initiation. The TAG form of chloroplast RNA polymerase appears to preferentially transcribe the rRNA operon in the inverted repeat region. The basis for the differences observed between the two RNA polymerase activities is still unclear. The nuclear RNA polymerases are structurally more complex than their counterpart in prokar- yotes. They contain large numbers of subunits and accessory factors that are required for correct initiation of transcription. In addition to having two large subunits with molecular weights of 190-150 kDa, purified nuclear RNA polymerases contain at least six or more additional subunits. Interestingly, the sequence of the two large subunits for Pol II and Pol III have been found to contain regions homologous with that of prokar- yotic RNA polymerase (Allison et al., 1985). Thus, it is very likely that these enzymes have evolved from a common ancestor and mechanistically, the reactions catalyzed by these different enzymes are likely to be very similar. In the plastid genome, three genes with good homologies to bacterial RNA polymerase subunits a , f3, and /3' are found in the genome of tobacco but are absent in a nonphotosynthetic parasitic plant Epifagus virgini- ana (Wolfe et al., 1992). Thus, like the case of mitochondrial RNA polymerase, the plastids of £. virginiana must import the subunits of its RNA polymerase from the cytoplasm. (a) Organellar transcription Transcription of human and yeast mitochondria DNA have been studied in detail. The genome of human mitochondria is apparently tran- scribed only from two neighboring points. Two polycistronic mRNAs are generated with one of them covering the entire circular genome. These pre-mRNAs are then spliced to generate the various tRNAs and rRNAs. In addition, the few mRNAs are also polyadenylated. In yeast mito- chondria, which are about five times the size of human mitochondria, individual promoters are used to transcribe each gene. In addition, multiple introns are found within some of the genes. An interesting observation in yeast mitochondria is that intron 2 of the cytochrome b gene encodes a protein called maturase. Mutations in intron 2 that disrupt the maturase also inhibited proper splicing of the cytochrome b transcript. In plant mito- chondria, the situation appears more complicated than the other systems since multiple forms of the genome are present at any one time. In addition, it is difficult to determine whether the 5' termini of the multiple transcripts seen with some genes are the results of transcription initiation or RNA processing. However, most plant mitochondrial transcripts appear to begin within a consensus motif: TAAG(T/A)GA. Whether this apparent motif for transcription initiation has any funct- ional significance has not been rigorously tested directly. In contrast to the case of plant mitochon- dria, transcription in the plastids has been well- characterized due to the availability of in vitro transcription systems that can accurately produce plastid RNAs (Hanley-Bowdoin & Ghua, 1988; Gruissem, 1989). Structurally, the plastid RNA polymerase consists of two large subunits with the apparent molecular weights of 180 to 140 kDa, which correspond to the P' and (3 subunits of prokaryotic RNA polymerase, in addition to about 10 other smaller subunits. Aside from the a , /^, P' subunits, which are likely to be encoded in the plastid genome, the other subunits must all be imported. Although fractions from chloroplast extracts have been reported to have a-like activity in vitro (Jolly & Bogorad, 1980; Lerbs et al, 1988; Tiller et al., 1991), the protein(s) involved has not been purified and its relationship to the well- known a factors of E. coli remains elusive. Many mRNA and tRNA genes are transcribed as polycistronic transcripts in the plastid and their promoters are strikingly similar to that in bacteria. The consensus promoter element for the gene encoding the large subunit of ribulose-l,5-bisphos- phate carboxylase {rbcL) from different species consisted of a TTGGGG and a TAGAAT motif at the —35 and —10 region, respectively. These elements are strikingly similar to the —35/—10 elements found in many prokaryotic promoters. In fact, E. coli RNA polymerase can readily NUCLEIC ACIDS AND PROTEINS 331 transcribe these plastid genes accurately in vitro. Deletion analyses have shown that a spacing of 18 bp between the —35 and —10 elements is crucial for optimal expression of a number of plastid promoters in vitro. Although there are many similarities between the transcription system of the plastid and bacteria, there are also a number of important differences. The gene encoding trnS has bee reported to be transcribed in the absence of its 5' promoter region (Gruissem et al., 1986). Also, a blue-light regulated promoter for the psbD gene does not contain obvious —35 or —10 motifs (Sexton et al., 1990). Plastid RNA polymerase has been found to prefer supercoiled, circular DNA templates in vitro whereas E. coli enzyme can transcribe equally well from linearized DNA (Lam & Chua, 1987). These observations suggest that there may be plastid-specific factors or functional properties that distinguish the RNA polymerase from its bacterial counterpart. One additional difference in this regard is the insensitivity of the plastid enzyme toward rifampicin, which is a potent inhibitor of bacterial RNA polymerase. It should be pointed out that most of the detailed promoter analyses for plastid-encoded genes have so far been carried out in vitro only. Thus, if labile upstream factors are indeed required for their regulated expression, one may not be able to detect requirements of sequences in addition to the —35/—10 regions. Indeed, site-specific DNA-bind- ing activities in chloroplast extracts have been reported in the last few years (Lam et al., 1988; Baeza et al., 1991). The function of the binding sites for these factors, however, has not been determined due to the lack of an in vivo assay. With the recent advent of a plastid transformation system in higher plants as well as in the alga Chlamydomonas (Svab et al., 1990; Boynton et al., 1988), these problems may be resolved in the near future. Transcriptional rate measurements with iso- lated spinach plastids have suggested that the relative promoter strength of plastid-encoded genes does not vary significantly (Deng & Gruissem, 1987) in different tissues and environ- mental conditions. Regulation at the level of RNA stability and mRNA translation is concluded to be the predominant mechanisms that affect plastid gene expression. However, in addition to psbD, preferential transcription of psbA and petE genes in illuminated barley plants have been documented (Klein & Mullet, 1990; Haley & Bogorad, 1990). Methylation has also been reported as a mecha- nism for differential gene repression during the transition from chloroplasts to chromoplasts in tomato fruits (Kobayashi et al., 1990). These results suggest that plastid gene expression may be regulated in a species-specific fashion. (b) Nuclear transcription control The initiation site of nuclear gene transcription is determined by promoter sequences that are distinct for each class of RNA polymerases and are very different from the —10/—35 element of prokaryote promoters. In spite of their differences in the sequence motifs used, these promoters are all organized in a roughly bipartite manner. In each case, a region of the promoter close to the start site of initiation is dedicated to the binding of crucial proteins that will recruit RNA polymerase to the correct start site. As one will predict from this function, the activities of these sequence motifs are distance- and orientation-dependent. The promoter element for RNA polymerase I has been shown to reside in the region from —45 to +20 of the human rRNA promoter (Jantzen et al., 1990) and it does not appear to be conserved among species. A factor SLl from human nuclear extracts can apparently override this species- specificity when added to a mouse extract. Thus, species-specific factors may have evolved to selectively activate the promoters of rRNA genes. In the case of the wheat rRNA operons, the intergenic regions have been found to contain multiple repeats of a 135 bp sequence upstream of a putative promoter that resembles a sequence of the Xenopus rRNA promoter (Flavell et al., 1986). The number of this A-repeat is variable among different rRNA encoding genes and may account for their variation in expression in vivo. In addition, methylation of the intergenic region may also modulate the activity of this class of promoters. Protein factors that bind to the upstream regions of rRNA genes have been detected in maize, radish and cucumber (Schmitz et al., 1989; Echeverria