J. Biomedical Science and Engineering, 2010, 3, 1125-1132
doi:10.4236/jbise.2010.312146 Published Online December 2010 (http://www.SciRP.org/journal/jbise/
JBiSE
).
Published Online December 2010 in SciRes. http://www.scirp.org/journal/JBiSE
Development of a novel protein multi-blotting device
Amin M. Hagyousif1, Voon J. Chong2, Hi roki Yokota3, Stanley Y. P. Chien1
1Department of Electrical and Computer Engineering, Indiana University Purdue University Indianapolis, Indianapolis, USA;
2Department of Biomedical Engineering, Indiana University Purdue University Indianapolis, Indianapolis, USA;
3Department of Biomedical Engineering, Rensselaer Polytechnic Institute, New York, USA
Received 10 Octorber 2010; revised 18 Octorber 2010; accepted 20 Octorber 2010.
ABSTRACT
Blotting is a common technique widely used for mo-
lecular analysis in life sciences. The Western blot, in
particular, is a process of transferring protein sam-
ples from a polyacrylamide gel to a blotting mem-
brane and detecting the levels of specific proteins
through reactions with primary and secondary anti-
bodies. The state-of-the-art of Western blotting usu-
ally generates one blotting membrane per gel. How-
ever, multiple copies of blots are useful in many ap-
plications. Two blotting copies from a single protein
gel, for instance, can be used for identifying a total
amount of proteins of interest as well as its specific
subpopulation level such as a phosphorylated isoform.
To achieve this multi-blotting operation from a single
gel, we modified a blotting procedure and developed
a novel blotting device. The device consisted of a
multi-anode plate and a microcontroller. It was de-
signed to generate a well-controlled electrophoretic
voltage profile, which allowed a quasiuniform trans-
fer of proteins of any size. The prototype device was
built and its operation procedure was described. The
experimental results clearly supported the notion
that the described device was able to achieve multiple
blotting from a single gel and reduce time and cost
for protein analysis.
Keywords: Western Blotting; Protein Transfer;
Multi-Blotting; PWM
1. INTRODUCTION
For detecting the amounts of specific proteins in various
biological samples, the Western blot is commonly used
among life scientists [1]. This blotting technique utilizes
gel electrophoresis to fractionate native or denatured
proteins based on their migration speed (mobility) in a
gel under an electrical field. The proteins, trapped and
size-fractionated in a gel, are transferred and immobi-
lized to a positively-charged membrane [2-4]. Using
primary antibodies specific to the target proteins as well
as secondary antibodies for signal amplification and vi-
sualization, expression and modification of the target
proteins can be investigated. Although the procedure is
well established and many tools are available for gel
fractionation and blotting, one of the bottlenecks is its
limited efficiency and controllability in the blotting pro-
cedure.
In most applications of Western blotting, there is a
need to evaluate expression levels of multiple protein
targets. Signal transduction pathways, for instance, are
often activated by protein modifications such as glyco-
sylation and phosphorylation [5,6]. It is important to
determine the amount of signaling molecules that are, in
many cases, phosphorylated. However, the current blot-
ting technique can generate only a single membrane per
gel and thus life scientists usually have to prepare multi-
ple gels. When the results from multiple gels are com-
pared, there is a concern about potential variations
among gels. Furthermore, since the preparation of pro-
tein samples, running gels, and transferring to mem-
branes are time consuming and costly, it is desirable to
develop a blotting procedure that enables multiple blot-
ting from a single gel. Such a procedure can reduce po-
tential inconsistencies during sample loading, electro-
phoresis, and blotting, which are affected by varying
factors including operational time, temperature, and
chemical compositions.
A specific aim of this study was to develop a novel
protein blotting device with two goals. First, the blotting
system would be able to generate three blotting mem-
brane per gel with equal quality (multi-blotting). Second,
it would allow transferring proteins with varying sizes
(for instance, 20 to 150 kD) in the same percentage rate
in the same blotting procedure without altering transfer-
ring time (uniform transferring).
