Leaf wax extracted from cauliflower waste shows antitranspirant efficacy

19 Purpose : Excessive transpiration of water from plant leaves can damage crop productivity 20 during droughts, but commercial antitranspirants are expensive. The aim of this research was 21 to characterise extracted wax from brassica leaf waste, and determine its antitranspirant 22 efficacy and economics. Methods : Yield of wax extracted with dichloromethane from six types of brassica waste was 24 measured and the highest yielding waste was selected for bulk extraction with supercritical CO 2 . Wax was compared with a commercially-available terpene antitranspirant (di-1- p - 26 menthene) for efficacy in reducing leaf water vapour loss, measured as stomatal conductance, 27 in three experiments on rapeseed and in one experiment on wheat. Cost of wax under different 28 production scenarios was calculated. 29 Results : Cauliflower leaf waste gave the highest wax yield, with the concentration varying 30 from 1.31% (m/m) to 5.85% (m/m) in different batches of dried leaves. Nonacosane was the 31 main component of the wax. In two of the three rapeseed experiments and in the wheat 32 experiment, stomatal conductance was significantly reduced to similar extents by wax and by 33 di-1- p -menthene, despite the wax being formulated and applied at a much lower 34 concentration. Economic analysis showed that a high wax concentration in the cauliflower 35 leaves would be needed to produce a commercially-viable leaf wax antitranspirant. 36 Conclusion : The results demonstrate biological efficacy as an antitranspirant of extracted 37 cauliflower leaf wax. Further research is needed on variation in wax yield to reliably source 38 high wax concentration leaves and reduce cost of production, and also to understand the 39 greater efficacy of wax than di-1- p -menthene.

the leaf wax was expressed as % m/m yield and GC-MS as described above was used to 140 analyse the wax. 141 142 A 6% (m/m) cauliflower leaf wax formulation was prepared and supplied to Harper Adam 143 University for in vivo plant assessments by adding into hot water wax and glycerol 144 monostearate at a ratio of 1:2 along with 1.0% of Tween 20. This mixture was homogenised 145 at 4000 rpm using a IKA T18 basic Ultra-Turrax until an emulsion was formed. 146 The drying of the leaves was the most time-consuming process during the production of the 147 cauliflower leaf wax and so a final experimental extraction was carried out to evaluate 148 whether the wet leaves could be directly extracted to recover the waxes using the Stage 1 149 conditions in Fig. 1, as previously. 150 151

Plant material and experimental design for rapeseed experiments 153
All the rapeseed experiments were carried out inside an environmentally controlled 154 greenhouse at Harper Adams University. Seeds of rapeseed (cv. Excalibur) were sown in 1 L 155 pots filled with ~ 600 g of John Innes #2 compost at 22 ± 1% volumetric water content 156 (VWC) analysed with a soil moisture probe (ML2X theta probe, Delta-T-device, Cambridge, 157 UK). Three seeds per pot were sown and the pots were thinned to contain one plant at the 2nd 158 leaf stage. The pots were manually watered approximately to saturation on the day that the 159 seeds were sown and no water was applied until the seedlings appeared. After the seedlings 160 appeared, until the application of the watering and AT treatments, the pots were manually 161 watered approximately to saturation every other day. 162 All three experiments consisted of a 2 x 3 factorial design with two watering levels (well-164 watered, WW and water stressed, WS) and three spray treatments (water, di-1-p menthene 165 and wax) in six or eight randomised blocks. The measurements for the three experiments were 166 conducted in November 2016, January 2017 and February 2017 respectively. 167 168

Water management and treatment application for rapeseed experiments 169
The available water content (AWC) in mL of the pots was calculated by plotting a volumetric 170 water content (VWC) -pot weight curve: three pots (filled with ~600 g of compost at 22 ± 1% 171 VWC) were water-saturated and then dried over ten days at 30°C. The VWC by soil moisture 172 probe (ML2X theta probe, Delta-T-device, Cambridge, UK) and the weight by balance (0.1 g 173 resolution, PCB 2500-2, Kern and Sohn GmbH, Balingen, Germany) were recorded daily. For 174 John Innes #2 compost the permanent wilting point and the pot capacity were ~7% VWC and 175 ~45% VWC respectively according to [9]. The total AWC in mL was then calculated as the 176 difference between the weight of the pot at pot capacity (~1000 g) and the weight of the pot at 177 7% VWC (~650 g) measured by moisture probe. 178 179 Before the spray treatments were imposed the surface of each pot was covered by 100 g of 180 plastic beads, so that the water evaporation from the soil surface was minimised. Then the 181 pots were watered until the weight of each pot was 1000 g, so that the pots are at pot capacity. 182 After this date, the plots belonging to the WS regime were not watered. The pots belonging to 183 the WW regime were watered every other day to maintain the pot weight at 1000 g. 184

