1
A systematic genetic analysis in Neisseria meningitidis defines the Pil proteins required
2
for assembly, functionality, stabilization and export of type IV pili
3
4
5
Etienne Carbonnelle 1, 2, 3, Sophie Helaine 1, 2, Xavier Nassif 1, 2, 3 and Vladimir Pelicic 1, 2*
6
7
1
8
9
2
Université Paris Descartes, Faculté de Médecine René Descartes, UMR-S570, Paris, F-
10
11
INSERM, U570, Paris, F-75015 France
75015 France
3
AP-HP, Hôpital Necker-Enfants Malades, Paris, F-75015 France
12
13
14
*To whom correspondence should be addressed.
15
Phone: (33 1) 40 61 54 82; Fax: (33 1) 40 61 55 92; E-mail: pelicic@necker.fr
16
17
18
Running title: contribution of N. meningitidis Pil components to Tfp biogenesis
19
1
20
Summary
21
Although type IV pili (Tfp) are among the commonest virulence factors in bacteria, their
22
biogenesis by complex machineries of 12-15 proteins, and thereby their function, remains
23
poorly understood. Interestingly, some of these proteins were reported to merely antagonize
24
the retraction of the fibers powered by PilT, since piliation could be restored in their absence
25
by a mutation in the pilT gene. The recent identification of the fifteen Pil proteins dedicated to
26
Tfp biogenesis in N. meningitidis offered us the unprecedented possibility to define their exact
27
contribution in this process. We therefore systematically introduced a pilT mutation into the
28
corresponding non-piliated mutants and characterized them for the rescue of Tfp and Tfp-
29
mediated virulence phenotypes. We found that in addition to the pilin, the main constituent of
30
Tfp, only six Pil proteins were required for the actual assembly of the fibers, since apparently
31
normal fibers were restored in the remaining mutants. Restored fibers were surface-exposed,
32
except in the pilQ/T mutant in which they were trapped in the periplasm, suggesting that the
33
PilQ secretin was the sole Pil component necessary for their emergence on the surface.
34
Importantly, although in most mutants the restored Tfp were not functional, the pilG/T and
35
pilH/T mutants could form bacterial aggregates and adhere to human cells efficiently,
36
suggesting that Tfp stabilization and functional maturation are two discrete steps. These
37
findings have numerous implications for understanding Tfp biogenesis/function and provide a
38
useful groundwork for the characterization of the precise function of each Pil protein in this
39
process.
40
2
41
Introduction
42
Pili, or fimbriae, are hair-like appendages extending out from the surface of most bacterial
43
types, which frequently mediate attachment to various surfaces (Soto and Hultgren, 1999).
44
Among these organelles, Tfp are likely to be the most widespread based on their detection in
45
numerous and diverse Gram negative species (Mattick, 2002) and unexpectedly some Gram
46
positive ones such as Ruminococcus albus in which their presence has been experimentally
47
confirmed (Rakotoarivonina et al., 2002). In addition, genes specifically involved in Tfp
48
biogenesis were discovered in tens of genome sequencing projects. Significantly, Tfp are
49
found in human pathogens as diverse as enteropathogenic Escherichia coli (EPEC),
50
pathogenic Neisseria species, Pseudomonas aeruginosa or Vibrio cholerae, where they play a
51
key role in virulence (Soto and Hultgren, 1999; Mattick, 2002). Unlike other types of pili,
52
however, Tfp not only participate in adhesion to host cells, but are also involved in a
53
surprising number of functions (Merz and So, 2000), including competence for DNA
54
transformation, the formation of bacterial aggregates and a form of surface translocation
55
known as twitching motility, which is powered by fiber retraction (Merz et al., 2000).
56
Tfp are thin (50-80 Å), long (up to several µm) and flexible but strong fibers, which
57
often interact laterally to form typical bundles (Craig et al., 2004). These highly dynamic
58
organelles are mainly composed of thousands of copies of one protein, the pilin. Pilin is
59
synthesized as a precursor and presents distinctive N-terminal sequence pattern and three-
60
dimensional structure with an extended N-terminal spine followed by a globular head domain
61
(Parge et al., 1995). This has led to helical assembly models for Tfp where the conserved N-
62
terminal spines formed the hydrophobic core of the pilus and the variable globular heads were
63
exposed on the surface. These shared sequence and structural characteristics, together with the
64
ability of P. aeruginosa to assemble the N. gonorrhoeae pilin into fibers (Hoyne et al., 1992),
65
suggested that the machineries used by different bacteria to assemble Tfp were closely
3
66
related, which was later confirmed by the identification of Tfp biogenesis genes. This process
67
requires a strikingly large set of dedicated proteins, between 12 and 15, including the pilin
68
(Yoshida et al., 1999; Anantha et al., 2000; Carbonnelle et al., 2005). Some of these have
69
homologues in the type II secretion system used by a variety of Gram negative bacteria to
70
transport exoproteins to the outside milieu (Nunn, 1999; Peabody et al., 2003). This suggested
71
that these two systems have a common evolutionary origin and possibly similar modes of
72
function, which fueled important transfers of knowledge between the two fields of research
73
(Pugsley, 1993; Nunn, 1999; Sauvonnet et al., 2000; Durand et al., 2005). However, a
74
function could be determined for only a few of the above proteins and Tfp biogenesis remains
75
a poorly understood process at a molecular level. Nevertheless, it is clear that upon synthesis,
76
the prepilin is inserted in the inner membrane as a bitopic protein owing to its hydrophobic N-
77
terminus, leaving the leader peptide on the cytoplasmic side (Strom and Lory, 1987). This
78
leader peptide is then cleaved by a polytopic inner membrane protein (Lory and Strom, 1997),
79
the prepilin peptidase (GspO in the unifying type II secretion nomenclature). This step is
80
required for subsequent fiber assembly (Strom and Lory, 1991), which is powered by an
81
oligomeric GspE-family cytoplasmic ATPase (Turner et al., 1993; Sakai et al., 2001). Finally,
82
Tfp emerge on the cell surface through rings formed in the outer membrane by multimers of a
83
protein belonging to the GspD-family of secretins (Collins et al., 2004; Chami et al., 2005),
84
which sometimes depend on small outer membrane lipoproteins, known as pilotins, for their
85
proper localization and stability (Hardie et al., 1996; Crago and Koronakis, 1998).
86
It seems obvious that a better understanding of Tfp biogenesis relies on our ability to
87
determine the role in this process of each of the remaining proteins, ie. an overwhelming
88
majority of them. In this regard, several recent studies aimed at the identification of the
89
interactions that occur among these proteins using different approaches (Ramer et al., 2002;
90
Hwang et al., 2003; Crowther et al., 2004). This has led to a better knowledge of the
4
91
corresponding machineries and some relevant functional information. For example, it was
92
found in the EPEC that interactions of the GspE-family ATPase with two membrane proteins
93
dramatically stimulated its activity (Crowther et al., 2005). This led the authors to propose a
94
model in which the produced energy was converted into mechanical force allowing one of the
95
interacting membrane proteins to act as a piston pushing the pilin out from the inner
96
membrane, preparing it for pilus assembly. A powerful genetic approach also generated
97
crucial information by providing evidence that the piliation defect in several N. meningitidis
98
and N. gonorrhoeae mutants could be suppressed by a mutation in pilT. This demonstrated
99
that Tfp biogenesis can be resolved in discrete steps and that at least some proteins were
100
actually dispensable for Tfp assembly and served to counteract PilT-mediated fiber retraction
101
(Wolfgang et al., 1998b; Carbonnelle et al., 2005; Winther-Larsen et al., 2005). In other
102
words, the corresponding mutants were non-piliated because pilus homeostasis was shifted
103
towards retraction (Morand et al., 2004). Interestingly, the restored fibers were not functional,
104
hence the definition of this step as a Tfp stabilization/functional maturation step. Similarly,
105
fibers could be restored in a N. gonorrhoeae pilQ/T mutant but remained within the cell,
106
providing the first evidence that Tfp were assembled in the periplasm before emerging on the
107
surface through the PilQ secretin (Wolfgang et al., 2000).
108
In order to improve our understanding of Tfp biogenesis, we defined at which step of
109
this process each Pil protein acts by implementing systematically, for the first time, the
110
previous genetic approach. This recently became possible in N. meningitidis with the
111
identification of all the piliation genes in this species (Carbonnelle et al., 2005). We therefore
112
characterized in details a set of mutants in which each piliation gene was mutated together
113
with pilT. This led to several important observations, which might have an impact on our
114
understanding of the biology of one of the most widespread virulence factors in bacteria.
115
5
116
Results
117
Only PilD, PilF, PilM, PilN, PilO and PilP are required to assemble the pilin into Tfp, while
118
PilQ is the single protein required for the emergence of the fibers on the surface
119
We recently identified, by screening an almost exhaustive collection of N. meningitidis
120
mutants, the genes that were necessary, and very likely sufficient, for Tfp biogenesis in this
121
species (Carbonnelle et al., 2005). The complete set, which also comprises the double
122
pilC1/C2 mutant in which the two alleles encoding PilC were sequentially mutated, consists
123
of fifteen non-piliated mutants (Table 1). Piliation status was unambiguously determined by
124
immunofluorescence (IF) microscopy using a monoclonal antibody specific for the fibers
125
(Carbonnelle et al., 2005). In order to genetically define the specific contribution in Tfp
126
biogenesis for each of the above Pil proteins, we systematically introduced a pilT mutation
127
into each of the corresponding non-piliated mutants and characterized them for the possible
128
restoration of Tfp and Tfp-mediated virulence phenotypes. Since non-piliated mutants are not
129
competent for DNA transformation, the introduction of a pilT mutation in the above fifteen
130
mutants was performed by transforming the wild-type (WT) strain simultaneously with the
131
chromosomal DNAs extracted from a pilT mutant and one of these non-piliated mutants. As
132
above, the triple pilC1/C2/T mutant was constructed sequentially by transforming the pilC1
133
mutant simultaneously with chromosomal DNAs extracted from the pilC2 and pilT mutants.
134
We then determined whether the double (triple) mutants were piliated or not, by scoring the
135
presence or absence of Tfp by IF microscopy (Fig. 1). This showed that while variable
136
amounts of Tfp could be easily seen in seven mutants (pilC1/C2/T, pilG/T, pilH/T, pilI/T,
137
pilJ/T, pilK/T and pilW/T), not even trace amounts of fibers could be detected in the
138
remaining eight mutants (pilD/T, pilE/T, pilF/T, pilM/T, pilN/T, pilO/T, pilP/T and pilQ/T).
139
The presence of pilQ/T among these eight apparently non-piliated double mutants,
140
which was previously reported to harbour intracellular fibers in N. gonorrhoeae (Wolfgang et
6
141
al., 2000), suggested that the method we used to detect Tfp might not discriminate between
142
mutants in which the fibers were actually absent from those in which they were restored but
143
trapped within the cells. Consequently, no conclusions could be drawn at this point on
144
whether the corresponding Pil proteins were involved in the assembly of the fibers or their
145
emergence on the surface. Therefore, we designed a method allowing the distinction between
146
these two classes of mutants. Since the fibers in a N. gonorrhoeae pilQ/T mutant were very
147
likely periplasmic (Wolfgang et al., 2000), we reasoned that similar fibers should be
148
detectable after submitting the N. meningitidis mutants to a slightly modified cold osmotic
149
shock treatment usually used to release periplasmic proteins (Neu and Heppel, 1965). As
150
could be expected, although there was a clear effect of this treatment on cellular integrity as
151
evidenced by the diffuse ethidium bromide staining of the bacteria (Fig. 2), no labelling was
152
observed in non-piliated mutants such as pilD, which produces the main pilus subunit but is
153
unable to assemble it in Tfp (Strom and Lory, 1991). In contrast, numerous short Tfp were
154
visible by IF microscopy after submitting the pilQ/T mutant to osmotic shock treatment (Fig.
155
2), which confirmed the efficacy of this procedure for releasing intraperiplasmic fibers and
156
demonstrated that N. meningitidis and N. gonorrhoeae pilQ/T mutants presented the same
157
phenotype. We then re-analyzed the piliation of the eight mutants in which no fibers could be
158
detected in classical IF microscopy conditions (pilD/T, pilE/T, pilF/T, pilM/T, pilN/T, pilO/T,
159
pilP/T and pilQ/T) after submitting them to this osmotic shock procedure. This demonstrated
160
that pilQ/T was the only mutant in which Tfp could be released in this way and subsequently
161
detected (Fig. 2).
162
Next, in order to determine whether the seven mutants in which no Tfp could be
163
restored (pilD/T, pilE/T, pilF/T, pilM/T, pilN/T, pilO/T and pilP/T) still expressed and matured
164
the prepilin PilE, the main pilus component, whole-cell protein extracts were prepared,
165
separated by SDS-Page and subjected to Western blotting using a monoclonal antibody
7
166
directed against this protein. Besides the pilE mutant that obviously produced no pilin and the
167
prepilin peptidase pilD mutant in which only the prepilin could be detected, no differences in
168
the expression and/or maturation of PilE were observable in the pilF/T, pilM/T, pilN/T, pilO/T
169
and pilP/T mutants (Fig. 3). This finding indicated that the absence of Tfp in these mutants
170
was not due to an impaired expression or maturation of the main pilus component PilE and
171
suggested that the corresponding proteins could actually be involved in the first step of Tfp
172
biogenesis, ie. the assembly of the fibers. However, since the pilM, pilN, pilO and pilP genes
173
are likely to be co-transcribed, the possibility that at least some of the phenotypes we
174
observed were due to polar effects of the mutations on distal gene expression could not be
175
excluded at this point. Therefore, we complemented the pilM, pilN, pilO and pilP mutants by
176
inserting in trans in their genomes an intact copy of the corresponding genes under the
177
transcriptional control of an IPTG-inducible promoter. IF microscopy observation of the
178
complemented mutants showed that they were all piliated (data not shown). This
179
demonstrated that piliation in the pilM, pilN, pilO and pilP mutants could be restored by intact
180
copies of the mutated genes, ruling out the possibility that the non-piliated phenotypes we
181
observed resulted from polar effects on downstream gene expression and strenghtening the
182
hypothesis that the corresponding proteins were indeed involved in Tfp assembly.
183
Taken together, these results suggested that a minority of the fifteen N. meningitidis Pil
184
proteins (PilD, PilF, PilM, PilN, PilO and PilP) were involved in the assembly stricto sensu of
185
PilE into Tfp and that PilQ was likely to be the sole Pil protein involved in the emergence of
186
the fibers on the surface.
