Structural Basis for the Interaction of Lipopolysaccharide with Outer Membrane Protein H (OprH) from Pseudomonas aeruginosa

Structural Basis for the Interaction of Lipopolysaccharidewith Outer Membrane Protein H (OprH) from Pseudomonasaeruginosa*□S

Received for publication, July 8, 2011, and in revised form, August 17, 2011 Published, JBC Papers in Press, August 24, 2011, DOI 10.1074/jbc.M111.280933

Thomas C. Edrington‡§, Erica Kintz‡¶, Joanna B. Goldberg‡¶, and Lukas K. Tamm‡§1

From the ‡Center for Membrane Biology and Departments of §Molecular Physiology and Biological Physics and ¶Microbiology,University of Virginia, Charlottesville, Virginia 22908

Background: Pseudomonas aeruginosa outer membrane protein OprH has been hypothesized to confer antibiotic resist-ance by interaction with LPS.Results:The structure ofOprHwas solved and LPS interactionwas demonstrated by solutionNMR supported by pulldown andbiochemical assays.Conclusion:OprH forms a �-barrel in membrane and interacts with LPS in vivo and in vitro.Significance: Structure and lipid interactions may help understand antibiotic resistance.

Pseudomonas aeruginosa is a major nosocomial pathogenthat infects cystic fibrosis and immunocompromised patients.The impermeability of the P. aeruginosa outer membrane con-tributes substantially to the notorious antibiotic resistance ofthis human pathogen. This impermeability is partially impartedby the outer membrane protein H (OprH). Here we have solvedthe structure of OprH in a lipid environment by solution NMR.The structure reveals an eight-stranded �-barrel protein withfour extracellular loops of unequal size. Fast time-scale dynam-ics measurements show that the extracellular loops are disor-dered and unstructured. It was previously suggested that thefunction of OprH is to provide increased stability to the outermembranes ofP. aeruginosa by directly interactingwith lipopo-lysaccharide (LPS) molecules. Using in vivo and in vitro bio-chemical assays, we show that OprH indeed interacts with LPSin P. aeruginosa outer membranes. Based upon NMR chemicalshift perturbations observed upon the addition of LPS to OprHin lipid micelles, we conclude that the interaction is predomi-nantly electrostatic and localized to charged regions near bothrims of the barrel, but also through two conspicuous tyrosines inthe middle of the bilayer. These results provide the first molec-ular structure of OprH and offer evidence for multiple interac-tions between OprH and LPS that likely contribute to the anti-biotic resistance of P. aeruginosa.

Pseudomonas aeruginosa is themost common cause of pneu-monia in cystic fibrosis patients (1). Pseudomonal infection in

the cystic fibrosis lung results in antibiotic-resistant biofilmsand is the leading cause of mortality in cystic fibrosis patients(2). It is also responsible for amajority of urinary tract and burnand wound infections and is the major nosocomial pathogen inhospital settings. The unusually high antibiotic resistance ofGram-negative P. aeruginosa is partially imparted by itsextremely tight and stable outer membrane (OM),2 whichmakes infections with these bacteria clinically very difficult totreat (3). Additionally, the OM is densely packed with lipopo-lysaccharide (LPS), or endotoxin, that contributes substantiallyto biofilm formation in the pathogenesis of P. aeruginosa infec-tions (4, 5). LPS molecules are found in the outer leaflet of theOM where they form a protective extracellular barrier againstthe penetration of potentially noxious molecules by divalentcation-mediated LPS-LPS interactions (3). Displacement ofdivalent cations from LPS by polycationic antibiotics such aspolymyxins and aminoglycosides through the self-promoteduptake pathway can lead to destabilization of the OM andincreased susceptibility of these bacteria to antibiotics (6, 7). Inlaboratory settings, the increased susceptibility of P. aeruginosato antibiotics caused by divalent cation displacement can beachieved by treatment with chelators such as EDTA.OprH is genetically linked to the PhoP-PhoQ two-compo-

nent regulatory system that is up-regulated in response toMg2�-limited growth conditions (8, 9). As a member of thecomplete P. aeruginosa Mg2� stimulon, the oprH-phoP-phoQoperon reinforces resistance to common antimicrobial cationicpeptides such as polymyxin B and aminoglycosides (10, 11).Under Mg2�-deficient growth conditions, OprH is up-regu-lated and overexpressed so that it becomes a major componentof the P. aeruginosaOM (8). Based on these correlations, OprHhas been hypothesized to play a critical function in antibioticresistance by acting as a surrogate for depleted Mg2� ions on

* This work was supported, in whole or in part, by National Institutes of HealthGrant RO1 GM51329 (to L. K.T.).

The atomic coordinates and structure factors (code 2LHF) have been deposited inthe Protein Data Bank, Research Collaboratory for Structural Bioinformatics,Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

The NMR assignments reported in this paper were submitted to the BiologicalMagnetic Resonance Databank under BMRB ID 17842.

□S The on-line version of this article (available at http://www.jbc.org) containssupplemental Figs. 1–7.

1 To whom correspondence should be addressed: P. O. Box 800886, Char-lottesville, VA 22908-0886. Tel.: 434-982-3578; Fax: 434-982-6814;E-mail: [email protected].

2 The abbreviations used are: OM, outer membrane; OprH, outer membraneprotein H; DHPC, 1,2-dihexanoyl-sn-glycero-3-phosphocholine; FhuA, fer-ric hydroxamate uptake A; TROSY, transverse relaxation optimized spec-troscopy; HSQC, heteronuclear single quantum coherence; r.m.s., rootmean square; r.m.s.d., root mean square deviation.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 286, NO. 45, pp. 39211–39223, November 11, 2011© 2011 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39211

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

by guest on D

ecember 3, 2015

http://ww

w.jbc.org/

Dow

nloaded from

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

by guest on D

ecember 3, 2015

http://ww

w.jbc.org/

Dow

nloaded from

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

by guest on D

ecember 3, 2015

http://ww

w.jbc.org/

Dow

nloaded from

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

by guest on D

ecember 3, 2015

http://ww

w.jbc.org/

Dow

nloaded from

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

by guest on D

ecember 3, 2015

http://ww

w.jbc.org/

Dow

nloaded from

the surface of the cell, thus stabilizing the LPS network on thesurface of the cell (12). This hypothesis is supported by evidencethat OprH stabilizes the OM by increasing its protection frommembrane perturbation and thereby increasing its antibioticresistance (11). Although numerous studies have hinted atOprH-LPS interactions, direct evidence of these interactionshas not been demonstrated previously.OprH is a 21-kDa, 200-residue, basic (theoretical pI �9.0)

protein that is integral to the OM of P. aeruginosa. Based onhydropathy analysis, OprH has been proposed to form a �-bar-rel with eight transmembrane strands and four extracellularloops (13). Circular dichroism data have further confirmed the�-sheet secondary structure of the protein. However, no high-resolution structure of OprH has yet been determined, and themolecular determinants of its postulated interaction with LPSremain elusive. Therefore, to better understand the contribu-tion of OprH to the antibiotic resistance of P. aeruginosa, wehave determined and evaluated its structure in a lipid environ-ment in the presence and absence of LPS by solution NMRmethods. We further show that LPS interacts directly withOprH in vitro and in vivo, and we identify interacting residuesthat could potentially be targeted in future studies aimed atfurther advancing our understanding of pseudomonal antibi-otic resistance.

