Separating distinct structures of multiple macromolecular assemblies from cryo-EM projections.

Area of Science:

• Structural biology
• Biophysics
• Biochemistry

Background:

• Single particle analysis in cryo-electron microscopy (cryo-EM) typically requires highly purified samples.
• Current reconstruction algorithms struggle with heterogeneous mixtures of macromolecular complexes.
• Stringent purification can be labor-intensive and may lead to sample loss.

Purpose of the Study:

• To develop and validate a computational method for partitioning distinct macromolecular complexes from heterogeneous cryo-EM datasets.
• To demonstrate that analyzing common lines in 2D projection images is sufficient for sorting.
• To reduce the reliance on extensive sample purification for structural studies.

Main Methods:

• Utilized a graphical framework to relate 2D projection images by their common lines.
• Tested the approach on synthetic reprojections from 35 unique macromolecular structures.
• Applied the algorithm to experimental cryo-EM data from a mixture of five protein complexes.
• Employed existing methods for ab initio three-dimensional structure determination.

Main Results:

• Successfully demonstrated the feasibility of partitioning distinct structures from synthetic data.
• Showcased the algorithm's ability to sort a mixture of five protein complexes from experimental cryo-EM data.
• Enabled the ab initio determination of multiple 3D structures from a heterogeneous sample.

Conclusions:

• The developed method effectively sorts single particle cryo-EM data from extremely heterogeneous mixtures.
• This approach alleviates the need for stringent purification, making structural studies more accessible.
• Paves the way for investigating complex biological samples containing numerous unique structures.