back to Biotech Outreach site



Assessment of the potential for gene flow from agricultural crops to native plant relatives in the Hawaiian Islands.

Peter Münster and Ania M. Wieczorek

The Hawaiian archipelago is the most isolated island group in the world, separated by almost 1,600 km to other islands and nearly 4,000 km to the next major land mass. Due to this isolation, and 1-5 million years of evolution (McDonald et al., 1986), Hawai‘i has become home to more than 1,000 native flowering plant species. There are 88 families and 211 genera represented by these species, of which over 90% are endemic (Sohmer et al., 1987). Many of these plants are threatened by extinction. Rare species have a high priority in conservation programs because of their vulnerability to extinction by natural processes and human disturbances (Levin et al., 1996).

Island populations have a much higher risk of extinction than mainland populations (Frankham, 1998; Smith et al., 1993) due to human activities, over-exploitation, habitat destruction, and the introduction of competitive species (Reid et al., 1989; Levin et al., 1996; Rieseberg, 1991); as well as being susceptible to hybridization because of their small numbers and their frequent proximity to numerically larger congeners (Crawford et al., 1987; Carlquist, 1974). Insular species are more prone to interbreeding because they tend to be less genetically (allozymatically) divergent and to have weaker crossing barriers than their continental counterparts (Crawford et al., 1987). Another factor that increases hybridization rates in insular populations is that they often have poorly differentiated floral architecture, which facilitates interspecific cross-pollination (Carlquist, 1974) or restricts self pollination (Fryxell, 1979). Furthermore, a long flowering season in many tropical islands may also promote hybridization (Carlquist 1974). Rare, insular species are also particularly vulnerable to hybridization because of gametic wastage, reduced seed-set, and the production of ill-fit progeny on the one hand, and the swamping effect of gene flow on the other, are not likely to be counterbalanced by immigration from other conspecific populations (Levin et al., 1996).

Concern has been voiced that gene flow and hybridization between agricultural crops and native plant species may exacerbate their precarious position. Of particular concern is that gene flow occurs from agricultural products that have been modified through recombinant-DNA or biotechnological techniques to display traits such as insect or herbicide resistance, and the potential creation of “super plants” which could disperse freely into native ranges.

A key point that needs to be made is that gene flow occurs widely in nature. This gene flow is independent of whether transgenes are involved (Haygood et al., 2003). Pollen from agricultural crops often reaches wild plants growing nearby, and when the wild species are closely related to the crops, hybridization often ensues (Ellstrand et al., 1999). The transfer of genes to wild populations from related crops depends on a variety of factors (Keeler et al., 1990).

In this report, we considered the conventional crops as well as genetically engineered organisms (GEOs) currently grown in Hawai‘i. The questions being asked include which agricultural crops are related to the native species, and what is the likelihood that hybridization between them may occur?  Horizontal gene transfer (HGT) is also considered as a possible mode of introgression. Once we have determined the native species at greatest risk, a distribution analysis of agricultural products compared to the native species will be used to determine if pollen dispersal distances may affect hybridization and gene flow.

HGT is defined as the transfer of genetic material from one organism (the donor) to another organism (the recipient) that is not sexually compatible with the donor (Gay, 2001). As plants do not have any identified mechanisms to facilitate broad host-range gene transfer (except for pollen hybridization with related species), the possibilities and barriers of HGT from plants to bacteria have been approached within the framework of known mechanisms of HGT within bacteria (Nielsen et al., 1998). HGT between transgenic plants and soil microorganisms is strongly restricted by biological and physical barriers (Bertolla et al., 1999). Several groups have tried to transform naturally competent bacteria with transgenic plant DNA from different plants, but these attempts have been unsuccessful (Gebhard et al., 1998). Consequently, HGT is highly unlikely to play a role in gene transfer between cultivated crop species and wild relatives or the native plant genera on the Hawaiian Islands.

Vertical gene transfer (VGT) is a natural phenomenon that may occur when the pollen of one plant is transmitted to another plant or closely related plant. Hybridization in plants in this way depends on a variety of factors (Grant, 1975). To assess the potential ecological impact of field or commercial releases of crops in a given region, the likelihood and impact of vertical gene flow for that crop in that region should be taken into consideration (Connor et al., 2003). 12 of the 13 most important food crops of the world hybridize with wild relatives in some part of their agricultural distribution (Ellstrand et al., 1999).

