**3. Results**

### *3.1. Basal and Nitrogen Starvation-Induced Autophagy are not A*ff*ected by Mmi1*

A previous study in mammalian cells indicated that TCTP promotes autophagy under hypoxia and starvation conditions [10]. Later, however, the results were challenged by a study demonstrating TCTP as a negative regulator of non-selective autophagy [11]. To investigate the role of the yeas<sup>t</sup> TCTP ortholog, Mmi1, in non-selective autophagy, we examined autophagy in wild-type (WT) and *mmi1*Δ cells upon shift to a nitrogen starvation medium that stimulates autophagy induction [12]. As a negative control we used autophagy deficient strains *atg8*Δ depleted for the key autophagy molecule Atg8 [50]. The non-selective autophagy was determined by the quantitative Pho8Δ60 assay [43]. The Pho8Δ60 assay is an enzymatic assay that utilizes a truncated version of the alkaline phosphatase Pho8Δ60 that lacks the targeting sequence to endoplasmic reticulum, and thus remains in the cytosol. Upon autophagy induction, Pho8Δ60 is delivered to the vacuole, gets activated by the proteolytic cleavage, and serves as a marker of the amount of cytosol delivered through the non-selective autophagy [43]. Yeast strains were grown in the YPD medium until the early logarithmic phase, washed, and shifted to the nitrogen starvation (SD-N) medium. As shown in Figure 1A, at 0 h upon shift to SD-N media low values of the phosphatase activity were detected in all tested strains, indicating that basal levels of autophagy were not affected. Upon prolonged nitrogen starvation, phosphatase activities increased. However, similar values were detected for the WT and the *mmi1*Δ strains at all tested time points. On the other hand, the control *atg8*Δ strain exhibited only a low level of autophagy, demonstrating that delivery of Pho8Δ60 to the vacuole depends on autophagy. These results indicated that Mmi1 did not affect the nitrogen starvation-induced non-selective autophagy.

To confirm the results by an independent approach we also performed GFP-Atg8 processing assay [28]. The GFP-Atg8 assay is based on Western blot detection of a free GFP moiety released from a core autophagy protein Atg8 that is resistant to vacuolar proteases [28]. As a negative control we used autophagy deficient *atg1*Δ depleted for the key autophagy molecule Atg1 [51]. As shown in Figure 1B, no free GFP band was detected by Western blot at 0 h time point of nitrogen starvation in WT, *mmi1*Δ, and *atg1*Δ strain demonstrating that basal level of autophagy was not affected in the strains. Upon prolonged nitrogen starvation, similar levels of the free GFP band were detected in the WT and the *mmi1*Δ cell lysates, indicating that Mmi1 did not influence GFP-Atg8 cleavage upon nitrogen starvation. Further, no free GFP band was detected in the negative control *atg1*Δ strain at all tested time points, demonstrating that GFP-Atg8 cleavage was dependent on the autophagic degradation. These results, together with the results obtained by the phosphatase assay, demonstrate that the basal and the nitrogen starvation-induced autophagy are not influenced by Mmi1 in yeas<sup>t</sup> *S. cerevisiae*.

**Figure 1.** Basal and nitrogen starvation-induced autophagy normally occurin*mmi1*Δcells. (**A**) Exponentially growing WT, *atg8*Δ, and *mmi1*Δcells expressing Pho8Δ60 (OD600 ≈ 0.8) were shifted to nitrogen starvation medium (SD-N). Samples were taken in indicated time points, proteins extracted, and the specific Pho8Δ60 activity was determined. Results are means ± SD of three independent experiments performed in duplicates (n = 6). The statistical evaluation was performed by using two way analysis of variance (ANOVA). The threshold for significance was set as *p* ≤ 0.01. NS; not significant; (**B**) Western blot detection of GFP-Atg8 cleavage in the WT, *atg1*Δ, and *mmi1*Δcells. Cells expressing GFP-Atg8 were grown until the logarithmic growth phase (OD600 ≈ 0.8) and then shifted to the SD-N medium. Samples were taken at indicated time points and the cleavage of GFP-Atg8 was analyzed by Western blot detection with antibodies against GFP. Detection of Pgk1 was used as a loading control. Quantification of the blots is presented below. GFP ratio (free GFP/Pgk1) was calculated. Error bars reflect SD from the two independent experiments.

