1 Introduction

1.1  Protein Targeting

↓1

Proteins are synthesized in the cytosol by ribosomes which are large macromolecular machines consisting of RNA and 50-80 proteins. Eukaryotic and prokaryotic ribosomes are very similar in design and function. Both are composed of small (30S in prokaryotes, 40S in eukaryotes) and large ribosomal subunit (50S in prokaryotes, 60S in eukaryotes) which join on the translated mRNA molecule. The mRNA codons are translated into amino acid sequence using the cognate tRNAs as adapters to add the correct amino acids to the growing polypeptide chain. The large subunit of the ribosome provides the peptidyl transferase activity while the small subunit binds mRNA and is the site of decoding. When the stop codon is recognized on the mRNA ribosomes release the protein, and subunits separate again.

Many proteins have not the cytoplasm but organelles or even the environment outside the cell as a destination and have to be transported there. Even though most secretory and membrane proteins are targeted to the endoplasmic reticulum (ER) membrane to be cotranslationally translocated, some are translated completely in the cytosol, and later, postranslationally, transported to their destination. This applies also to mitochondrial, chloroplast or nuclear proteins. The main advantage of cotranslational targeting is that the coupling of translation and translocation prevents misfolding of the proteins in the cytoplasm.

The essential signal for correct sorting of such proteins are hydrophobic N-terminal signal sequences typically comprising 15-20 amino acids: a short positively charged N-terminal region, a central hydrophobic core and a more polar C terminal part which has a cleavage site for signal peptidase. Signal sequences are very divergent and have two general features – hydrophobicity and α-helical conformation of the hydrophobic core. Disruption of one of these features leads to a non-functional signal sequences for cotranslational targeting [1]. While in eukaryotes signal sequences are usually located at the extreme N-terminus, in prokaryotes like E. coli SRP-dependent signal sequences are often a transmembrane helix within membrane proteins of the plasma membrane.

↓2

Nascent chains carrying signal sequences will be recognized by a signal sequence “binding factor” [2], later identified in a mammalian system as an 11S ribonucleoprotein particle and named signal recognition particle (SRP) [3]. SRP will recognize any signal sequence with a critical level of hydrophobicity. The question how SRP recognizes and binds almost any hydrophobic α-helix is currently unanswered. Binding of SRP will arrest elongation of the nascent peptide chain and target the complex to the membrane via GTP dependent interaction with the SRP receptor (SR). After the SRP-SR-nascent chain-ribosome complex interacts with the translocon, the signal sequence is released; the SRP-SR complex dissociates after GTP hydrolysis and translation resumes (fig.1).

Figure 1: Signal sequence recognition and cotranslational targeting by SRP.

(a) Schematic overview of cotranslational targeting of proteins destined for secretion or membrane insertion. SRP interacts with the signal sequence as soon as it emerges from the ribosomal polypeptide exit tunnel (step I). In eukaryotes peptide elongation pauses upon SRP / ribosome nascent chain (RNC) complex formation and the RNC complex is targeted to the ER membrane by the interaction with the SR (step II). GTP binding to SRP and SR has been shown to be a prerequisite for SRP/SR complex formation. The RNC is then transferred to the protein-conducting channel in the membrane (the translocon) (step III) and triggered by GTP hydrolysis in SRP and SR the SRP/SR complex dissociates (step IV). (b) Schematic overview of the mammalian SRP bound to the signal sequence carrying 80S ribosome (RNC) based on a cryo-EM structure. The SRP core as part of the S-domain is positioned near the tunnel exit of the large ribosomal subunit. The 40S and 60S ribosomal subunits are yellow and grey, respectively. The SRP RNA is shown in red and the SRP proteins are labelled as follows: SRP54NG (turquoise), SRP54M (dark blue), signal sequence (green), SRP19 and SRP68/72 (pink), SRP9/14 (turquoise/dark blue).

1.2 Signal recognition particle (SRP)

↓3

The signal recognition particle displays three main activities in the process of cotranslational targeting: (I) binding to signal sequences emerging form the translating ribosome, (II) pausing of peptide elongation, and (III) promotion of protein translocation through docking to the membrane-bound SRP receptor (FtsY in prokaryotes) and transfer of the ribosome nascent chain complex (RNC) to the protein-conducting channel [4].

