INTERNATIONAL SCIENCE

 - THE BIOGENESIS OF MEMBRANES AND ORGANELLES (*) -
- Part I I-
David D. Sabatini and Milton B. Adesnik
Department of Cell Biology, New York University School of Medicine, New York, NY 10016-6481
MICROSCOPEYeast as a Model Organism to Study Intracellular Protein Trafficking (18, 18a)
Although the emphasis here is on membrane and organelle biogenesis in the mammalian cell, it will frequently be necessary to discuss information on these processes obtained from studies with the yeast Saccharomyces cereviseae. This organism serves as a useful and widely employed model to study fundamental functions of the eukaryotic cell since with it it is possible to combine the powerful approach of molecular genetics with physiological and biochemical analyses. Using a variety of elegant but relatively simple selection procedures, a large number of conditionally lethal (temperature sensitive) yeast mutants have been identified which, at the non permissive temperature, fail to grow because they are defective in various steps along the pathway of protein secretion, or in the transport of newly synthesized proteins to a given organelle.
The availability of such mutants has permitted the cloning of the respective genes by complementation approaches in which mutant cells are transfected with a wild type yeast genomic library and transformants containing a plasmid with the gene of interest are selected by virtue of their growth at the restrictive temperature. In addition, the isolation of extragenic suppressor mutations has led to the identification of other genes whose products, usually when overexpressed or mutated, can compensate for the defect caused by the original mutation. The products of suppressor genes function in the same pathway and probably interact with the product of the original mutant gene. These and other genetic approaches, have facilitated the elucidation of the role of many gene products within the protein targeting and transport machinery of the cell. Moreover, analysis of the sequence of a cloned yeast gene has sometimes revealed the existence of a mammalian homologue and, in some instances, it has even been possible to correct the yeast defect by expressing in the yeast cell a gene encoding the mammalian protein (e.g. 19). Conversely, in some cases, the yeast product has been shown to substitute for the mammalian one in an in vitro transport reaction (20). Yeast homologues of mammalian proteins involved in transport have also been identified, or cloned, on the basis of their expected sequence homology to already known mammalian proteins (e.g. the clathrin heavy and light chains, and the ADP-ribosylation factor, Arf1p; see 18). Finally, the essentiality of a specific yeast protein whose gene has been cloned can always be ascertained from the viability of cells in which the gene has been "knocked out" (disrupted) by homologous recombination procedures, which are easy to carry out in this organism.
Studies on a large set of mutants defective in secretion, designated sec mutants*, have illuminated aspects of essentially all steps of transport along the endomembrane system. Sec mutants were first recognized because they accumulated precursor forms of exported proteins and it is now clear that some of the corresponding gene products are required for protein insertion into the ER (e.g. Sec61p, Sec62p, Sec63p, Sec65p), while others participate in transport from the ER to the Golgi apparatus (e.g. Sec12p, Sec13p, Sec17p, Sec18p, Sec20p, Sec21p, and Ypt1p), through the Golgi cisternae (e.g. Sec7p, Sec14p, Arf1p, Arf2p), or from the Golgi to the cell surface (e.g. Sec2p, Sec4p, Sec15p). A very large number of mutants (vps) defective in transport to the vacuole, the yeast equivalent of the mammalian lysosome, are also available, as are mutants in genes (ERD1 and ERD2) required for the retention in the ER of proteins that normally reside in this organelle. Other classes of mutants have also been isolated which are defective in the importation of proteins into the nucleus, mitochondria or peroxisomes.
GTP-Binding Proteins Control Many Steps Along the SecretoryPathway(21-27)
A very important class of yeast genes (e.g. YPT1 and SEC 4) involved in protein traffic encode proteins that bind and hydrolyze GTP (guanine nucleotide binding proteins). Studies of the corresponding mutants have contributed greatly to focus attention on the role of GTP-binding proteins as "molecular switches" that control the directionality of a wide variety of individual steps in intracellular protein transport.
GTP-binding proteins constitute a superfamily that includes: the heterotrimeric (Gabg) G proteins that transduce extracellular hormonal and sensory signals into intracellular changes; the protein synthesis elongation factor EF-Tu, that delivers aminoacyl-tRNA to the ribosome (28); the tubulin subunits of microtubules (29); a large number of low molecular weight (20 to 25 KDa) proteins related to the product of the ras oncogene, many of which are now known to be involved in protein transport (25-27); and subunits of the signal recognition particle (SRP) and its ER membrane receptor (SR) (see below) that participate in the targetting of nascent polypeptide chains containing insertion signal sequences to the protein translocation apparatus in the ER membrane (30, 31).
The essential feature of GTP-binding proteins, that allows them to serve as molecular switches, is that they have a slow or latent GTPase activity and, therefore, can exist in two different conformational states, depending on whether they contain bound GTP or GDP, with the GTP-bound form being referred to as the "active" one. In general the conformational state of a GTP-binding protein determines its capacity to associate with a downstream effector. In the case of the heterotrimeric G proteins (32) (Gabg) that are coupled to plasma membrane receptors with seven transmembrane domains, such as rhodopsin and the b adrenergic receptor, activation of the receptor catalyzes the exchange of GDP by GTP in the Ga subunit, which in this "activated state" dissociates from the b and g subunits, and exerts its effect on an effector, such as a channel (e.g. the muscarinic receptor-activated potassium channel) or an enzyme (e.g. phosphodiesterase, or adenylcyclase). There are two main types of G proteins that associate with different receptors, those that contain stimulatory a subunits (Gsa) and activate the effector, and those (Gia) that contain inhibitory a subunits (Gia) and exert the opposite effect. In both cases, the Ga subunit remains active until the switch is turned off by the spontaneous hydrolysis of its bound GTP (32). Although the best understood role of heterotrimeric G proteins is in signal transduction at the plasma membrane, very recent evidence indicates that proteins of this class, not yet fully characterized, are also associated with intracellular membranes and function in regulating various steps of intracellular vesicular transport (see 33, 34, and our discussion below).
The ras related small molecular weight GTP-binding proteins involved in intracellular protein transport, and probably the GTP-binding proteins in the signal recognition particle (SRP) and its cognate receptor (SR) that function in the targetting of newly synthesized polypeptides to the ER (see below), do not appear to control an enzymatic reaction but rather to serve as switches that confer unidirectionality to a sequence of molecular associations. The paradigm for this mode of action is the elongation factor EF-Tu (22, 24, 28), which in its GTP-bound or "active" form binds aminoacyl-tRNA in the cytosol, but releases it when GTP hydrolysis takes place (see Fig. 7). In this case, the molecular switch associated with GTP-hydrolysis controls the unidirectional transport of the aminoacyl t-RNA from the site of its charging with the amino acid (the cytosol) to its site of utilization (the ribosome). The cyclic function of EF-Tu requires that the released factor be recharged with GTP by another cytosolic elongation factor, EF-Ts, which functions as a guanine nucleotide exchange protein. The EF-Tu GTPase paradigm has served to inspire hypotheses (24-26, 35) for the role of small GTP-binding proteins in the vectorial insertion of proteins into the ER, and in the vesicular transport of proteins from one intracellular compartment to another, in which many ras-related proteins of the rab family have been found to play key regulatory roles.
The low molecular weight (20-29 kDa) GTP-binding proteins of the ras superfamily, are divided into several families (ras, rap, rab, rho, rac, ran and arf), based on sequence similarities (36). Many proteins of the rab family (for rat brain, from where the original DNAs were cloned), of which at least 25 members have now been identified, have been localized to several intracellular organelles in a variety of cell types and have been shown to be involved in different steps of transport at the endomembrane system (see 26, 26a, 27, and our discussion below). In some of these processes arf proteins (for ADP ribosylation factor, the name given to the original member of this family, identified as a cofactor in the cholera toxin-induced ADP ribosylation of Gsa proteins) have also been implicated.
GTP-binding proteins of the rab and ras families share a common domain structure which is reflected in their function. In addition to the cysteine-containing C-terminal region, that is required for their prenylation and membrane binding (see below), they contain three highly conserved segments that participate in GTP binding and a segment (residues 32 to 40 in the rab proteins), known as the effector domain (because it corresponds to effector domain of ras, i.e. the region on the surface of ras that interacts with the protein, called raf, that regulates its GTP hydrolysis). Studies with ras have shown that its effector domain undergoes a conformational change during the GTP cycle and, therefore, is part of the switch mechanism essential for the function of the protein. Finally, a hypervariable region near the C-terminus of the rab proteins appears to determine their specific distribution in different organelles (see 27, 37). Many different amino acid substitutions in the GTP-binding or effector domains of the ras protein have been shown to lead to its activation, independently of nucleotide binding (38), and the same effects are expected to result from similar mutations in other ras-related proteins.
