Repository logo

Towards macromolecular scaffold assisted crystallography




Huber, Thaddaus, author
Snow, Christopher, advisor
Ackerson, Christopher, committee member
Henry, Charles, committee member
Fisk, Nick, committee member

Journal Title

Journal ISSN

Volume Title


The current, dominant method for structure determination in atomic detail is X-ray crystallography; but, this method requires a brute force search through non-physiological solution conditions looking for the "needle-in-a-haystack" condition in which the target protein crystallizes. Unfortunately, despite exhaustive screening, most proteins of interest do not form crystals. Other proteins are difficult to obtain in sufficient quantities to make the attempt. Finally, even successful crystals reveal a structure adopted under artificial conditions-a single snapshot that dramatically underrepresents the protein mobility. The motivational insight for this work is the recognition that materials diffract X-rays if they consist of a highly-ordered, repeating lattice, but that the lattice need not be composed only of target protein. Instead of growing conventional protein crystals, we will take the unprecedented step of attaching target proteins (guests) to specific sites within pre-existing, crystalline scaffolds for a new technique called scaffold assisted crystallography. This approach circumvents the haphazard nucleation and growth process that underlies conventional crystallography. Instead, we face novel challenges. We must engineer scaffolds that have very large pores (>10 nm), withstand significant solution condition changes, yet still diffract to high resolution. We must also ensure that guest proteins tightly tethered to the crystalline scaffold adopt a coherent structure visible via X-ray diffraction. Instead of taking on the challenge of de novo design of porous protein crystals, we decided to search the protein databank for a suitable scaffold. Algorithms for identifying highly porous protein crystals are covered (Chapter 1) and a select few representative examples are presented. Constructs for high priority candidates were obtained and crystallization of the targets were attempted. One of the candidates crystallized rapidly and presented a platform for developing methods and identifying roadblocks for second generation scaffolds. Working extensively with a single scaffold, a putative periplasmic protein from Campylobacter jejuni (CJ), allowed for robust method development that enabled highly optimized expression and extensive knowledge of its crystallization space. CJ requires high salt for crystallization. Crystals quickly degrade outside of the growth conditions. Most guest macromolecules will have low solubility in the high salt required to preserve the CJ crystalline lattice. Therefore, methods for chemical crosslinking of CJ crystals were developed to withstand significant solution condition changes, yet still diffract to high resolution. The most ubiquitous crosslinking agent glutaraldehyde effectively stabilized the crystal, but resulted in a dramatic loss of diffraction. Three alternative crosslinkers, formaldehyde, glyoxal, and EDC, were tested for their ability to stabilize CJ crystals. The three alternative crosslinkers all stabilized CJ crystals in challenging conditions (no salt) with little degradation in diffraction quality. The crosslinked crystals were subjected to x-ray diffraction; the resulting electron density demonstrates the first known atomic resolution modifications from formaldehyde, glyoxal, or EDC crosslinks in a protein crystal. In contrast to the weak, noncovalent interactions that hold together typical protein crystals, guest domains can be attached to the host scaffold using strong interactions. For maximum programmability, affinity tags for the desired assembly can be genetically encoded on the guest and scaffold monomers. We demonstrated that non-covalent, metal-mediated capture and genetically encoded histidine tags provide a significant level of control. Loading and release of guest molecules were fine-tuned to spatially segregate multiple guest proteins. Similarly, by controlling the diffusion of crosslinking agents we engineered a crystalline shell that still diffracts well. Scaffold assisted crystallography techniques were demonstrated with small molecule guests in CJ crystals. Guest molecules were installed via a single covalent bond to reduce the conformational freedom and achieve high occupancy structures. We used four different conjugation strategies to attach guest molecules to three different cysteine sites within pre-existing protein crystals. In all but one case, the presence of the adduct was obvious in the electron density. The above methods led to preliminary attempts of scaffold assisted crystallography with macromolecules. Guest mini-proteins variants were obtained with solvent exposed cysteines. These were covalently attached in vitro to CJ with an engineered surface thiol. We attempted to crystallize the resulting CJ-mini-protein conjugates. One of the CJ-mini-protein conjugates crystallized and the structure was determined. While the presence of the guest mini-protein was obvious, the electron density past the attachment point was ambiguous. Still, this result demonstrates feasibility of fusing target proteins to engineered CJ monomers for "chaperoned crystallization". For targets that fail to crystallize when pre-installed, we can perform asynchronous crystallization and by attaching the guest mini-protein to a preformed CJ crystal. Techniques for in crystallo conjugation and quantification are developed. Finally, present strategies for realizing macromolecular scaffold assisted crystallography are presented.


Rights Access


structural biology


Associated Publications