Intrinsically Disordered Proteins: The 'Shapeless' Molecules Scientists Ignored for 50 Years

Key Takeaways 

• A third of all human proteins have a disordered region (also called an intrinsically disordered region, or DR), about 5% of proteins are almost completely disordered and this is a significant component of biology not an exception. 
• A new IDR-ome map of ~20,000 human disordered regions show they cluster by conserved physical features and those clusters predict function, location and interaction partners. 
• Disordered regions play a key role in DNA/RNA binding, cell signalling, and biomolecular condensates formed through liquid-liquid phase separation, which are membrane less organelles that help cells organize reactions in a rapid and efficient way. 
• The mutations that affect these areas are associated with developmental disorders, certain cancers and neurodegenerative disease. Such a change has already started to emerge opportunities for new avenues of drug design, clinical genetics, and protein engineering.

Until the latter part of the twentieth century, molecular biologists operated on a very basic model that the protein folds into one shape and shape determines its role. Lysozyme is a wedge-shaped enzyme that breaks down bacterial walls. Haemoglobin has a pocket to which it binds and transports oxygen. Structure equals function. However, that is only half the tale, a significant portion of the human proteome does not have a fixed conformation. These regions are referred to as intrinsically disordered regions, and proteins consisting almost exclusively of these regions are called intrinsically disordered proteins, which are constantly changing shape. They were considered for decades as little more than biochemical noise, the floppy leftovers that interfered with real proteins. That reputation has not aged well.

Gordon Research Conferences. (2024). Intrinsically Disordered Proteins (GRS)

1. What Disordered Actually Means: Defining Intrinsically Disordered Regions

All proteins are initially formed as a chain of amino acids linked together in a sequence determined by a gene. Most chains fold in a tight conformation that is maintained by hydrogen bonding, hydrophobic packing, and charge interactions, and that remains relatively stable. It was around this time after 2018, that Alpha Fold developed the ability to predict protein shapes with a level of accuracy that had previously been impossible to achieve, thus finally solving the protein folding problem that had eluded biochemists for almost 50 years. A region of disorder does not bend like this. It does not take a single shape, but rather is more like a piece of wet spaghetti in water constantly changing shape in a great number of different forms and conformations without ever locking into a single fixed structure. About a third of human proteins have a significant disordered part and approximately 5% of the proteome is disordered nearly throughout its entire length.

Read More: Scientists Uncover the Genetic Connection Between Cannabis Use Disorder and Mental Health Illness

A few traits tend to set these regions apart from their folded neighbours: 

• Simpler amino acid composition: Rich in proline, glycine, and charged residues, and poor in the bulky non-polar blocks which promote tight folding. 
• Fast sequence drift: The sequence changes rapidly between species for a long time causing them to be viewed as evolutionarily unimportant. 
• Shape-shifting on contact: The same stretch can assume a variety of local shapes, depending on the molecule that it is in contact with at that instant.

 That last quality as it turns out, is the whole story.

2. Why This Was Overlooked for So Long: The Limits of Structural Biology

One reason for the historical neglect is reflecting the limitation of what tools are available. The two techniques, X-ray crystallography and Cryo-electron microscopy require a stable, reproducible shape to clearly image the structure. In a field of science that focuses on explaining function by solving the structure, disorder didn't seem like a worthy topic for study either rather it just appeared to be a technical problem. There was also a deeper assumption at work. Until now, researchers have used sequence conservation as their primary indicator of the biological significance of an important stretch, because mutations in these locations are typically detrimental to organisms. 

Read More: Remarkable Breakthrough: Man Unexpectedly Cured of HIV After Stem Cell Transplant

By that standard, disordered sequences were not well-lettered and by that measure they seemed insignificant. These sections are far from random but only when the measurement was changed from the actual letters to the bulk physical and chemical nature of the region.

