BBRC

Biochemical and Biophysical Research Communications 297 (2002) 401405

www.academicpress.com

Comparative phenotypic analyses of human plasma and urinary

retinol binding protein using mass spectrometric immunoassay
Urban A. Kiernan, Kemmons A. Tubbs, Dobrin Nedelkov,
Eric E. Niederkoer, and Randall W. Nelson*
Intrinsic Bioprobes, Inc., 625 South Smith Road, Suite 22, Tempe, AZ 85282, USA
Received 13 August 2002

Abstract
Mass spectrometric immunoassay (MSIA) is a proteomics technology that combines the selectivity of anity capture with the
sensitivity and resolution of mass spectrometric detection. This unique approach allows for intact protein identication therefore is
readily capable of discriminating between protein variants, i.e., mutations, posttranslational modications, and truncations. In this
work, MSIA is used in the comparative analyses of retinol binding protein (RBP) from the plasma and urine of a small study
population. Detailed RBP proles were obtained from both biological uids, resulting in the identication of several catabolic RBP
products (present in urine) that have not been previously reported. In addition, comparative analysis of urine samples taken from
healthy and renally impaired individuals revealed dierent breakdown proles. These results illustrate the use of MSIA for the rapid,
sensitive, and accurate proling of RBP both within and between individuals. 2002 Elsevier Science (USA). All rights reserved.
Keywords: MALDI-TOF MS; Humans; Posttranslational modications; Retinol binding protein; Urine protein

Plasma and urine are viewed as signicant biological

uids for protein related studies due to their relative ease
of collection (non- or minimally invasive) and diverse
protein composition. Historically, the detection of uidborne proteins has been performed predominantly by
classical immunoassay techniques [1], which are capable
of selectively targeting a specic protein marker for
analysis. However, even though these techniques are
ideal for quantication and possess very low detection
limits; without the use of mono-specic antibodies, the
classical immunoassay techniques lack the ability to
discriminate between dierent structural variants of a
target protein. Thus, each structural variant of the target
protein requires its own separate antibody-based assay,
which can be costly when multiple structural variants
are under investigation. Moreover, the use of the specic
antibody approach leaves no room for the discovery of
other variants (that will not be recognized by the monospecic antibodies).

Corresponding author. Fax: 1-480-804-0778.

E-mail address: rnelson@intrinsicbio.com (R.W. Nelson).

Lately, proteomics approaches have been applied in

the analysis of bio-uids and are capable of producing
protein expression proles for use in comparative studies [25]. These methods excel in the global identication
of gene products present in biological uids and extracts, and when coupled with ICAT [6] technologies are
capable of the semi-quantitative proling of multiple
proteins in a single analysis. The approaches, however,
are generally hindered by slow analysis times (detracting
from high-throughput analysis), relatively poor sensitivities (compared to the immunoassay techniques) and,
importantly, the inability to readily give a full-length
read of a specic protein, which is ultimately needed to
distinguish structural variants of the protein (e.g., point
mutations, posttranslational modications or catabolism products) as present in multiple individuals.
Previously, we have introduced a hybrid proteomics
approach termed mass spectrometric immunoassay
(MSIA) [7], which more recently has been scaled for the
high-throughput analysis of target plasma proteins in
multiple individuals [810]. MSIA uses immuno-anity
recognition, by selectively concentrating proteins via
anity-pipettor tips, combined with matrix assisted laser

0006-291X/02/$ - see front matter 2002 Elsevier Science (USA). All rights reserved.
PII: S 0 0 0 6 - 2 9 1 X ( 0 2 ) 0 2 2 1 2 - X

