Expanding the substrate tolerance of biotin ligase through exploration of enzymes from diverse species

Article abstract of DOI:10.1021/ja076655i

Technologies that enable the site-specific conjugation of chemical probes onto proteins are extremely useful for applications in cell biology and proteomics. We cloned, expressed, and purified biotin ligases from nine different species and screened them for the ability to ligate unnatural analogues of biotin onto the human p67 biotin acceptor domain. We discovered that the biotin ligases of Saccharomyces cerevisiae (yeast) and Pyrococcus horikoshii could accept alkyne and azide derivatives of biotin, respectively. HPLC, gel-shift, and mass spectrometry assays confirmed that these ligation reactions were ATP- and enzyme-dependent, as well as site-specific. We used the P. horikoshii-catalyzed azide ligation reaction to site-specifically introduce a phosphine probe onto p67 using the Staudinger ligation. These new ligation reactions demonstrate the differential substrate specificities of biotin ligases from different organisms and open the door to novel protein labeling applications. Copyright

Full text of DOI:10.1021/ja076655i

Published on Web 01/03/2008
Expanding the Substrate Tolerance of Biotin Ligase through Exploration of
Enzymes from Diverse Species
Sarah A. Slavoff, Irwin Chen, Yoon-Aa Choi, and Alice Y. Ting*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
Received September 4, 2007; E-mail: ating@mit.edu
Escherichia coli biotin ligase (BirA) has been harnessed for
numerous biotechnological applications, including protein labeling,1-3
purification,4,5 and immobilization.6 BirA catalyzes the ATP-
dependent covalent ligation of biotin onto a lysine side chain of a
15-amino acid recognition sequence called the “acceptor peptide”
(AP).7 Previously, we showed that BirA could be applied to site-
specific protein labeling on the surface of living cells by making
use of a ketone isostere of biotin that could be ligated to AP fusion
proteins, then derivatized with hydrazide or hydroxylamine probes.8
Since that work, we have focused on expanding the small-molecule
specificity of BirA to incorporate other useful functional groups,
such as azides and alkynes, which can be derivatized with even
greater chemoselectivity than the ketone.9 Because we found that
BirA does not ligate diverse structural analogues of biotin to the
AP, in this study, we explored the activities of biotin ligases from
nine other organisms. We discovered that yeast biotin ligase accepts
an alkyne derivative of biotin, while Pyrococcus horikoshii biotin
ligase utilizes both alkyne and azide biotin analogues. These new
ligation reactions demonstrate the differential substrate specificities
of ligases from different species and open the door to novel protein
labeling applications.
Due to our interest in technologies for targeting chemical probes
to proteins in the cellular context, and the previous work of our
lab in exploiting E. coli BirA for this purpose,1,8 we synthesized
or purchased each of the biotin analogues shown in Figure 1.
Desthiobiotin azide (DTB-Az) 2 and propargyl biotin (PB) 3, whose
syntheses are shown in Figure 2, contain unique functional group
handles. Azides are naturally absent from cells and can be
selectively derivatized with strained alkynes or phosphines under
physiological conditions,9 while alkynes, also absent from cells,
Figure 1. Screening biotin ligases against biotin analogues. Top, structures
can be selectively derivatized with azides via [3 + 2] cycloaddition
of biotin analogues used in this study. Bottom, table showing the hits from
in the presence of a copper catalyst.10 Iminobiotin 4 and diamino-
screens: 1 µM of each enzyme was incubated with 1 mM probe and 100
biotin 5 exhibit pH-dependent binding to streptavidin and can be
used for protein purification.11 Nitrobenzoxadiazole γ-amino butyric
acid 6 is a fluorophore, and iodouracil and thiouracil valeric acid
probes (7 and 8) are photoactivatable cross-linkers.12
µM p67 acceptor protein for 14 h at 30 °C. Formation of product (indicated
by “+”) was detected by HPLC or native gel-shift assay. For screening,
probe 3 (PB) was provided as a mixture of cis and trans isomers.
