Index | Search | Home | Table of Contents

link to pdf file

Lata, H., R.M. Moraes, A. Douglas, and B.E. Scheffler. 2002. Assessment of genetic diversity in Podophyllum peltatum by molecular markers. p. 537–544. In: J. Janick and A. Whipkey (eds.), Trends in new crops and new uses. ASHS Press, Alexandria, VA.

Assessment of Genetic Diversity in Podophyllum peltatum by Molecular Markers*

Hemant Lata, Rita M. Moraes, Andrew Douglas, and Brian E. Scheffler

*This work was funded by USDA Seed grant 97-35501-4886, Strengthening Award grant 99-01739, and the Specific Cooperative Agreement No. 58-6408-7-012.

The quality and quantity of medicinal plants are serious issues for the pharmaceutical and dietary supplement industries. Traditionally, this material has been harvested from the wild, and while cultivation procedures have been developed for supplements such as St. John’s Wort, this is still the general method employed for the majority of medicinal plants. Increasing public demand for these products is creating a serious environmental problem as demand is out pacing the supply and endangering the survival of many of these species in the wild.

There are two basic uses for medicinal plants; direct use as dietary supplements or as “chemical factories” for the production of plant-derived drugs such as podophyllotoxin, a lignan produced by the phenylpropanoid pathway. Thus, the parameters used to evaluate the source differ slightly depending on the application.

As chemical factories, yield is a combination of biomass and drug content, which defines the quality of the plant material. In cultivation, yield is measured by the amount of compound produced per hectare and not on a single plant basis. Growing genetically defined plant material improves biomass quality and helps to protect the world’s germplasm from extinction. Unfortunately, there are few breeding programs for medicinal plants. We are presently trying to achieve these goals for several medicinal plants using a variety of quantitative methods.

One of the medicinal plant species we are adapting to cultivation is Podophyllum peltatum L., Berberidaceae, also known as North American mayapple. The genus Podophyllum is a rich source of podophyllotoxin, an aryltetralin lignan that has important biological activities (Leander and Rosen 1988) and is the precursor of semisynthetic chemotherapeutic drugs such as etoposide, teniposide, and etopophos (Stähelin and Wartburg 1991; Imbert 1998). Currently, the patent for etoposide has expired and new clinical trials on various cancer therapies are under way (Ekstrom et al. 1998; Holm et al. 1998; Ajani et al. 1999). These factors have caused the sales of podophyllotoxin in the US to double from 1994 to 1999. In addition, new improved etoposide derivatives such as NK 611, GL 331, and TOP 53 are in pre-clinical development and are showing good results (Huang et al. 1996; Pagani et al. 1996; Utsugi et al. 1996; Raßmann et al. 1999).

Podophyllum emodi is presently the commercial source of podophyllotoxin for the pharmaceutical industry. It has been so intensively harvested from the wild that it is now endangered in its natural habitat, an area of the Himalayas (Foster 1993; Bhadula et al. 1996). In P. emodi, only rhizomes and roots are rich in podophyllotoxin (Jackson and Dewick 1984), and it takes seven or more years to establish a well-developed underground root/rhizome system. The majority of the roots and rhizomes are harvested resulting in the total destruction of the plants. In contrast, P. peltatum stores podophyllotoxin 4-O-b-D-glucopyranoside in leaf blades, and recent developments have shown that this compound can be easily converted to podophyllotoxin (Canel et al. 2000, 2001). The amount of podophyllotoxin 4-O-b-D-glucopyranoside varies from plant to plant, but some contain as much as 3% on a dry weight basis (Canel et al. 2001). These findings suggest that P. peltatum is a better candidate for development as a sustainable crop than P. emodi because as a perennial, it generates leaves annually from the rhizome and the leaves can be harvested late in the season without damaging the plant. Therefore, it would be desirable to identify superior genetic material that could be adapted for mayapple cultivation.

