Cytoplasmic contractile injection systems mediate cell death in Streptomyces

Contractile injection systems (CIS) are bacteriophage tail-like structures that mediate bacterial cell–cell interactions. While CIS are highly abundant across diverse bacterial phyla, representative gene clusters in Gram-positive organisms remain poorly studied. Here we characterize a CIS in the Gram-positive multicellular model organism Streptomyces coelicolor and show that, in contrast to most other CIS, S. coelicolor CIS (CISSc) mediate cell death in response to stress and impact cellular development. CISSc are expressed in the cytoplasm of vegetative hyphae and are not released into the medium. Our cryo-electron microscopy structure enabled the engineering of non-contractile and fluorescently tagged CISSc assemblies. Cryo-electron tomography showed that CISSc contraction is linked to reduced cellular integrity. Fluorescence light microscopy furthermore revealed that functional CISSc mediate cell death upon encountering different types of stress. The absence of functional CISSc had an impact on hyphal differentiation and secondary metabolite production. Finally, we identified three putative effector proteins, which when absent, phenocopied other CISSc mutants. Our results provide new functional insights into CIS in Gram-positive organisms and a framework for studying novel intracellular roles, including regulated cell death and life-cycle progression in multicellular bacteria.

Bacteria exist in highly competitive environments that require them to interact with a range of organisms. To respond to potential stressors, bacteria have evolved complex strategies to mediate potential antagonistic interactions 1 . One such response is the deployment of cell-puncturing nanodevices called contractile injection systems (CIS), which are large macromolecular protein machines that can translocate cytotoxic effectors into the extracellular space or directly into target cells [2][3][4][5] . In general, CIS are composed of a contractile sheath that surrounds an inner tube loaded with effectors, fitted with a baseplate complex. A conformational change in the baseplate complex triggers the contraction of the outer sheath, which leads to the propulsion of the inner tube into the target 6,7 .
Phylogenetic analyses have indicated that these CIS are conserved across diverse microbial phyla, including Gram-negative and Gram-positive bacteria, as well as archaea 8,9 . CIS are commonly classified as Type VI secretion systems (T6SS) or extracellular CIS (eCIS) on the basis of their mode of action. Anchored at the host's cytoplasmic membrane, T6SSs function via a cell-cell contact-dependent mechanism wherein the T6SS injects effectors directly into a neighbouring cell [10][11][12][13][14] . In contrast, eCIS are assembled in the bacterial cytoplasm of the donor cell and are subsequently released into the extracellular space where they can bind to the surface of a target cell, contract and puncture the cell envelope [15][16][17][18] . Recently, a third mode of action was described in multicellular Cyanobacteria 19 . This system is also assembled in the bacterial cytoplasm and then attaches to the thylakoid membrane where it potentially induces lysis of the cell upon stress, resulting in the formation of 'ghost cells', which may in turn proceed to interact with other organisms 19 .
Article https://doi.org/10.1038/s41564-023-01341-x homologues (cis1a and cis1b), as well as additional proteins of unknown function. Cis10, a PAAR-repeat containing protein, is only present in S. venezuelae. Absent from both gene clusters are the following factors: a gene encoding a conventional tail fibre protein (Afp13) 24 , which mediates eCIS binding to target cells; a typical tape measure protein (Afp14) 25 , which is involved in regulating eCIS length 17 ; and a ClpV homologue, which is implicated in recycling some T6SSs 26 .
To test whether S. coelicolor and S. venezuelae produced CIS, we purified sheath particles from crude cell lysates, followed by negative-stain electron microscopy (EM) imaging. We observed typical contracted sheath-like particles in crude extracts from wild-type (WT) S. coelicolor and S. venezuelae; no such assemblies were seen in strains carrying a deletion in the gene that encodes the sheath Cis2 (ΔCIS, Fig. 1d). Mass spectrometry analysis of the purified WT particles detected peptides from Cis1a (inner tube) and Cis2 (sheath) (Supplementary Table 1), confirming that the CIS gene clusters from Streptomyces encode CIS-like complexes. We noticed that S. coelicolor produced approximately 50 times more sheath particles compared with S. venezuelae (Extended Data Fig. 1a,b). Therefore, we focused on the characterization of CIS from S. coelicolor (CIS Sc ) in subsequent experiments.
To test whether CIS Sc displayed a mode of action similar to canonical eCIS, we investigated whether CIS Sc were released from cells into the extracellular space. Using automated western blotting, we analysed the culture supernatant and whole-cell extracts from WT and ΔCIS S. coelicolor cells that were grown for 48 h in liquid medium. Interestingly, we detected the two key CIS Sc components Cis1a (inner tube) and Cis2 (sheath) only in whole-cell lysates but not in the supernatant of cultures of the WT or the complemented ΔCIS mutant ( Fig. 1e and Extended Data Fig. 1c). These findings suggest that the entire CIS Sc assembly is retained in the cytoplasm, unlike typical T6SS (inner tube protein translocated into the medium) and unlike eCIS (full assemblies released into the medium) 10,15 . Next, to visualize the localization of CIS Sc in situ, we imaged hyphae of S. coelicolor and S. venezuelae by cryo-electron tomography (cryoET). While intact S. coelicolor hyphae could be imaged directly, S. venezuelae hyphae were too thick and had to be thinned by cryo-focused ion beam (FIB) milling before imaging. We predominantly found extended CIS that appeared to be free-floating in the cytoplasm, a behaviour that is inconsistent with a T6SS mode of action (Fig. 1f,g and Extended Data Fig. 1d,e). Taken together, these results indicate that CIS from Streptomyces may play a role in intracellular processes, which would be distinct from the previously described functions for T6SS and eCIS.

