Impact of mobile and stationary phases on siRNA duplex stability in liquid chromatography

This study explores the impact of chromatographic conditions and mobile phase composition on the stability of chemically modified siRNA duplexes. Using differential scanning calorimetry (DSC) and various chromatographic techniques, we measured the melting temperature under different buffer and solvent conditions.

The findings reveal that factors such as ion-pairing reagents, buffer concentration, and organic solvent content significantly influence duplex stability. The study highlights how the formation of an organic/aqueous solvent layer on the chromatographic sorbent surface plays a crucial role in duplex stability, beyond mobile phase composition alone.

The original article

Impact of mobile and stationary phases on siRNA duplex stability in liquid chromatography

Martin Gilar, Samuel Redstone, Alexandre Gomes

Journal of Chromatography A, Volume 1733, 27 September 2024, 465285

https://doi.org/10.1016/j.chroma.2024.465285

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

1. Introduction

Analysis of oligonucleotides in liquid chromatography (LC) is typically performed under denaturing conditions, using elevated temperature [[1], [2], [3]]. High column temperatures ensures that oligonucleotides with propensity to form intra- (hairpin loops) and inter-molecular (G-quadruplexes) interactions are denatured and have a predictable retention behavior [[4], [5], [6]].

The denaturing conditions in LC are achieved above the melting temperature (T_m). T_m is defined as the temperature at which half of the double stranded duplex is dissociated into the complementary strands or alternatively when half of the secondary structure (intramolecular hairpin) is linearized. High T_m value indicates that nucleic acids can form stable duplexes or secondary (self-complementary) structures. This is the case for nucleic acids with high cytidine and guanosine content or for modified nucleic acids, such as locked nucleic acids or 2′-modified oligonucleotides.

Liquid chromatography separation can be also performed at non-denaturing (native) conditions when desired. Examples are sizing of double stranded DNA (dsDNA) restriction fragments by non-denaturing ion-pair reverse-phase (IP RP) LC [7] or analysis of silencing RNA (siRNA) duplexes with IP RP LC [8,9]. The goal of siRNA duplex analysis is to quantify the presence of the single stranded siRNA constituents in the siRNA formulations. Non-denaturing LC conditions must be maintained to avoid false positive detection of single stranded peaks. Selecting a column temperature close to the duplex melting conditions results in on-column duplex melting and peak smearing [9].

A special case of LC separation performed near the DNA duplex melting temperature is denaturing HPLC (dHPLC). This technique can resolve DNA homoduplexes (∼200–500 bp) from the partially melted heteroduplexes containing a minor sequence mismatch. dHPLC method is utilized for single nucleotide polymorphism discovery [6,[10], [11], [12]] and requires careful optimization of column temperature with high accuracy to ensure partial denaturation of heteroduplexes [13]. Similar methods utilizing partial duplex melting were developed for gel electrophoresis [14] or capillary gel electrophoresis [15,16] and used for DNA polymorphism discovery.

The melting temperature of short DNA or RNA oligonucleotides can be predicted in-silico; several calculators are available on-line [13,[17], [18], [19], [20], [21], [22], [23]] and could be used for polymerase chain reaction (PCR) primer design. However, the calculators predict T_m for aqueous conditions and a limited buffer selection. The effect of ion-pairing buffers and organic modifiers on T_m values has not been implemented. This presents a challenge for LC method development since the effect of mobile phase on T_m is unknown. To our knowledge the impact of LC separation conditions on T_m has not been systematically investigated.

The melting temperature of nucleic acids is affected by several factors: (i) concentration, type, and charge of the buffer cation (e.g. Li⁺ vs Na⁺, K⁺ … or Na⁺ vs Mg²⁺), (ii) nucleotide composition in the sequence, (iii) length of the duplex (iv) nucleic acid concentration, and (v) nucleic acid type (RNA/DNA) or its chemical modifications.

The melting temperature is often reported for 100 mM [Na⁺] concentration [24]. The role of the cation in hybridization is to counteract the negative charge of the nucleic acid strands, which have the tendency to repel each other. Both cation concentration and the type influence the T_m; compact ions such as Li⁺, or doubly charged cations such as Mg²⁺, stabilize the duplex more effectively than bulkier and singly charged cations [25,26].

Adenine (A) pairs with thymine (T, or uridine, U) by 2 hydrogen bonds while guanine (G) - cytosine (C) pairing involves 3 hydrogen bonds. Therefore, the C-G rich sequences have higher T_m [17]. Short DNA duplexes (16 base pairs, bp) with equal A, T, G, and C content have T_m approximately 51.9 °C, while T_m of a 24 bp duplex is about 60.9 °C (calculated for 100 mM [Na⁺]) [Anon., 22]. The melting temperature of DNA/RNA heteroduplexes of the same sequence is comparable to DNA/DNA; RNA/RNA duplexes have >10 °C higher stability (see Supplemental materials, Table S1).

Antisense oligonucleotide drugs or siRNA are often chemically modified to enhance the oligonucleotide stability in-vivo [27,28]. Modification of the ribonuclease ring with 2′-fluorination (2′-F) or 2′-O-methylation (2′-OMe) is known to increase the duplex stability. T_m of siRNA consisting of two modified complementary strands can increase by 10 °C or more. The T_m also depends on the chemical modification in the following order: RNA < 2′-OMe RNA < 2′-F RNA [28]. Similar T_m improvements were observed for locked nucleic acid (LNA) [29] or peptide nucleic acid analogs [[30], [31], [32]].

