Cys–Cys and Cys–Lys Stapling of Unprotected Peptides Enabled by Hypervalent Iodine Reagents

Abstract Easy access to a wide range of structurally diverse stapled peptides is crucial for the development of inhibitors of protein‐protein interactions. Herein, we report bis‐functional hypervalent iodine reagents for two‐component cysteine‐cysteine and cysteine‐lysine stapling yielding structurally diverse thioalkyne linkers. This stapling method works with unprotected natural amino acid residues and does not require pre‐functionalization or metal catalysis. The products are stable to purification and isolation. Post‐stapling modification can be accessed via amidation of an activated ester, or via cycloaddition onto the formed thioalkyne group. Increased helicity and binding affinity to MDM2 was obtained for a i,i+7 stapled peptide.


Introduction
Protein-protein interactions (PPI) mediate awide array of signaling pathways in the cell. Many of such interactions involve binding through a-helical sequences.I dentification and synthesis of these fragments therefore represents apowerful starting point for developing PPI inhibitors for drug discovery. [1,2] Nevertheless,t he helical conformation of short peptides is less stable than when the sequences are part of proteins,a nd they are rapidly degraded by proteases. Stapling-covalently linking two amino acids residues of peptides on the same face of an a-helix-has been shown to enforce the helical conformation and improve the stability of peptides.I ta lso enhances their cell membrane permeability and therefore their potential to be used as drugs. [3] When developing ab ioactive stapled peptide,t he structure,l ength, lipophilicity and reactivity of the introduced linker is important. [4,5] Thel inkers can be used to improve the solubility and the overall binding affinity of the peptide,n ot only by inducing helical conformation, but also through direct interaction with the binding protein. Therefore,during the search of prospective drug candidates,a ne asy access to alibrary of stapled peptides with different covalent linkers is important to increase the chance of finding active inhibitors and tune the properties of the lead compounds.T he introduction of linker variability has been mainly addressed by two-component stapling strategies between modified peptides and reactive small organic molecules (Scheme 1). [3] Natural or non-natural amino acids bearing reactive functional groups are introduced on the peptide during solid-state synthesis,then reaction with ab i-functional linker yields the desired stapled peptide. Further modification is then possible using an orthogonal functional group on the linker.I na ddition, the stapled peptides can not only be used as pharmaceuticals,but also as chemical biology tools. [6] In the latter case,l inkers with biorthogonal handles enabling in vivo visualization or target identification are especially useful.
While the use of non-natural amino acids provides good selectivity and reactivity,the required building blocks are less readily available and stapling often requires metal catalysis, which can be inconvenient and challenging on peptides. [7] For these reasons,s tapling methods using natural amino acids have been developed. Cysteine has been the most broadly used amino acid, due to its rare presence and high reactivity. [8] Them ain reagents used have been either Michael acceptors or benzyl, allyl or alkyl halides 1 and dichloroacetone 2 (Scheme 2, A). [9] More recently,t hioyne/ene (B,r eagents 3), [10] and cysteine arylation (C,r eagents 4 and 5) [11] stapling techniques have been reported by the Chou and the Pentelute laboratories,r espectively.D espite this important progress, there is still astrong need for the introduction of structurally diverse linkers reacting efficiently and with high selectivity towards cysteines.
When using two cysteines for stapling, the method is limited to symmetric linkers.T of urther expand linker variability,a symmetric reagents have been recently developed to staple two different natural amino acids.This strategy has been less explored, with only few reports on stapling between Lysa nd Glu or Asp [3,8b] or Cys. [12] Similarly to cysteine,l ysine displays as trong nucleophilicity and is therefore ideally suited for labelling.H owever,w hen both residues are present, the selective introduction of an onsymmetrical tether becomes challenging.P revious approaches for Cys-Lys stapling therefore required the use of non-standard protecting groups [12a] or the pre-functionalization of either Cys [12b] or both amino acids, [12c] which led to multi-step syntheses.M ore efficient approaches have been reported recently:P entelute,B uchwald and co-workers developed ahighly selective palladium complex 6 containing an active carbamate for Cys-Lys stapling [12d] [Scheme 2, B, Eq. (1)].T he need to synthesize palladium complexes is ad rawback of this approach and limits structural diversity. TheL ia nd Perrin groups independently reported am ethod using ortho-phthalaldehyde (7)a sac heap and broadly available Cys-Lys linker [Scheme 2, B,Eq. (2)]. [12e,f] Interestingly,t he obtained isoindole could be readily functionalized by reaction with maleimide derivatives [12e] or provided fluorescent peptides directly. [12f] However,i nstability of the linker was observed, which could limit applications in vivo. [12e] Other amino acids have been targeted for stapling with Lysby means of multicomponent reaction-based techniques, [13, 8b] including the use of propargylated azaglycine-Lysine A 3 , [13a] N and C termini Ugi-based cyclization, [13b] or the stapling of two lysines by aP etasis reaction. [13c] There is still ac lear demand for simple bifunctional linkers reacting with high selectivity on non-protected peptides that are well-suited for structural modifications.
Our group reported in 2013 am ild alkynylation of thiols using ethynyl-1,2-benziodoxol-3(1H)-ones (EBXs) hypervalent iodine reagents. [14] We later demonstrated that these reagents were able to selectively label cysteine on peptides and proteins. [15] Building on this work, we envisioned that the observed reactivity and selectivity towards Cys could be applied to the stapling of a-helices without the need of protecting or pre-functionalizing amino acid residues (Scheme 3). ForCys-Cys stapling EBX-based dimer reagents 8 with two reactive hypervalent iodine warheads were designed. ForC ys-Lys stapling an activated ester was introduced on the EBX reagent providing asymmetric linkers 9.A fter reaction of the hypervalent iodine reagent with cysteine,t he ester group of the reagent 9 would then react with an earby amine.T he tether that links the two electrophilic functional groups could be used to introduce structural variability in the stapled peptides and as areactive group for further functionalization. Furthermore,t he thioalkyne group present on the tether may open the possibility of further functionalization via [3+ +2] cycloaddition.
Herein, we report the use of EBX-derived reagents 8 and 9 for Cys-Cys and Cys-Lys stapling of unprotected peptides. Theobtained stapled peptides were stable,allowing purification and characterization. Post-stapling modifications were possible using either an additional activated ester or the formed thioalkyne (a and b in Scheme 3). Thei mpact of stapling on helicity was studied showing improved helicity with several peptides.F inally,t his technology was applied to staple an a-helical p53 derived peptide that binds to MDM2, an important cancer target. [16] One stapled peptide was demonstrated to be an efficient inhibitor of MDM2 with a K d of 29 AE 4nMs howing a1 2t imes increase of potency compared to the linear peptide,e mphasizing the substantial effect of the enhanced helicity.

Results and Discussion
We started our study with the investigation of potential Cys-Cys stapling reagents.A fter several attempts,i tw as concluded that the frequently used more reactive 2-iodobenzoic acid based-EBX reagents [14,15] were not suitable for the synthesis of dimeric reagents.F irst, their synthesis was long and low yielding.S econd, they exhibited poor solubility in most solvents.T hird, even when better solubility could be achieved, they were excessively strong oxidants,l eading to disulfide formation as major pathway.C onsequently,w e turned our focus to the bis-CF 3 substituted hypervalent  iodine reagents (1-ethynyl-3,3-bis(trifluoromethyl)-1,3-dihydro-1l3-benzo[d] [1,2]iodoxole), which exhibit al ower oxidation potential and reactivity,a sw ell as higher solubility. [17] First, the only reported aryl-derived bis-CF 3 benziodoxole dimer 8a was synthesized (Scheme 4A). [17d] Reagent 8a has been previously used for the synthesis of multi-substituted furans,b ut has never been applied for the alkynylation of thiols.T he meta substituted aryl reagent 8b was also synthesized, as the geometry of the linker was expected to have astrong influence on stapling.Inaddition, new reagents 8c-e bearing silicon linkers of different lengths could also be accessed. Silicon-substituted EBX reagents were more reactive towards thiols than alkyl-or aryl-substituted analogues and led more efficiently to the alkyne product in our previous work. [14,15] All reagents were obtained in one step from the known hypervalent iodine reagent 10 [18] in 40-94 %yield. The structure of 8c was further confirmed by X-ray analysis (Scheme 4B). [19] Reagent 8c reacted cleanly with N-Acetyl-L-cysteine methyl ester (12)togive bis-thioalkyne 13 in 63 % yield, demonstrating that this class of benziodoxole reagents can also be used for thioalkynylation (Scheme 4C).
Next, potential reagents for Cys-Lys stapling were prepared. In order to enable reaction with Lys( K), ac ommonly used pentafluoro phenyl (PFP) ester was selected. [20] PFP esters are easily accessible and stable,b ut still reactive enough to efficiently yield amides.H aving as ingle hypervalent iodine center allowed us to move back to the more reactive and more efficiently synthesized EBX core based on 2-iodobenzoic acid. We also anticipated that reactivity with these EBX reagents should be sufficient with aryl substituents and decided to focus on substituted phenyl rings as tethers. Furthermore,t he geometry and length of the linker can be readily modified with different substitution patterns at the phenyl ring. The para-reagent 9a and the meta-reagent 9b were accessed in 74 %and 45 %yield, respectively.The orthosubstitution pattern was more difficult to synthesize,b ut the EBX 9c could nevertheless be obtained in 16 %yield. Finally, the highly sensitive reagent 9d bearing two activated esters could be accessed in 27 %y ield. While the best result was obtained when aT MS protected alkyne was used, the reagents could also be accessed in slightly lower yields using directly the terminal alkynes (see Supporting Information, Section 4b).
To optimize the conditions reagent 8c was chosen. Reaction of 16 with one equivalent of 8c afforded stapled peptide 17 a in 55 %y ield after 4hours (Table 1, entry 1). Formation of the disulfide bridge was identified as the major side-product of the reaction. Excess of the reagent resulted in yields up to 73 %, but little or no difference was found beyond 3.0 equivalents (entries 1t o4 ). Raising the temperature to 37 8 8C, increased the yield to 78 %a nd allowed to better solubilize the stapling reagents (entry 5). At 5mM, the reaction was completed in 30 minutes albeit with aslight loss of yield (entry 6). Ther eaction worked also well in DMSO, but the yield decreased to 62 %( entry 7). In other solvents, the solubility of the linear peptide 16 and reagent 8c was not sufficient and very low to no conversion was observed (entries 8t o1 0).
