Alpha-helices and beta-sheets are the two dominant shapes that a protein chain folds into on its way to becoming a functional molecule. Proposed in 1951 by Linus Pauling, Robert Corey, and Herman Branson, these structural motifs form the backbone scaffolding of tens of thousands of known proteins.1PubMed Central. The discovery of the alpha-helix and beta-sheet, the principal structural features of proteins Together they account for a large share of the ordered architecture inside every cell, and understanding how they form, what stabilizes them, and what happens when they go wrong explains a surprising amount about health, disease, and modern protein engineering.
What an Alpha-Helix Actually Looks Like
Picture a ribbon spiraling around a central axis like a spiral staircase. That is an alpha-helix. The protein backbone winds tightly so that every amino acid’s backbone oxygen atom forms a hydrogen bond with the backbone nitrogen atom four residues ahead in the sequence. This repeating i-to-i+4 hydrogen-bonding pattern locks the chain into a right-handed coil with roughly 3.6 amino acid residues per full turn. The side chains of each amino acid stick outward from the helix, like spokes on a wheel, leaving the hydrogen-bonded core compact and regular.
Crystallographic studies of helices in real proteins show that the average backbone angles for alpha-helices cluster around −62° and −41° for the two main rotational degrees of freedom along the chain.2PubMed. Helix geometry in proteins Those numbers matter mostly to structural biologists, but the takeaway for everyone else is that nature converges on a very narrow geometric sweet spot. Alpha-helices across wildly different proteins look almost identical in their backbone geometry; what differs is the identity of the side chains poking out, and that is what gives each protein its unique function.
What a Beta-Sheet Actually Looks Like
A beta-sheet forms when two or more stretches of the protein chain, called strands, line up side by side and hydrogen-bond to each other laterally. Instead of a coil, think of a pleated ribbon. The backbone zigzags, and the hydrogen bonds run between neighboring strands rather than within a single stretch. The strands can run in the same direction (parallel) or in opposite directions (antiparallel), and many real sheets contain a mix of both.
Beta-sheets tend to have a characteristic right-handed twist when viewed along the strand direction, and they can stack on top of one another through hydrophobic contacts between their faces, forming sandwich-like layered structures.3PubMed Central. The supramolecular chemistry of β-sheets Side chains from adjacent layers can interdigitate in a knob-and-hole fashion, adding further stability. This layered architecture shows up in many structural proteins, enzymes, and antibodies.
The Glue That Holds It All Together
Both alpha-helices and beta-sheets owe their stability primarily to hydrogen bonds between backbone atoms, not side chains. This is a point worth emphasizing: the backbone is the same for every amino acid, so the basic capacity to form these structures is universal to all protein chains. What the side chains do is modulate how easily a given stretch of backbone slips into one shape versus another.
In an alpha-helix, each hydrogen bond runs roughly parallel to the helix axis, connecting partners four residues apart. In a beta-sheet, the hydrogen bonds run roughly perpendicular to the strand direction, stitching neighboring strands together. The difference in hydrogen-bond geometry is why helices look like coils and sheets look like flat pleated surfaces. Beyond hydrogen bonds, van der Waals contacts, hydrophobic packing of side chains, and electrostatic interactions all contribute. In beta-sheets especially, the face-to-face stacking of hydrophobic side chains between layered sheets adds significant stabilization.3PubMed Central. The supramolecular chemistry of β-sheets
Why Some Amino Acids Prefer One Shape Over the Other
Not all amino acids are equally happy in a helix or a sheet. Alanine, for example, is the best helix-former. Experiments using synthetic peptides with systematic amino acid substitutions have ranked all 20 amino acids by their intrinsic preference for helical structure.4PubMed Central. Helix propensities of the amino acids measured in alanine-based peptides without helix-stabilizing side-chain interactions After alanine, uncharged glutamic acid is among the strongest helix formers, and interestingly, it is a substantially better helix promoter when its side chain is not carrying a charge.5Biophysical Journal. A Helix Propensity Scale Based on Experimental Studies of Peptides and Proteins The likely reason is that a charged carboxyl group can form competing hydrogen bonds with the backbone in the unfolded state, pulling the equilibrium away from helix formation.
Proline, by contrast, is a helix-breaker. Its rigid ring locks the backbone into an angle incompatible with the helical spiral, so stretches rich in proline tend to adopt extended or irregular conformations instead. Glycine is a mixed case: it has so much backbone flexibility that it rarely sits in the interior of a helix (it would be too floppy), but it excels in certain structural roles at the edges of helices and in tight turns.
