Fractionation of the human plasma proteome for monoclonal antibody proteomics-based biomarker discovery

Research Article

Fractionation of the human plasmaproteome for monoclonal antibodyproteomics-based biomarker discovery

mAb proteomics, a reversed biomarker discovery approach, is a novel methodology to

recognize the proteins of biomarker potential, but requires subsequent antigen identi-

fication steps. While in case of high-abundant proteins, it generally does not represent a

problem, for medium or lower abundant proteins, the identification step requires a large

amount of sample to assure the proper amount of antigen for the ID process. In this

article, we report on the use of combined chromatographic and precipitation techniques

to generate a large set of fractions representing the human plasma proteome, referred to

as the Analyte Library, with the goal to use the relevant library fractions for antigen

identification in conjunction with mAb proteomics. Starting from 500 mL normal pooled

human plasma, this process resulted in 783 fractions with the average protein concen-

tration of 1 mg/mL. First, the serum albumin and immunoglobulins were depleted

followed by prefractionation by ammonium sulfate precipitation steps. Each precipitate

was then separated by size exclusion chromatography, followed by cation and anion

exchange chromatography. The 20 most concentrated ion exchange chromatography

fractions were further separated by hydrophobic interaction chromatography. All chro-

matography and precipitation steps were carefully designed aiming to maintain the

native forms of the intact proteins throughout the fractionation process. The separation

route of vitamin D-binding protein (an antibody proteomics lead) was followed in all

major fractionation levels by dot blot assay in order to identify the library fraction it

accumulated in and the identity of the antigen was verified by Western blot.

Keywords:

Chromatography / Dot blot screening / Fractionation / Human plasma / mAbproteomics DOI 10.1002/elps.201100018

1 Introduction

Plasma accounts for approximately 55% of the total human

blood volume, and contains a wide concentration range of

proteins with diverse biological functions. Human serum

albumin (HSA) and immunoglobulins are the most

abundant constituents, representing about 75–80% of all

plasma proteins [1]. Besides several other high-abundant

proteins, such as fibrinogen, transferrin, haptoglobin, a1-

acid glycoprotein, etc. (up to �18–20%), it contains

thousands and thousands of medium- and low-abundant

proteins only representing approximately 5% of the total

plasma proteome [2]. Because of its high complexity,

characterization and identification of any medium- or low-

abundant ingredients of the plasma proteome represents a

difficult and time-consuming task. The lower the protein

concentration in the plasma, the higher the sample volumes

it requires for the analysis. In addition, tedious sample

preparation methods are usually needed before individual

proteins or groups of proteins can be adequately purified by

various bioanalytical techniques [3], whose methods some-

times alter the native state of the proteins. Selective

fractionation by precipitation, immunoaffinity partitioning,

chromatography, and electrophoresis or their combinations

are usually used to enrich proteins of interest for down-

stream characterization [4]. In addition to the difficulties

related to the extremely wide concentration range, plasma

biomarker identification should address the presence of

protein isoforms and post-translationally modified species,

which may represent subtle but biologically significant

changes.

The history of human plasma/serum fractionation goes

back more than a hundred years. In the 19th century,

Hofmeister and coworkers reported the effect of alkali and

Andras Kovacs1

Edit Sperling1

Jozsef Lazar2

Attila Balogh2

Janos Kadas2

Akos Szekrenyes1

Laszlo Takacs2

Istvan Kurucz2�

Andras Guttman1,2�

1Horvath Laboratory ofBioseparation Sciences,Research Centre for MolecularMedicine, Medical and HealthScience Center, University ofDebrecen, Debrecen, Hungary

2BioSystems International Kft,Debrecen, Hungary

Received January 9, 2011Revised January 31, 2011Accepted February 2, 2011

Abbreviation: VDBP, vitamin D-binding protein �These authors have contributed equally to this publication.

Correspondence: Professor Andras Guttman, Horvath Labora-tory of Bioseparation Sciences, University of Debrecen, 98Nagyerdei krt, Debrecen H-4032, HungaryE-mail: [email protected]: 136-52-321-562

& 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2011, 32, 1916–19251916

magnesium salts on precipitation based fractionation of

serum proteins, establishing the salting out mechanism [5]

that was one of the most frequently used methods for

decades to obtain albumin as well as a-, b-, and g-globulin

fractions [6, 7]. Another approach used cold ethanol preci-

pitation under precisely controlled conditions (pH, buffer

ionic strength, temperature) producing the major plasma

protein fractions for such clinical applications as transfusion

[8]. Isolation of important human plasma ingredients by the

addition of polyethylene glycol was used for industrial scale

applications [9]. More specific techniques such as affinity

chromatography by Cibacron Blue F-3-GA-conjugated

Sepharose for albumin removal were also developed [10].

This so-called Blue Sepharose chromatography became a

widely applied technique since the approach provided a

versatile tool for partitioning not only albumin but also

lipoproteins, blood coagulation factors, nucleotide-requiring

enzymes, interferon [11, 12], and other plasma proteins [13].

Berglof et al. combined anion and cation exchange chro-

matography and gel filtration to obtain 99% pure albumin

with o1% in aggregated forms [14]. Combination of the

ethanol precipitation process with chromatographic meth-

ods further improved the selectivity of fractionation [15].

Thiophilic adsorption chromatography was reported in the

1980s as an efficient method to remove immunoglobulins

from human plasma [16] while maintaining their native

structures and biological activities [17]. Affinity matrices

based on protein A, protein G, or protein L are also

commonly used to purify various immunoglobulins from

human plasma [18].

The advent of modern, high-performance chromato-

graphic techniques, such as HPLC made more effective

fractionation approaches possible. Hochstrasser and

coworkers reported on comprehensive fractionation of

human plasma proteins using reversed-phase chromato-

graphy separations, yet in their approach, keeping the native

state of proteins was no priority [19]. Others also demon-

strated large-scale fractionation of pooled depleted serum by

anion exchange and reversed-phase chromatography

followed by MS-based proteomic analysis [20]. Antibody-

based depletion methods have also been emerging to parti-

tion-specific proteins [21]. An interesting recent effort used

parallel application of separation methods (anion exchange

chromatography, hydrophobic interaction chromatography,

chromatofocusing, and IEF) on cell lysates from a breast

cancer cell line as a proteomic approach to identify cognate

antigens associated with the disease [22]. Chromatofocusing

in combination with reversed-phase chromatography was

introduced as an automated partitioning system for protein

characterization [23], providing intact proteins but not in

their native form.

