Ethoxylate Polymer-Based 96-Well Screen for Protein Crystallization

: Crystallization is the limiting step in X-ray structure determination of biological macro-molecules. As crystallization experiments can be largely automatized, the diversity of precipitant solutions is often the determinant factor to obtain crystals of high quality. Here, we introduce a 96-well screening kit of crystallization conditions, centered on three ethoxylate-based organic polymers as precipitants and various additional compounds to promote crystal formation. This crystallization screen was tested on various non-standard proteins from bacteria and archaea. Structure determination succeeded for seven out of thirteen targets based on crystals that frequently diffracted to a higher resolution than those obtained with commercially available screening kits. Crystallization hits were rarely similar among the three ethoxylate-based organic polymers and, in comparison, with already available crystallization screens. Hence, the presented crystallization screen is an efﬁcient tool to complement other screens and increase the likelihood of growing crystals suitable for X-ray structure determination.


Introduction
Although challenged by single-particle cryo-electron microscopy (cryo-EM [1]) and structure-predicting artificial intelligence approaches (Alphafold2 [2]), X-ray crystallography is still the most widely used and productive method in structural biology [3].In particular, it is an appropriate tool for the structural characterization of protein-substrate complexes, sufficiently long-living intermediates, turnover processes in a time-resolved manner at room temperature (via X-ray free-electron laser methods), for identification and localization of metals in proteins and for high-throughput screening of a broad palette of compounds relevant for drug development due to rapid synchrotron data collection and full-automatic data processing protocols.On the other hand, X-ray crystallography suffers from different bottlenecks, the major one being the time-consuming trial-and-error process to compel a protein into a highly ordered crystalline form, which is indispensable for X-ray diffraction experiments.Together with the properties of a protein with respect to its size, purity, homogeneity, flexibility and stability, the composition of the precipitate solution is crucial for the success of crystallization.Since the first crystallization screen developed by Jancarik and Kim in 1991 [4], sophisticated screenings with a wide range of compounds have been continuously developed and are often commercially available [5][6][7][8].The automatization of the screening process and the miniaturization of crystallization experiments using robots allow the handling of a large number of crystallization conditions, save time and protein and thereby increase the success rate.The physicochemical properties of precipitant solutions supplemented with protein are decisively governed by the central precipitant.Precipitants are usually categorized into inorganic salts, organic solvents and organic polymers.Polyethylene glycols (PEGs) are the most applied agents [9], but alternative organic polymers are also powerful candidates [7,[10][11][12] whose potential is not completely worked out yet.The efficiency of the precipitate solution is significantly increased using a wide range of additional agents, for example, chaotropic and kosmotropic salts, amino acids, alcohols, polyamines, sugars, detergents, reducing agents and specific bound compounds like regulators or substrates, supplemented in small amounts [13][14][15].Finally, crystallization behavior is influenced by pH, which is adjusted using a buffer.
Thanks to our long-term experience, we developed a 96-well crystallization screen which is based on three soluble ethoxylate polymers as precipitants, various additives and buffers.The crystallization screen was tested on thirteen prokaryotic proteins that embody a wide range of characteristics regarding their origin, size, flexibility, purification strategies and cofactor content.Its success rate was in the range of other applied screens without indicating any apparent redundancy.The obtained crystals were of sufficient quality to determine structures without further improvement steps.In some cases, the crystals were the best diffracting specimens or even the only ones when compared with commercially available screens performed in parallel.

Crystallization Method and X-ray Diffraction Measurements
Crystallization was performed using the sitting-drop method with the protein and precipitant solutions listed in Table 1.Attempts under anaerobic conditions were achieved using an Oryx Nano pipetting robot (Douglas Instruments, Hungerford, UK) or manually with SWISSCI MRC 2-well plates and drop sizes of 300 + 300 or 600 + 600 nl, respectively (BCR-QONP, PCD, DCCP, CaCODH/ACS).For oxygen-insensitive proteins, crystallization screens were set up with a Phoenix Rigaku robot system using CrystalMation Intelli-plates with low profile (Art Robbins, Sunnyvale, CA, USA) and a drop size of 100 + 100 nl (RosC, S25-PSL, S25-PT, CML, CH1enDH) or manually with a drop size of 600 + 600 nl (MsKARI, MsFruA, MsMch and MsMer).Crystallization experiments were normally performed at a temperature of 18-20 • C, while for RosC, S25-PSL, S25-PT, CML and CH1enDH, additional experiments were performed at 4 • C. Obtained crystals were either frozen directly or after supplementing with a cryoprotectant.Their X-ray diffraction patterns were recorded at the Swiss-Light synchrotron source PXII (SLS, Villigen, Switzerland) and the Deutsches Elektronen-Synchrotron (DESY, Hamburg, Germany).