(**B**)

### *3.2. Mmi1 Negatively Affects Rapamycin-Induced Autophagy When the Cells Enter Post-Diauxic Growth Phase*

A previous study on HeLa cells demonstrated a negative effect of TCTP on rapamycin-induced non-selective autophagy [11]. Therefore, we also examined the effect of Mmi1 on autophagy induced by rapamycin, a potent inhibitor of the TOR pathway [33,34] that stimulates autophagy [32]. Yeast cells were grown in rich YPD medium until the early logarithmic growth phase. Autophagy was induced by the addition of rapamycin and determined by the phosphatase assay. As shown in Figure 2A, very low autophagy levels were detected at 0 h after rapamycin addition in all strains, demonstrating consistently with our previous results that the basal autophagy is not affected by Mmi1. After prolonged incubation with rapamycin, autophagy levels increased but no difference between the WT and the *mmi1*Δ strains was detected 2 and 4 h after the autophagy induction. However, an increased autophagy was detected 18 and 24 h after rapamycin addition to *mmi1*Δ cells (Figure 2A). Low levels of autophagy were detected in the control *atg8*Δ cells in all tested time points. These results indicated that rapamycin-induced autophagy was increased in the *mmi1*Δ strain after a long term incubation with rapamycin.

**Figure 2.** Non-selective autophagy is promoted in the *mmi1*Δ strain after prolonged incubation with rapamycin. (**A**) Exponentially growing WT, *atg8*Δ, and *mmi1*Δ cells expressing Pho8Δ60 (OD600 ≈ 0.8) were treated with rapamycin 200 nM. Samples were taken at indicated time points, proteins extracted, and the specific Pho8Δ60 activity was determined. Results are means ± SD of two independent experiments performed in duplicates (n = 4). The statistical evaluation was performed by using two way analysis of variance (ANOVA). NS; not significant; \*\*\* *p* < 0.001. (**B**) Western blot detection of GFP-Atg8 cleavage in WT, *atg1*Δ, and *mmi1*Δ cells after rapamycin addition. Cells expressing GFP-Atg8 were grown until log phase (OD ≈ 0.8) and then rapamycin was added to a final concentration 200 nM. At indicated time points proteins were extracted and the protein lysates were examined by Western blot. Pgk1 was used as a loading control. Quantification of the blots is presented to the right. GFP ratio (free GFP/Pgk1) was calculated. \*\*\* *p* < 0.01. Error bars reflects SD from three independent experiments.

To corroborate the results, we also performed GFP-Atg8 processing assays. Consistently, similar levels of the free GFP were detected in the WT and *mmi1*Δ cell lysates at early time points upon autophagy induction, and no free GFP was detected in the control *atg1*Δ cell lysates at all tested time points (Figure 2B). However, an increased accumulation of the free GFP was detected in the *mmi1*Δ cell lysate 24 hours after rapamycin addition. The increased GFP-Atg8 processing in *mmi1*Δ strain 24 hours upon rapamycin addition could be also confirmed by fluorescence microscopy (Figure 3). The majority of the GFP-Atg8 signal in the *mmi1*Δ cells was present in the vacuole compared to WT cells, while in the *atg1*Δ cells the majority of the GFP signal was present outside of the vacuole (Figure 3A, B). These results, together with the results of the phosphatase assay demonstrate the increased autophagy in the *mmi1*Δ strain at later time points after rapamycin addition.

**Figure 3.** Fluorescence microscopy detection of a higher GFP-Atg8 processing in *mmi1*Δ cells upon rapamycin treatment. (**A**) The same samples of WT, *atg1*Δ, and *mmi1*Δ cells treated with rapamycin for 24 h and analyzed by Western blot in Figure 2B were examined under fluorescence microscope. Vacuoles were labelled by the red fluorescence probe FM4-64 (1 μg/mL, 1h) and GFP signal was taken under the same exposure time in all tested strains. DIC, differential interference contrast. Scale bar, 5 μm. (**B**) Quantification of the GFP-Atg8 cellular distribution 24 h after rapamycin addition in WT, *mmi1*Δ, and *atg1*Δ strains. Results are means from two independent experiments and each bar represents 250 cells.

To verify that the phenotype seen in the *mmi1*Δ mutant strain was not due to an unknown secondary mutation, we created a new strain (CRY2959) possessing the wild-type *MMI1* gene under control of its endogenous promotor in the *mmi1*Δ strain. As shown in Figure 4, the presence of *MMI1* wild-type gene decreased the autophagy activity of *mmi1*Δ strain to the WT strain, suggesting that the mutant phenotype is indeed the result of deleting the *MMI1* gene.

**Figure 4.** The increased rapamycin-induced autophagy in the *mmi1*Δ strain can be rescued by wild-type *MMI1* gene. Exponentially growing cells of the WT, *mmi1*Δ, and *mmi1*Δ *MMI1* strains expressing Pho8Δ60 (OD600 ≈ 0.8) were treated with 200 nM rapamycin for 24 h and the specific Pho8Δ60 was determined. Results are expressed as means ± SD from two independent experiments performed in duplicates (n = 4). The statistical evaluation was performed by using two way analysis of variance (ANOVA). \*\*\* *p* < 0.001; NS; not significant.

Our results demonstrated the increased autophagy in *mmi1*Δ cells after a long-term incubation with rapamycin. However, no similar effect was detected in the case of the nitrogen starvation-induced autophagy. Nitrogen starvation and rapamycin treatment are widely used approaches to induce autophagy in yeas<sup>t</sup> [31,32]. In the nitrogen starvation conditions, yeas<sup>t</sup> cells complete division and arrest in the G1/G2 quiescence state [52], increase their volume due to enlarged vacuole by autophagy induction, and remain viable for two days [12]. Rapamycin treatment is believed to mimic nutrient deprivation, including greatly reduced cell growth [33] and autophagy induction [32]. However, it has been demonstrated independently by several groups that even at high concentration of rapamycin, yeas<sup>t</sup> cells maintain their proliferative ability [47,53–56]. We used batch cultivation, and cells were grown in rich YPD media to OD600 ≈ 0.8 before rapamycin was added to a final concentration 200 nM, or cells were washed and shifted to nitrogen starvation media. Both approaches induced autophagy as shown earlier (see Figures 1 and 2). Nevertheless, as shown in Figure 5, rapamycin and nitrogen starvation-treated cells highly differed in the cell growth. In nitrogen starvation medium the WT and the *mmi1*Δ strains reached only OD600 ≈ 2. In fact, the OD600 value in the SD-N medium seems to be overestimated likely due to an increased size of the cells. Indeed, when the number of cells per ml was measured, the increase of about 20% only was detected (Supplementary Figure S1). On the other hand, WT and *mmi1*Δ strain treated with 200 nM rapamycin exhibited higher growth rates and the cell cultures reached OD600 ≈ 7 after 24 h of cultivation after rapamycin addition. These results demonstrate a striking difference between used nitrogen starvation conditions and rapamycin treatment in batch-cultivated yeas<sup>t</sup> cells. To characterize the cell growth during nitrogen starvation and rapamycin treatment, we measured concentration of glucose in media. As showed in Figure 5B, glucose was still present even 24 h upon the shift to the SD-N media. However, glucose was completely exhausted in the YPD media at latest 15 h after rapamycin addition (Figure 5B). Since *S. cerevisiae* cultivated in the glucose-containing media undergoes a growth arrest (diauxic shift) after glucose depletion, no presence of glucose in the media demonstrates that the cells have already entered the post diauxic growth phase. To correlate glucose depletion with autophagy induction, we measured phosphatase activity in indicated time points I and II as shown in Figure 5A,B. The time point I represented an exponential phase, where glucose was present in media and the time point II represented the post-diauxic growth phase; where also glucose was absent from the media (Figure 5B). Our results indicate that autophagy is promoted in the *mmi1*Δ strain after rapamycin treatment, when glucose is already not present in the media (Figure 5C). These results altogether demonstrate that the rapamycin-induced autophagy is increased in the *mmi1*Δ strain after diauxic shift.