In order to satisfy the above design goals, we first
evaluated efficiency of protein transport by establishing

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1126
a cumulative signal intensity function, P(t), as a function
of transferring time, t. For instance, P(t1) = 1/3 implies
that one third of proteins is transferred for a duration of
t1. We then derived an electrophoretic voltage profile
along the direction of protein migration, in which one
third of proteins regardless of their molecular size would
be transferred in t1. This voltage profile was applied to a
multi-anode plate using a microcontroller based pulse
width modulated (PWM) voltage generator [7].
2. MATERIALS AND METHODS
2.1. Western Blotting
2.1.1 Protein Samples
MC3T3 osteoblast-like cells were cultured in MEM me-
dium containing 10% fetal bovine serum and antibiotics
(50 units/ml penicillin and 50 μg/ml streptomycin; Invi-
trogen) [8]. Cells were incubated at 37℃ in a humid
chamber with 5% CO2 and prepared for experiments at
70-80% confluency. Protein samples were isolated by
sonicating cells using a sonic dismembrator (Model 100,
Fisher Scientific) and lysed in a RIPA lysis buffer [9].
2.1.2 Electrophoresis and Semi-Dry Transferring
Isolated proteins were size-fractionated using 10% SDS
gels (1.5 mm thickness). Electrophoresis was conducted
using 100 V for 10 min (stacking), followed by 150 V
for various durations (separation). Proteins, immobilized
in SDS gels, were electro-transferred to Immobilon-P
membranes (Millipore) using a semi-dry transferring
apparatus (BioRad). The standard transferring condition
was 15 V for 40 min. To evaluate the effects of transfer-
ring conditions such as voltages and running times, va-
rying transferring conditions were also examined.
2.1.3. Antibody Reactio ns a nd Im a ge Analysis
The Immobilon-P membrane after protein blotting was
incubated overnight in a blocking solution (1% milk in a
PBS buffer). The membrane was then incubated for 1
hour with monoclonal β-actin antibodies (Sigma) fol-
lowed by 45 min incubation with goat anti-mouse IgG
conjugated horseradish peroxidase (Amersham Biosci-
ences). The protein levels were assayed using an ECL
Western blotting detection kit (Amersham Biosciences),
and signal intensities were quantified using a lumines-
cent image analyzer (LAS-3000, Fuji Film) [10].
2.2. Protein Multi-Blotting Device
2.2.1. Predicti on of the Bl o tt i ng Voltage Profile
In order to uniformly transfer proteins regardless of their
size in given transferring time, our approach was to re-
gulate blotting voltages for individual proteins. Based on
well-known observation in size fractionation, we as-
sumed that the mobility of proteins was proportional to
an applied electrophoretic voltage and transferring time.
Through experimentation, we predicted the blotting vol-
tage profile along the direction of protein migration to
achieve a uniform transfer of proteins of any size. In the
described blotting device, a protein with a normalized
mobility of r would receive a local blotting voltage that
was inversely proportional to r (Appendix A).
2.2.2. D es ign of th e M ultiple Anodes and the
Microcontroller Based PWM Generator
The described multi-blotting device has a plate with
multiple anodes (Figure 1). With a properly regulated
blotting voltage profile, proteins of the identical mo-
lecular size would receive the same voltage, while larger
proteins would be transferred with the higher voltage.
Here, we used a single DC voltage source and a com-
puter controlled voltage generator to provide voltages
for all anodes (Figure 2).
This generator was designed to provide multiple pulse
width modulated (PWM) DC signals of different duty
cycles for all anodes, in which each PWM signal was
individually amplified by a PWM amplifier and applied
to one anode. The voltage amplification gains are the
same to all PWM signals. The duty cycle was set to
make the average voltage of an amplified PWM equal to
Figure 1. Schematic illustration of the multi-anode plate with a
gel. The black stripes designate an array of anodes, and the gel
contains size-fractionated proteins (white circles – not in
scale).The top anodes receive higher voltage than the bottom
anodes.
Figure 2. Systems block diagram for PWM voltage control in
multi blotting, in which “n” designates the number of anodes.

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1127
the desired driving voltage of the corresponding anode.