402

U.A. Kiernan et al. / Biochemical and Biophysical Research Communications 297 (2002) 401405

desorption/ionization time-of-ight mass spectrometry

(MALDI-TOF MS) detection (see Fig. 1). This combination allows for the rapid purication/concentration
and mass spectrometric analysis of specic proteins
present in biological uids. Moreover, the approach
yields a detailed structural analysis of full-length proteins with sensitivities of the order of those found in
immunoassay approaches. This ability allows for the
discrimination of mass-shifted structural variants (retrieved by a pan-antibody) that would otherwise be
viewed as a single protein using conventional immunoassay, or possibly missed using conventional proteomics
approaches.
More recently, we have used MSIA to dierentiate
between variants of transthyretin and serum amyloid P
component present in the plasma of dierent individu-

als. During these studies, point mutations and alterations in glycosylation were observed between individuals
and subsequently characterized using mass spectrometry
[9,10]. Moreover, MSIA has been applied to the quantitative measurement of urine- and plasma-borne b-2microglobulin (b2m) [11]. Again, variants of the target
protein, in particular glycated-b2m, were observed
during the routine course of the analyses. Thus, it is
apparent the MSIA can be employed in the analysis/
discovery of protein variants present in the urine or
plasma of individuals.
Here, we present the development of urinary and
plasma retinol binding protein (RBP) mass spectrometric immunoassays for the rapid qualitative proling
of RBP isolated from multiple individuals. In previous
work, we have used a complementary technique, biomolecular interaction mass spectrometry (BIA/MS), in
the study of urine-borne RBP. These studies, although
promising, were unable to fully resolve urine-borne RBP
variants with the condence needed for unambiguous
identication [12]. In these studies presented here, which
achieve a higher level of performance, MSIA was used
to qualitatively compare RBP isolated from both the
plasma and urine of individuals within a small study
population. Samples from four healthy controls and one
individual suering from chronic pyelonephritis (stemming from diabetes mellitus) and renal failure were
donated for use in this study. The rst objective of the
study was to verify that no genetic or transcriptional
variants, which could potentially inuence urinary RBP
proles, existed between individuals. This object was
accomplished by analyzing RBP from the plasma of the
individuals. The subsequent objective of the study was
to use MSIA to prole urinary RBP for possible differences related to the (renal) health state of the individual.

Materials and methods

Fig. 1. Schematic of MSIA process. Biological uid (i.e., urine, plasma)

is repetitively pipetted through the MSIA-Tip derivatized with an afnity ligand for a specic protein target. The target protein is retrieved
and concentrated within the MSIA-Tip, which is then puried with the
application of multiple rinse steps to remove non-specically bound
material. Retained protein is eluted from the MSIA-Tip with MALDI
matrix and applied directly onto a MALDI hydrophilic/hydrophobic
contrasted target. Air-dried samples are then interrogated with MADLI-TOF MS.

Study subjects. Plasma and urine samples were donated by four

healthy male subjects, ages ranging from 2649, and from one elderly
female suering from chronic pyelonephritis and diabetes mellitus,
aged 94. All subjects participated in the study under an informed
consent form approved under IRB00001399, which is in compliance
with FWA00000559. Appropriate safeguards were used in collecting
samples using universal precautions.
Sample preparation. Plasma samples were collected from ve individuals and processed for immediate analysis as follows. Whole blood
(50 lL per individual) was acquired under sterile conditions through a
lancet punctured nger using a 50 lL microcolumn (Drummond Scientic, Broomall, PA).
Each whole blood sample was immediately combined with 200 lL
Hepes buered saline (10 mM Hepes, 150 mM NaCl, 3 mM EDTA,
and 0.005% polysorbate 20 (v/v), pH 7.4 (HBS)) containing 2 lL
protease inhibitor cocktail [AEBSF (100 mM); aprotin (80 lM); bestatin (5 mM); E-64 (1.5 mM); leupeptin (2 mM); and pepstatin A
(1 mM)added to prevent any enzymatic breakdown] and centrifuged
for 2 min (at 7000 rpm) to pellet red blood cells. The supernatant

U.A. Kiernan et al. / Biochemical and Biophysical Research Communications 297 (2002) 401405
(diluted plasma) was decanted from each sample and interrogated
using MSIA.
The ve plasma samples were addressed in parallel using an eightbarrel repeating pipettor outtted with anti-RBP MSIA-Tips (Intrinsic
Bioprobes). Sample incubation consisted of repeatedly cycling (aspiration and dispense; 50 cycles; 3 s/cycle; 150 lL/cycle) the samples
through individual anity tips. After incubation, tips were rinsed using
HBS (10 cycles, 150 lL), doubly distilled water (5 cycles, 150 lL), and
20% acetonitrile/1 M ammonium acetate wash (10 cycles, 150 lL), and
nally with doubly distilled water (15 cycles, 150 lL). Retained species
were eluted by drawing 4 lL MALDI matrix solution (saturated
aqueous solution of sinapic acid (SA) in 33% (v/v) acetonitrile and
0.4% (v/v) triuoroacetic acid) into each tip and depositing directly
onto a 96-well formatted hydrophobic/hydrophilic contrasting MALDI-TOF target [8]. Samples were allowed to air dry, prior to insertion
into the mass spectrometer. The total time required for the preparation
of the ve plasma samples was approximately 10 min.
Urine samples, 25 mL mid-stream voids, from the same ve individuals were also collected. The urine samples were immediately
combined 1:1 (v/v) with 2 M ammonium acetate (to adjust the pH to

6.87.2) and 50 lL protease inhibitor cocktail described above. Due
to their larger volume, the urine samples were interrogated individually
with anti-RBP MSIA-Tips. Incubation of each urine sample consisted
of 300 cycles (150 lL of sample) through a MSIA-Tip. After incubation, tips were treated using the same rinse/elution protocols described
for the plasma analyses. Because of the larger sample volumes of urine,
time spent in sample preparation was
20 min/sample. MALDI-TOF
MS and data analysis were performed on all samples as described
previously [9].

Results and discussion

RBP is a member of the lipocalin family and serves as
the primary carrier of retinol (vitamin A) from the liver
to peripheral tissues [13]. This small hepatic protein,
molecular weight 21,065.6, is believed to escape glomerular ltration by non-covalently complexing, in its
holo-(retinol bound) form, with transthyretin [14]
whereas free RBP is known to pass through the
glomeruli where it is then believed to be catabolized in
the tubular cells [15]. Previous studies have shown that
individuals with renal diseases known to impair ltering
function have an increased concentration of plasma
RBP [16]. Moreover, the presence of truncated RBP
variants, by the loss of one or both C-terminal leucines,
may also be markers of renal failure by their increased
plasma concentrations, resulting in varied RBP-totruncated variant ratios [17].
Fig. 2 displays the results of the plasma anti-RBP
MSIA of the ve individuals involved in this study. The
purpose of the plasma RBP analysis was mainly that of
establishing homogeneity of the protein species between
individuals, i.e., screening for the presence of point
mutations, or other overt structural modications (none
of which were observed). The mass spectra show a high
level of reproducibility with wt-RBP present in all
spectra. With external calibration, all wt-RBP signals
were within 0.1% of the known molecular mass of RBP
Fig. 2 (traces AD) are from healthy individuals and
display very similar RBP phenotypic proles. Each mass

403

Fig. 2. Results of the anti-RBP MSIA analysis of plasma. Traces AD

are from four healthy individuals and display a similar plasma retinol
binding protein prole, which consists of wt-RBP (MW 21,065) and
small amounts of RBP (-L; C-terminal) (MW 20,917). Trace E is
from an individual who suers from dysfunctional kidneys. Increased
signal intensity of RBP (-L), relative to wt-RBP, is observed and even
trace amounts of RBP (-LL; C-terminal) (MW 20,804) are detected.

