Chargaff's Rules and Base Composition

Chargaff's rules are two empirical generalizations about the base composition of DNA, established through direct chemical analysis of DNA from many organisms. The first rule (often called the "1:1 rule") states that within any double-stranded DNA sample, the molar percentage of adenine equals that of thymine, and the molar percentage of guanine equals that of cytosine: %A = %T and %G = %C. A direct corollary is that the total amount of purines (adenine + guanine) equals the total amount of pyrimidines (thymine + cytosine), because each purine base is matched one-for-one with a pyrimidine base. The second rule describes variation between species: the ratio of (A+T) to (G+C)—sometimes written as the "AT/GC ratio"—differs enormously from one organism to another, ranging from roughly 0.45 in some GC-rich bacteria to nearly 2.7 in some AT-rich organisms, yet this ratio is essentially constant across all the tissues and cell types of a single organism, and constant with the age or nutritional state of the organism. In other words, base composition is a fixed, heritable, species-specific property of the genome, not something that varies from a liver cell to a neuron to a sperm cell in the same animal. It is worth being precise about what the rules do not say. They do not claim that all four bases occur in equal amounts (that would require %A = %T = %G = %C = 25%, which is false for nearly every real genome). They also do not, by themselves, prove that DNA is double-stranded or reveal the helical geometry—Chargaff's numbers are consistent with several structural explanations. What they demonstrate, ironclad and reproducible, is a strict pairing constraint between specific bases, a constraint that any correct structural model of the genetic material had to satisfy exactly.

Step 1: The four bases and their chemical classes. DNA is built from four nitrogenous bases attached to the deoxyribose-phosphate backbone (see Nucleotides and the Sugar-Phosphate Backbone). Two are purines—adenine (A) and guanine (G)—built on a fused double-ring structure. Two are pyrimidines—thymine (T) and cytosine (C)—built on a single six-membered ring. Purines are physically larger than pyrimidines.

Step 2: Extracting and hydrolyzing DNA. To measure base composition, DNA is first isolated from cells (Chargaff worked with calf thymus, human sperm, yeast, tubercle bacilli, and many other sources) and purified away from protein and RNA. The purified DNA is then hydrolyzed—broken down by acid or enzymatic digestion—into its individual free bases (or nucleotides), releasing A, T, G, and C as separate chemical species in solution.

Step 3: Separating the bases by paper chromatography. The mixture of hydrolyzed bases is spotted onto paper and run in a solvent system. Because each base has a distinct polarity and shape, they migrate at different rates and separate into distinct spots on the chromatogram. This lets each base be isolated and identified independently.

Step 4: Quantifying each base by UV spectrophotometry. Each purine and pyrimidine has a characteristic absorbance spectrum in the ultraviolet range (all four bases absorb strongly near 260 nm because of their aromatic ring systems, but each has a distinguishable peak and extinction coefficient). By eluting each separated spot and measuring its UV absorbance, Chargaff's group could calculate the molar quantity of each base present in the original sample.

Step 5: Computing ratios. From the measured molar amounts, the lab calculated %A, %T, %G, %C for each DNA source, and derived the A/T and G/C ratios (which came out at or very near 1.0 in every double-stranded DNA tested) and the (A+T)/(G+C) ratio (which varied by species but was constant within one species' different tissues).

Step 6: The chemical logic that explains the numbers. The ratios are not a coincidence of nature but a direct consequence of molecular complementarity. In the eventual double helix, each base on one strand hydrogen-bonds to a base on the opposite strand, and the geometry of the helix only accommodates a purine paired with a pyrimidine (a purine–purine pair would be too bulky to fit the uniform helix diameter; a pyrimidine–pyrimidine pair would leave too large a gap). Chemically, adenine's hydrogen-bond donor/acceptor pattern matches thymine's exactly (forming two hydrogen bonds), and guanine's pattern matches cytosine's exactly (forming three hydrogen bonds). Because every A on one strand demands a T directly across from it, and every G demands a C, the total count of A in a genome must equal the total count of T, and the total count of G must equal the total count of C—regardless of how much A/T pairing versus G/C pairing the genome uses overall. That "how much" is exactly the species-specific quantity captured by the (A+T)/(G+C) ratio.

