Chemical Synthesis of HMGA1a Proteins with Post-translational Modifications via Ser/Thr Ligation

Article abstract of DOI:10.1021/acs.orglett.6b03056

The first chemical synthesis of nuclear protein HMGA1a via Ser/Thr ligation is reported. Notably, Hmb (2-hydroxy-4-methoxybenzyl) exhibits crucial improvement of both the difficult coupling during solid phase peptide synthesis and the poor ligation encountered in protein synthesis. These efforts led to preparation of HMGA1a analogs with well-defined phosphorylation and methylation patterns (9 synthetic proteins in total), thus overcoming the heterogeneous and combinatory problems inherent to protein post-translational modifications (PTMs), and facilitating the study of the regulatory roles of such PTMs.

Full text of DOI:10.1021/acs.orglett.6b03056

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Letter
pubs.acs.org/OrgLett
Chemical Synthesis of HMGA1a Proteins with Post-translational
Modifications via Ser/Thr Ligation
Tianlu Li, Heng Liu, and Xuechen Li*
Department of Chemistry, State Key Lab of Synthetic Chemistry, The University of Hong Kong, Pokfulam, Hong Kong
S
* Supporting Information
ABSTRACT: The first chemical synthesis of nuclear protein
HMGA1a via Ser/Thr ligation is reported. Notably, Hmb (2-
hydroxy-4-methoxybenzyl) exhibits crucial improvement of
both the difficult coupling during solid phase peptide synthesis
and the poor ligation encountered in protein synthesis. These
efforts led to preparation of HMGA1a analogs with well-
defined phosphorylation and methylation patterns (9 synthetic
proteins in total), thus overcoming the heterogeneous and
combinatory problems inherent to protein post-translational
modifications (PTMs), and facilitating the study of the regulatory roles of such PTMs.
n addition to histone proteins as chromosomal components,
the “high mobility group” (HMG) proteins (subfamilies:
The multifunctionality of HMGA1a is achieved through its
high adaptability.13 Apart from its intrinsically disordered status,
HMGA1a is subject to diverse post-translational modifications
(PTMs) by different kinases, acetyl- and methyltransferases,
and other modifying enzymes.14 Accumulating evidence has
indicated the regulatory roles of such PTMs, through which
HMGA1a is proposed to function as a “molecular switch”. A
well-studied example is the acetylation of HMGA1a at distinct
Lys residues (Lys64 and Lys70), coordinately regulated by two
histone acetyltransferases, CBP and P/CAF. HMGA1a
acetylation dynamically controls the transcription activation
process of interferon-β gene, giving rise to enhanced or
suppressed effects.15
Unfortunately, given its various PTMs and the extensive
biological processes, the PTM code of HMGA1a is yet to be
explored. For example, current studies indicated HMGA1a
phosphorylation as a biological event with various kinases
involved in a cell-type-specific and cell-cycle-dependent
manner;14 regretfully, the entire cellular network involving
HMGA1a phosphorylation remains elusive. Furthermore, these
results were mostly obtained either from nuclear extracts or
from in vitro kinase treatment of recombinant HMGA1a. Such
approaches could hardly prepare HMGA1a in large quantities,
which limits further application; moreover, the in vitro
enzymatic incorporation of PTMs could unlikely to be site-
specific, potentially undermining the studies’ credibility.
We therefore decided to apply chemical synthesis to the
preparation of homogeneous HMGA1a proteins, expectedly
with site-specific modifications installed at a stoichiometric
level. Unlike histone protein research,16 investigation of HMG
proteins with chemical approaches remains underexplored. In
this study, we have developed useful methods and strategies to
I
HMGA, HMGB, and HMGN) play vital roles in modulating
chromatin structure, regulating genomic function, and orches-
trating participation of other proteins, which are closely related
to nuclear activities such as transcription, replication, and DNA
repair.1 HMGA1a is a prototype member of the HMGA protein
family, which is recognized as a hub of nuclear functions to
affect a plethora of normal biological processes2,3 including
growth, proliferation, differentiation, and death; they are also
involved in some common diseases, including benign and
malignant tumors,4 osteoarthritis,5 diabetes,6,7 and atheroscle-
rosis.8
HMGA1a consists of 106 amino acids with three “AT hooks”
as DNA binding domains and an acidic tail at the C-terminus
(Figure 1).9 The AT-hook is a unique protein signature and
Figure 1. HMGA1a with diverse post-translational modifications.
