Molecular Systems Biology Peer Review Process File
High-throughput CRISPRi phenotyping identifies new
essential genes in Streptococcus pneumoniae
Xue Liu, Clement Gallay, Morten Kjos, Arnau Domenech, Jelle Slager, Sebastiaan P van Kessel,
Kevin Knoops, Robin A Sorg, Jing-Ren Zhang, Jan-Willem Veening
Corresponding author: Jan-Willem Veening, University of Lausanne
Review timeline:
Submission date:
Editorial Decision:
Revision received:
Editorial Decision:
Revision received:
Accepted:
17 November 2016
14 December 2016
03 March 2017
03 April 2017
12 April 2017
13 April 2017
Editor: Maria Polychronidou
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
14 December 2016
Thank you again for submitting your work to Molecular Systems Biology. We have now heard back
from the three referees who agreed to evaluate your study. As you will see below, the reviewers
appreciate that the presented CRISPRi screen is a useful resource for the field. However, they raise
several concerns, which we would ask you to address in a revision.
I think that the reviewers' recommendations are rather clear so there is no need to repeat the points
listed below, but please let me know in case you would like to further discuss any specific point.
One of the more fundamental issues raised by the reviewers, refers to the need to include additional
controls.
-----------------------------------------------------REFEREE REPORTS
Reviewer #1:
The study by Liu and coauthors is an elegant example of the adaptation of CRISPRi technology to
the phenotypization of bacterial genes. This technology has previously been adapted to a few other
bacteria, and the scale or the current work certainly merits significant attention. The authors focus
their work on genes predicted to be essential by their transposon mutagenesis experiments,
designing a library of sgRNAs that spans ~15% of the protein coding genes. From careful
phenotypic analysis of several candidates the authors help elucidate genes involved in peptidoglycan
and teichoic acid synthesis, and natural competence. The latter experiments are particularly striking,
in part because of the quality of the negative controls provided by the non-participating subunits.
My enthusiasm for the study is mollified by the general lack of key controls present in the study. For
© European Molecular Biology Organization
1
Molecular Systems Biology Peer Review Process File
instance, an adequate number of targets predicted to be dispensable in their transposon experiment
could have served to assess the precision of their CRISPRi approach. Such controls could be easily
included and would significantly improve the study, helping to assess the reliability of the CRISPRi
method, and increasing the confidence of a few key observations. Provided with the adequate
controls, I have no doubt the work will have a tremendous impact on the study of Streptococcus
pneumonia.
MAJOR COMMENTS:
Fig. 2. Given that polar effects have been well documented when using CRISPRi in bacteria, the
authors should analyze their screening results in the context of the operon architecture. When
targeting the same operon, what is the agreement of different sgRNAs? Such an analysis would be
facilitated by including selected sgRNAs against operons that contain both essential and dispensable
genes. Are there any cases where the CRISPRi effect on a known dispensable operon, extends into
an essential operon? Quantifying and displaying the magnitude of such effects, for instance by
performing RNAseq experiments on a couple of other genes, would help determine the degree of
confidence that readers should have on the authors' data set. However, a simple analysis of the
existing data is probably sufficient to satisfy this point, since there are a lot of genes that did not
show a phenotype in their analysis.
Fig. 2C. Were any of the genes predicted to be dispensable further analyzed? Presumably some of
these were particularly implicated in essential functions by the transposon experiments, yet appeared
dispensable by CRISPRi. Understanding the source of the discrepancy, could help establish the
reliability of the methods and the ultimate relevance of the screening results.
Fig. 3. Measuring growth curved is a central part of the study's phenotypic assessment, and would
benefit from a set of genes pre-selected as controls. Is there a set of genes that are known to be
dispensable and could be used as controls? This is particularly important given the documented
polar effects of the CRISPRi system. Can sgRNAs against dispensable genes at the end of an operon
be used to repress such genes without affecting the rest of the operon?
Fig. 4. Depletion of the GFP-tagged proteins should be demonstrated, to confirm the effects of the
genetic manipulations. This could be ideally performed by immunoblotting against the GFP portion
of the fusion proteins, compared to a control. Alternatively the strains could be compared by FACS,
provided an adequate counterstain to gate on the bacteria.
Fig. 5E. I recognize that the monoclonal antibody against phosphorylcholine has previously been
used; however, to my knowledge, the small species found in the TarP/Q knockdowns have not been
documented. If I am wrong on this point, please reference such an observation appropriately.
Alternatively, the authors should include a mutant of an upstream member of the pathway to fully
identify the bands they claim to be repeating units, and thereby validate their assay.
Fig. 6B. As mentioned above, I believe the strength of these experiments is the lack of an effect in
the case of the clpC, clpE and clpL strains. Finding a way to summarize the results from the growth
curves in a bar graph could be used to bring these important controls into the main text.
MINOR COMMENTS:
lines 222-224: The sentence needs clarification. Was a different construct than the one described
above used? It is not clear from the context what is being introduced by erythromycin selection.
line 266: Inability to replace the genes merely suggests essentiality; it does not "confirm" it.
Reviewer #2:
Summary
Liu et. al. constructed a genome scale library of essential gene knockdown strains by using CRISPR
interference (CRISPRi) in S. pneumoniae, and identified phenotypes for the knockdown strains
© European Molecular Biology Organization
2
Molecular Systems Biology Peer Review Process File
using a systematic microscopy screen. The importance of this study is that it is the first CRISPRi
screen in non-rod-shaped bacteria, as well as the first one in a pathogen, and will be a valuable
resource for the research community. It further shows the utility of using CRISPRi as for studying
essential gene function. The authors follow up on a number of phenotypes to ascribe function to
essential genes without previous functional description. Several aspects of the manuscript should be
clarified or improved, as described below.
General Comments
1. The overarching theme of this work is to ascribe function to poorly characterized genes that are
essential to the organism. Therefore, it is incumbent on the authors to carefully describe both
existing annotation and how the gene set to be analyzed was chosen.
a. which databases were used?1. The au
b. Line 347: please expand on the source and nature of the mis-annotation
c. In general, Tn-seq screens overestimate the number of essential genes. What criteria did you use
for deciding on the set of genes to profile?
d. Lines 207-208: Please explain in the text how you determined which subset to phenotype.