Current MS-based proteome profiling technologies are

far from covering the dynamic range necessary for global

proteome analysis due to sensitivity, reproducibility, and

throughput problems [24]. The advent of antibody proteo-

mics [25], a reversed proteomics approach, addresses these

issues by utilizing a large number of antibodies that

recognize most proteins in the human proteome at the low

level using simple mAb capture immunoassays. The

global generation of antibodies specific to various diseases

making the en mass screening of biomarkers possible [26].

Using this approach, discriminating biomarkers are recog-

nized and selected before their cognate antigen identity is

known; however, using this discovery procedure, all the

protein variants (isoforms, PTMs, etc.) present in the plas-

ma are addressed and translating the biomarkers into

bedside clinical assays is straightforward. Biomarkers

discovered by the antibody proteomics approach, on the

other hand, required the means of subsequent antigen

identification, which can represent a challenge in case of

lower abundant proteins, especially when samples are

limited.

To address sample limitation issues, we report here on

the generation of a human plasma proteome library (refer-

red to as Analyte Library) using a combination of chroma-

tographic and precipitation techniques. Starting from a

large volume of human plasma and using alternative puri-

fication procedures, the aim was to obtain hundreds of

fractions to accommodate easier identification of medium-

and low-abundant proteins in their intact and native forms

and/or modified polypeptides for the protein identification

step in mAb proteomics. SDS-PAGE was used to follow the

change of complexity/composition during the fractionation

process. As a representative example, the separation path of

vitamin D-binding protein (VDBP, a medium-abundant

protein in the human plasma with a molecular weight of

52–59 kDa), one of the leading antigens discovered by mAb

proteomics was followed by dot blot assay during the frac-

tionation process. VDBP is an important plasma carrier

protein for vitamin D and its metabolites. Apart from its

specific sterol-binding capacity, this glycoprotein takes part

in actin scavenging and fatty acid transport [27], also inter-

acting with several other important proteins [28]. VDBP is a

member of the albumin, a-fetoprotein, and a-albumin/

afamin gene family [27], and it has been shown to be a

biomarker in several pathologies [29, 30]; therefore, to

illustrate the applicability of the Analyte Library, it was of

special interest to follow this molecule during the fractio-

nation process. We expect that the Analyte Library concept

proves to be useful to simplify the protein identification step

in mAb proteomics as a tool of an alternative discovery

workflow to find plasma-born molecular markers specific to

different diseases.

2 Materials and methods

2.1 Chemicals

Normal pooled human plasma was purchased from

Innovative Research (Novi, MI, USA). Potassium chloride,

guanidine hydrochloride, Tris, glycine, SDS, Brilliant Blue

R, potassium sulfate, sodium phosphate, BSA, polyvinyl-

pyrrolidone, Tween-20, methanol, and Bradford Reagent

Electrophoresis 2011, 32, 1916–1925 Proteomics and 2-DE 1917


were from Sigma-Aldrich (St. Louis, MO, USA). Ammo-

nium sulfate was obtained from Scharlau (Sentmenat,

Spain). Phosphate-buffered saline (PBS), pH 7.4, was from

Invitrogen (Carlsbad, CA, USA). Pierce ECL Western

Blotting Substrate for the dot blot and Western blot analyses

was purchased from Thermo Fisher (Rockford, IL, USA).

Amido Black was from Amresco (Cleveland/Akron, OH,

USA). Water was purified with a Direct-Q 3 Ultrapure Water

System (Millipore, Billerica, MA, USA).

2.2 Methods

For all chromatographic steps, except the immunoglobulin

depletion, an AKTA Purifier system (GE Healthcare,

Uppsala, Sweden) was used with the following LC columns:

XK 26/40 for Blue Sepharose chromatography; XK 16/70

for size exclusion chromatography; HiTrap SP HP (1

and 5 mL size) for cation exchange chromatography; HiTrap

Q HP (size, 1 and 5 mL) for anion exchange chromato-

graphy; and HiTrap Phenyl Sepharose HP (size, 1 mL) for

hydrophobic interaction chromatography (all from GE

Healthcare). UV (280 nm) trace data were collected and

processed with the Unicorn 5.1 software package (GE

Healthcare). Immunoglobulin depletion by thiophilic

adsorption chromatography was conducted on a preparative

size column (20 cm, Spektrum-3D, Debrecen, Hungary)

loaded with Pierce Thiophilic Adsorbent (Thermo Fisher).

Sample concentration and buffer exchange steps were

carried out by Centricon Plus-70 (Ultracel-PL Membrane;

cutoff, 10 kDa), Amicon Ultra (cutoff, 10 kDa) or

Microcon (cutoff, 10 kDa) centrifugal filter units (all from

Millipore). SnakeSkin Pleated Dialysis Tubes (cutoff,

10 kDa) and Slyde-A-Lyzer Dialysis Cassettes were

purchased from Thermo Fisher. SDS-PAGE analyses were

performed in an XCell SureLock Mini-Cell electrophoresis

system (Invitrogen). The protein concentrations of the

interim fractionation levels and the final fractions were

determined by Bradford assay [31] using BSA as standard,

mixing 5 mL protein samples with 250 mL Bradford Reagent

in microtiter plates and the absorbance was measured at

595 nm with Model 680 Microplate Reader (Bio-Rad,

Hercules, CA, USA).

2.3 Albumin depletion from 500 mL normal human

plasma

HSA and associated proteins from 500 mL normal pooled

human plasma were depleted by Blue Sepharose 6 Fast Flow

affinity chromatography using an XK 26/40 column in the

AKTA Purifier system. Twenty times 25 mL plasma was

loaded and separated using a flow rate of 2 mL/min (PBS

buffer) and the flowthrough was collected for downstream

fractionation. The adsorbed albumin (with the associated

proteins) was eluted by 1 M KCl (dissolved in PBS) and

stored at �701C. Between each run, the column was

regenerated by 8 M guanidine hydrochloride and the pooled

elution and regeneration fractions were kept at �701C. All

fractions were analyzed by SDS-PAGE.