Structure, engineering and subcellular localization of CIS Sc
To obtain insights into the structural details of the CIS Sc contractile sheath-tube module, we performed single-particle cryoEM (helical reconstruction) of purified sheath particles from WT S. coelicolor, Of the hundreds of putative CIS gene clusters detected in bacteria, all well-characterized examples have come from two closely related clades and have been exclusively examined in Gram-negative bacteria. Characterized CIS representatives include 'metamorphosis-associated contractile structures' (MACs) from Pseudoalteromonas luteoviolacea 16 , the 'T6SS subtype iv' (T6SS iv ) in Candidatus Amoebophilus asiaticus 7 , 'antifeeding prophages' (AFPs) from Serratia 17 , 'Photorhabdus virulence cassettes' (PVCs) from P. asymbiotica 18 and two newly characterized CIS from the marine bacteria Algoriphagus machipongonensis 15 and cyanobacteria 19 .
Strikingly, 94 of 116 sequenced Gram-positive actinomycetes of the genus Streptomyces were shown to encode a potential CIS gene cluster 8,9 . A previous report suggested that CIS from Streptomyces lividans were involved in microbial competition; however, the mechanism remains unknown 20 . Streptomyces species are multicellular soil bacteria, renowned for their complex developmental life cycle and their ability to produce an array of clinically relevant secondary metabolites 21 . The Streptomyces life cycle begins with the germination of a spore and generation of germ tubes, which grow by apical tip extension and hyphal branching to form a dense vegetative mycelium. Upon nutrient depletion, non-branching aerial hyphae are erected, which synchronously divide into chains of uni-nucleoid spores 22 . Notably, the production of important secondary metabolites is tightly coordinated with the developmental life cycle 21 .
Here we provide evidence that CIS from the model organism Streptomyces coelicolor (CIS Sc ) function intracellularly and belong to a new class of CIS that exist as free-floating fully assembled particles in the cytoplasm and mediate cell death in response to stress conditions. Additionally, we find that the absence of CIS affects the coordinated cellular development and secondary metabolite production of S. coelicolor, indicating a wider role of CIS from Streptomyces in the multicellular biology of these important bacteria. Our data are consistent with a recent paper 23 .

Streptomyces express cytoplasmic CIS in vegetative growth
Previous bioinformatic studies revealed that the majority of sequenced Streptomyces genomes harbour a highly conserved cluster of eCIS genes related to the poorly studied CIS IId subtype 8,9 . This was further confirmed by our phylogenetic analyses using sheath protein sequences from known producers of CIS and from two representative Streptomyces species, namely S. coelicolor and Streptomyces venezuelae (Fig. 1a).
Closer inspection of the Streptomyces CIS gene clusters from S. coelicolor (sco4242-sco4260) and S. venezuelae (vnz_28865-vnz_28935) suggested that both species encode 10 or 11 core structural components of the phage tail-like systems, respectively 8,9 (Fig. 1b,c). On the basis of this sequence similarity, we renamed the genes from Streptomyces to cis1-16. Both CIS gene clusters encode two inner tube   19,67 . Asterisks indicate gene products that were detected by mass spectrometry after CIS purification (Supplementary Table 1). c, The schematic illustrates a putative CIS assembly from Streptomyces. Colour code is based on the predicted gene function shown in b. d, cis2 is required for CIS assembly. Shown are negative-stain EM images of crude sheath preparations from WT and ΔCIS mutant strains of S. coelicolor and S. venezuelae. White arrowheads indicate contracted sheath-like structures. Shown are representative micrographs of three independent preparations. Scale bars, 80 nm. e, CIS Sc proteins are detected in the cell lysate but not secreted into the supernatant. Shown is the automated western blot analysis of cultures of WT S. coelicolor, ΔCIS mutant and a complementation (ΔCIS/CIS + ). The presence of the sheath protein (Cis2) and the inner tube protein (Cis1a) in whole-cell lysates and concentrated culture supernatants was probed using polyclonal antibodies against Cis1a/2. Experiments were performed in biological triplicates. For the control sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel, see Extended Data Fig. 1c. f, Cryo-electron tomogram of a WT S. coelicolor hypha revealing two cytoplasmic extended CIS Sc assemblies (arrowhead). Eighty tomograms were collected from 10 independent datasets. PG, peptidoglycan; CM, cytoplasmic membrane; CP, cytoplasm. Putative structural components are indicated on the right. Scale bars, 75 nm and 12.5 nm (magnified inset). g, Tomogram of a cryoFIB-milled WT S. venezuelae hypha revealing one cytoplasmic extended CIS Sv assembly (arrowhead). Free-floating extended CIS Sv were observed in 12 tomograms from 3 independent datasets. Scale bar, 75 nm. Article https://doi.org/10.1038/s41564-023-01341-x which had a homogeneous length for the contracted sheath of ~140 nm (Fig. 2a,b and Supplementary Table 2). The resulting map reached a resolution of 3.6 Å ( Fig. 2c and Extended Data Fig. 2a,b). Contracted sheath proteins adopt a right-handed helical array with an inner diameter of 115 Å and an outer diameter of 233 Å (Fig. 2b). Similar to the recently described sheath structures observed in AlgoCIS 15 Fig. 2c).
From the resulting map, it was possible to build de novo domains 1 and 2, which contribute to the sheath wall (Fig. 2d). The additional domain 3, which is located on the sheath surface, seems to be highly flexible and could not be resolved. The overall contracted structure of Cis2 is similar to sheaths of previously characterized systems 18,27,28 .
To be able to purify the extended form of the CIS Sc sheath-tube module from S. coelicolor cell lysates, we engineered non-contractile CIS Sc , on the basis of the information from the contracted Cis2 structure and similar approaches used previously for Vibrio cholerae 29 and enteroaggregative Escherichia coli 30 (Extended Data Fig. 3a). Different sets of two (IE), three (IEG) and five (IEGVG) amino acid residues were inserted into the N-terminal linker of Cis2 after position G25, resulting in the mutants CIS-N2, CIS-N3 and CIS-N5, respectively. For the CIS-N2 and CIS-N3 mutants, less than 30% and 50% were found in extended form, respectively (Extended Data Fig. 3b,c). For the CIS-N5 non-contractile mutant, more than 95% of the complexes were seen in the extended conformation (Extended Data Fig. 3d). In vitro, the length of the CIS-N5 non-contractile mutant was homogeneous at ~230 nm (Fig. 2e). Moreover, mass spectrometry analyses confirmed the presence of most CIS Sc components, indicating the stability of the complex ( Fig. 1b and Supplementary Table 1).
Next, we optimized the purification of CIS-N5 particles and performed cryo-EM. Helical reconstruction was used to generate an EM map, which we then used to build de novo the sheath-tube (Cis2-Cis1a) module in the extended conformation at 3.9 Å resolution ( Fig. 2f- Table 2). Domain 3 of the extended sheath (Cis2) was again too flexible to be resolved. The tube (Cis1a) structure and fold are highly similar to the tube structures already described for other CIS ( Fig. 2h and Extended Data Fig. 3g). The comparison of the sheath (domains 1/2) in the extended versus contracted states revealed an increase in diameter and shortening of the length upon contraction, similar to other CIS (Fig. 2b,f) 15,18,19,28,31 . The introduction of five extra amino acids in the CIS-N5 mutant induced conformational changes at the Cis2 N terminus, similar to a previous report (Extended Data Fig. 3h,i) 31 .
Guided by the high-resolution structure of the sheath module, we engineered a fluorescently tagged CIS Sc by inserting YPet at position I274 in the Cis2 monomer. Subsequently, we used this Cis2-YPet sandwich fusion to complement the S. coelicolor Δcis2 mutant in trans (Extended Data Fig. 4a). Using negative-stain EM and cryoET, we confirmed that YPet-tagged CIS Sc (CIS Sc -YPet) were able to assemble into extended particles and to contract, suggesting that these fluorescently labelled CIS Sc particles were functional (Extended Data Fig. 4b,c). This construct enabled us to visualize the subcellular localization of CIS Sc in vegetative hyphae using time-lapse fluorescence light microscopy (fLM). Multiple CIS Sc -YPet foci were found inside the hyphae but not in extracellular space. The foci were largely static or displayed short-range movements within the hyphae (Fig. 2i,g and Supplementary Video 1). CIS Sc -YPet foci were stable over the course of the experiment and did not reveal notable changes in the shape or intensity of the fluorescence. While this invariability indicates the absence of firing events during the experiment, the resolution in fLM and the relatively short length of the CIS Sc may hamper the detection of firing events (in contrast to the much longer T6SSs 10,30 ).
Taken together, our structural data allowed us to engineer noncontractile and fluorescently tagged CIS Sc , which revealed the presence of scattered CIS Sc in S. coelicolor hyphae.

