J Electrophoresis
2009 ; 53 : 13
[Technical Note]
A simple and highly reproducible method for discovering potential disease markers in low abundance serum proteins Yusuke Kawashima1, Tomoyuki Fukuno1, Mamoru Satoh2, Hiroki Takahashi1, Takashi Matsui1, Tadakazu Maeda1, 3 and Yoshio Kodera1–3 SUMMARY
Serum provides a link between many human organs, tissues, and cells, and is one of the most informative body fluids. However, the presence of 22 abundant proteins and a large dynamic range of numerous other proteins make quantitative analysis of low abundance proteins challenging. Here, we describe simple and easy to use pretreatment techniques for serum combined with high abundant protein removal and reverse-phase high-performance liquid chromatography (RP-HPLC) separation, each step of which was optimized to minimize the loss of proteins and increase reproducibility. This method can be used to discover disease-specific biomarker proteins in concentrations of the ng/mL range, and furthermore, be used as a base strategy to comparatively analyze the deep proteome in the pg/mL range. Key words: biomarker protein, proteomics, serum, reverse-phase chromatography. INTRODUCTION Serum is a vast medium containing thousands of different types of proteins and peptides1–5). This body fluid contains information pertaining to the many processes taking place within the body. Quantitative analysis of proteins in serum can provide information as to how the body reacts to changes, such as disease. For example, increased serum levels of prostate-specific antigen6) and CA1257) are routinely used for the detection of cancer in the prostate and ovary, respectively. Determination of disease-specific targets found in serum could aid in the early detection of disease and in drug discovery for new medicinal treatments. However, serum proteins are present in an extremely wide range of concentrations that are likely to extend by more than 10 orders of magnitude. Just 22 major protein species comprise approximately 99% of total serum proteins4, 5), which interferes with the identification of low 1
abundance protein biomarkers. Therefore, the removal of high abundance proteins and the reduction in complexity of serum samples is essential prior to any analysis aimed at determining proteins present in small quantities, which potentially include protein biomarkers. Several strategies for serum proteomics have been reported2, 8–19). Representative and practical strategies use a high abundance protein removal method followed by 2-DE or shotgun proteomics2, 8–10). To analyze the lower abundance proteins, pretreatment is combined with a high abundance protein removal method and separation by sizeexclusion and/or ion-exchange chromatography followed by two-dimensional gel electrophoresis (2-DE) or shotgun proteomics11–13). Because the latter needs relatively high quality technology to discover disease related proteins through comparative analyses of more than ten samples, the former has been mainly used. However, detecting and identifying low abundance proteins, including tissue leak-
Laboratory of Biomolecular Dynamics, Department of Physics, Kitasato University School of Science. Clinical Proteomics Research Center, Chiba University Hospital. 3 Center for Disease Proteomics, Kitasato University School of Science. Correspondence address: Yoshio Kodera; Laboratory of Biomolecular Dynamics, Department of Physics, Kitasato University School of Science, 1-15-1 Kitasato, Sagamihara-shi, Kanagawa 228-8555, Japan. E-mail:
[email protected] Fax: +81-42-778-9541 Abbreviations: RP-HPLC, reverse-phase high-performance liquid chromatography; 2-DE, two-dimensional gel electrophoresis; TFA, trifluoracetic acid; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; CBB, Coomassie brilliant blue; IEF, isoelectric focusing; TCA, trichloroacetic acid. (Received January 7, 2009, Accepted February 2, 2009, Published March 15, 2009) 2
J Electrophoresis
2009 ; 53 : 14
