3D4/21 cells were cultured in a complete medium of RPMI-1640 (Gibco, Carlsbad, CA, United States). Human embryo kidney (HEK-293T) cells were grown in DMEM (Gibco, Carlsbad, CA, United States). Both cells required the addition of 10% fetal bovine serum (Gibco, Carlsbad, CA, United States). They were subcultured upon reaching 90% confluence and incubated in a 37 °C incubator containing 5% CO₂. The bacterial strains and plasmids used in this study are listed in Table 1. All strains were grown in broth with shaking at 180 rpm at 37 °C.

Antibodies for caspase-3 (ab32351), EEA1 (ab109110), Rab5 (ab218624), and RCAS1 (ab200348) were purchased from Abcam (Cambridge, Cambridgeshire, Britain). Antibodies for Flag (66008-4-Ig), and His (66005-1-IG) were from Proteintech Group (Wuhan, Hubei, China). Na⁺/K⁺-ATPase (P06685) and γH2AX (9718T) were purchased from Cell Signaling Technology (Danver, MA, United States). β-actin (High Dilution) (AC026) and the secondary antibodies including HRP Goat Anti-Mouse IgG (H + L) (AS003), HRP Goat Anti-Rabbit IgG (H + L) (AS014), Alexa Fluor 488-conjugated Goat anti-Rabbit IgG (H + L) (AS053) and Alexa Fluor 594-conjugated Goat anti-Rabbit IgG (H + L) (AS039) were bought from Abclonal (Wuhan, Hubei, China). Cell Plasma Membrane Staining Kit with DiI (Red Fluorescence) (C1991S) was bought from Beyotime (Shanghai, China).

A previous report has been published on the production of Rab4b antiserum. Simply put, Rab4b antiserum was produced by immunizing mice with the C-terminal peptide of Rab4b conjugated to KLH (Zhang et al., 2024).

Construction and expression of plasmids containing cdtA, cdtB, and cdtC genes have been reported previously. In brief, the constructed plasmids were transformed into Rosetta (DE3) pLysS. Recombinant clones were induced with 0.2 mM IPTG for 12 h at 25 °C to achieve optimal expression. These three His-tag fusion recombinant proteins were purified by Ni affinity chromatography (Bio-Rad, Hercules, CA, United States) and then confirmed by SDS-PAGE electrophoresis. In this study, the three subunits were reconstituted as a holotoxin at 4 °C overnight.

The expression of anti-GpCdtA, anti-GpCdtB, and anti-GpCdtC antisera was also as previously described. Each mouse was immunized subcutaneously on days 0, 14, and 21 with 0.1 mg of GpCdtA, GpCdtB, or GpCdtC (200 μL) and 20 μL of water adjuvant Montanide Gel 01 (SEPPIC, France). Twenty-eight days later, blood was collected from the mice and left at 4 °C overnight. The serum was then collected and stored in the refrigerator (Yang et al., 2023).

HEK-293T cells at 50% confluence in 6-well plates were cotransfected with pcDNA-3.1-Flag-Rab4b. At 36 h post-transfection, the cells were incubated with 10 μg/mL GpCDT for 30 min. Then, coimmunoprecipitation assays were performed using Protein A/G Magnetic Beads (MedChemExpress, United States). Western blot analysis was performed using antisera against GpCdtA, GpCdtB, and GpCdtC.

To analyze the interaction between GpCDT subunits and the host protein Rab4b, HEK-293T cells were seeded into 6-well culture plates and transfected with the corresponding expression plasmid. Transfected cells were harvested at 48 h post-transfection and lysed in cell lysis buffer containing one mM protease inhibitor. After centrifugation at 14,000 g for 10 min, the lysate supernatant was incubated with 2.5 μg of His-GpCDT subunits for 4 h with gentle rocking at 4 °C. It was then incubated overnight with mouse monoclonal antibodies (mAbs) against the Flag or His tag, also with gentle rocking at 4 °C. Protein A/G Magnetic Beads washed with cell lysate were added to the supernatants and incubated with gentle rocking for 4 h at 4 °C. The beads were washed four times with cold cell lysate and then boiled in SDS loading buffer for 10 min, followed by Western blot analysis.

The Rab4b and EEA1 small guide RNA (sgRNA) were inserted into the lentiCRISPR-V2 plasmid (Table 2). Using Lipofectamine 3000, the latter were transfected into HEK-293T cells. After a 40-h incubation, the lentivirus-containing supernatants were harvested. 3D42/1 cells were infected with the harvested lentiviruses by incubation for 24 h, and 8 μg/mL puromycin was added to the screen, followed by Rab4b and EEA1 knockout screening for stable Rab4b and EEA1 knockout (“Rab4b-KO” and “EEA1-KO”) cells.

After cells were infected with GpCDT, they were lysed in RIPA buffer (Proteintech, Wuhan, China) and centrifuged. The supernatant was then collected for Western blotting. Samples were then moved to PVDF membranes after being separated on 12.5% SDS-PAGE. The membranes were blocked with 5% nonfat dry milk before the primary antibody was incubated. Then, the membranes were treated with a second antibody after being washed with PBS. The Clarity Max Enhanced Chemiluminescence (ECL) (Bio-Rad, Hercules, CA, United States) was added to membranes and captured by ImageJ software (National Institutes of Health).

Using the UNIQ-10 column total RNA purification kit (Sangon, China), RNA was extracted from 3D4/21 cells that had been treated with GpCDT. The PrimeScript TM RT kit with gDNA Eraser (Takara, Japan) was used to conduct a two-step RT-PCR. Using SYBR premix EX Taq^™ II (Tli RNaseH Plus; Takara, Japan), transcripts were analyzed by qRT-PCR. The 2^−∆∆CT method was used to quantify gene expression, and results are presented relative to expression of β-actin. Table 2 lists the primer sequences that were employed.

Cells were seeded in 6-well tissue culture plates (5 × 10⁵ cells per well). After GpCDT exposure for 48 h, static bright field images of cells were captured using light microscopy (Olympus America, Center Valley, PA).

For detecting cell viability, cells were pre-seeded into 96-well plates and treated with GpCDT (10 μg/mL) for 0, 12, 24, 36, 48, and 60 h. A routine CCK-8 assay was used to examine cell viability according to the manufacturer’s protocol.

In 6-well plates, cells were grown to around 90% confluence. Purified GpCDT (10 μg/mL) was then added to each well, and the cells were incubated for 12 and 24 h at 37 °C. Following three PBS washes, the cells were fixed with 4% paraformaldehyde for 15 min before being rinsed with PBS once more. After that, cells were blocked for an hour at 37 °C in a BSA (3%) solution. Anti-H2AX primary antibody (1:1,000; Abcam, MA, United States) was incubated with cells overnight at 4 °C. Samples were then incubated at 37 °C for 1 h in the dark with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG (Proteintech, Beijing, China). To identify the nuclei, DAPI (Beyotime, Shanghai, China) was utilized.

Trafficking studies were performed as follows. Cells were inoculated in 6-well plates and incubated with 10 μg/mL GpCDT for 30 min at 4 °C to promote binding, followed by incubation at 37 °C for 45 min to stimulate uptake. Cells were subsequently immunostained for GpCdtB and intracellular markers.

Statistical analyses were performed using GraphPad Prism version 8.0 (CA, United States). Statistical significance was assessed using Student’s t-test, one-way ANOVA, or two-way ANOVA. Significant differences between groups are indicated by ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, and ^****p < 0.0001.

