`Quantitative validation of different protein
`precipitation methods in proteome analysis of
`blood platelets
`
`2481
`
`For the preparation of proteins for proteome analysis, precipitation is frequently used
`to concentrate proteins and to remove interfering compounds. Various methods for
`protein precipitation are applied, which rely on different chemical principles. This study
`compares the changes in the protein composition of human blood platelet extracts
`after precipitation with ethanol (EtOH) or trichloroacetic acid (TCA). Both methods
`yielded the same amount of proteins from the platelet preparations. However, the
`EtOH-precipitated samples had to be dialyzed because of the considerable salt con-
`tent. To characterize single platelet proteins, samples were analyzed by two-dimen-
`sional fluorescence differential gel electrophoresis. More than 90% of all the spots
`were equally present in the EtOH- and TCA-precipitated samples. However, both pre-
`cipitation methods showed a smaller correlation with nonprecipitated samples
`(EtOH 74.9%, TCA 79.2%). Several proteins were either reduced or relatively enriched
`in the precipitated samples. The proteins varied randomly in molecular weight and
`isoelectric point. This study shows that protein precipitation leads to specific changes
`in the protein composition of proteomics samples. This depends more on the specific
`structure of the protein than on the precipitating agent used in the experiment.
`
`Keywords: Ethanol / Platelets / Protein precipitation / Proteomics / Trichloroacetic acid
`DOI 10.1002/elps.200410262
`
`Maria Zellner1
`Wolfgang Winkler1
`Hubert Hayden1
`Michael Diestinger1
`Maja Eliasen1
`Bernd Gesslbauer2
`Ingrid Miller3
`Martina Chang4
`Andreas Kungl2
`Erich Roth1
`Rudolf Oehler1
`
`1Surgical Research Laboratories,
`Medical University of Vienna,
`Vienna, Austria
`2Institute of Pharmaceutical
`Sciences,
`University of Graz,
`Graz, Austria
`3Department of Natural Sciences,
`Institute of Medical Chemistry,
`University of Veterinary Medicine,
`Vienna, Austria
`4Institute of Applied Microbiology,
`University of Natural Resources
`and Applied Life Science,
`Vienna, Austria
`
`Proteomicsand2-DE
`
`filtered platelets are more quiescent, and are generally in
`better condition than platelets prepared by multiple
`steps of centrifugation [3]. However, it is impossible to
`concentrate the platelets using this technique. Therefore,
`the protein concentration is low and direct proteomic
`analysis of this platelet isolation is not possible. In addi-
`tion, cells contain proteases and high levels of non-
`protein impurities. An appropriate sample preparation is
`essential for obtaining reliable results in a proteomic
`analysis. Such a preparation should work by inhibiting
`protease activities and quantitatively enriching proteins,
`while leaving behind substances, such as salts, lipids,
`and nucleic acids, which would interfere with any further
`proteomic analysis. Protein precipitation followed by
`dissolving the pellet in IEF compatible sample solution is
`generally employed to concentrate and selectively
`separate proteins in the sample from the interfering sub-
`stances.
`In addition,
`the protein denaturation during
`precipitation leads to an inhibition of proteases [4].
`
`The precipitation of proteins from biological fluids had
`been observed for hundreds of years (e.g., the precipita-
`tion of casein from milk by dilute acid). However, the mo-
`lecular basis of protein solubility began to receive serious
`attention only in the middle of the last century. Fractiona-
`tion of human plasma by precipitation carried forward
`
`1 Introduction
`
`Proteome analysis of clinical samples aims at character-
`izing disease-specific changes in the protein expression
`profile of an affected tissue. Such changes are potential
`diagnostic markers or may help to identify drug targets.
