Squaramate‐Modified Nucleotides and DNA for Specific Cross‐Linking with Lysine‐Containing Peptides and Proteins

Abstract Squaramate‐linked 2′‐deoxycytidine 5′‐O‐triphosphate was synthesized and found to be good substrate for KOD XL DNA polymerase in primer extension or PCR synthesis of modified DNA. The resulting squaramate‐linked DNA reacts with primary amines to form a stable diamide linkage. This reaction was used for bioconjugations of DNA with Cy5 and Lys‐containing peptides. Squaramate‐linked DNA formed covalent cross‐links with histone proteins. This reactive nucleotide has potential for other bioconjugations of nucleic acids with amines, peptides or proteins without need of any external reagent.


Protein-DNAinteractions are of crucial importance in
DNAp ackaging,r eplication, transcription, epigenetic modifications,and repair. [1] Transcription factors (TFs) are particularly important DNA-binding proteins that regulate gene expression through sequence-specific binding to promoter sequences.A mong the approximately 1600 known human TFs,t he detailed biological role and binding motifs are fully understood only for as mall fraction. [2] Although, there is an umber of methods for studying of protein-DNAi nteractions and for identification of DNA-binding proteins, [3] there is still an urgent need of other alternative methods,i n particular for weakly binding proteins.Covalent cross-linking is one of the most promising methods for identification of DNA-binding proteins,b ut covalent protein-DNAc onjugates are also useful for other applications in chemical biology or biosensing. [4] There are some general non-specific cross-linking methods based on photochemical generation of radicals (from 5halouracils) [5] or carbenes (from diazirine-linked nucleobases) [6] in DNAwhich bind randomly to neighboring amino acids through C À Ha ctivations.M ore challenging but potentially very useful are reactions specific for one or several amino-acid side-chains,b ut so far av ery limited number of them have been reported for DNA-protein cross-linking. Thiol-linked DNAc an cross-link with Cys-containing proteins through disulfide formation. [7] Vinylsulfonamide-linked DNAw as reported to cross-link with Cys, [8] whereas chloroacetamide cross-linked with proteins through Cys or His. [9] In both cases,p roximity effect was crucial for efficient formation of the covalent cross-link between modified DNA and protein. Most frequent were reports on cross-linking of aldehyde-linked DNAwith Lyseither through inefficient and reversible Schiff-base formation [10] or (more often) through irreversible reductive amination, [11,12] which, requires an additional stoichiometric reductant (for example,t oxic NaBH 3 CN,w hich complicates any in cellulo or in vivo usage). Lys-DNAi nteractions are very frequent and important, in particular in histones.Sofar,noreactive nucleobasemodification in DNAh as been reported to form irreversible cross-links with Lyswithout an external reagent.
Mono-amides of squaric acid (squaramates) are often used for bioconjugations with Lysa nd other amines. [13] Diamides (squaramides) have been used as ap hosphate surrogate in nucleotide [14] or oligonucleotide (ON) analogues. [15] Ac hemically synthesized 2'-sugar-linked squaramate-RNAc onjugate,p repared through reaction of 2'amino-modified RNAwith diethyl squarate,w as reported to cross-link to aminoacyl-transferase FemX Wv , [16] as the only example of its use in nucleic-acid conjugation. Within the framework of our program aimed at base-functionalized nucleic acids for applications in chemical biology, [17] we designed novel squaramate-linked cytosine 2'-deoxyribonucleoside triphosphate (dNTP) for the enzymatic synthesis of modified DNAa nd cross-linking with proteins.
Thesynthesis of the desired modified nucleotides started with the preparation of 5-(3-aminopropynyl)-2'-deoxycytidine (1)bydeacylation of known trifluoroacetylamide, [18] see Scheme S1 in Supporting Information. Ther eaction of amine 1 with 2equiv of diethyl squarate gave the squaramate-linked nucleoside dC ESQ in 80 %y ield (Scheme 1). Standard Yoshikawa phosphorylation [19] with POCl 3 gave the monophosphate dC ESQ MP in 37 %y ield, whereas the triphosphorylation [20] with POCl 3 followed by pyrophosphate and triethylammonum bicarbonate (TEAB) gave the triphosphate dC ESQ TP in 7% yield. Them onophosphate dC ESQ MP served as model compound for reactions with Lysa nd peptides (Scheme 1). Ther eaction of dC ESQ MP with Ac-Lys or aL ys-containing tripeptide proceeded at room temperature in borate buffer (pH 9) overnight to give the desired conjugates dC ESQLys MP or dC ESQ3 pept MP in 54 and 63 %, respectively.T hese nucleotide-peptide conjugates were isolated by HPLC in pure form and fully characterized by NMR spectroscopy and mass spectrometry (MS) to confirm the expected formation of the amide bond with Lys.
