Báo cáo khoa học: Functional importance of a conserved sequence motif in FhaC, a prototypic member of the TpsB/Omp85 superfamily ppt

FhaC, a prototypic member of the TpsB/Omp85superfamily Anne-Sophie Delattre1–4, Bernard Clantin5, Nathalie Saint6,7, Camille Locht1–4, Vincent Villeret5and Franc¸oise Jacob-Dubuisson1–4

FhaC, a prototypic member of the TpsB/Omp85

superfamily

Anne-Sophie Delattre1–4, Bernard Clantin5, Nathalie Saint6,7, Camille Locht1–4, Vincent Villeret5and Franc¸oise Jacob-Dubuisson1–4

1 Inserm U1019, Center for Infection and Immunity of Lille, France

2 Institut Pasteur de Lille, France

3 Universite´ Lille Nord de France, Lille, France

4 CNRS UMR8204, Lille, France

5 Institut de Recherche Interdisciplinaire, USR3078 CNRS – Universite´ de Lille 1 et 2, Villeneuve d’Ascq, France

6 INSERM U554, Universite´ de Montpellier 1 et 2, France

7 UMR5048 CNRS, Universite´ de Montpellier 1 et 2, France

Keywords

Bordetella; outer membrane protein; protein

structure; protein transport; two-partner

secretion

Correspondence

F Jacob-Dubuisson, 1, rue Calmette, 59019

Lille Cedex, France

Fax: +33 320 87 11 58

Tel: +33 320 87 11 55

E-mail: francoise.jacob@ibl.fr

Database

Structural data are available at the Protein

Data Bank under the accession number

2QDZ (FhaCWT)

(Received 23 July 2010, revised 8

September 2010, accepted 13 September

2010)

doi:10.1111/j.1742-4658.2010.07881.x

In Gram-negative bacteria, the two-partner secretion pathway mediates the secretion of TpsA proteins with various functions TpsB transporters specifi-cally recognize their TpsA partners in the periplasm and mediate their trans-port through a hydrophilic channel The filamentous haemagglutinin adhesin (FHA)⁄ FhaC pair represents a model two-partner secretion system, with the structure of the TpsB transporter FhaC providing the bases to deci-pher the mechanism of action of these proteins FhaC is composed of

a b-barrel preceded by two periplasmic polypeptide-transport-associated (POTRA) domains in tandem The barrel is occluded by an N-terminal helix and an extracellular loop, L6, folded back into the FhaC channel In this article, we describe a functionally important motif of FhaC The VRGY tetrad is highly conserved in the TpsB family and, in FhaC, it is located at the tip of L6 reaching the periplasm Replacement by Ala of the invariant Arg dramatically affects the secretion efficiency, although the structure of FhaC and its channel properties remain unaffected This substitution affects the secretion mechanism at a step beyond the initial TpsA–TpsB interaction Replacement of the conserved Tyr affects the channel properties, but not the secretion activity, suggesting that this residue stabilizes the loop in the resting conformation of FhaC Thus, the conserved motif at the tip of L6 represents an important piece of two-partner secretion machinery This motif is conserved in a predicted loop between two b-barrel strands in more distant relatives of FhaC involved in protein transport across or assembly into the outer membranes of bacteria and organelles, suggesting a conserved function in the molecular mechanism of transport

Structured digital abstract

l MINT-7996294 : Fha30 (uniprotkb: P12255 ) binds ( MI:0407 ) to FhaC (uniprotkb: P35077 ) by filter binding ( MI:0049 )

Abbreviations

ECL, enhanced chemiluminescence; FHA, filamentous haemagglutinin adhesin; POTRA domain, polypeptide-transport-associated domain; TPS, two-partner secretion; WT, wild-type.

Gram-negative bacteria possess a complex cell

enve-lope composed of two membranes The outer

mem-brane, which provides the bacterium with significant

protection against toxic agents [1], represents a barrier

for the secretion of proteins destined for the cell

sur-face or the extracellular milieu Thus, Gram-negative

bacteria have developed a number of pathways

specifi-cally devoted to protein secretion Among these, the

two-partner secretion (TPS) pathway is widely

repre-sented The ‘TpsB’ transporters mediate the secretion

across the outer membrane of their ‘TpsA’ exoprotein

partners, which serve as adhesins, cytolysins, invasins,

proteases, etc., to the bacterial cell surface or in the

extracellular milieu [2] The TpsB proteins belong to

the TpsB⁄ Omp85 superfamily of protein transporters,

also called polypeptide-transporting b-barrel proteins

[3–5] This superfamily includes transporters located in

the outer membranes of Gram-negative bacteria and

of organelles of endosymbiotic origin [3,6–12] These

proteins, such as BamA (formerly YaeT) in Escherichia

coli, Toc75 and Oep80 in chloroplasts, and Sam50 in

mitochondria, are essential parts of complexes involved

in protein transport across, or assembly into, the outer

membranes of their respective organisms or organelles

The mechanistic principles of transport in the TPS

pathway remain to be deciphered The current model

of secretion is as follows Export of the TpsA

precur-sor across the cytoplasmic membrane is mediated by

the Sec machinery TpsA proteins harbour, at their

N-terminus, a conserved ‘TPS’ domain, approximately

250 residues long, required for secretion In the

peri-plasm, the TPS domain in an extended conformation

is recognized by the periplasmic domain of its TpsB

partner [13] This molecular interaction is then

fol-lowed by the initiation of TpsA translocation through

a hydrophilic channel formed by the transporter [14]

