# How to Read Circular Dichroism Graphs: A Student's Guide

## Introduction to Circular Dichroism Spectroscopy

### What is Circular Dichroism?

Circular dichroism (CD) spectroscopy measures the difference in absorption of left-handed versus right-handed circularly polarized light by a chiral sample. When plane-polarized light passes through a solution of optically active molecules, the left and right circular components are absorbed to different extents. This differential absorption, expressed as ΔA = A_L − A_R, arises because chiral chromophores interact asymmetrically with the two polarization states.

In proteins, the relevant chromophores are the peptide backbone amide bonds, the aromatic side chains of phenylalanine, tyrosine, and tryptophan, and disulfide bonds. The peptide bond absorbs light in the far-ultraviolet (far-UV) region (approximately 180–250 nm), where the electronic transitions (n→π* and π→π*) are sensitive to the local backbone geometry. Because secondary structures such as α-helices and β-sheets impose characteristic dihedral angles on the backbone, each structure produces a distinctive CD signature. This makes CD spectroscopy a rapid, solution-phase method for estimating secondary structure content and monitoring conformational changes.

The CD graph itself is a plot of ellipticity (or differential absorption) on the y-axis versus wavelength on the x-axis. The shape, sign, and magnitude of the bands encode structural information. Learning to read these graphs is a core skill in protein biochemistry, analogous to interpreting a [Read Melt Curve qPCR](/knowledge/molecular-biology/read-melt-curve-qpcr) output for nucleic acid analysis—both require pattern recognition and an understanding of the underlying physics.

### The CD Spectrophotometer and Sample Requirements

A CD spectrophotometer consists of a light source (typically a xenon arc lamp), a monochromator, a photoelastic modulator that alternates between left- and right-circularly polarized light, and a photomultiplier detector. The instrument measures the difference in absorbance between the two polarization states at each wavelength and reports it as ellipticity in millidegrees (mdeg) or as Δε in units of M⁻¹ cm⁻¹.

Sample requirements are stringent. The buffer must be transparent in the wavelength range of interest; common choices include phosphate-buffered saline (10–20 mM phosphate, pH 7.4) or Tris-HCl at low concentration. Reducing agents such as dithiothreitol (DTT) absorb strongly below 220 nm and should be avoided or used at minimal concentrations. The protein concentration typically ranges from 0.1 to 1.0 mg/mL for far-UV measurements, and the pathlength is 0.1 cm or less to minimize solvent absorption. For near-UV CD (250–350 nm), higher concentrations (0.5–2.0 mg/mL) and longer pathlengths (0.5–1.0 cm) are used because aromatic side chains have weaker CD signals than the peptide backbone.

## The CD Spectrum: Axes and Units

### Wavelength Range and Far-UV vs Near-UV CD

The x-axis of a CD graph is wavelength in nanometers (nm). The far-UV region (180–250 nm) reports primarily on secondary structure because the amide chromophore of the peptide backbone absorbs here. The near-UV region (250–350 nm) reports on tertiary structure through the aromatic side chains and disulfide bonds, which absorb at longer wavelengths due to their extended π-electron systems.

Far-UV CD is the workhorse for secondary structure analysis. The spectrum is typically recorded from 260 nm down to 190 nm or lower, provided the buffer and instrument allow transmission below 200 nm. Near-UV CD is recorded from 350 nm down to 250 nm and provides a "fingerprint" of the protein's tertiary fold. A well-packed protein shows a structured near-UV spectrum with multiple peaks and shoulders, whereas a molten globule or unfolded protein shows a nearly flat baseline.

### Mean Residue Ellipticity Calculation

The raw output of a CD instrument is ellipticity in millidegrees (mdeg). This value depends on the protein concentration, the pathlength, and the number of peptide bonds. To compare spectra between proteins or between laboratories, the data are normalized to mean residue ellipticity, [θ], with units of deg·cm²·dmol⁻¹.

The conversion formula is:

[θ] = (θ_obs × 100) / (c × l × n)

where θ_obs is the observed ellipticity in millidegrees, c is the protein concentration in mg/mL, l is the pathlength in centimeters, and n is the number of amino acid residues (peptide bonds) in the protein. The factor of 100 converts from millidegrees to degrees and accounts for the molar concentration of residues.

