Structural Relationships to Efficacy for Prazole‐Derived Antivirals

Abstract Here, an in vitro characterization of a family of prazole derivatives that covalently bind to the C73 site on Tsg101 and assay their ability to inhibit viral particle production is presented. Structurally, increased steric bulk on the 4‐pyridyl of the prazole expands the prazole site on the UEV domain toward the β‐hairpin in the Ub‐binding site and is coupled to increased inhibition of virus‐like particle production in HIV‐1. Increased bulk also increased toxicity, which is alleviated by increasing flexibility. Further, the formation of a novel secondary Tsg101 adduct for several of the tested compounds and the commercial drug lansoprazole. The secondary adduct involved the loss of the 4‐pyridyl substituent to form an irreversible species, with implications for increasing the half‐life of the active species or its specificity toward Tsg101 UEV. It is also determined that sulfide derivatives display effective viral inhibition, presumably through cellular sulfoxidation, allowing for delayed conversion within the cellular environment, and identify SARS‐COV‐2 as a target of prazole inhibition. These results open multiple avenues for the design of prazole derivatives for antiviral applications.


Synthesis of Compounds 4-19
Compounds used in this study were synthesized as detailed in the following schemes I-III.The initial step involved, either the formation of the aryl ethers (Scheme I) starting from the known alcohol 11 or alkylation of the commercially available 4-chloro-2,3dimethylpyridine-1-oxide with the corresponding phenol/alcohol under basic conditions.Low yields were observed for 10a-c (5%-10%), though moderate yields were obtained (40-50%) for compounds 3a-c under Mitsunobu conditions.Compounds 10a-c were subjected to Suzuki coupling under standard conditions to yield the coupled biaryls 11a-h2 in respectable yields (50-60%).Rearrangement of the alkylated/coupled products from Scheme I and II by heating with acetic anhydride provided the pyridyl 2acetoxymethyl derivatives that were taken to next step without further purification.Stirring the crude acetates with potassium carbonate in methanol provided the respective alcohols.It was determined that purification was not necessary at this stage, and the crude alcohols from 4a-c, and 12a-h were converted to the chloromethyl derivatives with thionyl chloride.The chlorides/hydrochlorides from the above step were pure enough for further alkylation with 2-thiobenzimidazole.Alkylation of the chloro-derivatives with 2-thiobenzimidazole in DMF, and in the presence of sodium bicarbonate yielded the precursors to the final sulfoxides.Purification was carried out at this stage by silica column chromatography using ethyl acetate and chloroform as eluents.Overall yields of 45-50% were observed for 4 steps (7a-c and 14a-h).Oxidation of the sulfides 7a-c, and 14a-h, yielded the sulfoxides with m-CPBA in ethyl acetate.Purification of final sulfoxides was carried out by preparative HPLC under mildly basic conditions to yield the sulfoxides 8a-c and 15 a-h (10-20%).Hydrolysis of the methyl esters containing the sulfoxide with lithium hydroxide in dioxane followed by preparative HPLC resulted in the sulfoxide-acids 9a,9b,16a, and 16c (25-30%).
Tetrazoles reported here were prepared as described in Scheme III.Conversion of the cyano to the tetrazole moiety was achieved by heating the cyano precursors (7c, 14b and 14d) with pyridine-hydrochloride (10.0 equiv.), and large excess of sodium azide (20.0 equiv.) in DMF to 70oC for 20h followed by silica gel column purification (chloroform/methanol with 0.1% ammonium hydroxide a s eluents; 20%-30% yields, 17a-c).Oxidation of the sulfides to sulfoxides was effected with m-CPBA in ethyl acetate at 0oC for 30 min, although low yields were observed after preparative HPLC purification under basic pH of 9.0 (10-24%, 18a-c