`Blood platelets are small enucleated cellular particles
`that play a fundamental role in hemostasis, contributing
`to the formation of vascular plugs. Pathologically, they
`are involved in thrombosis and atherosclerosis. A
`detailed analysis of the proteome and signaling cas-
`cades in platelets from patients is expected to aid the
`development of new therapeutic agents that may help to
`treat thrombotic diseases [1, 2]. Platelets can be isolated
`from peripheral blood in a few preparative steps to a high
`cellular purity. To prevent activation of platelets,
`it is
`advisable to avoid strong centrifugation and to separate
`platelets from plasma protein by gel filtration. The gel-
`
`Correspondence: Professor Rudolf Oehler, Surgical Research
`Laboratories, General Hospital Vienna, Medical University of
`Vienna, Waehringer Guertel 18–20, A-1090 Vienna, Austria
`E-mail: rudolf.oehler@meduniwien.ac.at
`Fax: 143-1-40400-6782
`
`Abbreviations: DIGE, differential gel electrophoresis; EtOH,
`ethanol; GFP; gel-filtered platelet; PRP, platelet-rich plasma;
`TM3, tropomyosin a 3
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`Page 1 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`2482
`
`M. Zellner et al.
`
`Electrophoresis 2005, 26, 2481–2489
`
`during World War II made possible the preparation of
`many protein components of the human plasma [5]. The
`solubility of proteins is determined by four variables: pH,
`ionic strength, temperature, and protein concentration.
`Numerous different techniques have been developed for
`protein precipitation by modifying one or more of these
`parameters (for review see [6]). Several methods have
`been adapted to the needs of proteome analysis [4, 7, 8]
`and today various protein extraction kits are commercially
`available which apply precipitation techniques. The pres-
`ent study evaluates the applicability of two methods of
`protein precipitation for clinical proteomics: ethanol
`(EtOH) precipitation and TCA precipitation. Both methods
`are commonly used in the preparation of protein extracts
`for proteomic analysis, although they rely on different
`chemical principles. EtOH causes precipitation of pro-
`teins mainly because it significantly lowers the dielectric
`constant of the aqueous solution (relative dielectric con-
`stant at 207C for H2O is 18 and for C2H5OH is 26). In
`general, ionic compounds are more soluble in solvents
`with high dielectric constants. Through its polar groups,
`EtOH interacts with the polar group of the protein in
`competition with water.
`In addition, the hydrophobic
`groups may disrupt
`the intramolecular hydrophobic
`interaction. Finally, a large volume of EtOH reduces the
`effective concentration of water,
`leaving only a small
`amount for hydration of the protein. Upon dehydration by
`EtOH, protein molecules attract each other to a sufficient
`degree by van der Waals forces and thus become insol-
`uble in the EtOH-water mixtures [9]. TCA, in contrast,
`leads to a strong decrease in pH, resulting in denatura-
`tion and consequently precipitation of the protein. A
`recent study showed that the three chloro groups in the
`molecule also play an important role in protein precipita-
`tion, which is not clear [10]. TFA, which is a stronger acid
`than TCA and possesses three fluoro groups instead of
`chloro groups as in TCA, is not such a potent protein
`precipitation-inducing agent.
`
`The precipitation efficiency of both EtOH and TCA
`depends also on the physicochemical characteristics of
`the protein. Therefore, they are also applied for the frac-
`tionation of protein mixtures. Especially EtOH is broadly
`used for protein fractionation even at industrial scale.
`Proteomics comprises the analysis of thousands of dif-
`ferent protein species simultaneously. Applying EtOH or
`TCA precipitation in the analysis of such complex protein
`mixtures may result in depletion of particular protein spe-
`cies and in a relative enrichment of other protein species.
`The present study investigates qualitative and quantita-
`tive changes in the protein composition of platelet
`extracts after protein precipitation. It describes the pro-
`tein yield, the linearity, and the protein selectivity of EtOH
`and TCA precipitation. By comparing these character-
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`istics with nonprecipitated samples, the study evaluates
`the suitability of protein precipitation in the proteome
`analysis of clinical samples of this type.
