Industrial Crops and Products 30 (2009) 9–16

Contents lists available at ScienceDirect

Industrial Crops and Products

journal homepage: www.elsevier.com/locate/indcrop

Improved methods for extraction and quantification of resin and rubber from guayule

Michael E. Salvucci a,∗, Terry A. Coffelt a, Katrina Cornish b a US Department of Agriculture, Agricultural Research Service, Arid-Land Agricultural Research Center, 21881 N. Cardon Lane, Maricopa, AZ 85238, United States b Yulex Corporation, Maricopa, AZ 85238, United States article info abstract

Article history: Guayule, a shrub native to the Chihuahuan desert, is a natural source of high quality, hypoallergenic rubber. Received 19 September 2008 Unlike rubber trees that produce rubber in laticifers, the rubber in guayule is produced in parenchyma cells Received in revised form 11 December 2008 of the bark tissue of stems and roots. Consequently, guayule tissue must be mechanically broken before the Accepted 18 December 2008 rubber can be extracted and analyzed. Since rubber extraction and analysis is time-consuming, progress towards increasing the rubber content of guayule through breeding or better cultivation practices has been Keywords: limited by the slow rate of sample processing. To address the need for faster and more efficient sample Accelerated solvent extraction throughput, conditions were optimized for automated extraction of dried guayule tissue using accelerated Evaporative light scattering Guayule solvent extraction (ASE) and rapid methods were developed to replace gravimetric determination of resin Parthenium argentatum and rubber content. For resin analysis, ultraviolet absorbance was used to determine resin concentration Resin after ASE of the tissue with acetone or acetonitrile. For rubber analysis, evaporative light scattering (ELS) Rubber was used to determine the amount of rubber recovered after ASE of the tissue with cyclohexane. Extraction of guayule tissue with high latex rubber content verified that the amounts of resin and rubber determined by these methods were similar to the amounts determined gravimetrically. Since these methods automate extraction and increase the speed of resin and rubber quantification, they could be used in combination with ASE to increase the throughput and efficiency of guayule evaluation in germplasm enhancement and agronomic improvement programs. Published by Elsevier B.V.

1. Introduction Whitworth and Whitehead, 1991). In fact, considerable effort was expended during the Second World War towards the development Natural rubber is an economically important biopolymer pro- of guayule as a domestic (US) source of rubber. That effort ended duced by plants. Alternatives to natural rubber can be synthesized after the war, was temporarily revived during the oil embargo of from petroleum products, but the properties of synthetic rubber the 1970s and 1980s, and has only recently been renewed with the are inferior to those of natural rubber for many applications (Beilen commercialization of guayule as a source of hypoallergenic latex and Poirer, 2007). Almost all of the world’s natural rubber comes by Yulex Corporation (Cornish et al., 2001). While a number of from the Para rubber tree (Hevea brasilinesis Muell. Arg). This tree barriers to successful widespread commercialization remain, the is cultivated in tropical regions, with the Asian nations of Malaysia, most significant biological hurdle is the development of high yield- Thailand and Indonesia accounting for over 90% of the world’s pro- ing, stress-tolerant germplasm (Estilai and Ray, 1991; Ray et al., duction of natural rubber. The importance of natural rubber for the 2005). In the past, efforts to improve guayule germplasm through global economy and concerns about the spread of South American classical breeding and selection were complicated by the ability Leaf Blight to Asia and Type I latex allergy have spurred a renewed of guayule to reproduce both sexually and by apomixis. However, interest in developing alternative sources of natural rubber (Beilen with the advent of molecular approaches, including transgenics and Poirer, 2007). (Veatch et al., 2005), marker-assisted breeding and ploidy analy- Guayule, a desert shrub native to the Chihuahuan desert, has sis by flow cytometry, new tools are now available for improving long been regarded as a promising alternative to the rubber tree guayule germplasm (Estilai and Ray, 1991; Ray et al., 2005; Veatch for production of natural rubber (Hammond and Polhamus, 1965; et al., 2005). Evaluation of guayule germplasm for industrial product produc- tion requires high throughput methods for determining the resin ∗ Corresponding author. Tel.: +1 520 316 6355; fax: +1 520 316 6330. and rubber content of the shrub. While NIR methods have been E-mail address: [email protected] (M.E. Salvucci). proposed for measurement of the rubber content in intact plants

0926-6690/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.indcrop.2008.12.006 10 M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16

