The ribosomal components bound by the Alu-domain are well conserved in all ribosomes and comprise the elongation factor binding site [124]. It is intriguing that all the contact sites used by the SRP Alu-domain are used by eEF2 [125,126] as well, allowing us to interpret the Alu domain ribosome interaction as elongation factor mimicry (fig. 29). A tRNA-like interaction, however, is taking place concomitantly: the RNA-RNA interaction between the loop L2 of the SRP 5’-RNP and helix 43 of the stalk base is reminiscent of the interaction of the tRNA T-loop with the same ribosomal helix in the context of the EF-Tu*tRNA*GTP ternary complex bound to the ribosome [127].

Binding of the Alu-domain in this position directly competes with elongation factors entering their binding site. Therefore, sufficiently high affinity of the Alu-domain for this site explains its elongation arrest activity. Variations of the Alu-domain affinity could explain different efficiencies observed in different systems [17,128]. In the cryo-EM structure the Alu-domain is bound to a ribosome in the post-translocational conformation (POST state), as defined by the presence of the peptidyl-tRNA in the P site and an unoccupied A-site. Thus, it is possible that the POST state of the ribosome is the preferred conformation for SRP binding during the elongation cycle [129].

Figure 29: SRP Alu-domain interaction with 80S ribosome.

a, Top: 80S RNC density without SRP, showing conserved contact sites of Alu-domain. B6, bridge 6. Bottom: Comparison of 80S RNC-SRP complex with 80S ribosome-eEF2 complex [126] in same orientation (SRP and eEF2 shown in red). Subunits of the Alu-domain and domains of eEF2 involved in similar contact sites are labelled. Colour code as in figure 12.

How can elongation arrest by the Alu-domain be controlled by the event of signal sequence binding, which occurs more than 250 Å away? Its binding has to induce or at least stabilize bending [130] of the particle at hinge 1 in order to promote Alu-domain binding (fig. 30). One possibility is that high-affinity binding of the S-domain simply tethers the Alu-domain in an appropriate position on the ribosome, and thereby favours the bound state conformation of SRP. However, another model can be suggested where recognition of a signal peptide produces specific positioning of the S-domain on the ribosome; this may lead to conformational changes of SRP68/72 that results in the hinge 1 stabilising a 90º angle and the Alu-domain closing into the factor binding site (fig. 30). In agreement with this idea is the brace-like localisation of SRP68/72 covering the hinge 1 region, its participation in connection 4 and the finding that SRP reconstituted without SRP68/72 lacks elongation arrest activity [18].

Figure 30: Signal sequence-dependent SRP-ribosome interaction.

Upon signal sequence binding by SRP54, a kinked conformation of SRP is stabilized involving possibly SRP68/72 and a rotation around hinge 1. As a result, SRP interacts with the ribosome, stretching from the peptide exit (S-domain) to the elongation factor binding site in the intersubunit space (Alu-domain), where it causes elongation arrest by competition with elongation factors. Colour code as in figure 12; signal, signal sequence (green); EXIT, peptide tunnel exit; EFS, elongation factor binding site.

In homologous systems only a slowdown of translation, and not the complete arrest, is observed [16,17,55]. In order to continue translation it is necessary that the Alu-domain relocates on the ribosome, so that the required elongation factors can access the ribosome and translation can resume. It is not clear where the Alu-domain is located in this situation. It could be that it interacts with the ribosome in different position by forming new contacts or that it is simply flexible and disordered. The necessary plasticity of the Alu-domain could be also a result of flexibility around the hinge 2.

Based on those structural and previous biochemical data a more detailed mechanistic model can be presented describing the conformational changes within the SRP core which facilitate first steps of the SRP cycle (fig. 31).

(i) The free SRP core in solution most likely adopts a compact conformation similar as observed in the recent X-ray structure, and might be stabilized by an interaction of SRP54G with the SRP RNA [53]. A direct contact between the distal loop of SRP54N and helix αM1b of SRP54M is present and the finger loop of SRP54M is covering the hydrophobic groove to protect it from aqueous solvent. In this free state, SRP is indicated to be incapable of signal sequence binding [131] and the GTP affinity of SRP54 is low.