`
`2 Materials and methods
`
`2.1 Blood sampling and platelet preparation
`
`Peripheral venous blood was drawn without stasis from
`60 healthy volunteers aged between 20 and 90 years
`(30 female/30 male; average age 53 6 21 years). The
`study was approved by the local ethics committee. Blood
`was drawn from the antecubital vein of the subjects into
`vacutainer tubes containing 0.129 mol/L trisodium citrate
`(Vacuette system; Greiner, Kremsmuenster, Austria). The
`first 3 mL of blood was discarded as usual for platelet
`studies. For exclusion of erythrocytes and leukocytes, the
`citrated whole blood was centrifuged at 50 3 g for 20 min
`at room temperature. The resulting supernatant was the
`platelet-rich plasma (PRP). A plasma-free platelet sus-
`pension was prepared by passing PRP through a size-
`exclusion chromatography (SEC) column. One milliliter
`PRP was applied onto 11 mL packed Sepharose 2B
`(Sigma, Steinheim, Germany) column (BioRad, Hercules,
`CA, USA; 15 mm diameter) equilibrated in calcium-free
`Dubecco’s PBS (GIBCO, Paisley, Scotland, UK). Platelet
`fractions (1.5 mL) of each individual were collected after
`an elution volume of 2.5 mL, and platelet concentration
`was counted on a MicroDiff 18 Blood Analyzer (Coulter
`Electronics, Miami, FL, USA).
`
`2.2 TCA precipitation
`
`Fifteen-hundred microliters of platelet suspension was
`ice-cold 6.1 N TCA solution
`mixed with 500 mL of
`(Sigma) containing 80 mM DTT (Roche Diagnostics,
`Mannheim, Germany). The mixture was incubated for
`1 h at 47C to allow the protein precipitation to complete.
`Then the extract was centrifuged at 10 000 3 g for
`10 min at 47C. The supernatant was discarded, and the
`pellet was washed four times with 1500 mL of ice-cold
`acetone (p.a. grade; Merck, Darmstadt, Germany) each,
`containing 20 mM DTT; the pellet was regained in each
`step by centrifugation at 10 000 3 g. Thereafter,
`the
`centrifuged pellet was dried by air evacuation. For 2-
`DE, the pellet was resolubilized in denaturing 2-D sam-
`ple buffer containing 7 M urea, 2 M thiourea, 4% CHAPS,
`30 mM Tris-HCl
`(pH 8.5) by shaking overnight at 47C.
`Seventy microliters of the sample buffer was used per
`100 6 106 platelets. Alternatively,
`the samples were
`stored in the last wash aliquot of acetone at 2707C until
`further use.
`
`Page 2 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Electrophoresis 2005, 26, 2481–2489
`
`Validation of different protein precipitation methods
`
`2483
`
`2.3 EtOH precipitation
`
`Nine-hundred microliters of platelet suspension was
`mixed with 8000 mL EtOH (99.9% Uvasol p.a; Merck).
`Proteins were allowed to precipitate in this EtOH solution
`for 3 h at room temperature followed by storage at 2207C
`overnight. To collect the proteins, the samples were cen-
`trifuged at 10 000 3 g for 20 min at 47C and the super-
`natant was removed. The pellet was washed once with
`pure EtOH before it was dried (as above). For 2-DE, the
`pellet was resolubilized in 100 mL of 2-D sample buffer as
`above per 100 6 106 platelets. Alternatively, the samples
`were stored in the 90% EtOH solution at 2207C until fur-
`ther use.
`
`2.4 Dialysis
`
`To reduce the salt contamination of 2-D samples from
`EtOH-precipitated platelet proteins, they were dialyzed
`against a 406 sample volume of identical 2-D sample
`buffer for 3 h at room temperature with PlusOne Mini
`Dialysis Kit (molecular mass cut-off 1 kDa) (Amersham
`Biosciences,Uppsala, Sweden).
`
`2.5 Preparation of proteins from
`nonprecipitated platelets
`
`Fifteen-hundred microliters of gel-filtered platelet (GFP)
`suspension was centrifuged at 1500 3 g for 10 min at
`room temperature. The supernatant was discarded and
`the platelet pellet was solubilized in 100 mL of 2-D sample
`buffer as above per 100 6 106 platelets.
`
`2.6 Determination of protein concentration
`
`The protein concentration in resolubilized samples was
`determined in triplicate using a CBB protein assay kit with
`BSA as the standard protein (Pierce Biotechnology,
`Rockford, IL, USA). With appropriate predilution, the 2-D
`sample buffer components do not interfere with the pro-
`tein assay. Therefore, the samples were diluted 1:20 with
`PBS and 5% of the 2-D sample buffer was added in the
`BSA standards.