et al., 1992; Zentgraf & Hemleben, 1992). Their precise function and structure, however, is still not well understood. In the case of Pol II promoters, two different non-exclusive elements are found in mRNA- encoding genes in diverse species. These are the TATA motif that is usually found at about 30 nucleotides upstream from the start site of transcription, and an initiator (Inr) motif that is absolutely required in Pol Il-transcribed promoters that lack a TATA box. The Inr has a consensus sequence of - ^ C T C A N T C T + ^ where the central A residue is the start site of transcription and N is either T or C (Smale et al., 1990). The basal factors that interact with Pol Il-specific promoters have been especially well-characterized in the last few 334 ERIC LAM 4. Covalent modification: protein phosphoryla- tion has been the most well-studied mechanism of transcription factor modulation by post- translational mechanisms (Hunter & Karin, 1992). In some cases, the sites of phosphoryla- tion have been mapped and the consequences on transcription varies from inhibitory to enhancement. Several plant DNA-binding pro- teins have also been shov^n to be phosphory- lated in vitro by plant kinases. Their in vivo functional significance, how^ever, remains unclear. (c) Transcription elongation and termination In addition to transcription initiation, the pro- cesses of transcript elongation and termination site selection are also possible points of gene control. In bacteria, w^here these processes have been best studied, two major classes of termination activities are known. These are called the rfeo-dependent and r/70-independent types. In the first type, a cellular factor rho is involved in sequence-specific transcription termination. For the r/70-independent process, transcription usually terminates at regions that have the following characteristics: a GC-rich region that can form a stable hair-pin structure in the RNA product followed by a stretch of A residues on the template strand. Studies with bacteriophages have shown that the action of rho can be counteracted by antitermination factors (Yager & von Hippel, 1987). For example, the Q protein of lambda phage can bind to the elongat- ing RNA polymerase complex and by doing so, allows transcription to proceed beyond either rho- dependent or -independent termination sites. The binding of Q to the elongation complex is enhanced by an elongation factor, NusA. Another lambda protein, N, can also act as antiterminator for r/70-dependent termination sites. In this case, co-operation with at least six other bacterial proteins and binding of the N protein to a specific site on lambda DNA are required. In eukaryotes, the characterization of factors that influence the elongation and termination of transcription have made significant advances in recent years (Spencer & Groudine, 1990). Two different types of elongation factor activities have been identified for mammalian Pol II transcription. The basal transcription factor, TFIIF, has been found to increase the rate of transcript elongation by the ternary complex between RNA polymerase, tem- plate DNA and nascent RNA. The other type of elongation factors, typified by the elongation factor SII, enhances he progression of the ternary complex through intrinsic termination sites. This process has been especially well-studied with the adenovirus system where transcription by Pol II in vitro are known to be blocked at specific sites along the major late gene. Addition of factor SII apparently releases the halted ternary complex and stimulates read-through of the natural termination sites (Izban & Luse, 1992). Interestingly, PPR2, a yeast gene identified as a transcription regulator has been found to be homologous to factor SII. This fact makes it very likely that elongation factors for transcription are involved in the regulated expression of certain endogenous genes. Characterization of SII and TFIIF homologs in plants will enhance our understanding of this process in the future. 8.3.3 RNA processing Nuclear and some organellar RNAs are synthe- sized as larger precursors and are usually asso- ciated with protein particles. These preRNAs are polyadenylated at their 3' ends, for mRNAs, as well as capped in their 5' end. Before their exit into the cytoplasm, however, these transcripts have to be properly processed into mature RNAs and then transported through the nuclear membranes. From pulse-labeling studies, it has been found that about 95% of the nuclear pre-RNA, called heteronuclear RNAs (hnRNAs), are degraded within about an hour after their synthesis and thus never make it to the cytoplasm. Although the process of nucleo- cytoplasmic RNA transport is poorly understood in plants, some of the RNA processing reactions are beginning to be characterized at the molecular level. One of the most important of these is the splicing of introns from pre-RNAs. For example, regulation of intron splicing has been demon- strated in animal systems to be important for the pathway of sex determination in Drosophila (Baker, 1989). It is reasonable to expect that the generation of different mRNAs from a primary transcript will be an important point of regulation for some pathways in plants as well. There are two major types of intron splicing mechanisms. The self-splicing introns of Tetra- hymena rRNA and some fungal mitochondrial genes are typical 'group I' introns. These introns were among the first RNA molecules shown to have enzymatic activities. Aside from the RNA molecule itself, a guanosine is necessary for this sphcing reaction. As shown in Fig. 8.6, the first step involves the attack of the phosphodiester bond of a conserved uridine residue at the 5'-splice site by the guanosine nucleotide cofactor. This is followed by another transesterification step in NUCLEIC ACIDS AND PROTEINS 335 Group I Introns 5' 3' ] j ) -0-P—aG-O-P-d \ yjoH G-OH Group II Introns V W - O - G U E 2* OH / "CZ 5' 3" ZIC Figure 8.6 Mechanisms of group I and group II self-splicing introns. In group I introns, a guanosine cofactor attacks a phosphate group located in the 5' splice-site of the pre-RNA. This results in a transient break in the RNA and the linkage of the guanosine cofactor to the 5' end of the intron. Subsequent transesterification between the uridine residue at the 5' splice site and the linkage at the 3' splice site then results in the excision of the intron. In group II introns, the 2' hydroxyl group of an internal adenosine attacks the phosphodiester linkage between the 5' splice site and a conserved-GU-dinucleotide. This results in an unusual lariat structure at the 5' end of the intron. Subsequent transesterification then results in the joining of the two exons and the excision of the intron in the lariat form. The region of the RNA that contains the intron is shaded. which the 3'-hydroxyl group of the upstream exon displaces the 3'-hydroxyl group of a conserved guanosine residue of the intron at the 3'-splice site. A consequence of this reaction pathway is that a guanosine is added to the 5' end of the excised intron sequence. 'Group IF self-splicing introns are found in some structural genes of chloroplasts, plant mitochondria and yeasts. In contrast to group I introns, cofactors are not required for the self- catalyzed splicing reactions of this group. Instead, a conserved AU dinucleotide sequence within the intron near the 3'-splice site is used to catalyze the first step of the reaction. This so-called branch point involves the formation of a 2',5,-phospho- diester bond between the adenosine residue and a guanosine residue at the 5' end of the intron. The product of this reaction is the formation of a lariat-like RNA molecule. Subsequent attack of the 3'-splice site by the 3'-hydroxyl group of the upstream exon links the exons together and also releases the intron in the lariat form. In addition to the self-splicing type, group II introns also include the introns of mRNA in the nucleus. However, the splicing process of these intervening sequences is dependent on a complex structure called the spliceosome. This is a ribonucleoprotein (RNP) complex that contains snRNAs as essential components in addition to proteinaceous components. Although the structure and function of the different components are beginning to be elucidated in yeast and animal model systems, the components of this important process in plants are less defined. Since introns of mRNAs from animals are not spliced, or only very poorly, in yeast and plants, it is likely that they may have very different requirements. The work of Goodall & Fillipowicz (1989, 1990) has established some important characteristics of the splicing system in higher plants. Their work using an