Experimental design. This work focuses on ribosome-bound nascent chains (RNCs) derived from Escherichia coli flavohemoglobin (Hmp, Fig. 1a) and from the phosphorylated insulin receptor interacting region (PIR) of the growth factor receptor-bound protein 14 from Rattus norvegicus (Fig. 1b). The Hmp protein comprises three domains, an N-terminal heme-binding (domain 1), a flavin adenine dinucleotide-binding (domain 2) and a C-terminal nicotinamide adenine dinucleotide-binding domain (domain 3), as shown in Fig. 1c 68. Several RNC chain lengths were examined, and all pertinent constructs are shown as solid bars in Fig. 1c. Hmp plays a key role in O2, NO and CO transport in E. coli, and is involved in a variety of signaling pathways 69,70. The cofactor-free form of Hmp is denoted as apoHmpH. Importantly, previous studies established that the N-terminal globin domain of Hmp is stable and folded even in its apo form 71. Our second target protein, PIR, is intrinsically disordered 72, i.e., an IDP (Fig. 1b). The specific nascent-chain constructs of both proteins analyzed in this work are schematically illustrated in Fig. 1c,d.
Ribosome-nascent-protein interactions were probed with via the well-characterized zero-length chemical crosslinker carbodiimide 1-ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) 51,73,74. Nascent chains were chemically crosslinked via known procedures 51 involving a combination of low-pH SDS-PAGE 75 and Western blotting in the absence and presence of the trigger factor (TF) chaperone. Notably, EDC enables the detection of existing noncovalent interactions and it does not provide an accurate quantitation of interacting populations, as discussed at length by Guzman-Luna et al.51. Yet, in the presence of appropriate controls, EDC is an extremely valuable tool to detect the existence of protein-protein interactions within the ribosome-nascent-chain complex. In addition, relative changes in the extent of the interactions, for any given ribosome-nascent-chain complex (RNC) as a function of environmental changes (e.g., variable urea or chaperone concentrations) were also qualitatively assessed. In general, due to the established presence of crosslinking-incompetent populations in RNC/r-protein complexes detected via EDC 51, interacting populations tend to be underestimated. Site-specific fluorescence labeling of nascent proteins at their N terminus enables focusing exclusively on interactions involving the nascent protein. Low-pH-gel and Western-blot were collected to explore interactions between nascent chains and r-proteins. It is worth noting that EDC does not have high accessibility within the exit-tunnel core 51. Therefore, detection of interactions within the tunnel core is not expected, within our experimental setup.
It is also important to mention that, under our experimental conditions, EDC does not report on interactions involving nascent protein chains and ribosomal RNA (rRNA). In the presence of imidazole, crosslinks between RNA 5' phosphate and aliphatic amines of proteins are known to take place 73. However, our samples did not contain imidazole, and this chemical would anyways be unable to detect interactions not involving the 5’ end of RNA. Therefore, even in the presence of imidazole, EDC would likely underestimate all potential interactions with RNA. Thus, interactions between nascent proteins and rRNA are beyond the scope of this study.
The compaction, tumbling rates, size (expressed in terms of approximate number of residues)and local-motion amplitude of non-interacting RNC regions were assessed via fluorescence depolarization in the frequency domain. In this way, it was possible to gain complementary and more comprehensive insights into RNC conformational characteristics. The apparent thermodynamic stability of r-proteins was collectively assessed by Trp fluorescence emission spectroscopy as a function of urea concentration. The apparent stability of the peptidyl transferase center (PTC) of the ribosome in the presence of a variety of RNCs was evaluated by urea titrations upon detection via a puromycin-release assay. Finally, the empty-ribosome and RNC assembly status of the ribosome, in terms of 30S, 50S and 70S subunits, was assessed via sucrose-gradients and negative-staining transmission electron microscopy.
ApoHmp RNCs of increasing length interact with ribosomal protein L23. We elected to probe whether apoHmpH RNCs interact with the L23, L24 and L29 r-proteins, which reside within the vestibule of the ribosomal exit tunnel and the adjacent outer surface of the ribosome. We explored the interaction patterns of three representative nascent chains, namely apoHmp1 − 55, apoHmp1 − 140 and apoHmp1 − 189. The data for these RNCs are shown in Fig. 2. A side-by-side comparison between low-pH SDS-Page gels 75 and Western blots indicate that all three nascent proteins interact with ribosomal protein L23. Western blotting carried out with antibodies against ribosomal proteins L24 and L29, shown in the Supplementary Information (Fig. S1), indicates no evidence for interactions between the L24 and L29 r-proteins and the target RNCs. Therefore, apoHmpH RNCs interact exclusively with ribosomal protein L23. In contrast, intrinsically disordered PIR1 − 91 RNCs, analyzed in previous studies51, interact with both the L23 and L29 ribosomal proteins.
Under our experimental conditions, the fraction of interacting RNCs is different, for nascent chains derived from apoHmp1 − 55, apoHmp1 − 140, apoHmp1 − 189 (Fig. 2a-c) and intrinsically disordered PIR1 − 9151. Indeed, apoHmp1 − 55 and PIR1 − 91 crosslink only in part, unlike apoHmp1 − 140 and apoHmp1 − 189 RNCs, which are nearly 100% crosslinked. On the other hand, the larger extent of crosslinking of the foldable apoHmp1 − 140 and apoHmp1 − 189 RNCs relative to apoHmp1 − 55 and PIR may be mainly a consequence of the greater number of EDC-reactive residues of apoHmp1 − 140 (25 EDC-reactive residues, ca. 20 beyond the tunnel core) and apoHmp1 − 189 (36 EDC-reactive residues, ca. 30 beyond the tunnel core) relative to apoHmp1 − 55 (11 EDC-reactive residues, ca. 5 beyond the tunnel core) and PIR (14 EDC-reactive residues, ca. 12 beyond the tunnel core). In support of this argument (see sections below), the urea sensitivity of the L23 / RNC complexes is similar for all RNCs, suggesting comparable interaction strengths.
Importantly, given that fluorescence anisotropy-decay data (see later sections) show that apoHmp1 − 140 and apoHmp1 − 189 RNCs have dynamic and independently tumbling N-terminal compact regions, it is clear that the RNC regions interacting with the ribosomal surface cannot include any significant fraction of N-terminal residues belonging to the compact region.
In the case of the longest RNCs analyzed in this work, corresponding to apoHmp1 − 189, we found an additional interacting complex of higher molecular weight, which we denote as RP2 (Fig. 2c). The corresponding population includes r-protein L23, according to Western blotting, see Fig. 2c, and one additional unidentified protein of c.a. 6–10 kDa, according to molecular weight arguments. Our Western blots indicate that L29 (7 kDa) is not present in the RP2 band (Fig. S1f). Yet, other cytoplasmic E. coli chaperones and ribosome interactors (GroEL, GroES, SecB, DnaK/DnaJ/GrpE, SRP and ClpB; MW range: 48–80 kDa) are ruled out, as they would appear well above the RP2 complex in our gels (Fig. 2c and d). Due to its close spatial proximity to L23 (Fig. 3a,b,c and d) and based on the above-mentioned molecular-weight arguments, it is possible that RP2 comprises both L23 and L29. However, our monoclonal antibodies against r-protein L29 were unable to capture an L29 epitope, as part of the crosslinked complex adsorbed onto the PVDF membrane. To further test for the possibility that L29 being part of the RP2 interacting protein pair, additional experiments in the presence of polyclonal antibodies against L29 will be carried out in the future.
The data in Figs. 2 and S2 also show that a fraction of the apoHmp1 − 55, apoHmp1 − 140 and apoHmp1 − 189 RNCs interacts with the trigger factor (TF) chaperone. The presence of these contacts was assessed upon comparing denaturing gels for data collected with wild-type (WT) and TF-depleted (Dtig) E. coli cell strains. Indeed, RNC / TF interactions are already known to exist from previous literature, especially for nascent proteins longer than ca. 100–110 residues8 42,76−79 Previous studies also showed that TF interacts with client proteins that bear a fairly expanded conformation, in their bound state80–82.
On the other hand, the small observed fraction of apoHmp1 − 55 interacting with TF (Fig. 2a) is unexpected. This result implies the presence of a highly stretched conformation of this short (55-residue) RNC, which must reside mostly within the 80–100 Å-long ribosomal exit tunnel. Yet, apoHmp1 − 55 manages to reach out to the TF chaperone, which is known to dock onto the outer surface of the ribosome via the L23 and L29 r-proteins. This conformational stretching experienced by a small fraction of the short apoHmp1 − 55 RNC is fascinating and unprecedented. Indeed, the presence of L23-docked and TF-docked apoHmp1 − 55 nascent-chains suggests that cotranslational conformational sampling can take place even in the case of a fairly short RNC.
TF and L23 are known to interact with one another on the ribosome 77,83,84, though we presently cannot explicitly discriminate whether the nascent chains interact with L23 and TF, or if the nascent chain interacts with TF, which in turn interacts with L23. Here, we propose the simplest scenario namely that RNCs interact with TF only and, in turn, TF interacts with L23, which is known to be the TF docking site on the ribosome 2.
Finally, in this work we only analyzed the behavior of RNCs in the presence of moderate concentrations of the 70 kDa Hsp70 chaperone. Hsp70 was studied in the context of the DnaK/DnaJ/GrpE chaperone system, denoted here as K/J/E. Now, the wild-type (WT) cell-free system used in Fig. 2a,b,d contains K/J/E at 0.5, 0.04, 0.05 mM concentrations, respectively, which are significantly lower than physiologically relevant values. Interestingly, at these low K/J/E concentrations, none of these chaperones is bound to the apoHmpH1 − 55, apoHmp1 − 140 and apoHmp1 − 189 resuspended RNCs, as shown in Fig. 2a,b,d. Therefore, the Hsp70 chaperone does not bind the RNCs studied in this work. The effect of higher, more physiologically relevant (20–50 mM) K/J/E concentrations will be studied elsewhere 85.
In all, our data show that apoHmp1 − 55, apoHmpH1 − 140 and apoHmp1 − 189 RNCs interact with either the L23 r-protein alone (apoHmp1 − 55 and apoHmpH1 − 140, Fig. 3e), with L23 and another ribosomal protein (apoHmpH1 − 189), or with the TF chaperone (all RNCs, including apoHmp1 − 55). We propose that these two classes of interactions (i.e., with r-proteins and with TF) play a similar chaperone-like role. This concept is consistent with previous studies, which showed that the ribosome serves as a nascent-chain solubilizing agent even in the absence of chaperones 7. The fairly solvent-exposed nonpolar patch of the L23 r-protein, highlighted in Fig. 3c-d, is also consistent L23 being able to interact with nonpolar regions of RNCs. Future work will focus on genomic E. coli r-protein modifications aimed at disrupting the detected interactions.
Ribosome-bound apoHmp nascent chains of variable length have a compact N-terminal region. Next, we performed fluorescence depolarization decay experiments in the frequency domain 86–88 to probe the rotational dynamics of nascent chains encoding foldable sequences. This technique has been previously employed to assess the rotational correlation time (tc) and amplitude of rotational motions of RNCs 7,20,21,52,89. The goal of this experiment was to determine whether RNCs harboring long nascent chains display any degree of compaction. We focused on RNCs of apoHmp1 − 140, corresponding to the N-terminal domain 1 of Hmp (Fig. 1a), and RNCs of apoHmp1 − 189, which comprise Hmp’s domain 1 and an additional 49 C-terminal residues belonging to domain 2 (Fig. 1c). Nascent proteins were site-specifically labeled at their N terminus with the BODIPY-FL fluorophore as described 20. Once information on nascent-chain compaction is in hand, the interplay between ribosome and nascent-chain interactions, and their sensitivity to urea denaturation can be more rationally explored and understood, as apparent in the sections below.
Representative data for apoHmp1 − 140 and apoHmp1 − 189 are shown in panels a and b of Fig. 4, respectively. Both RNCs display informative frequency-domain anisotropy decay profiles. As shown in Fig. 4c and consistent with the very low reduced c2 values, the fits that include 3 rotational-tumbling components give the best results. Importantly, panels c and d of Fig. 4 show that both apoHmp1 − 140 and apoHmp1 − 189 RNCs are characterized by an N-terminal compact domain that tumbles independently from the ribosome. This conclusion was reached upon applying known procedures based on a combination of microscale viscosity and fluorescence depolarization in the frequency domain. In both cases, this domain spans ca. 63 to 94 residues, depending on the exact shape. Note that RNC shape assessment is beyond the scope of this work. Regardless of the actual overall morphology of the compact domains, the fact that a compact domain of identical size is observed for both apoHmp1 − 140 and apoHmp1 − 189 suggests that both constructs undergo a similar degree of partial folding on the ribosome. Surprisingly, the observed size of the compact domain of apoHmp1 − 189 RNCs is significantly smaller than the size of the entire apoHmp domain 1, which comprises 140 residues (Fig. 1c). Therefore, biosynthesis of the additional 49 C-terminal amino acids belonging to domain 2 is not sufficient to lead to complete folding of the N-terminal domain domain 1, for this protein.
In addition, cone semi-angle analysis of the fluorescence anisotropy decay data (Fig. 4c) shows that the compact domain of Hmp1 − 189 RNCs spans a slightly wider cone semi-angle (26.5⁰ ± 0.5⁰) than Hmp1 − 140 RNCs (20⁰ ± 0.2⁰), consistent with the fact that the latter construct likely projects slightly further out from the ribosomal surface than the shorter Hmp1 − 140 construct. In all, our fluorescence anisotropy data show that the Hmp1 − 140 and Hmp1 − 189 nascent chains are both comparably compact and no more than partially folded, while on the ribosome, with Hmp1 − 189 spanning a slightly wider cone semi-angle.
All the above information on fluorescence anisotropy decays is pictorially recapitulated by the cartoons of Fig. 4e. The images presented in this figure also show a variety of compact species that take into account the r-protein-interacting and non-interacting populations deduced from the SDS-Page gels and Western blotting data of Fig. 2. In order to recapitulate the RNC/r-protein interaction profiles and nascent-protein conformation knowledge gained so far, a model highlighting the leading trends is shown in Fig. 5. The RNCs displayed in this figure highlight the evolution of foldable apoHmpH nascent-chain interactions with r-proteins as a function of chain elongation. Briefly, when the nascent chain reaches a 55-residue length, the main detected interactions are with r-protein L23. No compact region is shown at this chain length, consistent with known fluorescence anisotropy-decay data collected on RNCs of a related globin 20. As the nascent chain gets longer and reaches a length of 140 residues, interactions with L23 are still present, but the chain also features a non-interacting compact region that spans a cone semi-angle of ca. 20⁰. As the nascent chain reaches 189-residue length, two classes of RNC populations interacting with r-proteins are present. The former interacts only with L23 and the other one also interacts with an additional ribosomal protein. In both cases, an N-terminal compact region encompassing 65–94 residues is also detected. Further chain elongation and ribosome-release processes, which are beyond the scope of this study, are expected to give rise to the full-length ribosome-released folded protein.
Nascent chain-L23 complexes have the same apparent stability regardless of RNC sequence. To further explore the nature of the interactions between the L23 r-protein and nascent chains of increasing length and variable sequence, we performed urea titrations with chemical crosslinking detection (Fig. 6a,b). EDC readily reacts with amines and carboxylic acid functional groups, and there is no loss of EDC reactivity even in the presence of high urea concentrations 90. It is worth noting that the interactions identified in this work are not induced by the covalently N-terminal-linked BODIPY-502 fluorophore, as previous work has shown that this fluorophore does not interact with resuspended ribosomes under conditions like those of the present study 20. Therefore, by unfolding the complex in the presence of urea and subsequently adding EDC, we expect to gain insights into the urea sensitivity of nascent chain-L23 complexes. While different RNC constructs are expected to bear a different number of EDC-reactive residues, denaturant titration of RNC complexes always examine the same nascent chain at variable urea concentration. Therefore, it is not necessary to normalize the data on a per-EDC-reactive-residue basis, as done in other studies 51.
After collecting gel data on representative apoHmp and PIR nascent chains (Fig. 6c), we estimated the apparent stability (ΔG°app,unfold) of nascent chain-L23 complexes following a known extrapolation method which is further described in the SI methods 91. Representative EDC-mediated urea titrations are shown in Fig. 6d. Corresponding plots and apparent-stability data are displayed in Fig. 6e,f. The matching two-tailed Student’s t-test is provided in Fig. 6g. As shown in Fig. 6h, the apparent stability values for the apoHmp and PIR nascent-chain/L23 complexes (RP1 complexes) range between DG0app,unf of 2.8 ± 1.3 and 5.8 ± 1.2 kcal mol− 1. As shown in the t-test of Fig. 6g, all complexes display the same apparent stability within error. The corresponding values for the apparent unfolding equilibrium constants Kapp (Table S2) are within the 590 ± 340 mM to 58 ± 41 mM range. These values, if regarded as estimates of the lower limits of the expected dissociation constants of r-protein/RNC complexes, suggest that the binding affinity of the apoHmp and PIR nascent-chain/L23 complexes (RP1 complexes) is overall rather weak. This qualitative estimate is consistent with the need for the interactions to be continuously remodeled during translation elongation. Interestingly, the observed trends apply even though the nascent-chain portions emerging from the ribosomal exit-tunnel core have widely different nonpolar and net-charge-per-residue (Fig. 6i) as well as widely different total nonpolar surface accessible surface-area values (Fig. 6j).
In summary, the urea-titrations in Fig. 6 show that the urea sensitivity of r-protein-nascent-chain complexes is similar regardless of the nature and length of the nascent chain, across the short and long (55- to 189-residue) chains examined here. In other words, RNC/-r-protein complexes have the same apparent stability, even though the RNCs have widely different physical properties and compaction (as discussed above and) and in the case of PIR, lack of compaction (discussed in previous work). 51,52 Given that the amino-acid sequences of the interacting regions of apoHmp1 − 55 and apoHmp1 − 140 must be different yet the interactions are of comparable apparent strength, the contacts are likely to be of nonspecific nature (Fig. 6f-h). This scenario, again, is consistent with the fact RNC-r-protein interactions likely need continuous remodeling during translation elongation.
Finally, the urea titrations described in Fig. 6 are highly informative, as they also display the urea dependence of complexes between nascent chains and the trigger factor chaperone (RNC/TF complexes), e.g., see selected upper bands in Fig. 6d. The quality of the data for the RNC/TF complexes was rather poor due to unreliable pre-transition baselines, therefore we did not deduce apparent stability values. On the other hand, as shown in Fig. 6d and in the plots of Fig. S3, the complexes with the TF chaperones are consistently less stable than the corresponding complexes with the L23 protein. This result suggests that the interactions between RNCs and the TF chaperone are even weaker than the interactions between RNCs and the L23 r-protein. Hence, nascent chains interacting with TF may in general be allowed more extensive conformational sampling in their bound state than nascent chains interacting with r-proteins. Additional future work will be devoted to further explore this hypothesis.
Nascent chain and r-protein interaction strength does not vary in the presence of one or more molecular chaperone. Next, we explored the effect of molecular chaperones TF and Hsp70 on the RNC-r-protein interactions via the same type of EDC-mediated urea titrations employed in the last section. The effect of Hsp70 was examined in the context of the K/J/E chaperone system. TF is known to associate with prokaryotic ribosomes 92 and K/J/E works in cooperation with TF 93 to promote nascent-protein folding and prevent nascent-protein aggregation 8,84,94,95.
First, we evaluated apoHmp1 − 189 devoid of both TF and the Hsp70 chaperone system (K/J/E), apoHmp1 − 189 in the presence of low concentrations of TF (2–15 nM) only, apoHmp1 − 189 in the presence of low concentrations of K/J/E (0.5, 0.04 and 0.05 mM, respectively) only, and apoHmp1 − 189 in the presence of both chaperones at low concentration (Fig. 7a). Urea titrations were carried out with increasing concentrations of urea (Fig. 7b), and the intensities of the crosslinked fractions were plotted (Fig. 7c). We then obtained a ΔG°app, unfold values for each of these constructs (Fig. 7d) and evaluated them with a two-tailed Student’s t-test (Fig. 7g), similarly to what done for the data in Fig. 6. Interestingly, the apparent strength of the L23-nascent chain complex was found to be statistically similar in all cases, regardless of chaperone concentration (Fig. 7f, h). Given that this effect is not due to a variation in the fraction of crosslinked nascent chains to r-proteins (Fig. 7e, f), via Western Blot analysis, we conclude that the extent of interactions between nascent chains and L23 remains similar in the absence and presence of the TF and K/J/E chaperones (Fig. 2c). This finding suggests that nascent chains interact with ribosomal L23 in a structurally similar fashion regardless of the absence or presence of chaperones.
RNC/r-protein interactions are attenuated at high chaperone levels in a chain-length-dependent manner. To further elucidate the nature of RNC/chaperone complexes, we performed experiments at low (2–15 nM) and high (8 µM) TF concentrations (Fig. 8). The high concentration values are representative of physiologically relevant TF concentrations, upon taking into account the differences in the concentrations of actively translating ribosomes in our cell-free system and in live E. coli cells 51. Interestingly, interactions with r-protein L23 are mostly displaced by interactions with TF, at high TF concentrations (Fig. 8). This effect, however, is more pronounced for longer RNCs, as shown by the representative gels of Fig. 8a,c,e, Fig. S2, and by the comprehensive analysis of the interacting populations shown in Fig. 8b,d,f. The shortest nascent chains of apoHmp1 − 55 only show c.a. 25% interactions with TF, even at high TF concentrations. We attribute this result to the fact that apoHmp1 − 55 is likely too short to form extensive interactions with the TF chaperone.
To summarize, in the absence of TF (Figs. 7 and S2), the nascent chain either interacts primarily with r-proteins, mainly, L23. At higher, physiologically relevant concentrations of TF (8 µM TF), RNC interactions with r-proteins are displaced by interactions with this molecular chaperone. It is worth noting that TF is shared with thousands of additional cellular proteins in vivo, unlike in the experiments shown here, which include purified resuspended RNCs. Further, our RNC concentrations are only 20–30 nM. Hence, the TF chaperone is in large excess over RNCs even at the low chaperone concentrations employed here. This scenario differs from the cellular environment where both RNCs and molecular chaperones are at comparable concentrations, within the low uM range. Therefore, we propose that the actual cellular milieu likely involves RNC populations that interact in part with TF and in part with r-proteins.
In all, our findings highlight the prominent role of the ribosome as an RNC interactor and suggest that the ribosome may have played a primordial chaperone role in Nature, before the evolution of the TF molecular chaperone.
The presence of very short nascent chains stabilizes the 70S ribosomal complex. After exploring nascent chain and r-protein interactions, we investigated the potential effect of these contacts on the bacterial ribosome. We began by performing a series of qualitative sucrose-gradient studies on E. coli empty ribosomes and nascent-chain-loaded ribosomes. Our results, detailed in Fig. S4 and S5, showed that empty-70S ribosomes are more sensitive to urea denaturation than ribosomes bearing tRNAs linked to longer nascent chains. These results agree with previous sucrose gradient studies on RNCs 37. It appears that the snc-tRNA is responsible for most of the stabilizing effect (Figs. S4 and S5). Interestingly, these data suggest that length and amino-acid sequence of the nascent protein does not influence the urea sensitivity of ribosome-RNC complexes.
The peptidyl transferase center site is largely unaffected by nascent-chain sequence and length, beyond 32 residues. Next, we probed whether nascent chains of different length, amino-acid sequence and foldability affect the apparent stability of specific regions of the ribosome. We directed our focus on the peptidyl transferase center (PTC) of the E. coli ribosome, and we explored its urea sensitivity via a nascent-chain ribosome-release assay mediated by puromycin. These experiments employed a larger set of RNCs than in the previous sections.
The results of puromycin-release-detected urea titrations are shown in Fig. S8 and further described in the Supplementary Information. Overall, the data show that the apparent stability of the ribosomal PTC is not affected by the presence of nascent chains longer than 32 residues.
The global urea sensitivity of ribosomal proteins is largely unaffected by nascent-chain sequence and length. Next, we explored the effect of nascent-chain properties on the overall apparent stability of r-proteins via urea titrations based on Trp fluorescence emission. Trp is a well-known fluorescent reporter, and its emission properties are highly environmentally sensitive. Urea titrations were carried out and Trp fluorescence emission was monitored (Fig. S9b,c). Spectral shifts were regarded as reporters of r-protein folding, and centers of mass of emission spectra were assessed to generate titration curves reporting on the urea sensitivity of r-proteins. Note that incubation time totaling the measurements from beginning and end of experiments did not change the spectral center of mass (Fig. S9d). Urea titration data were processed according to Santoro and Bolen 91,96. Individual representative titration curves are shown in Fig. S9e.
The ΔG°app, unfold for each construct are plotted in Fig. S9f and corresponding t-test values are tabulated in Fig. S9g. Nearly all the constructs show statistically similar results, with ΔG°app, unfold values ranging from 2 to 5 kcal•mol− 1 Hence, the presence of peptidyl tRNA, regardless of nascent-chain characteristics, does not affect the urea sensitivity of r-proteins. As shown in previous sections, some nascent chains interact with the specific ribosomal protein L23. On the other hand, these interactions are not sufficiently strong to be detected via this assay, which monitors the overall sensitivity to urea of all r-proteins.
Conclusions. The presence of interactions between nascent chains bearing a foldable amino-acid sequence (with no signal or arrest tags) and specific ribosomal proteins has been suggested but never experimentally demonstrated, to date. Here, we identify the ribosomal protein L23 as a specific nascent-chain-interacting partner. L23 establishes noncovalent contacts with nascent chains of the multi-domain foldable model protein apoHmp, which lacks signal/arrest sequences. As nascent chains elongate, the RNC interaction network expands to another ribosomal protein. A non-interacting N-terminal compact RNC region comprising 63–95 residues has also been identified for nascent chains bearing both 140 and 189 residues. A model recapitulating the presence of both RNC/r-protein interactions and non-interacting N-terminal regions is shown in Fig. 5. Interactions with the TF take over, at high TF chaperone concentrations. Interestingly, ribosomal-protein/nascent-chain complexes have a similar weak apparent stability regardless of nascent-chain sequence, length and degree of foldability. Therefore, we propose that r-proteins shield nascent foldable proteins from aggregation before intramolecular folding becomes thermodynamically favorable, during and(or) immediately after translation. These findings are significant because they unveil the presence of interactions between a foldable nascent chain and the L23 ribosomal protein. In addition, the data reveal that these interactions coexist with nascent-chain compaction across the N-terminal region, suggesting ribosome-facilitated aggregation-prevention and conformational sampling.