`
`2.7 Analysis of platelet proteins by 1-DE and
`2-DE
`
`IEF was performed loading 120 mg (for gels to be silver-
`stained) or 150 mg (for gels with CyDye-labeled proteins)
`of platelet proteins by in-gel rehydration in a volume of
`450 mL, denaturating the 2-D buffer (7 M urea, 2 M thiou-
`rea, 4% CHAPS, 70 mM DTT, 0.5% Servalyt pH 3–10;
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`Serva, Heidelberg, Germany) onto 24 cm IPG DryStrips,
`pH 3–10 linear and pH 4–7 linear
`(Amersham Bio-
`sciences) and focused for 50 and 30 kVh, respectively,
`using an Amersham IPGphor unit. Before loading onto
`SDS-polyacrylamide gels, IPG strips were incubated for
`15 min in equilibration buffer (50 mM Tris-HCl, pH 8.8, 6 M
`urea, 30% glycerol, 2% SDS) containing 1% DTT and
`then for another 15 min in equilibration buffer containing
`2.5% iodoacetamide. The SDS-polyacrylamide gels
`(26 6 20 cm 6 1 mm, T = 11%, C = 2.6%) were cast
`according to Laemmli [11]. The second dimension was
`performed using an Ettan DALT six System (Amersham
`Biosciences) according to the manufacturer’s instruc-
`tions. 1-D SDS-gel electrophoresis was performed with
`13 6 16 cm gels.
`
`2.8 Protein staining and image analysis of
`silver-stained gels
`
`Silver nitrate staining for analytical gels was performed
`according to Heukeshoven and Dernick [12], while gels
`for MS analysis were stained with mass-compatible silver
`stain according to Shevchenko [13]. The silver-stained gel
`images were digitized using a Molecular Imager FX
`(BioRad). Computer-aided 2-D image analysis was car-
`ried out using the MELANIE 3 software (GeneBio, Ge-
`neva, Switzerland).
`
`2.9 Protein labeling and image analysis of DIGE
`gels
`
`Fifty micrograms of resolubilized platelet protein pre-
`parations was labeled with 333 pmol of CyDye DIGE Fluor
`minimal dyes (Amersham Biosciences). Pre-electropho-
`retic labeling was performed according to the manu-
`facturer’s instructions. The gels with separated labeled
`proteins were scanned using the Typhoon 9410 imager
`(Amersham Biosciences), and the protein patterns were
`displayed with the IQTools software. All sample gel ima-
`ges were processed by the DeCyder DIA (differential in-
`gel analysis) software (Amersham Biosciences) module to
`codetect and differentially quantify the protein spots.
`
`2.10 In-gel protein digestion
`
`The spots of interest were excised from the gels, chopped
`into pieces, and transferred into 0.5 mL tubes (Axygen,
`Union City, CA). The gel pieces were washed twice with
`200 mL of 50 mM NH4HCO3 buffer (pH 8.5) (Sigma) for
`10 min and afterwards with 200 mL of 50 mM NH4HCO3 in
`50% ACN (HPLC-grade; Merck) for 10 min. Subsequently,
`the gel pieces were dehydrated by adding 50 mL ACN and
`
`Page 3 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`2484
`
`M. Zellner et al.
`
`Electrophoresis 2005, 26, 2481–2489
`
`in 180 mL of 10 mM DTT in 50 mM
`allowed to reswell
`NH4HCO3 buffer in order to perform reduction (567C,
`30 min). After cooling to room temperature, the solution
`was replaced by 150 mL of 50 mM iodoacetamide in
`50 mM NH4HCO3 buffer and the gel pieces were incu-
`bated in the dark for 20 min at room temperature. The gel
`pieces were washed three times with 200 mL of 50 mM
`NH4HCO3 buffer and three times with 50 mM NH4HCO3 in
`50% ACN for 10 min at room temperature, dehydrated
`with 50 mL ACN, and dried in the Eppendorf Con-
`centrator 5301 (Eppendorf, Hamburg, Germany)
`for
`5 min. The enzymatic digestion of the proteins was car-
`ried out on ice by a stepwise addition of 0.5–2 mL of
`trypsin
`(sequenzing-grade
`unmodified;
`12.5 ng/mL
`Roche, Basel, Switzerland) in 50 mM NH4HCO3 buffer until
`they were totally rehydrated. Finally, enough 50 mM
`NH4HCO3 buffer was added to keep the gel pieces cov-
`ered during digestion at 377C overnight. After digestion,
`the supernatant was removed and the peptides were
`extracted once with 20 mL 50 mM NH4HCO3 buffer and
`twice with 20 mL 5% formic acid (Sigma) by sonification
`for 5 min at room temperature in an ultrasonic water bath
`(Sonorex RK 255 H; Bandelin, Berlin, Germany).