These activities can be assigned to the two main domains of SRP separable by the micrococcal nuclease treatment [5]: the first domain, called S-domain, binds to signal sequences and promotes translocation [6]. In mammalian SRP, it includes about half of 7S RNA of SRP (~nucleotides 100-250) as well as the essential proteins SRP19, SRP54 (Ffh in prokaryotes), and the SRP68/72 heterodimer (fig. 2, 3a). While SRP19 is required for SRP assembly [7], SRP54 is the functionally most significant protein subunit of the S-domain: it recognizes the signal sequence [6] and interacts with the SRP receptor in a GTP-dependent manner [8]. SRP54 is composed of an N-terminal domain (N), a central GTPase domain (G) and a methionine-rich C-terminal domain (M) [9], which anchors SRP54 to SRP RNA [10]. In addition, together with a part of the RNA backbone [11], the M-domain carries out the principal function of the signal sequence recognition [12] near the peptide tunnel exit site of the large ribosomal subunit [13].

The second main domain of SRP, called Alu-domain, mediates the elongation arrest activity [14]. It is supposed to enable efficient targeting by providing a time window during which the nascent chain can be targeted to the translocation site [15,16,17]. The Alu-domain contains the 5’- and 3’-part of 7S RNA (including Alu-like sequences) as well as the SRP9/14 heterodimer, which is essential for its activity [18].

1.2.1  SRP RNA

↓4

The presence and necessity of RNA in SRP is not completely understood yet. It seems that 4.5S RNA in E.coli stabilises the structure of the Ffh M-domain[19,20]. In addition, kinetic studies show that RNA enhances association and dissociation of Ffh-FtsY complexes, and the positively charged N-terminal part of the signal sequence probably interacts with the negatively charged RNA backbone.

Figure 2: Secondary structure of the SRP RNA with RNA domains indicated

Structurally, the 300 nucleotide long human SRP RNA can be divided into four domains[21,22]. Domain I consists of 5’- and 3’-ends of the molecule (helices 2-4 in mammalian SRP) and it builds the SRP Alu-domain. It seems that helix 1 is absent in eukaryotes. Domain II (helix 5) is the linker between the Alu-domain and the S-domain which consists of two RNA domains, domain III (helix 6) and domain IV (helix 8). Eukaryotes contain an additional helix 7 at the interface of domains III and IV. Only domain IV is conserved in all SRP RNA molecules.

1.2.2 Evolutionary conservation

↓5

Although SRP is essential and present in all kingdoms of life maintaining its general function, structurally it shows high diversity. Minimal SRP, as found in gram negative eubacteria such as E. coli consists of a 106 nucleotides long RNA (domain IV) and SRP54 homolog (Ffh) only (fig. 3c). E. coli SRP also completely lacks an Alu-domain, and there is no evidence for translational arrest induced by SRP binding. In gram-positive eubacteria, like Bacillus subtilis, SRP has a more complex structure. The RNA is longer and it consists of domains I, II and IV and, in contrast to the gram-negative, has an Alu-domain. The presence of the SRP9/14 heterodimer has not been shown in gram-positive eubacteria, but a histone-like protein HBsu has been found to bind to Alu part of SRP RNA [23] (fig. 3c). HBsu is not a homolog of SRP9/14, but it has been suggested that it can fold into a tertiary structure similar to SRP9/14 [24], and might serve as their functional analog.

Figure 3: Evolutionary conservation of SRP.

a) Eukaryotic SRP consists of six proteins and RNA divided into two main domains, S- and Alu-domain.
b) Archaeal SRP with similar RNA structure, but with only two proteins, SRP54 and SRP19.
c) Eubacterial SRP in gram-negative bacteria like E. coli represents a minimal SRP consisting of a SRP54 homolog (Ffh) and RNA domain IV only. Some gram-positive bacteria like B. subtilis have an Alu-domain with histone-like HBsu proteins instead of SRP9/14 [25]

Archaeal SRP is more similar to eukaryotic SRP (fig. 3b). Archaeal RNA secondary structure shows high similarity to the mammalian; it contains four domains divided in eight helical elements folding into S- and Alu-domain. It has only two proteins, SRP54 and SRP19, which is, like in eukaryotes, responsible for the assembly [26]. Moreover, archaeal SRP54 can functionally interact with mammalian signal sequences [27], showing its high evolutionary conservation.

↓6

Even certain eukaryotic SRPs display a high level of structural variety. For example SRP of S. cerevisiae has a 600 nucleotide long RNA, twice as long as the mammalian. Instead of the SRP9/14 heterodimer, S. cerevisiae SRP contains a SRP14 homodimer, and, in addition, the unique SRP protein SRP21. In contrast to the secondary structure of S-domain RNA, which is comparable to mammalian, the Alu-domain is far larger and more complex [28].

