Characterization of TelE, a T7SS LXG Effector Exhibiting a Conserved C-Terminal Glycine Zipper Motif Required for Toxicity

ABSTRACT Streptococcus gallolyticus subsp. gallolyticus (SGG) is an opportunistic bacterial pathogen strongly associated with colorectal cancer. Here, through comparative genomics analysis, we demonstrated that the genetic locus encoding the type VIIb secretion system (T7SSb) machinery is uniquely present in SGG in two different arrangements. SGG UCN34 carrying the most prevalent T7SSb genetic arrangement was chosen as the reference strain. To identify the effectors secreted by this secretion system, we inactivated the essC gene encoding the motor of this machinery. A comparison of the proteins secreted by UCN34 wild type and its isogenic ΔessC mutant revealed six T7SSb effector proteins, including the expected WXG effector EsxA and three LXG-containing proteins. In this work, we characterized an LXG-family toxin named herein TelE promoting the loss of membrane integrity. Seven homologs of TelE harboring a conserved glycine zipper motif at the C terminus were identified in different SGG isolates. Scanning mutagenesis of this motif showed that the glycine residue at position 470 was crucial for TelE membrane destabilization activity. TelE activity was antagonized by a small protein TipE belonging to the DUF5085 family. Overall, we report herein a unique SGG T7SSb effector exhibiting a toxic activity against nonimmune bacteria. IMPORTANCE In this study, 38 clinical isolates of Streptococcus gallolyticus subsp. gallolyticus (SGG) were sequenced and a genetic locus encoding the type VIIb secretion system (T7SSb) was found conserved and absent from 16 genomes of the closely related S. gallolyticus subsp. pasteurianus (SGP). The T7SSb is a bona fide pathogenicity island. Here, we report that the model organism SGG strain UCN34 secretes six T7SSb effectors. One of the six effectors named TelE displayed a strong toxicity when overexpressed in Escherichia coli. Our results indicate that TelE is probably a pore-forming toxin whose activity can be antagonized by a specific immunity protein named TipE. Overall, we report a unique toxin-immunity protein pair and our data expand the range of effectors secreted through T7SSb.

• Upload a compare copy of the manuscript (without figures) as a "Marked-Up Manuscript" file. • Each figure must be uploaded as a separate file, and any multipanel figures must be assembled into one file. • Manuscript: A .DOC version of the revised manuscript • Figures: Editable, high-resolution, individual figure files are required at revision, TIFF or EPS files are preferred For complete guidelines on revision requirements, please see the journal Submission and Review Process requirements at https://journals.asm.org/journal/Spectrum/submission-review-process. Submissions of a paper that does not conform to Microbiology Spectrum guidelines will delay acceptance of your manuscript. " Please return the manuscript within 60 days; if you cannot complete the modification within this time period, please contact me. If you do not wish to modify the manuscript and prefer to submit it to another journal, please notify me of your decision immediately so that the manuscript may be formally withdrawn from consideration by Microbiology Spectrum.
If your manuscript is accepted for publication, you will be contacted separately about payment when the proofs are issued; please follow the instructions in that e-mail. Arrangements for payment must be made before your article is published. For a complete list of Publication Fees, including supplemental material costs, please visit our website.
In this manuscript, Teh et al performed comparative genomics between subspecies of S. gallolyticus and found that subspecies gallolyticus (Sgg) uniquely encodes T7SS genes, often in a genomic arrangement similar to that found in S. intermedius. They further identified six Sgg T7SS effectors including a novel LXG-toxin, TelE, which promotes membrane integrity loss and can be mitigated by immunity factor TipE. Overall, this study provides an interesting and informative elaboration on the Sgg T7SS as well as identification and characterization of a novel T7SSb LXG toxin. Concerns exist that direct experimental evidence is not provided for claims of TelE poreforming activity, that the nomenclature chosen for Sgg T7SS proteins may cause confusion in the literature, and that relevant T7SSb literature is not sufficiently discussed. Additional concerns are listed below.
Major: 1. Claims of TelE pore-formation are not experimentally substantiated as pore-forming activity/membrane insertion of TelE is not directly shown via planar lipid bilayers or TEM pore visualization of the protein/in liposomes, etc. The description of TelE activity should be modified in the text to indicate that TelE promotes loss of membrane integrity, as indicated by propridium iodide staining. It is very possible this is due to pore formation; however, the evidence provided here is indirect and pore formation is not the only cause of membrane integrity loss. • Citation should be included on line 90.

