Anchor sequences determine protein orientation

  • An anchor (stop-transfer) (often referred to as a "transmembrane anchor") is a segment of a transmembrane protein which resides in the membrane.
  • An anchor sequence halts the passage of a protein through the translocon. Typically this is located at the C-terminal end and results in a group I orientation in which the N-terminus has passed through the membrane.
  • A combined signal-anchor sequence can be used to insert a protein into the membrane and anchor the site of insertion. Typically this is internal and results in a group II orientation in which the N-terminus is cytosolic. 

Proteins that are secreted from the cell pass through a membrane while remaining in the aqueous channel of the translocon. By contrast, proteins that reside in membranes start the process in the same way, but then transfer from the aqueous channel into the hydrophobic environment. The challenge in accounting for insertion of proteins into membranes is to explain what distinguishes transmembrane proteins from secreted proteins, and causes this transfer. The pathway by which proteins of either type I or type II are inserted into the membrane follows the same initial route as that of secretory proteins, relying on a signal sequence that functions co-translationally. But proteins that are to remain within the membrane possess a second, stop-transfer signal. This takes the form of a cluster of hydrophobic amino acids adjacent to some ionic residues. The cluster serves as an anchor that latches on to the membrane and stops the protein from passing right through.

A surprising property of anchor sequences is that they can function as signal sequences when engineered into a different location. When placed into a protein lacking other signals, such a sequence may sponsor membrane translocation. One possible explanation for these results is that the signal sequence and anchor sequence interact with some common component of the apparatus for translocation. Binding of the signal sequence initiates translocation, but the appearance of the anchor sequence displaces the signal sequence and halts transfer.

Membrane insertion starts by the insertion of a signal sequence in the form of a hairpin loop, in which the N-terminus remains on the cytoplasmic side. Two features determine the position and orientation of a protein in the membrane: whether the signal sequence is cleaved; and the location of the anchor sequence.

The insertion of type I proteins is illustrated in Figure 8.33. The signal sequence is N-terminal. The location of the anchor signal determines when transfer of the protein is halted. When the anchor sequence takes root in the membrane, domains on the N-terminal side will be located in the lumen, while domains on the C-terminal side are located facing the cytosol.

A common location for a stop-transfer sequence of this type is at the C-terminus. As shown in the figure, transfer is halted only as the last sequences of the protein enter the membrane. This type of arrangement is responsible for the location in the membrane of many proteins, including cell surface proteins. Most of the protein sequence is exposed on the lumenal side of the membrane, with a small or negligible tail facing the cytosol.

Type II proteins do not have a cleavable leader sequence at the N-terminus. Instead the signal sequence is combined with an anchor sequence. We imagine that the general pathway for the integration of type I proteins into the membrane involves the steps illustrated in Figure 8.34. The signal sequence enters the membrane, but the joint signal-anchor sequence does not pass through. Instead it stays in the membrane (perhaps interacting directly with the lipid bilayer), while the rest of the growing polypeptide continues to loop into the endoplasmic reticulum.

The signal-anchor sequence is usually internal, and its location determines which parts of the protein remain in the cytosol and which are extracellular. Essentially all the N-terminal sequences that precede the signal-anchor are exposed to the cytosol. Usually the cytosolic tail is short, ~6-30 amino acids. In effect the N-terminus remains constrained while the rest of the protein passes through the membrane. This reverses the orientation of the protein with regard to the membrane. 

The combined signal-anchor sequences of type II proteins resemble cleavable signal sequences. Figure 8.35 gives an example. Like cleavable leader sequences, the amino acid composition is more important than the actual sequence. The regions at the extremities of the signal-anchor carry positive charges; the central region is uncharged and resembles a hydrophobic core of a cleavable leader. Mutations to introduce charged amino acids in the core region prevent membrane insertion; mutations on either side prevent the anchor from working, so the protein is secreted or located in an incorrect compartment.
The distribution of charges around the anchor sequence has an important effect on the orientation of the protein. More positive charges are usually found on the cytoplasmic side (N-terminal side in type II proteins). If the positive charges are removed by mutation, the orientation of the protein can be reversed. The effect of charges on orientation is summarized by the "positive inside" rule, which says that the side of the anchor with the most positive charges will be located in the cytoplasm. The positive charges in effect provide a hook that latches on to the cytoplasmic side of the membrane, controlling the direction in which the hydrophobic region is inserted, and thus determining the orientation of the protein.

Proteins reside in membranes by means of hydrophobic regions

  • The transmembrane region (transmembrane domain) is the part of a protein that spans the membrane bilayer. It is hydrophobic and in many cases contains approximately 20 amino acids that form an α-helix. It is also called the transmembrane domain.
  • A transmembrane protein (Integral membrane protein) extends across a lipid bilayer. A hydrophobic region (typically consisting of a stretch of 20-25 hydrophobic and/or uncharged aminoa acids) or regions of the protein resides in the membrane. Hydrophilic regions are exposed on one or both sides of the membrane.
  • Group I proteins have the N-terminus on the far side of the membrane; group II proteins have the opposite orientation.
  • Some proteins have multiple membrane-spanning domains.  

