The EMBO Journal Review Process File - EMBO-2009-71593
Manuscript EMBO-2009-71593
MOLECULAR ARCHITECTURE OF STREPTOCOCCUS
PNEUMONIAE TIGR4 PILI
Markus Hilleringmann, Philippe Ringler, Shirley A. Müller, Gabriella De Angelis, Rino Rappuoli,
Ilaria Ferlenghi, Andreas Engel
Corresponding author: Andreas Engel, Basel University
Review timeline:
Submission date:
1st Editorial Decision:
1st Revision received:
2nd Editorial Decision:
2nd Revision received:
Accepted:
12 June 2009
02 July 2009
29 September 2009
13 October 2009
29 October 2009
06 November 2009
Transaction Report:
(Note: With the exception of the correction of typographical or spelling errors that could be a source of ambiguity,
letters and reports are not edited. The original formatting of letters and referee reports may not be reflected in this
compilation.)
1st Editorial Decision
02 July 2009
Thank you for submitting your manuscript for consideration by The EMBO Journal. It has now been
seen by three referees whose comments to the authors are shown below.
As you will see while referee 3 expresses more reservations the other two referees are very positive
and would support publication of the paper here. Referee 1 still feels strongly that the effect of the
rrgA and rrgC mutants on pilus structure and anchoring to the cell wall should be analysed in more
depth (see his/her point 3). Also, it becomes clear from referee 1 and 3 that the manuscript text
would benefit greatly from some major rewriting to focus the manuscript more to the state of the
field before your novel structural data and to how your data take the field forward. I can recognise
that this message is certainly there already now, but it could certainly be put forward in a clearer
manner. All in all and in the light of the positive referees' comments I would like to invite you to
revise your manuscript along the lines suggested by the referees and as detailed above. I should
remind you that it is EMBO Journal policy to allow a single round of revision only and that,
therefore, acceptance of the manuscript will depend on the completeness of your responses included
in the next, final version of the manuscript.
Thank you for the opportunity to consider your work for publication. I look forward to your
revision.
Yours sincerely,
Editor
The EMBO Journal
© European Molecular Biology Organization
1
The EMBO Journal Review Process File - EMBO-2009-71593
_____
REFEREE REPORTS:
Referee #1 (Remarks to the Author):
The manuscript by Hilleringmann and colleagues examines the structure of Streptococcus
pneumonia pili (rlrA pili) in strain TIGR4, a research topic that has recently seen a flurry of
publications, but also confusing and even conflicting results. A chromosomal locus encodes three
sortases (SrtBCD), as well as three pilin subunits (RrgABC). In agreement with the sortase-pilus
hypothesis of Ton-That and Schneewind, RrgB, the protein with the YPK pilin motif, polymerizes to
form the shaft of TIGR4 pili. Earlier work claimed that the minor subunits, RrgA and RrgC, are
incorporated along the length of RrgB pili and that multiple sortases fulfill redundant roles. These
reports arrived at complicated hypotheses that largely ignored a second prediction in the sortase
field - the transpeptidases are thought to accept a single type of nucleophile to resolve their acyl
enzyme intermediates with cleaved protein substrates. A third confusion are the phenotypes of rrgA
and rrgC mutants on pilus incorporation into the cell wall envelope.
By isolating pili from streptococcal preparations and viewing these structures by transmission
electron microscopy in the presence or absence of specific antibodies, Hilleringmann and colleagues
now show that RlrA pili assume a simple and easily comprehensible structure: a polymer of sortaselinked RrgB harbors the RrgA adhesin at its tip and the RrgC protein at its base. In addition to pilus
dimensions, one learns here of a beaded domain structure for all three pilins and of a "protruding
nose" for RrgB. The nose provides polarity and orientation to the fiber that will greatly facilitate
future studies on this topic. On balance, these are beautiful and clarifying observations that will
advance the field of the molecular biology of Gram-positive pili, in particular for pneumococcal pili.
Major comments
1. The text of this manuscript is far too long and its general write-up too complicated and bumpy to
match the beautiful electron microscopy. This paper is not about pathogenesis and vaccines and all
sections that deal with these issues can be trimmed down to a minimum. Further, the paper does not
synthesize the frontier of pneumococcal pili as is laid out above. Yes, there is a large collection of
data but only some of these observations can be correct, which means that others must be incorrect.
These views get completely lost in the paper. Finally, the well described and rather simple
architecture of pili must be reflected in a similarly simple mechanism of assembly. This is also not
discussed in the manuscript and the authors should look at available data for rrgABC and srtBCD
mutants for an interpretation.
2. The authors use no particular strategy to release pili from the envelope of pneumococci. Thus, if
RrgC sits at the base of the pilus, how or why did it come off? What happens when pneumococci are
treated with murein hydrolases? Obviously the authors cannot believe that RrgC tethered to RrgB
pili just sits in the membrane and eventually wiggles away?
3. If rrgC is mutated, is the rrgC protein absent from the pilus during TEM? Likewise, if rrgA is
mutated, are the tips of pili unoccupied by minor pilin proteins? The weakest part of this manuscript
is Figure S10, where essential information is buried for rrgC, rrgA, and rrgAC mutants. The authors
want to report that the variants lacking rrgC release more pili into the culture supernatant that wildtype or rrgA variants. However, the data provided lack all quantification of released and cell
associated proteins and the immunoblot signals are not quantified for intensity. At the very least the
samples should all be loaded on the same blot and the cells derived from cultures with equal
denisities! A little bit more work on this topic would be a dramatic improvement for this manuscript
and could elevate it to a standard that other pneumococcal pili papers have not achieved.
