www.proteomics-journal.

com

Page 1

Proteomics

Technical brief
An alternative method for serum protein depletion/enrichment by precipitation at
mildly acidic pH values and low ionic strength

Ann-Kristin Henning1, Dirk Albrecht2 , Katharina Riedel2, Thomas C. Mettenleiter1, Axel

Karger1*

Institute of Molecular Virology and Cell Biology, Friedrich-Loeffler-Institut,

Sdufer 10, 17493 Greifswald-Insel Riems, Germany

2

Institute for Microbiology, Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-

Jahn-Strae 15, 17487 Greifswald, Germany

* Corresponding author.
Dr. Axel Karger
Postal address: Friedrich-Loeffler-Institut, Sdufer 10, 17493 Greifswald Insel Riems,
Germany
Telephone number: +49 3835171247.
E-mail address: axel.karger@fli.bund.de
Keywords: equalization, ion strength, precipitate, proteome, serum
Total number of words: 2959

Received: 05-Jun-2014; Revised: 15-Dec-2014; Accepted: 28-Jan-2015

This article has been accepted for publication and undergone full peer review but has not been through the copyediting,
typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of
Record. Please cite this article as doi: 10.1002/pmic.201400345.
This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 2

Proteomics

Abstract
Serum proteome analysis is severely hampered by the extreme dynamic range of protein
concentrations, but tools for the specific depletion of highly abundant serum proteins lack for
most farm and companion animals. A well-established alternative strategy to reduce the
dynamic range of plasma protein concentrations, treatment with combinatorial peptide ligand
libraries (CPLL), is generally applicable but requires large amounts of sample. Therefore,
additional depletion/enrichment protocols for plasma and serum samples from animals are
desirable. In this respect we have tested a protein precipitate that formed after withdrawal of
salt from human, bovine, or porcine serum at pH 4.2. The bovine sample was composed of
over 300 proteins making it a potential source for biomarker discovery. Precipitation was
highly reproducible and the concentrations of albumin and other highly abundant serum
proteins were strongly reduced. In comparison to CPLL-treatment, precipitation did not
introduce any selection bias based on hydrophathy or pI. However, the composition of both
preparations was partially complementary. Salt withdrawal at pH 4.2 is suggested as
additional depletion/enrichment strategy for serum samples. Also, we point out that the
removal of precipitates from serum samples under the described conditions bears the risk of
losing a valuable protein fraction.
Plasma and serum analysis is severely hampered by the extreme dynamic range of protein
concentrations which spans 10 or more orders of magnitude [1, 2]. According to Tirumalai et
al. [3] the 10 most abundant proteins in human plasma constitute approximately 90% of the
total protein content. The most abundant serum protein is albumin, accounting for more than
50% of its protein content [3].

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 3

Proteomics

To facilitate serum proteome analysis or biomarker studies two strategies have been
developed to reduce the range of protein concentrations. Either highly abundant proteins are
depleted by specific extraction or protein concentrations are equalized e.g. by solid phase
extraction using immobilized combinatorial peptide ligand libraries (CPLL). The first
approach relies on affinity extraction of the targeted proteins. Removal of albumin from
human sera by Cibacron dyes [4] has been a cornerstone of plasma protein analysis.
Chromatographic maxtrices for the immunodepletion of the most abundant proteins have
been developed for human sera but also for some small laboratory animals [5, 6]. All
depletion techniques bear the risk of unintentional loss of untargeted proteins that bind to the
matrix in an unspecific way or are ligands of the highly abundant targeted proteins [7]. The
second strategy, equalization of protein concentrations, can be implemented by extraction
with libraries of immobilized combinatorial hexapeptides [8] or by sample displacement
chromatography [9]. Both approaches use modified solid supports that are treated with excess
plasma proteins. Under conditions of saturation, low abundance proteins are enriched and
high abundance components with low affinity to the support can be removed by washing. As
for the most farm and companion animals tools for the specific depletion of abundant plasma
proteins, e.g. by immunodepletion, are not available, equalization remains as the only option
to prepare sera or plasmas for proteome analysis. A major drawback of equalization is the
need for large amounts of serum to achieve saturating conditions and acceptable protein
yields which, in our laboratory, are approximately 0.6% for bovine sera. Also, elution
conditions have been discussed in the literature as the gold standard for efficient elution,
boiling in 4% SDS and 25 mM DTT [10], requires the removal of SDS for some downstream
applications as e.g. 2DE. Therefore, broadly applicable alternative protocols for the reduction
of the protein concentration range in animal sera are highly desirable in the veterinary
medicine.
This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 4