3. How You Study Something That Won't Hold Still: Techniques for Studying Disordered Proteins

Studying disorder requires a different toolkit than studying folded proteins. Nuclear magnetic resonance reads out an ensemble of conformations in solution rather than one fixed shape. Small angle X-ray scattering provides information on overall size and shape without requiring a crystal. Single-molecule techniques monitor the movement of the distance between two labelled points as a function of time, that is, the tracking of wiggling in real time.

In addition, computational and machine-learning based methods, including AI-driven protein structure prediction tools like AlphaFold, are able to look for patterns in sequences in terms of composition, charge and binding motifs on a whole proteome scale which the wet-lab methods could not in themselves achieve. This computational approach was heavily used in a recent large-scale effort led mostly by a group at the University of Toronto that was published in the Proceedings of the National Academy of Sciences. 

Read More: The Role of Biological Aging in the Genetic Evolution of Cancer

Instead of attempting to find the exact sequence letter by letter which the scientists say is impossible with the rate of the exact sequence drifting so quickly, the researchers measured conserved molecular features which include net charge, spacing of charged residues, aromatic content, and predicted binding motifs. Two disordered regions, from different species can have almost no sequence similarity yet have almost the same physical fingerprint and this is the fingerprint the team used to build their map.

4. Inside the Human IDR-ome: Mapping Disordered Regions with Machine Learning

The resulting atlas, which researchers have begun to refer to as the human IDR-ome, compared some two thousand genes of disordered regions and clustered them according to their conserved physical properties, not just their sequence. Physical feature clustering also indicates clustering by function and by cellular location and by interaction partners. Disorder, in other words, is not structureless in any meaningful sense. It contains an implicit logic that no traditional sequence-alignment program was designed to detect, and this is a very useful predictive tool: if the disordered part of a poorly studied protein has a feature profile that aligns with that of a well-characterized cluster, there is a now a testable hypothesis about what the poorly studied protein might be doing.

        

5. What These Regions Actually Do: Functions of Intrinsically Disordered Proteins

These regions turn out to be doing a great deal. The largest functional categories in the map include: 

• Chromatin and DNA binding: reading and organizing genetic material 
• RNA metabolism: Processing and regulating RNA molecules 
• Cytoskeletal organization: Shaping and maintaining the cell's internal scaffolding 
• Cell signalling: Relaying messages between and within cells 
• Transmembrane transport: Moving molecules across membranes 
• Reproduction along with smaller but still meaningful slices tied to histone modification, cell shape, immune response, and nuclear pore function

Coincidence or not, the greatest categories all have to deal with a lot of different partners over the years, whether that's a transcription factor that has to recognize dozens of different DNA sequences or a splicing factor that has to put together and take apart molecular machinery on the fly. 

Read More:  A Sinister Brain Protein and the Deep Mystery of Life’s Origins

A rigid protein with a single shape is not such a good fit for that kind of quick, promiscuous engagement, while a protein that changes shape upon contact is much better. This also accounts for the fact that so many disordered proteins are not specialists, but rather multitaskers, that is, proteins with multiple disordered regions, some of which may come from entirely different functional clusters in the same protein.

6. Why Disorder Gravitates Toward Condensates: Disordered Proteins and Biomolecular Condensates

One of the more intriguing threads that weaves through this research is biomolecular condensates, the droplet-like compartments cells form to organize reactions without enclosing them with a membrane, a process known as liquid-liquid phase separation. Stress granules, which appear and vanish under the stress of the cell, nuclear speckles in which RNA processing occurs and the nucleolus in which the assembly of ribosomes takes place all arise and disappear in a way that is similar to the way oil beads up in water.

The proteins that populate such condensates are predominantly disordered, and the mechanics makes sense if worked out. This means that these regions are skilled at creating a large number of weak, transient interactions instead of forming a single strong bond, which is exactly what causes them to separate into distinct regions: a collection of molecules that are constantly forming and breaking many weak interactions will naturally aggregate into droplets and break apart again just as quickly. A folded domain functions like a key cut for one particular lock, a disordered, condensate-forming region functions more like Velcro, designed to catch unsuspecting things over and over again rather than capture and hold things for eternity which is exactly what a membrane without any of the "walls" needs to do if it is going to assemble in seconds and then fall apart again.

