Quantum Tunneling in Biology: How Enzymes and Photosynthesis Cheat Classical Physics

PhysicsJul 25, 202613 min

A small glowing bead passes straight through a dense leafy hedge while a slightly larger twin bead waits at its foot.

Replace one hydrogen atom with deuterium, which is the same atom carrying one extra neutron, and a particular enzyme in soybeans slows down by a factor of about eighty. Not eighty percent. Eighty times. Classical chemistry has a perfectly good explanation for why heavier isotopes react more slowly, and that explanation runs out of road somewhere around seven. The other factor of ten or so is a hydrogen nucleus going through a wall instead of over it.

That fact is solid, and it is the strongest card quantum biology has. The trouble is that it usually gets dealt from the same deck as three other claims, about migrating birds, the sense of smell and photosynthesis, which are much shakier, and one of which has largely collapsed. The headline version folds all four into a single story about life secretly exploiting quantum weirdness. I think the useful version grades them separately, and the grades are very different.

One complaint about the framing first. Nothing here "cheats classical physics". Classical physics was always the approximation; the question is only where the approximation stops being good enough. Chemistry is quantum mechanics all the way down. Every bond in your body is an electron wavefunction. So "is biology quantum?" is a non-question. The interesting question is narrower: does a specifically quantum effect, one that classical intuition gets wrong, change a biological outcome in a way you can measure?

§Eighty times slower for one extra neutron

Tunneling is what happens when a particle finds itself on the far side of an energy barrier it did not have the energy to climb. The particle's wavefunction does not stop dead at the wall; it decays exponentially inside it, and if the wall is thin enough, some amplitude leaks through. The rough WKB estimate of the probability is:

T \approx \exp\!\left(-\frac{2d}{\hbar}\sqrt{2m\,(V - E)}\right)

The variable that matters is the one inside the square root. Mass $m$ sits in the exponent, which is why tunneling is a party trick for electrons, routine for hydrogen and fades fast for anything heavier. Double the mass, as you do going from hydrogen to deuterium, and the exponent grows by a factor of $\sqrt{2}$ . When the exponent is large, that single change can wipe out most of the rate. Tunneling is exquisitely sensitive to isotopes. That sensitivity is the fingerprint experimentalists look for.

The measurement is called a kinetic isotope effect: the rate with hydrogen divided by the rate with deuterium. Semiclassical theory, which accounts for the fact that a lighter atom sits higher in its vibrational well and so has less barrier left to climb, predicts this ratio should come in under about 7 at room temperature. Soybean lipoxygenase-1, the enzyme from the opening paragraph, has a measured value of 81, and it barely changes with temperature.^[1]1. S. J. Edwards, A. V. Soudackov and S. Hammes-Schiffer, Impact of Distal Mutation on Hydrogen Transfer Interface and Substrate Conformation in Soybean Lipoxygenase, J. Phys. Chem. B 114, 6653 (2010), which gives the wild-type value as 81 at room temperature, with only modest temperature dependence.

That second detail is the more damning one for the classical picture. If the hydrogen were getting over the barrier by thermal energy, a warmer enzyme should close the gap between hydrogen and deuterium. Instead, in many native enzymes the isotope effect is almost flat across temperature, while the reaction itself still speeds up with heat. Judith Klinman and Amnon Kohen's review of the field reads this as the signature of a two-step process: slow, temperature sensitive protein motions bring the hydrogen donor and acceptor close together, and then the hydrogen tunnels across in a step that is nearly indifferent to temperature.^[2]2. J. P. Klinman and A. Kohen, Hydrogen tunneling links protein dynamics to enzyme catalysis, Annual Review of Biochemistry 82 (2013). The semiclassical ceiling, the prefactor test, the 3.2 Å versus 0.7 Å gap and the distance-decay comparison are all from this review.