Then, we tested the squaramate-linked dNTP (dC ESQ TP) as substrate for DNA-polymerase-catalyzed synthesis of modified DNA ( Figure 1a). First, we performed the primer extension (PEX) in the presence of KODX LD NA polymerase with either 19-, 20-, 31-or 98-mer template and a13-, 15-or 25-mer primer (for the oligonucleotide sequences,s ee Tables S1 and S2 in the Supporting Information). In all cases (Figure 1b and Supporting Information, Figure S1), we observed the formation of full-length PEX products containing one,f our or eighteen dC ESQ modifications.P CR amplification with either 98-bp or 235-bp templates also proceeded well, giving as trong band corresponding to the modified amplicon (Supporting Information, Figure S2). This shows that the reactive dC ESQ TP nucleotide does not react with DNAp olymerase (neither during extension nor when modified DNAi su sed as template) and is ag ood substrate and building block for the enzymatic synthesis of reactive, modified DNAp robes.
Next, we tested the cross-linking reactions of dC ESQlinked DNAw ith amines and peptides (Figure 1a,c). The reactions of 20-bp DNA_C ESQ were performed at room temperature in borate buffer (pH 9). Thereaction with Sulfo-Cy5-NH 2 (100 equiv) gave the desired fluorescently labeled DNA_C ESQCy5 (Supporting Information, Figure S3). Analogous reactions of DNA_C ESQ with Ac-Lys and Lys-containing tripeptide or decapeptide were conducted with large excess (approximately 2500 equiv) of the peptides,s ince no proximity effect was expected in these non-DNA-binding peptides.U nder these conditions,t he reactions proceeded with moderate efficiency and gave the desired cross-linked conjugates DNA_C ESQLys , DNA_C ESQ3pept or DNA_C ESQ10pept with 50, 43, and 20 %c onversions,r espectively ( Figure 1c). All conjugates were characterized and confirmed by MALDI MS (Supporting Information, Table S4).
(BSA) as negative control of ap rotein containing 60 lysines that does not interact with DNA, the core-domain of p53 [21] as aD NA-binding protein containing lysines but not in the binding site,a nd as et of recombinant H2A, H2B,H 3.1, and H4 histones,a se xamples of Lys-rich proteins that strongly bind to DNA. Thec ross-linking reactions were performed with only 2equiv of the corresponding proteins.S ince the histones are known to form dimers and oligomers,weassume that this ratio is probably effectively close to equimolar.T obe closer to physiological conditions,w eu sed phosphate (or TRIS or HEPES,s ee Figure S8 in the Supporting Information) buffers (pH 7.4).
At first, we performed asimple kinetic study of the crosslinking reaction of DNA_C ESQ with histone H3.1 to show that the reaction reaches the maximum conversion in 16-24 h (Supporting Information, Figure S5). Therefore,weused 36 h reactions in other cases to ensure sufficient conversions. Figure 2s hows the results of the cross-linking experiments with proteins.T oour delight, the reactions of DNA_C ESQ with all four recombinant histones gave the covalent cross-linked conjugates with lower mobility on ad enaturing SDS-PAGE gel (Figure 2b). Theconversions of these reactions calculated from the SDS-PAGEwere 31-34 %(Supporting Information, Table S5). Thei dentity of the covalent DNA-protein conjugates with H2B,H 3.1, and H4 histones was also confirmed by SDS-PAGEw ith protein staining (Coomassie Blue, Supporting Information, Figure S7) and by HPLC-MS analysis using electrospray ionization (Supporting Information, Figures S16-S18). Also,alonger 98-bp PEX product containing 18 squaramate groups reacted with histone H3.1, though mixture of cross-linked products was obtained (Supporting Information, Figure S10). On the other hand, DNA_C ESQ did not cross-link with BSA or p53 (Supporting Information, Figure S9) or with DNApolymerase during the PEX or PCR. These results show that the proximity effect (presence of lysine(s) close to the DNA-binding site of protein) is crucial for efficient cross-linking in the absence of large excess of the peptide or protein.
In conclusion, we designed and synthesized an ovel squaramate-linked dNTP (dC ESQ TP)a nd demonstrated that it was avery good substrate for KODXLDNA polymerase in PEX or PCR synthesis of reactive DNAp robes.T he squaramate group reacts with amines to form as table covalent diamide (squaramide) linkage.W eh ave shown that the dC ESQ -modified DNAp robes reacted with amino-linked Cy5 to form fluorescently labeled DNA. Its reactions with Lys-containing peptides proceeded only when al arge excess of the peptide was present. On the other hand, in reactions with Lys-containing DNA-binding proteins,w here the proximity effect helps,t he reactions proceed with good conversions even in almost equimolar ratio.Compared to previously reported DNA-Lys conjugations based on reductive amination, [11,12] the squaramate modification and its transformation to astable amide proceeds under physiological conditions (at pH 7.4-9) and does not require any external reagent (i.e.toxic NaBH 3 CN used in reductive aminations [11,12] ). Therefore,this reactive modification and the presented methodology has good potential in the post-synthetic labeling of DNA, [22] bioconjugations of DNAw ith peptides,p roteins or other biomolecules, [4] as well as in cross-linking experiments to identify and study DNA-binding proteins.F urther research along these lines is under way in our lab.