age proteins less than 1 µg/mL in concentration, is difficult by the former method. Hence, it is necessary to develop a simple and widely usable pretreatment strategy to comparatively analyze low abundance proteins. Here, we describe simple and easy to use pretreatment techniques combined with an albumin affinity removal method and fractionation by reverse-phase highperformance liquid chromatography (RP-HPLC). Optimizing and minimizing protein loss at each step makes it possible to achieve a sensitive and reproducible analytical method. This method can be used to discover disease-specific biomarker proteins in concentrations of the ng/mL range, and furthermore, be used as a base strategy to comparatively analyze the deep proteome in the pg/mL range. MATERIALS AND METHODS Rat serum samples Blood samples were collected from Wistar rats. The samples were allowed to clot at room temperature for 2 h, and then centrifuged at 1,000×g for 20 min at room temperature. The serum (supernatant) samples were stored in aliquots at –80°C. Removal of albumin from serum samples The albumin affinity removal was performed using a ProteoExtract® Albumin Removal kit (Merck, Darmstadt, Germany) according to the manufacturer’s instructions. A 40-µL sample of serum was diluted 1:9 with 360 µL binding buffer, and was allowed to pass the column by gravity flow. The flow-through fraction was collected in a collection tube. To wash the column, 600 µL binding buffer was added to the column, and was allowed to pass the column by gravity flow. The flow-through fraction was collected in the same collection tube. Buffer-exchange/ concentration method by ultrafiltration The albumin depleted sample was buffer-exchanged and concentrated using 10-kDa molecular weight cut-off ultrafiltration (Sartorius, Gottingen, Germany). The sample was centrifuged at 6,000×g at 4°C until less than 50 µL and the buffer exchanged to 0.1% TFA by addition of 1 mL 0.1% trifluoracetic acid (TFA), with centrifugation at 6,000×g at 4°C until concentrated to less than 80 µL. The concentrated sample was adjusted to a final volume of 80 µL with 0.1% TFA. Acetone precipitation method Acetone, pre-cooled to –80°C, was added gradually (with intermittent vortexing) to the depleted sample to a final concentration of 80% (v/v), and incubated for 1 h at –80°C, then centrifuged at 19,000×g for 15 min at 4°C. The precipitate was taken up in 80 µL of 0.1% TFA and mixed at 4°C for 1 h, then centrifuged again at 19,000×g for 15 min at
4°C. Subsequently, the supernatant was collected. RP-HPLC An RP column (3.0 ID×150 mm, Intrada WP-RP; Imtakt Corp., Kyoto, Japan) was attached to an HPLC system (Nanospace SI-2; Shiseido Fine Chemicals, Tokyo, Japan). The column temperature was maintained at room temperature, about 25°C, and the flow rate of the mobile phase was 400 µL/min. The composition of the mobile phase was programmed to change in 65 min cycles with varying the mixing ratios r=[B]/([A]+[B])×100 of solvent A (0.1% TFA) and solvent B (90% acetonitrile in 0.08% TFA) as follows: constant mixing ratio (r=5%) from time t=0 to 5 min, linear gradient (r=5 to 51%) from t=5 to 28 min, linear gradient (r=51 to 95%) from t=28 to 29 min, constant (r=95%) from t=29 to 39 min, downward linear gradient (r=95 to 5%) from t=39 to 40 min and constant (r=5%) from t=40 to 65 min. Fractions were collected every 30 s from t=20.5 to 33 min, resulting in a total of 25 fractions per sample. The HPLC fractions were then lyophilized. SDS-PAGE analysis The lyophilized samples of HPLC fractions were dissolved in PAGE sample buffer (50 mM Tris-HCl, pH 6.8 containing 50 mM dithiothreitol, 0.5% SDS and 10% glycerol). The solution was then analyzed using SDS-PAGE (Perfect NT Gel W, 10–20% acrylamide, 28 wells; DRC Co., Ltd., Tokyo, Japan) according to the manufacturer’s protocol. The gel was stained with silver nitrate (Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan) or Coomassie brilliant blue (CBB, PhastGel Blue R; GE Healthcare, Little Chalfont, UK). The gel image was converted to a densitogram by Scion Image. Agarose 2-DE analysis The lyophilized samples of HPLC fractions were dissolved in 7 M urea, 2 M thiourea, 0.1 M dithiothreitol, 2.5% Pharmalyte (pH 3–10), 2% CHAPS, and Complete Mini EDTA-free (protease inhibitors; Roche Diagnostics, Mannheim, Germany). Agarose gels were prepared as described20). The agarose gel for the first dimension isoelectric focusing (IEF) was 70 mm in length and 2 mm in diameter in a glass tube. The 2-D polyacrylamide gels were 10–20% gradient slab gels, and were 140×80×1.0 mm (DRC Co., Ltd., Tokyo, Japan). The dissolved sample was applied to the agarose IEF gel, and first-dimensional IEF was conducted at 6,000 Vh at 4°C, followed by fixation in 10% trichloroacetic acid (TCA) and 5% sulfosalicylic acid for 45 min at room temperature. After washing with deionized water for 45 min, two agarose gels were then placed on top of a polyacrylamide gel side-by-side, and 2-D SDSPAGE was performed. The 2-DE gel was first incubated in 30% methanol and 10% acetic acid, and then stained with CBB.
