Evolutionary engineering of Wickerhamomyces subpelliculosus and Kazachstania gamospora for baking

The conventional baker’s yeast , Saccharomyces cerevisiae , is an indispensable baking workhorse of 15 all times. Its monopoly coupled to its major drawbacks such as streamlined carbon substrate 16 utilisation base and a poor ability to withstand a number of baking associated stresses prompt the 17 need to search for alternative yeasts to leaven bread in the era of increasingly complex consumer 18 lifestyles. Our previous work identified the inefficient baking attributes of Wickerhamomyces 19 subpelliculosus and Kazachstania gamospora as well as preliminarily observations of improving 20 fermentative capacity of potential alternative baker’s yeasts using evolutionary engineering. Here we 21 report the characterisation and improvement in baking traits in five out of six independently evolved 22 lines incubated for longer time and passaged for at least 60 cycles relative to their parental strains as 23 key strains. evolved the best characteristics attractive for alternative baker’s yeasts as compared to the evolved K. gamospora strains. These results demonstrate the robustness of evolutionary engineering in development of alternative baker ’ s yeasts.


Investigation of fermentative capacity 136
To investigate the improvement in fermentative capability of the evolved strains, fermentation was 137 carried out using YPM in 60 mL BD Luer-Lok™ syringes (BD® Syringes) as described in our 138 previous work (Zhou et al., 2017). A single colony (selected based on colony size) from each of the 139 evolved lineages of W. subpelliculosus (Ws_1, Ws_2, and Ws_3) was grown overnight in 2 mL of 5 140 % YPM in 5 mL culture tubes at 26 °C at 200 rpm on a shaker (Infors HT). The yeasts were then 141 harvested, washed and used to inoculate 5 mL YPM in syringes at an initial OD600nm of 1 and 142 incubated under the same conditions as above. The plunger movement, as CO2 was accumulated, was 143 6 recorded after every 2 hours for 20 hours. CO2 production yields were calculated by dividing the 144 amount accumulated at the end of fermentation by the biomass accumulated. In addition, CO2 145 production rate was calculated by determining the slope of the curve using the points at which the 146 accumulation of CO2 was the fastest. The same procedure was repeated with evolved lineages of K. 147 gamospora (Kg_1, Kg_2 and Kg_3), both ancestral strains and a control conventional baker's yeast. 148 These experiments were done in triplicates and repeated three times. 149

Investigation of leavening ability 150
The evolved yeast isolates (Kg_1, Kg_2, Kg_3, Ws_1, Ws_2, and Ws_3), the ancestral strains and the 151 control conventional baker's yeast were grown in 2 mL of 5 % YPM in 5 mL tubes and incubated 152 overnight at 26 ºC at 200 rpm on a shaker (Infors HT). The yeasts were then harvested by 153 centrifugation and inoculated into fresh 20 mL YPM and put back on the shaker for another 24 hours 154 to increase cell biomass. 2 mL of cells at an OD600nm of 10 was used to inoculate 10 g of flour dough 155 in Falcon tubes. The Falcon tubes were left to ferment for 1 hour in a 30 ºC incubator. The change in 156 leavening was assessed as dough increased in volume. Respective volume increases after incubation 157 were recorded by photographing. The experiment was done in triplicates and repeated three times. 158 159 2.7 Investigation of stress tolerance 160 The evolved yeast isolates (Kg_1, Kg_2, Kg_3, Ws_1, Ws_2, and Ws_3), the ancestral strains and the 161 control conventional baker's yeast were grown overnight in 5 % liquid YPM in 5 mL tubes at 26 ºC 162 at 200 rpm on a shaker (Infors HT) as above. Cells were then harvested by centrifugation and then 163 washed twice with sterile deionized water. The cells were then adjusted to an initial OD600nm of 0.2. 164 The cells were then serially diluted (2 folds dilution ranges) in sterile phosphate buffer saline and 165 triplicates and repeated thrice. The best representative plates were scanned and recorded. 177