All biological membranes contain proteins, which are held in the lipid bilayer by noncovalent interactions. The operational definition of an integral membrane protein is that it requires disruption of the lipid bilayer in order to be released from the membrane. A common feature in such proteins is the presence of at least one transmembrane domain, consisting of an α-helical stretch of 21-26 hydrophobic amino acids. A sequence that fits the criteria for membrane insertion can be identified by a hydropathy plot, which measures the cumulative hydrophobicity of a stretch of amino acids. A protein that has domains exposed on both sides of the membrane is called a transmembrane protein. The association of a protein with a membrane takes several forms (see 32.5 Membranes and membrane proteins). The topography of a membrane protein depends on the number and arrangement of transmembrane regions.
When a protein has a single transmembrane region, its position determines how much of the protein is exposed on either side of the membrane. A protein may have extensive domains exposed on both sides of the membrane or may have a site of insertion close to one end, so that little or no material is exposed on one side. The length of the N-terminal or C-terminal tail that protrudes from the membrane near the site of insertion varies from insignificant to quite bulky.

Figure 8.31 shows that proteins with a single transmembrane domain fall into two classes. Group I proteins in which the N-terminus faces the extracellular space are more common than group II proteins in which the orientation has been reversed so that the N-terminus faces the cytoplasm. Orientation is determined during the insertion of the protein into the endoplasmic reticulum.

Figure 8.32 shows orientations for proteins that have multiple membrane-spanning domains. An odd number means that both termini of the protein are on opposite sides of the membrane, whereas an even number implies that the termini are on the same face. The extent of the domains exposed on one or both sides is determined by the locations of the transmembrane domains. Domains at either terminus may be exposed, and internal sequences between the domains "loop out" into the extracellular space or cytoplasm. One common type of structure is the 7-membrane passage or "serpentine" receptor; another is the 12-membrane passage component of an ion channel.
Does a transmembrane domain itself play any role in protein function besides allowing the protein to insert into the lipid bilayer? In the simple group I or II proteins, it has little or no additional function; often it can be replaced by any other transmembrane domain. However, transmembrane domains play an important role in the function of proteins that make multiple passes through the membrane or that have subunits that oligomerize within the membrane. The transmembrane domains in such cases often contain polar residues, which are not found in the single membrane-spanning domains of group I and group II proteins. Polar regions in the membrane-spanning domains do not interact with the lipid bilayer, but instead interact with one another. This enables them to form a polar pore or channel within the lipid bilayer. Interaction between such transmembrane domains can create a hydrophilic passage through the hydrophobic interior of the membrane. This can allow highly charged ions or molecules to pass through the membrane, and is important for the function of ion channels and transport of ligands. Another case in which conformation of the transmembrane domains is important is provided by certain receptors that bind lipophilic ligands. In such cases, the transmembrane domains (rather than the extracellular domains) bind the ligand within the plane of the membrane.

Reverse translocation sends proteins to the cytosol for degradation

  • Retrograde translocation (Reverse translocation) is the translocation of a protein from the lumen of the ER to the cytoplasm. It usually occurs to allow misfolded or damaged proteins to be degraded by the proteasome.
  • Sec61 translocons can be used for reverse translocation of proteins from the ER into the cytosol. 

Several important activities occur within the endoplasmic reticulum. Proteins move through the ER en route to a variety of destinations (see 27 Protein trafficking). They are glycosylated and folded into their final conformations. The ER provides a "quality control" system in which misfolded proteins are identified and degraded. However, the degradation itself does not occur in the ER, but may require the protein to be exported back to the cytosol.
The first indication that ER proteins are degraded in the cytosol and not in the ER itself was provided by evidence for the involvement of the proteasome, a large protein aggregate with several proteolytic activities (see 8.32 The proteasome is a large machine that degrades ubiquitinated proteins). Inhibitors of the proteasome prevent the degradation of aberrant ER proteins. Proteins are marked for cleavage by the proteasome when they are modified by the addition of ubiquitin, a small polypeptide chain (see 8.31 Ubiquitination targets proteins for degradation). The important point to note now is that ubiquitination and proteasomal degradation both occur in the cytosol (with a minor proportion in the nucleus).
Transport from the ER back into the cytosol occurs by a reversal of the usual process of import (for review see Tsai, Ye, and Rapoport, 2002). This is called reverse translocation. The Sec61 translocon is used. The conditions are different; for example, the translocon is not associated with a ribosome. Some mutations in Sec61 prevent reverse translocation, but do not prevent forward translocation (Zhou and Schekman, 1999; Wilkinson et al., 2000). This could be either because there is some difference in the process or (more likely) because these regions interact with other components that are necessary for reverse translocation.

Figure 8.30 points out that we do not know how the channel is opened to allow insertion of the protein on the ER side. Special components are presumably involved. One model is that misfolded or misassembled proteins are recognized by chaperones, which transfer them to the translocon (for review see Johnson and Haigh, 2000). In one particular case, human cytomegalovirus (CMV) codes for cytosolic proteins that destroy newly synthesized MHC class I (cellular major histocompatibility complex) proteins. This requires a viral protein product (US2), which is a membrane protein that functions in the ER. It interacts with the MHC proteins and probably conveys them into the translocon for reverse translocation.
The system involved in the degradation of aberrant ER proteins can be identified by mutations (in yeast) that lead to accumulation of aberrant proteins. Usually a protein that misfolds (produced by a mutated gene) is degraded instead of being transported through the ER. Yeast mutants that cannot degrade the substrate fall into two classes: some identify components of the proteolytic apparatus, such as the enzymes involved in ubiquitination; other identify components of the transport apparatus, including Sec61, BiP, and Sec63. There is also a protein in the ER membrane that functions on the cytosolic side to localize the ubiquitination enzymes at the translocon. In fact, retrograde transport into the cytosol cannot occur in the absence of this protein, which suggests that there is a mechanical link between retrograde transport and degradation (Wiertz et al., 1996).

