AlphaFold2 (AF2) has emerged in recent years as a groundbreaking innovation that has revolutionized several scientific fields, in particular structural biology, drug design, and the elucidation of disease mechanisms. Many scientists now use AF2 on a daily basis, including non-specialist users. This chapter is aimed at the latter. Tips and tricks for getting the most out of AF2 to produce a high-quality biological model are discussed here. We suggest to non-specialist users how to maintain a critical perspective when working with AF2 models and provide guidelines on how to properly evaluate them. After showing how to perform our own structure prediction using ColabFold, we list several ways to improve AF2 models by adding information that is missing from the original AF2 model. By using software such as AlphaFill to add cofactors and ligands to the models, or MODELLER to add disulfide bridges between cysteines, we guide users to build a high-quality biological model suitable for applications such as drug design, protein interaction, or molecular dynamics studies.

Artificial intelligence has revolutionized the field of protein structure prediction. However, with more powerful and complex software being developed, it is accessibility and ease of use rather than capability that is quickly becoming a limiting factor to end users. LazyAF is a Google Colaboratory-based pipeline which integrates the existing ColabFold BATCH software to streamline the process of medium-scale protein-protein interaction prediction. LazyAF was used to predict the interactome of the 76 proteins encoded on the broad-host-range multi-drug resistance plasmid RK2, demonstrating the ease and accessibility the pipeline provides.

Cytokinins (CK) are one of the most important classes of phytohormones that regulate a wide range of processes in plants. A CK receptor, a sensor hybrid histidine kinase, was discovered more than 20 years ago, but the structural basis for its signaling is still a challenge for plant biologists. To date, only two fragments of the CK receptor structure, the sensory module and the receiver domain, were experimentally resolved. Some other regions were built up by molecular modeling based on structures of proteins homologous to CK receptors. However, in the long term, these data have proven insufficient for solving the structure of the full-sized CK receptor. The functional unit of CK receptor is the receptor dimer. In this article, a molecular structure of the dimeric form of the full-length CK receptor based on AlphaFold Multimer and ColabFold modeling is presented for the first time. Structural changes of the receptor upon interacting with phosphotransfer protein are visualized. According to mathematical simulation and available data, both types of dimeric receptor complexes with hormones, either half- or fully liganded, appear to be active in triggering signals. In addition, the prospects of using this and similar models to address remaining fundamental problems of CK signaling were outlined.

Protein structure determination is a critical aspect of biological research, enabling us to understand protein function and potential applications. Recent advances in deep learning and artificial intelligence have led to the development of several protein structure prediction tools, such as AlphaFold2 and ColabFold. However, their performance has primarily been evaluated on well-characterised proteins and their ability to predict sturtctures of proteins lacking experimental structures, such as many snake venom toxins, has been less scrutinised. In this study, we evaluated three modelling tools on their prediction of over 1000 snake venom toxin structures for which no experimental structures exist. Our findings show that AlphaFold2 (AF2) performed the best across all assessed parameters. We also observed that ColabFold (CF) only scored slightly worse than AF2, while being computationally less intensive. All tools struggled with regions of intrinsic disorder, such as loops and propeptide regions, and performed well in predicting the structure of functional domains. Overall, our study highlights the importance of exercising caution when working with proteins with no experimental structures available, particularly those that are large and contain flexible regions. Nonetheless, leveraging computational structure prediction tools can provide valuable insights into the modelling of protein interactions with different targets and reveal potential binding sites, active sites, and conformational changes, as well as into the design of potential molecular binders for reagent, diagnostic, or therapeutic purposes.

Magnesium (Mg2+) is the most abundant divalent cation in the cell and is critical for numerous cellular processes. Despite its importance, the mechanisms of intracellular Mg2+ transport and its regulation are poorly understood. MgtE is the main Mg2+ transport system in almost half of bacterial species and is an ortholog of mammalian SLC41A1 transporters, which are implicated in neurodegenerative diseases and cancer. To date, only MgtE from Thermus thermophilus (MgtETT) has been extensively characterized, mostly in detergent micelles, and gating-related structural dynamics in biologically relevant membranes are scarce. The MgtE homolog from Bacillus firmus (MgtEBF) is unique since it lacks the entire Mg2+-sensing N-domain but has conserved structural motifs in the TM-domain for Mg2+ transport. In this work, we have successfully purified this novel homolog in a stable and functional form, and ColabFold structure prediction analysis suggests a homodimer. Further, microscale thermophoresis experiments show that MgtEBF binds Mg2+ and ATP, similar to MgtETT. Importantly, we show that, despite lacking the N-domain, MgtEBF mediates Mg2+ transport function in the presence of an inwardly directed Mg2+ gradient in reconstituted proteoliposomes. Furthermore, comparison of the organization and dynamics of Trp residues in the TM-domain of MgtEBF in membrane mimetics, in apo- and Mg2+-bound forms, suggests that the cytoplasmic binding of Mg2+ might involve modest gating-related conformational changes at the TM-domain. Overall, our results show that the gating-related structural dynamics (hydration dynamics, conformational heterogeneity) of the full-length MgtEBF is significantly changed in functionally pertinent membrane environment, emphasizing the importance of lipid-protein interactions in MgtE gating mechanisms.

