Figure 1.
The flowchart of system designs for predicting and characterizing bioluminescent proteins (BLPs).
The SCMBLP method predicts BLPs and the SCMLFP method distinguishes luciferases from fluorescent proteins. The obtained scoring cards are used to further analyze BLPs.
Figure 2.
Heat map of the propensity scores of 400 dipeptides obtained from the SCMBLP method.
Color bar is obtained using Jet of Matlab.
Table 1.
The propensity scores of amino acids to be a bioluminescent protein (BLP) and amino acid composition (%) using BLP-TRN.
Figure 3.
The histogram of propensity scores for BLPs and non-BLPs in the training dataset BLP-TRN.
(A) Sequence scores before optimization (B) Optimized sequence scores.
Table 2.
The comparisons of SCMBLP with some comparable methods using BLP-TRN in terms of 5- or 10-fold cross-validation accuracy (%).
Table 3.
The test performance (%) of SCMBLP on various data sets.
Table 4.
The propensity scores of amino acids to be a bioluminescent protein (BLP) and four physicochemical properties of amino acids with Pearson’s correlation coefficient (the R value).
Figure 4.
Distribution of dipeptide scores on two typical sequences of BLPs.
The BLP with Pfam ID A8QZJ6 (length 96 with domain Oxidored shown in red) and the non-BLP with NCBI sequence entry GI: 20137223 (length 101 shown in blue) have high and low sequence scores 551.65 and 256.86, respectively.
Table 5.
The compositions of BLPs, integral membrane proteins, and nuclear proteins, and their corresponding R values using the training dataset BLP-TRN.
Figure 5.
The correlation coefficients between the propensity scores and various physicochemical properties of 20 amino acids.
(A) Transfer free energy (B) Average membrane preference (C) Hydrophobicity scale from native protein structures (D) Composition of amino acids in nuclear proteins.
Figure 6.
Heat map of the propensity scores of 400 dipeptides obtained from the SCMLFP method.
Color bar is obtained using Jet of Matlab.
Table 6.
The propensity scores of amino acids for distinguishing luciferases from FPs.
Table 7.
The performance (%) of SCMLFP and the compared SVM-based methods on the dataset consisting of 269 luciferases and 216 FPs.
Figure 7.
Bioluminescence system of the jellyfish Clytia with the GFP-clytin complex.
(A) Clytin is a Ca2+-regulated photoprotein comprising coelenterazine (CZ) as a substrate for its bioluminescence reaction. When Ca2+ binds to clytin, the bioluminescence reaction is triggered to produce the excited state product coelenteramide (CA) and CO2 with emission of a broad blue bioluminescence (maximum 470 nm). Due to energy transfer, Clytia GFP fluorescence is in green (maximum 500 nm). (B) The structure of clytin (22.4 kDa, PDB code 3KPX) and its representation using hydrophilic (outer part in blue) and hydrophobic (inner part) residues. (C) The structure of Clytia GFP (PDB code 2HPW) and its representation using hydrophilic and hydrophobic residues using Discovery Studio 3.5. The visible fluorophore of Clytia GFP (CSY) is a sequence of three amino acids (68S, 69Y, and 70G). To protect the chromophore fluorescence from quenching by water, the tightly packed nature of the barrel excludes solvent molecules resulting in that the environment comprising this fluorophore is also hydrophobic.
Figure 8.
Schematic presentation of the G protein-coupled receptor (GPCR) reporter cell line for drug discovery.
The drug discovery of high-throughput screening is monitored by activation of the calcium-sensitive photoprotein light production. Ligand-mediated activation of Gs-coupled receptors stimulates cAMP synthesis by adenylyl cyclase (AC) and opening of the cAMP-gated heteromultimeric cyclic nucleotide-gated (CNG) channel. Calcium ions from the extracellular solution enter the cell through the CNG channel and are detected by photoprotein luminescence measurements. Gq-coupled receptor activation stimulates the phospholipase C (PLC)/IP3 pathway detected via photoprotein luminescence stimulated by calcium released from the endoplasmic reticulum (ER).