artificial intron shows that although like all other introns, plant introns contain a GU and an AG dinucleotide at the 5'- and 3'-splice sites, respec- tively, the critical sequence elements are different. In vertebrates, a polypyrimidine tract is required between the branch point and the 3'-splice site whereas in yeast, a conserved UACUAAC sequence is found in the branch site of the intron. In plants, introns are found to be AU-rich (between 70 and 80%) and the branch site selection is not absolutely fixed. Introduction of GC-rich sequences within plant introns reduces splicing efficiency dramatically. A minimum length of 70-73 nucleotides (nt) for efficient splicing has 336 ERIC LAM also been defined for monocots and dicots. The fact that some monocot introns are not spliced efficiently in dicots indicates that there may also be species-specific mechanisms for intron recognition and processing (Keith & Chua, 1986). Studies of the mechanism of replication for certain plant RNA viruses have uncovered a third type of RNA self-splicing reaction. In the repli- cation cycle of the avocado sunblotch viroid and the satellite RNAs of tobacco ringspot virus, for example, RNA transcripts are initially synthe- sized as concatamers. Subsequent autocataly- tic cleavage then occurs without the requirement of any cofactors to generate unit-length RNA molecules. From deletion and mutational analyses, it has been shown that only a small region of these self-splicing viral RNAs is required for the catalytic activity. This region can potentially form a secondary structure that has been called a hammerhead. On the basis of this information, a 19-nucleotide RNA molecule has been demon- strated to contain the ability to catalyze sequence- specific RNA cleavage (Uhlenbeck, 1987). This type of catalytic RNA molecule has been dubbed ribozymes and will likely find broad applications in basic and applied biology. Two relatively new discoveries in RNA proces- sing are trans-splicing and RNA editing. Trans- splicing involves the joining of two separately transcribed RNA molecules. This phenomenon has been found in trypanosomes, Euglena, nema- todes, plant mitochondria and the chloroplast of Chlamydomonas reinhardtii and plants. Trans- splicing of nematode RNAs has also been demon- strated in mammalian cells (Bruzik & Maniatis, 1992), which indicates that the components of this reaction may be widely conserved. Based on systems that have been well characterized, such as that of the nematodes and trypanosomes, it is thought that trans-splicing involves reactions that are similar to those of group II introns of nuclear pre-mRNAs (Agabian, 1990). Similarities include: conserved GU and AG dinucleotides at the 5'- and 3'-splice sites, requirement of snRNAs, and for- mation of a Y-shaped intermediate that is struct- urally analogous to the lariat RNA. One of the most well-characterized trans-splicing systems is the assembly of the transcript for the gene encoding one of the reaction center subunits of photosystem I {psaA) in C. reinhardtii. Unlike in plant chloroplasts that have been studied to date, the coding region of psaA is transcribed in three separate units that are widely dispersed in the chloroplast genome of C. reinhardtii. Mutagenesis analysis has demonstrated that at least 14 nuclear genes and one chloroplast-encoded gene, tscA^ are required for the proper ^r^w5-splicing of psaA transcripts (Goldschmidt-Clermont et al., 1990, 1991). tscA appears to be a small chloroplast RNA that is required for the first step of the trans- splicing reaction that joins introns 1 and 2 of the psaA transcript. Future characterization of this system is likely to uncover new insights into this exciting new area of biology. From the complete sequence of the plastid genome of tobacco, rice and liverwort, the rpsll gene encoding a riboso- mal protein in the chloroplasts also appears to require trans-splicing for its assembly. The nadl gene, encoding subunit 1 of the NADH dehydro- genase complex, of Petunia^ wheat and Oenothera mitochondria, has been found to require at least three ^mws-splicing and one ds-splicing events to complete the assembly of five different exons (Chapdelaine 6c Bonen, 1991; Sutton et al., 1991; Wissinger et al., 1991). Thus, trans-splicing events are likely important reactions that take place in different compartments of the plant cell. It has long been assumed that the sequence information located in the DNA level will be directly transferred to the RNA during transcrip- tion. Although RNA is known to be modified subsequent to their synthesis, such as methylation, these reactions usually do not alter the informa- tion inherent in the nucleotide sequence. It is thus a surprise when discrepancies between the geno- mic DNA sequences and mRNA sequences were discovered in parasitic protozoans such as try- panosomes and later in other systems such as plant mitochondria and chloroplasts. RNA editing in the kinetoplastid mitochondria of trypanosomes has been found to involve the insertion or deletion of U residues at multiple sites along the coding region of mRNAs (Simpson, 1990). In plant mitochon- dria, the nadl gene product of petunia and wheat mitochondria as well as the coxll transcript of petunia, wheat, pea and maize mitochondria has been found to be edited (Hiesel et al., 1989; Sutton et al.y 1991). In the study with petunia mitochon- dria, evidence was obtained that indicates editing can proceed before the completion of cis- or trans- splicing. A total of 14 different C-to-U changes were found along the coxll coding region of petunia. Each of these changes causes an alteration in the amino acid encoded at that position. In the wheat mitochondrial nadl transcript, multiple editing sites are also found and most strikingly, the initiation codon for protein synthesis is created by RNA editing. Thus, a C-to-U change is required to generate the proper start site of translation. Evidence for RNA editing in higher plant chloro- plasts has been reported recently. In maize, a single C-to-U conversion is required to generate the NUCLEIC ACIDS AND PROTEINS 339 In monocots, a set of 38 frequently used codons are found and two classes of genes are revealed from the analysis. One class has the same U or A preference in the third position as that of dicots whereas the other class shows a clear bias toward codons with C and G in this position. Another observation is the avoidance of XCG and XUA codons, where X is any nucleotide. This may be the result of selective pressure during evolution of the genome. For example, the -CG- dinucleotide is a possible substrate for DNA- or RNA-methylases the action of which may lead to gene inactivation. In the organellar genomes of plants, codons ending in A or U are apparently preferred. Since optimal translation of the mRNA has been correlated wih the use of the proper codon bias of the particular system, this aspect of the gene structure is an important consideration for the transfer of genetic information from one source to another. An important reaction in the process of decod- ing mRNA sequences is the attachment, or charging, of the correct amino acid to each tRNA species with different anticodons. The fidelity of this process is vital to the proper translation of nucleotide sequences into proteins. Since tRNAs all have similar 3-dimensional structures, the specificity of the aminoacyl-tRNA synthetases that carry out this reaction must depend at least in part on subtle sequence differences among each tRNA. Interestingly, it has been found that multiple tRNAs, called isoacceptors, that have the same amino acid specificity but have different anticodons, are aminoacylated by the same aminoacyl-tRNA synthetase. This fact implies that this class of enzymes must be able to recognize common features in the isoacceptors. In recent years, mutagenesis experiments with the E. colt tRNA"'" have demonstrated that a single base- pair (G-U) in the acceptor helix of the tRNA determines the specificity of the aminoacyl- tRNA"'" synthetase (Hou et al., 1989). In fact, a synthetic RNA 'microhelix' with only seven base- pairs and without the anticodon can be properly aminoacylated if this major specificity determinant is included. In other tRNAs, notably the £. coli tRNA''"'and tRNA"'^^ the specificity determinants apparently reside within their anticodons since the simple exchange of these three bases between the two tRNAs causes a switch in their amino acid identities. Thus, distinct structural features in the different tRNAs may be involved in their recogni- tion by the proper aminoacyl-tRNA transferase. Lastly, overexpression of the Gln-tRNA synthetase in E. coli has been reported to alter its tRNA specificity and glutamine is found to be incorrectly acylated onto another tRNA (Swanson et al., 1988). This observation suggests that the relative level of tRNAs and the aminoacyl-tRNA trans- ferases are also important parameters in the specificity of their interactions. 8.4.2 Translation The process of translation takes place in a large, multimeric protein complex that is known as the ribosome. Ribosomes are found in both soluble and membrane-bound forms. They have been found to be associated with the endoplasmic reticulum (ER), thus giving rise to the so-called 'rough ER', and the thylakoid membranes of the chloroplasts. These membrane-associated ribo- somes are thought to facilitate in part the synthesis and assembly of proteins that are destined either for secretion or membrane-localized. The exact structural components and inhibitor sensitivity are different between ribosomes of eukaryotic and prokaryotic origins. In general, ribosomes consist of two subunits each comprising more than 50 polypeptides. In tobacco, for example, more than 70 distinct polypeptides associated with the 40S subunit and 47-50 with the 60S subunit of the cytoplasmic ribosome can be identified (Capel & Bourque, 1982). In addi- tion to the enormous number of distinct proteins, about half the weight of the ribosomes is con- tributed by the rRNA components. The eukaryotic type of ribosomes are typically characterized by a sedimentation coefficient of about 80S and is sensitive to cycloheximide. The prokaryotic ribo- somes, including those of the chloroplasts, have sedimentation coefficients of about 70S, are composed of fewer proteins, and are inhibited by chloramphenicol. In spite of these differences, the gross structural features of these ribosomes and many of the basic reactions involved in protein synthesis are remarkably conserved among differ- ent organisms. Ribosomes consist of two multimeric subunits that can dissociate and reassociate in vivo. Each of these two non-identical subunits contains a dif- ferent set of rRNAs that are likely to be involved in the organization of the associated polypeptides. mRNAs and tRNAs are associated with the small subunit whereas the formation of the peptide bonds between adjacent amino acids is catalyzed by the large subunit. There are three RNA binding sites in the small subunit that play important roles in translation. One of these sites binds to the mRNA that is to be translated while the other two, called P and A sites, interact with tRNAs 340 ERIC LAM that contain the proper anticodons for the parti- cular mRNA. The process of translation involves three distinct steps: initiation, elongation and termination. In the first step, multiple initiation factors (IFs) are required. These IFs are associated with the small ribosomal subunit in addition to an initiator tRNA"^^^ bound at the P site. To initiate transla- tion, this ribonucleoprotein complex has first to locate the proper AUG on the particular mRNA. In eukaryotic ribosomes, the cap structure at the 5' end of mRNAs is known to promote initial ribosome binding. Two essential eukaryotic IFs, eIF-4A and eIF-4B, act together as an ATP- dependent RNA helicase that will unwind second- ary structures at the 5' end of the mRNA and are thought to facilitate ribosome docking. An impor- tant factor eIF-2 is known to bind to and position the initiator tRNA on the ribosome. In some cases, modification of eIF-2 by phosphorylation has been found to control the rate of protein synthesis. After docking onto the 5' end of the mRNA, the small subunit of the ribosome is then thought to scan down the mRNA until an AUG codon is reached and translation can commence. The sequence requirement for an optimal initiation codon has been defined by mutational analyses. The consensus sequence found for animal transla- tion systems was ACCAUGG (Kozak, 1986) whereas that in plant systems was ACAAUGG (Lutcke et ai, 1987). Interestingly, efficient initia- tion in plant cells has been observed with codons that differ from AUG (Gordon et ai, 1992). This phenomenon has also been found in animal systems and thus implies that there may exist another undefined mechanism by which peptide chain initiation can occur in the cytosol. In contrast to the initiation process of eukaryotes, only three initiation factors, IF-1, -2 and -3, are required in E. colt. Moreover, instead of the 5' cap structure and ribosome scanning, the initiation site for translation is specified by a specific sequence about four to seven nucleotides upstream from the start site AUG. This initiator sequence, 5'-AGGAGG-3', is called the Shine-Dalgarno sequence and is known to base-pair with a conserved 5'-CCUCCU-3' sequence at the 3' end of the 16S rRNA located in the small subunit of prokaryotic ribosomes. In this way, multiple sites for translation initiation can be specified in polycistronic mRNAs, which are frequently found in prokaryotes, and protein synthesis for different gene products can be carried out simul- taneously on the same transcript. Although the 'ribosome scanning' model of translational initiation appears to hold true for most eukaryotic genes, a noteworthy exception has been reported by Pelletier & Sonenberg (1988). Working with the poliovirus RNA, they dis- covered that translation on this naturally uncapped, dicistronic transcript is initiated by ribosome binding to specific sequences within the transcript. Sequences required for this internal initiation process may be as many as 500 bp. Similar observations were also made with encephalomyocarditis RNA and thus this phenom- enon may be more widespread and a mechanism appears to exist in the cytoplasm to initiate trans- lation within a transcript. In the future, polycis- tronic mRNAs may yet be found in the cytosol. Once the initiation codon has been located by the smaller subunit of ribosomes, the large subunit is recruited with the help of eIF-5, a single polypeptide of 100-160 kDa. Concomitantly, a GTP that is bound to eIF-2 is hydrolyzed and the eIF-2GDP complex. Pi and eIF-3 are released from the small subunit. The catalysis of GTP hydrolysis by eIF-5 is a crucial step in the release of eIF-2 and eIF-3 as well as the joining of the two ribosomal subunits. If the GTPase reaction is inhibited by addition of non-hydrolyzable GTP analogs, these critical steps in translation are inhibited. The binding of the large ribosomal subunit and the release of the initiation factors thus complete the process of translation initiation. The final complex contains the two subunits of the ribosome, the Met-tRNA initiator and mRNA. The anticodon of the Met-tRNA is base-paired with the initiation codon AUG at the P-site whereas the A-site is empty. This ternary complex between mRNA, tRNA and ribosome is now ready for the elongation phase of translation. The process of translation elongation can be considered as a three-step cycle. In the first step, an appropriate aminoacyl-tRNA molecule interacts with the codon exposed at the A-site of the ribosome/mRNA complex. This reaction is mediated by the elongation factor EF-IQ,. The function of this factor is similar to that of eIF-2 and its prokaryote counterpart, EF-Tu. EF-l,^ first complexes with GTP before its association with aminoacyl-tRNAs. This ternary complex then interacts with the ribosome and the aminoacyl- tRNA is delivered to the A-site. Hydrolysis of the EF-la-associated GTP then leads to the formation of a E F - 1 Q ; G T P binary complex that dissociates from the ribosome. If this GTP hydrolysis reaction is inhibited by non-hydrolyzable GTP analogs, the subsequent reactions for peptide bond formation are inhibited. In this step of the cycle, the carboxyl group of the amino acid at the P-site is uncoupled from the tRNA and then joined by a newly formed NUCLEIC ACIDS AND PROTEINS 341 peptide bond to the amino acid at the A-site. This reaction is catalyzed by a peptidyltransferase activity that is thought to be associated with the rRNA molecule in the large ribosomal subunit. No soluble protein or cofactor is apparently required. The last step of this ribosome cycle involves the translocation of the tRNA from the A-site to the P- site. Simultaneously, the free tRNA molecule that used to occupy the P-site is released and the ribosome migrates on the mRNA by exactly three bases. This movement of the ribosome is driven by protein conformational changes that accompany hydrolysis of a molecule of GTP to GDP and Pi. An important factor involved in this process is EF-2 which is a target for modification by an inhibitor of eukaryotic protein synthesis called diphtheria toxin. A post-translationally modified histidine residue, called diphthamide, in EF-2 is specifically ADP-ribosylated by this toxin. The