2. Sequence similarity between investigated genes and other genes should be better described in the
body of the text, so that the reader can distinguish between cases where sequence similarity might
have sufficed to establish function and cases where the authors follow-up was essential. At present,
sequence similarity is discussed only in the supplement. For example in the investigation of
dnaB/dnaD and spd1405/spd1522 the sequence similarities and bioinformatics likely would have
sufficed to establish function, whereas for tarP/tarQ, the CRISPRi phenotype and follow-up was
essential. Example: In Line 202: "our data revealed" should be clarified, at least in part by bringing
supplemental data about sequence similarity into the body of the text.
3. For the genes explored in detail, care should be taken to explicate the genomic context and
enumerate the potential polar effects within operons. The recent work on the B. subtilis essential
gene set showed that CRISPRi exhibits polarity both on downstream and upstream genes in an
operon. Therefore, if the knockdown gene of interest is in an operon, the authors should mention the
possibility that several genes in addition to the targeted gene may be responsible for the phenotype.
One example of this issue is that the authors show the rpoA (alpha subunit of RNA polymerase) as
an indicator of transcription. However, rpoA is in an operon with ribosomal proteins and the
phenotype could be one resulting from those genes.
4. Various experiments are missing important controls. Examples:
a. Figure 1D: please add a control comparing WT (+luc, no-dCas9) to the dCas9 strain (+dCas9,
+luc) to verify that Cas9 itself is not responsible for changes.
b. Lines 149-150: Is the sequence targeted by sgRNAluc also found in the com operon (as a perfect
or imperfect match)? This might explain the slight-but reproducible -knockdown of com genes in
Fig. 1D. It would be a good idea to mention here that you ruled out perfect or near-perfect matches.
c. Lines 340-342: If you really are picking up suppressors during the growth of CRISPRi
knockdown strains, I would expect that most of these suppressors would be in the CRISPRi
machinery (e.g., dCas9, see Zhao et al., 2016, PMID: 27528508) and that the resulting strains would
grow like wildtype. Do the final populations, diluted back to starting concentration, display the
expected wildtype growth curve, or do they repeat the delayed lag phase, complicating the result?
d. While not required, attempting to delete the native spd1416/1417 genes in the presence of Zn+
(without gfp fusion backup) will rule out the possibility that Zn+ itself is suppressing the essential
phenotype.
Additional comments
1. Lines 60-64: needs to be rewritten for clarity (or the passage should simply be removed).
2. Lines 95-97: Saying the absence of CRISPR/Cas is "likely because it interferes" with natural
transformation would be better characterized as consistent with interference" or similar.
3. Figure 2C: Please consider whether this diagram is actually better than a simple Venn diagram.
The atypical representation for overlap of datasets is of unclear value, and makes rapid
© European Molecular Biology Organization
3
Molecular Systems Biology Peer Review Process File
comprehension of the data more difficult.
4. Figure 3: There is not sufficient information in the legend describing the various phenotypes that
are pointed out.
5. Lines 164-166: What was learned from the strains that could not be constructed? What
characterized the failure mode? Please say more about this.
6. Lines 183-184: "we first analyzed the effects of CRISPRi-based repression on cell morphology
using 69 genes that have been reported to be essential or crucial for normal pneumococcal growth"
Needs reference.
7. The supplement (Fig. S2) contains examples of growth defect and lysis phenotypes, but as this is
an important component of understanding the initial phenotype generation we would like to see
more examples (in the supplement) and in particular several examples of what "no phenotype" and
"both phenotypes"look like.
8. Figure 4D: gatD microscopy (bottom-left panel) is compelling on close inspection, but at current
resolution, printed out to paper, the graphic does not make membrane localization very apparent.
Reviewer #3:
In this study Xue Liu et al. perform a CRISPRi assay using an arrayed library of 348 sgRNA
targeting potentially essential genes of S. pneumoniae identified through a Tn-seq screen. The
authors establish a practical CRISPRi screen in S. pneumoniae (note however that CRISPRi was
already demonstrated in S. pneumoniae in PMID:23761437, this paper should be cited). By
combining CRISPRi with high-content microscopy they are able to phenotype individual
knockdowns and identify the function of several genes annotated as unknown. Overall this is a very
nice study, the experiments are performed with all the appropriate controls and the paper is well
written. One important point that I think is missing is a more detailed analysis of the targeted genes
that do not show a growth phenotype. Why were these genes identified as essential by other methods
and not CRISPRi?
Minor comments:
L215-225: this part is poorly written. It should be explained at the beginning of the paragraph that
both C and N-terminal fusions were attempted and the nomenclature of the plasmids explained,
otherwise the reader is lost when going over the first mention of Pzn-spd1416-gfp.
Figure2: I find the representation of panel D a little confusing. The scale of the axis are different and
the x-axis basically provides no information. Maybe showing a distribution plot would be easier to
read?
Table 1 is poorly formatted and hard to understand
1st Revision - authors' response
03 March 2017
Authors replies to the referees.
We would like to thank the reviewers for their supportive comments and suggestions.
We think that the changes and additional experiments we have made to address these
points have considerably improved the manuscript. We provide a detailed
point-by-point response and a list of the specific changes made to the Figures and
Supporting Information below. The changes made in the text in line with comments
from referees are highlighted.
The specific changes made to Figures and Supporting Information are as follows:
Main figures:
Figure 1: Minor formatting changes
Figure 2: Panel C was reformatted in line with the comment from Reviewer #2; Panel D
© European Molecular Biology Organization
4
Molecular Systems Biology Peer Review Process File
was reformatted and extended in line with the comment from Reviewer #3.
Figure 3: Minor formatting changes
Figure 4: Micrograph of GatD in panel D was replaced in line with the comment from
Reviewer #2.
Figure 5: Minor formatting changes
Figure 6 was extended to include results of all the tested clp genes (Panel C) in line with
the comment from Reviewer #1.