Nevertheless, high functional and structural conservation of the minimal SRP through all kingdoms of life has been experimentally revealed by replacement of subunits of mammalian SRP (SRP54) with bacterial homologs (Ffh) which leads to partially active chimeric SRP [29].

Recently, SRP has also been found in chloroplasts where it differs from the cytosolic in that it contains unique a 43kD [30] subunit and lacks an RNA. It is required for the posttranslational targeting to the thylakoid of chlorophyll proteins encoded in the nucleus, and the cotranslational targeting of proteins synthesized by the chloroplast ribosomes. Chloroplast SRP appears to be present in two forms [31]. Polytopic protein D1 is synthesized by the chloroplast membrane bound ribosomes and is cotranslationally integrated in the thylakoid membrane where it interacts with cpSecY, the translocation channel in thylakoid membrane [32,33]. The targeting to the thylakoid cpSecY involves cpSRP54 (chloroplast homolog of SRP54), and it is independent from cpSRP43 [34]. Chloroplast ribosomes have the ability to pause the translation involving a light dependent regulation mechanism different from the elongation arrest. CpSRP composed of cSRP43 dimer and cpSRP54 is responsible for the posttranslational targeting of nuclear encoded photosystem proteins (light harvesting complexes) [35], after they have been imported into the stroma. It interacts with the substrate and forms a soluble intermediate transit complex. Both cpSRP43 and cpSRP54 are involved in the substrate binding [36], and are necessary for the posttranslational targeting [30]. Insertion into the membrane requires GTP and cpFtsY (SRP receptor homolog). Both SRPs have no overlapping functions. No SRP or SRP receptor was found in mitochondria.

1.2.3 SRP assembly

↓7

SRP is partially assembled in the nucleus and partially in the cytoplasm a in agreement with that, nuclear localization for SRP proteins SRP9/14, SRP68, SRP72 and SRP19 has been determined [37]. After the transport into the nucleus the subunits bind SRP RNA and form a pre-SRP which is exported to the cytoplasm and binds SRP54 [7,37,38].

SRP assembly starts during 7S RNA transcription by RNA polymerase III in the nucleolus, by binding of the SRP 9/14 heterodimer and formation of Alu-domain. Prior to transportation to the nucleus SRP9 and SRP14 form the heterodimer in the cytoplasm which is a prerequisite for the binding to 7S RNA [39].

The structure of the Alu-domain, which includes 56 nucleotides of the 5’-region of 7S RNA and the SRP9/14 heterodimer does not resemble a tRNA-like structure as previously suggested. Both proteins, SRP9 and SRP14 are structurally related and have an αβββα fold. The heterodimer binds primarily to the highly conserved core of the 5’-region of 7S RNA which consists of a three way junction in which two helical hairpins are connected to a third helical stem by a conserved U-turn [40]. The 5’-region of 7S RNA contains highly conserved nucleotides and it is capable to bind SRP9/14 even in absence of 3’-region. The majority of protein-RNA interactions are made by the concave β-sheet of SRP14. The final assembly step requires the 3’-region of 7S RNA to which the 5’-region folds back. As a result, the stem leading to the S-domain emerges on the C-terminal side of SRP14 and the SRP9 β-sheet interacts with the 3’-region [40] (fig. 4). This backfolding of 7S RNA is probably the final but a reversible step of Alu-domain assembly. Intact Alu-domain is required for final 3’ processing of SRP RNA in mammalian cells, and its export to the cytoplasm [41]. A compactly folded 5’-region of 7S RNA and SRP9/14 is shown to be necessary for the efficient transcription of SRP RNA [42], probably by direct interaction with the transcriptional machinery. Notably, the central question of how the Alu-domain exerts its activity and mediates the arrest of peptide elongation remains unanswered so far. Since its structure apparently doesn’t resemble a tRNA-like fold it appears unlikely that the Alu-domain simply occupies the tRNA A-site of the ribosome as suggested before.

↓8

Figure 4: Mammalian Alu domain assembly.

In a first step the heterodimer consisting of SRP9 (red) and SRP14 (green, with the important C-terminal tail in cyan) binds to 5’-region of the Alu RNA. In a second step the 3’-region of the Alu RNA flips by up to 180 and it is clamped against 5’ Alu RNA. The RNA backbone is shown in yellow. Both 5’- and 3’-region of Alu RNA involved in protein binding are structurally conserved (orange) [25].

Sequences related to the Al-part of 7S RNA are dispersed through the mammalian genome. They originate from a reverse flow of the genetic information from 7S RNA to genomic DNA and amplification of 5’- and 3’-end of 7S RNA [43]. The Alu family of repetitive sequences is specific for rodent and primate genomes. The human genome has around 1 million copies of 300bp long Alu sequences which comprise over 10% of the genome. The function of these elements is unknown yet, but they are one of the major differences between human and other primate genomes.

One of these Alu elements is BC200, which is specifically expressed in human neural cells from a single gene by RNA polymerase III. BC200 localises in neurons specifically in the somatic and/or dendritic domains and it is still actively retrotransposing. The tertiary structure of BC200 is related to the SRP Alu-domain, however it permits the binding of SRP9/14 heterodimer [44].

↓9

A central role in the SRP assembly can be assigned to SRP19. From in vitro reconstitution experiments it is known that binding of the mammalian SRP54 to 7S RNA requires SRP19 binding first [7]. However, archaeal SRP54, unlike eukaryotic, has significant affinity for 7S RNA even in an absence of SRP19. SRP19 is a single domain αβ-type protein with a three-stranded antiparallel β-sheet packed on one side against two helices. With its flexible loops, it recognizes a particular shape of the rigid stem-loop RNA. SRP19 homologues have been found only in organisms which have helix 6. Helix 6 is closed with an unusual GNAR (N is any nucleotide; R is G or A) type loop, where A is strictly conserved and necessary for SRP19 binding. SRP19 interacts with a major groove of this tetraloop and with a minor groove of helix 8 tetraloop. In addition it interacts with parts of the highly conserved symmetrical loop of helix 8 that are not interacting with SRP54 M-domain [45]. In that way it brings together both helices in a side-by-side position (fig. 5, 6).

Figure 5: Structure of SRP S-domain assembly steps.

On left side the structure of SRP19 in complex with helices 5-8 is shown. SRP19 binds to both helices and brings them together as a prerequisite for SRP54 binding. On the right side the structure of a ternary complex consisting of SRP19, helices 5-8 and SRP54M domain is shown. Note the change in the asymmetric loop arrangement upon SRP54 binding [46].

Figure 6: Model of SRP S-domain assembly.

In a first step SRP19 brings helices 6 and 8 together. In a second step SRP54 binds to the symmetric and asymmetric loops and changes the asymmetric loop to a rigid conformation. In a last step the SRP68/72 heterodimer binds to the three -way junction [46].

↓10

The symmetric loop consists of four non-Watson-Crick base pairs which are conserved from eubacterial to human SRP RNA. Disruption of one of two base pairs, a sheared G-G pair or a reverse Hoogsteen A-C pair, eliminates protein binding and is lethal for the cell. SRP54 binds to these two loops and induces severe structural changes of the asymmetric loop. Two bases of it flip out and form an A-minor motif with helix 6 bases, which is only possible after SRP19 binding. In contrast to the well ordered symmetric loop the asymmetric loop is very flexible in free 7S RNA. SRP19 binding stabilizes the asymmetric loop of helix 8 via A-minor base triples with the helix 6 which is a prerequisite for SRP54 binding. The asymmetric loop has an evolutionary conserved 5’-side adenosine base which interacts with three universally conserved amino acids of the C-terminal part of SRP54 M-domain. This part of SRP54 M-domain forms a conserved arginine-rich HTH-motif which is involved in the RNA binding (αM2-αM5). SRP54 interacts with the minor groove of RNA with two helices αM3b and αM4 and in that way it brings asymmetric and symmetric loop together to form a stable protein-RNA interface (fig. 8). In contrast to RNA the SRP54 M-domain does not undergo large conformational changes upon binding [45,46,47].

In eubacteria without helix 6, the asymmetric loop is stabilized by magnesium ions. Currently no other function for SRP19 has been determined. It seems that helix 6/SRP19 is an evolutionary adaptation of SRP to enhance and control the kinetics of the assembly that cannot be done by metal ions.

SRP68 and SRP72 form a heterodimer in the nucleus only in a presence of 7S RNA, and as a dimer they bind to the three-way junction of the S-domain RNA. SRP68 binds first to RNA with its N-terminal region which is mainly positively charged. This is a prerequisite for the interaction of the C-terminal region of SRP72 with the C-terminal region of SRP68 in an hydrophobic manner [48]. SRP68/72 can be released from 7S RNA by high-salt treatment without dissociating into monomers.