With the exception of the ran family members, which are nuclear proteins, the different ras-related proteins are characterized by specific C or N terminal posttranslational modifications that are essential for their function. Thus, proteins in the ras and rho families are characterized by the C-terminal CAAX sequence motif (where A is an alphatic residue, and X any other amino acid), and ras proteins in which X is M or S, are modified by the addition of a C-15 isoprenoid moiety, farnesyl, to the cysteine residue. The modified proteins then undergo proteolytic removal of the AAX sequence, followed by carboxymethylation of the new C-terminal cysteine residue (see 39). Some of these proteins also undergo palmitoylation (which is a reversible modification) at a neighboring cysteine residue and, together, these modifications participate in anchoring the proteins to a membrane. Proteins in the rab family contain dicysteine motifs near their C-termini (CC, CXC, CCX, CCXX, CCXXX), which receive the C-20 isoprenoid moiety, geranylgeranyl (40-42). These proteins do not undergo proteolysis, but those with the CXC sequence are carboxymethylated (43).
In contrast to those in the ras and rab families, the arf proteins are modified only at their N-termini by removal of the initiator methionine and addition of a myristoyl group (a tetradecanoic acyl group) to a glycine residue that immediately follows it (44). The myristoyl moiety serves as a membrane anchor, but only if the arf protein carries bound GTP. An arf protein, therefore, cycles on and off a membrane (e.g. a Golgi membrane) during its GTP cycle (45).
The capacity of ras-related GTP-binding proteins to switch between their off (GDP bound) and on states (GTP-bound) is controlled by a set of regulatory proteins that interact with them (Fig. 8). These include (see 25), GTPase-activating proteins (GAPs) that, when the protein binds to its effector (which may be the GAP itself), increase the GTPase activity up to 100 fold (e.g. 46); guanine nucleotide-exchange proteins (GEF), also known as GDP/GTP dissociation stimulators (GDS), that accelerate the release of GDP - which in the GTP-rich cellular environment promotes the binding of GTP (47-49); and GDP dissociation inhibitors (GDI) that prevent the release of GDP and at least in some cases, after GTP hydrolysis, remove the cognate GTP-binding protein from the membrane, or prevent its membrane association (50, 51). GDI proteins may, therefore, maintain in the cytosol a pool of inactive GTP-binding proteins, ready to be activated by GDS proteins that may themselves be membrane-associated. In a plausible scheme (Fig. 9), activation of a GTP-binding protein by a specific guanine nucleotide exchange factor (GEF or GDS) located in a donor membrane would trigger the association of the GTP-binding with the membrane through its interaction with other membrane proteins, which would promote the formation of a carrier vesicle. Upon vesicle docking on the correct acceptor membrane GTP hydrolysis would occur, releasing into the cytosol the GTP-binding protein in its
Conversion of the inactive GTP-binding protein into the active form involves a guanine nucleotide exchange factor (GEF, rectangle) that stimulates the release of GDP, which leads to its replacement by GTP. The GEF may be located in a specific membrane or subcellular structure and its activation may require a signal from an upstream regulator, such as a heterotrimeric G protein, or a protein kinase.
A guanine nucleotide dissociation inhibitor (GDI, trapezoid) may also be an important regulator of the activity of a GTP-binding protein by binding to the GDP-containing form and preventing its recharging with GTP. A GDI has been shown to be capable of removing the GDP-containing (inactive) form of a GTP binding protein from membranes and to form a complex with it. Therefore, it is possible that for rab proteins involved in vesicular transport a GDI may play a role in allowing the cyclic function of the GTP-binding protein by transporting it from the acceptor to the donor membrane.
GDP-state. Docking of the vesicle would be followed by activation of the components of the molecular machinery which leads to membrane fusion.
The involvement of a GTP-binding protein in an intracellular transport process is most easily assessed through the effect on that process of the non hydrolyzable GTP analogue, GTPg-S, which, when bound to a GTP-binding protein maintains it in the "active" conformation. If a step in the transport process requires this conformation, addition of the analogue should stimulate it. If, on the other hand, the step requires hydrolysis of the bound GTP, it should be inhibited by GTP-g-S. Of course, hydrolysis of GTP is the step required for the GTP-binding protein to complete a cycle of function and to allow for its reutilization, as well as for the reutilization of those other components to which the protein binds specifically when in the activated state. Therefore, if a transport process - observed in a cell free system or in permeabilized cells - is regulated by a GTP-binding protein and the result of many cycles of function is being observed, addition of GTP-g-S should lead to inhibition of transport. Similarly, a GTP-binding protein carrying a mutation that maintains it in the active conformation, after binding to its effector, would not be able to function cyclically. Such a protein would block effector sites and inhibit subsequent cycles in the process, exerting a dominant negative effect.
AlF3-5 is also a useful reagent that can be used to assess specifically the participation of a heterotrimeric G protein in a cellular process (52). This complex ion can mimick the g phosphate of GTP and directly activate the Ga subunit of a G protein that contains bound GDP, but is not able to activate the low molecular weight GTP binding proteins. The specific role of a G protein may also be revealed by the effects of mastoparan, a cationic amphiphilic peptide from wasp venon that mimics the polypeptide segment of plasma membrane receptors that interacts with G proteins. This peptide preferentially binds to the C-terminal region of Gia subunits, uncouples them from their receptor, and triggers the replacement of GDP by GTP, thus leading to activation of the inhibitory G protein.
Two toxins capable of catalyzing the ADP-ribosylation of Ga subunits, the pertusis toxin and cholera toxin, have also been very useful in revealing the involvement of a heterotrimeric G protein in a cellular function. The pertusis toxin modifies (by ADP ribosylation) the C terminal ends of some inhibitory Gia subunits. This uncouples them from the receptors and, by preventing their activation through the GDP/GTP exchange, inhibits their function. The modification induced by pertussis toxin also prevents mastoparan from exerting its effect (53, 54). In contrast, cholera toxin ADP ribosylates the stimulatory Gsa subunits and leads to their constitutive activation (55).
Protein synthesis in the rough endoplasmic reticulum
(11-13, 30, 31, 56-59)
The endoplasmic reticulum is a complex system of intercommunicating membrane-bounded flattened sacs (cisternae) and tubules that is present in all eukaryotic cells and in many cases permeates large regions of the cytoplasm. Membranes of the rough portions of the ER contain receptors for ribosomes and for nascent polypeptides, as well as proteins that are involved in the cotranslational incorporation of these polypeptides into the organelle, or in their processing during or soon after their synthesis. In addition, both rough and smooth ER membranes contain enzymatic systems that carry out functions essential to all cells, such as steps in the synthesis of triglycerides, phospholipids and cholesterol, as well as systems that carry out specialized metabolic or biosynthetic functions and, therefore, are present only in specific cell types.
The rough ER is most prominent in cells that are engaged in protein secretion, such as pancreatic acinar and anterior pituitary cells - which synthesize digestive enzymes and polypeptide hormones, respectively -- plasma cells, which produce immunoglobulins, and hepatocytes -- which manufacture a wide variety of serum proteins. Because the rough ER plays a major role in the synthesis and assembly of membrane proteins, this organelle is prominent in cells, such as neurons, which maintain greatly expanded plasma membranes.
The degree of development of the smooth ER in different cell types usually reflects the participation of ER membrane enzymes in specialized activities of the cell. Thus, the smooth ER is very well developed in cells of steroid-secreting tissues where it contains enzymes that catalyze several of the hydroxylation reactions that modify the steroid nucleus. It is also highly developed in skeletal muscle cells, in which it is known as the "sarcoplasmic reticulum", an organelle equipped to sequester into its lumen calcium ions and to release them when the cells are stimulated to contract. In fact, the ER appears to play a major role in the control of cytoplasmic Ca2+ levels in almost all cell types.
Although the membranes of the rough and smooth portions of the ER are continuous, they usually adopt within the same cell different morphological configurations which must reflect differences in their protein and/or lipid composition. The rough cisternae are frequently arranged in stacks of interconnected flattened disks, whereas the smooth portions usually form an extensive system of thin convoluted tubules. Electron micrographs of grazing sections of rough cisternae reveal that the ribosomes attached to the membranes form rosettes, hairpins, or spiral patterns, which correspond to membrane-bound polysomes. Individual ribosomes within bound polysomes contact the membrane via their large subunits (60), which are known to contain the nascent polypeptide chains.
Much information on the biochemical composition and function of the ER has come from the analysis of rough and smooth microsomes, subcellular fractions derived from rough and smooth portions of the endoplasmic reticulum, respectively. Extensive fragmentation of the ER takes place during the tissue homogenization that must be carried out before cell fractionation. The broken ER membranes, however, rapidly reseal to form microsomal vesicles that still contain a large part of the luminal content of the intact organelle (61, 62). The rough microsomes retain the ribosomes bound to their membranes and can be separated from the smooth microsomes on account of their greater density.
The structural and compositional differences between rough and smooth ER membranes have been best studied in liver cells, where both portions of the organelle are well developed and can be isolated as rough and smooth microsomes, respectively, with high yields and relative purity. Many of the most abundant ER membrane proteins are present in both rough and smooth membranes, but rough microsomes contain several specific membrane polypeptides that are likely to be involved in functions associated with the synthesis and processing of proteins made in bound ribosomes, or with maintaining the characteristic structure of the rough cisternae (63-65).
