Articles 2-(2-Oxo-1,4-dihydro-2H-quinazolin-3-yl)- and 2-(2,2-Dioxo-1,4-dihydro-2H-2λ 6 -benzo[1,2,6]thiadiazin-3-yl)-N-hydroxy-acetamides as Potent and Selective Peptide Deformylase Inhibitors Christian Apfel, David W. Banner, Daniel Bur, † Michel Dietz, Christian Hubschwerlen, ‡ Hans Locher, ‡ Fre ´de ´ric Marlin, Raffaello Masciadri,* Wolfgang Pirson, and Henri Stalder Discovery Chemistry, F. Hoffmann-La Roche Ltd., CH-4070 Basle, Switzerland Received August 14, 2000 Potent, selective, and structurally new inhibitors of the Fe(II) enzyme Escherichia coli peptide deformylase (PDF) were obtained by rational optimization of the weakly binding screening hit (5-chloro-2-oxo-1,4-dihydro-2H-quinazolin-3-yl)-acetic acid hydrazide (1). Three-dimensional structural information, gathered from Ni-PDF complexed with 1, suggested the preparation of two series of related hydroxamic acid analogues, 2-(2-oxo-1,4-dihydro-2H-quinazolin-3-yl)-N- hydroxy-acetamides (A) and 2-(2,2-dioxo-1,4-dihydro-2H-2λ 6 -benzo[1,2,6]thiadiazin-3-yl)-N- hydroxy-acetamides (B), among which potent PDF inhibitors (37, 42, and 48) were identified. Moreover, two selected compounds, one from each series, 36 and 41, showed good selectivity for PDF over several endoproteases including matrix metalloproteases. However, these compounds showed only weak antibacterial activity. Introduction Peptide deformylase (PDF, EC 3.5.1.27) is believed to be an essential enzyme in both Gram-positive and Gram-negative bacteria. 1-3 In eukaryotes, gene se- quences similar to def (PDF gene) have been identified, but their functions have not yet been rigorously estab- lished. 4,5 PDF catalyzes the removal of a formyl group attached to the N-terminus of the leading methionine from nascent polypeptide chains. Initial studies with PDF were hampered by the fact that PDF was isolated accidentally as an almost inactive but thermodynami- cally stable zinc complex. 6 The native enzyme, however, turned out to be the oxidatively very labile Fe(II) complex. 7,8 The mechanism of action of PDF has been extensively studied. 9,10 The search for selective inhibi- tors of PDF has relied on modifying the thermolysin inhibitor thiorphane, 11-13 the calpain inhibitor calpep- tin, 14 the naturally occurring antibacterial hydroxamate actinonin, 15-18 the angiotensin II receptor antagonists from the biphenyl tetrazole class, 19 and the anticholes- teremic thyropropic acid. 20 Recently, the orally bioavail- able N-formyl hydroxylamine BB-3497 has given mo- mentum to the validity of PDF as a novel antibacterial target. 21 Screening, X-ray, and Modeling Results High-throughput screening of our compound reposi- tory in an Escherichia coli PDF based assay afforded the weakly binding (IC 50 ) 27µM) hydrazide 1 (Scheme 1), which was successfully cocrystallized with E.coli Ni- PDF (Figure 1): The hydrazide group of 1 was observed to chelate the Ni 2+ cation bidentally. The 5-chloro substituent of hydrazide 1 was localized in the hydro- phobic P 1 ′ pocket which accommodates the side chain of methionine in the natural enzyme/substrate complex. The urea carbonyl group of 1 did not undergo a hydrogen-bonding interaction with the enzyme. Com- pound 2 (Scheme 1), the R-methyl analogue of hydrazide 1, was devoid of PDF inhibitory activity, indicating an unfavorable steric interaction with the protein in the vicinity of the hydrazide function. From our X-ray structure of PDF complexed with actinonin, 18 we knew that the backbone NH of Ile44 forms an H-bond with the substrate-like inhibitor actinonin. We superimposed the two structures of PDF complexed with actinonin and hydrazide 1, respectively, and speculated that a tetra- hedral SO 2 group instead of the in-plane carbonyl group of 1 might undergo a similar hydrogen-bonding inter- action with the backbone (Figure 2). On the basis of our X-ray and modeling studies, we proposed the general structures A and B (Scheme 1), wherein the hydrazide group of 1 was replaced by a hydroxamic acid function resulting in stronger metal coordination. Substituent R 1 should be a small lipophilic substituent (F, Cl, Br, or CF 3 ), filling the hydrophobic methionine pocket optimally. Substituent R 2 , which points toward the surface of the enzyme, would probably allow broader structural variation. * To whom correspondence should be addressed: Raffaello Mas- ciadri, Ph.D., F. Hoffmann-La Roche Ltd., Pharmaceuticals Division, Discovery Chemistry PRBC-L, Bldg. 092/3.76, CH-4070 Basel, Swit- zerland. Phone: 0041 61 688 75 04. Fax: 0041 61 688 64 59. E-mail: raffaello.masciadri@roche.com. † Current address: Actelion, Gewerbestr. 16, CH-4123 Allschwil, Switzerland. ‡ Current address: Morphochem AG, Schwarzwaldallee 215, CH- 4058 Basel, Switzerland. 1847 J. Med. Chem. 2001, 44, 1847-1852 10.1021/jm000352g CCC: $20.00 © 2001 American Chemical Society Published on Web 05/05/2001