As secretion proceeds, the exoprotein folds

progres-sively at the cell surface into a long b-helix The TPS

domain itself adopts a right-handed b-helical structure

with short extrahelical segments [15–17] Two subtypes

of TPS system have been identified, which differ by the

sequences of the TPS domains of the TpsA proteins

and by those of their TpsB transporters [16,18]

Never-theless, the structure of the TPS domain is highly

con-served between the two subtypes [15–17], indicating

that the TPS pathway is dedicated to the secretion of

b-solenoid proteins [19]

Our model TPS system is the filamentous

haemag-glutinin adhesin (FHA)⁄ FhaC pair of the whooping

cough agent Bordetella pertussis FHA is a major

adhesin of this respiratory pathogen, and FhaC is its

specific TpsB transporter [20] The structure of FhaC has been solved by X-ray crystallography [21] FhaC is monomeric and comprises a 16-stranded b-barrel (height, 35 A˚) joined by short periplasmic turns and longer surface loops, called L1–L8 The N-terminus of the protein is located in the extracellular milieu and folds into a 20-residue-long a-helix, H1, that passes right through the transmembrane barrel The C-termi-nus of H1 emerges into the periplasm and is connected

to two tandem polypeptide transport-associated (PO-TRA) domains [22] via a 30-residue-long linker unde-fined in the crystal structure The extracellular loop L6 that joins b-strands 11 and 12 of the barrel is folded as

a hairpin in the barrel interior, with its tip reaching the periplasm The barrel of FhaC forms an ion-per-meable channel in lipid bilayers, and we have proposed that this pore represents the FHA-conducting channel [14,23] However, because it is almost totally occluded

in the structure, significant conformational changes must take place in the transport process Other pro-teins of the superfamily have also been shown to form ion-permeable channels in lipid bilayers [9,24–29] However, the role of the pore for the mechanism of integration of membrane proteins remains unknown Contrary to H1, which can be removed without sig-nificant loss of function, deletions of POTRA1, POTRA2 or L6 abolish FhaC activity [14,21] Interest-ingly, the L6 loop of FhaC harbours a conserved motif found in most members of the superfamily [3,4] In FhaC, the highly conserved V449RGY452 tetrad is located at the tip of L6 close to the periplasmic side of the barrel In this study, we demonstrate that the replacement of Arg by Ala in this conserved motif dra-matically affects the secretion activity of FhaC, but not the properties of the FhaC channel or its structure

In contrast, replacement of the conserved Tyr affects the pore properties of FhaC, but not its secretion activity, indicating a more subtle role for this residue This work thus provides the first identification of a functionally important motif for the molecular mecha-nism of TpsB transporters Because the VRGY⁄ F motif highlighted here is conserved in the TpsB⁄ Omp85 superfamily, it is also likely to be func-tionally relevant for other members of the superfamily

Results

Conservation of L6 sequence in the superfamily The VRGY motif is located at the tip of L6 of FhaC, which reaches the periplasm Alignments of the

predicted L6 sequences of a number of TpsB

trans-porters show that the VRGY⁄ F tetrad is conserved in

both TpsB subtypes, with Arg totally invariant and

only slight variations at the first, third and fourth

posi-tions of the motif ([18] andFig 1) In the rest of L6, a

few other residues are conserved between the two TpsB

subtypes, which otherwise appear to have distinct

sig-natures To enlarge our analysis, the sequence of L6

was also aligned with the corresponding regions of

rep-resentative members of the superfamily, including

E coli YaeT (BamA), Neisseria meningitidis Omp85,

Arabidopsis thaliana Toc75-III and Oep80,

Saccharo-myces cerevisiae Sam50, and TeOmp85 of the

cyano-bacterium Thermosynechococcus elongatus [30] (Fig 1)

In all proteins, a VRGY-related motif is found in a

segment predicted to form a loop between two strands

of the b-barrel and located close to, and at a conserved

distance from, the C-terminus Of note, the length of

the predicted loops varies between proteins [18] In all

proteins, the Arg residue of the VRGY-related motif is

invariant (Fig 1) The other three residues of the motif

are not strictly invariant, but their physicochemical

features are well conserved Additional similarities are

conspicuous in the same region (Fig 1) Furthermore, the analysis of the sequences of a large number of pre-dicted Toc75, Sam50 and Omp85 homologues indi-cated the presence of a closely related tetrad in the vast majority of these proteins (not shown) In most proteins, the first position of this motif harbours a hydrophobic residue (Val, Ile or Leu), although other residues occur occasionally Arg and Gly residues are found overwhelmingly at the second and third posi-tions, exceptionally replaced by Lys or Ser and by Ala

or Ser, respectively The fourth position of the tetrad

is occupied by Phe or Tyr in the vast majority of pro-teins or, occasionally, by other hydrophobic residues

or His This region is the best conserved in the super-family, which strongly suggests that it is important for the structure or function of these transporters

