schiffels et al-2015-biotechnology and bioengineering

A Flexible Toolbox to Study Protein-AssistedMetalloenzyme Assembly In Vitro

Johannes Schiffels, Thorsten Selmer

From the Aachen University of Applied Sciences, Campus Juelich, Department of

Chemistry and Biotechnology, Heinrich-Mussmann-Str. 1, D-52428 Juelich, Germany

ABSTRACT: A number of metalloenzymes harbor uniquecofactors, which are incorporated into the apo-enzymes viaprotein-assisted maturation. In the case of [NiFe]-hydrogenases,minimally seven maturation factors (HypABCDEF and a specificendopeptidase) are involved, making these enzymes an excellentexample for studying metallocenter assembly in general. Here, wedescribe an innovative toolbox to study maturation involvingmultiple putative gene products. The two core elements of thesystem are a modular, combinatorial cloning system and a cell-freematuration system, which is based on recombinant Escherichia coliextracts and/or purified maturases. Taking maturation of thesoluble, oxygen-tolerant [NiFe]-hydrogenase (SH) from Cupriavi-dus necator as an example, the capacities of the toolbox areillustrated. In total 18 genes from C. necator were analyzed,including four SH-structural genes, the SH-dedicated hyp-genesand a second set of hyp-genes putatively involved in maturation ofthe Actinobacterium-like hydrogenase (AH). The two hyp-sets wereeither expressed in their entirety from single vectors or split intofunctional modules, which enabled flexible approaches toinvestigate limitations, specificities and the capabilities ofindividual constituents to functionally substitute each other.Affinity-tagged Hyp-Proteins were used in pull-down experimentsto demonstrate direct interactions between dedicated or non-related constituents. The dedicated Hyp-set from C. necatorexhibited the highest maturation efficiency in vitro. Constituents ofnon-related maturation machineries were found to interact withand to accomplish partial activation of SH. In contrast tohomologues of the Hyp-family, omission of the SH-specificendopeptidase HoxW completely abolished in vitro maturation.We detected stoichiometric imbalances inside the recombinantproduction system, which point to limitations by the cyanylationcomplex HypEF and the premature subunit HoxH. Purification ofHoxW revealed strong indications for the presence of a putative[4Fe-4S]-cluster, which is unique among this class of maturases.Results are discussed in the context of [NiFe]-hydrogenasematuration, and in light of the capacity of the novel toolbox.Biotechnol. Bioeng. 2015;112: 2360–2372.� 2015 Wiley Periodicals, Inc.KEYWORDS: multigene expression; combinatorial cloning;metalloenzymes; maturation in vitro; [NiFe]-hydrogenases; Hyp-proteins

Introduction

Several enzymes require cofactors like organic compounds, metal ionsor metalloorganic groups to become activated. In many cases, thetransformation of an apo- into a holoenzyme is mediated by one ormore auxiliary proteins. Whereas a range of pleiotropic post-translational machineries are common to many organisms (e. g., theiron-sulfur-cluster machineries), some enzymes require highlyspecialized chaperones or maturases for assembly and insertion ofa prosthetic group (e. g., [FeFe]- and [NiFe]-hydrogenases) (B€ocket al.,2006; Forzi and Sawers 2007; Peters et al., 2014). Tight association andcomplex formation between individual maturation factors and thetarget enzyme are often essential requirements in order to achieveproper enzyme maturation and functional production.

[NiFe]-Hydrogenases catalyze the interconversion of molecularhydrogen into protons and electrons. Maturation of these enzymesrequires at least six gene products of the Hyp-family (HypABCDEF) andone specific endopeptidase for maturation (Peters et al., 2014) (Fig. 1A).Homologues of the hyp-genes are found in all genomes harboring[NiFe]-hydrogenase structural genes. Current biochemical models ofHyp-mediated maturation are composites based on the researchconducted predominantly on [NiFe]-hydrogenases from three differentorganisms, namely Escherichia coli, Cupriavidus necator (formerlyRalstonia eutropha H16), and Thermococcus kodakaraensis. Hyp-mediated [NiFe]-hydrogenase maturation in these organisms, althoughacting on different target enzymes, is highly similar (B€ock et al., 2006;Forzi and Sawers 2007; Peters et al., 2014; Watanabe et al., 2007).

The biochemical functions of Hyp-proteins are proposed to behighly conserved. Nevertheless, some organisms contain two ormore copies of hyp-genes. In E. coli, maturation of hydrogenases1 and 2 requires hybF and hybG, which are homologues of hypAand hypC (Blokesch et al., 2001; Hube et al., 2002). Cupriavidusnecator, a facultative chemolithoautotrophic “Knallgas” bacte-rium, harbors at least four [NiFe]-hydrogenases: a solublehydrogenase (SH), which links hydrogen oxidation to NADþ

reduction; a respiratory membrane-bound hydrogenase (MBH),shuttling H2-derived electrons directly into the quinone pool; aregulatory hydrogenase (RH), which acts as an H2 sensor involvedin transcriptional activation of both SH and MBH operons in thepresence of H2, and an Actinobacterium-like hydrogenase (AH)with an as yet unknown physiological function. Structural genesencoding these enzymes are organized in three clusters, which arelocated on megaplasmid pHG1 (Fig. 1B) (Schwartz et al., 2003).

Correspondence to: Thorsten Selmer

Received 19 February 2015; Accepted 11 May 2015

Accepted manuscript online 20 May 2015;

Article first published online 30 June 2015 in Wiley Online Library

(http://onlinelibrary.wiley.com/doi/10.1002/bit.25658/abstract).

DOI 10.1002/bit.25658

ARTICLE

2360 Biotechnology and Bioengineering, Vol. 112, No. 11, November, 2015 � 2015 Wiley Periodicals, Inc.

The MBH- and AH-cluster each contain an adjacent set of hyp-genes (hypA1B1F1C1D1E1X and hypF3C2D2E2A3B3, respec-tively). A partial set of hyp-genes (hypA2B2F2) is found as part ofthe SH-cluster. It has been demonstrated by knockout studies,that hypA2, hypB2, and hypF2 are not essential for maturation ofSH and that the hyp-genes located in the MBH-operon aresufficient to assemble and transfer [NiFe] active site constituentsto the large hydrogenase subunits HoxH, HoxG, and HoxC of SH,MBH, and RH, respectively (Wolf et al., 1998). Transcription of

the AH-operon appears to be uncoupled from SH-, MBH-, andRH-regulation, since hypABF-double mutants are impaired in H2

uptake, as are mutants lacking hypC1, hypD1, or hypE1. Twoindependent studies demonstrated, however, that the AH-operonencodes functional hydrogenase and accessory gene products(Sch€afer et al., 2013; Schiffels et al., 2013).We recently demonstrated by heterologous expression of 13–14

genes from C. necator, that the SH is functionally produced andactivated in recombinant E. coli- strains at high yields (Schiffels et al.,

Figure 1. Current model of Hyp-mediated [NiFe]-hydrogenase maturation (A) and genomic distribution of hyp-genes in C. necator and E. coli (B). Interactions and complex

formations between Hyp-proteins are required for stepwise assembly of the NiFe-(CN)2CO-cofactor (Blokesch and B€ock 2002). This is schematically represented by using the

example of HoxH, the hydrogenase subunit of the C. necator SH (A). Initially, the small chaperone HypC associates with the N-terminal region of HoxH, a complexwhich is maintained

until insertion of the [NiFe]-site is completed (step 1) (Drapal and B€ock 1998; Magalon and B€ock 2000). For coordination of the iron group, the scaffold protein HypD forms a complex

with a second copy of HypC (B€urstel et al., 2012; Stripp et al., 2013). The origin of the carbonyl ligand is presently unclear. Delivery of the two cyanide ligands to Fe-HypCD requires the

interaction of the latter with ‘‘modified’’ (cyanylated) HypE. CN--transfer from HypE to the HypCD-coordinated iron is proposed to require redox chemistry, which is effected by a

conserved [4Fe-4S]-cluster and two intramolecular disulfide bonds present in HypD (Stripp et al., 2013; Watanabe et al., 2007). Modification of HypE involves interaction with HypF,

which mobilizes a carbamate group from carbamoyl phosphate in an ATP-dependent step, and transfers it to HypE (Blokesch et al., 2004). ATP-dependent dehydration to the stable

thiocyanate is catalyzed by HypE (step 2). Nickel-delivery to HoxH is mediated by a complex between HypA and HypB and requires GTP (step 3) (Cai et al., 2011), followed by

dissociation of HypC. HoxW recognizes the incorporated nickel ion and a specific binding site (Thiemermann et al., 1996). It cleaves 24 residues off HoxH (step 4), which in turn

internalizes the [NiFe]-site next to the HoxY-facing contact site to afford the active holo-enzyme. The hexameric structure of the SH holoenzyme is completed by the subunits

HoxYUF(I2) (Burgdorf et al., 2005). || Hydrogenase-associated genes in C. necator are organized in three operons on the megaplasmid pHG1 (B) (Schwartz et al., 2003). Two copies of

hypCDE- and three copies of hypABF-genes are present. TheMBH/SH-associated genes hypA2B2C1D1E1F2 (termed ‘‘M1’’) and the AH-associated genes hypA3B3C2D2E2F3 (‘‘M2’’)

were cloned in this study. Beside the SH-genes hoxFUYH, the hoxW-gene, which encodes the SH-specific endopeptidase, and hypXwere included in all cases. The hoxI-gene was

omitted in this study. E. coli harbors one set of pleiotropic Hyp-proteins (hypABCDEfhlA operon). The hypF gene is located in the divergent hyc-operon encoding the subunits of the

formate-hydrogen-lyase/Hyd-3 (FHL) complex. Genes without relevance for this study are indicated in dark grey.

Schiffels and Selmer: A Toolbox to Study Metalloenzyme Maturation 2361

Biotechnology and Bioengineering

2013). Here, we explored different sets of Hyp-proteins for their basicrequirements, specificities and the capability to functionallysubstitute their respective homologues. The SH was used as themodel enzyme in a study aimed at the design of a toolbox for in vitroanalysis of enzyme maturation involving a large number of putativematuration factors in an E. coli based heterologous system (Fig. 2).Using [NiFe]-hydrogenase maturation as an example, we exploitedthe presented toolbox to address requirements, limitations andcomplementation capabilities of two independent hyp-gene sets.

Material and Methods

Gene Cloning and Expression Strategies

Cloning and assembly of C. necator genes was performed asdescribed previously (Schiffels et al., 2013) by using the E. colistrain DH5a (Invitrogen). The plasmid constructs used for theproduction and purification of complexes or for preparation of cell-free maturation (CFM) extracts are listed in Table I andSupplementary Table SI. For production and purification ofheterogenous Hyp-complex intermediates, respective genes werefused with the StrepII-tag sequence at their 3’- or 5’-end usingfusion plasmids of the pFxT7-series (Schiffels et al., 2013). Entryplasmids (pE-) carrying two or more genes under control ofindividual T7-promoters and -terminators were used as expressionconstructs (Table I and Supplementary Table SI). All genes exceptfor hoxW were expressed by employing a lactose-basedautoinduction strategy (Schiffels et al., 2013) at varying growthtemperatures (Supplementary Table SI). For production of StrepII-HoxW, the StarGate

1

expression vector pASG-5 (IBA, G€ottingen,Germany) was used, which places the hoxW-gene under control ofthe tetracycline-regulon. Expression of the gene was induced byaddition of anhydrotetracycline (AHT, 200mg � L�1) at an opticaldensity (OD578) of 0.4, following incubation at 22�C for 18 h inthe dark. In all cases, cells were pelleted by centrifugation, washedtwice with the buffer used for cell lysis and stored at –80�C untiluse.

Purification of HoxW and Hyp-Complexes

Wet cell pellets were resuspended in dedicated buffers (2mL buffer pergram cells; Supplementary Table I) and disrupted by sonication at 0�Cfor 10min with duty cycles of 50% and power set to 60%. Lysates werecleared by ultracentrifugation at 140,000 � g and 2�C for 45min. Thesoluble extracts were loaded onto pre-equilibrated 5mL StrepTactinSuperflow

1

HPR-columns (IBA, G€ottingen, Germany) by a peristalticpump. Following washing (3–5 column volumes), bound proteins orcomplexes were eluted with buffer containing 2.5mM D-desthiobiotin.Complexes containingHypD-homologueswere purifiedunder aerobicoranaerobic conditions. In the latter case, cell opening and purificationsteps were performed in an anaerobic glove box (Coy) with anatmosphere of 2.5–5% H2 in N2. All purifications were performed at20�C. Following purification, target proteins/complexes were concen-trated by using VivaSpin 6/20 devices (Sartorius, G€ottingen, Germany)withmolecular weight cut-offs of 10, 30, 50, or 100 kDa, respectively, andstored in aliquots at –80�C. Purified HypCDE-complex was furtheranalyzed by size-exclusion chromatography with a Superdex 200 HR 10/

300 column (GE Healthcare) to determine molecular masses. 50mMTris/Cl pH 8.0, containing 75mM NaCl and 0.1mM DTTwas used forpre-equilibration and analysis. Thyroglobulin (669 kDa), ferritin(440 kDa), catalase (232 kDa), aldolase (158 kDa), and bovine serumalbumin (BSA, 67 kDa) were used as standards (Gel Filtration HighMolecular Weight Calibration Kit, GE Healthcare).

Analytical Techniques

Preparations were routinely analyzed by SDS–PAGE (Laemmli 1970).Clear-Native PAGE (CN–PAGE) of purified HypCDE-complex wasperformed according to (Sch€agger et al., 1994) by using a NativeProteinMarker Kit (GE Healthcare) as standard. Proteins separated bySDS- or N-PAGE were stained with Coomassie Brilliant blue (Wilson1979). Protein species transferred to nitrocellulose membranes(Towbin et al., 1979) were selectively marked by using StrepTactin-horseradish-peroxidase-conjugate and visualized by chemilumines-cence detection according tomanufacturer’s protocols (IBA, G€ottingen,Germany). Colorimetric determination of non-heme iron in purifiedStrepII–HoxW preparations was carried out according to (Pierik et al.,1992). Protein concentrations were determined by the method ofBradford (Bradford 1976) using BSA as standard.

Preparation of Extracts for Cell-Free Maturation Assays

E. coli BL21StarTM (DE3) was used for the generation of sourcematerial for cell-free maturation (CFM) extracts. The plasmidconstructs transformed were either pE.M1 (harboring hoxW andhypA2B2C1D1E1F2X), pE.M2 (harboring hoxW and hypA3B3C2-D2E2F3X) or pE-derivatives with maturase gene pairs under controlof individual T7-promoters and -terminators (Table I) (Schiffelset al., 2013). Transformants were subjected to autoinductionaccording to (Schiffels et al., 2013) with a modification of theNiCl2 concentration in the medium (25mM instead of 1mM). Cellswere harvested 36 h post-induction (growth temperature: 20�C),washed twice with the CFM-assay buffer, and stored at –80�C. Forextract preparation, cells were resuspended in CFM-assay buffercontaining 5mM MgCl2 as well as 0.05mM PMSF and disrupted in5–6 consecutive freeze/thaw-cycles (–80/25�C, each 30min). Lysateswere cleared by ultracentrifugation at 140,000 � g and 2�C for 45min.The protein concentration of all soluble CFM-extracts was adjusted to20 mg �mL�1. Aliquots at 250–500mL were stored at –80�C.

Cell-Free Maturation Assays

Standard assays contained 15% (v/v) SH-extract and either 15% (v/v) of M1- or M2-extract or each 7.5% (v/v) of Hyp-pair extracts.Reactions were performed in a total volume of 100mL using 1.5mLpolypropylene tubes. The following buffer systems were tested:50mM KPi (pH 6.5–7.5), 50mM MOPS (pH 6.5–7.5), 50mMHEPES (pH 7.0–8.0), and 50mM Tris/Cl (pH 7.5–8.5). Theconcentration of small molecules was initially varied to determineoptimized maturation in vitro. Standard assays contained 2.5 mMDTT, 2.5mM ATP, 50mMcarbamoyl phosphate (CP), 50mM FeSO4,20mM NiCl2, and 10mM FMN in 50mM Tris/Cl buffer pH 8.0.5 mM MgCl2 were included in all assays. Physical parameters werevaried, too. Standard assays were run at 30�C (SH-dedicated set;

2362 Biotechnology and Bioengineering, Vol. 112, No. 11, November, 2015

M1) or 37�C (AH-dedicated set; M2). Reactions were stopped after50 minutes. Assays were started by addition of SH extract. Samplesof 10mLwere withdrawn and placed on ice, diluted were necessary,and measured for SH activity.

