Accurate quantification of protein content and composition has been achieved using isotope-edited surface enhanced resonance Raman spectroscopy. Synthesis of isotopomeric Rhodamine dye-linked bioconjugation reagents enabled direct labeling of surface lysines on a variety of proteins. When separated in polyacrylamide gels and stained with silver nanoparticles. The spectral signatures reflect the expected statistical distribution of isotopomeric labels on the labeled proteins in the gel matrix format without interference from protein features.
Accurate quantification of protein content and composition has been achieved using isotope-edited surface enhanced resonance Raman spectroscopy. Synthesis of isotopomeric Rhodamine dye-linked bioconjugation reagents enabled direct labeling of surface lysines on a variety of proteins. When separated in polyacrylamide gels and stained with silver nanoparticles. The spectral signatures reflect the expected statistical distribution of isotopomeric labels on the labeled proteins in the gel matrix format without interference from protein features.
Accurate and precise protein quantification remains an important
fundamental methodology and represents a critical component in identifying
biomarkers and early detection of diseases.(1) The capacity to quantify relative changes in proteins from complex
biological samples is a central challenge to understanding the response
of protein expression to genomic or environmental factors. Surface
Enhanced Resonance Raman Spectroscopy (SERRS) has proven to be a useful
spectroscopic tool for diagnostic applications,[2−5] explosive detection,(6) monitoring events
in enzyme catalysis,(7) detection of metabolites,(8) and detection of bioanalytes during capillary
electrophoresis.(9) However, SERRS suffers
from variability in enhancement of Raman intensity depending upon
the nanomorphology of the substrate (typically silver colloids), thereby
affecting the reproducibility of the measurement. In our previous
work, we have demonstrated the improved reproducibility and accuracy
of quantitative SERRS measurements obtained using isotopomeric rhodamine
6G (R6G) probes over a useful analyte concentration range (picomolar
to micromolar) of analytes.(10) Herein, we
demonstrate for the first time that these dyes may be transformed
into biomolecular labeling reagents and used for SERRS based quantification
of labeled proteins in a polyacrylamide matrix.The sensitivity of SERRS can rival that of fluorescence, a standard
for optical detection and quantification of antibodies and nucleic
acids.(11) Advantages of SERRS include the
narrow band widths and chemical fingerprint information of Raman spectral
lines, as well as the improved quantification capability of the isotopic
SERRS approach.(10) The SERRS of suitable
dye probes such as R6G adsorbed on roughened metallic nanostructures
dominates over other Raman scattering processes and enables the detection
of dilute analytes in mixtures.(12) These
chemical features provide suitable enhancements of the Raman signatures
for the dye to allow detection in the background of matrix immobilized
protein. By using isotope-encoded dyes for internal standardization
of SERRS one should be able to quantify protein concentration with
unprecedented accuracy.In order to establish the above strategy by incorporating isotope-encoded
Raman labels into proteins, a method for synthesis of modified R6G
was pursued through modification of the carboxylic functional group
(Scheme 1). Previous efforts showed that the
tetra-deuterio-R6G based upon 1A provided reliable differences in
SERRS signature and enabled quantification.[10,13] Using this pattern of hydrogen substitution, the precursor d4-R6G-acid
(1B) was obtained from condensation of 3-ethylamino 4-methylphenol
with d4-phthalic acid. R6G was hydrolyzed under basic conditions to
obtain the corresponding acid (1A). Installation of a side chain linker
was developed to enable efficient protein conjugation. The xanthene
chromophore has a propensity for cyclization under basic conditions.(14) To ensure retention of the xanthene chromophore,
a tertiary substituted amide linkage was selected to prevent cyclization
at the C7 position under basic conditions required for bioconjugation
reactions. A linker was derived by basic hydrolysis of N-methyl caprolactam and subsequent esterification of the terminal
carboxyl moiety. Amidation of 1A (and 1B) with the linker was followed
by ester interconversion by hydrolysis to the free acids and activation
to the corresponding NHS esters (2A and 2B). Since lysine tagging
is typically done at pH 8.5, we established that the dye linker was
stable in aqueous buffer at pH 9.0 for 2 d at room temperature.
Scheme 1
Reagents and conditions:
(a) HATU/DIPEA/NHMe(CH2)5CO2Me/DMF,
65%; (b) Ba(OH)2/MeOH−H2O, 45%; (c) (iii)
HATU/DIPEA/NHS/DMF, 80%. (NHS = N-hyrdoxysuccinimide
ester). The sites of deuterium are noted with an asterisk.
Reagents and conditions:
(a) HATU/DIPEA/NHMe(CH2)5CO2Me/DMF,
65%; (b) Ba(OH)2/MeOH−H2O, 45%; (c) (iii)
HATU/DIPEA/NHS/DMF, 80%. (NHS = N-hyrdoxysuccinimide
ester). The sites of deuterium are noted with an asterisk.To test the feasibility of the bioconjugation approach, protein
samples of lysozyme, glycogen phosphorylase, cytochrome c, casein,
and ferritin were labeled with the isotopomeric R6G probes and subsequently
quantified by SERRS detection: The labeling efficiency for reagents
2A and 2B was first tested using UV−vis spectra of the final
dye-labeled samples to standardize the protein and R6G content by
the relative absorbance’s at 280 and 540 nm, respectively (Supporting Information). Protein labeling on the
200 nmol dye per mg of protein scale gave a labeling efficiency of
approximately 5 dye molecules per protein chain but ranged from 0.5
to 10 for different proteins tested and depended upon lysine content
of the protein. Lysozyme labeling was further characterized by ESI
LC/MS and found to contain one (23% in the d0 sample or 29% in the
d4 sample) or two labels (13% d0 or 12% d4).Examples of the SERRS spectra of the d0/d4 labeled proteins are
shown for lysozyme and cytochrome c in Figure 1. In all cases, the R6G-labeled protein samples
showed identical SERRS spectra to those of the parent R6G labeling
reagents. The critical distinctions in the SERRS spectra of the isotopomeric
R6G-labeled proteins were also apparent. The Raman fine structure
is dominated by the features of the R6G. These colloidal suspensions
displayed a limit of detection at an approximately 300 pM protein
concentration, which was similar to the level of detection of R6G
dye alone. Despite the complexity of surface interactions, the reproducible
and predicted SERRS signatures in colloid suspensions with isotopomeric
mixtures of the R6G-labeled protein samples established a statistical
representation of the dye populations. No features of the protein
were observed in the Raman spectra, further supporting the dominant
role of the dye−nanoparticle exchange interactions as a mechanism
for the enhancement.(13)
Figure 1
SERRS
spectra of (left) 3 nM lysozyme-R6G-d0 (solid) and lysozyme-R6G-d4
(dashed) and (right) cytochrome C-R6G-d0 (solid) and cytochrome-C-R6G-d4
(dashed). The inset calibration curves show the relative amount of
d4 labeled protein determined from the ratio of peak areas of the
d0 (611 cm−1) peak and the d4 (600 cm−1) peak compared to the known mixture ratio with 95% C.I. error bars.
The solid line in the insets represents an ideal slope = 1.
SERRS
spectra of (left) 3 nM lysozyme-R6G-d0 (solid) and lysozyme-R6G-d4
(dashed) and (right) cytochrome C-R6G-d0 (solid) and cytochrome-C-R6G-d4
(dashed). The inset calibration curves show the relative amount of
d4 labeled protein determined from the ratio of peak areas of the
d0 (611 cm−1) peak and the d4 (600 cm−1) peak compared to the known mixture ratio with 95% C.I. error bars.
The solid line in the insets represents an ideal slope = 1.Polyacrylamide gel based protein separation and quantification
play a central role in protein analysis, proteomics, and diagnostics.