Podophyllum peltatum is found throughout the wooded landscape of the eastern half of North America (Meijer 1974). It has an extensive rhizogenous system that allows it to spread and survive as established colonies (de Kroon et al. 1991; Landa et al. 1992; Gerber et al. 1997). Studies have compared phenotypic characteristics between and within colonies, and while there can be variation between colonies, there appears to be little variation within smaller colonies (Parker 1989). Fertility studies between and within colonies indicate that seed set is lower when crosses are made within a colony (Laverty and Plowright 1988). It has been proposed that this decrease in seed set is due to self-incompatibility (Whisler and Snow 1992). The lack of phenotypic variation, possible self-incompatibility, seed mortality, and the extensive rhizome structure are all indications that colonies are the result of clonal propagation (Rust and Roth 1981; Parker 1989), but there are no genetic studies that conclusively support this hypothesis. In addition, the colonies appear in a semi-circular state indicating that development radiates outwards from a central point. Numerous colonies in close proximity to one another are common in wooded areas of the South. Some patches are so extensive that they are probably the result of either integration of several colonies or represent a single clonal population of ancient origins. Genetic analyses are needed to determine the composition of these large colonies and the relationship between patches within the same site. To our knowledge, no previous genetic analysis has determined how neighboring colonies are derived, from various seedlings or asexually from a single source.

Moraes et al. (2000) have shown that the podophyllotoxin content varies greatly between colonies (1.1 to 56.0 mg·g-1). Therefore, it would be inefficient, environmentally destructive, and economically unsound to randomly harvest mayapple. For the rapid and economical development of P. peltatum as an alternative crop, a more logical approach would be to evaluate natural populations of P. peltatum for podophyllotoxin content and agronomic traits. Desirable colonies of P. peltatum can be easily propagated through the use of rhizome cuttings. If a colony were truly clonal in nature, a larger colony would provide more material for agronomic evaluations and for quicker establishment of rhizome stock for producers. Therefore, it is important to know if a population of P. peltatum is composed of clones and if nearby populations are clones with the same origin. For evaluation purposes, determination of relationships between and within populations cannot be made by just chemotyping the populations as the production of podophyllotoxin can be influenced by environmental factors. Wooded areas can be highly variable in their growing conditions due to neighboring plants, soil type and shading due to overstory canopy. Therefore, it is possible that clones grown in the same-forested area might differ in their podophyllotoxin content.

RAPD (random amplified polymorphic DNA) and AFLP (amplified fragment length polymorphism) are genetic fingerprinting techniques suitable for the genetic evaluation of P. peltatum. These PCR (polymerase chain reaction) based technologies require small amounts of DNA. One method that is relatively accurate and reproducible from lab to lab is AFLP. This technique requires that the DNA is first digested by restriction enzymes and then DNA adapters are ligated to the ends of the DNA. Oligonuclotides corresponding to these adapters are then used for the PCR. RAPD is the other technique (Williams et al. 1990), which has been successfully used to genetically profile many different plant species, such as almond (Bartolozzi et al. 1998), Echinacea purpurea (Baum et al. 2001), hybrid strawberries (Degani et al. 1998), Digitalis obscura (Gavidia et al. 1996) and Artemisia annua (Sangwan et al. 1999). This technique uses small oligonucleotides (~10 bps) that differ in their DNA sequence. The binding of these oligonucleotides is fairly nucleotide specific but the sites are relatively random within the genome. DNA segments are amplified where two primers bind in close proximity to each other in the opposite orientation. RAPD is a lower cost technique when compared to AFLP but the results are highly dependent on the quality of the DNA and the PCR conditions, therefore reproducibility between labs is often difficult to achieve.

The goal of this preliminary study is to attempt to understand population structure of P. peltatum to better identify and ascertain individuals that could be used to establish cultivars for commercial level propagation of useful secondary compounds.


Plant Materials

Young leaves of 7 P. peltatum accessions were collected for RAPD studies from three different sites in Lafayette County, Mississippi (Fig. 1). Each patch was considered one accession and coded according to the site of the collection. Three samples each from two patches were collected for two distinct populations physically separated by the Old Taylor Road on The University of Mississippi Campus. Sampling within the patch was 2.5 m apart and coded as UM-MS West I-II, UM-MS East I-II. At Hwy 7, Lafayette County, one patch was collected from each side of the roadway and coded as LaMS-West, LA-MS-East. Two patches located along Jackson Avenue in Oxford were sampled and coded as OX-MS I-II. These young leaf samples were frozen in liquid nitrogen and stored at –80°C prior to DNA isolation.

Fig 1. The three sites of Podophyllum peltatum in Lafayette county, Mississippi designated for this study.

In order to determine genetic variation between and within colonies of P. peltatum, three primary locations in the area of Lafayette County, MS were selected. These locations were within 2–3 km of each other. The UM-MS and LA-MS locations consist of numerous colonies and both are transversed by a road. Two colonies on each side of the road were selected, and the separation distance between colonies was 50–100 m. Both the roads caused a separation distance of over 200 m. The colonies range in size from 10–20 m2. Instead of small independent colonies like those found at the UM-MS and LA-MS locations, the OX-MS site covers an area of 100–200 m2. It is not clear from its physical characteristics if this site represents a super colony or if it is the result of independent colonies coalescing. Approximately, 50 m separated the sample sites for this location. For all sampling sites, three collections within 2.5 m of each other were made.