Translocation requires insertion into the translocon and (sometimes) a ratchet in the ER

  • The ribosome, SRP, and SRP receptor are sufficient to insert a nascent protein into a translocon.
  • Proteins that are inserted post-translationally require additional components in the cytosol and Bip in the ER.
  • Bip is a ratchet that prevents a protein from slipping backward.
The translocon and the SRP receptor are the basic components required for co-translational translocation. When the Sec61 complex is incorporated into artificial membranes together with the SRP receptor, it can support translocation of some nascent proteins. Other nascent proteins require the presence of an additional component, TRAM, which is a major protein that becomes crosslinked to a translocating nascent chain. TRAM stimulates the translocation of all proteins.

The components of the translocon and their functions are summarized in Figure 8.28. The simplicity of this system makes several important points. We visualize Sec61 as forming the channel and also as interacting with the ribosome. The initial targeting is made when the SRP recognizes the signal sequence as the newly synthesized protein begins to emerge from the ribosome. The SRP binds to the SRP receptor, and the signal sequence is transferred to the translocon. When the signal sequence enters the translocon, the ribosome attaches to Sec61, forming a seal so that the pore is not exposed to the cytosol. Cleavage of the signal peptide does not occur in this system, and therefore cannot be necessary for translocation per se. In this system, components on the lumenal side of the membrane are not needed for translocation.
Of course, the efficiency of the in vitro system is relatively low. Additional components could be required in vivo to achieve efficient transfer or to prevent other cellular proteins from interfering with the process (Gorlich and Rapoport, 1993; for review see Walter and Lingappa, 1986; Rapoport, Jungnickel, and Kutay, 1996).

A more complex apparatus is required in certain cases in which a protein is inserted into a membrane post-translationally. The same Sec61 complex forms the channel, but four other Sec proteins are also required, and in addition the chaperone BiP (a member of the Hsp70 class) and a supply of ATP are required on the lumenal side of the membrane. Figure 8.29 shows that BiP behaves as ratchet (Matlack et al., 1999). In the absence of BiP, Brownian motion allows the protein to slip back into the cytosol. But BiP grabs the protein as it exits the pore into the endoplasmic reticulum. This stops the protein from moving backward. BiP does not pull the protein through; it just stops it from sliding back. (The reason why BiP is required for post-translational translocation but not for co-translational translocation may be that a newly synthesized protein is continuously extruded from the ribosome and therefore cannot slip backward.)

The translocon forms a pore

  • A translocon is a discrete structure in a membrane that forms a channel through which (hydrophilic) proteins may pass.
  • The Sec61 trimeric complex provides the channel for proteins to pass through a membrane.
  • A translocating protein passes directly from the ribosome to the translocon without exposure to the cytosol. 

There is a basic problem in passing a (largely) hydrophilic protein through a hydrophobic membrane. The energetics of the interaction between the charged protein and the hydrophobic lipids are highly unfavorable. However, a protein in the process of translocation across the ER membrane can be extracted by denaturants that are effective in an aqueous environment. The same denaturants do not extract proteins that are resident components of the membrane. This suggests the model for translocation illustrated in Figure 8.26, in which proteins that are part of the ER membrane form an aqueous channel through the bilayer. A translocating protein moves through this channel, interacting with the resident proteins rather than with the lipid bilayer. The channel is sealed on the lumenal side to stop free transfer of ions between the ER and the cytosol.
The channel through the membrane is called the translocon. Its components have been identified in two ways. Resident ER membrane proteins that are crosslinked to translocating proteins are potential subunits of the channel (Mothes, Prehn, and Rapoport, 1994). And sec mutants in yeast (named because they fail to secrete proteins) include a class that cause precursors of secreted or membrane proteins to accumulate in the cytosol (Deshaies and Schekman, 1987; Esnault et al., 1993). These approaches together identified the Sec61 complex, which consists of three transmembrane proteins: Sec61α,β,γ. Sec61 is the major component of the translocon. In detergent (which provides a hydrophobic milieu that mimics the effect of a surrounding membrane), Sec61 forms cylindrical oligomers with a diameter of ~85Å and a central pore of ~20Å. Each oligomer consists of 4 heterotrimers (Hanein et al., 1996).
Is the channel a preexisting structure (as implied in the figure) or might it be assembled in response to the association of a hydrophobic signal sequence with the lipid bilayer? Channels can be detected by their ability to allow the passage of ions (measured as a localized change in electrical conductance). Ion-conducting channels can be detected in the ER membrane, and their state depends on protein translocation (Simon and Blobel, 1991; Crowley, 1994). This demonstrates that the channel is a permanent feature of the membrane.
A channel opens when a nascent polypeptide is transferred from a ribosome to the ER membrane. The translocating protein fills the channel completely, so ions cannot pass through during translocation. But if the protein is released by treatment with puromycin, then the channel becomes freely permeable. If the ribosomes are removed from the membrane, the channel closes, suggesting that the open state requires the presence of the ribosome. This suggests that the channel is controlled in response to the presence of a translocating protein.
Measurements of the abilities of fluorescence quenching agents of different sizes to enter the channel suggest that it is large, with an internal diameter of 40-60Å. This is much larger than the diameter of an extended α-helical stretch of protein. It is also larger than the pore seen in direct views of the channel; this discrepancy remains to be explained (Simon and Blobel, 1991).
The aqueous environment of an amino acid in a protein can be measured by incorporating variant amino acids that have photoreactive residues. The fluorescence of these residues indicates whether they are in an aqueous or hydrophobic environment. Experiments with such probes show that when the signal sequence is first synthesized in the ribosome, it is in an aqueous state, but is not accessible to ions in the cytosol. It remains in the aqueous state throughout its interaction with a membrane. This suggests that the translocating protein travels directly from an enclosed tunnel in the ribosome into an aqueous channel in the membrane.