The ends of a helix present a special problem. At the first and last turns, some backbone hydrogen-bond donors or acceptors lack partners because the helix has started or ended. Certain amino acids are particularly good at “capping” these exposed positions. At the N-terminus of a helix, asparagine is the best capping residue, because its side chain can accept a hydrogen bond from an otherwise unsatisfied backbone NH group. Aspartate, serine, threonine, and cysteine also perform well. At the C-terminal end, capping preferences are more uniform across amino acids.6PubMed Central. N- and C-capping preferences for all 20 amino acids in alpha-helical peptides
Turns and Loops Connect the Pieces
Proteins are not just helices and sheets. Between these regular elements, the chain has to reverse direction, and it does so through structures called turns and loops. Beta-turns, the most common type, involve four consecutive residues making a tight reversal of the chain, often connecting adjacent beta-strands in a sheet. Two classical types, Type I and Type II, have strikingly different amino acid preferences. Type I turns favor aspartate, asparagine, and serine at several positions, while Type II turns strongly prefer proline at the second position and glycine at the third.7PubMed. Analysis and prediction of the different types of beta-turn in proteins
More recent analysis using high-resolution crystal structures has refined this picture. The old “Type II” category actually encompasses at least two distinct conformational clusters that differ in whether glycine occupies position 3. In one cluster glycine dominates, while in the other asparagine and aspartate take its place.8PLOS Computational Biology. A new clustering and nomenclature for beta turns derived from high-resolution protein structures The practical upshot is that the classic textbook categories, while useful, are simplifications. Real protein turns are more varied than two neat boxes suggest.
Helical Variants Beyond the Standard Alpha-Helix
The alpha-helix is the most common helical form, but it is not the only one. The 3₁₀-helix has a tighter winding with hydrogen bonds running i to i+3 instead of i+4, and the pi-helix is a wider, looser version with i-to-i+5 bonding. Both are rarer than the alpha-helix. Simulations of helix-to-coil transitions have found that 3₁₀-helix regions tend to be fleeting, whereas pi-helical segments can appear in substantial populations during the unfolding and refolding process.9PubMed Central. The role of alpha-, 3(10)-, and pi-helix in helix–>coil transitions The backbone angles that define a 3₁₀-helix (averaging around −71° and −18°) differ noticeably from those of the alpha-helix, giving it a distinct geometric signature.2PubMed. Helix geometry in proteins
Collagen, the most abundant protein in the human body, uses a completely different helical strategy. Its triple-helix motif consists of three chains, each adopting a polyproline II-like conformation, wound around each other into a supercoil. The first high-resolution structure of a collagen triple helix confirmed this arrangement and revealed a highly ordered network of water molecules that depends on the presence of hydroxyproline, a modified amino acid almost unique to collagen.10PubMed. Protein motifs. 8. The triple-helix motif in proteins
When Helices Wrap Around Each Other
Individual alpha-helices are useful, but many functional structures in nature involve multiple helices coiling around one another to form what are called coiled coils. The hallmark of a coiled-coil sequence is a repeating pattern of seven amino acids, with hydrophobic residues occupying the first and fourth positions. Those hydrophobic side chains line up along one face of each helix and interlock with a partner helix, creating a stable rope-like assembly. When additional hydrophobic positions are introduced at the flanking positions, higher-order structures emerge. Researchers have created stable four-helix bundles by mutating charged residues in the GCN4 leucine zipper to nonpolar amino acids, yielding antiparallel four-stranded coiled coils with a combination of knobs-into-knobs and knobs-into-holes packing.11PubMed Central. Antiparallel four-stranded coiled coil specified by a 3-3-1 hydrophobic heptad repeat
The principle scales impressively. A seven-helix coiled coil has been characterized, demonstrating that heptad repeats containing four hydrophobic positions can drive the assembly of complex higher-order structures with diverse packing geometries.12PubMed Central. A seven-helix coiled coil Coiled coils are not just structural curiosities; they appear in muscle proteins, transcription factors, viral fusion machinery, and many other biological contexts where a sturdy, rod-like structural element is needed.
Helices, Sheets, and Cell Membranes
Cell membranes are oily barriers, and proteins that sit inside them face a unique challenge: their backbone hydrogen-bond donors and acceptors need to be satisfied even though there is no water around to help. Alpha-helices solve this neatly because all backbone hydrogen bonds are internal. This is why most transmembrane proteins thread their chains through the membrane as bundles of alpha-helices. The outer face of each helix, the side chains pointing toward the lipid, tends to be hydrophobic, while polar residues can line the interior of the bundle to create channels or binding sites.