↓11

It is not completely clear what the function of SRP68/72 is. Biochemical experiments [49] indicate that SRP68/72 interacts with the SRP receptor, and that it possibly participates also in the elongation arrest. SRP with an alkylated SRP68/72 heterodimer fails to target to the ER. SRP missing SRP68/72 or SRP reconstituted with SRP68/72 alkylated as a free protein, fails not only in targeting, but also in the elongation arrest. It has been proposed that SRP68/72 has also a role as guanine nucleotide dissociation factor [50].

1.2.4 SRP54 and signal sequence recognition

Signal peptide is recognized by SRP54 which together with the RNA helix 8 builds the minimal SRP. These two components are present in all SRPs. SRP54 is a multidomain protein consisting of three domains; an N-terminal four-helix bundle (N-domain), a GTPase domain (G-domain) and a C-terminal methionine-rich domain (M-domain). The N- and G-domain are usually treated as one domain (SRP54NG) responsible for the GTP regulation of protein targeting and the interaction with SR. The M-domain can be divided into two parts, the evolutionary conserved and rigid C-terminal part which binds RNA helix 8, and the flexible N-terminal part which is responsible for the signal sequence recognition and binding as demonstrated by chemical cross-linking [12]. The M-domain contains typically a high percentage of methionine residues [10,12]. About 16% of all M-domain residues in E. coli are methionine, a frequency about 6 time’s higher than that of the average methionine occurrence in proteins. Methionine has a highly flexible hydrophobic side chain because it is unbranched and displays unique conformational properties of the thioether linkage [51]. This features led to the hypothesis that methionines and other hydrophobic residues are arranged so that their flexible side chains form the hydrophobic binding site for the signal sequence with sufficient plasticity to recognize the wide variety of signal sequences [9]. In T. aquaticus, methionine residues are often replaced by branched hydrophobic residues like Phe, Leu or Val, because at higher temperatures (optimal for T. aquaticus) increased thermal motion eliminates the need for more flexible methionine [4,26].

Compared to prokaryotic Ffh, the eukaryotic SRP54 M-domain contains additional 100 residues which play a role in the signal sequence binding. Deletion of this C-terminal region leads to the abolishment of cross-linking to signal sequences [52]. It is possible that the C-terminus increases the hydrophobic surface area. However, the necessity for that is not clear since eukaryotic signal sequences are often shorter then transmembrane regions which act as signal sequences in prokaryotic inner membrane proteins.

↓12

In the last few years several crystal structures of SRP54 M-domains have been solved [1,11,46,53]. While the C-terminal part of M-domain (Mc) is very similar and well ordered in all structures, the N-terminal part (Mn) shows differences, especially in the finger loop, which is closing the hydrophobic groove and significantly varies in diverse structures. Helices αM1, αM1b, αM2 and αM5 build the hydrophobic pocket closed by this flexible loop. In the T. aquaticus structure [1] the finger loop is well ordered as it binds into its hydrophobic pocket the helix of a neighbouring M-domain in the crystal resembling in that way the signal sequence binding. In the E. coli structure [11] the finger loop is disordered and in the S. solfactoricus [53] structure of the complete SRP core, the finger loop is closed (fig. 7, 8). These differences show the flexibility of the finger loop which might be the basis for binding a variety of signal sequences. The structures of the M-domain of T aquaticus and S solfatoricus may represent two functional states of M-domain; the open state occupied with the signal peptide, and the closed empty state. If this is true, the signal sequence binding induces severe structural changes in the N-terminal part of the M-domain (Mn) whereas the RNA binding C-terminal part seems to stay rigid (Mc). The N-terminal region of signal sequences contains positively charged residues which may contact RNA as suggested [11] (fig. 8).

Figure 7: A) Crystal structure of SRP54 in complex with helix 8 in a ribbons representation of the N- (green), G- (blue), Mn- (purple) and Mc-domains (red) and helix 8 (orange). The novel N-terminal part of the M-domain contains the linker helix (αML) and the closed finger loop is highlighted in purple. B) M-domain in a top view compared with A. The finger loop on top is folded into the hydrophobic groove, which is lined by helices αM1, αM1b, αM2 and αM5 [53].

Figure 8: A) Crystal structure of the SRP54 (Ffh) M-domain from T. aquaticus.

