Commercial drug norfloxacin as a novel ligand for the copper-catalyzed N-arylation of imidazole with aryl halides

Article abstract of DOI:10.1002/aoc.5195

Norfloxacin was used as an efficient ligand for the CuBr-catalyzed C-N coupling reaction of imidazole and aryl halides. The protocol presented good functional group compatibility, permitting many aryl halides to react with imidazole to form the desired products in good to excellent yields.

Full text of DOI:10.1002/aoc.5195

Technical Note
pubs.acs.org/ac
Cite This: Anal. Chem. 2019, 91, 10395−10400
Step-Wise Assessment and Optimization of Sample Handling
Recovery Yield for Nanoproteomic Analysis of 1000 Mammalian
Cells
Ruilin Wu,†,‡ Sansi Xing,†,‡ Maryam Badv,£ Tohid F. Didar,*,§,£,Ψ and Yu Lu*,†,‡
‡
£
†Department of Biochemistry and Biomedical Sciences, Stem Cell and Cancer Research Institute, School of Biomedical
Engineering, §Department of Mechanical Engineering, and ΨInstitute for Infectious Disease Research (IIDR), McMaster University,
Hamilton, ON L8S 4K1, Canada
S
* Supporting Information
ABSTRACT: Protein and peptide adhesion is a major factor
contributing to sample loss during proteomic sample preparation
workflows. Sample loss often has detrimental effects on the
quality of proteomic analysis by compromising protein identi-
fication and data reproducibility. When starting with a low sample
amount, only the most abundant proteins can be identified, which
often offers little insights for biological research. Although the
general idea about severe sample loss from low amount of starting
material is widely presumed in the proteomics field, quantitative
assessment on the impact of sample loss has been poorly
investigated. In the present study, we have quantitatively assessed
sample loss during each step of a conventional in-solution sample
preparation workflow using bicinchoninic acid (BCA) and
targeted LC/MS/MS protein and peptide assays. According to
our assessment, for starting materials of ∼1000 mammalian cells, surface adhesion, along with desalting and speed-vacuum
drying steps, all contribute heavily to sample loss, in particular for low-abundance proteins. With this knowledge, we have
adapted slippery liquid infused porous surface (SLIPS) treatment, commercial LoBind tubes, and in-line desalting during
sample processing. With these improvements, we were able to use a conventional in-solution sample handling method to
identify on average 829 proteins with 1000 U2OS osteosarcoma cells (∼100 ng) with 75-min LC/MS/MS runs, an 11-fold
increase in protein identification. Our optimized in-solution workflow is straightforward and also much less equipment- and
technique-demanding than other advanced sample preparation protocols in the field.
There is a widely accepted concept that the quality of
proteomic data is only as good as the quality of injected
samples. For both deep proteome coverage and nano-
proteomics, sample preparation is undoubtedly one of the
crucial factors to consider. Aside from conventional sample
preparation techniques,7 some of the effective strategies to
enhance protein identification include filter-aided sample
preparation (FASP),8 single-pot solid-phase enhanced sample
preparation (SP3),9 and StageTip.10 Focusing on nano-
proteomics, the most recent technological advances involving
nanodroplet processing11 and microfluidic chips12 allow for
proteomic analysis of 100 or fewer mammalian cells. Although
these methods have shown improved protein identification,
each has its own limitations. FASP has been criticized for its
sample recovery and robustness.13,14 Variations of FASP, such
as gel-aided sample preparation14 and enhanced FASP,15 seem
to have no significant improvement in protein coverage from
n the proteomics field, there are two major objectives
Idriving continuous technological advancements. One is to
obtain the highest number of protein identification. From mass
spectrometry (MS) instrumentation innovation1 to sample
fractionation method development,2 researchers aim to achieve
deep proteome representation in their findings. Deep
proteome coverage has the potential to encompass a broad
dynamic range of proteins in biological samples and reveal low-
abundance proteins, which are often key regulatory proteins
involved in cellular and disease processes.3 Another objective is
to utilize a minimal amount of starting biological materials, or
nanoproteomics.4 Nanoproteomics is critical when the
accessible sample is scarce, as it is often the case in clinical
settings or with primary cells that do not grow in vitro. In
addition, many research topics require proteome analysis at
subcellular or single-cell resolution.5,6 Although deep proteome
coverage and nanoproteomics clearly offer great value to
biomedical research, both remain challenging to the
proteomics field and require continuous technology and
methodology improvements.
