Vrije Universiteit Brussel

Structure of a prokaryotic fumarate transporter reveals the architecture of the SLC26 familyGeertsma, Eric R; Chang, Yung-Ning; Shaik, Farooque R; Neldner, Yvonne; Pardon, Els;Steyaert, Jan; Dutzler, RaimundPublished in:Nature Structural and Molecular Biology

DOI:10.1038/nsmb.3091

Publication date:2015

License:Unspecified

Document Version:Final published version

Link to publication

Citation for published version (APA):Geertsma, E. R., Chang, Y-N., Shaik, F. R., Neldner, Y., Pardon, E., Steyaert, J., & Dutzler, R. (2015). Structureof a prokaryotic fumarate transporter reveals the architecture of the SLC26 family. Nature Structural andMolecular Biology, 22(10), 803-808. https://doi.org/10.1038/nsmb.3091

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portalTake down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Download date: 24. Nov. 2020

nature structural & molecular biology  VOLUME 22 NUMBER 10 OCTOBER 2015 803

a r t i c l e s

The SLC26 or SulP proteins constitute a large family of anion trans-porters that are ubiquitously expressed in pro- and eukaryotes1–3. In humans, this family encompasses at least ten members whose malfunctioning is associated with severe diseases such as growth defects, chronic diarrhea and deafness1. Eukaryotic family members function as either coupled exchangers or channels4. The transported substrates include divalent anions, such as sulfate and oxalate, and the monovalent anions bicarbonate, formate, chloride and iodide1,4. The human protein Prestin is a particular SLC26 family member that works as a motor protein in cochlear outer hair cells. Variations in the membrane potential due to incoming sound waves, conferred to Prestin by an intracellularly bound anion, cause conformational changes of the protein that lead to rapid alterations in cell length and thereby result in the amplification of sound5,6. Characterized prokaryotic SLC26 proteins transport either bicarbonate3,7 or dicar-boxylic acids8 in a Na+- or H+-dependent manner. Despite their functional differences, all family members share a conserved struc-tural organization that consists of a transmembrane component followed by a cytoplasmic sulfate transporter and antisigma factor antagonist (STAS) domain1,9. Although structures of isolated STAS domains have been determined by X-ray crystallography and NMR spectroscopy10–12, the structure of a full-length SLC26 transporter has been elusive. To close this gap in the understanding of the fam-ily, we investigated the structural and functional properties of a prokaryotic SLC26 homolog, termed SLC26Dg, from the bacterium D. geothermalis. Transport experiments demonstrated that SLC26Dg functions as a proton-coupled fumarate symporter. Its crystal structure, determined at 3.2-Å resolution, reveals an inward facing,

substrate-free conformation of the transporter. The structure provides a common framework for the SLC26 family that underlies the diverse functional behavior among its members.

RESULTSFunctional characterization of SLC26DgTo identify proteins compatible with structural characterization by X-ray crystallography, we cloned 92 prokaryotic SLC26 homologs by the fragment exchange (FX) cloning method13 and subsequently selected well-expressed GFP-fused proteins that were stable in mild nonionic detergents14,15 (Supplementary Table 1). This broad screen allowed us to single out SLC26Dg on the basis of its superior biochemical properties.

SLC26Dg shares 46% residue similarity in the transmembrane (TM) region with human Prestin and 57% with the Escherichia coli transporter DauA (previously called YchM) (Supplementary Fig. 1). When purified in maltoside detergents, SLC26Dg was monomeric (Supplementary Fig. 2a,b), but within the lipid bilayer we observed dimers whose organization was consistent with that in other family members16,17 (Supplementary Fig. 2c). To characterize its transport properties, we reconstituted the purified transporter into liposomes and investigated the uptake of radiolabeled dicarboxylates, moti-vated by the observation that DauA transports succinate and other C4-dicarboxylates8. Our experiments showed a strong time-dependent accumulation of fumarate inside the proteoliposomes, thus identi-fying this important metabolic intermediate as candidate substrate in a biological context (Fig. 1a). The high substrate selectivity was manifested in the poor inhibition of fumarate uptake by structurally

1Department of Biochemistry, University of Zurich, Zurich, Switzerland. 2Institute of Biochemistry, Biocenter, Goethe-University Frankfurt, Frankfurt am Main, Germany. 3Structural Biology Research Center, Vlaams Instituut voor Biotechnologie, Brussels, Belgium. 4Structural Biology Brussels, Vrije Universiteit Brussel, Brussels, Belgium. 5Present address: Paul Scherrer Institute, Villigen, Switzerland. 6These authors contributed equally to this work. Correspondence should be addressed to R.D. (dutzler@bioc.uzh.ch) or E.R.G. (geertsma@em.uni-frankfurt.de).

Received 10 April; accepted 20 August; published online 14 September 2015; doi:10.1038/nsmb.3091

Structure of a prokaryotic fumarate transporter reveals the architecture of the SLC26 familyEric R Geertsma1,2, Yung-Ning Chang1,2,6, Farooque R Shaik1,5,6, Yvonne Neldner1, Els Pardon3,4, Jan Steyaert3,4 & Raimund Dutzler1

The SLC26 family of membrane proteins combines a variety of functions within a conserved molecular scaffold. Its members, besides coupled anion transporters and channels, include the motor protein Prestin, which confers electromotility to cochlear outer hair cells. To gain insight into the architecture of this protein family, we characterized the structure and function of SLC26Dg, a facilitator of proton-coupled fumarate symport, from the bacterium Deinococcus geothermalis. Its modular structure combines a transmembrane unit and a cytoplasmic STAS domain. The membrane-inserted domain consists of two intertwined inverted repeats of seven transmembrane segments each and resembles the fold of the unrelated transporter UraA. It shows an inward-facing, ligand-free conformation with a potential substrate-binding site at the interface between two helix termini at the center of the membrane. This structure defines the common framework for the diverse functional behavior of the SLC26 family.

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

804  VOLUME 22 NUMBER 10 OCTOBER 2015 nature structural & molecular biology

a r t i c l e s

related compounds (Fig. 1b and Supplementary Fig. 2d). Transport was strongly enhanced by an inwardly directed pH gradient but not by a Na+ gradient (Fig. 1a). We directly observed coupled transport of fumarate and protons by SLC26Dg by means of fumarate-dependent acidifica-tion of the lumen of proteoliposomes (Fig. 1c) and investigated the electrogenicity of the process by varying the membrane potential (Fig. 1d). Fumarate/proton symport was strongly increased by a positive and decreased by a negative membrane potential, thus indi-cating that the cumulative charge of the substrates per cycle was negative. Together, these results are consistent with a mechanism whereby fumarate transport by SLC26Dg is electrogenic and is coupled to the symport of protons. Whereas under the investigated conditions (pHoutside = 6.0) 97% of the fuma-rate was present in its dianionic state, it is currently not known whether the substrate is transported in its singly protonated mono-carboxylate form or as dicarboxylate together with a proton that is bound to a different site in the protein. The concentration depend-ence of fumarate uptake showed saturation that reached its half-maximum value at 2.7 mM, which is indicative of a weak affinity for the substrate (Fig. 1e and Supplementary Fig. 2e). Because of the mixed orientation of SLC26Dg in proteoliposomes and the result-ing difference in substrate access, we refer to this value as apparent Km. We also expressed the isolated TM domain of the transporter (residues 1–401) separately. Although this con-struct (SLC26Dg∆STAS) was stable, its transport

properties were strongly compromised, despite the similar mem-brane reconstitution efficiency and mixed orientation to those of WT (Supplementary Fig. 2f,g). It thus appears that the STAS domain

5

1001.0 * # 10

8

6

4

2

0

25

20

15

10

5

00 10 20 30 40

0.8

0.6

0.4

0 10 20 30 40

80

60

4

3

Tran

spor

t (nm

ol/m

g pr

otei

n)

Tran

spor

t (nm

ol/m

g pr

otei

n)A

ctiv

ity (

µmol

mg–1

min

–1)

Upt

ake

(%)

No

add.

Fum

arat

e Mal

eate

Suc

cina

te

Flu

ores

cenc

e ra

tio

2

1

00 2 4 6 8

No protein∆Na+ pH 6.0∆Na+ pH 7.5∆Na+ ∆pH∆pH

1 mM fumarate5 mM fumarate10 mM fumarate

No protein

K+out Na+

in val

K+out Na+

in

Na+out K

+in

Na+out K

+in val

Time (min)

0 2 4 6 8

Time (min)

Concentration (mM)

Time (min)

a b c d

eFigure 1 Transport properties of SLC26Dg. (a) Time-dependent uptake of [14C]fumarate (50 µM) into proteoliposomes containing SLC26Dg. ∆pH refers to a pH of 6.0 outside and 7.5 inside; ∆Na+ refers to 50 mM Na+ outside and 0 mM inside the liposome. ‘No protein’ refers to empty liposomes assayed in the presence of ∆Na+ and ∆pH. (b) Competition of [14C]fumarate uptake by a 70-fold excess of unlabeled dicarboxylates. ‘No add.’ refers to presence of 150 µM [14C]fumarate only. (c) Fumarate-dependent proton transport into the lumen of proteoliposomes, as measured with the pH-sensitive fluorophore pyranine. Asterisk indicates addition of fumarate; pound sign indicates addition of nigericin. Individual traces of duplicate experiments are shown. (d) Dependence of 14C-fumarate (50 µM) uptake on the membrane potential. ‘K+

out Na+in val’ refers to 50 mM of K+ outside

and 50 mM Na+ inside the liposome, which generates a positive membrane potential upon addition of valinomycin. This ratio is inverted in ‘Na+

out K+in val’. ‘K+

out Na+in’ and ‘Na+

out K+in’ refer to conditions in

which no valinomycin was added. Data in a, b and d represent mean and s.e.m. of 3 technical replicates. (e) Concentration dependence of fumarate uptake. Data were corrected for the passive diffusion of the substrate. The solid line shows the fit to a Michaelis-Menten equation with an apparent Km of 2.7 mM. Data represent mean and s.e.m. of 6 technical replicates from two independent batches of proteoliposomes. Errors are scaled to their respective values in the uncorrected curves.

Table 1 Data collection and refinement statistics

SLC26Dg–NB SLC26Dg–NB SeMet SLC26Dg∆STAS Nb5776

Data collection

Space group C2 C2 P1 P4322

Cell dimensions

a, b, c (Å) 127,7, 117.3, 84.4 127.6, 116.4, 82.3 80.6, 82.5, 166.4 139.5, 139.5, 196.9

α, β, γ (°) 90, 100.1, 90 90, 100.2, 90 96.3, 92.9, 119.2 90, 90, 90

Resolution (Å) 50–3.2 (3.3–3.2)a 50–3.7 (3.8–3.7) 50–4.2 (4.3–4.2) 50–2.4 (2.5–2.4)

Rmerge 5.3 (120.0) 19.5 (199.8) 4.0 (64.3) 10.1 (135.7)

I / σI 18.7 (2.0) 19.1 (2.1) 13.0 (1.8) 23.5 (2.2)

Completeness (%) 99.6 (99.6) 99.9 (99.7) 97.4 (97.8) 99.9 (99.7)

Redundancy 6.7 (6.6) 17.2 (16.7) 3.6 (3.6) 14.9 (14.7)

Refinement

Resolution (Å) 20–3.2 30–4.2 20–2.4

No. reflections 20,066 26,006 76,232

Rwork / Rfree 25.8 (30.8) 32.9 (32.6) 23.3 (24.2)

No. atoms 4,412 11,000 3,057

Protein 4,346 11,000 2,772

Detergent 66 – –

Water – – 285

B factors

Protein 144 212 57

Detergent 135 – –

Water – – 58

r.m.s. deviations

Bond lengths (Å) 0.003 0.005 0.002

Bond angles (°) 0.74 1.06 0.58

Slc26Dg–NB refers to the full-length transporter–Nb5776 complex, SLC26Dg∆STAS to the SLC26Dg TM domain and Nb5776 to the isolated nanobody. SeMet, selenomethionine.aValues in parentheses are for highest-resolution shell.

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature structural & molecular biology  VOLUME 22 NUMBER 10 OCTOBER 2015 805

a r t i c l e s

of SLC26Dg regulates the transport activity of the TM domain, as has been proposed for other family members9,18,19.

SLC26Dg structureTo gain insight into the architecture of the SLC26 family, we crystallized full-length and truncated forms of SLC26Dg. The full-length transporter yielded crystals diffracting to 7-Å resolution that could not be further improved. The truncated construct SLC26Dg∆STAS read-ily crystallized in several conditions, with the best crystal form of space group P1 diffracting up to 4.2 Å (Table 1). Only after generating nanobodies against SLC26Dg and using them as crystallization aids20 could we obtain improved crystals diffracting to 3.2-Å resolution. These crystals of full-length, monomeric SLC26Dg in complex with the nanobody Nb5776, are of space group C2, con-tain a single copy of the complex in their asymmetric unit and do not show any two-fold relationship with neighboring molecules that would reveal the dimeric organization observed in the membrane. We determined the structure of the SLC26Dg–Nb5776 complex with phases obtained by single anomalous dispersion (SAD) from crystals of a selenomethionine-derivatized transporter (Table 1 and Supplementary Fig. 3). In this structure, the membrane and STAS domains of SLC26Dg exhibit only a few noncovalent interactions. Although we expect a contribution of the crystalline environment to the observed relative arrangement of the two protein compo-nents, the arrangement is in line with a low-resolution envelope of a

related prokaryotic SLC26 transporter, described previously, which suggests a peripheral location of this cytoplasmic component within each subunit of a detergent-solubilized dimeric transporter17,21 (Fig. 2a). The STAS domain of SLC26Dg is compact and contains a core consisting of only four β-strands and three α-helices (Fig. 2b,c). Its structure is very similar to the same domain of the SLC26 trans-porter of Rhodobacter sphaeroides (Supplementary Fig. 4b). It also aligns well with equivalent regions of the STAS domains of Prestin12 and DauA11, which both contain extended structured elements at the C terminus that, in the case of DauA, have been found to interact with acyl carrier protein (ACP, Supplementary Fig. 4). Within the com-plex, the nanobody interacts exclusively with the cytoplasmic domain of SLC26Dg. It binds the STAS domain mainly via complementarity determining region 2, thereby burying 1,505 Å2 of the combined molecular surface (Supplementary Fig. 5a–e). In the crystals, the Nb5776–STAS domain complex adopts an orientation with respect to the membrane domain that would place it within the inner leaflet of the phospholipid bilayer (Supplementary Fig. 5f,g). This position is unlikely to represent a native orientation of the hydrophilic STAS domain and may result from the tight interactions with Nb5776 and their involvement in extensive crystal contacts (burying an interface of 1,451 Å2). The disturbed orientation of the STAS domain may thus be a consequence of the inherent flexibility of the linker joining both components of the protein.

Cβ2

Cβ3

Cβ1

Cβ4

Cα2Cα1

Cα3

C

α7

α6

α5

α4

α1

α

α

2

3

α10

α9

Cα3

Cβ4

Cα2

Cβ3

Cα1

Cβ2

Cβ1

C

α8

α11

α12

7

6 12

2

10

59

3 4

8

13

14

111

11′

4′

11′

4′

N

α13

α14

a b

c d

Figure 2 SLC26Dg structure. (a) Ribbon representation of the SLC26Dg–Nb5776 complex viewed from within the membrane. The N- and C-terminal halves of the TM domain are green and beige, respectively. The STAS domain is red, and the nanobody Nb5776 is blue. (b) Ribbon representation of the STAS domain with secondary-structure elements labeled. (c) Topology of SLC26Dg. (d) TM domain with α-helices shown as cylinders and labeled. The N-terminal half of the domain is green, and the C-terminal half is beige. Figures 2–6 were prepared with DINO (http://www.dino3d.org/).

90°

90°

α3

α6

α6

α6

α14

α8

α8

α8α14

α14α3

α10

N

C

C

C

11

1

712

6 5

10

29 4

8

3

1413

a b

c dFigure 3 TM domain and intracellular cavity. (a) View of the TM domain from the extracellular side, colored as in Figure 2d. (b) Organization of the TM helices. The view is as in a. Helices are shown as cylinders and labeled. The core domain of the protein is red-brown, and the gate domain is blue. (c) View of the TM domain from within the membrane. (d) View of the TM domain from the cytoplasm. The mesh in panels a, c and d shows the molecular surface of the intracellular cavity leading to the substrate-binding site.

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

806  VOLUME 22 NUMBER 10 OCTOBER 2015 nature structural & molecular biology

a r t i c l e s

Transmembrane domainThe TM domain of SLC26Dg consists of 14 α-helices of variable length, including several short helices that do not span the entire width of the lipid bilayer (Fig. 2c,d). It is organized into two struc-turally related halves of seven TM segments each. These halves are oriented in opposite directions, thereby forming an intertwined struc-ture of two ‘inverted repeats’ analogous to the general construction principle of inverted repeats observed for many transport proteins22 (Supplementary Fig. 6a,b). In the membrane plane, the dimensions are approximately 45 by 60 Å (Fig. 3a). The height of only 40 Å sug-gests that the TM region does not extend beyond the boundary of the phospholipid bilayer (Supplementary Fig. 6c). We obtained a similar view of the TM domain in the structure of SLC26Dg∆STAS determined at 4.2-Å resolution (Table 1). In its crystal form, the trun-cated protein adopts an equivalent conformation to that found in the

SLC26Dg–Nb5776 complex, thus suggesting that the crystal packing did not compromise the structural integrity of the membrane-inserted region. (Table 1 and Supplementary Fig. 6d,e). The organization of the TM domain resembles that of UraA, a member of the nucleobase cation symporter-2 (NCS2) family (also known as the nucleobase/ascorbate transporter family)23, which does not share any obvious sequence relationship. This resemblance is in agreement with recent predictions of structural conservation between the SLC26, NCS2 and SLC4 families, according to advanced remote homology detection methods24–28. Similarly to the organization of the UraA structure, a layered organization of the transmembrane part of SLC26Dg in two parallel subdomains is apparent (Fig. 3a,b). The pseudosymmetry-related helices 1–4 and 8–11 fold into a compact unit designated as the core domain. Helices 5–7 and 12–14, termed the gate domain, form an elongated structure that shields one side of the core domain.

N

N

C

CC

α3

α10

α10 α10

α8

α1 C

a11

11 2

3

88

9 42

17

7

10

12 12 5

5

6

6

14

13

13

b

d ec

Figure 4 Comparison of SLC26Dg and UraA. (a) Stereo view of a superposition of the SLC26Dg TM domain (green and beige) and the uracil transporter UraA23 (PDB 3QE7 (ref. 23) in blue and gray). Both proteins are shown as ribbons. (b) View of a superposition of SLC26Dg and UraA from the extracellular side. The TM helices of both proteins are represented as cylinders and numbered. The color coding is as in a. The boundary between the core (top) and gate domain (bottom) is indicated by a gray line. (c) Ligand-binding site in UraA. The protein is shown as Cα trace, and interacting side chains and uracil (brown) are shown as sticks. Aqueous cavity leading to the substrate-binding site of UraA (d) and SLC26Dg (e). The colors are as in a. The view is from the cytoplasm. The molecular surfaces are gray, and substrates are displayed as a CPK model.

α1

α8

α10

aSLC26A1

α1 α8 α10

SLC26A2SLC26A3SLC26A4SLC26A5SLC26A6SLC26A7SLC26A8SLC26A9DauABICASLC26WsSLC26AbSLC26AtSLC26CsSLC26Dg

b

α3

α10

cα8

α3α1

α10

E241E38

d 200

150

100

WT

E38A E241A

50

Act

ivity

(µm

ol m

g–1 m

in–1

)

0

Concentration (mM)

0 5 10 15 20

e

Figure 5 Potential substrate interactions. (a) View of the substrate-binding cavity from the cytoplasm. The protein is shown as ribbon, and side chains of polar residues in the cavity are shown as sticks. The N- and C-terminal halves of the protein are green and beige, respectively. The N termini of the two helices potentially involved in substrate interactions are blue. (b) Sequence alignment of selected regions facing the aqueous cavity. The alignment includes nine human (SLC26A1 (GI 209572674); SLC26A2 (GI 37590805); SLC26A3 (GI 19343676); SLC26A4 (GI 119603820); SLC26A5 (Prestin, GI 30348882); SLC26A6 (GI 94721253); SLC26A7 (GI 63102255); SLC26A8 (GI 119624268); and SLC26A9 (GI 153217499)) and seven prokaryotic transporters (DauA (GI 115512548); BicA (GI 109820126); SLC26Ws (GI 499451118); SLC26Ab (GI 110647915); SLC26At (GI 499274043); SLC26Cs (GI 91797972); and SLC26Dg (GI 499845065)). (c) View of the putative fumarate-binding site. Fumarate (shown as CPK model, brown) was modeled into the site on the basis of steric and electrostatic considerations. The molecular surface of the protein in the vicinity of the substrate is shown in gray. (d) Position of acidic residues with respect to the putative substrate-binding site. (e) Concentration dependence of fumarate uptake in the double mutant E38A E241A. The data were corrected for the passive diffusion of fumarate observed in empty liposomes. The solid line shows the fit to a Michaelis-Menten equation with an apparent Km of 1.3 mM. The kinetics of wild type (WT) is shown for comparison (dashed blue line). Data represent mean and s.e.m. of 3 technical replicates. Errors were scaled to their respective values in the uncorrected curves.

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature structural & molecular biology  VOLUME 22 NUMBER 10 OCTOBER 2015 807

a r t i c l e s

UraA, unlike SLC26Dg, contains a bound ligand, yet both structures align with an r.m.s. deviation of only 3.5 Å, thus indicating that they may represent nearby states in the transport cycle (Fig. 4).

The SLC26Dg structure shows an inward-facing conformation. As viewed from the cytoplasm, a large, 28-Å-long and 8-Å-wide aqueous cavity, located at the interface between the core and the gate domain, enables access to a putative substrate-binding site in the center of the transporter (Fig. 3c,d). Owing to the peripheral location of the helix pairs 6–7 and 13–14 plus the outward movement of helices 5 and 12 in the gate domain of SLC26Dg, this cavity is considerably larger than in UraA (Fig. 4). In its interior, it is predominantly lined by hydrophobic residues, with few exceptions (Fig. 5a,b). At the top of the cavity, the core domain contributes a pair of pseudosymmetry-related helices (i.e., α-helices 3 and 10) that point from opposite directions toward the center of the bilayer (Fig. 5c,d). The unpaired backbone amides at the N termini of these helices are potential constituents of a selec-tive substrate-binding site29. Within the cavity, they provide the only hydrogen-bond donors for interaction with the substrate, and they are located at an appropriate distance to bind the two carboxylates of fumarate (Fig. 5c). A similar binding mode, although with additional involvement of side chain interactions, has been observed in the trans-porter UraA23 (Fig. 4c). We attempted to investigate the interaction of fumarate with SLC26Dg by X-ray crystallography, but we observed no structural changes in the transporter in the presence of substrate, nor did we find excess electron density that could be attributed to the bound substrate. These results are in line with the comparably high apparent Km for fumarate and suggest that the protein-substrate interactions in the observed conformation of the transporter may be weak. A similar weak affinity for chloride and bicarbonate anions has also been described for Prestin5.

Among the few hydrophilic amino acids in the intracellular cavity, two pseudosymmetry-related glutamates (E38 and E241) are note-worthy. These residues lie adjacent to the potential ligand-binding site, with their carboxylate moieties oriented toward the cytoplasm (Fig. 5d). The first glutamate (E38) is conserved in some, and the second (E241) in most, prokaryotic SLC26 transporters. However, neither residue is found in the eukaryotic counterparts, which appear not to function as symporters (Fig. 5b). Because of their negative charges, we suspected that these residues may decrease the affinity of fumarate in the inward-facing state and thus facilitate substrate release into the cytoplasm and that they may be involved in proton cotransport by either serving as

acceptors or perturbing the pKa of the substrate. To clarify this hypoth-esis, we reconstituted the double mutant E38A E241A, in which both carboxylate moieties were removed. Interestingly, this mutant exhib-ited high fumarate-uptake activity, which was further enhanced by an inwardly directed pH gradient (Supplementary Fig. 6f). The con-centration dependence of transport showed a small decrease in Km but a strong seven-fold increase in Vmax (Fig. 5e and Supplementary Fig. 6g). These results suggest that the mutations did not prevent pro-ton coupling but instead alleviated an unknown rate-limiting step in substrate translocation. Although the proximity of these protonatable groups to the potential substrate-binding site affects the transport behavior of SLC26Dg, the role of these groups in transport in a physio-logical context remains to be elucidated.

DISCUSSIONOwing to the high degree of conservation, the SLC26Dg structure pro-vides a representative scaffold for the entire SLC26 family. Its architec-ture confirms the structural relationship with the NCS2 family and is thus in general agreement with results from a recent study on Prestin, which used homology modeling and functional experiments26. The structure reveals the location of a potential binding site for negatively charged substrates at the center of the transporter. The binding of anions between the N termini of α-helices has previously been observed in the structurally unrelated ClC family30,31. Although it can be assumed that this binding mode is conserved within the SLC26 family, it will be important to determine how the distinct structures of different members are able to coordinate substrates of varying size and negative net charge. It is thus conceivable that, in certain cases, side chain interactions with a positively charged residue on α-helix 10, which is conserved in most eukaryotic transporters (Fig. 5b), may contribute to anion binding, as has been suggested in the case of Prestin26.

The SLC26Dg structure shows a substrate-free inward-facing conformation with a binding site that is accessible from the intra-cellular side. Because UraA contains a bound uracil, differences between the two structures may illustrate movements after substrate binding. The transition to an outward conformation could involve a change in the relative arrangement between core and gate domains that would expose the substrate-binding site to the opposite side of the membrane, as previously proposed for UraA23. Rigid-body movements of the core and gate domain that have been observed in a recent study of UraA by molecular dynamics simulations support this

Inward OutwardOccluded

Gate domain Core domain

180°

Low High

a

b

Figure 6 Potential transport mechanism. (a) Sequence conservation mapped on the surface of the SLC26Dg membrane domain35. Conservation is plotted as a gradient with turquoise indicating the least and maroon the most conserved residues. Views of SLC26Dg from within the membrane on the gate domain (left) and the core domain (right). The calculation used amino acid sequences of 150 prokaryotic homologs with sequence identities to SLC26Dg ranging between 35% and 95%. (b) Schematic depiction (top) and model (bottom) of conformational changes underlying transport by the alternate-access mechanism in SLC26 transporters. The view is from within the membrane. Subunits of a hypothetical dimer interact via their respective gate domains (blue). The core domains (red) move, thereby changing the access of the substrate-binding site to either side of the membrane. A substrate is displayed as green spheres. The change of the access of the bound ion from an inward-facing toward an outward-facing conformation also changes its position in the transmembrane electric field, thus potentially explaining the voltage dependence of Prestin. An incomplete transition to an occluded state in this case may prevent release of the ion to the outside. Conformational changes could alter the cross-section of the protein within the densely packed membrane of cochlear outer hair cells and thus may underlie the observed change in cell length. It is currently not clear whether a stable occluded conformation is adopted by other SLC26 transporters.

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

808  VOLUME 22 NUMBER 10 OCTOBER 2015 nature structural & molecular biology

general mechanism32. The role of the dimeric organization of SLC26 transporters in the conformational changes required for transport is currently not clear. Because of the high sequence conservation of resi-dues on its surface (Fig. 6a), we suspect that the gate domain forms the interaction interface of the SLC26 dimer. In this case, transport may proceed in a similar way to that observed in dimeric Na+/H+ exchangers, in which a core domain that is responsible for the bulk of the interactions with the substrates oscillates around a stable gate domain that forms the dimer interface33,34 (Fig. 6b).

The observed architecture suggests that even in family members designated as channels (i.e., SLC26A7 and SLC26A9), ion transport may proceed by an alternate-access mechanism. Rapid conforma-tional changes that support high channel-like translocation rates are in agreement with the frequencies up to 20 kHz at which Prestin has been shown to operate6. Finally, the conformational change after sub-strate binding would also account for the voltage-dependent motor activity of Prestin, in which an incomplete transition may result in a decreased cross-sectional area of the protein within the membrane and prevent substrate release to the outside.

METhODSMethods and any associated references are available in the online version of the paper.

Accession codes. Coordinates and structure factors have been deposited in the Protein Data Bank under accession codes 5DA0 (SLC26Dg–Nb5776 complex) and 5DA4 (Nb5776).

Note: Any Supplementary Information and Source Data files are available in the online version of the paper.

ACkNowlEDGmENtSThis research was supported by a grant of the Swiss National Science Foundation through the National Center of Competence in Research (NCCR) Structural Biology (R.D.), a long-term fellowship from the Human Frontier Science Program (LT-00899/2008) (E.R.G.), the German Research Foundation through the Cluster of Excellence Frankfurt ‘Macromolecular Complexes’ (E.R.G.), INSTRUCT as part of the European Strategy Forum on Research Infrastructures (ESFRI) (J.S.) and the Hercules Foundation Flanders (J.S.). We thank B. Fakler and R.J.C. Hilf for input at initial stages of the project; the staff of the X06SA beamline for support during data collection; B. Blattman and C. Stutz-Ducommun (Protein Crystallization Center at University of Zurich) for their support with crystallization; B. Dreier for help with MALS experiments; and N. Buys and K. Willibal for help with nanobody selection. All members of the Dutzler laboratory are acknowledged for help in all stages of the project.

AUtHoR CoNtRIBUtIoNSF.R.S. and E.R.G. screened homologs, established purification of SLC26Dg and crystallized SLC26Dg∆STAS. E.P. performed immunization, cloned and expressed nanobodies, and guided E.R.G. during the initial selections. J.S. supervised nanobody production. E.R.G. and Y.N. established nanobody overexpression and purification for crystallization. Y.-N.C. and E.R.G. crystallized the SLC26Dg–nanobody complex and carried out crystallographic experiments and transport assays. R.D. assisted E.R.G. in structure determination. E.R.G. and R.D. jointly planned the experiments, analyzed the data and wrote the manuscript.

ComPEtING FINANCIAl INtEREStSThe authors declare no competing financial interests.

Reprints and permissions information is available online at http://www.nature.com/reprints/index.html.

1. Alper, S.L. & Sharma, A.K. The SLC26 gene family of anion transporters and channels. Mol. Aspects Med. 34, 494–515 (2013).

2. Dorwart, M.R., Shcheynikov, N., Yang, D. & Muallem, S. The solute carrier 26 family of proteins in epithelial ion transport. Physiology (Bethesda) 23, 104–114 (2008).

3. Price, G.D. & Howitt, S.M. The cyanobacterial bicarbonate transporter BicA: its physiological role and the implications of structural similarities with human SLC26 transporters. Biochem. Cell Biol. 89, 178–188 (2011).

4. Ohana, E., Yang, D., Shcheynikov, N. & Muallem, S. Diverse transport modes by the solute carrier 26 family of anion transporters. J. Physiol. (Lond.) 587, 2179–2185 (2009).

5. Oliver, D. et al. Intracellular anions as the voltage sensor of prestin, the outer hair cell motor protein. Science 292, 2340–2343 (2001).

6. Dallos, P. & Fakler, B. Prestin, a new type of motor protein. Nat. Rev. Mol. Cell Biol. 3, 104–111 (2002).

7. Price, G.D., Woodger, F.J., Badger, M.R., Howitt, S.M. & Tucker, L. Identification of a SulP-type bicarbonate transporter in marine cyanobacteria. Proc. Natl. Acad. Sci. USA 101, 18228–18233 (2004).

8. Karinou, E., Compton, E.L., Morel, M. & Javelle, A. The Escherichia coli SLC26 homologue YchM (DauA) is a C4-dicarboxylic acid transporter. Mol. Microbiol. 87, 623–640 (2013).

9. Sharma, A.K., Rigby, A.C. & Alper, S.L. STAS domain structure and function. Cell. Physiol. Biochem. 28, 407–422 (2011).

10. Sharma, A.K. et al. Solution structure of the guanine nucleotide-binding STAS domain of SLC26-related SulP protein Rv1739c from Mycobacterium tuberculosis. J. Biol. Chem. 286, 8534–8544 (2011).

11. Babu, M. et al. Structure of a SLC26 anion transporter STAS domain in complex with acyl carrier protein: implications for E. coli YchM in fatty acid metabolism. Structure 18, 1450–1462 (2010).

12. Pasqualetto, E. et al. Structure of the cytosolic portion of the motor protein prestin and functional role of the STAS domain in SLC26/SulP anion transporters. J. Mol. Biol. 400, 448–462 (2010).

13. Geertsma, E.R. & Dutzler, R. A versatile and efficient high-throughput cloning tool for structural biology. Biochemistry 50, 3272–3278 (2011).

14. Kawate, T. & Gouaux, E. Fluorescence-detection size-exclusion chromatography for precrystallization screening of integral membrane proteins. Structure 14, 673–681 (2006).

15. Geertsma, E.R., Groeneveld, M., Slotboom, D.J. & Poolman, B. Quality control of overexpressed membrane proteins. Proc. Natl. Acad. Sci. USA 105, 5722–5727 (2008).

16. Detro-Dassen, S. et al. Conserved dimeric subunit stoichiometry of SLC26 multifunctional anion exchangers. J. Biol. Chem. 283, 4177–4188 (2008).

17. Compton, E.L. et al. Conserved structure and domain organization among bacterial Slc26 transporters. Biochem. J. 463, 297–307 (2014).

18. Shibagaki, N. & Grossman, A.R. The role of the STAS domain in the function and biogenesis of a sulfate transporter as probed by random mutagenesis. J. Biol. Chem. 281, 22964–22973 (2006).

19. Chernova, M.N. et al. Acute regulation of the SLC26A3 congenital chloride diarrhoea anion exchanger (DRA) expressed in Xenopus oocytes. J. Physiol. (Lond.) 549, 3–19 (2003).

20. Pardon, E. et al. A general protocol for the generation of Nanobodies for structural biology. Nat. Protoc. 9, 674–693 (2014).

21. Compton, E.L., Karinou, E., Naismith, J.H., Gabel, F. & Javelle, A. Low resolution structure of a bacterial SLC26 transporter reveals dimeric stoichiometry and mobile intracellular domains. J. Biol. Chem. 286, 27058–27067 (2011).

22. Forrest, L.R., Kramer, R. & Ziegler, C. The structural basis of secondary active transport mechanisms. Biochim. Biophys. Acta 1807, 167–188 (2011).

23. Lu, F. et al. Structure and mechanism of the uracil transporter UraA. Nature 472, 243–246 (2011).

24. Wong, F.H. et al. The amino acid-polyamine-organocation superfamily. J. Mol. Microbiol. Biotechnol. 22, 105–113 (2012).

25. Västermark, A. & Saier, M.H. Jr. Evolutionary relationship between 5+5 and 7+7 inverted repeat folds within the amino acid-polyamine-organocation superfamily. Proteins 82, 336–346 (2014).

26. Gorbunov, D. et al. Molecular architecture and the structural basis for anion interaction in prestin and SLC26 transporters. Nat. Commun. 5, 3622 (2014).

27. Barneaud-Rocca, D., Etchebest, C. & Guizouarn, H. Structural model of the anion exchanger 1 (SLC4A1) and identification of transmembrane segments forming the transport site. J. Biol. Chem. 288, 26372–26384 (2013).

28. Cordat, E. & Reithmeier, R.A. Structure, function, and trafficking of SLC4 and SLC26 anion transporters. Curr. Top. Membr. 73, 1–67 (2014).

29. He, J.J. & Quiocho, F.A. Dominant role of local dipoles in stabilizing uncompensated charges on a sulfate sequestered in a periplasmic active transport protein. Protein Sci. 2, 1643–1647 (1993).

30. Dutzler, R., Campbell, E.B. & MacKinnon, R. Gating the selectivity filter in ClC chloride channels. Science 300, 108–112 (2003).

31. Lobet, S. & Dutzler, R. Ion-binding properties of the ClC chloride selectivity filter. EMBO J. 25, 24–33 (2006).

32. Kalli, A.C., Sansom, M.S. & Reithmeier, R.A. Molecular dynamics simulations of the bacterial UraA H+-uracil symporter in lipid bilayers reveal a closed state and a selective interaction with cardiolipin. PLOS Comput. Biol. 11, e1004123 (2015).

33. Lee, C. et al. A two-domain elevator mechanism for sodium/proton antiport. Nature 501, 573–577 (2013).

34. Paulino, C., Wohlert, D., Kapotova, E., Yildiz, O. & Kuhlbrandt, W. Structure and transport mechanism of the sodium/proton antiporter MjNhaP1. eLife 3, e03583 (2014).

35. Landau, M. et al. ConSurf 2005: the projection of evolutionary conservation scores of residues on protein structures. Nucleic Acids Res. 33, W299–W302 (2005).

a r t i c l e snp

g©

2015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature structural & molecular biologydoi:10.1038/nsmb.3091

ONLINE METhODSCloning and expression screening of prokaryotic SLC26 homologs. Open reading frames of 92 prokaryotic SLC26 homologs were amplified by PCR from genomic DNA. Primers were designed with an automated procedure available online (http://www.fxcloning.org/). Gel-purified PCR products were cloned into pBXC3GH with fragment exchange (FX) cloning13, transformed into E. coli MC1061 (ref. 36) and validated by sequencing. Plasmid pBXC3GH allows expression of target proteins with a C-terminal HRV 3C protease site, GFP and His10 tag. The compatibility of SLC26 homologs toward overexpression was inves-tigated in small-scale cultures. Cells were grown in 96-well plates containing 700 µl Terrific Broth (TB-amp) under vigorous aeration. After cultivation for 1.5 h at 37 °C, the growth temperature was decreased to 25 °C over the course of 1 h. Expression was induced by the addition of l-arabinose to 0.001% (w/v) and allowed to proceed for 16 h. The quality and quantity of the overexpressed SLC26 homologs were verified by determining the GFP fluorescence in whole cells and after cell disruption, with SDS-PAGE15. Detergent stability and mono-dispersity was monitored by fluorescence size-exclusion chromatograpy14. As a result, SLC26Dg, the sole SLC26 homolog from D. geothermalis (GI 499845065), part of a single gene operon, was identified as a promising can-didate for structural studies. A summary of the screening results is provided in Supplementary Table 1.

Protein expression and purification. E. coli MC1061 containing pBXC3GH-SLC26Dg was cultivated in 9–18 l TB/ampicillin in a fermenter. Cells were grown at 37 °C until an OD600 ≈ 2 was reached, after which the tempera-ture was gradually decreased to 25 °C over the course of 1.5 h. Expression was induced by the addition of 0.005% l-arabinose and proceeded for 16 h. Cell pellets were resuspended in 50 mM potassium phosphate (KPi), pH 7.5, 150 mM NaCl, and 1 mM MgCl2 and incubated for 1 h at 4 °C in the pres-ence of 1 mg/ml lysozyme and traces of DNase I before disruption with an Emulsiflex C3 (Avestin). The lysate was cleared by low-spin centrifugation, and membrane vesicles were obtained by ultracentrifugation. Vesicles were resuspended to 0.5 g/ml in 50 mM KPi, pH 7.5, 150 mM NaCl and 10% glycerol (buffer A). All subsequent steps were carried out at 4 °C. Membrane proteins were extracted for 1 h at a concentration of 0.1 g/ml buffer A sup-plemented with 1–1.5% (w/v) n-decyl-β-d-maltoside (DM, Anatrace). Solubilized SLC26Dg was purified by immobilized metal affinity chroma-tography (IMAC) and cleaved by HRV 3C protease during dialysis against buffer without imidazole. Histidine-tagged GFP and protease were removed by IMAC, and cleaved protein was concentrated and subjected to size- exclusion chromatography (SEC) on a Superdex S200 column (GE Healthcare) equilibrated with 10 mM HEPES, pH 7.5, 150 mM NaCl and 0.2% DM. Protein complexes were prepared by mixing SEC-purified and concentrated SLC26Dg with SEC-purified and concentrated Nb5776 (supplemented with 0.2% DM) at a molar ratio of 1:3 for 10 min. The complex was subjected to SEC, and peak fractions were concentrated and used for crystallization. Protein for nanobody selection with neutravidin capture was prepared identically but was expressed with a C-terminal AviTag preceding the HRV 3C protease site and in vitro biotinylated with E. coli BirA. SLC26Dg∆STAS was expressed and puri-fied by the same protocol except that the detergent was exchanged into 0.84% n-nonyl-β-d-maltoside (NM) during SEC.