A) The loop connecting helices αM1 and αM2 is open in this crystal structure [1]. B) Secondary structure of the domain IV sequence, and the solution structure of apo form of the domain IV from E. coli SRP RNA[54]. Nucleotides within the symmetric loop are highlighted in green, and nucleotides within the asymmetric loop in yellow. C) The SRP54 M-domain in complex with domain IV from E. coli [11] with a disordered finger loop. Note the conformational changes in the asymmetric loop (yellow) upon SRP54 binding. D) Model of T. aquaticus M-domain together with the structure of E. coli M-domain with the signal sequence (green) modelled in the hydrophobic groove of the M-domain. The signal sequence could simultaneously contact the backbone of domainm IV RNA with its positively charged residues [26].

1.2.5 Elongation arrest

↓13

The advantage of cotranslational targeting is that coupling of translation and translocation prevents misfolding of newly synthesized protein in cytoplasm. But protein translation can be faster then the diffusion of SRP-RNC complex to the membrane. To prevent that, SRP retards the translation and in that way it enlarges the time window during which a nascent chain can be targeted before it reaches a critical length prone to fold or aggregate. Elongation arrest was discovered in an heterologous system containing the canine SRP and the wheat germ ribosome. Here the SRP completely stops the translation of the nascent peptides [15]. Later, it was also observed in a homologous and more physiological systems from yeast [17,55] and also in the mammalian systems [16], and it is characteristic for all eukaryotic SRPs. In homologous systems the translational arrest is not very pronounced rather representing translation retardation, and it appears not to be essential for proper in vitro targeting [18]. It has been shown in yeast that defective translation arrest in vivo only slightly affects the translocation [17].

The Alu-domain of SRP, consisting of the 5’- and 3’-regions of 7S RNA and the SRP9/14 dimer, is responsible for the elongation arrest. SRP assembled without SRP9/14 is functional in protein targeting, but it lacks the elongation arrest feature. Even removal of the 20 C-terminal amino acids from SRP14 makes SRP non-functional in the elongation arrest. It has been suggested that the Alu-domain binds near the A-site on the ribosome, but the elongation arrest is still poorly understood. Since prokaryotic SRP lacks the Alu-domain it is one possibility that the elongation arrest is dispensable in prokaryotes. The elongation arrest main function is to enlarge the time window for targeting and prevent that the nascent chain reaches a length which can misfold or aggregate. Prokaryotic cells are in most cases significantly smaller then eukaryotic cells, and thus the time for diffusion to the membrane is smaller. Also, in prokaryotes bacterial DNA is anchored to the cytoplasmic membrane during coupled transcription/translation of membrane proteins which further reduces diffusion distance and, thus, this may allow cotranslational targeting without the need for the elongation arrest [4]. It has been shown that purified E. coli SRP is unable to arrest the translation in in vitro system although it properly binds to a signal sequence [56].

1.2.6 GTPase cycle and SRP receptor

The elongation arrest ability is abolished upon addition of microsomal membranes which led to discovery of the membrane bound SRP receptor (SR) [57]. SR is a heterodimeric complex formed by two subunits, the integral membrane protein SRβ and SRα. The assembly process includes cotranslational but SRP-independent targeting of SRα to the membrane. Within the SRα mRNA a stem loop structure similar to ribosomal frameshift structures causes pausing of the translation and allows folding of the N-terminal domain and interaction with SRβ before translation resumes [58]. In eukaryotes, SRα consists of three domains, the N-terminal X-domain which interacts with SRβ, the N-domain which builds a four helix bundle and the G-domain which binds GTP. The NG domain of the receptor is structurally and functionally homologous to the SRP54 NG domain. The bacterial homolog of eukaryotic SRα, FtsY, is a hydrophilic protein partially localized in the cytoplasm and partially at the membrane. However, a membrane anchoring protein homologous to SRβ has not been identified. Both, SRα and SRβ are GTPases.

↓14

GTPases are members of a protein family of highly conserved molecular switches responsible for the regulation of many complex functions such as cell cycling, protein synthesis and membrane trafficking. The general mechanism of GTPases (G-protein) is described in the molecular switch model [59,60] where the enzyme goes through three conformational steps: GTP-bound, GDP-bound and empty. The G-protein is initially in an empty and inactive state and it gets activated through a conformational change by GTP-binding. Such an active G-protein interacts with a target molecule (GTPase activating protein or GAP) which induces hydrolysis of GTP and inactivates the G protein. The remaining GDP is then released and the G-protein returns into an empty state which is only a transient intermediate during exchange of GDP to GTP. This exchange is regulated by the guanine nucleotide exchange factor (GEF) which switches the G-protein back to the active state.