spectrum is dominated by the wt-RBP, but also shows a

trace amount of a satellite protein at
113 Da lower in
molecular mass. This satellite signal has been identied
as the RBP-L (C-terminal) variant. Fig. 2 (trace E) is of
RBP isolated from the individual suering from renal
failure. Although wild-type RBP is still present, the
prole contains a noticeably larger contribution of RBPL. Additionally, a second lower intensity satellite signal
is observed (at dm 226 Da), which is identied as
RBP-LL (C-terminal). The detection of increased levels
of RBP truncated variants relative to wt-RBP in the
plasma of the renal failure patient is consistent with the
previous report of Jaconi et al. [17].
The results of the anti-RBP MSIA urine analyses are
shown in Fig. 3. Fig. 3 (traces AD) are from the same
healthy individuals used to generate the plasma RBP
data shown in Fig. 2 (traces AD). All four spectra
display a generally consistent phenotypic prole of urinary RBP. Similar to the plasma proles, wt-RBP and
RBP-L were common to the urine of all four healthy
subject samples. However, additional signals are observed that correspond to further C-terminally truncated
variants, namely RBP-LL, RBP-RNLL, and RBPRSERNLL. To date, these urinary RBP truncations
have not been reported. In contrast, the prole obtained
from the urine of the individual suering renal dysfunction is less complex (Fig. 3 trace E). The prole
noticeably lacks the presence of wt-RBP and is dominated by the presence of RBP-L. In addition, a small
amount of the RBP-RNLL is also observed, indicating
that this newly identied variant is still produced in
detectable amounts in the urine of individuals with
dysfunctional kidneys.

404

U.A. Kiernan et al. / Biochemical and Biophysical Research Communications 297 (2002) 401405

proles of individuals with healthy and diseased kidneys

were also observed. Therefore, a more concerted study
may be able to provide further insight into the catabolism of RBP, or at the very least verify proles capable
of serving as diagnostic indicators of renal function.
Thus, this preliminary demonstration of MSIA, as applied to RBP present in plasma and urine, serves as a
stepping-stone for the study of larger populations to
clinically dene normal and abnormal phenotypic
RBP proles and assist in determining the mechanistic
cause of dierences.

Acknowledgments
Fig. 3. Results of the anti-RBP MSIA analysis of urine. Traces AD
are from healthy study subjects and display a generally similar phenotypic prole. Wild-type RBP, RBP (-L), RBP (-LL), and RBP
(-RNLL) (MW 20,534) are present in all four mass spectra. Observable signal from RBP (-RSERNLL) (MW 20,162) is also present
in three of the four spectra. However, trace E is from the individual
with kidney impairment and mass spectrum is dominated by RBP-L
signal and devoid of wt-RBP.

Interestingly, repeating the analyses on urine sample

left at native conditions (no addition of pH buers or
protease inhibitors) and incubated at 37 C for 24 h did
not show any increased amount of RBP truncation (data
not shown). These observations indicate that the production of these C-terminally truncated forms of RBP is
not the result of proteolysis occurring in the urine. Thus,
the constant observation of C-terminally truncated
variants in urine suggests a catabolism mechanism involving exopeptidase truncation, prior to clearance into
the bladder (independent of health state). However, the
observation of two dierent truncation patterns further
suggests the involvement of multiple exopeptidase,
which may be dierentially regulated, dependent on the
health state of the individual.

Conclusion
This preliminary study demonstrates that mass
spectrometric immunoassay is a relatively easy and rapid method for isolating and detecting human retinol
binding protein from both plasma and urine. The use of
mass spectrometry for detection allows for the identication of variant forms of RBP that would possibly have
been overlooked by other methods. Even though only a
small study population was used in this study, dierences in RBP phenotypic proles from both plasma and
urine were readily observed. The sensitivity and resolution of MALDI-TOF MS detection were able to identify
RBP-RNLL and RBP-RSERNLL, two C-terminal
truncated urinary variants that have not been reported
previously. Overt dierences in the RBP phenotypic

This publication was supported in part by Grant No. R44

GM56603-01 and Contract No. N43-DK-1-2470 from the National
Institutes of Health. Its contents are solely the responsibility of the
authors and do not necessarily represent the ocial views of the National Institute of Health.

References
[1] D. Wild, The Immunoassay Handbook, second ed., Nature Pub.
Group, New York, 2001.
[2] C.S. Spahr, M.T. Davis, M.D. McGinley, J.H. Robinson, E.J.
Bures, J. Beierle, J. Mort, P.L. Courchesne, K. Chen, R.C. Wahl,
W. Yu, R. Luethy, S.D. Patterson, Towards dening the urinary
proteome using liquid chromatographytandem mass spectrometry. I. Proling an unfractionated tryptic digest, Proteomics 1
(2001) 93107.
[3] M.T. Davis, C.S. Spahr, M.D. McGinley, J.H. Robinson, E.J.
Bures, J. Beierle, J. Mort, W. Yu, R. Luethy, S.D. Patterson,
Towards dening the urinary proteome using liquid chromatographytandem mass spectrometry. II. Limitations of complex
mixture analyses, Proteomics 1 (2001) 108117.
[4] D.J. Hampel, C. Sansome, M. Sha, S. Brodsky, W.E. Lawson,
M.S. Goligorsky, Toward proteomics in uroscopy: urinary protein
proles after radiocontrast medium administration, J. Am. Soc.
Nephrol. 12 (2001) 10261035.
[5] E.F. Petricoin, A.M. Ardekani, B.A. Hitt, P.J. Levine, V.A.
Fusaro, S.M. Steinberg, G.B. Mills, C. Simone, D.A. Fishman,
E.C. Kohn, L.A. Liotta, Use of proteomic patterns in serum to
identify ovarian cancer, Lancet 359 (2002) 572577.
[6] S.P. Gygi, B. Rist, S.A. Gerber, F. Turecek, M.H. Gelb, R.
Aebersold, Quantitative analysis of complex protein mixtures
using isotope-coded anity tags, Nat. Biotechnol. 17 (1999) 994
999.
[7] R.W. Nelson, J.R. Krone, A.L. Bieber, P. Williams, Massspectrometric immunoassay, Anal. Chem. 67 (1995) 11531158.
[8] E.E. Niederkoer, K.A. Tubbs, K. Gruber, D. Nedelkov, U.A.
Kiernan, P. Williams, R.W. Nelson, Determination of b-2
microglobulin levels in plasma using a high-throughput mass
spectrometric immunoassay system, Anal. Chem. 73 (2001) 3294
3299.
[9] U.A. Kiernan, D. Nedelkov, K.A. Tubbs, E.E. Niederkoer,
R.W. Nelson, High-throughput analysis of human plasma proteins, Am. Biotech. Lab. 20 (2002) 2628.
[10] U.A. Kiernan, K.A. Tubbs, K. Gruber, D. Nedelkov, E.E.
Niederkoer, P. Williams, R.W. Nelson, High-throughput protein
characterization using mass spectrometric immunoassay, Anal.
Biochem. 301 (2002) 4956.

U.A. Kiernan et al. / Biochemical and Biophysical Research Communications 297 (2002) 401405
[11] K.A. Tubbs, D. Nedelkov, R.W. Nelson, Detection and quantication of b-2-microglobulin using mass spectrometric immunoassay (MSIA), Anal. Biochem. 289 (2000) 2635.
[12] D. Nedelkov, R.W. Nelson, Analysis of human urine protein
biomarkers via biomolecular interaction analysis mass spectrometry, Am. J. Kidney Dis. 38 (2001) 481487.
[13] M. Kanai, A. Raz, D.S. Goodman, Retinol-binding protein: the
transport protein for vitamin A in human plasma, J. Clin. Invest.
47 (1968) 20252044.
[14] H.M. Naylor, M.E. Newcomer, The structure of human retinolbinding protein (RBP) with its carrier protein transthyretin reveals

405

an interaction with the carboxy terminus of RBP, Biochemistry 38

(1999) 26472653.
[15] D.S. Goodman, Plasma retinol-binding protein, Ann. NY Acad.
Sci. 348 (1980) 378390.
[16] R. Beetham, A. Dawnay, J. Landon, W.R. Cattell, A radioimmunoassay for retinol-binding protein in serum and urine, Clin.
Chem. 31 (1985) 13641367.
[17] S. Jaconi, K. Rose, G.J. Hughes, J.H. Saurat, G. Siegenthaler,
Characterization of two posttranslationally processed forms of
human serum retinol-binding protein: altered ratios in chronic
renal failure, J. Lipid Res. 36 (1995) 12471253.