We previously demonstrated that Rab4b has the potential for a direct interaction with GpCDT (Zhang et al., 2024). The GpCDT subunits have molecular masses of approximately 42 kDa (GpCdtA), 47 kDa (GpCdtB), and 36 kDa (GpCdtC) (Yang et al., 2023). To further understand how Rab4b controls sorting events in the early endosomes of the GpCDT subunit, the present study aimed to investigate the possible role of Rab4b in this process using a coimmunoprecipitation assay. HEK-293T cells were transfected with expression plasmids to overexpress Flag-Rab4b. The expression of this construct was confirmed by Western blotting (Figure 1A). In addition, co-IP was performed with anti-GpCdtA antiserum, anti-GpCdtB antiserum, anti-GpCdtC antiserum, anti-Flag mAb, or anti-His mAb to capture protein complexes. As shown in Figure 1B, anti-GpCdtA and anti-GpCdtC antisera did not result in imprint bands, but GpCdtB could be observed in the IP samples by using anti-GpCdtB antisera. In the next series of experiments, we incubated Rab4b directly with the GpCdtA, GpCdtB, and GpCdtC subunits. Coimmunoprecipitations, performed using anti-His and anti-Flag mAb, showed that Rab4b coimmunoprecipitated with GpCdtB (Figure 1C) but not with GpCdtA and GpCdtC (Figures 1D,E). These results suggested that Rab4b can interact with the GpCdtB.

We used the CRISPR/Cas9 gene editing method to knock out Rab4b (see the Methods section). qRT-PCR and Western blotting results confirmed that Rab4b mRNA (Figure 2A) and protein (Figure 2B) were knocked out entirely in Rab4b-KO cells. Next, we examined the morphology of 3D4/21 cells and Rab4b-KO cells after treatment with GpCDT. Knockout of Rab4b resulted in significantly exhibited less cell distention after 48 h of GpCDT treatment (Figure 2C). Furthermore, Rab4b-KO cells demonstrated a considerable improvement in cell viability after GpCDT treatment (Figure 2D). Moreover, as shown in Figure 2E, GpCDT treatments for 24 h increased the number of gH2AX foci, a marker for DNA damage signaling, in 3D4/21 cells, when compared with Rab4b-KO cells. Moreover, in control 3D4/21 cells, GpCDT treatment induced caspase-3 cleavages (Figure 2F). These results indicate that Rab4b is necessary for GpCDT-induced cytotoxicity against the 3D4/21 cells.

Early endosomes and sorting endosomes, which are located where the endocytic and exocytic pathways converge, have been important hubs for membrane trafficking. We first demonstrated that in the presence of Rab4b, 3D4/21 cells took up GpCDT inwards (Figure 2G). Once internalized, CdtB first reaches the early endosomes. Rab4b, located in the early endosomes, then exerts its function of controlling early endosome sorting, regulating the translocation of internalized GpCdtB to the Golgi apparatus and endoplasmic reticulum (ER) in a retrograde manner in several cell types. 3D4/21 cells treated with GpCDT display fluorescence of GpCdtB (green), which co-localises with early lysosomes and the Golgi apparatus (red). This indicates that CdtB transport from early endosomes (Rab5+) to the Golgi apparatus (RCAS1+) also occurs in 3D4/21 cells (Figure 2H). Noteworthy is that we examined the vesicle trafficking of GpCDT in Rab4b-KO cells to investigate the impact of Rab4b on this process. The results revealed that in Rab4b-KO cells, the red fluorescence of GpCdtB did not colocalize with the green fluorescence of Rab5 and RCAS1 (Figure 2I), indicating that GpCDT failed to reach the early endosome after Rab4b knockout. In summary, these findings show that Rab4b influences vesicle trafficking in GpCDT.

To explore the role of Rab4b after GpCDT treatment, we conducted a comparative transcriptomic study on Rab4b-KO cells and 3D42/1 cells incubated with purified GpCDT for 12 h. Following quality control, all samples had Q20 and Q30 percentages of clean data greater than 97 and 93%, respectively. Each sample’s clean readings had a GC content ranging from 51.63 to 53.34% (Supplementary Tables S1). Transcriptome analysis revealed 1,420 differentially expressed genes (DEGs) after 12 h in the Rab4b-KO group compared with the 3D4/21 control group. Of these genes, 483 were upregulated and 937 were downregulated (Figure 3A). Six genes with elevated transcription levels and six with downregulated transcription levels were arbitrarily chosen for relative fluorescence measurement to corroborate the transcriptome sequencing results (Supplementary Tables S2). As shown in Figure 3B, the expression of most of the selected DEGs generated by RNA sequencing was consistent with the levels obtained using qRT-PCR. In addition, the correlation between the expression levels of the selected DEGs obtained using RNA sequencing and qRT-PCR was analyzed by calculating Pearson’s correlation coefficient. The results revealed that Pearson’s correlation coefficient was 0.3331, confirming that the data generated by RNA sequencing were reliable.