CIS contraction correlates with state of cellular integrity
Our initial cryoET data of S. coelicolor cells indicated that contracted CIS Sc were frequently found in hyphae that displayed a damaged cell membrane. To explore this correlation further, we first acquired low-magnification two-dimensional (2D) cryoEM images. On the basis of the contrast of individual hyphae in these 2D images (Fig. 3a), we classified the hyphae into three distinct groups: (1) 'intact hyphae' (dark appearance in 2D) with mostly intact cytoplasmic membrane and occasional vesicular membranous assemblies that are reminiscent of 'cross-membranes' 32   and subcellular localization of CIS Sc . a, Representative cryo-electron micrograph of a sheath preparation from WT S. coelicolor that was recorded for structure determination (505 micrographs were collected). All sheath structures were seen in the contracted state. Scale bar, 40 nm. b, Section of the CIS Sc sheath cryoEM structure in the contracted conformation. Left: side view. Right: top view. c,d, Side view ribbon representation of the Cis2 monomer in its contracted state superposed with the corresponding cryoEM map (c) and with dashed rectangles highlighting the positions of domains 1 (red), 2 (blue) and 3 (orange, not resolved because of high flexibility) (d). e, Representative cryo-electron micrograph of a sheath preparation from S. coelicolor expressing a non-contractile mutant of Cis2 (CIS-N5) (525 micrographs were collected). More than 95% of all structures were seen in the extended state. Scale bar, 40 nm. f, Ribbon representation of a section of the S. coelicolor Cis2 (sheath)-Cis1a (inner tube) cryoEM structure in the extended conformation that was solved using the non-contractile mutant. Left: side view. Right: top view. g,h, Side view ribbon representation of the Cis1a-Cis2 monomer (non-contractile mutant) in its extended state superposed with the corresponding cryoEM map (g) and with dashed rectangles highlighting the positions of domains 1 (red), 2 (blue) and 3 (orange, not resolved because of high flexibility) (h). i, Insights from the cryoEM structures enabled us to tag Cis2 with a fluorescent tag (YPet) for subsequent time-lapse imaging to determine the localization of assembled CIS Sc . Shown is a still image from Supplementary Video 1 showing scattered fluorescent foci inside vegetative hyphae. White rectangle highlights hypha shown in h. Scale bar, 10 µm. j, Fluorescently tagged CIS Sc remained largely static or showed short-range movements over time. Shown is an image montage of a representative growing S. coelicolor hypha from Supplementary Video 1. Note that the depicted hypha has been rotated. Images were acquired every 5 min. Scale bar, 10 µm.  As we observed previously for intact hyphae (Fig. 1f), individual CIS Sc particles were always found in the extended conformation and localized in the cytoplasm (Fig. 3b). In contrast, in partially lysed hyphae (Fig. 3c), the ratio of extended to contracted CIS Sc was 2:1. CIS Sc particles often appeared to cluster in the vicinity of membranous structures (Fig. 3d). Notably, we found that in some cases, the extended CIS Sc aligned perpendicular to membrane patches or vesicles with the baseplate complex facing the membrane, suggesting that CIS Sc may interact with the cytoplasmic membrane (Fig. 3c,d). In contrast, ghost cells only displayed CIS Sc particles in the contracted state, which were often clustered together (Fig. 3f).
Collectively, these results indicate that the conformational state of CIS Sc correlates with the integrity of the cell and that CIS Sc may play an intracellular role as a consequence of an unknown cellular signal and lead to cell death, either directly or indirectly. Consequently, we hypothesized that such a signal may be triggered by exogenous