2.4 Immunoglobulin depletion

After each individual Blue Sepharose affinity chromatogra-

phy step, the albumin-depleted flowthrough was mixed

with 20 mL Pierce Thiophilic Adsorbent resin in 0.5 M

potassium sulfate containing PBS and incubated

overnight with continuous stirring at 41C. The mixture

was then transferred to a preparative size chromatography

column (20 cm) and the flowthrough was collected. The

resin-filled column was washed with 0.5 M potassium

sulfate in water and the first two wash fractions (both

20 mL) were added to the flowthrough. The bound

immunoglobulins were eluted by 50 mM sodium

phosphate buffer (pH 8.0) and the column was regenerated

by flushing with 8 M guanidine hydrochloride.

All fractions (flowthrough, elution, and regeneration) were

analyzed by SDS-PAGE and kept at �701C until further

processing.

2.5 Concentration, buffer exchange, and ammonium

sulfate precipitation

The large volume of albumin and immunoglobulin-depleted

plasma solution (�3.6 L) was concentrated by 12.5-fold

using Centricon Plus-70 (cutoff, 10 kDa; Millipore) centri-

fugal filter units followed by buffer exchange to PBS by

dialysis (SnakeSkin Pleated Dialysis Tubes, 10 kDa cutoff,

Thermo Fisher). Prefractionation of the concentrated and

dialyzed sample was carried out by ammonium sulfate

precipitation at four consecutive saturation levels of 35, 45,

65, and 75%. The precipitation mixtures were incubated

with the corresponding salt concentration at 41C for 1 h with

continuous stirring. The precipitates were centrifuged at

4500 rpm for 15 min in a J2-HS centrifuge (Beckman

Coulter, Brea, CA, USA) and redissolved in PBS. The

fractions were analyzed by SDS-PAGE and kept at �701C

until further processing.

2.6 Chromatography fractionation steps

2.6.1 Size exclusion chromatography

The AKTA Purifier system with Sephacryl S-200 media

packed in an XK 16/70 column was used for the size

exclusion chromatography step of each PBS-solubilized

precipitates (0.7 mL/min flow rate, PBS). The injection

volume varied between 1 and 1.5 mL, depending on the

amount of loaded proteins. In total, 1 mL fractions was

collected, analyzed by SDS-PAGE, and the fractions with

similar profiles were pooled and kept at �701C.

Electrophoresis 2011, 32, 1916–19251918 A. Kovacs et al.


2.6.2 Cation exchange chromatography

Prior to cation exchange chromatography, the PBS buffer of

all size exclusion chromatography fractions was exchanged

to 50 mM phosphate buffer (pH 5.5) using Millipore

Amicon Ultra centrifugal filter devices (cutoff, 10 kDa) then

separated on a HiTrap SP HP column. Each flowthrough

was collected for consequent anion exchange chromatogra-

phy steps. The bound proteins were eluted by a linear

(0–0.5 M) KCl gradient in 50 mM phosphate buffer (pH 5.5),

and 0.5–1 mL fractions were collected depending on the

protein amount in the loaded samples. The collected

fractions were analyzed by SDS-PAGE and kept at �701C.

2.6.3 Anion exchange chromatography

Prior to anion exchange chromatography, the buffer of each

cation exchange chromatography flowthrough was changed

to 20 mM Tris-HCl (pH 8.5) using Amicon Ultra centrifugal

filter devices (cutoff, 10 kDa), then separated on a HiTrap Q

HP column. The bound proteins were eluted by a linear

(0–0.5 M) KCl gradient in 20 mM Tris-HCl (pH 8.5) and

0.5–1 mL fractions were collected depending on the protein

amount in the loaded samples. The collected fractions were

analyzed by SDS-PAGE and kept at �701C.

2.6.4 Hydrophobic interaction chromatography

The 20 most concentrated ion exchange chromatography

fractions (>8.0 mg/mL) were further separated by hydro-

phobic interaction chromatography using HiTrap Phenyl

Sepharose HP columns. For precipitates with o55%

ammonium sulfate saturation level, 1 M ammonium sulfate

containing 20 mM Tris-HCl buffer (pH 7.1), and for

precipitates with 455% ammonium sulfate saturation level,

2 M ammonium sulfate containing 20 mM Tris-HCl buffer

(pH 7.1) was used as sample/starting buffer, respectively.

The proteins were eluted by a descending ammonium

sulfate gradient (100–0%) in 20 mM Tris-HCl buffer (pH

7.1). An aliquot of 1 mL fractions was collected, analyzed by

SDS-PAGE, and kept at �701C.

2.6.5 Fraction coding

The reference codes for the chromatography fractions

comprise the mark for the % ammonium sulfate saturation

level (35, 45, 65, or 75), the number of the pooled gel

filtration fraction (G1–G4), specify the cation (C) or anion

exchange chromatography fractions (A) and, designate the

hydrophobic interaction chromatography fractions (H). CF

stands for cation-exchange chromatography flowthrough.

Numbers after C, CF, A, and H represent the position of

fractions in the collection tray. For example, 65%-G2-A19-

H9: codes a sample obtained by 65% ammonium sulfate

precipitation followed by gel filtration (fraction ]2), anion

exchange chromatography (fraction ]19), and hydrophobic

interaction chromatography (fraction ]9).

2.7 SDS-PAGE

All interim and final fractions during the library generation

process were analyzed by SDS-PAGE using 4–20% Novex

Tris-glycine gels (Invitrogen) and Tris-glycine-SDS running

buffer (25 mM Tris, 192 mM glycine, and 0.1% SDS, pH

8.6) by applying 150 V for 90 min at room temperature. The

gels were stained by 0.2% Brilliant Blue R in 7.5% acetic

acid containing 50% ethanol and visualized with UVIPro

Gold Gel Documentation System (Uvitec, Cambridge, UK).

2.8 Dot blot and Western blot assays

2.8.1 Dot blot

In total, 3 mg protein per fraction was spotted onto Protran

BA 85 nitrocellulose membrane (Whatman, Maidstone, UK)

in duplicates, placed in a DHM-96 dot blot vacuum

manifold (VWR International, Leuven, Belgium) with

10 kPa vacuum applied (XF54 230 50 vacuum pump,

Millipore). The dried membrane was treated for 30 min

with the blocking buffer of 0.5% polyvinylpyrrolidone in

0.05% Tween-20 containing PBS, following 1 h incubation

with the anti-VDBP-specific mAb (Biosystems International

Kft, Debrecen, Hungary) diluted to 1 mg/mL in the blocking

buffer. The membrane was washed for 5� 1 min in water,

5� 1 min in the blocking buffer, and 5� 1 min in PBS-T

(0.05% Tween-20 containing PBS). Goat anti-mouse IgG-

HRP (Southern Biotech, Birmingham, AL, USA), diluted

8000-fold in the blocking buffer, was used as secondary

antibody. The membrane was incubated with the secondary

antibody for 1 h and washed for 5� 1 min with PBS-T.

