Nucleotides and the Sugar-Phosphate Backbone

A nucleotide is the monomeric building block of DNA and RNA, and it always consists of three covalently joined parts: a nitrogenous base (a purine or pyrimidine), a five-carbon sugar (2′-deoxyribose in DNA, ribose in RNA), and a phosphate group. The sugar is the structural hub — its carbons are numbered 1′ through 5′ (the prime distinguishes sugar carbons from the numbered atoms of the base ring). The base attaches via a glycosidic bond to the 1′ carbon; the phosphate attaches at the 5′ carbon; and the 3′ carbon carries a free hydroxyl group that is not consumed until the nucleotide is added to a growing chain. A nucleotide missing its phosphate (base + sugar only) is technically called a nucleoside; the nucleotide is the nucleoside plus one, two, or three phosphates (as in ATP, dATP, GTP). The sugar-phosphate backbone is what you get when successive nucleotides are joined not through their bases but through their sugars and phosphates, in a repeating covalent chain. The linking bond is a phosphodiester bond: a single phosphate group is esterified twice, once to the 3′-oxygen of the sugar upstream and once to the 5′-oxygen of the sugar downstream. Because every sugar in the chain donates its 3′-OH to one bond and its 5′-phosphate to the neighboring bond, the backbone has an intrinsic, unbroken directionality — it is not symmetric like a chain of identical beads, but more like a chain of arrows all pointing the same way. This directionality is called 5′-to-3′ polarity. One end of any DNA or RNA strand terminates in a free (or capped) 5′-phosphate; the other end terminates in a free 3′-hydroxyl. Bases dangle off the sugars as side chains and carry the genetic information itself, but it is the invariant backbone — sugar–phosphate–sugar–phosphate — that gives the molecule its mechanical strength, its uniform diameter, and critically, its readable direction. Every enzyme that synthesizes, replicates, repairs, or translates nucleic acid information is built around this fixed polarity.

Step 1 — Anatomy of a single nucleotide. Take deoxyadenosine 5′-triphosphate (dATP) as a concrete example. Its deoxyribose sugar is a five-membered (furanose) ring with carbons numbered 1′ through 5′ around the ring and exocyclic 5′ carbon. Carbon 1′ bears the base adenine via an N-glycosidic bond (nitrogen of the base to carbon of the sugar). Carbon 2′ in DNA carries only hydrogens (hence "deoxy"); in RNA this same position carries a hydroxyl group, the single atomic difference that distinguishes ribose from deoxyribose. Carbon 3′ carries a hydroxyl group that will act as the nucleophile in chain extension. Carbon 5′, exocyclic to the ring, carries the triphosphate tail: three phosphate groups connected by two high-energy phosphoanhydride bonds.

Step 2 — Bond formation: how the chain grows. DNA and RNA polymerases build chains by nucleophilic attack. The free 3′-OH of the last nucleotide already in the chain (or of a primer) attacks the innermost (alpha) phosphate of the incoming nucleoside triphosphate. This forms a new phosphodiester bond between the 3′ oxygen of the growing strand and the 5′ phosphate of the incoming nucleotide, while the beta and gamma phosphates leave together as pyrophosphate (PPi). Pyrophosphate is then hydrolyzed by pyrophosphatase into two inorganic phosphates, which pulls the reaction forward thermodynamically (this hydrolysis step is essentially irreversible under cellular conditions, making chain elongation effectively unidirectional).

Step 3 — Why the direction never reverses. Because the incoming substrate always contributes its 5′ phosphate, and the growing chain always contributes its 3′ hydroxyl, the new bond can only ever add nucleotides onto the 3′ end. There is no enzymatic chemistry in normal replication that adds a nucleotide's 3′ end onto an existing chain's 5′ end. This is why every known DNA and RNA polymerase synthesizes strands 5′→3′, never 3′→5′. It is also why the 5′ end of a nascent strand needs a pre-existing primer with a free 3′-OH — polymerases cannot start a chain from scratch, they can only extend one.

Step 4 — The repeating backbone and its two chemically distinct ends. As synthesis proceeds, the strand becomes a chain of alternating sugars and phosphates: ``` 5′-phosphate–sugar(3′)–O–P–O–(5′)sugar–(3′)–O–P–O–(5′)sugar...-3′-hydroxyl ``` The result is a molecule with a 5′ end (typically a free or capped phosphate group) and a 3′ end (a free hydroxyl group). These ends are not just labels — they are chemically distinguishable and behave differently in reactions: only a 3′-OH can be extended by polymerase; only a 5′-phosphate can be ligated by DNA ligase to a neighboring 3′-OH to seal a nick.