(Cornish et al., 2004), these methods have not been refined suffi- 2.2. Accelerated solvent extraction (ASE) of guayule tissue ciently for reliable use. Consequently, measurement of resin and rubber content in guayule tissue requires destructive methods that Guayule stem tissue was extracted sequentially with organic extract the tissue with either an aqueous solvent (i.e., for latex rub- solvent using a Dionex ASE 200 (Bannockburn, IL). One gram of ber) or organic solvents (i.e., for resin and rubber). Quantification guayule tissue was mixed with Ottawa sand and loaded into an of rubber in latex has been successfully employed (Cornish et al., 11ml extraction cell. Samples were generally extracted with ace- 1999, 2000, 2005) but accurate quantification is dependent upon tone and then cyclohexane using two static extractions with each stringent post-harvest storage conditions and analysis within two solvent and a volume equivalent to 150% of the residual cell volume weeks to ensure that the rubber does not coagulate, but remains of about 7 ml. However, for some experiments, acetonitrile was used as an emulsion in the latex. In contrast, organic solvent extraction in place of acetone for the first extraction. Extractions were carried can be used with dried tissue samples, and provides information on out at either 24 or 100 ◦C and for the durations indicated in the text both rubber and resin amount. Extraction of dried and ground tissue using a pressure of 1500 psi. Twenty eight to thirty milliliters of with a soxhlet apparatus or a homogenizer has been used in past extract volume was recovered from the two static extractions plus studies with guayule (Spence and Caldwell, 1933; Black et al., 1983; flush and analyzed for resin (acetone- or acetonitrile-soluble) or Hammond and Polhamus, 1965; Jasso de Rodriquez and Kuruvadi, rubber (cyclohexane-soluble) as described below. For some experi- 1991; Coffelt et al., 2009a,b). Once extracted, resin and rubber are ments, additional extraction cycles using the same static conditions usually determined gravimetrically, although photometric meth- of temperature, pressure and duration were performed with each ods based on sample turbidity have been developed (Traub, 1946; solvent. Hammond and Polhamus, 1965; Sundar and Reddy, 2001). In recent years, instruments for accelerated solvent extraction 2.3. Gravimetric determination of resin and rubber content (ASE), also called pressurized liquid extraction, like the Dionex ASE 200 (Dionex Corp., Bannockburn, IL), have become available, pro- Four ml aliquots of the acetone or cyclohexane extract were viding a means for automating the sequential extraction of guayule transferred in triplicate to pre-weighed borosilicate glass vials tissue (Thurbide and Hughes, 2000). An important advantage of ASE (17 mm × 58 mm). The liquid was taken to dryness under vacuum at systems over traditional extraction methods is reduced handling of 65 ◦C in a Savant speed-vac centrifugal concentrator (Thermo Sci- organic solvents. In addition, ASE instruments increase the speed entific, Waltham, MA) and the vials were re-weighed to determine and efficiency of extraction by using high pressure, alone or in com- the dry weight of the extract. bination with high temperature, for solvent extraction (Thurbide and Hughes, 2000). 2.4. Determination of resin content by UV absorption In the present study, we describe the development of protocols for ASE of guayule tissue and for instrumental analysis of resin and To determine the maximum wavelength and extinction coeffi- rubber content. These methods can be used to increase the through- cient of unfractionated guayule resin, resin samples representing put of samples in screening programs designed to evaluate guayule the acetone-soluble extracts from different plant samples were co- germplasm. evaporated with and dissolved in HPLC-grade acetonitrile. The UV spectrum of the material was determined in acetonitrile and an 2. Materials and methods extinction coefficient at 272 nm was developed for both the UV–vis spectrophotometer (Cary 100, Varian Instruments, Inc., Palo Alto, 2.1. Plant material CA) and a Synergy HT microtiter plate reader (Biotek Instruments, Inc., Winooski, VT). Two-year old guayule (Parthenium argentatum Gray, line AZ-2) For routine determination of resin by absorbance at 272 nm, plants were harvested in April, 2008 from a commercial field near 1 ml of the acetone extract from ASE was transferred to a 1.5 ml Eloy, AZ. Plants were defoliated, chipped and screened through a centrifuge tube and centrifuged at 24 ◦C for 5 min at 10,000 × g. 3.8 cm mesh. The chipped material was dried by either lyophiliza- Aliquots of the supernatant (0.1 ml) were transferred to second tion for 24 h after freezing overnight at −80 ◦C, oven drying at 60 ◦C tubes containing 0.9 ml HPLC-grade acetonitrile. The acetone and for 2 days, or air-drying for 4 days at 23 ◦C. The dried material was acetonitrile were co-evaporated to near dryness (0.1 ml) under vac- stored at room temperature in a desiccator. After four weeks of uum for 25 min using a speed-vac at 65 ◦C. Co-evaporation was storage, the dried material was pulverized in a ball mill (SPEX Cer- repeated following the addition of a second 0.9 ml of acetonitrile. tiprep, Metuchen, NJ) by shaking for 3 min in a 30 ml hardened steel Following evaporation, acetonitrile was added to the residual liq- cup containing two 0.6 cm stainless steel ball bearings. The ground uid to a final volume of 1 ml and the solution was mixed using a material, equivalent to <1.7 mm, was stored at room temperature vortex mixer. The absorbance of the solution at 272 nm was deter- for one to two weeks in a sealed vial at room temperature before mined by transferring 0.2 ml aliquots to a 96-well UV-transparent extraction. microtiter plate and immediately measuring the absorbance with For some experiments, guayule plants were grown in growth the microtiter plate reader. The data presented are the means ± SEM chambers under controlled conditions of temperature and pho- of at least two separate ASE extractions for each sample and treat- toperiod to induce or suppress vegetative growth and subsequent ment. Duplicate aliquots of the acetone extract were co-evaporated rubber synthesis (Bonner, 1943; Sundar and Reddy, 2001). To for each ASE sample and the UV absorption of each aliquot was stimulate vegetative growth and flowering and suppress rubber measured in triplicate. synthesis, plants were grown under a 14 h, 30 ◦C light/10 h, 7 ◦C dark photoperiod. To suppress vegetative growth and flowering and 2.5. Analysis of rubber content by evaporative light scattering stimulate rubber synthesis, plants were grown under a 9 h, 16 ◦C (ELS) light/15 h, 7 ◦C dark photoperiod. Plant material was harvested and representative samples of the stem material were either freeze- Rubber content in the cyclohexane extracts was determined dried or immediately extracted for latex rubber (Cornish et al., using an evaporative light scattering detector (ELSD). The outlet 1999). Freeze-dried material from plants with either high (9.6%) of an HPLC pump was plumbed directly into a Varex MKIII ELSD or low (<1%) latex rubber content was pulverized and the ground (Alltech Associates, Inc., Deerfield, IL). A manual injection valve material was stored for six months at room temperature. with a 10 ␮l loop was inserted in-line to introduce the sample. For M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16 11

