Engineering a terbium-binding site into an integral membrane protein for luminescence energy transfer

Engineering a terbium-binding site into an integral membrane protein for luminescence energy transfer Jose´ Luis Va´zquez-Ibar, Adam B. Weinglass, and H. Ronald Kaback* Howard Hughes Medical Institute, Departments of Physiology and Microbiology and Molecular Genetics, Molecular Biology Institute, University of California, Los Angeles, CA 90095-1662

Luminescence resonance energy transfer with a lanthanide like Tb3ⴙ as donor is a useful technique for estimating intra- and intermolecular distances in macromolecules. However, the technique usually requires the use of a bulky chelator with a flexible linker attached to a Cys residue to bind Tb3ⴙ and, for intramolecular studies, an acceptor fluorophor attached to another Cys residue in the same protein. Here, an engineered EF- hand motif is incorporated into the central cytoplasmic loop of the lactose permease of Escherichia coli generating a high-affinity site for Tb3ⴙ 3ⴙ 3ⴙ (KTb ⬇ 4.5 ␮M) or Gd3ⴙ (KGd ⬇ 2.3 ␮M). By exciting a Trp residue in the coordination sequence, Tb3ⴙ bound to the EF-hand motif is sensitized specifically, and the efficiency of energy transfer to strategically placed Cys residues labeled with fluorophors is measured. In this study, we use the technique to measure distance from the EF-hand in the central cytoplasmic loop of lactose permease to positions 179 or 169 at the center or periplasmic end of helix VI, respectively. The average calculated distances of ⬇23 Å (position 179) and ⬇33 Å (position 169) observed with three different fluorophors as acceptors agree well with the geometry of a slightly tilted ␣-helix. The approach should be of general use for studying static and dynamic aspects of polytopic membrane protein structure and function. bioenergetics 兩 active transport 兩 lactose permease 兩 EF-hand motif

actose permease (LacY), encoded by the lacY gene of Escherichia coli (1), catalyzes galactoside兾H⫹ symport and is a paradigm for transport proteins from Archaea to the mammalian central nervous system that transduce free energy stored in electrochemical ion gradients into solute concentration gradients (reviewed in refs. 2–5). LacY has been solubilized and purified in a completely active state (reviewed in ref. 6) and functions as a monomer (see ref. 7). The molecule contains 12 transmembrane helices connected by hydrophilic loops with both the N and C termini on the cytoplasmic face of the membrane (Fig. 1A) (reviewed in refs. 8–10). In a functional mutant devoid of native Cys residues, each residue has been replaced with Cys (reviewed in ref. 10). Analysis of the mutant library in conjunction with a battery of techniques has led to the following developments (reviewed in refs. 11 and 12): (i) The great majority of the mutants are expressed normally in the membrane, exhibit significant activity, and only six side chains are clearly irreplaceable for active transport: Glu-126 (helix IV) and Arg-144 (helix V), which are indispensable for substrate binding, and Glu-269 (helix VIII), Arg-302 (helix IX), His-322, and Glu-325 (helix X), which are critical for coupling sugar and H⫹ translocation. (ii) Helix packing, tilts, and ligand-induced conformational changes have been determined. (iii) Positions that are accessible to solvent have been revealed. (iv) Positions where the reactivity of the Cys replacement is increased or decreased by ligand binding have been identified. (v) LacY has been shown to be a highly flexible molecule. (vi) A working model describing a mechanism for lactose兾H⫹ symport has been proposed. One powerful technique that has not been exploited for studying LacY is f luorescence resonance energy transfer (FRET) (13). This sensitive spectroscopic technique yields