characteristic functional sequence motif of HMGA family that
preferentially binds to the AT-rich stretches in the minor
groove of B-form DNA10 as well as four-way junction DNA.11
The acidic tail is a Glu-rich, 15 amino acid long sequence at the
C-terminus (Glu92-Gln106). The biological function of the
acidic tail is not yet fully understood.12
Received: October 11, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b03056
Org. Lett. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Organic Letters
Letter
synthesize HMGA1a proteins with well-defined post-transi-
tional modification(s) for the first time. Access to such proteins
would allow us to gain deeper insight into how PTMs of
HMGA1a affect its binding to DNAs and proteins.
Scheme 2. (a) Fmoc-Thr(HPO3Bn)-OH Failed To Be
Incorporated during SPPS. (b) Fmoc-pTP-OH and Fmoc-
pSP-OH Dipeptide Building Blocks Were Utilized to
Overcome the Difficulty of SPPS
Our first goal was to develop the total synthesis of HMGA1a
with three phosphates at in vivo cdc2 phosphorylation sites
(Ser35, Thr52, Thr77), which are abundant phosphorylated
sites most adjacent to the AT hooks. According to the principle
of Ser/Thr ligation (Scheme 1a),17,18 the retrosynthesis of
Scheme 1. (a) Ser/Thr ligation (STL). (b) Retrosynthetic
Plan of HMGA1a According to the Principle of STL
pTP-OH and Fmoc-pSP-OH in solution phase (Scheme 2 and
Schemes S1 and S2) followed by elongation onto the resin-
bound peptide chain. To our delight, this time, the
phosphorylated dipeptide Fmoc-pTP-OH was readily coupled
onto the resin-bound Lys54-Thr75 peptide under the HATU/
HOAt/DIEA/NMP/DMF conditions (Scheme 2b). Upon
solving this problem, the phosphorylated HMGA1a peptide
segments 1−3 could now be successfully synthesized via Fmoc-
SPPS.
Along our synthetic studies, the second challenge confront-
ing us was the unsatisfactory Ser/Thr ligation between the
peptide (Thr38-Gly62) salicylaldehyde (SAL) ester (2) and the
peptide 1 (Ser63-Gln106). Under various ligation conditions,
including different pyridine/AcOH ratios, peptide concen-
trations, ligation additives, and elevated temperature, the
desired ligation product was hardly observed and the peptide
SAL ester was eventually hydrolyzed. To understand the failure
in ligation, we used model peptides to test the reactivity and
properties of peptide SAL ester 2 and peptide 1, respectively.
While the ligation between 2 and SVAFKA or a truncated
fragment peptide (Thr76-Gln106) proceeded very smoothly,
peptide 1 failed to react with AIFPNAF-SAL ester and the
reaction mixture appeared as a slurry. These results indicated
that the failed ligation was due to the poor solubility of peptide
Ser63-Gln106 (1) under the ligation conditions (pyridine/
AcOH).
Toward this end, we reckoned that the backbone-protecting
group Hmb (2-hydroxy-4-methoxybenzyl) would help solve the
poor solubility problem by minimizing peptide aggregation
tendency. While Hmb(Ac) is stable to post-SPPS TFA
cleavage, Hmb can be readily removed by the TFA cocktail
to release the native peptide (Scheme 3a),20 making it a suitable
auxiliary for Ser/Thr ligation. Gratifyingly, Hmb(Ac)Gly67-
containing peptide 4 indeed exhibited better solubility and
could ligate with peptide SAL ester 2 to afford peptide (Thr38-
Gln106) 5 in 27% yield after HPLC purification (Scheme 3b).
