Structure, Volume 20
`
`Supplemental Information
`
`Staphylococcus aureus FabI: Inhibition, Substrate
`Recognition, and Potential Implications
`for In Vivo Essentiality
`Johannes Schiebel, Andrew Chang, Hao Lu, Michael V. Baxter, Peter J. Tonge,
`Caroline Kisker
`
`Inventory of Supplemental Information
`Inventory of Supplemental Information ................................................................................ 1 
`Supplemental Data ................................................................................................................... 2 
`Figure S1, related to Figure 1. Substrate and inhibitor structures. ............................ 2 
`Figure S2, related to Figure 2. Inhibitor and cofactor electron densities. .................. 3 
`Figure S3, related to Figure 6. Quaternary structure of saFabI in solution. .............. 4 
`Figure S4, related to Figure 5. Active site formation. ................................................... 5 
`Figure S5, related to Figure 6. Cooperativity of substrate binding to saFabI. ........... 6 
`Table S1, related to Experimental Procedures. ............................................................. 8 
`Table S2, related to Experimental Procedures. ............................................................. 9 
`Table S3, related to Table 3. .......................................................................................... 10 
`Table S4, related to Experimental Procedures. ........................................................... 11 
`Movie S1, related to Figure 5. Ordering of the substrate binding loop region. ........ 13 
`Movie S2, related to Figure 6. Dimer - tetramer transition. ...................................... 13 
`Supplemental Results ............................................................................................................. 14 
`Inhibition kinetics ............................................................................................................... 14 
`Reduction of straight-chain substrates by saFabI ........................................................... 14 
`Substrate specificity of saFabI .......................................................................................... 15 
`Supplemental Discussion ....................................................................................................... 16 
`Glutamate sensitivity .......................................................................................................... 16 
`Supplemental Experimental Procedures .............................................................................. 18 
`Cloning ................................................................................................................................ 18 
`Expression and purification .............................................................................................. 18 
`Crystallization, data collection and structure determination ........................................ 19 
`Steady-state kinetic assays ................................................................................................. 21 
`Inhibition kinetics ............................................................................................................... 23 
`Supplemental References ....................................................................................................... 24 
`
`1
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00001
`
`

`

`Supplemental Data
`
`
`
`
`
`Figure S1, related to Figure 1. Substrate and inhibitor structures. The chemical structures
`
`of the saFabI substrates and inhibitors, which were analyzed in this study, are shown.
`
`2
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00002
`
`

`

`
`
`
`
`Figure S2, related to Figure 2. Inhibitor and cofactor electron densities. (A-C) 2Fo-Fc
`
`omit maps of triclosan (CPP, EPP) and NADP+ in the TCL-2 (CPP, EPP) structure (subunit
`
`H) contoured at one time the standard deviation of the mean electron density.
`
`3
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00003
`
`

`

`
`
`
`
`Figure S3, related to Figure 6. Quaternary structure of saFabI in solution. (A) Dimeric
`
`and tetrameric structure of saFabI. Overlay of two different analytical SEC chromatograms
`
`using buffers at pH 5.6 (20 mM trisodium citrate pH 5.6, 500 mM NaCl and 100 mM EDTA)
`
`shown in red, or at pH 8.0 (20 mM Tris pH 8.0 and 200 mM NaCl) shown in blue,
`
`respectively. (B) Dimer - tetramer transition. Four different analytical SEC chromatograms
`
`and the respective cropped SDS-PAGE gel bands for the peak fractions (elution volumes of
`
`14.0 and 16.0 ml) running at the same height (~33 kDa, calculated monomeric saFabI
`
`molecular mass is 31 kDa) are shown with apo saFabI in black, and saFabI treated with
`
`NADP+, NADPH or NADP+ and TCL in red, orange or blue, respectively. The front peak
`
`corresponds to a molecular mass of 131 kDa, the middle peak to 47 kDa. Excess of unbound
`
`NADP+ and NADPH elutes at 19.8 ml.
`
`4
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00004
`
`

`

`
`
`
`
`Figure S4, related to Figure 5. Active site formation. Superposition of the apo-2 (blue,
`
`subunit A) and TCL-2 (yellow) structures. All three active site residues are depicted as stick
`
`representations for both structures. TCL and NADP+ are shown in grey. The three active site
`
`residues (Tyr147, Tyr157, Lys164) move significantly upon cofactor and inhibitor binding.
`
`Tyr147 is shifted by ~5 Å out of the nicotinamide binding pocket whereas Tyr157 rotates
`
`about 130° towards the TCL hydroxyl group (Priyadarshi et al., 2010; Rozwarski et al., 1999).
`
`At the same time, the 80° rotation of Lys164 allows the correct placement of the cofactors
`
`nicotinamide ring (Rozwarski et al., 1999).
`
`5
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00005
`
`