resultant EF-2 is unable to catalyze GTP hydroly- sis or ribosome translocation. The translocation of the ribosome to a new codon thus completes one cycle of the elongation reaction. The whole process requires about one twentieth of a second and for an average protein of 300 amino acid residues, about 15 s will be needed for its synthesis. Once the elongation phase of translation has begun and the ribosome starts to 'travel' down the mRNA, a new round of translational initiation with new ribosomes can take place. In this way, multiple ribosomes can be found to associate simultaneously with actively translated transcripts. These are called polysomes and can be isolated from free ribosomes based on the differences in their sedimentation coefficients. The final step in protein translation involves the termination of the ribosome cycle at the proper termination codon. As shown in Fig. 8.7, tRNAs with anticodons complementary to UAA, UAG and UGA are normally not present in eukaryotes. When the ribosome reaches one of these termina- tion codons, a protein called release factor (RF) binds to the A-site along with a molecule of GTP. The occupancy of the A-site stimulates the pep- tidyltransferase activity of the ribosome and the peptidyl-tRNA ester at the P-site is effectively hydrolyzed. This results in the release of the translated peptide and subsequently, the GTP bound to the RF protein is also hydrolyzed. Finally, RF dissociates from the ribosome and the free tRNA and ribosomal proteins are separated from the mRNA template. These com- ponents are then ready to participate in the synthesis of yet another protein. In rabbit reticu- locytes, RF has an apparent size of 110000 daltons and contains subunits of 55 000 daltons. Little is known about the in vivo regulatory mechanisms of translation in plants, although wheat germ extracts have been used for in vitro translation assays for many years. However, since the pattern of proteins in plant cytoplasmic ribosomes is quite similar to that of mammals and the green alga C. reinhardtii (Capel & Bourque, 1982), it seems likely that the basic function and properties of cytoplasmic ribosomes are well conserved through evolution. In recent years, the study of replication mechanisms in several plant RNA viruses have revealed transla- tional activation sequences. One of the best studied is a 68-base pair sequence called Q in the untranslated leader of tobacco mosaic virus. This sequence when fused upstream of heterologous transcription units can activate translation of the resultant mRNA in both plants and bacteria (Gallic et ai, 1987). Interestingly, the effects of the ^ sequence in bacteria is much more dramatic in the absence of a Shine-Dalgarno sequence in the transcript and its effects appear to be independent of its position relative to the initiation site (Gallic Sc Kado, 1989). The mechanism by which this cis- acting element enhances translation remains to be elucidated. An example of the more-characterized systems for study of translational regulation in eukaryotes is the transcript encoding ferritin, a protein conserved in eukaryotes (Klausner et al., 1993). This protein is involved in the sequestration of iron within cells and its synthesis is regulated by the concentration of free iron in the cytosol. A stem-loop structure, called IRE (Iron Responsive Element), at the 5'-untranslated region of the ferritin transcript has been shown to be involved in sequence-specific interaction with a cellular pro- tein, IRE-BP. Association of IRE-BP with the 5' region prevents translational initiation of the ferritin mRNA. The binding of IRE-BP to the ferritin transcript is inhibited when iron is bound to this protein, thus IRE-BP is displaced from the ferritin transcripts under high concentrations of iron and translation of ferritin is activated. Interestingly, a similar IRE and the same IRE-BP are used for the regulation of mRNA stability of another gene encoding transferrin receptor (TfR), a protein that is expressed under low intracellular concentrations of iron and is involved directly in iron transport. In this case, the IRE is situated at the 3'-untranslated region of the TfR transcript and the association of IRE-BP with this site prevents mRNA degradation. When the iron concentration within the cell rises, IRE-BP will dissociate from the IRE and TfR mRNA turnover will be accelerated. 344 ERIC LAM Moreover, the hydrolysis of ATP in the case of mitochondria occurs in the cytoplasmic side whereas that for chloroplast transport appears to take place inside the organelle. The binding of precursors to the chloroplasts also requires ATP whereas the interaction between mitochondria and the signal peptides appears to be spontaneous. Although these differences suggest that the process of protein recognition and transport may be different between these two organelles, a report by Hurt et al. (1986) suggests that there must also be common mechanisms involved. In this work, the transit peptide of C. reinhardtii RbcS (small subunit of ribulose-l,5-bisphosphate car- boxylase), a chloroplast-localized protein, was fused to mouse dihydrofolate reductase (a cyto- plasmic protein) and a truncated yeast mitochon- drial protein, subunit IV of cytochrome oxidase. The RbcS signal peptide was able to target both proteins into yeast mitochondria, although with less efficiency than an authentic mitochondrial signal peptide. These results argue that there must also be common functional properties between targeting signals for these organelles. The transport of proteins into mitochondria and chloroplasts may be cotranslational in vivo. Thus, as these proteins are being synthesized on cyto- plasmic ribosomes, their translocation into organ- elles may be initiated and both processes are carried out simultaneously. However, it is clear from in vitro experiments with isolated organelles that a cotranslation mechanism is not obligatory for the proper recognition and processing of these proteins. Upon uptake of the preproteins into the organelles, the signal peptide is cleaved by a peptidase activity. Several chloroplast proteins have been reported to be associated with this signal peptidase activity (Oblong & Lamppa, 1992). Their characterization may provide more information on the events and mechanisms involved in the processing of signal peptides. (ii) Translocation of proteins into the ER The endoplasmic reticulum (ER) is involved in the routing of many types of proteins within a cell to different destinations. Mechanistically, the two most well-characterized target compartments are the lumen of the ER and the vacuole. Like the case for mitochondria and plastids, the uptake of the preproteins into the ER involves a signal peptide at the N-terminal end. These are also short N- terminal extensions of about 13-30 amino acids long and are removed upon uptake into the ER by an endopeptidase. The composition of these ER- targeted signal peptides are similar to those for mitochondria and plastids as well. They contain basic regions near the amino terminus and also an abundance of hydrophobic residues. Once synthe- sized on cytoplasmic ribosomes, these signal peptides interact with a ribonucleoprotein called signal recognition particle (SRP) and an ER-bound receptor to initiate their transfer into the lumen of the ER. SRP has now been well-characterized in animals and yeast systems and is a protein complex of six different polypeptides. It contains a 7S RNA that appears to be highly conserved in eukaryotes, including plants. A 54kDa subunit of SRP is found to bind both the 7S RNA as well as the signal peptides destined for the ER. Upon binding to these signal peptides, SRP prevents any further translation of the mRNA by the associated ribosomes. This inhibition, however, is relieved once this complex is associated with a 'docking- protein' in the ER membrane. This protein is also known as the SRP receptor. Upon positioning the signal peptide at the SRP receptor site, SRP is released and translation of the mRNA is resumed with the concomitant translocation of the protein into the ER. This provides an elegant mechanism by which transfer into the ER is coupled with translation. The tight co-ordination between these two processes may be crucial for intrinsic mem- brane proteins that otherwise will not be soluble in the cytosol and thus will never reach the receptor sites. Once taken up in the ER, a protein is either secreted or is delivered to a specific subcellular compartment. For proteins known to be localized in the vacuoles, it has been shown that multiple signals are likely required for their proper target- ing. Thus, although their signal peptide is sufficient to affect ER-uptake when fused to a