J Electrophoresis
RESULTS AND DISCUSSION A flowchart of the developed method is depicted in Fig. 1. Highly abundant albumin was depleted from 40 µL serum using an albumin affinity removal column. Subsequently, the depleted sample was buffer-exchanged/concentrated using 10 kDa molecular weight cut-off ultrafiltration followed by separation into 25 fractions using RP-HPLC. The HPLC fractions were then lyophilized. The lyophilized samples were dissolved and analyzed using SDS-PAGE or 2-DE. Profiles of proteins in each fraction, corresponding to proteins from 5 µL serum, were analyzed by silver-stained SDS-PAGE (Fig. 2). Between fractions 1 and 17, proteins were efficiently separated. Approximately 500 bands were observed, most of which were detected in one or two fractions. In the latter fractions, 18–25, most of the proteins
Fig. 1.
were not well separated and observed in several fractions, and the elution profile of each protein seemed to broaden, perhaps because of high hydrophobicity. A total of about 400 proteins were observed. To discover biomarker proteins, it is essential that each treatment in the protocol be reproducible, though this is not easy, especially in very complex serum samples. We optimized each step of the protocol. The results are shown in Fig. 3. Four aliquots of serum were treated through the protocol shown in Fig. 1. The samples fractionated by RPHPLC were analyzed by silver-stained SDS-PAGE. In all fractions, the protein contents and abundances were exactly the same for all four aliquots. This result indicates that the reproducibility of each process is extremely high. To discover biomarker proteins with high sensitivity and comprehensibility, it is important to minimize the loss of proteins in each step of protocol. First we assessed albumin removal (Fig. 4A, lanes 2 and 3) and buffer-exchange/concentration by ultrafiltration (Fig. 4A, lanes 4 and 5). This result clearly shows that protein loss by these two steps is very low. We compared the treatment by ultrafiltration with the acetone precipitation treatment (Fig. 4A, lanes 4–7). The recovery rate from the ultrafiltration was much higher than that of acetone precipitation. The loss of proteins from each process was estimated by image analysis (Fig. 4B). The losses of the proteins by albumin removal were very small except for the protein as depicted by an arrow a and that of buffer-exchange/concentration was about 8%, mainly caused by a band indicated by an arrow b (Fig. 4A). Typically, acetone precipitation, TCA precipitation, and ultrafiltration are used as buffer-exchange/concentration methods. Since acetone precipitation denatures proteins
Flowchart for quantitative analysis of low abundance proteins in serum.
Fig. 3.
Fig. 2.
2009 ; 53 : 15
Silver-stained SDS-PAGE pattern of albumin removal serum proteins fractionated by RP-HPLC. Samples corresponding to those derived from 5 µL serum were loaded in each lane.
High reproducibility of the established method shown by silver-stained SDS-PAGE. Four aliquots of 40 µL from the same serum sample were treated with the albumin removal kit, solution-exchange/ concentration, and RP-HPLC fractionation independently. One-sixteenth of each sample, corresponding to proteins from 2.5 µL serum, was loaded in each lane.