EXPERIMENTAL PROCEDURES

FLAG Pulldown Assays—AnOprH deletion mutant strain ofP. aeruginosa PAO1 (PAO1�oprH) was created and trans-formed with the pHERD30T shuttle vector containing theOprH gene with a C-terminal FLAG tag, as described previ-ously (14–16). Plasmid-containing strainswith andwithout theclonedOprH genewere grown at 37 °C inM9minimalmediumwith either 20mM (low) or 500mM (high)MgSO4 for 8 h beforeinduction with 2% arabinose. Cells were pelleted by centrifuga-tion 12 h after induction and resuspended in 50mMTris-HCl atpH 8.0, 150 mM NaCl, 5% glycerol, and Complete EDTA-freeprotease inhibitor (Roche). Cells were lysed by sonication andultracentrifuged (75,000 � g for 60 min at 4 °C) to obtainedmembrane pellets. Membranes were solubilized by rotation for1 h at 4 °C in PBS with 2% �-octyl-glucoside, 2.5 mM EDTA andsupplemented to 500 mM NaCl. Insoluble membrane wasremoved by centrifugation (20,000 � g for 30 min at 4 °C), andthe final membrane lysate was used for immunoprecipitationwith 50 �l of FLAG magnetic matrix (Sigma). Protein waseluted using 50 �l of 3�FLAG peptide (150 ng/�l). The mem-brane proteins and LPS present in preimmunoprecipitationsamples were detected with polyclonal antisera specific forP. aeruginosa serogroup O5 (Accurate Chemical & Scientific).LPS present in the elutants was detected with a serotype O5specific monoclonal antibody (Rougier Bio-Tech), and OprHwas detected using the anti-FLAGM2 antibody (Sigma).Expression and Purification of OprH in Escherichia coli—

OprH without the N-terminal signal sequence (residues 1–22were replaced with Met-1 so that Ala-23 becomes Ala-2 in ournumbering system) was cloned from P. aeruginosa PAO1DNAinto a pET30a� vector (EMD Biosciences). Primers for theOprH loop deletion mutants (Operon) were designed, and thepET30a�-oprH vector was mutated using a QuikChange II

site-directed mutagenesis kit (Stratagene). The resulting plas-mids were transfected into BL21 (DE3) E. coli cells andexpressed the C-terminal His6-tagged construct to inclusionbodies under the control of the T7 promoter. Expression of theunlabeled proteins was performed in LB medium, whereas2H-,13C-,15N-labeled OprH was expressed in D2O (99.99%)minimal medium containing 2 g/liter 2H (98%)- and 13C (99%)-labeled D-glucose, 1 g/liter 15N (98%)-labeled (NH4)2SO4, and1% 2H-,13C-,15N-labeled Bioexpress (Cambridge Isotope Labo-ratories). Cells were grown at 37 °C to an A600 of 0.4–0.6 andinduced with 1 mM isopropyl �-D-1-thiogalactopyranoside.Cells were harvested by centrifugation (6000 � g, 15 min) 4–8h after induction and resuspended in 50 mM NaPO4 at pH 8.0,300 mMNaCl, and 25% sucrose. Cells were lysed by sonication,and the insoluble pellets were washed in 50 mM NaPO4 at pH8.0, 300 mM NaCl. The final inclusion body pellet was solubi-lized in 50 mM NaPO4 at pH 8.0, 300 mM NaCl, 20 mM imidaz-ole, and 8 M urea and loaded onto a Ni2�-nitrilotriacetic acidSuperflow (Qiagen) column by FPLC. OprH was eluted withelution buffer (50 mM NaPO4 at pH 8.0, 300 mM NaCl, 250 mM

imidazol and 8Murea). The puritywas increased by running theelutant through a second Ni2�-nitrilotriacetic acid column.Refolding and Sample Preparation—The refolding of OprH

into 1,2-dihexanoyl-sn-glycero-3-phosphocholine (DHPC)micelles followed the protocol previously described for E. coliOmpX with slight modification (17). Briefly, a sample contain-ing 0.4 mM OprH in elution buffer was diluted 10-fold into 20mM Tris-HCl at pH 8.5, 5 mM EDTA, and 0.6 M L-arginine(refolding buffer) with 3%DHPC (Avanti Polar Lipids Inc). Therefolding solution was incubated for 72 h at 37 °C before beingdialyzed (20min, room temperature) against 2.5 liters of 20mM

Tris-HCl at pH 8.5, 5 mM EDTA, and 50 mM KCl. The solutionwas concentrated, and the buffer was exchanged against 25mM

NaPO4 at pH 6.0, 50 mM KCl, 0.05% NaN3 with 5% D2O bydilution/concentration. Final NMR samples were concentratedto 1.0–1.3 mM OprH and contained 150–175 mM DHPC asdetermined by 1H NMR spectroscopy. Samples of OprH inDHPC:LPS mixed micelles were prepared by the addition ofpurified P. aeruginosa PAO1 LPS (Sigma) from stock solutionsdirectly to OprH-DHPC micelle samples. The LPS concentra-tion in the stock solutions was determined by thiobarbituricacid and Purpald assays for the measurement of 2-keto-3-de-oxyoctonate (18).Trypsin Susceptibility Assay—Trypsin from bovine pancreas

(Sigma)was added from stock solutions to samples of unlabeledOprH in DHPC or DHPC:LPS micelles in 20 mM Tris-HCl atpH 7.3 with either 5 mM EDTA or 2 mM MgCl2. After incuba-tion at 37 °C for 5 h, SDS-PAGE sample loading buffer wasadded to trypsin-treated and untreated samples that were sub-sequently boiled or not boiled at 100 °C for 15 min. Sampleswere then used for SDS-PAGE on 15%Tris-glycine gels supple-mented with 25 mM NaCl to increase band separation. Gelswere stained with Coomassie Brilliant Blue R-250 (Pierce), andgel densitometry analysis was performed using the NationalInstruments LabVIEW 2010 software.NMR Spectroscopy—All NMR experiments were recorded at

45 °C on a Bruker Avance III 800 spectrometer equipped with atriple-resonance cryoprobe. The one-dimensional TRACT

Structure and Interaction of OprH with LPS

39212 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 45 • NOVEMBER 11, 2011

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

pulse scheme utilized to determine rotational correlationtimes has been previously described (19). All double- andtriple-resonance experiments were from the Bruker Topspinversion 2.1.6 software suite. For detergent/lipid screeningand LPS studies, two-dimensional 15N-1H TROSY experi-ments were utilized. Sequential backbone assignmentswere obtained by recording TROSY versions of HNCA,HN(CO)CA, HN(CA)CB, HN(COCA)CB, HNCO, andHNCACO experiments. Both 15N-1H-1H NOESY-TROSYand 15N-15N-1H HSQC-NOESY-HSQC experiments withmixing times of 200 ms were recorded to obtain distanceconstraints. R1, R2, and {1H}-15N heteronuclear nuclearOverhauser effects (NOEs) were measured using two-di-mensional 15N-1H TROSY-based experiments. NMR datawere processed and analyzed with NMRPipe and Sparkysoftware, respectively (20).Structure Calculation—Distance constraints were calibrated

and calculated based upon an average distance of 3.3 Å between�-strands using Cyana version 2.1 (21). Backbone dihedralangle constraints were determined from chemical shifts cor-rected for deuterium andTROSY effects using TALOS (22, 23).Hydrogen bond constraints derived from 1H/2H exchangeexperiments were set at 2.5 and 3.5 Å for HN-O and N-O dis-tances, respectively. Structure calculations were performedusing CNS version 1.2 (24) (4,000 high temperature, 8,000 tor-sion slow-cool, and 8,000 Cartesian slow-cool annealing steps).A total of 200 structures were calculated, and the 20 lowestenergy structures were selected for ensemble analysis. Ram-achandran map analysis was performed with PROCHECK-NMR (25).

