Controlling Phosphorothioate Diastereomers in Oligonucleotide R&D
Replacing a non-bridging oxygen atom in the phosphodiester backbone with sulfur creates a phosphorothioate (PS) linkage, one of the most widely used chemical modifications in therapeutic oligonucleotides. PS modification can improve nuclease resistance, protein binding, tissue distribution, and other pharmacokinetic properties.
However, the same modification introduces a chiral center at each phosphorus atom. For an oligonucleotide containing $$ (n) $$stereogenic PS linkages, the theoretical number of stereoisomers can reach $$ (2^n) $$. As the number of PS linkages increases, the resulting diastereomeric complexity creates substantial challenges for structural characterization, analytical method development, process control, and quality assurance.
From a CMC perspective, stereochemical control is not limited to analytical testing. The synthesis process itself can influence the resulting diastereomeric distribution. In particular, the choice of phosphoramidite synthesis activator can introduce a reproducible stereochemical bias. Maintaining the same activator strategy across development and manufacturing is therefore important for achieving batch-to-batch comparability.
Purification can also alter the observed stereochemical profile. In particular, fraction-collection windows and cut points can influence the relative distribution of diastereomers in the collected product. Consequently, process parameters that affect stereochemical composition should be appropriately controlled across chemical R&D, manufacturing, and quality assurance.
A robust control strategy therefore requires two complementary elements:
- Process control to minimize uncontrolled changes in stereochemical composition.
- Fit-for-purpose analytical methods to characterize, compare, and continuously monitor the resulting diastereomeric profile.
This article examines these strategies from two analytical perspectives: spectroscopic characterization and chromatographic separation.
🧪 Regulatory Considerations #
As development of oligonucleotide therapeutics has expanded, regulatory expectations for characterization and control of stereochemical attributes have also become increasingly important.
Regulatory considerations should be evaluated in the context of the specific product, manufacturing process, analytical method, and jurisdiction. Guidance and expectations from agencies such as the European Medicines Agency (EMA) and China’s National Medical Products Administration (NMPA) provide relevant frameworks for demonstrating control of critical quality attributes.
For phosphorothioate oligonucleotides, stereochemical characterization can become particularly important when changes in synthesis chemistry, purification, scale, or raw materials have the potential to alter the diastereomeric distribution.
A practical CMC strategy should therefore establish:
- Appropriate characterization of the stereochemical profile.
- Defined analytical procedures for detecting meaningful changes.
- Controls for synthesis parameters that influence stereochemistry.
- Controls for purification and fraction-collection parameters.
- Appropriate reference standards and comparability criteria.
- Scientifically justified acceptance limits and tolerance intervals.
The analytical strategy should ultimately demonstrate that changes in stereochemical composition are understood, detectable, and controlled throughout the product lifecycle.
📐 Spectroscopic Characterization #
Spectroscopic techniques can provide powerful orthogonal approaches for monitoring phosphorothioate diastereomers.
Rather than relying exclusively on assignment of every individual diastereomer, spectral-comparison approaches can quantify the similarity between a reference standard and a test sample. Metrics such as Weighted Spectral Difference (WSD), the Pearson correlation coefficient, and Relative Peak Area Difference (RPAD) can be combined with method-specific detection limits and tolerance intervals to establish quantitative control strategies.
Two particularly relevant techniques are:
- Circular Dichroism (CD) spectroscopy for evaluating conformational and stereochemical differences.
- Phosphorus-31 Nuclear Magnetic Resonance ($(^{31}\mathrm{P})-NMR$) for directly probing the chemical environments of phosphorus atoms in PS linkages.
Weighted Spectral Difference (WSD) #
Weighted Spectral Difference (WSD) is a quantitative metric designed to measure the difference in the overall shape of two spectral curves, such as a reference standard and a test sample.
A generalized form can be expressed as:
$$ [ $$\mathrm{WSD}$$ #
$$\sqrt{ \frac{1}{n} \sum_{i=1}^{n} w_i(A_i-B_i)^2 } ] $$
where:
- $$ (A_i)$$ is the reference signal at spectral point (i).
- $$(B_i)$$ is the test-sample signal at spectral point (i).
- $$(w_i)$$ is the weighting factor assigned to point (i).
- $$(n)$$ is the total number of spectral points.
The calculation can be understood in three steps.
1. Calculate point-by-point spectral differences #
The term
[ (A_i-B_i)^2 ]
represents the squared difference between the reference and sample signals at each spectral point.
Squaring the difference prevents positive and negative deviations from canceling each other and increases the contribution of larger deviations.