Alternatively, if the protein concentration is known in molar terms, the molar ellipticity [θ]_molar can be calculated using the molar concentration of the protein. Mean residue ellipticity is preferred because it normalizes for protein size, allowing direct comparison of secondary structure content between proteins of different lengths. This normalization is conceptually similar to how [Read Length](/knowledge/molecular-biology/read-length) normalizes sequence features in genomics—both approaches remove size as a confounding variable.

## Characteristic CD Signatures of Secondary Structures

### Alpha-Helix: Double Minima at 208 and 222 nm

The α-helix produces the most distinctive CD spectrum of all secondary structures. It is characterized by two negative minima at approximately 208 nm and 222 nm, and a positive maximum near 192 nm. The 222 nm band arises from the n→π* transition of the amide group, while the 208 nm band and the 192 nm band arise from exciton splitting of the π→π* transition.

The ratio of the ellipticities at 222 nm and 208 nm ([θ]₂₂₂/[θ]₂₀₈) is often used as an empirical indicator of helix quality. For a canonical, well-formed α-helix, this ratio is close to 1.0. Values significantly less than 1.0 may indicate the presence of 3₁₀-helix or distorted helical segments. The magnitude of the signal is also proportional to helix content: a fully helical protein such as myoglobin (which is approximately 75% α-helical) shows a [θ]₂₂₂ of roughly −30,000 deg·cm²·dmol⁻¹, whereas a protein with no helix shows a much weaker signal.

### Beta-Sheet: Minimum near 218 nm and Maximum near 195 nm

The β-sheet CD spectrum is less intense than that of the α-helix and is characterized by a single negative minimum near 218 nm and a positive maximum near 195 nm. The negative band is broader and shallower than the α-helix minima, typically with [θ]₂₁₈ around −10,000 to −20,000 deg·cm²·dmol⁻¹ for a predominantly β-sheet protein.

The exact positions and intensities of β-sheet bands vary more than those of α-helices because β-sheets can be parallel, antiparallel, or mixed, and the twist of the sheet affects the electronic coupling between amide groups. Antiparallel β-sheets often show a slight red shift of the negative band compared to parallel sheets. Concanavalin A, a predominantly β-sheet protein, exhibits the classic β-sheet signature with a minimum near 218 nm and a maximum near 195 nm.

### Random Coil: Minimum near 200 nm

A disordered or random coil polypeptide produces a CD spectrum with a strong negative minimum near 200 nm and a weak positive or negative band above 210 nm. The negative band at 200 nm is typically intense, with [θ]₂₀₀ around −20,000 to −40,000 deg·cm²·dmol⁻¹ for a fully unfolded protein.

It is important to note that "random coil" is a misnomer; unfolded proteins are not truly random but sample a broad ensemble of backbone conformations. The CD spectrum reflects the average of this ensemble. Thermally denatured proteins, such as boiled lysozyme, show this characteristic random coil signature. The transition from a folded spectrum to a random coil spectrum as temperature increases is the basis for thermal denaturation studies monitored by CD.

## Estimating Secondary Structure Content from CD Data

### Deconvolution Software and Reference Databases

Quantitative estimation of secondary structure content from CD spectra requires computational deconvolution. The measured spectrum is a linear combination of the spectra of individual secondary structure elements, weighted by their fractional content. Deconvolution algorithms solve this inverse problem by fitting the experimental spectrum to a linear combination of reference spectra derived from proteins of known structure.

Common algorithms include SELCON3, CDSSTR, and CONTINLL. These programs use reference databases of CD spectra from proteins whose crystal structures are known, typically obtained from the [X Ray Crystallography](/knowledge/molecular-biology/x-ray-crystallography) database. The algorithms differ in how they handle the reference set and the fitting procedure, but all require input spectra with good signal-to-noise ratio and accurate concentration determination.

The output is a percentage estimate of α-helix, β-sheet, β-turn, and random coil content. For example, a protein might be estimated as 40% α-helix, 25% β-sheet, and 35% random coil. These estimates are generally accurate to within ±5–10% for the major secondary structure types when the spectrum is of high quality and the protein is well-behaved.

### Limitations of Quantitative Estimates

CD-based secondary structure estimates have inherent limitations. First, they provide a global average; they cannot tell you which residues are in which structure. Second, the estimates are model-dependent—different algorithms can give slightly different results for the same spectrum. Third, the reference databases are biased toward soluble, globular proteins; membrane proteins and [intrinsically disordered proteins](/knowledge/bioinformatics/intrinsically-disordered-proteins-and-computational-structural-classification) may not be well represented.