Extended view figures:
EV1: old Figure S1, a new panel B was added.
EV2 is the old Figure S2 with new panels D and E to show more examples of growth
defects, in line with the comment from Reviewer #2.
EV3 is new and shows the presence of suppression mutations of the CRISPRi system in
line with comment from Review #2.
EV4: same as old Figure S15.
Appendix Figures:
Appendix Figure S1 is new and shows the growth curves of CRISRPi strains targeting
dispensable genes (as negative controls), and examples of essential genes, in line with
comments from Reviewer #1.
Appendix Figure S2 is new and demonstrates the polar effect of CRISPRi with
RNA-Seq data and operons containing both essential and dispensable genes, as
suggested by Reviewer #1 and Reviewer #2.
Appendix Figure S3-S14: as before
Appendix Figure S15 is new and shows the TA western blotting results with a control
gene (tarI) in line with the comments from Reviewer #1.
Specific point-by-point responses to the referees’ comments and other changes made to
the manuscript are shown below.
Reviewer #1:
The study by Liu and coauthors is an elegant example of the adaptation of
CRISPRi technology to the phenotypization of bacterial genes. This technology
has previously been adapted to a few other bacteria, and the scale or the current
work certainly merits significant attention. The authors focus their work on genes
predicted to be essential by their transposon mutagenesis experiments, designing
a library of sgRNAs that spans ~15% of the protein coding genes. From careful
phenotypic analysis of several candidates the authors help elucidate genes
involved in peptidoglycan and teichoic acid synthesis, and natural competence.
The latter experiments are particularly striking, in part because of the quality of
the negative controls provided by the non-participating subunits.
My enthusiasm for the study is mollified by the general lack of key controls
present in the study. For instance, an adequate number of targets predicted to be
dispensable in their transposon experiment could have served to assess the
precision of their CRISPRi approach. Such controls could be easily included and
would significantly improve the study, helping to assess the reliability of the
CRISPRi method, and increasing the confidence of a few key observations.
Provided with the adequate controls, I have no doubt the work will have a
tremendous impact on the study of Streptococcus pneumonia.
We thank the referee for the overall positive evaluation of our study. To alleviate the
referee’s major concern, we now included CRISPRi data of several non-essential genes
as controls to assess the precision of the CRISPRi approach (new Appendix Figure S1).
For details, see our responses to the specific comments below.
MAJOR COMMENTS:
Fig. 2. Given that polar effects have been well documented when using CRISPRi
© European Molecular Biology Organization
5
Molecular Systems Biology Peer Review Process File
in bacteria, the authors should analyze their screening results in the context of the
operon architecture. When targeting the same operon, what is the agreement of
different sgRNAs? Such an analysis would be facilitated by including selected
sgRNAs against operons that contain both essential and dispensable genes. Are
there any cases where the CRISPRi effect on a known dispensable operon,
extends into an essential operon? Quantifying and displaying the magnitude of
such effects, for instance by performing RNAseq experiments on a couple of other
genes, would help determine the degree of confidence that readers should have on
the authors' data set. However, a simple analysis of the existing data is probably
sufficient to satisfy this point, since there are a lot of genes that did not show a
phenotype in their analysis.
We agree that this is an interesting question and we have tackled this in several ways.
First, we have reanalyzed our RNA-Seq data with the luc reporter strain and indeed
observed the polar effects on downstream of the luc gene as we did not include a
transcription terminator after the luc gene (new Appendix Figure S2A, Table EV2).
This also explains the majority of significantly downregulated genes upon dCas9
induction (Fig. 1D). To test the polar effect with real operon examples, we selected 6
operons that contain both essential and dispensable genes, based on a preliminary study
about identification of pneumococcal transcription start sites in our group (Slager et al.,
unpublished). As shown in new Appendix Figure S2, a polar effect could be observed
upon dCas9 induction in all cases. Thus, whenever an essential gene was downstream
of a non-essential gene, growth was still affected when the sgRNA targeted
transcription of the dispensable gene. To validate that these targeted dispensable genes
are truly non-essential, we made non-polar replacement mutants (Appendix Figure 2B
and 2C). This validated the Tn-seq data and showed that the genes can be readily
replaced by an antibiotic resistance marker (without promoter and without terminator).
Interestingly, it seems that in some cases there can even be a polar effect on upstream
genes (Appendix Fig S2D, operon number 5, spd_1895). However, when the operon
contains multiple genes and the targeted dispensable gene localizes at the end of the
operon, the polar effect on upstream genes is much less (Appendix Figure S2D, operon
number 6, lytC). This new data is now described in the revised manuscript (Lines
190-191).
Since this study mainly focused on essential genes, we didn’t have many CRISPRi
strains targeting dispensable operons. In our study, we didn’t observe any case where
the CRISPRi effect on a known dispensable operon extends into an essential operon
according to the growth analysis.
Fig. 2C. Were any of the genes predicted to be dispensable further analyzed?
Presumably some of these were particularly implicated in essential functions by
the transposon experiments, yet appeared dispensable by CRISPRi.
Understanding the source of the discrepancy, could help establish the reliability
of the methods and the ultimate relevance of the screening results.
Indeed, some genes were identified as essential in Tn-seq, but appeared dispensable by
CRISPRi, and this discrepancy can be caused by several reasons. These include
insufficient repression of transcription for instance because the sgRNA targets a PAM
site far away from the TSS, some proteins are sufficient at just a few molecule numbers,
poor access of the sgRNA-dCas9 complex to the target DNA, conditional essentiality
(growth on plates vs. growth in liquid), or polar effects within the operon alleviating the
essentiality. This is now clarified in the Discussion of the revised manuscript (Lines
384-398).
Fig. 3. Measuring growth curved is a central part of the study's phenotypic
assessment, and would benefit from a set of genes pre-selected as controls. Is there
a set of genes that are known to be dispensable and could be used as controls?
This is particularly important given the documented polar effects of the CRISPRi
system.
This is a good point, and we generated 8 new strains targeting non-essential genes.