SH Activity Assay

Soluble Hydrogenase (SH) activity (H2:NADþ) was measured

spectrophotometrically in an anaerobic glove box according to (Schiffelset al., 2013). Specific activities were calculated based on the SH extractfraction (final protein concentration in standard assays: 3mg �mL�1).

Results and Discussion

Experimental Design, Gene Selection and Cloning

Recently, we described the application of a combinatorial cloningand expression system for high-yield functional production of

recombinant SH from C. necator in E. coli (Schiffels et al., 2013).Based on the present knowledge on [NiFe]-hydrogenase maturationand taking into account the combinatorial power of the recentlydeveloped tool set, we designed three basic experiments: Firstly, acell-free system based on recombinant extracts was established,which is capable of producing functional tetrameric SH fromindividually produced, inactive apo-enzyme in vitro. This setup wasused to identify optimum conditions for hydrogenase maturation.Secondly, the capacity of non-native auxiliary proteins to contributeto SH maturation was analyzed. Here, assays were either enrichedby individual parts of the maturation machinery in order to identifylimitations (“spiking experiments”), or depleted of dedicatedmaturation factors to explore complementation capabilitiesexhibited by paralogous gene products from the AH-set (M2).Finally, individual affinity-tagged Hyp-proteins (baits) wereproduced together with potential complex partners and sub-sequently pulled down in order to investigate complex formationsand intermediate steps in the maturation sequence. Thus identifiedand purified complexes were functionally tested using the in vitro

Figure 2. Scheme highlighting the three different experimental sets conducted in this study. In experimental set 1, the general parameters for maximized in vitro maturation

efficiency were determined. The SH structural genes as well as the maturases were each produced from single plasmids (pSH4.wt, pE.M1 and pE.M2), yielding recombinant extracts

containing the complete sets (‘‘SH extract’’, ‘‘full M1,’’ ‘‘full M2’’). The conditions determined in experimental set 1 were subsequently applied for different maturation studies in

experimental set 2. Here, the maturases were not produced in one cell. Instead, plasmids each designed for expression of M1/M2-hyp-pairs were used to produce Hyp-pair extracts

used in spiking, deletion and substitution experiments. The XW-extract contained HypX and HoxW and was crucially required alongside the Hyp-analogues. ‘‘SH extract’’ was used

in all experiments. Spiking experiments contained the ‘‘full M1/M2’’ extracts from set 1 as well, which were supplemented with portions of Hyp-pair extracts. Experimental set 3was

conducted in order to gain insights into interactions between Hyp-proteins. Parts of theM1/M2 maturation sets were produced alongside an affinity-tagged Hyp-protein, which was

used as a bait. Following expression, the resulting cell-free soluble extracts were applied to StrepTactin-affinity chromatography, thereby enabling the isolation of tagged proteins

with tightly bound coelutes. Purified species were analyzed and functionally tested using in vitro maturation assays.



Table

I.Strainsandplasm

idsusedin

thisstudy.

Category

Genotype/relevant

characteristics

Source

Strains

E.coliDH5a

(cloning

strain)

fhuA

2D(argF-lacZ)U169phoA

glnV

44F80

D(lacZ)M15

gyrA96

recA1relA1endA

1thi-1

hsdR

17Invitrogen

E.coliBL21Star

(DE3)

dcm

ompT

hsdS(rB- m

B- )gal

Invitrogen

Plasm

ids

a.Expression

constructsused

forproductionandsubsequent

purificationof

Hyp-com

plexes

pE-[wt]hypA

2[5’]hypB2

hypA

2,StrepII-hypB2;Kan

R;ColE1ori

Thisstudy

pE-[3’]hypC1[wt]hypD

1hypC1-StrepII,hypD

1;Kan

R;ColE1ori

Thisstudy

pE-[3’]hypC1[wt]hoxH

hypC1-StrepII,hoxH

;Kan

R;ColE1ori

Thisstudy


1E1F2

hypC1-StrepII,hypD

1,hypE1,hypF2;Kan

R;ColE1ori

Thisstudy


2hypC1-StrepII,hypD

2;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypA

3[5’]hypB3

hypA

3,StrepII-hypB3;Kan

R;ColE1ori

Thisstudy


2hypC2-StrepII,hypD

2;Kan

R;ColE1ori

Thisstudy

pE-[3’]hypC2[wt]hoxH

hypC2-StrepII,hoxH

;Kan

R;ColE1ori

Thisstudy


2E2F3

hypC2-StrepII,hypD

2,hypE2,hypF3;Kan

R;ColE1ori

Thisstudy

pASG(5’-StrepII)-hoxW

hoxW

-StrepII;AmpR;ColE1ori;Tet-Regulon

Thisstudy

b.Expression

constructsused

forpreparationof

cell-free

maturationextracts

�

pSH4.wt(SHsubunitsalone;“SHextract“)

hoxY,hoxH,hoxU,hoxF;

Kan

R;ColE1ori

Schiffelsetal.,2013

pE.M1(CnmaturationsetM1;“fullM1extract“)

hypF2,hypE1,hypX,hoxW,hypC1,hypD

1,hypA

2,hypB2;Kan

R;ColE1ori

Schiffelsetal.,2013

pE.M2(CnmaturationsetM2þ

hypX

/hoxW;“fullM2extract“)

hypX,hoxW,hypE2,hypF3,hypC2,hypD

2,hypA

3,hypB3;Kan

R;ColE1ori

Schiffelsetal.,2013

pE-[wt]hypA

2B2(“M1-ABextract“)

hypA

2,hypB2;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypC1D

1(“M1-CD

extract“)

hypC1,hypD

1;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypE1F2(“M1-EF

extract“)

hypE1,hypF2;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypX/hoxW

(“XW

extract“)

hypX,hoxW;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypA

3B3(“M2-ABextract“)

hypA

3,hypB3;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypC2D

2(“M2-CD

extract“)

hypC2,hypD

2;Kan

R;ColE1ori

Thisstudy

pE-[wt]hypE2F3(“M1-EF

extract“)

hypE2,hypF3;Kan

R;ColE1ori

Thisstudy

� Nam

esof

thecorrespondingCFM-extractsaregivenin

parentheses.


system, too. Figure 2 gives an overview of the individualexperimental sets conducted.The hyp-genes of two distinct operons present on the pHG1

megaplasmid of Cupriavidus necator were studied: The geneshypA2B2C1D1E1F2, together with hoxW (encoding the SH-specificendopeptidase) and hypX (which encodes a protein putatively relevantfor the maturation of oxygen-tolerant hydrogenases), formed the SH-dedicated maturase set, termed “M1” (Dernedde et al., 1996; Schiffelset al., 2013; Wolf et al., 1998). A second set, comprising the putativelyAH-specific hypA3B3C2D2E2F3 together with hoxW and hypX, istermed “M2” (Fig. 1B). The endogenous hyp-genes of E. coli formedthe third independent maturase set. We recently observed that Hyp-activity is significantly stimulated under semi-anaerobic conditionspresent in late stages of growth when cells are subjected to lactose-based autoinduction (Schiffels et al., 2013). In the course of these trials,we detected reduced but significant levels of SH activity in extractsfrom strains lacking parts of the C. necator maturation machinery(DHypAB-, DHypCD-, and DHypEF-strains), indicating that E. coliHyp-homologues are present and capable to take over the function ofthe omitted maturases under these conditions. The autoinductionstrategy was applied for production of source cells for cell-freematuration (CFM) extracts in the present study, too.

