The strains were grown on LB or LB agar and cultured at 37°C under oxic conditions. Liquid cultures were incubated at 37°C in an oxic environment at 220 rpm. Plasmids were constructed using standard molecular biology techniques. Templates, primers, and enzymes were added to the PCR system to amplify the DNA fragments. The DNA was digested and ligated according to the reagent instructions, and the correct recombinant plasmid was screened through colony verification and confirmed by sequencing. The strains and plasmids used in this study are listed in Supplementary Table S1, and the primers are listed in Supplementary Table S2. The dual expression plasmid pBAD22a- ftsZ-icd was constructed, in which ftsZ and icd genes had their own ribosome binding site (rbs).

The monoclonal E. coli containing the pBAD22a plasmid was inoculated into 5 mL of LB liquid medium with the appropriate antibiotics. The spectrophotometer was calibrated, and once the bacterial culture reached an OD₆₀₀ of approximately 0.4, arabinose was added to a final concentration of 0.2% to induce gene expression for 3 h. The sample with 10 μL of the bacterial solution was then taken, spotted onto LB agar, and covered with a coverslip for observation under a microscope. Cell lengths were measured for each mutant, and the results were reported as the mean ± SD. Additionally, we spotted 5 μL of dilutions from a dilution series onto LB agar plates. The plates were dried, inverted, and incubated overnight at 37°C. The number of viable bacteria was counted the next day.

For the growth curve analysis, once the bacterial culture reached an OD₆₀₀ of approximately 0.4, 100 μL of culture was transferred into 20 mL of fresh LB medium. L-arabinose was added to a final concentration of 0.2%, and the culture was incubated in a constant temperature shaker at 37°C and 220 rpm. OD₆₀₀ was measured at hourly intervals, and the data was used to generate a growth curve.

Fluorescence image visualization of the Z ring used a mutant strain FtsZ-G55-mNeonGreen-Q56, which replaced genomic ftsZ with ftsZ-mNeonGreen under its normal genomic control (Moore et al., 2017). After the plasmid pBAD22a carrying the icd gene was transferred into the strain, the bacteria were cultured in LB medium with 100 μg/mL carbenicillin until the OD₆₀₀ reached ~0.4, followed by the addition of 0.2% arabinose for induction at 37°C for 3 h. A droplet with 10 μL of bacterial cells was placed onto an agarose spacer, and the cells were visualized using a Leica DMI3000B fluorescence microscope. A common light source was used to observe bacterial morphology, while a mercury lamp was employed for fluorescence observation, with an excitation wavelength of 488 nm for mNeonGreen. The lengths of individual cells were measured, and the results are presented as mean ± SD.

DAPI staining of bacterial nucleoids was performed using a stock solution of 20 mg/mL and a working concentration of 5 μg/mL. At room temperature (25°C), 1 mL of the working solution was applied to the bacterial cells for 8 min under dark conditions. The cells were then washed four times with PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH 6.5). The cells were visualized using a Leica DMI3000B fluorescence microscope with an excitation wavelength of 360 nm for DAPI.

The pET28a-His-ftsZ plasmid was transformed into E. coli BL21, plated on kanamycin-containing agar plates (50 μg/mL), and incubated overnight at 37°C in a constant temperature incubator. Monoclonal clones were selected and inoculated into 100 mL of TY medium (2.4% yeast extract, 1.2% tryptone, 0.94% wt/vol K₂HPO₄, and 0.02% wt/vol KH₂PO₄) under oxic conditions, and cultured for 24 h at 37°C with shaking at 220 rpm. The cultures were centrifuged at 3000 g for 20 min at ambient temperature, with the supernatant discarded. The precipitate was resuspended in the buffer containing 50 mM Tris, 300 mM NaCl at pH 7.5, with 1% PMSF, 1% lysozyme, and 1% Triton X-100 added, and incubated on ice for 30 min. The cell disruption was performed at 1000 Bar until the lysate was clear and transparent. The lysate was centrifuged at 18,000 g for 20 min at 4°C, and the supernatants and pellets were collected and analyzed for protein production by SDS-PAGE.

The pME6032-GST-icd recombinant plasmid was transformed into PAO1 by electrotransformation, grown to an OD₆₀₀ of approximately 0.6. Gene expression was induced with 0.8 mM IPTG at 16°C and 220 rpm for 24 h. Following cell lysis, the supernatant was analyzed by Western blot to examine protein production.