RESULTS

OprH Interacts with LPS in P. aeruginosa Outer Membranes—Using FLAG pulldown assays from membrane lysates ofP. aeruginosa bacterial strains containing a vector thatexpressed OprH with a C-terminal FLAG tag, we tested thehypothesis that OprH binds specifically with LPS in vivo. To dothis, we first generated amutantP. aeruginosaPAO1 strain thathad theOprH gene deleted (PAO1�oprH) and transformed themutant strain with either the empty pHERD30T vector as acontrol or the vector containing OprH-FLAG (pHERD30T-oprH-FLAG). Each strain was grown in minimal medium sup-plemented with limited concentrations of MgSO4 (low Mg2�),and PAO1�oprH (pHERD30T-oprH-FLAG) strains were addi-tionally grown in minimal medium supplemented with highconcentrations ofMgSO4 (highMg2�). After induction, the celldensity of the cultures was normalized to maintain equivalentconcentrations of membrane protein and LPS in all the finalmembrane lysates (Fig. 1, A and B). The lysates were then usedfor the immunoprecipitation of OprH-FLAG (Fig. 1, C and D).When OprH was precipitated from the membrane lysatesof PAO1�oprH (pHERD30T-oprH-FLAG), a considerableamount of LPS was co-immunoprecipitated. The amount ofLPS that was co-immunoprecipitated by OprH-FLAG fromthese cultures was also dependent on the concentration ofMg2� in the growth medium. Significantly less LPS was co-immunoprecipitated from the OM lysates of PAO1�oprH(pHERD30T-oprH-FLAG) cultures that were grown in high

Mg2�when comparedwith lowMg2�.When theOM lysates ofPAO1�oprH (pHERD30T) were used for the immunoprecipi-tation, only trace amounts of LPS were co-immunoprecipi-tated, indicating that co-immunoprecipitation of LPS byOprH-FLAGwas specifically due to lowMg2�-enhanced bind-ing of LPS to OprH in vivo.Protease Protection of OprH by LPS—To directly demon-

strate that LPS binds to OprH in vitro, the protein wasexpressed in E. coli and subsequently purified from inclusionbodies. This was achieved by deleting the membrane-targetingN-terminal signal sequence from OprH and attaching a C-ter-minal His6 tag to facilitate protein purification. The resultingOprH-His6-containing inclusion bodies were solubilized in 8 M

urea and purified in an unfolded form. The purified protein wasrefolded using numerous detergents and lipids includingdodecylphosphocholine, 1-myristoyl-2-hydroxy-sn-glycero-3-phosphocholine, N,N-dimethyldodecylamine-N-oxide, and DHPC.Each detergent or lipid was tested for the ability to producefolded protein for biochemical analysis and NMR experiments.The efficiency of refolding OprH was monitored by SDS-

PAGE. Like many other OM proteins, the apparent molecularmass of OprH on SDS gels shifts when going from an unfoldedto a folded form; the native protein runs at 18 kDa but shifts to21 kDa after boiling in SDS-PAGE loading buffer (26, 27). We

FIGURE 1. Association of LPS with OprH in Pseudomonas outer mem-branes. The membrane lysates from PAO1 �oprH containing either the clon-ing vector pHERD30T (Empty Vector) or pHERD30T-oprH-FLAG (OprH-FLAG)were prepared and solubilized in 20 mM NaPO4 at pH 7.4, 500 mM NaCl, 2%octyl-glucoside and 2.5 mM EDTA. A and B, Western blots of aliquots takenfrom the OM lysates. The aliquots were either left untreated (A) or treated (B)with proteinase K (10 �g/ml) prior to Western immunoblot analysis utilizing apolyclonal antibody specific for P. aeruginosa serogroup O5. C and D, afterimmunoprecipitation of OprH-FLAG, the resulting precipitate from eachrespective sample was immunoblotted with either anti-FLAG (C) or anti-LPS(D) monoclonal antibodies.

Structure and Interaction of OprH with LPS

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39213

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

also noted that the addition of the His6 tag causes a slightincrease in the apparent molecular weight from these previ-ously reported values. This reversible “heatmodifiability” of theSDS-gel mobility of OM proteins has been used frequently as areliable assay for their refolding. We found that OprH refoldedmost efficiently in the presence of the short chain lipid DHPC(supplemental Fig. 1). The quality and stability were furtherassessed by exchanging the refolded protein samples in differ-ent detergents or lipids into NMR buffer and acquiring 15N-1HTROSY spectra at 45 °C (supplemental Fig. 2). Again, the bestresults in terms of efficiency, quality, and stability were foundwhen OprH was refolded into DHPC micelles.To assay for LPS binding to OprH in DHPCmicelles in vitro,

we exploited the arginine and lysine cleavage sites in the pre-dicted extracellular loops ofOprH. Each predicted extracellularloop has at least one potential trypsin cleavage site. After diges-tion of refolded OprH in DHPCmicelles by trypsin, the appar-ent molecular weight of the protein on SDS gels before boilingthe sample remained similar to the undigested protein (Fig. 2A).However, in contrast to the undigested protein, unfolding thetrypsin-digested sample by boiling it in SDS loading bufferyielded several smaller protein fragments, the largest of whichwere�8 and 5 kDa (Fig. 2A, lane 4). Therefore, the extracellularloops of OprH in DHPC micelles were accessible to trypsindigestion. When LPS was included in the micelles (1:125, LPS:DHPCmolar ratio), some extracellular loops became protectedfrom trypsin digestion as evidenced by the appearance of twoprotein bands with apparent molecular masses of �18 and 15kDa after boiling the mixed micelle protein sample (Fig. 2A,lane 8). A schematic diagram in Fig. 2A shows maps of theobtained cleavage products.The addition of MgCl2 to the digestion buffer resulted in

complete reversal of this protease protection (Fig. 2A, lane 12).The relative fraction of each protected bandwas determined bygel densitometry and was found to be dependent on the LPSconcentration in the mixed micelles (Fig. 2, B and C). The frac-tion of protected protein reached saturation at an �10-foldexcess of LPS overOprH, atwhich point only 20%of the proteinremained unprotected. To ensure that the trypsin protection ofOprH was not a general consequence of adding charged lipidswith increased acyl chain lengths (12-carbon versus 6-carbonfor LPS and DHPC, respectively) to the protein-micelle com-plex, OprH in DHPC micelles was incubated with similar con-centrations of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocho-line, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine, or1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol. In each case,no protection from trypsin cleavage was observed (supplemen-tal Fig. 3). Thus, protection of the extracellular loops of OprHresults from specific binding to LPS and is dependent on theconcentration of Mg2�.

Mass spectrometry of the 18-kDa LPS-protected bandrevealed that it consisted of an OprH fragment comprising res-idues Asn-26–His-187 (His6 tag included) (data not shown).No signal was observed for residues Thr-13–Arg-25, and only aweak signal from a peptide corresponding to Ala-2–Glu-12wasobserved. Therefore, the predicted extracellular loops 2, 3, and4 of OprH were all at least partially protected from trypsincleavage by the presence of LPS. In contrast, residue Arg-25 in

predicted extracellular loop 1 was completely accessible totrypsin cleavage even in the presence of LPS. Based upon themolecular masses of the possible products resulting from tryp-

FIGURE 2. Trypsin protection of OprH by LPS. A, SDS-PAGE gel showing trypsindigestion products under different conditions. 16 mM OprH in 20 mM Tris-HCl atpH 7.3 and 25 mM DHPC were incubated at 37 °C under the conditions shown for5 h prior to running the gel. Samples without MgCl2 also contained 5 mM EDTA. Aschematic representation of the digestion products after treatment of OprH inDHPC micelles and DHPC:LPS mixed micelles with trypsin is shown on the bot-tom. Cleavage sites in the extracellular loops protected in both DHPC micellesand DHPC:LPS mixed micelles, as described under “Results,” are denoted (*) B, LPSconcentration dependence of OprH protection from trypsin cleavage shown bySDS-PAGE. Samples contained 16 mM OprH in 20 mM Tris-HCl at pH 7.3, 25 mM

DHPC, and LPS concentrations as indicated. C, the relative integral fractions of the18- (f), 15- (●), or 8-kDa (Œ) apparent molecular mass bands determined bySDS-PAGE gel densitometry are plotted against the LPS concentration. Error barsrepresent the standard deviations of five independent experiments.

Structure and Interaction of OprH with LPS

39214 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 45 • NOVEMBER 11, 2011

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

sin digestion of the extracellular loops, the 8- and 5-kDa frag-ments were identified as corresponding toVal-36–Lys-109 andAsp-114–Lys-159, respectively, which means that residuesLys-70 and Arg-72 in or near predicted loop 2 were protectedfrom trypsin cleavage even in the absence of LPS.Solution Structure ofOprH inDHPCMicelles—Todetermine

the structure of OprH by solution NMR, isotope-labeled OprHwas produced by expressing the protein in E. coli that was cul-tured inminimal medium supplemented with 15N-(NH4)2SO4-and 2H-,13C-labeled glucose. The protein was then purified andrefolded into DHPC as described above. The 15N-1H TROSYspectrum of 2H-,13C-,15N-labeled OprH in DHPCmicelles dis-played excellent dispersion of the backbone amide resonancesfrom 7.5 to 9.5 ppm, with a slightly greater amount of peakoverlap from 8.0 to 8.5 ppm (Fig. 3). 179 major cross-peakscould be identified. The overall correlation time of the OprH-DHPC micelle complex at 45 °C was determined to be 22 nsusing one-dimensional TRACT (19) (supplemental Fig. 4).Based upon the correlation times of theOmpX-DHPCcomplexat 30 °C (60 kDa, 21 ns) (19) and the PagP-dodecylphosphocho-line complex at 45 °C (50–60 kDa, 20 ns) (28), the molecularmass of the OprH-DHPC complex was estimated to be 60–65

kDa. Although 22 ns represents a lower limit to the actual over-all correlation time of the complex, these results encouragedthe acquisition of triple-resonance NMR experiments for theassignment and structure determination of OprH in DHPCmicelles.Assignment of nitrogen, hydrogen, C�, C�, and CO reso-

nances for 2H-,13C-,15N-labeled OprH in DHPC micelles wasachieved by using three pairs of TROSY-based triple resonanceNMR experiments to establish through-bond connectivity:HNCA, HN(CO)CA, HN(CA)CB, HN(COCA)CB, HNCO, andHN(CA)CO. By thismethod, 89% of the nitrogen and hydrogenresonances, 94% of the C� resonances, 91% of the C� reso-nances, and 93% of the CO resonances were assigned. Theremaining resonances were not assignable due to excessiveline broadening or ambiguity. Complete backbone chemicalshift assignments were obtained for 156 of the 180 residues(excluding the C-terminal His6 tag and lead methionine) inthe OprH-His6 construct. Only partial assignments could beobtained for 13 residues, which included the 2 prolines, and11 residues remained completely unassigned. Local second-ary structures determined from the secondary C� and C�chemical shifts (29) are shown in Fig. 4. Eight distinct regions

FIGURE 3. 15N-1H TROSY spectrum of 2H-,13C-,15N-labeled OprH in DHPC micelles collected at 800 MHz and 45 °C. The refolded protein sample wasexchanged into 25 mM NaPO4 at pH 6.1, 50 mM KCl, 0.05% NaN3, and 5% D2O before being concentrated to �1.0 mM for NMR experiments. Assignmentsdetermined as described under “Results” are shown. For some residues, all found in the micelle-solvent interfacial region of the protein, a second set of weakerpeaks could be assigned. These residues are denoted with an apostrophe.

Structure and Interaction of OprH with LPS

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39215

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

of large negative values characteristic of �-strands are sepa-rated by regions of values around zero, indicative of turns orrandom coils.The NOEs to generate structural distance constraints were

acquired through 15N-1H-1HNOESY-TROSY and 15N-1H-15NHSQC-NOESY-HSQC experiments. Redundancy from theacquisition of both spectra reduced any ambiguities resultingfrom overlapping peaks, and the assignments of many NOEscould be verified through two pairs of cross-peaks. A total of221 and 285 peaks were assigned and integrated in the 15N-1H-1H NOESY-TROSY and 15N-15N-1H HSQC-NOESY-

HSQC spectra, respectively. Distance constraints were prefer-entially calculated from NOEs derived from the 15N-1H-1HNOESY-TROSY spectrum unless there was severe peak over-lap, in which case the 15N-15N-1HHSQC-NOESY-HSQC spec-trum was utilized. The pattern of long-range NOEs in bothspectra was characteristic of antiparallel �-sheet secondarystructure. Many of the residue pairs responsible for thestronger long-range NOEs were identified as participating inhydrogen bonds through 2H/1H exchange experiments (sup-plemental Fig. 5). Based upon these 2H/1H exchange resultsand the pattern of strong NOEs, a total of 134 hydrogen bond

FIGURE 4. Three-bond averaged secondary chemical shifts of OprH in DHPC micelles. The secondary chemical shifts, where the deviation (�) of eachresidue-specific C� and C� chemical shift from random coil values was determined as (�C� � �C�) � 1/3 (�C�

i�1 � �C�i � �C�

i�1 � �C�i�1 � �C�

i � �C�i�1),

are plotted as a function of the amino acid sequence. Large negative values are indicative of �-sheet secondary structure, whereas large positive values areindicative of �-helical structure. Chemical shifts were corrected for both deuteration and TROSY effects prior to analysis (22). The secondary structure patternobserved in the solution structure is shown on the bottom.

FIGURE 5. Topology schematic of OprH. Residues that were partially assigned are colored light gray, and residues that were completely unassigned arecolored dark gray. For all other residues, complete nitrogen, hydrogen, C�, C�, and CO assignments were obtained. Residues that face the lumen of the barrelare colored light blue. �-Strand residues are denoted as squares and were determined from the solution structure using the Kabsch and Sander secondarystructure algorithm provided with MOLMOL software (46, 47). Loop and turn residues are denoted as circles. Inter-residue lines represent long- and medium-range NOEs observed in NOESY experiments. Hydrogen bond constraints that were identified through 2H/1H exchange experiments are denoted as blackinter-residue lines.

Structure and Interaction of OprH with LPS

39216 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 45 • NOVEMBER 11, 2011

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

constraints were generated (Fig. 5). Only two of theseconstraints between Phe-50–Gly-53 were short- ormedium-range.The structure of OprH in DHPC micelles was calculated

from 199 NOE-derived distance constraints, 188 chemicalshift-derived backbone dihedral angle constraints (23), and 134hydrogen bond constraints. The overall features of the lowestenergy conformer from the 20 lowest energy NMR structureensembles (Fig. 6) revealed an eight-stranded antiparallel�-barrel (�-strands �1–�8) connected by four extracellularloops and three smaller periplasmic turns. A common featureof �-barrel OM proteins is the presence of two girdles of aro-matic residues located at the�-barrel rim that partition into themembrane-water interface and define the membrane-embed-ded boundaries of the protein (30, 31). As shown in Fig. 6C,quite a large number of aromatic residues are located at anddefine the outer interfacial rim of OprH. However, these resi-dues are more sparse around the periplasmic �-barrel rim.Additionally, the side chains of the aromatic residues Tyr-55and Tyr-80 are more centrally located in the nonpolar mem-brane interior.The well defined �-barrel region of OprH had an average

�-strand length of 11 residues and an average pairwise back-bone root mean square deviation (r.m.s.d.) of 0.85 � 0.20 Å in

the NMR ensemble (Table 1).We also calculated the structureswithout hydrogen bond constraints, in which case the sameoverall fold, but with an average pairwise backbone r.m.s.d. of0.97 � 0.15 Å, was obtained. The shear number (32) of the�-barrel was 10, and the tilt anglewith respect to themembranenormal (31) was 43°. Nearly all of the residues that remainedunassigned are located in the extracellular loops of OprHexcept for theC-terminalHis6 tag and the leadmethionine (Fig.5). The small number of long-range NOEs in the extracellularloop regions resulted in a decreased structural definition inthese parts of the structure and an overall backbone r.m.s.d. of6.55 � 1.40 Å when the loops are included. The periplasmicturns were well defined and, when considered with the �-bar-rel, produced a backbone r.m.s.d. of 1.03 � 0.19 Å. Overall,these results indicate that the precision of the structural ensem-ble of the OprH global fold was comparable with that of otherOM �-barrel proteins determined by solution NMRspectroscopy.Dynamics of OprH in DHPC Micelles—To determine

whether the lack of structural information obtained from loopresidues was due to increased backbone dynamics in theseregions or due to other issues, we measured longitudinal (T1)and transverse (T2) relaxation times along with {1H}-15N het-eronuclear NOEs at 800 MHz and 45 °C. These data are sensi-

FIGURE 6. Solution structure of OprH in DHPC micelles. A, NMR ensemble of the 20 lowest energy structures calculated. B, top-down view of the lowestenergy conformer of OprH from the ensemble of 20 lowest energy structures. C, two side views of the lowest energy conformer with the side chains of aromaticresidues (red) located in the ordered �-barrel region are shown. The �-barrel and loop/turn regions are colored blue and gray, respectively, in B and C.