In fact, access to the pore is controlled (or "gated") on both sides of the membrane. Before attachment of the ribosome, the pore is closed on the lumenal side. Figure 8.27 shows that when the ribosome attaches, it seals the pore on the cytosolic side. When the nascent protein reaches a length of ~70 amino acids, that is, probably when it extends fully across the channel, the pore opens on the lumenal side. So at all times, the pore is closed on one side or the other, maintaining the ionic integrities of the separate compartments (Crowley, 1994; Liao et al., 1997).
The translocon is versatile, and can be used by translocating proteins in several ways:

The SRP interacts with the SRP receptor

  • The Alu domain comprises the parts of the 7S RNA of the SRP that are related to Alu RNA.
  • The S domain is the sequence of 7S RNA of the SRP that is not related to Alu RNA.
  • The SRP is a complex of 7S RNA with 6 proteins.
  • The bacterial equivalent to the SRP is a complex of 4.5S RNA with two proteins.
  • The SRP receptor is a dimer.
  • GTP hydrolysis releases the SRP from the SRP receptor after their interaction. 

The interaction between the SRP and the SRP receptor is the key event in eukaryotic translation in transferring a ribosome carrying a nascent protein to the membrane (for review see Keenan et al., 2001). An analogous interacting system exists in bacteria, although its role is more restricted.

The SRP is an 11S ribonucleoprotein complex, containing 6 proteins (total mass 240 kD) and a small (305 base, 100 kD) 7S RNA. Figure 8.24 shows that the 7S RNA provides the structural backbone of the particle; the individual proteins do not assemble in its absence (Walter and Blobel, 1981; Walter and Blobel, 1982).
The 7S RNA of the SRP particle is divided into two parts. The 100 bases at the 5 end and 45 bases at the 3 end are closely related to the sequence of Alu RNA, a common mammalian small RNA. They therefore define the Alu domain. The remaining part of the RNA comprises the S domain.
Different parts of the SRP structure depicted in Figure 8.24 have separate functions in protein targeting. SRP54 is the most important subunit. It is located at one end of the RNA structure, and is directly responsible for recognizing the substrate protein by binding to the signal sequence(Zopf et al., 1990). It also binds to the SRP receptor in conjunction with the SRP68-SRP72 dimer that is located at the central region of the RNA. The SRP9-SRP14 dimer is located at the other end of the molecule; it is responsible for elongation arrest (Siegel and Walter, 1988). 

The SRP is a flexible structure. In its unengaged form (not bound to signal sequence), it is quite extended, as can be seen from the crystal structure of Figure 8.24. Figure 8.25 shows that binding to a signal sequence triggers a change of conformation, and the protein bends at a hinge to allow the SRP54 end to contact the ribosome at the protein exit site, while the SRP19 swings around to contact the ribosome at the elongation factor binding site (Halic et al., 2004). This enables it to cause the elongation arrest that gives time for targeting to the translocation site on the membrane.
The SRP receptor is a dimer containing subunits SRα (72 kD) and SRβ (30 kD) (Tajima et al., 1986). The β subunit is an integral membrane protein. The amino-terminal end of the large α subunit is anchored by the β subunit. The bulk of the α protein protrudes into the cytosol. A large part of the sequence of the cytoplasmic region of the protein resembles a nucleic acid-binding protein, with many positive residues. This suggests the possibility that the SRP receptor recognizes the 7S RNA in the SRP.
There is a counterpart to SRP in bacteria, although it contains fewer components. E. coli contains a 4.5S RNA that associates with ribosomes and is homologous to the 7S RNA of the SRP. It associates with two proteins: Ffh is homologous to SRP54. FtsY is homologous to the α subunit of the SRP receptor. In fact, FtsY replaces the functions of both the α and β SRP subunits; its N-terminal domain substitutes for SRPβ in membrane targeting, and the C-terminal domain interacts with the target protein. The role of this complex is more limited than that of SRP-SRP receptor. It is probably required to keep some (but not all) secreted proteins in a conformation that enables them to interact with the secretory apparatus. This could be the original connection between protein synthesis and secretion; in eukaryotes the SRP has acquired the additional roles of causing translational arrest and targeting to the membrane.
Why should the SRP have an RNA component? The answer must lie in the evolution of the SRP: it must have originated very early in evolution, in an RNA-dominated world, presumably in conjunction with a ribosome whose functions were mostly carried out by RNA. The crystal structure of the complex between the protein-binding domain of 4.5S RNA and the RNA-binding domain of Ffh suggests that RNA continues to play a role in the function of SRP.
The 4.5S RNA has a region (domain IV) that is very similar to domain IV in 7S RNA (see Figure 8.24). Ffh consists of three domains (N, G, and M). The M domain (named for a high content of methionines) performs the key binding functions (Keenan et al., 1998). It has a hydrophobic pocket that binds the signal sequence of a target protein. The hydrophobic side chains of the methionine residues create the pocket by projecting into a cleft in the protein structure. Next to the pocket is a helix-turn-helix motif that is typical of DNA-binding proteins (see 12.12 Repressor uses a helix-turn-helix motif to bind DNA).
The crystal structure shows that the helix-loop-helix of the M domain binds to a duplex region of the 4.5S RNA in domain IV (Batey et al., 2000). The negatively charged backbone of the RNA is adjacent to the hydrophobic pocket. This raises the possibility that a signal sequence actually binds to both the protein and RNA components of the SRP. The positively charged sequences that start the signal sequence (see Figure 8.20) could interact with the RNA, while the hydrophobic region of the signal sequence could sit in the pocket.

