Index | Search | Home

link to pdf file

Keys, R.N., D.T. Ray, and D.A. Dierig. 1999. In vitro characterization of apomictic reproduction in guayule. p. 275–279. In: J. Janick (ed.), Perspectives on new crops and new uses. ASHS Press, Alexandria, VA.

In Vitro Characterization of Apomictic Reproduction in Guayule

Roy N. Keys, Dennis T. Ray, and David A. Dierig

    1. Plant Materials
    2. Culture Procedures
    3. In Vitro Tests
    4. Analytical Technique
    5. Making Crosses
    6. DNA Extraction and Evaluation with RAPD Markers
    7. Statistical Analyzes
    1. In Vitro Experiment 1
    2. In Vitro Experiment 2
    3. In Vitro Experiment 3
    4. RAPD Analyzes

Guayule (Parthenium argentatum Gray, Asteraceae) is a latex-producing perennial desert shrub native to southwestern Texas and north central Mexico. It is potentially an economically viable new crop for the desert Southwest, with the advantages of having low water requirements and producing non-allergenic latex (Thompson and Ray 1988). Diploid guayule plants reproduce predominantly sexually, and possess a sporophytic self-incompatibility system (Powers and Rollins 1945; Gerstel and Riner 1950; Gerstel 1950). Polyploid plants are self-compatible, and can also reproduce through facultative or obligate pseudogamous apomixis, predominantly through mitotic diplospory (Powers and Rollins 1945; Esau 1946). The frequency of apomictic reproduction can vary from plant to plant, and even from flower to flower on a single plant (Esau 1946). The factors that control the mode of reproduction are not known.

Breeding programs for P. argentatum would be facilitated if a relatively rapid and easy technique were available to characterize plants as to their mode of reproduction. Such a technique has been developed and tested on several members of the Poaceae that reproduce both apomictically and sexually (Matzk 1991; Mazzucato et al. 1996). Their test consists of the application of an auxin to the developing flowers. The hypothesis is that auxin will stimulate development of apomictic embryos in plants that are genetically disposed toward apomixis, but not in those that are truly sexual. In the Poaceae, apomictic embryos developed to apparent maturity in the absence of endosperm after auxin treatment. Such flowers were visually distinguished from normal filled and empty flowers (Matzk 1991; Mazzucato et al. 1996).

We tried the above auxin test in field trials in order to characterize the reproductive systems of P. argentatum breeding lines (data not presented in this paper). Initial field trials using 2,4-dichlorophenoxyacetic acid (2,4-D) showed inhibition of embryo production, so other auxins and concentrations were tested. Results of these trials were ambiguous, mainly because of problems with pollen release in isolation bags that could have resulted in self-pollination and the uncertainty that the auxins were actually penetrating the ovule. Therefore, a technique using in vitro floral culture was developed which provided better control of environmental factors and a better means of applying growth regulators. Auxin and other growth regulators inhibited embryo production in vitro, so that the best technique appeared to be in vitro floral culture without growth regulators, and this paper presents the results of these experiments. Seven breeding lines and a sexual diploid control were characterized for reproductive mode using this technique, and the results were substantiated with RAPD analyzes of progeny arrays of selected crosses. Thus, in vitro floral culture provided good characterization of reproductive mode in breeding lines of P. argentatum.


Plant Materials

Plant materials were obtained from one-year-old plants being grown at The University of Arizona, Marana Agricultural Center, 50 km north of Tucson, Arizona. Seven lines were tested (G7-11, G7-14, G7-15, N7-11, N9-5, P2-BK, and P10-13). Four of these lines have been released and registered as improved germplasm (Ray et al. 1999). These plants are all putative tetraploids. Although they have not been characterized as to ploidy level, Cho and Ray (unpublished data) found that other lines derived from the same parental stocks as the plants used in this study were predominantly tetraploid (2n = 72), with a small proportion being either triploid or polyhaploid. Six known diploid plants were used as the sexual control in these studies. The plants were grown in potting soil in 3- to 5-gallon pots and under greenhouse conditions. Irrigation was provided by an overhead mist system. Grolux supplemental lighting was used to provide a 14-h photoperiod during winter months to stimulate flowering. Osmocote fertilizer was applied as needed.

