Angewandte
Chemie
Ellmanꢀs reagent, consistent with reaction of the single Cys10.
Addition of 2-mercaptoethanol to product protein corrobo-
rated the formation of Dha.[33] Next, to install a mimic of
phosphoserine, sodium thiophosphate was added to the Dha
generated at position 10. Full conversion into phosphocys-
teine in 10 was observed by LC/MS after 1 hour at 378C at
pH 8.0 (Scheme 3).[33] The Western blot of 10 demonstrated
position 10 provides a modified histone that may be useful.
Accordingly, addition of GlcNAc glycosyl thiol to Dha
provided the GlcNAc-modified protein 11 (Scheme 3) con-
taining a mimic of GlcNAc-Ser that we have validated in
assays on several proteins.[43]
All modifications discussed so far were carried out at
single sites. Most histones, however, bear multiple
PTMs,[3,4,11,12] so we next tested our strategy for installation
of modifications at more than one site. Two alternative PTM
sites were explored; another H3 mutant (12) was prepared
with two Lys!Cys mutations at positions 4 and 79, two other
sites of Lys methylation and acetylation.[3,4,44] Reagent 1
readily converted both Cys residues into Dha, thus it was the
first demonstration that it can efficiently modify multiple Cys
residues. Both Dha residues could then be converted into
PTM mimics by addition of appropriate thiols. H3 derivatives
bearing two dimethyllysine mimics (13) or two acetyllysine
mimics (14) were prepared by this protocol (Scheme 4).
Scheme 3. One-pot installation of a phosphoserine mimic and a
GlcNAc-serine mimic at position 10 of H3. Western blot analysis
demonstrates that antibodies raised to phosphoserine at position 10
of H3 bind to phosphocysteine at the same position.
that phosphocysteine is sufficiently similar to phosphoserine
such that antibodies raised to natural H3 pSer10 protein
recognized our synthetically produced 10 (Scheme 3). This
mimic also showed no sign of degradation over several hours
(see the Supporting Information).
Scheme 4. Double modification of H3 at position 4 and position 79.
Although phosphorylated serine residues and related
analogues have been installed in proteins through NCL and
EPL,[37] this is the first demonstration that antibody recog-
nition of phosphoserine can be mimicked by phosphocysteine.
Notably, enhanced binding over a peptide mimic was
observed (see the Supporting Information). Moreover, there
are currently no alternative methods for site-selective chem-
ical phosphorylation of Cys or Ser residues in full-length
proteins.[23] We therefore propose the use of Dha and its
conversion into phosphocysteine to chemically access func-
tional mimics of proteins containing pSer residues. Moreover,
the facile generation of a PTM mimic (pS10) at another site in
H3 highlighted the flexibility of our approach. To test this
hypothesis we also examined the interaction of the synthetic
phosphoH3 10 with a so-called chromatin ꢁreaderꢀ protein, 14-
3-3z,[38] using phosphopeptide displacement in an amplified
(chemi)luminescent proximity homogeneous assay[39]
(ALPHA, see the Supporting Information). Consistent with
known modulations and interactions of naturally derived H3
and phosphoH3,[38,40] both H3 and 10 interacted with 14-3-3z;
the interaction of 10 was strongest.
Recently, Hart and co-workers have shown that O-linked
b-N-acetyl glucosamine (O-GlcNAc) is found at serine and
threonine residues in histones.[5] While modification sites
have not yet been identified, it has been shown that H3 bears
O-GlcNAc.[5] Since this modification is a relatively recent
addition to known histone PTMs, its influence on chromatin is
unexplored. The synthesis of histones bearing O-GlcNAc or
appropriate mimics will help elucidate roles. Since phospho-
rylation and GlcNAc often occur at the same residue and
impart reciprocal function,[41,42] the installation of GlcNAc at
Proteins containing multiple PTMs separated by more
than 50 residues are not easily prepared by NCL and EPL
since they require multiple ligations. Amber codon suppres-
sion is complicated by competing termination of translation,
especially when compounded by the need for suppression of
multiple stop codons. The chemistry in Scheme 4, however, is
comparatively simple from an operational standpoint. More-
over, this was performed on 10 mg scale, thus providing ample
quantities in short-order from recombinant protein.[33,45]
To demonstrate how these modified histones might be
used in the study of chromatin-modifying enzymes (“writers/
erasers”), the H3K9Ac mimic protein 8 was used as a
substrate in a direct histone deacetylase (HDAC) assay.
After incubation with either HDAC1 or HDAC2, Western
blot analysis showed only trace amounts of acetylated histone,
but LC/MS analysis of the reactions revealed 42% (ꢀ 3%)
and 55% (ꢀ 3%) deacetylation by HDAC1 and HDAC2,
respectively (Scheme 5).[46] When 8 was incubated under the
same conditions without HDAC, clear binding was observed
by Western blot analysis and no reaction observed by LC/MS,
thus ruling out adventitious hydrolysis of the natural l epimer
of 8 as well as suggesting selective binding to the l epimer by
the primary antibody used in the blot. While kinetic data
might be difficult to extract from such an assay, nominal
activity can be determined unambiguously. In fact, the
Western blot and LC/MS analysis shown in Scheme 5 are
the first direct evidence for HDAC activity on the acetyllysine
mimic in 8.[47] The only other acetyllysine mimic available, the
thiocarbamate analogue introduced by Cole and co-workers
(Scheme 1a), is not a substrate for HDAC enzymes.[25]
Angew. Chem. Int. Ed. 2012, 51, 1835 –1839
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1837