Mass‐Spec, Cryo‐EM and AI join forces for a close look at the transporter complex in cilia

Plain language summary

The intraflagellar transport (IFT) complex transports components between the cytoplasm and the ciliary tip. Two studies now report on the atomic structure of IFT‐B, the core of IFT, using cutting‐edge technology, such as cross‐linking mass spectrometry (MS), cryo‐electron tomography (cryo‐ET) and Alphafold2‐enabled AI‐based folding prediction. The 3D structure of IFT‐B reveals how the 15 component proteins are arranged to stabilize this gigantic complex and suggests a dynamic interplay between the proteins.

The long‐awaited atomic structure of the core of intraflagellar transport complex is revealed using cross‐linking mass spectrometry, cryo‐electron tomography and the AI‐based folding prediction program Alphafold2.

Cilia are appendage‐like organelles in eukaryotes, either generating beating motion (motile cilia) or responsible for sensory and intracellular transport (primary cilia). Judging from their highly conserved ultrastructure with nine microtubule doublets in a spoke‐like arrangement and component proteins, eukaryotes likely inherited this organelle from the last eukaryotic common ancestor. Ciliogenesis is initiated at the basal body (centriole) in the cytoplasm, starting from the recruitment of triplet microtubules into a nine‐fold symmetric arrangement by proteins forming a central hub and a bridge defining nine‐fold symmetry (Noga et al, 2022). At the distal end of the basal body, doublet microtubules extend to form a cilium. The extracellular part of cilia is called an axoneme. A short area called transition zone, located between the axoneme and the basal body at the level of the cilium's attachment to the plasma membrane. The transition zone shows distinct components decorating doublet microtubules, which serve to anchor the axoneme firmly to the cell body, as well as regulate entry and exit of components. Upon extension of the axoneme, “anterograde” IFT trains transport cargoes from cytoplasm and unload them at the ciliary tip. When the cell needs to resorb cilia, “retrograde” IFT carries cargoes from the tip back to the cytoplasm. IFT “runs” on track between doublet microtubules and ciliary membrane (Fig 1A). Ever since its discovery (Kozminski et al, 1993), cilia researchers have been fascinated by these nano‐scale trains and a number of diseases caused by defect of IFT have been identified.

Figure 1. 3D structure of intraflagellar transport (IFT)‐B.

(A) Schematic of the localization of IFT complexes and microtubule doublet in Cilia. (B) Close view of IFT‐B and IFT dynein. Surface rendered using EMDB4303 (Jordan et al, 2018), showing 11.5 nm periodicity of IFT‐B. One unit of IFT dynein, IFT‐B1 and IFT‐B2 are indicated. (C) 15mer IFT‐B structure, modelled by Petriman et al (2022). The colour code is the same as the article, but viewed from the opposite angle, to be compatible with Fig 1A and B.

IFT is made of two complexes (IFT‐A and IFT‐B) that are composed of 6 and 16 component proteins, respectively. Sequences of IFT components are also highly conserved from unicellular algae to human and show homology to coat proteins, hinting on the origin of this large transport complex (van Dam et al, 2013). For anterograde IFT, A and B complexes are arrayed with 6 and 11.5 nm periodicity (Jordan et al, 2018), while their arrangement is different in retrograde IFT. The train‐like structure of IFT moves towards the tip powered by kinesin‐2 and carrying IFT dyneins (homologous to cytoplasmic dynein), while it returns to the cytoplasm using IFT dyneins. Recently, the Engel group visualized the transition zone by cryo‐ET and observed that IFT‐B complexes form the backbone of IFT trains around the basal body, IFT‐A then attaches, followed by recruitment of IFT dynein near the transition zone, and finally kinesin‐2 (van den Hoek et al, 2022). Still, many questions remain unanswered. How does the IFT train load and release its cargo? How are the trains reorganized from the anterograde to the retrograde IFT conformation at the tip? 3D atomic structure of IFT is indispensable to answer these questions.

So far, 3D structural studies on IFT have not achieved much benefit from well‐known “Resolution revolution” in single particle cryo‐EM, which enables atomic resolution structural analysis without crystallization. This is probably due to that neither purification nor reconstruction of IFT complexes is straightforward. Instead, recent studies have used other cutting‐edge technologies to build an atomic model of IFT complexes. In this issue, the Lorentzen group reports 3D structure of IFT‐B (Petriman et al, 2022). IFT‐B consists of two subunits: IFT‐B1 (IFT52, IFT88, IFT70, IFT46, IFT74, IFT81, IFT22, IFT25, IFT27), which binds cargoes, and IFT‐B2 (IFT57, IFT38, IFT54, IFT20, IFT57, IFT80, IFT172), which binds IFT dynein and kinesin‐2 (Fig 1B). The work started from atomic models of IFT‐B proteins, that were previously solved by crystallography (Taschner & Lorentzen, 2016; Taschner et al, 2018). Applying cross‐linking MS, Petriman and colleagues mapped residues in proximity in the IFT‐B complex to reveal the topography of these components. For 3D structure mapping, they used Alphafold2. Alphafold2, a machine learning program that predicts protein fold, is another big revolution in structural biology, but prediction of complex formation, especially for complexes with many components, has just begun. Often the programme presents several possible models. The authors built models of subunits made by 4–6 components using the new Alphafold2‐multimer tool and validated their output models with cross‐linking MS (pull‐down assay as well). Next, this approach was extended to model the entire IFT‐B with 15 components. A number of coiled coils were found connecting adjacent proteins and at the interface to the cargo. These coiled coils become the backbone for complex formation, for example the IFT57/IFT38/IFT54/IFT20 tetramer. IFT57 and IFT52 show a long‐extended form as predicted by Alphafold2 and validated by cross‐linking MS (Fig 1C).

The Pigino group took a slightly different approach (preprint: Lacy et al, 2022). They improved spatial resolution of cryo‐ET and subtomogram averaging, with the help of a new subtomogram alignment algorithm. They reached outstanding resolution for thicker specimens (the diameter of cilia is ~ 300 nm, whereas high‐resolution cryo‐ET is usually limited to < 200 nm thickness), which allowed them to reconstruct the entire anterograde IFT train with the core area of IFT‐B1 below 1 nm resolution. They also employed Alphafold2, but only for monomers or oligomers with three or fewer molecules, following assembly of the entire IFT train based on the cryo‐ET map. The same structural features as seen in the works of Lorentzen's group are shared by Pigino's work. Furthermore, interaction between IFT‐B and IFT‐A and between repeating IFT‐B complexes in the array (Fig 1B) is visualized.

The most intriguing feature from Lorentzen's model concerns IFT81/IFT74 that bind to IFT22/IFT25/IFT27. Previous binding assays have supported a model, where the C‐termini of IFT81/IFT74 bind to C‐termini of IFT46/IFT52 (Taschner et al, 2014) (Fig 1C). The cross‐linking MS data can be explained by two mutually exclusive models. In the other model, N‐termini of IFT81/IFT74 bind to IFT88/IFT70. The model built by the Pigino group has a similar arrangement as the second model. The Lorentzen group interpreted that two conformations coexist in cilia, either by flexibility or by systematic change between two conformations – for example in anterograde and in retrograde IFT. Further work is needed to resolve, but both studies demonstrate the great potential of the used methods for structural biology at cellular scale.

The EMBO Journal (2023) 42: e113010