Bioorganic & Medicinal Chemistry 25 (2017) 897–911

Contents lists available at ScienceDirect

Bioorganic & Medicinal Chemistry

journal homepage: www.elsevier .com/locate /bmc

Design, synthesis and in vitro biological evaluation of oligopeptidestargeting E. coli type I signal peptidase (LepB)

http://dx.doi.org/10.1016/j.bmc.2016.12.0030968-0896/� 2016 The Authors. Published by Elsevier Ltd.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Abbreviations: dabcyl, 4-((4-(dimethylamino)phenyl)azo)benzoic acid; DCM,dichloromethane; DIEA, N,N-diisopropylethylamine; DMF, N,N-dimethylformamide;E. coli, Escherichia coli; EDANS, 5-((2-aminoethyl)amino)naphthalene-1-sulfonic acid;EDCI, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide; Fmoc, 9-fluorenylmethoxy-carbonyl; HBTU, N,N,N0 ,N0-tetramethyl-O-(1H-benzotriazol-1-yl)uronium hexafluo-rophosphate; HATU, 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate; HFIP, 1,1,1,3,3,3-hexafluoro-2-propanol;IC50, half maximal inhibitory concentration; MIC, minimal inhibitory concentration;SPPS, solid-phase peptide synthesis; TES, triethylsilane; SPase I, type I signalpeptidase; TFA, trifluoroacetic acid.⇑ Corresponding authors at: Uppsala University, Department of Medicinal

Chemistry, Organic Pharmaceutical Chemistry, BMC, Box 574, SE-751 23 Uppsala,Sweden (A. Karlén) and Department of Cell and Molecular Biology, BMC, Box 596,SE-751 24 Uppsala, Sweden (S.L. Mowbray).

E-mail addresses: Sherry.Mowbray@icm.uu.se (S.L. Mowbray), Anders.Karlen@orgfarm.uu.se (A. Karlén).

g Maria De Rosa and Lu Lu contributed equally to this work.

Maria De Rosa a,g, Lu Lu b,g, Edouard Zamaratski a, Natalia Szałaj a, Sha Cao c, Henrik Wadensten d,Lena Lenhammar e, Johan Gising a, Annette K. Roos f, Douglas L. Huseby c, Rolf Larsson e, Per E. Andrén d,Diarmaid Hughes c, Peter Brandt a, Sherry L. Mowbray b,f,⇑, Anders Karlén a,⇑aUppsala University, Department of Medicinal Chemistry, Organic Pharmaceutical Chemistry, BMC, Box 574, SE-751 23 Uppsala, SwedenbUppsala University, Department of Cell and Molecular Biology, BMC, Box 596, SE-751 24 Uppsala, SwedencUppsala University, Department of Medical Biochemistry and Microbiology, BMC, Box 582, SE-751 23 Uppsala, SwedendUppsala University, Department of Pharmaceutical Biosciences, BMC, Box 591, SE-751 24 Uppsala, SwedeneUppsala University Hospital, Department of Medical Sciences, Uppsala, SwedenfUppsala University, Science for Life Laboratory, Department of Cell and Molecular Biology, BMC, Box 596, SE-751 24 Uppsala, Sweden

a r t i c l e i n f o a b s t r a c t

Article history:Received 10 June 2016Revised 14 October 2016Accepted 3 December 2016Available online 7 December 2016

Keywords:AntibacterialsEscherichia coliOligopeptidesSolid-phase peptide synthesisType I signal peptidase

Type I signal peptidases are potential targets for the development of new antibacterial agents. Here wereport finding potent inhibitors of E. coli type I signal peptidase (LepB), by optimizing a previouslyreported hit compound, decanoyl-PTANA-CHO, through modifications at the N- and C-termini. Goodimprovements of inhibitory potency were obtained, with IC50s in the low nanomolar range. The best inhi-bitors also showed good antimicrobial activity, with MICs in the low lg/mL range for several bacterialspecies. The selection of resistant mutants provided strong support for LepB as the target of these com-pounds. The cytotoxicity and hemolytic profiles of these compounds are not optimal but the finding thatminor structural changes cause the large effects on these properties suggests that there is potential foroptimization in future studies.� 2016 The Authors. Published by Elsevier Ltd. This is anopenaccess article under the CCBY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

The frequency of antimicrobial resistance has increased signifi-cantly in recent decades.1,2 There is an urgent need to discover anddevelop new drugs with novel modes of action that are effective

against drug-resistant strains.1 Of particular interest from the clin-ical perspective are the ESKAPE pathogens (Enterococcus faecium,Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter bau-mannii, Pseudomonas aeruginosa, and Enterobacter species) thatcause the majority of healthcare-associated infections worldwide.3

In recent decades, type I signal peptidases (SPase I) have beenextensively explored as potential targets for new antibacterialagents. Signal peptidases play a crucial role in the bacterial Secand Tat secretion pathways, catalyzing the cleavage of a largelyhydrophobic N-terminal sequence (the signal or ‘‘leader” peptide)from their natural substrates (which include most pre-secretoryproteins).4 Inhibition of SPase leads to an accumulation of pre-pro-teins at the phospholipid bilayer and cell death.5,6 Bacterial SPasesare essential enzymes6,7 that differ significantly from their eukary-otic counterparts,8 making these enzymes attractive targets fordrug development.

SPases I (EC 3.4.21.89) are endopeptidases that belong to theserine protease family. They are ubiquitous and well conservedamong bacterial species. The biochemical properties of SPases Ifrom Gram-positive bacteria (B. subtilis, S. pneumoniae,9 and S.aureus10) as well as from Gram-negative bacteria have been

898 M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911

studied in detail. SPase I of E. coli, commonly known as LepB, andherein designated EcLepB, is the best characterized.11 SPase Ienzymes operate via a Ser/Lys catalytic dyad, which differs fromthe classical Ser-His-Asp triad seen in most serine proteases. Theactive site (including S91 and K146 in EcLepB)12 is located at theouter surface of the cytoplasmic membrane, making it accessible

N

S

O

Me

O

R1

O

HH

OOR2

1. Allyl (5S,6S)-6-[(R)-acetoxyethyl]-penem-3-carboxylateR1 = Acetyl, R2 = allyl

2. Allyl (5S,6S)-6-[(R)-hydroxyethyl]-penem derivativeR1 = H, R2 = p-nitrobenzyl

Carbamate derivatized penem3. R1 = -NHEt, R2 = p-nitrobenzyl4. R1 = -NHiPr, R2 = p-nitrobenzyl

N

S

O

Me

O

O

O

NO

O

5. Tricyclic pen

a) 5S-Penems

c) Oligopeptides

NO

HN

O

Me OH

HMe

NH O

HN

O

H2N

Me

NH

9. Decanoyl-PTANA-CHO

OO

MeO

Fig. 1. Classes of SP

Fig. 2. New PTANA oligo

to any inhibitor able to cross the outer membrane of Gram-nega-tives, or the cell wall of Gram-positives.11,13 SPases have evolvedrecognition criteria consistent with their need to cleave many,but not all, peptide sequences. The canonical signal peptide startswith a short positively-charged region, followed by a hydrophobicsegment of sufficient length to span the membrane in question in a

Me

O

em

b) Arylomycins

HNR4

OOH

O

NH

Me

O

HN

O

NMe

HN

O MeO

HNOH

O

OR2

R3

R1HO

Arylomycin A:R1 = R2= R3 = H; R4 = iso-C11,12,14, n-C12, anteiso-C13

6. Arylomycin A2: R4 = iso-C127. Arylomycin A-C16: R4 = n-C15

Arylomycin B:R1 = NO2; R2= R3 = H; R4 = iso-C11,12,14, n-C12, anteiso-C13,15

Arylomycin C (lipoglycopeptides):8: R1 = H; R2= deoxy-α-mannose; R3 = H or OH; R4 = iso- or n-C13,15

ase I inhibitors.

peptide analogues.

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 899

presumably a-helical conformation, then the actual recognitionsequence characterized by small, neutral residues in the �3 (P3)and�1 (P1) positions relative to the point of cleavage, the so-calledAla-X-Ala sequence. The fact that the substrates are pre-oriented,by virtue of the placement of their hydrophobic segment in thelipid bilayer, is thought to be a factor of importance for activity.12

To date, the search for effective drugs has uncovered three classesof SPase I inhibitors, namely penems (b-lactams),14,15 ary-lomycins16,17 and oligopeptides18 (Fig. 1). However, useful thera-pies targeting these enzymes have not yet been developed.

Our group is actively involved in the search for new startingpoints for antibiotic development, and targeting the LepB enzymehas been one of our areas of interest. A core premise in our workis the STOP (Same-Target-Other-Pathogen) strategy, in which one

e

FmocON

O

HN

O

Me OtBu

HMe

NH O

HN

O

O

NHT

OHMe

O Me

NH O

HN

O

O

NHTrt

O

HN

O

Me OtBu

HMe

NH O

HN

O

N

20

22

23

OHNO

HN

O

Me OtBu

HMe

NH O

HN

O

O

NHTRO

21a-c

N

BocHN

O

C6H13

13

17a-c

18

19

e

e

e

a

Fmoc-aa-COOHb, c

+

FmocN

O

HN

O

Me OtBu

HMe

NH

OHN

O

Me OtBu

HMe

NH O

HN

O

OFmoc

11: 2-CTC10: Fmoc-Asn-OH

13

15

1

FmocMe

NH O

HN

1412

OFmocHN

O

O

NHTrt

OHFmocHN

O

O

NHTrt

FmocHNO

N

C

Cl

Cl

NHTrt

Scheme 1. Solid-Phase Synthesis of Peptides 20, 21a–c, 22, 23. Reagents and conditions: a)rt; c) Fmoc-AA-OH, HBTU, DIEA, DMF, rt, 2 h or 2 h + 2 h or overnight; d) acetic anhydri

or more enzymes from different species are used to guide chem-istry efforts, generating compounds that are then tested against anumber of different pathogens. In this way, the benefits of organicsynthesis efforts are maximized. In the current study, we aimed atoptimizing the structure of the known oligopeptide decanoyl-PTANA-CHO 9 (Fig. 1, EcLepB IC50 = 13.4 lM, E. coli MIC > 500 lg/mL),18 in order to determine whether we could improve potencyagainst both the enzyme and bacterial cells.

In particular, we wished to investigate the influence ofmodifications at the N- and C-termini, while keeping thePro-Thr-Ala-Asn-Ala (PTANA) peptide sequence unchanged. Thistype of inhibitor was first reported by Bruton et al., and was basedon the consensus sequence of pre-proteins in S. aureus.19 In ourstudies, the aldehyde warhead at the C-terminus was first replaced

H

rt

OHO

HTrt

R = Me( )6

Me

C6H13(c)

(b)

(a)

rt

OO

HN

O

O

NHTrt

ONO

HN

O

Me OtBu

HMe

NH O

HN

O

O

NHTrt

b, dON

O

HN

O

Me OtBu

HMe

NH O

HN

O

O

NHTrtRO

OMe

O Me

NH

O

HN

O

O

NHTrt

b, d

b, c

O

17a-c

18

19

12

6

O

O

O

NHTrt

BocHN

C6H13

OO

HTrt

l

DIEA, DMF, rt, 2 h, then MeOH, rt, 15 min; b) 20% piperidine in DMF, 2 min + 10 min,de or decanoic anhydride, DIEA, DMF, rt, 20 h; e) HFIP:DCM 1:4, rt, 1 h, 60–88%

900 M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911

with chemically more stable but equally reactive bioisosters suchas a-keto-amides, boronic acid and ester (Fig. 2). We also intro-duced a proline at P10, which has been shown to convert serineprotease substrates into inhibitors.20,21 Finally, an analogue simplycarrying the unmodified alanine was prepared to determinewhether the electrophile moiety at the C-terminus is reallyrequired for antibacterial activity. At the N-terminus, the decanoyltail was replaced with a bulkier 4-(4-hexylphenyl)benzoyl group. Afew examples of derivatives without the fatty acid tail, but cappedat the N-terminus with an acetyl group, were also prepared, toevaluate the effects on the activity profile. In addition, a positivecharge was introduced at the P5 residue in an attempt to increasecell wall permeability22 and to improve the potency of the com-pound inside the bacterial cell. Herein, we describe the design, syn-thesis and in vitro evaluation of these new oligopeptides againstEcLepB. Other aspects, such as activity towards a panel of Gram-negative bacteria, cytotoxicity and hemolysis, are also explored.

2. Results and discussion

2.1. Chemistry

The peptides 20, 21a–c, 22, 23 were synthesized by manualsolid-phase peptide synthesis (SPPS) from 2-chlorotritylchloridepolymer resin 11 (Scheme 1, see Supplementary data for details).Standard 9-fluorenylmethoxycarbonyl amino-protected aminoacids (Fmoc-aa-OH) were used, and the protecting group wasremoved by treatment with 20% piperidine in N,N-dimethylfor-mamide (DMF). Coupling of the appropriate amino acid wasperformed in DMF using N,N,N0,N0-tetramethyl-O-(1H-benzotria-zol-1-yl)uronium hexafluorophosphate (HBTU) as the couplingreagent, in the presence of N,N-diisopropylethylamine (DIEA) asthe base, and shaking at room temperature for 2 h. Whena-branched amino acids (e.g. Fmoc-O-tert-butyl-l-threonine) orsecondary amines (Fmoc-l-proline) were involved, longer reactiontimes were preferred (14 h) or a double coupling was performed(2 h + 2 h, treating the resin with a fresh solution of reagents). Aftereach round of coupling and deprotection steps, the resin was

28

Cl

C6H13

O

2524OH

C6H13

O

d

a

N

O

C6H13

OTosMeO

O

29

C6H13

MeO

O

31

N

O

C6H13

MeO

O

NHBoc

Scheme 2. Synthesis of the modified cationic proline 16. Reagents and conditions: a) SOpyridine, anhy. DCM, 50 �C, 16 h, 90%; d) NaCN, DMSO, 50 �C, 16 h, 55%; e) Pd/C, PtO2, H2

g) LiOH, MeOH:THF:H2O (1:1:1), rt, 20 h, 97%.

extensively washed with DMF, methanol and dichloromethane,then dried under high vacuum.

Coupling of the tail at the N-terminus, using acetic anhydride,decanoic anhydride and 40-hexyl-biphenyl-carboxylic acid, respec-tively, afforded intermediates 17a–c, 18. In the case of the modifiedproline 16, the fatty acid tail was directly coupled to the commer-cially available 4-hydroxyproline 26 (Scheme 2).

Next, cleavage of resin-bound peptides 13, 17a–c, 18 and 19 bytreatment with a solution of 1,1,1,3,3,3-hexafluoro-2-propanol:dichloromethane (1:4),23 afforded peptides 20, 21a–c, 22 and 23.Scheme 2 depicts the synthesis of the modified proline 16. Startingfrom 4-hydroxyproline methyl ester hydrochloric salt 26, selectivecoupling of 40-hexyl-biphenyl carboxylic acid 24 at the nitrogenwas achieved in high yield, via generation of the correspondingchloride 25. The hydroxyl group was converted into a better leav-ing group via tosylation and subsequent cyanation of intermediate28 gave compound 29 as a mixture of two diastereomers in 55%yield. Reduction of the nitrile, using a mixture of platinum oxideand palladium on carbon in acidic medium at room temperatureunder hydrogen atmosphere, afforded the chloride salt 30, whichwas protected in high yield using di-tert-butyl dicarbonate in thepresence of triethylamine. Finally, treatment of the salt precursor31 with lithium hydroxide in a mixture of methanol-THF-water(1:1:1) afforded the desired modified proline 16. The synthesis ofthe modified alanine carrying the warheads is shown in Scheme 3.

First, the Weinreb amide 34 was prepared by coupling the com-mercially available N-tert-butoxycarbonyl-l-alanine 32 with N,O-dimethylhydroxyamine24 33, and then reduced to the correspond-ing aldehyde 35 by treatment with lithium aluminum hydride at�15 �C (Scheme 3-I). Reaction of the aldehyde 35 with the appro-priate isonitrile 38a–b in the presence of acetic acid (functioningboth as the catalyst and as the protecting group of the newlyformed secondary alcohol) at �78 �C, afforded the a-protectedalcohols 39a–b in 40–84% yields. Isonitrile 38a was prepared bymicrowave heating of a mixture of naphthylmethylamine 36 andethylformate 37 at 130 �C for 30 min. The obtained formamidewas dissolved in anhydrous THF and phosphorous oxychloridewas slowly added at �78 �C, stirring at room temperature over-

H2NOH

OMeO N

O

C6H13

b c

26

g

Cl

Cl

e f

OHMeO

O

27

N

O

CN30

N

O

C6H13

MeO

O

NH3

16

N

O

C6H13

HO

O

NHBoc

Cl2, 80 �C, 2 h, quantitative; b) NaHCO3, H2O:THF (1:2), rt, 18 h, 85%; c) p-TolSO2Cl,, EtOAc:MeOH (1:1), HCl, 1.5 h; f) Boc2O, Et3N, anhy. DCM, rt, 18 h, 70% over 2 steps;

O

O

HNHCl H2N

Me

O

O

HN

HN

MeBoc

MeO

O

O

HN

HN

MeBoc

OHHN

MeBoc

OMe

ONHMe

NHN

MeBoc

OOMe

HN

MeBoc

Oor

R1 R1R1

ab

c

d, e f

CN

40a-ba: R1 = tBu

b: R1 = 1-naphthylmethyl

NH2

O

O

Me

g

CN Me

MeMe

Me

O

O

HN

HN

Me

i Me

O

O

O

HNHCl H2N

Me

HN

Me

O

OHFmoca

N

OHN

Me

h

Fmoc

B O

OMe

H

MeMe

HCl H2N

Me

B O

OMe

H

MeMe

HN

MeO

Me

I) Synthesis of α-keto-amide warheads (41a-b)

II) Synthesis of a P1' proline analogue (45)

III) Acetylation of warheads (41b and 47)

i

32 34 35

36 37

38b

39a-ba: R1 = tBu

b: R1 = 1-naphthylmethyl

41a-ba: R1 = tBu

b: R1 = 1-naphthylmethyl

33

42 43 44 45

O NHMe

N

OH2N

Me

O NHMe

H

NHMeN

O

+

38a

41b 46

47 48

Scheme 3. Synthesis of modified alanine carrying the warheads. I) Synthesis of a-keto-amides warheads 41a–b; II) Synthesis of a P10 proline analogue 45; III) Acetylation ofwarheads 41b and 47. Reagents and conditions: a) CDI, DIEA, anhy. THF:anhy. DMF (2.5:1), rt, 18 h, 65–90%; b) LiAlH4, anhy. THF, �15 �C, 45 min, 87%; c) AcOH, ethyl acetate,�78 �C to rt, 18 h, 40–84%; d) K2CO3, H2O:MeOH (1:2), rt, 1 h, e) EDCI, dichloroacetic acid, anhy. DCM:anhy. DMSO (7:1), 0 �C to rt, 1 h, 25–63%, over two steps; f) HCl(g), 1,4-dioxane, rt, 5 min, quantitative; g) MW, 130 �C, 30 min then DIEA, POCl3, anhy. THF, �78 �C to rt, 18 h; h) 30% piperidine in ethanol, rt, 15 min, 73%; i) acetic anhydride, DIEA,anhy. DMF, rt, 3 h–overnight, 10–71%.

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 901

night. Deprotection of the obtained secondary alcohols 39a–b bytreatment with potassium carbonate and subsequent Pfitzner–Moffatt oxidation gave precursors 40a–b in 25–63% yields.Removal of the tert-butoxycarbonyl protecting group in acidicmedium afforded the final a-keto-amide hydrochloric salts 41a–bin quantitative yields. The P10 proline analogue 45 was preparedas described in Scheme 3-II. Coupling of commercially availableN-methylpyrrolidine-2-carboxamide 43 with Fmoc-l-alanine 42gave the dipeptide 44, which was treated with 30% piperidine inethanol to remove the Fmoc protecting group, yielding the desiredproduct 45. Further, the a-naphthylmethyl keto-amide 41b andcommercial (R)-boroAla-(+)-pinanediol hydrochloride 47 werecapped at the nitrogen with an acetyl group by treatment withacetic anhydride, yielding compounds 46 and 48, respectively(Scheme 3-III).

The obtained peptides 20, 21a–c, 22 and 23 were coupled at theC-terminus with the modified alanine carrying the warheads 41a–b, 45, 47 and 49. Finally, deprotection of the side chains by usingaq. 95% trifluoroacetic acid or a mixture of dichloromethane:triflu-oroacetic acid:triethylsilane (10:9:1) afforded the final oligopep-tides 50a–p (Scheme 4).

2.2. Biology and SAR

2.2.1. Characterization of expressed proteinEcLepB activity was investigated on two substrates. The first,

dabcyl-VGGTATAGAFSRPGLE-(EDANS)-OH, was based on theMycobacterium tuberculosis antigen 85A sequence. Thissubstrate gave a kcat of 135 s�1 and a Km of 2 � 10�5 M, resultingin a kcat/Km value of 6.7 � 106 M�1 s�1. The second,

Me

(HCl) NH2a, b or a, c, b

OHNO

HN

O

Me OtBu

HMe

NH O

HN

O

O

NHTrtOR1

NH

NO

HN

O

Me OH

HMe

NH O

HN

O

NH2

OR1

R2

Me

R2 =

R2

OHN

O

Me

Me

MeO H

N

O

N

O

BO

OMe

H

MeMe

BOH

OH

OH

O

20, 21a-c, 22, 23 41a-b, 45, 4749 R2: COOH

50a-p

O HN

MeR1 = H, acetyl, 4'-hexyl-biphenyl, nonyl, Fmoc

R3 = H or Boc NH

R3R3

O

Scheme 4. Coupling at C-terminus between peptides 20, 21a–c, 22 and 23 and modified alanine carrying warheads 41a–b, 45, 47 and 49. Reagents and conditions: a) HBTU orHATU, DIEA, anhy. DMF or DCM, rt, 2 h; b) 95% TFA in H2O or TFA:DCM:TES (10:9:1), rt, 2 h, 5–65% (over two steps); c) 30% piperidine in ethanol, rt, 3 h, 41–73%.

902 M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911

dabcyl-KLTFGTVKPVQAIAGYEILE-(EDANS)-OH, was based on thesequence of Streptococcus pyogenes streptokinase.25 This substrategave a kcat of 90 s�1 and a Km of 6 � 10�5 M, for a kcat/Km valueof 1.5 � 106 M�1 s�1; these values compare favorably with thereported activity of the S. pneumoniae SPase I on the same substrate(kcat 0.032 s�1, Km 1.18 � 10�4).25 The values for both substratesare in good agreement with earlier results for EcLepB.12 For furtherassay work, we decided to use the antigen 85A-based sequence.

As documented in the Supplementary data, cleavage of labeledand unlabeled substrates was investigated using mass spectrome-try. In each case, intact substrate could be detected, and was shownto decrease with time. Products were identified corresponding tothe C-terminal fragments predicted by SignalP,26 but not the N-ter-minal fragments (perhaps reflecting additional degradation of thelatter).

2.2.2. Evaluation of EcLepB IC50 and MICAll newly synthesized oligopeptides 50a–p were evaluated for

their inhibitory activity against the EcLepB enzyme, and antibacte-rial activity against a panel of strains; the results are reported inTables 1 and 2. Table 1 shows the percentage inhibition for the ref-erence compound 9, and for the new compounds with less than50% inhibition when tested at 100 lM. Table 2 presents the IC50

values (and MIC) for compounds that had more than 50% inhibitionof enzyme activity at 100 lM (% inhibition data not shown).

As seen in Table 1, our first compounds (50a–i) were very poorinhibitors of the enzyme. As expected, the analogue 50a, missingthe warhead at the C-terminus, was almost completely inactiveat a concentration of 100 lM (EcLepB percentage of inhibi-tion < 50%), indicating that a reactive electrophilic moiety at thisposition promotes inhibition. Our initial idea for optimization ofthe PTANA hit was to replace the aldehyde warhead of inhibitor9 at the C-terminus with a chemically more stable but equallyreactive group, while keeping the decanoyl tail at the N-terminusunchanged. We hypothesized that, similar to the 5S-penem typeof inhibitors, an electrophilic and highly reactive warhead couldcovalently bind S91 in the active site of LepB.27 Following the clas-sical approach of inhibiting serine protease enzymes with a-keto-amides,28 two moieties were chosen, namely tert-butyl– (50b)and a-naphthylmethyl-keto-amides (50c). Surprisingly, the inhibi-tory activity was severely compromised (compare 50b–c to inhibi-tor 9). We introduced a proline in the P10 position, which haspreviously produced effective inhibitors,20 but this was likewisenot a viable approach (50d). Removal of the fatty acid tail (50e–g) and/or truncation of the peptide PTANA sequence (50h–i) also

produced rather inactive compounds. Next, based on recent find-ings in the patent literature,29,30 we introduced a boronic esterand/or acid warhead at the C-terminus. Gratifyingly, a very goodinhibitor 50j was obtained, 25-fold more potent than the referenceinhibitor 9 (IC50 = 48 vs 1200 nM, see Table 2), that could serve as alead for further SAR studies. To start, the alkyl tail at the N-terminuswas replaced with a bulkier aromatic fatty acid group; the 40-hexyl-biphenyl moiety was chosen, as it has been shown to improveantimicrobial activity of the arylomycin series against a panel ofmicroorganisms.31 The resulting inhibitor 50k (IC50 = 12 nM) was4-fold more potent than the parent 50j. Surprisingly,29,30 thehydrolyzed analogue 50l having a boronic acid warhead showedlower potency (IC50 = 230 nM) than the corresponding boronicester derivative 50k. From these data, it was clear that modifica-tions of the warhead and tail both contributed to the significantincrease in potency of the new inhibitors compared to 9. Accordingto the literature,32,33 the fatty acid tail in the arylomycin series ispositioned close to the enzyme-membrane association surface,and the backbone is tightly bound via a hydrogen-bonding net-work. Consistent with this, replacement of the aromatic fatty acidtail with an acetyl moiety (50o) or removal of the entire fatty acidtail (50p) resulted in a large decrease in inhibitory activity(IC50 = 26,000 nM). When the warheads alone (in the N-acylatedform, 46 and 48) were tested in the same assay, inhibition was alsovery poor (Table 1). The active lipopeptides were then evaluatedagainst a panel of Gram-negative bacteria, as well as one Gram-positive organism, S. aureus (Table 2). Compound 9 lacked antibac-terial activity completely. Despite their low IC50 values, inhibitors50j–l did not show useful activity against the Gram-negativesbut some activity could be seen against S. aureus. We decided tointroduce a positive charge at the P5 proline residue of the lipopep-tides (compounds 50m and 50n), hoping to increase uptake andthereby the antimicrobial potency. Our docking study of inhibitor50n into EcLepB34 (Fig. 3), indicated that this group should fitnicely into the enzyme active site. For the modeling, the fatty acidtail of 50n was truncated to acetyl, as the enzyme lacks strongdeterminants for orienting the tail. As seen in Fig. 3, the P2, P4,and P5 side chains are exposed to bulk solvent and should toleratestructural variations without loss of target affinity. A positively-charged side chain at P5 could potentially interact with eitherE83 or D143.

The two new inhibitors, 50m–n, blocked the growth of most ofthe bacteria tested, and were much more potent in this regard thanthe reference inhibitor 9. It should be noted that these two inhibi-tors have both a bulky aromatic fatty acid tail, a key element for

Table 1Compounds showing less than 50% inhibition of EcLepB, when tested at a concentration of 100 lM. Compound 9 wasincluded as a reference. Measurements were made as triplicates.

Cmpds Structure % Inhibition

9a

O

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

Me 100

50aOH

O

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

Me 38.0 ± 2.5

50b

MeNH

O

O

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

Me MeMe

31.5 ± 4.5

50c

NH

O

O

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

Me 14.4 ± 0.6

50dN

O

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O NH

Me

Me

O

31.3 ± 2.6

50e

NH

O

O

Me

NH2

NH

O

OO

MeN

OMe OH

HN

O

NH

Me HN

HMe

MeMe

O

22.5 ± 0.9

50f

NH

O

O

Me

NH2

NH

O

OO

MeN

OMe OH

HN

O

NH

Me HN

H

O

32.3 ± 2.0

50gN

O

Me

NH

O

OO

Me

N

OMe OH

HN

O

NH

Me HN

H

O NH

Me

NH2

O

35.8 ± 0.7

50h

O

Me

NH2

NH

O

O

O

Me NH

Me HN

NH

O

Me

MeMe

O

13.3 ± 0.2

50i

O

Me

NH2

NH

O

O

O

Me NH

Me HN

NH

O

O

26.6 ± 0.3

50o

B

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

O Me

H

Me

MeMe O

26.0 ± 8.1

46

NH

O

O

Me

NH

O

Me

31.0 ± 0.2

48

B

Me

NH

O

MeO

O Me

H

Me

Me

18.0 ± 1.1

a Compound 9 is unstable, see footnote to Table 2.

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 903

Table 2Further characterization of compounds where in vitro inhibition was at least 50% in tests at 100 lM.

Cmpd Structure IC50 (nM) (95%confidence interval)

MIC (lg/mL) Cytoxicity on HepG2cellsi IC50 (lM)

Hemolysis(%)

E. c.wta

E. c.b E. c.c P. a.wtd

P. a.e K. p.wtf

A. b.wtg

S. a.wth

9j

O

Me

O

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

HMe

1200 (800–1800) >128 >128 >128 >128 128 >128 >128 >128 Ndk Ndk

50j B

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

HMe

O

O Me

H

Me

MeO

48 (30–76) >64 >64 >64 >64 >64 >64 >64 32 7.8 0.4 ± 0.1

50k B

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

O Me

H

Me

Me

MeO

12 (8.9–18) >64 64 16 >64 >64 >64 >64 4 4.7 1.1 ± 0.2

50l B

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

MeOH

OHO230 (150–360) >64 >64 >64 >64 >64 >64 >64 4 >64 1.1 ± 0.2

50mB

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

O

O Me

H

Me

Me

Me

NH3

O

18 (15–20) 32 4 1 64 8 16 32 0.5 18.6 5.7 ± 0.3

50nB

Me

NH2

NH

O

OO

N

OMe OH

HN

O

NH

Me HN

H

OH

OHMe

NH3

O

6 (4.8–8.4) 16 2 0.5 64 4 8 8 0.25 15.6 2.5 ± 0.1

50p B

Me

NH2

NH

O

O

NH

OMe OH

HN

O

NH

Me HN

H

O

O Me

H

Me

MeO

26,000 (17,000–40,000)

>64 >64 >64 >64 >64 >64 >64 >64 >64 Ndk

904M.D

eRosa

etal./Bioorganic

&Medicinal

Chemistry

25(2017)

897–911

Table2(con

tinu

ed)

Cmpd

Stru

cture

IC50(nM)(95%

confide

nce

interval)

MIC

( lg/mL)

Cytox

icityon

Hep

G2

cellsiIC

50(l

M)

Hem

olysis

(%)

E.c.

wta

E.c.b

E.c.c

P.a.

wtd

P.a.

eK.p

.wtf

A.b

.wtg

S.a.

wth

50ql

B

Me

N H

O

OO M

eO

H

H NO

N H

Me

H NH

O

OM

e

H

Me

Me

NH

3

N HNH

3

O

Me

110(80–

170)

84

132

82

32<0

.125

12.5

12.8

±0.1

aE.c.

wt=E.

coli(A

TCC25

922,

wild-type

).bE.c.=E.

coli(D

tolC

efflux-de

fectivemutant,isog

enic

toATC

C25

922).

cE.c.=E.

coli(D

22,lpx

Cmutant,drug-hyp

ersensitive

).dP.a.

wt=P.

aerugino

sa(PAO1,

wild-type

).eP.a.

=P.

aerugino

sa(PAO75

0efflux-de

fectivemutant,iso

genic

toPA

O1).

fK.p.w

t=K.p

neum

onia

(ATC

C13

833,

wild-type

).gA.b

.wt=A.b

auman

nii(A

TCC19

606,

wild-type

).hS.a.

wt=S.

aureus

(ATC

C29

213,

wild-type

).iDox

orubicinwas

usedas

thereference

compo

und,

IC50=0.23

lM.

jCom

pound9was

unstab

le;materialpu

rchased

aswellas

synthesized

,los

tactivity

withtime,

andga

vean

IC50of

6.5lM

afteron

ewee

k.kNd=not

determ

ined

.lSe

epa

tent.29

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 905

good affinity, and a cationic group added to the proline at P5,which is apparently required for permeability. It was comfortingto see that introduction of a positive charge at the proline didnot compromise the inhibitory activity. In fact, compounds 50m–nwere among the most potent EcLepB inhibitors we obtained, withIC50 values of 18 and 6 nM, respectively. The minor difference inthe IC50 values of the two compounds could potentially indicatehydrolysis of the ester into the corresponding acid, although inthe active site of EcLepB, there is space between residues S91and K146 to accommodate a bulky ester (see Fig. 3). Indeed,hydrolysis could potentially be catalyzed by the enzyme in a sim-ilar manner to that for the natural substrates. However, we notethat the observed reaction rates were linear over the 2 h timecourse of the kinetics experiment, and so if such hydrolysis isoccurring, it must be rapid (during the 10-min pre-incubationstep). In this case, the boronic ester derivative appeared to be morepotent than the boronic acid, in contrast to the 50k–l resultdescribed above.

2.2.3. Evaluation of cytotoxicity and hemolysisInhibitors 50j–p were also evaluated on the HepG2 cell line36

for cytotoxicity (for details see Supplementary data), and mostwere evaluated in a hemolysis assay; the results are reported inTable 2. It was clear that there were two pairs of potent enzymeinhibitors: 50j/50k and 50m/50n. Inhibitors 50j and 50k onlydiffer in the fatty-acid tail, and have similar EcLepB IC50s (48and 12 nM, respectively). Both are cytotoxic, with HepG2 IC50svalues of 7.8 and 4.7 lM, respectively, although each performedfairly well in the hemolysis assay. These compounds, in principle,lack activity against Gram-negative organisms, although resultswith S. aureus were more promising. Inhibitors 50m and 50n onlydiffer in the warhead, having either a boronic ester or boronicacid group. They also have good EcLepB IC50s (18 and 6 nM,respectively), although they gave poor results in both cytotoxicityand hemolysis assays. The introduction of a positive charge at P5of these two compounds appears to be correlated with the factthat they are active on most of the bacterial strains evaluated.However, these compounds effects on other properties do notfollow any clear pattern. In the case of the 50k/50m pair, theintroduction of the positively charged group is associated withless cytotoxicity, but more hemolysis; inhibition of the enzymeis not changed significantly. In the case of the 50l/50n pair,inhibition of the enzyme is better, while cytotoxicity andhemolysis are increased.

2.2.4. Profiling of LepB inhibitor from the patent literatureFinally, we synthesized one of the inhibitors from the WO

2015/023898 patent application to include it in our work as a ref-erence compound. We selected compound 50q, one of the morepotent compounds based on the reported EcLepB IC50 and E. coliMIC values, and which, furthermore, could be easily synthesized.Compound 50q is similar to our inhibitors (presented in Table 2)in structure and activity profile. The synthesis was performedaccording to the methods described in the patent, and the charac-terization of compound 50q is detailed in the Supplementary data.The most similar compound in our series 50m is 6-fold morepotent (EcLepB IC50 18 nM) than the reference compound 50q(EcLepB IC50 110 nM). However, 50q has a slightly better MIC onwild-type strains of E. coli and K. pneumonia (Table 2). Compound50q bears two positive charges as compared to one positive chargein the analogue 50m, which may explain its higher hemolyticactivity.

2.2.5. Selection of resistant mutantsThe spontaneous frequency of resistance, measured by plating

efflux-deficient DtolC E. coli at 8 �MIC, was <5 � 10�8 for 50m,

OH

OHB

Me

NH

O

O

OMe

N O

Me OH

HNO

NH

Me HN

H

ONH2

H3N

P1P2P3P4P5

Fig. 3. Molecular modeling of a covalently bound N-acetyl analogue of 50n (both diastereomers) into EcLepB (PDB code: 3IIQ),35 illustrating the direction of the hydrophobictail towards the membrane association surface. Boronic esters fit between S91 and K146, and the amine at P5 is in contact with E83 on the surface of the enzyme (thenumbering system for the protein has been adjusted to reflect the corrected sequence reported by Paetzel et al.).12 Mutations in amino acids in resistant strains from 50m arenumbered in red (vide infra). SiteMap-calculated volumes shown in yellow highlight the two hydrophobic regions contributing to the recognition of the Ala-X-Ala motif.

906 M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911

and <5 � 10�9 for 50q. Resistant mutants for whole genomesequencing were raised against both 50m and 50q by serialpassage of E. coli MG1655 DtolC at successively increasing concen-trations of each compound (Supplemental material). In each casethe great majority of the mutants (8/10 for 50m, 11/11 for 50q)had acquired mutations in lepB causing amino acid substitutions(Tables S3 and S4). The clustered location of most of the mutationsin the structure of LepB (Glu83, Phe85, Ile87, Pro88, and Ile145)supports the hypothesis that these inhibitors bind in the active siteof the enzyme (Fig. 3). The different mutations in LepB had similareffects on susceptibility, increasing MIC from 1–2 lg/mL up to 8–16 lg/mL (Tables S3, S4). The precise mechanism by which theLepB mutations cause resistance has not been investigated butthe nature of these binding-pocket mutations is consistent withthem acting to reduce compound affinity for LepB, and so supportsinhibition of LepB function as the whole cell mode of action ofcompounds 50m and 50q. Several mutants carried a second muta-tion in LepB at either Ala6 (50m) or Asp113 (50q). The significanceof these second mutations is not clear but the increase in MIC asso-ciated with mutants carrying only single amino acid substitutionsshows that a single substitution in the binding pocket is sufficientto fully account for the increase in MIC (Fig. 3, Tables S3, S4). Sev-eral mutants also carried 1 or 2 mutations in additional genes(Tables S3, S4). The significance of these additional mutations isunclear, but as 12 of 19 mutants carried only LepB mutations, wecan be confident that mutation in LepB is sufficient to confer theresistant phenotype.

3. Conclusions

The rapid development of antimicrobial resistance representsone of the major challenges in the search for new antibiotics. TypeI signal peptidases (SPases I) are for a number of reasons attractivetargets for the development of antibacterial agents with a novelmode of action. In this study, we explored the possibility of findingpotent EcLepB inhibitors by optimizing the existing hit compound

9. A set of 16 oligopeptides 50a–p structurally related to the decan-oyl-PTANA-CHO inhibitor 9 were designed, synthesized and evalu-ated. In general, good improvement of inhibitory potency wasobtained. Four inhibitors (50j, 50k, 50m and 50n) were found tobe very active, showing IC50s in the low nanomolar range. In partic-ular, inhibitor 50n was the most potent of this new series ofoligopeptides (IC50 = 6 nM). Two of the best inhibitors (50m–n)also showed promising antimicrobial activity, with MICs in thelow lg/mL range for several strains. The results obtained showedthat a bulky fatty acid tail at the N-terminus, an electrophilic andhighly-reactive warhead at the C-terminus, and a positivelycharged proline in the P5 position, are key features for obtaininggood inhibitors that can be used as potential starting points forantibiotic drug discovery. Raising mutants based on 50m and50q and performing whole genome sequencing of these enabledus to confirm the mode of action of the synthesized inhibitors.Out of 21 mutants, 19 carried mutations in LepB (12 had no addi-tional mutations in any other genes), which lends strong support tothe premise that LepB is the target for our inhibitors. Although it isalso clear that more optimization will be required to obtain inhibi-tors with favorable cytotoxicity and hemolytic profiles, we showedthat minor changes in the oligopeptides often produce large effectsin their biological profile and this information can be exploited infuture studies.

4. Experimental section

4.1. General methods

Analytical thin layer chromatography (TLC) was performedusing Merck aluminum sheets precoated with silica gel 60 F254.Column chromatography was performed on Merck silica gel 60(40–63 lm). The microwave reactions were performed in a BiotageInitiator producing controlled irradiation at 2450 MHz with apower of 0–300W. 1H and 13C NMR spectra were recorded on Var-ian Mercury Plus instruments; 1H at 399.9 MHz and 13C at

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 907

100.6 MHz at 25 �C. Exact molecular masses were determined onMicromass Q-Tof2 mass spectrometer equipped with an electro-spray ion source. Analytical RP-HPLC-MS was performed on a Gil-son RP-HPLC system with a Finnigan AQA quadrupole low-resolution mass spectrometer in positive or negative ESI modeusing an Onyx Monolithic C18 4.6 � 50 mm 5 lm (Phenomenex)with MeCN in 0.05% aqueous HCOOH as mobile phase at a flow rateof 4 mL/min. Preparative RP-HPLC was performed on a systemequipped with a Nucleodur C18 HTec 5 lm column(150 � 21.2 mm) or a Phenomenex C8 5 lm column(150 � 21.2 mm), using a H2O/CH3CN gradient with 0.1% TFA orH2O/CH3CN gradient with 20 mM TEAA, in both cases using UVdetection at 220 nm and 254 nm. The purity of the peptides wasdetermined by RP-HPLC using the columns: ACE 5 C18(50 � 4.6 mm) and Nucleodur C18 HTec 5 lm (4.6 � 50 mm),H2O/CH3CN gradient with 0.1% TFA or H2O/CH3CN gradient with20 mM TEAA and UV detection at 220 nm and 254 nm. All peptidesshowed purity above 95%.

4.2. Synthetic procedures

4.2.1. General method for the coupling between the peptides (20, 21a–c, 22, 23) and the warheads (41a–b, 45 and 47)

To a solution of the peptide (1 eq.) in anhydrous DMF (2 or5 mL) or anhydrous DCM (3 mL) HBTU or HATU or EDCI (1.2 or2 eq.) and DIEA (3 or 6 eq.) were added under N2 flow and the mix-ture was stirred at room temperature or 0 �C for 10 min, afterwhich the warhead (2 or 3 eq.) was added and the reaction mixturewas stirred at room temperature till completion (monitoring viaRPLC-MS). The solvent was evaporated under N2 stream and theresidue was used for the next step (deprotection or oxidation)without or after purification by RP-HPLC.

4.2.2. General method for the side chain deprotectionThe crude or pure protected peptides were dissolved in a solu-

tion of 95% TFA in H2O or in a solution of TFA:DCM:TES (10:9:1)and stirred at room temperature till completion (generally 1 or2 h). The solvents were removed under reduced pressure and theresidues were purified by preparative RP-HPLC.

4.2.3. Decanoyl-PTANA-OH, Decanoyl-l-prolyl-l-allothreonyl-l-alanyl-l-asparaginyl-l-alanine (50a)

The analogue was prepared according to the general procedureusing peptide 21b (0.022 mmol, 0.019 g), alanine (l-Ala-OH) 49(0.044 mmol, 0.004 g) in anhydrous DMF (2 mL), in the presenceof HBTU (0.044 mmol, 0.016 g) and DIEA (0.11 mmol, 0.014 g).The crude material was used directly for the deprotection step,after which preparative RP-HPLC purification was performed.3.5 mg (25%, over 2 steps) of desired oligopeptide 50a as a whitesolid were obtained. 1H NMR (400 MHz, DMSO-d6, 90 �C), d 12.56(s, 1H), 8.07 (d, J = 9.1 Hz, 1H), 7.89 (d, J = 6.3 Hz, 1H), 7.80 (d,J = 7.7 Hz, 1H), 7.70 (d, J = 8.3 Hz, 1H), 7.30–7.22 (m, 1H), 6.89(bs, 1H), 6.53 (bs, 1H), 4.88–4.84 (m, 1H), 4.54–4.50 (m, 1H), 4.38(dd, J = 7.8, 2.4 Hz, 1H), 4.30–4.08 (m, 2H), 4.06–4.00 (m, 1H),3.55–3.47 (m, 2H), 3.03–2.99 (m, 2H), 2.67–2.66 (m, 1H), 2.34–2.32 (m, 1H), 2.26 (t, J = 8.0 Hz, 2H), 2.20–2.12 (m, 1H), 1.91–1.87(m, 1H), 1.75–1.71 (m, 2H), 1.49–1.44 (m, 2H), 1.40–1.12 (m,16H), 1.05 (d, J = 6.8 Hz, 3H), 0.85 (t, J = 8.0 Hz, 3H). C29H50N6O9,MS (ESI): m/z 627.3 [M+H]+. RP-HPLC purity: C18 column > 95%.HRMS for C29H51N6O9 Calcd: 627.3718, Found: 627.3710.

4.2.4. (S)-N1-((S)-4-(tert-Butylamino)-3,4-dioxobutan-2-yl)-2-((S)-2-((2S,3S)-2-((S)-1-decanoylpyrrolidine-2-carboxamido)-3-hydroxybu-tanamido)propanamido)succinamide (50b)

The analogue was prepared according to the general procedureusing the peptide 21b (0.023 mmol, 0.020 g), the modified alanine

41a (0.046 mmol, 0.008 g) in anhydrous DMF (2 mL), in the pres-ence of HBTU (0.046 mmol, 0.017 g) and DIEA (0.115 mmol,0.015 g). The crude material was purified by preparative RP-HPLC,to get 9.5 mg (41%, MS (ESI) m/z 1008.8 [M+H]+) of desired pro-tected peptide as a white solid. Deprotection of the side chainswas carried out according to the general procedure (treating thecrude material with 95% TFA in H2O, 2 mL) after which the solventswere evaporated and the residue (yellow oil) was purified bypreparative RP-HPLC, affording 0.4 mg (6%) of desired oligopeptide50b as white solid. C34H59N7O9, MS (ESI): m/z 710.3 [M+H]+. RP-HPLC purity: C18 column > 95%. HRMS for C34H60N7O9 Calcd:710.4453, Found 710.4446.

4.2.5. (S)-2-((S)-2-((2S,3S)-2-((S)-1-Decanoylpyrrolidine-2-carbox-amido)-3-hydroxybutanamido)propanamido)-N1-((S)-4-((naphthalen-1-ylmethyl)amino)-3,4-dioxobutan-2-yl)succinamide (50c)

The analogue was prepared according to the general procedureusing the peptide 21b (0.024 mmol, 0.021 g), the modified alanine41b (0.048 mmol, 0.012 g) in anhydrous DMF (2 mL) in the pres-ence of HBTU (0.048 mmol, 0.018 g) and DIEA (0.12 mmol,0.015 g). The crude material was purified by preparative RP-HPLCaffording 14.7 mg (56%, MS (ESI) m/z 1092.4 [M+H]+) of desiredprotected peptide as a white solid. The deprotection of the sidechains was performed according to the general procedure andthe crude peptide was purified by preparative RP-HPLC, affording6 mg (58%) of desired oligopeptide 50c as white solid.C41H59N7O9, MS (ESI): m/z 794.3 [M+H]+. RP-HPLC purity: C18 col-umn > 95%. HRMS for C41H60N7O9 Calcd: 794.4453, Found:794.4427.

4.2.6. (S)-2-((S)-2-((2S,3S)-2-((S)-1-Decanoylpyrrolidine-2-carbox-amido)-3-hydroxybutanamido)propanamido)-N1-((S)-1-((S)-1-(methyl-carbamoyl)pyrrolidin-2-yl)-1-oxopropan-2-yl)succinamide (50d)

The analogue was prepared according to the general procedureusing the peptide 21b (0.031 mmol, 0.026 g), and the modified pro-line 45 (0.037 mmol, 0.007 g) in anhydrous DCM (3 mL), in thepresence of HBTU (0.037 mmol, 0.014 g) and DIEA (0.074 mmol,0.009 g). Deprotection of the side chains was carried out accordingto the general procedure (95% TFA in H2O, 2 mL) after which thesolvents were evaporated and the residue (yellow oil) was purifiedby preparative RP-HPLC, affording 0.014 g (62%, over two steps) ofdesired oligopeptide 50d as a white solid. 1H NMR (400 MHz,DMSO-d6, mixture of rotamers), d 8.11 (d, J = 8.1 Hz, 0.4H), 8.02–7.94 (m, 1.8H rotamer), 7.83 (d, J = 7.2 Hz, 0.6H), 7.73–7.66 (m,3.5H), 7.30 (s, 1H), 6.88 (s, 1H), 4.58–4.42 (m, 3.5H), 4.38 (dd,J = 8.1, 3.1 Hz, 1H), 4.32–4.22 (m, 5H), 4.11–3.99 (m, 1.3H), 3.61–3.27 (m, 3H), 2.56–2.54 (m, 3H), 2.35 (dd, J = 15.5, 7.3 Hz, 3H),2.26 (t, J = 7.4 Hz, 2.5H), 2.22–2.10 (m, 1H) 2.06–1.69 (m, 15H),1.49–1.44 (m, 5H), 1.28–1.18 (m, 35H), 1.04 (d, J = 6.4 Hz, 5H),0.92–0.81 (m, 5H). C35H60N8O9, MS (ESI): m/z 737.3 [M+H]+. RP-HPLC purity: C18 column > 95%. HRMS C35H60N8O9Na, Calcd:759.4381, Found: 759.4382.

4.2.7. (S)-2-((S)-2-((2S,3S)-2-((S)-1-Acetylpyrrolidine-2-carboxamido)-3-hydroxybutanamido)propanamido)-N1-((S)-4-(tert-butylamino)-3,4-dioxobutan-2-yl)succinamide (50e)

The analogue was prepared according to the general procedureusing the peptide 21a (0.033 mmol, 0.025 g), and the warhead 41a(0.101 mmol, 0.021 g) in anhydrous DMF (3 mL) in the presence ofHBTU (0.101 mmol, 0.038 g) and DIEA (0.198 mmol, 0.025 g). Thecrude material was purified by RP-HPLC obtaining 8 mg (26%) ofpure protected peptide as a fluffy white solid. C41H65N7O9, MS(ESI): m/z 896.4 [M+H]+. Deprotection of the side chains was car-ried out according to the general procedure (95% TFA in H2O,2 mL) after which the solvents were evaporated and the residue(yellow oil) was purified by preparative RP-HPLC, affording

908 M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911

3.5 mg (65%) of desired final oligopeptide 50e as white solid. 1HNMR (400 MHz, CDCl3 + 2 drops of DMSO-d6), d 4.53–4.51 (m,1H), 4.04–4.01 (m, 1H), 3.80–3.67 (m, 4H), 3.21–2.90 (m, 2H),2.05–1.96 (m, 5H), 1.53–1.36 (m, 2H), 0.78–0.55 (m, 15H), 0.24(t, J = 8.1 Hz, 3H). C26H43N7O9, MS (ESI): m/z 598.1 [M+H]+. RP-HPLC purity: C18 column > 95%. HRMS for C26H42N7O9 Calcd:596.2041, Found: 596.2061.

4.2.8. (S)-2-((S)-2-((2S,3S)-2-((S)-1-Acetylpyrrolidine-2-carboxamido)-3-hydroxybutanamido)propanamido)-N1-((S)-4-((naphthalen-1-ylm-ethyl)amino)-3,4-dioxobutan-2-yl)succinamide (50f)

The analogue was prepared according to the general procedureusing the peptide 21a (0.033 mmol, 0.025 g), and the warhead41b (0.101 mmol, 0.030 g) in anhydrous DMF (3 mL) in the pres-ence of HBTU (0.101 mmol, 0.038 g) and DIEA (0.198 mmol,0.025 g). The crude material was purified by RP-HPLC affording0.018 g (54%) of pure compound as a fluffy white solid.C56H65N7O9, MS (ESI): m/z 980.5 [M+H]+. Deprotection of the sidechains was carried out according to the general procedure (95%TFA in H2O, 2 mL) after which the solvents were evaporatedand the residue (yellow oil) was purified by preparative RP-HPLC,affording 5 mg (43%) of desired oligopeptide 50f as a white solid.1H NMR (400 MHz, CDCl3 + 2 drops of DMSO-d6), d 7.35–7.32 (m,1H), 7.24–7.20 (m, 1H), 7.15–7.12 (m, 1H), 6.80–6.68 (m, 4H),6.50–6.43 (m, 1H), 5.70–5.67 (m, 1H), 4.44–4.33 (m, 1H), 4.14(s, 2H), 4.01–3.87 (m, 1H), 3.68–3.53 (m, 4H), 2.96–2.80 (m,2H), 1.96–1.84 (m, 5H), 1.41–1.30 (m, 2H), 0.66–0.41 (m, 9H).C33H43N7O9, MS (ESI): m/z 682.0 [M+H]+. RP-HPLC purity: C18 col-umn > 95%. HRMS for C33H42N7O9 Calcd: 680.3044, Found:680.3031.

4.2.9. (S)-2-((S)-2-((2S,3S)-2-((S)-1-Acetylpyrrolidine-2-carboxamido)-3-hydroxybutanamido)propanamido)-N1-((S)-1-((S)-1-(methylcar-bamoyl)pyrrolidin-2-yl)-1-oxopropan-2-yl)succinamide (50g)

The analogue was prepared according to the general procedureusing the peptide 21a (0.027 mmol, 0.020 g), and the modifiedproline 45 (0.027 mmol, 0.005 g) in anhydrous DCM (3 mL) inthe presence of HBTU (0.027 mmol, 0.011 g) and DIEA(0.054 mmol, 0.007 g). Deprotection of the side chains was carriedout according to the general procedure (95% TFA in H2O, 2 mL)after which the solvents were evaporated and the residue (yellowoil) was purified by preparative RP-HPLC, affording 3 mg (18%,over two steps) of desired oligopeptide 50g as a white solid.RP-HPLC purity: C18 column > 95%. C27H44N8O9, MS (ESI): m/z625.3 [M+H]+. HRMS for C27H44N8O9Na Calcd: 647.3153, Found:647.3143.

4.2.10. (S)-2-((S)-2-Acetamidopropanamido)-N1-((S)-4-(tert-butyla-mino)-3,4-dioxobutan-2-yl)succinamide (50h)

The analogue was prepared according to the general procedureusing the peptide 22 (0.051 mmol, 0.025 g), and the warhead 41a(0.153 mmol, 0.032 g) in anhydrous DMF (3 mL) in the presenceof HBTU (0.153 mmol, 0.058 g) and DIEA (0.306 mmol, 0.039 g).The crude material was purified by RP-HPLC affording 0.013 g(40%) of pure protected oligopeptide as a fluffy white solid.C36H43N5O6, MS (ESI): m/z 642.1 [M+H]+. Deprotection of the sidechains was carried out according to the general procedure (95%TFA in H2O, 2 mL) after which the solvents were evaporated andthe residue (yellow oil) was purified by preparative RP-HPLC,affording 5 mg (62%) of desired oligopeptide 50h as a white solid.1H NMR (400 MHz, CDCl3 + 2 drops of DMSO-d6, mixture of rota-mers), d 7.66–7.64 (m, 1H), 7.51–7.46 (m, 1H), 7.34 (d, J = 6.4 Hz,1H), 6.86 (bs, 1H), 6.64 (bs, 1H), 5.95–5.94 (m, 1H), 4.83–4.78 (m,1H), 4.35–4.24 (m, 2H), 3.96–3.88 (m, 1H), 3.33–3.30 (m, 1H),2.21 (s, 3H), 1.60 (d, J = 8.3 Hz, 3H), 1.03–0.87 (m, 12H).C17H29N5O6, MS (ESI): m/z 400.2 [M+H]+, 799.2 [2 M+HM+H]+.

RP-HPLC purity: C18 column > 95%. HRMS for C17H30N5O6 Calcd:400.2196, Found: 400.2203.

4.2.11. (S)-2-((S)-2-Acetamidopropanamido)-N1-((S)-4-((naphthalen-1-ylmethyl)amino)-3,4-dioxobutan-2-yl)succinamide (50i)

The analogue was prepared according to the general procedureusing the peptide 22 (0.051 mmol, 0.025 g), and the warhead 41b(0.153 mmol, 0.045 g) in anhydrous DMF (3 mL) in the presenceof HBTU (0.153 mmol, 0.058 g) and DIEA (0.306 mmol, 0.039 g).The crude material was purified by RP-HPLC affording 0.018 g(49%) of pure compound as a fluffy white solid. C43H43N5O6, MS(ESI): m/z 726.1 [M+H]+. Deprotection of the side chains was car-ried out according to the general procedure (95% TFA in H2O,2 mL) after which the solvents were evaporated and the residue(yellow oil) was purified by preparative RP-HPLC, affording 5 mg(43%) of desired oligopeptide 50i as white solid. 1H NMR(400 MHz, CDCl3 + 2 drops of DMSO-d6), d 7.80–7.63 (m, 1H),7.62–7.32 (m, 2H), 7.13–7.07 (m, 3H), 6.88–6.80 (m, 1H), 5.95–5.94 (m, 1H), 4.83–4.78 (m, 1H), 4.55 (s, 2H) 4.35–4.24 (m, 2H),3.96–3.88 (m, 1H), 2.21 (s, 3H), 1.60 (d, J = 8.3 Hz, 3H), 0.87 (m,3H). C24H29N5O6, MS (ESI): m/z 484.2 [M+H]+. RP-HPLC purity:C18 column > 95%. HRMS for C24H28N5O6 Calcd: 482.2040, Found:482.2021.

4.2.12. (S)-2-((S)-2-((2S,3S)-2-((S)-1-Decanoylpyrrolidine-2-carbox-amido)-3-hydroxybutanamido)propanamido)-N1-((R)-1-((3aS,4S,6S,7aR)-3a,5,5-trimethylhexahydro-4,6-methanobenzo[d][1,3,2]dioxaborol-2-yl)ethyl)succinamide (50j)

The analogue was prepared according to the general procedureusing the peptide 21b (0.064 mmol, 0.055 g), and the warhead (R)-BoroAla-(+)-Pinanediol hydrochloride 47 (0.128 mmol, 0.033 g) inanhydrous DCM (3 mL) in the presence of EDCI (0.128 mmol,0.020 g) and DIEA (0.192 mmol, 0.025 g). DCM was evaporatedand deprotection of the side chains was performed according tothe general method (TFA:DCM:TES, 10:9:1, 3 mL). The crude yel-low oil was purified by preparative RP-HPLC (using TEAA, triethy-lammonium acetate, as the buffer, to avoid hydrolysis of theboronic ester to boronic acid), affording 0.010 g (20%, over twosteps) of desired oligopeptide 50j as a white solid. The 1H NMRshowed a mixture 4.5:1 of two rotamers. 1H NMR (400 MHz,DMSO-d6, 90 �C), d 8.77–8.64 (m, 0.5H rotamer), 8.60 (d,J = 2.9 Hz, 1H), 8.19 (dd, J = 17.6, 7.9 Hz, 0.5H rotamer), 8.13–8.10(m, 0.5H rotamer), 8.07–8.04 (m, 1H), 8.03–7.93 (m, 1H), 7.72 (d,J = 8.3 Hz, 1H), 7.37–7.29 (m, 1H), 6.93–6.94 (m, 1H), 5.02–4.99(m, 0.5H rotamer), 4.70–4.53 (m, 1H), 4.39 (dd, J = 8.0, 3.1 Hz,1H), 4.31 (dd, J = 8.6, 3.8 Hz, 0.5H rotamer), 4.27–4.15 (m, 2H),4.11 (dd, J = 8.0, 4.1 Hz, 0.5H rotamer), 4.07–4.04 (m, 2H), 3.55–3.45 (m, 2H), 2.44–2.39 (m, 1H), 2.26 (t, J = 8.6 Hz, 2H), 2.22–2.14(m, 1H + 1H rotamer), 2.04–1.76 (m, 9H), 1.65–1.61 (m, 1H),1.49–1.42 (m, 3H), 1.33–1.16 (m, 21H + 8H rotamer), 1.06–1.04(m, 3H + 0.5H rotamer), 1.02–0.98 (m, 3H), 0,92 (t, J = 8.2 Hz,0.5H rotamer), 0.87–0.84 (m, 3H + 1H rotamer), 0.80 (s, 3H).C38H65BN6O9, MS (ESI): m/z 761.3 [M+H]+. RP-HPLC purity: C8column > 95%. HRMS for C38H66BN6O9 Calcd: 761.4986, Found:761.4974.

4.2.13. (S)-2-((S)-2-((2S,3S)-2-((S)-1-(40-Hexyl-[1,10-biphenyl]-4-carbonyl)pyrrolidine-2-carboxamido)-3-hydroxybutanamido)propa-namido)-N1-((R)-1-((3aS,4S,6S,7aR)-3a,5,5-trimethylhexahydro-4,6-methanobenzo[d][1,3,2]dioxaborol-2-yl)ethyl)succinamide (50k)

The analogue was prepared according to the general procedureusing the peptide 21c (0.026 mmol, 0.025 g), and the warhead(R)-BoroAla-(+)-Pinanediol hydrochloride 47 (0.052 mmol,0.013 g) in anhydrous DCM (3 mL) in the presence of EDCI(0.052 mmol, 0.008 g) and DIEA (0.126 mmol, 0.020 g). DCM wasevaporated and deprotection of the side chains was performed

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 909

according to the general method (TFA:DCM:TES, 10:9:1, 3 mL).The crude yellow oil was purified by preparative RP-HPLC (usingTEAA as the buffer), affording 6 mg (27%, over two steps) ofdesired oligopeptide ester 50k. A small fraction (2.4 mg, 13%) ofboronic acid 50l was also isolated. Ester. 1H NMR (400 MHz,DMSO-d6, 90 �C), d 8.02–7.99 (bs, 1H), 7.89 (d, J = 3.9 Hz, 1H),7.77 (t, J = 4.5 Hz, 1H), 7.67 (d, J = 7.7 Hz, 2H), 7.60–7.52 (m,4H), 7.30 (d, J = 7.8 Hz, 2H), 4.61–4.59 (m, 2H), 4.29–4.20 (m,2H), 4.16 (d, J = 8.5 Hz, 1H), 4.08–4.02 (m, 1H), 3.62–3.59 (m,2H), 3.05–2.95 (m, 5H), 2.65 (t, J = 7.6 Hz, 2H), 2.28–2.20 (m,2H), 2.10–2.00 (m, 2H), 1.92–1.83 (m, 4H), 1.64–1.60 (m, 2H),1.34–1.24 (m, 18H), 1.05 (dd, J = 7.3, 3.4 Hz, 3H), 0.89 (t,J = 8.2 Hz, 3H), 0.83 (s, 3H). C47H67BN6O9, MS (ESI): m/z 871.5[M+H]+. RP-HPLC purity: C8 column > 95%. HRMS for C47H68BN6O9

Calcd: 871.5141, Found: 871.5136.

4.2.14. ((3S,6S,9S,12R)-9-(2-Amino-2-oxoethyl)-1-((S)-1-(40-hexyl-[1,10-biphenyl]-4-carbonyl)pyrrolidin-2-yl)-3-((S)-1-hydroxyethyl)-6-methyl-1,4,7,10-tetraoxo-2,5,8,11-tetraazatridecan-12-yl)boronicacid (50l)

Compound 50l was obtained (2 mg, 10.5%) as a by-product ofthe reaction between peptide 21c and (R)-BoroAla-(+)-Pinanediolhydrochloride 47. C37H53BN6O9, MS (ESI):m/z 701.2 [M+H�2H2O]+,719.2 [M+H�H2O]+. RP-HPLC purity: C8 column > 95%. HRMS forC37H53BN6O9Na Calcd: 759.3865, Found: 759.3861.

4.2.15. (2S)-2-((2S)-2-((2S,3S)-2-((2S)-4-(aminomethyl)-1-(40-hexyl-[1,10-biphenyl]-4-carbonyl)pyrrolidine-2-carboxamido)-3-hydroxybu-tanamido)propanamido)-N1-((R)-1-((3aS,4S,6S,7aR)-3a,5,5-trimethyl-hexahydro-4,6-methanobenzo[d][1,3,2]dioxaborol-2-yl)ethyl)succi-namide (50m)

The analogue was prepared according to the general procedureusing the peptide 23 (0.050 mmol, 0.056 g), and the warhead (R)-BoroAla-(+)-Pinanediol hydrochloride 47 (0.100 mmol, 0.026 g) inanhydrous DCM (5 mL), in the presence of HATU (0.100 mmol,0.038 g) and DIEA (0.15 mmol, 0.020 g). DCM was evaporated anddeprotection of the side chains was performed according to thegeneral method (TFA:DCM:TES, 10:9:1, 3 mL). The crude yellowoil was purified by preparative RP-HPLC (using TEAA, as the buffer),affording 6 mg (21%, over two steps) of desired oligopeptide ester50m. A small fraction (2.2 mg, 6.3%) of boronic acid 50n was alsoisolated, along with a second fraction of the de-boronation by-pro-duct 50o. Ester. C48H70BN7O9, MS (ESI):m/z 900.5 [M+H]+. RP-HPLCpurity: C8 column > 95%. HRMS for C48H71BN7O9 Calcd: 900.5406,Found: 900.5397.

4.2.16. ((3S,6S,9S,12R)-9-(2-amino-2-oxoethyl)-1-((2S)-4-(aminom-ethyl)-1-(40-hexyl-[1,10-biphenyl]-4-carbonyl)pyrrolidin-2-yl)-3-((S)-1-hydroxyethyl)-6-methyl-1,4,7,10-tetraoxo-2,5,8,11-tetraazatridecan-12-yl)boronic acid (50n)

Compound 50n was obtained (2.2 mg, 6%) as a by-product ofthe reaction between peptide 23 and (R)-BoroAla-(+)-Pinanediolhydrochloride 47. The compound was isolated as TFA salt afterpreparative RPHPLC (TFA buffer). 1H NMR (400 MHz, DMSO-d6) d8.06 (bs, 1H), 7.98 (bs, 1H), 7.86 (bs, 1H), 7.76–7.70 (m, 3H),7.63–7.58 (m, 3H), 7.53 (bs, 1H), 7.42 (bs, 1H), 7.31 (d, J = 7.6 Hz,2H), 7.16 (s, 1H), 6.98 (bs, 1H), 6.53 (s, 1H), 4.97–4.94 (m, 1H),4.67 (t, J = 8.9 Hz, 1H), 4.27–4.25 (m, 2H), 4.09–4.07 (m, 1H), 3.74(t, J = 9.7 Hz, 1H), 3.45–3.39 (m, 2H), 2.95–2.90 (m, 2H), 2.67–2.60 (m, 4H), 2.33–2.31 (m, 2H), 1.61–1.58 (m, 1H), 1.33–1.17(m, 11H), 1.10 (d, J = 6.0 Hz, 3H), 1.00 (d, J = 6.2 Hz, 3H), 0.87 (t,J = 7.8 Hz, 3H). C38H56BN7O9, MS (ESI): m/z 748.2 [M+H�H2O]+.RP-HPLC purity: C8 column > 95%. HRMS for C38H57BN7O9 Calcd:766.4311 Found: 766.4310.

4.2.17. (S)-2-((S)-2-((2S,3R)-2-((S)-1-acetylpyrrolidine-2-carboxamido)-3-hydroxybutanamido)propanamido)-N1-((R)-1-((3aS,4S,6S,7aR)-3a,5,5-trimethylhexahydro-4,6-methanobenzo[d][1,3,2]dioxaborol-2-yl)ethyl)succinamide (50o)

The analogue was prepared according to the general procedureusing the peptide 21a (0.054 mmol, 0.040 g), and the warhead (R)-BoroAla-(+)-Pinanediol hydrochloride 47 (0.108 mmol, 0.028 g) inanhydrous DCM (3 mL) in the presence of HATU (0.108 mmol,0.041 g) and DIEA (0.162 mmol, 0.021 g). DCM was evaporatedand deprotection of the side chains was performed according tothe general method (TFA:DCM:TES, 10:9:1, 2 mL). The crude yel-low oil was purified by preparative RP-HPLC (using TEAA as thebuffer), affording 5 mg (14%, over two steps) of desired oligopep-tide ester 50o. C30H49BN6O9, MS (ESI): m/z 649.1 [M+H]+. RP-HPLCpurity: C8 column > 95%.

4.2.18. (S)-2-((S)-2-((2S,3R)-3-hydroxy-2-((S)-pyrrolidine-2-carbox-amido)butanamido)propanamido)-N1-((R)-1-((3aS,4S,6S,7aR)-3a,5,5-trimethylhexahydro-4,6-methanobenzo[d][1,3,2]dioxaborol-2-yl)ethyl)-succinamide (50p)

The analogue was prepared according to the general procedureusing the peptide 20 (0.119 mmol, 0.110 g), and the commercial(R)-BoroAla-(+)-Pinanediol hydrochloride 47 (0.238 mmol,0.062 g) in anhydrous DCM (10 mL) in the presence of HATU(0.238 mmol, 0.091 g) and DIEA (0.357 mmol, 0.046 g). Afterpreparative RP-HPLC purification (TEAA buffer) 0.075 g (55%) ofpure peptide as white solid were obtained. C66H79BN6O10, MS(ESI): m/z 1127.6 [M+H]+. 0.050 g of protected compound wereused for the Fmoc protecting group removal (30% piperidine inethanol, room temperature, 3 h) and the subsequent deprotectionof the side chains (performed according to the general method inTFA:DCM:TES, 10:9:1, 3 mL). The crude yellow oil was purifiedby preparative RP-HPLC (using TEAA) affording 0.010 g (41%, overtwo steps) of desired acetate salt oligopeptide 50p as a white solid.1H NMR (400 MHz, DMSO-d6, 90 �C), d 8.27 (t, J = 6.8 Hz, 1H), 8.02(bs, 1H), 7.93 (d, J = 8.2 Hz, 1H), 7.81 (m, 1H), 4.67–4.48 (m, 1H),4.36–4.20 (m, 3H), 4.16–4.13 (m, 1H), 4.10–3.96 (m, 1H), 3.66–3.54 (m, 1H), 3.52 (s, 3H), 3.32–3.26 (m, 1H), 3.23–3.17 (m, 1H),3.10 (q, J = 7.3 Hz, 6H), 2.38–2.24 (m, 1H), 2.28–2.14 (m, 2H),2.07–2.01 (m, 1H), 1.94–1.87 (m, 5H), 1.83–1.81 (m, 1H), 1.77–1.74 (m, 1H), 1.70–1.69 (m, 1H), 1.67–1.65 (m, 1H) 1.28 (s, 3H),1.24–1.18 (m, 15H), 1.11 (d, J = 6.3 Hz, 3H), 1.05 (dd, J = 7.4,4.6 Hz, 3H), 0.81 (s, 3H). C30H51BN6O10, MS (ESI): m/z 607.1 [M+H]+. RP-HPLC purity: C8 column > 95%. HRMS for C30H52BN6O10

Calcd: 607.3627, Found: 607.3647.

4.3. Computational details

All calculations were performed using the Schrödinger SmallMolecule Drug Discovery Suite (Schrödinger Release 2015-4) usingthe OPLS-2005 force field.37 The E. coli crystal structure 3IIQ35 wasdownloaded from the PDB and prepared using Protein PreparationWizard.38 This implied the addition of hydrogens, the assignmentof protonation states and optimization of the hydrogen bond net-work followed by a restrained minimization. To model the bindingof both diastereomers of the N-acetyl analogue of 50n, theseligands were built manually into the active site of LepB, usingthe protein bound conformation of arylomycin as a template. A(R)-boroalanine was added to the C-terminal and the boron wasattached covalently to Ser91. The resulting complexes were furtherrefined by 1000 steps of conformational sampling by MCMM usingthe GB/SA continuum solvation model for water39 in the optimiza-tion step. Torsion angles sampled were those of the ligand and pro-tein side chains in close contact with the ligand. The backbone ofthe protein as well as amino acids more than 5 Å away from theligand were kept fixed. SiteMap40 was used to identify and visual-

910 M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911

ize the hydrophobic patches responsible for the Ala-X-Alarecognition.

4.4. Production of EcLepB

Details of EcLepB cloning, expression and purification are pro-vided in the Supplementary data. Briefly, a full-length constructwas designed that included a His6 tag between the two transmem-brane helices predicted at the N-terminal end, replacing residuesknown to represent a self-cleavage site in the wild typesequence.34 The resulting protein was expressed at high levels inE. coli, extracted from the membrane fraction with 1% reduced Tri-ton X-100, then purified by metal-chelation chromatography withgood yield in the presence of 0.5% detergent.

4.5. Assay

EcLepB activity and inhibition were determined using aFRET-pair assay, as detailed in the Supplementary data. Dabcyl-VGGTATA;GAFSRPGLE(EDANS)-OH, and dabcyl-KLTFGTVKPVQAIA;GYEILE(EDANS)-OH were used as substrates, where dabcyl wasthe fluorescence acceptor and EDANS was the fluorescence donor;the expected cleavage sites are indicated with arrows. The linearincrease in fluorescence, corresponding to substrate cleavage,was monitored at 22 �C for 2 h. For inhibition tests, EcLepB waspre-incubated with the compound for 10 min at 22 �C, then thereaction was followed at a substrate (dabcyl-VGGTATAGAFSRP-GLE(EDANS)-OH) concentration of 8 lM; final EcLepB concentra-tion was estimated to be 50 nM. Reaction rates were plotted as afunction of inhibitor concentration, and half maximal inhibitoryconcentration (IC50) values were determined by a non-linearregression analysis of the sigmoidal dose–response curves inGraphPad Prism (GraphPad software Inc., CA, USA).

4.6. Substrate cleavage sites

Cleavage sites were investigated for two substrates (one unla-beled peptide VGGTATAGAFSRPGL, as well as the dabcyl-VGGTA-TAGAFSRPGLE(EDANS)-OH substrate) by mass spectrometry (MS),as described in the Supplementary data. Briefly, a 5 fmol sampleof each intact substrate was loaded to the LC-MS for initial charac-terization. Subsequently, 10 lL of 5 lM substrate was incubatedwith 2.3 lM EcLepB. Samples (1 lL) were taken at 1, 2, 3, 4 and5 h, as well as overnight; each sample was diluted 1000 times with0.25% acetic acid, then loaded to the LC-MS to identify the prod-ucts, both qualitatively and quantitatively. The samples from theincubation were analyzed using nanoscale liquid chromatography(nanoAdvance, Bruker Daltonics, Bremen, Germany) coupled toan electrospray ionization quadrupole time-of-flight mass spec-trometer (Maxis Impact, Bruker Daltonics). They were injectedand desalted on a pre-column (100 lm i.d. � 20 mm, C18 Pep-Map100, 5 lm, 100 Å) at a flow-rate of 10 lL/min for 4 min. Afused silica needle emitter (15 cm, 75 lm i.d., 375 lm o.d., Picotip,New Objective, Woburn, MA, packed with Reprosil-Pur C18-AQ,3 lm resin) was used as spray emitter and analytical column.The mobile phases were 0.1% formic acid in water (buffer A) and0.1% formic acid in acetonitrile (buffer B). Each sample was ana-lyzed during a 20 min gradient (3–50% buffer B) at a flow rate of300 nL/min at 50 �C. Mass spectra were collected in full MS scanmode (m/z 50–2000).

4.7. Evaluation of minimum inhibitory concentration (MIC)

Compounds were evaluated against a panel of Gram-negativebacteria, including E. coli (ATCC 25922, wild-type; CH3130, anefflux-defective DtolC mutant isogenic to ATCC 25922; D22, a

drug-hypersensitive lpxC mutant), P. aeruginosa (PAO1, wild-type;PAO750, an efflux-defective mutant isogenic to PAO1), K. pneumo-nia (ATCC 13833, wild-type), A. baumannii (ATCC 19606, wild-type), and the Gram-positive S. aureus (ATCC 29213, wild-type).In vitro antimicrobial activity of each compound was determinedby measuring the MIC using the broth micro-dilution techniquein cation-adjusted Mueller-Hinton II medium (Becton-Dickinson,212322) according to EUCAST and CLSI guidelines (for details, seethe Supplementary data, 2.2).41

4.8. Evaluation of cytotoxicity

Compounds were evaluated in a fluorometric microculturecytotoxicity assay (FMCA) and the HepG2 cell line was used forthe study (see Supplementary data for more details).36 Cytotoxicitywas assessed after 72 h with cell survival presented as survivalindex (SI, %) defined as fluorescence in test wells in percent of con-trol cultures with blank values subtracted. The IC50 was deter-mined from log concentration-effect curves in Graph Pad Prismusing a non-linear regression analysis.

4.9. Hemolysis studies

Compounds were evaluated for hemolytic activity using redblood cells from heparinized human blood (for details, see the Sup-plementary data, 2.3).

4.10. Selection and sequence identification of resistant E. coli mutants

Mutants of E. coli DtolC resistant to compounds 50m and 50q29

were isolated by serial passage with selection for growth atincreasing concentrations of compound. Mutations were identifiedby whole genome sequencing of independently selected mutantswith increased MIC values (for details, see the Supplementary data,2.4).

Acknowledgements

This work was supported by the Swedish Research Council(521-2014-6711, 521-2013-2904, 521-2013-3105, 621-2014-6215), the Swedish Foundation for Strategic Research (no. RIF14-0078), a Swedish Institute scholarship (NS), Science for LifeLaboratory (SciLifeLab pilot infrastructure grant) and UppsalaUniversity. M.D.R. would like to thank Dr. Gunnar Lindberg forhis help with solid-phase synthesis.

Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.bmc.2016.12.003.

References

1. Payne DJ, Gwynn MN, Holmes DJ, Pompliano DL. Nat Rev Drug Discovery.2007;6:29.

2. Laxminarayan R, Duse A, Wattal C, et al. Lancet Infect Dis. 2013;13:1057.3. Boucher HW, Talbot GH, Benjamin Jr DK, et al. Clin Infect Dis. 2013;56:1685.4. Rao CVS, De Waelheyns E, Economou A, Anne J. Biochim Biophys Acta.

2014;1843:1762.5. Koshland D, Sauer RT, Botstein D. Cell. 1982;30:903.6. Dalbey RE, Lively MO, Bron S, van Dijl JM. Protein Sci. 1997;6:1129.7. Date T. J Bacteriol. 1983;154:76.8. Paetzel M, Karla A, Strynadka NCJ, Dalbey RE. Chem Rev. 2002;102:4549.9. van Roosmalen ML, Geukens N, Jongbloed JDH, et al. Biochim Biophys Acta.

2004;1694:279.10. Rao S, Bockstael K, Nath S, Engelborghs Y, Anne J, Geukens N. FEBS J.

2009;276:3222.11. Paetzel M, Dalbey RE, Strynadka NCJ. Pharmacol Ther. 2000;87:27.12. Paetzel M. Biochim Biophys Acta. 2014;1843:1497.

M. De Rosa et al. / Bioorganic & Medicinal Chemistry 25 (2017) 897–911 911

13. Auclair SM, Bhanu MK, Kendall DA. Protein Sci. 2012;21:13.14. Allsop AE, Brooks G, Bruton G, et al. Bioorg Med Chem Lett. 1995;5:443.15. Hu XE, Kim NK, Grinius L, et al. Synthesis. 2003;2003:1732.16. Holtzel A, Schmid DG, Nicholson GJ, et al. J Antibiot (Tokyo). 2002;55:571.17. Schimana J, Gebhardt K, Holtzel A, et al. J Antibiot. 2002;55:565.18. Buzder-Lantos P, Bockstael K, Anne J, Herdewijn P. Bioorg Med Chem Lett.

2009;19:2880.19. Bruton G, Huxley A, O’Hanlon P, et al. Eur J Med Chem. 2003;38:351.20. Barkocy-Gallagher GA, Bassford PJ. J Biol Chem. 1992;267:1231.21. Nilsson I, von Heijne G. FEBS Lett. 1992;299:243.22. Guilhelmelli F, Vilela N, Albuquerque P, Derengowski LdS, Silva-Pereira I, Kyaw

CM. Front Microbiol. 2013;4:353.23. Bollhagen R, Schmiedberger M, Barlos K, Grell E. J Chem Soc, Chem Commun.

1994;2559.24. Teleha C, Johnson E, Hubieki M, Combs A. Synthesis. 2011;2011:4023.25. Peng SB, Zheng F, Angleton EL, Smiley D, Carpenter J, Scott JE. Anal Biochem.

2001;293:88.26. Petersen TN, Brunak S, von Heijne G, Nielsen H. Nat Methods. 2011;8:785.27. Paetzel M, Dalbey RE, Strynadka NCJ. Nature. 1998;396:186.28. Perni RB, Pitlik J, Britt SD, et al. Bioorg Med Chem Lett. 2004;14:1441.

29. Smith PA, Roberts TC, Higuchi RI, Paraselli P, Koehler MFT, Crawford JJ,WO2015023898; 2015.

30. Higuchi RI, Roberts TC, Smith PA, Campbell D, Paraselli P, WO2013123456;2013.

31. Roberts TC, Schallenberger Ma, Liu J, Smith Pa, Romesberg FE. J Med Chem.2011;54:4954.

32. Liu J, Luo C, Smith PA, et al. J Am Chem Soc. 2011;133:17869.33. Smith PA, Roberts TC, Romesberg FE. Chem Biol. 2010;17:1223.34. Paetzel M, Strynadka NC, Tschantz WR, Casareno R, Bullinger PR, Dalbey RE. J

Biol Chem. 1997;272:9994.35. Luo C, Roussel P, Dreier J, Page MG, Paetzel M. Biochemistry. 2009;48:8976.36. Lindhagen E, Nygren P, Larsson R. Nat Protoc. 2008;3:1364.37. Banks JL, Beard HS, Cao Y, et al. J Comput Chem. 2005;26:1752.38. Sastry GM, Adzhigirey M, Day T, Annabhimoju R, Sherman W. J Comput Aided

Mol Des. 2013;27:221.39. Still WC, Tempczyk A, Hawley RC, Hendrickson T. J Am Chem Soc.

1990;112:6127.40. Halgren TA. J Chem Inf Model. 2009;49:377.41. Cockerill FR, Wikler MA, Alder J, et al. Approved Standard — Ninth Edition.

2012;32.