An Extract-Based Cell-Free SH Maturation System

In order to establish an efficient in vitro maturation system bycombination of SH- and maturase extracts, small molecules (ATP,GTP, carbamoyl phosphate, DTT, FeSO4, NiCl2, and FMN) wereprovided at defined concentrations. Omission of these moleculesexcept for GTP reduced overall maturation, although the includedcell extracts apparently provided sufficient amounts of thecompounds to achieve partial activation (0.18 U �mg�1 ascompared to 1.1–1.3 U �mg�1 achieved by addition of the smallmolecules mix; Tables II and III). The buffer composition hadsignificant effects on the maturation assay: Tris-HCl (50mM, pH8.0; 100%) gave the best results, followed by MOPS (50 mM, pH7.5; �85%), and KPi (50 mM, pH 7.0; �70%), whereasmaturation efficiencies dropped substantially at pH values ofbelow 7.0 or above 8.5.In contrast to previous reports on the E. coliHyp-proteins, whose

concerted activity was abolished upon oxygen exposure (Sobohet al., 2010; Soboh et al., 2014), anoxic conditions did not stimulateSH maturation using M1- or M2- maturases in vitro. In case of theE. coli Hyp-proteins, activity was increased by 20% when oxygenwas excluded (Table III). M1-mediated activities dropped by morethan 50%, however, when the reducing agent DTT (2.5 mM) wasomitted from the reaction mixture (Table II). Notably, in theirisolated form, C. necator Hyp-complexes containing HypD1 wereprone to inactivation by O2 as described later on.Stripp and coworkers proposed a sequence of events for

HypCDEF-mediated assembly of the iron-cofactor and suggestedthat CO-ligation precedes cyanylation (Stripp et al., 2014). Aninitial HypCD-coordinated oxygen-labile Fe-CO-intermediate isassumed, which is stabilized against O2-induced oxidation inthe course of cyanide ligation. In the experimental setupemployed in this study, the detrimental effects of oxygen mightbe minimized for two reasons: Firstly, the cells used for extract

preparation were harvested during semi-anaerobic stationarygrowth phases, when activity of the Hyp-machinery was foundto be substantially increased (Schiffels et al., 2013). Therefore, itis reasonable to assume that extracts derived from these cellscontained HypCDE-complexes loaded with fully assembled Fe-(CN)2CO-groups or intermediates therefrom. Heterologousexpression further accounts for high levels of the activecomplexes, thus buffering a time-dependent oxidative inacti-vation of constituents. The reduced extract environment incombination with DTT treatment obviously protects the labileHypCDE-complex, which is rapidly inactivated as aerobicallyisolated. A minor fraction of the HoxH-molecules present insidethe SH extract might harbor a functional [NiFe]-cofactor due tothe activity of the E. coli maturases, too. In line with thisnotion, the sole addition of HoxW without additional C. necatormaturases led to the partial activation of SH extract (0.15–0.18U �mg�1; Table III). Secondly, the oxygen-tolerance of the SHneeds to be taken into account. Since the enzyme is producedand active under aerobic conditions in C. necator, cofactorassembly takes place in the presence of fluctuating O2-levels,too. Whether HoxH, the acceptor of the cofactor, has a role inshielding the ‘construction site’ prior to its delivery or themechanism by which the iron group is assembled differs fromthe one in E. coli remains to be elucidated.The maturation efficiency of the M1 set peaked at 30�C, which

corresponds to the temperature optimum of C. necator (Fig. 3A). At30�C, SH maturation using the M1 set in a Tris-buffered assaycontaining ATP, CP, DTT, FeSO4, NiCl2, and FMN yielded 1.29U �mg�1 (Table III), which is in range of the specific activitiesobtained by SH production and maturation in vivo. The activitymaximum was reached after 45–50min (Fig. 3B).

Limiting Factors for Cell-Free SH Maturation

For SHmaturation in vitro, three sets of Hyp-proteins (“M1”, “M2”,and the E. coli endogenous system) were tested, each completed by

Table II. Effects of substrate omissions on SH maturation efficiency in

vitro.

SH activityAssay U �mg�1 a

Complete M1 aerobicb 1.18� 0.19Complete M1 (without small molecules) 0.18� 0.05Complete M1 without DTT 0.56� 0.12Complete M1 without ATP 0.71� 0.11Complete M1 without GTPc 1.15� 0.21Complete M1 without FeSO4 0.37� 0.07Complete M1 without NiCl2 0.55� 0.1Complete M1 without FMN 0.59� 0.12Complete M1 without CP 0.74� 0.08

aResults are mean values of three independent measurements. Indicated errorsrepresent standard deviations. 1 Unit¼H2-mediated reduction of 1mmol NAD

þ perminute. Specific activities were calculated based on the SH-extract protein content.

bGTP was included in the experimental setup here, but not in standard CFM-assays (see methods section). The following concentrations were routinely applied:5mM MgCl2, 2.5 mM DTT, 2.5 mM ATP, 0.5 mM GTP, 50mM CP, 50mM FeSO4,20mM NiCl2, and 10mM FMN. Reactions were performed in 50mM Tris/Cl pH 8.0(total volume: 100mL) for 50min at 30�C.



Table

III.

Main

cell-freeSHmaturationassays.

Assay

Extractfractions/com

ponentsa

SHactivity

c

U�m

g�1

Con

trol

assays

forM1-,M2-,an

dE.

colimaturases

CompleteM1(control)

15%

(v/v)SH

þ15%

(v/v)M1full

1.29

�0.16

CompleteM1anaerobicb

15%

(v/v)SH

þ15%

(v/v)M1full

1�0.13

M1usingcombinedHyp-pairextracts

15%

(v/v)SH

þeach

7.5%

(v/v)M1-AB,

M1-CD

,M1-EF,XW

1.34

�0.26

M1usingpurifiedM1complexes

15%

(v/v)SH

þeach

50mgpurifiedHypA2B2,HypC1D1E1(anaerobicallypurified),HoxW

1.26

�0.21

CompleteM2(control)

15%

(v/v)SH

þ15%

(v/v)M2full

0.44

�0.07

CompleteM2anaerobicb

15%

(v/v)SH

þ15%

(v/v)M2full

0.44

�0.08

M2usingcombinedHyp-pairextracts

15%

(v/v)SH

þeach

7.5%

(v/v)M2-AB,

M2-CD

,M2-EF,XW

0.45

�0.1

SHw/o

Cnmaturases

15%

(v/v)SH

<0.02

SHþ

HypX/HoxW

(E.coliHyp-protein

control)

15%

(v/v)SH

þ15%

(v/v)XW

0.15

�0.03

SHþ

HypX/HoxW

anaerobicb

15%

(v/v)SH

þ15%

(v/v)XW

0.18

�0.04

M1-assays

withvaryingextractfraction

sCompleteM1(1xSH

,2x

M1)

15%

(v/v)SH

þ30%

(v/v)M1full

1.18

�0.2

CompleteM1(2xSH

,1x

M1)

30%

(v/v)SH

þ15%

(v/v)M1full

0.86

�0.12

CompleteM1(3xSH

,3x

M1)

45%

(v/v)SH

þ45%

(v/v)M1full

0.99

�0.12

M1-

andM2-control

assays

withou

tHyp

Xan

dHox

WM1w/o

HypX/HoxW

15%

(v/v)SH

þeach

7.5%

(v/v)M1-AB,

M1-CD

,M1-EF

<0.02

M2w/o

HypX/HoxW

15%

(v/v)SH

þeach

7.5%

(v/v)M2-AB,

M2-CD

,M2-EF

<0.02

a For

details

regardingmaturationextractsandpurified

complexes,see

TableI,Supplementary

TableSIandthemethods

section.Thesm

allm

olecules

mixwas

addedin

each

case.

b Anaerobiosiswas

achieved

byanoxicpreparationofSH

-andfullM1-extracts.A

ssay

buffersandsm

allm

olecules

stocksolutions

wereequilibratedwith

thegloveboxatmosphereovernightpriorto

CFM-assays.

c Resultsaremeanvaluesofthreeindependentm

easurements.Indicated

errorsrepresentstandarddeviations.1

Unit¼

H2-mediatedreductionof1mmolNADþperm

inute.Specificactivitieswerecalculated

basedon

theSH

-extractprotein

content.


HypX and the SH-specific endopeptidase HoxW. When HoxW wasabsent, SH maturation did not exceed background values(Table III). Recent studies on SH maturation in vivo suggestedthat HoxW is the only specific part of the maturation machinerywithout functional substitutes (Schiffels et al., 2013). HypX, on theother hand, appears to be dispensable for in vitro maturation. Thiswas proven when SH extract was mixed with purified constituents,which were fully functional in absence of HypX (see below).Notably, it has recently been observed that the absence of HypXcauses a �25% activity decrease in vivo and further renders therecombinant SH present in cell extracts more instable towardsprolonged O2 exposure (Schiffels et al., 2013). Therefore, HypXmight only be required under unique conditions, which arecircumvented by the experimental setup described here.

