Sample preparation is the most critical part of a chromatographic assay. One can focus on the instrument, but the problems usually arise before that. Matrix effects are particularly troublesome, since they are often unanticipated and may never be understood. This is especially true for sample preparation for quantitative liquid chromatography-mass spectrometry (LC-MS), since one must work around the idiosyncrasies of the mass spectrometer as well as the sample. Nonvolatile salts must be avoided. Ionization efficiency, and hence signal, can be perturbed by the presence of other compounds in the mobile phase or sample matrix effects. The ability to get useful results from very small samples is attractive, especially for proteomics research, but new technologies are required for manipulating samples in the submicroliter range.
The rapid growth segments of quantitative LC-MS include food safety, biopharm, and clinical, including both research and diagnostics. Dr. Dale Schoener of the Intertek Group (El Dorado Hills, CA) lectured at a workshop entitled, “New Developments in Sample Preparation for LC-MS.”1 His focus was drugs and metabolites for pharmacokinetics (PK) and toxicokinetics (TK). Dr. Schoener reported that plasma is the preferred matrix, followed by urine and whole blood. Serum typically is a concentrated protein solution containing about 5000 compounds at ng/mL or higher concentrations. The concentration of albumin is 4.2 mg/mL. The immunoglobulin pool is about 17 mg/mL. This concentration range complicates the sample prep, since the therapeutic range of active pharmaceutical ingredients (APIs) usually falls in the low-ng/mL range.
Triple quadrupole mass spectrometers with electrospray ionization (ESI) are the most common detector. Recent design advances facilitate simplified sample preparation. Specifically, detection sensitivity has improved about 50×, which enables dilute samples. This reduces the need for sample concentration. Plus, dilute samples greatly reduce interferences, particularly ion suppression. Multiple reaction monitoring (MRM) is now much more robust; it simplifies quantitation of peaks that overlap or even coelute.
According to Dr. Schoener, phospholipids are the most common culprits in abnormal ESI response, but other surfactants, such as poly(ethylene glycol) (PEG) and polysorbates, also cause problems. Fortunately, the matrix effect on ionization is easy to measure: Simply compare the response in dilute solution to the response in a spiked matrix, or compare the analyzer response to an internal standard in the matrix and neat solution. The matrix effect should be less than 25%.
The European Medicines Agency recently issued guidance on matrix effects that takes effect in February 2012. Six disparate matrix lots should give matrix factors (MF) at low and high method limits of <15% CV. Further, the MF should be consistent from individual to individual.
Phospholipids are easy to detect with the mass spectrometer. The in-source decay can be set at m/e 184; with MRM, an m/e of 184 indicates compounds containing choline. These should not elute near the peaks of interest.
Dr. Schoener was also very positive about the use of solid-phase extraction (SPE) in a variety of formats from cartridges and plates. However, on-line SPE has the particular advantage of avoiding the problem-prone, time-consuming dry-down steps.
At the same workshop, Dr. Matt Hengle (University of California at Davis) addressed food safety, with a particular focus on hops for brewing beer. Hops is a leading legal crop in California. Many countries have strict restrictions on imported hops in order to protect the quality of the beer. In 1999, the regulations focused on six pesticide residues. Hops are different than most other food-stuffs: Water content is only about 10%. The target pesticides have a high affinity for the major constituents, i.e., pigments, waxes, and oils. Initially, sample prep was acetone extraction followed by gel permeation chromatography (GPC) and SPE followed by HPLC or GC-MS. It required about 24 hr to process 12 samples and used 1 L of solvent/sample.
In 2006, the list of target analytes increased to 16. Several required LC-MS. The method was modified by replacing acetone with acetonitrile for extraction. SPE with an NH2 phase provided a clear sample. Organic solvent consumption was reduced to 70 mL; 12 samples could be processed in 4 hr. In addition, the chromatographic windows were clear of matrix interferences. Limit of quantitation (LOQ) improved to 0.1 ppm, which is about 10 times better than required.