Received: May 3, 2019
Accepted: July 18, 2019
Published: July 18, 2019
© 2019 American Chemical Society
10395
DOI: 10.1021/acs.analchem.9b02092
Anal. Chem. 2019, 91, 10395−10400

Analytical Chemistry
Technical Note
low-amount samples.14,16 As for SP3, pH changes during
sample preparation could lead to protein loss during wash
steps.13 StageTip requires fine manipulation skills and
experience that can be challenging to many researchers.
Nanodroplet processing and microfluidic approaches rely on
specialized equipment, making them inaccessible to most
laboratories.
Although each of the protocols described above has its own
benefits and limitations, there is a common trait contributing
to their improved performance against conventional workflows.
All protocols focus on reducing sample processing steps or
volume to minimize sample loss. Protein and peptide adhesion
to labware surfaces is a well-known phenomenon limiting deep
proteome coverage and nanoproteomics.17 An early study18
showed that peptides are lost to labware surfaces nonspecifi-
cally, and even high-abundance peptides can suffer significant
signal loss. The nonspecific nature of sample loss can be
attributed to the roughness of labware surfaces where proteins
and peptides randomly occupy potential binding sites.19 In
addition, hydrophobic peptides are more prone to surface
adhesion,18 which is likely due to hydrophobic interaction as
common labware is made of hydrophobic materials.20 Several
studies have also examined the impact of sample adhesion to
MS signal intensity and reproducibility.21,22
Although these studies offer valuable insights to the field,
they still do not paint the full picture of sample loss in
proteomic analysis. Specifically, quantitative information on
the extent of sample loss seems to be poorly investigated. Most
sample loss studies only compare final protein identification
and lack details on protein and peptide loss during sample
processing.21 In addition, many studies on sample adhesion
have used simple peptide mixtures, such as bovine serum
albumin digest23 or standard peptides,18,20 which may not
recapitulate the sample loss pattern in complex lysates.
In this study, we have quantified protein and peptide sample
loss using the U2OS human osteosarcoma cell line with a
conventional in-solution sample preparation protocol, which is
often criticized for its low sample recovery. By quantitatively
understanding sample loss in each sample preparation step,
targeted adjustments were made to the in-solution workflow.
Specifically, we adapted two different modified tube surfaces to
improve sample recovery yield. Of these two surface
technologies, omniphobic slippery liquid infused porous
surface (SLIPS)24 treatment helps to address surface rough-
ness and reduce nonspecific protein adhesion,25,26 whereas
commercially available LoBind tubes (Eppendorf Corporate)
have a hydrophilic surface to mitigate protein and peptide loss
due to hydrophobic interactions.27 Together with in-line
desalting, we were able to minimize sample loss and enhance
protein identification from 1000 mammalian cells using a
simple workflow.
MS Data Analysis and Protein Identification. LC/MS/
MS raw files were searched using Proteome Discoverer V2.2
(Thermo Fisher) against the SwissProt human database
(released in June 2017), with the following parameters: full
tryptic specificity with up to two missed cleavage sites,
carbamidomethylation of cysteine residues as a fixed
modification, and protein N-terminal acetylation plus
methionine oxidation as variable modifications. The false
discovery rate was set to 0.01.