The geometry is the part I find genuinely beautiful. Crystal structures show the donor and acceptor atoms sitting more than 3.2 ångströms apart. Models of the tunneling step want something closer to 0.7. The enzyme closes that gap by flexing, sampling conformations until it hits one where the hydrogen's wavefunctions on both sides overlap. Hydrogen transfer is far more sensitive to distance than electron transfer, by an empirical factor Klinman and Kohen put at roughly 23 per ångström against about 1, so a fraction of an ångström matters enormously. Seen this way, an enzyme is less a lock-and-key than a machine for squeezing a barrier thin.

Grade: A. Hydrogen tunneling in enzymes is established, measured across many systems, and explains data that nothing classical does. The open argument is not whether it happens but how much of an enzyme's overall rate advantage it accounts for, and how tightly protein motion is coupled to it. That is a real scientific disagreement, not a hype problem.

§The robin's compass is spin chemistry, not tunneling

Bird navigation is usually filed under "quantum tunneling in biology", and that is simply the wrong label. The leading explanation for the light-dependent magnetic compass of migratory songbirds is the radical-pair mechanism, and its quantum ingredient is electron spin.^[3]3. P. J. Hore and H. Mouritsen, The radical-pair mechanism of magnetoreception, Annual Review of Biophysics 45, 299–344 (2016). A tutorial written for both biologists and physicists, and careful to call radical pairs the leading hypothesis rather than the answer.

The idea runs like this. Light hits a cryptochrome protein in the bird's retina and kicks an electron from one part of the molecule to another, leaving two radicals, each with an unpaired electron. The two spins start out correlated, and they then oscillate between configurations chemists call singlet and triplet. The two configurations lead to different chemical products. The oscillation is driven mostly by the magnetic nuclei nearby, but the Earth's field nudges it, and how much depends on the molecule's orientation relative to that field. Different orientation, different product yield, and in principle a different signal to the brain.

The obvious objection is scale. The energy of an electron spin in the Earth's field, roughly 50 microtesla, is about $\mu_B B \approx 5 \times 10^{-28}$ joules. Thermal energy at body temperature, $k_B T$ , is around $4 \times 10^{-21}$ joules, some ten million times bigger. Any mechanism that relied on the field shifting a thermal equilibrium would be hopeless. The radical-pair mechanism dodges this because it never reaches equilibrium. It is a race: the spin dynamics play out in the brief window before the radicals react or relax, and the field only has to tilt the odds in that window. That is a clever escape, and it is also exactly why the spins have to stay correlated long enough, which in a warm protein is the hard part.

The evidence is suggestive and getting better. In 2014, Henrik Mouritsen's group found that European robins in huts on the Oldenburg campus lost their magnetic orientation, and regained it when the huts were grounded and screened to cut electromagnetic noise in the 50 kHz to 5 MHz range. The tests were double-blind, and adding noise back in the screened huts broke the compass again.^[4]4. S. Engels et al., Anthropogenic electromagnetic noise disrupts magnetic compass orientation in a migratory bird, Nature 509, 353–356 (2014). The logic, as I understand it, is that noise that weak is hard to square with a magnetic-particle sensor and much easier to square with fragile spin dynamics. In 2021, a collaboration led from Oxford and Oldenburg showed that cryptochrome 4 from the robin is magnetically sensitive in the test tube, more so than the same protein from chickens and pigeons, neither of which is a night migrant, and used targeted mutations to identify four flavin-tryptophan radical pairs involved.^[5]5. J. Xu et al., Magnetic sensitivity of cryptochrome 4 from a migratory songbird, Nature 594, 535–540 (2021).

"In the test tube" is carrying weight in that sentence. No one has yet watched this happen inside a living bird's eye and traced the signal to behaviour. Grade: B. A real quantum effect, the best candidate mechanism by a wide margin, and still not closed. Also: not tunneling.

§The nose that may or may not do spectroscopy

Smell is where quantum biology gets its most romantic idea, and its weakest evidence. The standard theory is shape: an odorant binds to receptors whose pockets fit it, in a combinatorial code. The alternative, championed by Luca Turin in the 1990s, says receptors also read molecular vibrations, using something like inelastic electron tunneling spectroscopy. An electron tunnels across the receptor only if it can dump exactly the right quantum of energy into a vibration of the odorant molecule. Your nose, on this account, is a spectrometer.

