Architecture of a single membrane spanningcytochrome P450 suggests constraints that orient thecatalytic domain relative to a bilayerBrian C. Monka,b,1, Thomas M. Tomasiakc,1, Mikhail V. Keniyaa,b, Franziska U. Huschmanna,b,d, Joel D. A. Tyndalld,Joseph D. O’Connell IIIc, Richard D. Cannona,b, Jeffrey G. McDonalde, Andrew Rodriguezc, Janet S. Finer-Moorec,and Robert M. Stroudc,2

aSir John Walsh Research Institute, bDepartment of Oral Sciences, Faculty of Dentistry, and dSchool of Pharmacy, University of Otago, Dunedin 9054,New Zealand; cDepartment of Biochemistry and Biophysics, University of California, San Francisco, CA 94158; and eDepartment of Molecular Genetics,University of Texas Southwestern Medical Center, Dallas, TX 75390

Contributed by Robert M. Stroud, January 10, 2014 (sent for review September 30, 2013)

Bitopic integral membrane proteins with a single transmembranehelix play diverse roles in catalysis, cell signaling, and morpho-genesis. Complete monospanning protein structures are needed toshow how interaction between the transmembrane helix andcatalytic domain might influence association with the membraneand function. We report crystal structures of full-length Saccharo-myces cerevisiae lanosterol 14α-demethylase, a membrane mono-spanning cytochrome P450 of the CYP51 family that catalyzes thefirst postcyclization step in ergosterol biosynthesis and is inhibitedby triazole drugs. The structures reveal a well-ordered N-terminalamphipathic helix preceding a putative transmembrane helix thatwould constrain the catalytic domain orientation to lie partly inthe lipid bilayer. The structures locate the substrate lanosterol,identify putative substrate and product channels, and reveal con-strained interactions with triazole antifungal drugs that are impor-tant for drug design and understanding drug resistance.

Membrane proteins that span the lipid bilayer once consti-tute around 50% of all integral membrane proteins (1).

Although monospanning membrane proteins carry out numerouskey biological functions, including environmental sensing, organelle-specific catalysis, and the regulation of cell morphology, only in-dividual domains or subdomains are currently represented in theProtein Data Bank, and structural information about interactionsbetween their transmembrane domains and extramembranouscomponents is lacking. Cytochrome P450 proteins are prominentenzymes with orthologs found in all kingdoms of life. In eukar-yotes, microsomal members of this major family of mixed-func-tion mono-oxygenases contain a single transmembrane helix andcan be grouped in two broad functional categories: biodefense,such as the first phase of xenobiotic detoxification, and coremetabolism including reactions in sterol biosynthesis and fattyacid oxidation (2).The lanosterol 14α-demethylases or CYP51 enzymes, probably

the most genetically ancient of the cytochrome P450 families,play a central role in cholesterol or ergosterol biosynthesis (3).CYP51s carry out three consecutive mono-oxygenase reactioncycles to remove the 14α-methyl group from lanosterol to yield 4,4-dimethyl-cholesta-8,14,24-trienol, a key precursor in cholesteroland ergosterol biosynthesis, releasing water and formic acid (3).Because of the key roles that CYP51s play in yeast, filamentousfungi, and some parasitic protozoa, these enzymes are thera-peutic targets for antimicrobial agents, including fluconazole(FLC), voriconazole (VCZ), and itraconazole (ITC) (4). Fungalinfections play an increasingly significant role in disease, impactingagriculture ecosystems and human health, especially in immuno-compromised individuals (5–7) for whom antifungal resistancecontinually poses a threat (8). In humans CYP51 is being tested asa target for cholesterol-lowering drugs (9) and in antiangiogeniccancer therapies (10). A limited set of cytochrome P450 isoforms

(1A2, 2C8, 2C9, 2C19, 2D6, and 3A4) metabolize about 75% ofthe drugs currently used in medicine (11).Membrane-bound cytochrome P450s catalyze reactions in-

volving many hydrophobic substrates, including lipophilic drugsand sterols that partition from the lipid bilayer into the activesite. Identification of enzyme orientation relative to the lipidbilayer therefore is critical for understanding substrate entry intoand egress from the heme-containing active site (12). For theeukaryotic cytochrome P450 enzymes that contain transmembranesegments, most X-ray structures have been obtained after re-moval of the N-terminal transmembrane domain to facilitateexpression and crystallization (13–18), as is the case with manyother monospanning membrane protein structures. ResolvedCYP51 structures include the soluble enzyme fromMycobacteriumtuberculosis (19) and N-terminal truncated enzymes from Homosapiens (20), Trypanosoma cruzi (21), Trypanosoma brucei (22),and Leishmania infantum (23). The available eukaryotic cyto-chrome P450 structures show the catalytic subunit has a conservedfold surrounding a central prosthetic heme group. A proposedsubstrate channel to the membrane as well as interactions with

Significance

The absence in the Protein Data Bank of full-length structuresof bitopic membrane proteins with one transmembrane helix,probably because of difficulties with ordered crystallization,has limited understanding of how single-transmembrane heli-ces orient enzymes and sensors at the bilayer surface. X-raycrystal structures of full-length yeast lanosterol 14α-demethy-lase, a cytochrome P450, show how a helix spanning a singletransmembrane may lead to constraints on the orientation ofthe putative substrate entry portal from within the bilayer. Thecrystal structures also locate the substrate lanosterol, identifyputative substrate and product channels, and reveal constrainedinteractions with triazole antifungal drugs that are important fordrug design and understanding the drug resistance associatedwith orthologs of the enzyme found in fungal pathogens.

Author contributions: B.C.M., T.M.T., and R.M.S. designed research; B.C.M., T.M.T., M.V.K.,F.U.H., J.D.A.T., J.D.O., R.D.C., J.G.M., A.R., and J.S.F.-M. performed research; B.C.M. andT.M.T. contributed new reagents/analytic tools; B.C.M., T.M.T., M.V.K., F.U.H., J.D.A.T., J.D.O.,R.D.C., J.G.M., J.S.F.-M., and R.M.S. analyzed data; and B.C.M., T.M.T., and R.M.S. wrotethe paper.

The authors declare no conflict of interest.

Data deposition: The structures reported in this paper have been deposited in the ProteinData Bank, www.pdb.org [PDB ID codes 4LXJ (ScErg11p-6xHis + lanosterol) and 4K0F(ScErg11p-6xHis + itraconazole)].

See Commentary on page 3659.1B.C.M. and T.M.T. contributed equally to this work.2To whom correspondence should be addressed. E-mail: stroud@msg.ucsf.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1324245111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1324245111 PNAS | March 11, 2014 | vol. 111 | no. 10 | 3865–3870

PHARM

ACO

LOGY

SEECO

MMEN

TARY

several inhibitors, including ketoconazole, posaconazole, FLC,and substrate analogs such as 14α-methylenecyclopropyl-Δ7–24,25-dihydrolanosterol, have been visualized (19, 20, 22, 24).Although bacterial and mitochondrial orthologs of cytochrome

P450s and the M. tuberculosis CYP51 are soluble enzymes, bio-informatic analysis of many microsomal cytochrome P450s led tothe detection of a hydrophobic surface in the catalytic domain andprediction of an N-terminal monospanning transmembrane do-main that has been shown to be important for subcellular targetingand membrane insertion (25–28). Multiple in vitro approacheshave suggested that a transmembrane anchor is not always re-quired for the catalytic activity of some microsomal preparationsof cytochrome P450s (29, 30). For example,M. tuberculosis CYP51specifically converts detergent-solubilized lanosterol to its deme-thylated product (31). However, the in vivo role of the membraneanchor in catalysis by other CYP51s that use membrane-associatedlanosterol has yet to be established. Antibody-binding accessibilityassays showed that most of the surface of CYP2B4 (32) was sol-vent exposed with the exception of the putative transmembranehelix and the loop between the F and G helices. About 30% of thecatalytic domain of CYP11A1 was inaccessible to chemical mod-ification by fluorescent probes (33). The physical displacement ofphospholipids in Langmuir–Blodgett monolayers indicated a con-siderably greater area of protein insertion than could be accountedfor by a single N-terminal transmembrane helix (34). This findingwas supported by atomic force microscopy, which showed thecatalytic domain protrudes only 3.5 ± 1 nm above the membrane(35). A crystal structure of full-length aromatase at 2.9-Å resolu-tion has been modeled, but weak electron density meant that thestructures of 44 N-terminal and seven C-terminal residues couldnot be determined (36). Models that place the mouth of the sub-strate channel in direct contact with the hydrophobic environmentof the lipid bilayer required the catalytic domain helices A′ and A to

be imbedded and partially imbedded, respectively, in the lipid bi-layer (28, 37).