We found that wild-type BirA does not ligate analogues 2-8 to
the AP (data not shown). Because precedent exists for differential
substrate specificity among homologous enzymes from different
species,13,14 we decided to clone, express, and evaluate biotin ligases
from nine other species. All organisms express one or two biotin
ligases,15,16 which attach biotin to protein domains involved in
carboxyl group transfer. In some organisms, the ligase plays an
additional role in transcriptional regulation of biotin biosynthesis.17
To select our panel of novel biotin ligases, we first noted that
biochemical and/or structural data were available for the human,18
Saccharomyces cereVisiae (yeast),19 Pyrococcus horikoshii,16 and
Bacillus subtilis20 biotin ligases. We then selected five additional
evolutionarily distant species whose biotin ligase genes were
annotated in their sequenced genomes (Methanococcus jannaschii,
Leuconostoc mesenteroides, Trypanosoma cruzi, Giardia lamblia,
and Propionibacterium acnes). The biotin ligase genes from these
organisms were cloned into a bacterial expression vector and
overexpressed in E. coli. All enzymes were obtained in reasonable
yield and purity (Figure S1A), except for the human enzyme, which
degraded significantly during isolation. Nevertheless, biotinylation
activity was measurable for the human enzyme (vide infra), so we
made use of this sample.
To assay the biotin ligases, we required a protein or peptide
substrate. We and others previously found that the AP is recognized
only by E. coli BirA and not by biotin ligases from several other
species.4,21 However, it has been observed that endogenous biotin
acceptor proteins generally display cross-reactivity with biotin
ligases from other species.22 We expressed and purified a domain
of one of the endogenous biotin acceptor proteins of human biotin
ligase, called p67.23 By comparing the rates of biotin ligation to
9
1160
J. AM. CHEM. SOC. 2008, 130, 1160-1162
10.1021/ja076655i CCC: $40.75 © 2008 American Chemical Society

C O M M U N I C A T I O N S
Figure 3. HPLC detection of (A) DTB-Az and (B) cis-PB ligation to p67
acceptor protein by PhBL and yBL, respectively. Negative controls are
shown with ATP or enzyme omitted. The biotin-p67 conjugate has a shorter
retention time than the DTB-Az or cis-PB conjugates. The starred peaks
were collected and analyzed by mass spectrometry in a separate experiment
(Figure S3).
Figure 2. Synthetic routes to (A) desthiobiotin azide and (B) cis-propargyl
biotin. MeOH, methanol; AcCl, acetyl chloride; Me, methyl; TFAA,
trifluoroacetic anhydride; TEA, triethylamine; MsCl, mesyl chloride.
p67 for all the biotin ligases under identical conditions, we
confirmed that all enzymes except the G. lamblia ligase recognized
p67 (Figure S1B). The biotinylation rates measured under these
conditions spanned a 700-fold range of activity.
Using p67 as substrate, we assayed the eight new biotin ligases
with biotin and the seven analogues shown in Figure 1. In the
assays, formation of probe-p67 conjugate was detected either by
a change in the retention time of p67 on reverse-phase HPLC or
by a shift in p67 mobility on a native polyacrylamide gel.24 None
of the ligases incorporated probes 4-8 (Figure 1). However, we
detected product in the reactions of P. horikoshii biotin ligase
(PhBL) with DTB-Az and PB (a mixture of cis and trans isomers),
as well as yeast biotin ligase (yBL) with PB (Figure 1). Under
identical conditions, the reaction of PB was much faster with yBL
than with PhBL (data not shown), so we proceeded to determine
which regioisomer of PB was preferred by yBL. The cis and trans
isomers were separated by HPLC, then tested with yBL using the
native gel-shift assay. Using a reaction with biotin as a positive
control for product formation and mobility shift, we observed no
product with trans-PB but complete conversion to product after
14 h with cis-PB (Figure S2).