`
`2.11 NanoHPLC-MS/MS protein sequencing
`
`All nanoHPLC separations were performed on the Ulti-
`Mate system from LC Packings (Amsterdam, The Nether-
`lands). The in-gel digests were loaded onto a precolumn
`(PepMap C18 material, 300 mm ID 6 5 mm length; LC
`Packings) by the FAMOS m-autosampler and the Switchos
`loading pump operated at 20 mL/min using water with
`0.1% TFA (Pierce, Rockford, IL, USA) as mobile phase.
`The sample was eluted from the precolumn in a back flush
`mode. The dimensions of the separation column were
`0.075 mm ID 6 150 mm length, 3 mm particle size. The
`flow rate of the nanoHPLC system was set at 200 nL/min
`and the UV detector was operated at 214 nm using the
`nano UV-Z view flow cell (volume 3 nL). The mobile phas-
`es were A = 95% water (HPLC-grade, Supra-Gradient,
`Biosolve B.V., The Netherlands), 5% ACN (HPLC-grade,
`Supra-Gradient, Biosolve B.V.), 0.1% formic acid (Fluka,
`Buchs, Switzerland); and B = 30% water, 70% ACN, 0.1%
`formic acid. The HPLC gradient for separation was 0–
`50% B in 30 min and 50–100% B in 2 min. The nanoHPLC
`system was coupled to an IT mass spectrometer (LCQ
`Deca XPplus, Thermo Finnigan) via a nanoESI source
`using Pico Tip emitters (New Objective, Cambridge, MA,
`USA). The following ESI parameters were used: spray
`voltage, 1.8 kV; capillary temperature, 1857C; capillary
`voltage, 45 V; tube lens offset voltage, 25 V; and the elec-
`tron multiplier at 21050 V. The collision energy was set
`automatically depending on the mass of the parent ion.
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`Gain control was set to 5 6 107. The data were collected
`in the centroid mode using Dynamic Exclusion. One MS
`experiment
`(full-MS) was followed by three MS/MS
`experiments of the three most intensive ions (intensity at
`least 1 6 106). The analysis of MS/MS spectra with
`respect to the peptide identity was routinely performed by
`applying both the MASCOT (Matrix Science) and the
`SEQUEST (Thermo Finnigan) search engines. A peptide
`was reliably identified only if the individual peptide scores
`were .43 (MASCOT) and .3.2 (SEQUEST).
`
`3 Results
`
`3.1 Sample quality of precipitated platelets and
`compatibility with 2-DE
`
`To evaluate and compare the applicability of TCA precip-
`itation and EtOH precipitation for clinical proteomics, we
`used these methods for the preparation of proteins from
`blood platelets for 2-DE. GFP showed a mean con-
`centration of 117 6 42 6 106 Plt/mL. This suspension
`was subjected to a precipitation either with TCA or with
`EtOH, according to the procedures described in Sec-
`tion 2. The 2-D gel of TCA-precipitated proteins showed a
`clear protein pattern with about 1400 individual spots,
`which were found with good reproducibility (Fig. 1A).
`However, EtOH-precipitated proteins were only well
`separated in the lower pI range (Fig. 1B). In the pI range
`above 6.0, the proteins were only seen as horizontal
`streaks. This indicates a nonsufficient separation of
`EtOH-precipitated proteins in the first-dimensional IEF.
`Figures 1D–F show the electrical current and the voltage
`profile of the IEF separation. The high current peak in the
`first part of the IEF of EtOH-precipitated proteins indi-
`cates that this sample contained high salt concentrations
`(Fig. 1E, arrow). These salts had to be removed by dialysis
`in order to get a satisfactory separation by 2-DE (Fig. 1C
`and F). This dialysis step was included later on in all fol-
`lowing EtOH-precipitation experiments.
