High-speed atomic force microscopy tracks the dynamic parts of the ribosome

“In biology, use of the force microscope will probably become quite common because of its ability to deliver films of processes,” predicted Gerd Binnig, co-inventor and Nobel prize winner for the scanning tunneling microscope, and coinventor of the atomic force microscope (AFM), in a 1992 review (ref. 1, p. 14). This prediction was made based on the fact that by 1992 AFM operated in liquid and at ambient temperature and pressure, and thus should allow the direct observation of biological dynamics. However, the image acquisition speed of AFM was far too slow to visualize biomolecules in action at relevant timescales, until Ando et al. (2) presented a technological milestone in a 2001 PNAS paper, the development of the first high-speed AFM (HS-AFM) that allowed taking movies of biomolecules at subsecond imaging speed, in liquid and at ambient temperature and pressure. Further improvements were necessary to accomplish the direct visualization of myosin-V walking along actin filaments (3). Since then, HS-AFM has proven powerful to image a wide variety of biomolecules, including channels (4, 5), transporters (6), and protein–DNA complexes (7, 8). In all these examples, HS-AFM revealed molecular dynamics without inference, in contrast to methods where the response of labels or other secondary signals are measured. In the present work by Imai et al. (9), the first author who took the movies of walking myosin-V, as one of a group of lead authors, now presents movies showing how individual archaeal ribosomes use their P-stalk as a flexible landing platform for the recruitment of translational GTPases aEF1A and aEF2.

Translation, the process by which information encoded in messenger RNAs is used to assemble a growing polypeptide chain that will ultimately fold into a protein, is the second step of the central dogma of molecular biology, and a crucial process in every cell. Translation is mediated by the ribosome. The ribosome is likely one of the best characterized protein complexes, functionally, and structurally by means of X-ray crystallography (10) and electron cryomicroscopy (11). However, these structural techniques that take advantage of averaging over tens of thousands of molecules to calculate a high-resolution structure, are, therefore, essentially blind when it comes to visualizing highly dynamic structural elements and/or flexible complexes of changing molecular assembly. This is where the unique power of HS-AFM, capable to directly visualize single molecules at high spatiotemporal resolution, comes in.

The Ribosomal P-Stalk Recruits and Retains EFs

Ribosomes are constituted of a large and a small ribosomal subunit. Both are made up of several ribosomal proteins (RPs) and ribosomal RNAs (rRNAs). Ribosomes from prokaryotes, archaea, and eukaryotes resemble each other, but differ in size and overall RP and rRNA composition. The present study by Imai et al. (9) examines the archaeal ribosome. Archaeal ribosomes are similar in size to those in prokaryotes but resemble in terms of sequence and composition more eukaryotic ribosomes. A particular part of the ribosome is the P-stalk, which is poorly understood likely due to its extremely flexible nature (12, 13). The P-stalk is itself a heptamer, composed of aP0–(aP1–aP1)₃ subunits, where each subunit features a long C-terminal tail devoid of secondary structure. The authors draw a model picture of the P-stalk as a curved extension from the 50S ribosomal subunit that terminates in seven tentacle-like arms floating around the stalk. The P-stalk plays an essential role in recruiting and associating translational GTPases EF1A and EF2 to the ribosome, and delivering them to the factor-binding center (13). EF1A delivers aminoacyl-tRNAs to the ribosomal A-site, and EF2 catalyzes translocation (14). Thus, these factors fuel efficient translation elongation via GTP hydrolysis. HS-AFM revealed the flexibility of the ribosomal P-stalk and the recruitment of up to seven of these EF factors to the P-stalk with single-molecule accuracy.

In the HS-AFM movies, the P-stalk protruded out from the ribosomal 50S subunit. The P-stalk was not only found in its canonical conformation, but also in a flipped state, in which the peripheral part of the stalk changes its orientation over a hinge region located approximately halfway along the stalk. The canonical state was characterized by a kink of the hinge of about −50° with respect to a hypothetical straight extension, while the flipped state had a kink of about +30°. Given the fast time-lapse capability of HS-AFM, the authors estimated lifetimes of ∼0.5 and ∼0.3 s, respectively, for the two states. The authors do, however, hypothesize that these conformational changes are likely faster free in solution. Thus, the P-stalk, although adopting two favorable states in these experiments, is indeed a highly flexible structure.

Next, the authors supplemented the 50S–P-stalk complexes with archaeal EF factors, aEF2 or aEF1A, loaded with GTP or GDP. HS-AFM images revealed in both cases several EFs concomitantly associated to the stalk, direct evidence for the factor-pooling hypothesis (15). Probability density maps from single-molecule observations showed that GTP–aEF2s were found in a cloud of positions with average distance of ∼12.9 nm from the stalk base, while GDP–aEF2s distributed at ∼11.5-nm distance. The probability to find aEF2 close to the P-stalk was ∼6.6-fold (GTP–aEF2) and ∼4.3-fold (GDP–aEF2) increased compared to background. aEF1A was recruited at closer distance, ∼9.5 nm from the stalk base, with ∼4.2-fold (GTP–aEF1A) and ∼2.6-fold (GDP–aEF1A) increased probability. The authors refrain from speculating about the reasons why aEF1A was found somewhat closer associated than aEF2; they also do not speculate about the association probability differences between the GTP and GDP supplemented EFs, although this is evocative of the possibility that the EFs in presence of GTP were more strongly bound to the stalk than when with GDP in solution. However, the fact that GDP-bound translational GTPases were still detected with significantly increased probability indicated, and was interpreted as such, that the stalk not only recruits, but also retains the factors after use. All of these experiments were controlled against observations where 50S particles devoid of P-stalk were analyzed.

A Fully Loaded Toolbox

Finally, upon supplementing the 50S–P-stalk complexes with increasing concentrations of aEF2, increasing average numbers of bound EFs were observed. In the highest concentrations, the authors were indeed able to monitor particles with seven associated, clearly resolved aEF2s. Given the stoichiometry of the P-stalk, this should represent a fully loaded complex.

In the present work by Imai et al., the first author who took the movies of walking myosin-V, as one of a group of lead authors, now presents movies showing how individual archaeal ribosomes use their P-stalk as a flexible landing platform for the recruitment of translational GTPases aEF1A and aEF2.

The authors discuss possible concerns about the experiments being performed using a hybrid 50S subunit and that the HS-AFM observations took place on a mica sample support. Nevertheless, they highlight that the hybrid system was previously shown to accurately recapitulate the function of the ribosomal stalk in translation elongation (16). Additionally, as we have learned from the walking myosin-V and other studies, the HS-AFM’s touch is gentle enough to not significantly disturb biological processes (17). All these findings provide strong support for the so-called factor-pooling hypothesis (15). Accordingly, the P-stalk is used to increase and maintain an increased local concentration of translational GTPase elongation factors for highly efficient ribosomal function. Indeed, it has been shown biochemically that the recruitment of these factors occurs much more rapidly than expected through free diffusion in the crowded cytoplasm (18). The ribosome is thus the handyman of the cell that permanently carries the most crucial tools on its belt.

In this work, HS-AFM was diligently used to provide insights into a crucial biological process. HS-AFM’s ability, owing to the high signal-to-noise ratio in images, to directly visualize dynamic processes in real time and with nanometric precision, gives it a unique position besides other structural methods to “deliver films of processes.”