Team:Tartu TUIT/Design
Disruption of yeast cell wall
Yeasts are widely used bioproduction of various small-molecule compounds and proteins. In many cases, the disruption of cell wall is necessary to recover the compounds from the cell. On the industrial scale, yeast cell lysis is mainly performed using high pressure homogenization and bead milling, however both methods are costly and ineffective with samples with high cell concentration [1]. For this reason, we investigated the possibility of inducing self-lysis of yeast cells at a certain point of the bioproduction process.
We focused on the native remodeling of cell wall to see if the mechanisms could be used for inducible disruption of the cell wall. Yeast cell wall is a robust layered structure consisting of polysaccharides and some proteins that provides structural support and ensures protection of the inner cellular content. During normal yeast growth, the cell wall is constantly synthesized, remodeled and degraded locally to enable processes such as cell growth, cytokinesis and mating [2]. This dissolution is catalyzed by the enzymes from glucan hydrolase family – glucanases. Glucanases control hydrolysis of β-1,3-glucans, polysaccharides in the yeast cell wall [3] [4]. Many isoforms of these enzymes are found in the yeast cells and their activities are well-balanced and tightly regulated.
Inducible expression of glucanase genes
At the initial stage of our project, we attempted to unbalance the endogenous system responsible for wall remodeling through overexpression of genes encoding for glucan hydrolase. To be able to initiate cell wall degradation whenever necessary, we decided to regulate the expression of all our target genes with a synthetic promoter pLexA, which can be activated by synthetic transcription factor LexA-ER-B112 [5]. LexA is a DNA-binding domain of bacterial inhibitor protein LexA, ER is a β-estradiol receptor and B112 is a transcription activator. The synthetic factor binds to the promoter with embedded LexA binding sites and activates the transcription only after the conformational change caused by β-estradiol binding to the ER domain (Fig. 1). The advantage of this inducible system is that it does not interfere with metabolism compared to the commonly used galactose induction. Moreover, the level of gene expression depends on the estradiol concentration, providing additional possibility of dosing target enzyme expression [5].
Experimental plan to test the designs
The efficiency of induced cell lysis is tested for both yeasts growing in liquid and on solid medium. The liquid medium represents conditions more related to industrial applications, however, we run time-lapse microscopy experiments to follow the cells on solid media to better understand the dynamics of the changes caused by glucanase induction. In the microscopy experiments, both brightfield and GFP channels are imaged, as GFP enables visualization of dead cells due to drastically increased autofluorescence. From the images, the percentage of dead cells is counted. Also, as diffusion is limited on solid medium, it can lead to higher accumulation of the secreted proteins adjacent to the cells. All experiments are performed by growing the strains in parallel in media with and without estradiol. During the growth, the cell morphology and growth rate are followed. The efficiency of lysis in liquid medium is measured in two ways. First, when cells are lysed, the cellular contents such as proteins are released to the medium, and the total protein concentration in the medium can be measured using Bradford assay. Secondly, as suggested by synthetic biology experts Dr. Paul Grant and Dr. Maria Soloveychik, we run the autolysis experiments with strains expressing GFP, and measure the GFP released to the medium by fluorescence. Additionally, we perform viability assays to estimate how the glucanase induction affects cell wall strength by introducing the cells to hypoosmotic and membrane-permeabilizing stress conditions.
All experiments performed, parts used, successes and failures are documented and can be tracked down in case of need, so that the experiment would be possible to replicate.
Selection of genes for overexpression
Initially, after screening available information, we have chosen four S. cerevisiae genes for overexpression: Exg1, Exg2, Scw4 and Scw11. According to the Carbohydrate Active enzymes (CAZy) database, Exg1 and Exg2 are classified as exo-β-1,3-glucanases. It has been shown that Exg1 and Exg2 enzymes play a role in both β-1,3 and β-1,6-glucan remodelling [2], and overexpression of EXG1 results in the reduction of β-1,6-glucan in the cell wall [2] [6]. Based on sequence homology, Scw4 and Scw11 genes code for glucan hydrolases. It has been reported that deletions of either Scw genes led to cell separation defects during cytokinesis, pointing to their possible role in cell wall degradation [2] [6] [7].
In addition to endogenous glucan hydrolases, we decided to introduce heterologous genes from bacteria, as some bacteria are known to hydrolyze yeast cell wall. At first, we screened the part registry and found the BBa_K2711000 part, an endo-β-1,3-glucanase from Arthrobacter luteus (also known as Cellulosimicrobium cellulans) added by UiOslo_Norway. After further investigation, we found that A. luteus contains other glucanase isoforms, including β-1,3-glucan laminaripentao-hydrolase encoded by Glc1 gene. This enzyme is one of the main components of Zymolyase, a commercial enzyme mixture for yeast cell wall lysis. Glc1 catalyses glucan hydrolysis leading to formation of different products compared to BBa_K2711000. It has also been reported that Glc1 possesses high catalytic activity against yeast cell wall glucans [8]. Therefore, we decided to use these bacterial enzymes for overexpression in yeast cells as well.