Although cellular phospholipids are synthesized in the ER, the phospholipid composition of the ER membranes is not a simple reflection of their biosynthetic capacity. Thus, they are rich in phosphatidyl choline (PC), phosphatidyl ethanolamine (PE), phosphatidyl serine (PS), and phosphatidyl inositol (PI), but contain very small amounts of sphingomyelin (SM) and cholesterol, which are abundant in the plasma membrane. The fatty acids of ER phospholipids are usually highly unsaturated and this accounts for the high fluidity of the ER membranes (see 4).
Cotranslational Insertion of Polypeptides into ER Membranes: Role of Insertion Signals in Determining the Association withthe ER Membranes of Polysomes Synthesizing Specific Proteins
Ribosomes which are part of polysomes found free in the cytoplasmic matrix are structurally and functionally identical to those within polysomes bound to ER membranes (66, 67). Indeed, within the cell, after completion of each polypeptide chain, both polysomal populations may exchange ribosomal subunits (68). The attachment to the ER membrane of those ribosomes that synthesize secretory, lysosomal, or certain classes of membrane proteins is determined by information contained within the nascent polypeptide chains (69). Extensive studies with a wide variety of secretory proteins have demonstrated that, almost invariably, nascent secretory polypeptides contain aminoterminal peptide segments that are not present in the mature proteins and serve to determine the attachment of the ribosome bearing the nascent chain to the ER membrane (70-74). These segments consist of 15-30 amino acid residues and characteristically include a central hydrophobic core of at least 8 amino acids (see 75-78). Similar N-terminal peptides are found in nascent lysosomal proteins and in many membrane proteins. All these peptide segments are known as transient insertion signals or signal sequences or presequences. They serve to trigger the association of the ribosome with the membrane and initiate the complete or partial translocation of the nascent polypeptide through it, but are removed by proteolytic cleavage before synthesis of the polypeptide is completed. The translocation of proteins across the ER membranes mediated by signal sequences (Fig. 10) is frequently referred to as the vectorial discharge of nascent polypeptides (79).
The Process of Assembly of a Membrane-Bound Polysome
(Fig. 11)
As is the case with the assembly of a free polysome, the assembly of a membrane-bound polysome begins in the cytoplasm with the binding of a small ribosomal subunit to the 5' end of the mRNA. After the large ribosomal subunit joins the small subunit, at the initiation codon of the mRNA, synthesis of the polypeptide begins. It is only after elongation is in course and the polypeptide is long enough for the signal segment to emerge from the large ribosomal subunit, which normally encloses a 40 amino acid segment of the nascent chain (80, 81), that the mechanism that leads to translocation begins to operate. This mechanism, illustrated schematically in Fig. 11, includes fail-safe features which ensure not only that the nascent polypeptide is inserted into the ER during its synthesis but also that, if insertion cannot take place, synthesis of the polypeptide is halted soon after the signal segment emerges from the ribosome.
The process of targetting the ribosome to the membrane begins with the recognition of the emerging signal by a soluble macromolecular complex, the signal recognition particle (SRP) (82-91), which consists of six distinct polypeptides and a small RNA molecule (7SL RNA) of approximately 300 nucleotides in length. The SRP interacts not only with the signal but also with the ribosome in such a way as to lead to a temporary block in polypeptide chain elongation. This block is only relieved in a subsequent step, when the SRP binds to its cognate SRP receptor (SR) (87-93), also known as the docking protein (92), which is an integral membrane protein exposed on the cytoplasmic surface of the ER. The pause in translocation caused by SRP ensures that continued growth in the cytoplasm and subsequent folding of the polypeptide, which could prevent insertion in the membrane, do not take place.
Rough ER membranes also contain sites with high affinity for ribosomes, which may be regarded as ribosome receptors (94). Following docking of the SRP on the membrane, a firm attachment of the ribosome to its receptor takes place, which allows for the coupling of the processes of translation and membrane insertion. The exact sequence of events that occurs next is not known, but it is clear that binding of the SRP to its receptor displaces it from the signal and from the ribosome (95). The signal sequence must then enter the membrane where it is thought to interact with protein components of a translocation apparatus (96-100).
Figure 11.
The process of assembly of a membrane bound polysome and the mechanism for the cotranslational translocation of a nascent polypeptide.
An ordered series of molecular recognition events leads to the insertion of a nascent chain in the ER membrane. This involves an initial interaction of SRP with the ribosome (1) and with the emerging signal sequence (2,3), followed by binding of SRP to its receptor (4), which in turn leads to release of the SRP from the signal and the ribosome. The latter are then able to bind to their membrane receptors (5, 6). See the text for the detailed description of the role of the different components of the translocation machinery illustrated in this figure. Note that SRP and its receptor function catalytically and are only transiently associated with the site of translocation. In this figure, only two cotranslational modifications, signal cleavage (6) and core glycosylation (7), are shown, but several others, such as disulfide exchange and hydroxylation of side chains, are also known to take place cotranslationally. Note that signal cleavage occurs relatively early in translocation so that covalent linkage of the signal to the remainder of the polypeptide is not required for the continuation of translocation. After chain termination (8), it is presumed that the translocation apparatus disassembles and the ribosome detaches from the membrane. Translocation is depicted as taking place through a proteinaceous channel in the membrane, which becomes an extension of a tunnel within the large ribosomal subunit, where the nascent chain is contained.
Since neither the SRP nor its receptor appear to remain associated with the membrane-bound ribosome at the site of translocation (the number of SRP receptors in the ER membrane is much smaller than the number of active bound ribosomes), the essential role of the SRP/SRP receptor system appears to be to simply target the ribosome and its incipient chain to the endoplasmic reticulum, without participating in the translocation process itself.
Although it is not yet fully understood how the polypeptide actually traverses the ER membrane, it is clear that an interaction of the nascent chain with membrane proteins is necessary for translocation to occur. Considerable evidence supports a model in which passage of the nascent polypeptide across the hydrophobic barrier within the membrane occurs through the aqueous environment of a proteinaceous channel (73, 79, 101-104). The role of the insertion signal would be to open the channel or to trigger its assembly from dispersed membrane protein subunits. The open or assembled state of the channel could be stabilized by the interaction of its components with the large ribosomal subunit but the channel would be closed or disassembled when the ribosome is released following polypeptide chain termination, or a halt in translocation caused by a stop transfer signal (see below). In alternative models (105, 106), now not in favor (see below), following targetting to the membrane, the insertion signal and the nascent chain would interact directly with the membrane bilayer and no specific membrane proteins would be required to mediate translocation itself.
It is clear that, in addition to proteins that may participate directly in the translocation process, several membrane enzymes are also located near the site of translocation and are able to interact with the growing chain to modify it as it emerges on the luminal side of the ER membrane. Thus, a signal peptidase cleaves off the signal, a protein oligosaccharyl transferase links preformed mannose-rich oligosaccharide chains to selected asparagine residues within the nascent polypeptide, and a protein disulfide isomerase (PDI) enzyme catalyzes the formation of intramolecular disulfide bonds. These modifications, however, are not required for translocation to occur.
Experimental Analysis of Translocation (107)
The process of cotranslational insertion of nascent polypeptides into ER membranes has been best studied utilizing cell-free systems in which messenger RNAs from natural sources, or produced by in vitro transcription of cloned cDNA templates, are translated in the presence of rough microsomes. The most commonly used translation systems are derived from rabbit reticulocytes or wheat germ, and the most frequently used rough microsomal membranes are obtained from dog pancreas (108).
When messenger RNAs encoding secretory proteins are translated in the absence of membranes, primary translation products are obtained which contain the signal sequences and are devoid of any modifications of their primary structure . Such artificial products of in vitro translation, which are generally not produced in vivo, are called presecretory proteins or preproteins. These are not true precursors of the secretory proteins, since, in vivo, their signals are actually removed before synthesis of the polypeptide is completed. When rough microsomal membranes are present during the in vitro translation, a large fraction of the product synthesized is translocated into the lumen of the microsomes and undergoes signal removal. If the translocated protein does not contain sites for N-linked glycosylation, its electrophoretic mobility is higher than that of the primary translation product (the presecretory protein) by the 2-3 kDa that corresponds to the cleaved signal . The sequestration of the processed polypeptides in the lumen of the microsomes is demonstrated by the fact that they are protected from proteolysis when proteases are added to the reaction mixture after translation is completed . On the other hand, the presecretory proteins , which remain outside the microsomes, are completely digested . Destruction of the membranes by detergent solubilization, of course, leads to digestion of the translocated products by the exogeneous proteases.
When the messenger RNA utilized for in vitro translation experiments encodes a protein with sites for N glycosylation, translocation of the nascent chain is accompanied by both signal cleavage and addition of N-linked oligosaccharides (see below). In this case, the apparent size of the translocated product, when assessed by gel electrophoresis, may be higher than that of the primary translation product since the contribution of the added oligosaccharide chains may more than compensate for the size reduction resulting from signal cleavage . Upon treatment with proteases, only the glycosylated polypeptide remains undigested. The absence of the signal sequence in the translocated glycoproteins becomes apparent when, after dissolution of the microsomal membrane, they are treated with a glycosidase (endoglycosidase H) that removes the oligosaccharide chains (see below).
