Co-translational membrane targeting

SRP-dependent membrane targeting

E. coli possesses an SRP system typical of Gram-negative bacteria, consisting of Ffh, 4.5S RNA and FtsY. This system is responsible for the delivery of newly synthesized membrane proteins to the SecYEG translocation channel in the plasma membrane. Figure 1 shows the four main steps of the SRP targeting cycle, which starts  with binding of SRP (red sphere) to ribosome-nascent chain (RNCs; in yellow) complexes displaying a hydrophobic transmembrane domain (TMD). This is usually the first TMD, which is recognized by SRP, bound close to the ribosomal exit tunnel (5). In eukaryotes, binding of SRP to RNCs induces translational arrest, while in E. coli translation is not paused upon association of SRP with translating ribosomes (6). Following SRP binding, the SRP-RNC complex docks with membrane-bound FtsY (pink) located at the SecYEG channel (green). The presence of GTP in both Ffh and FtsY is required for the formation of a stable SRP-FtsY complex. Subsequently, the TMD is handed over from SRP to the translocation channel and hydrolysis of GTP results in dissociation of the SRP-FtsY complex (5). Thus, the GTPase activity of Ffh and FtsY controls the disassembly of the SRP-FtsY complex, thereby priming these factors for a new targeting cycle. 

Biochemical and structural properties of SRP and FtsY

Genetic, biochemical and crosslinking experiments demonstrated that membrane targeting of a few membrane proteins is SRP-dependent. Moreover, these studies revealed that deletion of the genes encoding Ffh, 4.5S RNA and FtsY are essential for E. coli viability (7). However, depletion of SRP only moderately affects E. coli cell growth and marginally inhibits membrane targeting. This raises the question why SRP is essential. Recent proteomic studies answered this question by assessing the global effects of SRP depletion and establishing the SRP interactome. These studies revealed that in E. coli cells devoid of SRP, cytoplasmic membrane proteins accumulate. Furthermore, sublevels of SRP induce a stress response, resulting in the upregulation of chaperones and proteases (8). Interestingly, deletion of the genes encoding Lon and Clp proteases induces synthetic lethality in E. coli cells depleted of SRP (9).  This suggests that in the absence of SRP, cytoplasmic proteases are required for the degradation of mistargeted membrane proteins. Collectively, these studies confirm that in vivo membrane proteins are the main substrates of SRP and its main physiological function is to deliver newly synthesized membrane proteins at the plasma membrane co-translationally, thereby preventing their toxic accumulation. SRP-dependent membrane targeting is therefore directly linked to maintaining protein homeostasis.

High resolution structures are available for most SRP components, including SRP, FtsY and a complex between these factors. Figure 2 shows the structure of E. coli SRP obtained at a resolution of 3.9 Å and contains Ffh and 4.5S RNA (10). Ffh is a multi-domain protein, comprising a C-terminal methionine-rich (M) domain, a helical N domain and a GTPase (G) domain. The M domain (in green) is responsible for high affinity RNA binding and recognizes hydrophobic signal sequences, while the N domain mediates interactions with the ribosome. The G domain binds and hydrolyzes GTP. The NG domain is in blue with a non-hydrolyzable GTP analog in red. The 4.5S RNA is in yellow. The G domains of SRP (Ffh) and SR (FtsY) are evolutionary conserved and comprise a subfamily of GTPases, which mediate co-translational membrane targeting.  

