The secondary structures alpha helices and beta sheets are both stabilized by hydrogen bonds. Which atoms share the hydrogen in these weak bonds: the backbone N–H donor and C=O acceptor

The hydrogen is shared between a nitrogen atom and an oxygen atom. In both alpha helices and beta sheets, the hydrogen donor is the amide nitrogen (N–H) of one amino acid, and the acceptor is the carbonyl oxygen (C=O) of another. The hydrogen atom sits closer to nitrogen but is drawn toward oxygen, creating the weak bond. Only backbone atoms participate; side chains are not involved in the defining hydrogen bonds of either structure.

The anatomy of a backbone hydrogen bond

Every amino acid contributes the same three-atom spine to a polypeptide: the amide nitrogen, the alpha carbon, and the carbonyl carbon with its oxygen. The N–H and C=O groups are the only polar handles available along that spine. A hydrogen bond forms when the partially positive hydrogen on nitrogen is attracted to the partially negative oxygen of a carbonyl group.

The geometry matters as much as the chemistry. The N–H···O atoms must lie nearly in a straight line, with the hydrogen-oxygen distance around 1.9 to 2.0 angstroms and the nitrogen-oxygen distance near 2.9 to 3.0 angstroms. The nitrogen-hydrogen-oxygen angle should approach 180 degrees. Deviations of more than about 30 degrees weaken the bond sharply.

This is why the two structures differ in which partners they choose, not in which atoms they use. Both recruit the same N–H donor and C=O acceptor. The difference is the sequence spacing between the partners.

In an alpha helix, each N–H hydrogen-bonds to the C=O of the amino acid four residues earlier in the chain. That i to i-4 pattern repeats every turn, producing a coiled arrangement with about 3.6 residues per turn. In a beta sheet, the partners are residues that may be far apart in sequence but sit adjacent in the folded structure, and the strands line up so their N–H and C=O groups face each other in alternation.

Plain language: the same two chemical groups do the work in both shapes; the shapes differ only in which distant neighbors they pair with.

Quantified context: bond lengths, counts, and energies

A single backbone hydrogen bond carries roughly 1 to 5 kilocalories per mole in the low-dielectric interior of a folded protein. That is much weaker than a covalent bond, which runs 50 to 100 kilocalories per mole or more. The weakness is the point: bonds must form and break reversibly.

An alpha helix of 20 residues contains roughly 16 backbone hydrogen bonds, because the first four N–H groups lack a partner four residues back. A helix shorter than about four turns, or roughly 14 residues, barely stabilizes at all. The terminal residues are often called frayed ends for this reason.

Beta sheets compensate with numbers. A sheet of parallel strands has hydrogen bonds perpendicular to the strand direction, one pair per residue between adjacent strands. Antiparallel sheets form tighter, more linear hydrogen bonds and are generally more stable. A two-stranded antiparallel sheet of 8 residues per strand yields 8 paired hydrogen bonds.

The energy difference between parallel and antiparallel geometry is modest per bond but compounds across a sheet. This is why antiparallel arrangements appear more often in stable protein cores and why beta sheets are frequently twisted rather than flat.

Side chains can contribute additional hydrogen bonds and packing forces, but they are not the defining stabilization described in the keyword. Removing backbone hydrogen bonds in a helix destabilizes the structure by several kilocalories per mole, enough to unfold a small domain.

Edge cases: where the simple picture breaks down

Not every backbone N–H participates. Proline has no amide hydrogen, so it cannot donate. When proline appears in a helix, it breaks the i to i-4 hydrogen-bond chain and often terminates or kinks the helix. Glycine, with its small side chain, is flexible and tolerates helix geometry poorly at certain positions.

Water competes for these bonds. Buried hydrogen bonds in a protein core are shielded from solvent and contribute the full stabilization energy. Exposed bonds on the surface often exchange with water and add little net stability. This is a professional caveat worth stating plainly: counting hydrogen bonds on a structure graphic overstates their energetic contribution.

Helices can be stabilized by terminal capping, where side chains or backbone groups hydrogen-bond to the exposed N–H and C=O at the helix ends, mimicking an extra partner. Sheets can be stabilized by edge-strand pairing with another sheet or by beta-bulges that interrupt an otherwise regular pattern.

A counterintuitive detail: some hydrogen bonds in proteins appear bifurcated, where one donor hydrogen points between two acceptors. These are common and can be nearly as strong as a linear bond, which complicates any simple distance-based count.

Misconceptions and what they get wrong

The most common error is saying hydrogen bonds form between side chains in alpha helices and beta sheets. They can, but the backbone N–H and C=O are the defining participants.

A second error is treating hydrogen bonds as true chemical bonds. They are electrostatic interactions with partial covalent character, not shared-electron bonds. The hydrogen remains attached to nitrogen; it is not transferred.

A third error is assuming an alpha helix is held together by hydrogen bonds along its length only. The coil geometry itself depends on backbone dihedral angles in a narrow allowed range, and hydrogen bonding reinforces rather than alone creates the shape.

Red-flag warnings when reading structural claims:

A figure labeling side-chain contacts as the key helix hydrogen bonds misrepresents the mechanism. A count of "one hydrogen bond per residue" ignores terminal residues and bifurcation. A claim that antiparallel sheets are always more stable ignores context and solvent exposure. A statement that hydrogen bonds are covalent contradicts basic chemistry and any structural biology text. A claim that proline stabilizes helices because it is rigid inverts its actual role.

Helix versus sheet: a comparison of the defining bonds

The table makes the core point explicit: the atoms are identical across all three arrangements. Only sequence spacing and geometry differ.

A timeline of how these bonds were characterized

The 1951 models were remarkable because Pauling and Corey predicted the structures from bond geometry before any protein structure was solved experimentally. The N–H···O distance and linearity constraints they used remain the standard.

Practical actions: how to check these bonds in a structure

If you are examining a protein model or structure file, these steps make the hydrogen-bond network readable:

Use a hydrogen-bond detection tool or a structural viewer that computes donor and acceptor distances, not just visual inspection. Set a distance cutoff of about 3.5 angstroms between nitrogen and oxygen for a plausible bond. Filter for N–H···O angle above roughly 120 degrees to exclude weak or spurious contacts. Confirm the donor is an amide N–H and not a side-chain nitrogen or oxygen unless you are studying side-chain interactions separately. Exclude proline from donor lists, since it has no amide hydrogen. Compare buried versus solvent-exposed bonds before claiming stabilization significance.

These thresholds are practical working ranges, not a universal standard. Different software uses slightly different cutoffs, and bifurcated bonds may fall outside a single simple rule. Reconcile any software output with the geometry described here rather than trusting a default setting.

For students, the most useful exercise is to take a helix and a sheet from the same structure, list each N–H and C=O pair, and confirm the partners differ in sequence position but not in atom type. That one comparison dissolves most confusion.

Limitations: what the simple answer leaves out

The keyword question asks which atoms share the hydrogen, and the direct answer is nitrogen and oxygen. But the real stabilization of any secondary structure involves more. Side-chain packing, van der Waals contacts, hydrophobic burial, and backbone dihedral preferences all contribute.

Hydrogen bond strength also depends on environment. A bond in a hydrophobic core can contribute several kilocalories per mole, while the same geometry at the surface may contribute nearly nothing. Numerical claims about total stabilization must specify burial and context.

This article describes general structural biochemistry. It is educational information, not a diagnostic or treatment guide, and it does not provide medical advice. For clinical or research questions about specific proteins, consult a qualified structural biologist or physician.