522
DOI 10.1002/prca.201400081
Proteomics Clin. Appl. 2015, 9, 522–530
RESEARCH ARTICLE
Capillary-channeled polymer (C-CP) fibers for the rapid extraction of proteins from urine matrices prior to detection with MALDI-MS Benjamin T. Manard, Sarah M. H. Jones and R. Kenneth Marcus Department of Chemistry, Clemson University, Clemson, SC, USA Purpose: While MS is a powerful tool for biomarker determinations, the high salt content and the small molecules present in urine poses incredible challenges. Separation/extraction methods must be employed for the isolation of target species at relevant concentrations. Micropipette tips packed with capillary-channeled polymer (C-CP) fibers are employed for the SPE of proteins from a synthetic and a certified urine matrix. Experimental design: Extractions are performed utilizing a very simple centrifugation method to spin-down species through the C-CP fiber tips. Proteins adsorb to the hydrophobic polypropylene fibers and are eluted in a solvent suitable for MALDI-MS analysis. Figures of merit are determined for representative compounds 2-microglobulin, retinol binding protein, and transferrin. Results: The optimum protein processing included a 100 L aqueous rinse and an elution solvent composition was 10 L of 55:45 ACN:water (with triflouroacetic acid). MALDI-MS responses for the target proteins are improved from nondetectable levels to eventually yield LOD ranging from 5 to 180 nM in 1 L aliquots. Conclusion and clinical relevance: C-CP fiber tips offer a plethora of advantages including low materials costs, high throughput, microvolume processing, and the determination of sub-nanogram quantities of analyte; allowing determination of biomarkers that are otherwise undetectable in urine matrices.
Received: August 4, 2014 Revised: October 10, 2014 Accepted: November 25, 2014
Keywords: Capillary-channeled polymer fibers / MALDI-MS / Proteins / Solid-phase extraction / Urine
1
Introduction
The field of clinical chemistry has shown a rapid expansion over the past decade in terms of biofluids analysis with MS. Analysis of biofluids (i.e., plasma, urine, saliva, etc.) for specific proteins and other biomolecules is necessary for the investigation of bodily functions and the presence or absence of disease. Of the biofluids, urine specimens are the most
Correspondence: R. Kenneth Marcus, Department of Chemistry, Clemson University, Clemson, SC 29634, USA E-mail: [email protected] Abbreviations: C-CP, capillary-channeled polymer; DI, deionized; FEP, fluorinated ethylene polypropylene; HAT, human African trypanosomiasis; MW, molecular weight; PP, polypropylene; RBP, retinol binding protein; RCF, relative centrifugal force; S/B, signalto-background; SBR, signal-to-background ratio; Tf, transferrin; UV-Vis, ultraviolet-visible spectroscopy C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
easily obtained and most commonly used samples in clinical analysis [1, 2]. Urinary proteins have potential to serve as excellent biomarkers not only for kidney and urological diseases, but also many other small proteins that can freely filter through the glomerular barrier as widely described [1, 3]. However, due to its low protein concentrations and matrix complexity, urine is considered a difficult biofluid to analyze. With the recent advancements in MS, urinary biomarkers can be screened rapidly with excellent sensitivity and specificity when compared to spectroscopic techniques (i.e., ultravioletvisible spectroscopy (UV-Vis) absorbance) [4, 5]. MS analysis of biological compounds commonly employs one of two soft ionization methods: (1) ESI and (2) MALDI. In ESI, the sample is ionized in a nebulization process in which multiply charged ions are produced [6–8]. While multiply charged ions decrease the required mass range to values easily accessed by Colour Online: See the article online to view Figs. 1, 3, 4 and 5 in colour. www.clinical.proteomics-journal.com
Proteomics Clin. Appl. 2015, 9, 522–530
523
Clinical Relevance The isolation of proteins or other target species from urine is probably one of the greatest roadblocks to the identification, quantification, and utilization of biomarkers. The salt/electrolyte matrix of urine makes direct protein determinations via MS nearly impossible. We present here a straightforward, low cost, and relatively high throughput methodology for SPE of proteins from urine matrices prior to MALDI-MS analysis. Polypropylene capillary-
quadrupole and ion trap analyzers, having multiply charged species is detrimental in the analysis of complex mixtures as spectral overlap of different proteins’ charged states complicates interpretation, on those MS platforms. To correct for this complexity, separation techniques (e.g., CE) must be employed prior to the ionization. Spectral overlap issues are alleviated on relatively common high-resolution spectrometers, such as quadrupole TOF (q-TOF) instruments. MALDI, while generally not as sensitive as ESI, is an excellent ionization source for complex samples as predominantly singly charged ions are generated via proton transfer from the added matrix. This fact makes MALDI-MS an excellent approach for rapid, top-down detection of disease-specific biomarkers [2, 6, 9]. Unfortunately, for both ionization sources, the analytes of interest (i.e., proteins) must be extracted from their native biofluids, as salts and small organic molecules can be detrimental to analysis [10–13]. With MALDI-MS, salts/organics are co-ionized with the target proteins, causing signal suppression up to the point that that proteins are not detectable at physiologically relevant concentrations [10]. This is a great limitation when analyzing biofluids such as urine. Protein concentrations in urine are very low in comparison to other biofluids (e.g., plasma), which places greater emphasis on sample preparation [14]. In order to overcome these issues, a variety of sample preparation techniques have been utilized to isolate the target species prior to MALDI-MS, as described by Zerefos and Vlahou [15]. Methods, or the combination of methods, that are typically employed for the extraction of proteins from biofluids for MS detection include HPLC [5], SPE [16–18], ultracentrifugation [19], ultrafiltration [2], and gel electrophoresis [20–23]. Among these options, a top-down approach utilizing an SPE extraction procedure prior to analysis with MS offers rapid, reproducible, and robust technique for biomarker. Current sample preparation trends have shifted toward miniaturization, in which rapid extractions (including some level of automation) can be performed on small sample volumes [13, 24, 25]. SPE works on a basic principle in which target species are adsorbed onto a stationary phase, similar to HPLC, while contaminants are either not retained or C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
channeled polymer (C-CP) fibers are assembled in micropipette tips, allowing for spin-down processing on benchtop microcentrifuges. Proteins adsorb to the hydrophobic fiber surface and are subsequently eluted in MALDI-friendly solvents. We demonstrate here analytical recoveries for three target proteins (2-microglobulin, retinol-binding protein, and transferrin) yielding sensitivities that are relevant toward biomarker development or clinical analyses.