FtsY was identified as a potential targeting factor based on sequence comparisons with the alpha subunit of eukaryotic SR. Biochemical experiments subsequently confirmed that FtsY is the bacterial SRP receptor and associates with SRP in an in vitro reconstituted system. Moreover, depletion of E. coli FtsY impaired SRP-dependent membrane targeting, while in vitro FtsY was essential for the targeting of nascent E coli membrane proteins (11). FtsY is a multi-domain protein similar to Ffh and contains an N-terminal negatively charged acidic (A) domain and the highly conserved N and G domains. The A domain is not conserved and is of variable length. For example, in E. coli FtsY this domain is 195 residues, while the same domain in T. aquaticus FtsY is only a few residues (3). The function of this domain is not clear because removal of the A domain does not interfere with FtsY function. The N and G-domains form a structural and functional unit, termed the NG domain and displays GTPase activity. Further, FtsY associates with the plasma membrane despite the absence of membrane anchoring signal. It was suggested that the A domain mediates interactions with the translocon and  is required for membrane binding (3). However, truncated versions containing the NG domain are fully functional and are able to associate with the membrane. Interaction of FtsY with the membrane is necessary for release of RNCs from the SRP-FtsY complex. Moreover, the GTPase activity of FtsY is stimulated through interaction with lipids in vitro, while genetic experiments provided in vivo evidence for a functional interaction of FtsY with lipids. Detailed structural information is available for E. coli FtsY but this has been mostly obtained with truncated apo forms of the protein. However, the high resolution structure of full length FtsY of T. aquaticus was solved in the apo and holo form at 2.24 and 2.3 Å, respectively (12). Figure 3 shows the structures of apo (upper left) and holo FtsY (upper right). The N-terminal helix is green, the N- domain in red, the G-domain in blue, the insertion box domain in yellow and non-hydrolyzable GTP analog is in pink. The insertion box domain is conserved throughout members of the SRP / SR subfamily of GTPases and contains catalytic residues important for GTP hydrolysis (12). A structural alignment of both structures is displayed in the bottom panel of figure 3 and shows no major conformational changes upon GTP binding (apo FtsY in green and GTP bound FtsY in blue), although moderate rearrangements of catalytic residues of the G-domain are observed to accommodate the nucleotide.

SRP-FtsY complex

SRP is only able to form a stable complex with FtsY when GTP is bound to both factors, resulting in a heterodimeric complex. Interactions between the NG domains mediate assembly of the complex and hydrolysis of GTP results in disassociation of the complex. The GTPase activity of the SRP-FtsY complex controls, therefore, the targeting cycle and mutations that block this activity interfere with protein targeting and translocation. It has been suggested that the conformational changes prior to GTP hydrolysis are crucial for handing over the RNC to the Sec-translocon – unloading of the cargo. 4.5S RNA is essential for cell viability and is indispensable for in vitro protein targeting (7). This RNA moiety plays a key biochemical role in the assembly of the SRP-FtsY complex,  namely it accelerates the interaction between Ffh and FtsY, stimulates the GTPase activity of the SRP-FtsY complex and enables conformational changes in Ffh and FtsY after binding of a signal sequence by Ffh (3). Several structures of the bacterial SRP-FtsY complex have been elucidated, including the structure of the E. coli complex (figure 4). This structure was solved at a resolution of 3.94 Å and reveals a heterodimeric assembly (upper panel) loaded with GTP (10). For clarity, a simplified version containing one Ffh and FtsY protomer as well as one 4.5S RNA molecule is also displayed (bottom panel). Ffh is shown in blue, RNA in yellow, FtsY in magenta and GTP in red or green (in ball and stick or spheres, respectively). The structure shows that GTP-loaded Ffh and FtsY bind to the distal of the RNA moiety, which is known to stimulate GTPase activity. This indicates that 4.5S RNA binds initially Ffh and FtsY at one end of the molecule and both are relocated towards the distal end upon GTP binding. Hence, the RNA moiety is a bifunctional binding platform for both apo and holo Ffh/FtsY. 

Complex of a translating ribosome and SRP

The SRP targeting cycle starts with binding of SRP to translating ribosomes. It has been established that eukaryotic and bacterial SRP display high affinity towards active ribosomes when compared to empty ribosomes (13, 14). In E. coli, SRP recognizes ribosomes translating membrane proteins and binds to hydrophobic TMDs as they emerge from the ribosome as indicated by photo-cross link studies. Interestingly, these studies also showed that SRP did not bind to ribosomes programmed with a nascent outer membrane protein or a cytosolic protein (15). Complementary evidence suggesting that hydrophobicity is a key determinant for SRP-dependency comes from studies with SRP conditional E. coli mutants. These showed that increasing the hydrophobicity of otherwise SRP independent targeting signals present in an outer membrane or small phage coat protein dramatically increase their in vivo SRP-dependency (16, 17). A recent proteomic study mapping the global SRP interactome confirmed that E. coli SRP preferentially binds to TMDs enriched in stretches of hydrophobic and bulky aromatic residues. Remarkably, SRP also associates with internal TMDs and appears able of binding multiple TMDs in polytopic membrane proteins (18). Combined, these results establish that E. coli SRP specifically binds to hydrophobic targeting signals present in nascent membrane proteins. The M domain of Ffh is required for the binding of targeting signals probably through a hydrophobic groove with a high number of methionine residues at the binding site. This groove is able to accommodate alpha helical structures such as a TMD or signal peptide in eukaryotes (3, 5).  The N-domain is required for ribosome association and photo-crosslinking studies established that SRP binds to L23 near the ribosomal exit site (19, 20). SRP is therefore ideally positioned to interact with nascent membrane proteins upon emergence from the ribosome. Moreover, SRP could only be crosslinked to nascent membrane proteins following the exit of a TMD from the ribosomal tunnel (15). This suggests that SRP is specifically recruited to the ribosome after a TMD has emerged from the ribosome. Although valuable, the high resolution structures of individual SRP components do not provide much detailed information on the assembly of the RNC-SRP complex. More insight into this was obtained from different cryo-EM reconstructions of isolated E. coli RNCs reconstituted with purified SRP (21). This complex is shown in figure 5 and reveals that SRP (with Ffh in green and 4.5S RNA in orange) binds near the exit tunnel (indicated by a star) at L23 (in yellow) and in close proximity to L29 (blue) and L24 (red). L23, L29 and L24 surround the surface of the exit tunnel. For clarity, rRNA and other ribosomal proteins are in grey. 