Furthermore, the far-UV CD spectrum is sensitive to the length of helical segments. Short helices (fewer than 10 residues) have reduced ellipticity at 222 nm compared to longer helices, which can lead to underestimation of helix content. Aromatic side chains and disulfide bonds can also contribute to the far-UV spectrum, particularly in the 220–230 nm region, complicating the analysis.

For these reasons, CD should be used as a complementary technique rather than a standalone structural method. It is excellent for comparing conformational states (e.g., native vs. denatured) and for monitoring structural changes, but it cannot replace high-resolution methods for detailed structural determination.

## Near-UV CD and Tertiary Structure Information

### Aromatic Side Chains and Disulfide Bonds

Near-UV CD (250–350 nm) probes the asymmetric environment of aromatic side chains and disulfide bonds. Phenylalanine shows sharp, weak bands between 255 and 270 nm. Tyrosine exhibits bands between 275 and 282 nm, with a characteristic shoulder near 278 nm. Tryptophan shows the most intense near-UV bands, with a broad peak between 280 and 300 nm. Disulfide bonds, which absorb in the 250–270 nm region, contribute broad, featureless bands.

The intensity and shape of these bands depend on the rigidity of the protein. In a folded protein, aromatic side chains are held in fixed orientations within the asymmetric environment of the protein, producing well-defined CD bands. In an unfolded protein, the side chains are mobile and experience a more isotropic environment, resulting in weak or absent near-UV CD signals.

### Using Near-UV CD to Monitor Conformational Changes

Near-UV CD is a sensitive probe for tertiary structure integrity. A protein that loses its native tertiary structure—for example, upon addition of a chemical denaturant such as 6 M guanidine hydrochloride—shows a dramatic loss of near-UV CD signal. This makes near-UV CD useful for monitoring folding/unfolding transitions, ligand binding, and protein stability.

For example, the enzyme lysozyme (14.3 kDa) shows a characteristic near-UV spectrum dominated by tryptophan bands between 280 and 300 nm. Upon thermal denaturation, these bands disappear, reflecting the loss of the rigid tertiary structure. Similarly, the binding of a ligand that changes the environment of an aromatic residue can be detected as a change in the near-UV CD spectrum. This approach is complementary to far-UV CD, which would report on secondary structure changes, and to techniques like [His Tag Protein Purification](/knowledge/molecular-biology/his-tag-protein-purification) that assess protein quality through binding behavior.

## Factors Affecting CD Spectra and Data Quality

### Buffer and Solvent Interference

Buffer components can absorb in the far-UV region and distort CD spectra. Common interfering agents include:

- **Dithiothreitol (DTT) and β-mercaptoethanol**: Strongly absorb below 220 nm; use at ≤1 mM or replace with Tris(2-carboxyethyl)phosphine (TCEP).
- **Imidazole**: Absorbs below 230 nm; remove by dialysis after affinity purification.
- **High salt concentrations**: Sodium chloride at >100 mM can cause baseline drift due to increased absorbance.
- **EDTA**: Absorbs below 200 nm; use at minimal concentrations.

The buffer baseline should always be recorded and subtracted from the sample spectrum. A good practice is to dialyze the protein into a minimal buffer such as 10 mM sodium phosphate, pH 7.4, before CD measurements.

### Concentration and Pathlength Optimization

The optimal sample conditions depend on the wavelength range. For far-UV CD, the absorbance of the sample should be kept below 1.0–1.5 to avoid artifacts from stray light and detector saturation. This typically requires a protein concentration of 0.1–0.5 mg/mL in a 0.1 cm pathlength cell. For near-UV CD, where the signals are much weaker, higher concentrations (0.5–2.0 mg/mL) and longer pathlengths (0.5–1.0 cm) are needed.

Accurate concentration determination is critical for quantitative analysis. A simple A₂₈₀ measurement using a UV spectrophotometer, with the extinction coefficient calculated from the [amino acid sequence](/blog/guides/amino-acid-sequence), is standard. Errors in concentration directly propagate to errors in mean residue ellipticity and hence to secondary structure estimates.

## Common Pitfalls in Reading CD Graphs

### Misinterpreting Noise as Real Features

CD spectra at low wavelengths (below 200 nm) are often noisy because the solvent and buffer absorb strongly, reducing the light reaching the detector. This noise can appear as spurious peaks or troughs. A common mistake is to interpret these artifacts as real spectral features. The signal-to-noise ratio can be assessed by comparing repeated scans; genuine features are reproducible, whereas noise is not.

