Team:Tartu TUIT/Results

Time-lapse Microscopy Shows Autolysis of Yeast Cells

We have tested six strains in time-lapse microscopy experiments following the cells for up to 20 hours after induction of glucanase genes from the estradiol-activated promoter. The strains were expressing the following glucanases: yeast Scw4 and bacterial enzymes Bba, α-Bba, Ost1-Bba, α-Glc1 and Ost1-Glc1.

The experiments revealed that an increase in cell size could be detected at around 5 hours after the start of glucanase induction. The origin of the slow responsiveness could be that the estradiol induction in general shows relatively slow target protein accumulation dynamics [1]. The overexpression of the yeast glucanase Scw4 had a minor effect on cell morphology, however, it did induce lysis of very few cells (Fig. 1A). A more defined effect can be seen in the strains that express bacterial genes (Fig. 1B-F). In the presence of estradiol these cells have abnormal morphology: the cells are enlarged, have swelled vacuoles, some are elongated, possibly due to cell wall defects.

In the later stages of the experiment, at around 8 hours from the start of the induction, cells start dying rapidly. There were cells with defined lysis, in which case the release of intracellular material could be detected (see Movies 1 and 2). Regrettably, we were unable to get all cells lysed even after 20 hours of induction. The most probable explanation is that the secretion signals we used were not very effective. 20 hours after the start of glucanase induction, about 7-13% of cells had died, whereas the percentage of dead cells was around 1% or less in the positions without estradiol (Fig. 2, Table 1). No significant difference was noticed between the four strains with the combinations of bacterial genes and secretion signals. However, the addition of secretion signals to Bba has enhanced its activity, as cells expressing Bba fused with secretion signals grew at a lower rate and the percentage of dead cells by the end of the experiment was increased (Fig. 2).


Movie 1. The effect of Ost1-GLC1 expression was followed with time-lapse microscopy. On the left uninduced cells, on the right Ost1-GLC1 expressing cells.


Movie 2. The effect of Ost1-BBa_K2711000 expression was followed with time-lapse microscopy. On the left uninduced cells, on the right Ost1-BBa_K2711000 expressing cells.

Interestingly, when one cell is lysed, the surrounding cells often lyse as well. This can be explained by a large amount of glucanases being released at once. This effect can be valuable as every lysed cell will speed up the process. Although the lysis efficiency is small, the time-lapse microscopy experiments show a proof of concept that controlled yeast self-lysis can be induced by expression of glucanases.







Figure 1. Time-lapse microscopy of six yeast strains grown in glucanase expression non-inductive (- panel, without ß-estradiol) or inductive (+ panel, with ß-estradiol) conditions. Brightfield (BF), GFP fluorescence and merged channel images are shown. Minutes indicate the time from ß-estradiol addition. Different strains contain different glucanase genes expressed from LexA promoter and are represented in following order: (A) SCW4; (B) BBa_K2711000; (C) α-factor pre-pro signal fused to BBa_K2711000; (D) Pre-Ost1-pro-α-factor signal fused to BBa_K2711000; (E) α-factor pre-pro-Glc1 fusion; (F) Pre-Ost1-pro-α-factor-Glc1 fusion.


Figure 2. The effect of glucanase expression on cell viability. The percentage of dead cells was counted from time-lapse microscopy experiments 20 hours after the start of experiment. The experiment was performed with six strains expressing different glucanases from the estradiol-responsive promoter. The plot shows the percentage of dead cells in the presence or absence of estradiol in the medium. For the number of cells used in the analysis, see Table 1.

Table 1. The number of cells counted for each condition in the calculation of the percentage of dead cells presented in Figure 2.

	Number of cells analyzed
Glucanase	Uninduced culture	Induced culture
α-Glc1	2587	1518
Ost1-Glc1	2355	1797
BBa	1097	379
α-BBa	1504	1383
Ost1-BBa	589	453
SCW4	1120	873

Autolysis Is Inefficient in Liquid Cultures

To test the effects of glucanase expression on yeast cells in liquid cultures, the cells were grown in synthetic complete media with or without of β-estradiol. We included four strains in the assay: the background strain NS202, that expresses only the estradiol-responsive transcription factor, and three strains that contain glucanase genes (Scw4, Ost1-Glc1 and Ost1-Bba) under the control of the estradiol-inducible promoter. In the presence of β-estradiol, the glucanases Scw4, Ost1-Glc1 and Ost1-Bba are induced. The induction led to a decreased growth rate and slightly enlarged cell size; however, no lysis of cells could be detected by measuring the total protein concentration in the media using Bradford assay (data not shown). This indicates that sufficient extracellular levels of the glucanases to induce cell lysis are not reached in liquid cultures.