(ii) In the sampling mode, SRP binds with low affinity to the ribosome in order to probe it for the presence of a signal sequence [3]. Although there are no structural data available for this mode, biochemical cross-links indicate that the signal sequence-independent interaction is similar to that observed in the cryo-EM structure: the different ribosomal cross-link patterns of the mammalian Alu domain [132] as well as of SRP54 in mammalian [13] or E. coli SRP [103] do not change extensively between sampling and targeting mode. Therefore, the hydrophobic contact between the distal loop of SRP54N and the M domain might be lost already due to the interaction of SRP54N with L23p/L29p [13,103,106], which might prime SRP for signal sequence binding on the ribosome. Notably, signal sequences of RNCs could be cross-linked to the same site of the ribosome, the L23p protein [106], making this site ideally suited for probing by SRP54M. L23p functioning as a transitory binding site for signal sequences would be in agreement with a stabile affinity of SRP signal sequences independently of chain length [133]. However, without a signal sequence available at the tunnel exit, the SRP core adopts a conformation that allows only transient binding and that can not retard elongation nor target the RNC to the membrane.

(iii) In the targeting mode, SRP binds with high affinity to the RNC adopting the conformation observed in the cryo-EM structure. The SRP core is positioned with the SRP54N and the C-terminal SRP54M domains on opposite sides of the tunnel exit and the hydrophobic groove with the bound signal sequence directly on top (fig. 12, 17, 20b). It remains unclear, whether the conformational changes within SRP54 open the hydrophobic groove, or if the signal sequence shapes its binding groove. Yet, a hydrophobic groove exposed to the aqueous solvent without substrate would be energetically unfavourable, and therefore, signal sequence binding is likely to follow an induced fit mechanism. The signal sequence locks the ribosome-bound SRP54 in the open conformation as seen in the EM structure and increases the affinities of SRP for both the ribosome [3,133] and GTP [71]. As a result, the SRP core is positioned on the ribosome in a way that allows the Alu domain to cause elongation arrest by binding in the orientation observed by cryo-EM.

(iv) Docking of the targeting complex to the SRP receptor is very likely to involve interaction of the NG domains and also direct interaction of SR with the ribosome [84,85,134]. The orientation of SRP54 on the ribosome is altered after SR binding as shown by cross-linking [13]. The RNC-SRP-SR complex formation will have to re-arrange at least the SRP54NG domain away from the exit tunnel which in turn would enable the transfer of the RNC to the translocon. Notably, in the X-ray structures of the interacting NG domains of bacterial SRP and SR [77,78] the N-terminal helix of both N domains is displaced and the C-terminal helix of the G-domain (adjacent to the ‘LGMGD’ linker) is shifted towards the NG interface. This might also be relevant for re-arranging the SRP54NG domain on the ribosome.

Figure 31: Model for the first steps of the SRP cycle.

A detailed scheme of the conformational events of the SRP core during the first steps of cotranslational targeting can be modelled as follows: (i) In the free state, the SRP core adopts a compact conformation. A distal loop of SRP54N contacts SRP54MN, which is later on involved in signal sequence binding. SRP54G interacts with helix 8 of the SRP RNA and the nucleotide affinity is low. (ii) In the sampling mode, SRP interacts with low affinity and transiently with translating ribosomes in order to scan for signal sequences. The contact involves at least the L23p/L29p region (L23p) of the large ribosomal subunit and the distal loops of SRP54N. Ribosome binding induces a change in the SRP core towards a more open conformation, which probably leads already to a loss of the contact between the distal loop of SRP54N and SRP54MN and renders SRP capable for signal sequence recognition. The reorientation of SRP54NG allows for GTP binding. (iii) In the targeting mode, SRP interacts with high affinity with the ribosome-nascent-chain complex (including the L23p/L29p region and ribosomal RNA helix 24, H24). Signal sequence binding occurs to the hydrophobic binding groove of SRP54M and the binding is transmitted to SRP54NG by inter-domain communication. As a result, the RNC-bound SRP core changes to the fully open conformation by adjusting the flexible linker region between SRP54NG and SRP54M. The SRP-RNC targeting complex induces the high affinity GTP binding to SRP54NG, which is now primed for the interaction with the SRP receptor. (iv) The successive docking of the SRP-RNC complex to the translocon via SR is structurally unresolved. However, the interaction of the SRP-RNC complex with SR may lead to the twin-like arrangement of the SRP/SR NG-domains resulting in relocalization of the SRP54 NG domain. Colour code as in fig. 20 and 22.

