Understanding Colour Tuning Rules and Predicting Absorption Wavelengths of Microbial Rhodopsins by Data-Driven Machine-Learning Approach

Abstract

The light-dependent ion-transport function of microbial rhodopsin has been widely used in optogenetics for optical control of neural activity. In order to increase the variety of rhodopsin proteins having a wide range of absorption wavelengths, the light absorption properties of various wild-type rhodopsins and their artificially mutated variants were investigated in the literature. Here, we demonstrate that a machine-learning-based (ML-based) data-driven approach is useful for understanding and predicting the light-absorption properties of microbial rhodopsin proteins. We constructed a database of 796 proteins consisting of microbial rhodopsin wildtypes and their variants. We then proposed an ML method that produces a statistical model describing the relationship between amino-acid sequences and absorption wavelengths and demonstrated that the fitted statistical model is useful for understanding colour tuning rules and predicting absorption wavelengths. By applying the ML method to the database, two residues that were not considered in previous studies are newly identified to be important to colour shift.

Exploration of natural red-shifted rhodopsins using a machine learning-based Bayesian experimental design

“…These were then introduced into E. coli cells, and the proteins expressed in the presence of 10 μM all-trans retinal. As a result, 40 E. coli cells showed significant colouring, indicating significant expression of folded protein, and their λ ≡ were determined by observing ultraviolet (UV)-visible absorption changes upon bleaching of the expressed rhodopsins through a hydrolysis reaction of their retinal with hydroxylamine, as previously reported 20 (Fig. 4).…”

“…Figure 3 shows a schematic of the updating procedure. First, we added 97 WT microbial rhodopsins and their variants for which the λ max had recently been reported in the literature or determined by our experiments, to a previously reported data set 20 . In other words, the new training data set consisted of the amino acid sequences and λ max of 893 WT microbial rhodopsins and their variants (Extended Data Table 1).…”

“…A significant portion (37.5%) of the genes with red-shifted absorption compared with the base wavelengths also had asparagine at this position (Extended Data Figure 2). The A-to-N mutation at this position had a smaller effect (4–7 nm) 20,35 than that of the A-to-S/T mutation; thus, the difference between alanine and asparagine is not so critical to explore red-shifted rhodopsins. Position 186 in BR is proline in most microbial rhodopsins (in 98.7% of the 3,064 candidate genes), and the mutation to non-proline amino acids induces red-shift of absorption 16 .…”

“…As an alternative approach, we recently introduced a data-driven machine learning (ML)- based approach 20 . In the previous study, we demonstrated how accurately the absorption wavelength of rhodopsins could be predicted based on the amino acid types on each position of the seven TM helices 20 . We constructed a database containing 796 wild-type (WT) rhodopsins and their variants, the λ max of which had been reported in earlier studies.…”

“…In this paper, we updated the ML-based model used in our previous study 20 so that it could properly compute expected red-shift gains and applied this new model to 3,064 ion-pumping rhodopsin candidates derived from archaeal and bacterial origins that can be easily expressed in Escherichia coli (Fig. 1b).…”