washed away in low-strength solvents/buffers. Target analytes are then eluted in an MS-friendly solvent for subsequent analysis. SPE methods differ from HPLC methods as they utilize lower sample volumes (50 mm/s) protein separations without penalties typical of van Deemter C-term broadening [40]. In terms of SPE, microbore columns can be used as inline desalting columns for ESI-MS [41] or simply cut into centimeter-long segments and affixed to standard micropipette tips [42–45]. In this mode, discrete samples in heavy matrices (buffer, saliva, and urine) can be processed in parallel, with load, wash, and elute steps performed in 1-min time frames using a table-top microcentrifuge. This approach allows for protein determinations via UV-Vis absorbance, ESI-MS, and MALDI-MS with high sensitivity [42–45]. In this study, PP C-CP fiber micropipette tips were employed as -SPE media to adsorb proteins from urine matrices prior to MALDI-MS analysis. Mock urine and certified human urine matrices were spiked with typical urinary proteins (2-microglobulin, retinol binding protein (RBP), and transferrin (Tf)) at biologically relevant concentrations. Method development was investigated in terms of the aqueous washing volumes and the elution solvent composition. Proteins possess different amino acid compositions, being more or less hydrophobic, potentially requiring a different organic solvent strength for the elution step. After a systematic evaluation of the methodology, the analytical figures of merit were evaluated for the target application in urinomics. Prior to extraction, the analytes of interest were virtually undetectable, but were readily identified following this simple SPE procedure. While this paper does not pertain to the identification of specific biomarkers, the results illustrate the utility of the stationary phase for the extraction of target species from microliter-size aliquots of urine. It is believed the present methodology illustrates great promise in terms of microscale biofluids analysis for biomarker determinations. Applications are foreseen to extend from neonatal subjects to forensics analysis as high quantities of biofluids may not be available in those instances.
Proteomics Clin. Appl. 2015, 9, 522–530
2.2 C-CP fiber tip construction The construction of the C-CP fiber packed tips remains relatively unchanged from previous publications [43]. C-CP fibers are packed with an interstitial fraction of i = 0.6 in 0.8 mm i.d. fluorinated ethylene polypropylene (FEP) tubing (Cole Parmer, Vernon Hills, IL, USA), following the studies previously performed with regards to chromatographic separation efficiency [37]. C-CP fibers are pulled through the FEP tubing such that the fiber capillary walls interdigitate to form collinear channels. In this study, a total of 658 PP C-CP fibers (55 m in diameter) were packed in each tip. The tips were cut with a surgical-grade scalpel with one end being flush with the fibers and the other end leaving 5 mm void space for slip-fitting onto the end of the 200 L pipette tips from Neptune (San Diego, CA, USA).
2.3 Adaptation for centrifugation fluidic processing In order to increase reproducibility, sample throughput, and run parallel experiments, centrifugation was employed for fluid manipulation as opposed to common aspiration methods. To adapt to a centrifugation format, the C-CP fiber packed tips were configured so they could be inserted into 15 mL centrifuge tubes (VWR, Radnor, PA, USA). The micro-SPE tip was placed inside of a 1.7 mL microtube (Genesse Scientific, San Diego, CA, USA) that was cut leaving 10 mm of the bottom removed. This assembly was then placed in the 15 mL centrifuge for use in a Symphony 4417 (VWR) centrifuge operating at 1900 × g relative centrifugal force (RCF), as described previously [41].
2.4 Extraction procedure employing C-CP micro-SPE tips
2
Materials and methods
2.1 Reagents and chemicals The three test proteins, 2-microglobulin (human urine), RBP (human urine), and Tf (human), were obtained from Sigma-Aldrich (St. Louis, MO, USA) and prepared (0.4– 10 M each) in mock urine and a certified, drug-free human urine control from UTAK laboratories (Valencia, CA, USA), and MilliQ water (18.2 M⍀•cm) derived from a NANOpure Diamond Barnstead/Thermolyne Water System (Dubuque, IA, USA) for comparison to ideal conditions. The elution solvents were prepared with MilliQ water, with various percentages of American Chemical Society grade ACN and 0.07% TFA, both from EMD Millipore (Billerica, MA, USA). Sinapinic acid (Sigma-Aldrich), prepared as 20 mg/mL in 50:50 ACN:H2 O with 0.1% TFA, was employed as the matrix required for the MALDI-MS analysis. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The complete extraction and sample spotting procedure is depicted in Fig. 1. After sample collection, protein-spiked urine samples (10 L) were loaded onto the PP C-CP micro-SPE tips and centrifuged for 30 s at 1900 × g RCF. Following the load step, an aqueous wash was performed wherein deionized (DI) water (50–150 L) was centrifuged through the tips for 60 s to remove nonadsorbed contaminants (i.e., salts). The PP fibers are employed for their hydrophobic characteristics that can be utilized for RP separations. To elute the proteins from the PP C-CP fiber surface, 10 L of the elution solvent (ACN:H2 O with 0.07% TFA) was centrifuged through the tips for 30 s and the purified proteins captured in sealable plastic vials (Fisherbrand, Pittsburgh, PA, USA) and stored at 5⬚C until being mixed with the MALDI matrix, target spotting, and analysis. The total sample extraction time is approximately 5 min for 12 extractions (only limited by centrifuge capacity). www.clinical.proteomics-journal.com
Proteomics Clin. Appl. 2015, 9, 522–530
Figure 1. Illustration of the procedure employing C-CP fiber packed -SPE tips for the extraction of urinary proteins from urine matrices prior to analysis by MALDI-MS.
2.5 MALDI-MS A Bruker Daltonics (Billerica, MA, USA) Microflex LRF, MALDI-TOF-MS was employed in positive ion, linear mode for the detection of extracted species. Bruker Compass software was employed for instrumentation control and data processing. Further data analysis was performed with Microsoft Excel (Redmond, WA, USA). The MALDI-MS employed a nitrogen (337 nm) laser with a pulse rate of 60 Hz for the sample ionization. Sample spectra were acquired with 200 laser shots at 60% power. Sinapinic acid (20 mg/mL) in 50:50 ACN:H2 O with 0.1% TFA was utilized as the MALDI matrix, where 1 L of matrix was spotted onto the MALDI target, allowed to dry, and then a 1 L aliquot of eluted sample was spotted on the matrix. Reported S/N were computed as the SD of the background signal across a 100 Da window, centered 2000 Da below the pseudo-molecular ion (i.e., (M+H)+ ) peaks of the target proteins.
3
Results and discussion
Presented here, C-CP fibers are investigated as a stationary phase for micro-SPE tips for the extraction of urinary proteins from urine matrices. Urinary proteomics is rapidly growing in terms of clinical studies for potential biomarker analysis. 2-microglobulin, RBP, and Tf were investigated as the “target” analytes due to their common presence in biofluids as well as being commonly cited as potential biomarkers [21, 46, 47]. These three proteins exhibit various physical characteristics particularly their molecular weight (MW), three-dimensional structure, and amino acid compositions. 2-microglobulin being the smallest of the three has a MW of 11.6 kDa and is 45 × 25 × 20 A⬚ in size containing 180 amino acids [48]. RBP is slightly larger (20 kDa) containing 182 amino acids. While similar in size, the structure of RBP C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
525 is more complex as it is composed of an N-terminal, -sheet, ␣-helix, and a C-terminal core compared to 2-microglobulin having only 2  structures [49]. Tf on the other hand is much larger (80 kDa) and 45 × 57 × 135 A⬚ in size and containing 678 amino acids [50]. These different physical characteristics could provide insight to the hydrophobic interaction with the PP C-CP fiber surface. When discussing biomarker employment, 2-microglobulin levels in cerebrospinal fluid have been studied extensively as human African trypanosomiasis (HAT) biomarkers as increased levels detected after Western blot and ELISA testing are positively correlated with HAT [21]. 2-microglobulin has also been suggested as a biomarker for acute rejection in renal transplantation [51]. Studies have noted RBP as a potential biomarker as the increased levels of RBP reflect its reduced uptake, reflective of proximal tubular dysfunction [46]. Recent studies have also suggested RBP levels in plasma could be employed for the early detection of coronary heart disease [47]. While Tf has been studied extensively for renal failure, recently, a correlation between increased levels of excreted Tf could be related to drug nephrotoxicity [52]. With much effort being devoted to the monitoring of these proteins, urine matrices spikes with this suite of proteins would serve as a useful demonstration of the utility of this technique. Based on literature reports, the certified human urine and representative mock urine matrices were spiked at concentrations of 7.1, 7.9, and 10.4 M of the respective target proteins (2-microglobulin, RBP, and Tf) and used as the primary test samples. The resultant MALDI-MS spectra obtained from these samples without any form of sample clean-up (1 L spotted onto 1 L of matrix) are shown in Fig. 2A and B, respectively for the human and mock urine samples. It is known that target proteins are virtually nondetectable by MALDI-MS in the presence of 100 mM [53] salts and organics, even more so for biofluids. As can be seen, there are no discernible spectra features that can be attributed to the target proteins in the case of the human urine, though a slight signal is seen that can be assigned to the 2-microglobulin in the mock urine solution. Indeed, the intensity of the spectral continuum above 15 000 Da is greater for the human urine than the mock urine. Based on the system data processing package, the background signal for the human urine has a peak-to-peak deviation of 40 counts versus to the mock urine deviation of 7 counts. This is not surprising the former is composed of a wide variety of small molecules/salts not included in the mock urine matrix. It is for this situation that the C-CP fiber SPE methodology is being developed.
3.1 Role of elution solvent composition on protein responses A key aspect to the simplicity of SPE techniques is that they utilize an isocratic approach in terms of the elution solvent composition, that is, a single solvent composition is employed to simultaneously remove the analytes from the www.clinical.proteomics-journal.com
526
B. T. Manard et al.
Proteomics Clin. Appl. 2015, 9, 522–530
Figure 3. Effect of elution solvent composition on the intensity of (M+H)+ signal levels for 2-microglobulin and RBP in human urine and mock urine matrices. Elution solvent volume = 10 L, aqueous wash volume = 100 L.
Figure 2. Benchmark MALDI mass spectra for the (A) certified human urine and (B) mock urine matrices spiked to levels of 7.1, 7.9, and 10.4 M of the respective target proteins 2-microglobulin, RBP, and Tf.
stationary phase surface. Various modes of chromatography (i.e., ion exchange, RP) and stationary phases (i.e., polymer, C18 ) have been employed for SPE of various compounds [29, 54]. Polymer phases are attractive due to their chemical robustness and immunity to harsh elution solvents [54, 55]. The PP C-CP fibers are employed here for RP separation of proteins [40], meaning that organic-based elution solvents (or organic-aqueous mixtures) are employed for the removal of adsorbed species. Due to the different degrees of hydrophobicity among proteins, a global optimum in the elution solvent strength is desired to affect the most uniform and efficient recovery. In this case, 10 L of the protein suite (in synthetic and natural urine matrices) was loaded onto the C-CP fiber packed -SPE tips. A single 100 L aqueous wash was then performed to remove the unwanted species (e.g., salt), as this was a reasonable volume based on previous works. Finally, the proteins were eluted with 10 L solvent aliquots ranging from 50 to 75% ACN in water with 0.07% TFA, representing C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
compositions which are common for RP separations of proteins from C-CP fibers and also compatible with MALDIMS. The study was performed by monitoring MALDI-MS (M+H)+ intensities of the test proteins for deposited spots for each of solvent compositions. As depicted in Fig. 3, there is an optimum range of solvent compositions that yield optimum MALDI-MS responses for the three proteins, which is not dependent on whether the mock of human urine matrices are applied. The highest signal recoveries were found when employing an elution solvent of 55:45 ACN:H2 O with 0.07% TFA. This result varies slightly from previous reports utilizing PP C-CP tips (60:40 ACN:H2 O), as different proteins were utilized [45]. This response is consistent with what is seen for other RP SPE media used in protein extractions, reflecting a balance in driving forces between the affinity of the proteins for the hydrophobic surface and their solubility in solvents of lower and higher organic (ACN) content. At low ACN compositions, the proteins favor the hydrophobic PP surface, and at higher values the solvent is too nonpolar for the proteins to be solubilized. It is noteworthy that the responses for each protein are greater for the mock urine matrices than the certified human urine. This is not surprising as the latter is a more complex matrix. At the same time, the background continuum for the human urine was 5× higher than the mock urine (consistent with Fig. 2A and B), implying the extraction procedure should be optimized in terms of buffer removal in the aqueous wash step. It is also interesting that the relative signal responses as a function of solvent composition for the two proteins track each other within each matrix. As a reference, identical concentrations of the three proteins prepared in milliQ-H2 O (as a “clean” sample) were also studied to monitor analyte recoveries. Proteins recovered from water and mock urine gave nearly identical spectra and yields as a function of solvent composition. www.clinical.proteomics-journal.com
Proteomics Clin. Appl. 2015, 9, 522–530
527
Figure 4. Effect of aqueous wash volume on the intensity of (M+H)+ signal levels for 2-microglobulin and RBP in human urine and mock urine matrices. Elution solvent (55:45 ACN:H2 O with 0.07% TFA) volume = 10 L.