The Composition of the Crystallization Screen
Organic polymers have proven to be valuable precipitants for protein crystallization [7,[10][11][12].They reduce the protein solubility through modifying the local ionic strength [28], decreasing the masking of proteins and/or neutralizing their charges.The interactions between a protein and its surrounding bulk solvent environment are lowered [13,29], and the magnitude of hydrophobic and hydrogen-bond forces among protein molecules is enhanced.In addition, the altered viscosity influences the kinetics of crystal growth.Preliminary experiments involving lysozyme as the test object pointed toward the organic polymers glycerol ethoxylate (GE 1000), trimethylolpropane ethoxylate (TMPE 1014) and glycerol ethoxylate-co-propoxylate triol (GEPT 2600) (Figure 1) as promising precipitants, which became the major components of the 96-well screening kit (Table 1).Trimethylolpropane ethoxylates with lower molecular masses, such as TMPE 170 and TMPE 450, were also tested for their use as precipitants, but they turned out to be less efficient.
As additional chemical compounds, we used different salts (MgCl 2 , CaCl 2 , KCl, (NH 4 ) 2 SO 4 and KSCN) that are well established to support crystal formation [30], as well as a set of small molecules chosen because of their favorable influence on protein stability and crystal nucleation.The cryoprotectant glycerol was used for its role in stabilizing proteins, suppressing their flexibility and aggregation and reducing their nucleation rate [31][32][33].The additive propane-1,2-diol is used as a stabilizer [34] and cryoprotectant in cryo-crystallography as an alternative to glycerol or ethylene glycol [35,36].L-Glu and L-Arg (50 mM each) had a synergistic effect on the solubility of the tested proteins [37], and L-Arg was found to prevent aggregation and increase protein stability [38,39].The disaccharide trehalose (Figure 1) was reported to enhance the nucleation of both lysozyme and thaumatin [40].The osmolyte trimethylamine N-oxide (TMAO, Figure 1) protects intracellular components against osmotic stress and hydrostatic pressure [41] and increases the stability and nucleation properties of proteins [42].The amino acid ectoine (2-methyl-1,4,5,6-tetrahydropyrimidine-4-carboxylic acid, Figure 1) [43] protects proteins from stress factors like heat, freezing, drying or mechanical forces [44], suppresses aggregation [45] and might increase the crystal size [46].Finally, NDSB-256 (Figure 1), a non-detergent zwitterionic sulfobetaine, stabilizes proteins, reduces their aggregation and favors the growth of ordered crystals [47,48].Three buffers at different pH (MES buffer pH 6.5, HEPES buffer pH 7.5 and TRIS buffer pH 8.5) were the final variables of the crystallization screen (Table 1).The pH measurements of the entire crystallization solution indicated ∆pH deviations from the buffer pH of up to 0.2 for GE 1000 and GEPT 2600 as well as about 0.5 for TMPE 1014.Precipitants, additives and buffers were mixed in a matrix of 96 different crystallization conditions, which were tested on different non-standard proteins to determine their practical suitability.

Application to Proteins
The quality of the ethoxylate-derived screening kit was assessed using a set of thirteen proteins with a wide range of different properties.The protein targets varied in size between 25 and 445 kDa with oligomeric state between dimers and dodecamers.They were natively extracted or heterologously expressed from mesophilic and thermophilic organisms belonging to archaea and bacteria.Some targets contained only the polypeptide, while others contained additional cofactors and individual substrates.Purification and crystallization were attempted under aerobic or anaerobic conditions.The success rate of the thirteen proteins applied was in the range found for the PACT++ and JBS pentaerythritol screens, tested for nine and eight proteins.A summary of each project is given in Table 2.For seven of the tested proteins (BCR-QONP, CODH/ACS, DCCP, S25-PT, RosC-E107Q, Ms-FruA and MsMch), the crystal quality was sufficient for structure determination (Figure 2, Table 2).For some projects, the crystals from the ethoxylate-based screening kit diffracted to the highest resolution.For one project (S25-PT), only the ethoxylate-based screening kit produced suitable X-ray crystals (Table 2).Furthermore, the overall efficiency may have been underestimated, as optimization was stopped when any of the applied kits provided high-quality crystals of the desired protein or protein state.Altogether, crystallization of the thirteen proteins succeeded at least once in sixty-three out of the ninety-six conditions.
The three precipitants provided suitable crystallization conditions in comparable amounts and rarely yielded identical hits under the same buffer and additive conditions.Crystal vitrification for data collection is usually straightforward, as their concentrations have to be increased to 45%-50%, further underscoring the utility of the three precipitants.An analysis of the impact of every additive for successful crystallization puts the salts and the L-Glu/L-Arg mixture at the top.According to limited statistics, the number of crystallization hits is significantly higher at pH values of 6.5 or 8.5 than at neutral pH, which perhaps reflects the extremophilic origin of some of the tested proteins and their accompanying protein stability and rigidity at non-neutral pH values.The statistically most successful solution was, however, obtained at a near-neutral pH with a composition of 35% GEPT 2600, 0.1 M HEPES, pH 7.5 and 0.2 M (NH 4 ) 2 SO 4 , resulting in five hits.For the applied proteins, the propensity to crystallize did not appear to be influenced by size but rather by the optimal growth temperature of the host organism (Table 2).The crystallization rate of proteins that originate from thermophilic organisms (M.thermoacetica and M. shengliensis, optimal growth temperatures of 50 • C) is higher than that from mesophilic species (Table 2).However, it is worth mentioning that a larger fraction of the thermophilic proteins is prepared from natural sources and not through heterologous expression.The impact of this factor on crystallization has not yet been explored.Our results again demonstrate the importance of using multiple crystallization screens.The newly developed ethoxylate-derived screen (Table 1) exhibited a high crystallization success rate and resulted in many crystals with good diffraction, which allowed the structure determination of several non-standard and never-characterized proteins.The ethoxylate-based screening kit is approximately as beneficial as other established commercial screening kits without apparent overlap.It is therefore a suitable tool to complement other kits to maximize the crystallization success of proteins, thereby indicating that the search for new precipitants for crystallization is still possible and worthwhile.

Figure 1 .
Figure 1.Chemical structures of the precipitants and some additives.

Table 1 .
Composition of the ethoxylate-based screen.
* All percentages are given as weight per volume (w/v), except for those indicated.

Table 2 .
Results of the tested proteins (x: few crystals, xx: several crystals, xxx: many crystals).