Culture Procedures

Inflorescences were collected on the day on which cultures were to be made. Individual flower heads were selected based on the stage of development. The best stage was considered to be when the ray flowers had begun to open and the stigma was at least partially visible, but before the stigma began to spread open into the two surfaces it presents at maturity. As an added safeguard against stray pollination, each flower head was examined under a dissecting scope for the presence of pollen on the stigmas. The entire flower head was then sterilized by placing it in 0.075% sodium hypochlorite (15% commercial bleach) and 0.1% Tween-20 for 3 min, with agitation. The heads were rinsed in sterile, deionized, distilled water and placed in the culture medium. Each flower head was cultured in 2.5 ml of liquid medium in 2.5 cm × 6.5 cm glass vials with plastic lids. The cultures were placed on a shaker at 80 rpm, and grown at room temperature and under room lighting.

In Vitro Tests

In Vitro Experiment 1. This was a preliminary test on the effect of growth regulators. Nitsch and Nitsch (1969) (NN) medium was used, either alone or supplemented with NAA, NAA with gibberellic acid (GA3), and NAA with kinetin (all growth regulators at a concentration of 1 × 10-6 M). The NN medium consisted of (mg/L): KNO3, 950; NH4NO3, 720; MgSO4·7H2O, 185; CaCl2, 166; KH2PO4, 68; FeSO4·7H2O, 27.8; Na2·EDTA, 37.3; MnSO4·4H2O, 25; H3BO3, 10; ZnSO4·7H2O, 10; CuSO4·5H2O, 0.025; Na2MoO4·2H2O, 0.25; myo-inositol, 100; nicotinic acid, 5; thiamine·HCl, 0.5; pyridoxine·HCl, 0.5. Sucrose (60 g/L) was added, and pH adjusted to 5.8 prior to sterilization. This test was non-replicated, using one flower head per treatment from each of five lines.

In Vitro Experiment 2. Two media were tested: NN and Woody Plant Medium (Lloyd and McCown 1980) (WP). WP medium consisted of (mg/L): NH4NO3, 400; Ca(NO3)2·4H2O, 556; K2SO4, 990; CaCl2·H2O, 96; KH2PO4, 170, H3BO3, 6.2; Na2MoO4·2H2O, 0.25; MgSO4·7H2O, 370; MnSO4·H2O, 22.3; ZnSO4·7H2O, 8.6; CuSO4·5H2O, 0.25; FeSO4·7H2O, 27.8; Na2·EDTA, 37.3; myo-inositol, 100; thiamine·HCl, 1.0; nicotinic acid, 0.5; and pyridoxine·HCl, 0.5. Both media contained 60 g/L sucrose and pH was adjusted to 5.8 prior to sterilization. This experiment used flower heads from two to seven plants in six of the polyploid lines and five diploid plants, for a total of 31 flower heads per treatment.

In Vitro Experiment 3. The reproductive mode of the seven breeding lines were characterized using NN medium without growth regulators. From 4 to 12 plants were sampled within lines, and the experiment was repeated 2 to 5 times per plant, depending on the availability of flowers.

All experiments used a completely randomized design. The flower heads were collected after 14 days in culture.

Analytical Technique

The ray flowers were removed from the flower heads, with note being made of the presence of pollen on the stigmas. For in vitro-grown flowers, floral development was recorded as: 0 (no obvious development); 1 (some development of either the ray or disk flowers); 2 (apparently complete development of both ray and disk flowers, even if pollen had not been released). The single ovule was excised from each flower and placed in water on a microscope slide. After covering with a cover slip, the ovule was carefully squashed and the contents examined using phase contrast microscopy. Proembryos and embryos could be detected visually, and were recorded when present.