Minor comments
1. Page 5, second paragraph. This isn't the first description of a simple pilus in Gram-positive
bacteria; it is the first description of a simple pilus structure in pneumococci.
2. The authors ignored actinomyces pili.
Referee #2 (Remarks to the Author):
© European Molecular Biology Organization
2
The EMBO Journal Review Process File - EMBO-2009-71593
This manuscript describes the architecture of a Streptococcus pneumoniae pilus assembly, formed
by the major pilin RrgB and two minor pilins RrgA and RrgC. There have been numerous previous
attempts to clearly visualise Gram-positive bacterial pili of this type, and to locate their components,
but they have been frustrated by their extreme thinness. Hilleringmann et al have obtained by far the
clearest and most exciting view of such pili using EM techniques with carefully designed
experiments.
The previously contentious issue on the location of the minor pilins has been finally resolved by
labelling the pili with antibodies strictly specific for each type of pilins. These experiments show
clearly that RrgB forms the shaft, with the two minor pilins located at each end of a pilus. They also
revealed the modular architecture of RrgA monomer and the distinct protrusion ('nose') in the RrgB
monomer; a very exciting step forward. The 'noses' allowed them to delineate individual RrgB
molecules from the pilus polymer and hence infer how the monomers are arranged relative to each
other in the pili.
The paper is well written, and will be a very authoritative one for this highly topical field. My only
reservations concern the arguments around pilus polarity. Using the 'noses' of RrgB to determine
pilus polarity is plausible but not fully convincing because of the limited resolution. The authors
might perhaps obtain more direct evidence by, for example, differentially labelling a single pilus
polymer with both RrgA and RrgC antibodies. Also, the argument (p.11) that RrgA and RrgC
cannot form higher order polymers because they lack a recognisable pilin motif lysine is not strictly
valid. The canonical 'pilin motif' is not always present even in some major pilins (eg. Spy0128). And
for RrgC to be the basal pilin, it would still require a lysine equivalent to that of the pilin motif.
One minor point: readers might be left wondering about the v-shaped assemblies of pili when
labelled with anti-RrgA antibodies. I understand that this is due to the fact that RrgA is at the end of
the pilus, and antibodies are bivalent, but this could be spelled out more clearly.
Referee #3 (Remarks to the Author):
The manuscript by Hilleringmann et al. describes the structural arrangement of the pilus of the
human pathogen Streptococcus pneumoniae in light of detailed TEM/STEM experiments. The
manuscript is well written and the work is timely, and the detailed visualization of the RrgB pilus,
including information on RrgA and RrgC localization, is interesting. However, some of the work
presented in the manuscript is not totally novel (and is presented as such), so in this reviewer's
opinion the paper contains a few pitfalls that preclude it from being publishable in EMBO J. in its
present form.
Major comments
p.4: The phrase '... because pneumococcal RrgB possesses the conserved motifs necessary for pilus
formation ... it has been proposed to form the backbone of the pneumococcal TIGR4 pilus.' This is
misleading; it is not only because it carries the (potentially) required motifs, but a variety of
laboratories, including the Henriques-Normark and Camilli groups, have shown that RrgB forms the
pilus backbone, and this should be clearly stated.
p. 5, first paragraph in results section: here authors describe their initial observations of pili,
including a statement regarding flexibility and measurements; similar results, however, were already
published for the isolated pili (by Dr. Hilleringmann and colleagues in PloS Pathogens). Authors
also mention that pili are intertwined to form tangles; the PloS work also discusses pilus intersection
and goes on to provide measurements on these forms.
p.6, middle paragraph: authors mention that ' ... a beaded structure was revealed ... making the pilus
shaft look like a string of pearls'. Manzano et al (2008) also showed by negative staining EM that
RrgB fibers produced by SrtC-1 activity appeared as 'beads on a string'. This should be mentioned
here.
p.6, same paragraph: analysis of pneumococcal pili by Western blotting has been published multiple
times by different laboratories and does not have to be included as a major figure in this paper.
© European Molecular Biology Organization
3
The EMBO Journal Review Process File - EMBO-2009-71593
Again, the phrase '... accordingly, the pilus filament must primarily be formed by RrgB ...' is
misleading, since this is information that has already been described before. Thus, this section of the
work seems to be mostly confirmatory of that of other groups, and not completely novel.
p.9, end of page: '... the HMW fractions observed by SDS PAGE confirmed that the RrgB
monomers are covalently linked.' : once again, this has been shown before by other laboratories.
p.19, figure 2: how are the relative specificities of the different antibodies for RrgA, RgB, and RrgC
accounted for?
Minor comments
p. 4, line 10: most authors chose to call the three sortases in the rlrA operon srtC-1, srtC-2, and srtC3, and not by their original names (srtB, srtC, srtD). The nomenclature should be updated here.
p.5: please add a reference to illustrate the phrase which states that pneumococci are generally found
in pars or short chains
1st Revision - authors' response
29 September 2009
We have accomplished the relevant experiments required by the referees and amended the
manuscript MOLECULAR ARCHITECTURE OF STREPTOCOCCUS PNEUMONIAE TIGR4
PILI, which we now would like to resubmit for publication in the EMBO Journal. In spite of the
additional text resulting from the requests of the three referees, the length of the manuscript has
remained practically the same. Below we answer the referee’s comments and indicate the changes
made to the manuscript.