heterologous protein, the final product is secreted rather than targeted to the vacuoles. The specific structural properties of vacuolar proteins that mediate their localization remain to be deter- mined. For proteins that are retained on the ER, a common carboxy-terminal motif is found in eukaryotes. This motif consists of either KDEL or HDEL. Although the mode of action for this motif is not known, its introduction to the C-terminus of a vacuolar protein results in sub- stantial retention of the chimeric gene product in the ER (Chrispeels, 1991). This result implies that this motif may play an important role in the sorting process of proteins within the ER. In addition to acting as a 'point-of-departure' for many proteins that are either secreted or tar- geted to specific compartments within a cell, the ER is also a place where many of these proteins are glycosylated. Typically, an asparagine residue of NUCLEIC ACIDS AND PROTEINS 345 the protein that is being translated is modified by glycans, such as mannose, through the action of oHgosaccharyl transferases. A preferred site for this process is NHa-Asn-X-Thr/Ser-COOH. If the translocation process for a protein involves pas- sing through the Golgi apparatus, the N-linked glycans can be further modified by glycosidases and glycosyItransferases. This results in the link- age of different types of sugar moieties, such as fucose and xylose, to the simple glycans that were added to the polypeptide in the ER. Glycosylation of some proteins has been implicated in facilitating their transport through the ER to their destina- tions. These protein-bound glycan residues may also be involved in protein folding or the turnover of the protein. (iii) Molecular chaperones and protein folding The synthesis of a functional protein requires that the polypeptide produced by translation can be folded in the proper conformation. In order for the active structure to be stable under normal condi- tions, it is thought that this should represent a thermodynamically favored state, an energy mini- mum. This idea then predicts that given the proper conditions, a newly synthesized or denatured polypeptide should be able to self-assemble into a functional form of unique structure. Indeed, this is the case for some small proteins, such as ribonuclease A. However, for most proteins, the probability of incorrect interactions among differ- ent regions of the polypeptide can occur. This is even more problematic for multimeric proteins with different subunits that require precise assem- bly. In the last 10 years, a new class of proteins has emerged from studies on requirements for protein assembly in vivo (reviewed in Ellis &: van der Vies, 1991). These are called molecular chaperones and are defined as proteins that prevent incorrect interactions between parts of other protein(s), but are not involved in the final assembled protein or protein complex. The most well-characterized chaperone in plants is the chaperonin 60 (chap60) located in the plastids. This protein was first discovered in 1980 as a protein that is associated with the large subunit of RbcL after its synthesis in the chloroplasts. It was subsequently found to bind to a variety of newly synthesized proteins, includ- ing the small subunit of ribulose-l,5-bisphosphate carboxylase (rbcS) and the ^-subunit of the chloroplast ATPase. Interestingly, chap60 does not appear to bind to the mature processed form of rbcS, thus indicating that the signal peptide and/or an unfolded structure is required for chaperone interaction. Plastid chap60 consists of two kinds of subunit, a and /3, that are related to each other in protein sequence. In addition, they are homologous to the bacterial chap60 protein (also known as groEL). The functional character- istics of the bacterial chap60 protein are much better defined than its plastid homolog. In E. coli, another chaperonin, chap 10, is known to interact with chap60 to facilitate assembly of various proteins. However, chap60 alone has been shown to be capable of facilitating transport of proteins across membranes. One likely mode of action for chap 10 is that by binding to chap60, it affects the release of the properly folded polypeptide. Both bacterial chap60 and chap 10 are oligomeric proteins with 14 and 7 identical subunits each, respectively. They are known to be arranged in a seven-membered ring configuration. In plants, chap60 is encoded by a small gene family and the expression of these genes is enhanced by light as well as heat shock. The bacterial and plastid chap60s have been shown to mediate the folding of denatured rbcL and rbcS subunits from Rhodo- spirillum rubrum. Elucidation of the mechanisms and components involved in chaperone function will provide new information on how protein assembly is carried out in vivo. (b) Regulation of proteins by covalent modification Post-translational modification of proteins plays an -important role in regulatory pathways of all cells. Covalent modifications such as phosphory- lation, acylation, acetylation and carboxymethyla- tion are some of the well-known examples. Pro- teins can also be regulated by processing events that are carried out by highly specific proteases. This section discusses two of the more widely studied protein modifications that are known to be involved in the modulation of protein functions. (i) Phosphorylation This is probably the best characterized and most widely applied mode of protein modification. A particularly versatile feature of protein phosphory- lation is its reversible nature, which is achieved by the competitive action of two classes of enzymes, kinases and phosphatases. In vivo, the sensitivity of these enzymes to external and internal signals allows for rapid as well as long-term cellular responses. The demonstration that the activity of some protein kinases and phosphatases are them- selves regulated by phosphorylation suggests that a cascade of phosphorylation events is involved in the transduction of extracellular signals or 346 ERIC LAM developmental decisions. Elegant examples of these phosphorylation cascades are those involved in the onset of mitosis in eukaryotes (Nurse, 1990) and the activation of oncogenesis in animal cells by mitogens (Roberts, 1992). Tw^o main types of protein kinases are known in eukaryotes: those that phosphorylate serine and threonine residues, and those that modify tyrosine residues. A phosphate group is attached by a phosphodiester bond to the hydroxyl oxygen of the R-group in these amino acids. The specificity of each kinase is governed by the particular amino acid sequence surrounding the target residue. The actions of these kinases are counteracted by two classes of phosphatase that act on phosphoserine/phospho- threonine and phosphotyrosine, respectively. In bacteria, phosphorylation of histidine and aspartic acid residues is used for the regulation of such diverse functions as chemotaxis and osmoregula- tion of transcription (Stock et ai, 1989). In higher plants, the study of protein phosphory- lation has been well documented (Ranjeva & Boudet, 1987; Roberts & Harmon, 1992). Protein phosphorylation responsive to stimuli such as light and phytohormones has been observed, although in most cases their functional significance remains unclear. However, there are a number of examples where the role of protein phosphorylation has been well characterized in plants. The multi- subunit enzyme pyruvate dehydrogenase (PDC), involved in the oxidative decarboxylation of pyruvate to acetyl-CoA, is a key enzyme for metabolite control. This enzyme complex has been purified from plant mitochondria and plas- tids. Like its counterpart in animal systems, phosphorylation of PDC leads to deactivation of this enzyme and a decrease in acetyl-CoA. The major light-harvesting chlorophyll binding protein in the chloroplasts, LHCPII, is also known to be reversibly phosphorylated at a threonine residue located in its N-terminus. In this case, phosphorylation is catalyzed by a thylakoid-bound kinase that is regulated by the redox potential of the plastoquinone (PQ) pool of the chloroplast membrane. This kinase is activated when the PQ pool is reduced, such as the case under high light intensities. The phosphorylation status of LHCPII apparently modulates the distribution of excita- tion energy between photosystems (PS) I and II complexes involved in photosynthesis. More light energy is diverted to PSI when LHCPII is phosphorylated, thus enhancing the rate at which the reduced PQ pool can be oxidized and electrons transferred to ferredoxin. A thylakoid-bound phosphatase is involved in the de-excitation pro- cess. This enzyme, however, appears to be constitutive and is not affected by the redox