A practical rule: if a peak or shoulder appears only in one scan or has an irregular, jagged shape, it is likely noise. The high-tension (HT) voltage of the photomultiplier tube is a useful diagnostic—if the HT exceeds approximately 600 V, the signal is unreliable and the data should be discarded.

### Overlooking Buffer Contributions

Buffer components can contribute to the CD signal, particularly in the far-UV region. Even after baseline subtraction, residual buffer absorbance can distort the spectrum. This is especially problematic when the buffer contains aromatic compounds, reducing agents, or high concentrations of salt. Always record a buffer blank under identical conditions and subtract it from the sample spectrum.

### Confusing Wavelength Units

The x-axis of a CD graph is wavelength in nanometers, not wavenumbers or frequency. Students sometimes confuse the positions of the α-helix minima (208 and 222 nm) with those of other techniques. Additionally, some older papers report CD data in wavenumbers (cm⁻¹) or as molar ellipticity rather than mean residue ellipticity. Always check the units before comparing spectra from different sources.

## Practical Steps for Analyzing a CD Graph

### Step-by-Step Interpretation Checklist

1. **Check the x-axis range**: Is this a far-UV (180–250 nm) or near-UV (250–350 nm) spectrum? This determines what structural information is available.
2. **Assess data quality**: Look for excessive noise, especially below 200 nm. Check that the baseline is flat and that the spectrum is reproducible.
3. **Identify the major bands**: Note the positions and signs of the main peaks and troughs. Compare these to the reference spectra for α-helix, β-sheet, and random coil.
4. **Evaluate the α-helix signature**: Look for the double minima at 208 and 222 nm. If present, estimate the helix content from the magnitude of [θ]₂₂₂.
5. **Check for β-sheet features**: A minimum near 218 nm with a maximum near 195 nm suggests β-sheet content.
6. **Consider the random coil contribution**: A strong negative band near 200 nm with weak features above 210 nm indicates significant disorder.
7. **Compare with a reference**: If possible, overlay the spectrum with a known protein of similar structure or with a deconvolution algorithm output.
8. **Interpret in context**: Consider the protein's size, the buffer conditions, and any prior knowledge about the protein's expected structure.

### Summary: From Graph to Structural Insight

Reading a CD graph is a systematic process of pattern recognition and contextual interpretation. The far-UV region tells you about secondary structure: the double minima at 208 and 222 nm indicate α-helix, the minimum near 218 nm with a maximum near 195 nm indicates β-sheet, and a strong minimum near 200 nm indicates disorder. The near-UV region tells you about tertiary structure: the presence of sharp bands between 250 and 300 nm indicates a well-packed protein, while a flat spectrum indicates a molten globule or unfolded state.

Quantitative analysis requires deconvolution software and careful attention to data quality. The estimates are global averages, accurate to within ±5–10% for major secondary structure types, but they cannot provide residue-specific information. CD is best used as a comparative and monitoring tool—for example, to compare the folded and unfolded states of a protein, to screen for optimal buffer conditions, or to monitor the kinetics of folding.

## Frequently Asked Questions

### How do I read a circular dichroism graph?

Start by identifying the wavelength range. If the x-axis spans 180–250 nm, you are looking at far-UV CD, which reports on secondary structure. Look for the characteristic signatures: α-helix shows two negative minima at 208 and 222 nm with a positive band near 192 nm; β-sheet shows a negative minimum near 218 nm and a positive band near 195 nm; random coil shows a strong negative minimum near 200 nm. If the x-axis spans 250–350 nm, you are looking at near-UV CD, which reports on tertiary structure through aromatic side chains and disulfide bonds. Assess data quality first—noisy spectra below 200 nm should be interpreted cautiously.

### What does a positive peak in CD mean?

A positive peak in a CD spectrum indicates that the sample absorbs left-circularly polarized light more strongly than right-circularly polarized light at that wavelength. The sign of the CD signal is determined by the chiral arrangement of the chromophores. For example, the positive band near 192 nm in an α-helix arises from the exciton coupling of the π→π* transitions of the amide groups. A positive peak is not inherently "good" or "bad"—it is simply a feature of the electronic structure and geometry of the chromophores in the protein.