Importance of conserved motif for FhaC function The complete deletion of L6 in FhaC has been shown

to abolish the secretion of FHA, indicating the impor-tance of this loop for transport activity [21] However, the channel properties of FhaC are also strongly

Fig 1 Sequence alignments of representative proteins of the TpsB ⁄ Omp85 superfamily The sequence around the conserved VRGY ⁄ F tet-rad is shown The first line shows the FhaC sequence, with the first six and last three residues belonging to b-strands 11 and 12 of the b-barrel, respectively Note that the loop corresponding to L6 is predicted to be longer in some other proteins of the superfamily [18] Only proteins that have been characterized to some degree were selected for the alignment, excluding predicted proteins The first 12 proteins belong to the TpsB family, with the first seven belonging to subtype I TpsB and the last five to subtype II TpsB transporters The other six proteins belong to other groups of the superfamily The more conserved motifs are highlighted At, Arabidopsis thaliana; Bp, Bordetella per-tussis; Ech, Erwinia chrysanthemi; Ec, Escherichia coli; Et, Edwardsiella tarda; Hd, Haemophilus ducreyi; Hi, Haemophilus influenzae; Nm, Neisseria meningitidis; Pf, Pseudomonas fluorescens; Pm, Proteus mirabilis; Sc, Saccharomyces cerevisiae; Sm, Serratia marcescens; Te, Thermosynechococcus elongatus; Ye, Yersinia enterocolitica.

affected by the 34-residue-long deletion, which suggests

that it might perturb significantly the structure of the

protein In order to probe more finely the function of

the conserved motif of L6, the VRGY tetrad was

tar-geted by site-directed mutagenesis to alter its

physico-chemical properties Because of the poor resolution of

L6 in the FhaC structure, no information is available

regarding putative interactions between the side chains

of these residues and the rest of the protein We thus

chose to replace Arg and Tyr based on their strong

conservation in the superfamily, whereas Gly was not

targeted because it might be involved in the structure

of the loop Arg450 and Tyr452 were replaced by

Ala separately or together, thus creating FhaCR450A,

FhaCY452A and FhaCR450A+Y452A The three FhaC

variants were co-expressed with a gene encoding a

secretion-competent, N-terminal FHA derivative called

Fha44 in E coli [31], and the secretion of Fha44 in

culture supernatants was determined by

semiquantita-tive immunoblotting using anti-FHA IgG1’s (Fig 2)

In parallel, the localization and abundance of FhaC in

the outer membrane were analysed by immunoblotting

of membrane extracts with an anti-FhaC serum

(Fig 2) The R450A substitution and the double

sub-stitution both reduced FHA secretion by

approxi-mately 90% relative to the wild-type (WT) control

(FhaCWT), whereas FhaC was present in similar

amounts in all strains In contrast, the Y452A

replace-ment appeared to have no significant effect on FHA

secretion Thus, the invariant Arg residue, but not the

conserved aromatic residue, is essential for FhaC

activity

Effect of the substitutions on FHA recognition

In the TPS pathway, the first step of secretion is a

spe-cific recognition between the two partners in the

peri-plasm This is followed by the translocation of the

TpsA partner through the channel formed by its TpsB

transporter We have shown that the

POTRA-contain-ing periplasmic portion of FhaC binds FHA in vitro

[13,21] Because the tip of the L6 loop reaches the

peri-plasm, it is conceivable that it also participates in the

initial interaction with the substrate

To determine whether the interaction between FHA

and FhaC is affected by the introduced substitutions,

the FhaC variants were tested for their ability to

rec-ognize an immobilized FHA fragment harbouring

the TPS domain [13,21] Using this overlay assay,

FhaCR450A and FhaCY452A bound to the FHA

frag-ment quite efficiently (Fig 3) To obtain a

semiquanti-tative assessment of their binding, we performed

densitometry scanning of the FhaC bands from several

overlays Using the WT band as a reference (100%), relative binding values of 84 ± 6% and 105 ± 25% were observed for FhaCR450A and FhaCY452A, respec-tively This indicates that the tip of the L6 loop does not appear to play a significant role for the initial recruitment of FHA in the periplasm

Structure of FhaCR450A

In order to test whether R450 is essential for the struc-tural integrity of FhaC, e.g for the position of func-tionally important elements such as the POTRA domains or L6, the structure of FhaCR450A was solved

by X-ray crystallography FhaCR450A crystallized in the same conditions as its FhaCWT counterpart [21] Data collection and refinement statistics are given in

A

B

Fig 2 Role of R450 of the VRGY tetrad for FhaC activity (A) Secretion activity of the FhaC variants Escherichia coli UT5600 har-bouring two plasmids, pFJD12 that encodes an efficiently secreted FHA derivative called Fha44 and pFcc3 (encoding FhaC), pFcc3-R450A, pFcc3-Y452Aor pFcc3-R450A+Y452A, was grown to mid-exponential phase, and expression of the recombinant fha44 gene was induced for 3 h Equal amounts of total membranes and non-concentrated culture supernatants from all recombinant strains were collected and analysed by immunoblotting using anti-FHA IgG1’s and an anti-FhaC serum (top and bottom panels, respec-tively) A representative experiment is shown (B) Quantification of the secretion efficiency The amounts of protein were quantified by densitometry scanning, and the Fha44 ⁄ FhaC ratio was calculated for each recombinant strain, with the secretion activity of the strain producing WT FhaC set to 100% The experiments were per-formed several times (> 3) for quantification.