Making Crosses

Crosses were made on plants that were considered to be predominantly apomictic, partially apomictic, and predominantly sexual, based on results of the in vitro trials. After 14 days, the crosses were harvested and the mature embryos excised, with care being taken not to include any parental tissue or endosperm.

DNA Extraction and Evaluation with RAPD Markers

DNA was extracted from newly developing leaves of all of the plants used in the study. The miniprep technique of Stewart and Via (1993) was used in the extraction. The extraction buffer consisted of 2% w/v CTAB, 2% w/v PVP-40, 1.42 m NaCl, 100 mm Tris·HCl, 20.0 mm EDTA, 5.0 mm ascorbic acid, 4.0 mm DIECA. A single leaf was placed in 450 µl of extraction buffer and 2.5 µl of 2-mercaptoethanol in a 1.5 ml microcentrifuge tube. The tissue was ground using a Kontes pestle with a handheld drive unit. After a 15 min incubation at 70°C, 375 µl of chloroform:isoamyl alcohol (24:1) were added and the mixture shaken for 5 min. The fractions were separated by centrifugation at 1000× g for 5 min. Isopropanol (0.7 vol) was added to the water fraction and the mixture was placed at 4°C overnight to allow precipitation of the DNA. The precipitate was pelleted at 14000× g for 20 min, the supernatant was discarded, and the pellets were air-dried overnight. After being dissolved in 60 µl of sterile water, the DNA was stained with Hoescht 33258 dye and quantified using a Hoeffer TKO-100 DNA fluorometer. All samples were diluted to 25 µg/ml. DNA of embryos derived from crosses was extracted in the same way, except that the initial amount of extraction buffer was 250 µl with 1.0 µl 2-mercaptoethanol, and the volume of all other components were adjusted accordingly.

RAPD markers were generated using Amersham Life Science components, and Operon 10-mer primer set OPB (Operon Technologies, Inc.). Each 50 µl reaction mix consisted of 1X PCR reaction buffer, with a total of 25 mm MgCl2, 100 mm each of dNTPs, 0.2 mm primer, 2 U Taq DNA polymerase and 50 ng template in a reaction. The reactions were heated to 95°C for 2 min, then subjected to 45 cycles of 1 min at 93°C, 1 min at 36°C, and 2 min at 72°C, followed by a 5 min extention cycle at 72°C using a Techne PHC-3 thermal cycler. A single master mix was used for each progeny array and its two parents. 20 µl of the reaction was mixed with 6 µl of loading buffer and run on a 1.2% agarose gel in 1X TBE at 3.5 V/cm for 4 hr, and stained with ethidium bromide for UV visualization. The particular primer used for each array was determined from prior screening of the parental templates for RAPDs. A primer was informative if it produced a band in the male parent that was absent in the female. Progeny were scored as apomictic or outcrossed, depending on the presence or absence of the male band, respectively.

Statistical Analyzes

The embryo counts were transformed using the arcsine of the square root to permit analysis of variance. Data were analyzed using the General Linear Models Procedure (SAS Institute, Inc. 1988). Mean separations among lines were accomplished using the Duncan Multiple Range Test, and orthogonal contrasts were performed for each breeding line against the bulked data for the diploid controls. Correlation coefficients were calculated between the percentage of outcrossed progeny based on RAPD analyzes and the percent of embryo production in vitro.


In Vitro Experiment 1

Flowers developed normally in vitro, even exhibiting the dark brown pigmentation of epidermal layers that occurs in vitro as the flowers mature. Although pollen developed in the anthers, the high humidity conditions in culture prevented pollen release. There were no significant differences among treatments, probably because of the small sample size (Table 1). However, addition of NAA to the medium reduced embryo production by half. Combination of either GA3 or kinetin with NAA resulted in a further four-fold reduction in embryo formation.

In Vitro Experiment 2

There were no significant differences between medium means for embryo production. On WP medium, 23% (standard error = 0.06) of the flowers produced embryos, while on NN medium 22% (standard error = 0.05) produced embryos.

