Oligonucleotide Purification: Methods and Selection Strategy

Oligonucleotide purification is not a single technique but a menu of methods, each with distinct strengths depending on the type of oligonucleotide, the impurities present, and the end use. A short PCR primer headed for a routine experiment can get away with simple desalting, while a therapeutic antisense drug destined for clinical trials demands multi-step chromatographic purification to reach purities above 95%. The challenge is that synthesis generates a complex mix of truncated sequences, chemical modifications gone wrong, and residual reagents, and no single purification approach handles all of them equally well.

What Synthesis Leaves Behind

Solid-phase oligonucleotide synthesis builds a chain one nucleotide at a time, and each coupling step falls slightly short of 100% efficiency. Over a 20-mer, even a 99% per-step yield means roughly one in five chains will have at least one error. The result is a crude product that contains the full-length target sequence mixed with shorter “failure sequences” that terminated early. Removing these truncated by-products is the primary reason purification exists at all, and it is required for applications ranging from PCR primers to DNA sequencing probes.

Beyond simple truncations, synthesis introduces subtler impurities. High-resolution mass spectrometry studies have identified failure sequences carrying terminal phosphate groups, incomplete backbone sulfurization products, desulfurization artifacts, high-molecular-weight adducts, and various chemical modifications left over from protecting-group chemistry.¹ For therapeutic oligonucleotides, these low-level impurities matter because even trace amounts of an unintended chemical species can affect safety or efficacy. The multistep nature of both synthesis and purification means that degradation products can also form during processing itself, adding yet another layer to the impurity profile.²

Desalting as a Baseline

Every oligonucleotide synthesis leaves behind salts and small-molecule reagents that interfere with downstream work. Desalting is the simplest form of purification, but it does not remove failure sequences. It just strips away salts and low-molecular-weight contaminants, giving you a cleaner full-length product mixed with cleaner truncated junk. For many routine laboratory applications, that is good enough.

The traditional approach uses alcohol precipitation with a salt promoter like sodium acetate. Recent work has shown that ammonium acetate combined with ethanol or isopropanol can handle oligonucleotide samples containing very high salt loads, reducing sodium ion levels by two to three orders of magnitude.³ In manufacturing settings, tangential flow filtration has largely replaced precipitation for isolation and desalting of therapeutic oligonucleotides, a shift that happened over the past decade or so as techniques originally developed for biologics were adapted for the smaller nucleic acid molecules.⁴ The advantage of tangential flow filtration is scalability and reproducibility, though it requires optimization for each oligonucleotide type to maximize yield and avoid co-precipitating salts.

Polyacrylamide Gel Electrophoresis

Denaturing PAGE separates oligonucleotides by size with single-nucleotide resolution, which makes it excellent at removing failure sequences. The technique is fast, simple to set up, and resolves full-length product from shorter truncations very cleanly. The tradeoff is yield: recoveries tend to stay below 50% of the material loaded onto the gel.⁵ For most molecular biology work, the amount recovered is still more than enough for cloning, sequencing, or hybridization experiments. PAGE becomes impractical, though, when you need milligram or gram quantities, since it does not scale well and the gel-extraction step introduces handling losses.

PAGE is particularly useful as a quality check or a secondary purification for oligonucleotides that need to be very clean but are not being produced in large volumes. It is less common in therapeutic manufacturing, where chromatographic methods dominate.

Ion-Pair Reversed-Phase HPLC

Ion-pair reversed-phase chromatography (IP-RP) is one of the two workhorses of oligonucleotide purification, alongside anion exchange. The principle is straightforward: you add an ion-pairing agent to the mobile phase that associates with the negatively charged backbone of the oligonucleotide, making the complex hydrophobic enough to interact with a standard C18 reversed-phase column. The oligonucleotide then elutes based on a combination of its length and its hydrophobicity.

The classic ion-pairing system uses triethylammonium acetate, which works well but produces separations that are somewhat sequence-dependent, meaning two oligonucleotides of the same length but different base compositions may co-elute or resolve differently than expected. An alternative system using triethylamine buffered with hexafluoroisopropanol provides more uniform, less sequence-dependent separations.⁶ This hexafluoroisopropanol-based system has become popular for analytical LC-MS work and is increasingly used at preparative scale as well.

