Phosphoregulation of tropomyosin-actin interaction revealed using a genetic code expansion strategy

Tropomyosins are coiled-coil proteins that regulate the stability and / or function of actin cytoskeleton in muscle and non-muscle cells through direct binding of actin filaments. Recently, using the fission yeast, we discovered a new mechanism by which phosphorylation of serine 125 of tropomyosin (Cdc8), reduced its affinity for actin filaments thereby providing access for the actin severing protein Adf1/Cofilin to actin filaments causing instability of actin filaments. Here we use a genetic code expansion strategy to directly examine this conclusion. We produced in Escherichia coli Cdc8-tropomyosin bearing a phosphate group on Serine-125 (Cdc8 PS125), using an orthogonal tRNA-tRNA synthetase pair that directly incorporates phosphoserine into proteins in response to a UAG codon in the corresponding mRNA. We show using total internal reflection (TIRF) microscopy that, whereas E.coli produced Cdc8 PS125 does not bind actin filaments, Cdc8 PS125 incubated with lambda phosphatase binds actin filaments. This work directly demonstrates that a phosphate moiety present on serine 125 leads to decreased affinity of Cdc8-tropomyosin for actin filaments. We also extend the work to demonstrate the usefulness of the genetic code expansion approach in imaging actin cytoskeletal components.


Introduction
The actin cytoskeleton plays a vast array of physiological roles, ranging from cell morphogenesis and cell division to developmental pattern formation and muscle contraction (Cheffings et al., 2016;Pollard & Wu, 2010;Szent-Gyorgyi, 2004). The coiled-coil protein tropomyosin is a key actin filament binding protein, which regulates actin cytoskeletal architecture and function (Gunning et al., 2015;Khaitlina, 2015;Longley, 1975). In its well-characterized role in muscle contraction, tropomyosin regulates the interaction between the motor protein myosin II and filamentous actin (F-actin), in a calcium and troponin dependent manner (Chalovich et al., 1981;Gergely, 1974;Khaitlina, 2015;Perry, 2001;Spudich & Watt, 1971;Szent-Gyorgyi, 1975). Recently, we have shown that in non-muscle cells tropomyosin function is regulated in part by phosphorylation (Palani et al., 2019). We have shown that fission yeast Cdc8-tropomyosin is phosphorylated on Serine-125 in vivo and that a phosphomimetic mutant protein (Cdc8S125E) shows reduced affinity for F-actin in sedimentation assays as well as in TIRF microscopy-based assays. We also showed that incubation of unphosphorylated Cdc8 with its kinase, Pom1 and adenosine triphosphate (ATP), caused release of pre-existing Cdc8-tropomyosin from actin filaments. We proposed that Cdc8-tropomyosin protected actin filaments from Adf1/cofilin mediated severing, and that phosphorylation exposed actin filaments for severing by Adf1/ cofilin, thereby providing a mechanism for actin filament turnover (Palani et al., 2019). While these in vitro experiments (using best available current strategies) strongly pointed to phosphoregulation of Cdc8-tropomyosin interaction with actin, they still left open some caveats, since the Cdc8 proteins used in sedimentation and / or TIRF assays were not singly phosphorylated on S125, but were either phosphomimetic or present in complex mixtures containing the kinase and ATP. Here using a genetic code expansion strategy (Neumann, 2012; Neumann-Staubitz & Neumann, 2016; Wang, 2017), we generate Cdc8 that bears a single phosphate group on Serine 125 and provide direct evidence for regulation of Cdc8-tropomyosin interaction with actin filaments via Serine 125 phosphorylation.

