Microfluidics and microarray based approaches to biological analysis 4

Chapter 4
CHAPTER 4 INTEIN-MEDIATED BIOTINYLATION OF PROTEINS AND
ITS APPLICATION IN A PROTEIN MICROARRAY

4.1 Introduction

4.1.1 Developing Microarrays of Functionally Active Proteins
DNA microarray is currently the method of choice for high-throughput analysis of
nucleic acids at their transcriptional level. However, it has been shown that the mRNA
expression level in a cell does not correlate well with the abundance of proteins.
1
To
gain more insights into protein functions, a number of techniques
2
and binding
chemistry have been developed to immobilize small molecules,
3,4
peptides,
5,6,7
and
proteins
8,9,10
in a microarray for high-throughput protein studies.
11

As mentioned in chapter 3, various strategies have been developed for the site-specific
immobilization of kinase substrates onto functionalized slides. However, proteins are
more difficult to handle than peptides. Indeed, they are delicate, denature in ‘harsh
environments’ and protein arrays are stable and useful only for a very short period of

time. Consequently, different approaches have been developed to ensure they retain
their activity. Schreiber et al dissolved proteins in a 60 % solution in order to keep
them hydrated;
8
whereas Zhu et al attached them to the surface of PDMS micro wells.
9

However, in most cases, protein immobilization was achieved via their nucleophilic
residues, resulting in random orientations of proteins on the glass surface, whereas an
oriented immobilization on the slide surface is desired for optimal protein activity.
Indeed, in order for arrayed proteins to retain their full biological activity, they need to
be arrayed in a proper orientation to ensure accessibility of their active sites with
interacting molecules.

98


Chapter 4
Thus far, there has only been one report of site-specific attachment of proteins on glass
slides.
12
Approximately 6000 yeast proteins were expressed as His-tag fusions, spotted
onto Ni-NTA functionalized slides, and >80 % were found to retain their full
biological activities, presumably as a result of site-specific immobilization which
ensures most proteins on the slide to be oriented correctly. However, the binding
between Ni-NTA and His-tag proteins is neither very strong, nor very stable and
susceptible to interference by many commonly used chemicals,
13
making this
immobilization method incompatible with many protein screening assays.

Besides the site-specific immobilization, another problem when working with protein
microarrays is the high throughput expression and purification of proteins. So far, all
protein arrays involve commercially available proteins or recombinant proteins.
9
In the
second case, proteins need to be purified before spotting, implying long, and labor-
intensive protocols. The in vitro expression of proteins appears as a promising solution
and a protein array of in situ synthesized proteins has been reported.
14
He at al
developed a new protein array production procedure termed PISA (protein in situ
array). PISA is designed to generate protein arrays directly from PCR generated DNA
via cell-free protein synthesis and simultaneous in situ immobilization of the generated
proteins on a surface. However, this array consists of microwells rather than a real
array and this strategy has not been applied to the array format yet. In addition, post
translation modifications may be lacking for in vitro synthesized proteins, and new
strategies are required for the in vivo synthesis, purification and functionalization of
proteins for site-specific immobilization.




99


Chapter 4
4.1.2 Site-Specific Protein Biotinylation
On the contrary to the interaction between His tag and Ni-NTA previously used for
protein microarrays, the biotin-avidin interaction is one of the strongest known non-
covalent interaction.

15
It is very stable toward a variety of harsh conditions,
16
and has
been widely used in standard biochemical assays for immobilization purposes. The in
vivo and in vitro biotinylation of proteins have previously been reported,
17,18
but with
limited success due to low yields and non-specific nature of the biotinylation
reaction.
19
In some cases biotinylation resulted in the addition of a long peptide to the
target protein, which may interfere with proper folding of the protein.
20
Furthermore,
avidin is known to be toxic to cells, making the expression of avidin-fused proteins a
difficult task.
21


4.1.3 Intein-Mediated Protein Engineering
Inteins are naturally occurring proteins that are involved in the precise cleavage and
formation of peptide bonds in a process known as protein splicing. The mechanism of
protein splicing (Figure 4.1) was elucidated using a combination of site directed
mutagenesis and chemical analysis and allowed the rational engineering of inteins for
use in protein chemistry. The process of protein splicing involves the excision of an
intervening protein sequence (the intein) from a precursor protein with the concomitant
fusion of the two flanking protein regions through a native peptide bond. Controlled
cleavage at single intein splice junctions led to the development of fusion protein
purification on chitin columns.