J Electrophoresis
2009 ; 53 : 16
Fig. 4. Assessment of protein loss from albumin removal and solution-exchange/concentration. (A) The same serum was treated twice independently, then analyzed by CBB-stained SDS-PAGE. Lane 1: crude serum; lanes 2 and 3: depleted serum using an albumin removal column; lanes 4 and 5: solutionexchanged/concentrated (in 0.1% TFA) samples of lanes 2 and 3 by ultrafiltration, respectively; lanes 6 and 7: supernatant of acetone precipitate dissolved in 0.1% TFA; lanes 8 and 9: supernatant of precipitate of lanes 6 and 7 dissolved in PAGE sample buffer. An equivalent of 0.25 µL serum was loaded in each lane. (B) Densitometric analysis of gel shown in (A). Lanes 1–9 in (B) correspond to lanes 1–9 in (A), respectively.
Fig. 5.
Assessment of protein loss from RP-HPLC fractionation. (A) Two aliquots of 40 µL from the same serum sample were treated with the albumin removal kit, solution-exchange/ concentration by ultrafiltration, and RP-HPLC fractionation independently. Samples corresponding to proteins from 0.25 µL serum were loaded in each lane. Lanes 1 and 2: solution-exchanged/concentrated serum; lanes 3 and 4: mixtures of all RP-HPLC fractionations from samples of lanes 1 and 2, respectively. (B) Densitometric analysis of gel shown in (B). Lanes 1–4 in (B) correspond to lanes 1–4 in (A), respectively.
and precipitates most of proteins (Fig. 4A, lanes 8 and 9), redissolving is difficult and most proteins may be lost (Fig. 4, lanes 6 and 7). Therefore, we used ultrafiltrations that can concentrate without precipitating the protein. Furthermore, we assessed the loss by RP-HPLC (Fig. 5). Proteins in all fractions were gathered and samples corresponding to 0.25 µL serum were analyzed by CBB-stained SDS-PAGE (Fig. 5, lanes 3 and 4). Protein loss by RPHPLC separation was minimal. We compared the method established here (Fig. 6B) with
a typical serum proteomic method, which is a combination of albumin affinity removal and 2-DE (Fig. 6A). When we analyzed proteins by 2-DE with good separation, the amount of applied proteins was limited. In Fig. 6A, proteins from 5 µL serum treated by an albumin removal column were analyzed. Hence in Fig. 6B, as proteins in the albumin depleted sample were separated by RP-HPLC, proteins from 40 µL serum were applied to 2-DE. Many protein spots that were not detected in Fig. 6A were detected in the gel that used the method established in this study (Fig. 6B). When the typical method was used, approximately 100 spots were observed, while in the developed method, when all fractions were analyzed, approximately 1000 spots were observed. The most important analytical aspects for discovering potential biomarkers are sensitivity, reproducibility, and comprehensibility. The newly established method was extremely reproducible. This excellent feature was demonstrated by the fact that low abundance proteins only detected by silver staining were commonly observed with the same intensities for all four samples in each fraction as shown in Fig. 3. This appears to be due primarily to the high reproducibility and stability of the RP-HPLC separation step. The selection of the column, the flow rate of solvent, and stability of the HPLC system are very important. Before sample separation, a standard peptide mixture was analyzed five times and agreement of retention times was confirmed. Additionally, we examined the protein quantity limit that could be analyzed by RP-HPLC with high reproducibility, and protein quantities less than the limit were separated. We confirmed highly reproducible separations for more than 16 samples. The 16 serum samples were treated with an albumin removal column for two hours,
J Electrophoresis
Fig. 6.
CBB-stained agarose 2-DE patterns of proteins in serum. (A) The albumin depleted serum using an albumin removal column. (B) Mixture of fractions 12 and 13 fractionated by RP-HPLC from a sample of (A) as shown in Fig. 2. Equivalents of 5 µL and 40 µL serum were loaded in (A) and (B), respectively.