Table 1 Although determined at a limited resolution

of 3.5 A˚, this structure allows us to compare the

R450A variant with WT FhaC Overall, the structure

of the FhaCR450A variant is very similar to the

FhaCWTstructure This is exemplified by the rmsds for the Ca superimposition of the barrel (0.44 A˚) and the two POTRAs (0.61 A˚) The overall rmsd is 0.53 A˚ (Fig 4A) The relative orientation of the two POTRA domains is the same in both structures, as well as the presence and orientation of helix H1 and loop L6 inside the b-barrel The moderate resolution of both the WT and R450 structures does not allow the fine analysis of differences between them However, H1 clearly occupies a similar position in both structures (Fig 4B) With regard to L6, the electron density of the R450A variant allows the unambiguous positioning

of residues Q433 to I441 and G463 to T469 inside the b-barrel at positions similar to those observed in the

WT structure Other residues of the L6 loop are not seen in the FhaCR450A structure, as a result of a high mobility of the loop and⁄ or the limited resolution Nevertheless, the conformational constraints imposed

by residues Q433 to I441 and G463 to T469 on L6 demonstrate that this loop occupies similar positions inside the b-barrel in both FhaCWTand the FhaCR450A variant Thus, our data strongly argue that the R450A substitution has no significant effect on the structure

of the FhaC barrel and the POTRA domains, or on the position of L6 in the channel, and therefore the Arg450 residue is probably conserved for a functional rather than a structural purpose

Channel properties of the FhaC variants Because L6 is not clearly defined in the FhaCR450A structure, we tested whether the low secretion activity

of FhaCR450A could be explained by pore alteration using electrophysiological techniques Indeed, we have observed previously that FhaC variants with low secre-tion activities generally have altered pore properties [14] The electrophysiological properties of FhaCR450A and FhaCY452A inserted in lipid bilayers were analysed

in comparison with those of FhaCWT Initial character-ization was performed by inserting a large number (about 100) of FhaC molecules in a membrane submit-ted to slow ramps of voltage Asymmetric I⁄ V record-ings, similar to those observed previously with FhaCWT (Fig 5A, part a), were obtained for both variants (Fig 5B, part a; Fig 5C, part a) This asym-metric I⁄ V profile indicated that both proteins have a preferred sense of insertion into the lipid bilayers, simi-lar to FhaCWT The two mutants showed a linear I⁄ V relationship from +120 mV to around )60 mV, indi-cating a voltage-independent ion conductance in this voltage range (Fig 5B, part a; Fig 5C, part a) From )60 to )120 mV, the I ⁄ V curve of both variants lost its linearity and the current recorded at the two

Table 1 Data collection and refinement statistics.

FhaCR450A Data collection

Cell parameters (A ˚ ) 107.55, 139.39, 113.08

Refinement

Rmsd

a Number in parentheses is the statistic for the highest resolution

shell. bR factor = R||F o | ) |F c || ⁄ R|F o |, where |F o | and |F c | are the

observed and calculated structure factor amplitudes, respectively.

A

B

Fig 3 Overlay assay showing FHA–FhaC interactions The FHA

derivative Fha30, immobilized on separate nitrocellulose strips, was

used as bait for the indicated FhaC myc variants Following

incuba-tion of the strips with each of the FhaC variants, the Fha30–FhaC

complexes were detected with an anti-c-myc IgG1 followed by ECL

reaction The strips were aligned and the ECL reaction was carried

out simultaneously for all strips, using a single autoradiographic

film A representative experiment is shown (A) Amounts of Fha30

used as bait as analysed in a duplicate electrophoresis gel The gel

was stained with Coomassie blue (B) FhaCmycdetected after the

overlay.

voltage sweep directions revealed an hysteresis Thus,

the curves of the current measured (I) as a function of

the voltage applied (V) do not superpose when V

increases or decreases, indicating a delay in the

con-ductance in response to voltage changes This

hystere-sis was similarly observed with FhaCWT(Fig 5A, part

a) However, the two variants displayed different

behaviour in response to voltage, as revealed by a

vari-ation of the hysteresis shape, suggesting that they may

have different channel characteristics (Fig 5D, part a)

The electrical properties of the two FhaC variants

were further examined in single-channel experiments

by measuring their conductance values FhaCR450A

showed discrete current transitions at positive and

neg-ative applied potentials (Fig 5B, part b) The

calcu-lated conductance value from the two major peaks of

the current amplitude histogram was 1240 ± 130 pS

at positive voltage (Fig 5D, part b), very similar to

that of FhaCWT [14] Previous single-channel analyses

of FhaCWT have revealed different behaviours at

posi-tive and negaposi-tive potentials, with noisier current and

the appearance of conductance substates at negative

polarity (Fig 5A, part b) Moreover, WT channels at

negative potential usually display different opening and closing kinetics than the channels recorded at positive voltage (Fig 5A, part b) FhaCR450A shares these characteristics as, at negative potential, its cur-rent recordings showed more rapid oscillations between the open and closed states of the channel, and substates giving rise to smaller conductance values (1010 ± 160 pS) Interestingly, at both polarities,

