Protein synthesis and post-translational modification

August 2024, Model of the Month by Dr Eva-Maria Geissen, Henrik M. Hammarén’s group, EMBL, Germany

Introduction

Post-translational modifications (PTMs), such as the addition and removal of chemical groups, influence various functions and properties of proteins, including proteolytic stability. Pulsed stable isotope labelling of amino acids in cell culture (pSILAC) allows us to assess protein turnover, i.e. the dynamics of the degradation of unlabelled (old) proteins and replacement by labelled (new) proteins [1,2]. Turnover data have traditionally been interpreted in terms of protein proteolytic stability, with a higher turnover (i.e. faster replacement) indicating lower protein stability. When combined with PTM-specific enrichment, pSILAC allows for assessing the replacement dynamics of modified and unmodified proteforms separately. Intuitively, the interpretation of differences in turnover as differences in stability has been extended to data from these PTM-resolved experiments [3,4]. However, for these data, the measured dynamics are a consequence not only of degradation and synthesis of the protein, but also of the interconversion dynamics of modified and unmodified proteoforms.

While the theory underlying pSILAC is well understood when measuring the turnover of proteins that exist in only a single form, no framework exists for interpreting pSILAC data, for the case of multiple, interconvertible proteoforms. To aid the interpretation of the observed dynamics, we model the biological process in the context of pSILAC experiments based on knowledge of the underlying process of PTM addition and removal. With the help of the model we demonstrate that for interconvertible protein species, like proteoforms defined by phosphorylation, metabolic labelling does not provide information on differences in protein stability as previously suggested based on intuition. Rather, these data suggest hypotheses on the relative order of PTM addition and removal along a protein’s lifetime [5].

Model

The model encodes the general biological processes of protein synthesis, reversible post-translational modification and degradation of a protein, put in the specific context of pSILAC experiments. This results in a two species reaction rate model of the dynamics of the old (unlabelled) protein (Figure 1a). In the model, a protein is synthesised as the unmodified species and subsequently reversibly modified into the modified species. Both species are subject to degradation. Different scenarios for the influence of a PTM on protein stability are encoded in the relative magnitude of species-specific degradation rate constants, with PTM degrons having a higher degradation rate constant for the modified species than for the unmodified species and vice versa for PTM stabilons.

The initial conditions are available as functions of all model parameters and correspond to the steady state of the model for a non-zero value of the synthesis rate constant. The model observables correspond to 1) the modified species normalised by its initial condition and 2) the sum of both species (i.e. the entire protein pool) normalised by the sum of their initial conditions, with the synthesis rate set to zero. The observables represent the fraction of modified proteins and the entire protein pool that have not been replaced yet. For analysis we use the natural logarithms of the observables and term them ‘clearance profiles’.

Figure 1: a) Kinetic scheme of the model. b) Typical clearance profiles for different parameter scenarios. In addition to the profiles of the observables the profile for the unmodified species is shown in blue.

Result

The comparison of simulated clearance profiles for PTM degron and stabilon parameter scenarios reveals that in either scenario the curve of the clearance profile of the modified species lies above the curve of clearance profile of the entire protein pool (Figure 1b), i.e. the modified species is always replaced slower than the entire protein pool. It can be shown analytically that this holds in this model for all possible parameter values. Therefore, the relative order of experimental clearance profiles does not allow to infer the effect of a PTM on protein stability. On the other hand, the simplest model that agrees with pSILAC data showing modified peptides disappearing faster than the entire protein pool, is a model with the reverse relative order of the two species, i.e. with the protein being synthesised in the modified form. Biologically this would correspond to a cotranslational modification of the protein that can reversibly be removed later on. In conclusion, there is no direct information about the effect of a PTM on protein stability contained in the order of the clearance profiles, in contrast to classic turnover experiments. Rather, our model shows that the order of clearance profiles allows for hypotheses about the order of PTM events relative to the time protein synthesis.

We also assessed expected characteristics of profiles when our model represents the underlying biological process. For this we analytically derived the intercepts of the extension of the linear part of the clearance profiles with the y-axis. These intercepts quantify the curvature of a profile before it becomes linear and the maximum vertical distance between any profiles. Systematically assessing the influence of the parameters on the profiles allowed for the conclusion that very different biological scenarios would lead to very similar characteristics in the experimental data (Figure 2). The curvature of the clearance profile of the entire protein pool is indicative for the stability effects of a modification (degrons concave down; stabilons concave up), but most parameter sets lead to linear profiles, even in cases of stability effects.

Figure 2: Curvature of the clearance profile of the entire protein pool for varying parameters, assessed via the intercept of the linear part of the clearance profile with the y-axis.

Discussion

Our model prevents the intuitive misinterpretation of proteomics data by enabling the interpretation of these data in a way that is impossible without a model. When a protein exists in multiple proteoforms, such as when measuring unmodified and modified proteoforms of the same protein, the biological interpretation of pSILAC data becomes non-trivial. We demonstrate that for interconvertible proteoforms, metabolic labelling does not provide direct information on differences in protein stability as previously suggested based on intuition alone. Measured clearance profiles rather tie into the relative order of modifications along the lifetime of a protein, with early modification exhibiting faster clearance than later ones.

References

[1] Schwanhäusser, B. et al. Global quantification of mammalian gene expression control. Nature 473, 337–342 (2011).

[2] Mathieson, T. et al. Systematic analysis of protein turnover in primary cells. Nat. Commun. 9, 689 (2018).

[3] Wu, C., Ba, Q., Lu, D., Zhou, H. & Fornasiero, E. F. Global and site- specific effect of phosphorylation on protein turnover global and site-specific effect of phosphorylation on protein turnover. Developmental Cell 56, 111–124 (2021).

[4] Zecha, J. et al. Linking post-translational modifications and protein turnover by site-resolved protein turnover profiling. Nat. Commun. 13, 165 (2022).

[5] Hammarén, H.M., Geissen, EM., Potel, C.M. et al. Protein-Peptide Turnover Profiling reveals the order of PTM addition and removal during protein maturation. Nat Commun 13, 7431 (2022).