`

`
`
`
`
`Figure S5, related to Figure 6. Cooperativity of substrate binding to saFabI. Reaction
`
`velocities with 22 nM saFabI are plotted as a function of substrate concentration, holding the
`
`other substrate concentration constant. (A) In the absence of potassium glutamate, cooperative
`
`binding of NADPH at 100 µM trans-2-octenoyl-CoA can be observed. To approximate the
`
`Hill coefficient (h) despite the truncated data set, best-fit curves to h = 1, 2, 3 and 4 are
`
`displayed. Note that h = 2.2 was previously reported for the binding of NADPH to saFabI
`
`(Heath et al., 2000). (B) In the presence of 1 M potassium glutamate, h = 0.9  0.1 for the
`
`binding of NADPH at 100 µM trans-2-octenoyl-CoA. (C) In the absence of potassium
`
`glutamate, h = 2.3  0.3 for the binding of trans-2-octenoyl-CoA at 350 µM NADPH (R2 =
`
`6
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00006
`
`

`

`0.99). (D) In the presence of 250 mM, 500 mM and 1 M potassium glutamate, h = 2.4  0.4,
`
`1.4  0.3 and 1.1  0.3, respectively, for the binding of trans-2-octenoyl-CoA at 350 µM
`
`NADPH (R2 = 0.99 for each curve-fit). In addition, note the rise in Vmax as the concentration
`
`of glutamate increases. (E) Potential glutamate binding site. The ambiguous density of the
`
`2Fo-Fc omit map (CPP structure, subunit G) contoured at one time the standard deviation of
`
`the mean electron density is shown between Arg103 located on SBL-2 (cyan) and Asn205
`
`located on SBL (red). The remainder of the protein is omitted for clarity.
`
`7
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00007
`
`

`

`Table S1, related to Experimental Procedures. Data collection and refinement statistics
`
`(apo structures).
`
`
`Data collection
`Cell dimensions
`a, b, c (Å)
`α, β, γ (°)
`Space group
`Resolution1 (Å)
`Observed reflections
`Unique reflections
`Completeness (%)
`Average redundancy
`2 (%)
`Rmerge
`<I / σ(I)>
`Monomers per AU
`Refinement
`Resolution (Å)
`3
` (%)
`Rcryst
`Rfree (%)
`Twin fraction
`Number of atoms
`rmsd bond lengths (Å)
`rmsd bond angles (°)
`Average B-factor (Å2)
`Ramachandran-plot4
`Favored (%)
`Allowed (%)
`Outliers (%)
`Maximum likelihood based
`estimated coordinate error (Å)
`PDB code
`
`Apo-1
`
`
`123.8, 123.8, 190.1
`90.0, 90.0, 120.0
`P32
`50.0-3.05 (3.20-3.05)
`160,521 (21,802)
`61,623 (8,296)
`99.3 (99.7)
`2.6 (2.6)
`11.7 (49.9)
`6.8 (2.1)
`12
`
`44.3-3.05
`20.3
`24.6
`0.33
`20,062
`0.015
`1.482
`67.1
`
`91.5
`7.9
`0.6
`0.28
`
`Apo-2
`
`
`87.5, 87.5, 307.2
`90.0, 90.0, 90.0
`P43212
`50.0-2.45 (2.60-2.45)
`319,536 (52,121)
`44,583 (7,185)
`98.9 (99.8)
`7.2 (7.3)
`8.8 (78.0)
`16.4 (2.8)
`4
`
`48.2-2.45
`18.5
`25.1
`-
`7,820
`0.007
`0.980
`62.9
`
`96.7
`3.1
`0.2
`0.30
`
`4ALN
`
`4ALM
`
`
`
`1Values in parenthesis refer to the highest resolution shell
`
`2
`
`R
`merge
`
`
`
` 
`I
`
`hkl
`
`i
`
`i
`
`I
`
`
`
`hkl
`
`i
`
`I
`
`i
`
`
`
`3
`
`R
`cryst
`
`
`
` 
`F
`obs
`
`hkl
`
`F
`calc
`
`F
`obs
`
`
`
`
`
`hkl
`
`4According to Molprobity (Davis et al., 2007)
`
`
`
`8
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00008
`
`