In vitro translocation experiments of the type just described, with mRNAs that encode lysosomal enzymes (most of which are glycoproteins), or type I transmembrane proteins have demonstrated that these polypeptides too, contain transient N-terminal signals which mediate their cotranslational insertion into the ER.
Characterization of Insertion Signals (75-78, 109-111)
Insertion signals are necessary to initiate the translocation of nascent polypeptides across the ER membrane. Secretory polypeptides from which the signal is deleted by modification of the corresponding cloned gene can no longer be translocated across ER membranes, in vivo or in vitro. Moreover, in some cases, attachment of a cleavable insertion signal to the amino terminus of a cytosolic protein has been shown to be sufficient to confer on it the capacity to be translocated. Although interaction of the signal with the membrane is necessary to initiate translocation, it is clear that covalent attachment of the signal to the rest of the polypeptide need not be maintained throughout translocation, since signal cleavage generally occurs much before elongation of the nascent chain is completed (73). However, it is possible that the signal could, even after it is severed from the body of the nascent chain, be necessary for translocation to continue. If this were the case, degradation of the cleaved signal segments by the yet to be discovered signal peptide peptidase, would, of course, occur only after translocation is completed.
Comparison of the amino acid sequences of different insertion signals shows that there is considerable variation in their primary structure. This suggests that general properties of the signals, including conformational features, rather than specific sequence information, are recognized by the various components of the translocation machinery - such as the SRP, a putative signal sequence receptor, and the signal peptidase - that interact with the signal. Indeed many random sequences of the human genome that encode peptide segments of relatively high but varying degrees of hydrophobicity were shown to be capable of serving a signal function when linked to the yeast secretory protein invertase, from which the aminoterminal signal was removed (112).
Insertion signals are, in general, 15 to 30 amino acids in length. Preprotein sequences are conventionally numbered so that the first residue after the cleavage site of the signal is designated +1 and the last residue of the signal is -1. Three segments can be recognized in all cleavable signals (76) (Fig. 13): 1) a hydrophobic core (the h-region) 8 to 16 residues in length, which generally ends at residue -6, 2) a hydrophilic N-terminal segment that precedes the core (the n-region) and usually contains, in addition to the positively charged N-terminus, one basic amino acid, 3) an approximately five residue-long C-terminal segment (the c-region) which defines the cleavage site and usually begins with a helix-breaking glycine or proline residue. In some signals, such as the one in rat growth hormone, the amino terminal segment bears no net charge due to the presence of a negatively charged amino acid residue.
The charges in the n-region of the signal are likely to play a role in initiating the association of the signal with the membrane that triggers the insertion of the nascent polypeptide. Deletion of the n-region from preproparathyroid hormone does not prevent the elongation arrest caused by SRP or its relief by the SRP receptor, but translocation of the nascent polypeptide is impaired (113). These observations are in accord with the notion that, as the signal begins to enter the membrane itself, the charges in the n-region associate directly with the polar head groups of the phospholipids (114, 115). If this association is maintained as translocation proceeds, the nascent chain would adopt a loop disposition in the membrane (Fig. 14), with its N-terminus on the cytoplasmic face, the hydrophobic core of the signal within the
Loop model for the disposition during translocation of the cleavable signal of a polypeptide synthesized in a membrane-bound ribosome.
In this model, the extreme N-terminus of the signal remains on the cytoplasmic face of the membrane and the nascent chain has a looped disposition during the initial stages of translocation. In most secretory, lysosomal, and type I transmembrane proteins, the signal is N-terminal and is cleaved during translocation, as depicted in this figure. In type II transmembrane proteins, the signal is not cleaved and the looped configuration would be maintained throughout translocation, until the extreme C-terminus of the polypeptide is released into the lumen.
membrane, and the cleavage site on the luminal surface. The formation of this loop would be facilitated by the helix-breaking nature of the residues immediately following the hydrophobic core.
Direct evidence for the loop model has been obtained from an analysis of the behavior of a genetically engineered protein, expressed in transfected cells, whose aminoterminal cleavable insertion signal was extended by the addition of upstream sequences and whose cleavage site was abolished by mutation (116). Translocation of this protein proceeded normally but, since the signal was not removed, it served as a membrane anchor for the final product, which was a transmembrane protein with the aminoterminal extension preceding the signal exposed on the cytoplasmic surface of the ER. The cytoplasmic exposure of the aminoterminal extension implies that throughout the course of translocation the signal was maintained in the loop configuration.
From the analysis of numerous insertion signal sequences, certain rules have emerged that, with a fair degree of certainty, allow the prediction of the site of cleavage of the signal within the sequence of a preprotein. Most notably, the (-1, -3 rule) (77, 111, 117) states that the -1 position is almost always occupied by small neutral amino acids, such as alanine, glycine, or serine and that the residue at -3 must not be aromatic (phe, his, tyr, try), charged (asp, glu, lys, arg), or large and polar (asn, gln). It is also apparent that residues following the cleavage site may contribute to its recognition by the signal peptidase. Thus, some point mutations, produced by genetic engineering techniques, that affect residues following the cleavage site have been shown to prevent cleavage. A role of the sequence following the cleavage site in determining signal cleavage may account for the fact that some secretory proteins, such as parathyroid hormone and albumin, contain a second transient amino terminal peptide segment, the propiece, that is removed from the proprotein during or after passage through the Golgi apparatus. In these instances, the N-terminal sequence of the mature protein, which may be important for the function of the protein, might not have permitted cleavage of the signal had it been immediately adjacent to the -1 residue (118). One function of the propiece could, therefore, be to satisfy the requirements for the creation of a signal peptidase cleavage site in the nascent preprotein.
The conformation that the signal segment attains within the interior of the membrane has not been established. Within a hydrophobic environment, an a helical conformation would be favored for the core region. However, the core is usually shorter than the approximately 20 amino acid residues that would be required for an a helix to completely span the membrane thickness of 2.5 to 3nm, whereas in a fully extended configuration a peptide segment of only 8 residues could span the membrane. It has, therefore, been suggested (110) that within the membrane the hydrophobic core of the signal may exist partially as an alpha helix and partially as a fully extended structure. The important role played by the hydrophobicity of the central core of the signal in translocation is apparent from the deleterious effects of mutations which in bacterial secretory proteins replaced some of the hydrophobic residues by charged ones, or introduced partial deletions covering core residues (119, 120).
Even though insertion signals are usually removed by cleavage from nascent secretory and lysosomal polypeptides and from many nascent membrane proteins (see below), signal cleavage is not required for translocation. Indeed, one secretory protein, ovalbumin (121-122), and several viral envelope glycoproteins (123-125) are known which contain signals near their amino termini that serve their function to mediate translocation, but are not cleaved and are themselves transferred with adjacent portions of the polypeptide into the ER lumen.
The Signal Recognition Particle (SRP) (30, 31, 58, 59)
The signal recognition particle plays a central role in selecting ribosomes for binding to the ER and in delivering the nascent chains to a receptor within the membrane. The distribution of SRP within the cell reflects its cyclic participation in these processes. SRP may be found free in the cytoplasm, weakly bound to inactive ribosomes or attached to its receptor in the ER membranes (126) (Fig. 11). The affinity of SRP for ribosomes, however, increases at least 6000 fold when the ribosome contains a nascent chain with an exposed signal sequence, to which the SRP also binds (84).
The most commonly used source of SRP for in vitro studies of its role in translocation is dog pancreas microsomes, from which SRP can be released by treatment with media of high salt concentration. Indeed, microsomes treated with high salt (KRM) are inactive in translocation unless supplemented with SRP (82). The pause in translation caused by SRP is best observed when SRP is added to a wheat germ translation system, which lacks endogenous SRP. In the absence of added microsomes, SRP leads to an effective block in the elongation of nascent proteins that contain a signal peptide, such as preprolactin, but the synthesis of cytosolic proteins, such as globin, proceeds unaffected (86). In several cases, it has been shown that in the presence of SRP, a ribosome-associated arrested fragment of the preprotein of approximately 80 amino acids accumulates in the translation system. The SRP-mediated arrest of polypeptide chain elongation is relieved by the addition of microsomes to the system, which leads to signal cleavage and translocation (85, 86), or even by the addition of purified SRP receptor (88).
The SRP obtained from dog pancreas microsomes by washing with high salt is a particle with a sedimentation coefficient of 11S that contains, in addition to the 7SL RNA molecule, six polypeptide chains of molecular weights 9, 14, 19, 54, 68, and 72 kDa (83, 127). The 7SL RNA is an abundant and metabolically stable molecule which contains at its 5' and 3' ends sequences of the alu family, one of the most highly repeated family of sequences in the genome, and in its middle region a core segment of 150 nucleotides, termed the S sequence, that is much less frequently repeated (128). Both protein and RNA components of SRP have been shown to be required for its function. SRP has been disassembled into its RNA and protein components by the removal of Mg2+ ions, which normally stabilize its structure, and it has been possible to reassemble a functional particle from the dissociated components (129). This has allowed studies on the role of the individual proteins on the different aspects of SRP function (130, 131).