3.2 Role of aqueous wash volume on protein responses Previous studies in the use of the C-CP fiber -SPE approach prior to MALDI-MS of proteins have illustrated the key characteristic that, while salts/buffer components have a low affinity for the PP surface, high concentrations require that more thorough aqueous wash steps must be performed. The protein molecular ion responses and continuum background differences discussed relative to Fig. 3 point clearly the fact that the human urine matrix poses a greater challenge in terms of contaminant species removal. In low concentrations, buffers can be readily removed with minute aqueous wash volumes (5 L). However, salt/buffer concentrations of up to 1 M or viscous biofluids such as saliva, require larger volumes [45]. Analogous to the studies of elution solvent strength, the respective (M+H)+ responses for 2-microglobulin and RBP in the two urine matrices are plotted in Fig. 4 as a function of the volume of aqueous wash employed. In each case, the applied sample volume was held at 10 L, and the optimum elution solvent composition identified above (55:45 ACN:H2 O with 0.07% TFA) was employed. Similar to the previous studies, the responses depicted in Fig. 4 suggest that matrix removal is most efficient when the aqueous wash is 10× the sample load volume for these matrix-heavy solutions. Not surprising, the signals for the proteins in the mock urine can be more effectively recovered at lower wash volumes (75–100 L) in comparison to the volume required for the more complex matrix of human urine (100 L). Indeed, those ion yields show a very steep increase with wash volume increases. Simply, the lower wash volumes do not effectively remove interfering salt species as the MALDI-MS analyte signals suffer from ion suppression. Based on the results depicted in Fig. 4, 100 L of water was employed to remove the urine matrix in the following experiments. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 5. MALDI mass spectra of mock urine specimen spiked to levels of 7.1, 7.9, and 10.4 M of the respective target proteins (2-microglobulin, RBP, and Tf), following PP C-CP fiber tip SPE. (A) full mass spectrum, (B) spectral background realized for the case of human urine before and after C-CP fiber clean-up. Elution solvent (55:45 ACN:H2 O with 0.07% TFA) volume = 10 L, aqueous wash volume = 100 L.
3.3 Analytical characteristics following C-CP fiber -SPE extraction Following the evaluations of the elution and aqueous wash steps, the composite analytical method includes initial loading of 10 L aliquot of the test solution onto a PP C-CP -SPE tip, followed by a 100 L aqueous wash, then elution with 10 L of 55:45 ACN:H2 O with 0.07% TFA. The application and elution steps are allowed to spin down in the microcentrifuge for 30 s and the wash for 1 min. Figure 5A presents the product mass spectrum for the analysis of the mock urine specimen, spiked with the three target proteins. The spectrum is directly comparable to that of the raw sample presented in Fig. 2B. The (M+H)+ ions for each of the three spiked proteins, 2-microglobulin (11 716 Da), RBP (20 581 Da), and Tf (81 791 Da) are all seen with high spectral clarity. Additional to the pseudo-molecular ions, the doubly charged ((M+2H)2+ ) species are also seen for Tf (40 078 Da) and RBP (10 253 Da). The lower response for RBP could be www.clinical.proteomics-journal.com
528
B. T. Manard et al.
Proteomics Clin. Appl. 2015, 9, 522–530
Figure 6. MALDI mass spectrum of a certified human urine specimen spiked to levels of 1.14, 1.26, and 1.66 M of the respective target proteins (2-microglobulin, RBP, and Tf), following PP C-CP fiber tip -SPE. Elution solvent (55:45 ACN:H2 O with 0.07% TFA) volume = 10 L, aqueous wash volume = 100 L.
explained by lack of purity (as noted in the supplier documentation) while Tf is approaching the optimized instrument mass limit of 100 000 Da. While the vast improvement in the target protein responses is dramatic, it is important to note the improvement as well in the continuum spectral background. This is demonstrated for the worst-case situation of the certified human urine matrix in Fig. 5B, where both the average value and the SD are much improved in the mass range of 60–70 000 Da. Specifically, the background for the preextraction case is 16.1 ± 3.9 counts, while following the extraction it is 7.1 ± 2.8. With the optimized procedure, LODs were assessed to determine whether or not the PP C-CP fiber -SPE tips would indeed be relevant for extraction of low mass quantities in human urine. The single-point LODs were determined by use of the “SBR-RSD” method originally described by Boumans and Vrakking [56]. In this approach, the LOD is defined as: LOD =
(0.01) k (RSD) m (S/B)
(1)
where RSD is the RSD of the adjacent spectral background, m is the concentration of the test element in the specimen, S/B is the SBR, and the k is a statistical factor equal to 3, which reflects a 99% confidence interval. In order for this method to be more accurate, the analyte concentrations in the test specimen were reduced to 1.14, 1.26, and 1.66 M for 2-microglobulin, RBP, and Tf, respectively, correlating to 0.13, 0.24, and 1.33 g total mass of each applied to each micropipette tip. The resultant mass spectrum is depicted in Fig. 6. Detection limits were determined to be 5.3, 30.3, and 184 nM, respectively. When referring to the 1 L extract spots placed on the MALDI targets (which are not totally ablated), these molar values equate to mass-based LODs of 60 pg for 2-microglobulin, 600 pg for RBP, and 15 ng for Tf. The discrepancies in LODs are readily explained in terms of the chemical nature of each, as well as the aforementioned C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
decrease in ion throughput/detector response as a function of mass for this particular instrument. (Additionally, it cannot be ignored that this particular instrument is that manufacturer’s lowest performance model.) All of these points made, it is quite clear that the performance demonstrated here is applicable to the field of urinomics, particularly for biomarker screening. Important biomarker concentrations are 0.05 mg/mL, corresponding to 500 ng per 10 L of volume, as processed here. This level is demonstrated for 2-microglobulin and RBP that were on the picogram level. It is very difficult to compare absolute performance among extraction methods used for urinary proteomics as the analytical platforms (e.g., ionization methods and mass analyzers) vary greatly. When compared to the aforementioned new SPE methods for biomarker analysis, the use of mesoporous silica beads [11] and magnetic nanoparticles [27] deal specifically with the urine matrix. Both methods require sample aliquots in the same regime (
DOI 10.1002/prca.201400081
Proteomics Clin. Appl. 2015, 9, 522–530
RESEARCH ARTICLE
Capillary-channeled polymer (C-CP) fibers for the rapid extraction of proteins from urine matrices prior to detection with MALDI-MS Benjamin T. Manard, Sarah M. H. Jones and R. Kenneth Marcus Department of Chemistry, Clemson University, Clemson, SC, USA Purpose: While MS is a powerful tool for biomarker determinations, the high salt content and the small molecules present in urine poses incredible challenges. Separation/extraction methods must be employed for the isolation of target species at relevant concentrations. Micropipette tips packed with capillary-channeled polymer (C-CP) fibers are employed for the SPE of proteins from a synthetic and a certified urine matrix. Experimental design: Extractions are performed utilizing a very simple centrifugation method to spin-down species through the C-CP fiber tips. Proteins adsorb to the hydrophobic polypropylene fibers and are eluted in a solvent suitable for MALDI-MS analysis. Figures of merit are determined for representative compounds 2-microglobulin, retinol binding protein, and transferrin. Results: The optimum protein processing included a 100 L aqueous rinse and an elution solvent composition was 10 L of 55:45 ACN:water (with triflouroacetic acid). MALDI-MS responses for the target proteins are improved from nondetectable levels to eventually yield LOD ranging from 5 to 180 nM in 1 L aliquots. Conclusion and clinical relevance: C-CP fiber tips offer a plethora of advantages including low materials costs, high throughput, microvolume processing, and the determination of sub-nanogram quantities of analyte; allowing determination of biomarkers that are otherwise undetectable in urine matrices.