`

`Table S2, related to Experimental Procedures. Data collection and refinement statistics
`
`(ternary complex structures).
`
`
`Data collection
`Cell dimensions
`a, b, c (Å)
`α, β, γ (°)
`Space group
`Resolution1 (Å)
`Observed reflections
`Unique reflections
`Completeness (%)
`Average redundancy
`2 (%)
`Rmerge
`3 (%)
`Rpim
`<I / σ(I)>
`Monomers per AU
`Refinement
`Resolution (Å)
`4
` (%)
`Rcryst
`Rfree (%)
`Twin fraction
`Number of atoms
`rmsd bond lengths (Å)
`rmsd bond angles (°)
`Average B-factor (Å2)
`Ramachandran-plot5
`Favored (%)
`Allowed (%)
`Outliers (%)
`Maximum likelihood based
`estimated coordinate error (Å)
`PDB code
`
`
`
`TCL-1
`
`
`83.5, 111.9, 111.6
`90.0, 90.0, 90.0
`P212121
`42.9-2.80 (2.95-2.80)
`119,875 (17,343)
`22,827 (3,361)
`87.2 (89.6)
`5.3 (5.2)
`15.3 (35.7)
`6.8 (16.2)
`6.5 (2.5)
`4
`
`79.0-2.80
`21.4
`26.2
`-
`8,101
`0.013
`1.591
`66.4
`
`96.2
`3.6
`0.2
`0.38
`
`TCL-2
`
`
`90.0, 94.8, 94.9
`98.1, 112.0, 97.3
`P1
`38.5-2.10 (2.21-2.10)
`640,244 (92,983)
`161,229 (23,385)
`97.8 (97.0)
`4.0 (4.0)
`14.1 (84.3)
`8.2 (48.8)
`8.0 (2.0)
`8
`
`38.5-2.10
`14.9
`19.8
`-
`18,131
`0.013
`1.839
`31.3
`
`96.3
`3.6
`0.1
`0.12
`
`CPP
`
`
`90.2, 94.8, 94.8
`98.0, 112.4, 97.4
`P1
`49.1-2.20 (2.32-2.20)
`249,699 (35,984)
`135,192 (19,491)
`94.4 (93.2)
`1.8 (1.8)
`8.2 (34.8)
`8.2 (34.8)
`5.9 (1.9)
`8
`
`49.1-2.20
`15.8
`21.9
`-
`17,947
`0.012
`1.788
`47.7
`
`96.8
`3.2
`0.0
`0.14
`
`EPP
`
`
`90.2, 95.1, 95.3
`98.2, 112.3, 97.3
`P1
`49.3-1.90 (2.00-1.90)
`427,322 (63,042)
`206,602 (30,427)
`92.1 (92.6)
`2.1 (2.1)
`6.9 (40.4)
`6.4 (38.0)
`6.4 (2.0)
`8
`
`40.9-1.90
`15.0
`19.3
`-
`18,573
`0.015
`1.767
`30.1
`
`97.0
`3.0
`0.0
`0.09
`
`4ALL
`
`4ALI
`
`4ALJ
`
`4ALK
`
`1Values in parenthesis refer to the highest resolution shell
`
`2
`
`R
`merge
`
`
`
` 
`I
`
`hkl
`
`i
`
`i
`
`I
`
`
`
`hkl
`
`i
`
`3
`
`R
`
`pim
`
`
`
`
`
`hkl
`
`
`1
`
`
`
`N
`
`
`
`2/11
`
`
`
` 
`I
`
`i
`
`i
`
`I
`
`i
`
`
`
`I
`
`
`
`hkl
`
`i
`
`I
`
`i
`
` (Weiss, 2001)
`
`4
`
`R
`cryst
`
`
`
` 
`F
`obs
`
`hkl
`
`F
`calc
`
`F
`obs
`
`
`
`
`
`hkl
`
`5According to Molprobity (Davis et al., 2007)
`
`
`
`9
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00009
`
`

`

`Table S3, related to Table 3. Kinetic parameters for ftFabI utilizing NADH and various
`
`trans-2-enoyl-CoA substrates.
`
`
`
`trans-2-butenoyl-CoA
`
`Km (µM)
`
`353  5
`
`kcat (min-1)
`
`kcat/Km,acyl-CoA (min-1 µM-1)
`
`1140  60
`
`3.2  0.2
`
`trans-2-octenoyl-CoA
`
`33.3  0.4
`
`1200  60
`
`36.0  1.9
`
`trans-2-decenoyl-CoA
`
`11.2  0.8
`
`1260  60
`
`112.5  9.7
`
`trans-2-dodecenoyl-CoA
`
`1.7  0.1
`
`1200  60
`
`705.9  54.5
`
`trans-5-methyl-2-
`
`hexenoyl-CoA (iso-)
`
`()-trans-4-methyl-2-
`
`
`
`
`
`hexenoyl-CoA (anteiso-)
`
`
`
`
`
`0.287  0.0071
`
`0.0040  0.00021
`
`NADH
`
`
`
`18.8  0.6
`
`1140  60
`
`60.0  6.0
`
`1Estimated based on linear slope of the Michaelis-Menten plot at low [NADH]
`
`
`
`10
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00010
`
`