ResultsArchitecture of the Cytochrome P450 Saccharomyces cerevisiaeLanosterol 14α-Demethylase. We report X-ray crystal structuresof full-length Saccharomyces cerevisiae lanosterol 14α-demethy-lase (ScErg11p-6xHis) with the membrane domain structurallydiscernable and two ligands cocrystallized: the substrate lano-sterol and the triazole inhibitor ITC. Clear electron density forthe 50 N-terminal amino acids revealed two helices orientedat ∼60° to each other (Fig. 1 A and B). The N-terminal helix[membrane helix 1 (MH1), residues 6–23] is amphipathic andmakes extensive crystal contacts with symmetry-related mole-cules (Fig. 2 A and B). The distribution of hydrophobic andhydrophilic side chains suggests that the hydrophobic face maycontact the inner leaflet of the endoplasmic reticulum (ER)membrane (Fig. 1B). MH1 connects to TMH1 via a short proline-containing turn (residues 24–26), a slightly kinked helix 37.5 Ålong (residues 27–51) that is of sufficient length to traverse thelipid bilayer. The side chains of amino acids R30 and Q29, ad-jacent to the turn, form a C-terminal cap for helix MH1 and anN-terminal cap for TMH1, respectively.Seventeen of the 25 residues in TMH1 are hydrophobic, with

a high concentration in the N-terminal portion, as is consistent withimmersion of TMH1 in the lipid bilayer. Beyond the proline kink atP38, TMH1 makes extensive polar contacts with the catalytic do-main (Fig. 1C). The helix then consists of opposing hydrophilic andhydrophobic faces that interact with the catalytic domain and lipidbilayer, respectively. An ionic interaction between the guanidiniumgroup of R52 and the carboxylate of D54 enables a sharp turnbetween TMH1 and the first loop of the catalytic domain.Paired polar residues constrain the orientation between

TMH1 and the catalytic domain. The e-amide group of Q46,which hydrogen bonds with the main-chain carbonyl and amideof Y61, is absolutely conserved in the fungal CYP51 family. Thehydroxyl of Y61 also makes a hydrogen bond with the δ-amidegroup of N42. Water molecules bridge the e-amide group of Q46with the main-chain amide oxygen of L58 and the ζ-hydroxyl ofY49 with the hydroxyl of S92. Other polar contacts in this region

Fig. 1. Overall fold of ScErg11p and predicted orientation in the lipid bilayer.(A) Helices are colored from the N terminus to the C terminus with a gradientfrom blue to red. The heme is shown as yellow sticks; lanosterol in the active-site cavity is shown in orange. (B) The structure of the amphipathic helix (MH1)and the transmembrane helix (TMH1). Hydrophobic side chains are coloredblue, and polar residues colored magenta, with oxygen atoms colored red, andnitrogen atoms colored blue. Difference electron density (Fo–Fc) contoured to2σ is depicted as a green mesh. (C) Interaction of the transmembrane domainwith the catalytic domain. Ordered water molecules are depicted as redspheres. Hydrogen bonding interactions are shown as dashed lines; distancesare in Å. (D) Distribution of hydrophobic and charged residues in the F/G loop.

Fig. 2. Crystal packing of ScErg11p-6xHis. (A) ScErg11p-6xHis crystals showalternating layers of transmembrane (TM) segments and the soluble seg-ments (S). (B) Crystal packing of two symmetry-related copies of ScErg11p-6xHis with the amphipathic helix S6–H21. The amphipathic helix is coloredgreen, and the symmetrically related copies are colored salmon or blue. (C)Overlay of lanosterol-bound (gray) and ITC-bound (salmon) ScErg11p-6xHis.

3866 | www.pnas.org/cgi/doi/10.1073/pnas.1324245111 Monk et al.

include the guanidinium group of R55 and the carbonyl of A399;the main-chain amide of K53 and the hydroxyl of S397; and themain-chain carbonyl of D54 and the amide of A399. These exten-sive contacts bury a surface area of 51.8 Å2 on TMH1 and appear toform a rigid architecture that is important for establishing the poseof the enzyme in the bilayer (Figs. 1A and 2 A and C). The energeticstabilization conferred by the extensive contacts between TMH1and the catalytic domain make it unlikely that the crystal contactsbetween the amphipathic MH1 (residues S6, E10, E13, and H21)and residues K270 and S267 of one symmetry-related molecule andG532 and E354 of the other (Fig. 2B) would have distorted theoverall structure of ScErg11p significantly. The ScErg11p-6xHisstructures show few significant differences between structures, withthe most extensive changes (0.36 Å rmsd on Cα) being in the F/Gloop (Fig. 2C) of the lanosterol and ITC costructures.

Inferred Orientation of ScErg11p Relative to the Membrane. To es-tablish the membrane orientation, ScErg11p-6xHis was insertedinto preformed lipid vesicles and labeled from the outside at freecarboxylates with glycinamide in the presence of N′N′-Dicyclo-hexylcarbdiimide (DCCD) at pH 6 (Fig. 3A) (38). Tandem massspectrometry of tryptic digests identified labeled residues only onthe catalytic domain, with no labeling on the N-terminal MH1(residues 1–26), as is consistent with topological labeling of othercytochrome P450s and with MH1 and the catalytic domain as-sociating with opposite sides of the bilayer.The orientation of ScErg11p also was explored using two

algorithms for membrane insertion: the all-atomistic positioningof proteins in the membrane (PPM) procedure (39) and theknowledge-based Ex3d procedure (40). Both calculated poseswere very similar to those inferred from our crystallographicanalysis, i.e., TMH1 residues 27–50 link the N-terminal amphi-pathic helix MH1 and the C-terminal catalytic domain that are

partially submerged in opposite leaflets of the lipid bilayer (Fig.1A). The OPM analysis, which calculated a partitioning freeenergy of −31.3 kcal/mol into the lipid bilayer, was the best fitwith our topological labeling, including the labeling of residueE249 on the F/G loop, and was used in this analysis (Fig. 3).The pose we infer for ScErg11p relative to the membrane

buries portions of the loop connecting helices F–F′ and G in thelipid bilayer (Figs. 1D, 2C, and 3). It is consistent with muta-genesis experiments showing this region is important for mem-brane binding (33, 41, 42), fluorescence quenching analysis (5),the predicted pose of aromatase (36), and computational simu-lations for human cytochrome P450s performed in the presenceof diolyeolphosphatidylcholine (DOPC) (37) or 1-palmitoyl-2-oleoyl-sn-glycerol-2-phosphocholine (POPC) (27, 32–34, 37, 43).Several aspects of the structure support the proposed orientation.Hydrophobic residues are distributed in a single band across thesurface of the structure. Poisson–Boltzmann electrostatics calcu-lations using Delphi software (44) reveal an electrostatic potentialsurface showing a distinct hydrophobic/hydrophilic boundary (Fig.3B). Ordered water molecules are present only in proposed sol-vent-accessible areas of the model and are absent from the bilayer-accessible surface of the structure (Fig. 1 B and C). As a corollaryto the positioning of the lipid bilayer, the protein crystallized withalternating layers of proposed membrane and soluble segments(Fig. 2A) corresponding to the placement of hydrophobic andhydrophilic elements in the crystal structure, including the surface-associated ligands.We hypothesize that a cluster of charged residues found at the

N-terminal end of the G helix forms a major membrane-bindingcontact with the negatively charged phosphate head groups ofthe lipid bilayer (Fig. 1D). This orientation places the end ofa substrate channel in direct contact with the cytoplasmic surfaceof the lipid bilayer (Figs. 1A, 2C, 3B, and 4C). This substratechannel is bifurcated near its mouth, yielding a secondary ves-tibule (Fig. 3C) not previously identified in CYP51s. Although anN-decyl-β-D-maltoside (DM) detergent micelle differs from anintact lipid bilayer, the proposed orientation supports a model inwhich the substrate first is abstracted directly from the surface ofthe membrane via the juxtaposed channel mouth. After hemereduction and complex formation with oxygen, an oxygen re-bound mechanism converts the C14 methyl to the first of threeproducts that result in 4,4-dimethyl-cholesta-8,14,24-trienol (2).

Substrate and Product Binding. The enzyme was purified andcocrystallized with the substrate lanosterol or with four otherligands (itraconazole, estriol, FLC, or VCZ). The lanosterol (Fig.4A) and estriol (Fig. S1A) cocrystal structures showed densityconsistent with the common four-ring sterol in the active-sitecavity. A mixed state probably involving both endogenous lano-sterol and added ligand complicated interpretation of electrondensity in the active site. This mixed state did not occur whenScErg11p was purified in the presence of ITC and crystallized.As a result, only the lanosterol and ITC structures are detailed herewith the four others reported as models in SI Discussion.The lanosterol (Fig. 4A) and estriol (Fig. S1A) cocrystal

structures showed density consistent with the common four-ringsterol in the active-site cavity. Electron density consistent withthe tail of lanosterol is located in the primary vestibule that leadsup to the active site (Fig. 5B). Similar electron density was ob-served in the same locations when no ligand was supplied (Fig.S1D), indicating that significant amounts of substrate copurifywith the enzyme. HPLC-MS in positive ion mode (45) showedthat lanosterol copurifies with ScErg11p-6xHis in the absence ofadded ligand, supporting this interpretation (Fig. S1E).Electron density proximal to the heme iron, consistent with

the presence of an oxygen molecule coordinated to the hemeiron, fits with the ferrous dioxo intermediate predicted for theCYP51 reaction cycle. This feature also was seen in the crystal

Fig. 3. Sequence and predicted membrane interactions of ScErg11p. (A)Regions highlighted in yellow are predicted by OPM to lie inside the lipidbilayer. Red squares indicate acidic residues that were accessible to thewater-soluble glycinamide probe when ScErg11p-6xHis was incorporatedinto artificial liposomes. (B) Electrostatic surface map of ScErg11p with cut-away showing the primary and secondary vestibules. Surface electrostaticpotential is displayed with the molecular graphics software Chimera (59)with blue as positive and red as negative.