Received: August 4, 2014 Revised: October 10, 2014 Accepted: November 25, 2014
Keywords: Capillary-channeled polymer fibers / MALDI-MS / Proteins / Solid-phase extraction / Urine
1
Introduction
The field of clinical chemistry has shown a rapid expansion over the past decade in terms of biofluids analysis with MS. Analysis of biofluids (i.e., plasma, urine, saliva, etc.) for specific proteins and other biomolecules is necessary for the investigation of bodily functions and the presence or absence of disease. Of the biofluids, urine specimens are the most
Correspondence: R. Kenneth Marcus, Department of Chemistry, Clemson University, Clemson, SC 29634, USA E-mail: [email protected] Abbreviations: C-CP, capillary-channeled polymer; DI, deionized; FEP, fluorinated ethylene polypropylene; HAT, human African trypanosomiasis; MW, molecular weight; PP, polypropylene; RBP, retinol binding protein; RCF, relative centrifugal force; S/B, signalto-background; SBR, signal-to-background ratio; Tf, transferrin; UV-Vis, ultraviolet-visible spectroscopy C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
easily obtained and most commonly used samples in clinical analysis [1, 2]. Urinary proteins have potential to serve as excellent biomarkers not only for kidney and urological diseases, but also many other small proteins that can freely filter through the glomerular barrier as widely described [1, 3]. However, due to its low protein concentrations and matrix complexity, urine is considered a difficult biofluid to analyze. With the recent advancements in MS, urinary biomarkers can be screened rapidly with excellent sensitivity and specificity when compared to spectroscopic techniques (i.e., ultravioletvisible spectroscopy (UV-Vis) absorbance) [4, 5]. MS analysis of biological compounds commonly employs one of two soft ionization methods: (1) ESI and (2) MALDI. In ESI, the sample is ionized in a nebulization process in which multiply charged ions are produced [6–8]. While multiply charged ions decrease the required mass range to values easily accessed by Colour Online: See the article online to view Figs. 1, 3, 4 and 5 in colour. www.clinical.proteomics-journal.com
Proteomics Clin. Appl. 2015, 9, 522–530
523
Clinical Relevance The isolation of proteins or other target species from urine is probably one of the greatest roadblocks to the identification, quantification, and utilization of biomarkers. The salt/electrolyte matrix of urine makes direct protein determinations via MS nearly impossible. We present here a straightforward, low cost, and relatively high throughput methodology for SPE of proteins from urine matrices prior to MALDI-MS analysis. Polypropylene capillary-
quadrupole and ion trap analyzers, having multiply charged species is detrimental in the analysis of complex mixtures as spectral overlap of different proteins’ charged states complicates interpretation, on those MS platforms. To correct for this complexity, separation techniques (e.g., CE) must be employed prior to the ionization. Spectral overlap issues are alleviated on relatively common high-resolution spectrometers, such as quadrupole TOF (q-TOF) instruments. MALDI, while generally not as sensitive as ESI, is an excellent ionization source for complex samples as predominantly singly charged ions are generated via proton transfer from the added matrix. This fact makes MALDI-MS an excellent approach for rapid, top-down detection of disease-specific biomarkers [2, 6, 9]. Unfortunately, for both ionization sources, the analytes of interest (i.e., proteins) must be extracted from their native biofluids, as salts and small organic molecules can be detrimental to analysis [10–13]. With MALDI-MS, salts/organics are co-ionized with the target proteins, causing signal suppression up to the point that that proteins are not detectable at physiologically relevant concentrations [10]. This is a great limitation when analyzing biofluids such as urine. Protein concentrations in urine are very low in comparison to other biofluids (e.g., plasma), which places greater emphasis on sample preparation [14]. In order to overcome these issues, a variety of sample preparation techniques have been utilized to isolate the target species prior to MALDI-MS, as described by Zerefos and Vlahou [15]. Methods, or the combination of methods, that are typically employed for the extraction of proteins from biofluids for MS detection include HPLC [5], SPE [16–18], ultracentrifugation [19], ultrafiltration [2], and gel electrophoresis [20–23]. Among these options, a top-down approach utilizing an SPE extraction procedure prior to analysis with MS offers rapid, reproducible, and robust technique for biomarker. Current sample preparation trends have shifted toward miniaturization, in which rapid extractions (including some level of automation) can be performed on small sample volumes [13, 24, 25]. SPE works on a basic principle in which target species are adsorbed onto a stationary phase, similar to HPLC, while contaminants are either not retained or C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
channeled polymer (C-CP) fibers are assembled in micropipette tips, allowing for spin-down processing on benchtop microcentrifuges. Proteins adsorb to the hydrophobic fiber surface and are subsequently eluted in MALDI-friendly solvents. We demonstrate here analytical recoveries for three target proteins (2-microglobulin, retinol-binding protein, and transferrin) yielding sensitivities that are relevant toward biomarker development or clinical analyses.