We then set out to expand the peptide scope using the reagent 8c under the optimized conditions (Table 2A). Based on our previous work with EBX, [14,15] we identified arginine (R) as the most likely residue to potentially raise chemoselectivity issues.H ence,w es elected the Ac-YGGEAAR-EACARECAARE Cys-Cys (i,i + 4) stapling model 18 re-ported by Greenbaum and co-workers,w hich contains three arginine (R) residues. [9a] Compared to the model system (Table 2A,e ntry 1), as imilar result was obtained (entry 2). Considering that the distance between the i th and i + 4 th amino acids in a-helix is 5.4 and the distance between iodine atoms in reagent 8c was 8.7 (according to X-ray analysis, Scheme 3B), we wondered if 8c could be used for two-loop stapling (i,i + 7, 10.8 ), taking into account the additional flexibility provided by the cysteine sidechains.Consequently, we synthesized another peptide model 20,with the sequence: Ac-QSQQTFCNLWRLLCQN.T his sequence has been introduced by Verdine and co-workers as amodification of the wild type helical binding domain of p53 that exhibits better cell permeability and helicity. [22] Thei nhibition of the interaction between p53 and MDM2 proteins has been linked to tumor suppression. They further improved the property of the peptide by stapling in at wo-loop fashion (i,i + 7) via metathesis,where non-natural olefinic amino acids were used in place of cysteines.Unfortunately,stapling with 8c was less efficient in this case (entry 3). In order to study the chemoselectivity of our reagents,w ea pplied 8c on peptide 22 containing the most nucleophilic amino acids;S er (S), Glu (E), Arg (R), Tr p( W), His (H), Gln (Q), Tyr( Y), Lys( K),  Asn (N) and Met (M) in no particular order,aswell as afree N terminus.T oo ur delight, reactivity and selectivity comparable with the ones of other peptide models were observed (entry 4). Next, we turned to the scope of reagents.T he cyclohexylbased reagent 8d afforded the stapled product with peptides 16, 18 and 20 as efficiently as the iPr-based reagent 8c (entries 5-7). Introduction of the longer silanol-based linker using reagent 8e was generally less efficient (entries 8-11), with the exception of aslightly better result obtained for the i,i + 7model (entry 10). This is in accordance with the longer distance between the two iodine atoms.Due to the observed low reactivity of 8e,t he monothioalkynylation intermediate was commonly detected as the major product. Longer reaction times did not increase conversion (Supporting Information, Table S5). Theu se of the para-and metasubstituted phenyl reagents, 8aand 8bwas then examined for both the one-loop model 16 and the two-loop model 20 (entries 12-15). In general, the phenyl-based linkers were less efficient than the corresponding silicon analogues.
Only the para-linker together with two-loop peptide model 20 provided ac omparable result (entry 13). [23] The isolation of complex pure peptides is often difficult and associated with significant loss in yield. We were therefore pleased to see that selected peptides could be isolated in pure form after preparative HPLC in 13-48 %yield (Entries 1-2, 7, 8-10).
In order to test the cysteine-lysine (CK) stapling reagents 9,w esynthesized the same peptide models as used for Cys-Cys stapling (16, 18, 20 and 22), but exchanging the second cysteine for al ysine-to give peptides 24, 26, 28 and 30 (Table 2B). In addition, the model 32,p reviously used by Buchwald and co-workers specifically for Cys-Lys stapling, [12d] albeit with af ree N-terminus was also synthesized. To our delight, the para-reagent 9a stapled the peptide models 24, 26,a nd 28 very efficiently in only 30 minutes (entries [16][17][18]. By HPLC,o nly the stapled products were observed. In this case again, higher absorbance is observed for the stapled products in comparison to the starting materials. To confirm that the relative absorbance was still reasonably correlated with the yield, the reaction was performed on larger scale.T he stapled products 25 a, 27 a and 29 a were isolated in 52, 65 and 87 %yields,respectively.Therefore,the stapling with the Cys-Lys system appeared to be more general and efficient than with Cys-Cys.T his could be due to the higher reactivity of the hypervalent iodine reagent, or the higher flexibility of the lysine side chain. Then ucleophilic peptide 30 with af ree N-terminus was stapled slightly less efficiently (entry 19). Interestingly,o nly one product was detected by HPLC analysis.M S/MS studies confirmed the formation of the Cys-Lys stapled product. This selectivity however, appears to be sequence dependent. In fact, when the unprotected Buchwald model 32 was examined, both stapled products-Cys-Lys (33 a)and Cys-N-terminus staple (33 a' ')were obtained.
The meta-reagent 9bshowed lower reactivity with peptide 24.W hen previously using reagent 9a,f ull conversion was observed, but meta-reagent 9bprovided only 63 %conversion of 24 in 30 minutes with a3 4% relative absorbance for the product (entry 21). Longer reaction times did not result in significant increase in conversion. Thei solation of pure product 25 b was also difficult and low yielding.H owever, good results were obtained with peptides 26 and 28 resulting in 77 %a nd 110 %r elative absorbance,r espectively (entries 22 and 23). In this case,the products 27 b and 29 b could also be isolated in good yields À44 %a nd 55 %r espectively. Stapling was also less efficient with peptide models 30 and 32 containing free N-terminus (entries 24 and 25). Only the stapling between Cys-Lys was observed by MS/MS analysis of both models.I nc ontrast, excellent reactivity and full conversion of the starting peptides 24, 26 and 28 was again observed with the ortho reagent 9c after 30 minutes (Table 2B,e ntries 26-28). However,a ll attempts to isolate the pure products 25 c, 27 c or 29 c were unsuccessful due to the observed instability of the products in solution. [24] Fort his reason, the peptide scope was not further investigated, as the instability of the ortho-linker made it not suitable for biological applications.O verall, the trends in reactivity suggest that the position of the electron withdrawing carboxy group on the reagent has astrong influence on the efficiency of stapling.
To further investigate the reaction mechanism and the origin of the observed selectivity,s tapling of peptide 26 with the para reagent 9a was chosen as it displayed quantitative conversion, and no other side-products were detected by HPLC.Kinetic MS experiments showed that the thiol attack onto the EBX core is very fast, complete in under 1m inute, yielding intermediate 34 (Figure 1). Thef ormation of an ynamine intermediate instead is highly improbable,a sE BX reagents are known to react rapidly with thiols and not with amines. [14,15] Overall, the reaction was finished in 10 minutes even at a0 .1 mM concentration providing the product 27 a. We were pleased to see that intermediates 35 and 36 arising

Angewandte Chemie
Research Articles from initial attack on the activated ester were not detected, suggesting that the Lysa ttack is proximity driven. If these intermediates were formed, al ack of selectivity would be expected. Theo bserved high selectivity and reaction rate of EBX reagents towards sulfur nucleophile is therefore believed to be the reason behind the high efficiencyobserved for peptide stapling.
To show that our Cys-Lys stapling can be performed in presence of additional Lys( K) residue,w es ynthesized am odified version of peptide 26 (Ac-YGGEAAREACAR-EKAARE) containing (4,4-dimethyl-2,6-dioxo-cyclohex-1ylidene)-3-methyl-butyl (ivDde) protected Lys-ivDde-26 a (Table 3, A). [25] Thes tapling using the reagent 9a proceeded efficiently and the ivDde protecting group could be removed by addition of hydrazine in one-pot manner,yielding the final product 26 a' ' in excellent rel. abs (Table 3, entry 1). While the reagent 9balso reacted well with ivDde-26 a,the deprotection proved to be more difficult as degradation of the peptide was observed (entry 2). Since the protecting group approach lacked generality,wewanted to investigate if unprotected Lys (K) could be tolerated in our stapling method. Based on the results of the kinetic experiment showing that thiol functionalization was orders of magnitude faster (Figure 1), we thought that formation of the staple would be favored over reaction with other Lys(K) not on the same side of the helix. In addition to 26 a,w ei ntroduced aL ys (K) in various positions relative to Cys (Table 3, B). After submitting the peptides 26 a, 26 b and 26 c to our Cys-Lys stapling reaction conditions and reagent 9a,the desired i,i + 4stapled peptides, as confirmed by MS/MS analysis,w ere obtained in excellent rel. abs (entries 3-5). Only minor formation of another stapled product or overreaction at the free lysine were observed (0-13 %a nd 0-23 %r el. abs,r espectively,s ee Supporting Information, Section 12b). Moreover,the product 26 a' '' ',which could not be obtained using the ivDde protecting group strategy,was now formed in 74 %r el. abs (entry 6).
We then explored the reactivity of the tris-functionalized reagent 9d in ao ne-pot stapling/labelling sequence (Scheme 5A). Thestapled intermediate 27 d containing aPFP ester was obtained in 70 %r elative absorbance using peptide 26.