solvent-exchanged/concentrated for five hours, and fractionated by RP-HPLC for 40 hours. Then the fractions were lyophilized and stored in a –80°C deep freezer, which maintains their quality for a long time without any changes. Therefore, it is possible to analyze them by SDS-PAGE, 2DE, or to conduct more detailed separations at any time. This feature is also important for analyzing proteins with high reproducibility. Hence, our method is a simple and relatively high throughput method which is adequate for discovering low abundance biomarker proteins. As for sensitivity and comprehensibility, our method is superior in these regards, as protein loss was very low in each step of the protocol both qualitatively and quantitatively, especially during solvent-exchange/concentration, and in RP-HPLC fractions (Figs. 4 and 5). Hence, the concentration detection limits of 40-µL serum samples using CBB-stained 2-DE (Fig. 6B) and silver-stain 2-DE are 500 ng/mL and 5 ng/mL, respectively; that is, the detec-
2009 ; 53 : 17
tion limits for CBB and silver staining are about 20 ng and 0.2 ng, respectively. This is sensitive enough to analyze tissue leakage proteins. For human serum, this protocol is possible to be applied and analyze with the same detection limits as mentioned above. Furthermore, by removing 12 high abundance proteins with a removal column, 250 µL human serum sample can be separated by RP-HPLC at once (data not shown). Thus, we can comparatively analyze serum proteins at concentrations of less than 100 ng/mL by CBB staining and less than 1 ng/mL by silver staining. We established a multi-dimensional separation technique based on a combination of affinity removal, RP-HPLC, and 2-DE to discover novel biomarkers in serum. This method has extremely high reproducibility and low protein loss. The method enables us to comparatively analyze low abundance proteins. Approximately 1,000 protein spots were observed by CBB staining of mini 2-DE with relatively high throughput from 40 µL serum samples. The highest level of gel-based serum proteome analysis to date was reported by Okano et al.12), who used a combination method of major protein removal and fractionation by a cation exchange column, making it possible to detect over 3500 protein spots using highly sensitive CyDye DIGE Fluor saturation dyes (GE Healthcare). In comparison, our method can detect about 1,000 protein spots using mini agarose 2-DE (7×8 cm) with CBB staining, which is approximately 1,000fold less sensitive than the saturation dyes detection. This suggests that the combination of the saturation dyes with a conventional size 2-DE gel (26×20 cm) may be able to detect extremely low abundance proteins on the order of pg/mL, as the total protein loss is very low and the 25 protein fractions have high resolution using RP-HPLC. Above all, this simple method established here is useful for searching for biomarker proteins in the range of ng/mL in serum. Furthermore, it can be used as a base method to discover the lower abundance biomarker proteins, if combined with other types of separation and detection methods or a shotgun method based on each RP-HPLC fraction. ACKNOWLEDGMENT This study was supported in part by the Ministry of Education, Culture, Sports, Science, and Technology of Japan. REFERENCES 1) Hortin GL, Sviridov D, Anderson NL. High-abundance polypeptides of the human plasma proteome comprising the top 4 logs of polypeptide abundance. Clin Chem 2008;54:1608–1616. 2) Liu X, Valentine SJ, Plasencia MD, Trimpin S, Naylor S, Clemmer DE. Mapping the human plasma proteome by SCX-LC-IMS-MS. J Am Soc Mass Spectrom 2007;18: 1249–1264. 3) Tanaka Y, Akiyama H, Kuroda T, Jung G, Tanahashi K, Sugaya H, Utsumi J, Kawasaki H, Hirano H. A novel
J Electrophoresis
4)
5)
6)
7)
8)
9)
10)
11)
12)
2009 ; 53 : 18