In Vitro Experiment 3

There were significant differences among lines for embryo production (Table 2). Orthogonal contrasts of polyploid lines against the bulked diploid controls, in combination with separations based on a Duncan Multiple Range Test and the actual mean values enabled characterization of lines as predominantly apomictic, facultatively apomictic, or predominantly sexual.

Although not statistically significant, there was important within-line variation in expression of apomixis. For example, in line N9-5, which was characterized as highly apomictic, one plant never produced embryos in vitro. In contrast, in line P10-13, characterized as being predominantly sexual, there were a few plants that produced embryos in vitro.

Of the flower heads placed in culture 37% did not develop. To determine if this phenomenon was related to reproductive mode, an c2 test for heterogeneity among the lines was conducted and was nonsignificant (c2 = 11.141, p = 0.084).

RAPD Analyzes

Primer OPB-15 produced one polymorphic band that was useful for determining outcrossing rates in progeny arrays. Outcrossing rates varied from 0% to 100%. The correlation of outcrossing rate with embryo production in vitro was –0.73 (p = 0.16).


The inhibitory effects of auxin application on embryo production in P. argentatum flowers grown in vitro were evident with the inclusion of NAA in the culture medium. Inclusion of NAA alone reduced this level by half, and in combination with other growth regulators resulted in a further four-fold decline in embryo production. Apparently, endogenous levels of growth regulators in the floral tissues are sufficient for embryo production. External application of growth regulators may cause either an inhibitory threshold to be exceeded, or a disruption of the balance of endogenous growth regulators so that the apomictic pathway is disturbed.

The lack of differences in embryo production when flowers were grown on the two media that were tested suggests that the particular nutrient combinations and concentrations are not crucial factors. NN and WP media differ considerably in their formulations. This lack of a medium effect was also reported for ovule and embryo culture of Helianthus hybrids in tests that also used NN medium as one of the media (Espinasse et al. 1991).

In the in vitro trials with NAA and other growth regulators, the embryos never developed to the extent that microscopic examination would not have been necessary. The embryos were allowed ample time for development in vitro, considering that embryos were mature enough after 14 days to fill the seed cavity in controlled crosses. Matzk (1991) and Mazzucato et al. (1996) reported that embryos in several members of the Poaceae developed enough after auxin treatment for the visual distinction of apomictic seeds, even without the presence of endosperm. The Poaceae possess an aposporic reproductive system, which may respond differently to auxin application than the predominantly mitotic diplosporic system in P. argentatum. Espinasse et al. (1991) incorporated NAA into the nutrient media in embryo rescue of hybrids in the related genus Helianthus to stimulate embryo elongation. Helianthus annuus also exhibited parthenogenesis in vitro (Yan et al. 1989). In this case, the highest rate of embryo formation occurred also without growth regulators in the medium and development advanced only to the early heart-shaped stage. It may be that parthenogenesis and subsequent embryo development have differing growth regulator requirements in the Asteraceae, as the result of being controlled by different genes or by different signals to the same genes.

The results with different media, and inclusion or exclusion of growth regulators, led us to the decision that in vitro floral culture in NN medium without growth regulators would be an adequate technique for the estimation of apomictic potential in P. argentatum. Using this technique, the reproductive systems of the seven lines used in this study could be characterized as being highly apomictic, moderately apomictic, or predominantly sexual. We expected the diploid controls to have a sexual reproductive mode (Powers and Rollins 1945; Gerstel and Riner 1950; Gerstel 1950). This proved to be the case, because they never produced embryos in vitro. RAPD analyzes of progeny arrays of selected matings of polyploid parents further substantiated the accuracy of this technique.

To conclude, the environmental factors that control the mode of reproduction in P. argentatum are unknown. The in vitro technique described here provides a means for studying the effects of environmental factors on floral development and expression of apomixis under controlled conditions. The presence of apomictic plants in predominantly sexual lines, and sexual plants in predominantly apomictic lines, may provide a pool of plant material adequate for the study of the molecular control of reproductive mode in P. argentatum.