CIS contraction mediates cell death under stress conditions
To test this hypothesis, we explored whether upon encountering stress, the presence of CIS Sc and their contraction could mediate cell death. To generate a marker for cell viability, we inserted sfgfp under the control of a constitutive promoter in trans in WT S. coelicolor, in the null mutant (ΔCIS) and in the non-contractile mutant (CIS-N5). To label intact and partially lysed hyphae, cells were incubated with the fluorescent membrane dye FM5-95. We first used correlated cryo-light and electron microscopy (CLEM) to confirm that the detected cytoplasmic and membrane fluorescence correlated with the physiological state of the hyphae (Extended Data Fig. 5). To assess the level of cell death in the imaged strains, we used fLM and quantified the ratio of the sfGFP signal (indicator of viable hyphae) to the FM5-95 signal (indicator of intact and lysed hyphae) in the different strains. Cells were grown for 48 h in liquid, a timepoint at which CIS Sc can be detected in hyphae (Fig. 1d,e).
During non-stress conditions, the WT, ΔCIS and CIS-N5 mutant strains displayed a similar sfGFP/FM5-95 ratio, indicating no important difference in viability between strains (Fig. 4a,c). In parallel, we challenged the same S. coelicolor strains with a sub-lethal concentration of the bacteriocin nisin (1 µg ml −1 ) for 90 min, which causes the formation of membrane pores and eventually disrupts the integrity of the cell envelope 33 . In the WT, we found that ~50% of the analysed hyphae displayed signs of cell death (Fig. 4b,d). Strikingly, in the CIS-deficient strain and the non-contractile CIS Sc mutant, nisin-stressed cells showed higher viability than WT cells (Fig. 4b,d).
To investigate whether other stress factors could induce cell death, we also challenged S. coelicolor with the membrane depolarizing agent carbonyl cyanide 3-chlorophenylhydrazone (CCCP) and with UV radiation to induce DNA damage (Extended Data Fig. 6a,b). In line with our previous results, the treatment of vegetative hyphae with both CCCP and UV radiation led to an increase in cell death by 25% in the WT. Treatment of the WT, the ΔCIS or the CIS-N5 mutant strain with 0.002% DMSO, which was used as a mock control in the CCCP stress experiments, did not impact hyphal viability (Fig. 4e,f). In parallel, we also purified CIS Sc from crude cell extracts obtained from non-stressed and stressed samples that were used for fLM imaging. By negative-stain EM imaging, we confirmed the presence of CIS Sc particles in hyphae of the WT and in the CIS-N5 mutant strain, and the absence of sheath particles in the ΔCIS mutant (Extended Data Fig. 7). The abundance of CIS Sc in non-stressed and stressed samples was comparable, which was also confirmed by the level of Cis1a/2 proteins in untreated and nisin-treated hyphae (Extended Data Fig. 8a,b). Taken together, these results indicate that CIS Sc contraction mediates cell death under exogenous stress conditions.

CIS contribute to multicellular development
Earlier studies indicated that the expression of the S. coelicolor CIS gene cluster is coordinated with the Streptomyces life cycle 34 . To follow the expression of CIS Sc during the developmental life cycle, we constructed a fluorescent reporter strain in which expression of ypet was driven by the cis2 promoter (P cis2 -ypet). Since S. coelicolor only completes its spore-to-spore life cycle when grown on solid medium, glass coverslips were inserted at a 45° angle into agar plates inoculated with spores. Coverslips with attached S. coelicolor hyphae were removed and imaged every 24 h for 4 d by fLM. Fluorescent signal indicated that the cis2 promoter was primarily active in vegetative hyphae at the 48-h timepoint (Fig. 5a). In parallel, we determined CIS Sc protein levels in surface-grown WT S. coelicolor over the life cycle via western blot analysis. Consistent with our fluorescence reporter experiment, Cis1a/2 levels were highest in vegetative mycelium that was collected after 30 h and 48 h of incubation (Extended Data Fig. 9a). These results are also in agreement with published transcriptomics data from S. venezuelae showing the specific induction of the CIS Sv gene cluster during vegetative growth (Extended Data Fig. 9b) 35 .
Since a previous study on S. lividans reported a putative role of CIS in interspecies interactions 20 , we performed a series of growth competition assays but did not observe differences in fitness between the WT and CIS Sc mutants (Supplementary Table 3). We then tested whether the expression of a functional CIS Sc influenced the timely progression of the S. coelicolor life cycle, using WT, ΔCIS, CIS-N5 and a complemented strain (∆CIS/CIS + ). First, we detected sporulating hyphae and spores by imaging surface imprints of plate-grown (R2YE agar) colonies at different timepoints. All strains consistently completed their life cycle and synthesized spores (Fig. 5b). Importantly, in contrast to the WT and the complemented strain, both ΔCIS and CIS-N5 mutants sporulated markedly earlier (72 h vs 96 h for the WT and the complemented mutant). These results were further corroborated by quantifying the number of spores produced by the individual strains under the same experimental conditions (Fig. 5c), indicating that CIS Sc play a role in regulating the S. coelicolor vegetative growth cycle.
In addition to the accelerated cellular development in the CIS Sc mutants, we also observed that the production of the two characteristic pigmented secondary metabolites in S. coelicolor, actinorhodin (blue) 36 and undecylprodigiosin (red) 37 , was notably reduced compared to the WT and the complemented ∆CIS/CIS + mutant (Extended Data Fig. 9c). This discrepancy was further confirmed by quantifying the total amount of actinorhodin (intracellular and secreted) produced by the different strains over a period of 72 h. Both ∆CIS and CIS-N5 mutants produced approximately 70% less actinorhodin compared to the WT and the ΔCIS complementation strain (Extended Data Fig. 9d). Moreover, in contrast to the observed delay in sporulation, actinorhodin production in the CIS Sc mutants was not only delayed but also never reached WT levels in the time frame of this experiment. All three proteins include one predicted transmembrane domain. Sco4257/4258 contain a ricin-like domain on the C terminus. b, Survival of E. coli expressing Sco4256, Sco4257, Sco4258 or YFP-GypA (used as a negative control of toxicity). Images are representatives of three independent experiments. c, CryoET slice of a hypha from the S. coelicolor effector-deficient mutant (∆sco4256-4258), showing empty CIS Sc but otherwise WT-like particles (arrowhead). Tomograms (90) were collected from 3 independent datasets. Scale bars, 75 nm. d, Fractions of empty CIS Sc particles are increased in the effector mutant compared with the WT (ScoWT, n total = 91; ∆sco4256-4258, n total = 73). e, Representative images of filled and empty CIS Sc particles. Tomograms (60) were collected from 2 independent datasets. Scale bars, 50 nm. f, fLM (shown are representative images) was used to determine the ratio between live cells (cytoplasmic sfGFP) and total cells (membrane dye FM5-95) after growth in the absence of stress or in the presence of nisin stress for S. coelicolor ∆sco4256-58/sfGFP. Scale bars, 10 µm. g, The quantification of the experiments in f showed no notable difference between the ∆sco4256-58/sfGFP strain grown without stress and with nisin. Superplots show the area ratio of live to total hyphae. The three different colours (green, grey and pink) indicate the three biological repeats. Black line indicates the mean ratio derived from biological triplicate experiments (n = 100 images for each experiment). Significance was calculated using one-way ANOVA and Tukey's post-test. h, Representative brightfield images of surface imprints of plate-grown colonies of WT S. coelicolor (from Fig. 5b) and the CIS Sc effector mutant strain ∆sco4256-4258. Images were taken at the indicated timepoints. Insets show magnified regions of the colony surface containing spores and spore chains. Note that strains with functional CIS sporulate later. Scale bars, 50 µm. i, Quantification of spore production (c.f.u.) in the same strains as above, revealing much higher c.f.u.s (spores) at 72 h in the effector-deficient CIS mutant. Data show mean ± s.d. obtained from biological triplicate experiments. Altogether, we have shown that deleting or expressing non-functional CIS Sc results in important changes in the S. coelicolor life-cycle progression, which also affects secondary metabolite production.