The goal of this work was to investigate the impact of mobile phase composition on the melting temperature of siRNA duplexes. We explored mobile phases commonly used in IP RP LC, hydrophilic interaction chromatography (HILIC), ion-exchange LC and size-exclusion chromatography (SEC). We first investigated siRNA melting temperature using the differential scanning calorimetry (DSC) method for solutions containing alkali metal cations, magnesium, ammonia, triethylamine or hexylamine. For the selected mobile phases, the effects of organic solvent addition (methanol, acetonitrile, ethanol, and hexafluoroisopropanol) were also investigated. In the second part of the study, we measured T_m of siRNA samples with chromatographic methods (RP, IP RP LC, SEC and HILIC) using different experimental column temperatures to observe the on-column duplex melting. The goal of this experiment was to investigate how the selected LC mode (besides the mobile phase) affects the duplex stability in chromatography.

2. Experimental

2.2 DSC and LC instrumentation, columns, and experimental conditions

DSC experiments were performed with a nano DSC instrument (Waters Corporation, Milford, MA, USA). Lyophilized siRNA duplex sample was dissolved in the selected buffers at concentration 1 mg/mL at room temperature; no additional annealing was performed. The DSC sample cell was filled with 0.7 mL of the solution and the reference cell contained the same buffer without the siRNA sample. The background signal was acquired with the buffer solutions placed in both sample and reference cells. Background signal was subtracted from the sample melting curve. Data were acquired with a temperature gradient of 2 °C/min from 15 to 100 °C followed by reversed gradient from 100 to 15 °C. The up-and-down gradient cycle was repeated twice, which yielded four melting curves. The averaged T_m values for each buffer condition are summarized in Table 1. Data analysis was performed with NanoAnalyze software version 3.12.5 using Voight 3 melting curve fitting. Data processing and plotting was performed in Microsoft 365 Excel.

Chromatographic measurements were performed using an ACQUITY™ Premier System equipped with binary solvent manager (BSM), Flow Through Needle Sample Manager (FTN SM), column heater manager (CM-A) with enabled active preheater (APH), and ACQUITY Tunable Ultraviolet (TUV) Detector with a 5 mm titanium flow cell. Sample manager temperature was set to 10 °C; 1–4 µL were typically injected for analysis. Gradient delay was 0.4 mL.

SEC experiments were performed using an ACQUITY UPLC™ Protein BEH™ SEC Column, 200 Å, 1.7 µm, 4.6 mm × 150 mm at 0.4 mL/min. Mobile phase was 25 mM sodium phosphate buffer, pH 7.5. UV detection wavelength was 260 nm. The experiments were run at temperatures between 30 - 85 °C using 5 °C increments. The LC system was thermally equilibrated for 45 min between experiments performed at different temperatures.

IP RP LC experiments were performed using an ACQUITY Premier Oligonucleotide C18 Column, 300 Å, 1.7 µm, 2.1 × 50 mm. Flow rate was 0.3 mL/min, UV detection 260 nm. Mobile phase A was 10/100 mM HA/HFIP solution in 25 % MeOH. Mobile phase B as 10/100 mM HA/HFIP solution in 75% acetonitrile. Gradient was from 30 to 100 % B in 10 min with 1 min wash at 100 % B and 7 min equilibration at initial gradient conditions before next sample injection. To investigate the on-column duplex melting the column temperature was raised by 5 °C increments for subsequent experiments. The experiments were run at the 30 - 70 °C temperature range. The LC system was thermally equilibrated for 45 min between experiments performed at different temperatures.

HILIC experiments were performed using ACQUITY Premier Amide Column, 1.7 µm, 2.1 × 50 mm. Flow rate was 0.3 mL/min, UV detection 260 nm. Mobile phase A was 10 mM ammonium acetate (native pH 6.9); mobile phase B was 10 mM ammonium acetate in 75% acetonitrile. Mobile phase gradient was 80 to 35 % B in 7.5 min (60 % to 26.25 % MeCN in 7.5 min). The experiment was performed at column temperatures 10, 30, 50, 70 and 80 °C. The LC system was thermally equilibrated for 45 min between experiments performed at different temperatures.

RP LC experiments were performed using ACQUITY Premier Oligonucleotide BEH C18 Column, 300 Å, 1.7 μm, 2.1 × 50 mm using 0.3 mL/min flow rate. Mobile phase A was 10 mM Ammonium acetate (native pH 6.9), mobile phase B was 10 mM AmAc in 50% acetonitrile. Mobile phase gradient was 0 to 50 % B in 10 min (0 % to 25 % MeCN). The experiment was performed at column temperatures of 10 and 30 °C. Additional experiments were performed with mobile phase A consisting of 50 mM ammonium acetate and mobile phase B consisting of 50 mM ammonium acetated in 25% acetonitrile. The gradient was 0 to 100 % B in 10 min; column temperature was 30 and 50 °C. Other experimental conditions were the same. The LC system was thermally equilibrated for 45 min between experiments performed at different temperatures.

All chromatographic columns were obtained from Waters Corporation (Milford, MA, USA). siRNA duplex samples were dissolved in the aqueous buffers (SEC, IP RP LC, RP LC) or in buffers containing 50 % acetonitrile (HILIC).

3. Results and discussion

3.3. siRNA duplex stability in IP RP LC

Quality control of siRNA duplexes and their separation from single stranded species is often performed by IP RP LC method in non-denaturing conditions [9,[43], [44], [45]]. The mobile phases compatible with MS detection permit peak confirmation using the molecular weight of RNA oligonucleotides [8,46]. However, as we found, mobile phases containing organic solvents can destabilize the siRNA duplex. To ensure the duplex stability in IP RP LC, the authors of published reports utilized ambient or sub-ambient column temperatures [8,43,46].