Baking trials 178
Leavened dough from the evolved yeast isolates (Kg_1, Kg_2, Kg_3, Ws_1, Ws_2, and Ws_3), the 179 ancestral strains and the control conventional baker's yeast from Section 2.6 were used as starter 180 cultures. 50 g of flour was added to each of the doughs and weighed before and after fermentation to 181 determine the percentage change in weight. After fermentation, the dough was kneaded and moulded 182 into greaseproof muffin moulds. The leavened and moulded dough was baked for 20 minutes at 250 183 ℃ and 15 minutes in a conventional oven until the bread developed a brownish crust. After baking, 184 the bread was weighed and the percentage change in weight was recorded. The overall texture and 185 pore sizes were photographed and recorded. 186

Statistical Analyses 187
To test whether independently evolved clones and the controls had significantly different 188 fermentative capacity as a function of CO2 production rate, CO2 yields and cell-densities, one-way 189 ANOVA was conducted. To test whether the same attributes of independently evolved lines 190 significantly differed from each other, we implemented a post-hoc Tukey's HSD. The significance 191 level was set at p < 0.05, p < 0.01 and p < 0.001. All the analyses were done using STATISTICA, 192 version 13.2 (Statsoft Inc., Tulsa, Oklahoma).  Figure 1A). In contrast, there was strangely no evident change in one of the six evolved clones from 216 one of the parallel lines analysed (Kg_2). Analyses of ITS -ITS4 amplicons suggested that the clone 217 was still K. gamospora, which probably lost the ability to ferment maltose efficiently. Overall, our 218 approach was very effective as we observed 4.8 times more CO2 accumulated when compared to the 219 control conventional baker's yeast (10.3 ± 1.15 mL) within the first 18 hours of incubation. These 220 results suggested that the ancestral strains and the control baker's yeasts are characterised by a longer 221 lag phase during the utilisation of maltose as they later on managed to accumulate more carbon 222 dioxide similar the amounts produced by the evolved clones (results not shown). 223 In addition to the ability to ferment maltose, we evaluated CO2 yield of the evolved clones as another 224 important attribute required in leavening the dough. The results showed that there was a significant 225 improvement in CO2 yield among 5 of the 6 evolved clones (Kg_1, Kg_3, Ws_1, Ws_2 and Ws_3) 226 when compared to the ancestral strains as well as the control baker's yeast (ANOVA, p < 0.001) 227 ( Figure 1B). The evolved clones exhibited a CO2 yield that was eight times higher than that of the 228 ancestral strains, which is a significant improvement. Again, Kg_2 was an outlier. Our approach 229 improved CO2 yield of 5 out of 6 evolved clones to two times more than that of the control baker's 230 yeast, suggesting that the evolved clones would be a preferable alternative baker's yeasts for the 231 baking industry. Interestingly, there was no significant difference in CO2 production, production rate 232 and yield among these evolved clones (ANOVA, p < 0.001) (see Appendices, Table 4-Table 6) 233 suggesting that the approach is independent of the background of the ancestral strain, which is a 234 9 positive attribute to adopt the same strategy to other yeast species of interest. Another important 235 attribute of a model baker's yeast, the gassing power (CO2 production rate), which reduces the time 236 taken to leaven dough, an important techno-economic factor (Giannone et al., 2010) was tested. We 237 observed a similar trend on the CO2 production rate, also known as the gassing power, an important 238 attribute determining the speed of dough leavening, among the 5 of the 6 evolved clones in 239 comparison to their ancestral strains and the control baker's yeast ( Figure 1C). There was a highly 240 significant improvement in the gassing power of the 5 of the 6 evolved clones when compared to 241 ancestral strains as well as the commercial baker's' yeast (ANOVA, p < 0.001). In addition, there 242 was no statistical difference on gassing power attribute tested among the 5 out of 6 evolved clones 243 (Tukey´s HSD, p < 0.001) (see Supplementary Materials, Table 8 - Table 10). at all this strain should be adopted for use in the baking industry. On the other hand, strains Ws_1, 254 Ws_2 and Ws_3 leavened the dough to double the volume when compared to their ancestor as well as 255 to the control baker's yeast ( Figure 2B). These strains evolved from the W. subpelliculosus lineage 256 and showed an even higher leavening ability as compared to the K. gamospora lineages (Kg_1, Kg_2, 257 Kg_3). 258

Evolved clones improved baking associated stress tolerance 259
Other than fermentative capacity, the ability to withstand baking associated stresses from biomass 260 production to baking is another desirable attribute of a baker's yeast. Baker's yeast may encounter 261 stresses during baking and storage high osmotic pressure, high oxidative stress, high/low 262 temperatures, ethanol stress among others (Attfield, 1999). We firstly investigated the ability to 263 withstand ethanol, a product of dough fermentation, as an important attribute that allows higher 264 10 efficiency of leavening ability. The evolved clones Ws_1, Ws_2 and Ws_3 were resistant to ethanol 265 up to 9 %, which is 2 % higher than the amount tolerable to the conventional baker's yeast (Figure  266 3). This was a huge improvement in ethanol stress tolerance, as their ancestor did not grow on 5 % 267 ethanol. On the other hand, the evolved lines from the K. gamospora background (Kg_1 and Kg_3) 268 tolerated only up to 5 % ethanol. An interesting observation was that Kg_2 in addition to its poor 269 utilisation of maltose as a carbon source it also did not tolerate ethanol when compared to its 270 ancestor. 271 Another important attribute of a baker's yeast is the ability to withstand high temperatures. baker's yeast was investigated. Our findings show that Kg_2 was more osmotolerant, a trait shared 280 by the parental strain (Anc Kg), than all its evolved counterparts ( Figure 5). In addition, a similar 281 trend was also noted for halotolerance ( Figure 6). Oxidative stress tolerance was also evaluated as a 282 critical attribute of a baker's yeast because yeasts are exposed to reactive oxygen species generated 283 during dough fermentation. Our results suggest that our approach improved oxidative stress tolerance 284 of all evolved lines (Figure 7). suggest that the highest loaf was attainable using W. subpelicullosus derived strains (Ws_1, Ws_2 and 296 Ws_3) (7.3 ± 0.36 cm) (Figure 8). 297 Another important factor in final quality of the bread is the pore sizes, which influences the texture of 298 bread. Bread baked with evolved clones had much more bigger and uniform pore sizes when 299 compared to bread baked with the ancestral strains as well as the control conventional baker's yeast 300 ( Figure 8). Our evolutionary engineering approach improved the baking traits as we observed an 301 improvement in loaf volume and overall appearance of the bread baked with 5 out of 6 evolved 302 clones when compared to both their ancestors and the baker's yeast. Bread baked with Kg_2 was in 303 agreement to poor attributes of other traits investigated above. 304 To further reveal the change in baking attributes, we investigated the change in weight of dough 305 before and after fermentation as well as that of bread after baking. Change in weight is considered a 306 desirable quality attribute for the best outcome of bread (Sanchez-Garcia et al., 2015). The results of 307 change in weight of dough after fermentation and weight of dough after baking are shown in Figure  308 9. It should be noted that the best producer of CO2 should be the best to leaven dough and, hence, the 309 best in producing the best quality of bread in terms of texture and size. Results show that leavening 310 of the dough and bread baked using the evolved clones lost more weight after baking as compared to 311 their ancestors and the control baker's yeast. yeast. Thermotolerance is one of the most relevant traits because yeasts are subjected to thermal 347 stress during preparation of biomass, transportation and during fermentation of dough (Panadero et 348 al., 2007). Here, we report that evolutionary engineering improved resistance to higher temperatures. 349 Another stress of importance is oxidative stress, which has a well-known effect on dough rheology 350 during bread making (Bonet et al., 2006). Since most yeast biological systems generate reactive 351 oxygen during growth (Sies, 2014), an alternative yeast should develop resistance to this stress. This 352 work suggests that the evolutionary approach exploited led to an improved resistance in oxidative 353 stress. In addition, yeasts fermentation of sugars in the flour dough produces ethanol which can 354 reduce rates of growth, fermentative capacity and cell viability (Nagodawithana et al., 1976). Ethanol 355 production also contributes to the increased rate of H2O2 diffusion into the cells and thereby 356 13 increasing oxidative stress during dough fermentation (Banat et al., 1998). In this sense, it is worth 357 mentioning that we observed an improved resistance to ethanol, extreme temperature and oxidative 358 stresses. In this way, the alternative baker's yeast reported in this work showed the improvement of 359 critical desirable phenotypes. In conclusion, our work highlighted that evolutionary engineering is an attractive tool to improve the 381 baking performance of non-conventional yeasts, which has been a major limitation for entrance into 382 the market monopolised by Saccharomyces cerevisiae. However, further studies are required to 383 reveal the molecular mechanisms behind the observed improvements. Other studies to investigate 384 more attributes such as the ability to withstand other baking associated stresses, changes in aroma 385 complexity after evolution as well as the ability to utilise other carbon sources are required to 386 develop more efficient alternative baker's yeasts.    Table 4). 532 Figure 1 Fermentative capability of evolved clones in comparison to their ancestors. A. CO2 production pro le B. CO2 yield C. CO2 production rate. Anc Kg is an ancestor for Kg_1 Kg_2 and Kg_3 strains, Anc Ws is an ancestor for Ws_1, Ws_2, Ws_3 strains and the baker's yeast as a control. The evolved clones show elevated fermentative capacity within 18 h of fermentation (pro le of fermentation after 18 h was excluded for brevity). This experiment was performed in triplicates and repeated twice. Error bars represent the standard deviation from the mean (see Supplementary Materials, Table 1-Table 3).