**Figure 5.** *Cont*.

**Figure 5.** Increased rapamycin-induced autophagy in *mmi1*Δ strain occurs after glucose exhaustion. (**A**) Exponentially growing WT and *mmi1*Δ cells expressing Pho8Δ60 (OD600 ≈ 0.8) in the YPD medium were either shifted to the SD-N medium or treated with rapamycin (200 nM). Optical density at 600 nm was measured at indicated time points. Results are presented as means ± SD of three independent experiments performed in duplicates (n = 6). (**B**) Concentration of glucose in the media as shown in A was measured after the addition of rapamycin (200 nM) to WT and *mmi1*Δ strains or after the shift of the strains to SD-N media. Results are means of two independent experiments performed in triplicates (n = 6). (**C**) Pho8Δ60 assay was measured in time points I and II as indicated in A, and B. Point I represents the exponential growth phase where glucose is still present in the medium. Point II represents the post-diauxic phase where glucose is already exhausted. Results are normalized to the WT strain (100%) and represent means ± SD from two independent experiments performed in duplicates (n = 4). The statistical evaluation was performed by using two way analysis of variance (ANOVA). \*\*\* *p* < 0.001.

### *3.3. In Post Diauxic Growth Phase Amount of Superoxide Radicals is Decreased in mmi1*Δ *Strain*

Our results demonstrated that Mmi1 negatively affects the rapamycin-induced autophagy in the post diauxic growth phase. Rapamycin forms a complex with FKBP (Fpr1 in yeast) to inhibit TORC1 [33]. Neklesa and Davis reported that superoxide anions regulate TORC1, and its ability to bind Fpr1:rapamycin complex in *S. cerevisiae* [47]. According to the authors, elevated levels of superoxide anions modify TORC1 that it is no longer able to fully bind Fpr1:rapamycin complex. To test the possibility that the increased rapamycin-induced autophagy in the *mmi1*Δ strain results from a lower pool of superoxide anions, we stained the WT and the *mmi1*Δ cells with dihydroethidium (DHE), a superoxide anions sensitive probe. The exponentially growing cells were treated with rapamycin for 18 h to reach the post-diauxic growth phase, then the cells were labeled with DHE, and analyzed by FACS flow cytometer. As shown in Figure 6, *mmi1*Δ cells exhibited significantly decreased amount of superoxide anions compared to the WT cells. This suggested that the decreased pool of superoxide anions in *mmi1*Δ cells might contribute to the stronger interaction between rapamycin/Fpr1 and TORC1 and, hence, it might promote the non-selective autophagy. To further test this hypothesis, we analyzed sensitivity of the WT and the *mmi1*Δ strains to rapamycin. We assume that the *mmi1*Δ strain should be more sensitive to rapamycin compared to WT strain if the decreased amount of ROS in *mmi1*Δ strain facilitate binding of Fpr1:rapamycin complex to TORC1. We cultivated the WT strain and the *mmi1*Δ strain in the presence of increasing concentrations of rapamycin, and calculated survival curves as described in [49]. As shown in Figure 7, the WT and the *mmi1*Δ strains exhibited the same sensitivity to rapamycin. These results indicate no correlation between the decreased ROS production and sensitivity to rapamycin in the *mmi1*Δ strain. This suggests that the decreased ROS production in the *mmi1*Δ strain likely does not facilitate the Fpr1:rapamycin binding to TORC1.

**Figure 6.** In post diauxic growth phase the *mmi1*Δ strain exhibits lower amount of superoxide anions compared to the WT strain. The amount of superoxide anions was determined in the WT and *mmi1*Δ strains 18 h after rapamycin treatment (200 nM). The production of superoxide anions was measured by dihydroethidium (DHE, 15 μg/mL, 1 h, 30 ◦C) and samples were analyzed by the FACS flow cytometer. Results are means ± SD of two independent experiments performed in triplicates (n = 6). The statistical evaluation was performed by using two way analysis of variance (ANOVA). \*\*\* *p* < 0.001.

**Figure 7.** Sensitivity to rapamycin is not influenced in the *mmi1*Δ strain. WT and *mmi1*Δ cells were incubated in the presence of an indicated concentration of rapamycin. Survival curves for the WT and the *mmi1*Δ strains were generated from outgrowth curves and represent the average viabilities of four biological replicates for each strain. Results are means ± SD of two independent experiments performed in duplicates (n = 4).
