Date Published: March 01, 2020
Publisher: International Union of Crystallography
Author(s): Adam D. Crawshaw, Arnaud Baslé, Paula S. Salgado.
A practical perspective on molecular-replacement (MR) structure-determination pipelines is presented, using PilA1 from Clostridioides difficile as an example of a difficult case. A manual approach informed by the biology of the system under study is described, together with ab initio MR structure determination.
Once complete diffraction data at an appropriate resolution (usually at least around 3–3.5 Å) have been recorded from a protein crystal sample, the next challenge is to solve the phase problem (Taylor, 2003 ▸). Phases are required to calculate an electron-density map, from which a model of the structure is built. These are missing from diffraction data experiments and can be experimentally derived using techniques such as single/multiple anomalous dispersion (SAD/MAD) or single/multiple isomorphous replacement (SIRAS/MIRAS) (Taylor, 2010 ▸; Wang et al., 2014 ▸). Alternatively, phases can be calculated using similar structures that have already been determined, which is known as molecular replacement (MR). Here, we review an outline of a number of different strategies (Fig. 1 ▸) starting with a single sweep of X-ray diffraction data to calculate an electron-density map using molecular replacement. We use the structure determination of the major pilin in type IV pili (TFP), PilA1, from Clostridioides difficile as a working example.
Here, we outline a simple manual approach to molecular replacement (Fig. 1 ▸) as initially attempted to determine the structure of PilA1.
The structure of R20291 PilA1Δ1–34 (Fig. 4 ▸) is typical of the TFP pilins, containing a long N-terminal α-helix (α1) which is linked by an α–β loop to an antiparallel β-sheet (β1–β2–β3) that encapsulates the D-region containing a shorter α-helix (α2), with the rest of this subdomain largely formed of loop regions. As there are no cysteine residues in PilA1, it lacks the disulfide bridge that delimits the D-region of Gram-negative pilins.
Despite collecting high-resolution data for R20291 PilA1Δ1–34, it was not possible to calculate phases and an interpretable electron-density map using our manually edited search models based on known pilin structures. As the native data set for R20291 PilA1Δ1–34 had high resolution (1.65 Å), high completeness, low Rmeas and high I/σ(I), it was an ideal candidate for ab initio phasing using ARCIMBOLDO (Millán et al., 2015 ▸). As our example shows, using a careful approach to molecular-replacement pipelines and exploring available options, combined with prior knowledge of both the biology and structural characteristics of the protein of interest, can lead to solution of the phase problem and the determination of novel protein structures.