Structure and Interaction of OprH with LPS

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39217

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

tive to fast time-scale (ps-ns) dynamics of each residue in theprotein. In general, T1 values are dependent on localized inter-nal motion, T2 values are inversely proportional to the rota-tional correlation time, and NOE values reflect the flexibility ofthe backbone amide N–H bond vectors (33, 34). As shown inFig. 7, A and B, residues in the extracellular loops of OprH hadsignificantly shorter T1 values and longer T2 values when com-paredwith values for residues located in thewell defined�-bar-rel region of the protein. Additionally, as shown in Fig. 7C,restrained motions in the �-barrel residues led to increased{1H}-15N NOE values relative to the less restrained motions ofthe loop and turn residues, which had values that decreasedwith increasing loop or turn length. Thus, the patterns of theT1,T2, and {1H}-15N heteronuclearNOE values followed the topol-ogy ofOprH. They consistently show that the extracellular loopregions of OprH are dynamic on the ps-ns time scale, are lessordered than the �-barrel region, and tumble independentlyfrom the micelle-embedded main body of the protein.Different Regions of OprH Interact with LPS—To further

determine which specific residues of OprH interact with LPS,purified LPS from P. aeruginosa PAO1 was added directly torefolded 2H-,13C-,15N-labeled OprH in DHPC micelles in thepresence of EDTA, which resulted in anNMR sample that con-tained a DHPC:LPS:OprH molar ratio of 150:10:1. Based uponthe results from the trypsin protection of OprH by LPS asshown in Fig. 2, this molar ratio should saturate binding of LPSto OprH. The 15N-1H TROSY spectrum acquired from thissample was still well dispersed. From the overlay of the spec-

trumonto the spectrumofOprH inDHPC, it was apparent thatthe addition of LPS resulted in a significant number of chemicalshift perturbations (supplemental Fig. 6A). The 15N-1HTROSYspectrum of OprH in DHPC:LPS was then assigned, utilizingthe assignments of OprH in DHPC-only micelles with an addi-tional triple-resonance HNCA experiment with the LPS-con-taining sample. The resulting chemical shift perturbationscould then bemapped onto the lowest energy conformer of theOprH structure ensemble (Fig. 8A). Themost significant chem-ical shift perturbations occurred for residues at the base ofextracellular loop 1 (Gly-11, Glu-12, Gly-40, Tyr-41, and Trp-42). These residues constitute a localized region of negativecharge on the surface of the barrel (Fig. 8B). Large chemicalshift perturbations were also observed in negatively chargedresidues located in turns 2 and 3 (Asn-90 and Glu-130). Thebackbone chemical shift perturbations for residues located inthe �-barrel region were mostly minor with the exceptions ofthe central tyrosine residues Tyr-55 and Tyr-80 (Fig. 8, A andB). Relatively small chemical shift perturbations were observedfor residues located within the more positively charged extra-cellular loops, namely: Arg-25 and Leu-27 in loop 1; Asn-69 inloop 2;Gly-110 andPhe-111 in loop 3; andLys-159 andLeu-163in loop 4. By direct comparison of a 15N-1H-1H NOESY-TROSY spectrum acquired for 2H-,13C-,15N-labeled OprH inDHPC:LPS mixed micelles to that of OprH in DHPC onlymicelles, �80% of the long-range distance constraints could beconfirmed through identical NOEs, whereas the remainingNOEs were unobservable, most likely due to line broadening.Thus, the addition of LPS to the micelle did not result inchanges to the global fold of the protein. When MgCl2 wasadded to OprH in DHPC:LPS, the TROSY spectrum was indis-tinguishable from the spectrum acquired in the presence ofEDTAexcept for a fewminor chemical shift differences in threeunassigned cross-peaks (supplemental Fig. 6B). The addition of6 mM 1,2-dimyristoyl-sn-glycero-3-phosphocholine, which hasan acyl chain length of 14-carbons, to OprH in DHPC micellesresulted in only minor chemical shift perturbations in a limitednumber of 15N-TROSY resonances, indicating that the pertur-bations resulting from the addition of LPS were not solely aconsequence of the increased acyl chain length of LPS relativeto DHPC (supplemental Fig. 6, C and D).Based upon the NMR and protease protection assay results

for OprH in DHPC:LPS mixed micelles, four OprH deletionmutants were made. Each mutant contained a deletion for adifferent extracellular loop of OprH to further examinewhether eliminating any one of these loops could abolish LPSprotection of trypsin digestion. The specific constructs were:OprH�17–38 (�L1), OprH�65–72 (�L2), OprH�108–114(�L3), and OprH�150–162 (�L4). Each OprH mutant inDHPC:LPS mixed micelles was treated with trypsin in theabsence or presence ofMgCl2 and boiled in SDS-PAGE loadingbuffer, and the resulting cleavage products were analyzed bySDS-PAGE (Fig. 9). LPS provided protection from trypsindigestion for all the OprH deletion mutants, indicating that nosingle loop deletionwas sufficient to abolishOprH-LPS bindingin the absence of Mg2�.The susceptibility of each specific cleavage site in loops 2, 3,

and 4 was assessed by qualitatively evaluating the apparent

TABLE 1NMR and refinement statistics for OprH structures in DHPC micellesResults were calculated among the 20 lowest energy CNS conformers of the struc-ture of OprH in DHPC micelles.

NMR distance and dihedral angle constraintsStructure calculationUnique HN-HN NOE distances 199Sequential 93Medium range 11Long range 95Intermolecular 0

Hydrogen bond constraints 134Dihedral angle constraints 188

ViolationsDistance constraints (Å) 0.017 � 0.001Dihedral angle constraints (°) 0.148 � 0.020Maximum distance constraint (�0.2 Å) 0Maximum dihedral angle (�2.0°) 0Deviations from idealized geometryBond lengths (Å) 0.001 � 0.00004Bond angles (°) 0.2780 � 0.0020Impropers 0.1106 � 0.0063

Ramachandran map analysisaMost favored regions (%) 66.0Additionally allowed regions (%) 29.3Generously allowed regions (%) 3.4Disallowed regions (%) 1.3

Ensemble r.m.s.d.Mean global backbone r.m.s. deviation (Å)

�-Sheet residuesb 0.85 � 0.20�-Sheet and turn residues 1.03 � 0.19All residues 6.55 � 1.40

Mean global heavy atom r.m.s. deviation (Å)�-Sheet residues 2.07 � 0.26�-Sheet and turn residues 2.20 � 0.23All residues 7.24 � 1.29

a Calculated using PROCHECK-NMR.b �-Sheet residues are defined as 3–12, 41–50, 53–60, 73–85, 92–104, 117–130,136–147, and 170–179 from the mean of the 20 conformers.

Structure and Interaction of OprH with LPS

39218 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 45 • NOVEMBER 11, 2011

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

molecular weight differences of the protected bands for eachrespective deletion mutant. The smallest observed band in theabsence and presence of MgCl2 for the OprH (�L3) mutationhad an apparent molecular mass of �13 kDa (Fig. 9A, lanes 5and 11). As shown in supplemental Fig. 7, this cleavage productwas also observed for the OprH (�L3) mutant in DHPCmicelles and could only result from the complete protection ofthe loop 2 residues Lys-70 and Arg-72 from trypsin digestion inboth DHPC micelles and DHPC:LPS mixed micelles. The dis-appearance of cleavage products 13 kDa also indicated thatthe concentration-dependent increase of trypsin protection

observed for full-lengthOprH in the absence ofMg2�, as shownin Fig. 2, B and C, was a result of increasing LPS protection ofresidues Lys-109 and Lys-112 in loop 3. Two protected bandswere observed for all the deletion mutants in the absence ofMgCl2 except for OprH (�L4) (Fig. 9A, lane6). Removal of thetrypsin cleavage site at residue Lys-159 in extracellular loop 4 bythe OprH (�L4) mutation resulted in only one protected band,indicating that residue Lys-159 in the full-length OprHremained partially accessible to trypsin cleavage in the presenceof LPS. The diagram in Fig. 9B shows the obtained cleavageproducts for the wild type and for each deletion mutant. Based

FIGURE 7. Backbone dynamics of OprH. A–C, the longitudinal relaxation times (A), transverse relaxation times (B), and {1H}-15N heteronuclear NOEs (C) of OprHin DHPC micelles determined at 800 MHz and 45 °C are plotted as a function of the amino acid sequence. Blue bars in the T1 and T2 plots are the upper limits ofthe standard deviations. Blue bars in the NOE plot represent the upper limits of the standard errors. The secondary structure pattern observed in the solutionstructure is shown on the bottom.

Structure and Interaction of OprH with LPS

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39219

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

upon these results, the susceptibility of the extracellular loopsof OprH to trypsin cleavage in the presence of LPS can bedefined as: loop 1 � loop 4 � loop 3 � loop 2.

DISCUSSIONThe newly determined structure of OprH consists of an

eight-stranded antiparallel �-barrel with four extracellularloops and three periplasmic turns (Fig. 6). To the best of ourknowledge, this is only the fifth polytopic integral membraneprotein structure that has been solved by NMR before crystalstructures were known. The others are PagP from E. coli (28)and the mitochondrial anion channel VDAC-1 (35), for both ofwhich a crystal structure was solved later, E. coli diacylglycerolkinase (36), and the murine mitochondrial uncoupling protein2 (37). The topology and secondary structure of OprH refoldedinto DHPCmicelles resemble a model that was previously gen-erated by hydropathy analysis and subsequently validated byinsertion and deletion mutagenesis on OprH that was overex-

pressed in P. aeruginosa OMs (13). Most �-strands terminatewithin �2 residues from the previously predicted model,except for �2 and �3, which extend for an additional 5 and 3residues in the extracellular direction, respectively. Six of theeight �-strands (�2–�7) also extend farther on the periplasmicside, resulting in tighter periplasmic turns in the structure thanin the predicted model. Because these differences are relativelysmall, interpretations of previous biological results requiringonly knowledge of the topology of OprH remain valid.Like in several other outer membrane proteins whose struc-

tures have been solved by NMR, the secondary C� and C�

chemical shifts, the longitudinal and transverse relaxationtimes, and the {1H}-15N heteronuclear NOE values of OprH inDHPC micelles show that the extracellular loops of OprH aredisordered and extended. Interestingly, many residues locatedat the barrel-loop interface have NOE values comparable withthose observed for the �-barrel residues, indicating restricted

FIGURE 8. Effect of LPS on the amide backbone resonances of OprH in DHPC:LPS mixed micelles. A, chemical shift perturbations between 15N-1H TROSYspectra of 2H-,13C-,15N-labeled OprH in DHPC micelles (150:1, DHPC:OprH molar ratio) and DHPC:LPS mixed micelles (150:10:1, DHPC:LPS:OprH molar ratio) inthe presence of 5 mM EDTA were determined. These differences are shown as compound chemical shift changes (��comp � [��2

HN � (��N/6.5) 2]1⁄2) (48)mapped color-coded onto the lowest energy structure of OprH in DHPC micelles. B, electrostatic surface potential plots of the lowest energy structure of OprHin DHPC micelles generated using the charge-smoothing algorithm in PyMOL (49).

Structure and Interaction of OprH with LPS

39220 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 45 • NOVEMBER 11, 2011

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

motions in these regions.When considering the shorter T2 val-ues indicative of intermediate conformational exchangeobserved in these residues (Fig. 7), it is clear that the structurehas unique slow dynamic properties in the interfacial regions atboth rims of the barrel. Similar observations have been madefor other outer membrane �-barrel proteins including OmpAand KpOmpA (38–40). Not unlike many other �-barrel mem-brane proteins, OprH has a girdle of aromatic residues at theouter barrel rim, but fewer such residues at the inner periplas-mic rim. Instead, OprH features two adjacent prominenttyrosines near the center of the lipid bilayer membrane (Fig. 6).

Previous studies provided only circumstantial indirect evi-dence that LPS might be interacting with OprH in Pseudomo-nas outer membranes, an interaction that has been hypothe-sized to contribute to the antibiotic resistance of these bacteriaunder lowMg2� conditions (11, 12). In this work, we utilized invivo and in vitro assays to demonstrate a direct interactionbetween the P. aeruginosa OM protein OprH and LPS. Theseresults are further bolstered and refined by NMR interactionstudies using the newly determined structure.The observation that significantly more LPS was co-immu-

noprecipitated by OprH from P. aeruginosamembrane lysates

FIGURE 9. Trypsin accessibility and protection of the OprH extracellular loops in the presence of LPS. A, trypsin digestion products after treatment of 16mM wild-type (WT) OprH and OprH loop deletion mutants (�L1–L4) in 20 mM Tris-HCl at pH 7.3, 25 mM DHPC, and 0.2 mM LPS with 0.7 mM trypsin in the presenceof either 5 mM EDTA or 2 mM MgCl2 and subsequent SDS-PAGE after sample boiling. B, schematic representation of the digestion products after treatment ofOprH and OprH deletion mutants in DHPC:LPS mixed micelles with trypsin. The digestion products after treatment of OprH in DHPC micelles with trypsin arealso shown for comparison. Potential trypsin cleavage sites in the extracellular loops are identified with arrows. Cleavage sites in the extracellular loopsprotected in both DHPC micelles and DHPC:LPS mixed micelles, as described under “Results,” are denoted (*).

Structure and Interaction of OprH with LPS

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39221

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

that were prepared from cultures grown in the presence of lowMg2� relative to cultures grown in high Mg2� suggests thatother genes that influence the affinity of OprH for LPS are up-regulated in a Mg2�-dependent fashion. Although OprHexpression in the PAO1�oprHmutant strain used in the immu-noprecipitation assays was abolished, the upstream promoterof the oprH-phoP-phoQ operon was not altered. Therefore,PAO1�oprH retained endogenous expression of the phoP-phoQ locus. This is important because phoPQ controls thepmrAB two-component regulatory system, which in turn reg-ulates structural alterations in LPS that increase resistance topolymyxin B (41–43). The addition of aminoarabinose to the1- and 4-phosphates of the LPS lipid A domain, as well as theaddition of palmitate to lipid A, resulting from P. aeruginosaPAO1 growth in low Mg2� medium, has been previously doc-umented (44). Therefore, the increase in the amount of LPSco-immunoprecipitated from the low Mg2� lysates by OprHindicates that such LPS structural alterations could lead to anincrease in the affinity of OprH for LPS.The trypsin digestion results with purified OprH that was

refolded into DHPCmicelles in vitro indicate that purified andreconstitutedOprH retains its native ability to bind unmodifiedLPS. Note that the LPS used in these assays was purified fromP. aeruginosa PAO1 grown in Mg2�-sufficient medium, andtherefore, the LPS lipidAdomain did not contain the additionalaminoarabinose and palmitatemoieties found under lowMg2�

conditions. Additionally, the complete reversal of LPS protec-tion from trypsin digestion upon the addition ofMg2� providesthe first direct evidence that OprH acts as a surrogate forMg2�

by cross-linking LPS, thereby tightening the outer membraneduring divalent cation deficiency.The notion that LPS binds directly to OprH is supported by

our NMR chemical shift perturbation experiments. The reso-nances of numerous residues were significantly shifted by LPS,but not by another more generic long-chain lipid (1,2-dimyris-toyl-sn-glycero-3-phosphocholine) that was added to themixed protein-detergent micelles. Because we were able toderive the structure of OprH in membrane-mimetic DHPCmicelles, these chemical shift perturbations could bemapped tospecific regions on this structure to identify LPS interactionsites. Multiple LPS interaction sites in different regions of thestructure were identified by these experiments (Fig. 8). Inter-estingly, most of these sites corresponded to the more nega-tively charged areas on the surface of the protein, most notablyat the base of loops 1 and 3 and in the periplasmic turns 2 and 3.Conspicuous LPS-specific chemical shift perturbations werealso found for the 2 unusual centrally located tyrosine residues.Chemical or physical reasons for these observed interactionsare presently not well understood, but will be the target offuture interaction studies.Detailed molecular interactions of LPS have been previously

reported with another OM protein, namely the ferric hydrox-amate uptakeA (FhuA) protein fromE. coliwhose crystal struc-ture was solved in complex with LPS (45). In this case, bindingis directed by numerous strong electrostatic interactionsbetween basic amino acid residues of FhuA with acidic phos-phate moieties located in the lipid A and inner core moieties ofLPS. Overall, the binding of FhuA to LPS encompasses a large

surface area including at least four �-strands and two extracel-lular loops. In addition to the interactions between OprH andLPS described above, our NMR results suggest a similar mech-anism of LPS binding to OprH. Although only weak chemicalshift perturbations were observed in these regions of the pro-tein, the lysines and arginines of the extracellular loops 2 and 3may form hydrogen bonds with the phosphate moieties of thelipid A domain of LPS. This binding mode would place theglucosamine backbone of lipid A in close proximity to the baseof extracellular loop 1, where the largest chemical shift pertur-bations were observed. Although this interpretation is also sup-ported by the protection of the OprH extracellular loops 2 and3 from trypsin digestion by LPS, it is not unlikely that OprHmay interact withmore than one LPSmolecule as its function isto replace Mg2� and stabilize the P. aeruginosa OM. It is alsoimportant to keep in mind that our studies so far only recordedperturbations at the polypeptide backbone level and that spe-cific interactions with side chains are beyond the reach of thepresent experiments. Therefore, a more definitive molecularmodel for interactions between OprH and LPS will have toawait the use of specific side-chain isotope labeling and possiblythe use of different LPS precursors in future efforts.In conclusion,wehave presented evidence for the interaction

between OprH and LPS both in native P. aeruginosa outermembranes and in a model membrane system using both bio-chemical and biophysical techniques. Beyond determining thestructure of OprH, our study also demonstrates that solutionNMR is a powerful tool to examine interactions of integralmembrane proteins with specific lipids in a fully solvated lipidicenvironment, which cannot be easily done by crystallography.Overall, the results of this study provide new insight into thestructure and role of OprH in P. aeruginosa outer membraneswhile offering new evidence for protein-lipid interactions thatlikely contribute to antibiotic resistance during P. aeruginosainfections.

Acknowledgments—We acknowledge the W. M. Keck BiomedicalMass Spectrometry Laboratory at the University of Virginia. Wethank Ming-Tao Pai, Ph.D. for initial work on this project.

REFERENCES1. Mesaros, N., Nordmann, P., Plésiat, P., Roussel-Delvallez, M., Van Eldere,

J., Glupczynski, Y., Van Laethem, Y., Jacobs, F., Lebecque, P., Malfroot, A.,Tulkens, P. M., and Van Bambeke, F. (2007) Clin. Microbiol. Infect. 13,560–578

2. Rajan, S., and Saiman, L. (2002) Semin. Respir. Infect. 17, 47–563. Delcour, A. H. (2009) Biochim. Biophys. Acta 1794, 808–8164. Rocchetta, H. L., Burrows, L. L., and Lam, J. S. (1999)Microbiol. Mol. Biol.

Rev. 63, 523–5535. Hunter, R. C., and Beveridge, T. J. (2005) J. Bacteriol. 187, 7619–76306. Hancock, R. E., and Chan, L. (1988) J. Clin. Microbiol. 26, 2423–24247. Vaara, M. (1992)Microbiol. Rev. 56, 395–4118. Nicas, T. I., and Hancock, R. E. (1980) J. Bacteriol. 143, 872–8789. Macfarlane, E. L., Kwasnicka, A., Ochs, M. M., and Hancock, R. E. (1999)

Mol. Microbiol. 34, 305–31610. Brown, M. R., and Melling, J. (1969) J. Gen. Microbiol. 59, 263–27411. Macfarlane, E. L., Kwasnicka, A., and Hancock, R. E. (2000)Microbiology

146, 2543–255412. Bell, A., Bains, M., and Hancock, R. E. (1991) J. Bacteriol. 173, 6657–666413. Rehm, B. H., and Hancock, R. E. (1996) J. Bacteriol. 178, 3346–3349

Structure and Interaction of OprH with LPS

39222 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 286 • NUMBER 45 • NOVEMBER 11, 2011

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

14. Choi, K. H., and Schweizer, H. P. (2005) BMCMicrobiol. 5, 3015. Flannagan, R. S., Linn, T., and Valvano, M. A. (2008) Environ. Microbiol.

10, 1652–166016. Qiu, D., Damron, F. H., Mima, T., Schweizer, H. P., and Yu, H. D. (2008)

Appl. Environ. Microbiol. 74, 7422–742617. Fernández, C., Adeishvili, K., and Wüthrich, K. (2001) Proc. Natl. Acad.

Sci. U.S.A. 98, 2358–236318. Lee, C. H., and Tsai, C. M. (1999) Anal. Biochem. 267, 161–16819. Lee, D., Hilty, C., Wider, G., andWüthrich, K. (2006) J. Magn. Reson. 178,

72–7620. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A.

(1995) J. Biomol. NMR 6, 277–29321. Güntert, P. (2004)Methods Mol. Biol. 278, 353–37822. Moseley, H. N., Sahota, G., and Montelione, G. T. (2004) J. Biomol. NMR

28, 341–35523. Cornilescu, G., Delaglio, F., and Bax, A. (1999) J. Biomol. NMR 13,

289–30224. Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P.,

Grosse-Kunstleve, R.W., Jiang, J. S., Kuszewski, J., Nilges,M., Pannu,N. S.,Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) ActaCrystallogr. D Biol. Crystallogr. 54, 905–921

25. Laskowski, R. A., Rullmannn, J. A., MacArthur, M. W., Kaptein, R., andThornton, J. M. (1996) J. Biomol. NMR 8, 477–486

26. Hancock, R. E., and Carey, A. M. (1979) J. Bacteriol. 140, 902–91027. Bell, A., and Hancock, R. E. (1989) J. Bacteriol. 171, 3211–321728. Hwang, P. M., Choy, W. Y., Lo, E. I., Chen, L., Forman-Kay, J. D., Raetz,

C. R., Privé, G. G., Bishop, R. E., and Kay, L. E. (2002) Proc. Natl. Acad. Sci.U.S.A. 99, 13560–13565

29. Metzler, W. J., Constantine, K. L., Friedrichs, M. S., Bell, A. J., Ernst, E. G.,Lavoie, T. B., and Mueller, L. (1993) Biochemistry 32, 13818–13829

30. Hong, H., Park, S., Jiménez, R. H., Rinehart, D., and Tamm, L. K. (2007)J. Am. Chem. Soc. 129, 8320–8327

31. Schulz, G. E. (2002) Biochim. Biophys. Acta 1565, 308–31732. Liu, W. M. (1998) J. Mol. Biol. 275, 541–54533. Kay, L. E., Torchia, D. A., and Bax, A. (1989) Biochemistry 28, 8972–897934. Jarymowycz, V. A., and Stone, M. J. (2006) Chem. Rev. 106, 1624–167135. Hiller, S., Garces, R. G., Malia, T. J., Orekhov, V. Y., Colombini, M., and