Assuming the amplification gain is G, the peak voltage
of all PWM after amplification is V, and the desired the
voltage for the anode i is vi, the duty cycle of the voltage
applied to anode i is [vi / (V/G)]*100%.
3. RESULTS
3.1. Experimental Evaluation and Validation
3.1.1. Blotting Time and Signal Intensities
Prior to generating multiple blotting membranes from a
single gel, we first conducted semi-dry blotting using
various transferring times and built a relationship of
signal intensities to transferring time. Using β-actin as an
example, representative blotting images are shown for
blotting for 5, 13, 20, 40, 60, 90, and 120 min at 15 V
(Figure 3(a)). Note that two protein bands in each image
were generated using the identical conditions. The signal
level after reactions with antibodies was initially in-
creased as the blotting time was lengthened. However,
the level was saturated for the transfers for 40 min or
longer durations.
The normalized signal intensities showed that this
temporal signal profile was approximated by a cumula-
tive Gaussian distribution function (mean = 22 min, and
standard deviation = 9 min) (Figure 3(b)). This cumula-
tive distribution function provided the basis for deter-
mining transferring time for each of the multiple blotting
membranes. Note that if this cumulative distribution
function is a linear function of transferring time, the
same blotting time is applied to all membranes. The
Gaussian distribution implies that for generating 3
membranes, the 1st and the 3rd transferring durations are
identical. The 2nd transferring duration is calculated 83%
of the standard deviation (approximately 7.5 min in our
experimental setup).
3.1.2. Multi pl e Transfers
Based on the relationship between blotting time and
signal intensity in the previous experiment, we next
conducted multiple transfers (generation of 3 blotting
membranes) using a single gel. Focusing on β-actin, the
transfer time was chosen to be 15 min (1st transfer), 7
min (2nd transfer), and 15 min (3rd transfer). The result
clearly shows that by properly choosing blotting dura-
tions it is possible to generate multiple blotting mem-
branes, although there are variations in signal intensity
among three membranes (Figure 4(a)). In each image,
two protein bands were processed using the same ex-
perimental condition.
3.1.3. Transfer under a Non-Uniform Voltage Profile
The experiments above support our first design goal for
achieving multiple blotting membranes from a single gel.
To examine the second design goal of simultaneously
(a)
(b)
Figure 3. Signal intensities of β-actin with various blotting
durations. (a) Representative blotting images after semi-dry
transferring for 5, 13, 20, 40, 60, 90, and 120 min. (b) Normal-
ized signal intensities as a function of 15 blotting durations.
The maximum intensity is set to 1.
(a)
(b)
Figure 4. Multiple transfers and a transfer using a non-uniform
voltage profile. (a) Blotting images of three transfers. The
transfer time was 14.5 min (1st transfer), 6.5 min (2nd transfer),
and 15 min (3rd transfer). (b) Differential transfer efficiency
using a non-uniform voltage profile. The white boxes indicate
the expected positions of -actin bands, in which 4 different
voltages are employed (12.0, 3.9, 2.3, and 1.7 V from left to
right).
transferring proteins with various sizes, we conducted
40-min semi-dry blotting using a non-uniform voltage
profile (Figure 4(b)). In this experiment all conditions
were identical to the condition employed for generating

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1128
the results depicted in Figure 3 except the voltages. The
voltages applied to four different bands (left to right)
were 12.0 V, 3.9 V, 2.3 V, and 1.7 V, respectively. The
result using β-actin as an example demonstrates that
transfer efficiency is controllable by an applied electro-
phoretic voltage. The left position under 12 V exhibits a
clear protein band, but the right position under approxi-
mately 1.7 V barely shows a detectable protein band.
3.2. Protein Multi-Blotting Device
3.2.1. D es ign and Prototypi ng the Mult i-Anode Plate
To demonstrate the feasibility of our design principle, a
prototype device was made. A computer printed circuit
board was etched to make a multiple anodes plate, where
the width of each anode stripe was set to be 4 mm and
the width of the gap between two anodes 1 mm (Figure
5). The size of both anode and cathode plates were 18
cm × 24 cm, and the prototype device was designed to
accommodate gels with size up to14 cm × 17 cm.