Fha-CR450Achannels displayed opening and closing kinetics faster than those of WT

In contrast with FhaCR450A, FhaCY452A exhibited a majority of very noisy channels at both polarities (Fig 5C, part b) Moreover, the current amplitude his-togram at positive voltages displayed a broader distri-bution of events compared with FhaCR450A, clearly indicating the presence of many conductance substates and impairing the determination of a precise value of conductance (Fig 5C, part c) At negative potentials, the channels were as noisy as those obtained at posi-tive voltages and had a tendency to also display reduced current amplitudes

Altogether, the substitution of Arg450 by Ala does not have a drastic effect on channel activity and thus

External

environment

Outer

membrane

Periplasm

POTRA 2

POTRA 1

L6 H1

VRGY

Fig 4 Superposition of the crystal structure of FhaCR450Adetermined at 3.5 A ˚ resolution (red) and the structure of FhaCWT(blue) (A) The proteins are oriented with their surface side at the top POTRA1 and 2, H1 and L6 are labelled All structural elements, including the barrel, the POTRA domains and H1 superpose neatly (B) The VRGY motif is positioned on the WT structure Because of the relatively low resolu-tion for FhaCR450A, the position of L6 inside the barrel cannot be traced until the tip of the loop However, the surface-proximal portions of L6, which are defined by residues Q433 to I441 and G463 to T469, superpose well between the two proteins, arguing against a major con-formational change of the loop Structural data are available in the Protein Data Bank under the accession numbers 2QDZ (FhaC WT ) and 3NJT (FhaCR450A).

does not appear to affect the FhaC pore In contrast,

the ion channel properties of FhaCY452A are markedly

altered, indicating that the pore is modified by the

sub-stitution, possibly because interactions between L6 and

the b-barrel are affected

Discussion

The structure of FhaC, the first and thus far the only

full-length structure available for the TpsB⁄ Omp85

superfamily, has provided useful insights into the

architecture of TpsB and related transporters [21] The

mechanistic principles that govern TPS remain to be

elucidated, however In this work, we provide the first

identification of a motif functionally important for

TPS transport Using a combination of sequence

align-ments and structure–function analyses of FhaC

vari-ants obtained by site-directed mutagenesis, we show

that a conserved motif at the tip of the long

extracellu-lar loop L6 is required for the activity of the

trans-porter Its strong conservation in TpsB proteins

strongly indicates a conserved role for this motif in

TPS systems

Altogether, our data show that Arg450 is essentially important for function Thus, although the FhaCR450A variant is strongly disabled with respect to secretion activity, its global structure is not affected The elec-trophysiological properties of its channels are similar

to those of FhaCWT In contrast, the complete deletion

of L6 both abolished the activity of FhaC and affected its channel properties, arguing that it significantly perturbed the structure of the protein [21] The FhaCR450A channels are, nevertheless, slightly less stable than those of FhaCWT, suggesting that the substitution may have caused minor structural changes

to the channel FhaCR450A appears to recognize its substrate as efficiently as FhaCWT, ruling out a role for this conserved Arg in the initial TpsA–TpsB inter-action We thus propose that Arg450 contributes to a later step in the secretion process related to the trans-location of the substrate The crucial position of L6 strongly argues that it is probably involved in the conformational changes expected to take place on translocation Understanding its molecular function in the translocation process will most probably involve the characterization of secretion intermediates

A

(c)

B

C

D

Fig 5 Electrophysiological properties of

FhaC variants (A) Characterization of the

channels formed by FhaCWT(figure

repro-duced from ref [14]) (B, C) Characterization

of the channels formed by FhaCR450Aand

FhaCY452A, respectively In (A–C), the I ⁄ V

curves between )100 and +100 mV are

shown in (a); the direction of the voltage

ramp is marked by the arrow near each

curve The single-channel recordings at

positive and negative applied voltages are

shown in (b) (BL, baseline) In (c), the

associated amplitude histograms at +60 mV

are shown (D) Comparison of the

electro-physiological behaviour of FhaCR450A(grey)

and FhaCY452A(black) with superposition of

the I ⁄ V curves in (a) and of the amplitude

histograms in (b) The perturbed behaviour

of FhaCY452Acan be seen most clearly from

the absence of defined conductance levels.

Interestingly, we showed that the replacement of

Tyr452 in the same conserved L6 motif modifies the

pore properties of FhaC The FhaCY452A channel has

no defined conductance levels, indicating that it is less

stable A similar behaviour has been observed

previ-ously with other variants harbouring peptide insertions

in L6 [14] Of note, in the latter cases, the secretion

activity of FhaC was abolished, whereas the single

Tyr452 to Ala substitution did not affect the secretion

activity in a significant manner Thus, Tyr452 most

probably participates in the positioning of L6 or the

regulation of L6 mobility in the channel in the

absence of substrate, but its substitution by Ala does

not prevent L6 from adopting its ‘active’

