Molecular replacement requires a search model with sufficient structural similarity to the target. Below what approximate sequence identity does molecular replacement typically fail?
ABelow 95% — only nearly identical structures work
BBelow approximately 25-30% sequence identity, because structural divergence becomes too great for the search model to provide useful phases
CMolecular replacement works at any sequence identity
DBelow 80% — high homology is always required
The success of molecular replacement depends on how closely the search model matches the target structure. Above 30-40% sequence identity, structures are generally similar enough that MR succeeds. Between 20-30%, success depends on the specific case — conserved core regions may be similar enough, but loops and insertions diverge. Below 20%, structural similarity is typically insufficient for MR, and experimental phasing (heavy atom or anomalous methods) is needed. The availability of AlphaFold predictions has extended MR to lower sequence identity targets by providing better search models than any single experimental structure.
Question 2 True / False
Single-wavelength anomalous dispersion (SAD) using selenomethionine-labeled protein has become the most common experimental phasing method because it requires only one data set from one crystal.
TTrue
FFalse
Answer: True
SAD phasing using selenomethionine (SeMet) has largely replaced MIR and MAD as the experimental phasing method of choice. Selenomethionine is incorporated biosynthetically by expressing the protein in methionine-auxotrophic bacteria grown on SeMet-supplemented media. The selenium atoms provide anomalous scattering signal at wavelengths near their absorption edge. SAD requires only one data set at one wavelength (near the selenium edge), while MAD requires data at multiple wavelengths. Modern computational methods (density modification, automated model building) can resolve the phase ambiguity inherent in SAD, making it robust enough that the additional wavelengths of MAD are usually unnecessary. The simplicity of SAD (one crystal, one data set) has made it the default experimental phasing approach.
Question 3 Short Answer
Explain why molecular replacement works — how can phases from a different (homologous) protein help solve the structure of the target protein?
Think about your answer, then reveal below.
Model answer: If a known structure is sufficiently similar to the target, their electron density distributions are similar. Placing the known structure in the correct position and orientation within the target's unit cell provides an approximate electron density map, from which approximate phases can be calculated. These phases are 'close enough' to the true phases that the resulting electron density map (computed with model phases and experimental amplitudes) reveals features of the target protein not present in the search model — new side chains, different loop conformations, bound ligands. Iterative refinement improves the phases and model simultaneously. The approach works because the phase information is dominated by the overall protein fold (which is conserved between homologs) rather than the sequence-specific details (which differ).
Molecular replacement's dominance reflects the growth of the PDB — with over 200,000 experimental structures and millions of AlphaFold predictions, the probability of finding a suitable search model for any new protein is high. The method is fast (minutes to hours) compared to experimental phasing (days to weeks), which is why it is tried first for every new crystal structure.