### Why does alpha-helix have two minima at 208 and 222 nm?

The two minima arise from two different electronic transitions in the amide chromophore. The 222 nm band is due to the n→π* transition, which is electric-dipole forbidden but magnetic-dipole allowed, giving it a negative CD signal. The 208 nm band arises from the π→π* transition, which undergoes exciton splitting into two components due to the helical arrangement of the amide groups. One component is polarized parallel to the helix axis and the other perpendicular to it, producing the negative band at 208 nm and the positive band at 192 nm.

### What is mean residue ellipticity and why is it used?

Mean residue ellipticity, [θ], is the observed ellipticity normalized by the protein concentration, the pathlength, and the number of amino acid residues. It has units of deg·cm²·dmol⁻¹. This normalization allows comparison of CD spectra between proteins of different sizes and concentrations. Without this normalization, a large protein at high concentration would show a larger raw ellipticity than a small protein at low concentration, even if their secondary structures were identical. Mean residue ellipticity removes these confounding variables.

### Can CD tell me the exact 3D structure of a protein?

No. CD spectroscopy provides global estimates of secondary structure content and a qualitative fingerprint of tertiary structure, but it cannot determine the three-dimensional arrangement of atoms. CD cannot tell you which residues are in which secondary structure, how the secondary structure elements are packed, or the overall fold of the protein. For high-resolution structures, you need [X Ray Crystallography](/knowledge/molecular-biology/x-ray-crystallography), NMR spectroscopy, or cryo-electron microscopy. CD is best used as a rapid, solution-phase method for estimating secondary structure, monitoring conformational changes, and comparing protein states.

### What is the difference between far-UV and near-UV CD?

Far-UV CD (180–250 nm) probes the peptide backbone amide chromophore and reports on secondary structure (α-helix, β-sheet, random coil). Near-UV CD (250–350 nm) probes the aromatic side chains (phenylalanine, tyrosine, tryptophan) and disulfide bonds, reporting on the asymmetric environment of these groups and hence on tertiary structure. Far-UV signals are relatively intense and require low protein concentrations (0.1–0.5 mg/mL) and short pathlengths (0.1 cm). Near-UV signals are much weaker and require higher concentrations (0.5–2.0 mg/mL) and longer pathlengths (0.5–1.0 cm).

### Why is my CD spectrum noisy at low wavelengths?

Noise below 200 nm is primarily due to absorption by the buffer, solvent, and oxygen. Water absorbs strongly below 190 nm, and common buffer components such as phosphate, Tris, and chloride also absorb in this region. This absorption reduces the light intensity reaching the detector, increasing the shot noise. The high-tension voltage of the photomultiplier tube rises as the light intensity falls, and when it exceeds approximately 600 V, the data are unreliable. To reduce noise, use a buffer with minimal absorbance, purge the instrument with nitrogen to remove oxygen, and use a shorter pathlength.

## Key Takeaways

- A CD graph plots ellipticity (y-axis) versus wavelength (x-axis); far-UV (180–250 nm) reports on secondary structure, while near-UV (250–350 nm) reports on tertiary structure.
- The α-helix signature is a double minimum at 208 and 222 nm with a positive band near 192 nm; the β-sheet signature is a minimum near 218 nm with a maximum near 195 nm; random coil shows a strong minimum near 200 nm.
- Raw ellipticity in millidegrees must be converted to mean residue ellipticity ([θ], deg·cm²·dmol⁻¹) using the formula [θ] = (θ_obs × 100) / (c × l × n) for quantitative comparison.
- Secondary structure content can be estimated using deconvolution algorithms (SELCON3, CDSSTR, CONTINLL), but these provide global averages accurate to only ±5–10%, not residue-specific information.
- Buffer components, concentration errors, and noise below 200 nm are the most common sources of artifacts; always subtract a buffer blank and check the high-tension voltage.
- Near-UV CD is a sensitive probe for tertiary structure integrity; loss of signal indicates unfolding or loss of rigid packing.
- CD is a comparative and monitoring tool, not a high-resolution structural method; use it alongside techniques like [X Ray Crystallography](/knowledge/molecular-biology/x-ray-crystallography) or NMR for detailed structural determination.

## Related Clinical & Scientific Guides

* [MAPK Pathway: Mechanism, Function, and Clinical Relevance](/knowledge/molecular-biology/mapk-pathway)
* [Mammalian Cell Culture Bioreactors: A Practical Guide](/knowledge/molecular-biology/mammalian-cell-culture-bioreactor)
* [Nucleotide Formation: Biosynthesis and Assembly of DNA/RNA Building Blocks](/knowledge/molecular-biology/nucleotide-formation)