`

`Table S4, related to Experimental Procedures. Details of the saFabI purification protocols.*
`
`A
`pET-16b(+)
`BL21 (DE3)
`pLysS
`100 µg/ml
`ampicillin and 34
`µg/ml
`chloramphenicol
`28 °C, 5 h
`
`B
`pETM-11
`BL21 (DE3)
`
`50 µg/ml
`kanamycin
`
`C
`pETM-112
`BL21 (DE3)
`
`50 µg/ml
`kanamycin2
`
`D
`pETM-11
`BL21 (DE3)
`
`50 µg/ml
`kanamycin
`
`E
`pETM-11
`BL21 (DE3)
`
`50 µg/ml
`kanamycin
`
`18 °C, 18 h
`
`18 °C, 18 h2
`
`18 °C, 18 h
`
`18 °C, 18 h
`
`
`Plasmid
`E. coli strain
`
`antibiotics
`
`cultivation
`temperature,
`time
`lysis buffer
`
`wash buffer
`
`elution buffer
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl, 10 mM
`imidazole
`
`100 mM
`trisodium citrate
`pH 5.6, 500 mM
`NaCl, 100 mM
`EDTA
`
`50 mM Tris-HCl
`pH 8.0, 1 M NaCl
`
`50 mM Tris-HCl
`pH 8.0, 1 M
`NaCl2
`
`50 mM Tris-HCl
`pH 8.0, 1 M NaCl
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl, 250 mM
`imidazole
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl, 250 mM
`imidazole
`
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl, 250 mM
`imidazole
`
`50 mM Tris-HCl
`pH 8.0, 200 mM
`NaCl, 2 mM β-
`mercaptoethanol,
`5% glycerol
`50 mM Tris-HCl
`pH 8.0, 500 mM
`NaCl, 2 mM β-
`mercaptoethanol,
`5% glycerol
`50 mM Tris-HCl
`pH 8.0, 200 mM
`NaCl, 2 mM β-
`mercaptoethanol,
`5% glycerol, 250
`mM imidazole
`no
`25 mM Tris-HCl
`pH 8.0, 200 mM
`NaCl, 1 mM DTT
`
`15.2 mg/ml (0.49
`mM)
`
`TEV cleavage
`storage buffer
`
`no
`20 mM trisodium
`citrate pH 5.6,
`500 mM NaCl,
`100 mM EDTA
`
`yes1
`20 mM trisodium
`citrate pH 5.6,
`500 mM NaCl,
`100 mM EDTA
`
`no
`20 mM trisodium
`citrate pH 5.6,
`500 mM NaCl,
`100 mM EDTA3
`
`9.8 mg/ml (0.32
`mM)
`
`38.6 mg/ml (1.24
`mM)
`
`13.0 mg/ml (0.42
`mM) 2,3
`
`no
`20 mM trisodium
`citrate pH 5.6,
`280 mM NaCl, 1
`mM EDTA, 280
`mM K-glutamate
`15.0 mg/ml (0.48
`mM)
`
`final
`concentration
`
`1After Ni2+ affinity chromatography the protein sample was incubated with TEV-protease in
`
`50 mM Tris pH 8.0 and 500 mM NaCl at 4 °C for 10 h, yielding a 50% removal of the His6-
`
`tag. Subsequently, the sample was re-applied to Ni-TED followed by size exclusion
`
`chromatography as described above.
`
`
`
`2Alternatively (C'), the saFabI mutants were expressed and purified (final concentrations of
`
`1.0 mg/ml) using the pET-16b(+) plasmid, 100 µg/ml ampicillin, a cultivation temperature of
`
`25 °C and the wash buffer described in procedure A.
`
`
`
`11
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00011
`
`

`

`3Alternatively (C''), the storage buffer of the pooled SEC-fractions was exchanged to 20 mM
`
`trisodium citrate pH 5.6, 150 mM NaCl and 1 mM EDTA before concentrating to 8.5 mg/ml
`
`(0.27 mM).
`
`
`
`*In general, the procedure described in the Supplemental Experimental Procedures was
`
`applied to obtain ≥ 95% pure saFabI protein. The crystals leading to the TCL-1, apo-1 and
`
`apo-2 structures were obtained using purifications A, B or C'', respectively. For the EPP,
`
`TCL-2 and CPP complex structures a FabI batch was used which was obtained by procedure
`
`D. The analytical size exclusion chromatography experiments were performed with protein
`
`batches E and C. The double and triple mutants were purified applying procedure C'.
`
`12
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00012
`
`