It is noteworthy that 7SL RNAs from such evolutionary distant species as Drosophila melanogaster and Xenopus laevis can replace the canine RNA in the reassembled particles. Reconstitution experiments have shown that the two smallest polypeptides of SRP are required for translational arrest but are not necessary for translocation which, of course, demonstrates that the arrest in translation is not essential for translocation to occur (130). In fact, SRP may cause only a slowdown in the translation of certain mRNAs for which arrested peptides have not been detected in the wheat germ system.
Treatment of SRP with nucleases generates two subparticles that may correspond to domains exerting the functions of SRP (128). One particle contains the two smallest polypeptides bound to the two ends of the 7SL RNA, and the other the four remaining ones bound to the central region of the RNA.
SRP can be purified by hydrophobic chromatography, which suggests that it interacts with the hydrophobic core of signal sequences. Indeed, replacement of leucine residues in a presecretory protein with -hydroxyleucine, a polar analogue, abolishes the high affinity binding of SRP for the ribosome, and hence the translational arrest and subsequent translocation (85). Moreover, it has been demonstrated that in a ribosome carrying SRP, the nascent chain is in close proximity to the 54 kDa polypeptide of SRP, since the two can be crosslinked through a photoactivatable group incorporated into the nascent chain (132). In this case, the elongation arrest was maintained after crosslinking and was relieved upon binding to the SRP receptor, but translocation could not occur.
Much progress has been made in identifying within the SRP, structural domains that carry out its three distinct sequential functions: signal sequence recognition, elongation arrest, and delivery of the nascent chain to the translocation machinery within the ER membrane. The two smaller SRP polypeptides (9 KDa and 14 KDa) form a heterodimer and are required for the elongation arrest to occur whereas the two largest ones (68 KDa and 72 KDa), which also form a heterodimer, are required for the binding of SRP to its membrane receptor (SR). The 54 KDa polypeptide, SRP54, is a GTP-binding protein that contains a putative GTPase segment and its primary function is in the recognition of the insertion signal in the nascent polypeptide. The C-terminal portion of SRP54 constitutes a methionine-rich "M Domain" ( 132a-c) which by itself shows affinity for the signal sequence, although in the native protein this appears to be regulated by the aminoterminal "G" domain (133).
It is noteworthy that genes encoding homologues of the SRP54 (133a, 134, 135) and of the a subunit of the SRP receptor (135, see below) have recently been identified in yeast and that, although their deletion leads to poor growth and to the accumulation of precursors of some secretory and membrane proteins in the cytosol, the cells, nevertheless, remain viable. This suggests that another mechanism, independent of SRP, can function in this unicellular eukaryote and effect the translocation of most essential proteins incorporated into the ER. Indeed, the posttranslational translocation of the a-mating factor precursor had previously been observed in yeast, both in vivo and in vitro (136-138). Evidence has also been presented that "molecular chaperones", encoded by the yeast Ssa1 and Ssa2 genes, facilitate the SRP-independent uptake of the a-factor precursor into the ER (139, 140). Molecular chaperones are members of a family of proteins that mediate the proper folding of other polypeptides and sometimes their assembly into oligomeric complexes (16). The chaperones involved in posttranslational translocation in the ER belong to the heat shock hsp70 (141) family of proteins and utilize ATP to confer on the polypeptide a conformation compatible with its transport across the membrane. These chaperones have also been shown (see below) to facilitate the uptake of polypeptides into mitochondria (139, 140).
It should be noted, that a yeast protein in the lumen of the ER, Kar2p that can be crosslinked to the nascent chain (142) is also necessary for translocation (143). Kar2p is the yeast homologue of the mammalian protein Bip (Grp 78), which is a heat-shock protein of the hsp70 family that serves as a molecular chaperone in the lumen of the ER and was first identified as an immunoglobulin heavy chain binding protein that functions in the assembly of immunoglobulin molecules (144, 145).
The Signal Recognition Particle Receptor (SR) or Docking Protein (30, 31, 59)
The SRP receptor is a heterodimeric (SRa: 72kDa; SRb:30kDa) protein complex exposed on the cytoplasmic surface of the ER membrane where it receives the SRP bound to a ribosome containing an exposed signal sequence (Fig. 11). This binding displaces the SRP from the ribosome and from the signal and allows the signal to insert into the membrane (95). The SRP receptor is present in the ER in low amounts (0.1% of the total membrane protein) and functions catalytically, remaining associated with SRP for only the very brief period required to displace it from the ribosome and to establish the ribosome-membrane junction. These reactions can occur at 0C in the absence of any polypeptide chain elongation (95).
The SRP receptor has been purified from solubilized rough microsomal membranes by affinity chromatography to immobilized SRP, using the relief of the translation arrest of preprolactin as an assay to detect functional receptor (87, 88). The SRP receptor obtained in this manner consists of two subunits, a 69 kDa glycoprotein a-subunit and a 30 kDa -subunit (146). Treatment of rough microsomes with the protease elastase renders the membrane inactive in translocation and leads to the release of a 52 kDa fragment of the a-subunit, which can be added back to the proteolysed membranes at low ionic strength to restore translocation competence (90, 91).
The complete primary structures of the SRP receptor subunits have been derived from the nucleotide sequence of cDNA clones (147, 31). Comparison of the sequence of the a-subunit with the N-terminal sequence of the 52 kDa fragment released by proteolysis shows that the protein is anchored to the membrane via a 155 amino acid amino terminal segment which contains two hydrophobic domains. The portion of the molecule exposed on the cytoplasmic surface contains three extremely hydrophilic regions rich in charged amino acids, with a predominance of basic residues which may interact directly with the 7SL RNA component of the SRP. This portion of the molecule also contains several additional hydrophobic segments which, clearly, do not interact permanently with the membrane and are probably buried within the protein.
Both the a and b subunits of the SR are GTP-binding proteins and, surprisingly, SRa and the 54 KDa subunit of SRP are sufficiently related in sequence to constitute a new subfamily of GTPases. The exact roles of these three GTPases (SRP54, SRa and SRb) have not yet become clear, but most likely, a series of GTP-dependent changes in the conformational states of these proteins controls the sequential macromolecular interactions which confer directionality to the polypeptide targetting and insertion processes. Using a non hydrolyzable GTP analogue (GppNHp) it has been established (148) that binding of GTP, but not its hydrolysis, is required for the displacement of the signal sequence from the SRP and for the insertion of the nascent chain into the membrane that takes place after docking on the SR. On the other hand, GTP hydrolysis is required for the dissociation of SRP from the SR (149), which allows both components to function cyclically. Although it is not known how many and which bound GTP molecules are hydrolyzed in each round of targetting and nascent chain insertion, it has been shown that a mutation in the GTP-binding consensus sequence of the SRa subunit reduces the efficiency of the GTP-dependent insertion of the nascent chain into the membrane and prevents the formation of the stable SRP/SR complex that occurs in the presence of the GppNHp (150).
Several plausible models can be proposed for the concerted action of the SRP and SR GTP-binding proteins in the targetting and membrane insertion processes. One of these assumes (Fig. 15) that binding of ribosome-associated SRP to an emerging signal sequence leads to exchange of GDP for GTP in SRP54 and that in this "active conformation" SRP binds more tightly to the ribosome in a manner that arrests translation. In its active conformation GTP-containing SRP would also have a higher affinity for the SR, to which it binds through its 68 and 72 kDa subunits. Unoccupied SR has at least its a subunit (SRa) in the GDP-bound state and the docking of SRP promotes the exchange of GTP for GDP in this subunit which, in some way, leads to release of the signal sequence from SRP54 and detachment of SRP from the ribosome. These events would not be followed immediately by hydrolysis of the GTP bound to SRP54, or else SRP would rebind to the signal at this point (149). Hydrolysis of the GTP in the SRa subunit would follow, leading to the dissociation of the SRP/SR complex and release of SRP into the cytosol where, after hydrolysis of its GTP, it would undergo another cycle of function. It has also been proposed that the GTPase activity of the SRb subunit, and the accompanying conformational changes, regulate the association of the SR with other components of the translocation machinery in the membrane, to which the nascent chain is delivered after its release from SRP (150).