Referee #1 (Remarks to the Author):
The manuscript by Hilleringmann and colleagues examines the structure of Streptococcus
pneumonia pili (rlrA pili) in strain TIGR4, a research topic that has recently seen a flurry of
publications, but also confusing and even conflicting results. A chromosomal locus encodes three
sortases (SrtBCD), as well as three pilin subunits (RrgABC). In agreement with the sortase-pilus
hypothesis of Ton-That and Schneewind, RrgB, the protein with the YPK pilin motif, polymerizes to
form the shaft of TIGR4 pili. Earlier work claimed that the minor subunits, RrgA and RrgC, are
incorporated along the length of RrgB pili and that multiple sortases fulfill redundant roles. These
reports arrived at complicated hypotheses that largely ignored a second prediction in the sortase
field - the transpeptidases are thought to accept a single type of nucleophile to resolve their acyl
enzyme intermediates with cleaved protein substrates. A third confusion are the phenotypes of rrgA
and rrgC mutants on pilus incorporation into the cell wall envelope.
By isolating pili from streptococcal preparations and viewing these structures by transmission
electron microscopy in the presence or absence of specific antibodies, Hilleringmann and
colleagues now show that RlrA pili assume a simple and easily comprehensible structure: a polymer
of sortase-linked RrgB harbors the RrgA adhesin at its tip and the RrgC protein at its base. In
addition to pilus dimensions, one learns here of a beaded domain structure for all three pilins and of
a "protruding nose" for RrgB. The nose provides polarity and orientation to the fiber that will
greatly facilitate future studies on this topic. On balance, these are beautiful and clarifying
observations that will advance the field of the molecular biology of Gram-positive pili, in particular
for pneumococcal pili.
Major comments
1. The text of this manuscript is far too long and its general write-up too complicated and bumpy to
match the beautiful electron microscopy. This paper is not about pathogenesis and vaccines and all
sections that deal with these issues can be trimmed down to a minimum. Further, the paper does not
synthesize the frontier of pneumococcal pili as is laid out above. Yes, there is a large collection of
data but only some of these observations can be correct, which means that others must be incorrect.
These views get completely lost in the paper. Finally, the well described and rather simple
© European Molecular Biology Organization
4
The EMBO Journal Review Process File - EMBO-2009-71593
architecture of pili must be reflected in a similarly simple mechanism of assembly. This is also not
discussed in the manuscript and the authors should look at available data for RrgABC and srtBCD
mutants for an interpretation.
The previous manuscript text has been shortened, streamlined and simplified. In particular, text
relating to pathogenesis and vaccines has been kept to the minimum. As suggested by this referee,
major progress in understanding the action of sortases is now properly summarized. Further,
contradictions in previously published results concerning the pilus architecture are now more clearly
pointed out. With this background, the progress provided by the EM analysis presented here is more
obvious, and is now summarized in a model.
2. The authors use no particular strategy to release pili from the envelope of pneumococci. Thus, if
RrgC sits at the base of the pilus, how or why did it come off? What happens when pneumococci are
treated with murein hydrolases? Obviously the authors cannot believe that RrgC tethered to RrgB
pili just sits in the membrane and eventually wiggles away?
In the procedure employed to isolate native pili, and pili of the ΔrrgA , ΔrrgC and ΔrrgAC mutants
for electron microscopy and the other analysis performed, the bacteria were treated with
mutanolysin, a murein-hydrolysing enzyme (Sigma M9901: N-Acetyl muramidase) to "liberate" the
covalently peptidogylcan- anchored pili into the supernatant. This fact is stated in the earlier paper
referred in the Materials and Methods. A sentence has now been included in both the Results and
Materials and Methods sections to make the reader fully aware of the method employed.
In contrast, no murein-hydrolysing enzyme was used in experiments studying the potential role of
ancillary protein RrgC (leading to figure S12): Rather the amount of high molecular weight pilus
material released from the wt and respective mutant bacteria "by themselves" into the "media", i.e.
the supernatant, was monitored. This is also now described in more detail in the Supplementary
Material (Supplementary Figure S12).
3. If rrgC is mutated, is the RrgC protein absent from the pilus during TEM? Likewise, if rrgA is
mutated, are the tips of pili unoccupied by minor pilin proteins? The weakest part of this manuscript
is Figure S10, where essential information is buried for rrgC, rrgA, and rrgAC mutants. The authors
want to report that the variants lacking rrgC release more pili into the culture supernatant that
wild-type or rrgA variants. However, the data provided lack all quantification of released and cell
associated proteins and the immunoblot signals are not quantified for intensity. At the very least the
samples should all be loaded on the same blot and the cells derived from cultures with equal
denisities! A little bit more work on this topic would be a dramatic improvement for this manuscript
and could elevate it to a standard that other pneumococcal pili papers have not achieved.
In control immuno-EM experiments made with ΔrrgA pili the tips of the pili were not labelled by
anti-RrgA-His antibodies.
In response to the referee’s request for proof that the RrgC and RrgA proteins were indeed not
produced by the respective ΔrrgA , ΔrrgC and ΔrrgAC mutants we have made a Western blot
analysis and include this data in the Supplementary Materials to document the absence of the
respective proteins (Supplementary Figure S6). Two things were done, (i) bacterial whole cell
lysates were examined by Western blots (SDS-PAGE) (Supplementary Figure S6A), and (ii)
isolated pili were collected and examined by Western blots (Dot Blot analysis) (Supplementary
Figure S6B) as well as by electron microscopy (EM) (Supplementary Figure S7). The blots clearly
show that RrgA and RrgC were completely absent in the respective mutants. Further, the EM
demonstrated that, as also shown for the ΔrrgA mutant (Supplementary Figure S7A), that both the
ΔrrgC and ΔrrgAC mutants were able to produce long pili with shafts identical to those of the wt.
Images of these pili are now also included in the Supplementary Material (Supplementary Figures
S7B, C).
To further improve our data several experiments were repeated and modified as follows:
1) Like before bacteria were grown in medium to the same optical density. Identical volumes of the
respective (growth medium) supernatants were then used to analyze the amount of HMW pilus
© European Molecular Biology Organization
5
The EMBO Journal Review Process File - EMBO-2009-71593
material released into the growth medium to study the potential role of proximal RrgC to attach pili
to the peptidoglycan layer (Supplementary Figure 12A).
2) Samples were all loaded on the same gel.
3) In addition to the standard SDS-PAGE Western blot analysis, the same material was analysed by
a dot blotting procedure, spotting different dilutions of the various supernatants and recombinant
RrgB-His (as a standard) onto a nitrocellulose membrane (Supplementary Figure S12B). This
allowed the quantification of the released pilus material to be improved; the densiometric signal of
the spots was determined and compared (Supplementary Figure S12C).
Minor comments
1. Page 5, second paragraph. This isn't the first description of a simple pilus in Gram-positive
bacteria; it is the first description of a simple pilus structure in pneumococci.
We have incorporated the reference Mandlik et al (2008) The molecular switch that activates the cell
wall anchoring step of pilus assembly in gram-positive bacteria. Proc Natl Acad Sci U S A
105(37):14147-52’ more prominently. There a simple model for the Corynebacterium diphtheriae
pilus is given, yet it is based mainly on biochemical and mutational analyses, not on the direct
visualization of the pilus structure.
2. The authors ignored actinomyces pili.
We have now included this work as requested. References added:
Actinomyces naeslundii: Mishra A, Das A, Cisar John O, Ton-That H (2007) Sortase-catalyzed
assembly of distinct heteromeric fimbriae in Actinomyces naeslundii. J Bact 189(8): 3156-3165.
General reference; Wu H, Fives-Taylor PM (2001) Molecular strategies for fimbrial expression and
assembly. Crit Rev Oral Biol Med 12(2): 101-115.
Referee #2 (Remarks to the Author):
This manuscript describes the architecture of a Streptococcus pneumoniae pilus assembly, formed
by the major pilin RrgB and two minor pilins RrgA and RrgC. There have been numerous previous
attempts to clearly visualise Gram-positive bacterial pili of this type, and to locate their
components, but they have been frustrated by their extreme thinness. Hilleringmann et al have
obtained by far the clearest and most exciting view of such pili using EM techniques with carefully
designed experiments.
The previously contentious issue on the location of the minor pilins has been finally resolved by
labelling the pili with antibodies strictly specific for each type of pilins. These experiments show
clearly that RrgB forms the shaft, with the two minor pilins located at each end of a pilus. They also
revealed the modular architecture of RrgA monomer and the distinct protrusion ('nose') in the RrgB
monomer; a very exciting step forward. The 'noses' allowed them to delineate individual RrgB
molecules from the pilus polymer and hence infer how the monomers are arranged relative to each
other in the pili.
The paper is well written, and will be a very authoritative one for this highly topical field. My only
reservations concern the arguments around pilus polarity. Using the 'noses' of RrgB to determine
pilus polarity is plausible but not fully convincing because of the limited resolution. The authors
might perhaps obtain more direct evidence by, for example, differentially labelling a single pilus
polymer with both RrgA and RrgC antibodies. Also, the argument (p.11) that RrgA and RrgC cannot
form higher order polymers because they lack a recognisable pilin motif lysine is not strictly valid.
The canonical 'pilin motif' is not always present even in some major pilins (eg. Spy0128). And for
RrgC to be the basal pilin, it would still require a lysine equivalent to that of the pilin motif.
The referee questions the negative stain TEM microscopy suggesting that the resolution attained is
not sufficient to reveal the nose-like structure. We now include new figures in the Supplementary
Material illustrating that the resolution is sufficient providing staining is optimum (Supplementary
Figure S11 for the wt and for the ΔrrgA, ΔrrgC and ΔrrgAC mutants, Supplementary Figure S7).
© European Molecular Biology Organization
6
The EMBO Journal Review Process File - EMBO-2009-71593
We now also show longer stretches of the pilus shafts in Figure 5 and include an additional figure in
the Supplementary Material (Supplementary Figure S11) to make the argument of pilus
directionality stronger. Since RrgA and RrgC antibodies cannot be differentiated, and because using
gold-labelled secondary antibodies would not provide the resolution required (see discussion), we
have not performed the differential labelling experiment proposed. Further as noted in the Material
& Methods section the antibody concentration must be optimised for these experiments (maximum
labelling, minimum free antibody concentration). The free antibody concentration would necessarily
be approximately double in the proposed experiment making interpretation of the data difficult.
We would like to thank the referee for the comment regarding canconical "pilin motifs" or
respective conserved lysine equivalents required for inter-molecular isopeptide bond formation. Our
argument that RrgA and RrgC cannot form higher order polymers is based on published results and
our own finding that no higher order RrgA, RrgC or RrgA-C polymers are observed in ∆rrgB
mutants (Figure S6A). The text has been changed. We refer to LeMieux et al (2008) J Bacteriol
190(17): 6002-6013. These authors showed that RrgA and RrgC form heterodimers of 170 kDa
when RrgB is absent, whereas no RrgA multimers are found in the rrgB rrgC double mutant. They
also suggest that the observed clustering of RrgA along pili shafts is due to RrgA proteins located at
the tips of different-length pili bundled together. We have reformulated this paragraph in the
discussion.