3.2.2. Design and Prototyping Multi-Blotting Control
Circuits
A PICF184515 microcontroller-based voltage generator
was designed and prototyped to provide the designated
voltage profile to individual anodes (Figure 6). This
microcontroller generated PWM voltage signals were
Figure 5. Prototype of a multi-anode plate. The vertical, paral-
lel blue lines indicate the direction of protein migration. The
anodes are arranged horizontally, and the top arrays receive
higher voltage than the bottom arrays.
amplified (Figure 7). The average voltage of the ampli-
fied PWM signal corresponded to the desired voltage for
the anode.
The maximum voltage of all PWM signals, generated
by the microcontroller, was 5 V with the current of 100 μA.
Since the PWM signals did not provide the necessary
power for driving a protein transfer, these signals were
amplified using a TC 4469 MOSFETS amplifier. The
maximum voltages of all PWM signals were set to 15 V.
Figure 8 shows the photograph of the prototype.
During blotting, the microcontroller device was de-
signed to request a user to provide the conditions used
for protein fractionation with the vertical gel as well as
the conditions for protein blotting. The former fractiona-
tion conditions included length and thickness of the gel,
electrophoretic voltage, and running time. The latter
transferring conditions were the power supply voltage,
the number of membranes to be blotted, and buffer solu-
tion type used in horizontal blotting and vertical gel. The
program in the microcontroller calculated the voltage
levels needed for each anode and the total time required
for transferring proteins. Based on the mean and the s.d.
values of Gaussian distributions, the blotting time re-
quired for each membrane was determined. The program
gave a sequence of instructions and guide to a user such
as “place the membrane on the device,” “start blotting,”
“wait,” “change the membrane,” etc.
4. DISCUSSION
We described a novel protein multi-blotting device con-
sisting of the multi-anode plate and the microcontrol-
ler-based voltage-regulating circuits. In experimental
evaluations, we modeled blotting signal intensities as a
cumulative Gaussian distribution function of transferring
durations and demonstrated that it was possible to gen-
erate 3 blotting membranes from a single gel. In the de-
vice development phase, an appropriate electrophoretic
voltage profile was derived and implemented using the
multi-anode plate and the microcontroller based PWM
voltage generator. The results using β-actin as an exam-
ple supported the notion that the described device was
capable of providing superior quality for comparing the
level of various proteins, reducing the required amount
of samples, time, and cost for protein blotting.
In generating a transferring voltage profile, we as-
sumed that protein mobility was proportional to applied
voltage and transferring time. In order to regulate dif-
ferential mobility of proteins with varying sizes, the new
multi-blotting device generated a well-controlled voltage
profile in which individual proteins received transferring
voltage inversely proportional to its intrinsic mobility.
Although we presented this particular design with multi-
anodes, it would be possible to build a device using

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1129
Figure 6. Microcontroller circuits for generating a series of PWM signals for establishing a well-controlled voltage profile to an ar-
ray of anodes.
JBiSE

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1130
Figure 7. Amplifier circuit for PWM voltage control.

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1131
JBiSE
Figure 8. Prototype device for multi bloting. The photograph
shows the multi anode plate and the microcontroller based
PWM voltage generator.
multi-cathodes and a common anode.
In predicting transferring time for each of the multiple
membranes, we built a protein mobility model using a
Gaussian function. Although average mobility can be
characterized by blotting voltage and running time, any
proteins of interest with same size present a variation in
their mobility. Based on our experimental result, this
variation was approximated by a Gaussian distribution
and the amount of proteins transferred to a membrane
was predicted from a cumulative Gaussian distribution.
By determining the mean and s.d. values, it was straight-
forward to divide the transfer process into multiple seg-
ments, each of which would receive the equal amount of
proteins of interest.
In our initial prototype design, copper was used as
material for the anode and stainless steel for the cathode.
We observed that electrical platting occurred and a no-
ticeable amount of cooper was transferred along with
proteins to the cathode. Other materials such as stainless
steel or platinum-titanium alloy should be used to avoid
the observed electrical platting during the blotting proc-
ess.
