Alpha Helix vs Beta Sheet: How Protein Secondary Structure Forms and How to Tell Them Apart

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-10-02

Protein secondary structure is the regular, local folding of the polypeptide backbone into repeating patterns. The two patterns you will meet constantly are the alpha helix, a coiled spring, and the beta sheet, a set of extended strands zipped together by hydrogen bonds. Both are held together by the same interaction, hydrogen bonds between backbone CO and NH groups, but they arrange those bonds in completely different geometries [1].

You need this distinction because it shows up everywhere. When you read a crystal structure, run a secondary-structure assignment tool, or interpret a circular dichroism spectrum, you are deciding which residues are helix, which are strand, and which are neither. Getting the alpha helix vs beta sheet call right changes how you read a binding site, a mutation, or a folding study. This article builds the concept from the backbone up, then shows how to assign secondary structure to a real structure with code.

Quick Answer

Secondary structure is the regular local folding of the polypeptide backbone; the main types are the alpha helix, the beta sheet, and turns and loops [1].
In an alpha helix, the CO group of each residue hydrogen bonds to the NH group four residues ahead (an i to i+4 pattern), with 3.6 residues per turn and a 1.5 Angstrom rise per residue [1].
In a beta sheet, two or more extended strands are linked by hydrogen bonds between strands, not within one local segment; adjacent residues in a strand are about 3.5 Angstrom apart [1].
The quickest visual rule: a helix is a compact coil with hydrogen bonds within one local segment, a sheet is an extended, pleated arrangement with hydrogen bonds between strands, which can belong to the same chain or to different chains [1].
Direction matters in sheets: antiparallel strands run opposite ways, parallel strands run the same way, and the hydrogen-bond patterns differ accordingly [1].
You can assign both types from atomic coordinates with DSSP, which reads hydrogen-bond patterns and geometry [4].

What Secondary Structure Actually Is

A polypeptide is a chain of amino acids joined by peptide bonds. The backbone has a repeating unit of nitrogen, alpha carbon, and carbonyl carbon, and the only groups available for regular hydrogen bonding are the backbone CO and NH. Secondary structure is what happens when those groups pair up in a repeating way along the chain [1].

The alpha helix was proposed by Pauling, Corey and Branson in a 1951 PNAS paper describing hydrogen-bonded helical configurations of the polypeptide chain [2]. In the same year, Pauling and Corey described pleated-sheet configurations of polypeptide chains [3]. These two papers set the framework that textbooks still use.

The helix solves a geometric problem. If each CO bonds to the NH four residues ahead, the backbone curls into a regular coil. That i to i+4 pattern is the defining feature [1]. The numbers follow from it: 3.6 residues per turn, a 1.5 Angstrom rise per residue along the axis, and a pitch of 1.5 times 3.6, which is 5.4 Angstrom per turn [1].

$$ \text{pitch} = 1.5\ \text{Å per residue} \times 3.6\ \text{residues per turn} = 5.4\ \text{Å per turn} $$

Here the pitch is the distance the helix advances along its axis in one full turn. Essentially all alpha helices found in proteins are right-handed [1]. Some proteins are dominated by this one pattern: about 75% of the residues in ferritin, an iron-storage protein, are in alpha helices [1]. Two or more helices can also entwine into a coiled coil, a very stable arrangement that can reach 1000 Angstrom (100 nm) or more in length [1].

The beta sheet works differently. A single beta strand is an extended chain, with adjacent amino acids about 3.5 Angstrom apart, far more stretched out than the 1.5 Angstrom rise per residue in a helix [1]. The side chains of adjacent residues point in opposite directions, alternating above and below the sheet, which is why the sheet looks pleated [1]. A sheet forms when two or more strands are linked by hydrogen bonds between the strands [1]. Many strands, typically 4 or 5 but as many as 10 or more, can come together in one sheet, and most sheets adopt a somewhat twisted shape instead of staying flat [1].

Parallel vs Antiparallel Beta Sheets

Direction is the detail that trips people up. In an antiparallel beta sheet, adjacent strands run in opposite directions, and the NH and CO groups of each amino acid hydrogen bond to the CO and NH groups of a single partner on the adjacent strand [1]. The pairing is direct and symmetric.

In a parallel beta sheet, adjacent strands run in the same direction. Each NH hydrogen bonds to the CO of one amino acid on the adjacent strand, while the CO hydrogen bonds to the NH of the amino acid two residues farther along the chain [1]. The bonds are slanted, not straight across, which is why parallel sheets look slightly different in a structure viewer.