Fig. 1. Recovery of resin and rubber after multiple cycles of ASE. Resin and rubber were extracted at the indicated temperatures with acetone and cyclohexane, respec- Fig. 2. Absorption spectrum of crude and partially purified guayule resin samples tively. Each cycle consisted of two 20 min static extractions at 150% cell volume. in acetonitrile. Resin from five different guayule plant preparations was partially Percent recovery is based on the amount extracted per cycle compared with the total purified by repeated extraction of dried guayule tissue with acetone (solid lines). amount extracted after three cycles. Resin and rubber amounts were determined Resin samples also were prepared by ASE of dried, ground plant tissue with acetone gravimetrically. (dashed line). All samples were co-evaporated with acetonitrile prior to analysis.

− measurements, cyclohexane was pumped at 0.5 ml min 1 through of the acetone extraction from 24 to 100 ◦C greatly increased the ◦ the ELSD, using a drift tube temperature of 98 Candaflowrate amount of material extracted from tissue with a high latex rub- of 1.6 lpm for the nitrogen gas stream. Triplicate 10 ␮l aliquots ber content. However, a portion of this material precipitated from from each extraction were injected manually into the flow stream solution once the acetone had cooled to room temperature (see and solute particles were detected by the ELS detector. The analog below). The white precipitate was partially, but not completely signal from the ELSD was converted to digital using a Chromtech soluble in water and 100% cyclohexane at 25 ◦C, indicating that (Alltech Associates, Inc.) A/D converter and associated software. it was not resin. The amount of rubber extracted by the ASE sys- The data presented are the means ± SEM of at least two separate tem decreased markedly when high latex tissue was first extracted ASE extractions for each sample and treatment. with acetone at 100 compared with 24 ◦C. Taken together, the data indicate that extraction of guayule tissue with hot acetone either 3. Results dissolves and/or degrades the rubber or makes the rubber unrecov- erable during the subsequent extraction with cyclohexane. 3.1. Accelerated solvent extraction 3.2. Resin analysis by UV absorbance Stem tissue from guayule plants with low and high latex rubber content, determined by aqueous extraction (Cornish et al., 1999), The resin from guayule is a complex mixture of terpenoid com- was used to refine the conditions for ASE. Multiple extraction cycles pounds that have a strong absorbance in the UV (Schloman et al., of dried stem tissue from plants with a high latex rubber content 1983; Sidhu et al., 1995). This property can be used to estimate the using the same solvent established that >90% of the acetone- amount of resin recovered from guayule tissue after ASE extraction soluble material and >88% of the cyclohexane-soluble material were with acetone. However, because of the high UV cut-off of acetone, it extracted at room temperature with a single extraction cycle, i.e., was necessary to evaporate the acetone and to determine the spec- two 20 min static extractions at 150% cell volume (Fig. 1). Increas- tra in acetonitrile, a solvent with a polarity index of 5.8, similar to ing the temperature used for the cyclohexane extraction from 24 to acetone at 5.1, but with a much lower UV cut-off. Fig. 2 shows the ◦ 100 C increased the efficiency of the first extraction cycle. Increas- absorption spectra in acetonitrile of five different samples of resin ◦ ing the extraction temperature from 24 to 100 C also increased the from the guayule bagasse remaining after aqueous latex extraction. total amount of rubber (cyclohexane-soluble material) recovered The samples were partially purified by repeated extraction and con- ◦ after three extraction cycles by about 10%. When 140 C was used centration in acetone, followed by co-evaporation with and dilution for cyclohexane extraction, a precipitant was visible in the extracts in acetonitrile. Also, shown is the UV spectrum of the acetone- and the amount of rubber in the soluble portion of the extract was soluble material from an ASE of two different guayule stem samples ◦ about 10% lower than after extraction at 100 C (data not shown). also in acetonitrile. For all samples, acetone-soluble material from Gravimetric determination of the cyclohexane-soluble mate- guayule had a pronounced peak of absorbance at 250–300 nm, with rial extracted from tissue with high and low latex rubber content maximum absorbance near 272 nm. showed large differences between the amounts of rubber extracted A relationship between the A272 in acetonitrile and the amount from the two types of tissue (Table 1). Increasing the temperature of acetone-soluble material extracted by ASE was developed in order to determine extinction coefficients for the crude resin frac- Table 1 tion from ASE of guayule tissue (Fig. 3). The relationship between Effect of acetone temperature on extraction of resin and rubber from guayule tis- the A272 and the amount of acetone-soluble material determined sue with low (∼1%) or high (∼9.6%) latex content. Resin and rubber content were gravimetrically was linear with both a spectrophotometer and a determined gravimetrically after ASE with acetone at the indicated temperatures ◦ microtiter plate reader, consistent with Beers–Lambert Law. The and cyclohexane at 100 C, respectively, for two 20 min static extracts at 150% cell ␧ volume. extinction coefficient ( ) for the crude resin fraction from the ASE was 8.81 cm−1 (mg ml−1)−1. With the volumes and geometry of the ◦ Latex content Temperature ( C) Resin (%) Rubber (%) microtiter plate reader, the absorbance readings with this instru- Low 24 6.2 ± 0.2 1.1 ± 0.1 ment were exactly half the values determined with the UV–vis High 24 9.1 ± 0.6 4.6 ± 0.1 spectrophotometer. Consequently, the ε value of the crude resin ± ± High 100 16.2 0.8 1.5 0.7 fraction in acetonitrile was 4.45 (mg ml−1)−1 or about 50% of the 12 M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16