Thea dditional activated ester was then further reacted with an amine-containing azido-PEG reagent 37,p roviding the final functionalized product 38 in 73 %r elative absorbance.N ext reagent 9d was applied to peptide 32 containing af ree N-terminus.A fter prolonged reaction time,t ricyclic peptide 39 was formed in 37 %yield.
With the successful synthesis and isolation of the stapled peptides,w et hen wanted to explore the use of the unique thioalkyne functionality for post-stapling modifications.T he metal-catalyzed cycloaddition between azides and terminal alkynes [26,27] has been broadly used as ab ioorthogonal reaction. [28] In contrast, internal alkynes are usually less reactive.I n2 017, MascareÇas and co-workers reported arare biocompatible Ru II -catalyzed azide-thioalkyne cycloaddition (RuAtAC) under aqueous conditions. [29] Thus,we decided to take advantage of this reactivity and apply it to the thioalkyne present in our linkers.I n order to apply the method to our stapled peptides,t he reported conditions needed to be adjusted. The reported 2:1ratio of thioalkyne and azide was changed to a1 :1 ratio to ensure the full conversion of the more precious stapled peptide. [30] Thec oncentration of 75 mM was unrealistic for our small reaction scale typically used for peptides and was decreased to 27 mM to ensure solubility.M oreover,t he solvent was changed from water or DCM Table 3: Cys-Lys stapling in presence of additional ivDde protected and unprotected Lys. to DMF,t oe nsure that ao ne-pot-stapling and cycloaddition-procedure could be done in the same solvent. After optimization using model stapled peptide 29 a in combination with 1equivalent of benzyl azide (40 a)(Supporting Information, Table S10), full conversion to yield two regioisomers of the triazole product in aratio of 10:1 was achieved in presence of 20 mol %ofRucatalyst at 27 mM concentration (Table 4, entry 1). Azides bearing fluorescent dyes (5/6-TAMRA-PEG 3 -N 3 )( 40 b)a nd 6-FAM-N 3 (40 c)) were then used. In both cases full conversion was observed (entries 2and 3). The effect of varying the linkers was explored next. Lower conversion, but higher regioselectivity was obtained when using staple 29 b bearing a meta linker.( entry 4). The optimized conditions were then applied to the para linked peptides 27 a and 25 a using benzyl azide (40 a). Excellent reactivity and selectivity were observed with staple 27 a (entry 5). Thedesired product was also successfully obtained using staple 25 a,b ut al ower conversion was observed (entry 6), possibly due to am ore hindered reaction site, created by leucine (L) and isoleucine (I) amino acid residues. Thec ycloaddition was not successful on the Cys-Cys staple 17 a,p robably due to the steric hindrance resulting from the iPr groups (Supporting Information, Section 14b). Theo ptimized reaction conditions were adjusted to enable one-pot stapling and subsequent RuAtAC(Scheme 6). Thec oncentration of the reaction was further reduced to 5mMt oe nsure intramolecular attack during the stapling step.T oreach full conversion, the ruthenium catalyst loading in the cycloaddition step needed to be increased to 50 mol %.
Thel inear peptide 28 could then be converted into triazole products 41 a and 41 a' ' in one-pot and 89 %conversion of the stapling intermediate.T he one-pot procedure was then used to isolate both regioisomers in 52 %and 4% yield.
To study the changes induced in the peptide conformations,w eo btained circular dichroism measurements of the helical linear peptides and the corresponding isolated stapled peptides ( Table 2, entries 1-3, 7-10, [16][17][18][21][22][23]. Both oneand two-loop staples 19 a and 21 a with the bis(isopropyl)silyl tether seemed to have lost partially the alpha helix conformation (Figure 2A,B). On the other hand, the staples 19 c and 21 c with the longer silanol linker showed similar helicity compared to linear peptide 18,asseen by the values at 222 nm (Table S11). [31] As imilar behavior was observed for peptide 17 c,while for 17 a obtaining aprecise measurement under the same conditions was difficult (See Supporting Information, Section 16, Figure S4 and S5). For Cys-Lys staples in general, little change or decrease in helicity was observed for one-loop staples 25 and 27 (see Supporting Information, Figure S6 and S7). In contrast, meta-a nd para-t wo-loop staples 29 a and 29 b as well as the corresponding triazole staple 41 a displayed higher helicity than linear peptide 28 ( Figure 2C). Overall, the greatest helicity increase was obtained for 29 b with a meta linker for Cys-Lys stapling of the two-loop model 28.
