Pergamon
`
`Bioorganic & Medicinal Chemistry Letters 8
`
`(1998) 1231-1236
`
`&
`BIOORGANIC
`MEDICINAL
`CHEMISTRY
`LETTERS
`
`SYNTHESIS AND ANTIBACTERIAL ACTIVITY OF [6,5,5] AND [6,6,5]
`TRICYCLIC FUSED OXAZOLIDINONES
`
`D. Mark Gleave,*'1'"'" Steven
`J.
`Kristine 1). Lovasz,' Douglas C. Rohrcr,^
`
`Peter R. Manninen,2'1 Debra A. Allwine,t
`Brickner,2'"!"
`
`John A. Tucker,^ Gary E. Zurcnko^ and Charles W.
`
`Ford^
`
`$ fMedicinal Chemistry Research, Computer-Aided Drug Discovery, § Cancer and Infectious Diseases Research.
`
`
`Pharmacia & Upjohn, Kalamazoo, MI 49001
`
`Received 9 April 1998
`
`Abstract: A series of conformationally restricted, [6,5,5] and [6,6,5] tricyclic fused oxazolidinones were
`synthesized and tested for antibacterial activity. Several compounds in the trans-[6,5,5] series demonstrated
`potent in vitro and in vivo activity. This work provides valuable information regarding the preferred
`conformational orientation
`
`of the oxazolidinones at
`the
`binding
`
`site. Elsevier Science Ltd. All © 1998
`
`
`rights reserved.
`
`Introduction: Oxazolidinones are
`in this
`an exciting new class of antibacterial agents.3,4 Key examples
`
`compound DuP 721 (I)4 and Pharmacia & Upjohn's clinical
`class are DuPont's seminal
`candidates U-100592
`(2) and U-100766 (3a).^h,c These compounds show promising activity against multiple antibiotic resistant
`in
`
`and vitro animal models.
`
`strains of gram-posilive
`bacteria
`in
`
`Q
`
`N
`
`O
`
`Y
`
`N-
`
`A
`
`R
`
`J N O
`
`- NHAc
`
`1 (DuP 721)
`
`X
`2: X=F, Y=NC(0)CH?0H
`3a: X=F, Y=0
`(U-100766)
`3b: X=H. Y=0
`
`NHAc
`Nil Ac
`(U-100592)
`4: n=l ([6,5,5] series)
`or n=2 ([6,6,5]
`series)
`(where R is defined
`in Schemes
`
`1-4)
`
`We became interested in determining whether there was a preferred relative orientation of the aryl and
`oxazolidinone
`rings at the binding site. To answer
`this question a series of rigid oxazolidinone analogs were
`the general structure 4. These tricyclic fused oxazolidinones
`synthesized having
`
`have the aryl and oxazolidinone
`rings joined together by either a one or two-carbon linker, giving rise to the [6,5,5] and [6,6,5] series,
`respectively.
`Chemistry and Biology: The original racemic sytheses of the [6,5,5] and [6,6,5] frameworks are
`shown in Scheme l.3a'5 The two synthetic
`routes closely parallel each other; both involve the formation of a
`
`
`either to the cis or trans
`diastereomeric pair of
`
`cyanohydrins ((±)-7 (±)-8) which are separated and or
`
`converted
`series of tricyclic oxazolidinones. The first tricyclic analogs to be made and tested were the four DuP 721
`analogs (±>9b,
`
`(±)-10b, (±)-llb
`and (±)-12b. The lrans-[6,5,5] compound (±)-10b exhibited antibacterial
`(Table I). In contrast, (±)-9b.
`(±)-llb
`potency approximately
`
`one-half that of the parent (±)-DuP 721
`12b were devoid of antibacterial
`
`activity at the concentrations
`tested.