In predicting and evaluating the capability of the de-
vice, we employed commonly used materials and condi-
tions such as 10% SDS gel, a transfer buffer with 70%
methanol, etc. Fine tuning is necessary in determining
the voltage profile and running time when other materi-
als and conditions are used. Although we focused on the
operation of the described device for semi-dry blotting,
the same principle should apply to dry blotting.
5. CONCLUSION
This paper described the novel multi-blotting device that
enabled transferring proteins of any sizes to multiple
blotting membranes from a single gel. Two design goals
were multiple blotting and a quasi-uniform transfer of
most proteins in 20-150 kD. These goals were achieved
by the selection of transferring time for each of the mul-
tiple blotting membranes, and the alteration of blotting
voltages for individual proteins with different sizes. The
device was designed and prototyped using the multi-
anode plate and the microcontroller based voltage gen-
erator, and a well-controlled voltage profile was gener-
ated.
6. ACKNOWLEDGEMENTS
The authors appreciated technical support of Chang Jiang.
REFERENCES
[1] Towbin, H., Staehelin, T. and Gordon, J. (1979) Electro-
phoretic transfer of proteins from polyacrylamide gels to
nitrocellulose sheets procedure and some applications.
Proceedings of the National Academy of Sciences (USA),
76, 4350-4354.
[2] Chen, H. (2003) Semi-dry electroblotter. US Patent
6,592,734 B2.
[3] Adcock, M. (1990) Field inversion elecroblotting &
electroElution. US Patent 4959133.
[4] Zahringer, H. (2008) Product survey: Blotting equipment,
low-tech at its best. lab times, May, 55-59.
[5] Ajit, V. (2009) Essentials of glycobiology. 2nd Edition,
Cold Spring Harbor Laboratories Press.
[6] Tarrant, M.K. and Cole, P.A. (2009) The chemical biol-
ogy of protein phosphorylation. Annual Review of Bio-
chemistry, 78, 797-825.
[7] Quercioli, V. (1993) Pulse width modulated (PWM)
power supplies, Vol. 45(Studies in Electrical and Elec-
tronic Engineering). Elsevier, Amsterdam.
[8] Wang, D., Christensen, K., Chawla, K., Xiao, G., Krebs-
bach, P.H., and Franceschi, R.T. (1999) Isolation and
characterization of MC3T3-E1 preosteoblastsubclones
with distinct in vitro and in vivo differentiation/miner-
alization potential. Journal of Bone Mineral Research, 14,
893-903.
[9] Hirasawa, H., Jiang, C., Zhang, P., Yang, F.C. and Yokota,
H. (2010) Mechanical stimulation suppresses phos-
phorylation of eIF2and PERK -mediated responses to
stress to the endoplasmic reticulum. FEBS Letters, 584,
745-752.
[10] Hamamura, K. and Yokota, H. (2007) Stress to endo-
plasmic reticulum of mouse osteoblasts induces apop-
tosis and transcriptional activation for bone remodeling.
FEBS Letters, 581, 1769-1774.

A. M. Hagyousif et al. / J. Biomedical Science and Engineering 3 (2010) 1125-1132 1132
APPENDIX
For a protein transferring in a gel, its moving distance, d,
can be modeled to be proportional to the electrophoretic
voltage, v, and blotting time, t:
d = kvt (1)
with k as a proportional factor. Suppose that for a fixed
time, T, a moving distance is expressed as d = rL, where
L=length of the gel, and r = normalized mobility ratio
between 0 (no mobility) and 1 (maximum mobility). In
the described apparatus, any proteins have the same mo-
bility along the thickness of the gel for a constant time
with a graded electrophoretic voltage for individual pro-
teins. Assume that a largest protein of interest moves dmin
at vmax with dmin = rminL. Then, v for any protein with a
normalized mobility ratio of r can be determined:
v = vmax*(rmin /r) (2)
In summary, the graded electrophoretic voltage in the
described apparatus is regulated to be inversely propor-
tional to its normalized mobility ratio.