AQUA Targeted Peptide Loss Assessment. Stable-
isotope labeled absolute quantification (AQUA) peptides for
eight proteins expressed in U2OS cells at various abundance
levels were synthesized by Thermo Fisher (Table S1). The
amounts of AQUA peptides for each sample of 10, 2, and 0.2
μg of peptides (∼100 000, ∼20 000, and ∼2000 U2OS cells)
are listed in Table S2. AQUA targeted protein and peptide
measurement conditions are described in the Supporting
Information.
SLIPS Treatment and Lubrication. Regular 1.5 mL
microcentrifuge tubes (VWR, USA) were placed vertically on a
Petri dish and oxygen plasma treated for 1 min (Harrick
Plasma Cleaner, PDC-002, 230 V). Immediately after, hydroxyl
terminated tubes were placed in a vacuum desiccator (Bel-Art
SP Scienceware, USA) containing 200 μL of tridecafluoro-
1,1,2,2-tetrahydrooctyl trichlorosilane (Sigma-Aldrich, USA).
The tubes were kept under vacuum for 5 h and silanized using
chemical vapor deposition (CVD). After the CVD treatment,
tubes were placed in an oven and heat treated overnight at 60
°C to complete the silanization procedure. SLIPS-coated tubes
were washed with ethanol. Immediately before use, 300 μL of
perfluoroperhydrophenanthrene (Sigma-Aldrich, USA) was
added into each SLIPS-coated tube to lubricate the SLIP
surface. Excess lubricant was pipetted out.
RESULTS AND DISCUSSION
■
Low Protein Identification and Severe Sample Loss
from ∼1000 Mammalian Cells Using a Conventional In-
Solution Sample Processing Method. We first tested
peptide and protein identification using a conventional in-
solution sample preparation workflow (Figure 1A) to assess
the overall sample loss. With our nanoflow LC/MS/MS
platform, in a 75 min LC gradient run, we were able to identify
on average 14 745 unique peptides and 2349 unique proteins
with an injection of 0.1% (∼1000 cells equivalent) from 106
U2OS cell (∼100 μg of total protein) tryptic peptides. In
comparison, when we started with 1000 U2OS cells (∼100 ng
of total protein), we were only able to detect 141 peptides and
76 proteins on average (Figure 1B,C). To ensure the result was
not cell line specific, we repeated the experiment using HL-60
myeloid leukemia cells and observed a similar trend in protein
and peptide identification (Figure S1). These results high-
lighted the severe sample handling loss with a small amount
(∼1000 mammalian cells) of starting material.
EXPERIMENTAL SECTION
■
Liquid Chromatography and Tandem Mass Spec-
trometry (LC/MS/MS). Peptides in mobile phase A (0.1%
formic acid) were injected and separated on homemade trap
and analytical columns with a 75 min reversed-phase gradient,
delivered by a Thermo Fisher Ultimate 3000 RSLCNano
UPLC system coupled to a Thermo QExactive HF quadru-
pole−Orbitrap mass spectrometer. LC/MS conditions are
detailed in the Supporting Information. MS raw files are
available via the PRIDE partner repository with the data set
identifier PXD013702.
To quantitatively assess sample handling loss, we used the
targeted AQUA method28 to track individual protein losses.
Two representative proteins (eight proteins in total) were
chosen from each of the 103, 104, 105, and 106 copy number
levels per cell in the U2OS cell line, based on a previous study3
(Table S1). We found that there was little protein loss (except
for CDC2) when we started from ∼100 000 U2OS cells (10 μg
of total protein), whereas protein recovery dropped signifi-
cantly for these eight proteins when we started from ∼20 000
U2OS cells (2 μg of total protein). Furthermore, when we
10396
DOI: 10.1021/acs.analchem.9b02092
Anal. Chem. 2019, 91, 10395−10400

Analytical Chemistry
Technical Note
Figure 2. Modified labware surfaces helped to mitigate protein
adhesion and improved sample recovery. (A) Schematic illustration of
in-tandem sample tube transfer test to assess protein loss to tube
surfaces. (B) Protein recovery decreased with extended exposure to
1.5 mL microcentrifuge tube surfaces with 50, 10, and 2 μg of U2OS
lysate, assessed by protein BCA assays. (C,D) Protein loss to tube
surfaces was greatly mitigated with both LoBind tubes and SLIPS
treated tubes, with 10 (C) and 2 μg (D) of U2OS lysate, assessed by
protein BCA assays. Error bars represent standard deviations of
triplicate repeats.