One minor point: readers might be left wondering about the v-shaped assemblies of pili when
labelled with anti-RrgA antibodies. I understand that this is due to the fact that RrgA is at the end of
the pilus, and antibodies are bivalent, but this could be spelled out more clearly.
This is indeed the case and the required details are now given.
Referee #3 (Remarks to the Author):
The manuscript by Hilleringmann et al. describes the structural arrangement of the pilus of the
human pathogen Streptococcus pneumoniae in light of detailed TEM/STEM experiments. The
manuscript is well written and the work is timely, and the detailed visualization of the RrgB pilus,
including information on RrgA and RrgC localization, is interesting. However, some of the work
presented in the manuscript is not totally novel (and is presented as such), so in this reviewer's
opinion the paper contains a few pitfalls that preclude it from being publishable in EMBO J. in its
present form.
We have thoroughly revised our manuscript and now clearly differentiate between what is novel in
the work, and what was done to characterise our preparations. We also now emphasize contradictory
results present in the literature to make it clear why certain, slightly modified, experiments had to be
performed even though they were not entirely novel. The negative stain electron microscopy
presented confirms some and contradicts other published data, and proceeds to reveal pilus structure
with unprecedented detail. Given the contradictions in the field and the widespread use of the
classical lower resolution immuno-EM it was important to first demonstrate the integrity of our
samples (prepared using a new modified version of published methods) and the relationship of our
images to published data, i.e., to start by showing that the pili appear as beads on a string and to
proceed from there. Importantly, the mass-per-length measurement carried out by STEM is the
first quantitative measurement proving that these ‘beads’ are indeed single RrgB monomers, a fact
assumed to date from the width of the pili. Our results provide the first direct electron microscopical
evidence for a simple Gram-positive pilus architecture, and resolve some of the open questions
concerning the location and function of RrgA and RrgC; the pilus shaft is shown to be formed
exclusively from a string of head-to-tail orientated RrgB subunits. Direct visualisation of antibody
binding, rather than standard immuno-EM, and TEM imaging of recombinant pilus protein
monomers allowed this conclusion to be drawn. We have taken care to indicate previous findings
when repeat experiments were made. We have now also written a concluding paragraph and
added a model that reflects the simple pilus architecture of TIGR4 indicated by our data.
Major comments
p.4: The phrase '... because pneumococcal RrgB possesses the conserved motifs necessary for pilus
formation ... it has been proposed to form the backbone of the pneumococcal TIGR4 pilus.' This is
© European Molecular Biology Organization
7
The EMBO Journal Review Process File - EMBO-2009-71593
misleading; it is not only because it carries the (potentially) required motifs, but a variety of
laboratories, including the Henriques-Normark and Camilli groups, have shown that RrgB forms
the pilus backbone, and this should be clearly stated.
This section of the introduction has been rewritten reporting these earlier findings and cites the
required references. We wish to mention that although it was "known" that RrgB forms pilus
backbone before the start of our work, the STEM data reported provide the first quantitative
evidence that pili are formed by a single string of RrgB monomers. Similarly, the negative stain EM
provides the first experimental evidence that the RrgB monomers are located head-to-tail one after
the other and the negative stain EM and antibody labelling dispels proposals in the literature that this
string is "interrupted" by RrgA or RrgC.
p. 5, first paragraph in results section: here authors describe their initial observations of pili,
including a statement regarding flexibility and measurements; similar results, however, were
already published for the isolated pili (by Dr. Hilleringmann and colleagues in PloS Pathogens).
Authors also mention that pili are intertwined to form tangles; the PloS work also discusses pilus
intersection and goes on to provide measurements on these forms.
In the present study we used a modified purification procedure to get native pili i.e., a different
buffer system to that used by Hilleringmann et al., PLoS 2008 (specified in M&M). Consequently
our first aim was to show that the native pili isolated by this new method still show flexibility and
that the three components RrgA, RrgB and RrgC are all present (Figure 2). Indeed, initially pili were
also isolated as reported in the PloS paper and compared by negative stain EM. Results from these
control experiments would be a repetition and are not shown. Imaging native pili isolated in this new
way by negative stain TEM and STEM allowed us to achieve much higher resolution than in the
initial PLoS work by Hilleringmann et al. (2008). In addition, STEM of unstained
samples yielded quantitative mass-per-length data, which allowed the number of RrgB monomers
per unit length to be calculated.
We explain what has been somewhat misleading in the PloS paper; no helical super pili are
systematically formed from single ones. Classic immuno-EM and cryo-EM were used in the PloS
work. The former technique readily visualised pilus bundles and tangles but did not allow single
6nm pilus filaments to be distinguished. This and similar results from other workers in the field has
lead to confusion. The smallest pilus diameter reported was 9.5nm, possibly corresponding to 2
intertwined pili. The resolution of our negative stain EM study clarifies what happens, clearly
showing the random bundling and tangling of pili attached to the bacteria. Already from
these images it becomes clear that helical super pili are not formed. The low contrast of cryo-EM
also hindered the visualisation of single pili at the time of the PloS paper; pili with minimum and
maximum diameters of 6.8nm and 9.5nm, were interpreted as helical super-pili formed by 2 pilus
filaments. The results presented here show that helical super-pili may exist at random in tangles and
bundles through intertwining, but do not represent the basic structure of TIGR4 pili as proposed in
the PloS paper.
p.6, middle paragraph: authors mention that ' ... a beaded structure was revealed ... making the
pilus shaft look like a string of pearls'. Manzano et al (2008) also showed by negative staining EM
that RrgB fibers produced by SrtC-1 activity appeared as 'beads on a string'. This should be
mentioned here.