A subtler point is that what drives a helix to insert into a membrane is not just raw hydrophobicity. Amphipathic helices, those with one hydrophobic face and one polar face, bind membranes with a strength that correlates linearly with their hydrophobic moment (a measure of how asymmetrically the hydrophobic side chains are distributed around the helix). Amphiphilicity turns out to be far more important for interfacial binding than simple overall hydrophobicity.13PubMed Central. Folding amphipathic helices into membranes: amphiphilicity trumps hydrophobicity
Beta-barrels offer the other membrane-spanning solution. In these proteins, beta-strands curve into a closed barrel, with hydrogen bonds satisfied between neighboring strands and hydrophobic side chains facing outward toward the lipid. Beta-barrel membrane proteins are common in the outer membranes of bacteria and mitochondria. The two strategies, helix bundle and beta-barrel, represent convergent solutions to the same thermodynamic problem: burying all backbone polar groups in a nonpolar environment.
When Secondary Structure Goes Wrong
The balance between alpha-helix and beta-sheet is not just an academic detail. In several devastating diseases, proteins that normally contain helical regions undergo a structural conversion to beta-sheet-rich forms, which then aggregate into insoluble fibers called amyloid. This helix-to-sheet switch is a central event in conditions like Alzheimer’s disease and type 2 diabetes.
The amyloid-beta peptide involved in Alzheimer’s has an N-terminal region (roughly residues 10 to 24) that exists in a dynamic equilibrium between alpha-helical and beta-strand conformations. Mutations that push this region toward more helix dramatically reduce amyloid formation, while mutations that reduce helical content accelerate it. In hereditary Dutch-type cerebral hemorrhage, a variant of Alzheimer’s, a single substitution at position 22 decreases the helix propensity of this domain and increases amyloid production.14PubMed. The alpha-helical to beta-strand transition in the amino-terminal fragment of the amyloid beta-peptide modulates amyloid formation An engineered mutation of valine 18 to alanine (a stronger helix-former) significantly increases helical content and dramatically diminishes fibril formation, confirming that the structural balance is what matters.14PubMed. The alpha-helical to beta-strand transition in the amino-terminal fragment of the amyloid beta-peptide modulates amyloid formation
A similar story plays out with human amylin, the peptide that forms amyloid deposits in the pancreas in type 2 diabetes. Simulations of the amylin fragment show that the helix-to-sheet transition can follow two paths: a direct conversion of the coil portion of a partially helical form into antiparallel beta-sheet, or a full unfolding of the helix followed by reorganization into sheet.15PubMed Central. Conformational distribution and α-helix to β-sheet transition of human amylin fragment dimer Both routes end at the same destination: the sticky, self-assembling beta-sheet structure that forms toxic aggregates. Understanding these pathways has become a major focus for drug design, since blocking the structural switch early could prevent fibril formation entirely.
How Scientists Measure Secondary Structure
If you cannot see individual atoms, you can still get a good estimate of how much helix and sheet a protein contains. The most widely used laboratory technique for this is circular dichroism (CD) spectroscopy, which measures how a protein solution absorbs left-handed versus right-handed circularly polarized light. Alpha-helices and beta-strands produce distinctly different CD spectra, so by comparing a protein’s spectrum to reference spectra from proteins of known structure, researchers can estimate secondary structure content.16PubMed. Prediction of protein secondary structure from circular dichroism using theoretically derived spectra
The catch is that the math behind this estimation works much more reliably for alpha-helices than for beta-sheets. Early analyses showed that the linear model used to decompose CD spectra was really only adequate for the helix class.17Analytical Biochemistry. Estimation of protein secondary structure and error analysis from circular dichroism spectra Beta-sheets are harder because their CD signal depends on factors like strand twist and whether they are parallel or antiparallel. A more recent method called BeStSel addresses this by explicitly accounting for the twist of beta-structures, enabling reliable distinction between parallel and antiparallel sheets across a broad range of proteins.18PubMed Central. Accurate secondary structure prediction and fold recognition for circular dichroism spectroscopy
On the computational side, predicting secondary structure from amino acid sequence alone has been a long-running challenge. Neural networks trained on evolutionary sequence information achieved sustained accuracy above 70% in three-state prediction (helix, sheet, or coil) as far back as the mid-1990s, outperforming alternatives by at least six percentage points.19PubMed. Combining evolutionary information and neural networks to predict protein secondary structure Modern deep-learning methods have pushed accuracy considerably higher, and tools like AlphaFold now predict full three-dimensional structures with remarkable reliability, but the secondary-structure prediction problem is where the field cut its teeth.