2. Apply weighting #
The weighting factor (w_i) allows selected spectral regions to contribute differently to the overall metric.
This can be useful when some regions contain stronger or more informative signals, while other regions are dominated by baseline noise or have limited analytical significance.
3. Average and calculate the root #
The weighted differences are averaged across the measured spectral range and the square root is taken to generate a composite similarity metric.
For oligonucleotide analysis, WSD can be used to evaluate overall CD spectral consistency, including changes associated with conformation and phosphorothioate diastereomeric composition.
Pearson Correlation Coefficient #
The Pearson correlation coefficient, commonly represented by (r) or (\rho), quantifies the degree of linear correlation between two continuous variables.
It can be expressed as:
[ \rho = \frac{ \sum_{i=1}^{n} (x_i-\bar{x})(y_i-\bar{y}) }{ \sqrt{ \sum_{i=1}^{n}(x_i-\bar{x})^2 } \sqrt{ \sum_{i=1}^{n}(y_i-\bar{y})^2 } } ]
where:
- (x_i) represents the reference-standard signal at point (i).
- (y_i) represents the sample signal at point (i).
- (\bar{x}) and (\bar{y}) are the corresponding mean signal values.
- (n) is the number of measured points.
For spectral-comparison applications, parallel measurements of a reference standard can be used to establish the method’s baseline repeatability.
The resulting distribution of reference-to-reference comparisons can then provide a basis for defining statistically justified acceptance regions for sample-to-reference comparisons.
This approach is applicable to multiple spectroscopic platforms, including NMR, infrared spectroscopy, Raman spectroscopy, and mass spectrometry, provided that the underlying signal representation and preprocessing are appropriately controlled.
Relative Peak Area Difference (RPAD) #
Relative Peak Area Difference (RPAD) provides a quantitative approach for comparing the relative distribution of resolved diastereomeric peaks.
A simple formulation is:
[ \mathrm{RPAD} #
\sum_{i=1}^{n} \left| p_{i,\mathrm{sample}} #
p_{i,\mathrm{ref}} \right| ]
where:
- (p_{i,\mathrm{sample}}) is the normalized area percentage of the (i)-th peak in the sample.
- (p_{i,\mathrm{ref}}) is the normalized area percentage of the corresponding peak in the reference.
- (n) is the number of resolvable peaks included in the comparison.
For example, a fragment containing two PS linkages has a theoretical maximum of four stereochemical configurations, assuming both linkages are independently stereogenic and all configurations are distinguishable.
RPAD is particularly useful when the analytical method resolves multiple stereochemical components but does not provide unambiguous absolute configuration assignments for every peak.
It can therefore serve as an overall stereochemical-profile comparability metric for methods such as enzymatic digestion followed by LC-MS or multi-peak AEX and IP-RP chromatographic analysis.
Spectral Limit of Quantitation (LOQ) #
In this context, the term LOQ requires careful definition.
Rather than representing the conventional concentration-based LOQ for an analyte, the proposed approach uses LOQ to describe the method’s ability to detect a defined change in the stereochemical profile.
A reference standard can first be analyzed through multiple parallel measurements to establish the baseline variability of the similarity metric. Variant or deliberately modified samples can then be compared with the reference to determine the magnitude of detectable profile change.
A conceptual calculation can be expressed as:
[ \mathrm{LOQ} #
\frac{ 10\times\sigma_{\mathrm{ref}} }{ \left| \bar{x}_{\mathrm{doped}} #
\bar{x}_{\mathrm{ref}} \right| } ]
where:
- (\sigma_{\mathrm{ref}}) represents the standard deviation of repeated reference-standard measurements.
- (\bar{x}_{\mathrm{doped}}) represents the mean result from repeated measurements of the intentionally varied or spiked sample.
- (\bar{x}_{\mathrm{ref}}) represents the mean result from repeated reference-standard comparisons.
The exact definition and calculation should be established according to the intended use of the method and its validation protocol. The terminology should also be clearly distinguished from the conventional analyte-concentration definition of LOQ to avoid ambiguity in regulatory documentation.
Tolerance Intervals #
Analytical control limits require an appropriate statistical treatment of the underlying data.
Spectral similarity metrics do not necessarily follow a normal distribution. Consequently, simply defining an acceptance range as:
[ \mathrm{mean}\pm k\times\mathrm{SD} ]
may not always be statistically appropriate.
The Wilks method provides a classical non-parametric approach for constructing tolerance intervals and can be useful when sample sizes are limited or the distributional assumptions required by parametric methods are not satisfied.