IP-RP excels when the target is a relatively short, unmodified or lightly modified oligonucleotide and when the main goal is removing truncated failure sequences. It is also the go-to technique for purifying oligonucleotides that carry a hydrophobic tag, such as those left with their dimethoxytrityl (DMT) protecting group intact during a “DMT-on” purification strategy. The DMT group dramatically increases the hydrophobicity of the full-length product relative to capped failure sequences, making the separation trivially easy on a reversed-phase column.

Anion-Exchange Chromatography

Anion-exchange chromatography (AEX) separates oligonucleotides based on their charge, which in practice means primarily by length, since each additional nucleotide adds another negatively charged phosphate group. Strong anion-exchange columns use quaternary amine functional groups, while weak anion-exchange media use tertiary or secondary amines. Both can resolve full-length product from shorter impurities and, in favorable cases, from closely related species that differ by a single nucleotide or a single chemical modification.

A recent head-to-head comparison of AEX versus IP-RP at preparative scale found that anion-exchange methods using agarose-based resins substantially outperformed IP-RP on silica media in both productivity and solvent consumption. At 95% target purity, AEX achieved more than twice the productivity of IP-RP; at 99% purity, the advantage grew to sevenfold. Solvent consumption was also dramatically lower, with AEX using only a third to a tenth of the solvent volume that IP-RP required.⁷ The catch is that AEX cycle times tend to be longer because of slower mass transfer in the agarose resin, but the high loadability compensates for it.

Weak anion-exchange methods have also been optimized specifically for analyzing and separating the diverse impurity profiles of therapeutic oligonucleotides, where distinguishing between closely related species matters for regulatory compliance.⁸ AEX is often the method of choice for a second “polishing” step after an initial capture by reversed-phase or hydrophobic interaction chromatography.

Mixed-Mode and Hydrophobic Interaction Approaches

Mixed-mode stationary phases combine two separation mechanisms on a single column, typically reversed-phase hydrophobicity and anion exchange. The benefit is selectivity that neither mechanism achieves alone. Studies on mixed-mode reversed-phase/weak anion-exchange columns have shown enhanced resolution for oligonucleotides with minor sequence variations, readily separating species that differ by just one, two, or three nucleotides in length.⁹ Selectivity was highest when the sequence difference fell near the 3′ end and lowest when it was at the 5′ end, which is useful to know when planning a purification strategy for a specific target.

Hydrophobic interaction chromatography (HIC) takes a different tack. Instead of using organic solvents, HIC uses high concentrations of a kosmotropic salt (typically ammonium sulfate) to drive oligonucleotide binding to a weakly hydrophobic resin, then elutes by lowering the salt concentration. HIC has been developed as a greener alternative to reversed-phase methods because the mobile phases are entirely aqueous, generating far less organic solvent waste.¹⁰ Because elution happens at low salt, the product can often be loaded directly onto an AEX polishing column without an intermediate desalting step, which avoids the 20–30% sample loss that desalting can introduce.

HIC has been applied to antisense oligonucleotides at semi-preparative scale for separating phosphorothioate targets from phosphodiester impurities, achieving purities in the range of 60–80% with recoveries of roughly 50–75%.¹¹ Those numbers are lower than what AEX achieves for the same separation, but HIC offers the practical advantage of working under non-denaturing conditions and tolerating high-salt feed streams, making it a strong complement rather than a replacement.

The Phosphorothioate Problem

Phosphorothioate oligonucleotides, in which one of the non-bridging oxygen atoms in the backbone is replaced by sulfur, are the backbone modification behind most approved antisense and siRNA therapeutics. The sulfur substitution makes the backbone resistant to nuclease degradation but introduces a new complication: each modified linkage creates a chiral center, producing a pair of diastereomers. A 20-mer with every linkage modified has 19 chiral centers and over half a million possible stereoisomeric forms. These diastereomers have the same length, the same sequence, the same charge, and the same mass, so separating them is extremely difficult.

Standard reversed-phase chromatography typically cannot resolve phosphorothioate diastereomers unless the oligonucleotide is very short. Pellicular anion-exchange chromatography has shown some ability to separate diastereomers arising from a small number of phosphorothioate linkages. In one study, a 37-nucleotide aptamer with just two phosphorothioate linkages resolved into four components on a monolithic pellicular anion-exchange column, consistent with the four possible diastereomeric combinations.¹²

Ion-pair reversed-phase methods can achieve diastereomer resolution under carefully optimized conditions, but the separation is highly sensitive to the choice of ion-pairing reagent. Triethylammonium, for instance, provides meaningful diastereomer selectivity, while tributylammonium effectively abolishes it. Diastereomer selectivity also depends on the position of the modified linkage: substitutions in the middle of a pentamer showed higher selectivity than those at either end, and the nucleobase identity at the modified site matters as well.¹³ Both reversed-phase and strong anion-exchange protocols can be made to work for separating the diastereomer mixture that conventional synthesis produces.¹⁴

For most therapeutic phosphorothioate programs, the practical goal is not to isolate individual diastereomers but to ensure that the diastereomeric mixture is consistent from batch to batch. Stereopure synthesis, where each chiral center is set deliberately, is an emerging alternative that avoids the diastereomer separation problem entirely, though it adds complexity to the synthesis step.

Purifying Conjugated Oligonucleotides

Many current oligonucleotide therapeutics are conjugated to targeting ligands, with N-acetylgalactosamine (GalNAc) clusters being the most common example. GalNAc directs the oligonucleotide to hepatocytes via the asialoglycoprotein receptor, and the triantennary GalNAc cluster used in approved drugs adds substantial bulk and hydrophilicity to the molecule. This changes the chromatographic behavior enough that standard purification methods may need reworking.

A direct comparison of mixed-mode and anion-exchange approaches for GalNAc-conjugated oligonucleotides found that a mixed-mode column delivered roughly double the recovery and higher purity (about 95%) compared to the AEX column (about 91%) for one siRNA conjugate. Across a broader panel of GalNAc-modified oligonucleotides, the mixed-mode resin achieved purities of 90–94% with recoveries in the range of 30–60%.¹⁵ The wide recovery range reflects the fact that some conjugate purifications are straightforward while others involve closely eluting impurities that force tighter fraction cuts and lower yields.

Continuous chromatography techniques offer a way to recover more product from these difficult separations. A twin-column continuous process applied to a GalNAc-conjugated DNA-LNA gapmer boosted yield from about 53% in standard batch chromatography to over 91%, while maintaining comparable purity, productivity, and buffer consumption.¹⁶ The idea behind continuous chromatography is that fractions containing partially pure product at the front and tail of a peak are recycled back into a second column rather than being discarded. For expensive conjugated oligonucleotides, the yield improvement can substantially change the economics of manufacturing.

Long Oligonucleotides and Guide RNAs

As CRISPR-based gene editing has scaled up, demand for high-purity single-guide RNAs (sgRNAs) has grown. These are substantially longer than typical therapeutic oligonucleotides, often exceeding 100 nucleotides. Length amplifies every purification challenge: coupling efficiency drops compound over more steps, generating a broader distribution of failure sequences, and the resolution between full-length product and the nearest n-1 truncation shrinks as the percentage length difference becomes smaller.

IP-RP chromatography can be applied to sgRNA purification, but late-eluting contaminants linked to incomplete deprotection of natural RNA bases remain a bottleneck. Even with optimized column chemistries, these deprotection-related impurities are difficult to resolve chromatographically, which means that synthesis and deprotection conditions need to be improved alongside the purification method.¹⁷

Non-chromatographic alternatives are emerging for applications where HPLC is too complex or time-consuming. One such method uses a five-step chemical process rather than column chromatography and has been reported to yield guide RNA at over 80% purity, which meets the threshold the FDA sets for gene-editing applications.¹⁸ For research-grade material, this kind of approach trades peak purity for speed and simplicity, which is often the right call when the alternative is weeks of HPLC method development for a sequence you will use once.

High-Throughput Purification at Microscale

Most purification methods handle one oligonucleotide at a time, but some applications need hundreds or thousands of distinct sequences purified simultaneously. Array-based gene assembly, multiplexed FISH probes, and encoded DNA libraries all fall into this category. Traditional methods are too slow and expensive when the number of sequences scales into the hundreds.

