Putting on molecular weight: Enabling cryo-EM structure determination of sub-100-kDa proteins

Abstract

Significant advances in the past decade have enabled high-resolution structure determination of a vast variety of proteins by cryogenic electron microscopy single particle analysis. Despite improved sample preparation, next-generation imaging hardware, and advanced single particle analysis algorithms, small proteins remain elusive for reconstruction due to low signal-to-noise and lack of distinctive structural features. Multiple efforts have therefore been directed at the development of size-increase techniques for small proteins. Here we review the latest methods for increasing effective molecular weight of proteins <100 ​kDa through target protein binding or target protein fusion - specifically by using nanobody-based assemblies, fusion tags, and symmetric scaffolds. Finally, we summarize these state-of-the-art techniques into a decision-tree to facilitate the design of tailored future approaches, and thus for further exploration of ever-smaller proteins that make up the largest part of the human genome.

Keywords: BRIL, cytochromeb562 RIL; DARPin, Design Ankyrin Repeat Protein; Fab, antigen binding fragment; GFP, Green Fluorecent Protein; GPCR, G protein-coupled receptor; MW, molecular weight; Mb, megabody; Nb, nanobody; SNR, signal-to-noise ratio; SPA, single particle analysis; TM, transmembrane; cryo-EM, cryogenic electron microscopy; kDa, kiloDalton; κOR ICL3, κ-opiod receptor intracellular loop 3.

Fig. 1

a. Size distribution of the UniProtKB human proteome compared to size distribution of EMDB entries of human proteins resolved below 4 ​Å resolution. Small proteins (<100 ​kDa) are under-represented in the SPA electron density database relative to their known abundance in the human proteome b. Low signal-to-noise ratio and lack of morphological features of small proteins (depicted as penguins) in electron micrographs complicates proper alignment of particles. c. Augmenting small proteins by appending extra molecular weight (depicted as seals) increases the signal-to-noise ratio and adds distinctive features to the assembly. The appended asymmetric body (missing flipper of the seal) allows different orientations of the assembly to be distinguishable for unambiguous alignment.

Fig. 2

a. NabFab enabled 3.8 ​Å-resolution reconstruction of 50-kDa target protein ScaDMT (EMD-13438, PDB:7PIJ). The target is rigidly bound by a Nb, which in turn is bound by a Nb-specific Fab. An additional Nb is bound to minimize the Fab internal flexibility. b. Example of a Megabody enabling a 2.7 ​Å-resolution reconstruction of 62-kDa target protein HHAT (EMD- 13764, PDB:7Q1U). A Mb, consisting of a target-specific Nb fused with a large protein, is stably bound to the target protein. c. Use of a Legobody enabled 3.6 ​Å-resolution reconstruction of 22-kDa target protein SARS-CoV-2 RBD (EMD-24729, PDB:7RXD). The target is rigidly bound by a Nb, which in turn is bound by a Nb-specific Fab. This assembly is stabilized by a MBP-PrAC fused to Fab-binding PrAD-PrG which binds Nb as well as Fab. d. Nanobit enabled high-resolution reconstruction of 46-kDa target protein GLP-2R (EMD-30590, PDB:7D68). LgBit-fused GPCR is bound to its HiBit-fused natural G protein ligand. The LgBit-HiBit interaction stabilizes the physiological interaction between GPCR and G protein. The originally unresolved LgBit and HiBit structures were predicted using AlphaFold (Jumper et al., 2021). NanoBit schematic not to scale. e. A so-called BRIL insert enabled 3.7 ​Å-resolution reconstruction of 65-kDa target protein Fzd5 (EMD-21927, PDB:6WW2). BRIL is a two-point insertion between two anti-parallel α-helices of the target protein. BRIL serves as binding site for a Nb-stabilized Fab. f. Schematic of a PGS-insertion strategy that enabled the 3.7 ​Å-resolution reconstruction of ∼61-kDa target protein SMO. PGS is fused to the target protein via a flexible two-point α-helical insertion, whereby the structural rigidity of the target-PGS fusion is ensured by a hydrophobic interaction between the target and PGS. g. Use of a κOR-ICL3 insertion with anti-ICL3 Nb6 enabled the 2.4 ​Å-resolution reconstruction of ∼46 ​kDa NTSR1 (EMD-26589, PDB:7UL2). κOR-ICL2 is a two-point insertion between two anti-parallel α-helices of the target protein, and is bound by a Nb (Nb6). Left of each panel: Schematics of assemblies. Middle: Colored electron density maps of full assemblies. Right: Zoom-ins of area indicated with black dashed line in middle panels. Colored dashed lines in panels c–d denote unresolved fusion connections. The molecular weight of the target protein (MW_target) and the total assembly (MW_total) are indicated above the schematics. Electron densities and structural models were produces using UCSF ChimeraX (Goddard et al., 2018). Note that the local target resolution may differ from the published global resolutions indicated here.

Fig. 3

a. Benefits of using a symmetric scaffold for SPA as compared to non-scaffolding size augmentations. A schematic is shown of a generic symmetric scaffold with a single in-plane 6-fold symmetry axis and with target proteins docked at the periphery. b. Use of a target-specific DARPin fused with a cage subunits enabled the 3.8 ​Å-resolution reconstruction of 26-kDa GFP upon self-assembly of the DARPin-cage symmetric scaffold (EMD-9373 and EMD-9374, PDB:6NHV and 6NHT). c. Direct fusion with apoferritin enabled the 2.6 ​Å-resolution reconstruction of 11-kDa KIX domain upon self-assembly of the apoferritin symmetric scaffold (EMD-25791, PDB:7TB3. Left of panels b–c: Schematics of assemblies. Middle: Colored electron density maps of full assemblies. Right: Zoom-ins of area indicated with black dashed line in middle panels. d. Scaffold parameters corresponding to the examples given in panels b–c, including the molecular weight of the target protein (MW_target) and the total assembly (MW_total). Electron densities and structural models were produces using UCSF ChimeraX (Goddard et al., 2018). Note that the local target resolution may differ from the published global resolutions indicated here.

Fig. 4

To enable cryo-EM SPA of small target proteins of interest (center vignette), the characteristics of the protein can guide choice of strategy for protein size-augmentation. A target binding (left) or a target fusion (right) strategy can be employed, each having multiple variants that together cover a wide range of small protein targets. Boxes in dashed blue lines are additional suggestions. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)