C-phycocyanin as a highly attractive model system in protein crystallography: unique crystallization properties and packing-diversity screening

C-phycocyanin, a photosynthetic antenna protein from the cyanobacterium Thermosynechococcus elongatus, was crystallized in hundreds of different conditions, producing protein crystals with various sizes and morphologies. Among the many diffraction data sets that were collected, high-resolution X-ray structures with novel symmetries were solved and are discussed with the aim of providing a highly adaptable experimental model system.

Table S2 Indexing parameters on data collected from crystals larger than 70 µm.
The protein was in 20 mM TRIS pH 8.0, 100 mM NaCl.The crystallization screens used is mentioned on the first column and the number from A1 to H12 corresponds to the drops on a 96 well plate MRC2.The crystallization experiments were set up manually with the mixing of 1 µl protein and 1 µl precipitant.All the plates were stored at 20 °C unless is mentioned otherwise.The crystals were randomly picked and froze in 25% PEG400 when cryoprotectant was necessary.
No Screen SPG Table S3 Indexing parameters on data collected from crystals larger than 70 µm.
The protein was in 20 mM MES pH 6.5 100 mM NaCl.The crystallization screens used is mentioned on the first column and the number from A1 to H12 corresponds to the drops on a 96 well plate MRC2.The crystallization experiments were set up manually with the mixing of 1 µl protein and 1 µl precipitant.All the plates were stored at 20 °C unless is mentioned otherwise.The crystals were randomly picked and froze in 25% PEG 400 when cryoprotectant was necessary.
No Screen SPG

C
-Phycocyanin microcrystals appear to have an advantage as a model protein for easy visualization in serial crystallography experiments.As shown in figureS1A and B, the crystal density on a fix target chip can be easily adjust.The microcrystals shown in (A, B) were grown with the batch method and appear in few hours.

Figure
Figure S1 (a and b) Microcrystalline material loaded on a silicon chip (Lieske et al., 2019) and used in a serial crystallography setup.The natural blue color makes the crystal density on the chip easily adjusted.Microcrystals diffracted up to 2.3 Å resolution (Meents et al., 2017).The diameter of a single hole is 30 µm and crystals do not exceed 10 µm diameter.(c) Self-assembled crystalline material of larger particle size (formed in solution in acetate buffer pH4) loaded onto an XtalTool-HT (Feiler et al., 2019) and data were collected with the serial crystallography approach.(d) Application of C-PC microcrystals for serial crystallography on silicon chip directly at the beam.The mother liquid and nanocrystal are blotted before the data collection.

Figure
Figure S2 (a) Images of randomly picked crystals in UV-TEF and SONICC.The images show that the SHG signal is not enhanced by the chromophore present in the protein therefore there is not a false positive signal during imaging.(b) Examples of droplets were two crystal sizes are appearing, therefore in these cases, we included both sizes in the statistics shown in Figure 2(a).

Figure S3 Figure S4
Figure S3 Images of another three 96 well plates in visible (left) and the SHG imaging mode (right) in continuation of figure 3. The details on the plates are described in table 3 (protein buffered at pH 6.5) and the text.To examine the possibility of a false positive SHG signal due to the presence of the chromophore in C-PC or a higher symmetry space group, UV-TEF imaging was additionally utilized, which is based on intrinsic tryptophan fluorescence (see examples in figure S2).

Figure S5
Figure S5 The size distribution of crystals as they appear under different conditions.Naturally, within one crystallization drop, crystal sizes may vary, the results are determined by the observations of the majority of crystals.In some cases, when the crystals have two distinct size regimes, both are included, see examples in figure S2.(b) Three categories of crystal morphology in the C-PC crystallization experiments utilizing three different screens as shown in table 1, with C-PC in Tris buffer at pH 8.0.Please note that the morphologies are reported for crystals bigger than 10 µm.

Figure S6 AFigure S7
Figure S6 A total number of 118 datasets were analysed.The data were categorized into six different resolution classes.The cumulative percentage of individual categories is reflected.The number of individual datasets populating the different maximum resolution bins in Å is provided.

Figure S8
Figure S8 Calculated Surface areas are plotted for each observed space group and indicate the solvent-accessible protein surface area of individual assemblies.The buried area depicts the solventaccessible surface area of monomeric units buried upon the assembly of hexameric or dodecameric structures.All surface areas are calculated in square Å.

Figure
Figure S9 (a) Superposition of all structural models.The average Cα rmsd calculated to 0.57 Å is slightly higher than the structural coordinate error.All models are individually coloured and the positions of the phycocyanobilin cofactor are numbered through-out all panels.(b) A magnified stick representation of the cofactor and its binding region in cartoon representation is shown for each of the three ligand molecules.(c) The ligand binding pockets were calculated and are shown for each structure is superposition with the occupying ligand in stick representation.The dotted line in the middle panel indicates a 2.8 Å difference in the ligand binding position.

Table S1
Crystal structures of C-phycocyanin from cyanobacteria.

Table S4
Structural comparison of all superposed protein models.The RMSDs are plotted against the different models and calculated on the bases 4958 atoms.