Generation of nanobodies targeting SLC26Dg. SLC26Dg-specific nanobodies were generated as previously described20,37. In brief, one llama (Lama glama) was immunized six times with 50 µg of DM-solubilized SLC26Dg. 4 d after the final antigen boost, peripheral blood lymphocytes were extracted, and their RNA was purified and converted to cDNA by reverse-transcription PCR. The nanobody repertoire was cloned into phage-display vector pMESy4 (ref. 20) containing a C-terminal His6-tag followed by the CaptureSelect C tag (with sequence Glu-Pro-Glu-Ala). 36 nanobody families that bound to SLC26Dg were identified in two rounds of biopanning: SLC26Dg was coated directly on a solid phase or immobilized via neutravidin capturing. Antigen-bound phages were recovered from antigen-coated wells without affecting phage infectivity by proteolysis with trypsin. After two rounds of selection, ELISAs were performed on periplasmic extracts of 72 individual colonies from each selection condition to screen for SLC26Dg-specific nanobodies. Nb5776 (clone MP326/G9) was selected by neutravidin-captured biotinylated SLC26Dg.

Expression and purification of nanobodies. All 36 nanobody families identified in the initial screen were recloned into pBXNPHM3 by FX cloning and expressed in E. coli MC1061. Plasmid pBXNPHM3 allows expression of target proteins with an N-terminal pelB signal sequence, His10-tag, MBP and HRV 3C protease site. Cells were cultivated in 5–9 l TB/ampicillin in a fermenter. Cells were grown at 37 °C until an OD600 ≈ 2 was reached. Expression was induced by the addition of l-arabinose to 0.01% (w/v) and was allowed to proceed for 6 h. Cell pellets were resuspended in 50 mM KPi, pH 7.5, 150 mM NaCl, and 1 mM MgCl2 and incubated for 1 h at 4 °C in the presence of 1 mg/ml lysozyme and traces of DNase I before disruption with an Emulsiflex C3 (Avestin). The lysate was cleared by ultracentrifugation, and nanobodies were purified with IMAC and SEC at 4 °C as described for SLC26Dg but in the absence of detergent.

Expression of selenomethionine-labeled protein. For preparation of selenom-ethionine-labeled SLC26Dg, an overnight culture grown in TB-amp was diluted 1:100 into 20 l M9 medium supplemented with trace elements, Kao and Michayluk vitamin solution (Sigma), 0.75% glycerol and 100 mg/l ampicillin. Cells were cultivated in a fermenter at 37 °C until the OD600 reached 0.4. Subsequently, the temperature was gradually decreased to 25 °C over the course of 1 h. Amino acids l-lysine, l-threonine and l-phenylalanine (each at a concentration of 125 mg/l) and l-leucine, l-isoleucine and l-valine (each at a concentration of 62.5 mg/l) were added to inhibit the methionine synthesis pathway38. Depletion of l-methionine was allowed for 1 h and was followed by the addition of 50 mg/l l-selenomethionine (SeMet). Expression was induced 30 min later by addition of 0.005% (w/v) l-arabinose. During the overnight expression, the temperature was kept at 22 °C. Cells were harvested by centrifugation and lysed by sonication. Extraction and purification were performed as described above.

Crystallization. The protein complex consisting of SLC26Dg and Nb5776 (9–12 mg/ml) was crystallized in sitting drops at 4 °C by vapor diffusion after mixture of protein and reservoir solutions in a 1:1 ratio. The best crystals, diffract-ing to 3.2 Å, were obtained for protein purified in DM and reservoir solutions containing 1 M ammoniumformate, 50 mM Na acetate, pH 4.5, and 45–50% PEG 400 (v/v). SeMet-labeled SLC26Dg complexed with Nb5776 was crystallized under identical conditions, except that lower PEG 400 concentrations were used (35–40%). SLC26Dg∆STAS was crystallized in the detergent NM at 4 °C from reser-voir solution containing 50 mM Mg-acetate, 50 mM glycine, pH 9.3, and 22–25% PEG 400 (v/v). Nb5776 (16–25 mg/ml) was crystallized at 4 °C from a reservoir solution containing 0.1 M Na-HEPES, pH 7.5, 200 mM CaCl2, and 27–29% PEG 400 (v/v). For cryoprotection, the PEG 400 concentration was increased stepwise to 35%. All crystals were flash frozen in liquid propane.

Structure determination. All data sets were collected on frozen crystals on the X06SA beamline at the Swiss Light Source (SLS) of the Paul Scherrer Institut (PSI) on a PILATUS 6M detector (Dectris; Table 1). The data were indexed, integrated and scaled with XDS39 and further processed with CCP4 programs40 (Table 1). The structure was determined by the single-wavelength anomalous dispersion (SAD) method with data collected at a wavelength of 0.978 Å from crystals con-taining a selenomethionine-derivatized membrane protein. The selenium sites were identified with SHELX C and D41,42 and refined in SHARP43. Phases were improved by solvent flipping in Solomon44. Models were built with O45 and COOT46. The correct register of the protein was confirmed by 18 methionine positions defined in the SeMet data set (excluding the methionine at the N termi-nus and Met496, which was not defined in the electron density). Five additional positions were observed in anomalous data of point mutants (i.e., I40M, L62M, W113M, I271M and L377M) in which methionine residues were introduced in regions lacking this amino acid in the native protein. The structure was refined in PHENIX47. R and Rfree were monitored throughout. Rfree was calculated by selection of 5% of the reflection data that were omitted in refinement. The final model has R/Rfree values of 26.2% and 30.7%, good geometry and no outliers in disallowed regions of the Ramachandran plot (Table 1). Regions not defined in the electron density include residues 1–13, 347–350 and 485–499 of SLC26Dg and 1–8, 30–33, and 125 of the nanobody. The structure of the nanobody was determined at 2.4 Å by molecular replacement in Phaser48, built in Coot46 and refined in PHENIX47. The structure of SLC26Dg∆STAS was determined by molecular replacement in Phaser48 and refined maintaining strong structural constraints in PHENIX47 (Table 1).

