Toward Biomarker Development in Large Clinical Cohorts: An

Subscriber access provided by Northern Illinois University

Article

Towards biomarker development in large clinical cohorts: An integrated high-throughput 96-well plate based sample preparation workflow for versatile downstream proteomic analyses Zeyu Sun, Xiaoli Liu, Jing Jiang, Haijun Huang, Jie Wang, Daxian Wu, and Lanjuan Li Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b01333 • Publication Date (Web): 29 Jul 2016 Downloaded from http://pubs.acs.org on July 31, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Towards biomarker development in large clinical cohorts: An integrated

high-throughput

96-well

plate

based

sample

preparation workflow for versatile downstream proteomic analyses

Zeyu Sun1#, Xiaoli Liu1#, Jing Jiang1, Haijun Huang2, Jie Wang1, Daxian Wu1, Lanjuan Li*1

AUTHOR ADDRESS 1

State Key Laboratory for Diagnosis and Treatment of Infectious Diseases,

Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, The First Affiliated Hospital, College of Medicine, Zhejiang University, 310003 Hangzhou, People's Republic of China 2

Department of Infectious Diseases, Zhejiang Provincial People's Hospital,

310024 Hangzhou, Zhejiang, People's Republic of China

Fax: +86-571-8723-6459. Email: [email protected] #

These two authors contributed equally to this work.

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 30

ABSTRACT

We describe a cheap, robust, fast, high-throughput and flexible proteomic sample processing method based on regular 96-well plate by acetone precipitation under low centrifuge speed (96PACS), which enables pre-digestion processing of 96 samples within 2 h. Tested on a complex Huh-7 total lysate, 96PACS produced comparable proteome coverage and even showed better reproducibility than FASP. Quantitative performance of 96PACS was further tested using data-independent acquisition and parallel reaction monitoring quantitation in a set of 6 benchmark samples consist of 6 serial dilution of BSA spiked in complex E.coli proteome background. The protocol was also successfully modified for automation, and was validated in a comparative label-free proteomic study to develop serum markers for early detection of liver fibrosis and necroinflammation in patients chronically infected with hepatitis B virus.

INTRODUCTION The coming age of precision medicine underscores the importance of robust clinical proteomic techniques for biomarker discovery, validation, and even LC-MS based clinical assays. Towards these aims, significant technological advances have been made in instrumentation, including chromatography, high resolution mass spectrometry, as well as innovative bioinformatics tools. A plethora of novel LC-MS workflows, such as data-independent methods including DIA, MSE or SWATH, and targeted proteomic methods including PRM or MRM have been developed to quantify proteins in individual samples from large cohort. Such new developments require a simple, reproducible and cost-effective sample preparation routine in the upfront of any downstream proteomic to profile large sample sets. Ideally, the preparation routine should be accommodated to 2


Page 3 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the 96-well plate format so that can be universally practiced by common clinical laboratories, and modified readily for automation. However, the widely adopted filter assisted sample processing (FASP) methods are carried out by individually packed 10 or 30 kDa MWCO spin filter units with multiple reagents loading and prolonged centrifugation steps, making this method only suitable to process small size of samples. There were several attempts to develop high-throughput sample preparation method in 96-well plate format, yet all with specially designed or modified 96-well apparatus 1-3. Alternatively, in-solution digestion protocols have long been used due to their simple setup. However, these methods often require additional clean up procedures such as acetone precipitation to remove unwanted contaminants. The key hurdle preventing in-solution preparation in 96-well plate format is its requirement of high speed centrifuge for protein precipitation, as typical 96-well plate can only withstand a g-force of around 2,000xg. In addition, it is difficult to add high speed centrifugation apparatus into sample automation platforms. In this study, we developed a straightforward and robust protocol via protein precipitation with acetone and centrifugation at slow speed in regular 96-well plates, dubbed 96PACS. Using complex protein mixtures, we illustrated the 96PACS provides higher proteome coverage with better reproducibility as compared to FASP. We tested the performance of this protocol in versatile downstream LC-MSMS experiments such as DIA and PRM. We showed this method can allow low sample load as little as 20 µg, and only cost minimum reagents and consumables. Finally, the protocol was modified for automation, and was used in a proof of concept study focusing on developing serum markers to foresee liver fibrosis and necroinflammation at early stage in patients chronically infected with hepatitis B virus (CHB).