Proteomics

Studying the bovine serum proteome we had observed that salt removal under certain
experimental conditions lead to the precipitation of a protein fraction (Fig. 1) exhibiting
interesting features. The precipitation yields, 4% of the input protein, were very stable
irrespective of how the salt was withdrawn experimentally, by dilution or dialysis. The
protein pattern after SDS-PAGE was very reproducible and strongly reminiscent of CPLLextracted material (Fig. 2A) with well-balanced protein concentrations over a wide range of
molecular weights (Fig.2, Supporting Information Fig. 1 ). The protein pattern of the
remaining supernatant was very similar to that of raw serum, suggesting that the precipitate is
enriched in less abundant serum proteins (Fig. 2A). Finally, under the same experimental
conditions precipitations with similar appearance also formed from human and porcine sera
(Supporting Information, Fig. 2).
To characterize this precipitate, we optimized precipitation conditions, identified the
constituting proteins by mass spectrometry, analyzed their physico-chemical properties and
determined the depletion of the most abundant serum components using SDS-PAGE, 2DE
and quantitative nLC-MALDI-TOF/TOF MS. For comparison, a protein fraction obtained by
extraction with a commercially available CPLL was analyzed in parallel.
Serum samples were collected from healthy cows by standard procedures as described
elsewhere [11]. Equalization was performed using ProteoMiner beads (BioRad) as described
in the Supporting Information Material and Methods file. This preparation is referred to as
equalized proteins.

Conditions for efficient precipitation were optimized with respect to pH and salinity (Fig. 1).
Variation of the pH at 1mM buffer concentration identified a pH of 4.2 to be optimal with
respect to precipitation yields and balanced protein concentrations (Figure 1A). At the
optimal pH, increasing buffer concentrations (Figure 1B) or addition of NaCl (Figure 1C)
gradually inhibited precipitation probably by a salting-in effect. For the following
This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 5

Proteomics

experiments, the precipitate was prepared by dialysis (SnakeSkin Dialysis Tubing, 7000 Da
molecular weight cutoff; Thermo Scientific) against 1 mM Na2HPO4/citric acid (NPC) pH
4.2 for 24 h at 4 C and recovered by centrifugation at 16,000 g for 10 min. It is referred to
simply as precipitate. Very similar results were obtained when the precipitation experiments
were repeated with human or porcine sera (Supporting Information, Fig. 2). Replacement of
NPC by other buffer systems that are useful at pH 4.2 produced virtually the same
precipitates (not shown).
Comparison of the protein profiles of equalized and precipitated proteins after SDS-PAGE
revealed marked differences in protein composition (Fig. 2A). For a more detailed view 150
g samples of raw serum, equalized proteins and precipitate were analyzed by 2DE. Pairwise
two-channel color overlay images of the 2D electrophoretic gels of raw, equalized and
precipitated proteins are shown in the Supporting Information, Fig. 3. A number of protein
spots were shared between both fractions but there also was remarkable complementarity.
Comparison with the 2DE of raw serum showed that albumin was efficiently depleted in both
preparations (Fig. 2B).
Reproducibility of precipitation was tested using a quantitative approach. Two independent
preparations of the precipitate from the same serum were trypsin digested, peptides labelled
with isotope-tagged dimethyl groups and 1:1 mixes analyzed by quantitative nLC-MALDITOF/TOF MS as described in the Supporting Information Material and Method file. The
experiment was carried out in triplicate using sera from three different animals. The same
experimental scheme was applied to three 1:1 mixtures each of differentially labelled digests
of raw serum and of equalized proteins. The resulting relative protein abundances are
presented as quantile plots in Fig. 3A. In comparison with raw serum both treatments
introduced minimal additional experimental variation. However, relative abundances of
>95% of the identified proteins were within a 1.5 fold range after equalization and
precipitation, indicating reproducibility of both treatments was very similar and satisfactory.