conforma-tion during translocaconforma-tion Therefore, a possible

func-tion for Tyr452 may be to stabilize L6 in the channel

when the protein is in the ‘resting’ conformation, i.e

not actively translocating FHA The aromatic

charac-ter of this residue is well conserved in the family,

which suggests a conserved role for this residue in the

channel

Together with L6, the H1 helix also runs through

the channel Unlike L6, however, H1 is not conserved

in the TpsB family, with at least one TpsB transporter

being devoid of a helix before the POTRAs [18] We

have also shown that it is not important for the

secre-tion activity of FhaC [14] One possible funcsecre-tion of H1

may be to plug the resting channel between two cycles

of secretion We have obtained indications that

one FhaC molecule secretes several molecules of FHA

(F Jacob-Dubuisson, unpublished data) If FhaC

cycles between several conformations, we have most

probably captured its ‘resting’ conformation by

crystal-lography, whereas structural rearrangements of the

channel are expected when FhaC is in action If H1

moves out of the pore in vivo, this may trigger a

repo-sitioning of L6 for the secretion mechanism Further

work will aim to test this hypothesis

So far, our studies have identified two major players

in the TPS pathway: the TPS domain of the TpsA

pro-tein and L6 of the TpsB transporter [13,21] (this

work) A strong argument that these pieces of the TPS

machinery function together comes from sequence

alignments Thus, it is striking that the two subtypes

of TPS system can be distinguished by specific

signa-tures in both elements [16,18] (this work) Similarly,

the sequences of the two POTRA domains differ

between the two subtypes of TpsB transporter [18]

The POTRAs are also essential for FhaC activity

[13,21], and therefore they also constitute essential

parts of the TPS machinery All members of the

super-family share similar structural features, namely a

C-terminal b-barrel preceded by 1–7 POTRA domains

[30] In addition, the VRGY⁄ F sequence motif is par-ticularly well conserved, and its position relative to the C-terminus of the barrel appears to be similar in all proteins In all cases, it is predicted to be part of an extracellular loop between two b-strands of the barrel

If, as suggested by its conservation, the VGRY⁄ F tetrad is also functionally relevant in other members of the TpsB⁄ Omp85 superfamily, it implies that the barrel actively participates in the mechanism of these trans-porters Given the size of the substrates handled by Omp85 transporters, however, these mechanisms remain to be understood Similar to that proposed for FhaC, it is possible that the loop harbouring the con-served motif is involved in critical conformational changes in these other transporters

Materials and methods

Plasmids and constructions

PCRII-TOPO12TM was constructed as follows: the fhaC fragment encoding residues Pro275 to Phe454 was amplified

by PCR using pFcc3 [32] as template and the oligonucleo-tides Fc12UP (5¢-TTAGATCTCCGCTGGGGCGTACGC G-3¢) and FcA2Lo (5¢-CCAAGCTTCCGGGCTCAGAA ACTGAGG-3¢) as primers The amplicon was inserted into pCRII-TOPO (Invitrogen, Cergy-Pontoise, France) and sequenced, yielding pCRII-TOPO12TM The point muta-tions were introduced using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies, La Jolla,

CA, USA), according to the manufacturer’s instructions PCRII-TOPO12TM was used as a template with the following primers : R450AUp (5¢-GACGAGTACACGGT GGCCGGATACAACCTCAGGA-3¢) and R450ALo (5¢-TC CTGAGGTTGTATCCGGCCACCGTGTACTCGTC-3¢); Y452AUp (5¢-ACACGGTGCGCGGAGCCAACCTCAAG ACGTC-3¢) and Y452ALo (5¢-GACGTCCTGAGGTTGG CTCCGGCCACCGTGT-3¢); RA+YAUp (5¢-ACACGGT GGCCGGAGCCAACCTCAGGACGTC-3¢) and RA+YA

Lo (5¢-GACGTCCTGAGGTTGGCTCCGGCCACCGTG T-3¢) After mutagenesis and sequence verification, the BsiWI-HindIII fragments of these vectors were exchanged for the WT fragment of pFcc3, yielding pFcc3-R450A,

pFcc3-Y452A and pFcc3-R450A+Y452A The BsiWI-HindIII frag-ments were similarly exchanged into pFJD118 [14], yielding pFJD118-R450A and pFJD118-Y452A, which were used to produce 6-His-tagged FhaC variants for electrophysiology analyses [14] pFJD118 encodes full-length FhaC with an N-terminal 6-His tag For crystallography, pT7FcA450-noHis was used It was generated by restricting pFJD118-R450A with BamHI and re-ligating to eliminate the 6-His tag coding sequence

pFJD140 [13] was used for the production of c-myc-tagged FhaCWTfor the overlay assay The PstI-HindIII

frag-ments of pFcc3-R450A and pFcc3-Y452A were introduced

into the same sites of pFJD140, replacing the WT fragment,

generating pFJD140-R450A and pFJD140-Y452A

Secretion assay

E coli UT5600 harbouring pFJD12 (encoding an 80-kDa

N-terminal portion of FHA called Fha44 which can be

effi-ciently secreted in E coli [31]) was transformed with pFcc3,

pFcc3-R450A, pFcc3-Y452A or pFcc3-R450A+Y452A The

cells were grown at 37C in liquid Luria–Bertani medium

until the cultures reached the late exponential phase

(A600= 0.8), and the expression of fha44 was induced with

1 mm isopropyl thio-b-d-galactoside for 3 h at 37C

Bacte-ria were then harvested by centrifugation at 10 000 g for

15 min at 4C The culture supernatants were separated by

SDS⁄ PAGE and analysed by western blot with a mixture of

anti-FHA monoclonal IgG1’s (F1, F4 and F5) [33] The cell

pellets were resuspended in 50 mm Tris⁄ HCl (pH 8.0),

150 mm NaCl, 10 lgÆmL)1DNAse I (Sigma, Lyon, France)