Monk et al. PNAS | March 11, 2014 | vol. 111 | no. 10 | 3867

PHARM

ACO

LOGY

SEECO

MMEN

TARY

structure without added ligand, which also had bound lanosterol(Fig. S1D). To interpret this peak, the electron density near theheme was assessed using crystallographic difference-map re-finement. Fo–Fc maps calculated after several rounds of re-finement showed that a single water (10 electrons) would generateinsufficient positive density, whereas imidazole (32 electrons)would give a strong negative density peak. Placing O2 (16 elec-trons) gave a flat Fo–Fc electron density map consistent with O2being at the heme site. Although unusual for P450s, this result isnot without precedent (46) and may be caused by reduction afterX-ray radiolysis (47) or purification of ScErg11p-6xHis trappedin the oxygen-bound state. The distance of ∼8 Å between oxygenand C14 methyl is consistent with a lanosterol binding in a pre-catalytic rather than a catalytically active state. Visual inspectionof the electron density maps and a high real-space R-factor in-dicated the need for considerable care in placement of the ligandin the lanosterol costructure. Minimally biased Fo–Fc electrondensity omit maps, calculated with lanosterol omitted to placethe lanosterol correctly, resulted in a ligand placement withreasonable correlation coefficient (0.813). Bound lanosterol isoriented differently from other lanosterol-like molecules in re-lated CYP51s (24), and the electron density offers no evidence ofthese alternate orientations.Density consistent with a second lanosterol-like molecule

projects from the secondary vestibule (Fig. 5 A and B). In allScErg11p structures examined, a hydrophobic tail fills the nar-row channel from the secondary vestibule to the surface ofErg11p, whereas a multiring head that projects into solutionappears to lack the C14 methyl group. Lack of density for theC14 methyl group suggested that the lanosterol-like moleculewas the reaction product 4,4-dimethyl-cholesta-8,14,24-trienol ora subsequent metabolite that likely copurified with the enzyme.Mass spectrometry of the purified ScErg11-6xHis detected lan-osterol, larger quantities of zymosterol, and traces of 4,4-dimethyl-cholesta-8,14,24-trienol (Fig. S1E). If a product egress chan-nel exists, the narrow channel off the secondary vestibule mayprovide a selectivity filter for the product; that is, removal of theC14 methyl group of lanosterol may be required for the multiringhead group to traverse the channel (Fig. 5B). Erg11p has beenshown to interact physically with Erg25p-Erg27p and possiblywith Erg24p (48). Presentation of 4,4-dimethyl-cholesta-8,14,24-

trienol on Erg11p then would allow concerted reduction byErg24p (C14 sterol reductase) and demethylation at carbon 4by Erg25p (C4 sterol methyl oxidase), Erg26p (C3-sterol de-hydrogenase), and Erg27p (3-keto sterol reductase) in complexwith the scaffold protein Erg28p to give zymosterol in the er-gosterol biosynthesis pathway (48). Alternatively, the retentionof 4,4-dimethyl-cholesta-8,14,24-trienol and zymosterol by Erg11pmay play roles that include conferral of structural stability ora feedback control mechanism. Two regions of weak excesselectron density located near the heme and adjacent to K220bordering helix G are likely hydrophobic molecules, possiblydetergent or phospholipid copurified with the enzyme, (Fig.S2 A and B).

DiscussionAzole Binding and Antifungal Resistance. Our structures identifyimportant ligand-binding constraints. The antifungal triazoledrug ITC extends from the active site to just beyond the mouthof the entry channel, similar to posaconazole in the T. bruceiCYP51 structure (49). The triazole group of the ITC makes acoordination bond with the heme iron. The space occupied byITC (Fig. 4B) fits closely with that occupied by lanosterol andbound O2 (Fig. 4A), with the triazole head group displacing theO2 and the di-halogenated headgroup replacing the first sterolring. The expected chlorine atoms on the dichlorophenyl groupconsistently showed strongly negative Fo–Fc density, suggestingthat the chloro groups had been damaged through photoreduc-tion upon X-ray data collection. As a result, the chloro groupshave been refined to zero occupancy in the model. In confir-mation of our findings, the triazole head groups in the smallertriazole drugs FLC and VCZ coordinate with the heme iron anddisplace the presumed O2, and their unmodified difluorinatedhead groups fill the space occupied by the first sterol ring oflanosterol and the sterol analog estriol (Fig. 4 A and B and Fig.S1 A–C). The long tail of ITC fills the entry channel and appearsto exclude all but one water molecule. The tail of ITC lies withina hydrophobic pocket lined by helices F and F′ (F236, P238, andF241), consistent with previous structures of truncated CYP51family members (50), and by the residues L380HSL383 andF506TSMV510 (Fig. 4B and see Fig. S4A).

Fig. 4. Lanosterol and ITC binding in ScErg11p. (A) Lanosterol is depicted with carbon atoms colored cyan and heme with carbon atoms colored yellow. Selectedoxygen atoms are colored red, nitrogen blue, sulfur gold, and iron brown. Electron density is depicted as a simulated annealing 2Fo–Fc omit map contoured to0.7σwith the ligand, heme, and oxygen omitted from the electron density calculation. A bond links the heme–oxo complex. (B) ITC in a coordination complex withactive-site heme. The color scheme is as in A, with ITC carbons colored magenta. Electron density is depicted as an Fo–Fc map contoured to 2.5σ with the liganddensity omitted from the calculation. (C) Conserved amino acids in fungi that are commonly mutated in antifungal resistance are depicted with carbon atomscolored orange. Amino acids conserved in S. cerevisiae and in themain fungal pathogens of humans but not in human CYP51, and not mutated, are depicted withcarbon atoms colored green. Amino acids that are similar only in the pathogenic fungi are shown in cyan. Other atoms are colored as in A.

3868 | www.pnas.org/cgi/doi/10.1073/pnas.1324245111 Monk et al.

Although less than 24% of the amino acids in ScErg11p areconserved among fungi (50), there is an especially high con-centration of conserved residues near the heme (Fig. S4A) andactive-site cavity. These conserved residues include the abso-lutely conserved coordination of the C470 sulfur to the hemeiron and the ionized K151, R385, and H468 side chains. Thesequence G310VLMG314 in helix I is a consensus sequence(GXXXG) for sterol (cholesterol) binding, with the carbonyl ofG310 making the only polar interaction with lanosterol and aslightly kinked face toward ligand.The structure of the active site and substrate channel may

explain the susceptibility of some azole-resistant clinical isolates,including Candida albicans mutants F126L, Y132F/H, P230L,G307S, and F380S, Aspergillus fumigatus G54E/F and I301,Cryptococcus neoformans Y145F, and Ajellomyces capsulatusY136F (51, 52) to ITC or FLC (Fig. S3). For example, Y140 inS. cerevisiae (Fig. S3C) is equivalent to Y132 in C. albicans, Y145in C. neoformans, and Y136 in A. capsulatus. The absence of theheme-binding hydrogen bond in S. cerevisiae Y140F/H mutationscould modify the orientation of the heme and reduce suscepti-bility to both FLC and ITC. In contrast, G73E/F mutations inScErg11p, equivalent to G54E/F in A. fumigatus Cyp51A, are inthe mouth of the substrate channel and might be expected tomodify susceptibility to the long-tailed triazoles ITC and pos-aconazole but not short-tailed VCZ.Numerous azole-resistance mutations occur outside the ac-

tive site. Significantly, the N22D azole-resistance mutation inTMH1 of A. fumigatus Cyp51A (ScErg11p N42D) supports theproposition that interactions between TMH1 and the catalyticdomain can affect active-site function. CYP51s from S. cerevisiaeand the fungal pathogens of humans (Fig. S4) contain a group ofconserved/similar aromatic residues (F60, Y61, W62, Y72, Y77,F79, F80, Y87, F90, H381) bordered by TMH1 that form a com-pact, buried structure. In concert with neighboring conserved/similar hydrophobic residues (V59, L95, L96 L383, L407) and thepolypeptide backbone (A69, G73, S382), this grouping supportsa narrow entry channel into the active site and egress vestibules.The tight vestibules orient the lanosterol and may provide gatedentry of demethylated reaction product into the egress channel.Finally, mutation conferring azole resistance at residues homol-ogous to ScErg11p L95, F241, H381, and S382 has yet to be re-ported. These residues along with the chemically similar residueI139 could provide fungal-specific side-chain contacts for the design

of novel antifungals directed against the major fungal pathogensof humans (Fig. 4C and Fig. S3B).

Bitopic Membrane Proteins. Full-length structures of bitopicmembrane proteins with one transmembrane helix are absent inthe Protein Data Bank, probably because of the difficulties withordered crystallization. This deficiency has limited understandingof how single-transmembrane helices orient enzymes/sensors atthe bilayer surface (1). The X-ray structure of ScErg11p-6xHis,a full-length fungal cytochrome P450 enzyme, shows that thetransmembrane domain not only tethers the protein to the ERbut also may orient it relative to the bilayer. The surprisingly rigidinteraction between the transmembrane domain and catalyticdomain may establish the pose required for catalysis and in-teraction with other protein partners such as Erg24p and Erg25p–Erg27p. With monospanning membrane proteins constitutingabout half of all membrane proteins, structural information sup-porting properties beyond tethering and localization in ScErg11pprovides clues about constraining polar residues in other mono-spanning enzymes. It also may define a motif that may be a featureof other membrane-bound enzymes.The ScErg11p structures identify lanosterol binding in the

active site of a CYP51, a substrate channel linking to the lipidbilayer, and a secondary vestibule with a product exit channel.The model also more fully maps how triazole antifungals interactwith a fungal cytochrome P450 to block catalysis and identifiespossible interactions that confer azole resistance applicable topathogenic fungal species. Like other CYP51 structures, the in-tact enzyme displays less conformational heterogeneity betweenstructures than seen in many truncated ectodomain eukaryoticcytochrome P450s, suggesting that the transmembrane domainitself or associated hydrophobic molecules such as lipids or re-action product may decrease conformational heterogeneity andprotect the active site from bulk water. The structures shownhere may enable the development of new molecular models tofacilitate drug design that targets fungal CYP51s or othereukaryotic cytochrome P450s and may provide a practical basisfor the design of therapeutics with minimized off-target effects.

MethodsS. cerevisiae ScErg11p was overexpressed homologously as a 6×His derivative(ScErg11p-6xHis) from the PDR5 locus in the host yeast strain ADΔ (Table S1)(55). ScErg11p-GFP expressed in ADΔ was found in nuclear-associated ER(Fig. S5A). ScErg11p-6xHis was purified from strain MMLY941 by extractionof a crude membrane fraction with the detergent DM, Ni-NTA-agarose af-finity chromatography, and size-exclusion chromatography (SEC) in thepresence or absence of ITC (Fig. S5B). The purified protein behaved asa detergent micelle-protein monomer during SEC, as is consistent witha molecular mass for the protein of ∼62 kDa (Fig. S2C). It bound ergosteroland showed type II binding of known antifungals (Fig. S5D). The purifiedprotein was authenticated by mass spectrometry of tryptic fingerprints atthe Stanford University Mass Spectrometry facility (Fig. S5E).

Red crystals were obtained initially by hanging-drop vapor diffusion at 18 °Cin 200-nL drops with a MemGold (Molecular Dimensions Limited) screen setusing a Mosquito robot (TTP Labtech). Crystallization conditions were opti-mized, and large, red, boat-shaped crystals (100 μm) were obtained in 2- to 6-μLdrops with the purified protein at ∼20 mg/mL in 10% (wt/vol) glycerol, 150 mMNaCl, 4× critical micellar concentration DM, and 20 mM Hepes (pH 7.5) mixed1:1 with 43–45% (wt/vol) PEG 400 in 100 mM glycine (pH 9.3–9.5) (Fig. S5F).

Complete datasets were obtained for crystals at the Berkeley AdvancedLight Source BL 8.3.1 with a beam at wavelength 1.1587 Å (Table S2). Datawere processed, integrated, and scaled using the HKL2000 (53) package. Aninitial model was determined by molecular replacement using the solubledomain of the HsCYP51 model 3LD6 (20) in Phaser (54). Refinement wasperformed using both Refmac5 in the CCP4 (55) suite of programs andPhenix (56), and model building was performed using Coot (57). Modelvalidation was performed in Phenix and with Molprobity (55) and PROCHECK(58). Display of electrostatic maps was performed in Chimera (59).

ACKNOWLEDGMENTS. We thank James Holton and George Meigs at beam-line 8.3.1 of the Advanced Light Source at Lawrence Berkeley National

Fig. 5. Multiple ligands bind to ScErg11p. (A) Electron density of lanosterol-like molecule. Fo–Fc density contoured to 2.0σ is shown in green witha theoretical model of 4,4-dimethyl-cholesta-8,14,24-trienol placed. The 4,4-dimethyl-cholesta-8,14,24-trienol is depicted with carbon atoms coloredcyan. (B) Hypothetical model of ligand binding with surface representationof the primary and secondary vestibules. Lanosterol bound in the active siteis depicted with carbon atoms colored cyan. The second lanosterol-like mol-ecule is depicted with carbon atoms colored purple. Oxygen atoms are col-ored red, nitrogen blue, and iron brown. The outside surface of the vestibulesand active site are depicted as a light mesh surface colored dark gray, andtheir visible inside surface is colored light gray.

Monk et al. PNAS | March 11, 2014 | vol. 111 | no. 10 | 3869

PHARM

ACO

LOGY

SEECO

MMEN

TARY

Laboratory; Bill Harries, Rebecca Robbins, and Erwin Lamping for sup-port; Paul Ortiz de Montellano and John Cutfield for critical reading; andAvner Schlesinger for helpful computational suggestions. This research was

supported by National Institutes of Health Grants U54GM094625, GM24485,and GM073210, the Marsden Fund of the Royal Society of New Zealand, andthe Otago Medical Research Foundation.

1. Hubert P, et al. (2010) Single-spanning transmembrane domains in cell growth andcell-cell interactions: More than meets the eye? Cell Adhes Migr 4(2):313–324.

2. Guengerich FP (2001) Common and uncommon cytochrome P450 reactions related tometabolism and chemical toxicity. Chem Res Toxicol 14(6):611–650.

3. Lepesheva GI, Waterman MR (2011) Structural basis for conservation in the CYP51family. Biochim Biophys Acta 1814(1):88–93.

4. Lepesheva GI, et al. (2008) CYP51: A major drug target in the cytochrome P450 su-perfamily. Lipids 43(12):1117–1125.

5. Lepesheva GI, et al. (2007) Conformational dynamics in the F/G segment of CYP51from Mycobacterium tuberculosis monitored by FRET. Arch Biochem Biophys 464(2):221–227.

6. Brown GD, Denning DW, Levitz SM (2012) Tackling human fungal infections. Science336(6082):647.

7. Iliev ID, et al. (2012) Interactions between commensal fungi and the C-type lectinreceptor Dectin-1 influence colitis. Science 336(6086):1314–1317.

8. Monk BC, Goffeau A (2008) Outwitting multidrug resistance to antifungals. Science321(5887):367–369.

9. Harwood HJ, Jr., et al. (2005) Dual-action hypoglycemic and hypocholesterolemicagents that inhibit glycogen phosphorylase and lanosterol demethylase. J Lipid Res46(3):547–563.

10. Nacev BA, Liu JO (2011) Synergistic inhibition of endothelial cell proliferation, tubeformation, and sprouting by cyclosporin A and itraconazole. PLoS ONE 6(9):e24793.

11. Stoll F, Göller AH, Hillisch A (2011) Utility of protein structures in overcoming ADMET-related issues of drug-like compounds. Drug Discov Today 16(11-12):530–538.

12. White SH, Ladokhin AS, Jayasinghe S, Hristova K (2001) How membranes shapeprotein structure. J Biol Chem 276(35):32395–32398.

13. Williams PA, et al. (2004) Crystal structures of human cytochrome P450 3A4 bound tometyrapone and progesterone. Science 305(5684):683–686.

14. Scott EE, et al. (2003) An open conformation of mammalian cytochrome P450 2B4 at1.6-A resolution. Proc Natl Acad Sci USA 100(23):13196–13201.

15. Yano JK, Hsu MH, Griffin KJ, Stout CD, Johnson EF (2005) Structures of human mi-crosomal cytochrome P450 2A6 complexed with coumarin and methoxsalen. NatStruct Mol Biol 12(9):822–823.

16. Williams PA, et al. (2003) Crystal structure of human cytochrome P450 2C9 with boundwarfarin. Nature 424(6947):464–468.

17. Schoch GA, et al. (2004) Structure of human microsomal cytochrome P450 2C8. Evi-dence for a peripheral fatty acid binding site. J Biol Chem 279(10):9497–9503.

18. Poulos TL, Finzel BC, Howard AJ (1987) High-resolution crystal structure of cyto-chrome P450cam. J Mol Biol 195(3):687–700.

19. Podust LM, Poulos TL, Waterman MR (2001) Crystal structure of cytochrome P45014alpha -sterol demethylase (CYP51) from Mycobacterium tuberculosis in complexwith azole inhibitors. Proc Natl Acad Sci USA 98(6):3068–3073.

20. Strushkevich N, Usanov SA, Park H-W (2010) Structural basis of human CYP51 in-hibition by antifungal azoles. J Mol Biol 397(4):1067–1078.

21. Lepesheva GI, et al. (2010) Structural insights into inhibition of sterol 14alpha-demethylase in the human pathogen Trypanosoma cruzi. J Biol Chem 285(33):25582–25590.

22. Lepesheva GI, et al. (2010) Crystal structures of Trypanosoma brucei sterol 14alpha-demethylase and implications for selective treatment of human infections. J BiolChem 285(3):1773–1780.

23. Hargrove TY, et al. (2012) CYP51 structures and structure-based development ofnovel, pathogen-specific inhibitory scaffolds. Int J Parasitol Drugs Drug Resist 2:178–186.

24. Hargrove TY, et al. (2012) Structural complex of sterol 14α-demethylase (CYP51) with14α-methylenecyclopropyl-Delta7-24, 25-dihydrolanosterol. J Lipid Res 53(2):311–320.

25. Williams PA, Cosme J, Sridhar V, Johnson EF, McRee DE (2000) Mammalian micro-somal cytochrome P450 monooxygenase: Structural adaptations for membranebinding and functional diversity. Mol Cell 5(1):121–131.

26. Wade RC, et al. (2005) Multiple molecular recognition mechanisms. CytochromeP450—a case study. Biochim Biophys Acta 1754(1-2):239–244.

27. Zhao YH, et al. (2006) Structure of microsomal cytochrome P450 2B4 complexed withthe antifungal drug bifonazole: Insight into P450 conformational plasticity andmembrane interaction. J Biol Chem 281(9):5973–5981.

28. Black SD (1992) Membrane topology of the mammalian P450 cytochromes. FASEBJ 6(2):680–685.

29. Sandhu P, et al. (1994) Expression of modified human cytochrome P450 1A2 inEscherichia coli: Stabilization, purification, spectral characterization, and catalyticactivities of the enzyme. Arch Biochem Biophys 309(1):168–177.

30. Scott EE, Spatzenegger M, Halpert JR (2001) A truncation of 2B subfamily cyto-chromes P450 yields increased expression levels, increased solubility, and decreasedaggregation while retaining function. Arch Biochem Biophys 395(1):57–68.

31. Pietila MP, et al. (2006) Cloning and characterization of CYP51 from Mycobacteriumavium. Am J Respir Cell Mol Biol 35(2):236–242.

32. Black SD, Martin ST, Smith CA (1994) Membrane topology of liver microsomal cyto-chrome P450 2B4 determined via monoclonal antibodies directed to the halt-transfersignal. Biochemistry 33(22):6945–6951.

33. Headlam MJ, Wilce MCJ, Tuckey RC (2003) The F-G loop region of cytochromeP450scc (CYP11A1) interacts with the phospholipid membrane. Biochim BiophysActa 1617(1-2):96–108.

34. Shank-Retzlaff ML, Raner GM, Coon MJ, Sligar SG (1998) Membrane topology ofcytochrome P450 2B4 in Langmuir-Blodgett monolayers. Arch Biochem Biophys359(1):82–88.

35. Bayburt TH, Sligar SG (2002) Single-molecule height measurements on microsomalcytochrome P450 in nanometer-scale phospholipid bilayer disks. Proc Natl Acad SciUSA 99(10):6725–6730.

36. Ghosh D, Griswold J, Erman M, Pangborn W (2009) Structural basis for androgenspecificity and oestrogen synthesis in human aromatase. Nature 457(7226):219–223.

37. Berka K, Hendrychová T, Anzenbacher P, Otyepka M (2011) Membrane position ofibuprofen agrees with suggested access path entrance to cytochrome P450 2C9 activesite. J Phys Chem A 115(41):11248–11255.

38. Izumi S, Kaneko H, Yamazaki T, Hirata T, Kominami S (2003) Membrane topology ofguinea pig cytochrome P450 17 alpha revealed by a combination of chemical mod-ifications and mass spectrometry. Biochemistry 42(49):14663–14669.

39. Lomize AL, Pogozheva ID, Mosberg HI (2011) Anisotropic solvent model of the lipidbilayer. 2. Energetics of insertion of small molecules, peptides, and proteins inmembranes. J Chem Inf Model 51(4):930–946.

40. Schramm CA, et al. (2012) Knowledge-based potential for positioning membrane-associated structures and assessing residue-specific energetic contributions. Structure20(5):924–935.

41. Murtazina D, et al. (2002) Membrane-protein interactions contribute to efficient 27-hydroxylation of cholesterol by mitochondrial cytochrome P450 27A1. J Biol Chem277(40):37582–37589.

42. Nakayama K, Puchkaev A, Pikuleva IA (2001) Membrane binding and substrate accessmerge in cytochrome P450 7A1, a key enzyme in degradation of cholesterol. J BiolChem 276(33):31459–31465.

43. Cojocaru V, Balali-Mood K, Sansom MSP, Wade RC (2011) Structure and dynamics ofthe membrane-bound cytochrome P450 2C9. PLOS Comput Biol 7(8):e1002152.

44. Li L, et al. (2012) DelPhi: A comprehensive suite for DelPhi software and associatedresources. BMC Biophysics 5(9).

45. McDonald JG, Smith DD, Stiles AR, Russell DW (2012) A comprehensive method forextraction and quantitative analysis of sterols and secosteroids from human plasma.J Lipid Res 53(7):1399–1409.

46. Nagano S, Poulos TL (2005) Crystallographic study on the dioxygen complex of wild-type and mutant cytochrome P450cam. Implications for the dioxygen activationmechanism. J Biol Chem 280(36):31659–31663.

47. Schlichting I, et al. (2000) The catalytic pathway of cytochrome p450cam at atomicresolution. Science 287(5458):1615–1622.

48. Mo C, Bard M (2005) A systematic study of yeast sterol biosynthetic protein-proteininteractions using the split-ubiquitin system. Biochim Biophys Acta 1737(2-3):152–160.

49. Chen CK, et al. (2010) Structural characterization of CYP51 from Trypanosoma cruziand Trypanosoma brucei bound to the antifungal drugs posaconazole and flucona-zole. PLoS Negl Trop Dis 4(4):e651.

50. Lepesheva GI, Virus C, WatermanMR (2003) Conservation in the CYP51 family. Role ofthe B’ helix/BC loop and helices F and G in enzymatic function. Biochemistry 42(30):9091–9101.

51. Morschhäuser J (2002) The genetic basis of fluconazole resistance development inCandida albicans. Biochim Biophys Acta 1587(2-3):240–248.

52. Becher R, Wirsel SG (2012) Fungal cytochrome P450 sterol 14α-demethylase (CYP51)and azole resistance in plant and human pathogens. Appl Microbiol Biotechnol 95(4):825–840.

53. Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected in os-cillation mode. Macromolecular Crystallography. Pt A 276:307–326.

54. McCoy AJ, et al. (2007) Phaser crystallographic software. J Appl Cryst 40(Pt 4):658–674.55. Chen VB, et al. (2010) MolProbity: All-atom structure validation for macromolecular

crystallography. Acta Crystallogr D Biol Crystallogr 66(Pt 1):12–21.56. Adams PD, Mustyakimov M, Afonine PV, Langan P (2009) Generalized X-ray and

neutron crystallographic analysis: More accurate and complete structures for bi-ological macromolecules. Acta Crystallogr D Biol Crystallogr 65(Pt 6):567–573.

57. Emsley P, Lohkamp B, Scott WG, Cowtan K (2010) Features and development of Coot.Acta Crystallogr D Biol Crystallogr 66(Pt 4):486–501.

58. Laskowski RA, Macarthur MW, Moss DS, Thornton JM (1993) PROCHECK - A Pro-gram to check the stereochemical quality of protein structures. J Appl Cryst 26:283–291.

59. Pettersen EF, et al. (2004) UCSF Chimera—a visualization system for exploratory re-search and analysis. J Comput Chem 25(13):1605–1612.

3870 | www.pnas.org/cgi/doi/10.1073/pnas.1324245111 Monk et al.

Supporting InformationMonk et al. 10.1073/pnas.1324245111SI MethodsExpression of Saccharomyces cerevisiae Lanosterol 14α-Demethylase-6xHis in Saccharomyces cerevisiae. Saccharomyces cerevisiae lano-sterol 14α-demethylase (ScErg11p) was overexpressed as a C-terminal 6xHis derivative from the PDR5 locus in a host yeaststrain (ADΔ) (1, 2) that is hypersensitive to xenobiotics becauseof the deletion of seven ATP-binding cassette (ABC) trans-porters, including pumps responsible for the efflux of azoledrugs (Table S1). The strain used to express ScErg11p-6xHis(MMLY941) is multiply deleted of the Pleiotropic Drug Re-sistance (PDR) class of drug efflux pumps, the YOR1 drug effluxpump, and the PDR3 transcriptional regulator but retains a sin-gle copy of its endogenous ERG11 gene. It also contains thegain-of-function pdr1–3 mutation in the Pdr1p transcriptionalregulator that renders transcription from the PDR5 promoterconstitutive (3). Placement of the PGK1 transcription terminatorand the URA3 gene immediately downstream of the ORF in-serted at the PDR5 locus (4) enables termination of transcriptsfrom the PDR5 locus and provides a selection marker, re-spectively. Control EGFP derivatives of the S. cerevisiae plasmamembrane ABC transporter ScPdr5p (ScPdr5p-GFP) and thecytoplasmic Candida albicans orotidine-5′-phosphate decarboxylase(CaUra3p-GFP) expressed in ADΔ were visualized by confocalmicroscopy almost exclusively in the plasma membrane and thecytoplasm, respectively (Fig. S5A). The ScErg11p-GFP de-rivative appeared enriched in the endoplasmic reticulum (ER),particularly in a region surrounding or capping the DAPI-stained nucleus (Fig. S5A). Overexpressed proteins that nor-mally localize to the ER often are concentrated in layers ofthe ER around much of the nucleus (5). The overexpressedScErg11p-6xHis protein from strain MMLY941 was recoveredin a crude membrane fraction, readily solubilized using the de-tergent N-decyl-β-D-maltoside (DM), and purified by Ni-NTAaffinity chromatography and size-exclusion chromatography (SEC)(Fig. S5 B and C).

Preparation of Purified ScErg11p-6xHis. Twelve liters of MMLY941cells were grown in 1-L volumes of YPD medium [1% yeastextract, 2% (wt/vol) peptone, and 2% (wt/vol) D-glucose] inbaffled 3-L flasks at 30 °C with shaking at 200 rpm to ensurevigorous aeration. The cells were harvested by centrifugationwhen OD600nm = 18–20, were washed once with distilled water,and were resuspended to a final volume of about 250 mL in ice-cold 2% (wt/vol) glycerol, 0.5 mM EDTA, 0.5 mM PMSF, and50 mM Tris adjusted to pH 7.5 at room temperature. The PMSFwas prepared as a 100-mM stock in isopropanol and stored atroom temperature. The ice-cold cell suspension was adjusted topH 7.2 with 1 M Tris and either stored at −80 °C until requiredor immediately homogenized in a Beadbeater (Biospec) using six1-min bursts, with 4-min pauses on ice to maintain temperaturebelow 10 °C. The pH of the homogenate was readjusted to 7.2 onice, and the cell debris was removed by centrifugation twice at5,000 × g for 10 min. The supernatant was centrifuged at 181,000 × gfor 70 min in a Beckman Ti45 rotor. The pelleted membranes werewashed once by centrifugation in 20% (wt/vol) glycerol, 10 mMTris, 0.5 mM EDTA, 0.5 mM PMSF adjusted to pH 8 at 4 °C andwere resuspended in this medium. The protein concentration ofthe crude membrane preparation was determined using the Lowrymethod with BSA as standard (5). The crude membranes werestored at ∼50 mg/mL at −80 °C.Samples of crude plasma membranes (1 g) were washed with

60 mL ice-cold solubilizing buffer [10% (wt/vol) glycerol, 250 mM

NaCl, 20mMTris, 0.5 mMPMSF containing Rocheminus EDTAprotease inhibitors and adjusted to pH 8 at 4 °C] by centrifugationat 92,400 × g in a Beckman Ti70 rotor for 70 min. The mem-branes (2.5 mg/mL protein) were extracted using the same buffercontaining 16 mM (∼10 times critical micelle concentration,cmc) DM for 1 h with stirring on ice. The DM-soluble fractionwas obtained by centrifugation at 92,400 × g for 70 min at 6 °C.Two milliliters (1/50th volume) of 1-M imidazole (20 mM final),pH 8, was added to the supernatant, which then was mixed witha 2-mL packed volume of preequilibrated Ni-NTA-agarose resinand incubated overnight with mixing at 6 °C. The affinity matrixwas preequilibrated by washing with 20 volumes of solubilizingbuffer containing 16 mM DM and 20 mM imidazole. Afterovernight incubation with the detergent extract, the Ni-NTA-agarose affinity matrix was placed in a 20-mL Bio-Rad Econo-Column, was washed four times with 10 volumes (20 mL) ofsolubilizing buffer containing 16 mM DM and 20 mM imidazole,and then was eluted using eight 1-mL samples of the same buffercontaining 200 mM imidazole. The affinity-purified fraction wasenriched for a 62-kDa band detected by SDS/PAGE with the sizepredicted for ScErg11p-6xHis (Fig. S5B). The preparation gavean absorbance peak at 424 nm consistent with the type II bindingof imidazole (6). In some cases 250 nmol of itraconazole (ITC)(125 μL, 2 mM in DMSO) was added to 5 mL of the pooledeluted protein. After 5 min on ice this preparation gave an ab-sorbance peak at 422 nm, which was indicative of the displace-ment of the imidazole in the active site by the ITC. The bulk ofnon–heme-containing ScErg11p-6xHis was precipitated by standingovernight at 6 °C. The soluble supernatant recovered after cen-trifugation at 16,000 × g for 10 min was concentrated by centrif-ugal filtration to about 1.0 mL using a 30-kDa cutoff membrane.Centrifugation at 16,000 × g for 10 min or filtration through a0.22-μm filter was used to remove residual insoluble material.Samples (0.5 mL) of the concentrated affinity-purified proteinwere separated by SEC using a Superdex 200 10/300 column (GEHealthcare) equilibrated and eluted with filtered and degassed10% (wt/vol) glycerol, 150 mMNaCl, 6.4 mMDM (approximately4× cmc), 0.5 mM PMSF, 1 Roche minus EDTA protease inhibitortablet per 400 mL, 20 mM Hepes (pH 7.5) ± 2 mM ITC. The 62-kDa ScErg11p-6xHis protein coeluted with the heme as a singlepeak at about 14 mL (Fig. S5C) and then was concentrated bycentrifugal filtration to >20 mg/mL in ∼200 μL for spectralanalysis (Fig. S5D) and crystallization.

Properties of ScErg11p-6xHis. SEC in the presence of DM ± ITC(2 μM) gave a heme-containing protein monomer/detergent micellecomplex (Fig. S5C) with a relative migration consistent witha molecular mass of 96 kDa by comparison with molecular massstandards (BenchMark; Invitrogen). Excess DM was detected byrefractive index measurements as a micelle with an apparentmolecular size of 35 kDa during SEC. This result (96 − 35 kDa =61 kDa) is consistent with the size expected for ScErg11p-6xHis(62 kDa). Before SEC or the addition of ITC, the affinity-puri-fied protein had an absorbance peak at 424 nm. Preparations ofthe SEC-purified apo-enzyme obtained in the absence of ITChad an absorbance peak at 418 nm, whereas preparations ob-tained by SEC in the presence of ITC had an absorbance peak at421 nm (Fig. 1D), as expected for type II binding of the azoledrug (7). After centrifugal concentration to >20 mg/mL, pre-cipitation of a colorless 62-kDa protein, which appeared to beScErg11p-6xHis depleted of the heme, ceased once the 418 nm/280 nm or 421 nm/280 nm absorbance ratio reached ∼0.65. The

Monk et al. www.pnas.org/cgi/content/short/1324245111 1 of 8

purified soluble protein then could be stored for at least 1 mo at4 °C. Mass spectrometry of tryptic fragments of the purifiedprotein obtained from an SDS/PAGE gel slice gave ∼70% cov-erage of the protein primary sequence (Fig. S5E).

Preparation of Proteoliposomes and Analysis of Amino Acid Accessibility.Azolectin (25 mg L-α-Phosphatidylcholine; Sigma) dissolved in 2 mLof chloroform was dried onto the surface of a Corex tube usinga flow of nitrogen gas. The lipid film was emulsified in 5 mL offiltered and degassed 20-mM Hepes (pH 7.5) buffer by vortexingand was transferred to a 15-mL Falcon tube. The emulsion (2.5mL) was sonicated on ice until optically clear. Samples of 0.5 mLwere extruded 19 times using a miniextruder (Avanti PolarLipids, Inc.) equipped with a polycarbonate membrane con-taining 100-nm pores. Samples (1.2 mL) of the resultant lipidvesicles were recovered by centrifugation at 100,000 × g for 45min using a Beckman mini-ultracentrifuge, were washed twice with1 mL of the Hepes buffer, and were resuspended in the samebuffer. A 20-μL sample of 23 mg/mL ScErg11p-6×His containing 6mM DM was added to 0.9 mL of the lipid vesicle preparation andincubated overnight at 6 °C to enable spontaneous proteoliposomeformation. The preparation was centrifuged at 5,000 × g for 5 minto remove a small amount of precipitated ScErg11p-6×His. Theproteoliposomes then were pelleted at 100,000 × g for 45 min andwashed twice by centrifugation in 1 mL of the Hepes buffer. Theslightly colored pellet was resuspended in 1 mL 20-mM Hepes, pH7.5, and stored on ice. Accessible acidic residues were glycinated atpH 5.0, and accessible lysine residues were acetylated at pH 9.0according to the method of Izumi et al. (8) after washing in theappropriate buffer. The modified ScErg11p-6×His residues in thetreated, washed, and acetone-precipitated proteoliposomes wereidentified by mass spectrometry of tryptic fingerprints of SDS/PAGE-separated protein bands at Stanford University Mass Spec-trometry facility.

SI DiscussionStructural Analysis. Use of the Constraint-based Multiple Align-ment Tool (Cobalt; National Center for Biotechnology In-formation), together with someminor adjustment of the alignmentsby eye, allowed structural alignment of the primary sequences ofScErg11p with the Erg11ps from 10 major fungal pathogens ofhumans (C. albicans, Candida glabrata, Candida dubliniensis,Candida tropicalis, Cryptococcus neoformans, Aspergillus fumigatus,Coccidioides immitis, Pneumocystis carinii, Ajellomyces capsulatus,and Malassezia globosa), including the Cyp51A and Cyp51B se-quences from Aspergillus fumigatus. This alignment showed thatonly 24% of the amino acid residues in the structural alignmentwith ScErg11p were absolutely conserved, and a further 25% ofthe amino acid residues were similar i.e., fell into the groupsSTA, NEQK, NHQK, NDEQ, QHRK, MILV, HY, FWY, CSA,ATV, SAG, STNK, STPA, SGND, SNDEQK, NDEQHK,NEQHRK, MILVF, HFY. Three separate views of the ScErg11p-6xHis structure (Fig. S4) show the distribution of conserved andsimilar amino acid residues for S. cerevisiae and the major fungalpathogens as red and white sticks, respectively. The most highlyconserved regions, which contain the bulk of the conservedresidues, surround and stabilize the heme, appear within theactive-site cavity in support of enzyme function, contribute to thewalls of the channel to the active site, occur in a strip runningfrom the proximal surface of the heme beside helix C and towardthe product exit site, and are found in a buried region of aro-matic groups that links between the C-terminal half of thetransmembrane segment TMH1 and the active-site channel. Thesimilar amino residues are found in regions enriched for abso-lutely conserved residues (and therefore are likely to be impor-tant for function) and also tend to be more widely dispersedthroughout the structure and appear likely to contribute tothe conservation of overall enzyme architecture. In contrast, the

highest concentrations of dissimilar residues are found on thefront and side surfaces of the catalytic domain and in membrane-embedded portions of the molecule that interact with the lipidbilayer. We propose that the strip of conserved residues besidehelix C, which reaches toward the product egress channel andwhich includes a number of charged residues, is the site thatenables productive interaction with NADPH-cytochrome P450reductase. Previous studies have suggested a role for helix C ininteractions with the reductase (9). We also propose that theregion of conserved and similar aromatic groups (F39, F60, Y61,W62, Y72, Y77, F79, F80, Y87, F91, H381) surrounded by twoconcave β-sheet surfaces, the A and A′ helices, and one of thebifurcating faces of the substrate entry and product egresschannel provides a compact, buried structure when it is boundedby TMH1. With the assistance of neighboring conserved andsimilar hydrophobic residues (V59, L95, L96, L383, L407) andthe polypeptide backbone (A69, G73, S382), this structure con-tributes to a narrow entry channel into the active site and egresspathway. We hypothesize that this structure could help ensure anefficient oxidation reaction by orienting lanosterol in the activesite, allowing entry of only the demethylated reaction productinto the egress channel and blocking external water from trav-eling through the occupied entrance and exit channels to theactive-site heme.

Analysis of the Catalytic Site and Entry and Egress Channels forAntifungal Discovery. A detailed knowledge of the active site,the substrate entry channel, and the product egress channel forScErg11p, together with reactants and antifungal inhibitors,should expedite structure-directed drug design. The unambiguousplacement of both the head groups and the detection of theconformation of the long tail of ITC in the crystal structureidentifies putative anchor points in ScErg11p for a pharmaco-phore that could enable the design of antifungals with broad-spectrum activity against the dominant fungal pathogens andminimal activity against the human homolog HsCYP51. Aminoacid residues within 6 Å of ITC are illustrated in Fig. S5A. Theseresidues include the previously identified fungal substrate rec-ognition sequences SRS-1, SRS-2, SRS-4, SRS-5, and SRS-6 (7).The helix between G67 and P76 and the loop between F92 andR98 also contribute amino acid residues within 6 Å of the ITC.The amino acid residues visualized in Fig. S5A can be catego-rized as conserved only in the S. cerevisiae and the major fungalpathogens of humans (red sticks), conserved in the fungalErg11ps and HsCYP51 (pink sticks), similar only in fungi (whitesticks), similar in fungal Erg11ps and HsCYP51 (gray sticks),dissimilar in fungi (blue sticks), or dissimilar in fungi but withpolypeptide backbone carbonyl groups within 4 Å of ITC (stickswith atoms colored by element), i.e., within possible hydrogen-bonding distance of the antifungal. The ScErg11p X-ray struc-ture shows that the amino acid residues L95, F241, H381, andS382 conserved in fungi and not in humans have side-chainatoms within 4.1 Å of ITC. In addition the backbone carbonylbut not the side chain of the similarly conserved M313 is within4.5 Å of ITC. Other side chains within 6 Å of ITC are likely toconfer lower specificity because they are conserved in the fungalErg11ps and HsCYP51. These include Y126, L129, T130, F133,Y140, and F236, whereas G73 and G315 offer polypeptidebackbone carbonyl contacts. I136 and F506 are similar in fungalErg11ps but not in HsCYP51. I136 provides a side-chain contact,and F506 provides a backbone carbonyl contact. A69, Y72, L307,V311, T318, L380, and V510 are similar amino acid residues infungal Erg11ps and HsCYP51. These amino acid residues pro-vide possible side-chain contacts within 6 Å of the ITC, whereasthe similar residues A69, G310, and G314 provide carbonylcontacts within 4 Å of the drug. Of these residues, A69 providesthe only direct hydrogen bond between the polypeptide and thetail of ITC. The backbone carbonyl of H381 and the side-chain