Protein targeting involves tree different GTPases (SRP54, SRα, SRβ) in eukaryotes and two in prokaryotes (FFH, FtsY). The GTPase cycles of SRP54 and SR do not follow the general model of the GTPase cycle but have several unique properties which led to the concerted switch model for SRP GTPases [60]. SRP54, SRα and their prokaryotic homologues constitute a new subfamily of small Ras-like GTPases [59] with relatively low affinity for nucleotides, and, in contrast to canonical GTPases, they are stable in the absence of the nucleotide. Biological relevance of this apo-form is not clear yet, but its stability is reflected by the fact that both GTPases have been crystallized in the empty state [61,62]. Structurally, they are more similar to ATP-binding proteins than to other GTPases. Biochemical evidence shows that SRP54 and SRα do not depend on external GEFs in order to dissociate GDP [63,64], but they have a built-in nucleotide exchange ability. It has been proposed that this activity is located in the unique insertion box domain (IBD) in the effector region of the GTPase [65,66]. The IBD is a unique structural motif characteristic for the subfamily of SRP GTPases.

Mutation of a conserved glycine in the interface region between N- and G-domain of Ffh and FtsY severely weakens their ability to interact with each other. The same mutations in a conserved N-domain motif (ALLEADV) produced significant defects in signal sequence binding that correlate with the severity of the mutation [67]. It has been suggested that this interface motif has a function in the communication between N-, G-, and M-domain and that it communicates signal-sequence binding by the M-domain to the NG-domain, thereby priming SRP for the subsequent interaction with SR.

↓15

The SRP-SR interaction takes place primarily via their NG-domains [68], but it is further modulated by the SRP RNA which catalyses complex formation [69]. Mutations of the 4.5S RNA which do not affect Ffh binding nor the SRP interaction with the ribosome affect the interaction between SRP and FtsY [70], which is in agreement with the proposed model,

GTP binding to SRP54 and SRα is a prerequisite for their complex formation, and GTP hydrolysis leads to complex dissociation. According to nucleotide cross-link data, GTP affinity of SRP54 is increased upon interaction with a ribosome carrying a signal sequence [71] and SRP is then in the activated GTP-bound form ready to interact with SRα. SRα is primed for complex formation by the interaction with translocon components [71,72,73,74], since GTP binding of SRα is stimulated by addition of purified Sec61 which probably serves as GEF for it. When both, SRP54 and SRα are in GTP from, the complex can be formed.

The isolated NG domains are necessary and sufficient to form the complex in the presence of non hydrolysable nucleotides, although with slow kinetics [75], and it has been shown that they act as GAPs for each other [76]. The recent crystal structures of the interacting NG domains of SRP54/Ffh and FtsY in the GTP state showed that both N-domains rearrange towards the G-domains and that the complex forms an active site at the interface of the two proteins [77,78] (fig. 9 The individual sites are so closely intertwined that the two nucleotides are hydrogen-bonded to each other via γ-phosphates and ribose moieties. In comparison to GTP-bound free NG domains, NG domains undergo severe conformational changes in highly conserved motifs upon complex formation. Catalytic residues in the IBD loop rearrange and align with respect to the bound substrate. Water molecules in the active site are in an ideal position for GTP hydrolysis, although a non-hydrolysing transition state is stabilized in the observed conformation. It is crucial in the physiological RNC-SRP-SR complex that GTP hydrolysis by SRP and SR is blocked until the signal peptide is released. This prevents complex dissociation prior to delivery of the nascent chain to the translocon. It is not clear how the release of the signal sequence in M domain is communicated to the NG domains to allow GTP hydrolysis. One possibility is that the presence of the signal sequence stabilizes a conformation of the NG twin that has no GTPase activity. To that end, the existence of several activation states has recently indeed been shown [79]. An exhausting mutational analysis of the NG twin structure interface resulted in a model describing discrete structural changes during NG interaction and reciprocal GTPase activation. According to that model the NG twin goes through several conformational states without GTPase activity which could serve as control points before reaching the activated state and complex dissociation.

↓16

Figure 9: Structure of the heterodimeric FFH/FtsY NG domain complex.

A) Ribbon representation viewed perpendicular to the dimer axis, which is vertical in the figure. The N domain (blue) and the C-terminal helices (golden) are at the top, and their IBD domains are at the bottom (purple). The two active sites are brought into direct apposition to form an active site chamber at the centre of G domain (grey) where the GMPPCP ligands are buried. The motif I P-loops of the two proteins pack adjacent to each other (*). The structure is highly symmetric with the exception of the smaller N domain of FtsY, and all secondary structure elements adopt the same orientation in both proteins. B) The structure viewed along the two-fold axis further highlights the symmetry of the complex. The viewpoint is toward IBD [77].