Bacterial two hybrid screening has been modified based on previous reports (Karimova et al., 1998). The plasmids pUT18-icd and pKNT25-ftsZ were simultaneously electrotransformed into E. coli BTH101. The bacteria were cultured on solid medium containing 40 μg/mL X-gal (5-bromo-4-chloro-3-indolyl-β-D-galactoside) as a chromogenic substrate and 50 μg/mL kanamycin and 100 μg/mL carbenicillin as antibiotics. IPTG (isopropyl-β-D-thiogalactoside) was typically added to the medium at a concentration of 0.5 mM to induce the expression of the full hybrid protein and the β-galactosidase reporter gene. To exclude false-positive clones, putative positive clones were selected and reseeded on LB agar plates containing 50 μg/mL kanamycin, 100 μg/mL carbenicillin, 0.5 mM IPTG, and X-gal, with true positive clones expected to yield blue colonies on these plates once again.

Bacterial cells with β-galactosidase activity to be detected were added with 0.5 mM IPTG and appropriate antibiotics in 3–5 mL LB and stirred 6 h at 220 rpm at 37°C. 50 μL bacterial suspension was added to 370 μL of Z buffer (100 mM Na₂HPO₄, 10 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, 5.4 μL/mL β-galactosidase, pH 7.0), followed by the addition of 20 μL chloroform and 10 μL of 0.1% SDS solution. After incubation on ice for 30 min, 4 mg/mL ONPG (o-Nitrophenyl β-D-galactopyranoside) was added, and the mixture was well mixed before being placed at 30°C for the reaction. After 30 min, a yellow color was observed. The reaction was terminated by adding 250 μL of 1 M Na₂CO₃ solution, and the time of color change was accurately recorded. Absorbance at 420 nm and 550 nm was measured. Miller Units were calculated using the formula: Miller Units = 1,000 x (OD₄₂₀–1.75 x OD₅₅₀)/t(min) x V(mL) x OD_600.

After lysing the P. aeruginosa PAO1 strain overexpressing GST-icd or the E. coli BL21 strain overexpression His-ftsZ, the samples were centrifuged at 18000 g for 20 min, and the supernatants were collected separately. After mixing the supernatants containing 1 mL of GST-Icd and 50 μL of His-FtsZ, the mixture was added to the cleaned 50 μL GST resin and incubated for 2 h then washed 5 times with PBS buffer. The mixture of GST protein and His-FtsZ protein, as well as a supernatant containing only GST-Icd or His-FtsZ were used as controls. After the protein bound to the resin, the resin was centrifuged at 3000 g for 2 min, and the supernatant was discarded. The samples were washed three times with 1 mL PBS buffer, and the final centrifugation step ensured maximum removal of the supernatant.

The samples were checked by western blotting. Proteins first were separated using SDS-PAGE gel electrophoresis and transferred from the gel to polyvinylidene fluoride (PVDF) membranes at 220 mA for 90 min and then blocked with 5% skim milk in TBST for 1 h at room temperature on a shaker. The membranes were incubated with primary antibodies overnight at 4°C, washed three times with TBST for 10 min each, and then incubated with secondary antibodies, an HRP-conjugated Goat Anti-Rabbit, for 1 h at room temperature. The membranes were washed three times with TBST for 10 min. The results of Western Blots were detected using the ECL Plus kit (Sangon Biotech). Luminol and hydrogen peroxide in the kit were mixed at a ratio of 1:1 and applied to the PVDF membrane. The membrane was then placed in the dark chamber of the chemiluminescent imaging system and exposed for 10 s to obtain the images. The ECL developing solution reacted with the HRP - labeled antibody to produce a fluorescent signal with a maximum emission wavelength of 425 nm. The primary antibodies used were Anti-GST tag and Anti-His tag antibodies, while the secondary antibody was HRP-conjugated Goat Anti-Rabbit. All antibodies were sourced from Shanghai Sangon Inc.