Monk et al. www.pnas.org/cgi/content/short/1324245111 2 of 8

hydroxyl of S382 form a hydrogen bond network with watermolecules that are hydrogen bonded to ITC. Finally, the dis-similar amino acid residues T507, S508, and M509 have poly-peptide backbone carbonyls within 4 Å of the ITC. We concludethat the four fungal-specific amino acid side-chain contacts (L95,F241, H281, and S282) around ITC are augmented by a largenumber of lower-specificity contacts that could enable the de-velopment of either azole or nonazole antifungals with broadspecificity for fungal pathogens and minimal opportunity foreffective interaction with HsCYP51. We also note that azoleresistance has yet to be conferred at any of the four amino acidsthat contribute a fungal-specific side-chain contact despite asignificant number of single-amino acid changes that conferazole resistance in C. albicans, A. fumigatus, C. neoformans, orA. capsulatus mapping to residues within the active site andsubstrate channel (Fig. S5C). These include ScErg11p A69(C. albicans A61V), V70 (A. fumigatus Cyp51A S52T), G73(A. fumigatus Cyp51A G54E/K/V/R/W), A127 (C. albicansK143E/R), F134 (C. albicans F126L), V138 (C. albicans V130I),Y140 (C. albicans Y132F/H, C. neoformans Y145F, A. capsulatusY136F), T237 (C. albicans T229A), P239 (C. albicans P230L,A. fumigatus P216L), V242 (A. fumigatus M220K/I/T/V), G314(C. albicans G307S), T318 (A. fumigatus S297T), and P504(C. albicans P503L). Consideration of such mutations will beimportant in pharmacophore definition and antifungal design.The detection of the putative reaction product in all ScErg11p-

6xHis crystal structures adds to pharmacophore definition.Lanosterol is within 6 Å of the bulk of the amino acid residuesthat are within a similar distance of ITC but has similar proximitywith less of SRS-4 (helix I) and G7–G67 (Fig. S3 A and B). Theputative reaction product detected in the egress channel, however,

has amino acid residues within 6 Å that extend SRS-1 (backbonecarbonyls of A124 and A125), SRS-2 (V242 side chain similar infungi and human CYP51s), SRS-5 (backbone carbonyl of R385 andthe dissimilar K386 sidechain), F92-R98 to V102 (similar in fungiand HsCYP51 R98 side chain and the dissimilar M100 side chain)and adds the G403-L407 β-sheet (H405 is aromatic in fungi butdissimilar in HsCYP51). The backbone carboxyl and the aromaticface of the conserved in fungi F241 are adjacent to the tail of theputative reaction product, and the edge of its aromatic side chainabuts the tail of lanosterol. These results indicate additional fea-tures that could contribute to the design of a ScErg11p-basedpharmacophore.

Conformational Homogeneity.Cytochrome P450s involved in sterolbiosynthesis show remarkable structural conservation and rigidity,especially as compared with microsomal xenobiotic modifyingcytochrome P450s (10–12). The ScErg11p-6xHis structure isconsistent with a more restrained model of substrate bindingobtained using fast X-ray crystallography (13), NMR, and fluo-rescent spectroscopy (5). For example, we find single con-formations for the substrates lanosterol and the triazole drugsITC (Fig. 2C) and FLC (Fig. S1B), unlike the multiple con-formations found with truncated CYP51s from eukaryotic par-asites (14). Furthermore, molecular dynamics simulations ona CYP3A4 without a transmembrane helix gave an open-statestructure that may not be physiologically relevant when bound tothe lipid bilayer (11). This property may have important con-sequences for the design of inhibitors that specifically targetCYP51 family members or other membrane-bound cytochromeP450s and has been considered in the design of novel inhibitoryscaffolds specific for Trypanosoma and Leishmania CYP51s (14).

1. Lamping E, et al. (2007) Characterization of three classes of membrane proteinsinvolved in fungal azole resistance by functional hyperexpression in Saccharomycescerevisiae. Eukaryot Cell 6(7):1150–1165.

2. Lamping E, et al. (2005) Characterization of the Saccharomyces cerevisiae sec6-4mutation and tools to create S. cerevisiae strains containing the sec6-4 allele. Gene361:57–66.

3. Balzi E, Wang M, Leterme S, Van Dyck L, Goffeau A (1994) PDR5, a novel yeastmultidrug resistance conferring transporter controlled by the transcription regulatorPDR1. J Biol Chem 269(3):2206–2214.

4. Shakoury-Elizeh M, et al. (2010) Metabolic response to iron deficiency in Saccharomycescerevisiae. J Biol Chem 285(19):14823–14833.

5. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein measurement with theFolin phenol reagent. J Biol Chem 193(1):265–275.

6. Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected inoscillation mode. Macromolecular Crystallography. Pt A 276:307–326.

7. Warrilow AGS, et al. (2010) Expression, purification, and characterization ofAspergillus fumigatus sterol 14-alpha demethylase (CYP51) isoenzymes A and B.Antimicrob Agents Chemother 54(10):4225–4234.

8. Izumi S, KanekoH, Yamazaki T, Hirasa T, Kominami S (2003)Membrane topology of guineapig cytochrome P450 17 alpha revealed by a combination of chemical modifications andmass spectrometry. Biochemistry 42(49):14663–14669.

9. Mo C, Bard M (2005) A systematic study of yeast sterol biosynthetic protein-proteininteractions using the soft-ubiquitin system. Biochim Biophys Acta 1737(2–3):152–160.

10. Strushkevich N, Usanov SA, Park H-W (2010) Structural basis of human CYP51inhibition by antifungal azoles. J Mol Biol 397(4):1067–1078.

11. Denisov IG, Shih AY, Sligar SG (2012) Structural differences between soluble andmembrane bound cytochrome P450s. J Inorg Biochem 108:150–158.

12. Lepesheva GI, et al. (2007) Conformational dynamics in the F/G segment of CYP51from Mycobacterium tuberculosis monitored by FRET. Arch Biochem Biophys 464(2):221–227.

13. Schlichting I, et al. (2000) The catalytic pathway of cytochrome p450cam at atomicresolution. Science 287(5458):1615–1622.

14. Hargrove TY, et al. (2012) CYP51 structures and structure-based development ofnovel, pathogen-specific inhibitory scaffolds. Int J Parasitol Drugs Drug Resist 2:178–186.

Monk et al. www.pnas.org/cgi/content/short/1324245111 3 of 8

Fig. S1. The ScErg11p active-site models in the presence or absence of estriol, FLC, and VCZ. (A) Plus estriol. (B) Plus FLC. (C) Plus VCZ. (D) Ligand free. Theestriol, triazoles, and heme are depicted with oxygen atoms colored red, nitrogen atoms blue, and iron brown. The carbon atoms of the heme are coloredyellow, the carbon atoms of the estriol are colored salmon, the carbon atoms of fluconazole are colored orange, and the carbon atoms of voriconazole arecolored white. Electron densities for the estriol and the triazoles were calculated as Fo–Fc maps contoured to 2.5σ for estriol and FLC and to 3σ for the emptystructure and VCZ with the ligand density omitted from the calculation. A covalent bond links the heme–oxo complex. (E) Mass spectrometry profile showingthe lanosterol and zymosterol identified in ScErg1p-6xHis preparations. The standards including fluid follicular meiosis activating sterol (FMAS) and testicularmeiosis activating sterol (TMAS) are shown.

Fig. S2. Nonmodeled regions of electron density associated with ScErg11p. (A) Electron density extending from near M313 in the catalytic site toward theproposed membrane boundary. (B) Electron density detected in the neighborhood between K220 and helix G.

Monk et al. www.pnas.org/cgi/content/short/1324245111 4 of 8

Fig. S3. Features in the active site and entry and egress channels of ScErg11p. (A) Amino acid residues within 6 Å of ITC in ScErg11p. Amino acid residueswithin 6 Å of ITC are shown as colored sticks on a cartoon polypeptide backbone that extends (in cyan) the primary sequence two amino acids N-terminal andC-terminal of these residues. Structural features previously identified as possibly contributing to the active site are marked as SRS-1, SRS-2, SRS-4, SRS5, and SRS-6(7). The G67-P76 α-helix and the F92-R96 loop also contribute residues within 6 Å of the ITC. The heme is shown in brown, and the molecular dimensions of ITCare shown with yellow lines and black dots. Amino acid residues conserved in ScErg11p and Erg11ps of the major pathogenic fungi of humans and not inHsCYP51 (L95, F241, M313, H381, S382) are shown in red. Residues that are conserved in fungal Erg11ps and HsCYP51 (G73, Y126, L129, T130, F133, Y140, F236,G315) are shown in pink. Amino acid residues that are similar in fungi but not in humans are shown as white sticks (I139, F506); residues that are similar inhumans and fungi are in gray. Amino acid residues that are dissimilar in humans and fungi are shown as blue sticks. Dissimilar residues that project thepolypeptide backbone carbonyl to within 4 Å of the ITC (T507, S508, and M509) are shown by element (oxygen, red; nitrogen, blue; carbon, green; sulfur,yellow). (B) Amino acid residues within 6 Å of lanosterol and the lanosterol-like product in ScErg11p. Amino acid residues within 6 Å of lanosterol and thelanosterol-like product are shown as colored sticks on a cartoon polypeptide backbone that extends the primary sequence (in cyan) two amino acids N-terminal

Legend continued on following page

Monk et al. www.pnas.org/cgi/content/short/1324245111 5 of 8

and C-terminal of these residues. The V70-K75 α-helix, the F92-V102 loop, and the G403-L407 β-sheet also contribute residues within 6 Å of the lanosterol andits reaction product. Amino acid residues that are conserved in the Erg11ps of S. cerevisiae and the major pathogenic fungi of humans and not in HsCYP51 (L95,F241, M313, H381, S382) are shown in red. Amino acid residues that are conserved in fungal Erg11ps and HsCYP51 are shown in pink (G73, Y126, L129, T130,F133, Y140, F236, G315, R385). Amino acid residues that are similar in fungi but not in humans are shown in white (I139, F507, H405); residues that are similar inhumans and fungi are in gray. Amino acid residues that are dissimilar in humans and fungi are shown as blue sticks. Dissimilar residues that project thepolypeptide backbone carbonyl to within 4 Å of the lanosterol or the reaction product (A124, A125, T507, S508, and M509) are shown by element (oxygen, red;nitrogen, blue; carbon, green; sulfur, yellow). (C) Mutations in amino acid residues neighboring the active site and substrate entry channel that confer FLCresistance to fungal pathogens. The annotation is as in A, with mutated residues shown in yellow and annotated in black.