GTP hydrolysis plays an important role in inserting the signal sequence into the membrane. Both the SRP and the SRP receptor have GTPase capability. The signal-binding subunit of the SRP, SRP54, is a GTPase. And both subunits of the SRP receptor are GTPases. All of the GTPase activities are necessary for a nascent protein to be transferred to the membrane. Figure 8.23 shows that the SRP starts out with GDP when it binds to the signal sequence. The ribosome then stimulates replacement of the GDP with GTP. The signal sequence inhibits hydrolysis of the GTP. This ensures that the complex has GTP bound when it encounters the SRP receptor.

For the nascent protein to be transferred from the SRP to the membrane, the SRP must be released from the SRP receptor. Figure 8.22 shows that this requires hydrolysis of the GTPs of both the SRP and the SRP receptor. The reaction has been characterized in the bacterial system, where it has the unusual feature that Ffh activates hydrolysis by FtsY, and FtsY reciprocally activates hydrolysis by Ffh (Powers and Walter, 1995).

The signal sequence interacts with the SRP

  • Protein translocation describes the movement of a protein across a membrane. This occurs across the membranes of organelles in eukaryotes, or across the plasma membrane in bacteria. Each membrane across which proteins are translocated has a channel specialized for the purpose.
  • The signal recognition particle (SRP) is a ribonucleoprotein complex that recognizes signal sequences during translation and guides the ribosome to the translocation channel. SRPs from different organisms may have different compositions, but all contain related proteins and RNAs.
  • Signal peptidase is an enzyme within the membrane of the ER that specifically removes the signal sequences from proteins as they are translocated. Analogous activities are present in bacteria, archaebacteria, and in each organelle in a eukaryotic cell into which proteins are targeted and translocated by means of removable targeting sequences. Signal peptidase is one component of a larger protein complex.
  • The signal sequence binds to the SRP (signal recognition particle).
  • Signal-SRP binding causes protein synthesis to pause.
  • Protein synthesis resumes when the SRP binds to the SRP receptor in the membrane.
  • The signal sequence is cleaved from the translocating protein by the signal peptidase located on the "inside" face of the membrane.
Protein translocation can be divided into two general stages: first ribosomes carrying nascent polypeptides associate with the membranes; and then the nascent chain is transferred to the channel and translocates through it.
The attachment of ribosomes to membranes requires the signal recognition particle (SRP). The SRP has two important abilities:
  • It can bind to the signal sequence of a nascent secretory protein.
  • And it can bind to a protein (the SRP receptor) located in the membrane.
  The SRP and SRP receptor function catalytically to transfer a ribosome carrying a nascent protein to the membrane. The first step is the recognition of the signal sequence by the SRP. Then the SRP binds to the SRP receptor and the ribosome binds to the membrane. The stages of translation of membrane proteins are summarized in Figure 8.21.
The role of the SRP receptor in protein translocation is transient. When the SRP binds to the signal sequence, it arrests translation (Walter and Blobel, 1981). This usually happens when ~70 amino acids have been incorporated into the polypeptide chain (at this point the 25 residue leader has become exposed, with the next ~40 amino acids still buried in the ribosome).
Then when the SRP binds to the SRP receptor, the SRP releases the signal sequence. The ribosome becomes bound by another component of the membrane. At this point, translation can resume. When the ribosome has been passed on to the membrane, the role of SRP and SRP receptor has been played. They now recycle, and are free to sponsor the association of another nascent polypeptide with the membrane (for review see Walter and Johnson, 1994).
This process may be needed to control the conformation of the protein. If the nascent protein were released into the cytoplasm, it could take up a conformation in which it might be unable to traverse the membrane. The ability of the SRP to inhibit translation while the ribosome is being handed over to the membrane is therefore important in preventing the protein from being released into the aqueous environment.
The signal peptide is cleaved from a translocating protein by a complex of 5 proteins called the signal peptidase. The complex is several times more abundant than the SRP and SRP receptor. Its amount is equivalent roughly to the amount of bound ribosomes, suggesting that it functions in a structural capacity. It is located on the lumenal face of the ER membrane, which implies that the entire signal sequence must cross the membrane before the cleavage event occurs. Homologous signal peptidases can be recognized in eubacteria, archaea, and eukaryotes (Tjalsma et al., 1998).