Overexpression of icd has a serious impact on the growth of E. coli. When we introduced the pET plasmid fused with the icd fragment into E. coli strain BL21 or C41, the bacteria had difficulty forming clones and growing in liquid culture, possibly due to leakage production of the protein. After the pET28a-GST-icd plasmid was transformed into E. coli strain BL21 or C41, it took almost 48 h to form visible colonies on the plate, furthermore, when the bacteria in a single colony were transferred to liquid LB medium, the bacteria hardly grew. To study the effect of Icd on E. coli division, we selected the pBAD22a plasmid with stricter arabinose-inducible expression control, constructed the recombinant plasmid pBAD22a-icd, and transformed it into E. coli DH5α. Overexpression of icd induced by 0.2% L-arabinose resulted in severe filamentation formation of E. coli. The average length of bacteria was 15 ± 1.1 μm after 3 h of overexpression and 53 ± 4.9 μm after 6 h of overexpression (Figures 1A,B).

To further investigate this, we examined the sensitivity of two mutant strains, PS106 (ftsZ84) and PS2342 (ftsA*), to Icd. PS106 (ftsZ84), a ftsZG105S mutant, is a temperature-sensitive mutation that causes E. coli cells to divide severely defectively and grow into filaments at the restrictive temperature (42°C), whereas they grow normally at 37°C. This mutation FtsZG105S protein is impaired in GTP binding and GTPase, as well as in its polymerization (RayChaudhuri and Park, 1992; RayChaudhuri and Park, 1994). Another E. coli mutant, PS2342 (ftsA*), is a mutant of ftsA. Compared to wild-type FtsA, the FtsA* (ftsAR286W, a zipA-bypass ftsA mutant allele) interacts more strongly with FtsZ than FtsA, accelerating the turnover of FtsA* in the ring (Geissler et al., 2007).

As shown in Figures 1A,B, Icd still inhibited cell division in PS106 (ftsZ84), whereas PS2342 (FtsA*) was resistant to Icd. PS106 grew normally at 37°C with an average length of 2.04 ± 0.04 μm. After 3 h of icd gene expression, the average length of PS106 increased to 10.1 ± 1.24 μm. At 42°C, PS106 had already exhibited division defects (27.4 ± 1.7 μm), and icd expression exacerbated these defects, resulting in an average length of 59.2 ± 3.4 μm. Surprisingly, the expression of icd did not affect the growth of strain PS2342 (ftsA*). It is possible that FtsA* enhanced the polymerization of FtsZ, thereby reducing the inhibitory effect of Icd.

Growth curve analysis showed that E. coli overexpressing icd reached its maximum OD₆₀₀ value after 5 h and was followed by a decrease in OD₆₀₀ value, signaling possible bacterial death or degradation (Figure 1C). Viable counts indicated that Icd caused severe damage to E. coli, resulting in a 1,000-fold reduction in viable cell numbers (Figure 1D). The inhibition of bacterial growth by Icd protein was species-specific. When the pME6032-icd recombinant plasmid was introduced into P. aeruginosa and the icd was induced for expression, it had no significant effect on growth (Supplementary Figure S1).

Due to the overexpression of icd leading to bacterial elongation, which may indicate inhibition of bacterial division, we then investigated the effect of icd expression on the formation of Z-ring, the bacterial divisome. Here, we used the E. coli mutant strain, FtsZ-G55-mNeonGreen-Q56 strain (Moore et al., 2017), in which the fluorescent protein mNeonGreen was inserted between positions 55 and 56 of FtsZ to replace the original FtsZ, to observe the Z-ring localization.

FtsZ-mNeonGreen was always located in the middle of the cell, indicating that the Z-ring was assembled. After the pBAD22a-icd plasmid was transferred, the cell appeared filamentous and FtsZ-mNeonGreen diffused throughout the cell, denoting that Z-ring was not assembled correctly. It has been reported that protein Gp0.4 directly inhibited FtsZ assembly in E. coli (Kiro et al., 2013), therefore, we used E. coli with a plasmid overexpressing gp0.4 as a control. We could observe the same phenomenon with overexpression of gp0.4 or icd (Figure 2A). We preliminarily speculate that Icd may directly or indirectly act on FtsZ or Z-ring binding proteins in E. coli.

To determine whether Icd affects bacterial DNA replication, DAPI staining was used to visualize the bacterial DNA. The results showed that although cell division was inhibited, DNA replication remained intact and unaffected (Figure 2A).