However, the preparation of peptide (63−106) 4 via SPPS
was very difficult because the incomplete coupling at EKEEEE
residues of the Glu-rich acidic C-terminal tail region was very
significant and the deletion product was difficult to separate
from the desired product, affording <2% isolated yield based on
resin loading. On the other hand, another problem was
encountered when the obtained peptide 5 underwent Fmoc
removal and the subsequent ligation with peptide (Ser1-Gly37)
SAL ester 3. After TFA treatment, a ligation product was found
with m/z corresponding to peptide (Ser1-Gln106) minus Hmb
HMGA1a was disconnected as peptide Ser1-Gly37 (3), Thr38-
Gly62 (2), and Ser63-Gln106 (1) with lengths of 37, 25, and 44
amino acid residues, respectively, each of which contains a
phosphorylated residue (Scheme 1b).
To prepare the phosphorylated peptide segments by Fmoc-
SPPS (solid-phase peptide synthesis), the building block Fmoc-
Thr(HPO3Bn)-OH was synthesized accordingly.19 However, its
coupling onto resin-bound Pro53-Thr75 peptide turned out to
be unexpectedly difficult (Scheme 2a). Various coupling
reagents and conditions were carefully attempted, including
HATU/DIEA/DMF, HATU/HOAt/collidine/DMF, PyBOP/
DIEA/DMF, PyBOP/HOBt/DIEA/DMF, PyBOP/HOBt/
DIEA/DMSO/NMP, and DEPBT/DIEA/DMF/DCM; un-
fortunately, little desired product, if any, was observed. It was
then clear to us that the coupling between the phosphorylated
Thr52 and the secondary amine of the Pro53 residue was not
properly efficient, probably due to the steric effect and low
reactivity. It is noted that pSer35 and pThr77 are also
connected after the Pro residue; thus, a general strategy to
overcome this obstacle had to be devised prior to the total
synthesis.
Considering the low proneness to epimerization at the C-
terminal Pro residue during the coupling reaction, we decided
to prepare phosphorylated dipeptide building blocks Fmoc-
B
DOI: 10.1021/acs.orglett.6b03056
Org. Lett. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Organic Letters
Letter
Scheme 3. (a) Backbone Hmb Protection. (b) Hmb Served
To Improve Ligation Efficiency
but plus an Ac group. It is likely that the Fmoc removal
condition caused the acyl group of Hmb(Ac) to transfer.21
To solve the incomplete coupling problem at the acidic C-
terminal tail during SPPS, we changed the position of
Hmb(Ac)Gly from Gly67 to Gly96 within the problematic
region to disrupt the interchain association during peptide
segment 6 synthesis. This change not only was effective for
minimizing the incomplete coupling of SPPS, affording peptide
6 in a higher yield (16% based on resin loading), but also
improved the ligation between peptide (Thr38-Gly62) SAL
ester 2 and peptide Ser63-Gln106 (6), affording 38% ligated
product 7 after HPLC purification (Figure 2a,b). To avoid the
potential acyl-transfer problem of ligation product 7, we
removed the Ac group of Hmb with 10 equiv of N2H4
monohydrate22 followed by one-pot N-terminal Fmoc group
removal with Et2NH23 in CH3CN/H2O. An additional workup
with 5% aqueous TFA proved to prevent the potential of
hydrazinolysis of the SAL ester in the next ligation. This
operation sequence produced the peptide 8 (Thr38-Gln106)
cleanly in 60% isolated yield over two steps. Subsequently, the
ligation between the resultant peptide 8 and peptide (Ser1-
Gly37) SAL ester 3 went very smoothly and surprisingly even
faster than the first ligation, with completion within 7 h (Figure
2c). Upon acidolysis to convert the N,O-benzylidene acetal
intermediate to the native peptidic linkage, the Hmb group was
removed concomitantly, and the full length of HMGA1a with
three phosphates 9 was obtained in 43% yield after HPLC
purification (Figure 2d).
Having established the Ser/Thr ligation strategy to produce
synthetic HMGA1a protein 9 with three phosphates, in a
similar fashion, we synthesized an unmodified HMGA1a 10
(Scheme S5) as well as three monophosphorylated HMGA1a
proteins at Ser35, Thr52, and Thr77, respectively (11−13)
(Scheme S6−S8), which serve to elucidate the effect of the
individual phosphorylation on the HMGA1a function. Syn-
thesis of HMGA1a 14 bearing asymmetric dimethylation at
Arg25 within the first AT-hook was also achieved (Scheme S9).
The generality of this synthetic strategy was further
demonstrated by the preparation of biotinylated HMGA1a
proteins including an unmodified one, an acidic tail truncated
one, and one with triphosphates at the acidic tail (15−17)
(Figure 3a and Schemes S10−S12). We examined how these
Figure 2. (a) Synthetic scheme of full-length HMGA1a. (b)
Representative HPLC profile of the first ligation. (c) Representative
HPLC profile of the second ligation. (d) HPLC−MS characterization
of the purified final product HMGA1a 9.
PTMs affect the secondary structure of HMGA1a. As shown in
Figure 3b, the incorporation of the PTM(s) did not affect the
intrinsically disordered nature and structural behavior of
HMGA1a.
In summary, we have developed an attractive and practical
route to chemically synthesize HMGA1a proteins for the first
time via Hmb-assisted Ser/Thr ligation. During the synthetic
C
DOI: 10.1021/acs.orglett.6b03056
Org. Lett. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Organic Letters
Letter
ACKNOWLEDGMENTS
■
This work was supported by the Research Grants Council−
Collaborative Research Fund of Hong Kong (C7038-15G), the
National Basic Research Program of China (2013CB836900),
and the CAS interdisciplinary Innovation Team.
REFERENCES
■
(1) Reeves, R. Biochim. Biophys. Acta, Gene Regul. Mech. 2010, 1799,
3.
(2) Reeves, R. Gene 2001, 277, 63.
(3) Cleynen, I.; Van De Ven, W. J. M. Int. J. Oncol. 2008, 32, 289.
(4) Fusco, A.; Fedele, M. Nat. Rev. Cancer 2007, 7, 899.
(5) Gasparini, G.; De Gori, M.; Paonessa, F.; Chiefari, E.; Brunetti,
A.; Galasso, O. Arthritis Res. Ther. 2012, 14, R207.
(6) Foti, D.; Chiefari, E.; Fedele, M.; Iuliano, R.; Brunetti, L.;
Paonessa, F.; Manfioletti, G.; Barbetti, F.; Brunetti, A.; Croce, C. M.;
Fusco, A.; Brunetti, A. Nat. Med. 2005, 11, 765.
(7) Chiefari, E.; Iiritano, S.; Paonessa, F.; Le Pera, I.; Arcidiacono, B.;
Filocamo, M.; Foti, D.; Liebhaber, S. A.; Brunetti, A. Nat. Commun.
2010, 1, 40.
(8) Schlueter, C.; Hauke, S.; Loeschke, S.; Hans Wenk, H.;
Bullerdiek, J. Pathol., Res. Pract. 2005, 201, 101.
(9) Zhang, Q.; Wang, Y. Biochim. Biophys. Acta, Proteins Proteomics
2008, 1784, 1159.
(10) Reeves, R.; Nissen, M. S. J. Biol. Chem. 1990, 265, 8573.
(11) Hill, D. A.; Reeves, R. Nucleic Acids Res. 1997, 25, 3523.
(12) Maurizio, E.; Cravello, L.; Brady, L.; Spolaore, B.; Arnoldo, L.;
Giancotti, V.; Manfioletti, G.; Sgarra, R. J. Proteome. Res. 2011, 10,
3283.
(13) Sgarra, R.; Zammitti, S.; Lo Sardo, A.; Maurizio, E.; Arnoldo, L.;
Pegoraro, S.; Giancotti, V.; Manfioletti, G. Biochim. Biophys. Acta, Gene
Regul. Mech. 2010, 1799, 37.
Figure 3. (a) Synthetic HMGA1a proteins (total amount and overall
yields after HLPC isolation in parentheses). (b) Circular dichroism
spectra of synthetic HMGA1a proteins.
(14) Zhang, Q.; Wang, Y. Biochim. Biophys. Acta, Gene Regul. Mech.
2010, 1799, 28.
(15) Munshi, N.; Agalioti, T.; Lomvardas, S.; Merika, M.; Chen, G.;
study, we discovered the poor coupling of the phosphorylated
Ser/Thr to the proline residue at SPPS, which was solved by
using preassembled phosphorylated Ser-Pro and Thr-Pro
dipeptide building blocks. In addition, solubility issues
represent a practical challenge in protein chemical ligation,
which can be tactically overcome.24 In our study, the utility of
Hmb crucially assisted the Ser/Thr ligation to allow the
HMGA1a proteins to be successfully synthesized, normally in
14 working days, on multimilligram scales. This Hmb-assisted
Ser/Thr ligation may find general applications for synthesis of
proteins of such type. Overall, the obtained HMGA1a proteins
with different phosphorylation and methylation patterns will
provide useful chemical probes to study how post-translational
phosphorylation(s) of HMGA1a affect the protein−DNA and
protein−protein binding events, which is ongoing and will be
reported in due course.
Thanos, D. Science 2001, 293, 1133.
(16) Muller, M. M.; Muir, T. W. Chem. Rev. 2015, 115, 2296.
̈
(17) Li, X.; Lam, H. Y.; Zhang, Y.; Chan, C. K. Org. Lett. 2010, 12,
1724.
(18) (a) Zhang, Y.; Xu, C.; Lam, H. Y.; Lee, C. L.; Li, X. Proc. Natl.
Acad. Sci. U. S. A. 2013, 110, 6657. (b) Lee, C. L.; Liu, H.; Wong, C. T.
T.; Chow, H. Y.; Li, X. J. Am. Chem. Soc. 2016, 138, 10477.
(19) Vorherr, T.; Bannwarth, W. Bioorg. Med. Chem. Lett. 1995, 5,
2661.
(20) Johnson, T.; Quibell, M.; Sheppard, R. C. J. Pept. Sci. 1995, 1,
11.
(21) Acyl groups on tyrosine phenolic OH have been observed to
transfer to an adjacent amine via 12−14-membered rings; see: Panda,
S. S.; Hall, C. D.; Oliferenko, A. A.; Katritzky, A. R. Acc. Chem. Res.
2014, 47, 1076.
(22) Johnson, T.; Packman, L. C.; Hyde, C. B.; Owen, D.; Quibell,
M. J. Chem. Soc., Perkin Trans. 1 1996, 719.
ASSOCIATED CONTENT
(23) Ogunkoya, A. O.; Pattabiraman, V. R.; Bode, J. W. Angew.
Chem., Int. Ed. 2012, 51, 9693.
■
S
* Supporting Information
(24) For examples, see: (a) Kochendoerfer, G. G.; Salom, D.; Lear, J.
D.; Wilk-Orescan, R.; Kent, S. B. H.; DeGrado, W. F. Biochemistry
1999, 38, 11905. (b) Clayton, D.; Shapovalov, G.; Maurer, J. A.;
Dougherty, D. A.; Lester, H. A.; Kochendoerfer, G. G. Proc. Natl. Acad.
Sci. U. S. A. 2004, 101, 4764. (c) Valiyaveetil, F. I.; MacKinnon, R.;
Muir, T. W. J. Am. Chem. Soc. 2002, 124, 9113. (d) Liu, L. P.; Deber,
C. M. Biopolymers 1998, 47, 41. (e) Johnson, E. C. B.; Kent, S. B. H.
Tetrahedron Lett. 2007, 48, 1795. (f) Zheng, J.-S.; Yu, M.; Qi, Y.-K.;
Tang, S.; Shen, F.; Wang, Z.-P.; Xiao, L.; Zhang, L.; Tian, C.-L.; Liu, L.
J. Am. Chem. Soc. 2014, 136, 3695. (g) Jacobsen, M. T.; Petersen, M.
E.; Ye, X.; Galibert, M.; Lorimer, G. H.; Aucagne, V.; Kay, M. S. J. Am.
Chem. Soc. 2016, 138, 11775.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b03056.
Experimental procedures and characterization data
(PDF)
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: xuechenl@hku.hk.
Notes
The authors declare no competing financial interest.
D
DOI: 10.1021/acs.orglett.6b03056
Org. Lett. XXXX, XXX, XXX−XXX