To test if the increased cell size phenotype detected in the cultures with glucanase expression is accompanied by a weakened cell wall, we tested the resistance of these cells to hypo-osmotic stress and to the presence of DMSO (Fig. 3). The cultures were diluted either in water, to induce hypo-osmotic stress or in YPD with 10% DMSO. DMSO increases membrane permeability, leading to a hypothesis that DMSO might promote lysis of cells with weakened cell walls. However, we did not see significant differences in cell viability in the tested stress conditions (Fig. 3). We note a decrease in the growth rate of cells diluted in H2O, but as this is similar for all tested strains and conditions, this is presumably due to cell cycle arrest caused by nutrient depletion.
We presume that the inability to achieve cell lysis by glucanase induction in liquid cultures is due to low secretion efficiency. We detected some lysis in cells grown on agar pads (Fig. 1), suggesting that when cells are grown on agar, the local concentration of the secreted enzymes increases around the cells due to limited diffusion. In liquid cultures, however, the enzymes are diluted, and the lysis is inefficient.


Figure 3. Viability assay to test the effect of glucosidase expression on resistance to hypo-osmotic stress and DMSO. Yeasts were grown to stationary phase in the absence or presence of 1 µM β-estradiol, which induces the expression of glucanases Scw4, Ost1-Glc1, and Ost1-Bba. The cultures were first diluted to 106 cells/ml either in YPD, H2O or YPD with 10% DMSO, followed by serial dilutions of 4x, 20x and 200x. The dilutions series were plated on YPD plates and grown for 24 hours to estimate the effect of stress conditions on cell viability.

Western Blotting Confirms Low Secretion Efficiency

We used Western blotting to check whether the target proteins are expressed in yeast cells after induction of the synthetic promoter with β-estradiol and to test the efficiency of their secretion. For this, we created strains expressing HA-tagged versions of Glc1 fused to different secretion signals (Ost1 or α-factor). Prior to addition of estradiol to the medium, no expression of Glc1 could be detected, whereas 6 hours after the induction high concentration of Glc1 had accumulated within the cells (Fig. 4). This confirms that the inducible expression is efficient and controllable. We note that Ost1-Glc1 is expressed at significantly higher levels compared to α-Glc1. However, no Glc1 was detected in the supernatant of the centrifuged yeast cell culture (data not shown), although several methods such as acetone precipitation of proteins, immunoprecipitation of 3HA-Glc1 and ultrafiltration were used to concentrate HA-tagged Glc1 from supernatant. This shows that the extracellular concertation of HA-tagged Glc1 protein was below the detection limit, presumably due to inefficient secretion of the Glc1 protein even with the added yeast secretion signals.

Surprisingly, both full-length secretion signal-Glc1 fusion and Glc1 with cleaved secretion signal were detected in cell lysates (Fig. 4). The pre-pro secretion signals are cleaved off in two stages: first, the 19 (α) or 22 (Ost1) amino acid pre signal is cleaved in ER, later, the 66 amino acid pro-signal is cleaved in Golgi [2]. Based on the relatively small difference in electrophoretic mobility, the shorter detected Glc1 fragment likely still contains the pro-signal, indicating that the protein accumulates in the secretory pathway. It has been noted before that overexpression of recombinant proteins can lead to overloading the secretion machinery and specific strains have been designed to overcome this problem [3].


Figure 4. Western Blot analysis to test the expression and secretion of Glc1. Western blot image showing the expression of α-3HA-Glc1 and Ost1-3HA-Glc1. The image shows the amount of the protein in cell lysate. The lysate was diluted 20 and 40 times to visualize the two fragments and to avoid signal leak to neighboring lanes.