We also constructed the recombinant plasmid pBAD22a-icd-mCherry and introduced it into E. coli FtsZ-G55-mNeonGreen-Q56. After induction with a low concentration of arabinose (0.02%) for 0.5 h, both Icd-mCherry and FtsZ-mNeonGreen were unable to be located correctly and instead dispersed throughout the entire bacteria (data not shown).

Overexpression of ftsZ simultaneously could reduce the inhibitory effect of Icd (Figures 2B–D). After 3 h of simultaneous overexpression of icd and ftsZ using plasmid pBAD-ftsZ-icd and 0.2% arabinose, the length of the bacteria became longer, but was significantly shorter than that of the strain overexpressing only icd using plasmid pBAD22a-icd (Figure 2B). The average cell length was 7.8 ± 0.39 μm, which was shorter than the 15 ± 1.1 μm of a strain overexpressing icd alone (Figure 2C). The results of the E. coli growth curve further confirmed that simultaneous overexpression of ftsZ significantly weakened the inhibitory effect of icd (Figure 2D). When icd was overexpressed, the OD₆₀₀ value slowly increased in the first 5 h and then began to decrease. When ftsZ and icd were overexpressed simultaneously, the OD₆₀₀ value continued to increase and reached a maximum value at around 8 h.

The bacterial two-hybrid system is an effective method for studying the protein–protein interactions in vivo (Karimova et al., 1998). We recombinantly fused Icd to the T18 fragments of the adenylate cyclase, FtsZ to the T25 fragments, and transformed them into E. coli BTH101. We used pKNT25 and pUT18C empty plasmids as negative controls and pKNT25-zip and pUT18C-zip plasmids as positive controls. Thus, the Zip proteins could polymerize to form a dimer, resulting in β-galactosidase activity in the bacteria. Additionally, co-transformations of pKNT25-ftsZ with the pUT18C empty plasmid and pUT18C-icd with the pKNT25 empty plasmid into E. coli BTH101 served as two other negative controls (data not shown).

After the recombinant plasmid was transformed into E. coli BTH101 and cultured on solid medium containing IPTG and X-gal (5-bromo-4-chloro-3-indolyl-β-D-thiogalactopyranoside) for 26 h, most of the bacteria cells containing pKNT25-ftsZ and pUT18C-icd turned blue, while the negative control group remained white (Supplementary Figure S2). Figure 3A shows a single clone randomly selected from the initial culture medium, in which the cells containing pKNT25-ftsZ and pUT18C-icd and the positive control containing pKNT25-zip and pUT18C-zip were blue, while the negative control strain was white. The activity of β-galactosidase was measured by UV–VIS spectrophotometer to quantify the results of the bacterial two-hybrid measurement, further confirming that Icd and FtsZ interacted in the cells (Figure 3B).

A pull-down assay was also conducted to confirm the interaction between Icd and FtsZ. Since Icd inhibited E. coli growth, the production level of Icd protein expressed in E. coli was low, therefore, we used P. aeruginosa PAO1 and the recombinant plasmid pME6032-GST-icd to overexpress Icd protein. The cell lysates containing Icd protein were purified and enriched by GST resin, and mixed with His-FtsZ sample. We used a mixture of GST protein and His-FtsZ as a control. After elution and immunoblotting with anti-GST and anti-His antibodies, it was found that GST-Icd could pull down His-FtsZ, while the GST tag alone could not (Figure 3C). This suggested that Icd interacts directly with FtsZ.

Amino acid sequence alignment revealed that Icd proteins exist not only in the P1 phage, but also in some other Enterobacteriaceae phages, although there are some sequence differences (Supplementary Figure S3). Based on the alignment, we constructed recombinant plasmids containing various lengths of icd gene deletions and determined the antibacterial core sequence region of Icd by observing bacterial growth. Icd sequences of different lengths are listed in Figure 4A. Icd sequences of different lengths were inserted into the pBAD22a plasmid and transformed into E. coli DH5α.

Firstly, we observed the effect of mutated Icd on bacterial length under a microscope (Figure 4B). We found that when the N-terminal truncation of Icd proteins exceeds 11 amino acids or the C-terminal exceeds 24 amino acids, the overexpressed icd mutant gene loses its antibacterial effect. Among the Icd mutants we constructed, Icd12-51 appeared to be the shortest protein with the same ability to inhibit bacterial division as full-length Icd. After overexpressing icd12-51 for 6 h, the average length of E. coli reached 48 ± 5.5 μm, and the longest bacteria we observed reached 220 μm. However, the average length of bacteria overexpressing icd12-50 decreased to 7 ± 0.34 μm, indicating a significant reduction in inhibitory ability (Figures 4B,C).