`

`Movie S1, related to Figure 5. Ordering of the substrate binding loop region. This model
`
`is based on the TCL-2 structure (one monomer is shown in cyan, TCL in yellow and NADP+
`
`in green). Helix α7 was rearranged in PyMol according to the different geometries of the apo-
`
`2 structure. The motion of helix α7 was visualized by interpolating between these different
`
`arrangements. The loop region between α7 and α8 was omitted for clarity (the associated
`
`termini are shown in red). The initial states of the SBL (shown in green), cofactor and
`
`inhibitor were not obtained directly in our apo structures. Conformational changes in
`
`additional regions (SBL-2 and ASL) were omitted in the animation. This animation was
`
`prepared using PyMol, VirtualDub and Xvid (DeLano, 2002).
`
`
`
`Movie S2, related to Figure 6. Dimer - tetramer transition. This animation was prepared
`
`using the structures 3GNT, 3GNS, apo-1, apo-2 and TCL-2. One dimer is shown in blue, the
`
`other in red. All structures were superimposed using the lower left monomer. LSQMAN was
`
`used to interpolate between Cα positions of consecutive structure pairs for the amino acids
`
`which were ordered in both structures (Kleywegt and Jones, 1996). Cα traces of these
`
`consecutive states were visualized using PyMol, VirtualDub and Xvid (DeLano, 2002).
`
`13
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00013
`
`

`

`Supplemental Results
`
`Inhibition kinetics
`
`Based on the relative Ki*,app measurements, the gain in affinity from the additional buried
`
`surface area (BSA) for the inhibitors CPP and EPP (BSATCL = 72.9%, BSACPP = 74.7%,
`
`BSAEPP = 76.2%) does not completely compensate for the loss of the halogen bond observed
`
`for TCL (Table 1 and Figure 4). The relative affinities of the diphenyl ether analogues for
`
`saFabI increase in the order CPP < EPP < TCL. Ki*,app includes the affinity of the more potent
`
`enzyme-inhibitor complex (E-I*) formed following slow isomerization of the initial, rapid-
`
`reversible enzyme-inhibitor complex (E-I; affinity described by Ki
`
`app), according to a two-step
`
`slow-binding mechanism of inhibition. Since saFabI exhibited greatly improved activity and
`
`stability in the presence of glutamate (see Supplemental Discussion), we were able to perform
`
`a more reliable progress curve-based analysis of steady-state inhibition as compared to
`
`previous studies (Heath et al., 2000; Xu et al., 2008).
`
`Reduction of straight-chain substrates by saFabI
`
`Among the straight-chain fatty acyl substrates, we observed an increase in the specificity
`
`constant as the chain length increases from 4 to 8 carbons (Table 3), consistent with previous
`
`results noting a preference for longer chain substrates in the case of S. aureus FabI (Xu et al.,
`
`2008). However, in contrast to other FabI homologues, the specificity constant decreases
`
`beyond a chain length of 8 carbons. These variations are driven by both an increase in kcat and
`
`a decrease in Km, as opposed to ftFabI (Table S3) and ecFabI, where changes in straight-chain
`
`substrate specificity are driven primarily by changes in Km (Rafi et al., 2006; Ward et al.,
`
`1999).
`
`14
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00014
`
`

`

`Substrate specificity of saFabI
`
`The ratio of specificity constants amongst the first-round straight- and branched-chain
`
`substrates of saFabI was approximately 1:24:1 (straight : iso : anteiso) (Table 3). Since the
`
`ratio of velocities vX : vY for competing substrates SX and SY is equal to the ratio of [SX]
`
`kX
`
`cat/KX
`
`m to [SY] kY
`
`cat/KY
`
`m, the available substrate pool also determines the membrane fatty
`
`acid composition (Cornish-Bowden, 2004). FabH, which catalyzes the initial condensation
`
`reaction in the FAS II pathway, controls the pool of substrates that enter the FAS II cycle
`
`(Figure 1). Accordingly, S. aureus and B. subtilis FabH exhibit a preference for branched-
`
`chain acyl-CoA primers over the typical acetyl-CoA primer (Choi et al., 2000; Qiu et al.,
`
`2005). Since the S. aureus membrane is composed mostly of anteiso-C15 fatty acids (Parsons
`
`et al., 2011), our results are consistent with the current thought that FabH is the main
`
`determining factor in branched-chain fatty acid biosynthesis. For the SCFA family of
`
`organisms, FabH tends to prefer the acetyl-CoA primer, resulting in a smaller pool of
`
`available branched-chain substrates (Qiu et al., 2005). Although FabIs from these organisms
`
`are capable of utilizing branched chain (iso-) substrates (Choi et al., 2000; Smirnova and
`
`Reynolds, 2001), the relative rate of BCFA substrate reduction catalyzed by SCFA FabIs like
`
`ftFabI is much lower than for the BCFA family member saFabI (Table S3). Thus, as expected,
`
`the pool of substrates and specificity constants both heavily favor the synthesis and
`
`incorporation of straight-chain fatty acids by this organism.
`
`15
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00015
`
`