approach and protocol for discovering extremely lowabundance proteins in serum. Proteomics 2006;6:4845– 4855. Tirumalai RS, Chan KC, Prieto DA, Issaq HJ, Conrads TP, Veenstra TD. Characterization of the low molecular weight human serum proteome. Mol Cell Proteomics 2003;2:1096–1103. Anderson NL, Anderson NG. The human plasma proteome: History, character, and diagnostic prospects. Mol Cell Proteomics 2002;1:845–867. Grossklaus DJ, Smith JA,Jr, Shappell SB, Coffey CS, Chang SS, Cookson MS. The free/total prostate-specific antigen ratio (%fPSA) is the best predictor of tumor involvement in the radical prostatectomy specimen among men with an elevated PSA. Urol Oncol 2002; 7:195–198. Whitehouse C, Solomon E. Current status of the molecular characterization of the ovarian cancer antigen CA125 and implications for its use in clinical screening. Gynecol Oncol 2003;88:S152–157. Liu W, Liu B, Xin L, Zhang Y, Chen X, Zhu Z, Lin Y. Down-regulated expression of complement factor I: A potential suppressive protein for gastric cancer identified by serum proteome analysis. Clin Chim Acta 2007;377: 119–126. Kim HJ, Cho EH, Yoo JH, Kim PK, Shin JS, Kim MR, Kim CW. Proteome analysis of serum from type 2 diabetics with nephropathy. J Proteome Res 2007;6:735–743. Fujii K, Nakano T, Kanazawa M, Akimoto S, Hirano T, Kato H, Nishimura T. Clinical-scale high-throughput human plasma proteome analysis: Lung adenocarcinoma. Proteomics 2005;5:1150–1159. Pieper R, Gatlin CL, Makusky AJ, Russo PS, Schatz CR, Miller SS, Su Q, McGrath AM, Estock MA, Parmar PP, Zhao M, Huang ST, Zhou J, Wang F, Esquer-Blasco R, Anderson NL, Taylor J, Steiner S. The human serum proteome: Display of nearly 3700 chromatographically separated protein spots on two-dimensional electrophoresis gels and identification of 325 distinct proteins. Proteomics 2003;3:1345–1364. Okano T, Kondo T, Kakisaka T, Fujii K, Yamada M, Kato H, Nishimura T, Gemma A, Kudoh S, Hirohashi S. Plasma proteomics of lung cancer by a linkage of multi-
13)
14)
15)
16)
17)
18)
19)
20)
dimensional liquid chromatography and two-dimensional difference gel electrophoresis. Proteomics 2006;6:3938– 3948. Kakisaka T, Kondo T, Okano T, Fujii K, Honda K, Endo M, Tsuchida A, Aoki T, Itoi T, Moriyasu F, Yamada T, Kato H, Nishimura T, Todo S, Hirohashi S. Plasma proteomics of pancreatic cancer patients by multidimensional liquid chromatography and two-dimensional difference gel electrophoresis (2D-DIGE): Up-regulation of leucine-rich alpha-2-glycoprotein in pancreatic cancer. J Chromatogr B Analyt Technol Biomed Life Sci 2007; 852:257–267. Qiu Y, Patwa TH, Xu L, Shedden K, Misek DE, Tuck M, Jin G, Ruffin MT, Turgeon DK, Synal S, Bresalier R, Marcon N, Brenner DE, Lubman DM. Plasma glycoprotein profiling for colorectal cancer biomarker identification by lectin glycoarray and lectin blot. J Proteome Res 2008; 7:1693–1703. Jin WH, Dai J, Li SJ, Xia QC, Zou HF, Zeng R. Human plasma proteome analysis by multidimensional chromatography prefractionation and linear ion trap mass spectrometry identification. J Proteome Res 2005;4:613–619. Martosella J, Zolotarjova N, Liu H, Nicol G, Boyes BE. Reversed-phase high-performance liquid chromatographic prefractionation of immunodepleted human serum proteins to enhance mass spectrometry identification of lower-abundant proteins. J Proteome Res 2005;4:1522– 1537. Li RX, Chen HB, Tu K, Zhao SL, Zhou H, Li SJ, Dai J, Li QR, Nie S, Li YX, Jia WP, Zeng R, Wu JR. Localized-statistical quantification of human serum proteome associated with type 2 diabetes. PLoS ONE 2008;3:e3224. Petricoin EF, Belluco C, Araujo RP, Liotta LA. The blood peptidome: A higher dimension of information content for cancer biomarker discovery. Nat Rev Cancer 2006;6:961– 967. Hanash SM, Pitteri SJ, Faca VM. Mining the plasma proteome for cancer biomarkers. Nature 2008;452:571– 579. Oh-Ishi M, Satoh M, Maeda T. Preparative twodimensional gel electrophoresis with agarose gels in the first dimension for high molecular mass proteins. Electrophoresis 2000;21:1653–1669.