187
188
PilW is the only protein that affects the stability of PilQ multimers
189
One possibility to learn more about the possible interactions between Pil proteins is to study
190
the stability of one or more of these proteins in the absence of the others (Ramer et al., 2002).
8
191
We previously reported that the PilQ multimers were strongly destabilized in the absence of
192
PilW (Carbonnelle et al., 2005), which was recently confirmed in Myxococcus xanthus with
193
the orthologous protein Tgl (Nudleman et al., 2006). This prompted us to test systematically
194
by Western blotting, using an antibody directed against the N. meningitidis secretin, whether
195
the absence of the other Pil components might have a negative effect on the stability of the
196
PilQ multimers. These high molecular weight species species of PilQ can be readily detected
197
because they are SDS-resistant and retained in the stacking gel after SDS-PAGE separation of
198
whole-cell protein extracts (Fig. 4). This systematic analysis indicated that an effect on
199
secretin multimers could be seen only in the absence of PilW. Indeed, while no high
200
molecular weight species species of PilQ can be detected in a pilW mutant, as previously
201
reported (Carbonnelle et al., 2005), all the other mutants presented amounts of both
202
multimeric and monomeric forms of PilQ similar to those that could be detected in the WT
203
strain (Fig. 4). This showed that, apart from PilW, no other N. meningitidis Pil protein was
204
important for the stability of the PilQ multimers, suggesting that a specific interaction occured
205
between these two proteins.
206
207
PilG and PilH are the only proteins involved in Tfp stabilization that are, at least partly,
208
dispensable for fiber functionality
209
As seen above (Fig. 1), Tfp could be detected by IF microscopy in classical conditions in
210
seven mutants (pilC1/C2/T, pilG/T, pilH/T, pilI/T, pilJ/T, pilK/T and pilW/T), including the
211
previously characterized pilW/T (Carbonnelle et al., 2005). This indicated that the fibers in
212
these mutants were surface exposed unlike the fibers in the pilQ/T mutant. As far as can be
213
appreciated by transmission electron microscopy at high magnification, the fibers in these
214
mutants were morphologically indistinguishable from those seen in the WT strain (Fig. 5). In
215
each case, few long Tfp could be seen extending out from the cells as large bundles of
9
216
laterally interacting fibers, although the bundles observed in the pilC1/C2/T and pilW/T
217
mutants seemed somewhat thinner. These findings clearly indicated that the corresponding
218
proteins (PilC, PilG, PilH, PilI, PilJ, PilK and PilW) were not canonical assembly components
219
since biogenesis of apparently normal Tfp was possible in their absence. The availability of N.
220
meningitidis strains expressing Tfp in the absence of the above proteins offered the unique
221
opportunity to address their possible importance for Tfp-mediated phenotypes, as we
222
previously did with the pilW/T mutant (Carbonnelle et al., 2005). We therefore characterized
223
the ability of the above piliated mutants to promote adhesion to human umbilical vein
224
endothelial cells (HUVEC) and to mediate the formation of aggregates, the two Tfp-linked
225
poperties that are not abolished in the absence of PilT unlike competence for DNA
226
transformation and twitching motility (Wolfgang et al., 1998a).
227
First, we tested the adhesive abilities to HUVEC of each of the pilC1/C2/T, pilG/T,
228
pilH/T, pilI/T, pilJ/T, pilK/T and pilW/T mutants using a classical 4 hours adhesion assay (Fig.
229
6A). Despite their piliated phenotype, most of these mutants (pilC1/C2/T, pilI/T, pilJ/T,
230
pilK/T and pilW/T) were unable to adhere to HUVEC, which was further confirmed by
231
counting the adherent colony-forming units (cfu) (Fig. 6A). The adhesive abilities of these
232
mutants were found to be as low as that of a non-piliated mutant with mean 1,000-fold
233
reduction in adherence when compared to the WT strain or a pilT mutant (approx. 105 versus
234
108 adherent cfu). Strikingly, the pilG/T and pilH/T mutants presented a phenotype previously
235
never reported, ie. Tfp that were restored were capable of mediating efficient adhesion to
236
human cells. Although the adhesive abilities of the pilG/T and pilH/T mutants were slightly
237
lower than that of the WT strain or a pilT mutant, they were able to adhere to HUVEC at least
238
two orders of magnitude better than a non-piliated mutant, with 100-fold and 400-fold
239
increases respectively (Fig. 6A). This was further evidenced by quantifying pilG/T and pilH/T
240
adherence in more details over the course of the infection, which was found to present
10
241
kinetics similar to that of the WT strain, from the very beginning of the adherence assay (Fig.
242
6B). Interestingly, however, the three-dimensional bacterial micro-colonies that these two
243
mutants formed on the cells, which did not disappear during the adhesion due to the absence
244
of PilT (Pujol et al., 1999), were morphologically different from those seen with a pilT
245
mutant. They seemed less compact and firm, which suggested that although adhesion to
246
HUVEC occurred in the absence of PilG and PilH, these proteins might be required for full
247
Tfp functionality.
248
Next, by observing liquid cultures of the above seven mutants by phase-contrast
249
microscopy, we found that most of them were unable to form multicellular aggregates, with
250
the notable exception of the pilG/T and pilH/T mutants that however formed aggregates
251
morphologically different from those highly irregular ones produced by a pilT mutant
252
(Helaine et al., 2005). Using a method we developed recently, based on the measuring of the
253
decrease in optical density (OD) that occurs in non-agitated liquid cultures upon
254
sedimentation of the bacterial aggregates (Helaine et al., 2005), we precisely quantified the
255
aggregative abilities of all these mutants (Fig. 7). The kinetics of aggregation that were
256
measured for the various strains were consistent with the phenotypes that were observed by
257
phase-contrast microscopy. Indeed, while pilC1/C2/T, pilI/T, pilJ/T, pilK/T and pilW/T
258
presented a nil aggregation, just as a control non-piliated mutant, the pilG/T and pilH/T
259
mutants presented aggregative abilities comparable to that of the WT strain (Fig. 7). This was
260
coherent with the results of the adhesion assays in which pilG/T and pilH/T were the only
261
mutants able to adhere to HUVEC and once again underscored the invariable direct link
262
between Tfp-mediated aggregation and adhesion to human cells (Helaine et al., 2005).
263
However, despite their PilT-negative background, the aggregative abilities of pilG/T and
264
pilH/T were lower than that of the pilT mutant, which almost completely and rapidly
11
265
sedimented (Fig. 7). Again, as with the adhesion assays, this indicated that PilG and PilH
266
contributed to full Tfp functionality.
267
Together, these results indicated that as much as seven Pil proteins (PilC, PilG, PilH,
268
PilI, PilJ, PilK and PilW) were involved in regulating pilus homeostasis by counteracting the
269
PilT-mediated retraction of the fibers (Morand et al., 2004), which emphasized the extreme
270
importance of this step in Tfp biogenesis. Importantly, although most of these proteins were
271
also essential for Tfp functionality as previously reported for PilW (Carbonnelle et al., 2005),
272
the absence of PilG and PilH was less deleterious, suggesting that the tightly linked events of
273
stabilization and functional maturation could be genetically resolved.