General Procedure for Preparation of 3
To a solution of the N-oxide 1 35 (1.0 mmol), and the phenol/alcohol (1.5 mmol) in anhydrous THF (5.0 ml) was added triphenylphosphine (1.6 mmol) and stirred at 0 0 C under argon.Diisoproyl azodicarboxylate was added dropwise over 5 min to the stirred solution at 0 o C, and allowed to come to RT, and stirring was continued for 20h more.The reaction was quenched with 10.0 ml of water and extracted with 3x25.0 ml of EtOAc.The combined organic layers were washed with water (3 x25 ml), and dried (sodium sulfate).The solution was filtered, concentrated under reduced pressure, and the residue was purified by flash silica column (40.0 g).Elution with 2% methanol in DCM, yielded the products as colorless gum.Yields: 25-45%

Alkylation of 4-Chloro-2,3-dimethylpyridine-N-oxide
A solution of the alcohol/phenol (5.0 mmol) was dissolved in 5.0 ml of anhydrous N,N-Dimethylacetamide (5.0 ml), and sodium hydride (60% in oil) was added (5.0 mmol), and stirred under argon for 30 min.Solid 4-chloro-2,3-dimethylpyridine-1-oxide (1.0 mmol) was added and heated to 110 o C for 72h. the reaction mixture was diluted with ethyl acetate (25.0 ml) and filtered through a pad of celite.The filtrate was concentrated under reduced pressure to a paste, and the crude product was purified by a flash silica column (40.0g).Elution with 12% methanol in DCM yielded the products as colorless gum.Yields: 5-10%

General procedure for Suzuki coupling of 10a-c with the aryl boronic acids 36
To N-oxides 10a-c (1.0 mmol), and aryl boronic acid (3.0 mmol) in anhydrous dioxane (10.0 ml), was added anhydrous solid potassium phosphate (18.0 mmol) and stirred while being degassed by bubbling argon through the solution.
Tetrakistriphenylphosphine palladium (5 mol%) was then added and heated to 100 o C with stirring under argon for 20h.The reaction mixture was diluted with 50.0 ml of ethyl acetate, filtered, and the filtrate was concentrated to a paste that was purified by flash silica gel (120.0g).Elution with 5% methanol in dichloromethane yielded the products as reddish-brown gum.

Boekelheide Rearrangement of N-oxides 3 & 11
The N-oxide derivatives were dissolved in (1.0 mmol) 5.0 ml of acetic anhydride and heated to 110 o C with stirring for 4h.All the volatiles were removed under reduced pressure and the crude acetates were taken to next step without further purification.

Hydrolysis of acetates 4 & 12:
The acetates were dissolved in methanol (5.0 ml for 1.0 mmol), and solid potassium carbonate (1.38g, 10.0 mmol) was added, and stirred at RT until LC/MS indicated the complete consumption of the acetates.The solution was filtered, concentrated under reduced pressure, and the residue was purified by flash column chromatography.Elution with 5% methanol in DCM yielded the 2hydroxymethyl derivatives as colorless gum.In most cases, the crude alcohol was taken to next step without any further purification.

Conversion of 2-hydroxymethylpyridyl derivatives to the corresponding chloro-derivatives
The hydroxymethyl derivatives from 4 & 12 (1.0 mmol, crude) were dissolved in 5.0 ml of anhydrous dichloromethane followed by the addition of 0.5 ml of thionyl chloride.The reaction mixture was heated to reflux under argon for 4h.All the volatiles were removed, and the residues were co-evaporated with toluene (2 x 5 ml) at RT and under reduced pressure to remove traces of thionyl chloride, triturated with anhydrous ethyl acetate (3x5.0 ml), decanted, and the remaining solid residue was taken to nest step without further purification.

Alkylation of chloromethyl pyridine derivatives 5 & 13 with 2-thiobenzimidazole (7a-c & 14a-h)
The crude chloro-hydrochloride salt (1.0 mmol) was heated with 2-thiobenzimidazole (1.5 mmol) in anhydrous DMF (10.0 ml) at 60 o C, and in the presence of solid sodium bicarbonate (5.0 mmol) under argon, and with stirring.After 20h, the mixture was diluted with 50.0 ml of ethyl acetate, filtered and the filtrate was concentrated under reduced pressure to a paste.The crude paste was purified by silica column (40.0g).Elution with 20% ethyl acetate in chloroform yielded the products as pale-yellow foam.Yield: 45-50%.

Oxidation of sulfides to sulfoxides
The sulfides 7 & 14 (0.1 mmol) were dissolved in ethyl acetate (5.0 ml), and solid sodium bicarbonate (0.138 g, 1.0 mol) was added and stirred at 0 o C. m-CPBA (77%, 26.0 mg, 0.15 mmol) was added in portions (5.0 mg at a time) over 5.0 min.After 30 min at RT, the reaction mixture was filtered, washed with ethyl acetate (3x 5.0 ml) and the combined filtrates were concentrated and purified by preparative HPLC.

Conversion of cyano substituted derivative to the corresponding tetrazole 17
The cyano-derivative (0.2 mmol) and pyridine hydrochloride (2.0 mmol), and sodium azide (4.0 mmol) were heated with stirring at 70 o C in anhydrous DMF (5.0 ml) for 20h.The crude reaction mixture was diluted with 25.0 ml of water and extracted with 3 x 25.0 ml of ethyl acetate.The combined organic layers were washed with water (2 x 20.0 ml) and dried (sodium sulfate).The solution was filtered, concentrated and the residue was purified by flash silica column (24.0 g).Elution with dichloromethane/methanol with 0.1% ammonium hydroxide (v/v; 85:15) yielded the products as off-white solids.Yield: 20-30% Conversion of tetrazole-sulfides 17a-c to the sulfoxides 18a-c: Repeated as described above for other sulfoxides and purified by silica gel column.Yields: 20-30%

Hydrolysis of sulfoxide methyl esters to acids (9a, 16a & 16c)
The sulfoxide-ester (0.1 mmol) in dioxane (200uL) was cooled in an ice bath, and 1M LiOH in methanol (1.0 ml, 1.0 mmol) was added and stirred until all the starting material was completely consumed (LC/MS).The reaction mixture was adjusted to pH 8.0 with acetic acid, 1.0 ml of methanol was added and filtered through 0.45µ filter.The filtrate was purified by preparative HPLC.

Synthesis of Compound 20
To a solution of 2,3-dimethyl-4-nitropyridine 1-oxide (1.68 g, 10 mmol) in anhydrous ethanol (4.1 ml, 70 mmol) at 0 o C was added slowly dropwise acetyl chloride (1.8 ml, 25 mmol) over a period of 30 min.After the addition was complete ice bath was removed and reaction was heated to reflux for 5 hours.After this all the volatiles were removed in vacuo and residue was dissolved in water and basified with saturated potassium carbonate solution in water to a pH of ~ 10.Aqueous layer was then extracted with dichloromethane (3 x 15 ml), combined organic layer was dried over anhydrous sodium sulfate and concentrated.Solid residue was then crystallized from toluene/heptane to give pure desired product.Yield 1.5 g 98 % To a solution of 4-chloro-2,3-dimethylpyridine 1-oxide (314 mg, 2 mmol) and triethylene glycol (165 mg, 1.1 mmol) in 6 ml of anhydrous 1,4-dioxane (6 ml) was added powdered potassium hydroxide (224 mg, 4 mmol) and resulting reaction mixture was heated to 110 o C for 24 h.After this time LCMS of the reaction indicated formation of product, all volatiles were removed in vacuo and residue was taken in minimum amount of water extracted with dichloromethane (5 x).Combined organic layer was dried anhydrous sodium sulfate concentrated and residue was purified over silica gel using 0 to 25 % methanol in dichloromethane to give desired product.Yield 380 mg 88 % To solid 4,4'-(((ethane-1,2-diylbis(oxy))bis(ethane-2,1-diyl))bis(oxy))bis(2,3-dimethylpyridine 1-oxide) (380 mg, 0.97 mmol) was added 6 ml of acetic anhydride and resulting solution was heated to 100 oC for 12 h.Reaction was then cooled to room temperature and excess acetic anhydride was quenched by dropwise addition of methanol.All then volatiles were then removed in vacuo and residue was taken basified with saturated sodium bicarbonate solution in water and aqueous layer was extracted with dichloromethane (3 x 20 ml).Combined organic layer was dried over anhydrous sodium sulfate, concentrated and chromatographed over silica gel using 0 to 20 % methanol in dichloromethane to give desired product.Yield 328 mg, 71 % A solution of ((((ethane-1,2-diylbis(oxy))bis(ethane-2,1-diyl))bis(oxy))bis(3-methylpyridine-4,2-diyl))bis(methylene) diacetate (328 mg, 0.69 mmol) in 5 ml 6 M hydrochloric acid in water was heated to 70 o C for 5 h.After this time LCMS indicated complete hydrolysis of starting material, all the hydrochloric acid was then removed in vacuo and residue was further dried in vacuum to give the p roduct as HCl salt which was used in next step without purification.Yield 312 mg, 98 %

Cellular Screening of Efficacy for Prazole Derivatives against SARS-COV-2
Tested compounds were sent to NIAID contractors at the University of Pennsylvania for analysis according to the following protocol.Calu3 (ATCC, HTB-55) cells were pretreated with test compounds for 2 hours prior to continuous infection with SARS-CoV-2 (isolate USA WA1/2020) at a MOI=0.5.Forty-eight hours post-infection, cells were fixed, immunostained, and imaged by automated microscopy for infection (dsRNA+ cells/total cell number) and cell number.Sample well data was normalized to aggregated DMSO control wells and plotted versus drug concentration to determine the IC50 and CC50.

HIV-1 VLP Assays
293 T (ATCC CRL-3216) or HeLa (ATCC CCL-2) cell lines were used to assess susceptibility to prazole exposure as previously described [5a] .Briefly, cells were grown to 70% confluency at 37 o C in Dulbecco's modified Eagle medium supplemented with fetal bovine serum (10%) and antibiotics (1%).Unless otherwise stated in figure legends, prior to drug treatment and transfection or toxicity assays, the tissue culture media was aspirated and replaced with control or treatment media.Transfection employed XtremeGene reagent (Roche) to facilitate DNA uptake.For production of virus particles, cells were transfected with pNL4-3-ΔEnv and pIIIB Env3-1 plasmids.At 24-hr post-transfection, tissue culture media was collected and passed through a 0.45 micron filter; cells were scraped off the tissue culture plate with a rubber policeman, rinsed with PBS and pelleted.For virus isolation, the filtered media was centrifuged through a 20% sucrose cushion at 22,000 × g for 90 min at 5 °C and the pellet fraction saved for analysis.For cell lysate preparation, cell pellets were lysed with Triton X-100 buffer (50 mMTris, pH 7.4, 137 mM NaCl, 1.5 mM MgCl2, 1 mM EDTA, 1% Triton X-100) containing il cOmplete Protease Inhibitor (Sigma).Virus and cell lysate samples were analyzed by Western blotting.Primary antibodies used were: Rabbit anti-CA; anti-Actin (Sigma).Secondary antibodies used were: goat anti-mouse IgG Alexa Fluor 680 (Molecular Probes); goat anti-rabbit IRDye800 (Rockland).Protein bands were visualized using an infrared-based imaging system (Odyssey, LI-COR Biotechnology) and band intensities measured using the Li-Cor Odyssey software version 2.1.15.  1.In addition to tenatoprazole, compounds 1, 2, 3, 6, 12, 14, 17, and 18 show significant evidence of labeling.Of these, compounds 1, 2, 12, 14, 17, and 18 have apparent rates comparable to or greater than tenatoprazole.Error bars are derived from the noise level of the spectra estimated using the Estimated Noise function in NMRPipe and NMRDraw.

Figure S2. Principal Component Analysis of the time-dependence of prazole-adduct formation for tested prazole derivatives
shows a bifurcated relationship.Plots of the first and second (left) and third and second (right) principal components obtained from PCA on height-normalized HSQC spectras used to follow prazole attachment.PCA results were obtained using NMRPipe and used the spectral region 7.5-9.4ppm ( 1 H) and 112-130 ppm ( 15 N).The first two PCs separate strongly on attachment, where eg: compounds 14, 1, and 3 appear to have the greatest labeling extent over the tested interval.The third PC, however, bifurcates the compounds into two groups, where those with negative PC3 values may be further subdivided.Low values of PC3 feature the rabeprazole derivatives, where 17+18 have large PC2 values and 15+16 smaller ones.Values near zero include compounds 12+14+15, which have the intermediate linker length and biphenyls.Large PC3 values include tenatoprazole, the three smallest derivatives (1, 2, 3) and interestingly compound 6, which had the shortest linker length but with a bulky biphenyl.Separation based on PC3 appears to be tied to chemical shifts rather than intensity differences, as compounds 17 and 18 display markedly different rates and behavior at C73 while having similar PC3 coordinates.shown to represent the primary adduct.The observed secondary adduct species formed with lansoprazole and several prazole derivatives involves apparent loss of the 4-pyridyl substituent, consistent with a nucleophilic attack at this position.The three closest nucleophilic residues to C73 on the Tsg101 surface are lysines 33, 36, and 90.The sidechain amine nitrogen of K33 is >1.5 nm from the 4-pyridyl carbon of the prazole, while K90 and K36 are <1 nm away.To determine which residue was likely to contribute to secondary adduct formation, we first generated the allowed K36 and K90 sidechain rotamers with the mutagenesis wizard in pymol (shown as blue spheres for the side chain nitrogen).For the prazole, rotations of the final three chi angles (I-III, inset) were sampled in 30° increments, with the cloud of resulting 4-pyridyl carbon positions shown as gray spheres.Several K90 rotamers overlap with the 4-pyridyl cloud, indicating that this position could likely be attacked by the K90 sidechain.For K36, the minimum distance between 4-pyridyl carbon and the sidechain amino nitrogen is 4.5 Å, and this orientation would likely be sterically blocked without rearrangement of other beta-sheet sidechains, including those of V38 and T56 (dark gray sticks).Thus, the most likely source of the secondary adduct is the sidechain amino nitrogen of K90.Having observed the formation of an irreversible, secondary adduct between Tsg101 and several prazoles, we also tested for the formation of an equivalent adduct in a simplified system having only cysteine and lysine sidechain nucleophiles.Tested peptides are given in (A), with the general sequence GCGnK, where the expected adduct mass was taken from the observed delta mass for the primary lansoprazole adduct with Tsg101 UEV domain.All peptides were N-terminally acetylated and C-terminally amidated to prevent the previously characterized adduct formation with the peptide amino-terminus [peptide citation].Lansoprazole was added to each peptide, where the resulting LC/MS results are given for each peptide in (B).In each case the chromatogram for the addition of stoichiometric lansoprazole is shown at top, followed by the chromatogram with 3x excess lansoprazole, and then the observed masses for peaks corresponding to the primary and secondary adducts (bottom) in the presence of excess lansoprazole.The secondary adduct is present in each case, having the same delta mass of 100 units corresponding to loss of the TFE group as seen for Tsg101 UEV domain.Compared to the protein, however, there is much less formation of the secondary adduct, with the most seen for the 4mer GCGK sequence.Reduced formation is likely reflective of the peptide flexibility reducing preferred orientations, and potential alterations in the cysteine or lysine pKas in the protein environment. .Small phenyl derivatives appear unstable in neutral conditions.(A) LC/MS results from labeling 15 N Tsg101 UEV domain with excess compound 1 in either NMR buffer (left) or PBS (right), to reproduce the conditions of the two assays.In both cases, compound 1 was added at 2-fold excess to 10 μM Tsg101 UEV domain.In the mildly acidic NMR buffer (pH 5.8), there is >90% labeling with a mixture of the expected (17102.6Da) and secondary (16978.4)adducts.In the higher pH PBS, however, there is <50% attachment, with a mixture of species including the expected adduct at 17102.5 Da, secondary adduct at 16978.5 Da, and an apparently double-labeled species at 17477.8 Da. (B) 1 H NMR traces of 1 mM Compound-1 in 20% d6-dmso and 80% NMR buffer (left) or 80% NMR buffer adjusted to pH 7.4 (right) after 30 minutes (black) and 15 hours (red).The loss of the primary, pro-drug species is slower at high pH, implying reduced conversion outside of the mildly acidic conditions, but neither condition shows secondary peaks consistent with build-up of activated drug.(C) As in (B) for tenatoprazole, highlighting the conversion to a secondary species with detectable peaks under both conditions.We showed previously that a N-acetyl cysteine (NAC) excess (3 mM) can compete for tenatoprazole over a broad range where the prazole is effective and suppress its ability to inhibit VLP production [5a] .Here, we show that NAC competes similarly with the sulfide derivative compound 19, indicating that its ability to inhibit must also involve Cys and, thus presumably is sulfoxidated to the sulfoxide pro-drug and converts to the same active sulfenamide.
Scheme II
Figure S1.Monitoring prazole attachment by loss of intensity at C73 in the Tsg101 UEV domain.(top-left), a view of the Tsg101 UEV domain (PDB ID: 5VKG) in gray, highlighting the location of C73, which forms a disulfide with the activated prazole in orange.Intensity plots as a function of time after prazole addition are shown for tenatoprazole and 18 tested derivatives, where loss was fit to a simple exponential decay function:   =  − +   and the values of k and If are given in Table1.In addition to tenatoprazole, compounds 1, 2, 3, 6, 12, 14, 17, and 18 show significant evidence of labeling.Of these, compounds 1, 2, 12, 14, 17, and 18 have apparent rates comparable to or greater than tenatoprazole.Error bars are derived from the noise level of the spectra estimated using the Estimated Noise function in NMRPipe and NMRDraw.

Figure S4 .
Figure S4.Overlays comparing compounds that formed secondary adducts with the Tsg101 UEV domain.(A) Five compounds: lansoprazole (lansoprazole), 3 (green), 4 (red), 6 (purple), and 7 (blue) share identical major shifts and give rise to the same secondary adduct by LC / MS, with a delta of +252 Da consistent with loss of the R3 substituent from the pyridine ring.(B) Comparison of lansoprazole (orange) with compound 13 (teal).Compound 13 also forms a secondary adduct by LC / MS, but with a mass of +238 Da, corresponding to the lack of R2 methyl relative to the other five compounds.The addition of the R2 methyl in lansoprazole appears to lead to additional interactions and multiple conformers on the UEV surface.(C) Comparison of compound 1(blue) with an early timepoint of lansoprazole (orange), taken 7 hours after prazole addition.When labeled in PBS, compounds 1 and 2 produced evidence of multiple species which correspond poorly to the eventual major state of the secondary adduct.At this early timepoint, however, there are minor peaks that match the dominant ones seen for compound 1, indicating that the compound 1 adducts formed at high pH include an intermediate conformation or species.

Figure S5 .
Figure S5.The secondary adduct formed for certain prazoles is irreducible.Shown are LC / MS results after labeling 15 N Tsg101 UEV domain with excess prazole (left) and the impact of excess DTT on the labeled adducts (right).(A) Compound 6, which has the shortest tested R3 linker length, shows approximately 90% labeling under these conditions, split between the expected (16978.4Da) adduct mass and an anomalous +251 Da species.Addition of DTT successfully reduces the expected adduct, but the unexpected +251 Da species is retained.(B) Compound 12, which differs from 6 only in the presence of a single carbon extension in the R3 linker, shows complete labeling under these conditions to the expected (17206.6Da) prazole adduct, which is nearly fully reduced in the presence of DTT.(C)The commercial lansoprazole compound, like compound 6 also shows the presence of both the expected adduct and the +252 Da species, where the latter once again cannot be reduced in the presence of excess DTT.Prazoles were added to 10 μM 15 N Tsg101 UEV in NMR buffer at 2-fold excess and incubated for 24 hours.For reduction, DTT was added to 10 mM final concentration after adjusting samples to neutral pH and incubated for 30 minutes.