L

www.pnas.org兾cgi兾doi兾10.1073兾pnas.052703599

information regarding inter- and intramolecular distances. Essentially the method analyzes how the lifetime and quantum yield of a fluorescent molecule (the donor) is influenced by another fluorescent molecule (the acceptor) and vice versa. By using Fo ¨rster theory, which states that the efficiency of energy transfer is inversely proportional to the sixth power of distance, distances between two fluorophors can be measured (14). The sensitivity and range of distances that can be calculated reliably from FRET depends on the spectral characteristics of the dyes and their relative orientations, which together determine R0, the distance at which energy transfer efficiency is 50%. The technique is typically most sensitive for distances between 20 and 100 Å (14, 15). Luminescence resonance energy transfer (LRET) is a particular type of FRET where a lanthanide atom (Tb3⫹ or Eu3⫹) transfers energy to an organic fluorescent acceptor. The technique exploits the remarkable luminescence properties of lanthanides (i.e., millisecond to submillisecond lifetimes, narrow and multiple emission bands in the visible spectrum, and unpolarized emission). As a consequence, lanthanide luminescence overcomes some of the limitations of conventional FRET experiments [e.g., the unpolarized long lifetime allows random orientation of both the donor and acceptor during the energy transfer process, thereby simplifying determination of the orientation factor between donor and acceptor (␬2)] (14). Current LRET studies use a thiol-reactive chelator to bind a Tb3⫹ ion at an engineered Cys residue in the protein (16–20). The chelator binds Tb3⫹ with high affinity, shielding the cation from nonradiation deexcitation processes (primarily solvent quenching) (19, 21). The resulting Tb3⫹ emission has a high quantum yield, sometimes approaching unity, and a long lifetime. Although these long-lived lanthanide chelates have been used successfully for estimating inter- and intramolecular distances in the range of 45 Å or more (16–20), they are problematic for short distance measurements in the range desired for most intramolecular interactions because: (i) reactive and highly luminescent chelates are not usually available commercially and require synthesis (21); (ii) the bulk of the chelator may perturb the local environment of the protein; (iii) the mobility inherent in the use of a bulky group with a linker makes distance calculations less precise, particularly under 40 Å; and (iv) both the chelator and the fluorophor are usually directed toward Cys residues, sample heterogeneity is introduced. An alternative possibility for binding Tb3⫹ at a defined position within a protein is to introduce an EF-hand motif, which Abbreviations: LacY, lactose permease; FRET, fluorescence resonance energy transfer; R0, the distance at which energy transfer efficiency is 50%; LRET, luminescence resonance energy transfer; ␬2, the orientation factor between donor and acceptor; DDM, dodecyl ␤,D-maltopyranoside; ␶, lifetime; Q, quantum yield; J, overlap integral between donor and acceptor. *To whom reprint requests should be addressed at: Howard Hughes Medical Institute兾University of California, 5-748 MacDonald Research Laboratories, Box 951662, Los Angeles, CA 90095-1662. E-mail: [email protected]. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

PNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6 兩 3487–3492

BIOCHEMISTRY

Contributed by H. Ronald Kaback, December 27, 2001

consistent with the proposed ␣-helical structure of this region and indicate that the approach may be of general use. Experimental Procedures Materials. TbCl3 (hexahydrate), BODIPY-N-(2-aminoethyl)-

maleimide, fluorescein-5-maleimide, and Alexa Fluor 594 C5 maleimide (sodium salt) were purchased from Molecular Probes. GdCl3 (hexahydrate) was purchased from Aldrich. 1-Palmitoyl2-oleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (sodium salt) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine were purchased from Avanti Polar Lipids. Deoxynucleotides were from Sigma–Genosys, and restriction endonucleases, T4 DNA ligase, and Taq polymerase were from Fisher Scientific. All other materials were reagent grade obtained from commercial sources. Construction of LacY Mutants. A DNA linker encoding the metal coordination site of an EF-hand motif (Fig. 1B) was inserted into a unique KasI restriction endonuclease site of pT7–5兾cassette lacY encoding wild-type LacY with residues 206–210 deleted from the middle cytoplasmic loop (27). Subsequently, the DNA fragment encoding the coordination site was subcloned by using BtgI and BsmI restriction endonucleases into pT7–5兾cassette lacY encoding Cys-less LacY with a C-terminal biotin acceptor domain followed by a stretch of six contiguous His residues (Cys-less LacY-EF). Mutants containing a single-Cys replacements for Val-169 or Ile-179 were generated by two-step PCR (28) and subcloned as PstI兾SpeI fragments into parental Cys-less LacY-EF, generating LacY-EF兾V169C or LacY-EF兾I179C. Each mutant was sequenced by using the dideoxynucleotide method (29). Bacterial Strains, Growth, and LacY Purification. E. coli HB101

Fig. 1. (A) Secondary structure model of LacY. Transmembrane helices are shown in boxes connected by hydrophilic loops represented by solid lines. Positions 169 and 179 are highlighted, and the arrow in the central cytoplasmic loop represents the site of insertion of the EF-hand motif. (B) Schematic representation of helix VI and the engineered EF-hand motif. The residues involved in Tb3⫹ coordination in the EF-hand motif are numbered, the Trp residue used to sensitize Tb3⫹ luminescence is emboldened, and the position of the Cys residues modified with fluorophor (169 and 179) are highlighted.

usually binds Ca2⫹. However, because of the similar ionic radii of Ca2⫹ and Tb3⫹ (1.06 and 0.98 Å), Tb3⫹ also binds well to these motifs (22–24). Typically, an EF-hand motif is composed of two ␣-helices connected by a loop containing 12 amino acids that form the metal coordination site (Fig. 1B). The amino acids at positions 1, 3, 5, 7, 9, and 12 of the metal-binding site, and often a coordinating water molecule, provide seven coordination oxygens for the cation, conferring high selectivity for Ca2⫹ or Tb3⫹ (25). Database analysis reveals that high-affinity EF-hand motifs contain acidic amino acids at most or all of the coordinating positions, with the exception of position 7 (Trp), where the coordination oxygen is provided by the main chain (26). In this paper, a novel approach for estimating distances in a polytopic membrane transport protein is presented. The coordination site of an EF-hand motif is engineered into the central cytoplasmic loop of LacY, generating a lanthanide-binding site. This site is then exploited for LRET experiments by using Tb3⫹ as a donor and fluorescent-labeled Cys residues at either of two positions in helix VI as acceptors. The distances measured are 3488 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.052703599

[hsdS20 (rB⫺, mB⫺), recA13,ara-14 proA2, lacY1, galK2, rpsL20 (Smr), xyl-5, mtl-1, supE44, ⌬⫺兾F⫺] was used as a carrier for the plasmids described and qualitative assessment of transport activity on MacConkey indicator plates containing 20 mM lactose. For protein purification, the appropriate plasmids were transformed into E. coli T184 (lacZ ⫺Y⫺). Cells were grown in 10 liters of Luria–Bertani broth at 37°C containing ampicillin (100 ␮g兾ml) to an OD600 of 0.8 and induced with 0.3 mM isopropyl 1-thio-␤,D-galactopyranoside. Cells were disrupted by passage through a French pressure cell and the membrane fraction isolated by centrifugation. Membranes were solubilized by adding n-dodecyl ␤-D-maltopyranoside (DDM) to a final concentration of 2%, and LacY was purified by Ni2⫹-NTA affinity chromatography as described (30). Purified LacY was subjected to SDS兾12% PAGE followed by silver nitrate staining and immunoblotting with anti-C-terminal antibody (31). Protein was assayed by using a microBCA kit (Pierce). Protein Labeling. Purified LacY-EF兾V169C or LacY-EF兾I179C (⬇10 ␮M) was incubated at 4°C overnight with 100 ␮M of a given maleimide-conjugated fluorophor while protecting the samples from light. Free fluorescent probe was removed by passage through a Sephadex G-25 spin-column (Pharmacia). Reconstitution into Proteoliposomes. After fluorescent labeling,

purified protein was reconstituted into proteoliposomes as described with minor modifications (32). Briefly, solubilized protein was added to a 3:1 (wt兾wt) mixture of 1-palmitoyl-2-oleoylsn-glycero-3-phosphoethanolamine and 1-palmitoyl-2-oleoyl-snglycero-3-[phospho-rac-(1-glycerol)] (sodium salt) dissolved in 0.5% DDM to a final lipid兾protein ratio of 12:1 (wt兾wt). The mixture was incubated at room temperature for 15 min with gentle agitation. DDM was then removed slowly by three successive extractions with polystyrene beads (Bio-Beads SM-2, Bio-Rad). The first two extractions were done at room temperVa´ zquez-Ibar et al.

Binding Experiments. Tb3⫹ binding to purified LacY-EF was

measured using Tb3⫹ luminescence sensitized by the Trp residue at position 7 of the coordination site. Aliquots (1 ␮l) of a concentrated stock solution of TbCl3 in 0.01 M HCl were added to 5 ␮M DDM solubilized Cys-less LacY-EF over a range of Tb3⫹ concentrations from 0 to 100 ␮M. All experiments were carried out at 25°C in 50 mM Pipes (pH 6.5). After excitation of the sample at 295 nm, Tb3⫹ emission was recorded at 542 nm. Because Tb3⫹ absorbs very weakly in solution, the signal reflects only bound Tb3⫹. The data were fit to Eq. 1 by using SIGMAPLOT nonlinear curve-fitting software (Jandel, San Rafael, CA): I⫽

I max关Tb3⫹兴F 3⫹ , 关Tb3⫹兴F ⫹ KTb

[1]

where I is the Tb3⫹ emission at a particular Tb3⫹ concentration, Imax corresponds to the Tb3⫹ emission when the system is saturated, [Tb3⫹]F is the free or nonbound Tb3⫹ concentration 3⫹ is the equilibrium dissociation after each addition, and KTb 3⫹] was determined from the difference: [Tb3⫹] constant. [Tb F F ⫽ [Tb3⫹]T ⫺ [Tb3⫹]B, where [Tb3⫹]T and [Tb3⫹]B correspond to the total and bound Tb3⫹ concentration, respectively. Assuming that Imax corresponds to a 100% of protein saturation (5 ␮M), [Tb3⫹]B was calculated from: [Tb3⫹]B ⫽ (I) (5兾Imax). Experimental data from studies assessing inhibition of Tb3⫹ luminescence by Gd3⫹ were fit to Eq. 2 by using SIGMAPLOT nonlinear curve-fitting software (Jandel): I⫽

I maxKapp ⫹ I0 , Kapp ⫹ 关Gd3⫹兴

[2]

Tb3⫹

where I is the emission at 542 nm measured at different concentrations of Gd3⫹, [Gd3⫹], Imax refers to the initial Tb3⫹ emission, and I0 is the Tb3⫹ emission at infinite Gd3⫹ concentration. Kapp is the apparent equilibrium dissociation constant, which was introduced into the following equation to calculate the 3⫹ ) (33): equilibrium dissociation constant for Gd3⫹ (KGd K 3Gd⫹ ⫽

Kapp 3⫹ , 1 ⫹ 共关Tb3⫹兴兾KTb 兲

Resonance Energy Transfer Theory and Distance Measurements. The efficiency of energy transfer for a single donor–acceptor pair is calculated experimentally from Eq. 4,

␶ DA , ␶D

[4]

where ␶D is the donor lifetime in the absence of acceptor, and ␶DA is the donor lifetime in the presence of acceptor. In LRET with Tb3⫹ as donor, ␶DA is the lifetime of the sensitized acceptor emission, which is in the millisecond range. Because acceptor emission lifetimes are normally in the nanosecond range, sensitized acceptor emission must be the result of energy transfer from Tb3⫹ (34). Va´ zquez-Ibar et al.

E⫽

R 60 共R 60

⫹ R 6兲

,

[5]

where R0, the Fo ¨rster distance, is the distance at which energy transfer efficiency is 50% (E ⫽ 0.5) and is characteristic for a given donor–acceptor pair. Fo ¨rster theory applies to LRET, because energy transfer occurs via spin–spin interactions (34). Here, R0 is related to the spectroscopic behavior of the donor and the acceptor and can be calculated by using Eq. 6: R 0 ⫽ 共8.785 ⫻ 10⫺11QD␩⫺4␬2J兲1/6nm ,

[6]

where ␩ is the refractive index of the medium between donor and acceptor, which is taken to be 1.4 for proteins in aqueous solution (14). ␬2 describes the relative orientation of transition dipoles in both donor and acceptor probes. The value of ␬2 used in this study (2兾3) assumes a randomized orientation of both donor and acceptor. This is a reasonable approximation, because Tb3⫹ emission is unpolarized, and the millisecond lifetime of Tb3⫹ allows averaging of the acceptor orientation (for more detail, see refs. 13 and 14). QD is the donor quantum yield in the absence of acceptor. With Tb3⫹, the very weak extinction coefficient makes it difficult to measure QD by using conventional fluorescent molecules as a standards. However, there are several indirect approaches to calculate QD. In particular, QD can be calculated from the relation QD ⫽ ␶D兾␶I, where ␶D is the experimental lifetime of Tb3⫹ in the EF-hand, and ␶I corresponds to the intrinsic lifetime of Tb3⫹ decay [4.75 ms (35)]. J is the overlap integral between donor and acceptor and is described in Eq. 7: J⫽

冘

关F D共␭兲␧共␭兲␭4⌬␭兴兾

⌬␭

冘

关FD共␭兲⌬␭兴 ,

[7]

⌬␭

where FD(␭) is the Tb3⫹ emission spectrum and ␧ (␭) is the molar extinction coefficient of the acceptor as a function of the wavelength (␭). J was calculated by using SIGMAPLOT and Microsoft EXCEL. Spectroscopy. Steady-state fluorescence measurements were per-

[3]

where [Tb3⫹] is the Tb3⫹ concentration preincubated with the 3⫹ is the equilibrium dissociation constant sample (30 ␮M), and KTb 3⫹ for Tb determined in Eq. 1.

E⫽1⫺

Fo ¨rster theory (14) describes the relationship between the efficiency (E) of long-range dipole–dipole coupling of a fluorescent donor to an acceptor chromophore and the distance between them (R) as shown in Eq. 5:

formed in a SLM–Aminco 8100 (Urbana, IL). Absorbance measurements were carried out in Spectramax plus (Molecular Devices). Time-resolved luminescence decay curves were acquired in a Spex Fluorolog lifetime phosphorimeter (Spex Industries, Metuchen, NJ) by using a xenonflash lamp as the excitation source. The lamp pulse has maximum intensity at 0.003 ms and continues for 0.01 ms. Data were acquired with an initial delay of 0.05 ms and at intervals of 0.06 ms for Tb3⫹ decay measurements or at 0.03 ms for acceptor-sensitized emission decay. Measurements were made in 5 ⫻ 5 mm quartz cuvettes at 25°C. Experimental decay curves were analyzed by nonlinear regression by using SIGMAPLOT software (Jandel) according to a multiexponential decay described in Eq. 8: I⫽

冘

I i e 共 ⫺t/␶i兲 ⫹ C,

[8]

where Ii and ␶i refer to the amplitude and lifetime, respectively, of every i component in the decay. C is the background signal. Both donor and sensitized acceptor emission decays fit a twoexponential decay. PNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6 兩 3489

BIOCHEMISTRY

ature for 2 h each, and the last extraction was performed at 4°C overnight. The proteoliposomes were harvested at 100,000 ⫻ gmax for 45 min and resuspended in a small volume of 50 mM Pipes (pH 6.5), followed by three cycles of freezing and thawing. Finally, the sample was stored at ⫺80°C at a final protein concentration of 1 mg兾ml. Immediately before use, the samples were subjected to a brief period of sonification in a bath sonifier.

Fig. 2. Tb3⫹ binding to purified Cys-less LacY-EF. Experiment was carried out as described in Experimental Procedures. Excitation and emission wavelengths were 295 and 542 nm, respectively. (Inset) Inhibition of Tb3⫹ luminescence by Gd3⫹. After incubation of purified Cys-less LacY-EF with 30 ␮M Tb3⫹, the decrease in Tb3⫹ emission at given concentrations of Gd3⫹ was measured. The 3⫹ ) was calculated by using Eqs. corrected dissociation equilibrium constant (KGd 2 and 3 as described in Experimental Procedures.

Results and Discussion Engineering a Functional Tb3ⴙ-Binding Site into the Central Cytoplasmic Loop. Ca2⫹-binding proteins have evolved to bind Ca2⫹ with

a wide range of affinities, reflecting their regulatory role. This is achieved by variation in the sequence of the EF-hand motifs (26). In this study, an EF-hand motif was engineered taking into account three considerations: (i) site-directed mutagenesis of EF-hand motifs has shown that substitution of the five negatively charged amino acids at positions 1, 3, 5, 9, and 12 with neutral amino acids dramatically reduces ion-binding affinity (36–38). Therefore, all of the negatively charged amino acids were retained in the engineered EF-hand sequence to ensure high affinity. (ii) An Asp residue was introduced at position 3 because of a known role in increasing Tb3⫹ affinity and decreasing Ca2⫹ affinity (37). (iii) One spectroscopic characteristic of Tb3⫹ is weak absorption. To sensitize Tb3⫹ luminescence, an aromatic ‘‘antenna’’ group with a high absorption efficiency must be introduced in close proximity (⬍5 Å), so that once excited,

Fig. 4. Luminescent decay of Tb3⫹ bound to Cys-less LacY-EF. The excitation wavelength was 295 nm, and Tb3⫹-sensitized emission was recorded at 487 nm. Tb3⫹ luminescent decay was fitted with a two-exponential model (see Table 1). The line represents a curve fit of the experimentally determined intensities. (Inset) Residuals (i.e., deviation of experimental from ideal) from fit to a two-exponential decay. The protein and Tb3⫹ concentrations were 5 and 10 ␮M, respectively.

energy is transferred efficiently to the lanthanide. The nature of this energy transfer obeys the Dexter mechanism of electron exchange, and the triplet excited state (phosphorescence) of the antenna is involved (39). Therefore, a Trp residue is introduced at position 7 of the EF-hand motif to serve this purpose (24, 40). Examination of the binding profile of Ca2⫹ or Ni2⫹ and Tb3⫹ or Gd3⫹ by Cys-less LacY-EF reveals that the purified protein 3⫹ binds Tb3⫹ with somewhat lower affinity [KTb ⬇ 4.5 ␮M at 25°C (Fig. 2)] than other proteins containing a natural EF-hand motif 3⫹ ⬍1 ␮M at 25°C (36)]. Competition assays using nonlumi[KTb nescent Gd3⫹, a trivalent cation with an ionic radius similar to Tb3⫹, demonstrate that there is a concentration-dependent 3⫹ decrease in Tb3⫹ luminescence by Gd3⫹ (Fig. 2 Inset) with KGd of ⬇2.3 ␮M. Although there is only a relatively small change in the affinity for Gd3⫹ with respect to Tb3⫹, the divalent cations Ca2⫹ and Ni2⫹ decrease Tb3⫹ luminescence to 50% at millimolar concentrations, thereby illustrating a lower affinity for divalent cations (data not shown). Thus, both the affinity and specificity of the EF-hand motif engineered into LacY are appropriate for LRET studies. Luminescence Properties of Bound Tb3ⴙ. The emission spectra of

Tb3⫹ bound to Cys-less LacY-EF in either DDM micelles or proteoliposomes shows characteristic narrow bands at 487 and 542 nm (Fig. 3). Tb3⫹ luminescence decay at both 487 and 542 nm fits two exponentials (Fig. 4 and Table 1). The overall decay is composed of two components, a population of 24% with a shorter lifetime of 0.3 ms and a population of 76% with a longer lifetime of 1.066 ms (Table 1). Given the low quantum yield and consequentially low efficiency of energy transfer, the shorter lifetime is not useful in LRET measurements. It is likely that the shorter lifetime represents a quenched excited state of the bound Tb3⫹, because Tb3⫹ chelates also exhibit a component with a similar short lifetime (18, 21). In contrast, the longer lifetime of 1.066 ms with a calculated quantum yield (QD) of 0.22 (Table 1) represents the meaningful lifetime for distance calculations Fig. 3. Luminescence emission spectrum of Tb3⫹ and absorption spectra of three fluorophors. LacY-EF兾V169C was purified, labeled individually with each of three different acceptor fluorophors, and reconstituted into proteoliposomes as described in Experimental Procedures. Emission of Tb3⫹ (broken lines) after excitation at 295 nm is compared with the absorption spectra of BODIPY (␧504 ⫽ 23 ⫻ 103 M⫺1䡠cm⫺1) (A) Fluorescein (␧492 ⫽ 76 ⫻ 103 M⫺1䡠cm⫺1) (B) or Alexa Fluor 594 (␧589 ⫽ 120 ⫻ 103 M⫺1䡠cm⫺1) (C) (solid lines). The overlap between Tb3⫹ emission and fluorophor absorption was used to calculate overlap integrals (J) (Table 2). The absorption spectra of each acceptor are normalized to the specific Tb3⫹ emission band that transfers the energy. 3490 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.052703599

Table 1. Tb3ⴙ lifetimes (␶D) and quantum yield (QD) in Cys-less LacY-EF Sample Micellar solution Proteoliposomes

␶D,1, ms

␶D,2, ms

QD*

0.26 (24%) 0.33 (24%)

1.062 (76%) 1.062 (76%)

0.22 0.22

*Calculated from the ␶D,2 decay according to Experimental Procedures. Va´ zquez-Ibar et al.

Table 2. Spectroscopic parameters and calculated distances from the EF-hand motif in LacY-EF to position 169 or 179 in helix VI V169C

BODIPY Fluorescein Alexa 594

I179C

␧␭max,* M⫺1䡠cm⫺1

J†

R0, ņ

␶DA, ms‡

E†

R, ņ

␶DA, ms‡

E†

R, ņ

23 ⫻ 103 76 ⫻ 103 120 ⫻ 103

9.3 ⫻ 1014 3.6 ⫻ 1015 3.8 ⫻ 1015

38 48 48

0.248 0.086 0.150

0.766 0.919 0.859

33 33 36

0.047 0.011 0.013

0.956 0.990 0.988

23 22 23

derived from LRET measurements. The population of the quenched state is higher in the EF-hand motif (24%) than that observed in lanthanide chelates (10%). Thus, the engineered EF-hand motif is likely more solvent exposed, the main factor contributing to quenching of the excited state (41). LRET Between the EF-Hand Motif and Helix VI. The main purpose of this study is to test a new approach to obtain static and dynamic information regarding the tertiary structure of LacY and possibly other proteins. To demonstrate feasibility, LRET experiments were performed with the aim of measuring distance between the EF-hand motif described and positions 169 and 179 in helix VI (Fig. 1). Therefore, the native residue at position 169 or 179 was replaced with Cys to generate LacY-EF兾V169C and LacY-EF兾I179C to allow conjugation of appropriate fluorophors. The efficiency of energy transfer (E) depends on both the distance between the donor (Tb3⫹) and the acceptor (BODIPY, ¨rster distance fluorescein, or Alexa Fluor 594) and R0, the Fo characteristic of a particular donor–acceptor pair. Experimentally, the efficiency of energy transfer can be determined by comparison of the emission lifetime of Tb3⫹ in the absence of acceptor with either the emission lifetime in the presence of acceptor or the sensitized lifetime of the acceptor (Eq. 4). The spectral properties of the three dyes used in these experiments are different. Specifically, J, the main component of Fo ¨rster distance, was calculated from the emission spectrum of Tb3⫹ and the excitation spectra of the dyes (Fig. 3), which have a strong dependence on the extinction coefficients of the acceptor at the wavelengths of donor emission (Table 2). Thus, the higher extinction coefficients (␧␭max) of fluorescein and Alexa Fluor 594 relative to BODIPY causes these dyes to have overlap integrals that are 10-fold higher than BODIPY. Conversely, BODIPY has a Fo ¨rster radius that is 10 Å lower (Table 2). The maximum emission peaks of all three acceptors (BODIPY, 510 nm; fluorescein, 518 nm; Alexa Fluor 594, 615 nm) are clearly situated between the sharp Tb3⫹ emission peaks at 487 and 542 nm (Fig. 5), thereby allowing accurate monitoring of the sensitized emission of these dyes, because there is no significant Tb3⫹ emission at these wavelengths. Therefore, a sensitized emission of the acceptor in a millisecond range as the result of Tb3⫹ excitation reflects the rate of Tb3⫹ relaxation because of energy transfer (34). This approach simplifies accurate calculations of energy transfer, because it is not necessary to analyze multiple decays of Tb3⫹ due to incomplete labeling with the acceptor fluorophor. Using this methodology, LacY-EF兾V169C and LacYEF兾I179C were studied unmodified or modified with a given acceptor fluorophor to estimate distance between the EF-hand motif and the fluorophor. After excitation, measurements were taken after a 50-␮s initial delay to avoid direct excitation of the acceptor. The sensitized decays of all three dyes at each position exhibit two lifetimes. The rapid decay of each fluorophore is the component relevant to energy transfer for the following reasons: Va´ zquez-Ibar et al.

(i) As expected from the Fo ¨rster distances (R0), the fast decay of fluorescein is more rapid than that of BODIPY at position 169, because energy transfer efficiency increases with R0 at a fixed distance (Table 2 and Fig. 6A). (ii) The fast decay of BODIPY emission decreases from 0.248 to 0.047 ms when the acceptor is closer to the EF-hand motif [position 179, which is closer to the Tb3⫹-binding site relative to position 169 (Fig. 6B)], whereas the second decay remains constant at ⬇0.5 ms. The second decay plays no role in the energy transfer process and may result from residual Tb3⫹ emission at this wavelength or detector background. The average calculated distances between the EF-hand motif and positions 169 or 179 are ⬇33 and ⬇23 Å, respectively (Table 2), which correlate well with the distances predicted for a slightly tilted ␣-helix (Fig. 1). The distances determined are based specifically on the distance between the donor and the acceptors. However, it is important to note that movement of the acceptor fluorophor relative to the donor during energy transfer introduces heterogeneity in transfer efficiency. Because the rate of energy transfer is inversely proportional to distance, the sensitized lifetime of the acceptor reflects the shortest distance between donor and acceptor (42). The strong agreement in the calculated distances using three acceptors with different Fo ¨rster radii argues convincingly for the ability of the technique to estimate intramolecular distances with reasonable accuracy. Furthermore, the excellent agreement confirms the random orientation of the acceptor probes during the energy transfer process, which supports the value given for the orientation factor (␬2 ⫽ 2兾3) used in the calculations. Significantly, this approximation cannot be made with conventional FRET techniques, because the dipole orientation of the

Fig. 5. Overlap between donor and acceptor emission spectra. Emission of Tb3⫹ (broken line) and steady-state emission spectra of BODIPY (␭max ⫽ 510 nm) (A), Fluorescein (␭max ⫽ 518 nm) (B), and Alexa Fluor 594 (␭max ⫽ 615 nm) (C) (solid lines) was measured with LacY-EF兾V169C reconstituted into proteoliposomes, as described in Experimental Procedures. PNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6 兩 3491

BIOCHEMISTRY

*␭max refers to the maximum absorption wavelength of each fluorophor (see Fig. 3). †The overlap integral ( J) and the Fo ¨ rster distances (R0) were calculated by using Eqs. 6 and 7 taking ␬2 ⫽ 2⁄3, ␩ ⫽ 1.4 and QD ⫽ 0.22. E and R were calculated by using Eqs. 4 and 5. ‡Lifetimes correspond to the faster decay of the sensitized emission of each acceptor, which is the relevant lifetime for energy transfer from Tb3⫹.

Fig. 6. Lifetime decays of the sensitized emissions of fluorophor acceptors. (A) Sensitized decay of BODIPY (1) or Fluorescein (2) at position 169 of LacY-EF兾V169C. (B) Sensitized decay of BODIPY in LacY-EF兾V169C or Cys-less LacY-EF兾I179C. Protein and Tb3⫹ concentrations were 5 and 10 ␮M, respectively. All sensitized acceptor decay curves were fit to a two-exponential decay. The sensitized lifetimes and calculated distances are given in Table 2.

fluorophor is different for each acceptor as shown by steadystate anisotropy measurements (data not shown). As judged by the sensitized lifetime values and the calculated energy transfer efficiency (E) (Table 2), BODIPY is the most sensitive acceptor for the range of distances measured here (23–33 Å). The efficiency of energy transfer to this dye shifts from 0.766 to 0.956 between positions 169 and 179, corresponding to a distance change of ⬇9 Å using the calculated Fo ¨rster distance (R0 ⫽ 38 Å). Hence, BODIPY should prove to be particularly useful in analysis of structural dynamics during the catalytic cycle of LacY.

motifs have also been introduced into other cytoplasmic, as well as periplasmic loops, without abolishing LacY function. Therefore, by LRET measurements from EF-hand motifs on either side of the membrane to BODIPY-labeled Cys residues in transmembrane helices, it should be possible to determine helix packing on both surfaces of the membrane and thus helical tilts, as well as changes induced by ligand binding. Finally, the affinity of the EF-hand motif for Gd3⫹ may be exploited for distance measurements by an independent spectroscopic technique, sitedirected spin labeling (see ref. 43).

Conclusion and Future Directions. Introduction of an EF-hand motif into LacY provides exciting new prospects for the study of static and dynamic aspects of structure. Clearly, LRET measurements yield distance information in the range of 20–50 Å and, although data have not been presented here, EF-hand

We express particular thanks to Miguel Garcia-Garibay and the members of his laboratory for the use of the phosphorimeter lifetime instrument and technical assistance. We are also indebted to Francisco Bezanilla, Joseph Falke, Christopher Miller, Miklo ´s Sahin-To ´th, and Ernest Wright for advice and helpful suggestions. Finally, we thank Gillian Verner for superb technical assistance. This work was supported in part by National Institutes of Health Grant DK51131:06.