22

In so-called “native chemical ligation”
23
described in chapter 3, an N-terminal cysteine
containing peptide is chemically ligated to a second peptide possessing a thioester
group with the resultant formation of a native peptide bond at the ligation junction.

100


Chapter 4
Peptide thioesters for use in native chemical ligation are generated solely through
chemical synthesis. The intein-based purification protocol has found wide applications
in protein engineering where the expressed protein ligation (EPL) strategy is utilized to
incorporate non coded amino acid sequences into a protein sequence (Figure 4.2 a).
24,25

This strategy has been extended to modifications of proteins at their C-termini with a
number of chemical tags.
26
Recently, Tolbert et al reported the use of TEV protease to
generate N-terminal cysteines from affinity-tagged fusion proteins (Figure 4.2),
27
but
unfortunately, the simultaneous protease cleavage and thioester labeling of the affinity-
tagged fusion proteins was unsuccessful since the thioester labels are inhibitors of the
TEV protease, possibly because the TEV protease is a cysteine protease with an active
site cysteine that can be acylated by the thioesters.

HN
HS
HS
H
2
N
S
HS
O
N
H
O
S
NH
2
O
O
Intein
N-extein
C-extein

Intein
O
O
N
H
O
S
NH
2
O
O
H
2
N
S
N
H
HS
O
Figure 4.1. Principle of intein-mediated protein splicing.



101


Chapter 4

+





Protein
Intein
Protein
Intein
R
Protein
R
His
6
ENLYFQHis
6
ENLYFQ
Protein
Protein
R
Protein
R
N
H
O
HS
O

+
NH
3
O
O-
NH
3
O
O-
S
O
NH
3
+
O
HS
+NH
3
O
+
NH
3
N
H
O
HS
O
+
NH
3

N
H
HS
O
+
H
3
N
HS
O
N
H
HS
O
O
O
O-
O
O-
O
O-
a
b
SR
O
Figure 4.2. Generation of N-terminal cysteine proteins using TEV protease for (a)
expressed protein ligation, (b) Generation of N-terminal


4.2 Results and Discussion

4.2.1 Intein-Mediated Biotinylation of Proteins at their C-Terminal
1

As described in chapter 3, the native chemical ligation developed by Kent et al allows
for the site-specific reaction between the N-terminal cysteine and the thioester function
of any other compound. EPL represents a novel semisynthetic approach that has
greatly expanded the utility of native chemical ligation chemistry and allows for the
site-specific incorporation of non-coded amino acids into the protein of interest. By
expressing a protein of interest as fusion to an intein, which also contains a chitin-
binding domain for purification on chitin column, one can both purify and label at its
C-terminal by flushing the column with the cysteine containing labeling agent.


1
The expression and biotinylation of proteins was performed by Lue Yee Peng Rina, and cysteine biotin
was synthesized by Dr Zhu Qing

102


Chapter 4
Therefore, by reacting cysteine biotin with intein fused protein, one can in a single
step, purify and site-specifically biotinylate proteins at their C-terminal (Figure 4.3).

S
HS
S
O
O
O
S
O
Target protein Intein ta g
N
Chitin
column
N
N
a) In vivo
expression
c) Biotinylation
Spontaneous
rearrangement
DNA
b) Purification
N
Intein tag

SS
HS
S
O
O
O
S
O
Target protein Intein ta g
N
Chitin
column
N
N
a) In vivo
expression
c) Biotinylation
Spontaneous
rearrangement
DNA
b) Purification
N
Intein tag
H
2
N
N
H
H
N

SH
O
O
S
HN
NH
O
H
2
N
H
N
N
H
O
O
S
HN
NH
O
S
H
N
N
H
O
O
S
HN
NH

O
N
H
HS
H
2
N
N
H
H
N
SH
O
O
S
HN
NH
O
H
2
N
H
N
N
H
O
O
S
HN
NH

O
S
H
N
N
H
O
O
S
HN
NH
O
N
H
HS
Figure 4.3. Intein-mediated site-specific biotinylation of proteins