`Due to the promising activity of (±)-10b, we were encouraged
`[6,5,5] tricyclic analogs. Sulfonyl analogs were considered
`suitable
`
`to
`
`
`synthesize
`
`
`
`additional
`
`targets work from DuPont since earlier had
`
`and
`
`(±)-
`
`trans-
`
`0960 894X/98/$ 19.00 © 1998 Rlsevier Science Ltd. All rights reserved
`PI I: S0960-894X(98)00194-2
`
`MYLAN - EXHIBIT 1022
`
`

`
`1232
`
`D. M. Cleave et ai/Bioorg. Med. Chcm. Utt. S(IWH) 1231-1236
`
`shown that some compounds of this type demonstrate good activity.4 The chlorosulfonyl [6,5,5] tricyclic
`oxazolidinone (±)-13 was synthesized from the parent (±)-10a by treatment with chlorosulfonic acid (Scheme
`2). This compound was then used as an intermediate in the synthesis of the methanesulfonyl analog (±)-14a and
`nine sulfonamides ((±)-14b-j). (±)-14a and the N,N-dimethylsulfonamide (±)-14b exhibited only weak in
`vitro antibacterial activity (Table 1). None of the other sulfonyl analogs exhibited antibacterial activity at the
`levels tested.
`
`Scheme 1
`
`NH
`
`a-c
`
`COjEi
`
`5
`|6.5^| senes
`
`J
`
`X
`
`NHAc
`
`(±)-9a:R=H
`: [
`'—^(±>-9b: R=Ac
`
`'Q
`
`(±).10M: K=H
`±)-IOb: R-Ac
`
`O X
`
`OR
`
`N
`
`OH
`
`NC
`(±)-7: n=l, K=Bn
`(±)-8: ii=2, R=Mc
`
`il
`
`a'.b'.c*
`
`I
`
`v OMc
`
`CN
`
`(±H»
`(6,6^1 series
`
`%
`
`NHAc
`
`E
`
`J
`
`(±)-l2a: R-H
`(±)-12b: R=Ac
`(±)-12c; R=3-pyridyl
`
`Nil Ac
`
`.[C
`
`(±)-lla: R=H
`(±)-llb: K«Ac
`(±)-l Ic: R=3-pyridyl
`
`(a) Mg. MeOH, 92%; (b) CBZCI. NaHCOv 90%; (c) LIBH4, 75%; (d) oxalyl chloride. DMSO. 94%; (e) (CH^QCN^H. K2C03.
`75%; (a1) II2. Pd/C. 99%; (b) Ra-Ni, NaH2P02. 79%. (c*) TMSCN (0 DH3.DMS; (g) Ac20; (h) K2COv CH,CN 41% for 9a (3
`steps: f.g and h). 40% lur 10a (3 steps). 37% for I la (4 steps: c'.f.g and h). 43% for 12a (4 steps). (1) Ac20, M^OH, Ms20. 55%
`for 9b. 67% for 10b, 87% for I lb. 79% for I2b. (j) NBS, benzoyl peroxide then ArB(UMe)2. Pd(PPhj)4. K2HPO4. 64% for 11c.
`55% for 12c.
`
`Scheme 2
`
`(±)-10a
`
`o
`a-5
`o
`
`A
`
`NHAc
`
`(±)-I3
`
`bore
`
`9
`R-S
`o
`
`O A
`
`Nil Ac
`
`(±)-14a-j
`
`(a) ClSOjOH, 72%; (b) (i) Na^Oj, NaHCOj (ii) Mel; (c) R|R2NH. Et^N. R groups (yield from 10a): 14a, Me (28%); 14b,
`NMe2 (87%); 14c, C-CJH9NH (89%); 14d. EtNH (37%); 14e. PhCH?NH (36%); 14f, PhNH (50%); 14g, 0(CH2CH2)2N (25%);
`14h, CH2(CH5CH2)2N (64%); 14i, UOiCHJJIMt (82%); 14J. NH, (65%).