The proteins identified in these nLC-MALDI-TOF/TOF MS experiments were used to

confirm the differences in protein composition between the equalized material and the
precipitate that were observed in the electrophoretic analysis (Fig. 2A, Supporting
Information Fig. 3). Panels of proteins identified in the three different preparations (raw
serum, equalized or precipitated proteins) are listed in the Supporting Information Table 1
and were compared in the Venn diagram [12] in Figure 3B. We observed some overlap but
This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 6

Proteomics

also a substantial degree of complementarity that confirmed SDS-PAGE and 2DE and
indicated that equalization and precipitation may be used as complementary techniques for
serum analysis. Under the experimental conditions applied, a number of proteins were
reliably identified in the precipitate (three out of three replicates) that were not identified at
all in raw serum, e.g. coagulation factors V and X, complement factors H and I, protease
inhibitors like serpin A10, serpin F1, and inter-alpha-trypsin inhibitor heavy chain H3, and a
number of complement-related proteins like subcomponents r and s of complement
component 1, or the complement component 8 alpha and gamma chains.
The physicochemical properties of the proteins identified by nLC-MALDI-TOF/TOF MS in
the three replicates of the equalized and precipitated samples were analyzed. Particularly, the
grand averages of hydropathy (GRAVY) [13], molecular weights, and the isoelectric points
were calculated and compared to those of the bovine proteins currently annotated to the
extracellular space (GO-term GO:0005615) in the EMBL database [14], release 67
(Supporting Information Table 2), for details see the Supporting Information Material and
Methods file. Figure 3C shows the representation of experimentally identified proteins in the
two preparations over GRAVY, Mr and pI in comparison to the reference proteins. Over the
entire range of GRAVY values, distribution of experimentally identified proteins from both
preparations matched the expected values of the reference proteins indicating that neither
protocol introduced a hydrophobicity based selection bias. For both protocols, proteins with
Mr < 12,000 Da were found to be underrepresented, presumably as a consequence of the
dialysis step during sample preparation. Thus, both protocols may have to be adapted for
studies targeting low molecular weight serum proteins. This effect was slightly more
pronounced for the precipitated proteins. Profiles of both fractions over the pI were very
similar and indicated that proteins with pI > 8.0 were not adequately represented in neither
fraction. Most importantly, the profiles of both fractions were very similar showing that, in
comparison to equalization, precipitation did not introduce any notable bias preferring
proteins on basis of their physicochemical parameters, with the exception of a more
pronounced underrepresentation of smaller proteins.
Next, the depletion of individual highly abundant serum proteins by both treatments was
assessed. To this purpose tryptic digests of raw serum, equalized and precipitated proteins
were isotope labelled by reductive dimethylation [15] using conventional or deuterated
formaldehyde to introduce a 4 Da mass tag per dimethyl group. Peptides were mixed at 1:1
ratios and analyzed as described in the Supporting Information, Material and Methods. All
experiments were done in duplicate with reciprocal labeling using independent preparations
of one serum . For experimental details see the Material and Method file in the Supporting
Information.

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 7

Proteomics

Depletion or enrichment of proteins by precipitation was assessed in mixtures of isotope

labelled precipitates and raw serum preparations. Efficient depletion was found for some of
the most abundant serum proteins like albumin, immunoglobulins, apolipoprotein A-I,
serotransferrin, -2-HS-glycoprotein, or hemopexin, while a number of less abundant
proteins were enriched like e.g. some complement-related proteins. For the most efficiently
depleted or enriched proteins, detailed results are given in Supporting Information Table 4.
For direct comparison of the depletion efficiency of the most abundant proteins by
equalization and precipitation, 1:1 mixtures of both preparations were analyzed by
quantitative MS. In the following, abundance ratios of two experiments are given with ratios
below 1 indicating a more effective depletion by equalization than by precipitation and vice
versa. Only values within the measurable range of approximately 20-fold are specified.
Albumin was depleted with very similar efficiency (1.0/0.71) whereas immunoglobulins
(0.43/0.31), Alpha-2-macroglobulin (0.28/0.36), Complement C3 (0.39/0.42), and Interalpha-trypsin inhibitor heavy chain H4 (0.30/0.65) were removed more efficiently by
equalization. However, Apolipoprotein A-I (>20/>20), and also the less abundant
apolipoproteins A-II (>20/>20), A-IV (>20/14), and C-III (9.4/9.5) were highly enriched by
equalization. Depletion of Serotransferrin (2.8/2.7) by precipitation was more efficient than
by equalization.

From these quantitative experiments we conclude that precipitation efficiently depletes some
of the most abundant serum proteins while other less abundant proteins are enriched.
Precipitation was beneficial for MS analysis as substantially more proteins could be identified
from precipitations than from raw serum preparations. The enrichment/depletion
characteristics of precipitation and equalization markedly differed so that both techniques
may be favorably combined.