Three proteins of interest, namely MBP (Maltose Binding Protein), EGFP (Enhanced
Green Fluorescent Protein) and GST (Glutathione S-Transferase) were chosen as
models and expressed in vivo as fusion proteins with an intein tag (intein fused to
chitin binding domain) at their C-termini. The proteins were purified and biotinylated,
in a single step (Figure 4.3), by first loading the crude cell lysate onto a column packed
with chitin beads, then flushing the column with biotinylated cysteine (Figure 4.4), to
obtain the C-terminally biotinylated proteins.


103


Chapter 4

Figure 4.4. Cysteine Biotin used for Intein-Mediated Site-Specific Biotinylation


The site-specific biotinylation of the proteins was unambiguously confirmed by SDS-
PAGE (Figure 4.5, a) and western blotting (Figure 4.5, b). Based on SDS-PAGE, the
biotinylation reaction took place with 90-95 % efficiency, generating proteins in
sufficient purity (> 95 %). Labeling of proteins expressed as intein fusion is very
simple since the ligation reaction can be combined with affinity purification, allowing
C-terminally modified proteins to be obtained from crude bacterial lysates in a single
step.


1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10
(b)
NH
2
H
N
N
H
HS
O
O
S

NH
HN
O


(a)






Figure 4.5. MBP purification and biotinylation. (a) SDS-PAGE. (1) protein marker,
(2) uninduced cell extract, (3) induced cell extract, (4) flow-through from column
loading, (5) flow-through from column wash, (6) proteins bound to chitin column
before cleavage, (7) flow-through from quick flush of cleavage agent, (8-9) first two
elution fractions after overnight incubation at 4 °C with cysteine biotin (10) remaining
proteins bound to chitin column after cleavage. (b) Western blotting of biotinylated
MBP. Biotinylated MBP was run on SDS PAGE and after transfer was detected using
Streptavidin-HRP.

104


Chapter 4
4.2.2 Microarrays of Site-Specific Immobilized Active Proteins
After biotinylation, without any further purification, the three biotinylated proteins
were spotted directly, without any further treatment, onto an avidin-functionalized
slide to obtain the corresponding protein array. (Figure 4.6)

NNNN



O
Immobilization
STA slide
OO
Immobilization
STA slideSTA slide
OO
Immobilization
STA slideSTA slide
OO
Immobilization
STA slideSTA slide

H
N
N
H
O
O
S
HN
NH
O

N
H
HS
H
N
N
H
O
O
S
HN
NH
O
N
H
HS
H
N
N
H
O
O
S
HN
NH
O
N
H
HS
H

N
N
H
O
O
S
HN
NH
O
N
H
HS
H
N
N
H
O
O
S
HN
NH
O
N
H
HS
H
N
N
H
O

O
S
HN
NH
O
N
H
HS
H
N
N
H
O
O
S
HN
NH
O
N
H
HS
Figure 4.6. Site specific immobilization of functionally active proteins

A protein array was generated with the biotinylated EGFP, MBP and GST, and probed
with Cy3-anti-EGFP, Cy5-anti-MBP and FITC-anti-GST, respectively. Three
corresponding non-biotinylated proteins were also spotted onto the same slide, as
controls, and the array was incubated with either individual antibodies (Figure 4.7), or
a mixture of all three antibodies (Figure 4.8). Only specific binding between the
biotinylated proteins and their corresponding antibodies were observed, regardless of
the presence of other proteins (Figure 4.7) and antibodies (Figure 4.8), indicating the

specific immobilization and versatility of this new protein array. Furthermore, no
fluorescence signal was observed with the non-biotinylated control proteins (data not
shown), confirming the essence of biotinylation for protein immobilization.