Figure 1. Challenge for nanoproteomic analysis of 1000 mammalian
cells. (A) Schematic illustration of conventional in-solution proteomic
sample preparation procedure. (B,C) Totals of 106 and 1000 U2OS
cells were processed with the conventional in-solution sample
preparation method. Of the 106 cell and 1000 cell samples, 0.1 and
100% were injected for LC/MS analyses, respectively. Peptide (B)
and protein (C) identification were compared. (D) Protein recovery
yields for eight selected proteins (Table S1) throughout the
conventional in-solution sample preparation procedure with 10 μg,
2 μg, and 200 ng starting U2OS protein amounts, assessed using
AQUA assays. Error bars represent standard deviations of triplicate
repeats.
2B). Specifically, at the 2 μg total protein level, 56.3% of the
initial sample content was lost even in the first microcentrifuge
tube, whereas only 1.8% was lost at the 50 μg total protein
level. These findings demonstrated that even single-tube
strategies (FASP, SP3, and StageTip) could suffer from
significant sample loss at low protein amount. Although the
percentages of proteins left in Tubes 2−4 seemed constant at
around 10% for 2 μg, this could be caused by falling below the
detection limit of the BCA assay. The experiment was repeated
using 0.5 mL microcentrifuge tubes (Figure S2). Less sample
loss incurred in 0.5 mL tubes than in 1.5 mL tubes, which is
attributed to the decreased surface area of 0.5 mL tubes.
Next, we tested the performance of two modified tube
surfacesomniphobic SLIPS24 and the hydrophilic LoBind
surface. While LoBind tubes are made of proprietary polymer
materials27 and are commercially available (Eppendorf
Corporate), we treated regular microcentrifuge tubes in-
house to make SLIPS tubes (Figure S3). After successive
tube transfer of samples with 10 μg of total proteins, the
protein recoveries in Tube 4 for SLIPS and LoBind tubes were
40.4 and 46.8% (Figure 2C), both of which were significantly
higher than those in the regular tubes (28.2%). The
experiment was repeated with 2 μg of total proteins samples,
in which the percentages of proteins left in Tube 4 for SLIPS
and LoBind tubes were 44.2 and 52.2%, respectively (Figure
2D). These were significantly higher when compared with
10.9% for regular tubes. Even in the first tube, SLIPS and
LoBind tubes showed significant improvement as compared to
regular tubes, which suggested potential application of these
modified tube surfaces for the aforementioned single-tube
sample processing strategies (FASP, StageTip, SP3). Addi-
tionally, sample preparation using SLIPS and LoBind tubes was
more reproducible when compared with regular tubes as
shown by the variability of measurements. SLIPS treatment
offered flexibility, as it could be applied to all types of labware
started from ∼2000 U2OS cells (200 ng of total protein), only
three proteins (GSK3B, MYL6, and H4) were still detectable,
among which only the most abundant protein H4 remained at
similar recovery yield to that acquired from ∼20 000 U2OS
cells (Figure 1D). These results clearly demonstrated the
severe sample loss that incurred during the sample handling
process from ∼1000−2000 U2OS cells, which explained the
aforementioned low protein identification coverage from
∼1000 U2OS cells. To improve the protein identification
rate from such a small sample amount, we decided to assess
and rectify sample loss in each of the conventional proteomic
sample preparation steps.