Sco4256, Sco4257 and Sco4258 are CIS-associated effectors
Most T6SS and eCIS gene clusters encode effector proteins, which are delivered to kill or manipulate the cellular activity of the corresponding target 38 . We noticed that the CIS Sc gene cluster encodes three uncharacterized proteins Sco4256-4258. All three proteins consist of a single transmembrane domain, a highly similar N terminus and a C-terminal domain, which in the case of Sco4257/4258, consists of a predicted carbohydrate-binding moiety (RICIN-type beta trefoil) (Fig. 6a). Interestingly, we detected at least one homologue of such potential effectors in 47 out of 94 sequenced Streptomyces strains that also encode a CIS gene cluster 8,9 , while CIS-negative strains typically do not encode these potential effectors (Supplementary Table 4). To characterize these proteins, we expressed the individual genes in E. coli from an inducible promoter and assessed cell viability. Upon induction, the Sco4256-4258-producing strains showed notably reduced viability, indicating toxic activity of the proteins (Fig. 6b).
To explore whether Sco4256-4258 also played a role in CIS Scmediated cell death, we generated an S. coelicolor triple mutant (Δsco4256-4258). We first confirmed that the mutation did not affect the assembly of CIS Sc particles by negative-stain EM (Extended Data Fig. 10a) and cryoET (Fig. 6c). The general in situ structure of CIS Sc particles in the Δsco4256-4258 mutant was similar to WT, with a notable exception regarding their luminal density. While the tube lumen was mostly filled in WT CIS Sc (~97%, n total = 91), the tube lumen appeared mostly empty in mutant CIS Sc (~85%, n total = 73) (Fig. 6d,e).
Using this triple mutant, we repeated the fLM-based viability assay upon nisin stress (Fig. 6f,g). Unlike the drop in viability of the WT, we found no notable difference in the viability of the Δsco4256-4258 mutant upon nisin stress, which was consistent with the pattern observed in ΔCIS and CIS-N5 mutants (Fig. 4b,d). Finally, Δsco4256-4258 cells also sporulated markedly earlier (Fig. 6h,i) and produced about 79% less actinorhodin (Extended Data Fig. 10c), which is again similar to the previously analysed CIS Sc mutants (Fig. 5b,c and Extended Data Fig. 9c,d). Taken together, these results show that Δsco4256-4258 cells phenocopy the ΔCIS and CIS-N5 mutants, supporting the idea of Sco4256-4258 being CIS Sc -associated effectors.

Discussion
Here we show that CIS particles from Streptomyces are functionally distinct from related eCIS and T6SS. Our data from fLM imaging, cryoET and western blotting all consistently indicate that CIS Sc were assembled free floating in the cytoplasm; however, under our experimental conditions, they were neither found to be released into the medium, nor were they observed attached to the cytoplasmic membrane. This argues against a typical eCIS mode of action, and it is also inconsistent with a typical T6SS mode of action; in particular, it is difficult to conceptualize how a T6SS would fire through the thick peptidoglycan cell wall in a Gram-positive host organism. Therefore, our data points to an intracellular function, which is supported by further observations that are discussed below.
CryoET imaging revealed a notable fraction of partially or fully lysed cells in a vegetative culture. Interestingly, the degree of cell lysis strongly correlated with the presence of contracted CIS Sc assemblies. These results were further supported by fLM imaging, demonstrating that under different types of exogenous stress conditions, cell death was induced in the WT at a much higher percentage than in mutants that expressed non-functional CIS Sc . Thus, CIS Sc contraction is required for inducing cell death once a culture encounters stress.
We propose that cell lysis is mediated by cell envelope-targeting effectors that are released from the CIS Sc upon contraction; however, how are these effectors delivered? This could happen by the firing of a free-floating CIS Sc , releasing loaded effectors into the host cytoplasm. An alternative mechanism could be (transient) CIS Sc binding to the host cytoplasmic membrane, followed by contraction. Membrane binding in eCIS is typically mediated by tail fibres that are attached to the baseplate complex. Oftentimes, tail fibre genes are found just downstream of baseplate components 15,18,24 . Interestingly, we detected the conserved protein Sco4242 just downstream of the putative baseplate genes cis11 and sco4243 (Fig. 1b). Bioinformatic analyses revealed the presence of Sco4242 homologues in 79% of CIS-positive Streptomyces strains and the absence of such homologues in CIS-negative strains (Supplementary Table 4). Sco4242 consists of a predicted transmembrane segment and a conserved sugar-binding domain. Together with the absence of typical tail fibre-like proteins and structures, we speculate that Sco4242 may play a role as an adaptor that mediates binding of CIS Sc to the cytoplasmic membrane upon a stress signal, followed by firing.
In addition to mediating death of the host cell in response to stress, we showed that CIS Sc contraction also plays a role in the timely progression of the Streptomyces life cycle, as evidenced by the earlier onset of sporulation in the CIS Sc and effector mutants. Cell death has been proposed as a distinct process in the developmental programme of Streptomyces 39 . However, the underlying molecular mechanism remains unclear. We speculate that contracting CIS Sc could induce hyphal cell death, which may lead to the release of nutrients that result in the postponement of subsequent life-cycle stages, thereby impacting multicellular development. Notably, increased cell death has been reported to occur at the centre of colonies 40,41 ; these regions are thought to be limited in nutrient and/or oxygen supply, which in turn may be perceived as stress and trigger CIS Sc -mediated cell death.
Notably, the morphological differentiation of Streptomyces colonies is tightly coordinated with the production of secondary metabolites, which are often secreted into the environment where they can provide a competitive advantage 21 . We showed that CIS Sc mutants were not only affected in the timing of the onset of sporulation, but also in the production of the secondary metabolite actinorhodin. We speculate that the delay of sporulation in the WT (and the complemented strain) may be advantageous to allow the coordinated production and release of key secondary metabolites such as toxins, proteases or signalling molecules. The lack of functional CIS Sc in both mutant strains could lead to improper timing of cell cycle progression, resulting in early sporulation, which may in turn lead to lower amounts of actinorhodin production.
In conclusion, our data provide new functional insights into CIS in a Gram-positive model organism and a framework for studying new intracellular roles of CIS, including regulated cell death and life-cycle progression.