`

`Supplemental Discussion
`
`Glutamate sensitivity
`
`We investigated the glutamate sensitivity of saFabI as glutamate is known to be present at
`
`high intracellular concentrations (> 100 mM) in Gram-positive organisms, such as S. aureus,
`
`and has been found to increase the in vitro activity of S. aureus gyrase and topoisomerase IV
`
`(Björklind and Arvidson, 1978; Blanche et al., 1996). Glutamate markedly improved the
`
`catalytic activity of saFabI (~5-fold for 1 M potassium glutamate, ~4-fold for 500 mM
`
`potassium glutamate), such that the maximum kcat obtained was similar to other FabI
`
`homologues with CoA-based substrates (kcat > 500 min-1; Table 3) (Heath et al., 2000; Parikh
`
`et al., 1999; Rafi et al., 2006; Ward et al., 1999; Xu et al., 2008). In contrast, the catalytic
`
`activity of ftFabI is unaffected by glutamate. The reaction velocity of 25 nM ftFabI with 10
`
`µM oct-CoA and 250 µM NADH was 11.25 µM/min both in the presence and absence of 1 M
`
`glutamate. To screen for salt effects, we performed single-point reaction velocity
`
`measurements. Unlike the S. aureus gyrase and topoisomerase IV enzymes (Blanche et al.,
`
`1996), the activation of saFabI by glutamate is not cation-specific though K+ is slightly
`
`preferred over Na+. Among the anions tested, glutamate provided the greatest rate
`
`enhancement. At 50 nM saFabI with 30 µM oct-CoA and 200 µM NADPH, the reaction
`
`velocities relative to sodium glutamate were 0, 0.16, 0.22, 0.49 with sodium chloride, sodium
`
`aspartate, sodium acetate and sodium ADA, respectively. Thus, enhancement of activity is not
`
`a general salt effect but appears to be mediated by a carboxylate motif. Interestingly, there is
`
`evidence of this effect in the literature. To avoid protein precipitation, previous saFabI
`
`purification protocols required buffers containing 100 mM EDTA (Heath et al., 2000; Xu et
`
`al., 2008). Furthermore, some saFabI inhibition assays were performed in the presence of 100
`
`mM sodium ADA pH 6.5 (Payne et al., 2002).
`
`16
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00016
`
`

`

`Notably, there is a physiological link between glutamate and BCFA synthesis - the first step
`
`of BCFA biosynthesis is the transamination of branched-chain amino acids to α-ketoglutarate,
`
`forming the corresponding α-keto acid and glutamate (Figure 1) (Mendoza et al., 2002). We
`
`have shown that glutamate not only enhances enzyme activity but also affects the cofactor
`
`specificity as well as the cooperativity of substrate binding (Figure S5). In fact, the dimer-
`
`tetramer transition may provide a mechanism of regulation by glutamate or similar
`
`metabolites. Based on ambiguous densities in the saFabI P1 crystal structures (Figure S5E)
`
`obtained in the presence of the saFabI-stabilizing anions glutamate, citrate and EDTA which
`
`share the glutaric acid motif, a potential glutamate binding site is located between Arg103 of
`
`SBL-2 and Asn205 of SBL. Molecules occupying this site could interconnect the SBLs,
`
`thereby enhancing substrate binding and stabilizing the tetrameric state. However, future
`
`experiments should clarify the role of glutamate and similar metabolites for saFabI
`
`functioning in vivo.
`
`17
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00017
`
`