It is a lovely idea and it makes a sharp prediction. Deuterium barely changes a molecule's shape but changes its vibrational frequencies a lot. So deuterated and ordinary versions of the same odorant should smell different. In 2004 Andreas Keller and Leslie Vosshall at Rockefeller ran Turin's proposed tests on human subjects. None of them could tell deuterated acetophenone from the ordinary kind, and the other predictions failed too. In 2011 Turin and colleagues reported that fruit flies could be trained to tell deuterated odorants apart, and even generalised the distinction to an unrelated molecule with a vibration near the carbon-deuterium stretch.^[6]6. Rockefeller University, Rockefeller University scientists take on controversial and widely publicized "vibration theory" of smell (2004), on A. Keller and L. B. Vosshall, Nature Neuroscience 7, 337 (2004). Then M. I. Franco, L. Turin, A. Mershin and E. M. C. Skoulakis, Molecular vibration-sensing component in Drosophila melanogaster olfaction, PNAS 108, 3797 (2011).

Then a 2013 study reported that people could discriminate deuterated musks, and in 2015 a large team went to the receptor directly. The human musk receptor OR5AN1 responded identically to normal and deuterated versions of its musk ligands in vitro, as did a mouse thiol receptor and several others. They also argued that the proposed electron-transfer mechanism would be swamped by vibrations of the receptor itself.^[7]7. E. Block et al., Implausibility of the vibrational theory of olfaction, PNAS 112, E2766 (2015). It also summarises the 2013 human study by Gane et al. that it was responding to.

My read: an animal telling isotopes apart is evidence against a purely shape-based model, but it is not evidence for electron tunneling specifically. Deuteration changes more than vibrational spectra, and a behavioural result sits a long way from a mechanism. The receptor-level data cuts against the theory where it should be easiest to see. Grade: C, a minority view with an interesting fly result and no confirmed mechanism. I would be pleased to be wrong.

§Photosynthesis: the beautiful result that did not survive

This is the one that made quantum biology famous, and the one where honesty matters most. In 2007, Graham Fleming's group published two-dimensional spectroscopy of the Fenna-Matthews-Olson complex, a pigment-protein wire in green sulphur bacteria, showing oscillating signals they interpreted as long-lived electronic quantum coherence. The paper suggested this wavelike transfer could explain the system's efficiency, letting it sample many paths at once to find the best one. The measurements were at 77 kelvin, about minus 196 °C.^[8]8. G. S. Engel et al., "Evidence for wavelike energy transfer through quantum coherence in photosynthetic systems," Nature 446, 782–786 (2007). Then H.-G. Duan et al., Nature does not rely on long-lived electronic quantum coherence for photosynthetic energy transfer, PNAS 114, 8493 (2017). The story that followed, of plants running quantum computations to route sunlight, wrote itself.

It did not hold up. In 2017, a team led by R. J. Dwayne Miller repeated the measurement on the same complex in water at room temperature and found electronic coherence decaying in about 60 femtoseconds, which is the old, boring, orthodox picture. They attributed the longer-lived oscillations to molecular vibrations rather than electronic coherence, and concluded there was no hint of a functional role.^[8]8. G. S. Engel et al., "Evidence for wavelike energy transfer through quantum coherence in photosynthetic systems," Nature 446, 782–786 (2007). Then H.-G. Duan et al., Nature does not rely on long-lived electronic quantum coherence for photosynthetic energy transfer, PNAS 114, 8493 (2017). A 2020 review titled "Quantum biology revisited" went further. Its authors, who include several people whose experiments built the original case, concluded that the coherences between excitons are "too short lived to have any functional significance" for energy transfer, and that the long-lived signals come from impulsively excited vibrations.^[9]9. J. Cao et al., Quantum biology revisited, Science Advances 6, eaaz4888 (2020). Note the author list: it includes several of the people whose spectroscopy built the original case.

