Crystal structure of bacterial succinate:quinoneoxidoreductase flavoprotein SdhA in complexwith its assembly factor SdhEMegan J. Mahera,1, Anuradha S. Heratha, Saumya R. Udagedaraa, David A. Dougana, and Kaye N. Truscotta,1

aDepartment of Biochemistry and Genetics, La Trobe Institute for Molecular Science, La Trobe University, Melbourne, VIC 3086, Australia

Edited by Amy C. Rosenzweig, Northwestern University, Evanston, IL, and approved February 14, 2018 (received for review January 4, 2018)

Succinate:quinone oxidoreductase (SQR) functions in energy me-tabolism, coupling the tricarboxylic acid cycle and electron transportchain in bacteria and mitochondria. The biogenesis of flavinylatedSdhA, the catalytic subunit of SQR, is assisted by a highly conservedassembly factor termed SdhE in bacteria via an unknown mecha-nism. By using X-ray crystallography, we have solved the structureof Escherichia coli SdhE in complex with SdhA to 2.15-Å resolution.Our structure shows that SdhE makes a direct interaction with theflavin adenine dinucleotide-linked residue His45 in SdhA and main-tains the capping domain of SdhA in an “open” conformation. Thisdisplaces the catalytic residues of the succinate dehydrogenase ac-tive site by as much as 9.0 Å compared with SdhA in the assembledSQR complex. These data suggest that bacterial SdhE proteins, andtheir mitochondrial homologs, are assembly chaperones that con-strain the conformation of SdhA to facilitate efficient flavinylationwhile regulating succinate dehydrogenase activity for productivebiogenesis of SQR.

SdhE | flavinylation | structure | SdhA | assembly

Succinate:quinone oxidoreductase (SQR) is a multisubunitmembrane-associated enzyme found in the cytoplasm of bacteria

and in the matrix of mitochondria (where it is commonly termedcomplex II). The enzyme is central to cellular metabolism andenergy conversion, contributing to the tricarboxylic acid cycleand the electron transport chain. It catalyzes the oxidation of suc-cinate to fumarate, which is coupled to electron transfer throughflavin adenine dinucleotide (FAD) and three Fe–S clusters, resultingin the reduction of the electron carrier ubiquinone to ubiquinol. Theoverall architecture of the bacterial and mitochondrial complexes ishighly conserved, with both enzymes forming a heterotetramericprotein complex (1, 2). The catalytic flavoprotein (SdhA in bacteria,Sdh1 in yeast, and SDHA in humans) contains a binding site fordicarboxylic acid and an N3-histidyl-8α-FAD linkage (1, 2). Thecovalent attachment of FAD to SdhA is essential for the oxidation ofsuccinate by the enzyme (3). In the mature complex, SdhA connectsto the inner membrane through interaction with the Fe–S proteinSdhB, which makes direct contact to both integral membrane sub-units, SdhC and SdhD. Ubiquinone binds at the interface of SdhBand the two membrane proteins (SdhC and SdhD), with the cofac-tors in SQR forming a direct path (∼40 Å long) for the transfer ofelectrons into the respiratory chain (2). Significantly, loss of complexII activity in humans is associated with Leigh syndrome and tumorsyndromes including hereditary paraganglioma (PGL) (4–8).The assembly of SQR is assisted by several assembly factors

involved in cofactor biogenesis and regulation of assembly in-termediates (9–12). The most widely conserved SQR assemblyfactor (10, 13, 14) is termed SdhE in bacteria [also known as Sdh5in yeast and SDH assembly factor 2 (SDHAF2) or SDH5 in hu-mans]. Initially characterized in Saccharomyces cerevisiae, Sdh5 isessential for the assembly of active complex II promoting the co-valent attachment of FAD to Sdh1 (10). Likewise, bacterial SdhE isalso required for SQR and fumarate reductase (FRD) activitypromoting flavinylation of SdhA and FrdA (subunit A of the

quinol:FRD), respectively (13, 15). The importance of this proteinfamily, in normal cellular metabolism, is manifested by the iden-tification of a mutation in human SDHAF2 (Gly78Arg), which islinked to an inherited neuroendocrine disorder, PGL2 (10). Cur-rently, however, the role of SdhE in flavinylation remains poorlyunderstood. To date, three different modes of action for SdhE/Sdh5 have been proposed, suggesting that SdhE facilitates thebinding and delivery of FAD (13), acts as a chaperone for SdhA(10), or catalyzes the attachment of FAD (10). Moreover, the re-quirement for SdhE in SdhA biogenesis remains controversial, as re-cent studies have demonstrated that flavinylation of bacterial, archaeal,and mitochondrial SdhA can still occur in the absence of the assemblyfactor (16–19). In this study, we have determined the crystal structureof the Escherichia coli flavinylation factor SdhE in complex with itsclient protein SdhA to 2.15-Å resolution. This three dimensionalstructure of an SQR assembly intermediate provides valuable insightsinto the evolutionary conserved process of flavoprotein assembly.