Fig. 3. Relationship between resin concentration and absorbance at 272 nm. The Fig. 5. Differences in rubber concentration detected by ELS after ASE. Evaporative A272 of dilutions of the acetone-soluble fraction from ASE (Fig. 2) were determined light scattering was determined in duplicate aliquots of the cyclohexane extract from using a spectrophotometer (solid line) or a microtiter plate reader (dashed line) plants with high (first four peaks) and low latex (second four peaks) content. The after co-evaporation with acetonitrile. The absorbances were compared with the numbers above the peaks correspond to the rubber content determined gravimetri- concentration of resin in solution determined gravimetrically. cally. The last four peaks correspond to the signals from plants with high latex that were extracted with acetone at 100 ◦C prior to extraction with cyclohexane. value determined with the spectrophotometer. The ε for resin from the partially purified preparations of resin was 5.13, only slightly response for a series of injections. Samples with high (>4% dry higher than the value for the crude acetone extract from the ASE. weight) and low (<1% dry weight) latex were easily differentiated. The ELS response of high latex samples that were extracted with 3.3. Rubber analysis by ELS acetone at 100 ◦C resembled the response of samples with low latex, consistent with the results based on gravimetric determi- Rubber can be extracted from dried guayule tissue using nation (Table 1). When ELSD and gravimetric determinations of non-polar solvents like hexane, benzene, petroleum ether or cyclo- rubber content were compared directly, a linear relationship was hexane. Once in solution, the dissolved rubber can be detected established with a slope of 1.15 (Fig. 6). using ELS, which detects the scattering of light caused by particles released into a heated drift tube by nebulization and evaporation of 3.4. Verification of the methods for resin and rubber the mobile phase. Fig. 4 shows the relationship between the area of determination the ELS peak and rubber concentration for a sample of guayule rub- ber dissolved in cyclohexane. The data show that the light scattering Guayule shrubs with a latex rubber content of 5.3% were ana- signal was linear over rubber concentrations from 0 to 2 mg/ml and lyzed for resin and rubber using UV absorption and ELS and the departed slightly from linearity between 2 and 5 mg/ml. results compared with those determined gravimetrically (Table 2). The cyclohexane extract from plants with high and low latex Whereas the plant material used in the experiments described rubber content (Table 1) was analyzed by ELS to determine the above was stored as a powder for six months at room tempera- effectiveness of ELS in detecting differences in the amount of ture prior to analysis, the material used here was analyzed within cyclohexane-soluble material with plant samples extracted by ASE. two weeks of pulverizing. The use of fresher tissue provided a more Fig. 5 shows the raw data from this analysis, i.e., the detector

Fig. 4. Relationship between rubber concentration and ELS. One hundred milligrams Fig. 6. Correlation between gravimetric and ELS methods for determining rubber of solid rubber were dissolved overnight at room temperature in 10 ml of cyclohex- concentration in guayule. Rubber samples were prepared by ASE of ground plant ane and diluted for use as standards, 0–5 mg ml−1. Light scattering was determined tissue with acetone and then cyclohexane. The cyclohexane fraction was analyzed by injecting 10 ␮l of the diluted rubber solutions into the ELSD at a flow rate of in triplicate by ELS or taken to dryness and weighed for gravimetric determination. 0.5 ml min−1. (Inset) Linear relationship between rubber concentration and ELS for To obtain the lowest values, tissue with a high latex content was re-extracted using rubber concentrations from 0 to 2 mg ml−1. The line is a linear regression of the data; a second extraction cycle (see Fig. 1). The line is a linear regression of the data; r2 = 0.995. r2 = 0.989. M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16 13

Table 2 Resin and rubber content of guayule tissue determined gravimetrically and by instrumental analysis after ASE under different conditions. Guayule tissue containing 5.3% latex was dried and extracted with acetone at the indicated temperatures and cyclohexane at 100 ◦C for two 20 min static extracts at 150% cell volume using ASE. The resin and rubber contents were determined gravimetrically and by UV absorption at 272 nm (resin) and ELS (rubber).

Acetone temperature Resin (%) Rubber (%)