Assessment and Improvement of Protein Loss to
Labware Surfaces. To assess protein loss to the labware
surface during the sample handling workflow, we analyzed 50,
10, and 2 μg of total proteins from U2OS lysate (∼500 000, ∼
100 000, and ∼20 000 cells, respectively). Each sample was
transferred in tandem into four 1.5 mL microcentrifuge tubes
to assess total protein loss (Figure 2A), which mimicked
surface exposure during the conventional in-solution work-
flowinitial cell sample tube, desalting well, elution well, and
speed-vacuum tube. As expected, samples with lower protein
amounts tend to incur higher total protein loss in Tube 4, with
sample recovery yields in Tube 4 for 50, 10, and 2 μg total
protein levels at 84.6, 28.2, and 10.9%, respectively (Figure
10397
DOI: 10.1021/acs.analchem.9b02092
Anal. Chem. 2019, 91, 10395−10400

Analytical Chemistry
Technical Note
surfaces and connection tubing inner surfaces, whereas
commercially available LoBind tubes were more convenient
and time-saving.
peptide loss) but not with 10 μg and higher amounts of
starting material (Figure S5). In the AQUA assays, all eight
selected peptides experienced more than 90% recovery when
10 μg of U2OS tryptic peptides were dried using speed-
vacuum. However, sample recovery yields for all eight selected
peptides dramatically decreased from 10 μg (ranging from 89
to 100%) to 2 μg of U2OS peptide samples (ranging from 70
to 88%) and went further down with 200 ng samples (ranging
from 53 to 68%). In particular, selected peptides representing
the two lowest abundant proteins, NEK9 and SCYL2, were not
detected during the tests at the 200 ng total peptide level
(Figure 3B). These results clearly demonstrated severe sample
loss during the sample drying step for samples starting from
∼1000−2000 mammalian cells.
Assessment and Improvement of Peptide Loss to
Sample Desalting and Drying Steps. The desalting step
removes cell lysis byproducts and ensures the quality of LC/
MS analysis. However, it is inevitable to experience peptide
loss due to the additional surface exposure, C18 resin binding,
and wash steps. In our peptide BCA assays, 72.2% of 50 μg,
64.0% of 10 μg, and 67.4% of 2 μg total U2OS tryptic peptides
were recovered after desalting (Figure S4) using wells
containing 10 mg of SOLA C18 resin for 50 μg of peptides
and 2 mg of resin for 10 and 2 μg of peptides. In the AQUA
assays, we further confirmed that recovery yields for selected
peptides from individual proteins (Figure 3A) during C18
To mitigate the above losses for samples with ∼1000−2000
total cells, we decided to test in-line desalting with the C18
trap column. In the test, 200 ng of U2OS tryptic peptides
(from ∼2000 U2OS cells) were injected onto a one-time use
homemade trap column, whereas desalted peptides were then
separated on the downstream analytical column. This setup
alleviated labware surface exposure during regular desalting
and speed-vacuum steps. It also shortened the time of sample
treatment. As shown in Figure 3C, in-line desalting successfully
enhanced AQUA peptide recovery, especially for selected
peptides that represented low-abundance proteins NEK9,
SCYL2, GSK3B, and NTKL. Conversely, selected peptides
representing higher-abundance proteins (SPTA2, CDC2,
MYL6, and H4) showed little recovery improvement between
the two desalting strategies. These results suggested that we
could use in-line desalting to improve sample recovery yield for
low-abundance peptides in samples from ∼1000 to 2000
mammalian cells.
MS Analysis with Improved Sample Handling Work-
flow. After stepwise assessment and optimization of sample
handling recovery yields, we went on to analyze a small
amount of U2OS cells. This time, we first processed ∼10 000
U2OS cells, followed by 10% injection of the resulted peptides
(i.e., ∼1000 cell equivalent) and 75 min LC/MS/MS runs with
each sample handling method. With LoBind tubes plus in-line
peptide desalting, unique protein identification for a 1000 cell
equivalent injection from 10 000 U2OS cells reached 1277 and
a unique peptide identification of 5453. This represented ∼2.5-
and ∼2.9-fold increases over the conventional workflow.