Although the resolution of the structure is not sufficient yet for complete interpretation and model building, a preliminary interpretation is already possible. When SRP-RNC complex interacts with SR in presence of GTP, the NG-domains of SRP54 and SRα form a twin-like heterodimeric complex. Apparently this leads to delocalization of both NG-domains which dissociates SRP from binding site C1. This binding site involves ribosomal proteins L23e and L35 which provide, in addition, the binding site for the translocon and, thus, clearing of the binding site C1 allows spatial access of the translocon.

In the crystal structure of the yeast SRX-SRβ complex [81], SRβ is in the GTP-bound state, the same state as in the complex used for this cryo-EM structure. Nevertheless, it is not possible to dock the crystal structure of SRβ into the density without causing steric clashes with other structures. This leads to the assumption that SRβ may dissociate from SRα. To dissociate SRβ has to switch from the GTP-bound state to the GDP-bound state in which the affinity for SRα is several orders of magnitude lower [81]. Since SRβ is the only component in this ternary complex which is not loaded with GMP-PNP, but with GTP, nucleotide hydrolysis and the subsequent dissociation is possible. As a second prerequisite for the GTP hydrolysis, SRβ needs a GAP to function as a GTPase switch as it belongs to the Ras-like Arf-family of GTPases [81]. According to earlier biochemical studies, it may be indeed the ribosome which provides the GAP activity for SRβ [84]. Therefore, it appears to be the most likely scenario that after SRβ interaction with the ribosome-SRP complex, GTP hydrolysis takes place, followed by the dissociation of SRβ from SRα. It is intriguing that despite the dissociation from SRα, SRβ stays bound to the ribosome. This is in agreement not only with our binding assays (fig. 23) but also with earlier cross-linking experiments showing that SRβ, in its GDP or nucleotide free state, interacts with a 21kD protein of the large ribosomal subunit [85]. In the presence of GTP or GMP-PNP the cross-links were not observed. Thus, in its GDP form, SRβ would provide a second anchor point of the targeting complex on the ER membrane. More recent biochemical studies show that SR interacts strongly with both, SRP and the ribosome, and that SR by this dual recognition rejects ribosomes that lack bound SRP [134]. Taken together, it appears that in eukaryotes two different contacts are required for functional targeting, the contact between SRP and SRα, and the contact between the ribosome and SRα and SRβ. Here, SRβ could serve in a proofreading mechanism or simply by ensuring that the SRP-RNC complex is targeted to the correct membrane in the cell.

Although the structural characterization of the first steps of the SRP cycle makes significant progress, many questions remain to be answered: Does the SRP54N domain in the ribosome-bound state indeed rotate with respect to the G domain and what is the exact conformation of the linker region? Is the signal sequence indeed bound to the hydrophobic groove of SRP54M as proposed, and what is the significance of the newly defined dynamic regions of the SRP core in different functional states? Mutational and biochemical analysis, more high resolution X-ray and cryo-EM structures should provide the missing details of what happens when SRP meets ribosome.

Following the SRP cycle further, the comparison of the SRP-RNC structure with the previously reported cryo-EM structure of the RNC in complex with the translocon [99] shows that both, SRP and the translocon, bind at the tunnel exit of the ribosome in a mutually exclusive way. Here, the question remains of how the transfer of the signal sequence and the entire RNC from SRP to the translocon is facilitated. The preliminary cryo-EM structure of SR-SRP-RNC complex provides more insights in the next step of protein targeting. Visualization of a Sec61-SR-SRP-RNC intermediate will be one of the next challenging tasks in order to build a more complete model of the SRP cycle.