`
`3.2 Protein yield
`
`The protein precipitation methods were evaluated in the
`GFP preparations of 60 different volunteers. The amount
`of proteins, which was extracted from 100 6 106 platelets
`after TCA precipitation, was equal to that extractable after
`EtOH precipitation from the same number of platelets
`(Fig. 2A). The protein amount increased linearly with the
`number of platelets used in the experiments (Fig. 2B). As
`an alternative approach for increasing the platelet con-
`centration of GFP, we centrifuged the cell suspension and
`extracted the proteins from the pellet in the 2-D sample
`
`Page 4 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Electrophoresis 2005, 26, 2481–2489
`
`Validation of different protein precipitation methods
`
`2485
`
`Figure 1. 2-D electrophoretic analysis of gel-filtered human platelet suspension following precipitation with (A) TCA and
`(B) EtOH resolubilized in 2-D sample buffer. In (C) EtOH-precipitated and -solubilized platelet proteins were dialyzed
`against fresh 2-D sample buffer to remove salt contamination. Twelve-hundred microliters of prepared platelet proteins was
`loaded by passive rehydration onto 24 cm pH 3–10 IPG strips. Current and voltage profile were recorded of (D) TCA-,
`(E) EtOH-, and (F) EtOH-precipitated and -dialyzed platelet samples during the 1-D. Gels were stained with analytical silver
`nitrate stain.
`
`buffer. The supernatant contained almost no protein (2 mg
`proteins per 100 6 106 platelets). The amount of proteins
`extracted from these nonprecipitated platelets was simi-
`lar to that found after precipitation (124 6 10 mg proteins
`per 100 6 106 platelets). These data show that both pre-
`cipitation methods have a similar protein yield and indi-
`cate that precipitation followed by resolubilization is not
`accompanied by a major protein loss.
`
`3.3 Specificity of protein precipitation
`
`To characterize potential protein specific differences in
`precipitation efficiency, we performed a comparative
`electrophoretic analysis. Nonprecipitated, TCA-pre-
`cipitated, and EtOH-precipitated platelet proteins were
`separated in 1-D SDS-PAGE. The results are shown in
`Fig. 3A. The protein band pattern is similar in all three
`samples. Only the nonprecipitated platelets show some
`slight bands in the upper molecular weight range, which
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`are barely detectable in the precipitated samples. TCA-
`and EtOH-precipitated proteins showed a nearly identical
`pattern of protein bands.
`
`For definitive comparison of different precipitation meth-
`ods on platelet protein pattern, we switched to the differ-
`ential gel electrophoresis (DIGE) technology. TCA-pre-
`cipitated proteins, EtOH-precipitated proteins, and non-
`precipitated proteins were labeled with three different
`fluorescent dyes, combined and separated by 2-DE. The
`overlay of the three signals is shown in Fig. 3B. Most pro-
`tein spots are equally present in all three samples (black
`spots). However, some protein spots which are present in
`the nonprecipitated samples are diminished in the TCA-
`precipitated and EtOH-precipitated samples (magenta
`spots). These proteins are equally distributed all over the
`2-D gel. Almost no TCA-specific protein spots (blue) or
`EtOH-specific protein spots (yellow) are visible. However,
`some green protein spots are present in the 2-D gel, indi-
`cating proteins which are enriched by both precipitation
`
`Page 5 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`2486
`
`M. Zellner et al.
`
`Electrophoresis 2005, 26, 2481–2489
`
`the nonprecipitated sample
`cipitated sample against
`(Figs. 3C and D) and of the EtOH-precipitated sample
`against
`the TCA-precipitated sample (Fig. 3E). About
`75% of all protein spots are present at the same levels in
`the EtOH-precipitated samples and nonprecipitated
`samples, whereas 11.6% are over-represented in the
`EtOH samples and 13.5% in nonprecipitated samples
`(Fig. 3C). Similarly, 79% are present at the same level in
`the TCA-precipitated samples and nonprecipitated sam-
`ples (Fig. 3D). However, when comparing the EtOH-pre-
`cipitated samples with the TCA-precipitated samples,
`91% of all protein spots turned out to be equally present
`in both samples (Fig. 3E). This scatter graph shows a
`substantially narrower Gaussian distribution of the protein
`spots than the other two graphs. This indicates a much
`stronger similarity between the two precipitation methods
`than between one single precipitation method and the
`nonprecipitated samples. Almost all of the 8.7% over-
`represented in TCA samples were overlapping spots
`within the two spikes left and right of the b-actin spots
`(chains of light blue spots indicated by arrows in Fig. 3B).