1. Mu ¨ller-Hill, B. (1996) The lac Operon: A Short History of a Genetic Paradigm (de Gruyter, Berlin). 2. Kaback, H. R. (1976) J. Cell Physiol. 89, 575–593. 3. Kaback, H. R. (1983) J. Membr. Biol. 76, 95–112. 4. Kaback, H. R. (1989) Harvey Lectures 83, 77–103. 5. Poolman, B. & Konings, W. N. (1993) Biochim. Biophys. Acta 1183, 5–39. 6. Viitanen, P., Newman, M. J., Foster, D. L., Wilson, T. H. & Kaback, H. R. (1986) Methods Enzymol. 125, 429–452. 7. Sahin-To ´th, M., Lawrence, M. C. & Kaback, H. R. (1994) Proc. Natl. Acad. Sci. USA 91, 5421–5425. 8. Kaback, H. R. (1996) in Handbook of Biological Physics: Transport Processes in Eukaryotic and Prokaryotic Organisms, eds. Konings, W. N., Kaback, H. R. & Lolkema, J. S. (Elsevier, Amsterdam), Vol. II, pp. 203–227. 9. Kaback, H. R. & Wu, J. (1997) Quart. Rev. Biophys. 30, 333–364. 10. Frillingos, S., Sahin-To ´th, M., Wu, J. & Kaback, H. R. (1998) FASEB J. 12, 1281–1299. 11. Kaback, H. R. & Wu, J. (1999) Acc. Chem. Res. 32, 805–813. 12. Kaback, H. R., Sahin-To ´th, M. & Weinglass, A. B. (2001) Nat. Rev. Mol. Cell. Biol. 2, 610–622. 13. Selvin, P. R. (1995) Methods Enzymol. 246, 300–334. 14. Lakowicz, J. R. (1999) Principles of Fluorescence Spectroscopy (Kluwer Academic兾Plenum, New York). 15. Selvin, P. R. (2000) Nat. Struct. Biol. 7, 730–734. 16. Cha, A., Snyder, G. E., Selvin, P. R. & Bezanilla, F. (1999) Nature (London) 402, 809–813. 17. Xiao, M., Li, H., Snyder, G. E., Cooke, R., Yount, R. G. & Selvin, P. R. (1998) Proc. Natl. Acad. Sci. USA 95, 15309–15314. 18. Getz, E. B., Cooke, R. & Selvin, P. R. (1998) Biophys. J. 74, 2451–2458. 19. Heyduk, T. (2001) Methods 25, 44–53. 20. Root, D. D. (1997) Proc. Natl. Acad. Sci. USA 94, 5685–5690. 21. Chen, J. & Selvin, P. R. (1999) Bioconjugate Chem. 10, 311–315. 22. Brittain, H. G., Richardson, F. S. & Martin, R. B. (1976) J. Am. Chem. Soc. 98, 8255–8260.

23. Horrocks, W. D. & Albin, M. (1984) Prog. Inorg. Chem. 31, 1–104. 24. Hogue, C. W., MacManus, J. P., Banville, D. & Szabo, A. G. (1992) J. Biol. Chem. 267, 13340–13347. 25. Falke, J. J., Drake, S. K., Hazard, A. L. & Peterson, O. B. (1994) Q. Rev. Biophys. 27, 219–290. 26. Kawasaki, H. & Kretsinger, R. H. (1994) Protein Profiles 1, 343–517. 27. Weinglass, A. B. & Kaback, H. R. (2000) Proc. Natl. Acad. Sci. USA 97, 8938–8943. 28. Jung, K., Jung, H., Colacurcio, P. & Kaback, H. R. (1995) Biochemistry 34, 1030–1039. 29. Sanger, F., Nicklen, S. & Coulsen, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463–5467. 30. le Coutre, J., Narasimhan, L. R., Patel, C. K. & Kaback, H. R. (1997) Proc. Natl. Acad. Sci. USA 94, 10167–10171. 31. Carrasco, N., Herzlinger, D., Dechiara, S., Danho, W. & Kaback, H. R. (1985) Ann. NY Acad. Sci. 456, 305–306. 32. Jung, H., Rubenhagen, R., Tebbe, S., Leifker, K., Tholema, N., Quick, M. & Schmid, R. (1998) J. Biol. Chem. 273, 26400–26407. 33. Cheng, Y. & Prusoff, W. H. (1973) Biochem. Pharmacol. 22, 3099–3108. 34. Selvin, P. R. & Hearst, J. E. (1994) Proc. Natl. Acad. Sci. USA 91, 10024–10028. 35. Root, D. D., Shangguan, X., Xu, J. & McAllister, M. A. (1999) J. Struct. Biol. 127, 22–34. 36. Drake, S. K. & Falke, J. J. (1996) Biochemistry 35, 1753–1760. 37. Drake, S. K., Zimmer, M. A., Kundrot, C. & Falke, J. J. (1997) J. Gen. Physiol. 110, 173–184. 38. Blumenschein, T. M. A. & Reinach, F. C. (2000) Biochemistry 39, 3603–3610. 39. Clark, I. D., MacManus, J. P., Banville, D. & Szabo, A. G. (1993) Anal. Biochem. 210, 1–6. 40. Schlyer, B. D., Steel, D. G. & Gafni, A. (1995) J. Biol. Chem. 270, 22890–22894. 41. Xiao, M. & Selvin, P. R. (2001) J. Am. Chem. Soc. 123, 7067–7073. 42. Heyduk, T. & Heyduk, E. (2001) Anal. Biochem. 289, 60–67. 43. Voss, J., Wu, J., Hubbell, W. L., Jacques, V., Meares, C. F. & Kaback, H. R. (2001) Biochemistry 40, 3184–3188.

3492 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.052703599

Va´ zquez-Ibar et al.