washed away in low-strength solvents/buffers. Target analytes are then eluted in an MS-friendly solvent for subsequent analysis. SPE methods differ from HPLC methods as they utilize lower sample volumes (50 mm/s) protein separations without penalties typical of van Deemter C-term broadening [40]. In terms of SPE, microbore columns can be used as inline desalting columns for ESI-MS [41] or simply cut into centimeter-long segments and affixed to standard micropipette tips [42–45]. In this mode, discrete samples in heavy matrices (buffer, saliva, and urine) can be processed in parallel, with load, wash, and elute steps performed in 1-min time frames using a table-top microcentrifuge. This approach allows for protein determinations via UV-Vis absorbance, ESI-MS, and MALDI-MS with high sensitivity [42–45]. In this study, PP C-CP fiber micropipette tips were employed as -SPE media to adsorb proteins from urine matrices prior to MALDI-MS analysis. Mock urine and certified human urine matrices were spiked with typical urinary proteins (2-microglobulin, retinol binding protein (RBP), and transferrin (Tf)) at biologically relevant concentrations. Method development was investigated in terms of the aqueous washing volumes and the elution solvent composition. Proteins possess different amino acid compositions, being more or less hydrophobic, potentially requiring a different organic solvent strength for the elution step. After a systematic evaluation of the methodology, the analytical figures of merit were evaluated for the target application in urinomics. Prior to extraction, the analytes of interest were virtually undetectable, but were readily identified following this simple SPE procedure. While this paper does not pertain to the identification of specific biomarkers, the results illustrate the utility of the stationary phase for the extraction of target species from microliter-size aliquots of urine. It is believed the present methodology illustrates great promise in terms of microscale biofluids analysis for biomarker determinations. Applications are foreseen to extend from neonatal subjects to forensics analysis as high quantities of biofluids may not be available in those instances.
Proteomics Clin. Appl. 2015, 9, 522–530
2.2 C-CP fiber tip construction The construction of the C-CP fiber packed tips remains relatively unchanged from previous publications [43]. C-CP fibers are packed with an interstitial fraction of i = 0.6 in 0.8 mm i.d. fluorinated ethylene polypropylene (FEP) tubing (Cole Parmer, Vernon Hills, IL, USA), following the studies previously performed with regards to chromatographic separation efficiency [37]. C-CP fibers are pulled through the FEP tubing such that the fiber capillary walls interdigitate to form collinear channels. In this study, a total of 658 PP C-CP fibers (55 m in diameter) were packed in each tip. The tips were cut with a surgical-grade scalpel with one end being flush with the fibers and the other end leaving 5 mm void space for slip-fitting onto the end of the 200 L pipette tips from Neptune (San Diego, CA, USA).
2.3 Adaptation for centrifugation fluidic processing In order to increase reproducibility, sample throughput, and run parallel experiments, centrifugation was employed for fluid manipulation as opposed to common aspiration methods. To adapt to a centrifugation format, the C-CP fiber packed tips were configured so they could be inserted into 15 mL centrifuge tubes (VWR, Radnor, PA, USA). The micro-SPE tip was placed inside of a 1.7 mL microtube (Genesse Scientific, San Diego, CA, USA) that was cut leaving 10 mm of the bottom removed. This assembly was then placed in the 15 mL centrifuge for use in a Symphony 4417 (VWR) centrifuge operating at 1900 × g relative centrifugal force (RCF), as described previously [41].
2.4 Extraction procedure employing C-CP micro-SPE tips
2
Materials and methods
2.1 Reagents and chemicals The three test proteins, 2-microglobulin (human urine), RBP (human urine), and Tf (human), were obtained from Sigma-Aldrich (St. Louis, MO, USA) and prepared (0.4– 10 M each) in mock urine and a certified, drug-free human urine control from UTAK laboratories (Valencia, CA, USA), and MilliQ water (18.2 M⍀•cm) derived from a NANOpure Diamond Barnstead/Thermolyne Water System (Dubuque, IA, USA) for comparison to ideal conditions. The elution solvents were prepared with MilliQ water, with various percentages of American Chemical Society grade ACN and 0.07% TFA, both from EMD Millipore (Billerica, MA, USA). Sinapinic acid (Sigma-Aldrich), prepared as 20 mg/mL in 50:50 ACN:H2 O with 0.1% TFA, was employed as the matrix required for the MALDI-MS analysis. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The complete extraction and sample spotting procedure is depicted in Fig. 1. After sample collection, protein-spiked urine samples (10 L) were loaded onto the PP C-CP micro-SPE tips and centrifuged for 30 s at 1900 × g RCF. Following the load step, an aqueous wash was performed wherein deionized (DI) water (50–150 L) was centrifuged through the tips for 60 s to remove nonadsorbed contaminants (i.e., salts). The PP fibers are employed for their hydrophobic characteristics that can be utilized for RP separations. To elute the proteins from the PP C-CP fiber surface, 10 L of the elution solvent (ACN:H2 O with 0.07% TFA) was centrifuged through the tips for 30 s and the purified proteins captured in sealable plastic vials (Fisherbrand, Pittsburgh, PA, USA) and stored at 5⬚C until being mixed with the MALDI matrix, target spotting, and analysis. The total sample extraction time is approximately 5 min for 12 extractions (only limited by centrifuge capacity). www.clinical.proteomics-journal.com
Proteomics Clin. Appl. 2015, 9, 522–530
Figure 1. Illustration of the procedure employing C-CP fiber packed -SPE tips for the extraction of urinary proteins from urine matrices prior to analysis by MALDI-MS.
2.5 MALDI-MS A Bruker Daltonics (Billerica, MA, USA) Microflex LRF, MALDI-TOF-MS was employed in positive ion, linear mode for the detection of extracted species. Bruker Compass software was employed for instrumentation control and data processing. Further data analysis was performed with Microsoft Excel (Redmond, WA, USA). The MALDI-MS employed a nitrogen (337 nm) laser with a pulse rate of 60 Hz for the sample ionization. Sample spectra were acquired with 200 laser shots at 60% power. Sinapinic acid (20 mg/mL) in 50:50 ACN:H2 O with 0.1% TFA was utilized as the MALDI matrix, where 1 L of matrix was spotted onto the MALDI target, allowed to dry, and then a 1 L aliquot of eluted sample was spotted on the matrix. Reported S/N were computed as the SD of the background signal across a 100 Da window, centered 2000 Da below the pseudo-molecular ion (i.e., (M+H)+ ) peaks of the target proteins.