Replacing the LoBind tubes with SLIPS-treated tubes, we
identified 758 unique proteins and 2936 unique peptides, both
∼1.5-fold increases over the conventional sample handling
workflow (Figures 4A and S6A). When we directly started
from ∼1000 U2OS cells using LoBind tubes and in-line
desalting, we identified 829 unique proteins and 2957 unique
peptides (Figures 4B and S6B), marking ∼11- and ∼21-fold
increases over the conventional workflow. With SLIPS treated
tubes, from ∼1000 U2OS cells, we identified 294 proteins and
792 peptides, still ∼3.8- and ∼5.6-fold increases over the
conventional workflow (Figures 4B and S6B). The under-
performance of SLIPS tubes versus LoBind tubes could be due
to the incompatibility of SLIPS with the in-solution workflow.
Lubricant used in the SLIPS treatment procedure could
evaporate during incubation at 37 °C, which compromised the
nonadhesion ability of the SLIPS surface. The evaporation
issue can be mitigated by keeping the lubricant infused SLIP
surfaces in closed environments, such as in microfluidic devices
and fluidic connection tubing. These measures can help to
Figure 3. In-line desalting improved peptide recovery. (A,B)
Recovery yields for individual peptides representing the eight selected
proteins during desalting (A) and speed-vacuum drying steps (B),
with 10 μg, 2 μg, and 200 ng of U2OS tryptic peptides, assessed by
AQUA assays. (C) In-line desalting of 200 ng of U2OS tryptic
peptides led to improved recovery yields for peptides representing
low-abundance proteins (NEK9, SCYL2, GSK3B, NTKL), assessed
by the AQUA method. Error bars represent standard deviations of
triplicate repeats.
desalting remained relatively stable, when we started with 10,
2, and 0.2 μg of U2OS tryptic peptides. The selected peptides
that represented the two lowest-abundance proteins (NEK9
and SCYL2) did display reduced recovery with 0.2 μg (∼2000
cells) samples (Figure 3A).
For the speed-vacuum drying step, high peptide loss was
observed with 2 μg of U2OS tryptic peptides (26.2% of
10398
DOI: 10.1021/acs.analchem.9b02092
Anal. Chem. 2019, 91, 10395−10400

Analytical Chemistry
Technical Note
Figure S6. Optimized nanoproteomic analysis of
mammalian cells (PDF)
AUTHOR INFORMATION
Corresponding Authors
*E-mail: yu.lu@mcmaster.ca. (Y.L.)
*E-mail: didar@mcmaster.ca. (T.F.D.)
■
ORCID
Tohid F. Didar: 0000-0002-8757-8002
Yu Lu: 0000-0003-4522-6823
Author Contributions
R.W. and Y.L. designed research, coordinated the project, and
wrote the manuscript. R.W., S.X., and Y.L. performed
experiments. M.B. and T.F.D. helped to prepare labware
SLIPS surfaces. All authors have given approval to the final
version of the manuscript.
Figure 4. Optimized analysis of 1000 mammalian cells. (A,B) Unique
proteins identifed with conventional, SLIPS tubes plus in-line
desalting and LoBind tubes plus in-line desalting workflows,
respectively, from 1000 cell equivalent injection of 10 000 U2OS
cells (A) and from 1000 U2OS cells (B). Error bars represent
standard deviations of triplicate repeats.
Notes
achieve optimal recovery with samples of ∼100 to ∼1000
mammalian cells.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
CONCLUSION
■
Support for this work was provided by a Canadian Natural
Science and Engineering Research Council Discovery Grant
(RGPIN-2017-06159), the Canada Foundation for Innovation
JELF (35544), the Ontario Ministry of Economic Develop-
ment, Job Creation and Trade ORF-RI (35544), and the
Marta and Owen Boris Foundation.