Results
To firmly establish the role of phosphorylation of serine-125 on Cdc8-tropomyosin on actin binding, we used a genetic code expansion strategy to produce Cdc8 PS125 ( Figure 1A). In this strategy, we used an orthogonal tRNA-tRNA synthetase pair from Methanocaldococcus jannaschii (tRNA Cys ) and Methanococcus maripaludis aminoacyl tRNA synthetase for O-phosphoserine (P Ser ) (Pirman et al., 2015). Further, the anticodon loop in tRNA Cys was altered such that it would base-pair with the amber codon. This system for expression of P Ser bearing proteins has been pioneered by Soll and colleagues (Park et al., 2011) and further improved by Rinehart and colleagues (which we have used in this study) (Pirman et al., 2015). To produce Cdc8 bearing P Ser at position 125, we made an E. coli expression construct in which the codon for Serine-125 was replaced with an amber (UAG) codon. This construct was expressed in an engineered E. coli devoid of UAG-codons, and expressing the orthogonal tRNA sep -tRNA synthetase pair for P Ser and also bearing a mutation in translational elongation factor (EF-Tu) to facilitate incorporation in response to an UAG codon (Pirman et al., 2015). We also replaced the codon for Aspartic Acid position 142 with a codon for Cysteine as described previously to facilitate fluorescent labelling of the recombinant Cdc8 (Christensen et al., 2017). We established that the presence of a Cysteine residue at position 142 did not impair Cdc8 function, since Cdc8D142C was able to rescue a cdc8-110 mutant for colony formation at the restrictive temperature of 36ºC ( Figure 1B (Palani, 2020a)). Cdc8 and Cdc8 PS125 expressed in E. coli were labelled with Atto-565 ( Figure 1C; top panel (Palani, 2020b)). Although polyclonal antibodies against Cdc8 recognized both proteins ( Figure 1C; middle panel (Palani, 2020b)), an antibody against phosphorylated RXXS (which we have shown previously recognizes Cdc8 PS125 ), only detected the Cdc8 PS125 produced in E. coli and did not recognize unphosphorylated Cdc8 produced in E. coli ( Figure 1C; bottom panel (Palani, 2020b)). Note that all Cdc8-tropomyosins expressed in E. coli had an N-terminal acetylation mimicking sequence, as described previously (Christensen et al., 2017; Skoumpla et al., 2007), due to the importance of acetylation in tropomyosin function. We then tested the ability of Cdc8 PS125 to bind actin filaments in TIRF assays that we have described previously (Palani et al., 2019). Consistent with our previous experiments with Cdc8S125E, 0.3µM Cdc8 PS125 failed to bind actin filaments ( Figure 1D and E (Palani, 2020c)) (Palani et al., 2019). Importantly, treatment of Cdc8 PS125 with λ-phosphatase allowed its binding to actin filaments, again consistent with previous conclusions that Cdc8-bound actin filaments more efficiently in a Serine-125 unphosphorylated state ( Figure 1D and E (Palani, 2020c)).
Given the strength of the genetic code expansion approach, we attempted to further its use by introducing a fluorescent label on Cdc8 to facilitate its imaging by total internal reflection fluorescence microscopy (TIRFM). For this purpose, we used a tRNA-tRNA synthetase pair that introduced azido-phenylalanine (AzF) in response to an amber-UAG codon ( . Purified Cdc8-I76AzF was reacted with a strainedalkyne coupled to Alexa-647 in an azide-alkyne cycloaddition reaction, which generated a fluorescently labelled Cdc8 ( Figure 1F and G (Palani, 2020d)). Cdc8 76AF647 was tested for its ability to bind actin filaments using TIRFM by mixing it with F-actin. In this assay we found that Cdc8 76AF647 strongly bound actin filaments ( Figure 1H (Palani, 2020e)).

Conclusions
In this work we have used genetic code expansion to unequivocally establish that interaction between Cdc8-tropomyosin and actin is inhibited by the presence of a phosphate group on Serine-125.  investigations of phosphoregulation of the actin cytoskeleton in vitro. In other work reported herein, we have fluorescentlylabelled Cdc8 tropomyosin using a combination of genetic code expansion and azide-alkyne click chemistry. This method should be broadly applicable and should facilitate introduction of the fluorochrome into proteins, both for TIRFM as well as for homo-FRET experiments. In particular, this approach circumvents difficulties caused by an inability to generate single-cysteine bearing proteins, a routine approach used in protein labelling. We conclude that the expanded genetic code can be used as a powerful tool to further investigate the actin cytoskeleton and its post-translational modifications.

Unnatural amino acid incorporation, protein purification and labelling
For phospho-serine (pSer) incorporation: To genetically encode phosphorylated Cdc8, pGEX-ASCdc8 was used to introduce the amber codon (TAG) at position S125. Phosphorylated version of Cdc8 PS125 was expressed in C321. ∆A; ∆serB or BL21 ∆serB cells
Images were acquired using a Nikon Eclipse Ti-E/B microscope equipped with perfect focus system, a Ti-E TIRF illuminator (CW laser lines: 488nm, 561nm and 640nm) and a Zyla sCMOS 4.2 camera (Andor, Oxford Instruments, UK) controlled by Andor iQ3 software.
Quantification of F-actin decoration by Cdc8 PS125 -D142C For the generation of the actin loading graphs (Fig. 1D), the Cdc8PS125-D142C decoration length and F-actin length were measured manually for each filament in ImageJ (version 1.52p; NIH, USA) and their ratio was computed and plotted.

Statistical analysis
Data was plotted as box plots depicting individual data points, the mean values (black lines) and standard deviation (whiskers) using Graph pad Prism version 6. I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.