To start with, we focused on induced expression of glucan hydrolases to target the main component of cell wall, however, Zymolyase also contains protease activity that targets the mannoproteins in the cell wall [9]. We note that co-expression of glucanases with the proteases might be necessary for efficient cell wall dissolution.
After the initial planning, three target genes (Scw4, BBa_K2711000 and Glc1) were inserted into W303 yeast strain in leu2-3 (SCW4) and trp1-1 (BBa_K2711000 and Glc1) loci. As a result, we obtained three yeast strains with either Scw4, BBa_K2711000 or Glc1 gene under the control of pLexA promoter (Fig. 1). However, microscopy analysis of yeast cells after induction pLexA with estradiol and subjecting them to osmotic stress did not reveal any difference between initial W303 strain and our transformants. Therefore, it was decided to make yeast strains carrying two different glucanase isoforms under pLexA in different loci in following combinations: 1) BBa_K2711000 and Scw4, 2) Glc1 and Scw4. Unfortunately, no changes were observed for the new strains either.
Figure 1. Inducible expression of glucanase genes in S. cerevisiae. We used estradiol-induction for controlled expression of different glucanase genes. When estradiol is added to the media, it diffuses to the nucleus of yeast cells, where it binds to the LexA-ER-B112 transcription factor (TF). This activates the TF, leading to induction of genes-of-interest (GOI) that are under the control of pLexA promoter. The expression of bacterial glucan hydrolases in yeast did not lead to dissolution of cell wall, presumably due to the enzymes being constrained to the cytoplasm, where they cannot act on the cell wall.
Secretion of recombinant proteins in yeast
We did further literature examination and discussed our problems with the advisors and came to the conclusion that the main concern with the bacterial enzymes could be their potential mislocalization in S. cerevisiae cells (Fig. 1). A. luteus glucanase genes contain an N-terminal Tat-type secretion signal. In the bacterial cell, these enzymes are secreted out of the cells into the environment, but Tat-type secretion signals are not used in eukaryotes, suggesting that Glc1 and BBa_K2711000 may not be secreted after translation, and, therefore, might stay in cytoplasm, where they cannot promote cell wall degradation. To achieve efficient lysis, it was important to get the bacterial enzymes out of the cells to enable them to reach cell wall; thus, it was decided to fuse these bacterial genes with yeast secretion signals to make sure the enzymes are secreted.
After reviewing the literature, we have found various secretion signals from different proteins. Considering that likely not all of them would be efficient when fused to different proteins, two different secretion signals were chosen for further work:
1.α-factor pre-pro signal
2.Pre-Ost1-pro-α-factor signal
α-factor pre-pro signal is the most widely used secretion signal in yeast synthetic biology. It works in a way that after translation, the protein to be secreted is transported to the endoplasmic reticulum (ER) in an unfolded form. So, if folding of the glucanases takes place in the cytoplasm right after the translation, this signal might be inefficient. For this reason, we also tried pre-Ost1 signal, which promotes translocation to the ER during translation [10]. We tried both types of secretion signals expecting that at least one of them will ensure efficient secretion of the target enzymes (Fig. 2).
Figure 2. Addition of secretion signals to promote secretion of recombinant glucanases.
Two different types of secretion signals were added to the bacterial glucanases expressed under the control of estradiol-responsive TF. Following translation, the secretion signals direct the proteins to the secretory pathway. During the pathway, the pro- and pre-signals are cleaved and the glucanases are released to the media, where they can reach the cell wall and hydrolyze the glucans.
This approach appeared to be successful. After induction of glucanase expression with estradiol, the strains showed cell wall defects (significantly enlarged cells, some cells were popping out).
Inactivation of cell wall synthesis genes
Remodeling yeast cell wall is a complex process. In this balanced system, cell wall degradation can be counteracted by parallel cell wall synthesis, as the cell responds to the stress caused by glucan hydrolysis [11]. Due to this, our approach of overexpressing glucanases may not be effective due to cell wall synthases activity. To improve the system, we decided to combine glucanase overexpression with a knock-out of the gene encoding for the enzyme responsible for the wall biosynthesis. We have chosen GAS1 gene as its deletion does not affect cell growth but makes cells more sensitive to Zymolyase treatment. The enzyme Gas1 (1,3-β-glucanosyltransferase) is involved in cell wall biosynthesis, catalyzing elongation of 1,3-β-glucan chains in the cell wall [12]. In addition, deletion of the gene led to decrease in the degree of cross-linking of 1,3-β-glucans to other cell wall polymers, weakening the cell wall [7] [13].
Modeling of bioproduction process and lysis induction
We created a model to follow the product accumulation dynamics in a batch culture. The purpose of this model is to determine the time point of glucanase induction with the aim to lyse the cells at the time of maximal product to growth time ratio. The model included experimentally measured constants and the induced lysis timings measured in time-lapse microscopy.
In order to predict the best time point for the induction of yeast autolysis, β-carotene production, selected as an example product, was modelled. The model consists of two parts. First one includes biomass growth rates, glucose uptake, ethanol accumulation and consumption, and β-carotene production. It allows us to find a time point when cell autolysis will be most beneficial. Second part calculates the time point for induction to result in the majority of cells being lysed by the optimal time, using experimental data.
References: