(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



The dynamic architecture of Map1- and NatB-ribosome complexes coordinates the sequential modifications of nascent polypeptide chains [1]

['Alexandra G. Knorr', 'Department Of Biochemistry', 'Gene Center', 'Ludwig-Maximilians University Munich', 'University Of Munich', 'Munich', 'Timur Mackens-Kiani', 'Joanna Musial', 'Otto Berninghausen', 'Thomas Becker']

Date: 2023-04

Taken together, our analysis shows that Map1 is bound to the 80S ribosome in the immediate vicinity of the ribosomal tunnel exit mainly via a flexible association with the dynamic rRNA ES27a. This brings Map1 in an ideal position to act on nascent polypeptide chains as soon as they emerge from the ribosomal tunnel into the cytoplasm.

To confirm the major contribution of ES27a to Map1 ribosome binding, we performed in vitro binding assays with purified Map1 and ribosome nascent chain complexes (RNCs) or RNaseI-treated RNCs (rtRNCs), as done before for NatA [ 25 ]. In rtRNCs, rRNA ESs are clipped off by the RNAseI treatment, as previously shown [ 25 ]. Map1 binding to rtRNCs was significantly reduced by about 60% when comparing to untreated RNCs, again confirming that ES27a is a major binding site for Map1 recruitment to the ribosomal exit site ( Fig 1I ).

As stated above, ES27a is the main contact site for Map1 to the 60S ribosomal subunit. ES27 consists of three A helices, which can change their position flexibly around the three-way junction connecting rRNA helices H63, ES27a, and ES27b (nomenclature according to Petrov and colleagues [ 39 ]). The longest helix, ES27a, thereby undergoes the most severe conformational changes. In yeast, so far, two main positions are known, one with the tip of ES27a facing towards the L1 stalk of the 60S (L1-position) and one facing towards the peptide exit tunnel (exit-position) [ 31 ]. Interestingly, we observed ES27a-exit in two novel stabilized conformations, when bound to Map1. Compared to ES27a-bound NatA complexes, in Map1 complexes, ES27a is rotated with the three-way junction as a pivot by 31 degrees for conformation 1 and by 19 degrees for conformation 2 ( S6 Fig ).

Thus, despite the rather low resolution, the reconstructions allowed us to fit a model of Map1 generated by AlphaFold 2 (AF2) [ 37 ] into the respective densities, thereby providing an idea of the overall positioning of Map1 with respect to the ribosome ( Fig 1D ). The fits were guided by high-resolution cryo-EM structures of ribosome-bound EBP1 [ 34 , 35 , 38 ] ( S4 Fig ). After fitting the human 80S-EBP1 models into our densities for an overall orientation determination of Map1, we superimposed the AF2 model for Map1 and rigid body fitted it separately into isolated densities (Figs 1D and S5A–S5D ). This resulted in positioning of the globular amino peptidase domain (APD) of Map1 below the peptide exit tunnel contacting ES27a and H59. Density for the APD spans from UAS II (comprising eL31 and eL22) to UAS I (comprising uL23 and uL29) ( Fig 1E–1H ) [ 29 ] and is thus in agreement with published cross-linking data [ 16 ]. In addition, density for the nascent chain was visible, representing a broad variety in composition and lengths of nascent chains obtained through the native pullout. Therefore, a possible influence of differences in the nature and length of the nascent chain on the two observed Map1 states cannot be addressed.

In both classes, the position of Map1 at the exit tunnel is similar to that of the homologs of Map1 involved in 60S subunit biogenesis, yeast Arx1 (associated with ribosomal export complex protein 1) and human EBP1 (ErbB3-binding protein 1) [ 32 – 35 ]. Moreover, Map1-C1 superimposes well with the bacterial Map visualized in a PDF-Map-70S ribosome complex from Escherichia coli [ 36 ] ( S4 Fig ).

While both classes could be refined to an overall resolution of 3.8 and 3.9 Å (referred to as Map1-C1-80S and Map1-C2-80S), respectively, local resolution of Map1 bound to ES27a was limited to 7 Å and below ( S3 Fig ). This indicates a high degree of flexibility of the Map1-ribosome interaction, most likely owing in part to the flexibility of its binding partner ES27a, which can cover a continuous conformational space between tunnel exit site and L1 protuberance [ 31 ]. However, since local refinement attempts using the multibody approach in RELION also failed, we concluded that the flexibility of ES27a does not solely prevent higher resolution but that the binding of Map1 itself to ES27a is flexible. Accordingly, we were not able to gain higher local resolution required to build a molecular model for the regions comprising ES27a and Map1.

( A ) Concentrated eluate obtained from Map1-TAP affinity purification shown on a 12% Nu-PAGE. Mass spectrometry analysis confirmed the presence of Map1 and ribosomal proteins. A contamination from a viral protein is marked with an asterisk. See S1 Raw Images for the raw gel image. ( B ) Cryo-EM structures of Map1 in complex with a translating 80S ribosome in two different conformations (left: Map1-C1-80S; center: Map1-C2-80S; right: bottom view on the peptide exit tunnel). The maps were filtered according to local resolution. Isolated densities were extracted after the final CTF refinement. ( C ) Cartoon representation of bottom views with overlay (left) and separate views (right) as shown in ( B ). ( D ) An AlphaFold 2 model for Map1 was fitted into the density, shown as front view. The maps were filtered to 20 Å using a Gaussian low-pass filter. ( E - H ) Two views on the tunnel exit highlighting the position of Map1-C1 ( E , G ) and Map1-C2 ( F , H ) with respect to tunnel exit surrounding ribosomal proteins coloured as indicated in the legend below. ( I ) In vitro binding assay addressing the contribution of ES27a to ribosome-Map1 binding. Samples from the pelleting assay using recombinant Map1 and purified RNaseI-treated (rtRNC) or nontreated RNCs were applied to a 15% SDS-PAGE. For Map1 alone the supernatant (SN) fraction and for all other samples the pellet (P) fraction is shown. Co-pelleting of Map1 with the ribosome was quantified by densitometry. When ESs were digested by RNAseI, Map1 binding was significantly decreased to about 40% compared to Map1 binding to untreated 80S ribosomes. TE: tunnel exit. ZF: zinc finger domain. APD: aminopeptidase domain. Map1-C1: light green, Map1-C2: dark green, eL22: purple, H59: orange, 40S SU: light yellow, 60S SU: grey, ES27a: cyan, tRNAs: dark blue, nascent chain (NC): pink. See S1 Raw Images for raw gel images and S1 Data for numerical data underlying the quantification.

Cryo-EM samples for endogenous MetAP-ribosome complexes were obtained from native pullouts of ribosome-bound TAP-tagged Map1 essentially as described before [ 25 , 30 ] (Figs 1A and S1 ). After elution by cleavage of the Map1-tag using tobacco etch virus (TEV) protease, the Map1-ribosome complexes were stabilized by treatment with the chemical crosslinker glutaraldehyde prior to cryo-grid preparation. As observed before in native pullouts with cycloheximide-treated samples [ 25 , 30 ], 3D classification revealed classes with programmed ribosomes predominantly in the pre-translocational state with tRNAs present in the canonical A and P sites, but also classes representing termination/pre-recycling complexes (with eRF1 and ABCE1) and idle (tRNA-free) ribosomes ( S2 Fig ). Notably, in the majority of classes, ES27a was found in the position below the peptide exit tunnel (ES27-exit), and attached to ES27a, we observed an additional density reaching to the peptide exit tunnel. We joined those classes and subjected them to focused 3D classification using a soft mask for the exit tunnel/ES27a region. Two classes were enriched, showing a prominent globular, nonribosomal extra density attached to ES27a in different conformations below the peptide exit. These two classes differed mainly in the position of ES27a and its attached nonribosomal density, but not in the overall ribosomal state. Although the additional density could not be better resolved due to its flexibility, based on its position, overall shape, and dimension, we assigned it to Map1 (Figs 1B , 1C , and S2 ). ES27a in the exit position serving as the principal binding site for Map1 is in agreement with deletion experiments where the tip of ES27a was shortened, leading to reduced levels of ribosome associated Map1 [ 17 , 18 ]. Furthermore, our assignment agrees with biochemical findings based on chemical cross-linking, showing Map1 close to uL29, a protein located adjacent to the tunnel exit [ 16 ].

Cryo-EM structure of the NatB-ribosome complex

To gain further insight into the coordination between methionine cleavage and N-acetylation, we followed an in vitro reconstitution approach using purified components. We purified RNCs carrying a well-established NatB substrate as nascent chain, in which the free N-terminus ends with the amino acid sequence MDEL (RNC MDEL ). The same sequence was used in form of a CoA-Ac-MDEL inhibitor for co-crystallization with Chaetomium thermophilum Naa20 [28]. High salt-washed RNC MDEL were reconstituted with an 18× molar excess of recombinantly purified NatB (Naa25/Naa20) and subjected to cryo-EM and single particle analysis (S7 Fig). 3D variability analysis in CryoSPARC and focused sorting on the exit tunnel region revealed classes with extra density accounting for the NatB complex associated with the 80S ribosome, yet displaying a high degree of compositional and conformational heterogeneity (S8 Fig). Classes containing additional density corresponding to NatB could be divided into two sets: one set with extra density for only one copy of NatB (NatB-1; consisting of Naa20-1 and Naa25-1) flexibly attached to ES27a and one set with additional density for a second NatB complex bound to UAS-II [29] (Fig 2A). The second NatB complex (henceforth referred to as NatB-2) generally exhibited low conformational variance in classes where it was present, and its interaction with the ribosome was well resolved. To address the fact that ES27a-bound NatB-1 exhibited greater conformational heterogeneity, we performed focused sorting on this density, revealing one class (9.645 particles) in which both NatB complexes showed secondary structure resolution. Here, the ES27a-bound NatB-1 complex was positioned in direct vicinity to NatB-2 and exhibited much lower conformational flexibility than in other classes. This class (class I) was refined to a final overall resolution of 3.8 Å (local resolution ranging from approximately 4 to 9 Å for the NatB complexes; S9 Fig, left panel), which unambiguously revealed the architecture of both NatB complexes. In addition, we subjected all particles containing NatB-2 (and flexible NatB-1) to 3D variability analysis focusing on the NatB-2 area, yielding a class (class II) containing particularly well-resolved NatB-2 (45.530 particles) and refined this class to an overall resolution of 3.1 Å (local resolution ranging from below 3 to 6 Å for NatB-2; S9 Fig, right panel).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Cryo-EM structure and model of the NatB-ribosome complex. (A) Composite map showing cryo-EM structures of NatB in complex with translating 80S ribosomes (RNC MDEl ) filtered according to local resolution. Isolated densities of NatB-1 (from class I) and NatB-2 (from class II) were extracted after the final refinement. Views are shown on the peptide tunnel exit (left; bottom view), rotated 70° horizontally (middle; front view), and rotated 60° vertically (right; side view). (B) Zoom on the peptide exit site showing the NatB-ribosome molecular model (NatB-1 and NatB-2) in front (left) and side views (right) as indicated in (A). Overview (upper left panel (C)) and zoomed views (D-F) focusing on the Naa25-2 ribosome contact sites. Interactions of (D) helix α35 at the C-terminus of Naa25-2 with the H100/101-and H94/98 junctions, (E) the Naa25-2 α34-α35 loop (Lys720) with Asp6 in the N-terminus of eL31, and (F) Lys791 and Arg794 at the very C-terminus (α36) of Naa25-2 with U3153 and U3293 within H94/98-junction are shown. NatB-1 was omitted for clarity. (G) Same view as (C) showing the model for the NatB-2 ribosome complex docked in density and highlighting Naa25-2 C-terminus (orange). (H) Model of the Naa25 C-terminus highlighting the position of the four positive patches (PPs). All patches contain two charged amino acids as indicated. Charge inversion double mutants were generated (PP1, PP3, PP4, and PPall). (I) Western blot analysis of sedimentation assays (triplicates) using recombinant wild-type or mutant NatB complex and idle (80S) or RNaseI-treated 80S ribosomes (rt80S). Top: fraction of NatB bound to ribosomes quantified by densitometric analysis of western blot images. Bottom: representative western blot displaying supernatant (SN) and pellet (P) fractions of two such experiments. See S1 Raw Images for all raw western blot images and S1 Data for numerical data underlying the quantification. https://doi.org/10.1371/journal.pbio.3001995.g002

This revealed α-helical secondary structure in regions proximal to the ribosome and allowed us to unambiguously identify the disc-shaped α-helical tetratricopeptide repeat (TPR) containing Naa25 subunit for both ribosome-bound NatBs in class I (S5E–S5J Fig). We further noticed that the globular catalytic Naa20 subunits are less well resolved (when compared to Naa25), indicative of flexibility, especially in the ES27a-bound NatB-1. We then performed rigid-body fitting of an AF2 model which is highly similar to the crystal structure of C. albicans NatB (PDB 5K18) [26]. In brief, the 12 tetratrico (TPR-) repeats (α0-α29) of Naa25 together with its C-terminal helical domain (α30-α36) form a ring-like structure with N-and C-termini in close vicinity. Naa20 is located in and protrudes from a circular pocket formed by the Naa25 TPR repeats. This structure could be fitted with only minor adjustments into both NatB densities (S5E–S5J Fig and S1 Table).

Overall, the two NatB densities cover the area below the 60S exit site spanning from ES27a to the second UAS for exit factors (eL31 and uL22) [29]. NatB-1 is anchored between H59 of 25S rRNA and the long arm of the ES27a A-helix (Fig 2B). Here, contacts are established by the loops of Naa25-1 N-terminal TPRs (TPR1 and 2) that are sandwiched between the two rRNA elements. Another contact to the ES27a tip is established by the TPR-helices of the Naa25-1 C-terminus (α34-α36). In this conformation, the catalytic subunit Naa20-1 faces towards the exit tunnel, but density is only visible at the well-conserved contact interface with Naa25-1 (including highly conserved Thr2 and Glu48 of Naa20 and Arg296 of Naa25) [40], indicating that it is largely delocalized.

NatB-2 is anchored to the ribosomal surface somewhat offset from the tunnel exit of the 60S and is arranged such that the two catalytic Naa20 subunits face each other. In contrast to NatB-1, ribosomal contacts are in this case established mainly via rRNA but also to the ribosomal protein eL31, yet involving only the C-terminal TPRs of Naa25-2 (Fig 2C). In detail, we identified three distinct interaction sites of Naa25-2 with the ribosomal exit site in the map after focused refinement on NatB-2 (class II): The first site was established around the junction of H100 and H101 of 25S rRNA and the N-terminal part of α35 of Naa25-2 (Fig 2D) that contains a series of positively charged amino acids (Lys725, Lys729, Lys732, and Lys736). The second contact site is established between the Naa25-2 Lys720, located in the loop between α34 and α35, and Asp6 of the N-terminus of ribosomal protein eL31 (Fig 2E).

The third site comprises bases at the junction of H94 and H98 that bind the very C-terminus (α36) of Naa25-2 (Fig 2F). Here, the bases U3153 and U3293 within the H94/98-junction were contacted by Lys 791 and Arg794 at the very C-terminus of Naa25-2, likely via a cation-π stack.

To test the contribution of the before-mentioned residues to ribosome binding, we selected positive patches (each patch containing two closely spaced basic amino acids) at the C-terminus of Naa25 (Fig 2G) and an unrelated positive patch in the same area (Lys 747 and Lys 751). We created double charge inversion mutants (Lys or Arg to Glu) for each patch (PP1 to PP4; Fig 2H) or for all patches (PPall) similar to as described in ref [41]. Purified wild type (wt) and mutant NatB complexes carrying an N-terminal His-tag on Naa25 (for western blot detection) were used for in vitro binding assays. To prevent any bias by a specific nascent chain, we chose purified idle 80S ribosomes over RNCs in these assays (S7 Fig). The western blot analysis showed that binding of NatB to 80S ribosomes was significantly reduced by mutation of K723E, K725E (positive patch PP1; 47% of wt binding) and K791E, R794E (PP4; 31%), whereas K747E, K751E that in our structure are not directly involved in ribosome binding showed only a very weak effect (PP3; 83%). When all positive patches (PPall) were mutated, binding was almost entirely abolished (6% of wt binding), confirming the contribution of these positive patches to the interaction of NatB with the ribosome (Fig 2I). This indicated that the positive charges on the surface of the Naa25 C-terminus have an additive effect on ribosome binding by establishing a composite binding patch for rRNA interaction. The results of these binding assays confirm the significance of the described interaction patches between Naa25 and the ribosome. Whereas in Naa25-1, these positively charged amino acids interact with ES27a, in Naa25-2, they enable the binding to H94/H98 and H100/H101 junctions (Fig 2D–2F).

Interestingly, as observed for NatA and also Map1, in the class showing the stable assembly with two NatB complexes (class I), ES27a adopts a specific conformation. Compared to the Map1-C1 position of ES27a that is closest to the tunnel exit, the NatB-bound position is rotated 37° away from the tunnel exit (S6 Fig).

We thus assessed the contribution of ESs to NatB binding by performing quantitative binding assays using empty 80S or RNAse-I-treated 80S (depleted of ES as described in [25] and for Map1 in Fig 1I) (S7 Fig). The absence of ESs indeed reduced the binding of NatB to 8% compared to NatB binding to untreated 80S ribosomes, confirming an important role of the ES for recruitment of both NatB copies to the ribosome (Fig 2I).