KEGG pathway analysis and GO categorization were applied to the DEGs. According to the GO analysis, the most prevalent categories were leukocyte migration, the G-protein-coupled receptor signaling pathway, the extracellular space, the cell surface, and chemokine activity (Figure 3C). KEGG analysis revealed that the significantly enriched pathways were associated with cytokines and cytokine receptors and pathways involved in lipid metabolism and atherosclerosis (Figure 3D). GO and KEGG enrichment analyses suggested that Rab4b influences several biological functions in 3D4/21 cells infected with GpCDT. Rab4b is associated with the endocytosis pathway in the KEGG signaling pathway, which is closely related to cell vesicle trafficking. Fourteen DEGs were identified in the endocytosis pathway (Figure 3E), and only EEA1, like Rab4b, is located in the early endosomes, which can mediate early endosome fusion, as well as the capture of vesicles derived from clathrin-coated pits by early endosomes. EEA1 may be an important protein involved in Rab4b-mediated GpCDT cytotoxicity and vesicle trafficking.

In this study, transcriptomic sequencing revealed that the expression level of EEA1 changed significantly after 12 h of cell infection with GpCDT. The expression levels of EEA1 in 3D4/21 and Rab4b-KO cells, as detected by qRT-PCR and Western blot, after GpCDT treatment. The results showed that the mRNA level of EEA1 increased after GpCDT treatment in WT cells. After the elimination of Rab4b, the mRNA level of EEA1 did not change significantly (Figure 4A). Western blotting results also revealed that the EEA1 protein level increased after 24 h of treatment of 3D4/21 cells with GpCDT, whereas the EEA1 protein level did not change significantly after Rab4b knockout (Figure 4B).

The EEA1 knockout cells were subsequently constructed via CRISPR/Cas9 technology. The western blot results showed that the expression level of the EEA1 protein in EEA1-KO cells was knocked out compared with WT cells, indicating that the EEA1 knockout (“EEA1-KO”) cells were successfully constructed (Figure 4C). Subsequently, using Western blotting and qRT-PCR, the Rab4b mRNA levels in EEA1-KO cells treated with GpCDT were determined. These results confirmed that when EEA1 was knocked out, Rab4b mRNA and protein levels did not alter considerably (Figures 4D,E). These findings suggested that in GpCDT-treated cells, EEA1 did not affect Rab4b transcription levels.

It has been confirmed that knockout of Rab4b can inhibit the expression of the EEA1 protein after GpCDT treatment. The following steps were taken to confirm whether the EEA1 protein influences Rab4b-mediated GpCDT-induced cytotoxicity and vesicle trafficking in 3D4/21 cells. Using indirect immunofluorescence, we were able to identify the fluorescence of GpCdtB and get insight into the uptake of GpCDT by cells following EEA1 deletion. The findings showed that GpCdtB fluorescence was not seen in EEA1-KO cells (Figure 5A). Next, observe the morphological changes that occur as a result of exposure to GpCDT. EEA1-KO cells were more resistant to GpCDT and exhibited less cell distention (Figure 5B). Meanwhile, the study of the EEA1 effect on cell survival rate showed that the survival rate of EEA1-KO cells treated with GpCDT was greater than that of WT cells (Figure 5C). Similarly, we demonstrated that GpCDT reduced the number of γH2AX foci in EEA1-KO cells compared to 3D4/21 cells, a marker for DNA damage signaling (Figure 5D). Western blotting was used to detect the activation level of caspase-3, a key apoptotic factor, after GpCDT treatment. The results showed that the cleaved caspase-3 bands appeared at 48 h after GpCDT treatment, while the caspase-3 cleavage bands appeared at 60 h in EEA1-KO cells (Figure 5E). Overall, the experimental data support our hypothesis that EEA1 affects the Rab4b-mediated internalization and intoxication of 3D4/21 cells by the GpCDT.