Gravimetric UV Gravimetric ELS

24 ◦C 8.3 ± 0.5 7.9 ± 0.8 7.7 ± 0.2 7.7 ± 0.5

100 ◦C17.0± 0.6 – 4.8 ± 0.09 5.5 ± 0.2 11.5 ± 0.2a 10.5 ± 0.5a

a After centrifugation to remove insoluble material. reliable estimate of rubber recovery from samples of known latex rubber content. Under standard extraction conditions, i.e., acetone at 24 ◦C and cyclohexane at 100 ◦C, the amounts of resin and rubber determined using UV absorption for resin and ELS for rubber were similar to the amounts determined gravimetrically. The amount of rub- ber recovered from the dried tissue was 7.7%, which was higher than the latex rubber content measured in a separate aliquot of this tissue, i.e., 5.3%, indicating that some of the latex rubber had coagulated to solid rubber in the plant before or after harvest. Gravimetric determination showed that increasing the temperature of the acetone extraction to 100 ◦C doubled the amount of mate- rial extracted with acetone, but decreased the amount of material recovered after extraction with cyclohexane. As discussed above, a significant portion of the material extracted with hot acetone was no longer soluble when the acetone cooled to room temper- Fig. 7. Effect of extraction time on the efficiency of resin and rubber extraction by ASE. Resin and rubber were extracted for one cycle at 24 and 100 ◦C with acetone and ature. When insoluble material was removed by centrifugation, cyclohexane, respectively. Each cycle consisted of two static extractions at 150% cell the amount of acetone-soluble material was only slightly greater volume for the indicated amount of time. Resin and rubber content were determined when hot acetone was used for extraction. Pre-extraction of dried by UV absorption and ELS, respectively. guayule tissue with water at 100 ◦C, reduced the amount of resin and rubber recovered from the tissue by about 20 to 25% (data not above was extracted for different lengths of times and the amount shown). of resin and rubber determined by UV absorption and ELS, respec- tively (Fig. 7). The results showed that two 5 min static extractions 3.5. Effect of drying method, extraction time and acetonitrile with acetone at 24 ◦C were sufficient for complete extraction of extraction on resin and rubber determination resin. In contrast, increasing the extraction time with cyclohexane at 100 ◦C from 5 to 20 min increased the amount of rubber recovered Using the high latex rubber content shrub described in the pre- by almost 30%. This result was consistent with the results from mul- vious section and the standard method of extraction, the effect of tiple extractions with the same solvent (Fig. 1) which showed that various drying conditions on the amounts of resin and rubber were a second round of two 20 min static extractions with cyclohexane determined both gravimetrically and by UV absorption and detec- recovered nearly 12% more rubber. tion by ELS (Table 3). The results showed that the three methods for The main drawback of the UV absorption method for high drying guayule stem, prior to pulverizing, gave comparable results. throughput analysis of resin is the requirement for co-evaporation Again, the results for resin and rubber content determined by UV with acetonitrile to remove the acetone. To obviate this problem, the absorption and ELS, respectively, were similar to those determined possibility of replacing acetone with acetonitrile for resin extraction gravimetrically. The only difference among the drying methods that was investigated. When guayule tissue was extracted with ace- was consistent with both methods of analysis was resin content tonitrile via ASE at 24 ◦C, the absorption spectrum of the extract ◦ after oven-drying. The analysis showed that oven-drying at 60 C was nearly identical to the spectrum of the acetone extract after caused a small, but significant decrease in the amount of resin co-evaporation with acetonitrile (Fig. 8). This result indicates that extracted compared with the other two methods of drying. acetonitrile extracts the same suite of UV absorbing compounds To determine if shorter extraction times could be used to from guayule tissue as acetone. Determination of resin content by increase throughput, the same high latex-containing tissue used UV absorption showed that the amount of resin extracted was sim- ilar when acetonitrile was substituted for acetone in the primary Table 3 ASE extraction (Table 4). Effect of drying method on the recovery of resin and rubber from guayule tissue. Guayule tissue with 5.3% latex was either freeze-dried, air-dried or oven dried and Table 4 then pulverized prior to ASE with acetone at 24 ◦C and cyclohexane at 100 ◦C for two Effect of primary solvent on resin and rubber content of guayule tissue. Guayule 20 min static extracts at 150% cell volume using ASE. The resin and rubber content of tissue containing 5.3% latex was oven-dried and extracted with either acetone or the tissue were determined gravimetrically and by UV absorption at 272 nm (resin) acetonitrile at 24 ◦C for two 5 min static extractions at 150% cell volume and then and ELS (rubber). extracted with cyclohexane for two 20 min static extractions at 100 ◦C using ASE. Drying conditions Resin (%) Rubber (%) The resin and rubber content were determined by UV absorption at 272 nm and ELS, respectively. Gravimetric UV Gravimetric ELS Primary solvent Resin (%) Rubber (%) Freeze-dried 8.2 ± 0.9 8.4 ± 1.3 8.0 ± 0.3 8.3 ± 0.3 Air-dried 8.8 ± 0.5 8.1 ± 0.3 7.3 ± 0.6 7.8 ± 0.9 Acetone 6.6 ± 0.2 8.0 ± 0.1 Oven-dried 7.4 ± 0.1 6.8 ± 0.3 7.7 ± 0.6 7.8 ± 1.2 Acetonitrile 6.2 ± 0.1 8.4 ± 0.3 14 M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16