Spectral detection methods for proteins in gels have been limited
to fluorescence. No prior studies have focused on using SERRS detection
in gels or applications of isotopic variants for quantitative measurements
in such a matrix. To obtain SERRS in a gel matrix, in situ formation of silver nanoparticles using a protein silver staining
protocol was pursued.[15,16] Ferritin samples labeled with
d0- or d4-R6G were mixed in varied ratios and separated on SDS PAGE
(Figure 2). The protein bands were imaged first
by fluorescence, followed by silver staining and SERRS spectral imaging
again using a 514.5 nm laser. The spectral shifts at 600 cm−1 were used to calculate the percent isotopic compositions for d0
and d4 R6G corresponding to the relative amounts of the two ferritin
samples. The SERRS spectral signature reflects that of the isotopomeric
label as shown (Figure 2E). The expected statistical
distribution of the two labeled proteins in the PAGE gel format was
observed.
Figure 2
SDS-PAGE
gel analysis of 0, 25, 50, 75, and 100% d4-R6G mixtures of isotopomeric
R6G-labeled ferritin, loaded with a total of 20 pmol of dye or 0.7
μg protein per lane (see Supporting Information for details of detection). Shown here is the 20 kDa protein band
imaged by silver stain (A), fluorescence (B), and SERRS intensity
normalized to the most intense signal in the image of the boxed region
in A and B (C). (D) A color coded representation of the percent of
R6G-d4 label (blue corresponds to 100% R6G-d4 label; red corresponds
to 100% R6G-d0-label). (E) Fluorescence baseline subtracted SERRS-spectra
of the five mixtures. The inset of E shows the percent d4 determined
by least-squares fitting of spectra from the ratio lanes. The error
bars represent the 95% C.I. for five measurements taken from each
ratio lane.
SDS-PAGE
gel analysis of 0, 25, 50, 75, and 100% d4-R6G mixtures of isotopomeric
R6G-labeled ferritin, loaded with a total of 20 pmol of dye or 0.7
μg protein per lane (see Supporting Information for details of detection). Shown here is the 20 kDa protein band
imaged by silver stain (A), fluorescence (B), and SERRS intensity
normalized to the most intense signal in the image of the boxed region
in A and B (C). (D) A color coded representation of the percent of
R6G-d4 label (blue corresponds to 100% R6G-d4 label; red corresponds
to 100% R6G-d0-label). (E) Fluorescence baseline subtracted SERRS-spectra
of the five mixtures. The inset of E shows the percent d4 determined
by least-squares fitting of spectra from the ratio lanes. The error
bars represent the 95% C.I. for five measurements taken from each
ratio lane.Standardization of SERRS measurements by use of isotopomers of
the same dye class overcomes many limitations of surface enhanced
spectroscopy.(10) Demonstrated here is the
feasibility of a bioconjugation approach for quantification of proteins
by SERRS detection enabled by isotopic substitution of a Raman active
label. Protein Raman signals were not observed in the surface enhanced
R6G spectra presumably due to larger SERRS enhancements for the dye−surface
interactions. Equally important is the fact that spectra in the immobilized
labeled protein accurately reflect the relative molecular content.
Formation of an SERRS active silver nanoparticle enhancement agent
within a polymeric matrix offers a practical approach to biomolecular
detection after capture or separation from biological mixtures. Optimization
of the methods to enhance SERRS intensity or minimize fluorescence
or other backgrounds offers additional promise.Future applications of isotopic SERRS-based quantification include
comparative proteomics, which would take advantage of labeling protein
mixtures from a distinct origin with different isotopomer dye reagents.(17) Additional quantitative biomolecular analyses
such as antibody−antigen complexes in semipermeable matrices
also represent promising applications for this detection strategy.
Highest accuracy measurements are expected when the isotopic ratios
are near unity, but preliminary results indicate the method may be
applicable with ratios up to a factor of 10.
Authors: D Cunningham; R E Littleford; W E Smith; P J Lundahl; I Khan; D W McComb; D Graham; N Laforest Journal: Faraday Discuss Date: 2006 Impact factor: 4.008
Authors: Giselle M Knudsen; Brandon M Davis; Shirshendu K Deb; Yvette Loethen; Ravindra Gudihal; Pradeep Perera; Dor Ben-Amotz; V Jo Davisson Journal: Bioconjug Chem Date: 2008-11-19 Impact factor: 4.774