In summary, we have provided quantitative information on
stepwise sample handling loss during the conventional in-
solution proteomic sample preparation workflow. For samples
of ∼1000 mammalian cells, we demonstrated that the major
sources of sample loss included protein surface adhesion,
peptide desalting, and speed-vacuum drying. Even single-tube
sample preparation strategies may not be sufficient for
processing such a low amount of samples. Instead, by
adaptation of modified labware surfaces (LoBind and
SLIPS), along with in-line desalting, we can achieve a
substantial increase in protein identification using the most
basic sample preparation workflow, without any complex
technology, equipment, and extensive training. The quantita-
tive analysis of sample loss provides valuable information to
researchers for future optimization and application of advanced
nanoproteomic sample preparation workflows to push the
boundary of detection sensitivity and proteome coverage with
1000 or fewer mammalian cells.
REFERENCES
■
(1) Maher, S.; Jjunju, F. P. M.; Taylor, S. Rev. Mod. Phys. 2015, 87
(1), 113−135.
(2) Mostovenko, E.; Hassan, C.; Rattke, J.; Deelder, A. M.; van
Veelen, P. A.; Palmblad, M. EuPa Open Proteomics 2013, 1, 30−37.
(3) Beck, M.; Schmidt, A.; Malmstroem, J.; Claassen, M.; Ori, A.;
Szymborska, A.; Herzog, F.; Rinner, O.; Ellenberg, J.; Aebersold, R.
Mol. Syst. Biol. 2011, 7, 549.
(4) Zhu, Y.; Piehowski, P. D.; Kelly, R. T.; Qian, W. J. Expert Rev.
Proteomics 2018, 15 (11), 865−871.
(5) Virant-Klun, I.; Leicht, S.; Hughes, C.; Krijgsveld, J. Mol. Cell.
Proteomics 2016, 15 (8), 2616−2627.
(6) Thul, P. J.; Akesson, L.; Wiking, M.; Mahdessian, D.; Geladaki,
̈
A.; AitBlal, H.; Alm, T.; Asplund, A.; Bjork, L.; Breckels, L. M. Science
2017, 356 (6340), eaal3321.
ASSOCIATED CONTENT
* Supporting Information
Supporting Information (experimental procedures, supplemen-
tary table and figures) is available on the ACS Publications
Web site. The Supporting Information is available free of
charge on the ACS Publications website at DOI: 10.1021/
acs.analchem.9b02092.
■
S
(7) Canas, B.; Pineiro, C.; Calvo, E.; Lopez-Ferrer, D.; Gallardo, J.
M. Journal of Chromatography A 2007, 1153, 235−258.
(8) Wisniewski, J. R.; Zougman, A.; Nagaraj, N.; Mann, M. Nat.
́
Methods 2009, 6 (5), 359−362.
(9) Hughes, C. S.; Foehr, S.; Garfield, D. A.; Furlong, E. E.;
Steinmetz, L. M.; Krijgsveld, J. Mol. Syst. Biol. 2014, 10, 757.
(10) Rappsilber, J.; Mann, M.; Ishihama, Y. Nat. Protoc. 2007, 2 (8),
1896−1906.
Experimental procedures; Supplementary Table S1:
Synthetic AQUA Peptides Selected for Sample Loss
Assessment; Supplementary Table S2: AQUA Peptide
Amounts Added into Protein and Peptide Samples for
Sample Loss Assessment Assays; Figure S1. Challenge
for nanoproteomic analysis of 1000 mammalian cells;
Figure S2. Protein recovery decreased with extended
exposure to 0.5 mL microcentrifuge tube surfaces with
50, 10, and 2 μg of U2OS lysate, assessed by protein
BCA quantification; Figure S3. Schematic of SLIPS
surface treatment on rough labware surface; Figure S4.
Peptide recovery yield for U2OS tryptic peptides after
desalting, measured by peptide BCA assay; Figure S5.