consuming, requiring up to 12 h of extraction in each solvent and considerable handling of organic solvents. In addition, the soxhlet method does not easily lend itself to automation. To increase the speed and efficiency of the extraction process, several investiga- tors have replaced the soxhlet extraction with homogenization in blenders or polytrons either in organic solvents or in ethyl alcohol (Hammond and Polhamus, 1965; Black et al., 1983; Coffelt et al., 2009a,b). However, exposure to solvents can be even greater with homogenization methods, especially if the residue is re-extracted with fresh solvent. In addition, homogenization methods, like the soxhlet method, are difficult to automate. Homogenization of fresh tissue in a mixture of water and ethyl alcohol to recover solid rub- ber is less toxic but requires considerable manipulation, including determination of moisture content and deresination of the solid rubber in acetone (Hammond and Polhamus, 1965). Fig. 8. Comparison of the absorption spectra of crude guayule resin fraction Measurement of rubber in latex after extraction in aqueous solu- extracted by ASE with either acetone or acetonitrile. Dried guayule tissue was tion at high pH avoids the problems with organic solvents, but the ◦ extracted at 24 C for two 5 min static extractions at 150% cell volume using ASE. extraction requires fresh material and is still time-consuming and The acetone extract was co-evaporated with acetonitrile prior to analysis. The ace- difficult to automate (Cornish et al., 1999). In general, the content tonitrile extract was diluted 1:10 and then analyzed without evaporation. of rubber in latex should be the same or lower than the total rubber content. For example, when considerable coagulation occurs, e.g., After extraction with acetonitrile, the cyclohexane extract some- in tissue with low water content, the latex rubber content greatly times contained a small amount of insoluble material whereas this underestimates the total amount of rubber in the tissue. In addition, fraction was usually clear after prior extraction with acetone. The extraction of tissue for latex rubber does not provide a measure of cyclohexane-insoluble material was soluble in acetonitrile, sug- the amount of resin in the tissue since resin determination is usually gesting that it was a solute that carried over into the cyclohexane made on a separate dried sample. fraction with the small amount of residual acetonitrile that was In contrast to the homogenization or soxhlet methods, meth- visible in the collection vial. That this material was not in the cyclo- ods based on solvent extraction systems like ASE provide a means hexane fraction after acetone extraction is probably related to the for automating the extraction of plant tissue with multiple sol- lower amount of residual solvent carried over into the cyclohex- vents (Thurbide and Hughes, 2000). The use of high pressure alone ane fraction after flushing with nitrogen. Because of the difference or in combination with high temperature increases the efficiency in boiling points, the volume of solvent carried over from the of solvent extraction, thus shortening extraction time. In addition, primary extraction was much less with acetone than with ace- worker safety is improved since handling of organic solvents during tonitrile. After removing this cyclohexane-insoluble material by extraction is essentially eliminated because extraction with each centrifugation, the rubber content was similar in acetone- and solvent is fully automated. The results from multiple extractions acetonitrile-extracted tissue. showed that most (88–95%) of the acetone- and cyclohexane- soluble material was eluted with just a single extraction regime 4. Discussion for each solvent consisting of two static extractions at 150% cell volume. For resin, the static acetone extractions could be as brief as Widespread commercialization of guayule as an alternative 5 min, while longer cyclohexane extraction times (i.e., 20 min) were source of natural rubber and resin for the bioproducts indus- needed for rubber. Pre-extraction of dried guayule tissue with water try will require the development of improved germplasm with at 100 ◦C did not aid in recovery of either resin or rubber. Instead, increased rubber and resin content (Estilai and Ray, 1991; Ray et al., recovery of both were reduced when tissue was initially extracted 2005). Germplasm enhancement, whether by traditional breeding with water at 100 ◦C, suggesting that the combination of the high and selection or through molecular approaches to trait improve- temperature and pressure generated during aqueous extraction ment, requires rapid and reliable methods for determining rubber with the ASE caused degradation of both the resin and rubber. and resin content in large numbers of plant samples. Unfortu- Temperature can be easily controlled with ASE, facilitating the nately, existing methods for processing and extracting the plant use of high temperatures and well as high pressures to increase the tissue and for analysis of the extracted material are extremely efficiency of extraction. Extracting guayule tissue with hot (100 ◦C) time-consuming, thus limiting the number of samples that can be acetone caused considerable loss of rubber in samples that were screened. In addition, the extraction methods require considerable known to contain high rubber based on latex rubber determinations handling of organic solvents, subjecting workers to organic solvents made with fresh tissue. This effect was noted in early studies and that are potentially dangerous to their health. was attributed to breakdown of rubber into lower molecular weight components (Meeks et al., 1950; Meeks and Feustel, 1951). Thus, key 4.1. Solvent extraction of guayule tissue to quantitative extraction of rubber was the use of low temperature for extraction of the resin with acetone. Using the two-solvent ASE In most early studies (Spence and Caldwell, 1933) and even method, 13 samples can be extracted into the 26-vial carousel. Since more recently (Jasso de Rodriquez and Kuruvadi, 1991), resin the entire extraction process is fully automated, solvent handling and rubber were sequentially extracted from dried guayule tis- is minimal during extraction. sue with polar (i.e., acetone or ethanol) and non-polar solvents (i.e., benzene, petroleum ether, hexane or cyclohexane), respec- 4.2. Analysis of resin and rubber tively, using a soxhlet apparatus. In some of the early studies, the plant material was first extracted with hot water to remove com- Gravimetric methods have generally been used for determining pounds that purportedly interfere with the subsequent extraction resin and rubber content after extraction in suitable solvents (Black of resin and rubber (Spence and Caldwell, 1933; Hammond and et al., 1983). These methods, while not specific for either com- Polhamus, 1965). The soxhlet extraction method is extremely time- pound, are generally acceptable because of the high content of these M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16 15 compounds in guayule tissue. Gravimetric determination requires system was used to provide rapid throughput for rubber analysis. evaporation of the solvent. Although equipment has been devel- Cyclohexane extracts can be analyzed immediately after extraction oped for this purpose, this step is still time-consuming when a high with ASE and the response of ELSD signal to rubber concentration boiling point solvent like cyclohexane is used to extract the rubber. is linear from 0 to 2 mg ml−1 of rubber. Extraction of 1 g of guayule Chromatographic separation provides a more accurate measure of tissue in an 11ml cell with two 20 min static extractions of 150% cell resin and rubber content (Proksch et al., 1981; Kang et al., 2000), volume produced about 30 ml of rubber extract that could be taken but these methods can be even more time-consuming, limiting the directly from the ASE vial and manually injected into the ELSD with rate of throughput. no concentration. For samples with rubber content greater than In the present study, resin content was determined from the about 7%, a simple 1:1 (v:v) dilution with cyclohexane kept read- absorbance of the resinous solution at 272 nm and an extinction ings within the linear detection range. Sampling rates can be as fast coefficient was determined empirically from the absorbance of as 1 per 2.5 min, allowing 24 samples to be analyzed per hour. known amounts of resin extracted from guayule plants. Although the UV absorption method requires repeated co-evaporation of the 4.3. Notes and recommendations about the methods sample with acetonitrile when acetone is used as the solvent, a large number of samples can be co-evaporated simultaneously using a Processing of the guayule tissue to determine resin and rub- vacuum concentrator. Also, measurement of UV absorbance does ber content involved drying freshly chipped shrub, pulverizing the not require precipitation of the resin from solution, making it more dried material in a ball mill and extracting the dried material in an amenable for microtiter plate analysis than photometric analysis ASE system. Good yields of rubber were obtained from dried tissue of sample turbidity (Traub, 1946). Finally, the UV method is more (i.e., stem pieces) stored for four weeks at room temperature prior specific for resin than gravimetric methods since the predominant to pulverizing, whereas significant losses occurred with material resin components in guayule, guayulin A and B and the argentatins that was pulverized and then stored for six months at room tem- (Proksch et al., 1981; Schloman et al., 1983; Sidhu et al., 1995)areUV perature (data not shown). Based on these findings, we recommend absorbing whereas non-resinous, but acetone-soluble compounds that dried tissue be extracted soon after pulverizing. like fatty acid triglycerides and low molecular weight rubber are A single extraction sequence of two static extractions at 150% not. cell volume extracted most of the resin and rubber. Thus, to increase Because of the low volatility of rubber compared with solvent, throughput, a single extraction sequence using a single collection ELS has been used to detect rubber during separation by gel perme- vial should be used for each solvent. The results of the present study ation HPLC (Kang et al., 2000). In the present study, a columnless showed that two 5 min static extractions with acetone at 24 ◦C were