`Since b-actin was found in the whole spike, it could not be
`determined whether this overrepresentation is derived
`from other proteins or is a b-actin effect.
`
`4 Discussion
`
`The present study evaluates the usability of two different
`protein precipitation methods, the EtOH precipitation and
`TCA precipitation, for the proteomic analysis of human
`platelets. It characterizes the precipitation efficiency on
`an overall as well as on a protein-specific level.
`
`Due to the distinct chemistry of EtOH and TCA, the pro-
`tein samples are treated in a different way for precipita-
`tion. For EtOH precipitation, platelets were diluted in a
`9-fold volume of EtOH and then incubated at room tem-
`perature. Studies with the human plasma showed that
`most proteins precipitate already at an EtOH concentra-
`tion of 40% v/v [5]. However, there were still some soluble
`proteins found under such conditions. The high surplus of
`EtOH used in our study facilitates the precipitation of all
`proteins. EtOH is readily miscible with water, but yields a
`significant heat of solution and has the tendency to
`denature proteins, especially at temperature above 07C
`[6]. In protein fractionation experiments, which are aiming
`at preparing nondenatured proteins, normally a “cold
`EtOH precipitation” protocol is used, in which the solvent
`temperature is always kept below 07C. In proteomics,
`however, denaturation is unproblematic as far as it is not
`accompanied by protein modification. It is, therefore, not
`necessary to decrease the temperature during EtOH pre-
`cipitation. In addition, EtOH has a lower dielectric con-
`
`Figure 2. Platelet protein yield for each precipitation
`method.
`(A) GFP preparations of different volunteers
`(n = 18) were precipitated with TCA as well as with EtOH
`and resolubilized in 2-D sample buffer (70 mL/100 6 106
`platelets for TCA and 100 mL/100 6 106 platelets for
`EtOH precipitation). EtOH-precipitated samples were
`dialyzed. Protein concentration was determined by Brad-
`ford. Values are expressed as mean value 6 SD. (B) Plate-
`let protein yield for each precipitation method of differ-
`ently concentrated GFP suspensions. GFP preparations
`of different volunteers were precipitated with TCA (n = 51
`volunteers; closed circle) or with EtOH (n = 18 volunteers;
`open triangle). Linear regression was calculated for pro-
`tein yield of differently concentrated GFP suspensions
`precipitated by TCA or EtOH.
`
`methods. Proteins which showed a modified abundance
`in precipitated samples in at
`least
`two independent
`experiments are indicated in Fig. 3B and their identity is
`shown in Table 1.
`
`To quantify the protein-specific properties of the three
`extraction methods a scatter plot analysis was per-
`formed. Therefore, we plotted the fluorescence intensity
`signal of every single spot of the EtOH- or TCA-pre-
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`Page 6 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Electrophoresis 2005, 26, 2481–2489
`
`Validation of different protein precipitation methods
`
`2487
`
`Figure 3. 1-D and 2-D DIGE-analysis of nonprecipitated, TCA- and EtOH-precipitated platelet samples. (A) Silver-stained
`1-D gel with 2 mg protein from each platelet preparation. (B) 2-D fluorescence overlay image of 50 mg protein from each
`simultaneously made platelet preparation are focused in one IPG strip (pH 4–7). TCA-precipitated platelet proteins are
`shown in blue, EtOH-precipitated platelet proteins in yellow, and nonprecipitated platelet proteins in magenta. Proteins
`which are equally present in each platelet protein preparation are displayed in black. The figure shows one of two inde-
`pendent experiments. (C–E) Scatter plot analysis of the 2-D gel shown in (B). Each dot represents one protein spot. Its
`position indicates the fluorescence intensity signal in the respective preparation group. Analysis includes 829 spots, which
`were detected in all preparation groups in two independent experiments. Regression line and confidential intervals indi-
`cating the twofold SD are shown. R indicates the linear correlation coefficient. (C) Nonprecipitated versus EtOH-pre-
`cipitated samples; R = 0.988, (D) nonprecipitated versus TCA-precipitated samples; R = 0.975, (E) EtOH-precipitated
`versus TCA-precipitated samples; R = 0.979.