CRISPRi knockdown of these randomly selected 8 dispensable genes didn’t
significantly influence growth of S. pneumoniae (Appendix Figure S1A), main text on
Lines 183-189.
© European Molecular Biology Organization
6
Molecular Systems Biology Peer Review Process File
Can sgRNAs against dispensable genes at the end of an operon be used to repress
such genes without affecting the rest of the operon?
To test this, we included two operons with dispensable genes at the end, while with
essential genes are upstream (Appendix Figure S2D, operon number 5 and number 6).
Interestingly, in operon number 5, we observed that CRISPRi repression of dispensable
gene spd_1895 led to a significant growth defect, probably due to the polar effect on the
upstream essential gene gltX. However, with operon number 6, which consists of 4
genes, it seems that the effect on upstream genes by CRISPRi repression of the
dispensable gene lytC cannot be detected in the growth analysis.
Fig. 4. Depletion of the GFP-tagged proteins should be demonstrated, to confirm
the effects of the genetic manipulations. This could be ideally performed by
immunoblotting against the GFP portion of the fusion proteins, compared to a
control. Alternatively the strains could be compared by FACS, provided an
adequate counterstain to gate on the bacteria.
The zinc-inducible system we used in this study has been reported to be suitable for the
fine control of gfp-fusion gene expression and for protein depletion experiments in S.
pneumoniae (Eberhardt et. al., 2009 Molecular Microbiology), and has been applied for
depletion in S. pneumoniae by several studies and confirmed as functional (Zapun, et.
al., 2013 ACS Chem. Biol; Kjos, et. al., 2016 Plos Pathogen.; Beilharz et. al., 2012
PNAS). Here, in this study, we did in-gel fluorescence of GFP and proved that the GFP
is successfully fused to the target protein (Appendix Fig S14), and we indeed observed
that the PZn depletion strains responded to the inducer, Zn2+ (Figure 4A, 5A and
Appendix Fig S13A). Nevertheless, we have performed a depletion assay as requested
by the reviewer, and, although the resolution is not great, the image below clearly
shows that the zinc-inducible system can be reliably used to deplete proteins.
Fig. 5E. I recognize that the monoclonal antibody against phosphorylcholine has
previously been used; however, to my knowledge, the small species found in the
TarP/Q knockdowns have not been documented. If I am wrong on this point,
please reference such an observation appropriately. Alternatively, the authors
should include a mutant of an upstream member of the pathway to fully identify
the bands they claim to be repeating units, and thereby validate their assay.
Good point. The monoclonal antibody used to detect teichoic acids can only recognize
molecules with choline modification, and in the teichoic acid biosynthesis pathway of S.
pneumoniae, choline modification of precursors happens just before polymerization
(Denapaite et al, 2012 Microbial Drug Resistance). So far, to the best of our knowledge,
no gene has been implicated in the polymerization of TA precursors in S. pneumoniae,
and thus there is no documentation of the small species observed in our gels. In the
revised manuscript, we included a CRISPRi strain targeting tarI of the lic1 locus,
which is involved in the very early step of precursor synthesis as a control, and repeated
© European Molecular Biology Organization
7
Molecular Systems Biology Peer Review Process File
the experiment (new Appendix Fig S15). Note that tarI is cotranscribed with the other 4
genes of the lic1 locus, including tarJ, licA, licB and licC. Theoretically, CRISPRi
knockdown of tarI will lead to repression of the whole lic1 locus, and thus block the
synthesis of TA precursors. In line with this, we observed reduction of teichoic acid
chains when tarI is repressed by CRISPRi, different from the band pattern of repression
of the studied tarP and tarQ. For text related with this point, see lines 337-343.
Fig. 6B. As mentioned above, I believe the strength of these experiments is the
lack of an effect in the case of the clpC, clpE and clpL strains. Finding a way to
summarize the results from the growth curves in a bar graph could be used to
bring these important controls into the main text.
Many thanks for this suggestion. We now added a new panel C in Figure 6
summarizing the results of all the studied clp genes.
MINOR COMMENTS:
lines 222-224: The sentence needs clarification. Was a different construct than the
one described above used? It is not clear from the context what is being
introduced by erythromycin selection.
Yes, different from the above used N-terminal GFP fusions, we used a strain
ectopically expressing C-terminal GFP fusions. For clarification, we have rewritten this
paragraph, Lines 245-258.
line 266: Inability to replace the genes merely suggests essentiality; it does not
"confirm" it.
Corrected. Line 301.
Reviewer #2:
Summary
Liu et. al. constructed a genome scale library of essential gene knockdown strains
by using CRISPR interference (CRISPRi) in S. pneumoniae, and identified
phenotypes for the knockdown strains using a systematic microscopy screen. The
importance of this study is that it is the first CRISPRi screen in non-rod-shaped
bacteria, as well as the first one in a pathogen, and will be a valuable resource for
the research community. It further shows the utility of using CRISPRi as for
studying essential gene function. The authors follow up on a number of
phenotypes to ascribe function to essential genes without previous functional
description. Several aspects of the manuscript should be clarified or improved, as
described below.
General Comments
1. The overarching theme of this work is to ascribe function to poorly
characterized genes that are essential to the organism. Therefore, it is incumbent
on the authors to carefully describe both existing annotation and how the gene set
to be analyzed was chosen.
a. which databases were used?1. The au
In this study, the annotation from NCBI (Version: CP000410.1, updated on
31-JAN-2015) was used. We have modified the text, and added this information in the
introduction (Lines 78-79) and result sections (Lines 222-223; 237) of the revised
manuscript.
b. Line 347: please expand on the source and nature of the mis-annotation.
The source has been added in the revised manuscript.
“Note that spd_2030 (dnaC) was mis-annotated as dnaB in several databases, such as in
NCBI (ProteinID:ABJ54728), KEGG (Entry: SPD_2030), UniProt (Entry:
A0A0H2ZNF7), which may be due to the different naming of primosomal proteins in E.
coli and Bacillus subtilis (Briggs et al, 2012; Smits et al, 2011)”, lines 403-406.