Results and DiscussionStructure of SdhA in Complex with Its Assembly Factor SdhE.To obtainE. coli SdhA in complex with its assembly factor E. coli SdhE, wecoexpressed untagged recombinant SdhA together with recombi-nant His6-tagged SdhE in E. coli (Fig. 1A). By using Ni2+-affinitychromatography, we isolated a mixture of free His6-SdhE and

Significance

Assembly factors play key roles in the biogenesis of many multi-subunit protein complexes regulating their stability, activity, orincorporation of essential cofactors. The bacterial assembly factorSdhE (also known as Sdh5 or SDHAF2 in mitochondria) promotescovalent attachment of flavin adenine dinucleotide (FAD) to SdhAand hence the assembly of functional succinate:quinone oxidore-ductase (also known as complex II). Here, we present the crystalstructure of Escherichia coli SdhE bound to its client protein SdhA.This structure provides unique insight into SdhA assembly,whereby SdhE constrains unassembled SdhA in an “open” con-formation, promoting covalent attachment of FAD, but rendersthe holoprotein incapable of substrate catalysis. These data alsoprovide a structural explanation for the loss-of-function mutation,Gly78Arg, in SDHAF2, which causes hereditary paraganglioma 2.

Author contributions: M.J.M., D.A.D., and K.N.T. designed research; M.J.M., A.S.H., andS.R.U. performed research; M.J.M., A.S.H., S.R.U., D.A.D., and K.N.T. analyzed data; andM.J.M., D.A.D., and K.N.T. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Published under the PNAS license.

Data deposition: The atomic coordinates and structure factors have been deposited in theProtein Data Bank, www.wwpdb.org (PDB ID code 6C12).1To whom correspondence may be addressed. Email: m.maher@latrobe.edu.au ork.truscott@latrobe.edu.au.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1800195115/-/DCSupplemental.

Published online March 7, 2018.

2982–2987 | PNAS | March 20, 2018 | vol. 115 | no. 12 www.pnas.org/cgi/doi/10.1073/pnas.1800195115

Dow

nloa

ded

by g

uest

on

June

20,

202

0

His6-SdhE in complex with untagged SdhA. A homogenous prepa-ration of the SdhA/SdhE (SdhAE) complex was then purified bysize-exclusion chromatography (Fig. 1B). The SdhAE complexeluted from an analytical size-exclusion column at a volume corre-sponding to the molecular weight of a heterodimer (Fig. 1C) andFADwas covalently bound (to SdhA) within the binary complex (10)

(Fig. 1D). To elucidate the atomic structure of the SdhAE proteinassembly, we crystallized the complex and determined its structure to2.15-Å resolution by X-ray crystallography (Fig. 2A and Table S1).The crystal contains two almost identical SdhAE complexes in

the asymmetric unit. The final SdhAE model includes residues1–583 from the SdhA protein, apart from two regions: residues52–67 and 107–137 in addition to residues 2–83 from the SdhEprotein. The structure of SdhA is composed of four domains (20):an FAD-binding domain (residues 1–245 and 351–431), which in-cludes a Rossmann-type fold and provides the binding site forFAD (Fig. 2A, pink); a capping domain composed of residues245–351 (Fig. 2A, gray); a helical domain composed of residues431–547 (Fig. 2A, blue); and a C-terminal domain composed ofresidues 547–583 (Fig. 2A, orange). FAD was well resolved in theelectron density map and, consistent with the in-gel assay (Fig. 1D),was covalently attached to SdhA via a N3-histidyl-8α-FAD linkage(1.6 Å) between the Ne2 atom of His45 and the C8 atom of theisoalloxazine group (Fig. 2B). Although a dicarboxylate is requiredfor the flavinylation reaction (21), the SdhAE structure presentedhere captures the assembly intermediate after flavinylation hasoccurred. Consistently, there was no evidence in the electrondensity for the presence of a dicarboxylate associated with SdhA.SdhE forms a single compact domain (Fig. 2A, green) composedof five α-helices and interfaces closely with SdhA, where it iswedged between the FAD and capping domains.The SdhE protein forms an intimate complex with SdhA, with

a buried surface area of 1,455 Å2 (Fig. 3 A–C). Three regions ofSdhE make contact with the SdhA protein, namely residues 5–25(which encompass helix α1, the N terminus of α2, and the loop thatconnects these two regions), residues 47–61 (which form helix α4),and residues 80–86 (which form the C terminus; Fig. 3A and Fig.S1). Overall, the interface is composed of electrostatic interactions,18 hydrogen bonds, and hydrophobic contacts and exhibits a highdegree of charge complementarity (Fig. 3D). Importantly, the in-terface defined by the SdhAE structure is consistent with a recentbiochemical study that identified two SdhE residues (Arg8 andMet17), which could be photocrosslinked to SdhA or FrdA (18).Consistent with an important functional role for residues locatedwithin this interface, mutagenesis of Arg68 and Tyr71 in yeast Sdh5(22) (equivalent to Arg8 and Trp11 in E. coli SdhE) preventedSdh1 flavinylation. Similarly, the inherited mutation coding forGly78Arg in human SDHAF2 (equivalent to Gly16 in E. coli SdhE,which is located in the loop between the α1 and α2 helices; Fig. S1)strongly destabilizes the SDHAF2 interaction with SDHA (10, 23).Collectively, these data suggest that the loop that connects theα1 and α2 helices of SdhE plays a crucial role in the stability ofthe complex. SdhA interacts with the SdhE protein in three main