Who did the work. Erwin Chargaff and his colleagues at Columbia University—particularly Ernst Vischer, Rollin Hodes, and Elizabeth Zamenhof, working through the late 1940s and published chiefly between 1949 and 1951—carried out the definitive measurements of DNA base composition across many species. This work followed and depended on a key earlier tool: the two-dimensional paper chromatography technique that had recently made it possible to separate and identify small quantities of the four bases quantitatively, something not reliably possible before.

What was done. DNA was extracted and purified from a wide range of biological sources—calf thymus, calf spleen, calf liver, human sperm, yeast, the bacterium Mycobacterium tuberculosis (the tubercle bacillus), wheat germ, and others. Each DNA sample was hydrolyzed with acid to release the free bases, and the resulting mixture was separated by paper chromatography into four distinct spots corresponding to A, G, C, and T. Each spot was eluted and its concentration determined by ultraviolet absorption spectrophotometry, exploiting the characteristic absorbance of purine and pyrimidine rings near 260 nm. From these measurements, molar percentages and ratios were computed for each sample.

What was observed. Two patterns emerged with striking consistency:

Across every organism tested, the A/T ratio and the G/C ratio came out at or extremely close to 1.0, even though the overall percentage of A+T versus G+C varied dramatically—from AT-rich genomes like human and yeast to strongly GC-rich genomes like the tubercle bacillus. Critically, when Chargaff's group compared DNA from different tissues of the same organism (calf thymus vs. calf spleen vs. calf liver), the (A+T)/(G+C) ratio stayed the same within experimental error, even though the tissues are cytologically and functionally completely different. This ruled out any simple model in which base composition reflected a cell's specialized function, and instead pointed to base composition as a fixed, genome-wide property—exactly what would be expected if DNA, not protein, were the hereditary material, since this was being established at the same time by the Avery-MacLeod-McCarty transformation experiments and, later, the Hershey-Chase blender experiment.

What it proved (and what it didn't). The consistent 1:1 purine-to-pyrimidine ratios told researchers that something in the chemistry of DNA forces adenine and thymine to occur in equal amounts, and guanine and cytosine to occur in equal amounts, in every organism regardless of overall GC content. It strongly suggested a fixed structural pairing rule linking A with T and G with C. It did not by itself specify a double helix, a single strand, three strands, or any particular geometry—Chargaff's numbers are a boundary condition on the correct structure, not a picture of the structure itself. Chargaff himself, having established the rules, did not propose a structural model; that step required combining his ratios with two other lines of evidence: Rosalind Franklin and Maurice Wilkins's X-ray diffraction images of DNA fibers (particularly Photograph 51, showing a helical diffraction pattern with defined spacings), and model-building of nucleotide geometry by Watson and Crick. When Watson and Crick built physical models of paired bases, they found that only A–T (via two hydrogen bonds) and G–C (via three hydrogen bonds) pairings produced a uniform-width double helix consistent with both the diffraction data and Chargaff's ratios simultaneously—no other pairing scheme (like G with T, or A with C) satisfied all three data sets at once. Watson later recounted that learning of Chargaff's ratios directly (Chargaff and Watson met in 1952) was a turning point that redirected the model-building effort toward specific base pairing rather than like-with-like pairing (which several earlier, incorrect models had assumed).

Example 1: Reconstructing full base composition from partial data. A newly sequenced viral DNA genome is reported to be 38% guanine. Assuming standard double-stranded Watson-Crick base pairing, find the percentages of all four bases and the (A+T)/(G+C) ratio. Because %G = %C by Chargaff's first rule: %C = 38%. Purines + pyrimidines must total 100%, and %G + %C already account for 38 + 38 = 76%. Remaining percentage for A + T = 100 − 76 = 24%. Since %A = %T, each is half of that remainder: %A = %T = 12%. Check: %A + %T + %G + %C = 12 + 12 + 38 + 38 = 100 ✓, and A/T = 12/12 = 1.0, G/C = 38/38 = 1.0, satisfying both parts of the first rule. (A+T)/(G+C) = 24/76 ≈ 0.32. This is a strongly GC-rich genome (compare M. tuberculosis above, at 0.60)—biologically plausible for some bacteriophages and GC-rich bacteria, where a high GC content is associated with greater thermal and mechanical stability of the double helix, because G–C pairs have three hydrogen bonds versus two for A–T. Note this calculation is a straightforward inference for double-stranded DNA; it would fail for a single-stranded genome (such as some bacteriophages, e.g. φX174), where Chargaff's first rule does not hold, because there is no complementary strand forcing A to pair with T.