Feature	Alpha helix	Beta sheet
Hydrogen-bond pattern	i to i+4 within one segment [1]	Between separate strands [1]
Residues per repeat	3.6 per turn [1]	No turns; side chains alternate above and below the sheet [1]
Rise per residue	1.5 Angstrom [1]	Extended, roughly 3.5 Angstrom spacing [1]
Handedness or direction	Almost always right-handed [1]	Strands parallel or antiparallel [1]
Side-chain arrangement	Point outward from the coil	Alternate above and below the sheet [1]

Turns, Loops, and the Residues That Break Regular Structure

Not every residue sits in a helix or strand. A beta turn, also called a reverse turn, reverses the chain direction, and its defining hydrogen bond runs from the CO group of residue i to the NH group of residue i+3 [1]. Turns and loops invariably lie on the surfaces of proteins and often take part in interactions with other molecules [1].

Amino acid identity shapes which structure forms. Alanine has the highest helix propensity, and excluding proline, glycine has the lowest, about 1 kcal/mol less favorable than alanine on the Pace and Scholtz experimental scale [7]. Proline tends to disrupt both alpha helices and beta strands because it lacks an NH group, so it cannot donate a backbone hydrogen bond [8]. Glycine readily fits into all structures and is well suited to reverse turns, while valine, threonine and isoleucine, which are branched at the beta carbon, destabilize helices through steric clashes, and serine and asparagine also tend to disrupt alpha helices [8]. These are the helix breakers people mean when they say proline and glycine.

Many alpha helices and beta strands are amphipathic, with a hydrophobic face pointing into the protein interior and a more polar face pointing into solution [8]. That split is what lets a helix sit at a membrane interface or pack against a hydrophobic core.

How to Assign Secondary Structure from Coordinates

The standard computational method is DSSP, published by Kabsch and Sander in 1983. It assigns secondary structure from hydrogen-bond patterns and geometry in atomic coordinates: repeating turns are helices, repeating bridges are ladders, and connected ladders are sheets [4].

DSSP uses one-letter codes: H for alpha helix, B for isolated beta bridge, E for extended strand in a beta ladder, G for 3-10 helix, I for pi helix, P for kappa (polyproline II) helix, T for hydrogen-bonded turn, and S for bend [5]. If you open a PDB file directly, the HELIX records carry a helix class (1 is right-handed alpha, 5 is right-handed 3-10), and the SHEET records carry a sense field: 0 for the first strand, 1 for parallel and -1 for antiparallel to the previous strand [9].

If you want to inspect a structure yourself, the Protein Structure Viewer on this site lets you load a PDB entry and look at the helix and strand assignments visually before you run any code.

Worked Example

Assign secondary structure to ubiquitin (PDB 1UBQ). Download https://files.rcsb.org/download/1UBQ.pdb, then run this in Python:

from Bio.PDB import PDBParser, DSSP
m = PDBParser(QUIET=True).get_structure('u', '1UBQ.pdb')[0]
d = DSSP(m, '1UBQ.pdb', dssp='mkdssp')
ss = ''.join(d[k][2] for k in d.keys())
print(ss)

The result for 76 residues is:

-EEEEEETTS-EEEEE--TTSBHHHHHHHHHHHH---GGGEEEEETTEEPPTTSBTGGGTPPTT-EEEEEE--S--

Counts: E 24, H 12, T 12, '-' 12, G 6, S 4, P 4, B 2. The elements are an alpha helix at residues 23-34 (12 residues, 3.3 turns), 3-10 helices at 38-40 and 57-59, and strands at 2-7, 12-16, 41-45, 48-49 and 66-71. So about 16% of residues are alpha helix (12/76) and 32% are strand (24/76).

The deposited PDB HELIX records agree on the main helix (Ile23-Glu34, class 1 right-handed alpha) and list Leu56-Tyr59 as class 5 (3-10). The SHEET records show a mixed sheet: 1-7 antiparallel to 10-17 (the N-terminal hairpin), 64-72 parallel to 1-7, 40-45 antiparallel to 64-72, and 48-50 antiparallel to 40-45.

Geometry checks confirm the textbook numbers. In the helix, the backbone O(i) to N(i+4) distances for i = 23 to 30 are 3.00, 3.09, 2.93, 2.98, 2.93, 2.87, 2.91, 2.91 Angstrom (mean 2.95), the signature of i to i+4 hydrogen bonds. Fitting an axis through the C-alpha atoms gives a mean rise of 1.48 Angstrom per residue (axial length 16.3 Angstrom over 11 steps), matching the textbook 1.5 Angstrom. In the strands, half the C-alpha(i) to C-alpha(i+2) distance is 3.27 to 3.36 Angstrom, close to the textbook 3.5 Angstrom; real strands are pleated and twisted.