DNA Extraction

DNA was extracted from frozen leaves by the CTAB method (Bousquet et al. 1990). Samples of 500 mg were ground to powder in liquid nitrogen, using a mortar and pestle. The powder was transferred to a 25 ml sterile Falcon tube with 10 ml of CTAB buffer. The extraction buffer consisted of 2% (w/v) CTAB (cetyltri-methyl ammonium bromide, Sigma), 1.4 m NaCl, 20 mM EDTA, 100 mM Tris-HCl pH 9.5, and 0.2% (v/v) b-mercaptoethanol. After incubating the homogenate for 1 hour at 65°C an equal volume of chloroform was added and centrifuged at 10,000 rpm for 20 min. DNA was precipitated with 1/10 volume (ml) of 3 M sodium acetate and an equal volume of isopropanol followed by centrifugation at 10,000 rpm for 10 min. The DNA pellet was washed with 70% ethanol, air-dried, and resuspended in TE-buffers (10 mM Tris pH 8.0 and 0.1 mM EDTA). DNA quantity was estimated spectrophotometrically by measuring absorbance at 260 nm.

PCR Amplification

Decamers from kits A, B, G, O, S, and T (Operon Technologies Inc., Alameda, California, USA) were used individually as primers. Each kit contains a total of 20 primers. DNA was amplified following the protocol of Williams et al. (1990). Amplification reactions were performed in volumes of 20 µl containing 1X reaction buffer (Promega Corporation, Madison, Wisconsin), 2.5 mM MgCl, 5 mM each of dNTP (Promega Corporation), 10 pmoles of primer, 50 ng of genomic DNA, and 0.2 unit of Taq polymerase (Promega Corporation). DNA amplification was performed in an Applied Biosystems Gene Amp PCR system 9700 thermal cycle programmed as follows: 1 cycle of 4 min at 94°C, 39 cycles of 1 sec at 94°C, 1 min at 37°C, and 1 min at 72°C. The last cycle was followed by a final incubation for 5 min at 72°C and the PCR products were stored at 4°C before analysis. To reduce PCR artifacts due to variation of various components, for each primer a PCR master mix of all the components, except the genomic DNA, was made and then aliquoted (Plant et al. 1993; Ellsworth et al. 1993). The amplified products were separated electrophoretically in a 1.2% agarose gel using 1X TBE buffer, and stained with Vistragreen (Amershram Life Science). The stained gels were scanned under Fluor Imager 595 (Molecular Dynamics, Sunnyvale, California). The molecular sizes of the amplification products were estimated using a 1 kb DNA ladder (Promega Corporation).

Amplified DNA Marker Scoring

Each genomic DNA sample was scored for the presence or absence of specific PCR bands (DNA marker) generated by a primer. PCR bands of low visual intensity were considered ambiguous and were not scored.

Data Analysis

Both distance and non-rooted parsimony analyses were executed to evaluate relationships within and between the patches using PAUP* (Swofford 2001). To evaluate relationships, branch and bound techniques were used with collapsing of zero-length branches in effect. Total and mean distances were evaluated using the Nei-Li assumptions and unambiguous changes supporting particular groups or patches were calculated by hand. Variances within and between patches were evaluated to determine degree of genetic divergence.


A total of 120 RAPD primers were screened on a subset of samples using only 1 PCR condition to ascertain which primers would provide useful information to determine genetic variation between the DNA samples. In most instances the primers did not produce any valuable data. Often no PCR bands were produced in any of the samples, or only 1–2 monomorphic bands were detected. Some primers resulted in a banding pattern that was not distinct. In all of these cases, the results were not used in the calculation of genetic distance. Table 1 represents all of the polymorphic and monomorphic bands detected for the listed primers. Fig. 2 shows the RAPD profiles generated by primer S-20. This data was used to calculate the genetic distance. A lack of resolution of banding among individuals within the LA-MS-W site resulted in over 20% unknowns and were thus eliminated from the final analysis. It should be noted that the large degree of uncertainty in the LA-MS-W group did not affect the overall of relationships.