185
The spray treatments were applied at 4th leaf stage, just after the pots of the WS regime were 186 watered (to pot capacity) for the last time. The three treatments were as follows: water (for 187 control); 1% v/v Vapor Gard (di-1-p-menthene 96%, Miller Chemical and Fertilizer LLC, Hanover, USA) in water; 1% v/v wax in water + 0.5% v/v Wetcit. For Expts 1 and 2, the 189 adaxial surface of the leaves was uniformly sprayed using a small hand-held sprayer until the 190 surface was fully covered. For Expt 3, the plants were sprayed using a custom-built automatic 191 pot sprayer with nozzles at 50 cm height from the plants, 3 bar pressure at 1 m/s speed using 192 was measured with a balance between 8.30 am to 9.30 am. It was assumed that the beads 205 completely blocked evaporation of water from soil. Therefore, plant water use was considered 206 equal to transpiration. Daily transpiration or water use of each pot belonging to the WS 207 regime was calculated as the difference between the weight of the pot on the day and the 208 weight of the pot after 24 hours. Daily transpiration or water use of each pot belonging to 209 WW regime was calculated as the difference between the weight of the pot on the day after 210 watering (which is ~1000 g) and the weight of the pot after 24 hours, before watering. Total adaxial and abaxial gs was measured between 08.30 and 15:00 on the flag leaf of 255 selected tillers on DAS 3 (n=5), 6 (n=5), 9 (n=6) and 12 (n=6) by using a WALZ GFS-3000 256 system (WALZ, Effeltrich, Germany) with a 4 cm 2 cuvette. Daily WU was estimated as the 257 difference in weight after 24 hours. Daily water use was summed to give cumulative water 258 use over the stress period. 259

Statistical analysis 260
All the data were analysed using Genstat (18 th Edition, VSNi, Hemel Hempstead, UK). Data 261 were checked for normality and homoscedasticity following visual assessment of residuals vs 262 fitted values plots. The sum of adaxial and abaxial gs were analysed with a three factor (watering regime x AT x time) randomised complete block repeated measures ANOVA. 264 Since there were no significant interactions between AT and time, the means over all 265 assessment dates are presented. Cumulative water use was analysed by a two factor (watering 266 regime x AT) randomised complete block ANOVA. 267

Wax yield and composition from solvent extraction of different types of brassica waste 269
The yield of wax from the six types of brassica trimming waste was: cauliflower leaves 1. showed that the most abundant compounds in all these samples of waxes were nonacosane, 275 15-nonacosanone and triacontane (Fig. 3). Other compounds detected in the waxes were free 276 fatty acids, long chain alcohols, long chain diols, long chain alkanes, wax esters and sterols. 277

Wax yield and composition from scCO2 extractions 278
The input and output quantities for the two stage sequential extraction of air-dried cauliflower 279 leaves using scCO2 are shown in Fig. 1. The wax obtained in Stage 1 had shown that 280 nonacosane, 15-nonacosanone, triacontane and γ-sitosterol were the principal components and 281 these compounds were again detected as the most abundant compounds in the wax obtained at 282 Stage 2 of the extraction. 283 The economics of the process was evaluated and it was concluded from the poor yield 284 obtained using higher operating pressures and temperatures at the second stage, that a second 285 stage was not economic. 286 In the batch of cauliflower leaves used in the scale-up trial, there was less leaf stalk and leaf 287 blades were larger, leading to a higher dry matter (Fig. 2). A lower yield of wax was obtained in this trial and the principal components were the same as in the two stage trial (Fig. 4). It 289 was seen that there was a decline in the extraction yield of wax throughout the project and this 290 may be due to the seasonal variation of the leaves. 291 When two fresh leaves (92 g) were extracted, the yield was 0.03 g of wax (0.033% m/m), and 292 analysis of this wax showed that the principal components were nonacosane, 15-293 nonacosanone and triacontane (Fig. 4). 294