EXPERIMENTAL SECTION Preparation of controlled benchmark sample sets. Human Huh-7 hepatocyte 3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 30

was grown to 90% confluency in Dulbecco's modification of Eagle's medium (DMEM; Gibico) containing 1% L-glutamine, 10% fetal bovine serum, 100 units/mL penicillin and 100 µg/mL streptomycin with 5% CO2 at 37°C. Wild-type E.coli strain K12 was grown in LB medium to midlog phase. The bacterial cells were collected by centrifugation and washed twice with PBS. The E.coli and Huh-7 proteins were extracted by B-PER and NP-40 lysis buffer (both Pierce), respectively, assisted by ultrasonication. Both lysates were centrifuged at 3,000×g for 10 minutes to remove cell debris. Protein concentration was determined by using the BCA Protein Assay (Pierce) following manufacturer’s protocol. Total E.coli protein was used to test modified acetone precipitation protocols with various incubation and low-speed centrifugation conditions. Total Huh-7 protein was used to compare 96PACS method with conventional FASP method in terms of proteome-wide identification coverage and method reproducibility. In addition, a set of BSA dilutions were added to an E.coli (20 µg) proteome background to test the sensitivity and quantitative performance of 96PACS in DIA-based LC-MSMS experiments. The BSA stock was mixed with E.coli protein at 1:10, 1:50, 1:100, 1:200, 1:500, and 1:1,000 ratios. Since 1 µg tryptic peptides were loaded per LC-MS injection, the amount of BSA peptides were loaded in series were 100, 20, 10, 5, 2 and 1 ng. The same set of samples were also tested using PRM targeting BSA for dynamic range and linearity assessment and 31 E.coli proteins with different abundance for reproducibility assessment. A list of 96 targeted peptides in PRM experiment were detailed in Table S-1. To process DIA and PRM data, spectral library of E.coli peptides was built by LC-MSMS experiments in DDA mode. To investigate possible sample loss during 96PACS processing, proteins retained in the acetone fraction were also analyzed by LC-MSMS as described in the supporting materials.

Serum Collection for Chronic HBV Infection Biomarker Study. Serum samples were collected from consenting patients visiting the Department of 4


Page 5 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Infectious Diseases, Zhejiang Provincial People's Hospital. Inclusion criteria were age of more than 30 years, positive HBsAg for more than 6 months, HBV DNA ≥ 103 copies/ mL and ALT ≤ 2 ULN (ULN = 50 U/L); ALT and HBV DNA were monitored monthly for 6 months prior to enrollment to ensure the persistent maintenance of ALT ≤ 2 ULN and HBV DNA ≥ 103 copies/mL. Liver necro-inflammation was assessed via liver biopsy and histopathological analysis. Two experienced clinical pathologists examined the biopsy sections for liver necro-inflammation activity (G0-G4) and liver fibrosis (S0-S4) according to Scheuer’s classiﬁcation system 4. In this study, patients (n=12) with early sign of liver necro-inflammation activity (G1 and S1) were enrolled. Patients co-infected with HIV, HCV or HDV, or having liver cirrhosis, alcoholic liver or non-alcoholic fatty liver diseases, autoimmune liver diseases, hepatocarcinoma were excluded from the study. Healthy individuals (n=12) who came to the Zhejiang provincial people’s hospital for medical evaluation were enrolled into control group, and were confirmed to have normal liver function without any type of liver diseases. Characteristics of all enrolled cases were summarized in Table 1. The study protocol in conformance with the ethical guidelines of the 1975 Declaration of Helsinki was granted by Ethics Committee of the Zhejiang provincial people’s hospital. Peripheral venous blood samples were collected and allowed to clot for 1 h at 4 °C. The serum fraction was prepared by centrifugation at 2,500xg for 15 min at 4 °C. The high abundant protein from 10 µL serum samples were depleted by the Top 2 abundant protein depletion spin columns (ThermoFisher) according to the manufacturer’s protocol. The resulting low abundant serum protein fractions were stored at -80 °C before further analysis.

Sample preparation using acetone precipitation with low centrifugation speed. Preliminary tests using 200 µg Huh-7 proteins were performed in centrifuge tubes to assess protein recovery rate at different conditions. Three key parameters were optimized: cold acetone incubation duration in ice-bath, 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 30

centrifugation speed and centrifugation duration. While fixing the ice incubation time for 2 h, we first investigated the protein recovery rates by setting centrifugation speed at 1,000 or 3,000xg for 30 or 60 min, and compared the results to that of the conventional method using 30 min centrifugation at 14,000xg. We then compared protein recovery rates of different incubation time (30 min, 2 h, 12 h) followed by centrifugation at 1,000xg for 30 min. In all tests, 6x volume of cold acetone was used. All centrifugation operations were done at 4 °C. The Huh-7 proteins in triplicates were reduced by 10 mM dithiothreitol (DTT) at 37°C for 20 min and alkylated by 30 mM iodoacetamide (IAA) in dark for 20 min, and then were precipitated at the sub-optimized condition (30 min acetone incubation and centrifugation at 1,000xg for 30 min). The protein pellets were resuspended in 200 µL 100 mM triethylammonium bicarbonate (TEAB). Trypsin (V5111, Promega) at enzyme to protein ratio of 1:50 was added for 14 h digestion at 37°C. Peptides were then desalted by Sep-Pak C18 SPE (Waters) and then fractionated using off-line HPLC before LC-MSMS analysis. The results were compared to that of FASP processed Huh-7 proteins. Peptide yield was determined by the ratio of resulting peptide amount to the total proteins amount loaded using the BCA method.

96-well plate with acetone precipitation with low centrifugation speed sample preparation (96PACS). For sample processing using 96PACS, 25 µL total protein (~20 µg) was added to each well and mixed with 2 µL 100 mM DTT. Protein reduction was performed at 37°C for 20 min. Cysteine alkylation was performed by adding 2 µL 300 mM IAA and incubated at room temperature in dark for 20 min. The well was then added with 200 µL pre-cooled acetone and ice-bathed for 30 min. The protein precipitate was collected by centrifugation at 1,000xg for 30 min. Acetone was discarded and the pellet was resuspended in 200 µL 100 mM TEAB. Trypsin at enzyme to protein ratio of 1:50 was added for 14 h digestion at 37°C. In all steps, the 96-well plate was simply shook by hand 6


Page 7 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


for 15 s to ensure proper reagent mixing, and was covered by parafilm to prevent contamination. This protocol was first tested on E.coli protein supplemented with BSA using DIA and PRM MS assays, and was later modified for automation to process serum samples for the CHB study as detailed in the supplementary methods.

Filter assisted sample preparation (FASP). The FASP method was performed as described previously 5. Briefly, Huh-7 proteins (200 µg) were denatured and reduced with 4% SDS and 100 mM DTT at 95°C for 5 min. The sample was centrifuged at 16,000xg for 5 min and then cooled to room temperature. Sample was then mixed with 200 µL of 8 M urea and applied to a 30 kDa MWCO spin filter (MRCPRT030, Millipore) and centrifuged. Another 200 µL of 8 M urea was added to wash the sample and removed by centrifugation. Protein alkylation was performed by incubate sample at dark with 100 µL 50 mM IAA in 8 M urea and for 30 min. After centrifugation, the sample was washed by 100 µL 8 M urea twice and 100 µL 50 mM TEAB twice. Trypsin in 200 µL 50 mM TEAB was added in 1:50 enzyme to protein ratio. After 16h digestion at 37°C, the tryptic peptides were collected in a new tube by centrifugation. Another 200 µL 50 mM TEAB wash, and a 200 µL 500 mM NaCl wash was performed. Peptide mixture was acidified by 10% formic acid (FA). All centrifugation steps in FASP protocol were carried out at 14,000xg for 15 min. Peptides were then desalted by Sep-Pak C18 SPE (Waters) and then fractionated using off-line HPLC before LC-MSMS analysis.

Off-line HPLC peptide fractionation. The peptide mixtures generated from either 96PACS or FASP processed Huh-7 total lysate were further fractionated with high pH reverse phase HPLC. Briefly, the first dimension RP separation was performed on a 1260 HPLC System (Agilent) with an XBridge RP column (5 µm, 150 Å, 4.6 mm X 250 mm, Waters). Mobile phase A contains 2% acetonitrile (ACN) and B contains 98% ACN. Both were adjusted to pH 10.0 using NH4OH. 7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 30

The solvent gradient was set as follows: 2–5% B in 2 min; 5–18% B in 11 min; 18–32% B in 9 min; 32–95% B in 1 min; maintained at 95% B for 5 min. The tryptic peptides were separated at a flow rate of 1 mL/min and monitored by UV at 214 nm. Eluent was collected every minute and was combined into 13 fractions via a concatenated fashion. The samples were dried by Speed Vac (Labconco) and reconstituted in 20 µL of 0.1% FA, 2% ACN in water for subsequent analyses. All fractions were analyzed by LC-MSMS in DDA mode.

Mass spectrometric acquisition. Peptides (~1 µg) were reconstituted in 0.1% FA and then enriched on a Symmetry C18 nanoACQUITY Trap Column (100 Å, 5 µm, 180 µm x 20 mm). Peptide separation was carried out by a BEH C18 nanoACQUITY Column (130Å, 1.7 µm, 75 µm X 250 mm) on an nanoACQUITY UPLC system (Waters, Milford, MA) at a flow rate of 200 nL/min. The gradient started with 2% of ACN and increased to 5% ACN in 10 min, then reached 19% ACN in 70 min and 30% ACN in 15 min. The gradient finally reached 98% ACN in 10 min and was then held for 5 min before it return to 2% ACN in 2 min and kept at the re-equilibration condition for 8 min until next injection. The total analysis time per injection was 120 min. The mobile phases were all supplemented with 0.1% FA. The nanoLC was coupled to an Orbitrap Q-Exactive mass spectrometer (Thermo). The source was operated at 1.8 kV. The data-dependent analysis (DDA) scheme included a full MS survey scan from 400 to 1,200 Th at the resolution of 70,000 FWHM (at m/z 200 Th) with automatic gain control (AGC) set to 3e6, followed by MS2 scans of 20 most intense peaks selected for fragmentation by higher-energy collision dissociation (HCD). The MS2 spectra were acquired at 17,500 FWHM resolution with AGC set to 2e5. The DIA scheme included a full MS survey scan from 500 to 900 Th at the resolution of 17,500 FWHM (at m/z 200 Th) with AGC set to 1e6, followed by HCD-MS2 spectra that covers 500-900 Th with 20 isolation windows each covers 20 Th in consecutive order. All MS2 spectra were acquired with 35,000 FWHM resolution with AGC set to 5e5. The PRM scheme included a full MS 8


Page 9 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


survey scan from 300 to 1,800 Th at the resolution of 17,500 FWHM (at m/z 200 Th) with AGC set to 3e6, followed by scheduled HCD events to collect MS2 spectra from a list of pre-defined m/z (Table S-1) with 2 Th isolation windows. All MS2 spectra were acquired with 35,000 FWHM resolution with AGC set to 1e6. In all analyses, the HCD normalized collision energy was set to 27%.

Data analysis. For Huh-7 dataset, the acquired .RAW files were loaded into MaxQuant (version 1.5.3.30) and searched against the human UniProtKB database (88,473 sequences, version 09-2015) with label free quantitation (LFQ) and intensity based absolute quantitation (iBAQ) function. Variable modifications were

set

to:

oxidation(M)

and

acetylation(protein

N-term).

Only

carbamidomethyl(C) was set as fixed modification. Trypsin with up to two missed cleavages was set. Mass tolerance of 20 ppm and 7 ppm were set for first and main search, respectively. Protein level 1% FDR was set to filter the result. For E.coli dataset, the same search strategy was used except against an E.coli UniProtKB database (4,305 sequences, version 09-2015) supplemented with BSA sequence. A spectral library for DIA and PRM analysis was generated by Skyline (v3.1.0.7382) based on MaxQuant search results. Precursor and all product peaks showing coeluting retention time, and identical peak shape within 10 min of the retention time of corresponding DDA peaks were retained. In addition, peak matches with dopt >0.8 and idopt >0.8, mass error

Toward Biomarker Development in Large Clinical Cohorts: An

Recommend Documents