When HoxW was present, each of the Hyp-sets tested was capableof at least partial SH activation in vitro. Compared to M1- controlassays containing the SH-dedicatedmaturase set, the AH-set (M2) orE. coli Hyp-homologues were less effective, yielding activities up to0.45 U �mg�1 (35%) and 0.18 U �mg�1 (14%), respectively(Table III). The reduced SH activities achieved with AH-derived orE. coli-maturases point towards a kinetic limitation originating fromdead-end catalytic intermediates, which are formed between non-native maturation factors and are unable to complete the fullsequence of enzyme activation. Notably, SH maturation using AH-maturases (M2) peaked at 37�C instead of 30�C observed for the SH-set M1, which supports the notion that the gene products encoded inthe AH-operon are uncommon to the cellular environment of C.necator. In line with this assumption, the AH-enzyme itself shows amaximum activity greatly above the temperatures naturallyencountered by C. necator (Sch€afer et al., 2013).

For practical reasons, we used fractions of 15% (v/v) of bothSH- and full maturation extracts in standard assays. Raising theM1-maturase fraction to 30% (v/v) did not yield higher SHactivities. By elevating the fraction of SH extract inside the assays,the overall activity was increased, whereas specific activities

(calculated for the amount of SH extract used) were onlymarginally affected (Table III). Hence, the levels of maturasespresent in the M1-extract are sufficient to completely activate theapo-enzyme provided, indicating that the level of non-assembledSH subunits and thus, the amount of SH extract, limits theactivity in vitro.The M1-extract contained the complete set of SH-dedicated,

native maturases produced by expression of the genes from a singleplasmid. Thus, modifying the amount of M1-extract may reveal akinetic limitation of the Hyp-machinery in its entirety, but does notcover possible bottlenecks caused by individual maturases presentin limiting, sub-stoichiometric amounts. Therefore, in addition tocomplete M1- or M2-sets being produced in one cell, we expressedhyp-pairs as binary modules. The corresponding cell extractscontained functional pairs of maturases, e.g., for Fe-clusterinsertion (M1-CD), nickel insertion (M1-AB), cyanylation (M1-EF), or sequence completion (XW) (Table I and Fig. 2). Byenrichment of the complete M1-extract in standard assays withdefined volumes of Hyp-pair extracts, limitations caused byindividual constituents were addressed. Apparently, only the Fe-cluster insertion complex (M1-CD) was sufficiently present in theoriginal assays, whereas the addition of other constituents improvedoverall SH maturation substantially (Fig. 4A). Addition of thecyanylation complex (M1-EF, HypE1F2) resulted in the mostpronounced stimulation, increasing SH-activity by a factor of 3.5(3.84 U �mg�1). Similar experiments were carried out for the AH-derived maturases (using M2-CD-, M2-AB-, M2-EF-extracts andXW-extract; Table I and Fig. 2). Contrary to the M1-spikingexperiments, minor effects were observed and activities were in thesame range as the control assay, regardless of the addedcomponents (Fig. 4B). These findings support the previous notionof a kinetic effect. Apparently, when non-native maturases associatewith HoxH, slowly reacting intermediates are formed, which existlong enough to react in a non-productive manner and causeirreversibly inactivated side products rather than activated enzyme.The possibility that Hyp-pairs derived from the AH-set were

Figure 3. Physical conditions for SHmaturation in vitro.A) Influence of the incubation temperature on SHmaturation efficiency. Reactions were stopped after 50 min by placing

the assay tubes on ice. B) Activity trends for maturation sets M1 (incubated at 30�C) and M2 (37�C). Samples of 10mL were withdrawn every 15min and measured for SH activity.

Legend: M1 complete¼ 15% (v/v) SH-extract, 15% (v/v) full M1-extract (total volume: 100mL); M2 complete¼ 15% (v/v) SH-extract, 15% (v/v) full M2-extract (total volume: 100mL).

Assays contained 5mM MgCl2, 2.5 mM DTT, 2.5 mM ATP, 50mM CP, 50mM FeSO4, 20mM NiCl2, and 10mM FMN in 50mM Tris buffer pH 8.0. Results are mean values of three

independent measurements. Indicated error bars represent standard deviations. 1 Unit is defined as the H2-mediated reduction of 1mmol NADþ per minute. Specific activities were

calculated based on the SH-extract protein content (3 mg �mL�1).



insoluble or inactive was further addressed in the experimental setdescribed below.

Analyses of Complementation Capabilities and InterplayBetween Sets of Hyp-Proteins

The C. necator M1-, M2-, and the E. coli Hyp-sets are associatedwith different hydrogenase enzymes (Fig. 1B). With the exceptionof HypA and HypB, the M2-proteins share less sequence identitywith the M1-, than with the E. coli homologues. Although codingfor fully functional gene products, the AH operon (M2) is onlyweakly expressed under physiological conditions. AH-mediatedH2 oxidation does not sustain lithoautotrophic growth in absenceof MBH and SH (Kleihues et al., 2000; Sch€afer et al., 2013). TheAH-operon is thought to be the relict of a horizontal gene transferevent and shares sequence similarity with genes fromStreptomyces spp. (Schwartz et al., 2003). The biochemical

properties of the M2-proteins and the capability of the three non-related Hyp-sets to complement each other were of particularinterest. By using the extracts containing SH- or AH-derived Hyp-pairs, experiments were designed, in which one of these pairs waseither omitted (“deletion”) or substituted by the correspondinghomologue of the other set (“substitution”). These experimentsaddressed the complementation capabilities by E. coli Hyp-proteins (“M1/M2 deletion experiments”) or between SH- andAH-maturases (“M1/M2 substitution experiments”) (see Fig. 2).

Control assays contained combined extracts each harboringindividual pairs of maturases (M1/M2-AB, -CD, -EF and XW,respectively, each at 7.5% v/v), which successfully substituted thecomplete M1- or M2-extracts used in previous experiments(yielding 1.34 U �mg�1 and 0.45 U �mg�1, respectively; Table III).In most cases, neither M2-, nor E. coli Hyp-homologues werecapable of restoring the SH activity obtained with M1 control assayscompletely (Fig. 5A). Full complementation was achieved, however,

Figure 4. Effects of extract- and complex-additions to full maturation assays in vitro. Indicated constituents were added to complete M1 (A) and M2 reactions (B) using

fractions of 15% (v/v) Hyp-pair extracts. For details regarding extract contents and names, see Table I. Purified complexes indicated in (C) were added to complete M1 assays at

portions of 10, 20 or 50mg. Complexes labeled with asterisks were purified under anoxic conditions. Legend: M1 complete¼ 15% (v/v) SH-extract, 15% (v/v) full M1-extract (total

volume: 100mL). Assays contained 5mMMgCl2, 2.5 mM DTT, 2.5 mM ATP, 50mMCP, 50mM FeSO4, 20mMNiCl2, and 10mM FMN in 50mM Tris buffer pH 8.0 and were incubated for

50 min at 30�C (A, C) or 37�C (B). Results are mean values of three independent measurements. Indicated error bars represent standard deviations. 1 Unit is defined as the H2-

mediated reduction of 1mmol NADþ per minute. Specific activities were calculated based on the SH-extract protein content (3 mg �mL�1).


when HypA2B2 (M1) was substituted by HypA3B3 (M2). Likewise,the omission of HypA2B2 (M1) showed minor effects on SHmaturation. Reconstitution efficiencies were unaffected by nickelconcentrations above 20mM (as applied in standard assays),regardless whether M1-HypAB or M2-HypAB were present(Supplementary Table SII). These findings indicate that the E.coli HypAB homologues were capable to take over the function ofnickel insertion. An undefined level of Ni2þ was introduced into theassays by the extracts used, however, since �50% residual SHactivity was measured when NiCl2 was omitted from the smallmolecules mix (Table II, Supplementary Table SII).The “deletion experiments” with the AH-derived maturase set

lacking individual components were almost identical to thepicture obtained for the SH-derived set (Fig. 5B). Substitution ofM2-homologues by M1-constituents, however, revealed higherspecificities of Hyp-sets for dedicated apo-enzymes thanpreviously anticipated. In fact, each M1-homologue substitutingits respective M2-homologue stimulated SH maturation sub-stantially (Fig. 5B). Notably, substitution of AH-derived HypCD-and HypEF-homologues by their M1 counterparts yieldedactivities very similar to the complete M1 control assay. Bycontrast, E. coli homologues did not fully compensate for M2-HypCD- and -HypEF-absence. In the case of the nickel-insertioncomplex HypAB, the E. coli homologues nearly compensated thelack of M2-HypAB, whereas substitution of the AH-derived nickelinsertion complex by M1-HypAB did not show the pronouncedeffect observed in M2-HypCD- and M2-HypEF-substitutionassays (Fig. 5B). This suggests that in particular, interactionsbetween HypC-, HypD-, HypE- and HypF-homologues and HoxHare impaired when non-native maturases are involved. Possibly,an unfavorably strong interaction of AH- or E. coli derived HypC-homologues with HoxH causes a partial blockade of the active sitefor Ni2þ-insertion. In line with this notion, pulldown experimentsrevealed that HoxH coeluted preferably with HypC2 (AH-set)

rather than with HypC1 (SH-dedicated set) (SupplementaryFig. S1A, E), indicating a tighter binding of the hydrogenasesubunit to the HypC2 protein.