We next compared ribosome-bound NatB-1 and NatB-2 with the NatA complex. Here, several observations were made. (i) The overall space occupied below the exit site is overlapping, indicating that in the observed conformations NatA and NatB can only bind exclusively (Fig 3A). (ii) The architecture of ribosome-bound NatB clearly differs from the NatA-ribosome complex and displays a distinct 60S binding mode. NatA mainly employs 25S rRNA ES for 60S binding. Here, ES27a and ES39 anchor the auxiliary Naa15 (Nat1) subunit to the exit site, and Naa50 (Nat5)—which has no equivalent in NatB—makes a third contact to ES7. While NatB-1 also binds to ES27a, NatB-2 engagement of the 60S differs compared to NatA binding and does not involve ES27a or other ESs. (iii) Like Naa10 (Ard1), the catalytic subunit of the NatA complex, both Naa20-1 and Naa20-2 of NatB have no direct contact to the ribosome. We further note that Naa20-2 is better resolved, while Naa20-1, apart from the contact site with Naa25-1, is largely delocalized. Notably, a rigid body fit of the NatB-2 model into NatB-1 would lead to a clash between Naa20-1 and Naa25-2, indicating that Naa20-1 needs to adjust its orientation with respect to Naa25 compared to NatB-2 (and the X-ray structure [26]). Nevertheless, in order to assess their principal potential to contribute catalytic activity, we compared this rigid-body fit of Naa20-1 with our models for Nat20-2 and Naa10 of NatA, since it represents a sufficiently accurate approximation of the overall position of Naa20-1 (Figs 3C and S10).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Comparison of ribosome-bound NatB and NatA complexes. (A) Bottom view (upper panel) and front view (lower panel) showing an overlay of the NatB-1 (pink) and NatB-2 (orange) ribosome structure with isolated densities for ribosome-bound NatA (bright green) (EMD-0201; [25]). (B) Comparison of positions for the NatA and NatB-2 catalytic subunits (Naa10 and Naa20) with respect to the 60S subunit shown as front and top view. The position of acetyl-CoA (Ac-CoA) and a putative model for the nascent chains is shown. For clarity, only Naa20 of NatB-2 is shown. (C) Left panel: cut front view of the NatB ribosome cryo-EM map highlighting the nascent polypeptide chain and the position NatB-1 and NatB-2 (left panel). Right panels: Zoom-in views highlighting the catalytic Naa20-2 subunit and illustration of the minimal distance that a nascent chain has to span to reach the active site of Naa20-2. https://doi.org/10.1371/journal.pbio.3001995.g003

Interestingly, it would require roughly the same minimum length of the nascent chain of about 55 amino acids to reach either one of the catalytic centers, assuming a direct path from the 3′-CCA end of the tRNA to the tunnel exit and from there into the Naa20 catalytic center (30 aa inside and 25 aa outside the exit tunnel) (Fig 3C). Naa20-2 is oriented similarly to Naa10 of NatA [25] with respect to the position of acetyl-CoA and accessibility for the nascent chain N-terminus (Fig 3B and 3C). While the substrate could enter Naa20-2 in a straight path from the tunnel exit, it would need to form a turn to reach into the center of Naa20-1, the entrance to which is located on the lateral side (S10 Fig). Thus, both copies of Naa20 (in NatB-1 and in NatB-2) could in principle be catalytically active. Yet, taking into account the delocalization and high degree of flexibility of Naa20-1 with respect to its auxiliary subunit in contrast to the more stably positioned Naa20-2, and given the more direct path that the nascent chain can take to enter Naa20-2, we speculate that Naa20-2 rather than Naa20-1 would act to N-acetylate most NatB substrates.

While binding of NatA and NatB appears mutually exclusive we wondered to what extent concomitant binding of Map1 would be sterically allowed. Unlike stated before [25], comparison of the binding modes reveals that Map1 in both C1 and C2 positions would possibly clash with NatA, although clashes between Map1-C1 and NatA would be rather minor. Yet, both NatB complexes, especially the ES27a-bound NatB-1, would severely overlap with both observed Map1 positions (S11 Fig). Thus, this comparison is rather suggestive for competitive binding of Map1 and NATs. This notion is further supported by the observation that ES27a orientations are apparently different for every ligand.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001995

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.