Non-conventional octameric structure of C-phycocyanin

C-phycocyanin (CPC), a blue pigment protein, is an indispensable component of giant phycobilisomes, which are light-harvesting antenna complexes in cyanobacteria that transfer energy efficiently to photosystems I and II. X-ray crystallographic and electron microscopy (EM) analyses have revealed the structure of CPC to be a closed toroidal hexamer by assembling two trimers. In this study, the structural characterization of non-conventional octameric CPC is reported for the first time. Analyses of the crystal and cryogenic EM structures of the native CPC from filamentous thermophilic cyanobacterium Thermoleptolyngbya sp. O–77 unexpectedly illustrated the coexistence of conventional hexamer and novel octamer. In addition, an unusual dimeric state, observed via analytical ultracentrifugation, was postulated to be a key intermediate structure in the assemble of the previously unobserved octamer. These observations provide new insights into the assembly processes of CPCs and the mechanism of energy transfer in the light-harvesting complexes.

solutions prepared by dissolving single crystals of TlCPC-6 or TlCPC-8 into 10 mM potassium phosphate buffer (pH 7.0). Note that the concentrations of these solutions (0.057 mg/mL for hexamer, 0.047 mg/mL for octamer), which were determined by measuring UV-vis spectra, were low enough to disassemble into monomers. b Temperature dependences of CD at 222 nm. The concentrations of the solutions were as follows; 0.034 mg/mL for hexamer, 0.027 mg/mL for octamer. The thermal denaturation midpoints were calculated as 71°C for both solutions.
obtain the final cryo-EM map of TlCPC-6. The particle images of TlCPC-6 and TlCPC-8 are extracted from the same micrographs, and thus, the processing steps with red texts are the same as TlCPC-8 ( Supplementary Fig. 17). Please see the "Details of the cryo-EM data processing in Supplementary   Fig. 15-18" section in the Supporting information for details. obtain the final cryo-EM map of TlCPC-8. The particle images of TlCPC-6 and TlCPC-8 are extracted from the same micrographs, and thus, the processing steps with red texts are the same as TlCPC-6 ( Supplementary Fig. 15). Please see the "Details of the cryo-EM data processing in Supplementary

Supplementary Table 1 Summary of the crystallization conditions of TlCPC.
Entry Space group    Supplementary Fig. 15-18.

Supplementary Note 1 Details of the cryo-EM data processing in
First, the movie fractions were aligned, dose-weighted, and averaged using MotionCor2 on 5 × 5 tiled fractions with a B-factor of 300 S1 . The non-weighted movie sums were used for Contrast Transfer Function (CTF) estimation with the Gctf program S2 . The dose-weighted sums were used for all subsequent steps of image processing. The subsequent processes of particle picking, two-dimensional (2D) classification, ab initio reconstruction, three-dimensional (3D) classification, 3D refinement, CTF refinement, and Bayesian polishing were performed using RELION-3.0 S3 . Initially, 1,903 particles were manually picked and performed 2D classification for preparing a 2D reference of template-matching-based auto-pick by RELION-3. From the first 91 micrographs, 19,367 particles were automatically picked and extracted while rescaling to 2.64 Å/pixel with 100-pixel box size. The extracted particle images were subjected to the second reference-free 2D classification (200 expected classes, 200 Å mask diameter). The resultants show small and large particle images, which were used for the subsequent analyses of TlCPC-6 and TlCPC-8, respectively.
For TlCPC-6, the 9,601 particles corresponding to the best 22 classes, which had around 100 Å diameter and displayed secondary-structural elements, were selected from the result of the second reference-free 2D classification, and then used for ab initio reconstruction (asymmetry, single expected class, 180 Å mask diameter). D3 symmetry was imposed on the generated ab initio map, which was used as an initial 3D reference for the 3D refinements (D3 symmetry, 160 Å mask diameter). The refined volume and particle images were rescaled to 0.88 Å/pixel with a 300-pixel box size and used for the 3D refinement (D3 symmetry, 160 Å mask diameter). The generated 3D refined map was used for 3D classification (4 expected classes, 160 Å mask diameter). The 3D volume and 5,192 particles of the best 3D class, which displayed the highest resolution, were used for the subsequent 3D refinements (D3 symmetry, 160 Å mask diameter).
The generated 3D refined map was used as a 3D reference for template-matching-based auto-pick and a stack of 283,980 particle images was extracted from the 2,036 dose-weighted sum micrographs while rescaling to 2.64 Å/pixel with 100-pixel box size. The images were subjected to two consecutive runs of reference-free 2D classification (200 expected classes, 160 Å mask diameter). Next, the 158,571 particles corresponding to the best 9 classes, which displayed secondary-structural elements, were selected for ab initio reconstruction (asymmetry, single expected class, 180 Å mask diameter).
D3 symmetry was imposed on the generated ab initio map, which was used as an initial 3D reference for the 3D classification with D3 symmetry (4 expected classes, 160 Å mask diameter). The 3D volume and 72,655 particles of the best 3D class, which displayed the highest resolution, were used as an initial 3D reference for the 3D refinements (D3 symmetry, 160 Å mask diameter). The generated 3D refined map was rescaled to 0.88 Å/pixel with a 400-pixel box size and used for the subsequent 3D refinements (D3 symmetry, 240 Å mask diameter). The 3D refined map was used for no-alignment 3D classification with D3 symmetry (2 expected classes, 240 Å mask diameter). The 3D volume and 28,120 particles of the best 3D class, which displayed the highest resolution, were used for the subsequent 3D refinements (D3 symmetry, 240 Å mask diameter). The cycle of CTF refinement and Bayesian polishing was repeated two times. The 3D refinement (D3 symmetry, 240 Å mask diameter) with a soft-edged 3D mask (15-pixel extension, 30-pixel soft cosine edge) was executed after each Bayesian polishing step.
For TlCPC-8, the 903 particles corresponding to the best 6 classes, which had around 130 Å diameter and displayed secondary-structural elements, were selected from the result of the second reference-free 2D classification, and then used for ab initio reconstruction (asymmetry, single expected class, 220 Å mask diameter). D4 symmetry was imposed on the generated ab initio map, which was used as an initial 3D reference for the 3D refinements (D4 symmetry, 200 Å mask diameter). The refined volume and particle images were rescaled to 0.88 Å/pixel with a 300-pixel box size and used for the 3D refinement (D4 symmetry, 200 Å mask diameter). The generated 3D refined map was used for 3D classification (2 expected classes, 200 Å mask diameter). The 3D volume and 786 particles of the best 3D class, which displayed the highest resolution, were used for the subsequent 3D refinements (D4 symmetry, 200 Å mask diameter).
The generated 3D refined map was used as a 3D reference for template-matching-based auto-pick and a stack of 129,653 particle images was extracted from the 2,036 dose-weighted sum micrographs while rescaling to 2.64 Å/pixel with 100-pixel box size. The images were subjected to two consecutive runs of reference-free 2D classification (200 expected classes, 200 Å mask diameter). Next, the 23,643 particles corresponding to the best 13 classes, which displayed secondary-structural elements, were selected for ab initio reconstruction (asymmetry, single expected class, 220 Å mask diameter). D4