Peptide recovery yield for U2OS tryptic peptides after
speed-vacuum drying, measured by peptide BCA assay;
(11) Zhu, Y.; Piehowski, P. D.; Zhao, R.; Chen, J.; Shen, Y.; Moore,
R. J.; Shukla, A. K.; Petyuk, V. A.; Campbell-Thompson, M.;
Mathews, C. E.; et al. Nat. Commun. 2018, 9, 882.
(12) Zhu, Y.; Clair, G.; Chrisler, W. B.; Shen, Y.; Zhao, R.; Shukla,
A. K.; Moore, R. J.; Misra, R. S.; Pryhuber, G. S.; Smith, R. D.; et al.
Angew. Chem., Int. Ed. 2018, 57 (38), 12370−12374.
(13) Sielaff, M.; Kuharev, J.; Bohn, T.; Hahlbrock, J.; Bopp, T.;
Tenzer, S.; Distler, U. J. Proteome Res. 2017, 16 (11), 4060−4072.
(14) Fischer, R.; Kessler, B. M. Proteomics 2015, 15 (7), 1224−1229.
(15) Erde, J.; Loo, R. R. O.; Loo, J. A. J. Proteome Res. 2014, 13 (4),
1885−1895.
(16) Nel, A. J. M.; Garnett, S.; Blackburn, J. M.; Soares, N. C. J.
Proteome Res. 2015, 14 (3), 1637−1642.
(17) Specht, H.; Slavov, N. J. Proteome Res. 2018, 17 (8), 2565−
2571.
10399
DOI: 10.1021/acs.analchem.9b02092
Anal. Chem. 2019, 91, 10395−10400

Analytical Chemistry
Technical Note
(18) Kraut, A.; Marcellin, M.; Adrait, A.; Kuhn, L.; Louwagie, M.;
Kieffer-Jaquinod, S.; Lebert, D.; Masselon, C. D.; Dupuis, A.; Bruley,
C.; Jaquinod, M.; Garin, J.; Gallagher-Gambarelli, M. J. Proteome Res.
2009, 8 (7), 3778−3785.
(19) Mermut, O.; Phillips, D. C.; York, R. L.; McCrea, K. R.; Ward,
R. S.; Somorjai, G. A. J. Am. Chem. Soc. 2006, 128 (11), 3598−3607.
(20) Weikart, C. M.; Klibanov, A. M.; Breeland, A. P.; Taha, A. H.;
Maurer, B. R.; Martin, S. P. SLAS Technology 2017, 22 (1), 98−105.
(21) de Graaf, E. L.; Pellegrini, D.; McDonnell, L. A. J. Proteome Res.
2016, 15, 4722−4730.
(22) Warwood, S.; Byron, A.; Humphries, M. J.; Knight, D. J.
Proteomics 2013, 85, 160−4.
(23) Bark, S. J.; Hook, V. J. Proteome Res. 2007, 6 (11), 4511−4516.
(24) Wong, T.-S.; Kang, S. H.; Tang, S. K. Y.; Smythe, E. J.; Hatton,
B. D.; Grinthal, A.; Aizenberg, J. Nature 2011, 477, 443−7.
(25) Badv, M.; Jaffer, I. H.; Weitz, J. I.; Didar, T. F. Sci. Rep. 2017, 7
(1), 11639.
(26) Hosseini, A.; Villegas, M.; Yang, J.; Badv, M.; Weitz, J. I.;
Soleymani, L.; Didar, T. F. Adv. Mater. Interfaces 2018, 5 (18),
1800617.
(27) Heussner, A. H.; Altaner, S.; Kamp, L.; Rubio, F.; Dietrich, D.
R. Chem.-Biol. Interact. 2014, 223, 87−94.
(28) Gerber, S. A.; Rush, J.; Stemman, O.; Kirschner, M. W.; Gygi, S.
P. Proc. Natl. Acad. Sci. U. S. A. 2003, 100 (12), 6940−6945.
10400
DOI: 10.1021/acs.analchem.9b02092
Anal. Chem. 2019, 91, 10395−10400