Step 5 — Polarity as the basis for antiparallel double helices and biological reading direction. When two strands pair to form a double helix (foreshadowing the double-helix model), Watson-Crick base pairing geometry requires the two backbones to run in opposite directions — one strand 5′→3′ left to right, its partner 3′→5′ left to right (i.e., 5′→3′ right to left). This antiparallel arrangement is a direct geometric consequence of how the sugar-phosphate backbone's asymmetric shape fits together with base-pair hydrogen bonding. Downstream, this same 5′→3′ rule dictates that replication forks synthesize a continuous leading strand and a discontinuous lagging strand (Okazaki fragments), that ribosomes read mRNA 5′→3′, and that geneticists write sequences left-to-right by convention starting from the 5′ end.

The idea that nucleic acids are linear polymers of nucleotides joined by phosphodiester bonds, with a defined 3′–5′ connectivity, was not simply assumed — it was built from decades of converging chemical, crystallographic, and enzymological evidence.

Phoebus Levene's chemical degradation studies (1910s–1930s). Levene hydrolyzed nucleic acids using nucleases and mild acid/base treatments and identified the breakdown products: bases, pentose sugars, and phosphate. By characterizing which bonds were cleaved by which conditions, he established that phosphate groups bridge adjacent sugars — the polynucleotide chain hypothesis — and correctly identified the phosphodiester linkage as the connecting chemistry, and distinguished ribose (RNA) from deoxyribose (DNA). His later, incorrect "tetranucleotide hypothesis" (that DNA was a simple repeating unit of four bases in fixed proportion) was wrong about composition, but his structural chemistry of the backbone bond itself was correct and stood the test of subsequent work.

X-ray fiber diffraction (Rosalind Franklin and Maurice Wilkins, early 1950s). Franklin's diffraction photographs, most famously "Photo 51," showed a helical diffraction pattern with a repeat distance and helical parameters consistent with a regular sugar-phosphate backbone on the outside of the molecule and a defined rise per nucleotide (~3.4 Å) and helical turn (~34 Å, about 10 base pairs per turn in B-form DNA). Critically, the diffraction data placed the electron-dense phosphates on the periphery of the molecule, confirming that the repeating backbone — not the bases — forms the structural "rails" of the helix, exactly as required if bases are strung off sugars that are themselves linked by phosphates.

Enzymology of DNA synthesis (Arthur Kornberg, 1956–1960). Kornberg isolated DNA polymerase I from E. coli and showed biochemically, using radiolabeled nucleoside triphosphates, that the enzyme incorporates nucleotides only onto a pre-existing primer strand's 3′-OH end, requires all four dNTPs, and releases pyrophosphate as a byproduct — a direct biochemical confirmation of the 3′→5′-directional phosphodiester bond-forming mechanism predicted from the structural chemistry. This work, for which Kornberg won the 1959 Nobel Prize, was the direct functional proof that synthesis proceeds 5′→3′: reactions run with primers lacking a free 3′-OH (blocked, or dideoxy-terminated) failed to extend, showing unambiguously which end of the chain is elongated.

Sanger sequencing and dideoxy chain termination (Frederick Sanger, 1977) later exploited this same directional chemistry as direct proof rather than just inference: dideoxynucleotides (ddNTPs), which lack the 3′-OH entirely, terminate chain growth the instant they are incorporated, because there is no hydroxyl left for the next phosphodiester bond to form. The fact that ddNTP incorporation reliably and immediately halts synthesis — used to generate sequencing ladders — is itself a clean experimental demonstration that chain extension absolutely requires a free 3′-OH, confirming the mechanism predicted from bond chemistry decades earlier. Together, chemical degradation defined the linkage, diffraction showed the backbone's physical location and helical geometry, and enzymology proved directionality was mechanistically enforced, not just structurally implied.

Example 1 — Counting phosphodiester bonds and backbone chemistry in a short oligonucleotide. Consider a synthetic single-stranded DNA primer 18 nucleotides long, written conventionally as: 5′-GATTACAGGCTTAGCCATG-3′ (count the letters: this is actually 19 nucleotides — let's use it precisely). Question: how many phosphodiester bonds link this 19-nucleotide strand, and what are the chemical identities of its two ends? Reasoning: in a linear strand of n nucleotides, each internal junction between adjacent sugars requires exactly one phosphodiester bond, and there are always n − 1 such junctions (the first and last nucleotide each contribute only one connecting bond, not two, because they are chain ends). Calculation: ``` n = 19 nucleotides phosphodiester bonds = n − 1 = 18 ``` The 5′-most nucleotide (G) retains a free (or in synthesis, triphosphate-derived monophosphate) group at its 5′ carbon — the 5′ end. The 3′-most nucleotide (G, last letter) retains a free 3′-OH — the 3′ end, the only site where DNA polymerase could extend this molecule if it were used as a primer. If this oligo were used as a sequencing or PCR primer, a polymerase would recognize its 3′-OH and begin adding nucleotides after position 19, extending the sequence in the 5′→3′ direction, never before position 1.