In eukaryotes the complexity of the targeting GTPase cycle is increased by one more GTPase, the Arf-like Srβ, the function of which is not entirely clear. The Arf subfamily of GTPases is absent in prokaryotes [80] and it has a higher affinity for nucleotides compared to the SRP family of GTPase. A recent crystal structure of the SRα-SRβ complex [81] revealed the spatial arrangement of the N-terminal SRαX1 domain and SRβ from Saccharomyces cervisiae (Fig. 4b). While the complex was purified in the presence of GDP, the structure clearly shows GTP in the active site suggesting catalytic inactivity of SRβ in complex with SRα. The interface between these two domains includes the entire switch 1 region of SRβ which is critical for GDP-GTP conformational switching. One important conclusion in agreement with previous findings [82] is that the interaction between the X domain of SRα and SRβ is nucleotide dependent and that it requires the GTP state of SRβ. Furthermore, the crystal structure confirms that in contrast to SRP54 and SRα, SRβ requires both, a GAP and a GEF, to function as a GTPase switch. Recent data show that a subunit of the translocon, Sec61β, can function as a GEF for SRβ [83], which points at a role for SRβ in sensing the availability of a translocon. Interestingly, the ribosome has been suggested to function as a GAP for SRβ [84] implying that the SRα-SRβ complex would dissociate upon interaction with the RNC-SRP complex and subsequent GTP hydrolysis by SRβ. The dissociated SRβ in the GDP-state most likely stays bound to the ribosome since close proximity between SRβ in the GDP-state and a ribosomal protein (21 kD) has been shown by chemical cross linking [85].

Figure 10: Structure of SRX-SRβ(GTP) complex.

A) Structure of SRX-SRβ(GTP) complex from yeast. The SRβ subunit is shown in cyan and the SRX1 domain of SRα subunit in magenta. The GTP nucleotide is drawn in ball-and-stick representation. The switch 1 region (residues 64-72) of SRβ (yellow) forms the main interaction site with the SRα. Secondary structure elements are labelled. Unstructured loop regions are coloured gray. B) Same as A but rotated around horizontal axis counter-clockwise by 90o [81].

↓17

In the current model of the SRP cycle, SRP54 can interact in an empty state with the ribosome carrying the signal sequence. Assembly of the SRP-RNC complex slows down the elongation of the nascent chain and it induces stable GTP binding to SRP54. This results in the primed state of SRP54 with a conformation that does not yet allow GTP to access the catalytic centre, but is ready to interact productively with the SR. On the membrane side, the contact of SRβ with the translocon induces GTP binding by SRβ, which results in formation of the SRα-SRβ complex. The SRP-RNC complex is then targeted to the ER membrane where it interacts with SR. SRP54 and SRα NG-domains interact in a GTP-dependent manner which brings GTP into the catalytic centre. However, simultaneous GTP hydrolysis is blocked until the signal sequence is released. With all three GTPases in the GTP-bound state, the ternary complex is stably assembled. The synchronized GTP hydrolysis follows the release of the signal sequence to the translocation channel and results in the dissociation of SRP and SR from the ribosome and translocon while peptide elongation resumes. The transfer of the nascent chain apparently precedes the GTP/GDP switch of all GTPases, since it also happens in the presence of non-hydrolysable GTP analogs.

In bacteria which lack SRβ, the SRα homolog FtsY exists in a soluble and a membrane-bound form. The soluble form of FtsY is not sufficient to dissociate SRP from RNCs but requires the context of the membrane. The membrane receptor for FtsY has not been identified yet, and it is possible, since there is only one target membrane in bacteria, that the FtsY ability to bind the membrane is sufficient for proper targeting to the bacterial SecYE translocon. Binding of FtsY to the membrane and the translocon induces GTP binding to FtsY and primes it for interaction with SRP.

1.3 Goals

Despite a large amount of functional data and a growing number of SRP sub-structures, several fundamental questions remain unresolved: How does SRP interact with the ribosome. How is translation arrest induced? How does SRP recognise and bind a signal sequence on the ribosome? How is signal-sequence binding coupled to GTP binding, a prerequisite for docking of SRP to its receptor (SR)? How does the SRP-RNC complex dock to the SR in GTP-dependent manner. And after docking to the membrane-bound receptor, how is the release of the signal sequence and transfer of the RNC to the translocon coordinated?


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
13.12.2006