Developing novel fusion tags to facilitate protein production and structure determination

Maurer, Sigrun (2021) Developing novel fusion tags to facilitate protein production and structure determination. PhD thesis, University of Nottingham.

[img] PDF (Final corrected version after viva) (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Available under Licence Creative Commons Attribution.
Download (27MB)


Covalently fused protein tags can be a helpful tool in protein structure determination by X-ray crystallography and Cryo-electron microscopy. Carrier-driven crystallisation can aid protein structure determination, but the currently available selection of crystallisation tags is limited. Extant tags aim to increase solubility, provide additional crystal contacts, and can help in model building by molecular replacement. In addition, scaffold tags for electron microscopy (EM) techniques can help visualisation and structure determination, when target proteins would otherwise be of insufficient size for EM approaches.

6 candidate tags, three for crystallisation, two for crystallisation and EM, and one for EM, were selected based on a Protein Data Bank (PDB) search for specific attributes, including size, accessibility of termini, and available high-resolution structures. Ribosomal protein L7Ae (archaeal, PDB: 1xbi) and adenylate kinase (bacterial, PDB: 4qbg) were selected as crystallisation tag candidates for N-terminal fusion to target proteins. Ribosomal protein L30e (archaeal, PDB: 1w41) was selected for its potential as a crystallisation tag to be inserted into loop regions. Orotidine monophosphate decarboxylase (archaeal, PDB: 3wjz) and trans-esterase HcgF (archaeal, PDB: 3wva) were selected as dimeric crystallisation tag candidates and as potential electron microscopy scaffolds. Isopentenyl-diphosphate isomerase (archaeal, PDB: 2zru) was selected as a tetrameric tag candidate, specifically for EM applications.

All 5 novel crystallisation tag candidates were purified and evaluated for yield, solubility and crystallisability. The candidates were subjected to the same three commercial crystallisation screens, and the thermal stability of the tags was determined by ThermoFluor assays. The best crystallisability was shown by 1w41 and His6-1xbi, with multiple crystallisation hits for each candidate except 3wva. A 1w41 structure was solved to a resolution of 1.08 Å, and a His6-1xbi structure to 2 Å. 4qbg showed high solubility. ThermoFluor assays showed very high melting temperatures for 1w41, His6-1xbi and His6-3wva, consistent with attributes of their thermophilic organisms of origin.

The model protein enhanced yellow fluorescent protein (eYFP) was tagged with all crystallisation tag candidates for purification and crystallisation trials, and the 2zru tetramer for EM approaches. In addition, a loop insertion fusion construct of eYFP with another crystallisation tag candidate, the receiver domain of social mobility protein FrzS (PDB: 2gkg), was created. All eYFP-crystallisation tag fusion constructs, as well as His-tagged eYFP, were purified and evaluated for yield, solubility and crystallisability by subjecting them to the same 3 commercial crystallisation screens. Fusions of eYFP-His6 to the EM scaffold tag 2zru and the dimeric tag 3wva were investigated for their suitability for visualisation by negative staining and cryogenic transmission electron microscopy (Cryo-EM) methods. eYFP-(1w41)-His6, 1xbi-eYFP-His6 and 3wva-eYFP-His6 were successfully crystallised, and atomic models of the protein fusions were solved by molecular replacement. An eYFP-(1w41)-His6 loop insertion structure was determined to a resolution of 2.08 Å, a 1xbi-eYFP-His6 structure was determined to a resolution of 1.54 Å and a dimeric 3wva-eYFP-His6 structure was determined to a resolution of 3.11 Å. 3wjz-eYFP-His6 and eYFP-(2gkg)-His6 did not yield diffraction-quality crystals and showed low solubility, and 4qbg-eYFP-His6 did not yield any crystals while showing high expression and solubility. Both 2zru-eYFP-His6 and 3wva-eYFP-His6 showed promise for EM applications, and two-dimensional (2D) class averages and an initial three-dimensional (3D) model were obtained for both fusions from negative stain data. Initial Cryo-EM data was obtained for 2zru-eYFP-His6. Further optimisations are necessary to collect improved Cryo-EM datasets.

The crystallisation tag candidates 1xbi, 1w41, 3wva, 2gkg and another crystallisation tag candidate, the beta-spectrin calponin homology domain (PDB: 1bkr) were further fused to the human Ubiquitin-specific protease 11 (USP11) catalytic core, for which no structural information was yet available. Purifications and crystallisation trials were performed for all fusion constructs, with the 1w41 and 2gkg loop insertion fusions showing high purification yields (~8mg and ~7 mg from 1 L bacterial culture on average respectively), while fusion with the dimeric tag 3wva resulted in very low purification yields (~0.06 mg from 1 L bacterial culture on average). Active site mutants of USP11 in complex with minimal substrate mimic ubiquitin-3G (Ub-3G) were also generated, purified, and subjected to crystallisation. The fusion construct complex USP11-(2gkg)-His6 C318S + Ub-3G crystallised successfully and the crystals diffracted well to a resolution of 2.4 Å. The novel structure of the USP11 catalytic core was subsequently solved and isothermal titration calorimetry (ITC) and enzyme activity assays were performed. This was the first catalytic core structure of a USP in the DUSP-UBL family that was solved in complex with a substrate. The structure will benefit further research into the function of this important cancer target. Fusions of the USP11 catalytic core and USP4 full-length (FL) with EM-scaffold tag candidates 3wva and 2zru were investigated for EM approaches, but did not result in more than low-resolution negative staining class averages.

In conclusion, 7 crystallisation tag candidates and two electron microscopy scaffold candidates were investigated for their suitability to aid protein structure determination efforts. In total, 6 tag constructs, 10 eYFP constructs, 16 USP11 catalytic core constructs and 2 USP4 FL complexes were generated, purified, and subjected to either crystallisation trials or EM experiments. Two tags were crystallised successfully without cargo protein, and three structures of eYFP with different novel crystallisation tags were also obtained. The principal result of this project is solving the USP11 catalytic core structure with help of the novel loop insertion crystallisation tag candidate 2gkg, and in complex with Ub-3G. The EM-scaffold results are promising at the stage of negative staining but will require further work to eventually obtain high-resolution Cryo-EM structures. Further research will include tagging relevant proteins of unknown structure, including other ubiquitin-specific proteases, with the most promising tag candidates identified in this project (i.e. 1w41, 2gkg, 1xbi, 2zru and 3wva).

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Dreveny, Ingrid
Keywords: fusion tags, proteins, crystallisation, protein structure determination
Subjects: Q Science > QD Chemistry > QD901 Crystallography
Q Science > QP Physiology > QP501 Animal biochemistry
Faculties/Schools: UK Campuses > Faculty of Science > School of Pharmacy
Item ID: 65498
Depositing User: Maurer, Sigrun
Date Deposited: 04 Aug 2021 04:42
Last Modified: 04 Aug 2023 04:30

Actions (Archive Staff Only)

Edit View Edit View