Webinar Highlights: Two Dimensional Liquid Chromatography (2D LC) for Impurity Analysis
Q&A from webinar on using targeted 2D LC (RP-RP) for effective impurity analysis to address peak coelutions
The Waters pharmacuetical team recently hosted a webinar discussion how two-dimentional LC can offer a solution to analyze all impurities, including coelutions, in one chromatography system setup that can save analysis time and increase productivity.
The webinar was given by Zhimin Li, Ph.D., a senior scientist at Waters. Below, she provides answers to a healthy number of questions that were raised during the presentation.
What separation mode is the preferred method for quantitation: heart-cutting or comprehensive and why?
That would depend on what type of sample you are analyzing, and what is your requirement for quantitation.
Heart-cutting is often used for separating difficult-to-resolve compounds such as regioisomers, stereoisomers, etc. In these types of samples, the number of compounds is limited. Another advantage to heart-cutting is that the 2nd dimension method is not constrained by 1st dimension. You can work on optimizing the 2nd dimension with better resolution and sensitivity. For these reasons, heart-cutting is often a preferred mode for quantification in the second dimension.
On the other hand, comprehensive 2D mode is normally used for complex mixtures, which are often qualitative. While it is uncommon to use comprehensive 2D mode for quantitation, there are some publications that outline this approach:
- Comprehensive two-dimensional liquid chromatography-tandem mass spectrometry for the simultaneous determination of wine polyphenols and target contaminants. Donato P et al. J Chromatography A. 2016 Aug 5;1458:54-62.
Is it practical to apply this 2D technique in the identification of mass of the coeluting peaks?
Yes, it would. However, if you are using nonvolatile salts in the mobile phase, make sure to desalt the mobile phase before sending sample to MS.
After using 2DLC, can we say there are no hidden peaks in a sample containing peptides?
2DLC provides a tool to discover more hidden peaks, but separation of all compounds is dependent on the quality of the separation in both the 1st and 2nd dimensions, no matter the sample composition.
Have you studied 2DLC in the biologics space, such as separation of proteins and antibodies?
Are there applications using 2DLC hyphenated with QTof MS? Yes, Waters has performed a variety of work using 2D LC applications for analyzing proteins and antibodies, and applications using 2D LC hyphenated with QTof MS. Please refer to the following application notes:
- Characterization of Biotherapeutics: ACQUITY UPLC H-Class Bio with 2D Part 1 of 3: On-line Desalting of Biotherapeutic Samples for Increased Productivity
- Characterization of Biotherapeutics: ACQUITY UPLC H-Class Bio with 2D Part 2 of 3: Rendering a Viable Interface for IEX with ESI-MS Analysis
- Characterization of Biotherapeutics: ACQUITY UPLC H-Class Bio with 2D Part 3 of 3: On-line Enrichment of Low Abundance Species
- High Resolution Separation of Phospholipids Using a Novel Orthogonal 2D UPLC/Q-Tof MS System Configuration
- 2D LC for Quantification and MS Analysis of Monoclonal Antibodies in Complex Samples
Can we use a nano column in the second dimension?
Yes, if you have the he nanoACQUITY UPLC System with 2D Technology, which is very different from the ACQUITY System with 2D Technology. NanoACQUITY UPLC System with 2D Technology is designed for a specific application with a preset plumbing configuration. This solution consists of a reversed phase (RP) separation at pH 10 in the first dimension, followed by a RP separation at pH 2 in the second dimension.
Here are some references for a 2DLC application using microflow LC-MS columns:
- Evaluating 2D-RP/RP Fractionation Capabilities of the ACQUITY UPLC M-Class System with 300-µm I.D. Configuration
- Assay for Identification and Quantification of Host Cell Protein Impurities in High-Purity Monoclonal Antibodies Down to 1 ppm: An Inter-Laboratory Study
What could be the possible problems, when applying a size exclusion chromatography (SEC) column as the 2nd dimension?