Purification of HoxW and Hyp-Intermediate Complexes inPulldown Experiments

By producing tagged baits together with selected parts of thematuration machinery followed by StrepTactin-affinity chromatog-raphy, intermediate complexes formed between maturation factorscan be captured and functional interactions can be studied (Maieret al., 1998). In this study, we isolated several Hyp-complexespredicted for both SH- and AH-specific maturases in this way (seeSupplementary Results and Supplementary Fig. S1). These includedcomplexes between HypC and HoxH, HypCD, HypCDE, HypAB, andthe specific endopeptidase HoxW, which is not predicted to formcomplexes with other maturation factors.HoxW, the specific endopeptidase required for maturation of the

SH, catalyzes the final step in the maturation sequence by cleaving asmall peptide (24 amino acids) off the C-terminus of HoxH(Thiemermann et al., 1996). The hoxW-gene was expressed in E.coli by using a tetracycline-inducible promoter (Table I andSupplementary Table SI). Upon aerobic purification of StrepII–HoxW by StrepTactin-affinity chromatography (Fig. 6A), we madethe unexpected observation that elution fractions showed a deep-brownish color. UV/Vis analysis further revealed an absorptionspectrum of the oxidized purified protein consistent with thepresence of iron-sulfur clusters. We observed broad charge-transferbands ranging over the full visible spectrum with maxima at 325–345, 420, and 460 nm, which disappeared upon reduction withdithionite under anaerobic conditions (Fig. 6B). Colorimetricdetermination of the iron-content gave values of 3.9� 0.1 non-heme iron atoms per mol protein. In line with these findings, theHoxW-sequence contains, aside from the conserved nickel-binding

Figure 5. Effects of Hyp-pair substitutions and deletions in SH maturation assays in vitro. Complementation capabilities were tested in M1- (A) and M2-based assays (B).

Control assays (M1/M2 combined) contained 15% (v/v) SH-extract and each 7.5% (v/v) of M1/M2-AB-, -CD-, -EF- and XW-extracts. Indicated Hyp-pair constituents were either

omitted or substituted by 7.5% (v/v) of the extracts containing M2- (substitutions in A) or M1-counterparts (substitutions in B). Assays contained 5mM MgCl2, 2.5 mM DTT, 2.5 mM

ATP, 50mM CP, 50mM FeSO4, 20mMNiCl2, and 10mM FMN in 50mM Tris buffer pH 8.0 and were incubated in a total volume of 100mL for 50min at 30�C (A) or 37�C (B). Results are

mean values of three independent measurements. Indicated error bars represent standard deviations. One Unit is defined as the H2-mediated reduction of 1mmol NADþ per minute.

Specific activities were calculated based on the SH-extract protein content (3 mg �mL�1).



domain, a C115xxC118xxC121-motif which, together with a fourthcoordinating residue, could coordinate a [4Fe-4S]-cluster (Supple-mentary Fig. S2). Alignment with homologous HoxW-sequencesavailable so far revealed that the motif is not conserved, and thatonly one homologue of the C. necator protein contains a completeCxxCxxC-motif (HoxW from Mycobacterium vanbaalenii PYR-1, astrictly aerobic gram-positive bacterium known for its capability toutilize polycyclic aromatic hydrocarbons).

This is the first report of an iron-sulfur cluster containingmaturation endopeptidase, which raises questions regarding thefunction of this structural element. The SH is among a few [NiFe]-hydrogenases, which retain their activity even at high levels ofoxygen (Schneider et al., 1979; Schneider and Schlegel, 1981). Themechanism responsible for the oxygen-tolerance of the SH iscurrently in dispute (Lauterbach et al., 2011a). HoxY, the smallsubunit of the hydrogenase moiety of the SH, harbors a flavin(FMN-a) proximal to the [NiFe]-site of HoxH, which is weaklybound and lost when the enzyme is exposed to prolonged reducingconditions (van der Linden et al., 2004). Concomitant to thedissociation of FMN-a, the SH loses its insensitivity towardsoxygen. Notably, the loss of FMN-a is only reversed under reducingconditions, too. Maturation of the SH must be adapted to aerobicconditions, as does the recruitment of FMN-a. Hence, HoxWmighthave a role in protecting the [NiFe]-site at the final stage ofmaturation, by delivering electrons for oxygen removal and/or forreduction of the FMN-a cofactor in the course of its incorporation.Studies highlighting these possibilities are currently underway.

Purified Hyp-complexes, together with HoxW, were functionallytested in cell-free maturation assays. Complexes between AH-maturases (M2), which were obtained in substantially less yield andpurity than homologues from the MBH/SH-operon (M1), were noteffective in stimulating SH maturation (data not shown). Amongthe SH-dedicated complexes pulled down, HypCDE (purifiedanaerobically), HypAB and HoxW were active and increased theoverall SH activity upon addition to assays (Fig. 4C). The largesteffect was achieved by addition of anaerobically purified HypCDE-

complex. In turn, the HypCD-complex without bound HypE,purified either aerobically or anaerobically, did not exert the sameeffect. This suggests that the initial formation of a complex betweenHypC, which is proposed to deliver iron as well as CO2 as a precursorfor the carbonyl ligand (Soboh et al., 2013), and HypD, which servesas the scaffold for assembly of the Fe-(CN)2CO-group, does notdetermine the overall rate of the maturation sequence. Followingthis notion, the more crucial events with respect to the overallreaction rate arguably take place after formation of the HypCD-complex, including the HypEF-mediated, ATP-dependent trans-carbamoylation and dehydration reactions, the binding ofcyanylated HypE to a preformed HypCD-complex and the transferof the Fe-(CN)2CO-group to HoxH. The molecular events leading tothe assembly of the iron-cofactor are currently under debate on thebasis of the E. coli Hyp-machinery (Soboh et al., 2014; Stripp et al.,2014; Stripp et al., 2013). Following entry of “modified” HypE into afunctional ternary complex with the HypCD- or the homologousHybG-HypD-complex, redox chemistry affords transfer of thecyanide ligands, yielding the fully assembled Fe-(CN)2CO-group.Whether HypE or even HypF are involved in channeling the iron-cofactor to the hydrogenase subunit is presently unknown, as is theelectron donor required for CO2-reduction and cyanylation at theHypD-scaffold.

As stated before, the unassembled SH-subunits represent alimiting constituent in cell-free maturation. Interestingly, theaddition of purified HypC1–HoxH complex to complete M1 assaysincreased SH activity (Fig. 4C), indicating that apo-HoxH suffersfrom reduced stability or solubility. Since SH-activity was measuredas the H2-mediated reduction of NADþ, which requiresstoichiometric proportions of the four subunits (Burgdorf et al.,2005; Lauterbach et al., 2011a,b; Schiffels et al., 2013; Schneider andSchlegel 1976), SH-extracts are likely to contain sub-stoichiometricamounts of the metal-free HoxH precursor.