Chemiluminescent signals were developed by the applica-

tion of Pierce ECL Western Blotting Substrate and detected

with Gel Logic 2200 PRO gel imaging system (Carestream

Health, Rochester, NY, USA). An aliquot of 0.05 mg of the

primary antibody was spotted on the membrane as a positive

control for the reaction and the blocking buffer was used as

a negative control. Increasing signal intensities were scored

from 0.5 to 5 (ns, no signal).

2.8.2 Western blot

Blocking/antibody diluting and washing buffers were the

same as described above for the dot blot assay, and identical

methods were applied for membrane wash after incubation

with the antibodies. About 2 mg protein per fraction was

separated by SDS-PAGE (as described above) and trans-

ferred onto Protran BA 85 nitrocellulose membrane in a

Mini Trans-Blot Cell system (Bio-Rad) using 25 mM Tris,

192 mM glycine, 20% methanol as blotting buffer. The

transfer process was carried out for 1 h applying a voltage of

135 V. The membrane was treated for 1 h with blocking

buffer of 0.5% polyvinylpyrrolidone in 0.05% Tween-20

containing PBS, incubated overnight at 41C with 5000-fold

diluted anti-vitamin D-protein binding mAb, then washed



and incubated at room temperature with 5000-fold diluted

goat anti-mouse IgG-HRP secondary antibody. The chemi-

luminescent signal was developed by Pierce ECL Western

Blotting Substrate and detected with Gel Logic 2200 PRO gel

imaging system (Carestream Health). The membrane was

stained with 0.05% Amido black (in 10% acetic acid) for 1 h

at room temperature, washed 2� 10 min with 5% acetic

acid, and visualized with UVIPro Gold Gel Documentation

System (Uvitec).

3 Results and discussion

As a continuation of our previously reported antibody

proteomics work [26], in this article, we report on the

generation of an Analyte Library containing a large set of

fractions representing the human plasma proteome. Start-

ing from 500 mL of normal human plasma, we have

obtained a large number of fractions by using the

combination of nondenaturing chromatographic and salt

precipitation steps, also aiming to maintain the native

nature of the proteins in the fractions. The library serves as

depository for mAb-based biomarker research offering the

option to provide sufficient amount of analyte for immune-

separation-linked LC-MS-based identification of medium-/

low-abundant plasma proteins (along with their modified

variants) of biomarker potential. The fractionation route of

VDBP (an antibody proteomics lead) was followed by a dot

blot assay during the library generation process.

3.1 Albumin and immunoglobulin depletion

The first step toward the generation of a comprehensive

Analyte Library containing mainly native proteins was the

removal of the large albumin content of human plasma.

Since the cost of modern antibody-based depletion methods

at the 500 mL plasma scale was prohibitory, Blue Sepharose

chromatography was applied. This affinity chromatography

step resulted in the albumin-depleted flowhtrough along

with the eluted albumin and regeneration fractions. All

fractions were analyzed by SDS-PAGE and as expected, the

albumin content in the flowthrough was significantly

decreased. The elution step resulted in a fraction containing

the majority of the depleted albumin, along with some

albumin-associated proteins and proteins showing affinity to

Blue Sepharose. The regeneration eluate still contained

some remaining HSA and other proteins not eluted by the

1 M KCl wash. The albumin removal step was followed by

the depletion of immunoglobulins (the second largest

component of human plasma) using Thiophilic Adsorbent

Chromatography (based on the same cost consideration as

above). The bound immunoglobulins were eluted from the

adsorbent beads and collected. After these two steps, a large

volume (�3600 mL) of albumin- and immunoglobulin-

depleted flowthrough was obtained, ready for further

fractionation. The elution and regeneration fractions of the

albumin and immunoglobulin depletion steps were stored

for further fractionation in a similar way as the flowthrough

fractions with the expectation to provide additional proteins

for the Analyte Library.

3.2 Ammonium sulfate precipitation

Prior to the ammonium sulfate precipitation steps, the

albumin- and immunoglobulin-depleted plasma solution was

preconcentrated and dialyzed against PBS. In this way, the

total volume of the albumin/immunoglobulin depletion

flowthrough was decreased to 335 mL. The salt concentrations

for the precipitation steps (35, 45, 65, and 75%) were carefully

chosen taking into consideration solubility and possible

denaturing issues. Please note that our preliminary experi-

ments showed that the interim 55% saturation level provided

identical protein distribution profile as of the 65% level (based

on chromatographic trace and SDS-PAGE pattern) and

therefore, this salt concentration step was skipped. At lower

ammonium sulfate concentrations (o55%), mostly the

hydrophobic and consequently less water-soluble proteins

were precipitated, whereas at higher ammonium sulfate

concentrations (>55%), the less hydrophobic, i.e., more

water-soluble proteins were salted out [32]. Since the super-

natant of the 75% precipitation step did not contain any

detectable amount of proteins, the resulting four fractions were

subjected for further separation (Flowchart 1).

3.3 Size exclusion chromatography

Further fractionation of the ammonium sulfate precipitates

started using size exclusion chromatography. The collected

fractions were analyzed by SDS-PAGE and the fractions

with similar profiles were pooled. Profiles were assessed by

comparing the number of bands of the same molecular

weight range and their approximate amounts. This step

generated 14 interim Analyte Library fractions (Flowchart

1): 3–3 from the 35 and 45% ammonium sulfate precipita-

tion steps, and 4–4 from the 65 and 75% ammonium sulfate

precipitation steps, respectively.

3.4 Cation exchange chromatography

Each pooled size exclusion chromatography fraction was

further separated by strong cation exchange chromatogra-

phy (SCX). The injected proteins were eluted by means of an

ascending linear potassium chloride gradient and 1 mL

fractions were collected. Figure 1 shows a representative

cation exchange chromatography separation trace (UV

280 nm, solid black line) of the third size exclusion fraction

of the 65% ammonium sulfate precipitate, referred to as

65%-G3. The dark gray line is the conductivity trace and the

gray line is the elution buffer concentration profile. The

large peak at the beginning of the chromatogram represents



the flowthrough, containing all unbound proteins

(0–110 min). Fractions were collected with the start of the

KCl gradient (138 min) and stopped at 182 min. Selected

SDS-PAGE images in the insets (fractions 65%-G3-17, -23,

and -28) show the representative SDS-PAGE patterns of the

major peaks in the chromatogram. The cation exchange

chromatography step resulted in 322 fractions (Flowchart 1).

3.5 Anion exchange chromatography

Each individual cation exchange chromatography flow-

through fraction was subsequently separated by strong

anion exchange chromatography (SAX) by means of an

ascending linear potassium chloride gradient. In contrast to

cation exchange chromatography, none of the anion

exchange chromatography flowthroughs contained any

detectable amounts of proteins. A typical anion exchange

chromatography trace (UV 280 nm, black line) is shown in

Fig. 2 (the cation exchange flowthrough of the fourth size

exclusion fraction of the 75% ammonium sulfate precipi-

tate; 75%-G4-CF). The dark gray line represents the

conductivity, whereas the gray denotes the elution buffer

concentration change in time. Fractions (1 mL) were

collected at the onset of the salt gradient, i.e. from 31 min

and stopped at 69 min. The insets show the selected SDS-

PAGE images of fractions 75%-G4-A14, -19, -25, and -28,

representing the separated peaks. Please note that anion

exchange chromatography provided a higher number of

Blue Sepharose chromatography of 500 mL normal pooled human plasma(total protein: cc. 24 g)

Flowthrough: HSA depleted plasmaHSA eluted by 1M KCl Regeneration by 8 M guanidine-HCl

Immunoglobulin depletion with Thiophilic Adsorbent

HSA and immunoglobulin depleted plasma(total protein: 3290 mg)Immunoglobulins eluted

Concentration, buffer exchange (PBS)

Ammonium sulfate precipitation(4 fractions, total protein: 1789 mg)

Size exclusion chromatography(14 fractions, total protein: 878 mg)

Cation exchange chromatography(322 fractions, total protein: 258 mg)

Anion exchange chromatography(316 fractions, total protein: 230 mg)

Hydrophobic interaction chromatography of the 20most concentrated ion exchange fractions

(145 fractions, total protein: 295 mg)

ANALYTE LIBRARY(783 fractions, total protein: 783 mg)

Flowchart 1. Analyte Library generation flowchart. Numbers in parenthesis show the number of fractions collected at the individualsteps and the total amount of proteins at that level of fractionation.

260160

MW

65%

-G3-

C17

65%

-G3-

C23

65%

-G3-

C28

160

11080605040

mAU

30

20

15

3000

2500

2000

1500

1000

500

0

Time (min)

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Figure 1. Cation exchange chromatography trace (UV 280 nm,black line) of the interim Analyte Library fraction of 65%-G3 (65%ammonium sulfate saturation, third pooled gel filtration frac-tion). Insets show the corresponding SDS-PAGE images of themajor peaks obtained. The dark gray and the gray lines show tothe conductivity and elution buffer concentration traces, respec-tively. MW: molecular weight marker. Conditions: HiTrap SP HPcolumn with 0 to 0.5 M KCl (in 50 mM phosphate buffer, pH 5.5)gradient elution, flow rate: 1 mL/min, 1 mL fractions.



peaks with more complex SDS-PAGE patterns and less

overlap between neighboring fractions. The anion exchange

chromatography step contributed 316 additional fractions to

the Analyte Library (Flowchart 1).

3.6 Hydrophobic interaction chromatography

The 20 most concentrated anion exchange chromatography

fractions (>8.0 mg/mL) were subjected to further separation

by hydrophobic interaction chromatography (HIC). Two

types of sample/starting buffers were used with the phenyl

type stationary phase to ensure that their ammonium sulfate

concentration level did not cause precipitation. For the

fractions precipitated at the 35–45% ammonium sulfate

saturation level, 1 M ammonium sulfate; whereas for the

fractions precipitated at 65–75% ammonium sulfate satura-

tion level, 2 M ammonium sulfate was used and applied as

sample/starting buffer, respectively. The proteins were

eluted by a descending ammonium sulfate gradient from

1 M to 0, as well as 2 M to 0 (NH4)2SO4 for the 35–45% and

65–75% ammonium sulfate-precipitated fractions, respec-

tively. One milliliter fractions were collected and analyzed by

SDS-PAGE. Separation of the more hydrophobic fractions

(35–45% ammonium sulfate saturation level) resulted in

better resolution than the rather hydrophilic ones (ammo-

nium sulfate saturation, 65–75%). A representative hydro-

phobic interaction chromatography trace (UV 280 nm, black

line) of the 65% ammonium sulfate saturation, third pooled

gel filtration fraction, anion exchange chromatography,

fraction ]17 (65%-G3-A17) is shown in Fig. 3. Dark gray

and gray lines represent conductivity and elution buffer

concentration changes in time, respectively. Actual fraction

collection took place between 16 and 31 min. The insets

show the SDS-PAGE images of fractions 65%-G3-A17-H8 to

-H12. The flowthroughs of hydrophobic interaction chro-

matography separations did not contain any detectable

amount of proteins. A total of 145 hydrophobic interaction

chromatography fractions have been collected, making a

total of 783 fractions as a result of the Analyte Library

generation process (Flowchart 1).

260160

MW

75%

-G4-

A14

75%

-G4-

A19

75%

-G4-

A25

75%

-G4-

A28

160

11080

6050

30

40

30

20

15

3000

mAU

2500

2000

1500

1000

500

Time (min)

0

0 5 10 15 20 25 30 35 40 45 50 55 60 65

Figure 2. Anion exchange chromatography trace (UV 280 nm,dark line) of the cation exchange chromatography flowthroughof fraction 75%-G4 (75% ammonium sulfate saturation, fourthpooled gel filtration fraction). Insets show the correspondingSDS-PAGE images of the major peaks obtained. The dark greyand the gray lines show the conductivity and elution bufferconcentration traces, respectively. MW, molecular weightmarker. Conditions: HiTrap Q HP column, 0–0.5 M KCl(in 20 mM Tris-HCl, pH 8.5) gradient elution; flow rate, 1 mL/min, 1 mL fractions.