In most 2DLC applications, the RP separation is typically used in the 2nd dimension, due to high plate counts and excellent peak capacity in gradient elution. Placing the SEC separation in the 2nd dimension is limited and challenging. Some of the points for consideration include:
- SEC is isocratic, and you don’t have the ability to retain the analyte at the head of the column
- SEC is impacted by the injection volume. Thus, there will be limitations on how much volume you can deliver to the 2nd dimension
Given these factors, the performance of the SEC in the 2nd dimension is limited.
Here are a few examples on using SEC in the 2nd dimension:
- IEX (1st dimension) – SEC (2nd dimension): Automated instrumentation for comprehensive two-dimensional high-performance liquid chromatography of proteins. Michelle M. Bushey and James W. Jorgenson, Anal. Chem., 1990, 62 (2), pp 161–167
- Protein A (1st dimension) – SEC (2nd dimension): Automated 2D-HPLC method for characterization of protein aggregation with in-line fraction collection device. Williams A et al. J Chromatogr B Analyt Technol Biomed Life Sci. 2017 Mar 1;1046:122-130.
Is 2DLC able to separate enantiomers? Can you do RP in one dimension and NP in the other?
As we know, most of the time normal phase (NP) is used to separate enantiomers or chiral compounds. However, typically RP and NP each require very different mobile phase compositions. Specifically in normal phase any presence of water can impact the separation. Therefore, the challenge would be to find the correct solvents for dilution (At-Column Dilution) and elution of the enantiomers in the 2nd dimension.
2DLC setup & methods
Is it more important to optimize the separation on the 2nd or the 1st dimension?
I would say that it is important to optimize both dimension methods. In 2DLC, the 2nd dimension adds orthogonality for the separation and increases the peak capacity, so you don’t want to comprise the resolution in the 1st dimension.
What are the challenges when developing a method in the 2nd dimension?
Aside from the same challenges as conventional LC, you will also face additional challenges that are unique to 2DLC setup. These include:
- Means to transfer the peak from 1stdimension to 2nddimension,
- If directly transferring an analyte from the 1st dimension column to the 2nd dimension column, system pressure may be an issue
- If using a sample loop, the selection of loop size, transfer volume, percentage of filling (the loop) need to be considered
- If using a trap column, the selection of the trap column chemistry is critical
- Mobile phase compatibility between both dimensions
- If the modes are RP- RP, the most common challenge is “strong solvent effects”, since the heart-cut from 1stdimension often contains a high portion of organic solvent. As we discussed in the presentation, we can use at column dilution (ACD) to reduce the organic composition before the sample reaches to 2nd dimension
- If the two methods contain incompatible mobile phases (such as IEX – RP), the compatibility of high salt with organic solvent needs to be considered. Again you might need to use ACD or the same concept to mitigate the issue
- Mobile phase compatibility with detection technique. For example,
- If MS is the downstream characterization for the 2nd dimension, you will need to remove MS unfriendly mobile phases before the effluent reaches the MS detector.
What about the configuration would be more relatable to a normal laboratory, i.e., a single detector configuration?
The configuration with a signal detector has limited applicability. You only need one detector for “Comprehensive 2DLC mode”, but it could require significant application development. For heart-cutting, targeted 2DLC approach, you will need two detectors so you know where your target is.
How many heartcuts can be done in one analysis?
In the typical heart-cutting scenario the peak is diverted from the 1st dimension to the 2nd dimension. As soon as the 2nd dimension separation is complete another peak can be diverted to the 2nd dimension. In this case, another peak or multiple peaks could be captured, depending on the timing of both dimensions and the location of the peaks one would like to capture.
Is there any requirement for the peak width for the interested peak in the 1D chromatogram?
When you add a trap column prior to 2nd dimension column to retain the “heart-cutting” from 1st dimension, there are technically no restrictions for the peak width of the interested peak (i.e. the transfer volume). The trap capacity (mass not volume) of a trap column is based on the mass of the packing bed. However, you do need to consider the mass load for the column in the 2nd dimension.
What is the total analyses time difference between running the two methods separately and in the 2D mode?
In the targeted 2DLC mode, after “heart-cutting” in the 1st dimension, the consecutive run time of 1st dimension overlaps with that of 2nd dimension. For instance, if both dimensions are 5 min methods and the heartcut is at 3 min in the 1st dimension, the total run time for the 2D mode would be 8 min, as compared to the additive total run time of 10 min. This comparison is under the assumption that you have two separate systems to run these two methods. If not, then you will need to add time to setup the system and equilibrium the column in order to run the second method.
How many column volumes were required to achieve the separation of cis-isomer?
Approximately 7 column volumes in my case study.
What is the protocol of 2D setup in MassLynx 4.1?
You may refer to the following document on configuring 2DLC system in MassLynx software: ACQUITY UPLC Systems with 2D Technology Capabilities Guide. I would recommend you to contact Waters for setting up the 2DLC system in MassLynx.
Use of a trap column in targeted 2DLC
What kind of column chemistry should we consider for the trap column?
What are the specifications for your trap columns as far as capacity? Also, is there ever a need to change the trap to a different stationary phase? How do you ensure that the heart-cut will retain on the trap column, particularly for compounds that elute late in the first dimension (high organic)?
The following columns have been frequently used for trapping. Particle size ranges from 10 to 20 µm. (Direct Connect HP 2.1 x 30 mm UPLC Columns for On-line SPE)
- Part Number 186005231 Oasis HLB 2.1x30mm Direct Connect HP, 20 µm
- Part Number 186005232 XBridge C18 2.1x30mm Direct Connect HP, 10 µm
- Part Number 186005233 XBridge C8 2.1x30mm Direct Connect HP, 10 µm
The XBridge packing materials were designed to provide excellent peak shape, high efficiency and excellent stability for acidic and basic mobile phases. The trapping capacity (mass not volume) is based on the mass of the packing bed. It is about the same as an 80mg (mass of packing bed) SPE cartridge.
To select the right chemistry for the trap column depends on the properties of the analyte of interest. Yes, there is a need to change the trap to a different stationary phase based on your application. You may place the trap column in the 1st dimension to test the trap and elute properties of the analyte.
For the high organic effluent from 1st dimension, I would also recommend screening the trap columns at different strength of mobile phases. There is a short time window (usually < 1 min based on tubing connections) between when the analyte is trapped on the trap column and before it elutes to the 2nd dimension. As long as the trap column provides the capability to trap the analyte within your specified time window, you should be good for your application. In my case study, I tested XBridge C8 2.1x30mm Direct Connect HP and it is sufficient for my application.
Are trap columns and guard columns the same? Can I use a guard column of my 2nd dimension column as a trap column?
Trap columns and guard columns are different.
A guard column is packed with the same packing materials as the analytical column, including the same particle size and the same frit size (0.2 m). It uses the same ID (inner diameter) and is usually ≤ 5 mm. It is used to protect the analytical column.
A trap column is typically packed with a different packing material than the analytical column. The range of packing materials vary, including RP, ion-exchange and mixed mode stationary phases with particle sizes from 10 m to 30 m. Waters’ current portfolio includes XBridge 10 m C18 and C8 , and Oasis 20 m HLC. A trap column has a larger frit ( 2.0 m) to allow high flow rate during loading process, up to 5 mL/min, while keep the back pressure in optimum range of the pump.
You do not want to use a guard column as a trap column, primarily because of its smaller particle size packing. A guard column would result in high back pressure at higher flow rates during the loading step. Also, the mass of the guard column would most likely not be sufficient for effective trapping.
What is the trap column matrix that allows retention in 1D and elution upon backflush?
Because of the frits on each end and the packing mechanism, the flow in a trap column can be in either direction.
Can a dead volume column be used to replace a trap column in the 2nd dimension to reduce the system pressure?
Compared with directly transferring an analyte from a 1st dimension column to 2nd dimension column, a trap column reduces system pressure at transfer due to larger particle size and shorter length than a typical LC column. The main use of trap column is to retain the analyte in a range of volumes to facilitate quantification in 2nd dimension. Along with at-column dilution, it can effectively focus the analyte and prevent break-through. A sample loop won’t have the same functions as a trap column.