Example 2: Using the second rule to identify a tissue-composition anomaly. A student measures DNA from four samples labeled Liver-A, Liver-B, Kidney-C, and "Sample-X" from the same rabbit, and obtains (A+T)/(G+C) ratios of 1.35, 1.34, 1.36, and 1.71 respectively. What should the student conclude? Chargaff's second rule predicts that all tissues from one organism should show the same (A+T)/(G+C) ratio, since genomic base composition is a fixed property of the species' genome, not of cell type. Liver-A, Liver-B, and Kidney-C cluster tightly around 1.35, consistent with this expectation and with the rabbit's genomic AT/GC ratio. Sample-X's value of 1.71 is a clear outlier. Reasoning through the possibilities: this discrepancy is far too large to be explained by ordinary measurement noise (the other three values agree to within 0.02). More likely explanations include:

• Sample-X is contaminated with DNA from a different, more AT-rich organism (e.g., a bacterial or viral contaminant, or even human DNA from handling).
• Sample-X was not pure DNA—if RNA (which also uses A, G, C but uracil instead of thymine, and typically shows somewhat different base ratios) was incompletely removed, or if mitochondrial DNA (which can have a distinct base composition from nuclear DNA in some species) dominated the sample, the ratio would shift.
• A technical artifact in the chromatography or hydrolysis step led to selective loss of one base.

This illustrates the actual power of Chargaff's second rule in practice: because within-organism constancy is so reliable, an outlier ratio is itself diagnostic of a technical or biological problem with the sample, a logic still used today as a quality-control check in genome sequencing pipelines (unexpected GC-content shifts flag contamination or assembly errors).

Erwin Chargaff was an Austrian-born biochemist who spent most of his career at Columbia University after fleeing Europe in the 1930s. Trained broadly in chemistry, he had worked on lipids and lipoproteins before turning, in the mid-1940s, to nucleic acid chemistry—a decision he later said was directly prompted by reading the 1944 Avery, MacLeod, and McCarty paper showing that DNA, not protein, was the transforming principle in pneumococcus. Chargaff recognized immediately that if DNA truly carried genetic information, its chemistry deserved the same rigorous quantitative scrutiny biochemists gave to proteins, and he set about establishing reliable methods to measure exactly how much of each base a given DNA sample contained. Chargaff was notoriously skeptical of what he saw as overreach in molecular biology, and he became a famously acerbic critic of Watson and Crick's model-building style, once describing them memorably (and unkindly) in his memoirs. Ironically, his own most enduring scientific legacy—the base ratios that bear his name—became indispensable raw material for the very double-helix model he later disparaged.

Chargaff's base-composition logic remains a working tool across biology and medicine, not merely a historical curiosity. In genome sequencing and assembly, GC-content profiling is used routinely as a quality-control signature: unexpected shifts in local or genome-wide GC content flag contamination, mis-assembly, or horizontal gene transfer events, and characteristic GC content is used to help identify the taxonomic origin of unknown sequences in metagenomic samples. In molecular diagnostics and PCR primer design, GC content directly determines the melting temperature of a DNA duplex (GC-rich regions, with three hydrogen bonds per base pair, denature at higher temperatures than AT-rich regions with two), a calculation biochemists perform every time they design primers for PCR-based tests, including clinical genotyping and pathogen detection assays. In microbial taxonomy, characteristic genomic GC content has historically been used (alongside DNA-DNA hybridization and now whole-genome sequencing) as a coarse species-delineation criterion, since GC content is a stable, heritable property of a given bacterial lineage, exactly as Chargaff's second rule predicted decades before genome sequencing existed. In synthetic biology and vaccine design, engineers deliberately tune GC content of synthetic constructs (including codon-optimized genes and, notably, some mRNA vaccine sequence designs) to control mRNA stability, secondary structure, and translation efficiency—an applied extension of the same base-pairing chemistry Chargaff's ratios first exposed. And pedagogically, Chargaff's rules remain the cleanest bridge available for students moving from nucleotide chemistry to the geometric logic of the double helix, because the numbers force the conclusion that specific, fixed base pairing—not any structure a student might otherwise guess—is the only chemistry consistent with the data.

Ask the tutor: How did Watson and Crick use hydrogen-bonding geometry together with Chargaff's ratios to rule out incorrect base-pairing schemes before arriving at the final double helix model?