`
`A 1990 DuPont patent disclosed a variety of 3-phcnyl-5-acetamidomethyl oxazolidinones having potent
`antibacterial activities, wherein aromatic and heteroaromatic moieties were coupled at the para position of the
`phenyl ring.6 The most active analogs reported were those having 3-pyridyl, 4-pyridyl and 4-cyanophcnyi
`substituents. We wished to explore the utility of these and related groups appended to the trans-[6,5,5] tricyclic
`
`

`
`D. M. Cleave el ul. / Bioorg. Med. Chem. Lett. 8(1998) 1231-1236
`
`1211
`
`oxazolidinone framework. The parent tricyclic oxazulidinone (±)-10a was selectively brominalcd upon treatment
`with NBS and benzoyl peroxide to give aryl bromide (±)-15 (Scheme 3). A series of racemic aryl- and
`heteroaryl-substituted [6.5,5] tricyclic oxazolidinoncs ((±)-16a-e) were synthesized by treating (±)-15 with
`various aryl boronic acids under Suzuki palladium-catalyzed cross coupling conditions.7 The cis- and trans- 3-
`pyridyl [6,6,5] tricyclic oxazolidinones (±)-llc and (±)-12c were synthesized in an analogous manner (Scheme
`1 ) .
`
`Scheme 3
`
`(±)-10n
`
`Br
`
`i N O
`
`NHAc
`
`(±)-15
`
`b
`
`Ai
`
`A
`
`NHAc
`
`(±)-16a-«*
`
`(a) NBS, benzoyl peroxide. 98%; (b) ArD(OH)2. (PhjP^Pd. KHPO^. Ar groups (yield from 10a): 16a. Ph (36%); 16b, 3-pyridyl
`(45-65%); 16c, 4 pyridyl (55%); 16d, 4-cyanophcnyl (80%); 16*% 5-pyrimidyl (40%).
`
`Several of the rran5-[6,5,5j analogs (16a-e) had very potent antibacterial activity (Table 1); the most
`potent being the 3-pyridyl compound (±)-16b. This compound had in vitro antibacterial activity approximately
`twice as potent as (±)-DuP 721. Unfortunately, (±)-16b was found to elicit toxic effects in the rat when
`Interestingly, the /ra/zi-3-pyridyl [6,6,5] tricyclic
`administered orally at 100 mg/kg b.i.d. for 30 days.K
`oxazolidinone (±)-12c also demonstrated weak antibacterial activity (Table 1).
`An asymmetric synthesis of the parent trans-16,5,5] tricyclic molecule (±)-10a was developed so that we
`would have access to these active [6.5.5] tricyclic analogs in enantiomerically enriched form 9 (-)-10a (98.8%
`ee) was converted into the optically active DuP 721 analog (+)-10b using the same conditions employed in the
`racemic series (Scheme 1). (+)-10b was at least twice as potent in vitro as the racemate (Table 1). This indicates
`that the enantiomer having the 5-(S)-acctamidomelhyl side chain is responsible for biological activity which is in
`accord with earlier findings, in the non-tricyclic series, at DuPont.4
`
`Scheme 4
`
`F
`
`17
`
`g.h
`
`a.b.c
`
`NO2
`
`CN
`
`W"
`
`18
`
`NO2
`
`d.e.f
`
`CC^Me
`
`VyN
`(±)-20
`
`NHCBZ
`
`v c
`O
`
`OTBS
`
`O
`
`N
`
`NHCBZ
`
`19
`
`O X
`
`N O
`
`N-
`
`OH
`
`(±)-2l: R=OTBS
`(±)-22: R=NHAc
`
`R
`
`(a) Morpholinc. MeCN, 93%; (b) DIBALH, PiiCHj. 46%; (c) (CF3CH20)2P(0)CH?C02Mc, KHMDS, THE, 76%; (d) DIBALH.
`PhCH,. 91%; (e) SnClj.HjO, EtOH, 93%; (f) CBZCl, NaHCOj, 76%; (g) m-CPBA. CHC1V NaHCO,. 76%; (h) TBSCl, imidazole,
`DMF, 62%; (i) LHMDS. TIIF. 70%; (j) Bu4NF, TUP, 82%; (k) MsCI, EI3N. CH2C12, 95%; (1) NaNj, DMF, 94%; (m) H2) Pd/C,
`MeOH, 100%; (n) Ac20, pyridine, 76%.