A 12-residue alpha helix spans 12 / 3.6 = 3.3 turns and about 12 x 1.5 = 18 Angstrom along its axis. The measured 16.3 Angstrom axial length is shorter mainly because 12 residues span only 11 rise steps between C-alpha atoms (11 x 1.48 = 16.3 Angstrom).

Reading Secondary Structure from Circular Dichroism

Far-UV circular dichroism gives you a fast, low-sample estimate of secondary-structure content. Alpha-helical proteins show negative bands at 222 nm and 208 nm and a positive band at 193 nm [6]. Beta-sheet structure shows a negative band at 218 nm and a positive band at 195 nm [6]. Disordered proteins have very low ellipticity above 210 nm and a negative band near 195 nm [6].

CD can be measured on samples of 20 micrograms of protein or less in physiological buffers in a few hours, but it gives no residue-specific information of the kind obtained by X-ray crystallography or NMR [6]. Treat the band positions as characteristic, not exact, and remember that estimated fractions depend on the reference set and algorithm you choose.

Common Mistakes

Treating a beta sheet as one continuous chain. A sheet is built from separate strands, and the hydrogen bonds run between strands, not within one local segment as in a helix [1]. If you look for i to i+4 bonds, you will miss it.
Assuming all helices are left-handed or mixed. Essentially all alpha helices found in proteins are right-handed [1]. A left-handed alpha helix is a red flag in a model.
Confusing 3.6 residues per turn with 3.6 Angstrom. Residues per turn is a count, not a distance. The rise per residue is 1.5 Angstrom, and the pitch is 5.4 Angstrom per turn [1].
Ignoring strand direction. Parallel and antiparallel sheets have different hydrogen-bond geometries [1]. The PDB SHEET sense field encodes this as 1 or -1 [9].
Reading CD bands as exact wavelengths. The 222, 208, 218 and 195 nm positions are characteristic, not fixed, and the estimated fractions depend on the reference set and algorithm [6].
Assuming proline and glycine behave the same. Proline lacks an NH group and cannot donate a backbone hydrogen bond, so it disrupts helices and strands, while glycine fits many structures and suits reverse turns [8].

Limitations

DSSP versions differ. DSSP 4 adds the P (polyproline II) code, so older DSSP output for 1UBQ would show those 4 residues as coil or turn. Different assignment methods (DSSP, STRIDE, author annotations) give slightly different helix and strand boundaries, so do not treat a boundary as exact.

The 3.6 residues per turn, 1.5 Angstrom rise and 5.4 Angstrom pitch quoted here are the modern textbook values [1][2]; when you cite them, cite a current source alongside the 1951 Pauling, Corey and Branson paper.

Helix propensity values from Pace and Scholtz apply to solvent-exposed residues in the middle of helices; context such as N-cap, C-cap and tertiary packing changes propensities [7]. Pace and Scholtz also report that the average globular protein contains 30% alpha helix, and other surveys give different averages [7].

CD band positions are characteristic, not exact, and secondary-structure fractions estimated from CD depend on the reference set and algorithm [6]. Check the current documentation for whichever tool you use, since codes and defaults change between releases.

Frequently Asked Questions

What is an alpha helix?

An alpha helix is a regular coil in which the CO group of each amino acid hydrogen bonds to the NH group of the amino acid four residues ahead, an i to i+4 pattern [1]. It has 3.6 residues per turn and a 1.5 Angstrom rise per residue, giving a pitch of 5.4 Angstrom per turn [1]. Essentially all alpha helices in proteins are right-handed [1].

What is a beta pleated sheet?

A beta pleated sheet forms when two or more extended strands are linked by hydrogen bonds between the strands [1]. Adjacent residues in a strand sit about 3.5 Angstrom apart, and their side chains alternate above and below the sheet, which creates the pleated look [1]. Sheets can be flat or, more often, somewhat twisted [1].

What is the difference between parallel and antiparallel beta sheets?

In an antiparallel sheet, adjacent strands run in opposite directions, and each NH and CO hydrogen bonds to a single partner on the neighboring strand [1]. In a parallel sheet, strands run the same direction, and each NH bonds to one CO on the adjacent strand while each CO bonds to an NH two residues farther along [1]. The PDB SHEET sense field marks these as 1 and -1 [9].

What is a beta turn?

A beta turn, or reverse turn, reverses the direction of the chain. Its signature hydrogen bond runs from the CO group of residue i to the NH group of residue i+3 [1]. Turns and loops sit on protein surfaces and often participate in interactions with other molecules [1].

Why do proline and glycine break helices?

Proline lacks an NH group, so it cannot donate the backbone hydrogen bond a helix needs, and it disrupts beta strands too [8]. Glycine is different: it fits into many structures and is well suited to reverse turns, but excluding proline it has the lowest helix propensity, about 1 kcal/mol less favorable than alanine [7][8].