As suggested by this referee, we have added Manzano et al (2008) in this context.
p.6, same paragraph: analysis of pneumococcal pili by Western blotting has been published multiple
times by different laboratories and does not have to be included as a major figure in this paper.
Again, the phrase '... accordingly, the pilus filament must primarily be formed by RrgB ...' is
misleading, since this is information that has already been described before. Thus, this section of the
work seems to be mostly confirmatory of that of other groups, and not completely novel.
Our aim was to confirm that the isolated pili analysed were still intact and covalently linked. We
feel that this is of particular importance as a new buffer system and slightly modified protocol was
used for their isolation. In this sense the Western blot (SDS-PAGE) shown in Figure 2B is not a
© European Molecular Biology Organization
8
The EMBO Journal Review Process File - EMBO-2009-71593
repetition and we would like to keep this figure; the information it provides is key to the paper.
p.9, end of page: '... the HMW fractions observed by SDS PAGE confirmed that the RrgB monomers
are covalently linked.' : once again, this has been shown before by other laboratories.
We meant to imply this by the word ‘confirmed’ but unfortunately forgot to include the reference.
This omission has now been corrected.
p.19, figure 2: how are the relative specificities of the different antibodies for RrgA, RrgB, and RrgC
accounted for?
The purification procedure described by Mueller C.A. et al (Mueller CA, Broz P, M¸ller SA, Ringler
P, Erne-Brand F, Sorg I, Kuhn M, Engel A, Cornelis GR (2005) The V-antigen of Yersinia forms a
distinct structure at the tip of injectisome needles. Science 310(5748): 674-676), was used to obtain
highly specific polyclonal antibodies against the respective recombinant pilus proteins i.e.,
antibodies showing no cross-reactivity towards other Rrg’s. Their specificity was tested and is now
documented by a new figure (Supplementary Figure S2). To make this clearer a sentence to this
effect as well as the pertinent reference has been included in the Results section and in the legend of
Supplementary Figure S2.
Minor comments
p. 4, line 10: most authors chose to call the three sortases in the rlrA operon srtC-1, srtC-2, and
srtC-3, and not by their original names (srtB, srtC, srtD). The nomenclature should be updated
here.
The nomenclature has been changed as suggested.
p.5: please add a reference to illustrate the phrase which states that pneumococci are generally
found in pairs or short chains
We have added a reference that describes pneumococci string and biofilm formation.
We hope that we have duly responded to the referee’s comments and that our work will be accepted
for publication in the EMBO Journal.
2nd Editorial Decision
13 October 2009
Thank you for sending us your revised manuscript. Our original referees 1 and 3 have now seen it
again, and you will be pleased to learn that in their view you have addressed their criticisms in a
satisfactory manner, and that the paper will therefore be publishable in The EMBO Journal.
Before this will happen, however, I was wondering whether you would like to consider addressing
the minor issues suggested by referee 1 (see below). Please let us have a suitably amended
manuscript as soon as possible. I will then formally accept the manuscript.
Yours sincerely,
Editor
The EMBO Journal
_____
REFEREE REPORTS:
Referee #1 (Remarks to the Author):
© European Molecular Biology Organization
9
The EMBO Journal Review Process File - EMBO-2009-71593
Overall:
This is a great article, very well described, the conclusions are justified. The descriptive results
presented here are significantly contributing to the understanding of the structure of Gram-positive
pili and are of significant novelty and amazing quality. "Accept for publication".
Minor points in the intro: rlrA pathogenicity islet should read rlrA pathogenicity islet;
- results: remove supplemental figure S1. The negative stain of Streptococcus pneumoniae TIGR4
cells without pili does not contribute to the article
- check whether all the µ are in times new roman;
- Explain the circular structures/background seen in Fig 1.
- 1st paragraph: "The aggregates resulting" should read "The resulting aggregates";
- 2nd paragraph: Change "In the transmission electron microscope they..." into "TIGR4 pili..."
- 2nd paragraph: a beaded structure is not convincing/clear from the inset in figure 2A;
- 2nd paragraph: is it really true that there is really NO crossreactivity. A small/thin band is visible
or is it overlay from the previous well?
- 2nd paragraph: write couple of sentences that the inset of 2D are supposed to be antibodies.
Materials and methods:
- TIGR4 pilus purification, second last sentence: "were applied: Harvested" should be "were
applied: harvested" with lowercase 'h'
- TEM, Eindhofen should be Eindhoven, as this is a Dutch city.
Figure captions:
- figure 2D: only one TEM image is shown, thus the sentence should start with "TEM image",
remove 's'. Furthermore, place the 'inset in D' text behind the text of 2d and not at the end of this
figure legend.
- Figure s6A ∆rrgB should be ∆rrgB
- S6+6b: dilutions should be 1:10,000 and not 1:10'000 and so on.
Figures:
1= ok, explain the circle structures of the background. Is this UAc negative stain?
2= ok, indicate the antibody cluster of 2D + 2E with an arrow. The inset of 2D could be anything.
3= ok, however, the STEM images are not that good/convincing. For 3B and 3C include the scale
bar as well.