HPLC assays showed that both the PhBL-catalyzed ligation of
DTB-Az and the yBL-catalyzed ligation of cis-PB were ATP- and
enzyme-dependent (Figure 3). The product peaks were purified by
HPLC and analyzed by mass spectrometry, which indicated that
exactly one copy of DTB-Az or cis-PB had been ligated to each
molecule of p67 by their respective enzymes (Figure S3). For
comparison, the product of yBL-catalyzed biotinylation of p67 was
also analyzed by mass spectrometry, and the product conjugate was
clearly distinguishable from the p67-cis-PB conjugate. We also
prepared a point mutant of p67, called p67(Ala), in which the lysine
61 modification site23 was mutated to alanine. Mass spectrometry
showed that p67(Ala) was not modified by either yBL or PhBL,
demonstrating the site-specificity of these labeling reactions (Figure
S3).
Figure 4. Functionalization of site-specifically ligated DTB-Az by
Staudinger ligation. PhBL was used to ligate DTB-Az to p67 acceptor
protein. The introduced azide was then functionalized with phosphine-
FLAG.25 The FLAG epitope was detected with anti-FLAG antibody staining
of the nitrocellulose membrane (R-FLAG). Negative controls show omission
of ligase, DTB-Az, or ATP from the reactions (lanes 2-4). Coomassie
staining (right) demonstrates equal protein loading in all lanes. Unmodified
p67 has a slightly lower molecular weight than FLAG-conjugated p67.
PhBL ligated DTB-Az to p67 at a rate of (1.34 ( 0.11) × 10-4
µM s-1, while biotin ligation occurred at a much faster rate of 0.20
( 0.02 µM s-1. We were able to measure the kcat values for yBL
ligation of cis-PB and biotin because it was possible to saturate
the yBL active site with 5 mM cis-PB (data not shown). We
obtained a cis-PB ligation kcat of (2.07 ( 0.10) × 10-2 s-1, and a
14-fold higher kcat for biotinylation of 0.28 ( 0.04 s-1. We
attempted to accelerate the DTB-Az ligation kinetics by cloning
one of P. horikoshii’s endogenous biotin acceptor proteins (the
biotinyl domain of acetyl-CoA carboxylase), but the rate of DTB-
Az ligation to this substrate was much slower than the rate with
p67 (data not shown).
Finally, to demonstrate the utility of the azide ligation reaction
for introduction of useful probes, we used PhBL to site-specifically
attach DTB-Az to p67, and then we functionalized the introduced
azide using a Staudinger ligation with a phosphine conjugate to
the FLAG peptide (Figure 4).25 Product was detected by immuno-
blotting with anti-FLAG antibody. Negative controls showed that
ATP, PhBL, and DTB-Az were all required for the FLAG
conjugation to p67.
To further characterize these ligation reactions, we compared
their kinetics to the kinetics of biotin ligation as catalyzed by the
same enzymes (Figure S4). Under identical conditions with 1 mM
probe (which was non-saturating for DTB-Az, data not shown),
In conclusion, by exploring biotin ligase enzymes from diverse
species, we discovered that azide and alkyne derivatives of biotin
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1161

C O M M U N I C A T I O N S
(2) Howarth, M.; Chinnapen, D. J.; Gerrow, K.; Dorrestein, P. C.; Grandy,
M. R.; Kelleher, N. L.; El-Husseini, A.; Ting, A. Y. Nat. Methods 2006,
3, 267-273.
(3) Hu, Q.; Tay, L. L.; Noestheden, M.; Pezacki, J. P. J. Am. Chem. Soc.
2007, 129, 14-15.
(4) Grosveld, F.; Rodriguez, P.; Meier, N.; Krpic, S.; Pourfarzad, F.;
Papadopoulos, P.; Kolodziej, K.; Patrinos, G. P.; Hostert, A.; Strouboulis,
J. Ann. N.Y. Acad. Sci. 2005, 1054, 55-67.
(5) de Boer, E.; Rodriguez, P.; Bonte, K.; Krijgsveld, J.; Katsantoni, E.; Heck,
A.; Grosveld, F.; Strouboulis, J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
7480-7485.
are accepted by P. horikoshii and yeast biotin ligases, respectively.