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 8

Proteomics

Finally, in-depth proteome analysis of the precipitate was carried out using a nLC-MALDITOF/TOF MS platform and a nLC-LTQ OrbitrapXL system. Before MS analysis, tryptic
peptides were fractionated by gel-free IEF to improve yields of identified proteins [11]. MS
data from 24 nLC runs (12 IEF fractions analyzed on each platform) were combined and a
joint list of identified proteins was compiled. For experimental details see the Supporting
Information Material and Methods file. In total, 306 proteins were identified on basis of a
minimum of 2 identified peptides per protein (Supporting Information Table 3). Gene
ontology (GO) analysis of the identified proteins using QickGO [16]
(http://www.ebi.ac.uk/QuickGO/GAnnotation) and Blast2GO software [17] showed that 184
(60.1%) identified proteins were annotated with GO-term GO:0005615 which is
recommended for plasma proteins. This is higher than we had observed in a recent study of
the bovine plasma proteome (41.7%) using equalized samples [11], indicating that numerous
typical serum proteins are efficiently precipitated under the described conditions.
In conclusion, we show that precipitation of serum proteins by salt withdrawal yields a
protein fraction which shows similarity to CPLL treated proteins. A number of highly
abundant serum proteins are efficiently depleted and thus protein identification by MS is
enhanced. Both treatments efficiently remove Albumin and Transferrin. Depletion and
enrichment of individual other highly abundant proteins is partially complementary. Protein
yields after precipitation are higher than after equalization making precipitation applicable in
cases where the serum samples are limited, e.g. when analyzing serum from smaller animals.
In-depth protein analysis showed that the precipitate is rich in proteins and the composition is
dominated by typical serum proteins. Precipitation is economic, rapid, and high-throughput
capable. No reagents, particularly detergents, are introduced that may be incompatible with
following analysis. We suggest precipitation as a species-unspecific protocol to reduce the
dynamic range of protein concentrations, not to replace but rather to complement CPLL
treatment in serum proteomic workflows. Also, we would like to point out that removal of
precipitations that occur under the specified conditions, e.g. when preparing serum samples
for ion exchange chromatography, may lead to the loss of this potentially valuable protein
fraction.
Acknowledgements
This study was supported by the Federal Ministry of Education and Research, Phenomics
Network of Excellence, Germany.
Conflict of interest statement
None of the authors of this paper has a financial or personal relationship with other people or
organizations that could inappropriately influence or bias the content of the paper.

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 9

Proteomics

References
[1] Anderson, N. L., Anderson, N. G., The human plasma proteome: history, character, and diagnostic
prospects. Molecular & cellular proteomics : MCP 2002, 1, 845-867.
[2] Mitchell, P., Proteomics retrenches. Nature biotechnology 2010, 28, 665-670.
[3] Tirumalai, R. S., Chan, K. C., Prieto, D. A., Issaq, H. J., et al., Characterization of the low molecular
weight human serum proteome. Molecular & cellular proteomics : MCP 2003, 2, 1096-1103.
[4] Travis, J., Pannell, R., Selective removal of albumin from plasma by affinity chromatography.
Clinica chimica acta; international journal of clinical chemistry 1973, 49, 49-52.
[5] Yadav, A. K., Bhardwaj, G., Basak, T., Kumar, D., et al., A systematic analysis of eluted fraction of
plasma post immunoaffinity depletion: implications in biomarker discovery. PLoS One 2011, 6,
e24442.
[6] Linke, T., Doraiswamy, S., Harrison, E. H., Rat plasma proteomics: effects of abundant protein
depletion on proteomic analysis. Journal of chromatography. B, Analytical technologies in the
biomedical and life sciences 2007, 849, 273-281.
[7] Bellei, E., Bergamini, S., Monari, E., Fantoni, L. I., et al., High-abundance proteins depletion for
serum proteomic analysis: concomitant removal of non-targeted proteins. Amino acids 2011, 40,
145-156.
[8] Thulasiraman, V., Lin, S., Gheorghiu, L., Lathrop, J., et al., Reduction of the concentration
difference of proteins in biological liquids using a library of combinatorial ligands. Electrophoresis
2005, 26, 3561-3571.
[9] Josic, D., Breen, L., Clifton, J., Gajdosik, M. S., et al., Separation of proteins from human plasma by
sample displacement chromatography in hydrophobic interaction mode. Electrophoresis 2012, 33,
1842-1849.
[10] Candiano, G., Dimuccio, V., Bruschi, M., Santucci, L., et al., Combinatorial peptide ligand libraries
for urine proteome analysis: investigation of different elution systems. Electrophoresis 2009, 30,
2405-2411.
[11] Henning, A. K., Groschup, M. H., Mettenleiter, T. C., Karger, A., Analysis of the bovine plasma
proteome by matrix-assisted laser desorption/ionisation time-of-flight tandem mass spectrometry.
Vet J 2014, 199, 175-180.
[12] Micallef, L., Rodgers, P., eulerAPE: drawing area-proportional 3-Venn diagrams using ellipses.
PLoS One 2014, 9, e101717.
[13] Kyte, J., Doolittle, R. F., A simple method for displaying the hydropathic character of a protein. J
Mol Biol 1982, 157, 105-132.
[14] Flicek, P., Ahmed, I., Amode, M. R., Barrell, D., et al., Ensembl 2013. Nucleic Acids Res 2013, 41,
D48-55.
[15] Boersema, P. J., Raijmakers, R., Lemeer, S., Mohammed, S., Heck, A. J., Multiplex peptide stable
isotope dimethyl labeling for quantitative proteomics. Nature protocols 2009, 4, 484-494.
[16] Binns, D., Dimmer, E., Huntley, R., Barrell, D., et al., QuickGO: a web-based tool for Gene
Ontology searching. Bioinformatics 2009, 25, 3045-3046.
[17] Conesa, A., Gotz, S., Garcia-Gomez, J. M., Terol, J., et al., Blast2GO: a universal tool for
annotation, visualization and analysis in functional genomics research. Bioinformatics 2005, 21,
3674-3676.
[18] Hortin, G. L., Sviridov, D., Anderson, N. L., High-abundance polypeptides of the human plasma
proteome comprising the top 4 logs of polypeptide abundance. Clin Chem 2008, 54, 1608-1616.