and a cocktail of protease inhibitors (Complete EDTA Free,

Roche, Rosny-sous-Bois, France), and the bacteria were

broken using a French pressure cell After clarification of

the lysates by centrifugation at 20 000 g for 20 min,

mem-brane proteins were harvested by ultracentrifugation at

100 000 g for 1 h at 13C Each pellet was resuspended in

200 lL of the same buffer as above Equal amounts of

sam-ples were separated by SDS⁄ PAGE and analysed by western

blot with an anti-FhaC serum This antibody was raised in

rats against the periplasmic domain of FhaC and prepared

by Eurogentec (Seraing, Belgium) The amounts of Fha44

and FhaC were quantified by densitometry scanning of the

immunoblots, followed by data analysis with imagequant

tlsoftware (GE HealthCare, Saclay, France)

Protein production and purification

The FhaC variants were produced and purified as described

previously [21]

Overlay assay

Fha30 is a 30-kDa, secreted FHA truncate encompassing

the TPS domain and first three repeats It was used as bait

and FhaCmycand its variants as prey Fha30 was produced

and purified as described previously [15] Identical amounts

of Fha30 (5 lg) were loaded onto several lanes of

SDS⁄ PAGE gels and, following electrophoresis, the protein

was blotted onto nitrocellulose The strips of nitrocellulose

were each incubated with purified FhaCmyc (WT or

mutant), and bound FhaCmyc was detected using an

anti-c-myc IgG1 and chemiluminescence as in ref [13] The

strips were aligned and the development reaction [enhanced

chemiluminescence (ECL); GE Healthcare] was carried out

simultaneously for all strips using a single autoradiography

film, for comparison purposes The assay was repeated three times

Crystallization, data collection, structure determination and refinement

FhaCR450Acrystals were obtained at 20C using the hang-ing drop vapour diffusion method The protein and precipi-tant solutions were mixed in a 1 : 1 ratio Crystals were grown at a protein concentration of 26 mgÆmL)1 in 28% poly(ethylene glycol) 1000, 1% b-octyl-glucoside and

500 mm imidazole (pH 6.5) The FhaCR450A crystals were similar, with regard to space group and asymmetric unit composition, to the native crystals reported previously [21] The diffraction data were collected at 100 K on beamline ID14-1 at the European Synchrotron Radiation Facility (Grenoble, France) The diffraction data were processed with xds [34] Data collection and refinement statistics are summarized in Table 1

The three-dimensional structure of FhaCR450Awas solved using the FhaCWT structure (PDB code: 2QDZ) as the starting model Rigid-body refinements were performed with cns [35,36] using H1, L6, POTRA1, POTRA2 and the b-barrel as independent bodies The final refinement steps were performed using the maximum likelihood algorithm, and grouped B-factor calculation was performed with cns The refinement led to Rwork of 33.0% and Rfreeof 37.3% using all data to 3.5 A˚ The final model does not comprise the first two residues, the loop between H1 and POTRA1 (residues 31–52) and the extracellular loops L1 (221–228), L3 (295–301), L4 (342–350), L5 (384–397) and L8 (532– 542), which are not visible in the electron density map In the FhaCWTstructure, the tip of L6 (residues 443–458) was not well defined in the electron density map and was there-fore built as a polyalanine chain In the FhaCR450A struc-ture, the residues 441–463 of L6 were not visible in the electron density map and are not included in the final model Structural data are available in the Protein Data Bank database under the accession number 3NJT

Channel analysis

The planar lipid bilayer recordings were performed as described in ref [14] Virtually solvent-free planar lipid bilayers were formed over a 125–200-lm hole in a 10-lm-thick polytetrafluoroethylene film pretreated with a mixture

of 1 : 40 (v⁄ v) hexadecane–hexane and sandwiched between two half glass cells Phosphatidylcholine from soy beans (azolectin from Sigma type IV S), dissolved in hexane (0.5%), was spread on the top of the electrolyte solution [1 m KCl, 10 mm Hepes (pH 7.4)] on both sides of the bilayer chamber Bilayer formation was achieved by lower-ing and then raislower-ing the levels in one compartment and monitoring capacity responses The trans chamber was con-nected to ground and the cis chamber to the input of a

BLM 120 amplifier (Bio-Logic, Halifax, Canada) The

puri-fied FhaC proteins were added to the cis side

(5–100 ngÆmL)1) of the bilayer chamber

For the macroscopic conductance experiments, doped

membranes were subjected to slow voltage ramps

(10 mVÆs)1), and the transmembrane currents were

ampli-fied (BBA-01; Eastern Scientific, Rockville, MD, USA)