Fig. 1. Isolation of the binary protein complex containing SdhE, SdhA, andcovalently bound FAD. (A) Schematic illustration showing His6-tagged SdhEand untagged SdhA constructs used in this study. The numbers reflect aminoacid residues of the authentic portion of the protein constructs. (B) Elutionprofile (Left) showing separation of the SdhAE complex (first peak) from freeSdhE (second peak) using size-exclusion chromatography. (Right) Confir-mation of purified SdhAE complex by Coomassie Brilliant Blue (CBB)-stainedSDS/PAGE. (C) Elution profile of purified SdhAE relative to indicated proteinstandards analyzed by size-exclusion chromatography. The molecular weightof the SdhAE complex is estimated to be 88.5 kDa. (D) Detection of FADfluorescence (Fluor.) in protein samples (Upper) of E. coli lysate (lane 1) andpurified SdhAE complex (lanes 2 and 3) following separation by SDS/PAGE.(Lower) The same gel following CBB staining. Only the region of the gelcontaining SdhA is shown.

Fig. 2. Structure of the SdhAE complex. (A) The SdhE protein (green) forms a 1:1 complex with SdhA. SdhE is positioned between the SdhA FAD-binding(pink) and capping (gray) domains. The helical and C-terminal domains of the SdhA protein are shown in pale blue and orange, respectively. FAD (yellowspheres) is covalently bound to the SdhA protein. (B) In the SdhAE complex, residue His45 (SdhA; pink) is covalently bonded to the FAD cofactor (yellow).The carbonyl group of SdhE residue Gly16 (green) participates in a hydrogen bonding interaction with SdhA His45.

Maher et al. PNAS | March 20, 2018 | vol. 115 | no. 12 | 2983

BIOCH

EMISTR

YBIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

June

20,

202

0

regions: residues 38–50, 138–143, and 210–217 of the FAD domain(which frame the FAD binding site); residues 247–251 of the uppersurface of the capping domain; and residues 505–515 at the tip ofthe helical region of the FAD domain (nomenclature defined inref. 18). Consistent with the SdhAE structure, FrdA residues 456–462, which are equivalent to the helical region in SdhA, wereshown to photo–cross-link to Met17 of SdhE (18).

Rotation of the SdhA Capping Domain Accompanies Formation of theSdhAE Complex. Superposition of SdhA from the SdhAE complexwith SdhA from the intact SQR enzyme [Protein Data Bank (PDB)ID code 2WDQ; chain A (24)] reveals a number of significant dif-ferences between the two structures (rmsd of 4.14 Å for 536 com-mon Cα positions). Although most of the structure of the FAD-binding domain is similar in both states, there are significantchanges to three distinct regions of SdhA, specifically that residues40–50 are displaced by as much as 7 Å, residues 88–105 are dis-placed by as much as 8 Å, and the capping domain (residues 246–350) is displaced by as much as 13 Å. In the SdhAE complex, thecapping domain of SdhA is rotated away from the FAD-bindingdomain by 43.5° [as analyzed by DynDom (25)] relative to thatobserved for the intact SQR enzyme (Fig. 4A). The rotation of the

capping domain, coupled with the apparent mobility of residues52–67 and 107–137, exposes the isoalloxazine group of the boundFAD in the SdhAE complex to solvent (solvent-accessible surfacearea is 64 Å2; Fig. 4B), whereas the cofactor is completely buriedin the intact SQR structure [PDB ID code 2WDQ (24); solventaccessible area is 5 Å2; Fig. 4C].Interestingly, in contrast to the SdhAE structure, in which the FAD

binding site remains exposed in the presence of covalently bound FAD,the FAD binding pocket of most FAD-bound enzymes is enclosed. Forexample, the incorporation of FAD into L-aspartate oxidase triggersstructural reorganization of the capping domain relative to the FAD-binding domain (26). In this case, the apo protein exists in an openconformation in which both domains are well separated. However,upon binding of FAD, the capping domain rotates 27° toward theFAD-binding domain, thereby closing the binding site. This type of“closed” conformation is commonly observed in these types of enzymesfollowing FAD incorporation (27–29). Hence, our FAD-bound struc-ture of SdhA in complex with SdhE is unique in presenting an FAD-bound, open arrangement of the FAD-binding and capping domains,and helps to reveal the role of SdhE in stabilizing this conformation.