`
`

`
`1234
`
`D. M. Cleave et ai / Bioorg. Med. Chem. Lett. H {1998) 1231-1236
`
`Tlie morpholino tricyclic molecule 22 (Scheme 4) was chosen as a synthetic target. A priori it was
`anticipated that this molecule would have good antibacterial activity. It is the tricyclic des-fluoro equivalent of the
`current clinical candidate U-100766 (3a). Previous work in our laboratories and at DuPont suggests thai a
`fluorine substituent provides only a modest potency enhancement compared to the corresponding des-fluoro
`compounds. Earlier, more direct attempts at synthesizing 22 from the parent 10a or the aryl bromide 15 had
`proved unsuccessful10 and so an independent synthesis was developed. A strategy similar to that used in the
`asymmetric synthesis of (-)-10a was chosen.9 A key reaction in the synthesis was a novel intramolecular lithio-
`carbamate/epoxide cyclization (20 to 21). This led to the formation of both five membered rings of the tricyclic
`oxazolidinone in one step with complete regiochemical control. Surprisingly, (+)-22 exhibited poor antibacterial
`activity (Table 1).
`Discussion: A novel series of 16,5,5] and [6,6,51 tricyclic oxazolidinones has been prepared and
`examined for antibacterial activity. All cis analogs tested were inactive. In the trans series the [6,5,5]
`compounds proved much more potent than the corresponding |6,6,5] compounds. Many of the trans-[6,5,5]
`analogs had potent activity (i.e. (±) llb and (±)-16a-c), similar to that of their non-tricyclic counterparts. The
`most potent was the 3-pyridyl analog (±)-16b. Some m2W5-[6,5,5| tricyclic analogs did not follow this trend and
`had surprisingly poor antibacterial activity (i.e. (±)-14a-j and (±)-22). To account for some of these
`observations, molecular modeling studies were performed.
`Energy minimization calculations (MM2) performed on DuP 721 (1) showed that a negative torsional
`angle between the aryl and oxazolidinone rings is favored (-29° calculated for 'average' low energy
`conformations).11 This is comparable to that found in the analogous /rani-[6,5,5] compound (+)-10b (-24°
`calculated) but opposite to that found in the cis compound (±)-9b (+21° calculated).12 The activity of (±)-10b
`(ca. half that of (±)-l) and the inactivity of (±)-9b suggest that (±)-l and related compounds may bind to their
`receptor via a low energy conformation in which the torsional angle about the aryl N bond adopts a value roughly
`similar to that found in (±) 10b and other trans analogs.
`A comparison of the energy minimized structures of 10b and DuP 721 (1) (Figure 1 A) reveals a disparity
`in the orientation of the aryl ring attached to the oxazolidinone nitrogen in each of these two compounds. This
`difference arises in part from bond angle distortions introduced by the presence of a second five-member ring in
`10b. Despite rhis difference m™.?-[6,5,5] compounds, such as 10b, are still able to bind to the receptor, as
`evidenced by the high potency of 10b and the aryl series of compounds (±)-16a-e. This trend does not hold true
`for the sulfonyl analogs (±)-14a-j or for the morpholino analog (+)-22 which have much weaker activity. It is
`assumed that the small size of the acyl substituent in 10b and the planarity of the aryl substituents in (±)-16a-e
`enable those tricyclic compounds to still comfortably fit the receptor. This is not the case when the bulkier
`morpholino group is attached to the tricyclic framework or when a sulfonyl substituent is used. Clearly, subtle
`forces are in effect, and it is hoped that the synthesis and testing of additional trans-[6,$95] tricyclic analogs will
`provide further enlightenment.
`Figure IB illustrates that in trans-[6,6,5] analogs such as (±)-12b the position of the aryl ring more
`closely approximates the position of the aryl ring found in active non-tricyclic oxazolidinones such as (±)-l. A
`priori it was therefore predicted that trans-[(>,(),5] analogs would have potent antibacterial activity. This proved
`not to be the case. Their lack of activity is assumed to be due to the additional steric bulk imparted by the ethylene
`linker.
`
`

`
`D. M. Gleave et al. / Bioorg. Med. Chenu Istt. 8 (1998) 1231-1236
`
`1235
`
`Table 1.