Quaternary structure of a targeting complex

Following targeting to the plasma membrane, the RNC-SRP complex docks with membrane-associated FtsY after which the nascent chain is handed over to the translocon. Based on this it was proposed that  SRP, FtsY and the translocon can bind simultaneously to active ribosomes, forming a so-called quaternary targeting complex to transfer the nascent chain from SRP to the translocon. Interestingly, photo-crosslinking studies, employing ribosomes programmed with nascent chains of different membrane proteins, have consistently shown that RNCs are in close proximity to SRP as well as the Sec-translocon during the early stages of membrane insertion (15,22). These results are therefore consistent with the formation of a targeting complex comprising a translating ribosome, SRP and the Sec-translocon. In addition to functioning as docking site for SRP and folding catalysts, L23 also interacts with the Sec-translocon during protein translocation or membrane protein insertion (23). L23 serves therefore as universal ribosomal docking site and seems of key importance for the formation of a targeting complex. Although no contacts between the RNC and FtsY were detected in aforementioned photo-crosslinking experiments, a direct interaction between FtsY and SecY was demonstrated (24). This suggests that SecY functions as receptor for FtsY. Moreover, the results of a recent study obtained with photo-crosslinking and FRET analyses pointed towards a quaternary targeting complex, consisting of SRP, FtsY and the Sec-translocon bound to translating ribosomes (25). Hence, several lines of biochemical evidence support the (transient) association of a RNC with SRP, FtsY and the Sec-translocon to form a targeting complex. Despite the biochemical evidence, structural information of this complex is scarce. The structure of this targeting complex was recently investigated in more detail and therefore purified SRP, FtsY and the Sec-translocon were assembled on isolated ribosomes loaded with a nascent membrane protein. Subsequently, these reconstituted complexes were analyzed by cryo-EM (26). The structure of the reconstituted targeting complex is shown in figure 6 with Ffh in red, 4.5S RNA in yellow, FtsY in blue, the translocon in brown, L23 in dark blue, L29 in aquamarine and the hydrophobic signal anchor of the nascent membrane protein in green. For clarity, rRNA and ribosomal proteins are shown in grey. The left panel shows a side view from the complex, while the right panel shows a view towards the exit tunnel. These results confirm that a translating ribosome is able to accommodate SRP and FtsY together with the translocon. As expected, Ffh is located near the exit tunnel through associations with L23, while its M domain is in contact with the signal anchor emerging from the exit tunnel. The GTP-loaded NG domain of Ffh is situated at the distal end of the RNA moiety together with GTP-bound FtsY. Thus, the SRP-FtsY complex is in the active state and its structure agrees well with other structural data on this complex (figure 4). The Sec-translocon, containing SecYEG, is bound to the ribosome near FtsY and the signal anchor emerging from the exit site. However, no direct contact between FtsY and the translocon is observed in this structure, while a subpopulation of the reconstituted complexes contains the translocon bound to the ribosome at the exit tunnel through association with FtsY (26), confirming previous biochemical data (24). These latter assemblies most likely represent the structure of the quaternary targeting complex in the pre-transfer state, while the aforementioned structure is presumably an intermediate state of the targeting complex in which the translocon stabilizes the SRP-FtsY complex to stimulate their GTPase activity (26).