Direct Contact Between SdhE and the Flavin Binding Site of SdhAFacilitates Covalent Flavinylation. In contrast to the significant struc-tural rearrangements in SdhA, which accompany formation of the

Fig. 3. SdhE forms an intimate contact with SdhA. (A) The SdhE protein israinbow-colored (blue at the N terminus to red at the C terminus). Secondarystructural elements are labeled (helices α1−α5). The capping domain of SdhAis colored gray and the remainder of SdhA is colored pink. The FAD cofactoris represented as yellow spheres. (B) Surface representation of the SdhAEcomplex with SdhE shown in cyan and SdhA in dark blue. The FAD cofactor isrepresented as yellow sticks. (C) “Open book” representation of the SdhAEcomplex from B (SdhE, cyan; SdhA, dark blue). The protein surfaces that areburied in the complex are highlighted in white. (D) The SdhAE complex asrepresented in C. The protein surfaces are colored according to the elec-trostatic potentials (red, negatively charged; blue, positively charged; white,uncharged). The protein–protein interface in the SdhAE complex showscharge complementarity.

Fig. 4. The position of the SdhA capping domain is markedly different inthe SdhAE assembly relative to SQR. (A) Comparison of the orientation ofthe SdhA capping domain in the structure of SQR (PDB ID code 2WDQ; red)(24) with that in the SdhAE complex (blue). In SdhAE, the capping domain isobserved in an orientation rotated 43.5° away from the FAD-binding do-main. The FAD-binding, helical, and C-terminal domain structures for theSdhA protein within SdhAE and SQR complexes are shown in gray. The FADcofactor is represented as yellow sticks. (B) Surface representation of theSdhAE complex (SdhE, green; SdhA capping domain, gray; FAD-binding,helical, and C-terminal domains, pink; FAD cofactor, yellow spheres). (C)Surface representation of the SdhAB structure within SQR (24) (PDB ID code2WDQ; SdhB, green; SdhA colored as in B).

2984 | www.pnas.org/cgi/doi/10.1073/pnas.1800195115 Maher et al.

Dow

nloa

ded

by g

uest

on

June

20,

202

0

SdhAE complex, the structure of SdhE is virtually unchanged. Su-perposition of the structure of SdhE alone (PDB ID code 1X6I)(30) with SdhE in the SdhAE complex reveals an rmsd of 0.39 Å(for 75 common core Cα atoms), with minor structural rear-rangements localized to the C terminus. Interestingly, the SdhEloop, between helices α1 and α2, does not change conformationupon complex formation, despite its central location at the SdhAEinterface. This part of the SdhE structure appears to be tailoredfor its interaction with SdhA and therefore may contribute directlyto catalysis of SdhA flavinylation. Nevertheless, the autocatalyticflavinylation of several FAD-dependent enzymes has been welldescribed in the literature. The mechanism of catalysis is proposedto involve five features (31): (i) abstraction of a proton from theFAD C8 methyl group (by bases near the residues that tetherthe flavin moiety); (ii) stabilization of the negative charge at theN1-C2=O2 locus of the isoalloxazine moiety; (iii) attack bythe histidyl-imidazolyl group at C8α to form a covalent bond be-tween the polypeptide chain and the reduced flavin; (iv) pro-tonation of the negative charge at the N5 position of the reducedisoalloxazine ring system by a neighboring amide group; and (v)reoxidation of the reduced flavin by electron transfer to oxygen.As detailed in Table S2, the majority of the crucial interactions thatsupport these steps are maintained not only in the final structures ofbovine and E. coli enzymes but also in the SdhAE complex de-scribed here. Therefore, these data suggest that SdhE does not playa direct role in the catalysis of SdhA flavinylation, but rather sup-ports a conformation of SdhA that facilitates autocatalysis.Part of the interaction network between SdhA and SdhE in-

cludes a hydrogen bond between the carbonyl oxygen of Gly16 inSdhE and the Nδ1 atom of SdhA residue His45, which is covalentlybound to the FAD cofactor (Fig. 2B). This glycine residue is highlyconserved in SdhE/Sdh5 proteins across a broad range of speciesincluding bacteria, fungi, plants, and animals (Fig. S1B) (14, 32).Significantly, a point mutation at the equivalent Gly residue inhuman mitochondrial SDHAF2 (Gly78Arg) is associated withPGL2 disease, and this mutation dramatically impairs its in-teraction with human SDHA (10, 23). Consistent with an impor-tant role for this highly conserved Gly residue in the function ofSDHAF2, tumor cells isolated from patients with PGL2 containextremely low levels of flavinylated SDHA (10). Similarly, theequivalent mutation (Gly16Arg) in bacterial SdhE inhibits flavi-nylation of SdhA (32). Therefore, we propose that the hydrogenbond formed between SdhE residue Gly16 and SdhA residueHis45 is crucial for efficient flavinylation of SdhA. This interactionensures that the Nδ1 atom at SdhA residue His45 is protonated,promoting nucleophilic attack by the deprotonated Ne2 atom atFAD C8α to facilitate covalent bond formation between thepolypeptide chain and the reduced flavin. This hydrogen bondmay also lock the side chain of His45 in an orientation conduciveto this reaction.

SdhE Stabilizes SdhA in a Nonactive Conformation During Assembly.The SdhAE complex also reveals that flavinylation of SdhA doesnot trigger release of SdhE. Instead, SdhE is wedged between theFAD-binding and capping domains of SdhA, relocating its activesite residues (24). Specifically, the side chains of SdhA residuesHis242, Thr254, Glu255, and Arg286 are displaced by distancesof as much as 9.0 Å away from the dicarboxylate-binding site (incomparison with their positions in the SQR structure) (24),which is located between the isoalloxazine ring of FAD and thecapping domain (2, 24) (Fig. 5). We propose that the significantpositional displacement of these residues during SdhAE assemblyprevents oxidation of succinate by the complex. This disruption ofthe active site is likely important in preventing electron “leakage”from free SdhA (via succinate dehydrogenase activity) and hence thesubsequent production of damaging reactive oxygen species. So howdoes SdhA progress in the assembly pathway and associate with theiron–sulfur protein SdhB? Given that the surface area of the SdhA/

SdhB interface in the SQR structure (2,400 Å2) (24) is significantlygreater than the interface between SdhA and SdhE in the SdhAEcomplex (1,455 Å2), it is likely that the binding affinity of flavinylatedSdhA is stronger for SdhB than it is for SdhE (33). Displacementof SdhE therefore likely occurs via competition with SdhB afterthe covalent attachment of the FAD, which allows “closure” of thecapping and FAD-binding domains and repositioning of the catalyticresidues, resulting in activation of the enzyme.

Mechanism of Action of SdhE in the Biogenesis of Holo-SdhA.Recently,it was proposed that SdhE docks onto its target flavoprotein fol-lowing closure of the capping domain over noncovalently boundFAD, and, as such, maintains the flavoprotein in the closed stateto facilitate flavinylation (18, 19). In contrast to this recentlyproposed model, our structure clearly demonstrates that SdhE, inseparating the FAD-binding and capping domains, holds holo-SdhA in an open conformation. Therefore, we propose an alter-native pathway, at least for the biogenesis of holo-SdhA (Fig. 6).In the absence of SdhE, FAD binds noncovalently to the FADbinding domain of folded apo-SdhA. At this point, in the presenceof dicarboxylate, spontaneous flavinylation may occur (Fig. 6, path I);however, under certain conditions and in certain cell types, thisprocess is very inefficient (10, 13, 18), possibly because of non-productive pathways of misfolding or unregulated assembly ofapo-SdhA (Fig. 6, path III). In path II, SdhE and FAD bind tofolded apo-SdhA sequentially (as illustrated in Fig. 6) or simul-taneously. The docking of SdhE to SdhA not only maintains thecapping domain of SdhA in an open conformation but also fa-cilitates the formation of a critical hydrogen bond between thecarbonyl oxygen of SdhE residue Gly16 and the Nδ1 atom ofSdhA residue His45, which stabilizes SdhA in a state that is con-ducive to highly efficient autocatalytic flavinylation. The limitedefficiency of flavinylation in the absence of SdhE (18) is most likelythe result of the loss of exclusive protonation of the Nδ1 atom andtethering of the side chain of His45 in SdhA that forms the N3

-histidyl-8α-FAD linkage. Following flavinylation of SdhA, SdhEremains bound to SdhA, thereby stabilizing holo-SdhA in an in-active conformation by displacement of catalytic residues (requiredfor oxidation of succinate). This is likely important to prevent elec-tron leakage and hence oxidative damage in the cell. Such a role maybe particularly important under aerobic conditions in E. coli in whichSQR is expressed (ref. 34 and references therein). Finally, wespeculate that holo-SdhB, by virtue of a higher affinity for SdhA,displaces SdhE from SdhA, allowing the capping domain of SdhAto close over the FAD binding domain. Within this model, wecannot exclude the possibility that other factors act to passivelyor actively displace SdhE from SdhA before SdhB docking.

Fig. 5. The influence of bound SdhE on catalytic residues in SdhA. Super-position of the SdhA proteins from the SdhAE (gray) and SQR [PDB ID code2WDQ (24); blue] structures. For clarity, the FAD cofactor from the SdhAEstructure only is represented in yellow. Active site residues (labeled) aredisplaced by as much as 9.0 Å away from the substrate binding site in theSdhAE complex relative to SQR.

Maher et al. PNAS | March 20, 2018 | vol. 115 | no. 12 | 2985

BIOCH

EMISTR

YBIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

June

20,

202

0

Intriguingly, mitochondria contain an additional assembly factor,Sdh8 (also known as SDHAF4), which forms a stable complex withflavinylated SdhA and seemingly facilitates its assembly with SdhB(11). Nevertheless, a bacterial homolog (or functional equivalent) ofSdh8 has yet to be described. The assembly intermediate SdhABsubsequently associates with membrane-integrated SQR subunits toform the functional complex. In summary, SdhE has a dual functionto promote efficient flavinylation of SdhA while simultaneouslyneutralizing its succinate dehydrogenase activity during assembly.