Interaction of the Signal Sequence and the Nascent Polypeptide with Membrane Protein Components
After the displacement from the SRP, induced by the SRP receptor, the insertion signal and the nascent chain enter the ER membrane where they interact with protein components (Fig. 11) of a putative translocation machinery, for which the term "translocon apparatus has been proposed (58). An association of the nascent chain with membrane proteins was first demonstrated by the observation that partially translocated, incomplete, nascent polypeptides could be removed from the microsomal membrane by treatment with agents, such as urea, that perturb protein-protein interactions but do not remove integral membrane proteins from membranes (102). Attempts to identify components of the translocation apparatus in the ER membrane have employed crosslinking agents to link a radioactive nascent chain to dog pancreas microsomes membrane proteins that are in its close proximity when it traverses the membrane. A 35-39 KDa glycoprotein, first thought to be a signal sequence receptor and, hence, termed SSRa was identified in this manner (96). It was later shown, however, that this protein, also termed mp39, can be crosslinked to various other portions of the translocating polypeptide and, therefore, is unlikely to serve only as a signal sequence receptor (97). Moreover, although SSRa/mp39 clearly resides in the neighborhood of the site of passage of the nascent chain throughout the course of elongation (99), this protein does not appear to be necessary for translocation, since translocation-competent microsomes can be reconstituted with detergent extracts from which it was removed (151). Other microsomal polypeptides have also been crosslinked to nascent chains, including a 34 KDa (imp34) nonglycosylated protein (98) and a 39 KDa multispanning membrane glycoprotein, termed TRAM (translocating chain-associating membrane glycoprotein), that appears to be required for the translocation activity of reconstituted vesicles (100). TRAM could only be crosslinked to short nascent polypeptide chains, which indicates that it is near the nascent chain only at the beginning of its insertion and that cleavage of the signal sequence may displace it from the passageway in the membrane.
Several translocation-deficient yeast mutants have been identified in genes that encode transmembrane glycoproteins (Sec61p, Sec62p, Sec63p) that are part of a multisubunit complex of the type expected to function in translocation (152). Of these, Sec61p and Sec62p could be crosslinked to nascent chains, the latter only when the nascent chains are short (153). The mammalian homologue of Sec61p (40 kDa) has recently been purified (154) and its sequence, derived from the cloning of its cDNA, reveals that it is likely to have 10 transmembrane domains which contain a number of hydrophobic and charged amino acid residues. Sec61p is the major membrane component that can be crosslinked to long nascent chains in both mammalian and yeast cells. This protein is also homologous to the SecYp product of E. coli that, together with SecE, are the only two integral membrane proteins required to effect translocation in a system of reconstituted liposomes (155). Moreover, Sec61p becomes tightly associated with ribosomes during the course of translocation. Taken together, these findings raise the strong possibility that Sec61p represents a major constituent of the channel through which the nascent chain traverses the membrane.
Further characterization of Sec61p, Sec62p and Sec63p and other recently identified SEC gene products (Sec70p, Sec71p, Sec72p) whose mutations affect translocation and/or membrane protein integration (156), and of their mammalian homologues, may soon yield a more complete picture of the molecular assembly that constitutes the translocation apparatus in the ER membrane.
The Signal Peptidase
The signal peptidase activity of microsomes can be demonstrated in detergent solubilized preparations (157), using as substrates certain completed preproteins synthesized in vitro, such as preprolactin and pregrowth hormone. Most preproteins, however, cannot be processed posttranslationally by microsomal extracts, presumably because they are folded in such a way as to sequester the signal cleavage site. This sequestration of the signal may occur before synthesis of the preprotein is completed and the incapacity of the masked signals to interact with SRP when this is added late in translation would account for the fact that, beyond a certain length, nascent polypeptides are no longer "translocation competent" (158).
The solubilized signal peptidase is only active in the presence of phospholipids (159), and its activity can be inhibited by agents such as chymostatin that inhibit zinc metallopeptidases (160). Because the peptidase activity cannot be demonstrated without detergent solubilization, and it is not destroyed by proteolysis of intact microsomes (157, 161), it can be concluded that the active site of the signal peptidase is located on the luminal side of the ER membrane (Fig. 11). This location is consistent with the observation that signal cleavage does not take place before the polypeptide attains a minimal length of 70 to 90 residues, which are required to bring the cleavage site to the luminal face of the ER.
A protein complex with signal peptidase activity has been purified from solubilized dog pancreas microsomes (162). It contains five polypeptide chains of apparent molecular weights ranging from 12 to 25 kDa, one of which is glycosylated. It is likely that only one of these polypeptides carries out the signal cleavage and that the remaining ones participate in other aspects of the translocation process, such as the degradation of the cleaved signal peptide (signal peptide peptidase) or cotranslational modifications of the nascent chain. An intriguing possibility is that some of the polypeptides in the complex may bipart the channel in the membrane through which translocation is likely to take place. It is noteworthy that neither the SSRa (mp39) nor proteins that have been implicated in ribosome binding (see below) are part of the signal peptidase complex. Two of the protein subunits of the mammalian signal peptidase display substantial sequence homology to the yeast sec11 protein (163), which is also part of a complex with signal peptidase activity (164) and when mutated leads to defective signal cleavage in vivo (165).
Ribosome Binding Sites on the ER Membrane
After displacement of the SRP by its receptor, binding of the ribosome to the ER membrane takes place (Fig. 11). During the subsequent translocation, the ribosome remains associated with the ER membrane via two types of bonds, direct ones between the large ribosomal subunit and a receptor in the membrane, and an indirect link that is provided by the nascent polypeptide chain (166). The latter is broken upon termination of polypeptide growth, or when a prematurely terminated polypeptide is released from the ribosome as a result of the incorporation at the C-terminus of the nascent chain of the chain-terminating peptidyl-tRNA analog, puromycin. Ribosomal subunits can then be detached from the membrane by exposure of the microsomes to media of high ionic strength, which disrupt electrostatic interactions between the large ribosomal subunit and its receptor. Ribosomes not containing nascent chains rebind in vitro and at low ionic strengths to rough microsomal membranes stripped of ribosomes but not to other cellular membranes, including those of smooth microsomes (94, 167, 168). The number of ribosome binding sites detected by this method is equivalent to the number of ribosomes originally present in the rough microsomes (169).
The ribosome receptors present in the rough ER contain proteinaceous components, since ribosome binding is abolished by mild proteolysis or heat treatment of the membranes (94). The specific polypeptides involved in ribosome binding, however, have not been definitively identified. Two transmembrane glycoproteins, ribophorins I and II, present only in rough microsomes, where they are found in amounts stoichiometrically related to the number of ribosomes, appear to be associated with the binding sites (63, 64, 170). Thus, these proteins, and only a few other membrane polypeptides, are recovered with the ribosomes when these are sedimented after certain non ionic detergents are used to solubilize the membranes. In this case, a membrane residue is obtained in which proteins appear to form a two dimensional network bearing ribosomes. On this basis, it has been proposed (63, 64) that the ribophorins also play a structural role in the rough ER, providing a scaffolding within the ER membrane that restricts the ribosome binding sites and their associated translocation apparatus to the rough domains and confers on the rough cisternae their characteristic morphology. The cDNAs for both ribophorins have been cloned (171, 172), and the primary structure of the polypeptides indicates that both proteins are type I monotopic proteins that cross the membrane only once, and have C-terminal segments of 150 and 70 amino acids, respectively, exposed on the cytoplasmic face of the membrane. It now seems clear, however, that proteins other than the ribophorins, must contribute to the ribosome binding sites in the ER membrane, since mild proteolysis that does not appear to cleave the cytoplasmically exposed portions of the ribophorins abolishes the ribosome binding capacity of the membrane (173, 174). Moreover, it has been reported (175) that membrane vesicles that are capable of efficient ribosome binding can be reconstituted from purified lipids and a microsomal protein fraction from which all glycoproteins (including ribophorins) were removed by lectin chromatography. However, a function of the ribophorins associated with translocation has been recently discovered (176) with the finding that a protein complex consisting of both ribophorins and a 48 KDa polypeptide isolated from microsomal membranes manifests oligosaccharyl transferase activity, i.e. catalyzes the transfer of a high mannose oligosaccharide from a dolichol pyrophosphate donor to asparagine residues in the peptide sequence Asn X Ser/Thr, a process (see below) which normally occurs as the nascent glycoprotein chain emerges on the luminal side of the ER membrane.
Two other candidate proteins for ribosome receptors, a 34 KDa (177), and a 180 KDa protein (178), have also been identified, but definitive evidence for their ribosome binding roles is yet to come.
A Protein Conducting Channel in the ER Membrane
As previously noted, it was originally proposed that a transient or permanent aqueous channel through the ER membrane, just under the ribosome, provides a passageway for the nascent polypeptide into the ER lumen and it seems likely that at least some of the ER membrane proteins that have been crosslinked to the nascent polypeptide chain are part of such a channel. Electrophysiological studies (103) of the properties of microsomal membranes fused to planar lipid bilayers have, in fact, demonstrated the existence of large ion conducting channels in the ER membrane, which appear to be occupied, and therefore occluded, by the nascent polypeptide chain, since the conductance of the membrane increased dramatically upon addition of puromycin, a drug that releases the nascent chain. The channel also appeared to be stabilized by the bound ribosomes since the conductance was markedly reduced when, after nascent chain release, the ribosomes were detached from the membrane. More recently, similar electrophysiological studies with E. coli plasma membranes -- through which translocation in vivo can occur cotranslationally, as well as posttranslationally (i.e. without ribosome binding) -- have demonstrated the existence in these membranes of channels with a conductance comparable to that of those in the mammalian ER. In this case, opening (or assembly) of the channels could be triggered by addition of a synthetic signal peptide (104) which, therefore, appears to be functional as a physiological ligand for channel opening. These exciting findings augur that within the next few years the full set of molecular components that constitute the prokaryotic and eukaryotic protein conducting channels will have been identified and it may be expected that the molecular interactions that control channel function will soon thereafter begin to be understood.
