Differential scanning fluorimetry (DSF) is one of the most useful tools for conformational stability assessment, and nano differential scanning fluorimetry (nanoDSF) is its label-free variant. Where classical DSF relies on an extrinsic dye whose fluorescence increases as it binds to exposed hydrophobic patches, nanoDSF uses the protein's own intrinsic fluorescence — chiefly from tryptophan and tyrosine residues. The advantage is that nothing is added: there is no dye to perturb the sample, no dye–protein interaction to confound the signal, and no interference from coloured or turbid formulations. The trade-off is that the readout depends entirely on the aromatic content and local environment of the protein being studied, which shapes both what the technique can measure and how its results should be interpreted.
The principle: what intrinsic fluorescence actually reports
A temperature ramp is applied to a protein solution, and the intrinsic fluorescence is monitored continuously. In the folded state, tryptophan residues are typically buried within the hydrophobic core, where the local environment quenches and shifts their emission. As the protein unfolds, those residues become exposed to solvent; the emission intensity drops and the peak position shifts (a red shift). The relationship between the change in fluorescence signal and the temperature gradient is used to obtain the apparent melting temperature (Tm) — the point at which half of the protein, or of the most temperature-sensitive domain, is unfolded — and the onset temperature (Tonset), the temperature at which the first conformational changes begin.
Tonset is the more sensitive early indicator of instability and is often the more discriminating parameter when comparing closely related formulations. Tm is the more robust, widely reported value. Because both are derived from the same scan, nanoDSF reports them together, usually alongside a scattering channel (backreflection) that flags aggregation as it occurs.
Monoclonal antibodies unfold in domains
For monoclonal antibodies (mAbs) and related formats, the unfolding transition is not a single event: the Fab and Fc regions, and the individual domains within them, melt at different temperatures. This produces a multi-transition thermogram rather than a single sigmoidal curve. The most temperature-sensitive domain is normally the CH2 domain of the Fc, which melts first and is therefore the transition most commonly used for the reported Tm. Because the CH2 domain sits at the interface that governs Fc effector function and much of the molecule's aggregation propensity, its stability is a meaningful proxy for overall conformational robustness. When interpreting a mAb thermogram, it is essential to know which transition the reported value corresponds to; a "Tm of 62 °C" that refers to CH2 is not comparable to one that refers to the Fab.
Scope of this protocol
This protocol describes a method to measure the apparent melting temperature (Tm) and onset temperature (Tonset) by nanoDSF in order to assess the conformational stability of unstressed proteins. Stir-stressed and light-stressed samples are not analysed by DSF; those stress modalities are assessed by other methods.
Important notes
- This protocol is not a Standard Operating Procedure. It assumes the required instrument is in good working order and that the analysis is performed by experienced user(s).
- Required calibration or check standards should be run according to the manufacturer's instructions.
- This protocol does not contain every detail of the analysis. The analyst should rely on their best judgment, routine practices, and knowledge of the technique; this document is a guideline, not a recipe.
- Samples should be analysed immediately after thawing and preparation.
Equipment and materials
- NanoDSF device.
- Single-use sample containers (e.g. capillaries for the Prometheus, UNi for the UNcle).
- Protein samples and formulation buffers supplied by the sample originators — buffers and aliquots of unstressed samples from Shipping #1 only.
- Pipettes and pipette tips.
Reagents and solutions
Samples and their buffers are provided by the sample originators and are nominally at 1 mg/mL. The two reference proteins used here are an Amgen IgG2 in 10 mM sodium acetate, pH 5.0, and NISTmAb (an IgG1 reference material) in 12.5 mM L-histidine / 12.5 mM L-histidine HCl, pH 6.0. Both are used as unstressed aliquots at 1 mg/mL.
| Protein | Formulation | Sample |
|---|---|---|
| Amgen IgG2 | 10 mM sodium acetate, pH 5.0 | Unstressed, 1 mg/mL |
| NISTmAb IgG1 | 12.5 mM L-histidine / 12.5 mM L-histidine HCl, pH 6.0 | Unstressed, 1 mg/mL |
Procedures
Instrument setup and sample preparation
- Thaw the samples from Shipping #1 and aliquot the Tier 2 vials (unstressed only) as described in the instructions into smaller aliquots for the various methods, then freeze until ready to use. Follow the procedure for vigorously mixing the sample prior to aliquoting.
- For this analysis, thaw the aliquots designated for nanoDSF measurements (prepared and frozen in step 1) by bringing them to room temperature for no more than 15 minutes.
- Using a pipette, fill the sample container with the volumes suggested by the manufacturer (e.g. 9 µL for UNi; 10 µL for Prometheus capillaries). No dilution is required. Use a maximum of 8 samples in triplicate per run — up to 24 wells/capillaries per run. Note: the 24-well/capillary ceiling exists to ensure a good and constant data-point density. Make sure there are no bubbles and that the sample order matches the one entered in the software.
- Insert the sample container into the support provided by the manufacturer, then insert it into the nanoDSF device.
- When setting up the method, use the parameters in the table below and start the measurements.