For a study designed to achieve, for example, 90% confidence that the interval covers 95% of the population, ranked observations can be used to determine appropriate upper and lower tolerance limits according to the relevant statistical tables or formulations.
The selected confidence level, population coverage, sample size, and statistical method should be predefined as part of the analytical validation or method-control strategy.
Circular Dichroism (CD) Spectroscopy #
Circular Dichroism (CD) spectroscopy measures the differential absorption of left- and right-circularly polarized light by chiral molecules.
In a simplified optical description, light from the source is first polarized and subsequently converted into circularly polarized components. When these components interact with a chiral sample, differences in absorption produce a measurable CD signal.
The resulting spectrum provides information about molecular chirality and conformation and can serve as a characteristic spectral fingerprint.
For nucleic acids, CD is widely used to investigate higher-order structure and conformation. In phosphorothioate oligonucleotides, the introduction of PS linkages creates (R_p/S_p) stereocenters at phosphorus.
Different stereochemical configurations can influence:
- Backbone torsion angles.
- Base-stacking interactions.
- Local molecular conformation.
- Chiral microenvironments.
- The interaction of the molecule with polarized light.
These differences can generate measurable changes in CD spectra, making CD potentially useful for monitoring changes in the overall stereochemical and conformational profile.
Activator-Dependent Stereochemical Effects #
Nicholas R. et al. investigated phosphorothioate-modified antisense oligonucleotides prepared using different synthesis activators, including DCI and BTT.
Through parallel testing and statistical comparison, the study evaluated the influence of activator selection on diastereomeric composition and batch consistency.
CD measurements indicated differences in the spectral profiles of samples prepared using different activator strategies. Quantitative analysis using WSD was used to evaluate these differences, with the reported study establishing a corresponding detection threshold of approximately 3.8% under its experimental conditions.
The broader CMC implication is that activator selection can become a stereochemistry-relevant process parameter. If the activator affects the resulting diastereomeric distribution, maintaining a consistent synthesis strategy becomes important for ensuring batch comparability.
Phosphorus-31 NMR #
Phosphorus-31 Nuclear Magnetic Resonance ((^{31}\mathrm{P})-NMR) provides a complementary approach because the chemical shift of phosphorus is highly sensitive to its local chemical environment.
For phosphorothioate oligonucleotides, the observed chemical environment can be influenced by:
- (R_p/S_p) configuration.
- Neighboring nucleotide sequence.
- Backbone chemistry.
- Chemical modifications.
- Local electronic and shielding effects.
Consequently, different phosphorothioate configurations can produce distinguishable resonance patterns in (^{31}\mathrm{P})-NMR spectra.
High-field (^{31}\mathrm{P})-NMR can therefore support qualitative stereochemical differentiation and, under appropriately validated conditions, semi-quantitative assessment of the overall stereochemical profile.
In work by Nicholas R. et al. involving antisense oligonucleotides synthesized using different activators, model-compound studies were used to support assignment of two major resonance regions in full-length ASOs. The reported chemical-shift ranges associated with the assignments were approximately 60.0–56.7 ppm for (R_p) and 56.7–53.0 ppm for (S_p).
Such assignments are method- and system-dependent and should be established using appropriate reference compounds and experimental conditions rather than assumed universally across sequences.
🔬 Chromatographic Separation and Characterization #
Spectroscopy provides a powerful means of monitoring overall stereochemical changes, but chromatographic methods offer a complementary capability: physical separation of stereochemical variants.
For phosphorothioate oligonucleotides, relevant separation platforms include:
- Ion-Pair Reversed-Phase High-Performance Liquid Chromatography (IP-RP-HPLC).
- Anion-Exchange Chromatography (AEX-HPLC).
- Capillary Electrophoresis (CE).
- Enzymatic digestion followed by LC-MS.
These techniques exploit differences in charge, hydrophobicity, molecular conformation, and effective hydrodynamic properties.
Combining chromatographic separation with UV or MS detection can provide both qualitative and quantitative information about the stereochemical composition of an oligonucleotide sample.
Anion-Exchange Chromatography #
Oligonucleotides contain negatively charged phosphate backbones, making Anion-Exchange Chromatography (AEX) a natural orthogonal separation mechanism.
In AEX, negatively charged oligonucleotides interact with positively charged ligands on the stationary phase. Differences in molecular charge and charge density influence retention and therefore enable separation of chemically distinct species.
AEX is particularly useful for monitoring process- and product-related impurities such as:
- PO/PS backbone variants.
- (n-1) and (n+1) length variants.
- Depurination products.
- Deamination products.
- Other charge-related impurities.