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`Page 7 of 9
`
`YEDA EXHIBIT NO. 2092
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`2488
`
`M. Zellner et al.
`
`Electrophoresis 2005, 26, 2481–2489
`
`Table 1. Identifications of representative precipitation-
`dependent and -independent proteins
`
`Spot
`ID
`
`Protein name
`
`Accession
`number
`
`Protein spots decreased in TCA- and EtOH-precipitated
`platelet samples
`A1
`Transgelin 2 isoform
`A2
`Transgelin 2 isoform
`TM3a) chain isoform
`C1
`C2
`Myosin heavy chain b-subunit
`D
`TM3, fibroblast
`E1
`Vinculin isoform
`E2
`Vinculin isoform
`F
`Nucleosome assembly protein
`1-like 1
`
`Q9BUH5
`Q9BUH5
`P06753
`Q14905
`NCBI 88928
`P18206
`P18206
`P55209
`
`Protein spots increased in TCA and EtOH-precipitated
`platelet samples
`B
`TM3 chain isoform
`G
`PDZ and LIM domain
`
`P06753
`O00151
`
`Protein spots equally present in all platelet samples
`H
`P02672
`Fibrinogen g chain A
`I
`Heat shock cognate 71 kDa
`P11142
`protein
`Chloride channel
`Integrin a IIb isoform
`Albumin
`TM3 chain isoform
`
`O00299
`P08514
`P02768
`P06753
`
`J
`K
`L
`M
`
`a) Tropomyosin a 3
`
`stant at room temperature than below 07C, which further
`facilitates the precipitation [9]. The platelets used in this
`study were suspended in a buffered isotonic salt solution
`in order to avoid hypotonic lysis. Due to the low solubility
`of salts in alcohols, EtOH precipitation is not suitable to
`remove salts from the platelet sample, which resulted in
`an insufficient protein separation in the IEF. This effect
`was also observed when the precipitation was performed
`at 2207C (data not shown). To reduce the salt concentra-
`tion, the resolubilized precipitate had to be dialyzed which
`is a time and cost intensive extension of the EtOH-pre-
`cipitation protocol. However, the quantitative validation of
`the EtOH-precipitated and dialyzed samples indicate that
`dialysis has almost no effects on the protein composition
`of the sample.
`
`For TCA precipitation, the precipitant was added to the
`platelet suspension to a final concentration of 25%. The
`TCA-induced protein precipitation curves are observed to
`be U-shaped [10]. All proteins are soluble at very low TCA
`concentrations and the first precipitation of serum albu-
`min can be observed when the concentration increases
`
` 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`above 5%. Between 15 and 40% TCA even the highly
`TCA-soluble protein cardiotoxin III precipitates. When the
`acid concentrations are raised above 50%, all proteins
`are found to redissolve back into the solution. This is
`thought to be mediated by acid-induced structural tran-
`sitions in the proteins. Such irreversible protein modifica-
`tion should be avoided in the preparation of samples for
`proteome analysis. Thus, the TCA concentration used in
`the present study is suited to precipitate almost all pro-
`teins with a minimum of protein modification. In order to
`avoid any acid-induced protein hydrolysis during storage,
`the precipitate is centrifuged and washed in acetone to
`remove TCA. The organic solvent, acetone, is on one
`hand, a protein precipitant with a similar dielectric con-
`stant as EtOH, and on the other hand, a good solvent of
`TCA.
`
`The present study shows that, in spite of the different
`chemistries and the different precipitation protocols,
`EtOH and TCA precipitation lead to similar results in the
`proteomic analysis. Both the methods seem to induce
`precipitation of almost all proteins of the platelet suspen-
`sion. The incubation of the precipitated proteins in the
`strongly denaturing 2-D buffer, which contained chao-
`tropic agents and detergents, resolubilized almost all
`proteins. The protein content of this solution was inde-
`pendent of the used precipitant. The amount of protein
`found after precipitation and resolubilization was the
`same as in a 2-D buffer extract from the same number
`of platelets, which were concentrated by centrifugation
`instead of precipitat