3
Results and discussion
Presented here, C-CP fibers are investigated as a stationary phase for micro-SPE tips for the extraction of urinary proteins from urine matrices. Urinary proteomics is rapidly growing in terms of clinical studies for potential biomarker analysis. 2-microglobulin, RBP, and Tf were investigated as the “target” analytes due to their common presence in biofluids as well as being commonly cited as potential biomarkers [21, 46, 47]. These three proteins exhibit various physical characteristics particularly their molecular weight (MW), three-dimensional structure, and amino acid compositions. 2-microglobulin being the smallest of the three has a MW of 11.6 kDa and is 45 × 25 × 20 A⬚ in size containing 180 amino acids [48]. RBP is slightly larger (20 kDa) containing 182 amino acids. While similar in size, the structure of RBP C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
525 is more complex as it is composed of an N-terminal, -sheet, ␣-helix, and a C-terminal core compared to 2-microglobulin having only 2  structures [49]. Tf on the other hand is much larger (80 kDa) and 45 × 57 × 135 A⬚ in size and containing 678 amino acids [50]. These different physical characteristics could provide insight to the hydrophobic interaction with the PP C-CP fiber surface. When discussing biomarker employment, 2-microglobulin levels in cerebrospinal fluid have been studied extensively as human African trypanosomiasis (HAT) biomarkers as increased levels detected after Western blot and ELISA testing are positively correlated with HAT [21]. 2-microglobulin has also been suggested as a biomarker for acute rejection in renal transplantation [51]. Studies have noted RBP as a potential biomarker as the increased levels of RBP reflect its reduced uptake, reflective of proximal tubular dysfunction [46]. Recent studies have also suggested RBP levels in plasma could be employed for the early detection of coronary heart disease [47]. While Tf has been studied extensively for renal failure, recently, a correlation between increased levels of excreted Tf could be related to drug nephrotoxicity [52]. With much effort being devoted to the monitoring of these proteins, urine matrices spikes with this suite of proteins would serve as a useful demonstration of the utility of this technique. Based on literature reports, the certified human urine and representative mock urine matrices were spiked at concentrations of 7.1, 7.9, and 10.4 M of the respective target proteins (2-microglobulin, RBP, and Tf) and used as the primary test samples. The resultant MALDI-MS spectra obtained from these samples without any form of sample clean-up (1 L spotted onto 1 L of matrix) are shown in Fig. 2A and B, respectively for the human and mock urine samples. It is known that target proteins are virtually nondetectable by MALDI-MS in the presence of 100 mM [53] salts and organics, even more so for biofluids. As can be seen, there are no discernible spectra features that can be attributed to the target proteins in the case of the human urine, though a slight signal is seen that can be assigned to the 2-microglobulin in the mock urine solution. Indeed, the intensity of the spectral continuum above 15 000 Da is greater for the human urine than the mock urine. Based on the system data processing package, the background signal for the human urine has a peak-to-peak deviation of 40 counts versus to the mock urine deviation of 7 counts. This is not surprising the former is composed of a wide variety of small molecules/salts not included in the mock urine matrix. It is for this situation that the C-CP fiber SPE methodology is being developed.
3.1 Role of elution solvent composition on protein responses A key aspect to the simplicity of SPE techniques is that they utilize an isocratic approach in terms of the elution solvent composition, that is, a single solvent composition is employed to simultaneously remove the analytes from the www.clinical.proteomics-journal.com
526
B. T. Manard et al.
Proteomics Clin. Appl. 2015, 9, 522–530
Figure 3. Effect of elution solvent composition on the intensity of (M+H)+ signal levels for 2-microglobulin and RBP in human urine and mock urine matrices. Elution solvent volume = 10 L, aqueous wash volume = 100 L.
Figure 2. Benchmark MALDI mass spectra for the (A) certified human urine and (B) mock urine matrices spiked to levels of 7.1, 7.9, and 10.4 M of the respective target proteins 2-microglobulin, RBP, and Tf.
stationary phase surface. Various modes of chromatography (i.e., ion exchange, RP) and stationary phases (i.e., polymer, C18 ) have been employed for SPE of various compounds [29, 54]. Polymer phases are attractive due to their chemical robustness and immunity to harsh elution solvents [54, 55]. The PP C-CP fibers are employed here for RP separation of proteins [40], meaning that organic-based elution solvents (or organic-aqueous mixtures) are employed for the removal of adsorbed species. Due to the different degrees of hydrophobicity among proteins, a global optimum in the elution solvent strength is desired to affect the most uniform and efficient recovery. In this case, 10 L of the protein suite (in synthetic and natural urine matrices) was loaded onto the C-CP fiber packed -SPE tips. A single 100 L aqueous wash was then performed to remove the unwanted species (e.g., salt), as this was a reasonable volume based on previous works. Finally, the proteins were eluted with 10 L solvent aliquots ranging from 50 to 75% ACN in water with 0.07% TFA, representing C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
compositions which are common for RP separations of proteins from C-CP fibers and also compatible with MALDIMS. The study was performed by monitoring MALDI-MS (M+H)+ intensities of the test proteins for deposited spots for each of solvent compositions. As depicted in Fig. 3, there is an optimum range of solvent compositions that yield optimum MALDI-MS responses for the three proteins, which is not dependent on whether the mock of human urine matrices are applied. The highest signal recoveries were found when employing an elution solvent of 55:45 ACN:H2 O with 0.07% TFA. This result varies slightly from previous reports utilizing PP C-CP tips (60:40 ACN:H2 O), as different proteins were utilized [45]. This response is consistent with what is seen for other RP SPE media used in protein extractions, reflecting a balance in driving forces between the affinity of the proteins for the hydrophobic surface and their solubility in solvents of lower and higher organic (ACN) content. At low ACN compositions, the proteins favor the hydrophobic PP surface, and at higher values the solvent is too nonpolar for the proteins to be solubilized. It is noteworthy that the responses for each protein are greater for the mock urine matrices than the certified human urine. This is not surprising as the latter is a more complex matrix. At the same time, the background continuum for the human urine was 5× higher than the mock urine (consistent with Fig. 2A and B), implying the extraction procedure should be optimized in terms of buffer removal in the aqueous wash step. It is also interesting that the relative signal responses as a function of solvent composition for the two proteins track each other within each matrix. As a reference, identical concentrations of the three proteins prepared in milliQ-H2 O (as a “clean” sample) were also studied to monitor analyte recoveries. Proteins recovered from water and mock urine gave nearly identical spectra and yields as a function of solvent composition. www.clinical.proteomics-journal.com
Proteomics Clin. Appl. 2015, 9, 522–530
527
Figure 4. Effect of aqueous wash volume on the intensity of (M+H)+ signal levels for 2-microglobulin and RBP in human urine and mock urine matrices. Elution solvent (55:45 ACN:H2 O with 0.07% TFA) volume = 10 L.
