Autophagy is an evolutionarily conserved process to protect cells against some unfavorable conditions, such as starvation, hypoxia, and nutrient deficiency. It is first activated by the autophagy-related genes (Atg) to form autophagosomes in the cytoplasm. Subsequently, some autophagic cargos like misfolded long-lived proteins and damaged organelles and nucleic acids are wrapped into autophagosomes and then degraded by autolysosomes. Increasing evidence confirmed that autophagy provides metabolic products for tumor cells, and thereby reduced the sensitivity of tumor cells to chemotherapeutic drugs and promoted metastasis and chemoresistance in these cells. In this regard, autophagy inhibition may represent a potential approach for cancer therapy. In this study, we firstly performed a preliminary screen in search of novel inhibitors of autophagic flux from eight commercially available pulchinenosides. We found that PSD, PSE2 and PSE4 can be identified as the most active autophagic flux inhibitors in MCF-7 and MDA-MB-231 human breast cancer cells, which is evidenced by a large number of LC3B punctate dots and increased levels of turnover of LC3II/LC3I and p62 proteins. These three active saponins were used as tool drugs for autophagy inhibition in the following studies. In order to identify the dominant structure of pulchinenosides that contributes to the autophagy-inhibitory activity, we preliminarily searched 17 analogues of oleanane-type pulchinenosides from the compound library. Analysis of the structure-activity relationship demonstrated that pulchinenoside analogues with one sugar unit attached to C-3 position probably could induce autophagic flux in the two breast cancer cells. However, the analogues with more than two sugar units attached to C-3 position possibly were more likely to inhibit autophagic flux in comparison to the analogues with more sugar units attached to C-3 and C-28 position. Then, we investigated the underlying molecular mechanisms of autophagy inhibition by PSE4 and its synergistic anti-breast cancer effects with 5-fluorouracil in the two breast cancer cells. It was found that PSE4 remarkably induced the formation of autophagic vesicles and increased the LC3II/LC3I ratio and p62 protein levels, indicating that PSE4 could inhibit autophagic flux. Moreover, a disruption of lipid rafts in lysosomes was observed in PSE4-treated cells, as evidenced by decreased fluorescence of CTxB staining at the lysosomal fraction of the cells. Interestingly, recovering lipid rafts with cholesterol decreased high levels of the conversion of LC3I into LC3II and accumulation of the p62 protein induced by PSE4, indicating that lipid raft disruption is required for PSE4-induced autophagy inhibition. Furthermore, PSE4 remarkably increased lysosomal pH and suppressed lysosomal cathepsins activation, leading to lysosomal dysfunction in the two breast cancer cell lines. However, the activities in lysosomes were recovered significantly when treatment with cholesterol, which is used to restore lipid rafts. These results demonstrated that PSE4-mediated disruption of lipid rafts in lysosomes impaired the function of lysosomes, blocked the autophagosome-lysosome fusion and thereby inhibited autophagic flux. Finally, we demonstrated that PSE4 synergistically enhanced the anti-breast cancer activity of 5FU in cultured breast cancer cells and in xenograft mouse models. Next, we explored blockade of the autophagosome-lysosome fusion and promotion of p62-medicated ubiquitinated protein aggregation by PSD and its synergistic antitumor effects in combination with camptothecin (CPT) in vitro and in vivo. PSD was found to inhibit the late-stage autophagic flux through interrupting the autophagosome-lysosome fusion, neutralizing the acidity of lysosomes and diminishing the activity of lysosomal protease enzymes. Furthermore, PSD strongly facilitated ubiquitinated protein aggregation upon high levels of p62 accumulation. These effects may confer