state of PQ. Several protein kinases have been identified and purified from thylakoid membranes. It remains to be demonstrated unequivocally which enzyme corresponds to that which actually phos- phorylates LHCPII in vivo. In the last few years, many plant genes that encode protein kinases or putative kinases have been reported. These are typically obtained by one of two methods. The first is a biochemical approach in which the enzyme activity of interest is purified to homogeneity. The corresponding gene is then cloned by using either degenerate oligonucleotides synthesized according to the partial protein sequence, or antibodies raised against the protein (Roberts & Harmon, 1992). The second is a direct screening procedure using a set of degenerate oligonucleotides that is based on the conserved sequences of protein kinases studied so far (Lawton et al., 1989). The study of calcium-dependent protein kinases (CDPKs) is especially interesting since many cellular pro- cesses in plants appear to involve calcium as a signal transduction intermediate. These include phytochrome-dependent seed germination and abscisic acid-induced stomatal closure. CDPK- like enzyme activities have been found in a variety of plants and algal systems. In general, they appear to be monomers of 40-90 kDa. The enzyme binds calcium and is activated by 50-100-fold. In contrast, calmodulin or phospho- lipids have no obvious effects on CDPK activities, unlike protein kinase C or the Ca^^/calmodulin- dependent kinases of animal systems. Like most kinases that have been studied, CDPK can also catalyze at a slow rate autophosphorylation on certain of its own serine and threonine resi- dues. The genes for CDPK of soybean, carrot and Arabidopsis have been cloned recently. They appear to be members of a multigene family in these plants and their sequences show significant similarities to the catalytic domain of Ca^"^/ calmodulin-dependent protein kinases of animal systems. Moreover, the deduced amino acid sequence of the cDNAs shows that the car- boxyl-terminus of CDPK is homologous to the calcium binding domain of calmodulin. This unique feature is consistent with the Câ "*"- dependent, but calmodulin-independent behavior of CDPKs. Although CDPKs are now well- characterized at the molecular level, their in vivo role and mechanism of regulation remains unknown. Future studies using a combination of biochemical and molecular approaches will hopefully elucidate the function of this class of proteins. NUCLEIC ACIDS AND PROTEINS 349 otherwise identical. This work introduced the so- called N-end rule for protein stability. The amino acids M, S, A, T, V and G (in the one letter amino acid code), when placed at the N-terminus of the bacterial protein /?-galactosidase, are found to result in stable proteins with half-lives of greater than 20 h. When substituted with the other amino acids at this position, half-lives of 30min to less than 2 min were found. These results demonstrate that the identity of the N-terminal amino acid is a critical signal for the destabilization of proteins. More recent work has shown that a second determinant of the N-end rule is a critical lysine residue within the protein itself. No specific sequence context for this critical lysine is appar- ently required. The optimal location of this lysine residue is within a relatively disordered region that is close to the N-terminal domain. The lack of a specific amino acid sequence that is required for the recognition of this lysine residue and the PEST sequence by the targeting components is remini- scent of the situation with transit peptides for protein translocation. In both cases, the structural characteristics of the peptide rather than its specific sequence act as the determinant of specifi- city. For the N-end rule, a family of related proteins called E3 is known to recognize and bind to the destabilizing amino acids at the N-terminus. Subsequently, the interaction with the internal lysine residue then targets the protein for degrada- tion by a proteolysis pathway that involves a protein component called ubiquitin (Bachmair &: Varshavsky, 1989). This pathway of targeted proteolysis is discussed in more detail in the following section. (ii) Ubiquitin: a universal targeting system of protein turnover Ubiquitin is a 76 amino acid-long polypeptide that is found with few changes in all eukaryotes examined to date. Its amino acid sequence is one of the best conserved in nature and differs by only two to three residues between yeast, mammals and plants. This protein can be detected in both the nucleus and the cytoplasm. However, it does not appear to be in the mitochondria or plastids. The role of ubiquitin in protein turnover was first elucidated in animal systems and subsequently verified in other eukaryotic systems, including plants (reviewed in Vierstra, 1989). Essentially, the main function of ubiquitin is to 'tag' the proteins which are destined to be proteolyzed quickly in the cytosol. It achieves this goal by covalently attach- ing its C-terminal glycine carboxyl group to the e-NH2 group of the N-terminal or internal lysine residues of the target protein. Once attached, the protein-ubiquitin conjugate is then degraded rapidly by an ATP-dependent protease complex and free ubiquitin is also released in the process. In order to participate in the conjugation reaction, free ubiquitin needs to be activated by an enzyme. E l . El catalyzes the ATP-dependent adenylation of the C-terminus of ubiquitin. Ubiquitin is first linked to El via a thiol ester bond and is then transferred to a carrier protein called E2 by a transesterification reaction. The E2-ubiquitin conjugate is then used as a substrate for the ligation of ubiquitin to lysine residues in proteins complexed with E3. This is catalyzed by a ubiquitin-protein lyase. Interestingly, this enzyme can also carry out the opposite reaction of specifically cleaving the amide linkage between ubiquitin and the target protein. This reverse reaction may be important in the regulation and proof-reading of the pathway. In any case, this lyase does not require ATP, unlike the ubiquitin- conjugate protease complex. The ubiquitin-dependent pathway of protein turnover appears to be the primary avenue of targeted proteolysis in cells. Greater than 90% of short-lived proteins have been shown to be degraded via the ubiquitin conjugate pathway in a mouse cell line. In plants, the photoreceptor phytochrome has been found to be proteolyzed in its active form via an ubiquitinated intermediate. However, the mechanisms involved in the differ- ential targeting of plant cell proteins for this and other pathways of proteolysis remain to be eluci- dated. Little is known about the mechanism for protein catabolism in plastids and mitochondria. A nuclear mutant of maize shows an accelerated rate of turnover for two polypeptides associated with photosystem II (Leto et aL, 1985). This suggests that there are likely to be nuclear-encoded components which are involved in the regulation of protein stability in the plastids. Although we are beginning to understand more about protein turnover in plants, it is quite clear that much more fundamental questions will need to be addressed in the near future in order for us to really appreciate the intricacies of the pathways involved. 8.5 SUMMARY In writing this chapter, I have tried to cover as much as possible of the general aspects of nucleic acids and proteins, with emphasis on regulatory mechanisms. Data obtained in various plant systems are compared with those of other eukaryotes to illustrate the striking conservation 350 ERIC LAM of many pathways and modes of regulation. Thus, the beauty of the apparent complexity of an organism lies within the simple principles that are reiterated in many different forms and used in a combinatorial fashion. This is illustrated in the common architecture of transcription factors and promoters, the role of RNA in translation and splicing, and the structural properties of peptides as determinants for protein transport and degra- dation. With a combination of molecular genetics and biochemical approaches, and applying the tools and knowledge gained from other systems, advances in plant biochemistry will more than likely increase its pace in the near future. REFERENCES Agabian, N. (1990) Cell 61, 1157-1160. Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. & Watson, J.D. (1989) Molecular Biology of the Cell, 2nd edn. Garland Publishing, New York. Allison, D.S. & Schatz, G. (1986) Proc. Natl. Acad, Set. U.S.A. 83, 9011-9015. Allison, L.A., Moyle, M., Shales, M. & Ingles, J.C. (1985) Cell 42, 599-610. Anderson, S., Bankier, A.T., Barrell, B.G., de Bruijn, M.H.L., Coulson, A.R., Drouin, J., Eperon, I.C., Nierlich, D.P., Roe, B.A., Sanger, F., Schreier, P.H., Smith, A.J.H., Staden, R. & Young, I.G. (1981) Nature 290, 457-465. Arrigo, A.