Minor comments:
• Sequences obtained for clinical isolates should be uploaded to a public repository and accession numbers provided in the manuscript. Information on the 38 Sgg isolates (e.g., body site isolated from) would also be helpful.
• Statistical analyses were not performed in this manuscript. This may not be needed for E. coli intoxication assays but would be useful for microscopy data in Figures 4 and 5.
• Introduction: The transition from S. gallolyticus phylogeny to T7SS in the opening paragraphs of the introduction would be clearer if comparative genomics were introduced sooner. Can the authors clarify why completion of the UCN34 genome provided "clear insights for its adaptation to the rumen" in lines 62-63? Lines 70-71: SGP is less pathogenic but is common in neonatal meningitis, which seems contradictory.
• Can the authors comment on the effector repertoires encoded downstream of T7SS machinery genes ( Fig  S2)? Are these the only two Sgg T7SS arrangements of effectors found across the 38 isolates? Are different effector repertoires associated with the different machinery arrangements and/or does effector repertoire/machinery arrangement correlate with strain isolation site/host?
• More information on the conservation/location/genetic arrangement of the telE variants would be helpful. Is telE always associated with the T7SS locus and separated from tipE by the same two genes? Is telE always preceded by the same two Lap genes? Of the strains encoding more than one TelE variant (in Fig S3), are those genes encoded proximally to each other/the T7SS locus, or are they orphaned genes encoded elsewhere in the genome?
• Of the TelE point mutant-expressing strains that did not confer toxicity in Fig 3E, were those plasmids sequenced following toxin induction to confirm that an additional suppressor mutation is not responsible for the observed loss of toxicity?
• Lines 88-89; 163-164: I am not aware of a role for some of these toxins in antibacterial activity (TspA, EsxX). LXG toxins have also been implicated in virulence.
• The mention of the "main" ATPase in line 137 sounds as if UCN34 encodes a second essC. Is this the case?
• Line 139 should indicate differential abundance in supernatant (rather than expression).
• Line 178: What was the identity cut-off used to identify TelE variants? As TelE4 exhibits very low homology to other TelE proteins ( Fig S4; solely in the glycine motif) it is a less convincing TelE homolog.
• Is TelE1 the TelE that is used for experiments in Fig 3E-Fig5?
• The protein gel in Fig 5C should be repeated as the inversion of well order between TelE and TelE-His is not ideal, it appears that two lanes may have run together in the anti-HA blot, and in the anti-His Western blot shows very minimal eluted TelE-His (especially compared to the input) as well as minimal captured TipE. These data would be strengthened if the reciprocal co-IP experiment was also performed, using Histagged TipE as bait.
• Is tipE encoded by any non-pathogenic S. gallolyticus strains (to prevent from killing by related subspecies?) • Line 298. As several T7SS toxins have cytosolic immunity proteins, I would not categorize TipE as an "atypical" or "non-canonical" T7SS immunity protein solely because it does not encode transmembrane domains. It seems more unusual based on previous literature that the immunity factor gene would be encoded so far downstream of the toxin gene. Can the authors comment in the Discussion on other examples of this kind of genetic toxin-immunity factor arrangement in the literature?

Minor comments on tables, figures, and legends:
• Some information may be remnant from a prior version of this manuscript and is no longer relevant to the current manuscript (Lines 380-382, antibiotics listed; Table S1 and lines 417-420, S. agalactiae/E. faecalis strains listed that are not used in the current manuscript).
• Fig S1B is missing panel A-C designations as mentioned in the legend. What identity cut-off was used to designate a "homolog" for T7SS machinery proteins in Fig S1b-panel B? It is not clear whether the "highly similar" indication in blue refers to the T7SS locus in panel C. Please indicate more clearly in the figure that esxA is not encoded directly adjacent to essA in the B196 genome.
• Figure S2: it would be helpful to label the UCN34 T7SS genes that encode the identified secreted proteins. Light pink and bright blue arrows in Figure S2 are not identified by the key. • Figure S3: Proteins are annotated as TelD. In Suppl Table 1, some genes are referred to as tipD.
• Figure  • Line 508: What is/which experiments utilized E. coli GM48? If used, it should be added to Table S1.
• Line 957: In Figure 5 legend, the AlphaFold model should be labeled as panel (B).

Summary:
Manuscript by Wooi Keong The et al., titled "Characterization of TelE, a T7SS LXG effector exhibiting a conserved C-terminal glycine zipper motif required for toxicity" investigates Type 7b Secretion System (T7SSb) loci encoded by the Streptococcus gallolyticus subsp. gallolyticus (SGG). Authors characterize the genetic organization of these loci and find SGG strain UCN34 to secrete 6 substrates via T7SSb. One of these secreted proteins, hereby named TelE, is found to be toxic when overexpressed in Escherichia coli. TelE is reminiscent of glycine zipper pore forming proteins and contains conserved glycine rich motif required for TelE toxic activity in E. coli. Toxicity of TelE was found to be partially alleviated by TipE that serves as an antitoxin.
This work is of interest to people studying Streptococcus gallolyticus or T7SS as well as to the broader audience studying bacterial physiology, pathogenesis and interbacterial antagonism.
Overall, this is a well written manuscript that presents compelling data characterizing a novel type of T7SSb-secreted polymorphic pore-forming toxin. The experimental methods and approaches are appropriate and support authors' conclusions.
Page 10, line 222: I believe it should be " Fig  Below we provide the predic ons using AlphaFold coloured based on the pLDDT score (pLDDT =predicted local distance difference test), a reliability per residue measure (blue for high confidence, red for low confidence).

Are UCN34 "Laps" homologous to S. intermedius LapC1 and LapC2?
LapC1 and LapC2 do share a certain degree of similari es with Gallo_1576 (45%) and Gallo_1575 (53%), respec vely. However, with the renaming of TelC2 back to Gallo_1574, we will not call these proteins as Lap homologs.

More discussion of relevant T7SSb literature is needed. Another group recently published on Sgg T7SS effectors encoded within a pathogenicity locus in strain TX20005 (Taylor/Xu et al; bioRxiv preprint, April 2022; now published in Scien fic Reports; April 2023), but this study is not men oned in the present manuscript. The authors should discuss this work as it is only the second published study on the Sgg T7SS to date. Are the T7SS proteins iden fied in TX20005 (SparG -SparL) homologous to any T7SS proteins in UCN34?
We agree and have now added a few lines about this paper in the discussion (lines 382-388). SparG-SparL are 99% similar at nucleo de level to the region annotated as Gallo_1066 to Gallo_1072.

Addi onal acknowledgement of previous T7SSb findings would also be helpful throughout the manuscript. The below concepts are related to this work and should be men oned/discussed. • Lines 131-133 indicate that puta ve T7SSb effectors differ across Sgg strains but, aside from the diagram in Fig S2, this is not elaborated on. Extensive diversity in effector repertoires within a given species has been shown in other T7SSb systems: Listeria monocytogenes (Bowran/Palmer, Microbiology, 2020), Staph aureus (Warne et al, BMC Genomics, 2016), Staph lugdunensis (Lebeurre et al, Fron ers in Microbiology, 2019), and Strep agalac ae (Spencer, Plos Path 2021).
We have now added references to previous T7SSb findings in the discussion (lines 301-303).
Can the authors provide more detail on diversity of Sgg T7SS effector repertoires/subtypes? It would be helpful to compare this to the previously observed intra-species T7SSb diversity. The diversity of the T7SSb machinery and repertoire will be detailed in a manuscript currently under prepara on describing the genomic comparison of these 40 SGG clinical isolates.
Agreed. We have added these bibliographical references in the discussion (lines 306-309).
• Similar to detec on of secreted Gallo_0559 and _0560 in this manuscript, a recent preprint indicates that S. aureus WXG100/WXG-like EsxBCD proteins interact with T7SS nuclease EsaD and are secreted (Yang/Palmer et al, bioRxiv, 2023 Sequences are uploaded to NCBI under the accession number PRJNA762634. This informa on is added in the method sec on.

• Sta s cal analyses were not performed in this manuscript. This may not be needed for E. coli intoxica on assays but would be useful for microscopy data in Figures 4 and 5.
The result of microscopy data quan fied in Fig. 5 is so clearcut that we do not think that sta s cal analyses is necessary.
• Introduc on: The transi on from S. galloly cus phylogeny to T7SS in the opening paragraphs of the introduc on would be clearer if compara ve genomics were introduced sooner. Can the authors clarify why comple on of the UCN34 genome provided "clear insights for its adapta on to the rumen" in lines 62-63? SGG genome was found to contain a very high number of enzymes involved in the degrada on of complex carbohydrates found in plants but also can detoxify the plants tannins in the rumen of herbivores. We have added a sentence in the introduc on.
Lines 70-71: SGP is less pathogenic but is common in neonatal meningi s, which seems contradictory. You are right. We have changed this in the revised version (line 77).
• Can the authors comment on the effector repertoires encoded downstream of T7SS machinery genes (Fig S2)? Are these the only two Sgg T7SS arrangements of effectors found across the 38 isolates? Are different effector repertoires associated with the different machinery arrangements and/or does effector repertoire/machinery arrangement correlate with strain isola on site/host? All the SGG clinical isolates were isolated from the blood of pa ents. The T7SS locus of each clinical isolate will be discussed in detail in a future paper on genomic comparisons of these isolates.
• More informa on on the conserva on/loca on/gene c arrangement of the telE variants would be helpful. Is telE always associated with the T7SS locus and separated from pE by the same two genes? Is telE always preceded by the same two Lap genes? Of the strains encoding more than one TelE variant (in Fig S3), are those genes encoded proximally to each other/the T7SS locus, or are they orphaned genes encoded elsewhere in the genome?
The genomes of the clinical isolates were only sequenced and assembled as dra genomes with mul ple con gs. Whereas most of the genomes carry the full region of Gallo0560 (genes preceding Gallo0562 TelE) to Gallo0565 (TipE), this region was not assembled into one single con g in some genomes. Therefore, we are unable to draw conclusion on the conserva on of the region across all the isolates we sequenced. In the same vein, we are unable to comment on the genomes carrying mul ple TelE homologs.
• Of the TelE point mutant-expressing strains that did not confer toxicity in Fig 3E, were those plasmids sequenced following toxin induc on to confirm that an addi onal suppressor muta on is not responsible for the observed loss of toxicity? No, the plasmids were not sequenced following toxin induc on. However, the experiment shown in Fig3E was repeated on separate days, using cultures prepared from single colonies. Plasmids isolated from the culture was sequenced to verify that only the desired point muta on was observed. It is unlikely that the observed loss of toxicity is due to suppressor muta on.
• Lines 88-89; 163-164: I am not aware of a role for some of these toxins in an bacterial ac vity (TspA, EsxX). LXG toxins have also been implicated in virulence.
Membrane depolarizing and intraspecies compe on ac vity of TspA had been demonstrated. ExsX is indeed implicated in virulence. The sentence is corrected for clarity.
• The men on of the "main" ATPase in line 137 sounds as if UCN34 encodes a second essC. Is this the case?
No there is only one copy of essC in SGG UCN34.

• Line 139 should indicate differen al abundance in supernatant (rather than expression).
Ok we have changed this.
• Line 178: What was the iden ty cut-off used to iden fy TelE variants? As TelE4 exhibits very low homology to other TelE proteins (Fig S4; solely in the glycine mo f) it is a less convincing TelE homolog.
All TelE homologs were iden fied using NCBI tblastn. Except TelE4, all TelE homologs share at least 95% similari es. TelE4 was iden fied based on similarity (55%) at the C-terminal with significant hit (e-value of 1.64e-11). Further analysis iden fied that TelE4 contains LXG domain and glycine zipper mo f. Since glycine zipper mo f was iden fied to be essen al of TelE's func on, we therefore included TelE4 as TelE homolog.
• Is TelE1 the TelE that is used for experiments in Fig 3E-Fig5? Yes

• The protein gel in Fig 5C should be repeated as the inversion of well order between TelE and TelE-His is not ideal, it appears that two lanes may have run together in the an -HA blot, and in the an -His
Western blot shows very minimal eluted TelE-His (especially compared to the input) as well as minimal captured TipE. These data would be strengthened if the reciprocal co-IP experiment was also performed, using Histagged TipE as bait.
We are sorry but the first author of this paper is no longer in the laboratory, and we are not able to repeat this experiment. OK thank you for your vigilance.
• Fig S1B is missing panel A-C designa ons as men oned in the legend. What iden ty cut-off was used to designate a "homolog" for T7SS machinery proteins in Fig S1b- Thank you. We have added A-C labels in Fig. S1B and have marked with two lines that esxA is not adjacent to essA.
• Figure S2: it would be helpful to label the UCN34 T7SS genes that encode the iden fied secreted proteins. Light pink and bright blue arrows in Figure S2 are not iden fied by the key.
We have omi ed non-essen al informa on in the legend. • Figure 4A: What do the pink boxes indicate? Pink boxes indicate low complexity region. We have added this informa on in the legend.
• The phylogene c tree in Fig S3 appears to be redundant with that shown in Fig S1. The Fig S3 legend should refer to C-terminal TelE sequences in Figure 3C instead of Fig 3B. We prefer to keep this as it stands. We have corrected the typo Fig. S3 legend and in the Fig. S3 replace TelD by TelE.
• Figure S3: Proteins are annotated as TelD. In Suppl Table 1, some genes are referred to as pD. Thank you. It has been corrected.
• Figure S5A does not currently show E. coli toxicity due to TelE-sfGFP expression as stated in lines 221-222. Correct we meant Fig. 4D the right part of the figure.
I do not see the current Fig S5A (OD600) referenced in the text at this me. Correct this has been added in the revised version.
• Is line 274 meant to reference Fig 5A? In Fig 5A, Table S1. It was used as host for expression of pTCV-TelEgfp or TelEG470V-gfp. We have add it in Table S1.
• Line 957: In Figure 5 legend, the AlphaFold model should be labeled as panel (B). Done. Thank you for your vigilance. Overall, this is a well wri en manuscript that presents compelling data characterizing a novel type of T7SSb-secreted polymorphic pore-forming toxin. The experimental methods and approaches are appropriate and support authors' conclusions.
T7SSb is the correct abbrevia on, to dis nguish it from the original one discovered in Mycobacteria, abbreviated as T7SSa.
Page 10, line 222: I believe it should be " Fig. S5C". We do not think so. The text correctly refers to Fig. 4D.
Page 11, line 229: Protein degrada on of the TelE por on of the fusion protein could s ll occur?
Maybe it can occur at the amino-terminal part. However, the func onal cataly c domain of TelE is located at the distal C-terminal part fused to GFP and that muta on of C-terminal Glycine at posi on 470 into Valine did not alter GFP fluorescence.
Page 37, line 957: In the Figure 5 legend, "(B)" is mislabeled as "(C)" Thank you. We have changed that. The HA input signal is detected as a very intense signal on the very le part of the panel 5C and thus I am not sure to understand the ques on. The other input lanes correspond to untagged TelE or 6Histagged TelE and thus HA is not detected.