An enzymatic approach called SNOP (Simultaneous Nuclease-mediated Oligonucleotide Purification) addresses this by encoding each target sequence within a longer precursor, then using restriction enzymes to cleave out the desired product while simultaneously destroying failure sequences. In a 256-plex experiment producing 70-nucleotide products from 94-nucleotide precursors, the median purity of SNOP products was about 82%, compared to roughly 57% for standard desalted oligonucleotides and about 60% for PAGE-purified oligonucleotides measured by next-generation sequencing.¹⁹ The ability to purify hundreds of sequences in a single tube, stoichiometrically, is something no chromatographic method can match.

How Purity Is Actually Measured

Choosing the right purification method is only half the problem. You also need an analytical method sensitive enough to confirm that the purification worked and to characterize whatever impurities remain. Different analytical techniques see different things, and a sample that looks pure by one method may reveal significant impurities under another.

Reversed-phase UHPLC coupled with high-resolution mass spectrometry can resolve dozens of distinct impurity species, including both truncated failure sequences and degradation products carrying modified phosphate groups.²⁰ Capillary gel electrophoresis provides single-nucleotide resolution and has been coupled on-line with electrospray mass spectrometry for characterizing both phosphodiester and phosphorothioate oligonucleotide mixtures, including the identification of terminally phosphorylated failure sequences.²¹ CGE has also found a role in characterizing longer species like mRNA poly(A) tails, where it achieves resolution comparable to LC-MS but is simpler to implement for routine quality testing.²²

For therapeutic oligonucleotides, the analytical picture feeds directly into the regulatory picture. Proposals exist for standardized reporting, identification, and qualification thresholds for product-related impurities, drawing on both chemistry and safety considerations to establish platform approaches that can be applied across different oligonucleotide drugs.²³ In practice, this means that the purification method you choose constrains not just the purity you achieve but also the analytical methods you need to validate and the regulatory documentation you file.

Choosing a Strategy in Practice

The selection of a purification method reduces, in most cases, to three questions: what impurities are you trying to remove, how much material do you need, and what purity does your application require?

Research-grade primers and probes: Desalting is sufficient for most standard PCR and sequencing primers. If your experiment is sensitive to truncated sequences, PAGE or a cartridge-based solid-phase extraction gives clean material with minimal effort.
Modified oligonucleotides for cell-based assays: IP-RP HPLC with a DMT-on strategy provides a convenient single-pass purification. The hydrophobic tag makes the separation straightforward and the yields are good.
Therapeutic manufacturing: A two-step process is typical. An initial capture step (IP-RP, HIC, or AEX depending on the modification chemistry) removes the bulk of failure sequences, followed by an AEX polishing step to reach purities above 95%. Continuous chromatography can dramatically improve yield on expensive conjugated species.
Phosphorothioate therapeutics: AEX tends to provide better resolution of closely related impurities than IP-RP, especially when diastereomer consistency matters. Careful optimization of ion-pairing reagent and gradient conditions is essential regardless of the technique chosen.
Long RNAs for gene editing: IP-RP remains the default, but synthesis and deprotection quality are often the limiting factor rather than chromatographic resolution. Non-chromatographic methods are emerging for applications where 80% purity is acceptable.
High-throughput pools: Enzymatic purification methods like SNOP can handle hundreds of sequences simultaneously at purities that outperform both desalting and PAGE.

Environmental Costs and Greener Alternatives

Oligonucleotide purification generates substantial solvent waste, particularly when IP-RP methods using acetonitrile are run at preparative scale. Manufacturing a single batch of a therapeutic oligonucleotide can consume hundreds of liters of organic solvent. AEX methods use far less organic solvent by design, since the mobile phases are aqueous salt solutions. The productivity comparison between AEX and IP-RP mentioned earlier showed that AEX consumed only a fraction of the solvent volume.⁷

HIC pushes the sustainability argument further. Because it relies on aqueous ammonium sulfate gradients rather than organic solvents, the waste streams after the initial flow-through fraction can be disposed of as aqueous waste rather than hazardous organic waste.¹⁰ As the oligonucleotide therapeutics pipeline grows and manufacturing volumes increase, the environmental and cost pressures of solvent-intensive purification will likely push more processes toward AEX and HIC as primary capture steps, reserving IP-RP for analytical applications and cases where its unique selectivity is genuinely needed.