Fig. 9. Flowchart of methods for extraction and analysis of resin and rubber from guayule, optimized for the maximum rate of throughput. 16 M.E. Salvucci et al. / Industrial Crops and Products 30 (2009) 9–16 sufficient for extraction of resin, but longer extraction times, i.e., Acknowledgements 20 min, are required for extraction of the rubber with cyclohexane at 100 ◦C. Thus, ASE of guayule was still a relatively time-consuming The authors wish to acknowledge the expert technical assistance process, requiring nearly 1 h for extraction of each sample with the of Gwen Coyle and Steve Vinyard (USDA-ARS, Maricopa, AZ). two solvents. However, by automating the extraction process, the rate of throughput is increased since the time normally devoted to References extraction can be used to process plant material or conduct analysis of the extracts. Beilen, J., Poirer, B.Y., 2007. Establishment of new crops for the production of natural rubber. Trends Biotechnol. 25, 522–529. Resin analysis by absorption at 272 nm and rubber analysis Black, L.T., Hamerstrand, G.E., Nakayama, F.S., Rasnik, B.D., 1983. Gravimetric analyses by ELS gave comparable results to gravimetric methods. Thus, for determining the resin and rubber content of guayule. Rubber Chem. Technol. gravimetric determination of resin and rubber content can be 56, 367–371. substituted for these more sophisticated methods in laboratories Bonner, J., 1943. Effects of temperature on rubber accumulation in guayule plant. Bot. Gaz. 105, 233–243. without one or more of the specialized instrumentation described Coffelt, T.A., Nakayama, F.S., Ray, D.T., Cornish, K., McMahan, C.M., 2009a. Post- here, i.e., speed-vac, UV spectrophotometer or microtiter plate harvest storage effects on guayule latex, rubber, and resin contents and yields. reader and ELSD. Instrumental analysis of resin and rubber con- Ind. Crops Prod., (2008), doi:10.1016/j.indcrop.2008.06.003. Coffelt, T.A., Nakayama, F.S., Ray, D.T., Cornish, K., McMahan, C.M., Williams, C.F., tent greatly reduced the time required for analysis by eliminating 2009b. Plant population, planting date, and germplasm effects on guayule latex, the need to pre-weigh vials and caps before extraction and by not rubber, and resin yields. Ind. Crops Prod. 29, 255–260. requiring complete sample dry-down. Cornish, K., Myers, M.D., Kelley, S.S., 2004. Quantification of rubber latex in homogenate and purified samples using near infrared spectroscopy. Ind. Crops By adapting resin analysis to a microtiter plate format, 96 sam- Prod. 19, 283–296. ples can be analyzed for resin concentration within a few minutes. Cornish, K., Chapman, M.H., Nakayama, F.S., Vinyard, S.H., Whitehand, L.C., 1999. The only drawback is the requirement for co-evaporation to elimi- Latex quantification in guayule shrub and homogenate. Ind. Crops Prod. 10, 121–136. nate acetone. By substituting acetonitrile for acetone it was possible Cornish, K., Chapman, M.H., Brichta, J.L., Vinyard, S.H., Nakayama, F.S., 2000. Post- to eliminate the co-evaporation step and measure resin directly in harvest stability of latex in different sizes of guayule branches. Ind. Crops Prod. the extract after a simple 1:10 dilution of the ASE vial contents. Simi- 12, 25–32. Cornish, K., Van Fleet, J.E., Chapman, M.H., Brichta, J.L., Knuckles, B.E., 2005. Latex larly, rubber analysis using ELS required no additional preparation, yield and quality during storage of guayule (Parthenium argentatum Gray) except when acetonitrile was used for the first solvent, in which homogenates. Ind. Crops Prod. 22, 75–85. case centrifugation or settling of the insoluble material in the cyclo- Cornish, K., Brichta, J.L., Yu, P., Wood, D.F., McGlothlin, M.W., Martin, J.A., 2001. hexane extracts is required. By introducing the cyclohexane extract Guayule latex provides a solution for the critical demands of the non-allergenic medical products market. Agro-Food-Industry Hi-Tech 12 (6), 27–31. directly into the detector without GPC separation, 24 samples can Estilai, A., Ray, D.T., 1991. Genetics, cytogenetics, and breeding of guayule. In: Whit- be analyzed per h. The method could be automated by addition worth, J.W., Whitehead, E.E. (Eds.), Guayule