Folding and Unfolding Mechanism of Highly Stable Full-Consensus Ankyrin Repeat Proteins

Abstract

Full-consensus designed ankyrin repeat proteins were designed with one to six identical repeats flanked by capping repeats. These proteins express well in Escherichia coli as soluble monomers. Compared to our previously described designed ankyrin repeat protein library, randomized positions have now been fixed according to sequence statistics and structural considerations. Their stability increases with length and is even higher than that of library members, and those with more than three internal repeats are resistant to denaturation by boiling or guanidine hydrochloride. Full denaturation requires their heating in 5 M guanidine hydrochloride. The folding and unfolding kinetics of the proteins with up to three internal repeats were analyzed, as the other proteins could not be denatured. Folding is monophasic, with a rate that is nearly identical for all proteins (∼ 400–800 s^− 1), indicating that essentially the same transition state must be crossed, possibly the folding of a single repeat. In contrast, the unfolding rate decreases by a factor of about 10⁴ with increasing repeat number, directly reflecting thermodynamic stability in these extraordinarily slow denaturation rates. The number of unfolding phases also increases with repeat number. We analyzed the folding thermodynamics and kinetics both by classical two-state and three-state cooperative models and by an Ising-like model, where repeats are considered as two-state folding units that can be stabilized by interacting with their folded nearest neighbors. This Ising model globally describes both equilibrium and kinetic data very well and allows for a detailed explanation of the ankyrin repeat protein folding mechanism.

Introduction

Repeat protein architecture does not rely on interactions between residues that are distant in sequence (long-range interactions), but stabilizing and structure-determining interactions are formed within a repeat and between neighboring repeats. This special feature, the modular nature of repeat proteins, makes them fundamentally different from globular proteins and, thus, interesting for testing experimental and theoretical views that have emerged from the study of globular proteins. Moreover, since repeat proteins are the only class of proteins that can be extended in size while still constituting a contiguous domain, unique questions about how folding and stability change with the number of repeats can be asked.

Repeat proteins constitute, next to immunoglobulins, the most abundant natural protein classes specialized in binding.1, 2 Ankyrin repeat (AR) proteins consist of repeating structural units (repeats) that stack together to form elongated nonglobular repeat domains. The AR is one of the most common protein sequence motifs. This 33-residue motif consists of a β-turn, followed by two antiparallel α-helices and a loop that reaches the turn of the next repeat³ (see Fig. 1a).

Stability and kinetic folding studies of mostly natural AR proteins have been performed. The tumor-suppressor protein p166, 7 unfolds in a sequential manner; first, both N-terminal repeats unfold, followed by the two C-terminal repeats. Molecular dynamics simulations have been carried out to study this in more detail.⁸ The tumor-suppressor protein p19 shows an equilibrium intermediate, as well as three folding phases.⁹ In a more detailed kinetic study, an on-pathway intermediate was detected, as well as a suggestion for its structure was made using NMR hydrogen/deuterium exchange.¹⁰ Similarly as observed for the p16 protein, both N-terminal repeats 1 and 2 unfolded first, while repeats 3–5 were still folded. When dissecting the Notch receptor ankyrin domain from Drosophila melanogaster,11, 12 constructs from four to seven repeats were made, in which multiple repeats were deleted from either end or both ends, resulting in the finding that stability increased with repeat number. The longest construct has been used for kinetic folding studies,13, 14 and it was found that refolding and unfolding kinetics are best described by a sum of two exponential phases. Equilibrium and kinetic studies were also conducted with myotrophin, a small four-repeat AR protein.15, 16 While the kinetics of the Notch ankyrin domain could be fitted by a sequential three-state model, myotrophin kinetics were assigned to a two-state model. Further analysis with single and double mutants showed that myotrophin follows parallel pathways, where folding is initiated either by the C-terminal repeat or by the N-terminal repeat.¹⁷

All these studies showed that the folding of AR proteins is not simply a cooperative process, but intermediate states do occur. However, they have all been carried out with natural proteins containing repeats of different sequences and stabilities. Hence, many results only describe the particular protein under study, and they can only partially and qualitatively test the effects of protein length on kinetics and thermodynamics. In addition, they give no indication on the intrinsic properties of the consensus AR.

We therefore intended to examine the folding and unfolding of designed ankyrin repeat proteins (DARPins) with identical repeats as a function of repeat number. The consensus AR represents an “average structure” of all of the natural ARs and will eliminate properties that only come about with particular sequences of individual repeats. The DARPins can thus be considered as generalized examples for the study of AR protein folding. By characterizing the thermodynamic and kinetic parameters of these consensus proteins using circular dichroism (CD) and fluorescence spectroscopy, the dependence of stability, as well as of folding and unfolding rate constants, on repeat number was investigated. Moreover, we intended to gain mechanistic insight into the folding pathway of the three smallest proteins consisting of three to five AR repeats.