274
12
275
Discussion
276
When compared to the well-characterized chaperone-usher pathway that relies on two
277
proteins to assemble structural subunits into a pilus on the bacterial surface (Soto and
278
Hultgren, 1999), the biogenesis of Tfp, which requires 11-14 dedicated proteins in addition to
279
the pilin, seems utterly complex and remains poorly understood. Interestingly, however, it has
280
been shown that in several Neisseria mutants the piliation defect could be suppressed by
281
introducing a mutation in pilT (Wolfgang et al., 1998b; Wolfgang et al., 2000; Carbonnelle et
282
al., 2005; Winther-Larsen et al., 2005). This suggested that the corresponding Pil proteins
283
acted after Tfp assembly either to stabilize the fibers by counteracting PilT-mediated
284
retraction or to allow their emergence on the bacterial surface. It should be noted that all the
285
proteins found to counteract PilT action were also necessary for the functionality of Tfp since
286
the restored fibers were incapable of mediating well-known Tfp-linked phenotypes, adding to
287
the complexity of Tfp biogenesis/function. Together, these findings led to a three-step model
288
for Tfp biogenesis (Wolfgang et al., 2000; Carbonnelle et al., 2005), where fibers were first
289
assembled in the periplasm, then altogether stabilized and functionally matured and finally
290
emerged on the surface through a channel formed by the secretin. The recent identification of
291
the complete set of N. meningitidis genes specifically involved in Tfp biogenesis (Carbonnelle
292
et al., 2005), whose number and genomic organization were found to be strikingly similar to
293
those in P. aeruginosa (Alm and Mattick, 1997), prompted us to implement, for the first time,
294
the above genetic approach on a systematic basis in order to define at which step of Tfp
295
biogenesis each of the corresponding proteins was involved.
296
One important finding in this study was that the first step of Tfp biogenesis, the actual
297
assembly of the fibers, is simpler than could be anticipated because it required, in addition to
298
the pilin, a core set of only six proteins (PilD, PilF, PilM, PilN, PilO and PilP). Since the pilM
299
to pilP genes are very likely to be co-transcribed, it is important to note that complementation
13
300
experiments excluded the possibility that the piliation defects observed in the corresponding
301
mutants were due to polar effects on the expression of downstream genes. Furthermore, the
302
non-piliated phenotypes of these mutants were neither due to reduced amounts of pilin nor to
303
an impaired maturation of this protein. Therefore, if the prepilin peptidase and PilF are only
304
necessary, respectively, for the maturation of the prepilin and to provide the energy necessary
305
to push the pilin outside the inner membrane, the four PilM, PilN, PilO and PilP proteins
306
might constitute the essence of the machinery necessary for the mechanical assembly of Tfp.
307
In accord with this hypothesis, these proteins were reported to be part of the core set of
308
conserved Pil proteins in the cyanobacteria and ß-, -, and ∂-proteobacteria harbouring Tfp
309
(Nudleman and Kaiser, 2004). These findings also point to the possibility of creating a
310
minimal Tfp biogenesis system in a heterologous non-piliated organism by transferring the
311
above genes on suitable expression vectors, which could be instrumental to the
312
characterization of the underlying molecular mechanisms of pilus assembly.
313
A corollary of the above findings is that a significant proportion of the Pil proteins
314
(PilC, PilG, PilH, PilI, PilJ, PilK and PilW) is involved in counteracting PilT action on the
315
fibers, since morphologically normal Tfp can be expressed in their absence. This underscores
316
the unexpectedly central role in Tfp biogenesis of pilus homeostasis, which results from the
317
balance between retraction and counter-retraction (Morand et al., 2004). In accord with
318
previous reports (Wolfgang et al., 1998b; Carbonnelle et al., 2005; Winther-Larsen et al.,
319
2005), we found that this stabilization step is most often linked to what we previously defined
320
as functional maturation of the fibers (Carbonnelle et al., 2005). Indeed, a precise
321
quantification of Tfp-linked properties clearly demonstrated that the fibers in most of these
322
mutants (pilC1/C2/T, pilI/T, pilJ/T, pilK/T and pilW/T) were not functional since they were
323
incapable of mediating adhesion to human cells and the formation of bacterial aggregates.
324
However, one of the most striking findings in this study was the discovery of a novel class of
14
325
double mutants (pilG/T and pilH/T) where the fibers that were restored were able to mediate
326
adhesion to human cells and the formation of bacterial aggregates. However, these mutants
327
did not behave as pilT mutants, which indicated that although PilG and PilH role in Tfp
328
functionality was minor when compared to the role of PilC, PilI, PilJ, PilK and PilW, these
329
proteins nevertheless participated in the modulation of fibers functionality. This suggested
330
that Tfp biogenesis is not a three-step pathway as previously thought because the counter-
331
retraction and functional maturation activities could be genetically separated. Interestingly,
332
one of the two proteins found to be associated with such a phenotype was the polytopic inner
333
membrane protein PilG (GspF), which is one of the four extremely conserved proteins in both
334
Tfp biogenesis and type II secretion (Peabody et al., 2003) and as such thought to be a key
335
player or even the corner stone in both systems (Py et al., 2001; Crowther et al., 2004). Our
336
data dot not support this possibility in Tfp biogenesis since piliation could be restored in N.
337
meningitidis in the absence of PilG. This clearly indicates that PilG is not a canonical pilus
338
assembly factor in N. meningitidis, which could be tested in other bacterial species using the
339
same approach. However, it is unclear whether and how this finding might apply to type II
340
secretion and the biogenesis of pilus-like structures by this machinery (Sauvonnet et al., 2000;
341
Durand et al., 2005) since no PilT-like protein is present in this system. Concerning PilH, the
342
second protein found to be associated with such a phenotype, our findings might be surprising
343
in the first place because this protein belongs to a group of four related proteins, PilH to PilK,
344
that were all shown to be cleaved by PilD in N. gonorrhoeae and because a pilH/T N.
345
gonorrhoeae mutant apparently failed to adhere to human cells (Winther-Larsen et al., 2005).
346
However, a possible clue as to PilH particular properties in N. meningitidis when compared to
347
PilI, PilJ and PilK comes from ClustalW multiple sequence alignments of these proteins,
348
which show that while PilI, PilJ and PilK cluster together, PilH is clearly more distant and
349
more closely related to PilE and the minor pilins (data not shown).
15
350
The design of a suitable method allowing the detection of intracellular fibers by IF
351
helped us confirm that fibers were indeed restored intracellularly in a pilQ/T mutant, which
352
was previously demonstrated in N. gonorrhoeae in which fibers embedded in membranous
353
protrusions were seen by transmission electron microscopy (Wolfgang et al., 2000).
354
Furthermore, the use of this method on a systematic basis demonstrated that intraperiplasmic
355
fibers were seen exclusively in the absence of the secretin. Together, these findings suggested
356
that PilQ's specific role was to provide a route to the surface for Tfp and that it was the only
357
Pil protein implicated in this step of Tfp biogenesis, which has several important implications.