Signal sequences initiate translocation

  • Protein sorting (targeting) is the direction of different types of proteins for transport into or between specific organelles.
  • A signal sequence is a short region of a protein that directs it to the endoplasmic reticulum for co-translational translocation.
  • Proteins associate with the ER system only co-translationally.
  • The signal sequence of the substrate protein is responsible for membrane association. 

Proteins that associate with membranes via N-terminal leaders use a hierarchy of signals to find their final destination. In the case of the reticuloendothelial system, the ultimate location of a protein depends on how it is directed as it transits the endoplasmic reticulum and Golgi apparatus. The leader sequence itself introduces the protein to the membrane; the intrinsic consequence of the interaction is for the protein to pass through the membrane into the compartment on the other side. For a protein to reside within the membrane, a further signal is required to stop passage through the membrane. Other types of signals are required for a protein to be sorted to a particular destination, that is, to remain within the membrane or lumen of some particular compartment. The general process of finding its ultimate destination by transport through successive membrane systems is called protein sorting or targeting, and is discussed in 27 Protein trafficking.

The overall nature of the pathway is summarized in Figure 8.18. The "default pathway" takes a protein through the ER, into the Golgi, and on to the plasma membrane. Proteins that reside in the ER possess a C-terminal tetrapeptide (KDEL, which actually provides a signal for them to return to the ER from the Golgi). The signal that diverts a protein to the lysosome is a covalent modification: the addition of a particular sugar residue. Other signals are required for a protein to become a permanent constituent of the Golgi or the plasma membrane. We discuss direction to these locations in 27 Protein trafficking.

There is a common starting point for proteins that associate with, or pass through, the reticuloendothelial system of membranes. These proteins can associate with the membrane only while they are being synthesized. The ribosomes synthesizing these proteins become associated with the endoplasmic reticulum, enabling the nascent protein to be co-translationally transferred to the membrane. Regions in which ribosomes are associated with the ER are sometimes called the "rough ER," in contrast with the "smooth ER" regions that lack associated polysomes and which have a tubular rather than sheet-like appearance (for review see Palade, 1975). Figure 8.19 shows ribosomes in the act of transferring nascent proteins to ER membranes.
The proteins synthesized at the rough endoplasmic reticulum pass from the ribosome directly to the membrane. Then they are transferred to the Golgi apparatus, and finally are directed to their ultimate destination, such as the lysosome or secretory vesicle or plasma membrane. The process occurs within a membranous environment as the proteins are carried between organelles in small membrane-coated vesicles (see 27 Protein trafficking.)

Co-translational insertion is directed by a signal sequence. Usually this is a cleavable leader sequence of 15-30 N-terminal amino acids. At or close to the N-terminus are several polar residues, and within the leader is a hydrophobic core consisting exclusively or very largely of hydrophobic amino acids (for review see von Heijne, 1985). There is no other conservation of sequence. Figure 8.20 gives an example.
The signal sequence is both necessary and sufficient to sponsor transfer of any attached polypeptide into the target membrane. A signal sequence added to the N-terminus of a globin protein, for example, causes it to be secreted through cellular membranes instead of remaining in the cytosol (Lingappa et al., 1984).
The signal sequence provides the connection that enables the ribosomes to attach to the membrane. There is no intrinsic difference between free ribosomes (synthesizing proteins in the cytosol) and ribosomes that are attached to the ER. A ribosome starts synthesis of a protein without knowing whether the protein will be synthesized in the cytosol or transferred to a membrane. It is the synthesis of a signal sequence that causes the ribosome to associate with a membrane (Blobel and Dobberstein, 1975; for review see Lee and Beckwith, 1986).

Hsp60/GroEL forms an oligomeric ring structure

  • Hsp60/GroEL forms an oligomeric structure consisting of 14 subunits arranged in two inverted heptameric rings.
  • A GroES heptamer forms a dome that caps one end of the double ring.
  • A substrate protein undergoes a cycle of folding in the cavity of one of the Hsp60/GroEL rings. It is released and rebound for further cycles until it reaches mature conformation.
  • Hydrolysis of ATP provides energy for the folding cycles.
Large (oligomeric) structures with hollow cavities are often used for handling the folding or degradation of proteins (for review see Horwich, Weber-Ban, and Finley, 1999). The typical structure is a ring of many subunits, forming a doughnut or cylinder. Figure 8.14 shows that the target protein is in effect placed in a controlled environmentthe cavitywhere it is closely associated with the surrounding protein. This creates a high local concentration of binding sites and supports cooperative interactions. In the case of folding, the closed environment prevents the target protein from forming wrongful interactions with other proteins, which may be important in driving folding along the proper pathway. In the case of degradation, isolation presumably makes for a more controlled process than would be possible in open cytosol (see 8.32 The proteasome is a large machine that degrades ubiquitinated proteins). The energy for these processes is provided by hydrolysis of ATPtypically the subunits of the ring are ATPases.

The Hsp60 class of chaperones forms a large apparatus that consists of two types of subunit. Figure 8.15 illustrates the structure schematically. Hsp60 itself (known as GroEL in E. coli) forms a structure consisting of 14 subunits that are arranged in two heptameric rings stacked on top of each other in inverted orientation. This means that the top and bottom surfaces of the double ring are the same. The central hole is blocked at the equator of each ring by the COOH ends of the subunits, which protrude into the interior. So the ends of the double cylinder form symmetrical cavities extending about half way into each unit.