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 10

Proteomics

Fig. 1 Precipitation of serum proteins under variation of pH (A), buffer concentration (B) and
salinity (C). (A) Precipitates formed after dialysis in a wide pH range, but the most balanced
band intensities and the highest protein yields were observed after dialysis against 1 mM
NPC buffer at pH 4.2. Precipitation yields at pH 4.2 decreased with increasing buffer
concentrations (B) or addition of NaCl to the 1 mM NPC buffer, pH 4.2, indicating a saltingin effect. M: molecular weight marker.

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 11

Proteomics

Fig.2 Comparison of the protein patterns of raw (R), equalized (CPLL) and precipitated
(PR) serum after SDS-PAGE (A) and 2DE (B). In panel A, 10 g of the indicated fractions
or the supernatant of the precipitate (Sn) were analyzed. Note the similarity of raw serum
and supernatant of the precipitation indicating that the precipitate may be rich in lowabundance proteins. The asterisks mark serum albumin. Panel B shows the corresponding 2D
electrophoretic patterns of 150 g aliquots. M: molecular weight marker.

This article is protected by copyright. All rights reserved.

www.proteomics-journal.com

Page 12

Proteomics

Fig. 3 Comparison of equalized and precipitated proteins by nLC-MALDI TOF/TOF MS

analysis. Isotope-labelled peptide preparations of raw serum, equalized or precipitated
samples were mixed in 1:1 ratios and quantitated by MS. The distributions of the relative
abundances of identified proteins were calculated and are presented as quantile plots (A) over
the medians of the isotope ratios of the respective protein-specific peptides. Horizontal dotted
lines represent the 1, 5, 10, 90, 95, and 99 % quantiles respectively. Distributions of relative
abundances are centered around 1.0 and very narrow indicating a low degree of experimental
variation of raw, equalized (dotted) and precipitated (dashed) protein preparations. (B) Areaproportional Venn diagram representing the number of proteins reliably (three out of three
replicates) identified in raw, equalized, and precipitated serum. (C) To test for any bias
introduced by equalization (plain lines) or precipitation (dashed lines) with respect to
hydrophobicity (GRAVY), molecular weight (log10(MW)), or charge (pI), these parameters
of the identified proteins were compared with those calculated for bovine serum proteins
listed in the EMBL database. For details see the Supporting Information Material and
Methods file. Positive and negative values indicate over- and underrepresentation of the
identified proteins in the respective parameter range in comparison to the reference proteins.
Binary logarithms of the ratios are given.

This article is protected by copyright. All rights reserved.