Finally,w et ested if modification of peptides with the novel stapling reagents can yield efficient inhibitors of PPIs.F or this we studied the interaction of the p53derived peptide sequence Ac-QSQQTFCNLWRLLC/KQN (20 and 28)with MDM2. We first tested the binding of the peptides to MDM2 using af luorescence polarization (FP) competition assay in which the displacement of al inear, fluorescein-labeled reporter peptide was measured (fluorescein-GSGSSQETFSDLWKLLPEN-  [a] Absorbanceratio(%) = [(UV absorbanceofproduct)/((combined UV absorbanceofstapled peptide and product)] *1 00, and is given as combinedy ield when both regioisomers where observed.O nly starting material, reagent and products were observed by HPLC. In brackets ratio of regioisomers is indicated where applicable. Products 41 b and 41 b' ' could not be fully separated, while only one peak corresponding to the m/z of the desired product was detected for 41 c and 44 by HPLC analysis. Only the structure of the likely major product is drawn.
NH 2 ). Thes tapled peptides 21 a, 29 a and 29 b efficiently displaced the probe at concentrations close to that of MDM2, suggesting high binding affinities and K d si nt he nanomolar range (Supporting Information, Figure S8). We decide to further investigate the binding of the staple 29 a as it exhibited good helicity (Figure 2), binding affinity in competition assay and could be isolated in high yield ( Table 2). In order to accurately determine the binding affinities we synthesized 28' ', and the corresponding stapled peptide 29 a' ',b oth as conjugates with a5 (6)-carboxyfluorescein at the N-terminus (Scheme 7). Thep resence of fluorescein did not affect the stapling reaction, and product 29 a' ' was isolated in 76 %y ield and could be used to measure the binding to MDM2 in adirect FP assay.T he stapled peptide bound to MDM2 with a1 2-fold higher affinity (K d = 29 AE 4nM) than the linear one (K d = 346 AE 45 nM) ( Figure 3). Interestingly,b oth the absolute value and the increase of affinity are higher than for the reported metathesis staple (K d = 55 AE 10 nM and 2-fold increase). [22] It is comparable to affinities observed for other reported highly optimized metathesis-based peptide staples. [32] In addition, asingle pure staple is obtained in contrast to the E/Z isomers formed in metathesis reactions,w hich display different affinities.T his result showed that the developed stapling reagents can substantially improve the binding affinity of a-helical peptides and yield high-affinity inhibitors of PPIs.

Conclusion
In summary,w eh ave reported the synthesis of new bifunctional hypervalent iodine reagents and their use for peptide stapling.R eagents bearing two reactive hypervalent iodine sites were developed for Cys-Cys stapling, whereas combining one hypervalent iodine group with an activated ester enabled Cys-Lys stapling.B oth one-loop (i,i + 4) and two-loop (i,i + 7) stapling was possible.T he stapling did not require the use of protecting or activating groups.T he geometry and length of the linker in the reagents greatly influenced the stapling efficiencyand helicity.P ost-modification of the stapled peptides was achieved either by introducing an additional reactive ester group,o rb yr utheniumcatalyzed cycloaddition of the formed thioalkynes with azides. Depending on the linker length and geometry,e ither an  increase or decrease of helicity was observed. Stapled peptide 29 a displayed both enhanced helicity and binding affinity to the MDM2 protein.