Next, we measured the bacterial growth curves to test their antibacterial activity. Overexpression of icd genes has a strong inhibitory effect on bacterial growth, which we used as a positive control, and the growth of bacteria containing an empty pBad22 plasmids as a negative control. The OD₆₀₀ values of E. coli overexpressing icd genes increased slowly in the first 5 h, which may be due to an increase in bacterial numbers or the formation of filamentous bacteria. Afterwards, the OD₆₀₀ value began to decrease and dropped to baseline after 9 h, indicating bacterial death (Figures 5A–C). Figure 5A examined the effect of overexpression of N-terminally truncated Icd mutant proteins on E. coli growth. The results showed that removing 7 or 11 amino acids from the N-terminus of the Icd mutant proteins still had partial inhibitory effects, while removing 16 amino acids from the N-terminus resulted in the disappearance of antibacterial activity. Figure 5B examines the effect of C-terminal truncation on the antibacterial activity of Icd proteins. Truncating the 21 amino acids at the C-terminus (Icd1-52) did not affect the antibacterial activity of the Icd protein. Figure 5C shows the antibacterial effect of Icd proteins with both N-terminal and C-terminal truncations to search for the shortest core region. The results showed that Icd12-52 had almost the same antibacterial activity as full-length Icd, with OD₆₀₀ increasing slowly in the first 6 h then decreasing. Overexpression of icd12-51 also had strong antibacterial activity, but no decrease in OD₆₀₀ was observed, indicating that the antibacterial activity was partially weakened. At the same time, the antibacterial effect of overexpression of icd12-50 was significantly weakened (Figure 5C). Interestingly, the antibacterial activity of overexpression of icd12-52 or icd12-51 was stronger than that of overexpression of icd8-73, icd12-73 or icd8-57, suggesting that the extra amino acids may weaken the antibacterial activity.

Not only does an increase in the number of bacteria cause an increase in OD₆₀₀ value, but a significant increase in bacterial length will also increase OD₆₀₀ value. To obtain more accurate results, we used the serial spot dilution assay to measure the number of viable bacteria and assess the antibacterial activity of Icd mutant proteins. Here, we measured the viable counts after 6 h of overexpression of icd and its mutant genes (Figure 5D). In the absence of arabinose, the number of E. coli colonies was nearly identical, around 4 × 10⁸ CFU/mL. When icd was overexpressed, the number of E. coli colonies decreases to 2 × 10⁵ CFU/mL. Meanwhile, the number of colonies overexpressing icd12-52 or icd12-51 was 2 × 10⁵ CFU/mL and 4 × 10⁵ CFU/mL, respectively. On the other hand, the colony count of overexpressed icd12-50 was only 4 × 10⁶ CFU/mL, which was significantly weakened. In addition, the measurable viable counts after overexpression of icd8-73, icd8-57 or icd12-73 were 2 × 10⁶, 2 × 10⁶ and 6 × 10⁶ CFU/mL, respectively, further demonstrating that their antibacterial activity was lower than that of icd12-51. The colony counts after overexpression of icd12-52 and icd12-51 were in the same order of magnitude, which concluded that Icd12-51 can be considered as the shortest core region with antibacterial effects similar to those of the full-length Icd.

Bacterial two-hybrid experiments confirmed the interaction between Icd12-51 and FtsZ (Figures 5E,F; Supplementary Figure S2). Most of the bacteria cells containing pKNT25-ftsZ and pUT18C-icd12-51 turned blue, but the color was lighter than that of bacteria containing pKNT25-ftsZ and pUT18C-icd. Moreover, the β-galactosidase activity was only 40% of that of the bacterial cells containing pKNT25-ftsZ and pUT18C-icd (Figures 5E,F). It can be inferred that Icd12-51 forms an effective but slightly weaker interaction with FtsZ. Meanwhile, the cells containing pKNT25-ftsZ and pUT18C-icd12-50 turned a lighter blue, and the β-galactosidase activity was approximately 12% of that of the cells containing pKNT25-ftsZ and pUT18C-icd (Figures 5E,F). This showed that Icd12-50 was still able to form a weaker interaction with FtsZ.