7. Modifications and Splicing: Post-Translational Modification of Disordered Regions

Disorder is not fixed permanently at the DNA level either. Many examples of such protein post-translational modifications after their synthesis have been identified and they are concentrated in disordered regions. Such tags can alter the binding state or charge of the region almost instantaneously making it possible for a cell to shift the function of a given stretch of protein depending on particular circumstances without needing to produce anything new. This pattern is also consistent with alternative splicing, as many of the disordered regions are located adjacent to exon boundaries. Splicing can introduce, delete or manipulate a disordered region altogether, altering what sorts of condensates a protein can interact with or what partners it can bind to and this allows cells to produce a lot of functional variety for a relatively small number of genes.

Read More Moringa: The Hidden Secret Behind Natural Energy, Beauty, and Healing

8. The Connection to Disease: Disordered Proteins in Cancer and Neurodegenerative Disease

These regions are critical in controlling various cellular functions, such as gene regulation, RNA handling, and cell signaling, so it's no wonder that disrupting them can present actual complications. There are several proteins involved in Alzheimer's and Parkinson's disease, both neurodegenerative diseases that have disordered regions that under the wrong conditions, can misfold into aggregation prone states. Cancer-causing transcription factors rely on disordered activation domains to inflict their damage. Mutations in the unstructured parts of chromatin-associated proteins are more frequent than would be predicted in some neurodevelopmental disorders. Because discovery of disease genes has traditionally been targeted to mutations within folded, well annotated domains, it is a reasonable assumption that damage in disordered regions had been overlooked for years, as it was much more difficult to interpret.

    

9. Where This Is Already Proving Useful: Drug Design and Protein Engineering Applications

This is quite practical: 

• Drugging the "undruggable": Numerous transcription factors and signaling proteins have long been deemed inaccessible to small molecule drugs due to their deep pockets. Small molecule drugs have long been considered untargetable because they prefer to target small pockets in proteins. Newer approaches focus on the short recognition motifs, and on the condensate-forming behavior, which are available as targets, which were not available before. 
• Better clinical judgment calls: Feature-based maps provide a better set of criteria for the clinician to consider when deciding of whether a newly discovered mutation in a disordered region is likely benign or truly pathogenic rather than labeling it "unknown significance. 
• Smarter protein engineering: For a synthetic biologist attempting to design an arbitrary, custom condensate or switchable protein assembly, it's now easy to copy an existing feature from a natural disordered region. 
• Sharper AI predictions: This type of feature information can be fed back into the structure prediction pipelines to mark areas where a single correct structure is unlikely to be the right answer to begin with.

10. Where the Field Goes from Here: The Future of Disordered Protein Research

The researchers behind the recent mapping effort see a few open directions that could drive the next few years of research, such as finding more refined ways of subdividing disorder into distinct functional categories, rather than lumping it all together; methods that would allow researchers to identify new and relevant molecular features without relying on human assumptions, using self-supervised approaches; and more mechanistic studies of how condensates achieve specificity in the first place, such as understanding why one disordered protein clusters in the nucleolus and another in a stress granule. The course of the whole thing is pretty obvious. The field is shifting its focus from the cataloging of disorder to predictions of disorder, and even towards the planned creation of disorder.

Read More:  The Hidden Secret to Strong Teeth: How Oral Probiotics Transform Your Dental Health Naturally

Conclusion:

Disorder turned out to be not a contradiction of function, but only a different way of functioning that these older tools could not see clearly. The image that is now emerging is a protein that does not have a fixed shape is as useful as a protein that does. With more and more mapping projects such as the IDR-ome filling in the details, these chaotic regions may soon find themselves in one of the main sections of the biochemistry textbooks and crucially in the developmental plans of new drugs.