© European Molecular Biology Organization
8
Molecular Systems Biology Peer Review Process File
c. In general, Tn-seq screens overestimate the number of essential genes. What
criteria did you use for deciding on the set of genes to profile?
The criteria for selection of genes targeted in our CRISPRi library are described in the
Materials and Methods section (lines 472-480) and Table EV1 in detail. To select genes
for our CRISPRi study, we used the output of three independent Tn-seq studies: our
Tn-seq study on strain D39 and two published Tn-seq studies using strain TIGR4 (van
Opijnen et al., 2009 Nat Methods; van Opijnen et al., 2012 Genome research). We first
assigned a score to each gene based on the fitness cost obtained from each Tn-seq study,
and then summarized the essentiality score of each gene of the 3 Tn-seq studies. Then
the genes with relatively higher essentiality scores were selected (Table EV1).
To clarify this, we have modified the main text of the revised manuscript (Lines
113-119). Moreover, to explain why Tn-seq screens overestimate the number of
essential genes, we added a paragraph in the Discussion part (lines 386-398).
d. Lines 207-208: Please explain in the text how you determined which subset to
phenotype.
To explore the function of unknown genes, we focused on a subset of the 348
CRISPRi knockdowns targeting genes that were annotated as “hypothetical protein”
in the NCBI database (Version: CP000410.1, updated on 31-JAN-2015). This has
now been clarified in the revised manuscript (Lines 235-238).
2. Sequence similarity between investigated genes and other genes should be
better described in the body of the text, so that the reader can distinguish between
cases where sequence similarity might have sufficed to establish function and
cases where the authors follow-up was essential. At present, sequence similarity is
discussed only in the supplement. For example in the investigation of dnaB/dnaD
and spd1405/spd1522 the sequence similarities and bioinformatics likely would
have sufficed to establish function, whereas for tarP/tarQ, the CRISPRi
phenotype and follow-up was essential. Example: In Line 202: "our data
revealed" should be clarified, at least in part by bringing supplemental data
about sequence similarity into the body of the text.
Agreed and we have clarified this in the revised manuscript (Lines 227-232).
3. For the genes explored in detail, care should be taken to explicate the genomic
context and enumerate the potential polar effects within operons. The recent
work on the B. subtilis essential gene set showed that CRISPRi exhibits polarity
both on downstream and upstream genes in an operon. Therefore, if the
knockdown gene of interest is in an operon, the authors should mention the
possibility that several genes in addition to the targeted gene may be responsible
for the phenotype. One example of this issue is that the authors show the rpoA
(alpha subunit of RNA polymerase) as an indicator of transcription. However,
rpoA is in an operon with ribosomal proteins and the phenotype could be one
resulting from those genes.
This is a good point also raided by Referee #1. To highlight the polar effects of
CRISPRi technology, we made a new figure (Appendix Fig S2), see our responses to
Referee 1. We have clarified this also in the text (Line 190-191).
4. Various experiments are missing important controls. Examples:
a. Figure 1D: please add a control comparing WT (+luc, no-dCas9) to the dCas9
strain (+dCas9, +luc) to verify that Cas9 itself is not responsible for changes.
b. Lines 149-150: Is the sequence targeted by sgRNAluc also found in the com
operon (as a perfect or imperfect match)? This might explain the slight-but
reproducible -knockdown of com genes in Fig. 1D. It would be a good idea to
mention here that you ruled out perfect or near-perfect matches.
We had included the proper controls but made this not clear enough. This has been
corrected in the revised manuscript. In the RNA-Seq study, we included three different
samples:
sample 1: WT(+luc, +sgRNAluc, +dCas9) without IPTG
sample 2: WT(+luc, +sgRNAluc, +dCas9) with 1 mM IPTG
© European Molecular Biology Organization
9
Molecular Systems Biology Peer Review Process File
sample 3: WT(+luc, -sgRNAluc, +dCas9) with 1 mM IPTG.
This shows that expression of dcas9 is stringently repressed by LacI in the absence of
IPTG in Figure EV1, and RNA-Seq data of the sample 1 also confirmed this, because
the RPKM-value of dcas9 gene is practically zero in this sample. Based on these
observations, the sample 1 essentially acts as a no-dCas9 control. What’s more, in the
supplementary table 2, we provided the comparative analysis of sample 2 vs. sample 1;
and sample 3 vs. sample 1. From the results, we can see that high expression of dcas9 in
the strain without sgRNA didn’t cause any significant changes of the luc and the
downstream genes, which was observed in the strain with sgRNA, confirming that the
changes are not caused by dCas9 itself but by the functional CRISPRi system which
requires expression of both dCas9 and sgRNA.
In addition, we observed a slight down-regulation of several competence-related genes
in both sample 2 and sample 3. This is probably due to the fact that the applied
experimental conditions allowed for natural competence to develop. Since the
expression of the competence regulon is strongly regulated, but also highly
time-dependent, small changes in their expression are almost inevitable. We have also
checked for perfect or imperfect matches of the luc sgRNA to com genes and have
excluded this. This observation supports the idea that the slight repression of com genes
does not rely on the CRISPRi system (addressing question b). These results are now
better explained in the revised manuscript (Line 145-155).
c. Lines 340-342: If you really are picking up suppressors during the growth of
CRISPRi knockdown strains, I would expect that most of these suppressors
would be in the CRISPRi machinery (e.g., dCas9, see Zhao et al., 2016,
PMID: 27528508) and that the resulting strains would grow like wildtype. Do the
final populations, diluted back to starting concentration, display the expected
wildtype growth curve, or do they repeat the delayed lag phase, complicating the
result?
To answer this question, we performed the suggested experiment. The results are
similar to the expectation of this reviewer (new Figure EV3). We grew 6 of the
CRISPRi strains targeting different essential genes in C+Y medium with 1 mM IPTG.
After long-time incubation (15 hours), bacterial cells were streaked onto agar plate to
purify single colonies. Single colonies were picked and used for a new round of growth
analysis. The data shows that all the 6 purified ‘suppressor’ strains do not respond to
IPTG anymore. Further sequencing analysis of the sgRNA and Plac-dcas9 of these
strains showed that most of the suppression mutations mapped to the coding sequence
of dcas9 (Figure EV3B). Description of this result has been added into the revised
manuscript, please refer to lines 191-198.
d. While not required, attempting to delete the native spd1416/1417 genes in the
presence of Zn+ (without gfp fusion backup) will rule out the possibility that Zn+
itself is suppressing the essential phenotype.
Good point and we agree that this was not clearly described. In fact, we did this control
for every essential gene confirmation in this study, and found that attempts to knockout
these genes all failed in the wild-type background with Zn2+, ruling out the possibility
of Zn2+ suppressing the essential phenotype. To clarify this, we have added this
information in the revised manuscript (Line 251-253).
Additional comments
1. Lines 60-64: needs to be rewritten for clarity (or the passage should simply be
removed).
Done (passage was removed).
2. Lines 95-97: Saying the absence of CRISPR/Cas is "likely because it interferes"
with natural transformation would be better characterized as consistent with
interference" or similar.
We have modified this as below:
“Note that S. pneumoniae does not contain an endogenous CRISPR/Cas system,
consistent with interference with natural transformation and thereby lateral gene
© European Molecular Biology Organization
10
Molecular Systems Biology Peer Review Process File
transfer that is crucial for pneumococcal host adaptation (Bikard et al, 2012).”
3. Figure 2C: Please consider whether this diagram is actually better than a
simple Venn diagram. The atypical representation for overlap of datasets is of
unclear value, and makes rapid comprehension of the data more difficult.
Thanks for this advice. We have improved this figure, please refer to the new Figure 2C.
4. Figure 3: There is not sufficient information in the legend describing the
various phenotypes that are pointed out.
Thanks for the advice. We have extended the description in the legend. Please refer to
the revised figure legend of Fig 3.
5. Lines 164-166: What was learned from the strains that could not be constructed?
What characterized the failure mode? Please say more about this.
The failed 43 sgRNA cloning may be caused by some technical reasons, for example,
quality of the primers that we ordered; technical failure of the primer annealing; the
limitation of cloning efficiency of the mentioned infusion-cloning, etc. Since we intend
to apply the designed CRISPRi technique as a high-throughput tool for genome-scale
study, we didn’t do multiple attempts to clone these sgRNAs to exclude the possibility
of technical reason caused failure. We thus cannot conclude on any common
characteristics of the failed sgRNAs. We have clarified this in the revised manuscript
(Lines 169-172).
6. Lines 183-184: "we first analyzed the effects of CRISPRi-based repression on
cell morphology using 69 genes that have been reported to be essential or crucial
for normal pneumococcal growth" Needs reference.
We have added references for this (van Opijnen et al, 2009; van Opijnen & Camilli,
2012), see lines 204-208 of the revised manuscript.
7. The supplement (Fig. S2) contains examples of growth defect and lysis
phenotypes, but as this is an important component of understanding the initial
phenotype generation we would like to see more examples (in the supplement) and
in particular several examples of what "no phenotype" and "both
phenotypes"look like.
We agree and have added additional examples, see the new Appendix Figure S2B-E.
8. Figure 4D: gatD microscopy (bottom-left panel) is compelling on close
inspection, but at current resolution, printed out to paper, the graphic does not
make membrane localization very apparent.
Agreed and we have improved this figure, see new Figure 4D.
Reviewer #3:
In this study Xue Liu et al. perform a CRISPRi assay using an arrayed library of
348 sgRNA targeting potentially essential genes of S. pneumoniae identified
through a Tn-seq screen. The authors establish a practical CRISPRi screen in S.
pneumoniae (note however that CRISPRi was already demonstrated in S.
pneumoniae in PMID:23761437, this paper should be cited). By combining
CRISPRi with high-content microscopy they are able to phenotype individual
knockdowns and identify the function of several genes annotated as unknown.
Overall this is a very nice study, the experiments are performed with all the
appropriate controls and the paper is well written. One important point that I
think is missing is a more detailed analysis of the targeted genes that do not show
a growth phenotype. Why were these genes identified as essential by other
methods and not CRISPRi?
We thank the reviewer and apologize for not citing the mentioned paper, this has now
been fixed.
We appreciate the advice to provide a more detailed analysis of the targeted genes that
do not show a growth phenotype (see also our reply to Referee 1). To explain this in
© European Molecular Biology Organization
11
Molecular Systems Biology Peer Review Process File
more detail, we added a paragraph in the Discussion (lines 384-398).
Minor comments:
L215-225: this part is poorly written. It should be explained at the beginning of
the paragraph that both C and N-terminal fusions were attempted and the
nomenclature of the plasmids explained, otherwise the reader is lost when going
over the first mention of Pzn-spd1416-gfp.
Agreed and we have rewritten this paragraph (Line 245-258).
Figure2: I find the representation of panel D a little confusing. The scale of the
axis are different and the x-axis basically provides no information. Maybe
showing a distribution plot would be easier to read?
We made a new panel D, providing negative controls compared with the CRISPRi
strains, to show the growth feature of the knockdown strains. Please refer to the new
panel D of Figure 2.
Table 1 is poorly formatted and hard to understand
The Table 1 was re-formatted. Please refer to the new Table 1.
2nd Editorial Decision
03 April 2017
Thank you again for sending us your revised manuscript. We have now heard back from the two
referees who were asked to evaluate your study. As you will see below they think that most issues
have been satisfactorily addressed. However, reviewer #2 lists a few remaining concerns, which we
would ask you to address in a minor revision.
In your revision, we would also like to ask you to address some remaining editorial issues listed
below.
--------------------------------------------------------------------------REFEREE REVIEWS
Reviewer #1:
The authors have provided extensive new data, figures and clarifications that together answer all of
my initial concerns. In my opinion, the paper has been greatly improved by these changes and will
undoubtedly have a significant impact on the field.
Reviewer #2:
The authors revisions address most of our concerns, but there are twopoints that still need work.