Wax evaluation experiments 295
The three rapeseed experiments differed in gs and WU (Fig. 5), probably linked to 296 environmental differences dependent on the time of year. For all three experiments, a 297 significant (p<0.001) reduction in both gs and WU from water stress was observed as 298 expected (data not presented). The interaction between spray treatment and watering regime 299 was not significant in all three experiments, therefore only the main effect of spray treatment 300 is presented in Fig. 5. In Expts 1 and 2, both AT and wax reduced gs to similar extents 301 (p<0.001). The effects of AT and wax in Expt 3 were not significant (p=0.267), possibly 302 because all the plants in this experiment had low values of gs. For WU in Expt 1, AT gave 303 only a small (non-significant) reduction, whereas wax significantly reduced water use by 304 17%. WU was not affected by either AT or wax in Expts 2 and 3. 305 306 For the wheat experiment (Expt 4), the spray treatment and watering regime interaction was 307 significant (p=0.039) for gs and the data to show this interaction is presented in Fig. 6. Both 308 AT and wax reduced gs to similar extents in the well-watered plants, but did not reduce gs in 309 the water-stressed plants which had very low values. WU was not affected by AT in Expt 4. 310 In our study this gave more than a fourfold variation in wax yield (from 1.31% m/m to 5.85% 321 m/m) over a 6-month period, which has implications for the economics of the extraction. 322 There appears to be no difference in composition linked to wax yield, consistent with the 323 findings of Baker et al. [10], so the highest concentration should always be selected. Further 324 research to decide the optimal time of year for leaf collection is necessary. 325

326
The cost of extracting functional extracts from biomass is largely determined by product 327 yield, extraction time and the volume to be processed. It appears from our work that the 328 highest wax content is in the leaf blade waste rather than in the stalk or leaf mid-rib, so that 329 there is a clear economic advantage in pre-sorting to remove stalk if possible. The wax yield 330 from fresh trimmings is very low and in addition the high moisture content appears to modify 331 the polarity of the scCO2 further reducing the yield. Fortunately, the leafy material can be 332 easily dried at low temperature (35 o C with high air flow) to give a more easily processed 333 material with a higher wax content. We consider that this step is essential for the process to be 334 economical as only 10% of the biomass is processed with a much higher wax yield compared 335 with fresh material. 336

337
The scCO2 process cost is very influenced by volume and given that the end application could 338 require high volumes of extract, commercial costs for large-scale drying and extraction should 339 be considered. If we consider a scenario where 1000 kg of wax extract is required and the 340 dried biomass has a wax content of 5% we need to extract 20 t of dried cauliflower leaf (133 t 341 wet mass). At this scale the cost/ton input material would be approximately £3,000 so the 342 1000 kg wax would cost £60,000 or £60/kg. If the dried leaf biomass contains only 1.5% then 343 66.7 t (445 t wet mass) needs to be extracted but this would be slightly lower cost due to the 344 higher mass processed (£2,500/t) but the overall cost/kg wax would rise to £167. Conversely 345 if the required mass needed of wax was to rise to 10,000 kg then extraction cost would fall to 346 approximately £1,500/t and so at 5% wax in dried biomass the wax cost/kg would fall to 347 £30/kg. 348 349 There is clearly some optimisation that could be achieved in the biomass selection and 350 preparation and it may be possible to shorten the extraction time a little by optimising the 351 extraction parameters. Using the above estimates for cost per kg, however, an approximate 352 cost of production for a formulated wax AT (based on the wax cost only) at 6% (m/m) as used 353 in this research, would vary from £1.80/l (10,000 kg batch @ 5% wax) to £10.02/l (1,000 kg 354 batch @ 1.5% wax). The active substance (di-1-p-menthene) in the commercial AT used in 355 our study is no longer sold in the UK because of high cost relative to other products, but 356 previously the retail price was £20/l (B. Lewis, Intracrop, personal communication). Only at 357 the lowest wax cost in the above range would this allow a commercially viable AT to be 358 produced (S. Adams, Plant Impact, personal communication), and if other costs are added e.g. 359 transport of raw material to processing site, formulation components, the economics may be 360 marginal. Conversely, the economics could be more favourable if the wax extraction was an 361 integrated part of biomass biorefining and electricity generation. 362