358
First, our data argue against the possibility that the fibers in the pilW/T mutant, which are
359
perfectly exposed on the surface despite an apparent absence of PilQ multimers, used a
360
completely different route to the bacterial surface. Otherwise that alternative route would have
361
also been available in the pilQ/T mutant and its fibers would have been exposed on the
362
surface instead of being trapped in the periplasm. PilQ multimers were therefore likely to be
363
still present in the absence of PilW, allowing the emergence of the fibers in the pilW/T
364
mutant, but they might be too unstable to be detected after SDS PAGE, which remains to be
365
actually proven. This also suggested that the unstability of the PilQ multimers in the absence
366
of PilW should merely be viewed as an indication that these proteins physically interact,
367
rather than evidence that PilW takes an active part in the assembly of a functional secretin
368
complex as recently proposed in M. xanthus for the orthologous protein Tgl (Nudleman et al.,
369
2006). A second implication of our results was that PilQ in N. meningitidis probably does not
370
rely on a dedicated pilot protein for its insertion in the outer membrane, similarly to the
371
situation in EPEC (Schmidt et al., 2001). Indeed, it is likely that the introduction of a pilT
372
mutation in a mutant missing such a pilot protein would lead to the restoration of
373
intraperiplasmic fibers, since no secretin channel would be present in the outer membrane.
374
However, this was never observed except for the pilQ/T mutant. Interestingly, it seems that in
16
375
N. meningitidis PilQ multimerization and insertion in the outer membrane is rather under the
376
dependance of Omp85, an evolutionarily conserved protein ubiquitous among Gram-negative
377
bacteria thought to be involved in the membrane insertion and/or multimerization of most, if
378
not all, bacterial outer membrane proteins (Voulhoux et al., 2003). Therefore, although this
379
has been a matter of speculation for some time based on the report that PilP affects the
380
expression of PilQ multimers in N. gonorrhoeae (Drake et al., 1997), which was not observed
381
in N. meningitidis nor in M. xanthus (Nudleman et al., 2006), a role as a pilotin can a priori
382
be excluded for this protein. Instead, PilP was found to be a member of the core set of
383
proteins necessary for the mechanical assembly of Tfp.
384
In conclusion, the findings here offer some clues as to the mechanisms of Tfp
385
biogenesis/function since they provide a four-step comprehensive scheme for the biogenesis
386
of functional Tfp on which each of the fifteen Pil proteins essential for piliation in N.
387
meningitidis could be placed (Fig. 8). According to this model, which could be tested in other
388
bacteria expressing Tfp, non-functional fibers are assembled in the periplasm by the PilD,
389
PilF, PilM, PilN, PilO and PilP subset of proteins. Although the chronology of the steps after
390
Tfp assembly remains to be precisely determined, one could speculate that these fibers Tfp are
391
then taken in charge by the PilC, PilI, PilJ, PilK and PilW subset of proteins that make Tfp
392
functional and then antagonize their retraction with the participation of PilG and PilH
393
(although these two proteins also play a minor role in functional maturation). Finally, the
394
fibers emerge on the surface through a channel formed by the secretin PilQ. This scenario
395
could be instrumental in designing approaches aimed at understanding the above steps at a
396
molecular level, such as for example testing whether the functional maturation of the fibers
397
might be the result of subtle alterations in Tfp structure and/or composition, which could have
398
an impact on our understanding of what makes Tfp extremely efficient virulence factors.
399
17
400
Experimental procedures
401
Bacterial strains: culture conditions and construction
402
E. coli TOP10 (Invitrogen) was used for cloning experiments. It was grown at 37°C in liquid
403
or solid Luria-Bertani medium (Difco), which contained 200 g/ml erythromycin and 50
404
g/ml kanamycin, when appropriate.
405
N. meningitidis was grown at 37°C in a moist atmosphere containing 5% CO2 on GCB
406
agar plates (Difco), which contained Kellogg's supplements and, when appropriate, 3 g/ml
407
erythromycin, 100 g/ml kanamycin and 60 g/ml spectinomycin (Pelicic et al., 1997).
408
Liquid cultures were performed in RPMI 1640 medium supplemented with 10% heat-
409
inactivated fetal calf serum (both from PAA laboratories GmBH) as previously described
410
(Helaine et al., 2005). The WT strain, a derivative of the serogroup C 8013 isolate, and most
411
of its non-piliated mutants were described previously (Nassif et al., 1993; Carbonnelle et al.,
412
2005). These mutants contain a mini-transposon, consisting of mariner Himar1-based
413
inverted repeats flanking a kanamycin resistance cassette, inserted in the N-terminal halves of
414
the respective pil genes (Pelicic et al., 1997; Geoffroy et al., 2003). In addition, we
415
constructed for this study the pilC1/C2 double mutant by transforming a pilC2 transposition
416
mutant with the chromosomal DNA of a pilC1 mutant, in which the corresponding gene is
417
partly deleted and replaced by a spectinomycin resistance cassette (Morand et al., 2001).
418
Chromosomal DNA preparation and transformation of N. meningitidis were done as
419
previously described (Geoffroy et al., 2003). Since all the above non-piliated mutants are not
420
competent for transformation, the introduction of a pilT mutation in the various mutant
421
backgrounds, where the pilT gene is interrupted by an erythromycin resistance cassette (Pujol
422
et al., 1999), was performed by transforming the WT strain simultaneously with the
423
chromosomal DNAs extracted from the pilT mutant and the respective non-piliated mutants.
424
Similarly to the above conditions, the pilC1/C2/T triple mutant was constructed by
18
425
transforming the pilC2 mutant simultaneously with the chromosomal DNAs extracted from
426
the pilT and pilC1 mutants. Primers specific for each pil gene, including pilT, were designed
427
when they were not available and used to check by PCR that the mutants contained the
428
expected mutant alleles. The pilC series mutants were also checked by Southern blotting, as
429
previously described (Pelicic et al., 1997), using as a probe a PCR fragment in the region
430
common to the pilC1 and pilC2 genes.
431
To complement the pilM, pilN, pilO and pilP mutants, the corresponding wild-type
432
genes
were
amplified
using
433
CCTTAATTAAGGAGTAATTTTATGCGCTTGTTTAAAAGCTTGA-3' and pilM-IndR 5'-
434
CCTTAATTAATTATAATCCCCGTACCGCCAA-3',
435
CCTTAATTAAGGAGTAATTTTATGAACAATTTAATCAAAATCAAC-3' and pilN-IndR
436
5'-CCTTAATTAATCAGTTTGCCTCCTGTGCGTT-3',
437
CCTTAATTAAGGAGTAATTTTATGGCTTCTAAATCATCTAAAAC-3' and pilO-IndR
438
5'-CCTTAATTAATTATTTTTGCTCGGCATTTTGTG-3'
439
CCTTAATTAAGGAGTAATTTTATGAAACACTATGCCTTACTCA-3' and pilP-IndR 5'-
440
CCTTAATTAATTAATTTTGTTCTGCGGCAGG-3', which contained overhangs with
441
underlined restriction sites for PacI. These PCR fragments were first cloned into pCRII-
442
TOPO (Invitrogen), generating plasmids pYU21, pYU22, pYU23 and pYU24 respectively,
443
and verified by sequencing to contain no errors. The fragments were excised from the above
444
plasmids by PacI digestion and cloned into PacI-cut pGCC4 vector, adjacent to the lacIOP
445
regulatory sequences, generating plasmids pYU25, pYU26, pYU27 and pYU28, respectively.
446
This placed the pil genes under the transcriptional control of an IPTG-inducible promoter
447
within a DNA fragment corresponding to an intragenic region of the gonococcal chromosome
448
conserved in N. meningitidis (Mehr et al., 2000). The inducible alleles were then introduced
449
into the chromosome of the corresponding mutants by homologous recombination. However,
19
primers
pilM-IndF
pilN-IndF
pilO-IndF
and
pilP-IndF
5'-
5'-
5'-
5'-
450
since these mutants are non-piliated and therefore not competent for transformation, this was
451
performed in a reverse fashion by first transforming the WT strain with the NotI-cut pYU25,
452
pYU26, pYU27 or pYU28 plasmids. Then the wild-type pil allele was interrupted in these
453
transformants by a second transformation with the chromosomal DNAs extracted from the
454
pilM, pilN, pilO or pilP non-piliated mutants, respectively.
455
456
Tfp detection and morphological analysis
457
N. meningitidis pili were detected by IF microscopy on bacteria fixed on coverslips as
458
previously described (Carbonnelle et al., 2005). Fibers were specifically labelled with the
459
anti-Tfp 20D9 mouse monoclonal antibody (Pujol et al., 1997) used at a 1/100 dilution, while
460
the bacteria were stained with ethidium bromide at a 1/6,000 dilution. The secondary
461
antibody, also used at a 1/100 dilution, was a goat anti-mouse antibody conjugated with Alexa
462
Fluor 488 (Molecular Probes). When indicated, bacteria were first submitted to a modified
463
cold osmotic shock treatment (Neu and Heppel, 1965). They were resuspended in 30 mM Tris
464
(pH 7.5), 20% sucrose, 2 mM EDTA, 1 mg/ml lysozyme and incubated 10 min at room
465
temperature under constant agitation. Cells were then pelleted by centrifugation at 9,000 g for
466
10 min and resuspended in an equal volume of cold 5 mM MgSO4. After a 10 min incubation
467
at 4°C, bacteria were fixed on coverslips and their fibers detected by IF microscopy as above.
468
Fine structure of the fibers was determined by TEM, after negative staining with 1%
469
phosphotungstic acid, using a JEOL JEM-100CX microscope operated at 80 kV as previously
470
described (Carbonnelle et al., 2005).
471
472
Phenotypic analyses: aggregation and adhesion to HUVEC
473
N. meningitidis ability to form aggregates was monitored by phase-contrast microscopy after
474
resuspending the bacteria in RPMI-serum, as previously described (Helaine et al., 2005).
20
475
Aggregation was quantified by measuring the changes in OD600 that occur in static cultures
476
upon sedimentation of the aggregates, as previously described (Helaine et al., 2005).
477
Adhesion of N. meningitidis to HUVEC was quantified as described previously
478
(Helaine et al., 2005). In brief, monolayers of 105 cells, in 24-wells plates, were infected with
479
approximately 3 x 107 cfu/ml. After 30 min of contact between the bacteria and the cells,
480
unbound bacteria were removed by performing several washes and the infection was
481
continued for 4 h with washes every hour. Adherent bacteria, recovered by scraping the wells,
482
were then counted by plating appropriate dilutions on GCB agar plates.
483
484
Preparation of N. meningitidis protein extracts
485
Whole-cell extracts were prepared as described elsewhere (Winther-Larsen et al., 2005),
486
using bacteria grown on GCB agar plates, which were first resuspended in cold PBS at an
487
OD600 of 1. Ten ml of this suspension were then pelleted by centrifugation at 9,000 g for 10
488
min and resuspended in 1 ml of cold lysis buffer containing 200 mM Tris (pH 7.5), 20%
489
acetone, 40 mM EDTA, 0.1% Triton X100. This suspension was incubated 15 min at 4°C and
490
the bacteria were briefly centrifuged for 20 sec using a bench mini-centrifuge. The
491
supernatant was recovered and solubilized proteins concentrations were determined using the
492
Bio-Rad Protein Assay (Bio-Rad) following the manufacturer's instructions.
493
494
SDS-PAGE and Western blotting
495
Separaration of the proteins by SDS-PAGE and subsequent transfer to nitrocellulose
496
membranes were done as previously described (Carbonnelle et al., 2005). Blocking,
497
incubation with primary/secondary antibodies and detection using ECL Plus (Amersham
498
Biosciences) were done as outlined in the ECL Plus Western Blotting Detection Manual.
499
When studying the oligomers formed by PilQ, which are resistant to SDS and are retained in
21
500
the stacking gel due to their high molecular weight, the stacking gel was kept and transferred
501
as well. All the antibodies were used at a 1/10,000 dilution. The class 4 outer membrane
502
RmpM protein, used as a control to check the protein preparations, was detected using a
503
mouse monoclonal serum (Morand et al., 2001). PilE and PilQ were detected, respectively,
504
using the 5C5 mouse monoclonal serum (Marceau et al., 1998) and a rabbit polyclonal serum
505
donated by T. Tønjum (University of Oslo, Norway). The secondary antibodies were anti-
506
mouse or anti-rabbit horseradish peroxidase-linked IgG (Amersham Biosciences).
507
22
508
Acknowledgements
509
We thank T. Tønjum (University of Oslo, Norway) for the kind gift of anti-PilQ antibody. We
510
are grateful to Guillaume Duménil, Patricia Martin and Chiara Recchi for critical reading of
511
the manuscript. This work was supported by INSERM, Université Paris Descartes and
512
Agence Nationale de la Recherche (JC05_44953 grant to V. Pelicic).
513
23
514
References
515
Alm, R.A., and Mattick, J.S. (1997) Genes involved in the biogenesis and function of type-4
516
fimbriae in Pseudomonas aeruginosa. Gene 192: 89-98.
517
Anantha, R.P., Stone, K.D., and Donnenberg, M.S. (2000) Effects of bfp mutations on
518
biogenesis of functional enteropathogenic Escherichia coli type IV pili. J Bacteriol 182:
519
2498-2506.
520
Carbonnelle, E., Helaine, S., Prouvensier, L., Nassif, X., and Pelicic, V. (2005) Type IV pilus
521
biogenesis in Neisseria meningitidis: PilW is involved in a step occuring after pilus assembly,
522
essential for fiber stability and function. Mol Microbiol 55: 54-64.
523
Chami, M., Guilvout, I., Gregorini, M., Remigy, H.W., Muller, S.A., Valerio, M. et al. (2005)
524
Structural insights into the secretin PulD and its trypsin-resistant core. J Biol Chem 280:
525
37732-37741.
526
Collins, R.F., Frye, S.A., Kitmitto, A., Ford, R.C., Tonjum, T., and Derrick, J.P. (2004)
527
Structure of the Neisseria meningitidis outer membrane PilQ secretin complex at 12 Å
528
resolution. J Biol Chem 279: 39750-39756.
529
Crago, A.M., and Koronakis, V. (1998) Salmonella InvG forms a ring-like multimer that
530
requires the InvH lipoprotein for outer membrane localization. Mol Microbiol 30: 47-56.
531
Craig, L., Pique, M.E., and Tainer, J.A. (2004) Type IV pilus structure and bacterial
532
pathogenicity. Nat Rev Microbiol 2: 363-378.
533
Crowther, L.J., Anantha, R.P., and Donnenberg, M.S. (2004) The inner membrane
534
subassembly of the enteropathogenic Escherichia coli bundle-forming pilus machine. Mol
535
Microbiol 52: 67-79.
536
Crowther, L.J., Yamagata, A., Craig, L., Tainer, J.A., and Donnenberg, M.S. (2005) The
537
ATPase activity of BfpD is greatly enhanced by zinc and allosteric interactions with other Bfp
538
proteins. J Biol Chem 280: 24839-24848.