At-column dilution (ACD)
ACD may impact the limit of quantitation (LOQ) of the method, how can we overcome it?
At-column dilution should only be used when it is needed, as in the example described in my seminar. In these instances, while ACD does dilute the peak going into the 2nd dimension, this does not impact the LOQ – primarily because ACD reduces strong solvent effects and improves the peak shape. ACD allows sample mass to be focused in a tight and narrow band on a given sorbent. For 2DLC applications, adding ACD to the interface between 1st and 2nd dimension also provides the versatility of separation modes for both dimensions.
Are there any limitations to at-column dilution? What is the regular dilution range?
The dilution factor of ACD depends on the flow rates of ACD pump and that of the 1st dimension pump. The desired organic portion is 5 ~ 10% in the effluent from 1st dimension. The flow rates ratio can be calculated based on the desired organic portion. If it is needed operate within the system’s operating pressure, you may reduce the flow rate for the 1st dimension during transferring.
What kind of mixer do you recommend for ACD?
You can use any standard mixer for your system. In my case study, I used Waters 50 µL mixer (p/n 700002911).
Does you need in-line check valves when using the ACD to prevent the diluting flow from restricting the flow of the 1st dimension input?
No, you don’t need to use in-line check valves. Since the trap column has larger particle size, there is no significant pressure build-up to push the flow back to 1st dimension.
2DLC and quality control
Can 2DLC be used in quality control and regulated environments?
This is an important question. In pharmaceutical industry, 2DLC has historically been used as an analytical tool to support the early stages of drug development or for troubleshooting. With the advancement in 2DLC technology, the data quality on 2nd dimension is greatly improved. You probably already learned from my today’s talk, that the data quality in the 2nd dimension can be comparable to that of the 1st dimension.
While this is a very complex and multifaceted question, some recently published papers (see below for references) suggest that 2DLC will play an important role in quality control and regulated environment. Specifically the publications discuss the versatility of a heart-cutting 2DLC-MS platform that enables characterization of mAbs (monoclonal antibodies) and ADCs (antibody-drug-conjugates) during all stages of the product life cycle from development to release and stability testing.
- Automated 2D-HPLC method for characterization of protein aggregation with in-line fraction collection device. Williams, E.K. et al. J. Chromatogr. B 1046, 122–130 (2017).
- 2D-LC–MS for the Analysis of Monoclonal Antibodies and Antibody–Drug Conjugates in a Regulated Environment. E. Largy et al. Curr. Trends Mass Spectrom. 14(2), 29–35 (2016).
- Profiling Analysis of Low Molecular Weight Heparins by Multiple Heart-Cutting Two Dimensional Chromatography with Quadruple Time-of-Flight Mass Spectrometry. Y. Ouyang et al. Anal. Chem. 87, 8957–8963 (2015).
How did you consider validation elements for robustness? Does USP methodology allow for substitution of a single 2D method for two separate 1D methods for the analysis of impurities?
Validation wasn’t done on the 2DLC separation described in my presentation. However, typically the validation elements for robustness in conventional LC can be used as a reference. Additional work includes – but is not limited to- evaluation of the elements from the interface between 1st and 2nd dimensions.
As for USP methodology, you should consult your regulatory agency in your company.
Can this 2D set up be run through Empower 3 Software so it is compliant for GMP analysis?
Waters Empower software is compliance ready software. If you already set up Empower to be compliant for GMP analysis, then the 2D setup through Empower is compliant. If you haven’t setup the Empower to be compliant for GMP analysis, please refer to this white paper, The role of Empower 3 chromatography data software in assisting with electronic records regulation compliance.
- See our latest work in impurity analysis and profiling at Waters Impurities Central: impurities.waters.com
- If you would like more about our solutions for impurity analysis, email Tilak Chandrasekaran, business development manager on our pharmaceutical team. Tilak[underscore]Chandrasekaran [at]waters.com.