1) "In general, Tn-seq screens overestimate the number of essential genes. What criteria did you use
for deciding on the set of genes to profile?"
The presented table (EV1) and explanations in the new text are
needlessly confusing. The provided "essentiality score" is not
justified or characterized. There is no obvious meaning for "core genes", much less any reason to
think that a 0.2 or a 0.5 is a good numerical score for a given characterization. Furthermore the
multiple cutoff thresholds for sum-of-scores across different data subsets are not justified and appear
to be internally inconsistent.
We suggest that something much much simpler would serve the authors far better. Something like
"We chose to include all genes that we found to be essential, and added in genes that were found to
be essential by these previous Tn-Seq studies", with any exceptions noted and briefly explained. The
tables are useful for comparison, but the quasi-quantitative scores and thresholds are not.
© European Molecular Biology Organization
12
Molecular Systems Biology Peer Review Process File
2) "For the genes explored in detail, care should be taken to
explicate the genomic context and enumerate the potential polar
effects within operons. The recent work on the B. subtilis essential gene set showed that CRISPRi
exhibits polarity both on downstream and upstream genes in an operon. Therefore, if the knockdown
gene of interest is in an operon, the authors should mention the possibility that several genes in
addition to the targeted gene may be responsible for the phenotype. One example of this issue is that
the authors show the rpoA (alpha subunit of RNA polymerase) as an indicator of transcription.
However, rpoA is in an operon with ribosomal proteins and the phenotype could be one resulting
from those genes."
rpoA is still the presented exemplar for "Transcription" phenotype in Figure 3, despite the polar
interactions with ribosomal genes. A different example should be used that does not share this
complication (e.g. rpoB, rpoC, from figure S4).
It is still not clear whether the phenotypes presented in figure 3 are general to the annotation class,
or specific to the presented gene. The text should address this question. The phenotypes presented in
supplemental figures S3-S10 need explanations of observed phenotypes as have now been added to
figure 3 to help clarify this point further. Ideally we would see representative exemplars of each
phenotype in figure 3, reinforced by the other examples in S3-S10, to support the
claim in 220-221 that this approach was useful for 'functional
verification' of uncharacterized genes.
2nd Revision - authors' response
12 April 2017
We have made the following changes in the second revised version:
1) Changed the descriptions of selection of genes to be involved in the CRISPRi library as suggested
by the reviewer #2;
2) Replaced the rpoA result with rpoC, which is cotranscribed with another transcription related
gene rpoB, in Figure 3, as suggested by reviewer #2;
3) Explained the observed phenotype in Figure 3 as suggested by reviewer #2;
© European Molecular Biology Organization
13
EMBO PRESS YOU MUST COMPLETE ALL CELLS WITH A PINK BACKGROUND 
PLEASE NOTE THAT THIS CHECKLIST WILL BE PUBLISHED ALONGSIDE YOUR PAPER
USEFUL LINKS FOR COMPLETING THIS FORM
Corresponding Author Name: Jan-­‐Willem Veening
Journal Submitted to: Molecular systems biology
Manuscript Number: MSB-­‐16-­‐7449
http://www.antibodypedia.com
http://1degreebio.org
Reporting Checklist For Life Sciences Articles (Rev. July 2015)
http://www.equator-­‐network.org/reporting-­‐guidelines/improving-­‐bioscience-­‐research-­‐reporting-­‐the-­‐arrive-­‐guidelines-­‐for-­‐r
This checklist is used to ensure good reporting standards and to improve the reproducibility of published results. These guidelines are consistent with the Principles and Guidelines for Reporting Preclinical Research issued by the NIH in 2014. Please follow the journal’s authorship guidelines in preparing your manuscript. http://grants.nih.gov/grants/olaw/olaw.htm
http://www.mrc.ac.uk/Ourresearch/Ethicsresearchguidance/Useofanimals/index.htm
A-­‐ Figures 1. Data
The data shown in figures should satisfy the following conditions:
http://ClinicalTrials.gov
http://www.consort-­‐statement.org
http://www.consort-­‐statement.org/checklists/view/32-­‐consort/66-­‐title
 the data were obtained and processed according to the field’s best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner.
 figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way.
 graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates.
 if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified
 Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation.
http://www.equator-­‐network.org/reporting-­‐guidelines/reporting-­‐recommendations-­‐for-­‐tumour-­‐marker-­‐prognostic-­‐studies
http://datadryad.org
http://figshare.com
http://www.ncbi.nlm.nih.gov/gap
http://www.ebi.ac.uk/ega
2. Captions
http://biomodels.net/
Each figure caption should contain the following information, for each panel where they are relevant:




http://biomodels.net/miriam/
http://jjj.biochem.sun.ac.za
http://oba.od.nih.gov/biosecurity/biosecurity_documents.html
http://www.selectagents.gov/
a specification of the experimental system investigated (eg cell line, species name).
the assay(s) and method(s) used to carry out the reported observations and measurements an explicit mention of the biological and chemical entity(ies) that are being measured.
an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner.
 the exact sample size (n) for each experimental group/condition, given as a number, not a range;
 a description of the sample collection allowing the reader to understand whether the samples represent technical or biological replicates (including how many animals, litters, cultures, etc.).
 a statement of how many times the experiment shown was independently replicated in the laboratory.
 definitions of statistical methods and measures:
 common tests, such as t-­‐test (please specify whether paired vs. unpaired), simple χ2 tests, Wilcoxon and Mann-­‐Whitney tests, can be unambiguously identified by name only, but more complex techniques should be described in the methods section;
 are tests one-­‐sided or two-­‐sided?
 are there adjustments for multiple comparisons?
 exact statistical test results, e.g., P values = x but not P values < x;
 definition of ‘center values’ as median or average;
 definition of error bars as s.d. or s.e.m. Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.
Please ensure that the answers to the following questions are reported in the manuscript itself. We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects. In the pink boxes below, provide the page number(s) of the manuscript draft or figure legend(s) where the information can be located. Every question should be answered. If the question is not relevant to your research, please write NA (non applicable).