Finally, the possibility that purified complexes could not onlystimulate, but also functionally substitute maturase-containingextracts, was analyzed. By combining SH-extract with each 50mg of

Figure 6. Purification and analysis of StrepII-HoxW.N-terminally StrepII-tagged HoxWwas purified from E. coliBL21StarTM (DE3) cells harboring the plasmid pASG(5’-StrepII)-

hoxW. Gene expression was performed as described in themethods section. (A) Purification by StrepTactin-affinity chromatography, visually followed by 12% SDS–PAGE stained by

Coomassie-Brilliant Blue. 25mg of cell-free extract and 6mg of purified StrepII–HoxWwere applied. Legend: M¼ SDS–PAGE marker/protein ladder; CFE¼ cell-free soluble extract;

E¼ elution fraction. (B) UV/Vis spectroscopic analysis of purified StrepII–HoxW (15mg �mL�1). The solid line represents the spectrum of the oxidized protein, whereas the dotted

line shows the spectrum after reduction with 1mM sodium dithionite under anaerobic conditions.


purified HypAB, HypCDE and HoxW, SH maturation was achievedyielding 1.26 U �mg�1, which is in the range of the specific activitiesachieved with a complete M1-extract (Table III). By contrast, thepurified AH-derived complexes were not functional in replacing thecomplete M2-extract. Aside from the lower yields and purities ofthe AH-Hyp-complexes, it remains to be elucidated whether thesematurases are more instable outside the cellular or extractenvironment than their SH-counterparts from the M1-set.

Conclusion

In total, the findings presented herein demonstrate that applicationof a combinatorial cloning system to provide individual constituentsof complex interacting biological systems provides a powerful toolto identify and systematically investigate functional interactions. Asystematic and ordered preparation of a large number ofrecombinant E. coli strains for the parallel analysis of a largenumber of gene products enables combinatorial analysis of multipleinteraction partners in short time and permits a truly alternativeapproach to in vivo knockout studies. Since the latter were and stillare hampered by the fact that tools for genetic manipulation are notavailable for all organisms, such ex vivo studies facilitate mattersand also permit the analyses of systems whose functionalimpairment is lethal or causes severe growth effects in the nativesystem. Likewise, using isolated components of a system,intermediates are readily pulled down without being affected byother parts of the native environment, which is frequentlyimpossible inside a native host. In particular, organisms harboringmultiple copies of related biological factors might elude themselvesfor such studies.Cell-free systems represent a direct approach for analysis and

manipulation of the physical conditions favoring protein-assistedmaturation of an enzyme in question (Smith et al., 2014). This hasbeen most convincingly demonstrated on the example of [FeFe]-hydrogenase maturation (Kuchenreuther et al., 2009, 2011, 2012;Posewitz et al., 2004). Such systems may utilize purified maturasesand apo-enzymes, recombinant extracts or a combination of both.Even a start from scratch (at the genetic level) using cell-free proteinsynthesis followed by maturation in vitro has been successfullytested (Boyer et al., 2008). In either case, a wealth of informationcan be obtained, which remains locked when common in vivoapproaches are used. In this study, we discovered stoichiometricimbalances for HypE, HypF and the SH-subunit HoxH, which couldnot be identified using cell-based studies before. Based on thisinformation, an overexpression system for hypE, hypF and hoxHrelative to the other genes could be installed in order to improve invivo maturation and the yield of active recombinant hydrogenase.Further, the conditions favoring in vitro maturation of the SH mightbe adapted to alter the conditions applied in cell-based approaches.[FeFe]-Hydrogenase research has reached a stage, where

functional [FeFe]-subcluster mimics can be synthesized andincorporated into apo-enzymes without involvement of additionalmaturation factors (Esselborn et al., 2013). Cell-free systems mayhelp to deepen the understanding of [NiFe]-hydrogenase matura-tion and, furthermore, hold the capacity to develop simple, scalableand low-cost systems for the synthesis of the active site cofactor.The specific activities obtained by maturation in vitro were in the

same range as previously achieved in cell extracts from SHproduction and maturation in vivo (Schiffels et al., 2013). We haveshown here that although non-dedicated Hyp-sets are adapted todifferent physical conditions, they are capable of transferring afunctional [NiFe]-cofactor to an apo-enzyme. Hence, the require-ments for the synthesis of an arbitrary [NiFe]-hydrogenase mightbe significantly simplified using a cell-free system with pleiotropic,highly active maturases. In the case of oxygen-tolerant enzymes, afew specific maturation factors might be additionally required.HoxW, for example, is an essential component for the SHmaturation machinery, whose function might have been under-estimated up to now. We discovered an unusual cofactor in HoxW,whose functional role has yet to be elucidated.Beyond its application in hydrogenase research, the methods

described herein are directly applicable to a large number ofmetalloenzyme maturation processes. Amongst putative targets forits application are the dedicated maturation machineries for anumber of complex systems, including machineries for superoxidedismutases, ureases, acetyl-CoA synthases/carbon monoxidedehydrogenases, molybdenum-containing enzymes and the syn-thesis of cytochromes (Jeon et al., 2001; Kuchar and Hausinger,2004). Further, the stepwise composition of metabolic pathwaysfrom predefined modules might be possible using the methodsdescribed. In this context, the system might enable theidentification of bottlenecks, the tuning of processes to favorrate-limiting steps and the determination of requirements usuallynot ascertainable in culture-based approaches.

Acknowledgments

We are indebted to Maximilian Schelden for skillful technical assistance, andto Professor Johann Heider for valuable discussions.

References

Blokesch M, B€ock A. 2002. Maturation of [NiFe]-hydrogenases in Escherichia coli: theHypC cycle. J Mol Biol 324(2):287–296.

Blokesch M, Magalon A, B€ock A. 2001. Interplay between the specific chaperone-likeproteins HybG and HypC in maturation of hydrogenases 1, 2, and 3 fromEscherichia coli. J Bacteriol 183(9):2817–2822.

Blokesch M, Paschos A, Bauer A, Reissmann S, Drapal N, B€ock A. 2004. Analysis ofthe transcarbamoylation-dehydration reaction catalyzed by the hydrogenasematuration proteins HypF and HypE. Eur J Biochem 271(16):3428–3436.

B€ock A, King PW, Blokesch M, Posewitz MC. 2006. Maturation of hydrogenases. AdvMicrob Physiol 51:1–71.

Boyer ME, Stapleton JA, Kuchenreuther JM, Wang CW, Swartz JR. 2008. Cell-freesynthesis and maturation of [FeFe] hydrogenases. Biotechnol Bioeng 99(1):59–67.

Bradford MM 1976. A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein-dye binding. AnalBiochem 72:248–254.

Burgdorf T, van der Linden E, Bernhard M, Yin QY, Back JW, Hartog AF, Muijsers AO,de Koster CG, Albracht SP, Friedrich B. 2005. The soluble NADþ-Reducing[NiFe]-hydrogenase from Ralstonia eutropha H16 consists of six subunits andcan be specifically activated by NADPH. J Bacteriol 187(9):3122–3132.

B€urstel I, Siebert E, Winter G, Hummel P, Zebger I, Friedrich B, Lenz O. 2012.A universal scaffold for synthesis of the Fe(CN)2(CO) moiety of [NiFe]-hydrogenase. J Biol Chem 287(46):38845–38853.

Cai F, Ngu TT, Kaluarachchi H, Zamble DB. 2011. Relationship between the GTPase,metal-binding, and dimerization activities of E. coli HypB. J Biol Inorg Chem16(6):857–868.



Dernedde J, Eitinger T, Patenge N, Friedrich B. 1996. Hyp gene products inAlcaligenes eutrophus are part of a hydrogenase-maturation system. Eur JBiochem 235(1-2):351–358.

Drapal N, B€ock A. 1998. Interaction of the hydrogenase accessory protein HypC withHycE, the large subunit of Escherichia coli hydrogenase 3 during enzymematuration. Biochemistry 37(9):2941–2948.

Esselborn J, Lambertz C, Adamska-Venkatesh A, Simmons T, Berggren G, Noth J,Siebel J, Hemschemeier A, Artero V, Reijerse E, Fontecave M, Lubitz W, Happe T.2013. Spontaneous activation of [FeFe]-hydrogenases by an inorganic [2Fe]active site mimic. Nat Chem Biol 9(10):607–609.

Forzi L, Sawers RG. 2007. Maturation of [NiFe]-hydrogenases in Escherichia coli.Biometals 20(3-4):565–578.

Hube M, Blokesch M, B€ock A. 2002. Network of hydrogenase maturation inEscherichia coli: role of accessory proteins HypA and HybF. J Bacteriol184(14):3879–3885.

Jeon WB, Cheng J, Ludden PW. 2001. Purification and characterization ofmembrane-associated CooC protein and its functional role in the insertion ofnickel into carbon monoxide dehydrogenase from Rhodospirillum rubrum.J Biol Chem 276(42):38602–38609.

Kleihues L, Lenz O, Bernhard M, Buhrke T, Friedrich B. 2000. The H2 sensor ofRalstonia eutropha is a member of the subclass of regulatory [NiFe]hydrogenases. J Bacteriol 182(10):2716–2724.