Table 1. RAPD results are presented. The primers used are indicated and the number of monomorphic bands is given after the primer number in brackets. A solid black box indicates a specific PCR band was present and an empty box indicates absences of the PCR product. Absence of any box indicates that the PCR reaction did not work for that particular sample. Each column represents one genomic DNA sample.

Fig 2. RAPD profiles generated by primer S-20. Lane M: 1 kb DNA ladder. Lanes 1-20 show populations tested (Lanes 1-3: UM-MS West, 4-7: UM-MS East, 8-10: LA-MS West, 11-13: LA-MS East, 14-20: OX-MS).

Based on total distance measures, there was no significant difference between patches because patches were collected in close distances and also no significant difference found within the UM-MS patch. (mean divergence: UM-MS-E/UM-MS-W = 0.177; UM-MS/OX-MS = 0.19; UM-MS/LA-MS-E = 0.168; OX-MS/LA-MS-E = 0.171). It should be noted that putatively different populations (geographically distant) need to be examined before an accurate estimation of the degree of variance can be interpreted.

Phylogenetic analysis of 47 parsimony informative characters resulted in 34 trees (Fig. 3; 84 steps; r.c.i. 0.49, r.i. 0.83). It should be noted that the strict consensus tree, not illustrated here, showed collapsed branches within some of the patches but no switching of taxa between patches. Each of the three sites is phylogenetically isolated from one another. Within the UM-MS site, the East and West groups separated by a total of 5 unambiguous differences. The LA-MS and OX-MS patches shared the most in common and the LA-MS-E patch has no unambiguous character support (Fig. 3).

Fig 3. One of 34 unrooted phylograms obtained from parsimony analysis of data illustrating relationships within and among patches. Circled clades represent areas and subareas. Numbers on branches represent the total number of changes on the top and unambiguous changes in bold on the bottom.

The results clearly show that the three sites share some relationship to one another with the OX-MS being more closely related to the LaMS site with 9 changes (3 unambiguous) (Fig. 3). It is actually the genetic relationship within each site and among colonies that provides the most enlightening information.

Minor genetic variation was detected between the patches collected at the OX-MS location, indicating that this location is composed of a super colony. The differences detected are probably the result of somaclonal variation. In this instance, the size of the colony is not necessarily indicative of the age of the colony. This colony is located near a major road and undoubtedly the soil has been disturbed by urban sprawl. Therefore, it is possible that the rhizomes of a smaller colony were disturbed by earth moving machines and distributed over a large area thus causing the development of this super colony. Unfortunately, there are insufficient records on this colony to make any determination about how it developed into a super colony.

The UM-MS site was composed of four patches and two of the patches were separated from each other by a road. The genetic variance analysis shows that all four patches are closely related to one another, but that there is some variation within and between patches. The low degree of variation suggests that all of these patches are clonally derived from the same maternal plant. It is difficult to believe that the rhizomes of P. peltatum could transverse the approximately 50 m roadway to establish new colonies on the other side. Therefore, the presence of the roadway and limited genetic variation between these patches might be an indication that the patches existed before in the installation of the road. It is possible that the spread of patches is due to the intervention of humans when the road was installed, but this is not likely as all the patches are at least 25 m from the roadway and are located in a established wooded area. Podophyllum peltatum is known to have a high degree of self-incompatibility and while it is possible that the slightly different genotypes might be the result of pollination events it is very unlikely. The strong self-incompatibility mechanism indicates the genetic composition of most P. peltatum is outcrossed origin and not due to inbreeding. If the original seed for the mother plant was derived from a hybrid cross then any progeny derived from selfing would show the significant variation of an F2 population. As in the OX-MS site, the most likely cause of genetic variation is due to somaclonal variation although additional sampling is necessary.

Looking at the results from other studies on P. peltatum (Parker 1989; Whisler and Snow 1992) and its extensive rhizome structure, it is not surprising to discover that patches in one area might be clones of one another. Nonetheless, this study was relatively small in the number of sites and patches sampled, and it would be logical to assume that there are other sites that are composed of colonies that are genetically diverse from one another. The sites tested cover an extensive area of land and this might be an indication they are of ancient origin. The UM-MS site covers an area of over 150 m × 200 m and it would be interesting to learn if all the patches within this location are related clones.

The results from this study provide some indication of the genetic constitution of P. peltatum colonies and what methods might be used to evaluated the genetic variation observed between colonies. More extensive studies will be needed to check if more sensitive methods such as AFLP can be used to measure genetic variance. More accessions, covering a broader area will also need to be evaluated to determine if the results of this study apply to colonies in general.