This structure associates with a heptamer formed of subunits of Hsp10 (GroES in E. coli). A single GroES heptamer forms a dome that associates with one surface of the double ring, as shown in Figure 8.16. The dome sits over the central cavity, thus capping one opening of the cylinder. The dome is hollow and in effect extends the cavity into the closed surface. We can distinguish the two rings of GroEL as the proximal ring (bound to GroES) or the distal ring (not bound to GroES). The entire GroEL/GroES structure has a mass ~106 daltons, comparable to a small ribosomal subunit. GroEL is sometimes called a chaperonin, and GroES is called a co-chaperonin, because GroEL plays the essential role in guiding folding, but GroES is required for its activity (Braig et al., 1994; Hunt et al., 1996; Xu, Horwich, and Sigler, 1997).
GroEL binds to many unfolded proteins, probably by recognizing a condensed "molten globule" state. Interaction with the substrate is based on hydrophobic interactions between surfaces of the substrate and residues of GroEL that are exposed in its central cavity. Substrates may be provided by proteins that have become denatured; or they may be transferred to GroEL by other chaperonesfor example, Hsp70 may assist a nascent protein in folding, but then passes it on to GroEL for the process to be completed when it is released from the ribosome.

The key reactions in substrate binding and folding are illustrated in Figure 8.17. The reaction starts when substrate and ATP are bound to the same ring of GroEL. This defines the proximal ring. Then GroES caps this ring. Binding of GroES induces a conformational change in the proximal GroEL ring, increasing the volume of the central cavity. This also changes the environment for the substrate. The hydrophobic residues in GroEL that had previously bound substrate are involved in binding to GroES. The result is that the substrate now finds itself in a hydrophilic environment that forces a change in its conformation (Chen et al., 1994; Mayhew et al., 1996).
ATP plays an important role in GroEL function. Each subunit of GroEL has a molecule of ATP. The presence of ATP on the subunits of the proximal ring is required for folding to occur. Hydrolysis is required for the transition to the next stage. Hydrolysis of the ATP in the proximal ring changes the properties of the distal ring in such a way as to allow substrate and ATP to bind to it. This in turn triggers the dissociation of the substrate and GroES from the proximal ring. Now the situation at the start of the cycle has been restored. The ring that was the distal ring in the previous cycle is bound to substrate and ATP, and becomes the proximal ring for the next cycle. So the rings of GroEL alternate as proximal and distal (Rye et al., 1997; Rye et al., 1999).
An important question in the action of this (and other macromolecular) chaperones is whether their action is processive. Does a substrate enter the central cavity, undergo multiple cycles of folding within it, and become released in mature form? Or does it undergo a single folding cycle, after which it is released; typically it will still have improperly folded regions, and therefore will be bound again for another folding cycle. This process will continue until the protein has reached a mature conformation that does not offer a substrate to the chaperone.
These models have been tested by using a mutant GroEL that can bind unfolded proteins but cannot release them. When this "trap GroEL" is added to wild-type GroEL that is actively engaged with a substrate, it blocks the appearance of mature protein. This suggests that the substrate has been released before folding was completed. The simplest explanation is that substrate protein is released after each folding cycle. One cycle of folding, ATP hydrolysis, and release takes about 15 sec in vitro (Weissman et al., 1994; Weissman et al., 1995).

The Hsp70 family is ubiquitous

  • Hsp70 is a chaperone that functions on target proteins in conjunction with DnaJ and GrpE.
  • Members of the Hsp70 family are found in the cytosol, in the ER, and in mitochondria and chloroplasts.
The Hsp70 family is found in bacteria, eukaryotic cytosol, in the endoplasmic reticulum, and in chloroplasts and mitochondria (for review see Frydman, 2001). A typical Hsp70 has two domains: the N-terminal domain is an ATPase; and the C-terminal domain binds the substrate polypeptide (Flaherty, DeLuca-Flaherty, and McKay, 1990; Zhu et al., 1996). When bound to ATP, Hsp70 binds and releases substrates rapidly; when bound to ADP, the reactions are slow. Recycling between these states is regulated by two other proteins, Hsp40 (DnaJ) and GrpE.

