You're staring at a protein structure. Maybe it's a sequence alignment where you're trying to figure out why that one mutation breaks everything. In real terms, maybe it's a ribbon diagram in PyMOL. Either way, you keep coming back to the same question: which amino acids actually hydrogen bond?
The short answer is more of them than you think. backbone, donor vs. Day to day, that depends entirely on context — side chain vs. acceptor, buried vs. exposed. The useful answer? And most textbooks oversimplify it.
What Is Hydrogen Bonding in Amino Acids
A hydrogen bond isn't a covalent bond. It's not a van der Waals interaction either. It sits in that weird middle ground: an electrostatic attraction between a hydrogen atom covalently bonded to an electronegative atom (N, O, or S) and another nearby electronegative atom with a lone pair.
In proteins, this happens two ways Worth keeping that in mind..
Backbone hydrogen bonds — every single amino acid participates. The carbonyl oxygen (C=O) of residue i accepts a bond from the amide hydrogen (N-H) of residue i+4 in an alpha helix. In beta sheets, it's between strands. This is the scaffolding. The grammar of protein structure Small thing, real impact..
Side chain hydrogen bonds — this is where the variation lives. Some side chains donate. Some accept. Some do both. Some do neither. And that difference drives specificity, catalysis, folding kinetics, and ligand binding.
The donors, acceptors, and both
Quick mental model: donors have a hydrogen attached to N, O, or S. Consider this: Acceptors have a lone pair on N, O, or S. Both have both features in the same functional group.
| Category | Amino Acids | Functional Group |
|---|---|---|
| Donor only | Trp (indole N-H), Arg (guanidinium), Lys (ε-NH₃⁺) | N-H, N-H, N-H |
| Acceptor only | Asp (carboxylate), Glu (carboxylate) | C=O⁻, C=O⁻ |
| Both donor & acceptor | Ser, Thr, Tyr, Asn, Gln, His (neutral) | -OH, -OH, -OH, -CONH₂, -CONH₂, imidazole |
| Neither (hydrophobic) | Ala, Val, Leu, Ile, Met, Phe, Pro, Gly | — |
No fluff here — just what actually works.
Cysteine's thiol (-SH) can hydrogen bond, but it's weak. Sulfur is less electronegative. " Methionine's thioether? In practice, in practice, it's often treated as "sometimes, weakly. Basically never.
Why It Matters / Why People Care
You might be here because you're designing a mutation. Or interpreting a crystal structure. Or trying to understand why your protein aggregates at 37°C but not 4°C Which is the point..
Hydrogen bonds are the specificity code of biology.
An alpha helix isn't held together by magic — it's 3.6 residues per turn, each carbonyl accepting from the amide four residues ahead. Break one backbone H-bond? The helix frays. Worth adding: break a side chain H-bond in the active site? Catalysis drops 1000-fold.
Real example: serine proteases. The catalytic triad (Asp-His-Ser) works because His accepts from Ser, then donates to Asp. Which means a proton relay. Mutate His to Ala? Dead enzyme. Mutate to Gln? Sometimes retains partial activity — because Gln can still H-bond, just differently The details matter here..
Or consider antibody-antigen interfaces. But they're not random — they're enriched in Tyr, Ser, Asn, Arg. The average interface has 15-20 hydrogen bonds. The "polar zipper" idea. Evolution selects for side chains that can form networks, not just single bonds.
And folding? Hydrogen bonds drive secondary structure formation early. The hydrophobic collapse gets the press, but without backbone H-bonds nucleating helices and sheets, you get molten globules, not native states.
How It Works: Side Chain by Side Chain
This is the reference section. Bookmark it.
Serine and Threonine: The Workhorses
-OH group. One donor (the H), two acceptor lone pairs on oxygen. Small. Flexible. Everywhere That's the part that actually makes a difference. Still holds up..
Ser is the most common phosphorylation site for a reason — that hydroxyl is nucleophilic and hydrogen bonds like crazy. Worth adding: in active sites, it's often the nucleophile (serine proteases, lipases, esterases). In binding pockets, it grips ligands via water-mediated H-bonds.
Thr adds a methyl group. Same chemistry, more steric bulk. That β-branch restricts φ/ψ angles — Thr loves beta sheets, hates alpha helices. Worth remembering when you're modeling loops.
Tyrosine: The Aromatic Hybrid
Phenolic -OH. pKa ~10, so it's protonated at physiological pH. Donor + acceptor like Ser/Thr, but the ring changes everything The details matter here..
The hydroxyl is more acidic than aliphatic alcohols. This makes Tyr a better H-bond donor — the H is more δ+. The ring withdraws electron density. And the π system? It can do CH-π stacking and hydrogen bond simultaneously.