GFP Leakage Assay To Measure Cell Lysis

To evaluate the efficiency of cell wall lysis after glucanase overexpression, we introduced pTDH3-EGFP cassette into four strains carrying different glucanase variants. It was expected that if induction of glucanase expression by addition of ß-estradiol leads to cell wall degradation, we might see an increase in the GFP fluorescence in the cell culture supernatant. This method was suggested by experts outside the team and was thought to be more sensitive than the Bradford assay used above.

Ost1-GLC1, BBA, Ost1-BBA and α-BBA carrying strains were pre-grown overnight in liquid cultures and then diluted to the same optical density with fresh medium with (inductive condition) or without (non-inductive conditions) 1 μM β-estradiol. After 8 hours of induction, GFP fluorescence was measured in liquid cultures, cells resuspended in the fresh medium after centrifugation and cell culture supernatant. All measurements were performed with the BioTek Synergy MX microplate reader.

Unfortunately, we did not observe any statistically significant differences in GFP fluorescence in the yeast culture supernatant between induced and non-induced cultures (data not shown). The reason behind this might be very low lysis efficiency, leading to very little release of GFP to the medium, which is under the detection limit of the fluorescence reader. After induction with estradiol, the glucanases are secreted outside of the cells and are diluted by the medium. In this case, the enzymes’ concentration in cell wall proximity does not reach the threshold necessary for cell wall hydrolysis. Evidence for this assumption comes from time-lapse microscopy experiments. When yeast cells are grown on agar pads, glucanases, secreted outside the cells, stay next to the cell wall due to limited diffusion, resulting in higher local concentration and activity. This led to the development of cell wall disruption phenotype (significantly enlarged or popped-out cells) in induced cells (Ost1-GLC1 and Ost1-BBa strains).

Future Plans

There are several aspects that can be improved in our project.

To begin with, LexA-ER-AD expression system has proved to be not the most suitable for this purpose, since the response is relatively slow and therefore it does not allow a quick outburst of the enzymes. Additionally, we found that estradiol also affects the growth of cells that do not contain the LexA promoter. This is a toxic of the expression system and can lead to misinterpretation of some results. One possibility is to use light inducible promoters, as these are not only faster but also easier to manipulate and more convenient for a large-scale production.

Another improvement that can be made is experimenting with other yeast secretion signals and other S. cerevisiae strains, which are optimized for protein secretion [3]. Western blotting revealed that in our strains, the enzymes were still mostly kept inside the cell and potentially had accumulated in ER or Golgi. The secretory efficiency of different recombinant proteins cannot be predicted, as it depends on the folding rate of the protein [4]. To overcome this limitation, a screening system that connects secretion of the protein of interest to cell growth has been developed [5]. This system enables identification of optimal secretion signals and could be used to improve the secretion of bacterial glucanases in yeast. Another possibility is to directly target the glucanases to the cell wall using GPI anchors [6], however, as for this the glucanase has to be fused with a GPI anchor, it might have a negative effect on the enzymatic activity.

Also, Zymolyase, commonly used for enzymatic dissolution of yeast cell wall, contains protease activity in addition to glucanase activity. The protease activity is needed to degrade the mannoproteins in the cell wall [7]. Induction of these protease genes in addition to glucanases might further improve the autolysis process.

In addition, some knockouts can be made to weaken cell response to the cell wall stress. Ideally, these should be conditional knockouts that are coregulated with the induced expression of glucanases. Some potential targets could be the genes from HOG (High Osmolarity Glycerol response) pathway or CWI (Cell Wall Integrity) signaling pathway, since these pathways have been previously reported to be responsible for the response to the cell wall stress caused by Zymolyase [8] [9].

Another possibility is to knock out the genes responsible for the synthesis of other cell wall components, as there is always a risk that cell will adapt to the stressful conditions via enhancing the production of other ones, or β-1,6-Glucan itself. Some examples are chitin synthases: Chs1, Chs2, Chs3 and β-1,3-Glucan synthases: Fks1, Gsc2 [10] [11].

We can also continue trying out the deletion of genes responsible for assembly and cross-linking, such as GAS1 (β -1,3-Glucanosyltransferase) [8] [10] [11].

References: