1. Notes & Tips
A visual chip-based coimmunoprecipitation technique for analysis
of protein–protein interactions
Qing Chen a,1
, Qiongming Liu a,1
, Zhoumin Li b
, Wenying Zhong b
, Wei He a,**, Danke Xu a,c,*
a
State Key Laboratory of Proteomics, Beijing Proteome Research Center, Beijing Institute of Radiation Medicine, Beijing 102206, China
b
Department of Analytical Chemistry, China Pharmaceutical University, Nanjing 210009, China
c
Key Laboratory of Analytical Chemistry for Life Science (MOE), School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210093, China
a r t i c l e i n f o
Article history:
Received 8 April 2010
Received in revised form 5 May 2010
Accepted 6 May 2010
Available online 10 May 2010
Keywords:
Protein–protein interactions
Silver enhancement detection
Protein microarrays
Coimmunoprecipitation
a b s t r a c t
Here we report a visual chip-based coimmunoprecipitation (vChip–coIP) platform for analysis of protein–
protein interactions (PPIs) by combining advantages of an antibody microarray, traditional coIP, and a
silver enhancement detection method. The chip was fabricated by spotting anti-Flag antibody on alde-
hyde-modified slides, and the resulting platform could assay immunoprecipitate from a small amount
of crude cell lysates containing Flag–bait and Myc–prey. The interaction signals are visible using biotin-
ylated anti-Myc antibody and colloidal gold-labeled streptavidin followed by a silver enhancement detec-
tion method. It is shown that vChip–coIP is a simple, cost-effective, and highly efficient platform for the
comprehensive study of PPIs.
Ó 2010 Elsevier Inc. All rights reserved.
The human genome consists of 20,000 to 30,000 genes encod-
ing more than 500,000 different proteins. Moreover, a cell can pro-
duce more than 10,000 proteins at any given time. It has been
estimated that more than 80% of proteins do not operate alone
but rather operate in complexes by which protein–protein interac-
tions (PPIs)2
are regulated by several mechanisms [1–3]. Most, if
not all, biological processes require the cooperation of at least two
proteins and may require additional proteins for highly complex
functions. Therefore, the analytical methods of PPIs, such as coim-
munoprecipitation (coIP), affinity chromatography, and two-hybrid
assays, are essential for the elucidation of biological processes
and/or networks [4–6]. However, the traditional resin-based coIP
is time-consuming, especially for high-throughput analysis. More
recently, protein arrays have been developed to assay PPIs [7–9],
and several detection strategies based on fluorescent or radiometric
labels have been compared [10].
In this article, we report a visual chip-based coIP (vChip–coIP)
technology for rapid analysis of PPIs, and its principle is illustrated
in Fig. 1A. The protein array chip was fabricated by spotting anti-
bodies with the Smart Arrayer 48 Spotting Robot (CapitalBio, Bei-
jing, China) on CSS aldehyde-modified slides (CEL Associates,
Pearland, TX, USA). Here 0.2 mg/ml murine immunoglobulin G
(IgG), 1 mg/ml anti-Flag M2 antibody (Sigma, St. Louis, MO, USA),
and 0.02 mg/ml biotinylated bovine serum albumin (Bio-BSA) were
spotted from left to right, respectively (Fig. 1B). All slides were
incubated overnight at 4 °C to allow maximum binding of antibod-
ies to the aldehyde slides, followed by blocking with 10 mg/ml BSA
in TBST (20 mM Tris–Cl [pH 8.0], 150 mM NaCl, and 0.05% [v/v]
Tween 20) at room temperature for 1 h. After blocking, the slides
were rinsed three times with TBST, followed by adding cell lysate
to the slide to analyze protein expression and concentration.
To study cell lysate protein expression and concentration on PPI
signals, two cell lysates with differing antibodies were employed.
First, Flag–p50 and c-Myc–p65 were used as positive controls,
and Flag–Jun and c-Myc–lacZ were used as negative controls.
Human embryonic kidney 293 (HEK293) cells were cultured in
Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen, Carlsbad,
CA, USA) supplemented with 10% fetal bovine serum (FBS, HyClone,
Logan, UT, USA), penicillin, streptomycin, and glutamine. Transfec-
tions were performed with Lipofectamine 2000 (Invitrogen) accord-
ing to the manufacturer’s instructions. After 30 h, cells were
harvested and lysed in EBC buffer (50 mM Tris–Cl [pH 8.0], 120
mM NaCl, 0.5% [v/v] NP40, and 1 mM ethylenediaminetetraacetic
acid [EDTA]) supplemented with 50 lg/ml phenylmethanesulfonyl
0003-2697/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.ab.2010.05.003
* Corresponding authors at: key Laboratory of Analytical Chemistry for Life
Science (MOE), School of Chemistry and Chemical Engineering, Nanjing University,
22 Hankou Road, Nanjing 210093, China. Fax: +86-10-80705199.
** Corresponding authors.
E-mail addresses: hewei1012@163.net (W. He), xudk@nju.edu.cn (D. Xu).
1
These authors contributed equally to this work.
2
Abbreviations used: PPI, protein–protein interaction; coIP, coimmunoprecipita-
tion; vChip–coIP, visual chip-based coIP; IgG, immunoglobulin G; Bio-BSA, biotinyl-
ated bovine serum albumin; HEK293, human embryonic kidney 293; DMEM,
Dulbecco’s modified Eagle’s medium; FBS, fetal bovine serum; EDTA, ethylenedi-
aminetetraacetic acid; PMSF, phenylmethanesulfonyl fluoride.
Analytical Biochemistry 404 (2010) 244–246
Contents lists available at ScienceDirect
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio

2. fluoride (PMSF) and protease inhibitor cocktail (Roche, Basel,
Switzerland) at room temperature for 10 min. After centrifugation
(4 °C, 12,000 rpm, 10 min), the protein concentration of the superna-
tant was determined by the Bradford method. Finally, 20 ll of cell ly-
sate was transferred into detection wells of the chip and incubated
for 2 h at room temperature.
The chips were washed three times with TBST to remove un-
bound proteins. The interacting partners on the microarrays were
incubated with 0.02 mg/ml biotinylated monoclonal anti-c-Myc
antibody (Sigma) in a humid chamber for 1 h. The excess unbound
anti-c-Myc antibody was removed, followed by washing with TBST.
Colloidal gold-labeled streptavidin (Sigma) and substrate silver en-
hancer solution (Sigma) were added to incubate with the interact-
ing partners sequentially for the desired time. Finally, the slides
were rinsed thoroughly with water to remove the excess substrate
solution, followed by scanning using a dual-media scanner (Scan-
Maker 8700, Microtek, China). The scanned images (Fig. 1C) reveal
black spots on the anti-Flag M2 antibody for the samples of the po-
sitive control, whereas the negative control samples containing
Flag–Jun and c-Myc–lacZ do not have black spots, thereby confirm-
ing the positive and negative controls, respectively. In addition, the
murine IgG spots do not show any dark images, suggesting that
there is a low cross-reactivity performance on the arrays and the
interacting protein pair could be visualized by this antibody
chip-based method.
To further extract the gray values of the spots, the images of the
antibody arrays were processed with Spot Data Pro 2.0 (Capital-
Bio), and the relative data are shown in Fig. 1D. Surrounding the
gray values is the background, which was subtracted from the
anti-Flag antibody spots. The protein pair of Flag–p50/c-Myc–p65
Fig. 1. (A) Principle and process of the vChip–coIP technique. (B) Pattern of spotting antibody: (1) murine IgG (as negative spots), (2) anti-Flag M2 antibody, and (3) Bio-BSA
(as positive spots). (C) Scanned images of the interacting protein pairs of Flag–p50 and c-Myc–p65 (positive control) and Flag–Jun and c-Myc–lacZ (negative control). Protein
concentrations were 0, 0.15, 0.30, 0.55, 1.10, 2.25, and 4.50 lg/ll (from left to right). (D) Relationship between the gray levels and protein concentrations of cell lysate. j,
Flag–p50 and c-Myc–p65; N, Flag–Jun and c-Myc–lacZ.
Notes & Tips / Anal. Biochem. 404 (2010) 244–246 245

3. was detectable at 0.25 lg/ll. Moreover, the cell lysate protein sig-
nals increased proportionally with the increase in protein concen-
tration, followed by signal saturation at 1.0 lg/ll, whereas the
negative controls of Flag–Jun and c-Myc–lacZ maintained rather
low gray values. These results indicate that the strength of inter-
acting protein signals was dependent on the protein concentration
of cell lysate below 1.0 lg/ll. To avoid this influence, the protein
concentration of 1.0 lg/ll in cell lysate was employed in the fol-
lowing analysis.
Furthermore, the concentration of biotinylated anti-c-Myc anti-
body and silver enhancement reaction time are crucial experimen-
tal parameters that affect the intensity of the final signal [11]. In
our experiments of optimizing analytical parameters, it can be
found that the gray levels of the spots increase with the concentra-
tion of the antibody and then level off at the concentration of
20 lg/ml. The relationship of signal intensity with silver enhance-
ment time indicates that the silver enhancement approaches satu-
ration and the gray levels reached a plateau at 18 min, at which
point the detection gives the best ratio of signal to noise. Thus,
the concentration of 20 lg/ml biotinylated anti-c-Myc antibody
and 18 min of silver enhancement time were employed for the fol-
lowing measurements.
Based on the above optimized conditions, this method was fur-
ther used to evaluate 5 pairs of noninteracting protein (Jun/c-
Term1, Jun/Jlip, Jun/c-Term2, Jun/DPH, and Jlip/LZPR) and 5 pairs
of well-known interacting protein (TRB3/ATF4, Mafk/Nrf2, Keap1/
Nrf2, MafG/Nrf2, and PLK1/TANK). The experiment was repeated
three times, and the values of the gray levels on the chips were cal-
culated according to the above method. The relative statistical
analysis was processed, and analysis of variance was used to eval-
uate the method’s reproducibility. The data showed no significant
difference for the 5 pairs of noninteracting protein pairs and the 5
pairs of well-known interacting protein pairs (P = 0.9880 and
P > 0.05, respectively). These results suggest that the method is
both accurate and reproducible.