260160

MW

65%

-G3-

A17

-H8

65%

-G3-

A17

-H9

65%

-G3-

A17

-H10

65%

-G3-

A17

-H11

65%

-G3-

A17

-H12

110

6080

4050

mAU

30

20

151800

1600

1400

1200

1000

800

600

0

400

200

Time (min)

8 10 12 14 16 18 20 22 24 26 28

Figure 3. Hydrophobic interaction chromatography trace (UV280 nm, dark line) of the anion exchange chromatographyfraction 65%-G3-A17 (65% ammonium sulfate saturation, thirdpooled gel filtrated fraction, anion exchange chromatographyfraction ]17). Insets show the SDS-PAGE images of fractions8–12. The dark grey and the gray lines show the conductivity andelution buffer concentration traces, respectively. MW, molecularweight marker. Conditions: HiTrap Phenyl Sepharose HPcolumn, 2–0 M descending ammonium sulfate (in 20 mM Tris-HCl buffer, pH 7.1) gradient elution; flow rate, 1 mL/min, 1 mLfractions.



3.7 Dot blot and Western blot analysis of selected

fractions

mAbs have been generated against pooled normal human

plasma as reported earlier [26]. Based on the considerations

described in Section 1, the dot blot assay followed the route

of VDBP (antigen) during the library generation process.

The specificity of the antibody was verified with Western

blot and also determined by LC-MS from the immunopre-

cipitated protein band of the SDS-PAGE gel (manuscript in

preparation). As Table 1 summarizes, fractions from all

levels of the separation process were selected on the basis of

their complexity (defined by SDS-PAGE) and protein

concentrations (>0.1 mg/mL). Selection included eight

interim fractions from three early levels of the library

generation process (Table 1, levels 1–3): the albumin- and

immunoglobulin-depleted plasma, the four ammonium

sulfate precipitates and three size exclusion chromatography

fractions (45%-G2, 65%-G2, and 75%-G3). The rest of the

samples were selected from the final fractions including 14

cation exchange (35%-G2-C13, 45%-G1-C24, 45%-G2-C11,

45%-G3-C16, 65%-G1-C9, 65%-G1-C23, 65%-G2-C15, 65%-

G3-C17, 65%-G4-C16, 75%-G1-C13, 75%-G2-C8, 75%-G3-

C19, 75%-G4-C12, and 75%-G4-C22), 16 anion exchange

(35%-G2-A26, 45%-G1-A21, 45%-G2-A30, 45%-G3-A13,

45%-G3-A25, 65%-G1-A22, 65%-G2-A17, 65%-G2-A22,

65%-G3-A12, 65%-G3-A23, 65%-G4-A11, 75%-G1-A14,

75%-G2-A10, 75%-G3-A26, 75%-G4-A17, and 75%-G4-A27)

and 8 hydrophobic interaction chromatography (45%-G2-

A21-H8, 65%-G2-A19-H9, 65%-G3-A12-H8, 65%-G3-A17-

H10, 75%-G3-A14-H8, 75%-G3-A17-H9, 75%-G4-A15-H9,

and 75%-G4-A29-H10) fractions representing the three

higher separation levels (Table 1, levels 4–6).

Figure 4A shows the chemiluminescent signals in the

dot blot membrane, showing the distribution and the path

of the target protein in the selected Analyte Library fractions.

Signal intensities were scored from 0.5 to 5 (ns, no signal)

after the reaction between the fractions of the different levels

of the fractionation process and the anti-VDBP mAb

summarized in Table 1. Out of eight interim fractions seven

proved to be positive (i.e. contained the antigen), whereas

out of 38 final fractions only 19 contained the antigen in low

concentrations, mainly the ones derived from the 45 and

65% ammonium sulfate precipitates, suggesting that during

the library generation process the complexity of the fractions

gradually decreased. High enrichment level of the antigen

was obtained in the hydrophobic interaction chromato-

graphy fraction 65%-G3-A17-H10 (65% ammonium sulfate

saturation, third pooled gel filtration fraction, anion

exchange fraction ]17, hydrophobic interaction fraction ]10)

(also apparent in Fig. 3, SDS-PAGE inset), which exhibited

high signal intensity (scored 5 at positions F3 and F4 in

Fig. 4A). Upstream fractions (depleted plasma, ammonium

sulfate precipitates, gel filtration, cation and anion exchange

chromatography fractions) revealed lower signal intensity.