The crystal structures of PhBL in complex with biotin and the biotin
adenylate ester have recently been solved,16 but comparison to BirA
crystal structures26,27 reveals a high degree of structural similarity
in the biotin binding pockets and no insight into the differential
small-molecule substrate specificities of the two enzymes. No
structure is yet available for yeast biotin ligase.
Azides and alkynes are useful functional group handles that have
been widely exploited in chemical biology for protein,25,28-30
DNA,31-33 sugar,9,34 small-molecule,35 and virus tagging36 in vitro,
on the surface of living cells, and in living organisms. To truly
harness the power of azide and alkyne-based bio-orthogonal ligation
reactions, however, it is desirable to couple them with general
methodology for site-specific introduction of azides and alkynes
onto proteins or other biomolecules, particularly inside living cells.
We note that both DTB-Az and cis-PB are more hydrophobic than
biotin (based on HPLC retention times, Figure 3), suggesting that
they should be at least as membrane-permeable as biotin, which
crosses mammalian cell membranes by passive diffusion at
concentrations greater than 2 µM.37 In addition, at low concentra-
tions, biotin is actively transported across the membrane by the
sodium-dependent multivitamin transporter (SMVT).38 The SMVT
has been shown to interact with biotin analogues such as desthio-
biotin,38 so DTB-Az and cis-PB may be actively transported across
mammalian membranes as well.
(6) Scholle, M. D.; Collart, F. R.; Bay, B. K. Protein Expr. Purif. 2004, 37,
243-252.
(7) Beckett, D.; Kovaleva, E.; Schatz, P. J. Protein Sci. 1999, 8, 921-929.
(8) Chen, I.; Howarth, M.; Lin, W.; Ting, A. Y. Nat. Methods 2005, 99-
104.
(9) Agard, N. J.; Baskin, J. M.; Prescher, J. A.; Lo, A.; Bertozzi, C. R. ACS
Chem. Biol. 2006, 1, 644-648.
(10) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew.
Chem., Int. Ed. 2002, 41, 2596-2599.
(11) Green, N. M. Biochem. J. 1966, 101, 774-780.
(12) Meisenheimer, K. M.; Koch, T. H. Crit. ReV. Biochem. Mol. Biol. 1997,
32, 101-140.
(13) Ja¨ger, S.; Rasched, G.; Kornreich-Leshem, H.; Engeser, M.; Thum, O.;
Famulok, M. J. Am. Chem Soc. 2005, 127, 15071-15082.
(14) Kopp, M.; Rupprath, C.; Irschik, H.; Bechthold, A.; Elling, L.; Mu¨ller,
R. Chembiochem 2007, 8, 813-819.
(15) Chapman-Smith, A.; Cronan, J. E., Jr. J. Nutr. 1999, 129, 477S-484S.
(16) Bagautdinov, B.; Kuroishi, C.; Sugahara, M.; Kunishima, M. J. Mol. Biol.
2005, 353, 322-333.
(17) Streaker, E. D.; Beckett, D. Biochemistry 2006, 45, 6417-6425.
(18) Campeau, E.; Gravel, R. A. J. Biol. Chem. 2001, 276, 12310-12316.
(19) Polyak, S. W.; Chapman-Smith, A.; Brautigan, P. J.; Wallace, J. C. J.
Biol. Chem. 1999, 274, 32847-32854.
A future challenge will be to improve the kinetics of both ligation
reactions and to demonstrate their utility with peptide rather than
protein substrates (for instance, with the yBL acceptor peptide that
we recently discovered by phage display21). These efforts may be
accomplished through a combination of rational mutagenesis and
in vitro evolution.
(20) Bower, S.; Perkins, J.; Yocum, R. R.; Serror, P.; Sorokin, A.; Rahaim,
P.; Howitt, C. L.; Prasad, N.; Erhlich, S. D.; Pero, J. J. Bacteriol. 1995,
177, 2572-2575.
(21) Chen, I.; Choi, Y.-A., Ting, A. Y. J. Am. Chem. Soc. 2007, 129, 6619-
6625.
(22) Cronan, J. E., Jr. J. Biol. Chem. 1990, 265, 10327-10333.
(23) Leo´n-Del-Rio, A.; Gravel, R. A. J. Biol. Chem. 1994, 269, 22964-22968.
(24) Chapman-Smith, A.; Turner, D. L.; Cronan, J. E., Jr.; Morris, T. W.;
Wallace, J. C. Biochem. J. 1994, 302, 881-887.
Acknowledgment. We thank the NIH (R01 GM072670-01 and
1PN2EY018244), MIT, the Sloan Foundation, the Dreyfus Founda-
tion (Teacher-Scholar Award), and Pfizer-Laubach for supporting
this work. S.S. was supported by a Whitaker Health Sciences Fund
graduate fellowship. I.C. was supported by graduate research
fellowships from the NSF and Wyeth Research/ACS Division of
Organic Chemistry. Y.-A.C. was supported by a Samsung Lee Kun
Hee Scholarship. We thank John E. Cronan, Jr., Roy Gravel, and
Mark Howarth for helpful advice, and Tanabe U.S.A. for biotin.
(25) Kiick, K. L.; Saxon, E.; Tirrell, D. A.; Bertozzi, C. R. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 19-24.
(26) Wilson, K. P.; Shewchuk, L. M.; Brennan, R. G.; Otsuka, A. J.; Matthews,
B. W. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 9257-9261.
(27) Wood, Z. A.; Weaver, L. H.; Brown, P. H.; Beckett, D.; Matthews, B.
W. J. Mol. Biol. 2006, 357, 509-523.
(28) Dieterich, D. C.; Link, A. J.; Graumann, J.; Tirrell, D. A.; Schuman, E.
M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 9482-9487.
(29) Dieters, A.; Cropp, T. A.; Mukherji, M.; Chin, J. W.; Anderson, J. C.;
Schultz, P. G. J. Am. Chem. Soc. 2003, 125, 11782-11783.
(30) Liu, W.; Brock, A.; Chen, S.; Chen, S.; Schultz, P. G. Nat. Methods 2007,
4, 239-244.
(31) Gierlich, J.; Burley, G. A.; Gramlich, P. M.; Hammond, D. M.; Carell, T.
Supporting Information Available: Experimental protocols for
synthesis of DTB-Az and cis-PB, cloning and purification of proteins,
biotinylation assays, analogue screens, HPLC and ESI-MS assays, and
kinetic measurements. Sources of expression plasmid gifts. SDS-PAGE
analysis and biotinylation activity of all enzymes toward p67, use of
cis-PB vs trans-PB, ESI-MS of DTB-Az and cis-PB ligated to p67 with
controls, kinetics of DTB-Az and cis-PB use compared to biotin. This
material is available free of charge via the Internet at http://pubs.acs.org.
Org. Lett. 2006, 17, 3639-3642.
(32) Weisbrod, S. H.; Marx, A. Chem. Commun. 2007, 1828-1830.
(33) Chandra, R. A.; Douglas, E. S.; Mathies, R. A.; Bertozzi, C. R.; Francis,
M. B. Angew. Chem., Int. Ed. 2006, 45, 896-901.
(34) Agard, N. J.; Prescher, J. A.; Bertozzi, C. R. J. Am. Chem. Soc. 2004,
126, 15046-15047.
(35) Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 2003, 125,
4686-4687.
(36) Wang, Q.; Chan, T. R.; Hilgraf, R.; Fokin, V. V.; Sharpless, K. B.; Finn,
M. G. J. Am. Chem. Soc. 2003, 125, 3192-3193.
(37) Gore´, J.; Hoinard, C.; Maingault, P. Biochim. Biophys. Acta 1986, 856,
357-361.
References
(38) Park, S.; Sinko, P. J. Drug. Metab. Dispos. 2005, 33, 1547-1554.
(1) Howarth, M.; Takao, K.; Hayashi, Y.; Ting, A. Y. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 7583-7588.
JA076655I
9
1162 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008