Figure 8.13 shows that Hsp40 (DnaJ) binds first to a nascent protein as it emerges from the ribosome. Hsp40 contains a region called the J domain (named for DnaJ), which interacts with Hsp70. Hsp70 (DnaK) binds to both Hsp40 and to the unfolded protein. In effect, two interacting chaperones bind to the protein. The J domain accounts for the specificity of the pairwise interaction, and drives a particular Hsp40 to select the appropriate partner from the Hsp70 family.
The interaction of Hsp70 (DnaK) with Hsp40 (DnaJ) stimulates the ATPase activity of Hsp70. The ADP-bound form of the complex remains associated with the protein substrate until GrpE displaces the ADP. This causes loss of Hsp40 followed by dissociation of Hsp70. The Hsp70 binds another ATP and the cycle can be repeated. GrpE (or its equivalent) is found only in bacteria, mitochondria, and chloroplasts; in other locations, the dissociation reaction is coupled to ATP hydrolysis in a more complex way.
Protein folding is accomplished by multiple cycles of association and dissociation. As the protein chain lengthens, Hsp70 (DnaK) may dissociate from one binding site and then reassociate with another, thus releasing parts of the substrate protein to fold correctly in an ordered manner. Finally, the intact protein is released from the ribosome, folded into its mature conformation (for review see Georgopoulos and Welch, 1993; Hartl, 1966; Bukau and Horwich, 1998).
Different members of the Hsp70 class function on various types of target proteins. Cytosolic proteins (the eponymous Hsp70 and a related protein called Hsc70) act on nascent proteins on ribosomes. Variants in the ER (called BiP or Grp78 in higher eukaryotes, called Kar2 in S. cerevisiae), or in mitochondria or chloroplasts, function in a rather similar manner on proteins as they emerge into the interior of the organelle on passing through the membrane.
What feature does Hsp70 recognize in a target protein? It binds to a linear stretch of amino acids embedded in a hydrophobic context (Flynn et al., 1991; Blond-Elguindi et al., 1993). This is precisely the sort of motif that is buried in the hydrophobic core of a properly folded, mature protein. Its exposure therefore indicates that the protein is nascent or denatured. Motifs of this nature occur about every 40 amino acids. Binding to the motif prevents it from misaggregating with another one.
This mode of action explains how the Hsp70 protein Bip can fulfill two functions: to assist in oligomerization and/or folding of newly translocated proteins in the ER; and to remove misfolded proteins. Suppose that BiP recognizes certain peptide sequences that are inaccessible within the conformation of a mature, properly folded protein. These sequences are exposed and attract BiP when the protein enters the ER lumen in an essentially one-dimensional form. And if a protein is misfolded or denatured, they may become exposed on its surface instead of being properly buried.

Chaperones are needed by newly synthesized and by denatured proteins

  • Chaperones act on newly synthesized proteins, proteins that are passing through membranes, or proteins that have been denatured.
  • Hsp70 and some associated proteins form a major class of chaperones that act on many target proteins.
  • Group I and group II chaperonins are large oligomeric assemblies that act on target proteins they sequester in internal cavities.
  • Hsp90 is a specialized chaperone that acts on proteins of signal transduction pathways. 
The ability of chaperones to recognize incorrect protein conformations allows them to play two related roles concerned with protein structure:
  • When a protein is initially synthesized, that is to say, as it exits the ribosome to enter the cytosol, it appears in an unfolded form. Spontaneous folding then occurs as the emerging sequence interacts with regions of the protein that were synthesized previously. Chaperones influence the folding process by controlling the accessibility of the reactive surfaces. This process is involved in initial acquisition of the correct conformation.
  • When a protein is denatured, new regions are exposed and become able to interact. These interactions are similar to those that occur when a protein (transiently) misfolds as it is initially synthesized. They are recognized by chaperones as comprising incorrect folds. This process is involved in recognizing a protein that has been denatured, and either assisting renaturation or leading to its removal by degradation.
Chaperones may also be required to assist the formation of oligomeric structures and for the transport of proteins through membranes. A persistent theme in membrane passage is that control (or delay) of protein folding is an important feature. Figure 8.9 shows that it may be necessary to maintain a protein in an unfolded state before it enters the membrane because of the geometry of passage: the mature protein could simply be too large to fit into the available channel. Chaperones may prevent a protein from acquiring a conformation that would prevent passage through the membrane; in this capacity, their role is basically to maintain the protein in an unfolded, flexible state. Once the protein has passed through the membrane, it may require another chaperone to assist with folding to its mature conformation in much the same way that a cytosolic protein requires assistance from a chaperone as it emerges from the ribosome. The state of the protein as it emerges from a membrane is probably similar to that as it emerges from the ribosomebasically extended in a more or less linear condition.
Two major types of chaperones have been well characterized (for review see Frydman, 2001). They affect folding through two different types of mechanism:

  • Figure 8.10 shows that the Hsp70 system consists of individual proteins that bind to, and act on, the substrates whose folding is to be controlled. It recognizes proteins as they are synthesized or emerge from membranes (and also when they are denatured by stress). Basically it controls the interactions between exposed reactive regions of the protein, enabling it to fold into the correct conformation in situ. The components of the system are Hsp70, Hsp40, and GrpE. The name of the system reflects the original identification of Hsp70 as a protein induced by heat shock. The Hsp70 and Hsp40 proteins bind individually to the substrate proteins. They use hydrolysis of ATP to provide the energy for changing the structure of the substrate protein, and work in conjunction with an exchange factor that regenerates ATP from ADP.
  • Figure 8.11 shows that a chaperonin system consists of a large oligomeric assembly (represented as a cylinder). This assembly forms a structure into which unfolded proteins are inserted. The protected environment directs their folding. There are two types of chaperonin system. GroEL/GroES is found in all classes of organism. TRiC is found in eukaryotic cytosol. 

The components of the systems are summarized in Figure 8.12. The Hsp70 system and the chaperonin systems both act on many different substrate proteins. Another system, the Hsp90 protein, functions in conjunction with Hsp70, but is directed against specific classes of proteins that are involved in signal transduction, especially the steroid hormone receptors and signaling kinases (for review see Moarefi and Hartl, 2001). Its basic function is to maintain its targets in an appropriate conformation until they are stabilized by interacting with other components of the pathway (Rutherford and Lindquist, 1998; Queitsch, Sangster, and Lindquist, 2002).
(The reason many of these proteins are named "hsp", which stands for "heat shock protein" is that increase in temperature causes production of heat shock proteins whose function is to minimize the damage caused to proteins by heat denaturation. Many of the heat shock proteins are chaperones and were first discovered, and named, as part of the heat shock response.)