`
`Antibacterial Activity of [6,5,5] and 16,6,5] Fused Tricyclic Ovazolidinoneb
`
`R
`
`N' O
`
`n
`
`NHAc
`
`n
`
`C.mpch
`9a
`10a
`1 1 u
`12a
`9b
`l i b
`1 2 b
`10b
`(*)10b
`14a
`14b
`14c-j
`16a
`16b
`
`1 6h.HCI
`16c
`
`16d
`16c
`1 1c
`1 2 c
`2 2
`
`C o m p a m l o r d i u ^ s
`DuP 721
`0
`(J±tll
`U-100766
`((-)-*»)
`(-)-3b
`vuiu'uinycin
`
`0
`
`0
`
`R
`II
`H
`H
`H
`Ac
`Ac
`Ac
`Ac
`Ac
`SO-Me
`SO.NMe;
`SO-NRiR:
`Ph
`3 pyridyl
`
`3-p\Tidvl.HCl
`4-pyridyl
`
`4-cyanophcnyl
`5 pyrimidyl
`3-pyndyl
`? -pyridyi
`morpholino
`
`Ac
`
`murphulino
`
`morpholino
`
`S.a.\<
`>32
`>32
`>64
`>64
`> 1 6
`>16
`> 1 6
`8
`4
`64
`64
`>64
`
`4
`
`2
`4
`
`4
`8
`>64
`32
`64
`
`2-4
`
`24
`
`4
`0.5
`
`S.a.2r
`>32
`>32
`>64
`>64
`>16
`>16
`>16
`
`4
`32
`64
`>64
`
`2
`
`2
`2
`
`4
`4
`>64
`16
`64
`
`4
`
`4
`
`4
`0.5
`
`MIC (MS/ml')*
`TP
`>^2
`>32
`>64
`>64
`>16
`>16
`> 1 6
`8
`2
`64
`64
`>64
`2
`2
`
`>32
`>32
`>64
`>64
`>16
`>16
`>16
`16
`4
`64
`.^4
`>64
`4
`2
`
`1
`2
`
`4
`>64
`Iti
`32
`
`4
`
`1
`
`2
`0 5
`
`2
`
`2
`4
`>64
`32
`128
`
`8
`
`2
`
`4
`4
`
`fefeJ £e:
`
`R f
`
`>32
`>32
`>32
`>32
`>16
`>16
`>16
`
`4
`>32
`>32
`>32
`32
`2
`
`4
`8
`
`32
`8
`>32
`>32
`>16
`
`8-16
`
`4
`
`4
`> 1 6
`
`CDcAh
`S.a.l
`
`39.5' (1.8 Vk
`
`>25J(2 2Vk
`9.4 (2.9)k
`4.4J (I.IV^
`6 0 ( 1 . 8 ) ^
`6.1 (2.1)^
`16 V (S.9V^
`17.1 (2.5 )k
`11.4 (2.5)k
`
`>20 (2.2)'
`
`9 4 (4.8^
`18 I1 (T^)1 k
`2.0-15
`
`1.1-4.8
`
`>32
`>32
`>64
`>64
`>16
`>16
`> 1 6
`
`4
`16
`64
`>64
`
`05
`
`2
`
`>64
`16
`32
`
`2
`
`0.5
`
`1
`0.5
`
`•Minimum inhibitory conccnirauon: lowest conccnirauon of diug lhai inhibits visible growth of the organism. bEffccUve dosc50:
`amount of drug required (mg/kg body weight/do^;) to cure 50% of infected mice subjected to a lethal systemic infection (drug
`administered orally) cDala IN for racemic maierial unlets otherwise noted. AStaphylococcus aureus UC (Upjohn Culture) 9213
`(mcthicillin susceptible). 9Staphylococcus aureus UCI2673 (methicillin resistant). Staphylococcus epidermidis UC 30031
`*Enterococcus faecalis UC9217. ^Streptococcus pneumoniae UC99I2. iBacteroides frag His UC12199. These values ai* for
`Staphylococcus aureus UC9271. kFigures in parenthesis arc for the vancomycin control. 'Figure in parenthesis is for the U 100592
`control, which is cquipotcnt to vancomycin.