BIBLIOGRAPHY
18. PRYER NK, WUESTEHUBE LJ, SCHEKMAN R: Vesicle-mediated protein sorting. Ann Rev Biochem 61:471, 1992.
18a. FRANZUSOFF A: Beauty and the yeast: compartmental organization of the secretory pathway. Semi
nars Cell Biol 3:309, 1992
19. HAUBRUCK H, PRANGE R, VORGIAS C, BALLWITZ D: The ras-related mouse ypt1 protein can functionally replace the YPT1 gene product in yeast. EMBO J 8:1427, 1989.
20. WILSON DW, WILCOX CA, FLYNN GC, CHEN E, HUANG WJ, HENZEL WJ,
BLOCK MR, ULLRICH A, ROTHMAN JE: Fusion protein required for vesicle-mediated transport in both mammalian cells and yeast. Nature 339:355, 1989.
21. GILMAN AG: G proteins: Transducers of receptor-generated signals. Ann Rev Biochem 56:615, 1987.
22. K
AZIRO Y, ITOH H, HOZASA T, NAKAFUKA M, SATON T: Structure and function of signal-transducing GTP-binding proteins. Ann Rev Biochem 60:349, 1991.
23. BOURNE HR, SANDERS DA, McCORMICK F: The GTPase superfamily: conserved structure and molecular mechanism.
Nature 349:117, 1991.
24. BOURNE HR: Do GTPases direct membrane traffic in secretion? Cell 53:669, 1988.
25. GOUD B, McCAFFREY M: Small GTP-binding proteins and their role in transport. Curr Opin Cell Biol 3:626, 1991.
26. GRUENBERG J, CLAGUE MJ: Re
gulation of intracellular membrane transport. Curr Opin Cell Biol 4:593, 1992.
26a. BALCH WE: Small GTP-binding proteins in vesicular transport. Trends in Biochemical Sciences 15:473, 1990.
26b. PFEFFER SR: GTP-binding proteins in intracellular transpo
rt . Trends in Cell Biology 2:41, 1992.
27. ZERIAL M, STENMARK H: Rab GTPases in vesicular transport. Curr Opin Cell Biol 5:613, 1993.
28. THOMPSON RC: EFTu provides an internal kinetic standard for translational accuracy. TIBS 13:91, 1988.
29. GELFAN
D VI, BERSHADSKY AD: Microtubules dynamics: mechanism, regulation, and function. Ann Rev Cell Biol 7:93, 1991.
30. GILMORE R: The protein translocation apparatus of the rough endoplasmic reticulum, its associated proteins, and the mechanism of transloca
tion. Curr Opin Cell Biol 3:580, 1991.
31. NUNNARI J, WALTER P: Protein targeting to and translocation across the membrane of the endoplasmic reticulum. Curr Opin Cell Biol 4:573, 1992.
32. STRYER L, BOURNE HR: G proteins: a family of signal transducer
s. Ann Rev Cell Biol 2:391, 1986.
33. BARR FA, LEYTE A, HUTTNER WB: Trimeric G proteins and vesicle formation. Trends Cell Biol 2:91, 1992.
34. BALCH WE: From G minor to G major. Tantalizing evidence is accumulating that trimeric G proteins play a mu
ch greater part than expected in signaling vesicle transport in endocytic and exocytic secretory pathways. Curr Biol 2:157, 1992
35. GRAVOTTA D, ADESNIK M, SABATINI DD: Transport of influenza HA from the trans-Golgi Network to the apical surface of MDCK
cells permeabilized in their basolateral plasma membranes: energy dependence and involvement of GTP-binding proteins. J Cell Biol 111:2893, 1990.
36. KAHN RA, DER CJ, BOKOCH GM: The ras superfamily of GTP-binding proteins: guildelines on nomenclature. FA
SEB J 6:2512, 1992.
37. CHAVRIER P, GORVEL J-P, STELZER E, SIMONS K, GRUENBERG J, ZERIAL M: Hypervariable C-terminal domain of rab proteins acts as a targeting signal. Nature 353:769, 1991.
38. BARBACID M: ras genes. Ann Rev Biochem 56:779, 1987.
39. CHOW M, DER CJ, BUSS JE: Structure and biological effects of lipid modifications on proteins. Curr Opin Cell Biol 4: 629, 1992.
40. KHOSRAVI-FAR R, LUTZ RJ, COX AD, CONROY L, BOURNE JR, SINENSKY M, BALCH WE, BUSS JE, DER CJ: Isoprenoid modification of r
ab proteins terminating in CC or CXC motifs. Proc Natl Acad Sci USA 88:6264-6268, 1991.
41. KINSELLA BT, MALTESE WA: rab GTP-binding proteins with three different carboxyl-terminal cysteine motifs are modified in vivo by 20-carbon isoprenoids. J Biol Ch
em 267: 3940, 1992.
42. SEABRA MC, GOLDSTEIN JL, SUDHOF TC, BROWN MS: Rab geranylgeranyl transferase. J Biol Chem 267:14497, 1992.
43. NEWMAN CMH, GIANNAKOUROS T, HANCOCK JF, FAWELL EH, ARMSTRONG J, MAGEE AI: Post-translational processing of Schizosacch
aromyces pombe YPT proteins. J Biol Chem 267:11329, 1992.
44. KAHN RA, GODDARD C, NEWKIRK, M: Chemical and immunological characterization of the 21-kDa ADP-ribosylation factor of adenylation cyclase. J Biol Chem 263:8282, 1988.
45. KAHN RA, KERN FG, CLA
RK J, GELMANN EP, RULKA C: Human ADP-ribosylation factors: a functionally conserved family of GTP-binding proteins. J Biol Chem 266:2606, 1991.
46. BURSTEIN ES, LINKO-STENTZ K, LU Z, MACARA IG: Regulation of the GTPase activity of the ras-like protein p
25rab3A. J Biol Chem 266:2689, 1991.
47. KAIBUCHI K, MIZUNO T, FUJIOKA H, YAMAMOTO T, KISHI K, FUKUMOTO Y, HORI Y, TAKAI Y: Molecular cloning of the cDNA for stimulatory GDP/GTP exchange protein for smg p21s (ras p21-like small GTP-binding proteins) and
characteriza tion of stimulatory GDP/GTP exchange protein. Mol Cell Biol 11:2873, 1991.
48. MOYA M, ROBERTS D, NOVICK P: DSS4-1 is a dominant suppressor of sec4-8 that encodes a nucleotide exchange protein that aids Sec4p function. Nature 361:460, 1993.
49. BURTON J, ROBERTS D, MONTALDI M, NOVICK P, De CAMILLI P: A mammalian guanine-nucleotide-releasing protein enhances function of yeast secretory protein Sec4. Nature 361:464, 1993.
50. ARKI S, KIKUCHI A, HATA Y, ISOMURA M, TAKAI Y: Regulation of rever
sible binding of smgp25A, a ras p21-like GTP-binding protein, to synaptic plasma membranes and vesicles by its specific regulatory protein, GDP dissociation inhibitor. J Biol Chem 265:13007, 1990.
51. SASAKI T, KIKUCHI A, ARAKI S, HATA Y, ISOMURA M, KUROD
A S, TAKAI Y: Purification and characterization from bovine brain cytosol of a protein that inhibits the dissociation of GDP from and the subsequent binding of GTP to smg p25A, a ras p21-like GTP-binding protein. J Biol Chem 265:2333, 1990.
52. KAHN RA:
Fluoride is not an activator of the smaller (20-25 kDa) GTP-binding proteins. J Biol Chem 266:15595, 1991.
53. ROSS EM: Signal sorting and amplification through G protein-coupled receptors. Neuron 3:141, 1989.
54. WEINGARTEN R, RANSNÄS L, MUELLER H, SK
LAR LA, BOKOCH GM: Mastoparan interacts with the carboxy terminus of the a subunit of Gi. J Biol Chem 265:11044, 1990.
55. KAHN RA, GILMAN AG: The protein cofactor necessary for ADP-ribosylation of Gs by cholera toxin is itself a GTP binding. J Biol Ch
em 261:7906, 1986.
56. PALADE G: Intracellular aspects of the process of protein synthesis. Science 189:347, 1975.
57. RAPOPORT TA. Protein transport across the endoplasmic reticulum membrane: facts, models, mysteries. FASEB J 5:2792, 1991.
58. WALTER
P, LINGAPPA VR: Mechanism of protein translocation across the endoplasmic reticulum membrane. Ann Rev Cell Biol 2:499, 1986.
59. SKACH WR, LINGAPPA VR. Intracellular trafficking of pre-(pro) proteins across RER membranes in Mechanisms of Intracellular t
rafficking and processing of proproteins, Y.P. Loh (ed). CRC Press, Boca Raton, pp 19-77, 1993.