`

`Supplemental Experimental Procedures
`
`Cloning
`
`SaFabI from the S. aureus strain NCTC 8325 was purified using the safabINCTC8325:pET-
`
`16b(+) plasmid (Xu et al., 2008). We also amplified the same safabI gene from S. aureus
`
`strain N315 via PCR (originating from safabIN315:pET-16b(+)) and cloned it into pETM-11
`(EMBL) using the NcoI and HindIII (New England Biolabs) restriction sites (primers: 5‘-
`CGAAGGTCGTCCCATGGTAAATCTTGAAAAC-3‘
`5'-
`and
`CCAATAACGTGAACAAAGCTTCTGAATG-3' (Biomers)). The saFabI isoform from S. aureus
`
`NCTC 8325 was used only for kinetic experiments. It differs from the S. aureus N315 isoform
`
`by a single amino acid (G191S). Although residue 191 is positioned in close proximity to the
`
`nicotinamide ring of NADPH, the Km of saFabI-G191 with respect to NADPH was identical
`
`to that reported previously under similar assay conditions for saFabI-S191 (Xu et al., 2008).
`
`Expression and purification
`
`E. coli BL21 (DE3) bacteria (Novagen) containing the safabI:pETM-11 plasmid were grown
`
`in Luria-Bertani (LB) medium (Lennox) supplemented with 50 µg/ml kanamycin at 37 °C.
`
`After induction with 0.5 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG) at an OD600 nm of
`
`0.6 the temperature was reduced to 18 °C. Cells were harvested after 18 hours post-induction,
`
`resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 500 mM NaCl) and lysed through a cell
`
`disruptor at 1.5 kbar. The lysate was centrifuged at 50,000 g and the supernatant loaded onto a
`
`Protino Ni-TED column (Macherey-Nagel). After washing (50 mM Tris-HCl pH 8.0, 1 M
`
`NaCl), the protein was eluted (50 mM Tris-HCl pH 8.0, 500 mM NaCl, 250 mM imidazole).
`
`A buffer exchange with PD-10 desalting columns (Amersham Biosciences) was followed by
`
`size exclusion chromatography (SEC) on a HiLoad 26/60 Superdex 200 pg column (GE
`
`Healthcare) pre-equilibrated with storage buffer (20 mM trisodium citrate pH 5.6, 280 mM
`
`18
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00018
`
`

`

`NaCl, 1mM EDTA, 280 mM potassium glutamate). Protein containing fractions were pooled
`
`and concentrated to 15 mg/ml prior to shock-freezing in liquid nitrogen and storage at -80 °C.
`
`For further details see also Table S4 (this protocol refers to procedure D in Table S4).
`
`Crystallization, data collection and structure determination
`
`In a vapor diffusion experiment containing equal amounts of protein and precipitant solution
`
`(0.1 M K/Na-phosphate pH 6.5, 35% 2-methyl-2,4-pentanediol (MPD)), a single TCL-1
`
`crystal (nomenclature according to Table S2) of space group P212121 grew within 3 weeks.
`
`Diffraction data of the directly flash-frozen crystal were collected to 2.8 Å at the MX
`
`beamline 14.1 at BESSY II (Berlin) utilizing a MarMosaic 225 detector (λ = 0.976 Å, T = 100
`
`K). Data were processed with iMosflm (Leslie, 1992) and Scala (Evans, 2006). Phases were
`
`determined with Phaser (McCoy et al., 2007) by molecular replacement using the FabI
`
`structure of B. anthracis (PDB code: 2QIO:A) as search model (Tipparaju et al., 2008).
`
`Alternate cycles of maximum likelihood refinement and model building were performed in
`
`Refmac 5 (Murshudov et al., 1997) and Coot (Table S2) (Emsley and Cowtan, 2004).
`
`Refinement was improved with the implementation of TLS-refinement and NCS-restraints
`
`excluding Met99. After the first rounds of refinement triclosan, NADP+ and water molecules
`
`were included into the model according to the Fo-Fc map.
`
`Rod-shaped apo-1 crystals of space group P32 were obtained in a vapor diffusion setup using
`
`0.61 M NH4H2PO4, 9% t-butanol, 0.01 M EDTA pH 7.0 as precipitant. For data collection at
`
`ID 29 (λ = 0.919 Å, T = 100 K) of the European Synchrotron Radiation Facility (ESRF,
`
`Grenoble) the crystals were briefly transferred into cryoprotectant solution (0.61 M
`
`NH4H2PO4, 9% t-butanol, 0.01 M EDTA pH 7.0, 25% glycerol) and cryocooled in liquid
`
`nitrogen. After data processing with XDS (Kabsch, 1993), twinning was detected using
`
`Phenix.xtriage (Zwart et al., 2008). Phaser molecular replacement was performed with the
`
`TCL-1 structure omitting regions 95-108, 149-158 and 190-221 of the protein as well as all
`
`19
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00019
`
`

`

`small molecules. To avoid model bias due to high non-crystallographic symmetry (12
`
`molecules per asymmetric unit) and twinning, Rfree flags were assigned in 20 thin resolution
`
`shells. Model building in Coot was alternated with a combined refinement using Phenix.refine
`
`(Adams et al., 2010) and Refmac 5 (Table S1). As no additional electron density for the
`
`cofactor and inhibitor was observed, this structure is referred to as the apo-form.
`
`Cube-shaped apo-2 crystals of space group P43212 grew in 1.15 M (NH4)2SO4, 0.1 M sodium
`
`acetate pH 5.0, 1% ethanol as precipitant in a vapor diffusion experiment. Crystals were
`
`cryocooled (30% glycerol) and diffraction data were collected to 2.45 Å at beamline ID 29
`
`(ESRF) using an ADSC Quantum Q315r detector (λ = 0.919 Å, T = 100 K). Data were
`
`analyzed as described for the apo-1 structure and refined (including TLS refinement) with
`
`Phenix.refine (Table S1). As in the case of apo-1, no interpretable cofactor or inhibitor density
`
`was present.
`
`Ternary saFabI-NADP+-EPP crystals were obtained with 0.1 M K/Na-phosphate pH 6.5 and
`
`36% MPD as precipitant. Data were collected with an ADSC Quantum Q315r detector at
`
`beamline ID 14-4 (ESRF, λ = 0.980 Å, T = 100 K) and were processed to 1.9 Å using
`
`iMosflm and Scala. Molecular replacement was performed with Phaser using the TCL-1
`
`structure lacking all small molecules as search model. Refinement (including TLS refinement
`
`(Painter and Merritt, 2006)) and model building were performed using Refmac 5 and Coot
`
`(Table S2). NADP+ and EPP were included in the model (Schuttelkopf and van Aalten, 2004)
`
`after molecular replacement and initial refinement. Additionally, the Fo-Fc map indicated an
`
`alternative conformation of the amino acid stretch 99-105. The second conformation was
`
`modeled for all subunits with clear densities for this feature.
`
`A vapor diffusion experiment with TCL-treated saFabI protein and a similar precipitant
`
`solution (0.1 M K/Na-phosphate pH 6.5 and 38% MPD) resulted in TCL-2 crystals of space
`
`group P1. Diffraction data were collected to 2.1 Å at beamline 14.1 (BESSY II) using a
`
`MarMosaic 225 detector (λ = 0.918 Å, T = 100 K) and processed with iMosflm and Scala.
`
`20
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00020
`
`

`

`The EPP structure without alternative conformations served as a search model for molecular
`
`replacement in Phaser. Refmac 5 and Coot were used for refinement (including TLS
`
`refinement) and model building (Table S2).
`
`Similarly, CPP crystals of saFabI were grown with a precipitant solution containing 0.1 M
`
`K/Na-phosphate pH 6.5 and 35% MPD. Diffraction data of a directly cryocooled crystal were
`
`collected at beamline ID 29 (ESRF) with a PILATUS 6M detector (λ = 0.976 Å, T = 100 K).
`
`Indexing and integration were performed with the XDS package. Scala was used for scaling.
`
`Structure determination and refinement followed the procedure as described for TCL-2 (Table
`
`S2).
`
`As saFabI crystallized in space group P1 with similar cell parameters for the different
`
`inhibitors, comparability is warranted (Table S2). For further analysis we used the subunits H
`
`of the different structures which display no significant crystal contacts at the active site and
`
`therefore approximate the natural state in solution despite the increased flexibility that results
`
`in weaker electron densities. Distances and angles were measured for all subunits and are
`
`given as mean values ± standard deviations.
`
`To avoid model bias, omit maps were calculated directly after molecular replacement and an
`
`initial refinement prior to inclusion of any small molecule (Figure S2). All structural figures
`
`were prepared with PyMol (DeLano, 2002).
`
`Steady-state kinetic assays
`
`Kinetic parameters for the wild-type saFabI catalyzed reduction of trans-2-octenoyl-CoA
`
`(oct-CoA) using NADPH were determined by varying the concentration of oct-CoA at several
`
`fixed concentrations of NADPH. Previous mechanistic studies suggest that saFabI catalyzes
`
`its reaction via an ordered Bi Bi mechanism with the fatty acyl substrate binding first (Xu et
`
`al., 2008). However, our structural studies suggest that NADP+ likely binds first, followed by
`
`the inhibitor. We believe this extends to substrate binding with NADPH binding before the
`
`21
`
`Chugai Exhibit 2305
`Pfizer, Inc. v. Chugai Pharmaceutical Co. Ltd.
`IPR2017-01357
`Page 00021
`
`

`

`fatty acyl substrate, in accordance with other FabI homologues. Thus, the data were fitted to
`
`equations 2 and 3, assuming an ordered B