The Pairing: Beyond Flatland
Draft for the author's reading. Established science throughout; the framing is the author's.
Many science writers describe DNA as a linear, digital set of instructions, but this frame limits our understanding in a very harmful way.
A line is the simplest reading: one strand, four letters, read one way. As a temporary frame it is useful — it lets you spell a gene and copy it. The harm begins the moment the temporary stop is forgotten and the line is mistaken for the molecule. What was a frozen slice becomes, by habit, the whole. An entire picture of life is then built on the assumption that the line is the thing — and the thing keeps moving underneath it.
The geometric view does not discard the sequence. It sets it back into the structure it came from. DNA is not a string that happens to twist; it is a twisted, charged, folding object whose sequence is one of its dimensions — the one along the strand. The other dimensions do the rest of the work, and the line cannot see any of them.
Geometry, not alphabet
The four letters are shapes. Each base pair is a hydrogen-bonded fit at a fixed angle — two bonds holding A–T, three holding G–C — and the rungs stack on one another at a repeat spacing that sets the helix its pitch. The complementarity that carries the code is a geometric fit, not a symbolic rule. The molecule does not read its letters; it meets them.
More than one dimension
The strand runs along its length, but it also runs around the helix, across to its paired strand, and through the space of the cell. A region millions of letters apart on the line can sit touching in three dimensions. On the line they are strangers; in the cell they are neighbours, and that neighbourhood is what brings an enhancer to its promoter and switches a gene on. The contact, not the order, is the event.
DNA can be read as a line of letters. That is a fair description of what one strand spells. It is not a description of what the molecule is. The line breaks at the first step: the moment a second strand faces the first.
What the line drops
The linear-digital frame is not wrong as a first abstraction. It is a temporarily stopped frame, and useful as long as we recognise we have stopped it. What follows is what that frame cannot hold — what is lost the moment the stop is forgotten and the slice is treated as the whole.
- Topology
- The strands wind around each other, and the whole molecule supercoils. How many times they wind — the linking number — is not written in any letter, yet it constrains everything the molecule can do. Topoisomerases cut and reseal to manage torsion the sequence never records.
- Torsion
- A wound helix stores mechanical energy. Negative supercoiling is tension held in the molecule, and that tension does work: it pries the strands apart so the code can be read and copied. A linear string has no tension; the living molecule does.
- The shape a protein reads
- Proteins do not read letters; they read shapes. The edges of the bases present a pattern in the major groove, the minor groove narrows or widens, a stretch bends — and the same sequence in a different shape is a different message. The code a protein fits against is three-dimensional [3, 4].
- Spatial relation
- The genome folds, and folding overrides order. Enhancers reach promoters across megabases; loops hold regions in contact; the genome partitions into domains of neighbourhood. What is adjacent in space is not what is adjacent on the line, and function follows the space.
- Electrostatics
- The phosphate backbone is a line of negative charge, and its field is three-dimensional — screened by ions, read by proteins, gathered by condensates. How the molecule bends, compacts, and binds is set by this charge landscape. The letters name the charges; the shape delivers them.
- Periodicity
- The helix has a repeat — roughly ten and a half pairs per turn, a fixed rise along the axis [6]. The stacking of bases is an oscillation the line flattens into a count. Resonance-like structure disappears into a number.
- Phase and motion
- The helix breathes: it opens and closes, it switches form — B-, A-, Z-DNA — under conditions the cell sets [6]. It is a moving, phase-changing structure, not a static tape. A frozen frame reads it as still; it is not still.
None of these losses is recoverable from the sequence alone. That is the point: the line is a projection, and a projection cannot be summed back into the object without the dimensions it threw away. The geometric view sets them back — the sequence restored to its twist, its charge, its contact, its motion — a standing wave in the fluid, not a string laid flat.
Freeze-framing a living process
And DNA is only the first case. Freeze-framing any biological process loses a great deal in the translation, because a living process is not an object held still — it is a trajectory, a passage through states. The snapshot is a projection of a motion, and the motion is where the meaning sits. Stop the frame to study it, by all means; but the frame stopped is a tool, not the thing.
This is sharpest where the structure is the motion: in the folding of proteins. A protein's fold is not a final pose reached and then held; it is a kinetic route through intermediate shapes, steered by the same fields this book follows, timed by the cell. A static picture of the folded chain shows the lock and never shows the lock learning to be a lock. The fold is a path, and freezing it discards the path. The next chapter takes up that motion in detail.
Two strands, one relation
A pairs with T; G pairs with C [1, 2]. Each strand is therefore enough to rebuild the other. The information no longer sits in either strand alone. It sits in the match between them — a relationship that must be satisfied in both directions at once. The strands also run opposite ways, so the same pair is read one way from one side and the other way from the other [1].
Four letters, not two symbols
A binary code would need only two symbols. Four letters in two complementary pairs give more than a count: every pair is a two-ring base facing a one-ring base, so the ladder stays the same width whatever sits in it [1]. A–T is held by two hydrogen bonds, G–C by three, so the letters also set how easily a stretch opens. Most of the helix's stability, though, comes from the stacking of the bases on top of one another, not from the pairing bonds [5].
From outside the helix, each of the four possible pair orientations presents a different pattern of chemical edges in the major groove. A protein can tell A–T from T–A, or G–C from C–G, without opening the molecule at all [3]. The sequence also changes the local width and shape of the minor groove, and proteins read that shape too [4].
Pick a letter on the top strand
- The pair
- G – C: a purine (two rings) facing a pyrimidine (one ring). Every pair spans the same width, so the ladder stays even whichever letters sit in it.
- Hydrogen bonds
- 3 — G–C holds more tightly, so G–C-rich stretches open less easily.
- Major-groove edge
- A · A · D · H
- The pattern a protein can feel from outside without opening the helix. Each of the four orientations is different.
- Direction
- The strands run opposite ways (5′→3′ against 3′→5′). The same pair reads differently from each side.
The helix
Pairing and stacking together wind the ladder into a helix, about two nanometres across, roughly ten and a half pairs per turn in its common form [6]. The turn is what makes the grooves, and the grooves are where the letters show themselves to the outside world. So the four-letter alphabet is doing work in every dimension: along the strand (order), across the strands (complement), around the helix (groove patterns), and in space (the shape a protein fits against).
A flat, one-way reading of that is a useful abstraction. It no longer suffices as a description.
The layers together
The full model, from the string to time, for stepping through:
The string
What it holds: Order. Which letter follows which.
What the layer below cannot: The description — the tape Dawkins reads. Useful, and already an idealisation.
References
- Watson, J. D. & Crick, F. H. C. (1953). Molecular structure of nucleic acids. Nature 171, 737–738. doi:10.1038/171737a0
- Chargaff, E. (1950). Chemical specificity of nucleic acids and mechanism of their enzymatic degradation. Experientia 6, 201–209. doi:10.1007/BF02173653
- Seeman, N. C., Rosenberg, J. M. & Rich, A. (1976). Sequence-specific recognition of double helical nucleic acids by proteins. PNAS 73, 804–808. doi:10.1073/pnas.73.3.804
- Rohs, R. et al. (2009). The role of DNA shape in protein–DNA recognition. Nature 461, 1248–1253. doi:10.1038/nature08473
- Yakovchuk, P., Protozanova, E. & Frank-Kamenetskii, M. D. (2006). Base-stacking and base-pairing contributions into thermal stability of the DNA double helix. Nucleic Acids Research 34, 564–574. doi:10.1093/nar/gkj454
- Dickerson, R. E. et al. (1982). The anatomy of A-, B-, and Z-DNA. Science 216, 475–485. doi:10.1126/science.7071593