24
539
Drake, S.L., Sandstedt, S.A., and Koomey, M. (1997) PilP, a pilus biogenesis lipoprotein in
540
Neisseria gonorrhoeae, affects expression of PilQ as a high-molecular-mass multimer. Mol
541
Microbiol 23: 657-668.
542
Durand, E., Michel, G., Voulhoux, R., Kurner, J., Bernadac, A., and Filloux, A. (2005) XcpX
543
controls biogenesis of the Pseudomonas aeruginosa XcpT-containing pseudopilus. J Biol
544
Chem 280: 31378-31389.
545
Geoffroy, M.-C., Floquet, S., Métais, A., Nassif, X., and Pelicic, V. (2003) Large-scale
546
analysis of the meningococcus genome by gene disruption: resistance to complement-
547
mediated lysis. Genome Res 13: 391-398.
548
Hardie, K.R., Lory, S., and Pugsley, A.P. (1996) Insertion of an outer membrane protein in
549
Escherichia coli requires a chaperone-like protein. EMBO J 15: 978-988.
550
Helaine, S., Carbonnelle, E., Prouvensier, L., Beretti, J.-L., Nassif, X., and Pelicic, V. (2005)
551
PilX, a pilus-associated protein essential for bacterial aggregation, is a key to pilus-facilitated
552
attachment of Neisseria meningitidis to human cells. Mol Microbiol 55: 65-77.
553
Hoyne, P.A., Haas, R., Meyer, T.F., Davies, J.K., and Elleman, T.C. (1992) Production of
554
Neisseria gonorrhoeae pili (fimbriae) in Pseudomonas aeruginosa. J Bacteriol 174: 7321-
555
7327.
556
Hwang, J., Bieber, D., Ramer, S.W., Wu, C.Y., and Schoolnik, G.K. (2003) Structural and
557
topographical studies of the type IV bundle-forming pilus assembly complex of
558
enteropathogenic Escherichia coli. J Bacteriol 185: 6695-6701.
559
Lory, S., and Strom, M.S. (1997) Structure-function relationship of type-IV prepilin peptidase
560
of Pseudomonas aeruginosa--a review. Gene 192: 117-121.
561
Marceau, M., Forest, K., Beretti, J.-L., Tainer, J., and Nassif, X. (1998) Consequences of the
562
loss of O-linked glycosylation of meningococcal type IV pilin on piliation and pilus-mediated
563
adhesion. Mol Microbiol 27: 705-715.
25
564
Mattick, J.S. (2002) Type IV pili and twitching motility. Annu Rev Microbiol 56: 289-314.
565
Mehr, I.J., Long, C.D., Serkin, C.D., and Seifert, H.S. (2000) A homologue of the
566
recombination-dependent growth gene, rdgC, is involved in gonococcal pilin antigenic
567
variation. Genetics 154: 523-532.
568
Merz, A.J., and So, M. (2000) Interactions of pathogenic Neisseriae with epithelial cell
569
membranes. Annu Rev Cell Dev Biol 16: 423-457.
570
Merz, A.J., So, M., and Sheetz, M.P. (2000) Pilus retraction powers bacterial twitching
571
motility. Nature 407: 98-102.
572
Morand, P.C., Tattevin, P., Eugène, E., Beretti, J.-L., and Nassif, X. (2001) The adhesive
573
property of the type IV pilus-associated component PilC1 of pathogenic Neisseria is
574
supported by the conformational structure of the N-terminal part of the molecule. Mol
575
Microbiol 40: 846-856.
576
Morand, P.C., Bille, E., Morelle, S., Eugène, E., Beretti, J.L., Wolfgang, M. et al. (2004)
577
Type IV pilus retraction in pathogenic Neisseria is regulated by the PilC proteins. EMBO J
578
23: 2009-2017.
579
Nassif, X., Lowy, J., Stenberg, P., O'Gaora, P., Ganji, A., and So, M. (1993) Antigenic
580
variation of pilin regulates adhesion of Neisseria meningitidis to human epithelial cells. Mol
581
Microbiol 8: 719-725.
582
Neu, H.C., and Heppel, L.A. (1965) The release of enzymes from Escherichia coli by osmotic
583
shock and during the formation of spheroplasts. J Biol Chem 240: 3685-3692.
584
Nudleman, E., and Kaiser, D. (2004) Pulling together with type IV pili. J Mol Microbiol
585
Biotechnol 7: 52-62.
586
Nudleman, E., Wall, D., and Kaiser, D. (2006) Polar assembly of the type IV pilus secretin in
587
Myxococcus xanthus. Mol Microbiol 60: 16-29.
26
588
Nunn, D. (1999) Bacterial type II protein export and pilus biogenesis: more than just
589
homologies? Trends Cell Biol 9: 402-408.
590
Parge, H.E., Forest, K.T., Hickey, M.J., Christensen, D.A., Getzoff, E.D., and Tainer, J.A.
591
(1995) Structure of the fibre-forming protein pilin at 2.6 Å resolution. Nature 378: 32-38.
592
Peabody, C.R., Chung, Y.J., Yen, M.R., Vidal-Ingigliardi, D., Pugsley, A.P., and Saier, M.H.,
593
Jr. (2003) Type II protein secretion and its relationship to bacterial type IV pili and archaeal
594
flagella. Microbiology 149: 3051-3072.
595
Pelicic, V., Jackson, M., Reyrat, J.-M., Jacobs, W.R., Jr., Gicquel, B., and Guilhot, C. (1997)
596
Efficient allelic exchange and transposon mutagenesis in Mycobacterium tuberculosis. Proc
597
Natl Acad Sci USA 94: 10955-10960.
598
Pugsley, A.P. (1993) The complete general secretory pathway in gram-negative bacteria.
599
Microbiol Rev 57: 50-108.
600
Pujol, C., Eugène, E., de Saint Martin, L., and Nassif, X. (1997) Interaction of Neisseria
601
meningitidis with a polarized monolayer of epithelial cells. Infect Immun 65: 4836-4842.
602
Pujol, C., Eugène, E., Marceau, M., and Nassif, X. (1999) The meningococcal PilT protein is
603
required for induction of intimate attachment to epithelial cells following pilus-mediated
604
adhesion. Proc Natl Acad Sci USA 96: 4017-4022.
605
Py, B., Loiseau, L., and Barras, F. (2001) An inner membrane platform in the type II secretion
606
machinery of Gram-negative bacteria. EMBO Rep 2: 244-248.
607
Rakotoarivonina, H., Jubelin, G., Hebraud, M., Gaillard-Martinie, B., Forano, E., and Mosoni,
608
P. (2002) Adhesion to cellulose of the Gram-positive bacterium Ruminococcus albus involves
609
type IV pili. Microbiology 148: 1871-1880.
610
Ramer, S., Schoolnik, G., Wu, C.-Y., Hwang, J., Schmidt, S., and Bieber, D. (2002) The type
611
IV pilus assembly complex: biogenic interaction among the bundle-forming pilus proteins of
612
enteropathogenic Escherichia coli. J Bacteriol 184: 3457-3465.
27
613
Sakai, D., Horiuchi, T., and Komano, T. (2001) ATPase activity and multimer formation of
614
PilQ protein are required for thin pilus biogenesis in plasmid R64. J Biol Chem 276: 17968-
615
17975.