Data acquisition
- Review the nanoDSF data template for the requested information before beginning the measurements, and input the setup information.
- Use the following experimental parameters to collect the data (default of the UNcle – quick start mode):
| Parameter | Setting |
|---|---|
| Number of samples | Up to 8 samples in triplicate per run, i.e. up to 24 wells/capillaries per run |
| Temperature ramp | Linear, 1 °C/min |
| Equilibration time | 5 min |
| Temperature range | 20–95 °C |
| Sample concentration range | 1–150 mg/mL — no dilution required |
| Fluorescence settings | Ex: 266 nm; Em: 300–450 nm (or equivalent device default) |
- Perform a discovery scan if necessary.
- Attribute the sample names in the software and proceed to the melting scan.
Understanding the results
- Data and selected instrument parameters should be entered in the nanoDSF data reporting template.
- Use the software default method to calculate Tonset and Tm automatically (e.g. the barycentric mean [BCM] over 300–450 nm for the UNcle).
- Average the temperatures of the triplicates to obtain the final values, and use the standard deviation to report the measurement variability.
Troubleshooting
- Ensure there are no bubbles in the UNi or capillaries — trapped air disrupts the optical path and produces erratic baselines.
- Ensure the sample holder is properly seated in the instrument — a misaligned holder shifts the readout and can invalidate the run.
- Watch the scattering/backreflection channel. A rise in scattering during the scan indicates aggregation occurring before or during unfolding; if it precedes the expected transition, the sample may already be compromised and the apparent Tm should be treated with caution.
- Mind buffer matching. Because the signal is intrinsic and small, a mismatch between sample and reference buffer can distort the baseline; use the formulation buffer supplied with the sample as the blank where the software allows it.
- Keep the ramp rate fixed. Tm is scan-rate dependent; never compare values collected at different ramp rates.
How to report
Report Tm and Tonset as the mean ± standard deviation of the triplicates, state the protein concentration, the formulation buffer, the ramp rate (1 °C/min), the scan range (20–95 °C), and the analysis method used to derive the values (e.g. BCM 300–450 nm). For mAbs, state explicitly which transition the reported Tm corresponds to — in most cases the CH2 domain. Note any aggregation observed on the scattering channel, and record whether a discovery scan preceded the melting scan. Transparency about these parameters is what makes a nanoDSF value comparable across labs and interpretable years later.
Further information
For specific questions regarding this method, contact the study coordinator, and copy the NIST interlaboratory study address on correspondence. This protocol originates from an interlaboratory comparison of biophysical methods for monoclonal antibody characterisation, and the values it produces are most useful when the reporting conventions above are followed consistently.
References
- Pantoliano MW, Petrella EC, Kwasnoski JD, et al. High-density miniaturized thermal shift assays as a general strategy for drug discovery. J Biomol Screen. 2001;6(6):429–440. doi:10.1177/108705710100600609
- Niesen FH, Berglund H, Vedadi M. The use of differential scanning fluorimetry to detect ligand interactions that promote protein stability. Nat Protoc. 2007;2(9):2212–2221. doi:10.1038/nprot.2007.321
- Boivin S, Kozak S, Meijers R. Optimization of protein purification and characterization using Thermofluor screens. Protein Expr Purif. 2013;91(2):192–206. doi:10.1016/j.pep.2013.08.002
- Alexander CG, Wanner R, Johnson CM, et al. Novel microscale approaches for easy, rapid determination of protein stability in academic and commercial settings. Biochim Biophys Acta. 2014;1844(12):2241–2250. doi:10.1016/j.bbapap.2014.09.016
- Cohrs LB, Davy AMJ, Van Ackere T, De Smedt SC, Braeckmans K, Epe B, et al. Intrinsic differential scanning fluorimetry for protein stability assessment in biopharmaceutical development. Mol Pharm. 2025;22(3):1697–1706. doi:10.1021/acs.molpharmaceut.4c01496
- Senisterra G, Chau I, Vedadi M. Thermal denaturation assays in chemical biology. Assay Drug Dev Technol. 2012;10(2):128–136. doi:10.1089/adt.2011.0390
- Schlebach JP, Narayan M. Membrane (and soluble) protein stability and binding measurements in the lipid cubic phase using label-free differential scanning fluorimetry. Anal Chem. 2018;90(20):12152–12160. doi:10.1021/acs.analchem.8b03176
- Garidel P, Hegyi M, Bassarab S, Weichel M. A rapid, sensitive and economical assessment of monoclonal antibody conformational stability by intrinsic tryptophan fluorescence spectroscopy. Biotechnol J. 2008;3(9-10):1201–1211. doi:10.1002/biot.200800091
- De Leoz MLA, Duewer DL, Stein SE, et al. NIST Interlaboratory Study on the Glycosylation of NISTmAb, a Monoclonal Antibody Reference Material. NIST IR 8186. National Institute of Standards and Technology; 2017. doi:10.6028/nist.ir.8186
- Martin A, Maschberger M, Breitsprecher D. Thermal unfolding of antibodies with the Prometheus NT.48. BioTechniques. 2015;59(4):251. doi:10.2144/000114347