In the SDS-PAGE image of fraction 65%-G3-A17-H10

in Fig. 3, the intensive protein band observed at approxi-

mately 52 kDa matched the molecular weight of the VDBP

reported in the literature. This was verified with Western

blot method, where in addition to 65%-G3-A17-H10 and its

neighboring fractions (65%-G3-A17-H8, 65%-G3-A17-H9,

65%-G3-A17-H11, and 65%-G3-A17-H12) several interim

Table 1. Fractions selected for dot blot assay-based screening

from the different levels of the library generation

process

Dot blot

position

Fractionation

level

Fraction Signal

intensity

A1-2. Negative control: blocking buffer

A3-4. Positive control: anti-VDBP mAb

A5-6. 1 HSA and immunoglobulin-depleted

plasma

2

A7-8. 2 35% Ammonium sulfate precipitate 1

A9-10. 4 35%-G2-C13 0.5

A11-12. 5 35%-G2-A26 ns

B1-2. 2 45% Ammonium sulfate precipitate 1

B3-4. 3 45%-G2 1

B5-6. 4 45%-G1-C24 2

B7-8. 4 45%-G2-C11 1

B9-10. 4 45%-G3-C16 0.5

B11-12. 5 45%-G1-A21 ns

C1-2. 5 45%-G2-A30 1

C3-4. 5 45%-G3-A13 1

C5-6. 5 45%-G3-A25 0.5

C8. 6 45%-G2-A21-H8 1

C10. 2 65% Ammonium sulfate precipitate 2

C11-12. 3 65%-G2 ns

D1-2. 4 65%-G1-C9 0.5

D4. 4 65%-G1-C23 1

D5-6. 4 65%-G2-C15 1

D8. 4 65%-G3-C17 0.5

D9-10. 4 65%-G4-C16 ns

D11-12. 5 65%-G1-A22 ns

E1-2. 5 65%-G2-A17 0.5

E3. 5 65%-G2-A22 1

E5-6. 5 65%-G3-A12 0.5

E7-8. 5 65%-G3-A23 1

E9-10. 5 65%-G4-A11 ns

E11-12. 6 65%-G2-A19-H9 ns

F1-2. 6 65%-G3-A12-H8 ns

F3-4. 6 65%-G3-A17-H10 5

F5-6. 2 75% Ammonium sulfate precipitate 2

F7-8. 3 75%-G3 1

F9-10. 4 75%-G1-C13 1

F11-12. 4 75%-G2-C8 0.5

G1-2. 4 75%-G3-C19 ns

G3-4. 4 75%-G4-C12 ns

G5-6. 4 75%-G4-C22 ns

G7-8. 5 75%-G1-A14 ns

G9-10. 5 75%-G2-A10 ns

G11-12. 5 75%-G3-A26 ns

H1-2. 5 75%-G4-A17 ns

H3-4. 5 75%-G4-A27 ns

H5-6. 6 75%-G3-A14-H8 ns

H7-8. 6 75%-G3-A17-H9 ns

H9-10. 6 75%-G4-A15-H9 ns

H11-12. 6 75%-G4-A29-H10 ns

Dot blot reaction intensities were scored from 0.5 to 5; ns, no

signal.



upstream fractions on the lineage such as HSA and

immunoglobulin-depleted plasma, 65% ammonium sulfate

precipitate, 65%-G3 and 65%-G3-A17 were also analyzed

(Fig. 4B). The enrichment of the antigen is apparent from

the increased signal intensity of VDBP. Please note that the

hydrophobic interaction chromatography fraction 65%-G3-

A17-H8 did not contain the VDBP, whereas the antigen

accumulated in fractions 65%-G3-A17-H9 to 65%-G3-A17-

H12. Figure 4C shows a part of the Amido Black-stained

blotting membrane, showing the detectable protein bands in

the fractions including the VDBP at approximately 52 kDa

in eight fractions out of the examined nine.

4 Concluding remarks

A 783-fraction containing Analyte Library, representing a

significant portion of the human plasma proteome, was

generated from 500 mL normal pooled human plasma using

protein precipitation and chromatographic techniques. The

main separation steps, exhibiting the numbers of interim

and final fractions (as well as their total protein concentra-

tion) obtained during the library generation process is

shown in Flowchart 1. It is important to note that this

Analyte Library was generated with the consideration to

maintain the native forms of the proteins and hence the

individual fractions can be screened by mAb proteomics.

The fractions cherry-picked-based hits during the screening

process are used for antigen identification either by Western

blotting or by LC-MS-based technology The complexity of

the interim and final fractions was checked by SDS-PAGE

and selected fractions representing different levels of the

fractionation process were interrogated by dot blot assay for

VDBP (an antibody proteomics screening hit) as an

exemplary demonstration of the fractionation path. As

expected, the initial fractionation steps (ammonium sulfate

precipitation and size exclusion chromatography steps) did

not result in discrete partitioning, making VDBP detectable

only at low levels in these fractions. However, later

fractionation steps in the library generation process

(anion exchange chromatography and hydrophobic interac-

tion chromatography) resulted in almost discrete accumula-

tion of this particular protein in fractions 65%-G3-A17-

H9–H12, as demonstrated by the SDS-PAGE and dot blot

assay. The identity of VDBP, as a major band, was

confirmed by the strong Western blot signal. Such

enrichment of any individual protein in the Analyte

Library fractions is expected to provide higher specificity

for the protein identification process than that of

more complex samples, such as whole plasma. The

possibility that individual antigens will be found in more

fractions, even in ones located in different lineages of the

library generation flow chart cannot be excluded due to

various PTMs of the plasma proteins (e.g. sialylation,

acetylation, etc.). Our future plan is to use these library

fractions in our antibody proteomics-based biomarker

discovery and identification endeavors for antigen identifica-

tion (protein ID) and also for limited immunization, to

generate more specific antibodies against the various

isoforms of the same proteins.

dep

lete

d p

lasm

a

65%

AS

pre

cip

itat

e

65%

-G3

65%

-G3-

A17

65%

-G3-

A17

-H8

65%

-G3-

A17

-H9

65%

-G3-

A17

-H10

65%

-G3-

A17

-H11

65%

-G3-

A17

-H12

BA

52 kDa

C52 kDa

Figure 4. (A) Dot blot assay analysis to follow the accumulation path of the VDBP during the library generation process. Membrane map:A1-2, negative control (blocking buffer, 3 mL); A3-4, positive control (0.05 mg VDBP mAb); A5-6, albumin- and immunoglobulin-depletedplasma fraction; A7-8, 35% ammonium sulfate precipitate; A9-10, 35%-G2-C13, A11-12, 35%-G2-A26; B1-2, 45% ammonium sulfateprecipitate; B3-4, 45%-G2; B5-6, 45%-G1-C24; B7-8, 45%-G2-C11; B9-10, 4%5-G3-C16; B11-12, 45%-G1-A21; C1-2, 45%-G2-A30; C3-4, 45%-G3-A13; C5-6, 45%-G3-A25; C7, no sample; C8, 45%-G2-A21-H8; C9, no sample; C10, 65% ammonium sulfate precipitate; C11-12, 65%-G2;D1-2, 65%-G1-C9; D3-4, 65%-G1-C23; D5-6, 65%-G2-C15; D7, no sample; D8, 65%-G3-C17; D9-10, 65%-G4-C16; D11-12, 65%-G1-A22; E1-2,65%-G2-A17; E3, 65%-G2-A22; E4, no sample; E5-6, 65%-G3-A12; E7-8, 65%-G3-A23; E9-10, 65%-G4-A11; E11-12, 65%-G2-A19-H9; F1-2,65%-G3-A12-H8; F3-4, 65%-G3-A17-H10; F5-6, 75% ammonium sulfate precipitate; F7-8, 75%-G3; F9-10, 75%-G1-C13; F11-12, 75%-G2-C8;G1-2, 75%-G3-C19; G3-4, 75%-G4-C12; G5-6, 75%-G4-C22; G7-8, 75%-G1-A14; G9-10, 75%-G2-A10; G11-12, 75%-G3-A26; H1-2, 75%-G4-A17; H3-4, 75%-G4-A27; H5-6, 75%-G3-A14-H8; H7-8, 75%-G3-A17-H9; H9-10, 75%-G4-A15-H9; H11-12, 75%-G4-A29-H10. (B) Westernblot assay of VDBP based on SDS-PAGE gel separation of selected library fractions. Fraction codes are given above the lanes. (C) AmidoBlack-stained section of the Western blot membrane showing the 52 kDa band of the VDBP. Fraction codes are the same is in (B). All dotblot and Western blot conditions are detailed in Section 2.8.