-P., Tanaka, K., Goldberg, A.L. & Welch, W.J. (1988) Nature 331, 192-195. Bachmair, A. & Varshavsky, A. (1989) Cell 56, 1019- 1032. Bachmair, A., Finley, D. & Varshavsky, A. (1986) Science 234, 179-186. Baeza, L., Bertrand, A., Mache, R. & Lerbs-Mache, S. (1991) Nucleic Acids Res. 19, 3577-3581. Baker, B.S. (1989) Nature 340, 521-524. Bernstein, P. & Ross, J. (1989) Trends Biochem. Sci. 14, 373-377. Blank, A. & McKeon, T.A. (1991) Plant Physiol. 97, 1409-1413. Boynton, J.E., Gillham, N.W., Harris, E.H., Hosier, J.P., Johnson, A.M., Jones, A.R., Randolph-Anderson, B.L., Robertson, D., Klein, T.M., Shark, K.B. & Sanford, J.C. (1988) Science 240, 1534-1538. Bruce, B.D. & Malkin, R. (1991) Plant Cell 3, 203-212. Bruzik, J.P. & Maniatis, T. (1992) Nature 360, 692-695. Campbell, W.H. & Gowri, G. (1990) Plant Physiol. 92, 1-11. Capel, M.S. & Bourque, D.P. (1982) / . Biol. Chem. ISl, 7746-7755. Cardenas, M.E., Dang, Q., Glover, C.V.C. & Gasser, S.M. (1992) EMBO J. 11, 1785-1796. Carignani, G., Groudinsky, O., Frezza, D., Schiavon, E., Bergantino, E. & Slonimski, P.P. (1984) Cell 35, 733-742. Carrington, D.M., Auffret, A. & Hanke, D.E (1985) Nature 313, 64-67. Cech, T.R. (1991) Cell 64, 667-669. Chapdelaine, Y. & Bonen, L. (1991) Cell 65, 465-^72. Chen, H-C. & Stern, D.B. (1991) Mol. Cell. Biol. 11, 4380-4388. Chrispeels, M.J. (1991) Annu. Rev. Plant Physiol. Mol. Biol. 42, 21-53. Coen, E.S. (1991) Annu. Rev. Plant Physiol. Plant Mol. Biol. 42, 241-279. Croston, G.E., Kerrigan, L.A., Lira, L.M., Marshak, D.R. & Kadonaga, J.T. (1991) Science 251, 643-649. Danon, A. & Mayfield, S.P. (1991) EMBO J. 10, 3993-4001. Dehesh, K., Hung, H., Tepperman, J.M. & Quail, P.H. (1992) EMBO } . 11, 4131-4144. Deng, X-W. & Gruissem, W. (1987) Cell 49, 379-387. Deng, X-W., Wing, R.A. & Gruissem, W. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 4156-4160. Dickey, L.F., Gallo-Meagher, M. & Thompson, W.F. (1992) EMBO J. 2311-2317. Doi, R.H. & Wang, L.-F. (1986) Microbiol. Rev. 50, 227-243. Dombroski, A.J., Walter, W.A., Record, Jr, T.M., Siegele, D.A. & Gross, C.A. (1992) Cell 70, 501-512. Dooner, H.K., Keller, J., Harper, E. & Ralston, E. (1991) Plant Cell 3, 473-482. Dunham, V.L. & Bryant, J.A. (1985) In The Cell Division Cycle in Plants (J.A. Bryant & D. Francis, eds), pp. 37-59. Cambridge University Press, London. Durrenberger, F. & Rochaix, J-D. (1991) EMBO J. 10, 3495-3501. Echeverria, M., Delcasso-Tremousaygue, D. & Delseny, M. (1992) Plant J. 2, 211-219. Ellis, J.R. & van der Vies, S.M. (1991) Annu. Rev. Biochem. 60, 321-347. Ferl, R.J. & Paul, A-L. (1992) Plant Mol. Biol. 18, 1181-1184. Flavell, R.B., O'Dell, M., Thompson, W.F., Vincent, M., Sardana, R. & Baker, R.F. (1986) Phil. Trans. R. Soc. London, B 314, 386-397. Gallego, C , Gupta, S.K., Winitz, S., Eisfelder, B.J. & Johnson, G.L. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 9695-9699. Gallie, D.R. & Kado, C.I. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 129-132. Gallie, D.R., Sleat, D.E., Watts, J.W., Turner, P.C. & Wilson, T.M.A. (1987) Science 236, 1122-1124. Gallie, D.R., Lucas, W.J. & Walbot, V. (1989) Plant Cell 1, 301-311. Gasser, S.M., & Laemmli, U.K. (1987) Trends Genet. 3, 16-22. Gill, G. (1992) Curr. Biol. 2, 565-567. Gilmartin, P.M., Memelink, J., Hiratsuka, K., Kay, S.A. & Chua, N-H. (1992) Plant Cell 4, 839-849. Goldschmidt-Clermont, M., Girard-Bascou, J., Choquet, Y. & Rochaix, J-D. (1990) Mol. Gen. Genet. 223, 417-425. Goldschmidt-Clermont, M., Choquet, Y., Girard- Bascou, J., Michel, F., Schirmer-Rahire, M. 6c Rochaix, J-D. (1991) Cell 65, 135-143. NUCLEIC ACIDS AND PROTEINS 351 Goodall, G.J. & Filipowicz, W. (1989) Cell 58, 473-483. Goodall, G.J. & Filipowicz, W. (1990) Plant Mo/. Biol. 14, 727-733. Gordon, K., Futterer, J. & Hohn, T. (1992) Plant J. 2, 809-813. Goto, T. & Wang, J.C. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 7178-7182. Gralla, J.D. (1989) Cell 57, 193-195. Grierson, D. & Covey, S.N. (1988) Plant Molecular Biology, 2nd edn. Blackie, Glasgow. Griffin, L.C. & Dervan, P.B. (1989) Scince 245, 967-971. Gruissem, W. (1989) In The Biochemistry of Plants, vol. 15 (P.K. Stumpf & E.E. Conn, eds), pp. 151-191. Academic Press, New York. Gruissem, W., Elsner-Menzel, C , Latshaw, S., Narita, J.O., Schaffer, M.A. & Zurawski, G. (1986) Nucleic Acids Res. 14, 7541-7556. Grunstein, M. (1992) Sci. Am. 267, 68-74B. Haley, J. & Bogorad, L. (1990) Plant Cell 2, 323-333. Hall, Jr, G., Allen, G.C., Loer, D.S., Thompson, W.F. & Spiker, S. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 9320-9324. Hanley-Bowdoin, L. & Chua, N-H. (1988) Trends Biochem. Sci. 12, 67-70. Harrison, S.C. (1991) Nature 353, 715-719. Hiesel, R., Wissinger, B., Schuster, W. & Brennicke, A. (1989) Science 246, 1632-1634. Hoch, B., Maier, R.M., Appel, K., Igloi, G.L. & Kossel, H. (1991) Nature 353, 178-180. Hou, Y-M., Francklyn, C. & Schimmel, P. (1989) Trends Biochem. Sci. 14, 233-237. Huez, G., Marbaix, G., Gallwitz, D., Weinberg, E., Devos, R., Hubert, E. & Cleuter, Y. (1978) Nature 271, 572-573. Hunter, T. & Karin, M. (1992) Cell 70, 375-387. Hurt, E.G., Slotanifar, N., Goldschmidt-Clermont, M., Rochaix, J-D. & Schatz, G. (1986) EMBO J. 5, 1343-1350. Inglese, J., Koch, W.J., Caron, M.G. & Lefkowitz, R.J. (1992) Nature 359, 147-150. Izban, M.G. & Luse, D.S. (1992) Genes Dev. 6, 1342-1356. Jantzen, H-M., Admon, A., Bell, S.P. & Tjian, R. (1990) Nature 344, 830-836. Jolly, S.O. & Bogorad, L. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 822-826. Kang, C., Zhang, X., Ratliff, R., Moyzis, R. & Rich, A. (1992) Nature 356, 126-131. Katagiri, F. & Chua, N-H. (1992) Trends Genet. 8, 22-27. Katagiri, F., Yamazaki, K., Horikoshi, M., Roeder, R.G. & Chua, N-H. (1990) Genes Dev. 4, 1899-1909. Keegstra, K., Olsen, L. & Theg, S.M. (1989) Annu. Rev. Plant Physiol. Plant Mo/. Biol. 40, 471-501. Keith, B. & Chua, N-H. (1986) EMBO J. 5, 2419-2425. Kieber, J.J., Tissier, A.F. ^ Signer, E.R. (1992) Plant Physiol. 99, 1493-1501. Klausner, R.D., Rouault, T.A. 6c Harford, J.B. (1993) Cell 72, 19-28. Klein, R.R. & Mullet, J.E. (1990) / . Biol. Chem. 265, 1895-1902. Kobayashi, H., Ngernprasirtsiri, J. & Akazawa, T. (1990) EMBO ;. 9, 307-313. Kolodner, R. 6c Tewari, K.K. (1975) Biochim. Biophys. Acta 402, 372-390. Kozak, M. (1986) Cell 44, 283-292. Kudla, J., Igloi, G.L., Metzlaff, M., Hagemann, R. & Kossel, H. (1992) EMBO /. 11, 1099-1103. Lam, E. & Chua, N-H. (1987) Plant Mo/. Biol. 8, 415^24. Lam, E., Hanley-Bowdoin, L. & Chua, N-H. (1988) / . Biol. Chem. 263, 8288-8293. Lawton, M.A., Yamamoto, R.T., Hanks, S.K. &: Lamb, C.J. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 3140-3144. Lerbs, S., Brautigam, E. & Mache, R. (1988) Mol. Gen. Genet. 211, 459-464. Leto, K.J., Bell, E. & Mcintosh, L. (1985) EMBO J. 4, 1645-1653. Levings, C.S. (1983) Cell 32, 659-661. Lin, Y-S., Carey, M., Ptashne, M. 6c Green, M.R. (1990) Nature 345, 971-981. Little, P. (1992) Nature 359, 367-368. Lonsdale, D.M., Hodge, T.P. & Fauron, C.M-R. (1984) Nucleic Acids Res. 12, 9249-9261. Luderus, E.M.E., de Graaf, A., Mattia, E., den Blaauwen, J.L., Grande, M.A., de Jong, L. & van Driel, R. (1992) Cell 70, 949-959. Lutcke, H.A., Chow, K.C., Mickel, F.S., Moss, K.A., Kern, H.F. & Scheele, G.A. (1987) EMBO J. 6, 43-48. Ma, H., Yanofsky, M.F. & Meyerowitz, E.M. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 3821-3825. Maier, R.M., Hoch, B., Zeltz, P. & Kossel, H. (1992) Plant Cell 4, 609-616. Marshallsay, C , Connelly, S. & Filipowicz, W. (1992) Plant Mol. Biol. 19, 973-983. Martinez-Zapater, J.M., Estelle, M.A. & Somerville, C.R. (1986) Mol. Gen. Genet. 204, 417-423. Mattoo, A.K. & Edelman, M. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 1497-1501. Mayfield, S.P. (1990) Curr. Opin. Cell Biol. 2, 509-513. McGurl, B., Pearce, G, Orozco-Cardenas, M. & Ryan, C.A. (1992) Science 255, 1570-1573. Mitchell, P.J. & Tjian, R. (1989) Science 245, 371-378. Murray, M.G., Cuellar, R.E. & Thompson, W.F. (1978) Biochemistry 17, 5781-5790. Nikolov, D.B., Hu, S-H., Lin, J., Gasch, A., Hoffmann, A., Horikoshi, M., Chua, N-H., Roeder, R.G. & Burley, S.K. (1992) Nature 360, 40-46. Nurse, P. (1990) Nature 334, 503-508. Oblong, J.E. & Lamppa, G.K. (1992) EMBO /. 11, 4401-4409. Oda, K., Yamato, K., Ohta, E., Nakamura, Y., Takemura, M., Nozato, N., Akashi, K., Kanegae, T., Ogura, Y., Kohchi, T. & Ohyama, K, (1992) Plant Mol. Biol. Rep. 10, 105-111. Ohad, I., Kyle, D.J. & Hirschberg, J. (1985) EMBO /. 4, 1655-1659. Palme, K., Diefenthal, T., Vingron, M., Sander, C. & Schell, J. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 787-791.