4= ok, make the subunits of 4E in one color, as it are the same subunits.
5= indicate the antibody clusters with a white arrow.
6= ok.
S1= not ok, not necessary, remove this figure
S2= is there really no cross reactivity?
S3-5= ok
2nd Revision - authors' response
29 October 2009
We are pleased to hear that our manuscript (EMBOJ 2009 71593R) is now considered suitable for
publication in The EMBO Journal. We have taken the opportunity offered to us to respond to the
few further minor points made by Referee 1 and thank this referee for the constructive comments
and suggestions. The manuscript has been amended accordingly in all but one point, as detailed in
the following rebuttal.
Minor points in the intro:
rlrA pathogenicity islet should read rlrA pathogenicity islet
This has been corrected as suggested and also in the reference LeMieux J, Hava DL, Basset A,
Camilli A (2006) RrgA and RrgB are components of a multisubunit pilus encoded by the
© European Molecular Biology Organization
10
The EMBO Journal Review Process File - EMBO-2009-71593
Streptococcus pneumoniae rlrA pathogenicity islet. Infect Immun 74(4): 2453 2456.
Results:
Remove supplemental figure S1. The negative stain of Streptococcus pneumoniae TIGR4 cells
without pili does not contribute to the article
The figure has been removed; the figure numbers have been corrected throughout.
In view of this change the initial sentence of this section ‘S. pneumoniae TIGR4 bacteria were
frequently found as pairs or short chains in the samples examined, an arrangement found to be the
first step in biofilm formation (Allegrucci et al, 2006).’ is out of place and has been removed;
Allegrucci et al has been removed from the reference list. The first sentence of the results section
now reads As documented earlier, the surface of the S. pneumoniae TIGR4 bacteria examined was
covered ……’
Check whether all the µ are in times new roman
Unfortunately some of this sentence could not be read in the e mail. We assume it refers to the fonts
used throughout the manuscript. All text, including the symbols Δ
and
µ, is now written in the font
Times New Roman, both in the main manuscript and Supplementary On line Material.
Explain the circular structures/background seen in Fig 1.
Considering how the EM grids were prepared, the circular structures seen in Figure 1 are most likely
impurities from the blood agar plates on which the bacteria were grown. To make this clear, precise
details of the grid preparation method are now included in the Materials & Methods; the following
text has been added to the TEM section:
‘For TEM of the whole bacteria, 100 200 µl of PBS was added to the blood agar growth plate and
agitated gently to softly remove bacteria from the agar. The plate was tilted and an aliquot of the
resulting bacterial suspension was removed from close to the liquid surface. Small aliquots of this
stock suspension were then directly loaded onto carbon coated Parlodion microscopy grids. The
bacteria were allowed to settle (5 min) and then stabilized by the addition of 2% paraformaldehyde
(40s). Grids were washed on droplets of water, negatively stained and examined. As dictated by grid
quality, the stock was sometimes centrifuged gently for several minutes (3000 rpm for 5 - 10 min.),
the pellet gently re suspended in PBS and grids prepared; if necessary these steps were repeated.’
1st paragraph: "The aggregates resulting" should read "The resulting aggregates"
This has been corrected as suggested.
2nd paragraph: Change "In the transmission electron microscope they..." into "TIGR4 pili..."
The sentence now reads ‘In the transmission electron microscope TIGR4 pili appeared ….’
2nd paragraph: a beaded structure is not convincing/clear from the inset in figure 2A;
We have adjusted the contrast of this image (inset, Figure 2A) to enhance the structural details. As
the beading will probably still be quite difficult to discern at the enlargement possible in the printed
manuscript we have also rephrased the sentence which now reads ‘The structure revealed in
enlarged views (Figure 2A, inset) bears some similarity to that of RrgB filaments assembled in vitro
(Manzano et al, 2008).’
2nd paragraph: is it really true that there is really NO crossreactivity. A small/thin band is visible
or is it overlay from the previous well?
Lanes 1, 3 and 5 clearly document where the individual Rrg-His proteins appear on the Westerns.
The small thin band in lane 6 is at the molecular weight of RrgC-His, as RrgC-His was not loaded in
this lane it can only be an overlay from lane 5 as suggested by the referee. There is no signal at all in
lane 6 at the positions where RrgAHis and RrgB-His would be, so these proteins are not labeled by
anti-RrgC-His. Similarly the small signal in lane 4 is at the molecular weight of RrgB-His, but
© European Molecular Biology Organization
11
The EMBO Journal Review Process File - EMBO-2009-71593
RrgB-His was not loaded in this lane. Therefore the signal can only be contamination from lane 3,
which in this case probably occurred when the initial SDS PAGE was loaded. There is no signal at
all in lane 4 at the positions where RrgA-His and RrgC-His would be expected, so these proteins
were not labeled by anti RrgB-His.
A comment has been added to the caption of Supplementary Figure S1. The text reads ‘The faint
signals in lanes 4 and 6 are overlays from the preceding wells; they are not at the molecular weights
of the proteins loaded.’
2nd paragraph: write couple of sentences that the inset of 2D are supposed to be antibodies.
The sentence ‘Views of single antibodies are shown in the inset of Figure 2D for comparison.’ has
been included.