616
Sauvonnet, N., Vignon, G., Pugsley, A.P., and Gounon, P. (2000) Pilus formation and protein
617
secretion by the same machinery in Escherichia coli. EMBO J 19: 2221-2228.
618
Schmidt, S.A., Bieber, D., Ramer, S.W., Hwang, J., Wu, C.Y., and Schoolnik, G. (2001)
619
Structure-function analysis of BfpB, a secretin-like protein encoded by the bundle-forming-
620
pilus operon of enteropathogenic Escherichia coli. J Bacteriol 183: 4848-4859.
621
Soto, G.E., and Hultgren, S.J. (1999) Bacterial adhesins: common themes and variations in
622
architecture and assembly. J Bacteriol 181: 1059-1071.
623
Strom, M.S., and Lory, S. (1987) Mapping of export signals of Pseudomonas aeruginosa pilin
624
with alkaline phosphatase fusions. J Bacteriol 169: 3181-3188.
625
Strom, M.S., and Lory, S. (1991) Amino acid substitutions in pilin of Pseudomonas
626
aeruginosa. Effect on leader peptide cleavage, amino-terminal methylation, and pilus
627
assembly. J Biol Chem 266: 1656-1664.
628
Turner, L.R., Lara, J.C., Nunn, D.N., and Lory, S. (1993) Mutations in the consensus ATP-
629
binding sites of XcpR and PilB eliminate extracellular protein secretion and pilus biogenesis
630
in Pseudomonas aeruginosa. J Bacteriol 175: 4962-4969.
631
Voulhoux, R., Bos, M.P., Geurtsen, J., Mols, M., and Tommassen, J. (2003) Role of a highly
632
conserved bacterial protein in outer membrane protein assembly. Science 299: 262-265.
633
Winther-Larsen, H.C., Wolfgang, M., Dunham, S., van Putten, J.P., Dorward, D., Lovold, C.
634
et al. (2005) A conserved set of pilin-like molecules controls type IV pilus dynamics and
635
organelle-associated functions in Neisseria gonorrhoeae. Mol Microbiol 56: 903-917.
28
636
Wolfgang, M., Lauer, P., Park, H.-S., Brossay, L., Hébert, J., and Koomey, M. (1998a) PilT
637
mutations lead to simultaneous defects in competence for natural transformation and
638
twitching motility in piliated Neisseria gonorrhoeae. Mol Microbiol 29: 321-330.
639
Wolfgang, M., Park, H.-S., Hayes, S.F., van Putten, J.P.M., and Koomey, M. (1998b)
640
Suppression of an absolute defect in type IV pilus biogenesis by loss-of-function mutations in
641
pilT, a twitching motility gene in Neisseria gonorrhoeae. Proc Natl Acad Sci USA 95: 14973-
642
14978.
643
Wolfgang, M., van Putten, J.P.M., Hayes, S.F., Dorward, D., and Koomey, M. (2000)
644
Components and dynamics of fiber formation define a ubiquitous biogenesis pathway for
645
bacterial pili. EMBO J 19: 6408-6418.
646
Yoshida, T., Kim, S.R., and Komano, T. (1999) Twelve pil genes are required for biogenesis
647
of the R64 thin pilus. J Bacteriol 181: 2038-2043.
648
29
649
Legends to table and figures
650
651
Table 1. Relevant characteristics of the corresponding gene products in the fifteen non-
652
piliated N. meningitidis pil mutants.
653
654
Fig. 1. Systematic immunofluorescence microscopy detection of Tfp in the N. meningitidis pil
655
mutants, which harbour mutations in genes specifically required for piliation, after the
656
introduction of an additional mutation in the pilT gene in order to abolish Tfp retraction (bars
657
represent 10 µm). The WT strain was included as a control. Bacteria (red) were stained with
658
ethidium bromide. Tfp (green) were detected using first the 20D9 anti-Tfp mouse monoclonal
659
antibody and then a goat anti-mouse antibody labelled with Alexa Fluor 488.
660
661
Fig. 2. Immunofluorescence microscopy detection of Tfp in representative N. meningitidis
662
strains before and after osmotic shock release of their periplasmic content (bars represent 10
663
µm). As for the pilP/T mutant, no pili could be detected after osmostic shock for the pilD/T,
664
pilE/T, pilF/T, pilM/T, pilN/T and pilO/T mutants (data not shown)
665
666
Fig. 3. Western blot detection of the pilin (PilE) in whole-cell extracts of the non-piliated
667
mutants in which fibers could not be restored in the absence of PilT. The WT strain was
668
included as a control. The protein concentrations in the extracts were quantified and adjusted,
669
which was verified by detecting the RmpM protein as a control (data not shown). Equal
670
amounts of proteins were present in each lane except for the pilD lane in which 5-fold more
671
protein was loaded in order to detect the prepilin readily.
672
673
Fig. 4. Western blot detection of the PilQ monomers and oligomers in whole-cell extracts of
30
674
each of the non-piliated mutants. The WT strain was included as a control. The protein
675
concentrations in the extracts were quantified and equal amounts of proteins were loaded in
676
each lane, which was verified by detecting the RmpM protein as a control (data not shown).
677
678
Fig. 5. Transmission electron microscopy analysis of Tfp in the double (triple) mutants in
679
which the fibers were restored. The WT strain was included as a control. Fibers were
680
visualized directly after negative staining with phosphotungstic acid (50,000 x magnification).
681
682
Fig. 6. Quantification of the adhesive abilities to human cells in the double (triple) mutants in
683
which the fibers were restored. The WT strain, a non-piliated pilD mutant and a hyper-piliated
684
pilT mutant were included as controls. After a 30-min contact during which bacteria adhered
685
to the cells, the cells were incubated with regular washes every hour and the numbers of
686
adherent bacteria were recovered by scrapping the wells at defined timepoints and counted.
687
A. Numbers of bacteria adhering to a monolayer of HUVEC after 4 hours of infection. Values
688
are the means ± standard deviations of 3-9 independent adhesion assays.
689
B. Kinetics of the adherence to HUVEC of the pilG/T and pilH/T mutants in which adhesion
690
was restored. The 0-h points in these graphs represent the sizes of the inocula that were
691
adjusted and counted prior the assay. Adherent bacteria were counted after the initial 30-min
692
contact and after 2 and 4 hours of incubation. Values are the means ± standard deviations of 3
693
independent assays.
694
695
Fig. 7. Quantification of the aggregative abilities in the double (triple) mutants in which the
696
fibers were restored. The WT strain, a non-piliated pilD mutant and a hyper-piliated pilT
697
mutant were included as controls. In brief, aggregation was quantified by measuring the
698
decrease in OD600 that occurs upon sedimentation of bacterial aggregates in static liquid
31
699
cultures. The aggregation of the non adhesive pilC1/C2/T, pilI/T, pilJ/T, pilK/T and pilW/T
700
mutants, were nil and indistinguishable from that of a non-piliated pilD mutant and we
701
therefore display only one curve for the sake of legibility. Values are the means of 3-9
702
independent experiments.
703
704
Fig. 8. Implication of the N. meningitidis Pil components, as inferred from this study, in the
705
four steps of Tfp biogenesis: fiber assembly, functional maturation (PilG and PilH are not
706
represented here because they only play a minor role in this step), counteraction of fiber
707
retraction powered by the PilT protein and emergence of the fibers on the cell surface. Note
708
that the order of the steps after Tfp assembly remains speculative.
32