602 / 658
2
Improving Throughput for Targeted Quantification Methods by Intelligent Acquisition
Overview
Automated remote multiplexed targeted protein quantification utilizing real-time
qual/quan processing for increased quantitative accuracy over large dynamic ranges.
Introduction
Targeted quantification has become a very popular technique to verify putative
biomarker candidates in large clinical cohorts of samples. These candidates are
usually generated following a biomarker discovery experiment or derived from a
biological hypothesis, for example, a pathway or biophysical interaction. These lists
are usually large, containing upwards of 100–1000 proteins spanning several orders
of magnitude dynamic concentration range. This presents analytical challenges for
conventional SRM assays both in terms of method development and throughput. We
propose using high-resolution, accurate-mass (HRAM) mass spectrometry (MS) and
MS/MS schemes in conjunction with validated spectral libraries for automated method
building, data acquisition, verification, and quantification in real-time using novel
acquisition schemes.
Methods
K562 colon carcinoma cells were grown in heavy and light media, collected and mixed
at different ratios to cover a 20-fold dynamic range. All samples were digested and
analyzed on a Thermo Scientific™ Q Exactive™ hybrid quadrupole-Orbitrap mass
spectrometer equipped with a nanospray ion source. Data was acquired in two steps
to simulate traditional workflows. Initial experiments employed unbiased data-
dependent MS/MS acquisition resulting in peptide/protein identification as well as
building of a spectral library. The spectral library contains relative retention time,
precursor charge state distribution, and product ion distributions, creating a unique
verification/quantification scheme. A highly multiplexed, targeted protein list was
created from the spectral library and used for automated data acquisition and
processing real time to facilitate changes to the acquisition scheme.
The scheme in Figure 1 describes the methodology in more detail. The first step is to
characterize the LCMS parameters using the Thermo Scientific™ Pierce™ PRTC
Mixture Kit. The next step is to build a list of proteins that we are interested in. This will
typically come from a pathway study or a discovery experiment. The next is to build a
spectral library for this list of proteins. This can be built via predictive algorithm or
empirical observations. This turns into a spectral library lookup table. The look-up
table includes the precursor
m/z
values for the defined charge state as well as the
expected retention time window, which are used to initiate product ion spectral
acquisition based on the presence of multiple precursor isotopes during the expected
elution window. Once the signal for multiple precursor isotopes surpasses the user-
defined intensity threshold, a higher-energy collision dissociation (HCD) spectrum is
acquired and immediately compared against the spectral library generating a dot-
product correlation coefficient to determine spectral overlap and to check if the
targeted peptide has been detected previously. If the calculated correlation coefficient
surpasses the user-defined acceptance value, HCD product ion spectra will continue
to be acquired across the elution profile. This is shown in Figure 2.
FIGURE 1. Strategy for large-scale targeted quantification based on high IQ
data acquisition scheme
LC-MS characterization using the PRTC
kit
to determine:
Scheduled retention time windows
Average chromatographic peak widths
Determine targeted
protein list:
Discovery experiments
Pathway determination
Functional groups
Build targeted acquisition methods from
Scheme
FIGURE 2. Pictorial re
targeted peptide quan
elution identification,
precursor and produc
analytical selectivity o
*
*
Measured Ion Intensity
Start time for
Trig
Thre
1.
Theoretical
Isotope
Experimenta
HR/AM MS
Spectrum
Results
Highly multiplexed targe
refinement prior to imple
straightforward based o
and corresponding
m/z
and quantitate the pepti
and acquisition windows
achieve robust quantific
development, we have c
analytically rigorous dis
contains both LC and M
methods requiring few r
To first test our methods
PTRC kit). Spectral libra
performed on the quadr
ion collection and detect
perform state-model dat
shows the CV distributio
area of top eight produc