This section of the text now reads:
‘Anti-RrgB-His antibodies decorated the pilus shaft at irregular intervals and, having two binding
sites, often linked pili together forming ladder and net like assemblies depending on the degree of
lateral cross linking (Figure 2C; Supplementary Figure S3). Views of single antibodies are shown in
the inset of Figure 2D for comparison. In contrast to anti-RrgB-His antibodies, antibodies against
RrgA-His only bound at the end of pili, generally clustering and, as they are divalent, frequently
linking 2 pili together in typical v shaped assemblies not otherwise observed (Figure 2D;
Supplementary Figure S4)’.
Materials and methods:TIGR4 pilus purification, second last sentence: "were applied: Harvested"
should be "were applied: harvested" with lowercase 'h'
The requested change has been made.
TEM, Eindhofen should be Eindhoven, as this is a Dutch city.
The correction has been made.
Figure captions:
figure 2D: only one TEM image is shown, thus the sentence should start with "TEM image", remove
's'. Furthermore, place the 'inset in D' text behind the text of 2d and not at the end of this figure
legend.
The caption to Figure 2D is now written in the singular; ‘images’ has been corrected to ‘image’. The
caption to Figure 2E has been similarly corrected.
The number of antibody orientations shown in the inset of Figure 2D has been reduced. Instead
traces now indicate the contours of the antibody projections retained to guide the eye of readers not
familiar with such electron microscopy images. Accordingly, the text to this inset has been changed
to ‘Inset: Various orientations of individual antibodies and traces indicating their outer contours.’ It
is now positioned behind the text of Figure 2D as requested.
Figure s6A ∆rrgB should be ∆rrgB S6+6b: dilutions should be 1:10,000 and not
1:10'000 and so on.
Unfortunately some of this sentence could not be read in the e mail. We think that the font used for
the symbol ∆ is referred to as this was incorrect; ∆ is now written in Times New Roman.
Dilutions are now written as 1:10,000 and 1:30,000 in the captions to S5A and S5B (previously S6A
& S6B) as requested.
Figures:
1= ok, explain the circle structures of the background. Is this UAc negative stain?
The circular structures seen in Figure 1 are impurities from the blood agar plates on which the
bacteria were grown. Please see above. The negative stain employed was 2% phosphotungstic acid.
This is stated in the figure caption.
2= ok, indicate the antibody cluster of 2D + 2E with an arrow. The inset of 2D could be anything.
© European Molecular Biology Organization
12
The EMBO Journal Review Process File - EMBO-2009-71593
Arrows have been added to Figures 2D and 2E as requested. For clarity, arrows have also been
added to the corresponding Supplementary Figures S3 and S4. All figure captions have been
changed accordingly.
The number of antibody orientations shown in the inset of Figure 2D has been reduced. Instead
traces now indicate the contours of the antibody projections retained to guide the eye of readers not
familiar with such electron microscopy images. The figure caption to this section now reads, ‘Inset:
Various orientations of individual antibodies and traces indicating their outer contours.’
3= ok, however, the STEM images are not that good/convincing. For 3B and 3C include the scale
bar as well.
Average projections are shown for the TEM microscopy; the signal to noise ratio is high, the low
signal to noise ratio of the individual projections having been much increased by the averaging step.
Single shot images are shown for the STEM microscopy. Their signal to noise ratio is low compared
to that of the TEM averages, but high compared to that of the individual TEM projections. Although
they are not as clear as the TEM averages, we wish to retain the STEM images as they document the
ability of the STEM to visualize the various protein orientations without averaging steps,
information that is relevant to Figure 4. The images also document more strongly the various
projections found on the EM grid. Scale bars are now shown on 3B and 3C as well as on 3A of this
figure as requested.
4= ok, make the subunits of 4E in one color, as it are the same subunits.
The different subunit grey levels are important to illustrate that adjacent subunits overlap. Therefore,
we wish to retain the original color scheme. The point made by the reviewer is however taken, text
has been added to the figure caption, explaining that both colors refer to RrgB. This section of the
figure caption now reads:
‘E) Highly contoured TEM images of RrgB His monomers (average length 12.2 (±0.5) nm)
matched, without straightening, to the subunits of pilus (D) illustrating that its RrgB subunits
overlap; the RrgB monomers are shown in alternating shades of grey and their ends are marked by
dotted lines to facilitate visualization.’
5= indicate the antibody clusters with a white arrow.
The requested arrows have been added to this figure and to the corresponding Figure of the
Supplementary Material, Supplementary Figure S10. Both figure captions have been changed
accordingly.
6= ok.
S1= not ok, not necessary, remove this figure
This figure has been removed as requested.
S2= is there really no cross reactivity?
This Figure is Supplementary Figure S1 of the revised version. As detailed above, there is no cross
reactivity. The faint bands in lanes 4 and 6 are at the molecular weight of the protein loaded in the
preceding lane, i.e., the protein with which the antibody under test should interact. They result from
an overlay as suggested by the referee, please see above. There is no interaction between the
antibody under test and the proteins actually loaded in wells 4 and 6; there are no bands at the
corresponding molecular weights.
S3 5= ok
Additional changes
© European Molecular Biology Organization
13
The EMBO Journal Review Process File - EMBO-2009-71593
1. LeMieux J, Woody S, Camilli A (2008) Roles of the sortases of Streptococcus pneumoniae in
assembly of the RlrA pilus. J Bacteriol 190(17): 6002 6013 ‘RlrA’ is now written ‘rlrA’
2. The author list of the reference Tetterlin et al (2001) has been shortened to ‘20 author names et al’
in compliance with the Instructions to Authors of The EMBO Journal.
3. Minor correction to the SOM: Change from lower to upper case; ‘table’ is now written ‘Table’.
© European Molecular Biology Organization
14