Because phosphorothioate and phosphodiester linkages have different chemical properties, changes in backbone composition can influence chromatographic retention.
Likewise, truncated oligonucleotides have different numbers of charged groups from the full-length product, producing differences in interaction with the stationary phase.
AEX therefore serves as a useful orthogonal analytical method alongside IP-RP chromatography, particularly when broader impurity profiling is required.
Capillary Electrophoresis #
Although Capillary Electrophoresis (CE) is less frequently used than liquid chromatography for therapeutic oligonucleotide analysis, it provides complementary separation selectivity.
Two important CE modes are Capillary Gel Electrophoresis (CGE) and Capillary Zone Electrophoresis (CZE).
Capillary Gel Electrophoresis #
CGE uses a polymeric or gel-like sieving matrix to separate molecules primarily according to size and chain length.
This makes CGE particularly useful for detecting length-related impurities, including:
- (n-1) products.
- (n+1) variants.
- Shorter fragments.
- Other chain-length-related impurities.
Its high size-based resolution can complement chromatographic methods that rely more strongly on charge or hydrophobicity.
Capillary Zone Electrophoresis #
CZE operates without a molecular sieving gel and separates species according to their electrophoretic mobility.
Factors affecting migration can include:
- Net molecular charge.
- Charge density.
- Molecular conformation.
- Hydrodynamic properties.
This makes CZE potentially useful for investigating charge variants, PS-related variants, and other polar impurities.
CGE and CZE therefore provide different but complementary separation mechanisms and can expand electrophoretic coverage for oligonucleotide purity and impurity characterization.
Ion-Pair Reversed-Phase Chromatography #
Ion-Pair Reversed-Phase (IP-RP) chromatography is one of the most widely used chromatographic approaches for oligonucleotide analysis.
Ion-pairing reagents such as triethylammonium acetate (TEAA) and tributylammonium acetate (TBAA) influence the balance between electrostatic interactions and hydrophobic interactions with the stationary phase.
The choice of ion-pairing reagent, organic modifier, stationary-phase chemistry, temperature, and gradient can therefore have a substantial effect on retention and resolution.
Common reversed-phase stationary phases include:
- C4.
- C8.
- C18.
- C30.
- Phenyl.
Among these, C18 phases are widely used for oligonucleotide IP-RP analysis.
Bagge et al. systematically investigated IP-RP chromatographic separation of phosphorothioate oligonucleotide diastereomers using 5–20-mer single-stranded DNA sequences containing four PS linkages.
The study compared Kromasil C4, C8, C18, and phenyl columns with 2.5 µm particles. Under the reported conditions, the C18 column provided the strongest resolving performance among the evaluated stationary phases.
For particularly complex stereochemical mixtures, multidimensional chromatographic strategies can provide additional resolving power. For example, combining a chiral reversed-phase separation with C18 IP-RP chromatography in a comprehensive two-dimensional liquid chromatography (2D-LC) workflow can increase the separation space available for complex diastereomeric mixtures.
🧬 Enzymatic Digestion Coupled with Mass Spectrometry #
Nuclease P1 (NP1) digestion coupled with LC-MS provides another powerful strategy for investigating phosphorothioate stereochemistry.
Rather than attempting to resolve every diastereomer in the intact full-length oligonucleotide, the method uses enzymatic digestion to generate shorter fragments. This reduces chromatographic complexity while retaining localized information about stereochemical differences.
The workflow can be summarized in three stages.
1. Stereoselective Enzymatic Digestion #
NP1 is used to digest the oligonucleotide into shorter fragments containing PS linkages.
By reducing the number of stereogenic centers present in each analytical fragment, the approach simplifies the theoretical stereoisomer space. Fragments containing one or two PS linkages can be substantially easier to resolve and characterize than the corresponding full-length molecule.
2. Chromatographic Separation #
The resulting fragments are separated using a chromatographic method such as IP-RP-HPLC.
Diastereomeric fragments can exhibit differences in polarity, conformation, and interactions with the stationary and mobile phases. These differences can produce distinct chromatographic retention times and partially or fully resolved peaks.
3. Mass Spectrometric Detection #
Mass spectrometry provides an additional dimension of structural confirmation.
MS1 can establish that chromatographically separated components have the expected molecular mass, helping distinguish stereochemical variants from unrelated mass variants. Retention time and normalized peak area can then be used to characterize the relative distribution of the detected isomeric fragments.
The combination of enzymatic digestion, chromatography, and MS is particularly valuable when the objective is to determine whether stereochemical differences are localized to specific regions of an oligonucleotide sequence.
Li, Z. et al. used NP1 digestion coupled with LC-MS to identify 11 fragment types, achieving approximately 95% sequence coverage under the reported analytical conditions.
Nicholas R. et al. similarly used NP1/LC-MS to compare oligonucleotides synthesized using different activator strategies. Their work demonstrated that the method can provide sequence-region-specific information about changes in diastereomeric distribution.
However, NP1/LC-MS is not necessarily a universal plug-and-play method. Significant method development may be required to optimize:
- Enzyme selection.
- Enzyme concentration.
- Digestion time.
- Digestion temperature.
- Buffer composition.
- Substrate-to-enzyme ratio.
- Chromatographic conditions.
- MS detection parameters.
These variables can affect digestion efficiency, fragment coverage, peak resolution, and quantitative reproducibility.
🏭 Integrating Stereochemical Control into CMC #
The analytical techniques described above address different aspects of phosphorothioate stereochemistry.
CD and (^{31}\mathrm{P})-NMR are useful for evaluating overall spectral or structural differences. AEX, IP-RP-HPLC, and CE provide chromatographic or electrophoretic separation of chemically distinct species. NP1/LC-MS adds sequence-localized information by reducing complex full-length molecules into analytically tractable fragments.
No single method necessarily provides complete stereochemical characterization across every oligonucleotide sequence.
A practical control strategy can therefore combine complementary methods:
| Analytical objective | Potential approach |
|---|---|
| Overall conformational and stereochemical consistency | CD |
| Phosphorus chemical-environment characterization | (^{31}\mathrm{P})-NMR |
| Broad impurity and charge-variant profiling | AEX-HPLC |
| Diastereomer and hydrophobic-variant separation | IP-RP-HPLC |
| Size and length-variant characterization | CGE |
| Charge and conformation-based electrophoretic separation | CZE |
| Sequence-region-specific stereochemical profiling | NP1/LC-MS |
| Quantitative overall profile comparison | WSD, Pearson correlation, RPAD |
The key point is that stereochemical control begins with process design rather than analytical testing alone.
If activator selection influences the (R_p/S_p) distribution, the activator becomes an important process parameter. If purification fraction boundaries influence which stereochemical populations are retained, collection cut points can also become relevant process parameters.
These variables should therefore be considered during process development and carried consistently into manufacturing.
🔗 Connecting Process Parameters with Analytical Control #
A robust phosphorothioate stereochemical-control strategy can be viewed as a chain connecting synthesis → purification → characterization → specification.
During synthesis, activator selection and reaction conditions can influence stereochemical composition. During purification, chromatographic selectivity and fraction-collection windows can enrich or deplete specific stereochemical populations.
Analytical methods then provide the evidence needed to determine whether the resulting profile remains within the predefined control strategy.
This creates several important CMC considerations:
- Define the stereochemical profile during development.
- Identify synthesis parameters that influence stereochemical composition.
- Control activator selection and relevant reaction conditions.
- Evaluate purification conditions and fraction-collection cut points.
- Establish suitable reference standards.
- Validate analytical methods for their intended purpose.
- Define statistically justified comparability or acceptance criteria.
- Monitor stereochemical consistency across development and manufacturing batches.
The objective is not necessarily to assign every individual (R_p/S_p) configuration in a highly stereochemically complex oligonucleotide. In many development settings, a more practical objective is to demonstrate that the overall stereochemical profile is controlled, reproducible, analytically detectable, and consistent with the established reference state.
🧩 Conclusion #
Phosphorothioate modification provides important pharmaceutical advantages for therapeutic oligonucleotides, but each PS linkage also introduces a phosphorus stereocenter. As the number of PS linkages increases, the resulting diastereomeric complexity can become a significant analytical and manufacturing challenge.
Effective control therefore requires more than a single characterization method.
CD and (^{31}\mathrm{P})-NMR can provide orthogonal spectroscopic information about stereochemical and conformational differences. IP-RP-HPLC, AEX-HPLC, and CE can provide separation-based characterization, while NP1 digestion coupled with LC-MS can provide localized information about stereochemical changes within specific sequence regions.
At the process level, activator selection and purification fraction-collection parameters can be important contributors to stereochemical consistency. These parameters should therefore be evaluated and appropriately controlled throughout chemical development, manufacturing, and quality assurance.
Ultimately, a fit-for-purpose combination of process controls, orthogonal analytical methods, statistical comparability tools, and predefined acceptance criteria provides a practical framework for maintaining phosphorothioate stereochemical consistency and supporting robust oligonucleotide CMC development.