I want to be careful not to overcorrect. Quantum mechanics obviously matters to photosynthesis. Excitations are delocalised across neighbouring pigments, and that delocalisation is quantum. The 2020 review's most interesting line of argument is that nature does not fight the warm, noisy environment at all. It tunes how the excitations couple to that environment, so that dissipation itself pushes energy downhill toward the reaction centre. What failed is the specific, exciting claim: that long-lived electronic coherence is doing functional work. Grade for that claim: it did not pass. Grade for the field, for publishing the correction in its own journals with its own names on it: excellent.

§Warm and wet is not the same as noisy

Line the four up and a pattern shows. The claim that survived best, enzyme tunneling, is the least exotic: a single light particle crossing a fraction of an ångström, a process that needs no long-lived coherence between distant parts of a molecule. The claims that depended on fragile quantum states persisting in a warm protein are the ones that either shrank or stalled. The radical pair is the interesting middle case, because it needs spin correlations to last, but spins tend to be better shielded from molecular jostling than electronic states are, and the mechanism only needs them for the short life of the radical pair.

That is a more useful heuristic than "life exploits quantum weirdness". Ask what the effect requires to survive, and for how long, and whether the biological environment would let it. Short and local wins. Long and delocalised mostly loses. And when someone tells you biology has figured out how to protect delicate quantum states from the heat of a living cell, notice that the most solid example of biology using quantum mechanics works the other way round: the enzyme uses its thermal motion to set the tunneling up.

So yes, some of this is happening inside you right now. Your enzymes are moving hydrogen through barriers it cannot climb, and you can watch the deuterium slow them down. The rest of the quantum biology catalogue deserves a seatbelt.

^S. J. Edwards, A. V. Soudackov and S. Hammes-Schiffer, Impact of Distal Mutation on Hydrogen Transfer Interface and Substrate Conformation in Soybean Lipoxygenase, J. Phys. Chem. B 114, 6653 (2010), which gives the wild-type value as 81 at room temperature, with only modest temperature dependence.
^J. P. Klinman and A. Kohen, Hydrogen tunneling links protein dynamics to enzyme catalysis, Annual Review of Biochemistry 82 (2013). The semiclassical ceiling, the prefactor test, the 3.2 Å versus 0.7 Å gap and the distance-decay comparison are all from this review.
^P. J. Hore and H. Mouritsen, The radical-pair mechanism of magnetoreception, Annual Review of Biophysics 45, 299–344 (2016). A tutorial written for both biologists and physicists, and careful to call radical pairs the leading hypothesis rather than the answer.
^S. Engels et al., Anthropogenic electromagnetic noise disrupts magnetic compass orientation in a migratory bird, Nature 509, 353–356 (2014).
^J. Xu et al., Magnetic sensitivity of cryptochrome 4 from a migratory songbird, Nature 594, 535–540 (2021).
^Rockefeller University, Rockefeller University scientists take on controversial and widely publicized "vibration theory" of smell (2004), on A. Keller and L. B. Vosshall, Nature Neuroscience 7, 337 (2004). Then M. I. Franco, L. Turin, A. Mershin and E. M. C. Skoulakis, Molecular vibration-sensing component in Drosophila melanogaster olfaction, PNAS 108, 3797 (2011).
^E. Block et al., Implausibility of the vibrational theory of olfaction, PNAS 112, E2766 (2015). It also summarises the 2013 human study by Gane et al. that it was responding to.
^G. S. Engel et al., "Evidence for wavelike energy transfer through quantum coherence in photosynthetic systems," Nature 446, 782–786 (2007). Then H.-G. Duan et al., Nature does not rely on long-lived electronic quantum coherence for photosynthetic energy transfer, PNAS 114, 8493 (2017).
^J. Cao et al., Quantum biology revisited, Science Advances 6, eaaz4888 (2020). Note the author list: it includes several of the people whose spectroscopy built the original case.