Tyr is overrepresented in protein-protein interfaces. It's the "sticky" aromatic It's one of those things that adds up..
Asparagine and Glutamine: The Amide Pair
-CONH₂. Two donor hydrogens (on nitrogen), one acceptor oxygen. Planar. Rigid.
Asn is shorter — one methylene. Think about it: gln has two. That extra length lets Gln reach further, but also increases entropy cost upon fixing the side chain But it adds up..
Key insight: the amide nitrogen is a weak donor. The hydrogens aren't very acidic. But the carbonyl oxygen is a strong acceptor. In practice, Asn/Gln often accept two H-bonds (bidentate) and donate one or two. They're the backbone-mimics — they love substituting for main chain interactions in turns and loops It's one of those things that adds up. That's the whole idea..
Asn also deamidates. Non-enzymatically. This leads to over time, Asn → Asp/isoAsp. So naturally, that changes H-bonding from donor/acceptor to acceptor-only. Aging proteins literally lose hydrogen bonding capacity Simple, but easy to overlook..
Histidine: The pH Sensor
Imidazole ring. pKa ~6.5. Right at physiological pH Small thing, real impact..
Protonated (HisH⁺): two donor N-H groups, no acceptor (both nitrogens protonated). Neutral (His): one donor (Nε-H or Nδ-H, tautomers), one acceptor (the other nitrogen). Deprotonated: two acceptors, no donors — but this basically doesn't happen at pH 7.
This pH-dependent switching is the point. Histidine in catalytic triads, metal binding sites, proton
shuttling, and pH-sensing loops. Mutagenesis studies often exploit its pKa to lock it in a specific state. In DNA-binding proteins, His can coordinate metal ions like Zn²⁺ or Mg²⁺, stabilizing structures.
Proline: The Helix Breaker
Cyclic structure forces φ angle to ~−60°. No side chain rotation — it’s locked. This rigidity disrupts α-helices but stabilizes β-turns. Proline’s structure also allows it to adopt φ angles favorable for reverse turns. Its lack of a hydrogen bond donor in the ring (N-H is absent) makes it a poor helix participant but a unique player in loop regions. Proline-rich regions often signal disordered or flexible segments.
Cysteine: The Disulfide Architect
Thiol group (-SH) is reactive. Oxidizes to form disulfide bonds (S-S) with another cysteine. These covalent links stabilize tertiary and quaternary structures. Redox-sensitive — disulfide bonds form in oxidizing environments (e.g., extracellular proteins) and break in reducing ones. Mutations disrupting disulfides often lead to misfolding. Cysteine’s small size allows it to fit into tight hydrophobic pockets, making it a frequent residue in active sites Easy to understand, harder to ignore. Surprisingly effective..
Methionine: The Thioether Outlier
Sulfur atom buried in a hydrophobic environment. Thioether (-S-CH₃) is non-reactive compared to cysteine’s thiol. Methionine’s sulfur has lone pairs but rarely participates in H-bonding. Its presence often signals buried hydrophobic cores. Methionine oxidation (to sulfoxide) is a common post-translational modification, altering local structure and function.
Tryptophan: The Aromatic Powerhouse
Largest aromatic side chain. Indole ring has a NH group (donor) and π system (acceptor). The NH is a strong H-bond donor, often interacting with carbonyl oxygens in β-turns or helix-helix interfaces. Tryptophan’s bulk makes it rare in tight spaces but critical in binding pockets for stacking interactions. Its fluorescence is exploited in FRET studies and tryptophan quenching assays.
Glycine: The Flexibility Enabler
No side chain — just a hydrogen. Allows extreme φ/ψ angles. Frequently found in turns, loops, and disordered regions. Glycine’s absence in α-helices (due to destabilizing steric clashes) and preference for β-turns make it a structural wildcard. That said, too much glycine can destabilize proteins by increasing conformational entropy.
Proline and Glycine: The Dynamic Duo
Together, they define β-turns. Proline’s fixed φ angle and Glycine’s flexibility allow the tight turns required for β-sheet formation. Their combinatorial use in turns is a hallmark of secondary structure prediction algorithms.
Conclusion: The Symphony of Side Chains
Every amino acid contributes uniquely to protein structure and function. Hydrophobic residues anchor cores, polar/charged ones mediate interactions, aromatics enable stacking, and reactive side chains (Cys, His) drive catalysis or regulation. Understanding these roles allows rational protein design — tweaking a single residue can flip a switch in activity, stability, or binding. From molten globules to native states, the code isn’t just about sequence; it’s a language of interactions, where every side chain tells a story.