Natural Rubber. Office of Arid Land of an autosampler, however this would require an additional step; Studies, University of Arizona, Tucson, AZ, pp. 47–91. Hammond, B.L., Polhamus, L.G., 1965. Research on guayule (Parthenium argentatum): i.e., transferring sample from the ASE vials to autosampler vials. A 1942–1959. U. S. Dept. Agric. Tech. Bull. 1327, 157. robotic liquid handling system would allow the process to be fully Jasso de Rodriquez, D., Kuruvadi, S., 1991. Comparison of soxlet and homogenizer automated, thus facilitating screening of large numbers of individ- extraction methods to determine rubber and resin content of Mexican guayule plants. Bioresour. Technol. 35, 179–183. ual plants for rubber content. A flow chart of the methods described Kang, H., Kang, M.Y., Han, K.-H., 2000. Identification of natural rubber and charac- here for the extraction and analysis of resin and rubber in guayule terization of rubber biosynthetic activity in fig tree. Plant Physiol. 123, 1133– tissue, optimized for the maximum rate of throughput, is shown in 1142. Meeks, J.W., Feustel, I.C., 1951. The molecular weight of guayule rubber as influ- Fig. 9. enced by extraction and processing treatments. India Rubber World 125, 187– 188. 5. Conclusions Meeks, J.W., Banigan, T.F., Planck, R.W., 1950. A low-molecular-weight fraction of guayule rubber. India Rubber World 122, 301–304. Proksch, P., Behl, H.M., Rodriguez, E., 1981. Detection and quantification of guayulins A protocol was developed for resin and rubber extraction using A and B in Parthenium argentatum (guayule) and F1 hybrids by high-performance the ASE that involved (1) grinding the dried sample in a ball mill; (2) liquid chromatography. J. Chromatogr. 213, 345–348. Ray, D.T., Coffelt, T.A., Dierig, D.A., 2005. Breeding guayule for commercial produc- extracting the resin with acetone or acetonitrile; and (3) extracting tion. Ind. Crops Prod. 22, 15–25. the rubber with cyclohexane. Rapid and highly accurate methods Schloman, W.W., Hively, R.A., Krishen, A., Andrews, A.M., 1983. Guayule byproduct were developed for quantification of both the resin and the rub- evaluations: extract characterization. J. Agric. Food Chem. 31, 873–876. ber components that elute with each solvent. For resin analysis, Sidhu, O.P., Ratti, N., Behl, H.M., 1995. Quantitative and qualitative variations in resin content and guayulins (A and B) among different guayule cultivars. J. Agric. Food ultraviolet absorbance was used to determine the resin content of Chem. 43, 2012–2015. the tissue in acetone or acetonitrile extracts. For rubber analysis, Spence, D., Caldwell, M.L., 1933. Determination of rubber in rubber-bearing plants. evaporative light scattering was used to determine the rubber con- Ind. Eng. Chem. Anal. Ed. 5, 371–375. Sundar, D., Reddy, A.R., 2001. Interactive influence of temperature and growth light tent of the tissue after extraction with cyclohexane. Extraction of intensity on rubber accumulation and rubber transferase activity in guayule guayule tissue known to contain a high latex rubber content verified (Parthenium argentatum Gray). J. Plant Physiol. 158, 1291–1297. that the amounts of resin and rubber determined by these meth- Thurbide, K.B., Hughes, D.M., 2000. A rapid method for determining the extractives content of wood pulp. Ind. Eng. Chem. Res. 39, 3112–3115. ods were similar to the amounts determined gravimetrically. Since Traub, H.P., 1946. Rapid photometric methods for determining rubber and resin in these methods increase the speed of resin and rubber analysis, they guayule tissue and rubber in crude-rubber products. U. S. Dept. Agric. Tech. Bull. could be used in combination with ASE to increase the throughput 920, 37. Veatch, M.E., Ray, D.T., Mau, C.J.D., Cornish, K., 2005. Growth, rubber and resin eval- of guayule evaluation in germplasm enhancement programs, agro- uation two-year-old transgenic guayule. Ind. Crops Prod. 22, 65–74. nomic studies, and other experiments involving the quantification Whitworth, J.W., Whitehead, E.E., 1991. Guayule Natural Rubber. Office of Arid Land of resin and rubber in large numbers of samples. Studies, The University of Arizona, Tucson, AZ, p. 445.