Fig. S4. Conservation of amino acid residues in ScErg11p and the Erg11ps of 10 prominent fungal pathogens. The cartoon of ScErg11p is shown with the hemeas blue sticks, ITC in yellow sticks, and with either conserved amino acid residues as red sticks or similar amino acid residues as white sticks. The figure comparesthe primary sequence of ScErg11p used in the present study with its homologs in 10 major fungal pathogens affecting humans: C. albicans (accession no. EMBLP10613.2), C. glabrata (EMBL P50859.1), C. dubliniensis (EMBL AAK57519.1), C. tropicalis (EMBL P14263.2), Cryptococcus neoformans (EMBL AAF35366.1),Aspergillus fumigatus (EMBL Q4WNT5 and EMBL Q96W81), Coccidioides immitis (RefSeqXP_001240306.1), Pneumocystis carinii (EMBL ABG91757), Ajellomycescapsulatus or Histoplasma capsulatum (EMBL AAU01158.1), and Malassezia globosa (RefSeqXP_001730619.1). The primary sequences were aligned using theConstraint-based Multiple Alignment Tool (Cobalt; www.ncbi.nlm.nih.gov/tools/cobalt/cobalt.cgi?link_loc=BlastHomeLink) plus some minor modifications byeye. (A) Side view of ScErgllp. (B) Opposite side view of ScErgllp. (C) Top view of ScErgllp.

Monk et al. www.pnas.org/cgi/content/short/1324245111 6 of 8

Fig. S5. Expression of ScErg11p-GFP and purification of ScErg11p-6xHis. (A) Confocal microscopy of GFP conjugates of the plasma membrane ATP-bindingcassette transporter ScPdr5p, the cytosolic marker CaUra3p, and ScErg11p. Yeast cells were grown in YPD medium to OD600nm = 3.0, harvested, and washedwith fresh medium, and suspensions at OD600nm = 1.0 were treated with 0.5 μg/mL DAPI for 1–20 min at room temperature before analysis by confocal mi-croscopy (Zeiss 710 confocal laser scanning microscope) using white light illumination and laser illumination at 405 and 488 nm. The fluorescence emission wasdetected at ∼455 nm (DAPI, false color blue) and at ∼507 nm (GFP, false color green) for slices of 0.4 μm. White light and fluorescent images are combined fora single slice. The images show that ScPdr5p-GFP is associated almost exclusively with the plasma membrane, Ura3p-GFP is associated with the cytosol and isexcluded from organelles, and ScErg11p-GFP surrounds the nucleus. (B) SDS/PAGE analysis of samples from the purification of the ScErg11p-6xHis used forcrystallization. Lanes 1 and 9: protein molecular weight markers (BenchMark, Invitrogen); lane 2: 50 μg crude yeast membranes; lane 3: DM extract from 50 mgcrude yeast membranes; lane 4: flow through of Ni-NTA column from 50 μg crude yeast membranes; lane 5: 10 μL nonconcentrated Ni-NTA affinity-purifiedScErg11p-6xHis; lane 6: 10 μL 5× concentrated Ni-NTA affinity-purified ScErg11p + 50 μM ITC; lane 7: 23 μg Ni-NTA affinity- and SEC-purified ScErg11p-6xHis;lane 8: 124 μg Ni-NTA affinity- and SEC-purified and concentrated ScErg11p-6xHis. (C) Preparation of purified ScErg11p-6xHis by SEC. Samples of 0.5 mL ofconcentrated Ni-NTA affinity-purified ScErg11p-6xHis in the presence of 50 μM ITC were separated using a Superdex 20 10/300 column (GE Healthcare)equilibrated with 20 mM Hepes (pH 7.5), 0.15 M NaCl, 10% glycerol (wt/vol), 2 μM ITC, 6.4 mM DM, 0.5 mM PMSF, and 1 Roche minus EDTA protease minipillper 400 mL at a flow rate of 0.4 mL/min. Protein was detected at 280 nm, and the heme detected at 420 nm was normalized relative to the peak maximum. Theprotein and heme peak coincided, and their relative shapes indicated that the bulk of the protein in the peak was heme-containing ScErg11p-6xHis, which ranwith a mobility expected of the 61-kDa protein inserted in a DMmicelle of 35 kDa. The enzyme purified in the absence of ITC showed the same behavior duringSEC. The molecular weight standards used were thyroglobulin, gamma globulin, ovalbumin, myoglobin, and vitamin B12. The DM micelle was detected usingrefractive index measurements. (D) Spectra of concentrated Ni-NTA affinity- and SEC-purified ScErg11p-6xHis ± 2 μM ITC. The spectra of 1-μL samples weredetermined using a Pearl Nanodrop with a 0.2-mm path length. The alpha and beta peaks were detected in these preparations at 570 and 540 nm, respectively.

Legend continued on following page

Monk et al. www.pnas.org/cgi/content/short/1324245111 7 of 8

(E) Mass spectrometry analysis of tryptic fingerprints of ScErg11p. The protein sequence of ScErg11p-6xHis used in the present study. A sample of the proteinused for crystallography was separated by SDS/PAGE as a 62-kDa protein band and analyzed by mass spectrometry of tryptic peptides at the Stanford UniversityMass Spectrometry facility. Peptides detected by tandem mass spectrometry are shown in bold. Coverage of 70% of the ScErg11p-6xHis primary sequence wasobtained. (F) Crystals of ScErg11p-6xHis. The crystals were formed by vapor diffusion in 6-μL drops by mixing the purified protein 1:1 with 45% (wt/vol) PEG400in 100 mM glycine at pH 9.3. The crystals are ∼100 μm in length.

Table S1. Yeast strains used in the present study

Strain number Genotype

MMLY663 (ADΔ) MATα his1 Δura3 pdr1-3 Δpdr3::hisG Δyor1::hisG Δsnq2::hisG Δycf1::hisG Δpdr10::hisGΔpdr11::hisG Δpdr15::hisG Δpdr5::hisG

MMLY941 ADΔ Δpdr5::pABC3ScERG11-6xHisMMLY942 ADΔ Δpdr5::pABC3ScERG-GFPMMLY939 ADΔ Δpdr5::pABC3ScPDR5-GFPMMLY1834 ADΔ Δpdr5::pABC3CaURA3-GFP

The MMLY941 strain was constructed by transforming the yeast host strain ADΔwith a linear transformation cassette comprising theAscI fragment from the pABC3ScERG11-6xHis at the PDR5 locus and selecting ura+ progeny (1). The strains MMLY942, MMLY939, andMMLY1834 were obtained using recombinant PCR to prepare constructs that replaced a C-terminal 6xHistidine tag with EGFP aspreviously described (1). The PDR5 locus in each yeast transformant was recovered by PCR of genomic DNA, and its DNA sequence wasdetermined.

Table S2. X-ray crystal structure dataset statistics

Erg11 datasets Empty pocket + ITC + FLZ + VCZ + Estriol + Lanosterol

Space group P21 P21 P21 P21 P21 P21dmin 2.4 Å 2.2 Å 2.5 Å 2.8 Å 2.1 Å 1.9 Åa 80.1 Å 79.6 Å 80.31Å 80.1 Å 79.4 Å 78.2Åb 67.7 Å 67.6 Å 67.3 Å 67.7 Å 67.5 Å 67.0 Åc 81.0 Å 80.7 Å 81.3 Å 81.2 Å 80.8 Å 80.4 Åα 90° 90° 90° 90° 90° 90°β 100.14° 99.75° 100.34° 100.14° 99.58° 99.54°γ 90° 90° 90° 90° 90° 90°

Completeness 95.8% (96.7%) 97.0% (99.0%) 91.7% (86.7%) 98.9% (99.7%) 97.0% (94.7%) 96.5% (72.8%)I/σI 31.3 (3.0) 25 (2.6) 22.7 (1.9) 31.3 (2.5) 32.2 (2.9) 26.6 (1.58)Rsym* 0.058 (0.486) 0.052 (0.410) 0.049 (0.472) 0.069 (0.577) 0.054 (0.541) 0.056 (0.577)Rfree

‡ 0.237 0.234 0.294 0.253 0.237 0.227Rcryst

† 0.196 0.201 0.227 0.191 0.198 0.195

*Rsym =P

hkl

Pi jIiðhklÞ− IðhklÞj=Phkl

Pi IiðhklÞ, where IiðhklÞ and IiðhklÞ are the ith and mean measurements of the intensity of

reflection hkl.†Rcryst =

PhklkFobsj−kjFcalck=

Phkl jFobsj, where Fobs and Fcalc are the observed and calculated structure factors for the reflection hkl and

k is a weighting factor.‡Rfree =

Phkl⊂ TkFobsj−kjFcalck=

Phkl⊂ T jFobsj, where Fobs and Fcalc are the observed and calculated structure factors for the reflection hkl

and k is a weighing factor. T is the test set of reflections.§Rmsd.

Monk et al. www.pnas.org/cgi/content/short/1324245111 8 of 8