The current–voltage curves were stored on a computer and

analysed using scope software (PowerLab, ADI

Instru-ments, Sydney, Australia) For single-channel recordings,

currents were amplified by a BLM 120 amplifier

Single-channel currents were monitored using an oscilloscope

(TDS 3012, Tektronix, Beaverton, OR, USA) and stored

on a CD recorder via a DRA 200 interface (Bio-Logic) for

off-line analysis CD data were then analysed by winedr

(Bio-Logic) and clampfit (Molecular Devices, Sunnyvale,

CA, USA) software Data were filtered at 1 kHz before

dig-itizing at 11.2 kHz for analysis

References

1 Nikaido H (2003) Molecular basis of bacterial outer

membrane permeability revisited Microbiol Mol Biol

Rev 67, 593–656

2 Jacob-Dubuisson F, Fernandez R & Coutte L (2004)

Protein secretion through autotransporter and

two-part-ner pathways Biochim Biophys Acta 1694, 235–257

3 Moslavac S, Mirus O, Bredemeier R, Soll J, von

Haeseler A & Schleiff E (2005) Conserved pore-forming

regions in polypeptide-transporting proteins FEBS J

272, 1367–1378

4 Gentle IE, Burri L & Lithgow T (2005) Molecular

architecture and function of the Omp85 family of

pro-teins Mol Microbiol 58, 1216–1225

5 Koenig P, Mirus O, Haarmann R, Sommer MS,

Sinning I, Schleiff E & Tews I (2010) Conserved

properties of polypeptide transport-associated (POTRA)

domains derived from cyanobacterial Omp85 J Biol

Chem 285, 18016–18024

6 Voulhoux R, Bos MP, Geurtsen J, Mols M &

Tommas-sen J (2003) Role of a highly conserved bacterial protein

in outer membrane protein assembly Science 299, 262–

265

7 Bos MP, Robert V & Tommassen J (2007) Functioning

of outer membrane protein assembly factor Omp85

requires a single POTRA domain EMBO Rep 8, 1149–

1154

8 Knowles TJ, Scott-Tucker A, Overduin M & Henderson

IR (2009) Membrane protein architects: the role

of the BAM complex in outer membrane protein

assembly Nat Rev Microbiol 7, 206–214

9 Paschen SA, Waizenegger T, Stan T, Preuss M,

Cyrk-laff M, Hell K, Rapaport D & Neupert W (2003)

Evo-lutionary conservation of biogenesis of beta-barrel

membrane proteins Nature 426, 862–866

10 Wiedemann N, Kozjak V, Chacinska A, Schonfisch B, Rospert S, Ryan MT, Pfanner N & Meisinger C (2003) Machinery for protein sorting and assembly in the mito-chondrial outer membrane Nature 424, 565–571

11 Gentle I, Gabriel K, Beech P, Waller R & Lithgow T (2004) The Omp85 family of proteins is essential for outer membrane biogenesis in mitochondria and bacte-ria J Cell Biol 164, 19–24

12 Schleiff E & Soll J (2005) Membrane protein insertion: mixing eukaryotic and prokaryotic concepts EMBO Rep 6, 1023–1027

13 Hodak H, Clantin B, Willery E, Villeret V, Locht C & Jacob-Dubuisson F (2006) Secretion signal of the fila-mentous haemagglutinin, a model two-partner secretion substrate Mol Microbiol 61, 368–382

14 Meli AC, Hodak H, Clantin B, Locht C, Molle G, Jacob-Dubuisson F & Saint N (2006) Channel proper-ties of TpsB transporter FhaC point to two functional domains with a C-terminal protein-conducting pore

J Biol Chem 281, 158–166

15 Clantin B, Hodak H, Willery E, Locht C, Jacob-Dub-uisson F & Villeret V (2004) The crystal structure of filamentous hemagglutinin secretion domain and its implications for the two-partner secretion pathway Proc Natl Acad Sci USA 101, 6194–6199

16 Yeo HJ, Yokoyama T, Walkiewicz K, Kim Y, Grass S

& Geme JW III (2007) The structure of the Haemophi-lus influenzaeHMW1 pro-piece reveals a structural domain essential for bacterial two-partner secretion

J Biol Chem 282, 31076–31084

17 Weaver TM, Hocking JM, Bailey LJ, Wawrzyn GT, Howard DR, Sikkink LA, Ramirez-Alvarado M & Thompson JR (2009) Structural and functional studies

of truncated hemolysin A from Proteus mirabilis J Biol Chem 284, 22297–22309

18 Jacob-Dubuisson F, Villeret V, Clantin B, Delattre AS

& Saint N (2009) First structural insights into the TpsB⁄ Omp85 superfamily Biol Chem 390, 675–684

19 Kajava AV & Steven AC (2006) The turn of the screw: variations of the abundant beta-solenoid motif in pas-senger domains of Type V secretory proteins J Struct Biol 155, 306–315

20 Jacob-Dubuisson F, Locht C & Antoine R (2001) Two-partner secretion in Gram-negative bacteria: a thrifty, specific pathway for large virulence proteins Mol Microbiol 40, 306–313

21 Clantin B, Delattre AS, Rucktooa P, Saint N, Meli AC, Locht C, Jacob-Dubuisson F & Villeret V (2007) Struc-ture of the membrane protein FhaC: a member of the Omp85-TpsB transporter superfamily Science 317, 957– 961

22 Sanchez-Pulido L, Devos D, Genevrois S, Vicente M & Valencia A (2003) POTRA: a conserved domain in the FtsQ family and a class of beta-barrel outer membrane proteins Trends Biochem Sci 28, 523–526