Example 2 — Predicting the fate of a replication reaction when a chain-terminating nucleotide is added. A biochemist adds E. coli DNA polymerase I, a primed single-stranded template, all four normal dNTPs, and a small amount of dideoxy-CTP (ddCTP, lacking a 3′-OH) to a reaction tube. The template requires a C to be incorporated at position 15 of the new strand relative to the primer's 3′ end. Question: what fraction of newly synthesized strands will terminate exactly at position 15, given that ddCTP is present at 1% the concentration of normal dCTP, assuming polymerase has no strong bias between ddCTP and dCTP incorporation (a reasonable simplification; in practice there is a modest kinetic discrimination)? Reasoning: at each position requiring a C, the polymerase "chooses" stochastically between dCTP and ddCTP in proportion to their relative concentrations. Since ddCTP has no 3′-OH, its incorporation is a dead end — no further nucleotides can be added after it, because there is no hydroxyl for the next phosphodiester bond. Calculation: probability of terminating specifically at the first "C" position (position 15) rather than reading through with normal dCTP: ``` P(terminate at position 15) = [ddCTP] / ([ddCTP] + [dCTP]) = 1 / (1+100) ≈ 0.0099 ≈ 1% ``` So about 1% of all strands reaching position 15 will terminate there (get a ddC incorporated and stop), while ~99% will incorporate normal dCTP and continue synthesizing further downstream, where they face the same probabilistic choice at the next C in the template. Summed over many templates and many C positions, this generates a population of strands of many different lengths, each ending precisely at a C — the physical basis of Sanger dideoxy sequencing, and a direct demonstration that 3′-OH availability, not base identity, is what permits chain continuation.

Arthur Kornberg trained as a physician before turning to enzymology, and his defining contribution was insisting that the mechanism of DNA replication had to be demonstrated in a test tube with purified components, not merely inferred from genetics or structure. Working at Washington University in St. Louis and then Stanford, Kornberg's laboratory purified an enzyme from Escherichia coli extracts capable of synthesizing DNA in vitro when supplied with a DNA template, a primer, magnesium ions, and all four deoxyribonucleoside triphosphates. He and his colleagues showed rigorously that this enzyme — DNA polymerase I — required a free 3′-hydroxyl group on an existing strand to add each new nucleotide, directly confirming that chain growth proceeds exclusively 5′→3′, exactly as bond chemistry predicted.

Understanding 5′→3′ backbone polarity is not an academic footnote — it is load-bearing infrastructure for modern biotechnology and medicine. PCR primer design depends entirely on placing primers with correctly oriented 3′ ends so that Taq polymerase can extend them toward each other; a primer with the wrong orientation simply cannot be extended, a common troubleshooting error in diagnostic and research labs. Sanger and next-generation sequencing chemistries (dye-terminator sequencing, Illumina's reversible terminator chemistry) are direct engineering applications of blocking or restoring the 3′-OH to control chain termination one base at a time. Antiviral and anticancer nucleoside analog drugs — AZT (zidovudine) for HIV, acyclovir for herpesviruses, and many chemotherapeutics — work by mimicking natural nucleotides but lacking a functional 3′-OH, so that once incorporated by a viral or cancer cell's polymerase, they act as chain terminators exactly like dideoxynucleotides, halting replication of the pathogen or malignant cell's genome. CRISPR-based genome editing and antisense oligonucleotide therapeutics (such as those treating spinal muscular atrophy) are synthesized with precise chemical control over their 3′ and 5′ termini to control stability, ligation, and enzymatic recognition. Molecular cloning relies on DNA ligase's absolute requirement for a 5′-phosphate adjacent to a 3′-OH to seal nicks — a fact used deliberately when engineers dephosphorylate vector ends to prevent unwanted self-ligation. Every one of these technologies is, at bottom, an exploitation of the same simple chemical fact: the backbone only grows one way.

Ask the tutor: How does 5′→3′ backbone polarity force DNA replication to be discontinuous on the lagging strand?