Figure S6 .
Figure S6.Modeling the Nucleophilic Environment around the Prazole bound to C73 of the Tsg101 UEV Domain.The previously determined pose (PDB ID: 5VKG) of tenatoprazole (sticks) on the surface of the Tsg101 UEV domain (gray cartoon) isshown to represent the primary adduct.The observed secondary adduct species formed with lansoprazole and several prazole derivatives involves apparent loss of the 4-pyridyl substituent, consistent with a nucleophilic attack at this position.The three closest nucleophilic residues to C73 on the Tsg101 surface are lysines 33, 36, and 90.The sidechain amine nitrogen of K33 is >1.5 nm from the 4-pyridyl carbon of the prazole, while K90 and K36 are <1 nm away.To determine which residue was likely to contribute to secondary adduct formation, we first generated the allowed K36 and K90 sidechain rotamers with the mutagenesis wizard in pymol (shown as blue spheres for the side chain nitrogen).For the prazole, rotations of the final three chi angles (I-III, inset) were sampled in 30° increments, with the cloud of resulting 4-pyridyl carbon positions shown as gray spheres.Several K90 rotamers overlap with the 4-pyridyl cloud, indicating that this position could likely be attacked by the K90 sidechain.For K36, the minimum distance between 4-pyridyl carbon and the sidechain amino nitrogen is 4.5 Å, and this orientation would likely be sterically blocked without rearrangement of other beta-sheet sidechains, including those of V38 and T56 (dark gray sticks).Thus, the most likely source of the secondary adduct is the sidechain amino nitrogen of K90.

Figure S7 .
Figure S7.The Secondary Adduct observed for Tsg101 and Lansoprazole is also observed in Short CK Peptides.Having observed the formation of an irreversible, secondary adduct between Tsg101 and several prazoles, we also tested for the formation of an equivalent adduct in a simplified system having only cysteine and lysine sidechain nucleophiles.Tested peptides are given in (A), with the general sequence GCGnK, where the expected adduct mass was taken from the observed delta mass for the primary lansoprazole adduct with Tsg101 UEV domain.All peptides were N-terminally acetylated and C-terminally amidated to prevent the previously characterized adduct formation with the peptide amino-terminus [peptide citation].Lansoprazole was added to each peptide, where the resulting LC/MS results are given for each peptide in (B).In each case the chromatogram for the addition of stoichiometric lansoprazole is shown at top, followed by the chromatogram with 3x excess lansoprazole, and then the observed masses for peaks corresponding to the primary and secondary adducts (bottom) in the presence of excess lansoprazole.The secondary adduct is present in each case, having the same delta mass of 100 units corresponding to loss of the TFE group as seen for Tsg101 UEV domain.Compared to the protein, however, there is much less formation of the secondary adduct, with the most seen for the 4mer GCGK sequence.Reduced formation is likely reflective of the peptide flexibility reducing preferred orientations, and potential alterations in the cysteine or lysine pKas in the protein environment.

FigureFigure S10 .
Figure S9.A sulfide prazole derivative does not form a covalent adduct with the Tsg101 UEV domain in vitro.(A) Comparison of the C73 intensity from 1-24 hours post-addition of stoichiometric compound 14 (black), or the sulfide variant compound 19 (blue).While Compound 14 rapidly forms a disulfide adduct with the UEV domain, evidenced by the near complete disappearance of C73 intensity by 4-6 hours, there is almost no change in intensity for compound 19 (< 5%) over the full 24-hour period.An overlay of the one (gray) and 24-hour (red) spectra also shows no evidence of covalent attachment, and a lack of minor shifts that imply minimal noncovalent affinity of the sulfide derivative for the UEV.Error bars in panel (A) are derived from the estimated noise level of the spectra obtained using the Estimate Noise function in NMRPipe and NMRDraw.