The financial support of the EU-FP6/COBRED and K-81839 OTKA grants is greatly appreciated. This work was alsopartially supported by the National Office for Research andTechnology of Hungary (TECH-09-A1-2009-0113; mAB-CHIC).

The authors have declared no conflict of interest.

5 References

[1] Burnouf, T., Transfus. Med. Rev. 2007, 21, 101–117.

[2] Anderson, N. L., Anderson, N. G., Mol. Cell. Proteomics2002, 1, 845–867.

[3] Zolotarjova, N., Mrozinski, P., Chen, H., Martosella, J.,J. Chromatogr. A 2008, 1189, 332–338.

[4] Fountoulakis, M., Juranville, J. F., Jiang, L., Avila, D.,Roder, D., Jakob, P., Berndt, P., Evers, S., Langen, H.,Amino Acids 2004, 27, 249–259.

[5] Kunz, W., Henle, J., Ninham, B. W., Curr. Opin. ColloidInterface Sci. 2004, 9, 19–37.

[6] Howe, P. E., J. Biol. Chem. 1921, 49, 93–107.

[7] Kibrick, A. C., Blonstein, M., J. Biol. Chem. 1948, 176,983–987.

[8] Cohn, E. J., Strong, L. E., Hughes, W. L., Mulford, D. J.,Ashworth, J. N., Melin, M., Taylor, H. L., J. Am. Chem.Soc. 1946, 68, 459–475.

[9] Polson, A., Ruiz-Bravo, C., Vox Sang 1972, 23, 107–118.

[10] Travis, J., Bowen, J., Tewksbury, D., Johnson, D.,Pannell, R., Biochem. J. 1976, 157, 301–306.

[11] Affinity Chromatography: Principles and Methods, GEHealthcare Bio-Sciences AB, Uppsala, 2007, pp. 70–73.

[12] Knight, E., Jr, Fahey, D., J. Biol. Chem. 1981, 256,3609–3611.

[13] Gianazza, E., Arnaud, P., Biochem. J. 1982, 201,129–136.

[14] Berglof, J. H., Eriksson, S., Curling, J. M., J. Appl.Biochem. 1983, 5, 282–292.

[15] Adcock, W. L., MacGregor, A., Davies, J. R., Hattarki, M.,Anderson, D. A., Goss, N. H., Biotechnol. Appl.Biochem. 1998, 28, 85–94.

[16] Porath, J., Maisano, F., Belew, M., FEBS Lett. 1985, 185,306–310.

[17] Hutchens, T. W., Porath, J., Anal. Biochem. 1986, 159,217–226.

[18] Greenough, C., Jenkins, R. E., Kitteringham, N. R.,Pirmohamed, M., Park, B. K., Pennington, S. R.,Proteomics 2004, 4, 3107–3111.

[19] Rose, K., Bougueleret, L., Baussant, T., Bohm, G.,Botti, P., Colinge, J., Cusin, I., Gaertner, H., Gleizes, A.,Heller, M., Jimenez, S., Johnson, A., Kussmann, M., Menin,L., Menzel, C., Ranno, F., Rodriguez-Tome, P., Rogers, J.,Saudrais, C., Villain, M., Wetmore, D., Bairoch, A., Hoch-strasser, D., Proteomics 2004, 4, 2125–2150.

[20] Faca, V., Pitteri, S. J., Newcomb, L., Glukhova, V.,Phanstiel, D., Krasnoselsky, A., Zhang, Q., Struthers, J.,Wang, H., Eng, J., Fitzgibbon, M., McIntosh, M., Hanash,S., J. Proteome Res. 2007, 6, 3558–3565.

[21] Kim, M. R., Kim, C. W., J. Chromatogr. B Analyt. Tech-nol. Biomed. Life Sci. 2007, 849, 203–210.

[22] Mun, J., Kim, Y. H., Yu, J., Bae, J., Noh, D. Y., Yu, M. H.,Lee, C., Electrophoresis 2010, 31, 3428–3436.

[23] Schlautman, J. D., Rozek, W., Stetler, R., Mosley, R. L.,Gendelman, H. E., Ciborowski, P., Proteome Sci. 2008, 6,26.

[24] Rifai, N., Gillette, M. A., Carr, S. A., Nat. Biotechnol.2006, 24, 971–983.

[25] Blow, N., Nature 2007, 447, 741–744.

[26] Csanky, E., Olivova, P., Rajnavolgyi, E., Hempel, W.,Tardieu, N., Katalin, E. T., Jullien, A., Malderez-Bloes, C.,Kuras, M., Duval, M. X., Nagy, L., Scholtz, B., Hancock,W., Karger, B., Guttman, A., Takacs, L., Electrophoresis2007, 28, 4401–4406.

[27] Speeckaert, M., Huang, G., Delanghe, J. R., Taes, Y. E.,Clin. Chim. Acta 2006, 372, 33–42.

[28] Thrailkill, K. M., Jo, C. H., Cockrell, G. E., Moreau, C. S.,Fowlkes, J. L., J. Clin. Endocrinol. Metab. 2011, 96,142–149.

[29] Disanto, G., Ramagopalan, S. V., Para, A. E., Handunnetthi,L., J. Neurol. 2011, 258, 353–358.

[30] Bortner, J. D., Richie, J. P., Das, A., Liao, J., Umstead,T. M., Stanley, A., Stanley, B. A., Belani, C. P.,El-Bayoumy, K., J. Proteome Res. 2011, 10, 1151–1159.

[31] Bradford, M. M., Anal. Biochem. 1976, 72, 248–254.

[32] Stryer, L., Biochemistry, W. H. Freeman and Company,New York 1995, pp. 45–74.



Documents

Fractionation of the human plasma proteome for monoclonal antibody proteomics-based biomarker discovery