Figure 2
Qualitative leavening abilities of strains after being adaptively evolved in our dough. A. Left to right: control with no yeast (NC), control baker's yeast (control), ancestral K. gamospora and evolved clones (Kg_1, Kg_2, Kg_3) B. Left to right: control with no yeast (NC), control baker´s yeast, ancestral W.
subpelliculosus and evolved strains. The evolved clones showed improved leavening ability. Images were taken after an hour of incubation at room temperature.

Figure 3
Ethanol stress tolerance of the evolved clones. Parental strains Anc Kg and Anc Ws, evolved clones and commercial baker's yeast were spotted (OD600nm 0.2, 0.1 and 0.05) for growth on YPM media supplemented with different ethanol concentrations (5 %, 7 %, 9 %, and 10 %). The evolved clones show improved ethanol tolerance as compared to the ancestral strains and conventional baker's yeast.

Figure 4
Thermotolerance of the evolved clones. Parental strains Anc Kg and Anc Ws, evolved clones and commercial baker's yeast were spotted (OD600nm 0.2, 0.1 and 0.05) for growth on YPM incubated at different temperatures (30 ºC, 37 ºC, 40 506 ºC, 42 ºC and 44 ºC). 5 out of 6 evolved clones show improved thermotolerance in contrast with the ancestral strains and the conventional baker's yeast. Kg_2 lost its ability to withstand thermal stress as compared to its ancestor.

Figure 5
Osmotolerance of the evolved clones. 18 Parental strains Anc Kg and Anc Ws, evolved clones and commercial baker's yeast were spotted (OD600nm 0.2, 0.1 and 0.05) for growth on YPS media supplemented with different concentrations of sucrose (50 % and 60 %). Ws_1, Ws_2, Ws_3 and Kg_2 evolved clones retained a similar osmotolerance capability compared to parental strains. Kg_1 and Kg_3 lost the osmotolerance trait.  Oxidative stress tolerance of the evolved clones. Parental strains Anc Kg and Anc Ws, evolved clones and commercial baker's yeast were spotted (OD600nm 0.2, 0.1 and 0.05) for growth on YPM media supplemented with different H2O2 concentrations (3 mM, 4 mM, 5 mM, 6 mM and 7 mM). All the evolved clones showed improved oxidative stress tolerance as compared to the ancestral strains and the conventional baker's yeast.   Table 4).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.