Bacterial strains, plasmids and oligonucleotides
Bacterial strains, plasmids and oligonucleotides can be found in Supplementary Tables 5 and 6. E. coli strains were cultured in LB, SOB or DNA medium. E. coli cloning strains TOP10 and DH5α were used to propagate plasmids and cosmids. E. coli strain BW25113/pIJ790 was used for recombineering cosmids 42 . For interspecies conjugation, plasmids were transformed into E. coli ET12567/pUZ8002. Where necessary, media were supplemented with antibiotics at the following concentrations: 100 µg ml −1 carbenicillin, 50 µg ml −1 apramycin, 50 µg ml −1 kanamycin and 50 µg ml −1 hygromycin.

Phylogenetic analysis
Phylogenetic analysis of the different contractile injection systems (from eCIS, T6SS, phage and CIS from Streptomyces) was conducted using the putative sheath proteins. Alignment and generation of the phylogenetic tree was performed as previously reported 7,15 . First, the amino acid sequences from 16 sheath proteins were aligned using the MUSCLE online tool 45,46 . Standard parameters were applied for multiple sequence alignment. Then, the MEGAX programme 47 was used to reconstruct phylogenetic trees using the maximum likelihood (ML) method and bootstrap values (1,000 resamples) were applied to assess the robustness of the tree.

Negative-stain electron microscopy
Purified sheath particles (4 µl) were adsorbed to glow-discharged, carbon-coated copper grids (Electron Microscopy Sciences) for 60 s, washed twice with milli-Q water and stained with 2% phosphotungstic acid for 45 s. The grids were imaged at room temperature using a Thermo Fisher Morgagni transmission electron microscope (TEM) operated at 80 kV.

Mass spectrometry analysis
To confirm the presence of predicted CIS components from Streptomyces, isolated sheath particles were subjected to liquid chromatography-mass spectrometry analysis (LC-MS/MS). First, the samples were digested with 5 µl of trypsin (100 ng µl −1 in 10 mM HCl) and microwaved for 30 min at 60 °C. The samples were then dried, dissolved in 20 µl double-distilled water (ddH 2 0) with 0.1% formic acid, diluted at 1:10 and transferred to autosampler vials for liquid chromatography with tandem mass spectrometry analysis. A total of 1 µl was injected on a nanoAcquity UPLC coupled to a Q-Exactive mass spectrometer (Thermo Fisher). Database searches were performed using the Mascot swissprot and tremble_streptomycetes search programme. For search results, stringent settings were applied in Scaffold (1% protein false discovery rate, a minimum of two peptides per protein, 0.1% peptide false discovery rate). The results were visualized using the Scaffold software (Proteome Software, v4.11.1).

Automated western blot analysis
Automated western blot analysis (WES) of liquid-grown Streptomyces strains was essentially performed as described previously 49 using the anti-rabbit secondary antibody detection module (Protein Simple DM-001). Cell pellets were resuspended in 0.4 ml sonication buffer (20 mM Tris, pH 8.0, 5 mM EDTA, 1× EDTA-free protease inhibitors (Sigma-Aldrich)) and subjected to sonication at 4.5 µm amplitude for 7 cycles of 15 s on/15 s off. Samples were centrifuged at 17,000 × g for 15 min at 4 °C. The supernatants were removed and subjected to a Bradford Assay (Biorad). Equivalent total protein concentrations (0.2 mg ml −1 ) were assayed using the automated western blotting machine WES (Protein Simple) according to the manufacturer's guidelines. For the detection of Cis1a and Cis2 proteins, antibodies for α-Cis1a (GenScript) and α-Cis2 (GenScript) were used at a concentration of 1:200. For detection of WhiA, 0.5 µg of total protein and anti-WhiA (Polyclonal) at 1:100 dilution was used 50 .
For the detection of Cis1a and Cis2 in culture supernatants, WT S. coelicolor, SS387 and SS395 were grown in duplicate in TSB medium for 48 h. Cultures were pelleted and 20 ml supernatants obtained from each culture were concentrated to approximately 1 ml using Amicon Ultra-15 10K spin column (Millipore). Total protein samples were further processed as described above. In parallel, an aliquot of each sample was loaded onto a 12% Teo-Tricine/SDS precast protein gel (Expedian) to demonstrate the presence of proteins in the culture supernatants. SDS gels were stained with InstantBlue (Sigma-Aldrich) and scanned.
For the automated western blot analysis of surface-grown S. coelicolor samples from R2YE plates, mycelium was scraped off sterile cellophane discs that had been placed on top of solid R2YE medium. Mycelia were removed at the described timepoints and washed with 1X PBS. The supernatant was discarded and the pellet frozen. Pellets were treated and WES run as above. All virtual western blots were generated using the Compass software for simple western (v6.0.0). Data on protein abundance were plotted using GraphPad Prism (v9.3.1).
For WES analyses of Cis1a and Cis2 abundance following nisin stress, WT S. coelicolor cultures were grown in TSB medium at 30 °C for 48 h, after which they were split and normalized to the same optical density. To one culture replicate, nisin was added to a final concentration of 1 µg ml −1 and to the other, the diluent (0.05% acetic acid) was added in equal volume. Next, 2 ml aliquots were removed from each sample and pelleted at 17,000 × g. Pellets were treated as above but were additionally probed with an α-WhiA antibody at 1:100 concentration. The band intensities for Cis1a and Cis2 were normalized against the band intensity of WhiA and plotted in GraphPad Prism (v9.3.1) with the mean ± s.d.

Fluorescence light microscopy and image analysis
For imaging protein localization and fluorescent promoter reporter fusion in S. coelicolor, a Zeiss Axio Observer Z.1 inverted epifluorescence microscope fitted with an sCMOS camera (Hamamatsu Orca FLASH 4), a Zeiss Colibri 7LED light source, a Hamamatsu Orca Flash 4.0v3 sCMOS camera and a temperature-controlled incubation chamber was used. Images were acquired using a Zeiss Alpha Plan-Apo ×100/1.46 Oil DIC M27 objective with YFP excitation/emission bandwidths of 489-512 nm/520-550 nm. Still images and time-lapse image series were collected using Zen Blue (Zeiss) and analysed using Fiji 51 .
To monitor the activity of the fluorescent sheath promoter fusion in S. coelicolor, strain SS484 spores were spotted onto solid R2YE medium and grown alongside a microscope coverslip that had been inserted into the agar at approximately a 45° angle. Plates were Article https://doi.org/10.1038/s41564-023-01341-x incubated at 30 °C for up to 4 d. At the indicated timepoints, glass coverslips with attached hyphae were removed, mounted onto slides affixed with 1% agar pads and imaged.
For time-lapse imaging of S. coelicolor expressing a fluorescently labelled sheath protein (SS389), cells were first grown in TSB-YEME for 40 h and a 2 µl sample of the culture was immobilized on a 1% agarose pad prepared with filtered culture medium and using a gene frame (Thermo Fisher). Experiments were performed at 30 °C and growing hyphae were imaged every 5 min. Image collection and analysis were performed using Zen Blue (Zeiss) and Fiji, respectively 51 .

Plunge freezing of Streptomyces hyphae
For cryoET, Streptomyces cells were mixed with 10 nm Protein A conjugated colloidal gold particles (1:10 v/v, Cytodiagnostics) and 4 µl of the mixture was applied to a glow-discharged holey-carbon copper EM grid (R2/1 or R2/2, Quantifoil). The grid was automatically blotted from the backside for 4-6 s in a Mark IV Vitrobot using a Teflon sheet on the front pad, and plunge-frozen in a liquid ethane-propane mixture (37%/63%) cooled by a liquid nitrogen bath.
For single-particle cryoEM (SPA), the S. coelicolor CIS particles (from WT CIS and non-contractile CIS) collected after sheath preparation were vitrified using a Vitrobot Mark IV (Thermo Fisher). Samples (4 µl) were applied on glow-discharged 200 mesh Quantifoil gold grids (R2/2). Grids were blotted for 5 s and plunged into liquid ethane-propane mix (37%/63%). Frozen grids were stored in liquid nitrogen until loading into the microscope.

Cryo-focused ion beam milling
A standard protocol was used to perform cryo-focused ion beam milling (CryoFIB milling) of S. venezuelae 52 . Plunge-frozen grids were clipped into cryoFIB-autoloader grids (Thermo Fisher), then transferred into a liquid nitrogen bath of a loading station (Leica Microsystems) and mounted into a 40° pre-tilted scanning electron microscope (SEM) grid holder (Leica Microsystems). The holder was transferred with a VCT100 cryo-transfer system (Leica Microsystems) into a Helios NanoLab600i dual beam FIB/SEM (Thermo Fisher). Grids were coated with platinum precursor gas for 6 s and checked with the SEM at 3-5 kV (80 pA) to evaluate grid quality and identify targets. Lamellae were milled in multiple steps using the focused gallium ion beam (43 nA to 24 pA) until a thickness of ~250 nm was achieved. The holder was returned to the loading station using the VCT100 transfer system. Unloaded grids were stored in liquid nitrogen before cryoET imaging.

CryoET
Intact or cryoFIB-milled Streptomyces cells were imaged by cryoET 53 . Images were recorded on Titan Krios 300 kV microscopes (Thermo Fisher) equipped with a Quantum LS imaging filter operated at a 20 eV slit width and with K2 or K3 Summit direct electron detectors (Gatan). Tilt series were collected using a bidirectional tilt scheme from −60 to +60° in 2° increments. Total dose was 130-150 e − Å −2 and defocus was kept at −8 µm. Tilt series were acquired using SerialEM 54 , drift-corrected using alignframes, reconstructed and segmented using the IMOD programme suite 55 . To enhance contrast, tomograms were deconvolved with a Wiener-like filter 'tom_deconv' 56 .

