In one of our recent posts, we touched upon the intrinsically disordered proteins and regions (IDPs/IDRs, I will use IDP abbreviation in further text). There we explained how their specific primary structure and the choice of very hydrophilic, charged, and small amino acids enables them to exist in an extended conformation thereby defying one of the main paradigms of molecular biology – in order for a protein to be functional, it needs to exist in the most stable fold (called also the native fold). Today we know that not to always be the case, as we have characterized many functions of IDPs, even though they lack globular structure.
Since IDPs exist in this extended, stretched conformation with no distinct secondary or tertiary structure, a huge portion of their surface is exposed to the solvent, meaning that IDPs posses a high entropic energy. However, as soon as they bind to their target protein (and thereby form weak intermolecular bonds and interactions), their entropic energy decreases following the decrease in their exposed surface. As always, nature loves the equilibrium, so this drop in the entropy can’t be ignored – we need to find a way to compensate for this loss so that the overall energy of the system remains more or less the same. IDPs realized that if they follow up this entropy loss with some enthalpy gain, things could work out and nature won’t be forced to abandon IDPs. Once bound, IDPs start forming some sort of a secondary structure by creating intermolecular bonds and interactions. So the entropy that we lost because of the weak intermolecular bonds that were created between the IDP and the target protein is being balanced with the enthalpy gain due to the formation of intramolecular bonds within the IDP. This process, in which IDPs exit as completely disordered and then become ordered when bound to the target protein is called folding upon binding and it’s one of the coolest mechanisms in biochemistry (in my humble opinion at least).
Life is all about balance
In most cases, folding upon binding is energetically favorable for the majority of IDPs, but not everything is as simple as we would want it to be. If the entropic penalty (or decrease) is well compensated with enthalpic gain (in translation: if enough non-covalent bonds are created within the IDP and between IDP and target protein), a strong binding with a high affinity is achieved. However, since only basic secondary structure is being induced in the majority of IDPs (for example simple alpha-helix or beta strand), the number of those newly formed interactions is theoretically limited, meaning that the enthalpic contributions are usually not enough to counteract the entropic penalty and therefore the binding is usually of lower affinity. This is where the story gets messy, but don’t worry, nature has a few tricks up its sleeve…
Firstly, not all IDPs fold upon binding. Even though we said that folding is energetically more favorable than having the highly exposed surface of the extended IDP conformation, this would be the case in a perfect scenario where the gain in enthalpy is sufficient to cover for the loss of entropy. As explained, those newly formed bonds are usually not that strong and therefore we can’t always count that the energies will end add up in the end. Because of this a substantial amount of IDPs retain their disorder even when bound to the binding partner and they never really fold. We call those are “fuzzy” IDPs and they create “fuzzy” complexes with their binding partners where the “fuzziness” depends on the degree of disorder that IDP retains, which is also based upon the relative entropy inherent to the complex (the fuzziness helps to minimize the entropic penalty caused by binding of IDP and the binding partner, which would otherwise be too high to compensate with the increase in enthalpic energy of a newly formed complex, in which case there would be no binding whatsoever). The main characteristic of fuzzy complexes is that the interactions between the IDP and the binding partner protein are transient and coupled with a weak binding affinity (they come and go fast).
Ways to feel fuzzy
The fuzziness can be divided into 4 classes. It’s important to note that they are not mutually exclusive and the disorder at the bound state is a continuum, meaning that at any point in time a protein complex may exhibit the character of more than one class.
- Polymorphic model : In this case the IDP adopts distinct well-defined extended conformations in the bound state, while remaining fully disordered. The 3D conformations of IDP and the protein that are a part of this fuzzy complex are unrelated, meaning that we can’t find traces of lock-key binding mechanism or induced-fit mechanism here.

2. Clamp model: Clamp model is possible when we are dealing with a protein that has two structured globular regions connected with one intrinsically disordered region (IDR) that serves as a linker segment in this case. Two structured regions bind following a regular lock-key mechanism of binding, while the function of the IDR is just to link them. Because IDR is very flexible, it enables different 3D positions of the two structured regions, meaning that they can easily adapt to the binding protein surface.