In 2009, the regulators proposed adding 17 additional registered compounds (for a total of 33) to the target list. Some were not compatible with amino SPE. A new MS reduced the on-column LOQ to 5 pg on column from 200. Thus, sample size could be reduced. Plus, the extract could be diluted, which reduced interferences. Solvent usage per sample was also reduced to 30 mL. Twelve samples can be processed in 2.5 hr.
As the method evolved over the years, Dr. Hengle tried several popular sample prep protocols including QuEChERS and a variety of SPE phases. Relying on his prior experience with hops, some sample prep approaches could be eliminated quickly since the color of the extract indicated that matrix interference was still present. The take-home message was: Know your sample, especially the matrix.
Applications of solid-phase extraction for proteomics
In the parallel universe of proteomics, the emphasis is on SPE LC-MS of much smaller samples of biologicals from body fluids and tissue. Body fluids, especially plasma, are a difficult matrix since the target analytes are only at trace concentrations and the most abundant proteins are 1010 more concentrated. In addition, the most abundant proteins may reversibly bind to many of the least abundant, and thus sequester them. Again, the sample matrix must be considered.
Solid-phase extraction is also attractive for proteins and peptides, especially when bioaffinty ligands give high selectivity. However, the sample sizes are small, which means the amount of extractant must also be small. One has a choice of on-line or off-line approaches. The Perfinity Biosciences (Lafayette, IN) LC-MS workstation from Shimadzu (Columbia, MD) uses up to six trap and analysis columns to digest serum proteins to peptides in 10 min. The stationary phases include a variety of affinity chromatography columns including Perfinity G, Tetravidin, and Monoavidin. The latter two ligands are designed to grab to biotin-labeled compounds. The particular configuration of columns depends on the focus of the study. For example, antibodies are trapped from serum with Perfinity G. Other serum proteins wash through and are discarded. The release of antibodies and passing to an immobilized trypsin column require less than 10 min for digestion to peptides. Next, the digest goes to the desalting column where peptides are trapped at the top. Compounds that interfere with MS detection are in the wash-through. Next, the peptides flow to a high-resolution reversed-phase liquid chromatography (RPLC) column with ESI MS-MS for peptide identification, usually with multiple reaction monitoring (MRM). The entire workstation, including LC-MS, is marketed by Shimadzu. The columns are from Perfinity Biosciences.
Glygen Corp. (in Columbia, MD) is a leader in nanoscale off-line SPE. Glygen exploits a novel SPE design to make SPE tips with over 30 different chromatographic phases, including group- and compound-specific affinity phases. For example, zirconia and titania are recommended for phosphopeptides, and graphitized carbon is selective for glycopeptides and proteins. Starting with a narrow-diameter microtip for air-driven pipets, column packings are mechanically imbedded (without glue or binder) along the interior capillary surface. These active tips facilitate SPE with volumes in the sub-μL range. The active pipet tips have received 300 literature citations over the last two years, most of which are preparing samples for LC-MS. Glygen has also imbedded particles into the walls of open tubular capillaries for on-line extraction. The “bed volume” of extractant of these tips is much lower than packed tips, which reduces sample loss due to nonspecific adsorption.
Future of LC-MS
Mass spectroscopists should expect further improvements in detection sensitivity, probably at 10× for five years. New techniques analogous to MRM should also be expected. This will increase the value of new instruments. Hopefully, we will also see new ionization modes as well as improvements to existing modes that will facilitate better quantitative analysis.
For chemists, the most interesting developments will arise from applying chemistry to the separation protocol. Reducing matrix effects is an obvious example. Perhaps a choline-specific affinity SPE cartridge could reduce suppression problems.
For chromatographers, one can already see that ultralow-dispersion instruments fitted with capillary columns will reduce the mobile phase load for the MS while improving mass detection. Capillary LC (cLC) is not easy or forgiving today, but good engineering will address and solve the rub points. The driving force will be that cLC tames the often cranky MS, making it more powerful and generally applicable.
- What’s New in Sample Preparation for LC and LCMS; Sept 14, 2011. Organized by CASSS, Emeryville, CA.
Robert L. Stevenson, Ph.D. is a Consultant and Editor of Separation Science for American Laboratory/Labcompare; e-mail: firstname.lastname@example.org.