As seen in Fig. 2, 5 pairs of noninteracting protein produced no
visible spots, and the average of gray values was 41.8. To evaluate
whether there is an interacting protein pair based on its gray value,
the cutoff value was calculated as the average value plus double
the standard error. Based on this principle, the cutoff value was
determined to be 162.5. The gray values of the well-known inter-
acting protein pairs were assayed to be 609.7 (TRB3/ATF4), 511.3
(Mafk/Nrf2), 411.3 (Keap1/Nrf2), 762.0 (MafG/Nrf2), and 4591.1
(PLK1/TANK). They are much higher than the cutoff value. As a
result, this method could provide a novel and simple assessment
for PPI.
Compared with traditional resin-based coIP, vChip–coIP re-
quires much smaller amounts of cell lysate and allows a large num-
ber of samples to be studied en masse for their PPI information
without the need for further manipulation of the slides such as
immunofluorescence staining or enzymatic amplification. Our re-
sults indicate that vChip–coIP could provide a highly accurate,
cost-effective, and highly efficient platform in practical applica-
tions to assay PPIs.
Acknowledgments
The authors acknowledge financial support from the
National Natural Foundation of China (20975050, 20575079) and
the National Basic Research Program of China (973 Program,
2006CB910803).
References
[1] U. Stelzl, U. Worm, M. Lalowski, C. Haenig, F.H. Brembeck, H. Goehler, M.
Stroedicke, M. Zenkner, A. Schoenherr, S. Koeppen, J. Timm, S. Mintzlaff, C.
Abraham, N. Bock, S. Kietzmann, A. Goedde, E. Toksöz, A. Droege, S. Krobitsch,
B. Korn, W. Birchmeier, H. Lehrach, E.E. Wanker, A human protein–protein
interaction network: A resource for annotating the proteome, Cell 122 (2005)
957–968.
[2] E. Phizicky, P.I. Bastiaens, H. Zhu, M. Snyder, S. Fields, Protein analysis on a
proteomic scale, Nature 422 (2003) 208–215.
[3] T. Berggad, S. Linse, Methods for the detection and analysis of protein–protein
interactions, Proteomics 7 (2007) 2833–2842.
[4] C. Lee, Coimmunoprecipitation assay, Methods Mol. Biol. 362 (2007) 401–406.
[5] S.C. Horn, M. Lalowski, H. Goehler, A. Dröge, E.E. Wanker, U. Stelzl, Huntingtin
interacts with the receptor sorting family protein GASP2, J. Neural Transm. 113
(2006) 1081–1090.
[6] P. Uetz, Y.A. Dong, C. Zeretzke, C. Atzler, A. Baiker, B. Berger, S.V. Rajagopala, M.
Roupelieva, D. Rose, E. Fossum, J. Haas, Herpesviral protein networks and their
interaction with the human proteome, Science 311 (2005) 239–242.
[7] J. LaBaer, N. Ramachandran, Protein microarrays as tools for functional
proteomics, Curr. Opin. Chem. Biol. 9 (2005) 14–19.
[8] R. Lehmann, C. Melle, Detection and identification of protein interactions of
S100 proteins by ProteinChip technology, J. Proteome Res. 4 (2005) 1717–
1721.
[9] Y. Merbl, M.W. Kischner, Large-scale detection of ubiquitination substrates
using cell extracts and protein microarrays, Proc. Natl. Acad. Sci. USA 106
(2009) 2543–2548.
[10] R. Hurst, B. Hook, M.R. Slater, J. Hartnett, D.R. Storts, N. Nath, Protein–protein
interaction studies on protein arrays: Effect of detection strategies on signal-
to-background ratios, Anal. Biochem. 392 (2009) 45–53.
[11] R.-Q. Liang, C.-Y. Tan, K.-C. Ruan, Colorimetric detection of protein microarrays
based on nanogold probe coupled with silver enhancement, J. Immunol.
Methods 285 (2004) 157–163.
Fig. 2. Evaluation of 5 noninteracting protein pairs (1: Jun/c-Term1; 2: Jun/Jlip; 3:
Jun/c-Term2; 4: Jun/DPH; 5: Jlip/LZPR) and 5 well-known interacting protein pairs
(6: TRB3/ATF4; 7: Mafk/Nrf2; 8: Keap1/Nrf2; 9: MafG/Nrf2; 10: PLK1/TANK) by the
vChip–coIP method. HEK293 cells were cotransfected with Flag–bait and Myc–prey
constructs, and 1 lg/ll cell lysates were used.
246 Notes & Tips / Anal. Biochem. 404 (2010) 244–246