3. Flanking model : This mechanism is similar to the clamp model. The difference is that in the clamp model the only disordered part is the linker segment, while the two domains that actually bind to the binding partner are structured (globular/ordered). In the case of a flanking model, the two domains that bind to the binding partner are disordered, as well as the linker segment. The two (or more) domains that bind to the binding partner are SLiMs (short linear motifs) that undergo folding upon binding. Because the linker segment is disordered, it retains its conformational freedom and therefore reduces the entropic penalty caused by binding of the SLiMs to the binding partner. Importantly, the flanking regions modulate the nature of the interaction between the SLiMs and the binding partner by contributing to the affinity and specificity of the interaction (and by “keeping the binding sites safe”).

Flanking model is very interesting because there is a folding upon binding and static disorder upon binding present in the same IDP chain. We can further describe the modes of binding within the flanking model as: avidity and allovalency. In short, avidity is a mechanism of binding which entails two or more binding sites present on an IDP (so two or more SLiMs connected with a linker segment), that are complementary to two or more binding sites present on its target protein (remember here that in the case of a polymorphic model the IDP conformations didn’t need to be complementary to the binding partner 3D structure). In order to achieve avidity, it’s important that the number of binding sites on the IDP is equal to the number of binding sites on a binding partner and at no point in time should the contacts interchange. An advantage of this binding mechanism is that once the initial contact is made, a cooperative effect takes place, facilitating further interactions. Moreover, the increased local concentration from the first binding event, combined with the lowered entropic penalty when binding only one IDP results in a greater binding affinity and the interaction is more thermodynamically favorable.
Conversely, allovalency models a system where there are multiple binding sites in tandem on an IDP that are complementary to a single binding site on its partner. Once the first contact is made, the rest of the complementary binding sites of the IDP start competing for this spot so that even when the first contact dissociates (due to low binding affinity which we discussed earlier), anoher complementary binding spot od an IDP is close enough to form a contatc. In this way, some part of an IDP is always bound to the binding partner, but the binding spots on the IDP are exchaning relatively quickly.

Back to the binding modes, finally we have
4. Random model: The random fuzzy model describes a situation where both the IDP and its target protein have a number of interaction sites and the interaction between each site is not restricted by any specificity. The random model works similar to the allovalency: first one nonspecific low affinity binding takes place between one binding site on the IDP and one on the binding partner protein. Even though this interaction dissolves quickly, because of the large number of putative binding sites in a close proximity, the binding is resumed even quicker. Random fuzziness represents the extreme case of disorder where there is little to no secondary structure induced upon binding of the IDP to a binding partner and the IDP retains a high degree of conformational freedom, making this model hard to characterise (both experimentaly and computationaly).

To be continued
Even though it might seem like we have it all figured out with the classes and groups we formed in an attempt to describe and understand the wild nature of IDPs when they interact with other proteins, the reality might not be so optimistic. We are still far from fully understanding these complex mechanisms, let alone predicting this kind of behaviour, but with the growing interest in this very special group of proteins, we will surely soon be in a much better position. I will do my best to cover a part of that amazing journey here so stay with us for more stories about the world of the mischievous IDPs 🙂
References
- Miskei M, Antal C, Fuxreiter M. FuzDB: database of fuzzy complexes, a tool to develop stochastic structure-function relationships for protein complexes and higher-order assemblies. Nuc Acid Res. (2017) 4;45(D1):D228-D235.
- Morris O, Torpey J, Isaacscon RL. Intrinsically disordered proteins: Modes of binding with emphasis on disordered domains. Open Biol. (2021) 11: 210222. https://doi.org/10.1098/rsob.210222
- Tompa P, Fuxreiter M. Fuzzy complexes: polymorphism and structural disorder in protein–protein interactions. Trends Biochem Sci. (2008) 33:1 (2-8). https://doi.org/10.1016/j.tibs.2007.10.003
- Wikipedia. Fuzzy complex. https://en.wikipedia.org/wiki/Fuzzy_complex (date of usage: 1.5.2023.)