60. SABATINI DD, TASHIRO Y, PALADE GE: On the attachment of ribosomes to microsomal membranes. J Mol Biol 19:503, 1966.
61. KREIBICH G, DEBEY P, SABATINI D
D: Selective release of content from microsomal vesicles without membrane disassembly. I. Permeability changes induced by low detergent concentrations. J Cell Biol 58:436, 1973.
62. KREIBICH G, SABATINI DD: Selective release of content from microsomal v
esicles without membrane disassembly. II. Electrophoretic and immunological characterization of microsomal subfractions. J Cell Biol 61:789, 1974.
63. KREIBICH G, ULRICH BL, SABATINI DD: Proteins of rough microsomal membranes related to ribosome binding.
I. Identification of ribophorins I and II, membrane proteins characteristic of rough microsomes. J Cell Biol 77:464, 1978.
64. KREIBICH, G, FREIENSTEIN CM, PEREYRA BN, ULRICH BL, SABATINI DD: Proteins of rough microsomal membranes related to ribosome bi
nding. II. Cross-linking of bound ribosomes to the specific membrane proteins exposed at the binding sites. J Cell Biol 77:488, 1978.
65. HORTSCH M, MEYER DI: Immunochemical analysis of rough and smooth microsomes from rat liver. Segregation of docking
protein in rough membranes. Eur J Biochem 150:559, 1985.
66. LEWIS JA, SABATINI DD: Accessibility of proteins in rat liver-free and membrane-bound ribosomes to lactoperoxidase-catalyzed iodination. J Biol Chem 252:5547, 1977.
67. LEWIS JA, SABATINI DD:
Proteins of rat liver free and membrane-bound ribosomes: modification of two large subunit proteins by a factor detached from ribosomes at high ionic strength. Biochim Biophys Acta 478:33, 1977.
68. BORGESE D, BLOBEL G, SABATINI DD: In vitro exchange o
f ribosomal subunits between free and membrane-bound ribosomes. J Mol Biol 74:415, 1973.
69. BLOBEL G, SABATINI DD: Ribosome-membrane interaction in eukaryotic cells, in Manson LA (ed): Biomembranes. New York, Plenum, 1971, vol 2, p 193.
70. MILSTEIN C,
BROWNLEE GG, HARRISON TM, MATHEWS MB: A possible precursor of immunoglobulin light chains. Nature New Biol 239:117, 1972.
71. SWAN D, AVIV H, LEDER P: Purification and properties of biologically active messenger RNA for a myeloma light chain. Proc Nat
l Acad Sci USA 69:1967, 1972.
72. SCHECHTER I: Biologically and chemically pure mRNA coding for a mouse immunoglobulin L-chain prepared with the aid of antibodies and immobilized oligothymidine. Proc Natl Acad Sci USA 70:2256, 1973.
73. BLOBEL G, DOBBERS
TEIN B: Transfer of proteins across membranes. I. Presence of proteolytically processed and unprocessed nascent immunoglobulin light chains on membrane-bound ribosomes of murine myeloma. J Cell Biol 67:835, 1975.
74. BLOBEL G, DOBBERSTEIN B: Transfer of
proteins across membranes. II. Reconstitution of functional rough microsomes from heterologous components. J Cell Biol 67:852, 1975.
75. PERLMAN D, HALVORSON HO: A putative signal peptidase recognition site and sequence in eukaryotic and prokaryotic sig
nal peptides. J Mol Biol 167:391, 1983.
76. VON HEIJNE G: Signal sequences: The limits of variation. J Mol Biol 184:99, 1985.
77. VON HEIJNE G: Towards a comparative anatomy of N-terminal topogenic protein sequences. J Mol Biol 189:239, 1986.
78. WATSON
MEE: Compilation of published signal sequences. Nucleic Acids Res 12:5145, 1984.
79. REDMAN CM, SABATINI DD: Vectorial discharge of peptides released by puromycin from attached ribosomes. Proc Natl Acad Sci USA 56:608, 1966.
80. MALKIN LI, RICH A: Pa
rtial resistance of nascent polypeptide chains to proteolytic digestion due to ribosomal shielding. J Mol Biol 26:329, 1967.
81. BLOBEL G, SABATINI DD: Controlled proteolysis of nascent polypeptides in rat liver cell fractions. I. Location of the polypep
tides within ribosomes. J Cell Biol 45:130, 1970.
82. WARREN G, DOBBERSTEIN B: Protein transfer across microsomal membranes reassembled from separated membrane components. Nature 273:569, 1978.
83. WALTER P, BLOBEL G: Purification of a membrane-associa
ted protein complex required for protein translocation across the endoplasmic reticulum. Proc Natl Acad Sci USA 77:7112, 1980.
84. WALTER P, IBRAHIMI I, BLOBEL G: Translocation of proteins across the endoplamsic reticulum. I. Signal recognition protein (
SRP) binds to in vitro assembled polysomes synthesizing secretory proteins. J Cell Biol 91:545, 1981.
85. WALTER P, BLOBEL G: Translocation of proteins across the endoplasmic reticulum. II. Signal recognition protein (SRP) mediates the selected binding t
o microsomal membranes of in-vitro-assembled polysomes synthesizing secretory proteins. J Cell Biol 91:551, 1981.
86. WALTER P, BLOBEL G: Translocation of proteins across the endoplasmic reticulum. III. Signal recognition protein (SRP) causes signal seq
uence-dependent and site-specific arrest of chain elongation that is released by microsomal membranes. J Cell Biol 91:557, 1981.
87. GILMORE R, BLOBEL G, WALTER P: Protein translocation across the endoplasmic reticulum. I. Detection in the microsomal mem
brane of a receptor for a signal recognition particle. J Cell Biol 95:463, 1982.
88. GILMORE R, WALTER P, BLOBEL G: Protein translocation across the endoplasmic reticulum. II. Isolation and characterization of the signal recognition particle receptor.
J Cell Biol 95:477, 1982.
89. WALTER P, GILMORE R, BLOBEL G: Protein translocation across the endoplasmic reticulum. Cell 38:5, 1984.
90. MEYER DI, DOBBERSTEIN B: Identification and characterization of a membrane component essential for the translocatio
n of proteins across the membrane of the endoplasmic reticulum. J Cell Biol 87:503, 1980.
91. MEYER DI, DOBBERSTEIN B: A membrane component essential for vectorial translocation of nascent proteins across the endoplasmic reticulum: requirements for its e
xtraction and reassociation with the membrane. J Cell Biol 87:498: 1980.
92. MEYER DI, KRAUSE E, DOBBERSTEIN B: Secretory protein translocation across membranes-the role of the "Docking Protein." Nature 297:647, 1982.
93. MEYER DI, LOUVARD D, DOBBERSTEI
N B: Characterization of molecules involved in protein translocation using specific antibodies. J Cell Biol 92:579, 1982.
94. BORGESE N, MOK W, KREIBICH G, SABATINI DD: Ribosomal-membrane interaction: in vitro binding of ribosomes to microsomal membrane
s. J Mol Biol 88:559, 1974.
95. GILMORE R, BLOBEL G: Transient involvement of signal recognition particle and its receptor in the microsomal membrane prior to protein translocation. Cell 35:677, 1983.
96. WIEDMAN M, KURZCHALIA TV, HARTMANN E, RAPOPORT T
A: A signal sequence receptor in the endoplasmic reticulum membrane. Nature 328:830, 1987.
97. KRIEG UC, JOHNSON AE, WALTER P: Protein translocation across the endoplasmic reticulum membrane: Identification by photocross-linking of a 39-kD integral mem
brane glycoprotein as part of a putative translocation tunnel. J Cell Biol 109:2033, 1989.
98. KELLARIS KV, BOWEN S, GILMORE R: ER translocation intermediates are adjacent to a nonglycosylated 34-kD integral membrane protein. J Cell Biol 114:21, 1991.
99. THRIFT RN, ANDREWS DW, WALTER P, JOHNSON AE: A nascent membrane protein is located adjacent to ER membrane proteins throughout its integration and translation. J Cell Biol 112:809, 1991.
100. GÖRLICH D, HARTMANN E, PREHN S, RAPOPORT TA: A protein of
t he endoplasmic reticulum involved early in polypeptide translocation. Nature 357:47, 1992.
101. SABATINI DD, BLOBEL G: Controlled proteolysis of nascent polypeptides in rat liver cell fractions II. Location of the polypeptides in rough microsomes. J C
ell Biol 45:146, 1970.
102. GILMORE R, BLOBEL G: Translocation of secretory proteins across the microsomal membrane occurs through an environment accessible to aqueous perturbants. Cell 42:497, 1985.
103. SIMON SM, BLOBEL G: A protein-conducting channel
in the endoplasmic reticulum. Cell 65:371, 1991.
104. SIMON SM, BLOBEL G: Signal peptides open protein-conducting channels in E. coli. Cell 69:677, 1992.


(*): Reprinted with permission from "The Metabolic and Molecular Bases of Inherited Disease", Vol. I, edited by C.R. Scriver, W.S. Sly and D. Valle, The McGraw-Hilll Companies, New York, N.Y. 10020-1095.

INDEX / INDICECONTRIBUTIONS