SPA data collection and image processing
CryoEM datasets of S. coelicolor contracted sheath and extended sheath-tube module were collected as movie stacks using the Seri-alEM programme on the Titan Krios EM operating at 300 kV and equipped with an energy filter and a K2 Summit camera. The movie frames of each collected stack were aligned and summed up into one single micrograph with dose weighting at the binning factor of 2 using MotionCor2 (ref. 57 ). The contrast transfer function (CTF) parameters of the micrographs were estimated using Gctf. Pixel size at specimen level was 1.4 Å and target defocus ranged from −1.5 µm to −3.5 µm.
Each stack contains 50 frames, and the accumulated electron dose rate was ~60 e − Å −2 .
The image processing of contracted sheath and extended sheath-tube from S. coelicolor was performed as previously reported 15 . The particles were picked manually using Relion 3.0 58 . The particle extraction was performed in 'Extract helical segments' mode to extract helical segments. The structural determination of the contracted sheath and the extended sheath-tube module was performed using helical reconstruction in Relion 3.0 59 .
For the extended sheath-tube module, the final 3.9 Å resolution structure of the extended sheath-tube module was determined from 18,822 particles calculated with 6-fold symmetry and helical parameters (rise = 38.50 Å, twist = 23.10°) (Extended Data Fig. 3e).
The resolutions of relative reconstruction maps were estimated on the basis of the gold-standard Fourier shell correlation (FSC) = 0.143 criterion 60 . The local resolution estimations of individual maps were performed using the local resolution module in Relion 3.0 and examined using UCSF Chimera 61 (Extended Data Figs. 2b and 3f).

Structure modelling
Proteins were built de novo using COOT 62 . Models were iteratively refined using RosettaCM 63 and real-space refinement implemented in PHENIX 64 . Sheath protein could only be partially modelled and in some cases side chains were not assigned. Final model validation was done using MolProbity 64 and correlation between models and the corresponding maps were estimated using mtriage 64 . All visualizations were done using PyMOL, UCSF Chimera 61 or ChimeraX 65 .

Correlative cryo-light and electron microscopy
For correlative cryo-light and electron microscopy, frozen grids containing WT S. coelicolor were transferred to a CMS196V3 Linkam cryo-stage and imaged using a ×100 numerical aperture 0.74 objective on an LSM900 Airyscan 2 Zeiss microscope driven by ZEN Blue software (v3.5). Fluorescence images of areas of interest were manually correlated with the corresponding TEM square montage using SerialEM 54,66 .

Fluorescence-based cell viability assay
To express sfGFP constitutively in Streptomyces strains, the coding sequence for sfGFP was introduced downstream of the constitutive promoter ermE* on an integrating plasmid vector (pIJ10257). The plasmid was introduced by conjugation to S. coelicolor strains (WT, ∆CIS and CIS-N5, ∆Sco4256-4258). These strains were inoculated into 30 ml of TSB liquid culture and incubated at 30 °C with shaking at 250 r.p.m. in baffled flasks for 48 h. Where appropriate, nisin and CCCP (or 0.002% DMSO) were added to a final concentration of 1 µg ml −1 and 10 µM, respectively. Cultures were incubated for a further 90 min. For UV exposure, 10 ml of the S. coelicolor cultures was transferred into a Petri dish and treated with Sankyo Denki Germicidal 68 T5 UV-C lamps for 10 min in a Herolab UV DNA crosslinker CL-1. Then, 1 ml aliquots were centrifuged for 5 min at 15,000 × g, washed twice with PBS and resuspended in 1 ml PBS with 5 µg ml −1 FM5-95 membrane stain. The cell suspension and membrane stain were mixed by vortexing and kept in the dark at room temperature for 10 min. The suspension was then centrifuged for 5 min at 15,000 × g, washed twice with PBS and resuspended in 50 µl PBS. Samples (10 µl) were immobilized on 1% agar pads and imaged on the Thunder Imager 3D cell culture microscope (Leica) at room temperature. First, tile scan images were acquired on the LasX Navigator plug-in of Leica Application Suite X (LasX) software (v3.7.4.23463), and 100 targets were picked manually. Then z-stack images with the HC PL APO ×100 objective were acquired at an excitation of 475 nm and 555 nm under GFP (green) and TRX (red) filters, respectively. Images Article https://doi.org/10.1038/s41564-023-01341-x were processed using LasX software to apply thunder processing and maximum projection, FIJI to create segmentation and quantify the live (sfGFP)/total cells (FM5-95) area ratio 51 , and statistical analysis was performed on GraphPad Prism 9 (v9.3.1).

Cover glass impression of Streptomyces spore chains
Spore titres of relevant strains were determined by standard techniques. S. coelicolor (10 7 colony forming units (c.f.u.s)) of strains WT, SS387, SS393, SS395 and SS540 were spread onto R2YE agar plates and grown at 30 °C. Sterile glass coverslips were gently applied to the top surface of each bacterial lawn after 48 h, 72 h and 96 h post inoculation. Cover slips were then mounted onto glass microscope slides and imaged using a ×40 objective on a Leica Thunder Imager 3D cell culture microscope. Images were processed using FIJI 51 .
Actinorhodin production assay S. coelicolor strains (WT, SS387, SS393, SS395 and SS540) were inoculated into 30 ml R2YE liquid media at a final concentration of 1.5 × 10 6 c.f.u. ml −1 . Cultures were grown in baffled flasks at 30 °C overnight. Cultures were standardized to an optical density (OD) 450 of 0.5 and inoculated in 30 ml of fresh R2YE liquid medium. For visual comparison of pigment production, images of the growing culture were taken between t = 0 and t = 72 h (as indicated in Extended Data Fig. 9c). For quantification of total actinorhodin production, 480 µl of samples were collected at the same timepoints when images were taken. KOH (120 µl of 5 M) was added, samples were vortexed and centrifuged at 5,000 × g for 5 min. The weight of each tube was recorded. A Synergy 2 plate reader (Biotek) was then used to measure the absorbance of the supernatant at 640 nm. The absorbance was normalized by the weight of the wet pellet.

E. coli killing assay
E. coli Rosetta (DE3) cells carrying the plasmids for the expression of the putative effector genes (sco4256, sco4257 or sco4258) or the control construct (yfp with a synthetic membrane anchor) were grown in liquid LB medium supplemented with 150 µg ml −1 carbenicillin and 0.2% glucose overnight. Optical density of the overnight cultures was adjusted to 1 and serial 10-fold dilutions were spotted on solid LB containing carbenicillin in addition to either 0.2% glucose (repressive conditions) or 100 µM isopropyl β-d-1-thiogalactopyranoside (IPTG) (induction conditions). Plates were scanned after an overnight incubation at 37 °C.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.