Post-Translational Modifications as Structural Switches
A protein’s secondary structure is not necessarily fixed after it folds. Cells routinely attach chemical groups to amino acid side chains after translation, and some of these modifications can stabilize or destroy local helical structure. Phosphorylation, the addition of a phosphate group, is a good example of how context-dependent the effect can be. When phosphorylation occurs at the very start of an alpha-helix, it actually stabilizes the helix by acting as a hydrogen-bond acceptor for exposed backbone NH groups, similar to how asparagine acts as a natural N-cap. Dianionic phosphothreonine at the N-terminus is one of the most helix-stabilizing modifications known.20PubMed Central. OGlcNAcylation and phosphorylation have similar structural effects in α-helices: post-translational modifications as inducible start and stop signals in α-helices, with greater structural effects on threonine modification
Put the same phosphate group in the middle of a helix, though, and the effect reverses. Interior phosphorylation destabilizes helices, and in extreme cases the modified peptide reverts to a random coil. O-linked sugar modifications (OGlcNAcylation) follow a parallel pattern: stabilizing at the N-terminus, destabilizing in the interior and at the C-terminus.20PubMed Central. OGlcNAcylation and phosphorylation have similar structural effects in α-helices: post-translational modifications as inducible start and stop signals in α-helices, with greater structural effects on threonine modification A study on myelin basic protein confirmed that phosphorylation at a specific threonine impedes alpha-helix formation, reduces the reversibility of folding, and alters the global structure of the peptide through changed electrostatic interactions.21PubMed Central. The effects of threonine phosphorylation on the stability and dynamics of the central molecular switch region of 18.5-kDa myelin basic protein
This position-dependent logic gives cells a way to use phosphorylation and sugar modifications as inducible start and stop signals for helical segments. A single enzyme adding a phosphate can flip a stretch of protein from ordered to disordered, or vice versa, depending on where the modification lands. It is an elegant layer of regulation built on top of the same hydrogen-bonding physics that makes helices possible in the first place.
Designing New Proteins from Scratch
One of the sharpest tests of our understanding of secondary structure is whether we can design entirely new proteins that fold into predicted shapes and stay stable. Researchers have recently used computational methods, including both physics-based approaches and deep-learning hallucination, to generate novel repeat proteins featuring mixed alpha-helix and beta-strand topologies. Twenty-five new, highly stable alpha-beta proteins were designed and experimentally validated using these approaches.22bioRxiv. De novo design of alpha-beta repeat proteins The fact that these designed proteins fold correctly and remain stable in the lab is strong evidence that the rules governing secondary structure formation, amino acid preferences, hydrogen-bonding patterns, and hydrophobic packing, are understood well enough to be applied constructively, not just descriptively.
This capability has practical implications. Designed proteins can serve as scaffolds for vaccines, as biosensors, as components of molecular machines, or as therapeutics. The ability to specify exactly where helices and sheets appear in a novel protein, and to have that design hold up in reality, represents a transition from reading nature’s structural language to writing in it.
Intrinsically Disordered Regions and Transient Structure
For decades, the assumption was that a protein needed a stable three-dimensional structure to function. That assumption has been overturned. A large fraction of proteins in complex organisms contain regions that remain disordered under normal conditions, lacking fixed alpha-helices or beta-sheets. These intrinsically disordered regions (IDRs) are not broken or incomplete; they are functional precisely because they are flexible.
Many disordered regions fold into defined secondary structures only when they encounter a binding partner, a process sometimes called folding upon binding. The structural outcome, whether a stretch becomes helical or adopts a strand conformation, can depend on the shape of the target molecule. The binding mechanisms vary widely: some disordered regions sample their folded state transiently even before contact (conformational selection), while others are essentially shapeless until the binding partner molds them into form (induced fit). Real binding events often combine elements of both.23PubMed Central. Features of molecular recognition of intrinsically disordered proteins via coupled folding and binding This means that secondary structure, in these proteins, is not a permanent architectural feature but a conditional response to molecular context. It is a useful reminder that alpha-helices and beta-sheets, as stable as they seem in crystal structures, are really just the lowest-energy states of a chain that is constantly sampling alternatives.