Wagner, G. (2008) Science 321, 1206–121036. VanHorn,W. D., Kim, H. J., Ellis, C. D., Hadziselimovic, A., Sulistijo, E. S.,

Karra, M. D., Tian, C., Sönnichsen, F. D., and Sanders, C. R. (2009) Science324, 1726–1729

37. Berardi, M. J., Shih, W. M., Harrison, S. C., and Chou, J. J. (2011) Nature476, 109–113

38. Arora, A., Abildgaard, F., Bushweller, J. H., and Tamm, L. K. (2001) Nat.Struct. Biol. 8, 334–338

39. Renault, M., Saurel, O., Czaplicki, J., Demange, P., Gervais, V., Löhr, F.,Réat, V., Piotto, M., and Milon, A. (2009) J. Mol. Biol. 385, 117–130

40. Liang, B., Arora, A., and Tamm, L. K. (2010) Biochim. Biophys. Acta 1798,68–76

41. Gunn, J. S., and Miller, S. I. (1996) J. Bacteriol. 178, 6857–686442. Gunn, J. S., Lim,K. B., Krueger, J., Kim,K., Guo, L., Hackett,M., andMiller,

S. I. (1998)Mol. Microbiol. 27, 1171–118243. Soncini, F. C., and Groisman, E. A. (1996) J. Bacteriol. 178, 6796–680144. Ernst, R. K., Yi, E. C., Guo, L., Lim, K. B., Burns, J. L., Hackett, M., and

Miller, S. I. (1999) Science 286, 1561–156545. Ferguson, A. D., Welte, W., Hofmann, E., Lindner, B., Holst, O., Coulton,

J. W., and Diederichs, K. (2000) Structure 8, 585–59246. Koradi, R., Billeter, M., andWüthrich, K. (1996) J. Mol. Graph. 14, 51–5547. Kabsch, W., and Sander, C. (1983) Biopolymers 22, 2577–263748. Mulder, F. A., Schipper, D., Bott, R., and Boelens, R. (1999) J. Mol. Biol.

292, 111–12349. DeLano, W. L. (2010) The PyMOL Molecular Graphics System, version

1.1, Schrödinger, LLC, New York

Structure and Interaction of OprH with LPS

NOVEMBER 11, 2011 • VOLUME 286 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 39223

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from

Supplemental Fig. 1 Purification of OprH and refolding into DHPC micelles. Arepresentative SDS-Page gel depicting the steps of OprH purification and refolding isshown. Lane 1, molecular weight markers; lane 2, uninduced culture of BL21(DE3) cellscontaining the pET30a+-OprH plasmid; lane 3, culture of BL21(DE3) cells containingthe pET30a+-OprH plasmid induced with 1 mM IPTG; lane 4, cell lysis supernatant frominduced culture; lanes 5-7, inclusion body wash steps; lane 8, solubilized inclusion bodiesin 8M urea that were subsequently used for Ni-affinity chromotography; lane 9, flowthrough from Ni-Nta column 1; lane 10, elutant from Ni-Nta column 1; lane 11, flowthrough from Ni-Nta column 2; lane 12, elutant from Ni-Nta column 2; lane 13, unboiledOprH refolded into DHPC micelles; lane 14, boiled OprH refolded into DHPC micelles.

Supplemental Fig. 2 15N-1H TROSY screening of OprH refolded into differentdetergents and lipids. The refolded samples of 15N-labeled OprH in LMPC, DPC, LDAO,and DHPC were exchanged into 25 mM NaPO4 at pH 6.1, 50 mM KCl, 0.05% NaN3, and5% D2O before being concentrated to approximately 1.0 mM. 15N-1H-TROSY spectrawere collected at 800 MHz and 45C.

Supplemental Fig. 3 SDS-PAGE gels showing the trypsin digestion products of OprHrefolded into DHPC micelles in the presence of increasing concentrations of the long-chain lipids POPC, POPE, and POPG. Samples containing 16 M OprH in 20 mM Tris-HCl at pH 7.3, 5 mM EDTA, 25 mM DHPC and lipid concentrations as indicated wereincubated with 0.7 M trypsin for 5 hrs at 37°C. SDS-PAGE gels were run after theaddition of SDS-PAGE sample loading buffer to each sample followed by boiling at100°C for 10 min. An SDS-PAGE gel of the trypsin digestion products of OprH refoldedin DHPC micelles in the presence of increasing concentrations of LPS, as described inFig. 2, is shown for reference.

Supplemental Fig. 4 1H NMR signal decay curves of OprH in DHPC micelles acquiredby 1D TRACT. The relative integral, Irel, from 6.5 – 10.5 ppm of each 1H NMR spectrumis plotted versus increasing relaxation delay times. The decay of the amide 1H NMRsignal is caused by the slow () and fast () relaxing 15N spin-states. The respectiverelaxation rates were determined by exponential curve fitting and an overall correlationtime of 22 ns was determined. Data were collected at 800 MHz and 45C.

Supplemental Fig. 5 Protection of backbone amide resonances from 1H=>2H exchange.A 1.0 mM sample of OprH in DHPC micelles was lyophilized and resuspended in 100%D2O. 15N-1H TROSY spectra were then acquired at 800 MHz and 45°C after specifictime points and analyzed to determine the residues protected from backbone amide1H=>2H exchange. Only the resonances that remained protected after 21 hours wereutilized to introduce hydrogen bond constraints in the structure calculations.

Supplemental Fig. 6

Supplemental Fig. 6 Chemical shift perturbations upon addition of LPS or DMPC to 2H-,13C-,15N-labeled OprH in DHPC micelles. All 15N-1H TROSY spectra were recorded in25 mM NaPO4 at pH 6.0, 50 mM KCl, and 5% D2O. (A) 15N-1H TROSY spectrum of 0.6mM OprH in DHPC:LPS mixed micelles (150:10:1, DHPC:LPS:OprH molar ratio) with5 mM EDTA (red) overlaid onto the spectrum of 1.0 mM OprH in DHPC micelles(black). (B) 15N-1H TROSY spectrum of 0.6 mM OprH in DHPC:LPS mixed micelleswith 2 mM MgCl2 (blue) overlaid onto the spectrum of OprH in DHPC:LPS mixedmicelles with 5 mM EDTA (red). (C) 15N-1H TROSY spectrum of 0.6 mM OprH inDHPC:DMPC mixed micelles (150:10:1, DHPC:DMPC:OprH molar ratio) with 5 mMEDTA (red) overlaid onto the spectrum of 1.0 mM OprH in DHPC micelles (black). (D)Compound chemical shift changes (comp=[2

HN + (N/6.5) 2]1/2) resulting from theaddition of 6 mM DMPC (red) and 6 mM LPS (black) relative to the chemical shifts ofOprH in DHPC only.

Supplemental Fig. 7 Trypsin protection of OprH loop deletion mutants by LPS. EachOprH deletion mutant at a concentration of 16 M in 20 mM Tris-HCl at pH 7.3 and 25mM DHPC was incubated at 37°C under the conditions shown on top for 5 hrs prior torunning the gel. Samples without MgCl2 also contained 5 mM EDTA.

Goldberg and Lukas K. TammThomas C. Edrington, Erica Kintz, Joanna B.  aeruginosa

PseudomonasProtein H (OprH) from Lipopolysaccharide with Outer Membrane Structural Basis for the Interaction ofMembrane Biology:

doi: 10.1074/jbc.M111.280933 originally published online August 24, 20112011, 286:39211-39223.J. Biol. Chem. 

  10.1074/jbc.M111.280933Access the most updated version of this article at doi:

  .JBC Affinity SitesFind articles, minireviews, Reflections and Classics on similar topics on the

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

  http://www.jbc.org/content/suppl/2011/08/24/M111.280933.DC1.html

  http://www.jbc.org/content/286/45/39211.full.html#ref-list-1

This article cites 48 references, 17 of which can be accessed free at

by guest on Decem

ber 3, 2015http://w

ww

.jbc.org/D

ownloaded from