As HPLC systems become faster and employ increasingly exotic stationary phases, sample preparation becomes essential for reliable, reproducible analyses.
Biological samples generally require concentration in the target analyte and, in the case of proteins, depletion of high-concentration interfering materials. The value of concentration/ depletion is illustrated by the dilemma facing proteomic analysis, particularly using blood-derived samples. In plasma, the concentration dynamic range between the most abundant protein, albumin, and the protein of interest may be as high as 10-13. Sample concentration may be carried out through dialysis or with specialized evaporation systems. Porvair Sciences (Leatherhead, UK) and other vendors sell devices that blow heated nitrogen into individual microwells to speed up evaporation and sample concentration. Depletion is somewhat more difficult to achieve, particularly when purifying a low-abundance protein in the presence of high-concentration species.
Depending on the sample type, common prep methods may include solid phase extraction (SPE), centrifugation, or liquid-liquid extraction. SPE is straightforward and available as single disposable cartridges filled by hand or through automated instruments. Protein prep for most LC-MS often involves an additional step, proteolysis, which breaks large molecules into smaller, more manageable peptides. Filtration is the simplest and perhaps most essential of LC sample prep techniques. Schleicher & Schuell MicroScience (S&S; Dassel, Germany) suggests filtering through a 0.45-micron microporous syringe membrane filter to break up emulsions and retain particulates that clog columns.
S&S recommends regenerated cellulose (RC) membranes, which are hydrophilic, are highly resistant to HPLC solvents, and show very low extractables. According to the company, RC is one of the weakest proteinbinding membranes, which means it’s suitable for proteomic analysis.
Automation: Not just for high throughput
Sample preparation is an oft-overlooked source of error and inconsistency in HPLC analysis. Column and mobile phase conditions may be optimal and steady from run to run, but if samples are prepared differently, they will run differently as well.
That is why Simon Robinson, HPLC product manager at Shimadzu (Columbia, MD), recommends automating sample prep whenever possible. “Otherwise, reaching a point where data becomes untrustworthy is quite likely,” he says. “Automated sample preparation eliminates human error.” It also works unattended, freeing up modest chunks of time for occasional chromatographers and several hours per day for heavy users. Automated sample prep may be integrated directly with an auto-injector for overnight operation.
One drawback of automated sample prep is that standards and samples need to be refrigerated, which introduces scheduling issues and somewhat reduces the attractiveness of hands-off operation.
Shimadzu has been working closely with Perfinity Biosciences (West Lafayette, IN) on a next-generation LC sample prep system. Through this collaboration, Shimadzu provides the LC hardware while Perfinity supplies the Perfinity Workstation, workflow, methods, user interface software, column technology, and reagents. Shimadzu is responsible for marketing, sales, installation, and hardware support, while Perfinity provides applications support and consumables.
The companies share more than a common product. Both agree that thinking in terms of throughput thresholds before adopting automation is a mistake. “We believe that automation is appropriate for all sample prep methodologies regardless of throughput demands,” says Nicholas Herold, research scientist at Perfinity. “With conventional prep, sample loss and contamination can be significant. Not so with direct, automated transfer between steps in a closed system. Automation keeps the focus on answers instead of workflows.”
The Perfinity Workstation couples directly with any ESI-MS through contact closure and operates the mass detector software in “acquisition only” mode. An optional fraction collector provides the capability of transferring samples manually for further downstream analysis, such as MALDI-MS.
Don’t confuse cost and value
Paul Boguszewski, product manager for sample preparation products at Agilent (Santa Clara, CA), notes that automated sample preparation plays into three significant trends in LC analysis:
- Productivity (or lack thereof), speed of analysis, and throughput, as typified by fast, high-pressure systems (UPLC and fused core technology)
- Sensitivity at ever-lower analyte concentrations, which is the justification for adopting mass detectors (particularly triple-quad)
- Increasing complexity, particularly for biological samples
Mr. Boguszewski compares the wide variation in complexity to “drinking water compared with fruit, compared with blood.”
To put this into perspective, consider that tap water holds a few dozen detectable contaminants, and fruit perhaps several thousand compounds and minerals. The human genome consists of 30,000 genes; human blood contains more than 100,000 individual components; and the human proteome includes about 1 million proteins, whose concentrations vary widely. Protein biomarkers currently under study occur at concentrations below 10 ng/mL and account for no more than 1 percent of blood proteins by weight. The 22 highest-abundance proteins comprise the remaining 99 percent.
Agilent has constructed a matrix of preparation methods suitable for small-molecule (e.g., metabolite or drug) HPLC analysis. Solid phase extraction is the most complete, versatile technique, removing particulates, proteins, lipids, oligomeric surfactants, and salts. At the other end of the efficiency scale, performing just a tad better than “dilute and shoot,” is liquid extraction, which removes proteins only partially and not much of anything else.
In between are dried matrix spotting, precipitation/filtration, and “smart” titration, which, successively, provide more efficient cleanup. Adoption of these methods, some of which are relatively new, has been driven by mass detection.
But as techniques become more efficient, they also raise the cost of sample preparation and complicate the workflow. Many lab managers view this as an absolute negative.
This is a mistake, says Mr. Boguszewski, who claims the higher cost is illusory; cleaner samples translate to more robust, reproducible LC analysis, with fewer repeat runs. “Plus,” he says, “if you use one of the more efficient preparation methods, you’re more likely to have a cleaner LC/MS system than if you take the less-efficient, lower-cost route.”
Angelo DePalma holds a Ph.D. in organic chemistry and has worked in the pharmaceutical industry. You can reach him at firstname.lastname@example.org.