B-­‐ Statistics and general methods
Please fill out these boxes  (Do not worry if you cannot see all your text once you press return)
1.a. How was the sample size chosen to ensure adequate power to detect a pre-­‐specified effect size?
NA
1.b. For animal studies, include a statement about sample size estimate even if no statistical methods were used.
NA
2. Describe inclusion/exclusion criteria if samples or animals were excluded from the analysis. Were the criteria pre-­‐
established?
NA
3. Were any steps taken to minimize the effects of subjective bias when allocating animals/samples to treatment (e.g. randomization procedure)? If yes, please describe. NA
For animal studies, include a statement about randomization even if no randomization was used.
NA
4.a. Were any steps taken to minimize the effects of subjective bias during group allocation or/and when assessing results NA
(e.g. blinding of the investigator)? If yes please describe.
4.b. For animal studies, include a statement about blinding even if no blinding was done
NA
5. For every figure, are statistical tests justified as appropriate?
Yes
Do the data meet the assumptions of the tests (e.g., normal distribution)? Describe any methods used to assess it.
NA
Is there an estimate of variation within each group of data?
NA
Is the variance similar between the groups that are being statistically compared?
NA
C-­‐ Reagents
6. To show that antibodies were profiled for use in the system under study (assay and species), provide a citation, catalog mAb-­‐TEPC15 (Sigma, catalog number:M1421); anti-­‐mouse IgG HRP antibody （GE Healthcare UK number and/or clone number, supplementary information or reference to an antibody validation profile. e.g., Limited， catalog number:NA931)
Antibodypedia (see link list at top right), 1DegreeBio (see link list at top right).
7. Identify the source of cell lines and report if they were recently authenticated (e.g., by STR profiling) and tested for mycoplasma contamination.
NA
* for all hyperlinks, please see the table at the top right of the document
D-­‐ Animal Models
8. Report species, strain, gender, age of animals and genetic modification status where applicable. Please detail housing and husbandry conditions and the source of animals.
NA
9. For experiments involving live vertebrates, include a statement of compliance with ethical regulations and identify the NA
committee(s) approving the experiments.
10. We recommend consulting the ARRIVE guidelines (see link list at top right) (PLoS Biol. 8(6), e1000412, 2010) to ensure NA
that other relevant aspects of animal studies are adequately reported. See author guidelines, under ‘Reporting Guidelines’. See also: NIH (see link list at top right) and MRC (see link list at top right) recommendations. Please confirm compliance.
E-­‐ Human Subjects
11. Identify the committee(s) approving the study protocol.
NA
12. Include a statement confirming that informed consent was obtained from all subjects and that the experiments conformed to the principles set out in the WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report.
NA
13. For publication of patient photos, include a statement confirming that consent to publish was obtained.
NA
14. Report any restrictions on the availability (and/or on the use) of human data or samples.
NA
15. Report the clinical trial registration number (at ClinicalTrials.gov or equivalent), where applicable.
NA
16. For phase II and III randomized controlled trials, please refer to the CONSORT flow diagram (see link list at top right) and submit the CONSORT checklist (see link list at top right) with your submission. See author guidelines, under ‘Reporting Guidelines’. Please confirm you have submitted this list.
NA
17. For tumor marker prognostic studies, we recommend that you follow the REMARK reporting guidelines (see link list at NA
top right). See author guidelines, under ‘Reporting Guidelines’. Please confirm you have followed these guidelines.
F-­‐ Data Accessibility
18. Provide accession codes for deposited data. See author guidelines, under ‘Data Deposition’.
Raw data of Tn-­‐seq are available at SRA archive with accession number SRR5298192; The RNA-­‐Seq data are available at GEO database with accession number GSE89763.
Data deposition in a public repository is mandatory for:
a. Protein, DNA and RNA sequences
b. Macromolecular structures
c. Crystallographic data for small molecules
d. Functional genomics data e. Proteomics and molecular interactions
19. Deposition is strongly recommended for any datasets that are central and integral to the study; please consider the Data were deposited in databases.
journal’s data policy. If no structured public repository exists for a given data type, we encourage the provision of datasets in the manuscript as a Supplementary Document (see author guidelines under ‘Expanded View’ or in unstructured repositories such as Dryad (see link list at top right) or Figshare (see link list at top right).
20. Access to human clinical and genomic datasets should be provided with as few restrictions as possible while NA
respecting ethical obligations to the patients and relevant medical and legal issues. If practically possible and compatible with the individual consent agreement used in the study, such data should be deposited in one of the major public access-­‐
controlled repositories such as dbGAP (see link list at top right) or EGA (see link list at top right).
21. As far as possible, primary and referenced data should be formally cited in a Data Availability section. Please state NA
whether you have included this section.
Examples:
Primary Data
Wetmore KM, Deutschbauer AM, Price MN, Arkin AP (2012). Comparison of gene expression and mutant fitness in Shewanella oneidensis MR-­‐1. Gene Expression Omnibus GSE39462
Referenced Data
Huang J, Brown AF, Lei M (2012). Crystal structure of the TRBD domain of TERT and the CR4/5 of TR. Protein Data Bank 4O26
AP-­‐MS analysis of human histone deacetylase interactions in CEM-­‐T cells (2013). PRIDE PXD000208
22. Computational models that are central and integral to a study should be shared without restrictions and provided in a NA
machine-­‐readable form. The relevant accession numbers or links should be provided. When possible, standardized format (SBML, CellML) should be used instead of scripts (e.g. MATLAB). Authors are strongly encouraged to follow the MIRIAM guidelines (see link list at top right) and deposit their model in a public database such as Biomodels (see link list at top right) or JWS Online (see link list at top right). If computer source code is provided with the paper, it should be deposited in a public repository or included in supplementary information.
G-­‐ Dual use research of concern
23. Could your study fall under dual use research restrictions? Please check biosecurity documents (see link list at top right) and list of select agents and toxins (APHIS/CDC) (see link list at top right). According to our biosecurity guidelines, provide a statement only if it could.
NA