Comparison with the E. coli FrdA–SdhE Complex. While this manu-script was under review, the crystal structure of E. coli SdhE cross-linked to flavinylated FrdA (denoted here as the FrdA–SdhEcomplex) was reported to 2.6-Å resolution by Iverson and co-workers (35). The FrdA–SdhE structure shares a number of fea-tures with that of SdhAE reported here: (i) SdhE is locatedbetween the FAD and capping domains of the flavinylated proteinin both structures, and makes a direct hydrogen bonding contactbetween Gly16 (in SdhE) and the histidyl FAD linkage in therespective flavoprotein; (ii) the SdhE binding site on the flaviny-lated protein overlaps with that of the FrdB/SdhB subunit withinthe final complex; and (iii) complex formation with SdhE ac-companies a rotation of the appropriate capping domain. Im-portantly, comparison of the two structures also reveals a numberof significant differences, which may reflect the relative stabilitiesof the protein complexes and/or the methods used to generatethem. The SdhAE structure reported here represents a stablecomplex, which was recovered following size-exclusion chroma-tography. In contrast, the FrdA–SdhE structure required thegeneration of a FrdA–SdhE complex by chemical cross-linking,and, hence, as suggested by Iverson and coworkers, “reflects animperfectly stabilized complex” (35). Interestingly, our compari-son of the two structures reveals that the positioning of SdhEdiffers. In the FrdA–SdhE complex, SdhE is tilted toward theproposed position of the cross-link. The result is a smaller rotationof the FrdA capping domain (10.6–10.8°, dependent on whethermalonate or acetate is present in the dicarboxylate binding site) inthe FrdA–SdhE complex compared with the SdhAE complex(rotation of 43.5°). In addition, the size of the protein–protein

interface within the FrdA–SdhE complex is smaller than that inthe SdhAE structure (1,085 and 1,455 Å2, respectively). Signifi-cantly, the structure of the two FrdA–SdhE assemblies within theasymmetric unit are nonidentical, and the average B-factor for theSdhE protein is approximately twice that of the FrdA protein,indicating significant disorder. The authors also reported that theelectron density for SdhE was “weaker” than for the FrdA subunit(35). In contrast, the crystal structure of both SdhAE complexesper asymmetric unit are almost identical, with consistent averageB-factors across the SdhA and SdhE proteins and ordered elec-tron density for the entire SdhE molecules (Table S1). Whetherthese differences are a result of the different experimental methodsused to generate these complexes for structural analyses will awaitfurther investigation.

Materials and MethodsProtein Expression, Purification, and Characterization. The SdhE and SdhA codingsequences were amplified by PCR by using E. coli (MC4100) genomic DNA andprimer pairs (Bam_E, 5′-TGTAGCGGATCCGGACATTAACAACAAAGCCCG-3′;E_Hind3, 5′-CTATCGAAGCTTAAGCTTAGATTGCCACAGGACCA-3′; and Nde_A,5′-ATGCGACATATGAAATTGCCAGTCAGAG-3′; Xho_A, 5′-TAGCACCTCGAGT-TAGTAAGTACGAATCTTCGG-3′; restriction enzyme recognition sites are in-dicated in bold) and cloned into pETDuet-1. His6-SdhE and SdhA werecoexpressed for 4 h at 37 °C in E. coli strain BL21(DE3) CodonPlus-RIL (Stra-tagene) transformed with sequence-verified pETDuet-1/His-sdhE/sdhA. Over-expressed His6-SdhE/SdhA was isolated from soluble lysate [in 20 mM Tris·HCl,pH 8.0, 200mMKCl, 10% (vol/vol) glycerol, 1 mM EDTA, 0.05% (vol/vol) Triton X-100, 0.1 mg/mL lysozyme, 5 mM MgCl2, 10 μg/mL DNase I] by affinity chro-matography by using Ni-NTA agarose (Qiagen) under standard conditions.Excess SdhE was separated from the SdhAE binary complex by using aSuperdex 200 HiLoad 16/60 column (GE Healthcare) run at 0.5 mL/min in GFbuffer (50 mM Tris·HCl, pH 8.0, 300 mM NaCl). The molecular weight of thepurified SdhAE complex was estimated by size-exclusion chromatography byusing a Superdex 200 Increase 10/300 GL column (GE Healthcare) calibrated byusing a gel filtration protein standard mix (Bio-Rad). To analyze flavinylation,an in-gel fluorescence detection method was used (10).

Protein Crystallization and Data Collection. SdhAE was crystallized at 20 °C byhanging-drop vapor diffusion with drops consisting of equal volumes (1 μL)of protein (6.3 mg/mL in 50 mM Tris·HCl, pH 8.0, 300 mM NaCl) and crys-tallization solution [0.1 M Hepes, pH 7.5, 0.2 M MgCl2.6H2O, 30% (wt/vol)

Fig. 6. Proposed SdhA assembly pathway. Path I represents flavinylation of SdhA occurring in the absence of SdhE. In this case, the yield of flavinylated SdhAand hence active SQR is very low. Path II represents SdhE-assisted flavinylation of SdhA, the optimal route for efficient flavinylation and SQR activity. Path IIIalso represents assembly of SdhA in the absence of SdhE. In this case, however, FAD is absent or remains associated with SdhA noncovalently. This leads tolimited progression in the pathway or assembly of a nonfunctional SQR. This model is based largely on the structural (this study) and cellular studies of E. coliSdhA flavinylation and assembly under aerobic conditions. CAP, capping domain of SdhA; FBD, FAD binding domain of SdhA; E, SdhE; G16, Gly16 residue inSdhE; H45, His45 residue in SdhA. FAD is indicated in yellow.

2986 | www.pnas.org/cgi/doi/10.1073/pnas.1800195115 Maher et al.

Dow

nloa

ded

by g

uest

on

June

20,

202

0

PEG 3350, and 40 mM NaF]. Crystals were cryoprotected in reservoir solutionwith 30% (vol/vol) glycerol before flash-cooling in liquid nitrogen. Diffrac-tion data were collected on an ADSC Quantum 315r detector at the Aus-tralian Synchrotron on beamline MX2 at 100 K and were processed withHKL2000 (36). Unit cell parameters and data collection statistics are pre-sented in Table S1.

Structure Solution and Refinement. The crystal structure of SdhAE was solvedby molecular replacement by using PHASER (37) with search models from thecrystal structures of the SdhA subunit from the E. coli SQR [chain A; PDB IDcode 2WDQ (24)] and the E. coli “hypothetical protein” YgfY [now known tobe SdhE; PDB ID code 1X6I (30)]. The resulting model was refined by iterativecycles of refinement with REFMAC (38) against anisotropically corrected

diffraction data (39) interspersed with manual inspection and correctionagainst calculated electron density maps by using COOT (40). The refinementof the model converged with residuals R = 0.199 and Rfree = 0.258 (Table S1).The structure was judged to have excellent geometry as determined byMOLPROBITY (41) (Table S1).

ACKNOWLEDGMENTS. Aspects of this research were undertaken on theMacromolecular Crystallography beamline MX2 at the Australian Synchrotron(Victoria, Australia), and we thank the beamline staff for their enthusiastic andprofessional support. This work was funded by Australian Research CouncilGrants DP150104639 (to K.N.T. and D.A.D.), a La Trobe University Full FeeResearch Scholarship (to A.S.H.), and a La Trobe University PostgraduateResearch Scholarship (to A.S.H.).

1. Sun F, et al. (2005) Crystal structure of mitochondrial respiratory membrane proteincomplex II. Cell 121:1043–1057.

2. Yankovskaya V, et al. (2003) Architecture of succinate dehydrogenase and reactiveoxygen species generation. Science 299:700–704.

3. Robinson KM, Rothery RA, Weiner JH, Lemire BD (1994) The covalent attachment ofFAD to the flavoprotein of Saccharomyces cerevisiae succinate dehydrogenase is notnecessary for import and assembly into mitochondria. Eur J Biochem 222:983–990.

4. Astuti D, et al. (2001) Gene mutations in the succinate dehydrogenase subunit SDHBcause susceptibility to familial pheochromocytoma and to familial paraganglioma.Am J Hum Genet 69:49–54.

5. Baysal BE, et al. (2000) Mutations in SDHD, a mitochondrial complex II gene, in he-reditary paraganglioma. Science 287:848–851.

6. Bourgeron T, et al. (1995) Mutation of a nuclear succinate dehydrogenase gene re-sults in mitochondrial respiratory chain deficiency. Nat Genet 11:144–149.

7. Burnichon N, et al. (2010) SDHA is a tumor suppressor gene causing paraganglioma.Hum Mol Genet 19:3011–3020.

8. Niemann S, Müller U (2000) Mutations in SDHC cause autosomal dominant para-ganglioma, type 3. Nat Genet 26:268–270.

9. Ghezzi D, et al. (2009) SDHAF1, encoding a LYR complex-II specific assembly factor, ismutated in SDH-defective infantile leukoencephalopathy. Nat Genet 41:654–656.

10. Hao HX, et al. (2009) SDH5, a gene required for flavination of succinate dehydrogenase,is mutated in paraganglioma. Science 325:1139–1142.

11. Van Vranken JG, et al. (2014) SDHAF4 promotes mitochondrial succinate dehydrogenaseactivity and prevents neurodegeneration. Cell Metab 20:241–252.

12. Na U, et al. (2014) The LYR factors SDHAF1 and SDHAF3 mediate maturation of theiron-sulfur subunit of succinate dehydrogenase. Cell Metab 20:253–266.

13. McNeil MB, Clulow JS, Wilf NM, Salmond GP, Fineran PC (2012) SdhE is a conservedprotein required for flavinylation of succinate dehydrogenase in bacteria. J Biol Chem287:18418–18428.

14. Huang S, Taylor NL, Ströher E, Fenske R, Millar AH (2013) Succinate dehydrogenaseassembly factor 2 is needed for assembly and activity of mitochondrial complex II andfor normal root elongation in Arabidopsis. Plant J 73:429–441.

15. McNeil MB, et al. (2014) The succinate dehydrogenase assembly factor, SdhE, is re-quired for the flavinylation and activation of fumarate reductase in bacteria. FEBSLett 588:414–421.

16. Bezawork-Geleta A, Dong L, Rohlena J, Neuzil J (2016) The assembly factor SDHAF2 isdispensable for flavination of the catalytic subunit of mitochondrial complex II inbreast cancer cells. J Biol Chem 291:21414–21420.

17. Kounosu A (2014) Analysis of covalent flavinylation using thermostable succinatedehydrogenase from Thermus thermophilus and Sulfolobus tokodaii lacking SdhEhomologs. FEBS Lett 588:1058–1063.

18. Maklashina E, et al. (2016) Binding of the covalent flavin assembly factor to the fla-voprotein subunit of complex II. J Biol Chem 291:2904–2916.

19. Starbird CA, et al. (2017) Structural and biochemical analyses reveal insights intocovalent flavinylation of the Escherichia coli complex II homolog quinol:fumaratereductase. J Biol Chem 292:12921–12933.

20. Lancaster CR, Kröger A, Auer M, Michel H (1999) Structure of fumarate reductasefrom Wolinella succinogenes at 2.2 A resolution. Nature 402:377–385.

21. Brandsch R, Bichler V (1989) Covalent cofactor binding to flavoenzymes requiresspecific effectors. Eur J Biochem 182:125–128.

22. Eletsky A, et al. (2012) Solution NMR structure of yeast succinate dehydrogenase

flavinylation factor Sdh5 reveals a putative Sdh1 binding site. Biochemistry 51:8475–8477.

23. Bezawork-Geleta A, Saiyed T, Dougan DA, Truscott KN (2014) Mitochondrial matrix

proteostasis is linked to hereditary paraganglioma: LON-mediated turnover of thehuman flavinylation factor SDH5 is regulated by its interaction with SDHA. FASEB J

28:1794–1804.24. Ruprecht J, Yankovskaya V, Maklashina E, Iwata S, Cecchini G (2009) Structure of Es-

cherichia coli succinate:quinone oxidoreductase with an occupied and empty quinone-

binding site. J Biol Chem 284:29836–29846.25. Hayward S, Berendsen HJ (1998) Systematic analysis of domain motions in proteins

from conformational change: New results on citrate synthase and T4 lysozyme.

Proteins 30:144–154.26. Bossi RT, Negri A, Tedeschi G, Mattevi A (2002) Structure of FAD-bound L-aspartate

oxidase: Insight into substrate specificity and catalysis. Biochemistry 41:3018–3024.27. Stoltz M, Brandsch R (1998) The conformational change induced by FAD in covalently

flavinylated 6-hydroxy-D-nicotine oxidase does not require (8alpha)FAD-(N3)histidyl

bond formation. J Biochem 123:445–449.28. Koetter JW, Schulz GE (2005) Crystal structure of 6-hydroxy-D-nicotine oxidase from

Arthrobacter nicotinovorans. J Mol Biol 352:418–428.29. Fraaije MW, van Den Heuvel RH, van Berkel WJ, Mattevi A (2000) Structural analysis

of flavinylation in vanillyl-alcohol oxidase. J Biol Chem 275:38654–38658.30. Lim K, et al. (2005) Crystal structure of the YgfY from Escherichia coli, a protein that

may be involved in transcriptional regulation. Proteins 58:759–763.31. Heuts DP, Scrutton NS, McIntire WS, Fraaije MW (2009) What’s in a covalent bond? On

the role and formation of covalently bound flavin cofactors. FEBS J 276:3405–3427.32. McNeil MB, Fineran PC (2013) The conserved RGxxE motif of the bacterial FAD as-

sembly factor SdhE is required for succinate dehydrogenase flavinylation and activity.Biochemistry 52:7628–7640.

33. Chen J, Sawyer N, Regan L (2013) Protein-protein interactions: General trends in the

relationship between binding affinity and interfacial buried surface area. Protein Sci22:510–515.

34. Cecchini G (2003) Function and structure of complex II of the respiratory chain. AnnuRev Biochem 72:77–109.

35. Sharma P, Maklashina E, Cecchini G, Iverson TM (2018) Crystal structure of an as-

sembly intermediate of respiratory complex II. Nat Commun 9:274.36. Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected in os-

cillation mode. Methods Enzymol 276:307–326.37. McCoy AJ, et al. (2007) Phaser crystallographic software. J Appl Cryst 40:658–674.38. Murshudov GN, et al. (2011) REFMAC5 for the refinement of macromolecular crystal

structures. Acta Crystallogr D Biol Crystallogr 67:355–367.39. Strong M, et al. (2006) Toward the structural genomics of complexes: Crystal structure

of a PE/PPE protein complex fromMycobacterium tuberculosis. Proc Natl Acad Sci USA103:8060–8065.

40. Emsley P, Cowtan K (2004) Coot: Model-building tools for molecular graphics. Acta

Crystallogr D Biol Crystallogr 60:2126–2132.41. Chen VB, et al. (2010) MolProbity: All-atom structure validation for macromolecular

crystallography. Acta Crystallogr D Biol Crystallogr 66:12–21.

Maher et al. PNAS | March 20, 2018 | vol. 115 | no. 12 | 2987

BIOCH

EMISTR

YBIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

June

20,

202

0