105


Chapter 4

(a) (b) (c)(a) (b) (c)





Figure 4.7. Site-specific immobilization of proteins via avidin-biotin interaction.
Three biotinylated proteins : (a) EGFP, (b) MBP and (c) GST were arrayed onto avidin
functionalized slides and individually detected with Cy3-anti-EGFP (green), Cy5-anti-
MBP (red) and FITC-anti-GST (blue), respectively

(a) (b) (c)(a) (b) (c)






Figure 4.8. Site-specific immobilization of proteins via avidin-biotin interaction.

Three biotinylated proteins : (a) EGFP, (b) MBP and (c) GST were arrayed onto avidin
functionalized slides and detected with a mixture of Cy3-anti-EGFP (green), Cy5-anti-
MBP (red) and FITC-anti-GST (blue)


4.2.3 Arrays of Functionally Active Proteins
The most critical issue in generating a protein array is to ensure that proteins maintain
their native activity, as it is previously known that proteins tend to denature on glass
surfaces. In order to confirm that biotinylated proteins immobilized on the avidin slide
retain their proper folding, the native fluorescence of EGFP on the slide was monitored

106


Chapter 4
(Figure 4.9). No loss of fluorescence intensity was observed after prolonged incubation
at 4
0
C, suggesting that folding of the protein was properly maintained on the slide.







Figure 4.9. Fluorescence from the native EGFP

In a separate experiment, a slide immobilized with EGFP, MBP and GST was
incubated with Cy3-labeled glutathione (Figure 4.10), a known natural ligand of GST.

The result showed exclusive binding between GST and glutathione (Figure 4.11),
further indicating full retention of the native GST activity.

N
H
H
N
OH
O
H
2
N
OH
O
SH
O
O






Figure 4.10. Structure of glutathione



107

Chapter 4






Figure 4.11. Functional activity of arrayed biotinylated GST. Biotinylated GST was
arrayed onto avidin functionalized slides and incubated with its specific natural ligand,
glutathione, labeled with Cy3.


Furthermore, all data gathered thus far indicates that the presence of avidin as a
molecular layer between the immobilized proteins and the glass surface also serves to
minimize nonspecific absorption of proteins.

Future improvement may be readily made
by using streptavidin as the immobilization agent on the slide in place of avidin, which
is a glycoprotein and known to have higher nonspecific binding characteristics.
28


4.2.4 Comparison of Avidin-Biotin Interaction Stability with Other Existing Site-
Specific Immobilization Strategies
Thus far, the only reported method for site-specific attachment of proteins in a
microarray has been the immobilization of His-tag proteins on slides functionalized
with Ni-NTA.
12
However, the binding between His-tag proteins and Ni-NTA complex
is not very strong, and incompatible with many commonly used chemicals such as

DTT, SDS, EDTA, etc. The binding is also depleted outside the 4 to 10 pH range, or
when the buffer contains high concentrations of common salts.
On the contrary, as mentioned in chapter 3, the binding between biotin and avidin is
one of the strongest known non covalent interaction. Avidin is also extremely stable,
16


108


Chapter 4
making it an ideal agent for slide functionalization since glass slides can be prepared in
advance and stored before being spotted. In addition, the interaction between avidin
and biotin is instantaneous, hence requiring no incubation for protein immobilization.
In order to confirm the benefit of avidin-biotin linkage, slides immobilized with GST
were first subjected to a number of harsh washing conditions, and then detected with
FITC-labeled anti-GST for any loss of GST on the surface. No loss of GST was
observed even after the slide had been treated with 1M acetic acid at pH 3.3, or 60
0
C
water or even 4 M GuHCl for prolonged time (Figure 4.12). For comparison, we have
also prepared Ni-NTA slides according to published protocols.
29
On the contrary to the
avidin slides, when these GFP-spotted slides were treated with any of the above harsh
conditions, the immobilization of the his-tag protein on the Ni-NTA was completely
removed.


(a) (b) (c) (d)

Figure 4.12. Biotin avidin interactions subjected to harsh conditions. Biotinylated
GST was arrayed on four avidin functionalized slides. Three slides were soaked in (a)
acetic acid solution pH 3.3, (b) 60 ºC water, (c) 4 M GuHCl and (d) no treatment for
30 min, and probed FITC-anti-GST.


4.3 Summary
Findings described here present a new strategy for site-specific protein biotinylation
and immobilization on a glass surface, generating a novel protein array on which

109


Chapter 4
proteins are oriented optimally, and able to retain their native activity suitable for
subsequent biological screenings. The advantage of avidin/biotin linkage over His-
tag/Ni-NTA strategies for protein immobilization is highlighted by its ability to
withstand a variety of chemical conditions, which may make this new protein array
compatible with most biological assays.


4.4 Materials and Methods

4.4.1 Chemicals
All chemical were obtained from normal suppliers. The cysteine biotin was
synthesized in Prof Yao’s lab, chemistry Department.

30
The cloning of target genes
into pTVB1 expression vector and on-column biotinylation were done by Rina Lue in
Prof Yao’s lab, department of biological sciences, NUS.
31

The sequence specific monoclonal antibodies, anti-EGFP and anti-MBP, were
obtained from available commercial sources, and labeled with Cy3-NHS and Cy5-
NHS (Amersham Pharmacia, USA) respectively. The antibody was reacted with the
dye for one hour in 0.1 M NaHCO
3
, pH 9, according to manufacturer’s protocols and
purified with a NAP5 column (Amersham Pharmacia, USA). The anti-GST was
purchased as a FITC-conjugate (Molecular Probes, USA).

4.4.2 Spotting of Biotinylated Proteins on Avidin Functionalized Slides and
Fluorescent Antibody-Based Detection
Glass slides were cleaned in a piranha solution and derivatized with a 1 % solution of
3-glyicidoxypropyltrimethoxisilane (95 % ethanol, 16 mM acetic acid) for 1 hr and

110


Chapter 4
cured at 150 °C for 2 hours. The epoxy slides were reacted with a solution of 1 mg/mL
avidin in 10 mM NaHCO
3
for 30 minutes, washed with water, air dried, and the
remaining epoxides were reacted with a solution of 2 mM aspartic acid in a 0.5 M
NaHCO

3
buffer, pH 9. The biotinylated proteins were dissolved in PBS buffer, pH 7.4,
and spotted onto the avidin functionalized slides using an ESI SMA arrayer
(Toronto, Canada). No incubation was necessary before the slide were further
processed by washing with PBS and drying in air.
The spotted slides were incubated with the labeled antibody (or mixture of antibodies)
for 1 hour, washed 4 times, each time for 15 min with PBST (PBS + 0.1 % Tween 20),
dried and scanned with an ArrayWoRx microarray scanner (Applied Precision,
USA).

4.4.3 GST Assay
The N-terminal amine group of glutathione was selectively labeled with Cy3-NHS by
reacting the molecule overnight with the dye in sodium phosphate buffer at pH 7. The
reaction was subsequently quenched with ethanolamine for 12 hours to degrade the
remaining Cy3-NHS Biotinylated GST was arrayed on avidin slides. Following
incubation with the Cy3-labeled glutathione for one hour, the slide was washed with
PBST as described earlier, and the specific binding between GST and glutathione was
visualized with an ArrayWoRx microarray scanner.

4.4.4. Treatment of Avidin Slides Under Harsh Conditions
Biotinylated GST was arrayed on four avidin functionalized slides. Three slides were
soaked in (1) 1M acetic acid solution at pH 3.3, (2) 60 ºC water, (3) 4 M GuHCl and
(4) no treatment for 30 min, and probed with FITC-anti-GST. For comparison, we

111


Chapter 4

112

have also prepared Ni-NTA slides according to published protocols. Briefly, epoxy
slides were incubated with NTA dissolved in NaHCO
3
. The slides were washed in
water and soaked in 100 mM NiSO
4
for at least 1 hour, washed with 0.2 M acetic acid,
100 mM NaCl to give the Ni-NTA slides. GFP was expressed with a his-tag, and
spotted it onto Ni-NTA slides. When this GFP-containing slide was treated with any of
the above harsh conditions, the immobilization of the his-tag protein on the Ni-NTA
was completely removed


4.5 References

1 Gygi, S. P.; Rochon, Y.; Franza, R. B.; Aebersold, R. Mol. Cell. Biol. 1999, 19,
1720
2 Lee, K B.; Park, S J.; Mirkin, C. A.; Smith, J. C.; Mrksich, M. Science 2002,
295, 1702
3 MacBeath, G.; Koehler, A. N.; Schreiber, S. L. J. Am. Chem. Soc. 1999, 121,
7967
4 Hergenrother, P. J.; Depew, C.; Schreiber, S. L. J. Am. Chem. Soc. 2000, 122,
7849
5 Falsey, J. R.; Renil, S.; Park, S.; Li, S.; Lam, K. S. Bioconjugate Chem. 2001,
12, 346
6 Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat. Biotechnol. 2002, 20,
270
7 Melnyk, O.; Duburcq, X.; Olivier, C.; Urbès, F.; Auriault, C.; Gras-Masse, H.
Bioconjugate. Chem. 2002, 13, 713
8 MacBeath, G.; Schreiber, S. L. Science, 2000, 289, 1760

Chapter 4

113

9 Zhu, H.; Klemic, J. F.; Chang, S.; Betone, P.; Casamayor, A.; Klemic, K. G.;
Smith, D.; Gerstein, M.; Reed, M.; Snyder, M. Nat. Genet. 2000, 26, 283
10 Wang, D.; Liu, S.; Trummer, B. J.; Deng, C.; Wang, A. Nat. Biotechnol. 2002,
20, 275
11 Robinson, W. H. et al Nat. Med. 2002, 8, 295
12 Zhu, H.; Bigin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.; Lan, N.;
Jansen, R.; Bidlingmaier, S.; Houfek, T.; Mitchell, T.; Miller, P.; Dean, R. A.;
Gerstein, M.; Snyder, M. Science 2001, 293, 2101
13 Paborsky, L. R.; Dunn, K. E.; Gibbs, C. S.; Dougherty, J. P. Anal. Biochem.
1996, 234, 60
14 He, M.Y.; Taussig, M. J. Nucleic Acids Res. 2001, 29, e73
15 Green, N. M.; Toms, E. J. Biochem. J. 1973, 133, 687
16 Reznik, G. O.; Vajda, S.; Cantor, C. R.; Sano, T. Bioconjugate Chem. 2001, 12,
1000
17 Cull, M. G.; Schatz, P. J. Methods Enzymol. 2000, 26, 430
18 Cronan, J. E.; Ree, K. E. Methods Enzymol. 2000, 27, 440
19 Taki, M.; Sawata, S. Y.; Taira, K. J Biosci. Bioeng. 2001, 92, 149
20 Smith, P. A.; Tripps, B. C.; DiBlasio-Smith, E. A.; Lu, Z. LaVallie, E. R.;
McCoy, J. M. Nucleic Acids Res. 1998, 26, 1414
21 Sano, T.; Cantor, C. R.; Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 142
22 Xu, M Q.; Evans, T. C. Methods 2001, 24, 257
23 Dawson, P.E.; Muir, T.W.; Clark-Lewis, I.; Kent, S.B.H. Science 1994, 266, 776
24 Evans, T. C.; Benner, J.; Xu, M Q. J. Biol. Chem. 1999, 274, 3923
25 Muir, T. W.; Sondhi, D.; Cole, P. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,

6705


Chapter 4

114

26 Tolbert, T.; Wong, C H. J. Am. Chem. Soc. 2000, 122, 5421
27 Tolbert, T.J.; Wong, C H. Angew. Chem. Intl. Ed. 2002, 41, 2171
28 Savage, D.; Mattson, G.; Nielander, G.; Morgensen, S.; Conklin, E. Avidin-
Biotin Chemistry: A Handbook, 2nd Ed.; Pierce Chemical Co.
29 Hochuli, Dobeli, Schacher, J. Chrom. A. 1987, 411, 177
30 Lesaicherre, M L.; Lue R. Y. P.; Chen, G. Y. J. ; Zhu, Q., Yao, S.Q. . J. Am.
Chem. Soc. 2002, 124, 8768
31 “Expression and Purification of recombinant proteins with intein”, Honors
Thesis, Lue Yee Peng, Rina, 2002, National University of Singapore