Plants are gaining traction as a cost-effective and scalable platform for producing recombinant proteins. However, expressing integral membrane proteins in plants is challenging due to their hydrophobic nature. In our study, we used transient and stable expression systems in Nicotiana benthamiana and Camelina sativa respectively to express SARS-CoV-2 E and M integral proteins, and target them to lipid droplets (LDs). LDs offer an ideal environment for folding hydrophobic proteins and aid in their purification through flotation. We tested various protein fusions with different linkers and tags and used three dimensional structure predictions to assess their effects. E and M mostly localized in the ER in N. benthamiana leaves but E could be targeted to LDs in oil accumulating tobacco when fused with oleosin, a LD integral protein. In Camelina sativa seeds, E and M were however found associated with purified LDs. By enhancing the accumulation of E and M within LDs through oleosin, we enriched these proteins in the purified floating fraction. This strategy provides an alternative approach for efficiently producing and purifying hydrophobic pharmaceuticals and vaccines using plant systems.

Jasmonates (JAs) are plant hormones with crucial roles in development and stress resilience. They activate MYC transcription factors by mediating the proteolysis of MYC inhibitors called JAZ proteins. In the absence of JA, JAZ proteins bind and inhibit MYC through the assembly of MYC-JAZ-Novel Interactor of JAZ (NINJA)-TPL repressor complexes. However, JAZ and NINJA are predicted to be largely intrinsically unstructured, which has precluded their experimental structure determination. Through a combination of biochemical, mutational, and biophysical analyses and AlphaFold-derived ColabFold modeling, we characterized JAZ-JAZ and JAZ-NINJA interactions and generated models with detailed, high-confidence domain interfaces. We demonstrate that JAZ, NINJA, and MYC interface domains are dynamic in isolation and become stabilized in a stepwise order upon complex assembly. By contrast, most JAZ and NINJA regions outside of the interfaces remain highly dynamic and cannot be modeled in a single conformation. Our data indicate that the small JAZ Zinc finger expressed in Inflorescence Meristem (ZIM) motif mediates JAZ-JAZ and JAZ-NINJA interactions through separate surfaces, and our data further suggest that NINJA modulates JAZ dimerization. This study advances our understanding of JA signaling by providing insights into the dynamics, interactions, and structure of the JAZ-NINJA core of the JA repressor complex.

The FtsQBL is an essential molecular complex sitting midway through bacterial divisome assembly. To visualize and understand its structure, and the consequences of its membrane anchorage, we produced a model of the E. coli complex using the deep-learning prediction utility, AlphaFold 2. The heterotrimeric model was inserted into a 3-lipid model membrane and subjected to a 500-ns atomistic molecular dynamics simulation. The model is superb in quality and captures most experimentally derived structural features, at both the secondary structure and the side-chain levels. The model consists of a uniquely interlocking module contributed by the C-terminal regions of all three proteins. The functionally important constriction control domain residues of FtsB and FtsL are located at a fixed vertical position of ∼43-49 Å from the membrane surface. While the periplasmic domains of all three proteins are well-defined and rigid, the single transmembrane helices of each are flexible and their collective twisting and bending contribute to most structural variations, according to principal component analysis. Considering FtsQ only, the protein is more flexible in its free state relative to its complexed state-with the biggest structural changes located at the elbow between the transmembrane helix and the α-domain. The disordered N-terminal domains of FtsQ and FtsL associate with the cytoplasmic surface of the inner membrane instead of freely venturing into the solvent. Contact network analysis highlighted the formation of the interlocking trimeric module in FtsQBL as playing a central role in mediating the overall structure of the complex.

Lipid transfer between organelles requires proteins that shield the hydrophobic portions of lipids as they cross the cytoplasm. In the last decade a new structural form of lipid transfer protein (LTP) has been found: long hydrophobic grooves made of beta-sheet that bridge between organelles at membrane contact sites. Eukaryotes have five families of bridge-like LTPs: VPS13, ATG2, SHIP164, Hobbit and Tweek. These are unified into a single superfamily through their bridges being composed of just one domain, called the repeating beta groove (RBG) domain, which builds into rod shaped multimers with a hydrophobic-lined groove and hydrophilic exterior. Here, sequences and predicted structures of the RBG superfamily were analyzed in depth. Phylogenetics showed that the last eukaryotic common ancestor contained all five RBG proteins, with duplicated VPS13s. The current set of long RBG protein appears to have arisen in even earlier ancestors from shorter forms with 4 RBG domains. The extreme ends of most RBG proteins have amphipathic helices that might be an adaptation for direct or indirect bilayer interaction, although this has yet to be tested. The one exception to this is the C-terminus of SHIP164, which instead has a coiled-coil. Finally, the exterior surfaces of the RBG bridges are shown to have conserved residues along most of their length, indicating sites for partner interactions almost all of which are unknown. These findings can inform future cell biological and biochemical experiments.

Reef-building corals play an important role in the marine ecosystem, and analyzing their proteomes from a structural perspective will exert positive effects on exploring their biology. Here we integrated mass spectrometry with newly published ColabFold to obtain digital structural proteomes of dominant reef-building corals.
Of the 8,382 homologous proteins in Acropora muricata, Montipora foliosa, and Pocillopora verrucosa identified, 8,166 received predicted structures after about 4,060 GPU hours of computation. The resulting dataset covers 83.6% of residues with a confident prediction, while 25.9% have very high confidence.
Our work provides insight-worthy predictions for coral research, confirms the reliability of ColabFold in practice, and is expected to be a reference case in the impending high-throughput era of structural proteomics.