3.2 Role of aqueous wash volume on protein responses Previous studies in the use of the C-CP fiber -SPE approach prior to MALDI-MS of proteins have illustrated the key characteristic that, while salts/buffer components have a low affinity for the PP surface, high concentrations require that more thorough aqueous wash steps must be performed. The protein molecular ion responses and continuum background differences discussed relative to Fig. 3 point clearly the fact that the human urine matrix poses a greater challenge in terms of contaminant species removal. In low concentrations, buffers can be readily removed with minute aqueous wash volumes (5 L). However, salt/buffer concentrations of up to 1 M or viscous biofluids such as saliva, require larger volumes [45]. Analogous to the studies of elution solvent strength, the respective (M+H)+ responses for 2-microglobulin and RBP in the two urine matrices are plotted in Fig. 4 as a function of the volume of aqueous wash employed. In each case, the applied sample volume was held at 10 L, and the optimum elution solvent composition identified above (55:45 ACN:H2 O with 0.07% TFA) was employed. Similar to the previous studies, the responses depicted in Fig. 4 suggest that matrix removal is most efficient when the aqueous wash is 10× the sample load volume for these matrix-heavy solutions. Not surprising, the signals for the proteins in the mock urine can be more effectively recovered at lower wash volumes (75–100 L) in comparison to the volume required for the more complex matrix of human urine (100 L). Indeed, those ion yields show a very steep increase with wash volume increases. Simply, the lower wash volumes do not effectively remove interfering salt species as the MALDI-MS analyte signals suffer from ion suppression. Based on the results depicted in Fig. 4, 100 L of water was employed to remove the urine matrix in the following experiments. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 5. MALDI mass spectra of mock urine specimen spiked to levels of 7.1, 7.9, and 10.4 M of the respective target proteins (2-microglobulin, RBP, and Tf), following PP C-CP fiber tip SPE. (A) full mass spectrum, (B) spectral background realized for the case of human urine before and after C-CP fiber clean-up. Elution solvent (55:45 ACN:H2 O with 0.07% TFA) volume = 10 L, aqueous wash volume = 100 L.
3.3 Analytical characteristics following C-CP fiber -SPE extraction Following the evaluations of the elution and aqueous wash steps, the composite analytical method includes initial loading of 10 L aliquot of the test solution onto a PP C-CP -SPE tip, followed by a 100 L aqueous wash, then elution with 10 L of 55:45 ACN:H2 O with 0.07% TFA. The application and elution steps are allowed to spin down in the microcentrifuge for 30 s and the wash for 1 min. Figure 5A presents the product mass spectrum for the analysis of the mock urine specimen, spiked with the three target proteins. The spectrum is directly comparable to that of the raw sample presented in Fig. 2B. The (M+H)+ ions for each of the three spiked proteins, 2-microglobulin (11 716 Da), RBP (20 581 Da), and Tf (81 791 Da) are all seen with high spectral clarity. Additional to the pseudo-molecular ions, the doubly charged ((M+2H)2+ ) species are also seen for Tf (40 078 Da) and RBP (10 253 Da). The lower response for RBP could be www.clinical.proteomics-journal.com
528
B. T. Manard et al.
Proteomics Clin. Appl. 2015, 9, 522–530
Figure 6. MALDI mass spectrum of a certified human urine specimen spiked to levels of 1.14, 1.26, and 1.66 M of the respective target proteins (2-microglobulin, RBP, and Tf), following PP C-CP fiber tip -SPE. Elution solvent (55:45 ACN:H2 O with 0.07% TFA) volume = 10 L, aqueous wash volume = 100 L.
explained by lack of purity (as noted in the supplier documentation) while Tf is approaching the optimized instrument mass limit of 100 000 Da. While the vast improvement in the target protein responses is dramatic, it is important to note the improvement as well in the continuum spectral background. This is demonstrated for the worst-case situation of the certified human urine matrix in Fig. 5B, where both the average value and the SD are much improved in the mass range of 60–70 000 Da. Specifically, the background for the preextraction case is 16.1 ± 3.9 counts, while following the extraction it is 7.1 ± 2.8. With the optimized procedure, LODs were assessed to determine whether or not the PP C-CP fiber -SPE tips would indeed be relevant for extraction of low mass quantities in human urine. The single-point LODs were determined by use of the “SBR-RSD” method originally described by Boumans and Vrakking [56]. In this approach, the LOD is defined as: LOD =
(0.01) k (RSD) m (S/B)
(1)
where RSD is the RSD of the adjacent spectral background, m is the concentration of the test element in the specimen, S/B is the SBR, and the k is a statistical factor equal to 3, which reflects a 99% confidence interval. In order for this method to be more accurate, the analyte concentrations in the test specimen were reduced to 1.14, 1.26, and 1.66 M for 2-microglobulin, RBP, and Tf, respectively, correlating to 0.13, 0.24, and 1.33 g total mass of each applied to each micropipette tip. The resultant mass spectrum is depicted in Fig. 6. Detection limits were determined to be 5.3, 30.3, and 184 nM, respectively. When referring to the 1 L extract spots placed on the MALDI targets (which are not totally ablated), these molar values equate to mass-based LODs of 60 pg for 2-microglobulin, 600 pg for RBP, and 15 ng for Tf. The discrepancies in LODs are readily explained in terms of the chemical nature of each, as well as the aforementioned C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
decrease in ion throughput/detector response as a function of mass for this particular instrument. (Additionally, it cannot be ignored that this particular instrument is that manufacturer’s lowest performance model.) All of these points made, it is quite clear that the performance demonstrated here is applicable to the field of urinomics, particularly for biomarker screening. Important biomarker concentrations are 0.05 mg/mL, corresponding to 500 ng per 10 L of volume, as processed here. This level is demonstrated for 2-microglobulin and RBP that were on the picogram level. It is very difficult to compare absolute performance among extraction methods used for urinary proteomics as the analytical platforms (e.g., ionization methods and mass analyzers) vary greatly. When compared to the aforementioned new SPE methods for biomarker analysis, the use of mesoporous silica beads [11] and magnetic nanoparticles [27] deal specifically with the urine matrix. Both methods require sample aliquots in the same regime (