`
`

`
`1236
`
`D. M. Cleave el al. / Bioorg. Med. Chem. ten. 8 (1998) I23I-I236
`
`Overall, these studies demonstrate the feasibility of preparing active oxazolidinone antibacterial agents
`which are conformationally constrained by a linker between the oxazolidinone ring and an attached phenyl group.
`The comparative SAR of these compounds and their non-tricyclic counterparts provides a number of interesting
`insights concerning the structural requirements for antibacterial activity.
`
`A
`
`B
`
`Figure 1: Superimposition of DuP 721 (1) with; (A) r7-a/w-[6,5,51 tricyclic analog 10b and (B) fran,j-[6,6,5]
`tricyclic analog 12b.
`
`Acknowledgment: The authors thank John W. Allison, Ronda D. Schaadt and Betty H. Yagi for in
`vitro data; Judith C. Hamel and Douglas Stapert for the in vivu data.
`References and Notes:
`1. Current address; NeoRx Corporation, 410 W. Harrison, Seattle, WA 98119.
`2. Current address: Central Research, Pfizer, Inc, Groton, CT 06340.
`(a) Brickner, S.J. Current Pharmaceutical Design 1996, 2, 175; (b) Brickner, S.J.; Hutchinson, D.K.;
`3.
`Barbachyn, M.R.; Manninen, P.R.; Ulanowicz, D.A.; Garmon, S.A.; Grega, K.C.; Hendges, S.K.;
`Toops, D.S.; Zurenko, G.E.; Ford, C.W. J. Med. Chem. 1996, 39, 673; (c) Barbachyn, M.R.;
`Hutchinson, D.K.; Brickner, S.J.; Cynamon, M.H.; Kilbum, J.O.; Klemens, S.P.; Glickman, S.E.;
`Grega, K.C.; Hendges, S.K.; Toops, D.S.; Ford, C.W.; Zurenko, G.E. J. Med. Chem. 1996, 39, 680.
`(d) Shinabarger, D.L.; Marotti, K.R.; Murray, R.W.; Lin, A.H.; Melchior, E.P.; Swaney, S.M.; Dunyak,
`D.S.; Demyan, W.F.; Buysse, J.M. Antimicrnh. Agents Chemother. 1997, 41, 2132.
`4. Gregory, W.A.; Brittelli, D.R.; Wang, C.-L.J.; Wuonola, M A.; McRipley, R.J.; Eustice, D.C.; Eberley,
`V.S.; Bartholomew, P.T.; Slee, A.M.; Forbes, M. J. Med. Chem. 1989, 32, 1673.
`(a) Brickner, S.J.; Manninen, P.R.; Ulanowicz, D.A.; Lovasz, K.D.; Rohrer, D.C. 206th ACS National
`Meeting, Chicago, IL. August 22-27, 1993. Abst. Pap. Am. Chem. Sac. 206 (1-2); (b) Brickner, S.J.;
`Manninen, P.R.; Ulanowicz, D.A.; Zurenko, G.E.; Schaadt, R.D.; Yagi, B.H.; Lovasz, K.D. 33rd
`Interscience Conference on Antimicrobial Agents and Chemotherapy, 1993, paper 72.
`6. Carlson, R.K.; Park, C.-H.; CJregory, W.A. US Patent 4,948,801 Aug. 14, 1990.
`7. Yanagi, T.; Miyaura, N.; Suzuki, A. Synth. Comm. 1981,11, 513.
`8. Unpublished results of Palmer, J.R.; Piper, R.C. and Platte, T.F. Pharmacia & Upjohn.
`9. Gleave, D.M.; Brickner, S.J. J. Org. Chem. 1996, 61, 6470,
`10. More recently, conditions have been found whereby the parent 10a can be selectively nitrated. This
`potentially provides an alternative, more direct, route to the morpholino and other amino analogs.
`11. The lowest energy conformation was obtained by employing the MM2 potential. A conformational space
`search was then performed to locate a large population of low energy conformations. These were grouped
`and analyzed to evaluate conformational populations and geometric parameters. An average torsional angle
`was calculated for the largest population of low energy conformations.
`12. Corresponding values for (±)-llb and (±)-12b are +15° and -19° respectively.
`
`5.