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature structural & molecular biology doi:10.1038/nsmb.3091

Preparation of proteoliposomes. Proteoliposomes were essentially prepared as described previously49. In brief, l-α-phosphatidylcholine pellets (from soybean, Sigma) were dissolved in chloroform, dried in a rotary evaporator, resuspended to 20 mg/ml and sonicated in buffer containing 50 mM KPi, pH 7.5. After three freeze-thaw cycles, large unilamellar vesicles were prepared by extrusion through a 400-nm-diameter polycarbonate filter. Liposomes were diluted to 4 mg/ml and destabilized beyond Rsat with Triton X-100. SEC-purified SLC26Dg in 0.2% DM was added to the liposomes at a weight ratio of 1:50 (protein/lipid), and deter-gent was subsequently removed by the addition of Biobeads. Proteoliposomes were harvested by centrifugation for 1.5 h at 250,000g and resuspended to a lipid concentration of 20 mg/ml. After three freeze-thaw cycles, proteoliposomes were stored in liquid nitrogen until use.

Radioisotope transport assays. For transport studies, proteoliposomes were thawed, resuspended in the appropriate internal buffer (either 50 mM KPi, pH 7.5, and 2 mM MgSO4, or 50 mM sodium phosphate (NaPi), pH 7.5, and 2 mM MgSO4) and pelleted by centrifugation for 20 min at 250,000g at 15 °C. Proteoliposomes were subsequently resuspended, pelleted, and after resuspen-sion subjected to three freeze-thaw cycles. After extrusion through a 400-nm polycarbonate filter and centrifugation, liposomes were resuspended to a final lipid concentration of 100 mg/ml. The sample was homogenized with a 26-gauge needle and stored at room temperature until use. Radioisotope transport studies were performed on stirred samples at 30 °C. To initiate transport, proteolipo-somes were diluted 40-fold into the appropriate external buffer (for example, 50 mM NaPi, pH 6.0 and 2 mM MgSO4, or 50 mM KPi, pH 6.0 and 2 mM MgSO4, and in certain cases 100 nM valinomycin) containing [14C]fumarate (American Radiolabeled Chemicals). At appropriate time points, 100-µl samples were taken and immediately diluted with 2 ml ice-cold external buffer; this was followed by rapid filtration on 0.45-µm nitrocellulose filters. After washing the filters with another 2 ml buffer, the radioactivity associated with the filter was determined by scintillation counting. Concentration dependence was assayed 10 s after substrate addition. Data were corrected for the passive diffusion of fumarate that was also observed in empty liposomes. Competition experiments were carried out at a [14C]fumarate concentration of 140 µM and 10 mM of the potential substrates. Other radiolabeled substances tested as potential substrates for SLC26Dg were [3H]succinate, [14C]oxalate and [14C]malonate.

Fluorescence-based transport assays. Fumarate-dependent proton transport was measured with the pH-sensitive fluorophore pyranine. To avoid bleaching of the fluorophore, the sample was shielded from direct light. Proteoliposomes were thawed, resuspended in 10 mM NaPi, pH 7.2, 2 mM MgSO4 and 2 mM pyranine and pelleted by centrifugation for 20 min at 250,000g at 15 °C. Proteoliposomes were subsequently treated as described for radioisotope transport assays. After extrusion through a 400-nm polycarbonate filter and centrifugation, liposomes were washed four times with buffer without pyranine and finally resuspended to a lipid concentration of 100 mg/ml. The sample was homogenized with a 26-gauge needle and stored at room temperature until use. For transport measurements,

proteoliposomes were diluted 114-fold into 10 mM KPi, pH 6.4, 2 mM MgSO4 and 100 nM valinomycin, and fluorescence was recorded at an excitation of 460 (F460) and 415 (F415) nm and emission of 509 nm in a Tecan Infinite M200 plate reader with black 96-well plates. After measurement of a stable baseline, fuma-rate was added to the sample and mixed by pipetting, and fluorescence readings were recorded. For analysis, the F460/F415 ratio was plotted. At the end of the experiment, nigericin was added to a final concentration of 500 nM to dissipate the proton gradient.

Multiangle light scattering. SEC-MALS experiments were carried out at 20 °C on an HPLC system (Agilent 1100) connected to an Eclipse 3 system equipped with a miniDAWN TREOS MALS detector and an Optilab T-rEX refractomoter (Wyatt Technology). 50 µg of purified protein (at 1 mg/ml) was injected onto a Superdex S200 column equilibrated in 10 mM HEPES, pH 7.5, 150 mM NaCl and 0.2% DM. Molecular weights and s.d. were determined with the Astra package (Astra 6.0, Wyatt Technology).

36. Casadaban, M.J. & Cohen, S.N. Analysis of gene control signals by DNA fusion and cloning in Escherichia coli. J. Mol. Biol. 138, 179–207 (1980).

37. Ehrnstorfer, I.A., Geertsma, E.R., Pardon, E., Steyaert, J. & Dutzler, R. Crystal structure of a SLC11 (NRAMP) transporter reveals the basis for transition-metal ion transport. Nat. Struct. Mol. Biol. 21, 990–996 (2014).

38. Van Duyne, G.D., Standaert, R.F., Karplus, P.A., Schreiber, S.L. & Clardy, J. Atomic structures of the human immunophilin FKBP-12 complexes with FK506 and rapamycin. J. Mol. Biol. 229, 105–124 (1993).

39. Kabsch, W. Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 26, 795–800 (1993).

40. Collaborative Computational Project, Number 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50, 760–763 (1994).

41. Schneider, T.R. & Sheldrick, G.M. Substructure solution with SHELXD. Acta Crystallogr. D Biol. Crystallogr. 58, 1772–1779 (2002).

42. Pape, T. & Schneider, T.R. HKL2MAP: a graphical user interface for phasing with SHELX programs. J. Appl. Crystallogr. 37, 843–844 (2004).

43. de La Fortelle, E. & Bricogne, G. in Methods in Enzymology (eds. Carter, C.W. & Sweet, R.M.) 492–494 (Academic, New York, 1997).

44. Abrahams, J.P. & Leslie, A.G. Methods used in the structure determination of bovine mitochondrial F1 ATPase. Acta Crystallogr. D Biol. Crystallogr. 52, 30–42 (1996).

45. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47, 110–119 (1991).

46. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).

47. Adams, P.D. et al. PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr. D Biol. Crystallogr. 58, 1948–1954 (2002).

48. McCoy, A.J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40, 658–674 (2007).

49. Geertsma, E.R., Nik Mahmood, N.A., Schuurman-Wolters, G.K. & Poolman, B. Membrane reconstitution of ABC transporters and assays of translocator function. Nat. Protoc. 3, 256–266 (2008).

npg

© 2

015

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.