Kuchar J, Hausinger RP. 2004. Biosynthesis of metal sites. Chem Rev 104(2)):509–525.

Kuchenreuther JM, Britt RD, Swartz JR. 2012. New insights into [FeFe] hydrogenaseactivation and maturase function. PLoS One 7(9):e45850.

Kuchenreuther JM, George SJ, Grady-Smith CS, Cramer SP, Swartz JR. 2011. Cell-freeH-cluster synthesis and [FeFe] hydrogenase activation: all five CO and CN-

ligands derive from tyrosine. PLoS ONE 6(5):e20346.Kuchenreuther JM, Stapleton JA, Swartz JR. 2009. Tyrosine, cysteine, and S-adenosyl

methionine stimulate invitro [FeFe] hydrogenase activation. PLoSONE4(10):e7565.Laemmli UK 1970. Cleavage of structural proteins during the assembly of the head of

bacteriophage T4. Nature 227(5259):680–685.Lauterbach L, Idris Z, Vincent KA, Lenz O. 2011a. Catalytic properties of the isolated

diaphorase fragment of the NAD-reducing [NiFe]-hydrogenase from Ralstoniaeutropha. PLoS ONE 6(10)):e25939.

Lauterbach L, Liu J, Horch M, Hummel P, Schwarze A, Haumann M, Vincent KA,Lenz O, Zebger I. 2011b. The Hydrogenase Subcomplex of the NADþ-Reducing[NiFe] Hydrogenase from Ralstonia eutropha - Insights into Catalysis and RedoxInterconversions. Eur Journal Inorg Chem 2011(7):1067–1079.

Magalon A, B€ock A. 2000. Analysis of the HypC-HycE complex, a key intermediate inthe assembly of the metal center of the Escherichia coli hydrogenase 3. J BiolChem 275(28):21114–21120.

Maier T, Drapal N, Thanbichler M, B€ock A. 1998. Strep-tag II affinity purification: anapproach to study intermediates of metalloenzyme biosynthesis. Anal Biochem259(1):68–73.

Peters JW, Schut GJ, Boyd ES, Mulder DW, Shepard EM, Broderick JB, King PW,Adams MW. 2014. [FeFe]- and [NiFe]-hydrogenase diversity, mechanism, andmaturation. Biochim Biophys Acta 1853(6):1350–1369.

Pierik AJ, Wolbert RB, Mutsaers PH, Hagen WR, Veeger C. 1992. Purification andbiochemical characterization of a putative [6Fe-6S] prismane-cluster-containingprotein fromDesulfovibrio vulgaris (Hildenborough). Eur J Biochem206(3)):697–704.

Posewitz MC, King PW, Smolinski SL, Zhang L, Seibert M, Ghirardi ML. 2004.Discovery of two novel radical S-adenosylmethionine proteins required for theassembly of an active [Fe] hydrogenase. J Biol Chem 279(24):25711–25720.

Sch€afer C, Friedrich B, Lenz O. 2013. Characteristics of a novel, oxygen-insensitivegroup 5 [NiFe]-hydrogenase in Ralstonia eutropha. Appl Environ Microbiol79(17):5137–5145.

Sch€agger H, Cramer WA, von Jagow G. 1994. Analysis of molecular masses andoligomeric states of protein complexes by blue native electrophoresis andisolation of membrane protein complexes by two-dimensional nativeelectrophoresis. Anal Biochem 217(2):220–230.

Schiffels J, Pinkenburg O, Schelden M, Aboulnaga el HA, Baumann ME, Selmer T.2013. An Innovative Cloning Platform Enables Large-Scale Production andMaturation of an Oxygen-Tolerant [NiFe]-Hydrogenase from Cupriavidusnecator in Escherichia coli. PLoS One 8(7):e68812.

Schneider K, Cammack R, Schlegel HG, Hall DO. 1979. The iron-sulphur centres ofsoluble hydrogenase from Alcaligenes eutrophus. Biochim Biophys Acta578(2):445–461.

Schneider K, Schlegel HG. 1976. Purification and properties of soluble hydrogenasefrom Alcaligenes eutrophus H 16. Biochim Biophys Acta 452(1):66–80.

Schneider K, Schlegel HG. 1981. Production of superoxide radicals by solublehydrogenase from Alcaligenes eutrophus H16. Biochem J 193(1):99–107.

Schwartz E, Henne A, Cramm R, Eitinger T, Friedrich B, Gottschalk G. 2003.Complete nucleotide sequence of pH G1: a Ralstonia eutropha H16megaplasmid encoding key enzymes of H2–based lithoautotrophy andanaerobiosis. J Mol Biol 332(2):369–383.

SmithMT, Wilding KM, Hunt JM, Bennett AM, Bundy BC. 2014. The emerging age ofcell-free synthetic biology. FEBS Lett 588(17):2755–2761.

Soboh B, Kruger S, Kuhns M, Pinske C, Lehmann A, Sawers RG. 2010. Developmentof a cell-free system reveals an oxygen-labile step in the maturation of [NiFe]-hydrogenase 2 of Escherichia coli. FEBS Lett 584(18):4109–4114.

Soboh B, Lindenstrauss U, Granich C, Javed M, Herzberg M, Thomas C, Stripp ST.2014. [NiFe]-hydrogenase maturation in vitro: analysis of the roles of the HybGand HypD accessory proteins. Biochem J 464(2):169–177.

Soboh B, Stripp ST, Bielak C, Lindenstrauss U, Braussemann M, Javaid M,Hallensleben M, Granich C, Herzberg M, Heberle J, Sawers RG. 2013. The[NiFe]-Hydrogenase Accessory Chaperones HypC and HybG of Escherichia coliare Iron- and Carbon Dioxide-Binding Proteins. FEBS Lett 587(16)):2512–2516.

Stripp ST, Lindenstrauss U, Granich C, Sawers RG, Soboh B. 2014. The influence ofoxygen on [NiFe]-hydrogenase cofactor biosynthesis and how ligation of carbonmonoxide precedes cyanation. PLoS ONE 9(9):e107488.

Stripp ST, Soboh B, Lindenstrauss U, Braussemann M, Herzberg M, Nies DH, SawersRG, Heberle J. 2013. HypD Is the Scaffold Protein for Fe-(CN)2CO CofactorAssembly in [NiFe]-Hydrogenase Maturation. Biochemistry 52(19):3289–3296.

Thiemermann S, Dernedde J, Bernhard M, Schroeder W, Massanz C, Friedrich B.1996. Carboxyl-terminal processing of the cytoplasmic NAD-reducing hydro-genase of Alcaligenes eutrophus requires the hoxW gene product. J Bacteriol178(8):2368–2374.

Towbin H, Staehelin T, Gordon J. 1979. Electrophoretic transfer of proteins frompolyacrylamide gels to nitrocellulose sheets: procedure and some applications.Proc Natl Acad Sci USA 76(9):4350–4354.

van der Linden E, Faber BW, Bleijlevens B, Burgdorf T, Bernhard M, Friedrich B,Albracht SP. 2004. Selective release and function of one of the two FMN groupsin the cytoplasmic NADþ-reducing [NiFe]-hydrogenase from Ralstoniaeutropha. Eur J Biochem 271(4):801–808.

Watanabe S, Matsumi R, Arai T, Atomi H, Imanaka T, Miki K. 2007. Crystalstructures of [NiFe] hydrogenase maturation proteins HypC, HypD, andHypE:;1; insights into cyanation reaction by thiol redox signaling. Mol Cell27(1):29–40.

Wilson CM 1979. Studies and critique of Amido Black 10B, Coomassie Blue R, andFast Green FCF as stains for proteins after polyacrylamide gel electrophoresis.Anal Biochem 96(2):263–278.

Wolf I, Buhrke T, Dernedde J, Pohlmann A, Friedrich B. 1998. Duplication of hypgenes involved in maturation of [NiFe] hydrogenases in Alcaligenes eutrophusH16. Arch Microbiol 170(6):451–9.

Supporting Information

Additional supporting information may be found in the onlineversion of this article at the publisher’s web-site.


schiffels et al-2015-biotechnology and bioengineering

Documents

maturation of thesoluble

set of hyp

shdedicated hypgenesand

maturation factors hypabcdef

highest maturation efciency

nifehydrogenases bocket

context of nife

case of nife