Ellstrand et al. (1999) states that for hybridization to occur between plants, they must be cross-compatible so that the pollen is able to germinate and effect fertilization. To visualize which native plants are related to agricultural products that possibly fall into this category as well as to determine which native species are potentially at risk, a dendogram of plant orders was obtained (APG, 1998). Phylogenetic relationships are used to infer the relatedness between genera or species with relationship at the tribe, family, and order level. It is unlikely for species belonging to different tribes to exhibit cross-compatibility. Consequently, we used tribal boundaries as a limit to the potential for gene flow between species. With this information we were able to make a comparison between the agriculturally relevant crops and the native plant species and in this way further hone the list of native species potentially at risk.

As a result four tribes were seen to have an overlap between agricultural crops and native plant species:

Heliantheae: Crop: Sunflower (Helianthus annuus L.); Native genera: Argyroxiphium (DC),Bidens (L.), Dubautia (Gaud.), Lipochaeta (DC), Wilkesia (A. Gray).
Cultivated sunflower (H. annuus) is known to hybridize with wild sunflowers (Rogers et al., 1982; Ellstrand, 2003; Arias et al., 1994; Burke et al., 2002; Ellstrand et al., 1999), however we found no information confirming intergeneric hybridization. The Hawaiian native genus Argyroxiphium has been known to hybridize with the closely related genera Dubautia and Wilkesia (Carr 1985). Bidens spp. exhibit interfertility amongst all species in this genus whilst Lipochaeta is a polyphyletic genus, members of which resulted from the hybridization of species with different numbers of chromosomes (Wagner et al., 1999). A number of these genera are pollinated by the same, or similar, pollinators. Sunflower pollen is spiny, and thus well adapted to being transported by insects. Since the pollen is relatively heavy, wind pollination is negligible (Fick, 1978). A hybrid frequency of 0.02 was detected at 1,000 m in a hybridization study between cultivated and wild sunflowers (Arias and Rieseberg, 1994). These are factors which could increase the likelihood of gene flow and hybridization. Other factors decrease the likelihood of hybridization; pollen dispersal distance and viability (especially for Argyroxiphium and Wilkesia), and the relatedness between the genera. Argyroxiphium, Dubautia and Wilkesia fall into a subtribe of the Compositae, Madiinae (Baldwin et al., 2000), according to their current classification.

Gossypieae: Crop: Cotton (Gossypium barbadense L.; G. hirsutum L.); Native genera: Gossypium (Nutt. ex Seem.), Kokia (Lewton).
The probability of gene flow from cultivated cotton to G. tomentosum is high, and has in fact been shown to occur. This is due to both cultivated and wild cotton being allotetraploids (2n = 52) (DeJoode et al., 1992), as well as being located in similar habitats (Fryxell, 1979). Kokia spp. (2n = 24), would be expected not to hybridize with cultivated cotton due the vast genotypic differences, as well as a result of the small number of naturally occurring Kokia populations. Numerous factors affect gene flow from cultivated cotton, including: pollen size and weight, typical self pollination in cotton, flower morphology, flowering times, genotypic differences, and pollinators. To decrease the possibility of gene flow, USDA-APHIS requires an isolation distance, currently suggested to be ~12 m, between cultivated transgenic cotton and wild or non-transgenic species.

Solaneae: Crop: Potato (Solanum tuberosum L.); Native genera: Solanum (L.)
Isolation barriers between S. tuberosum and non-tuber-bearing Solanum species are strong (Ellstrand, 2003). Native Hawaiian Solanum varieties are all non-tuber-bearing (Wagner et al., 1999) Pollen dispersal distances from S. tuberosum are low (0.017% seed set at 10 m) (McPartlan et al., 1994). No naturally occurring intergeneric hybrids have been reported between S. tuberosum and other members of the Solanaceae family, although artificial hybridization events have been documented (Schoenmakers et al., 1993; Wolters et al., 1995; Gilissen et al., 1992).


Phaseoleae: Crop: Soybeans (Glycine max L.); Native genera: Canavalia (Adans.), Erythrina (L.), Mucuna (Adans.), Strongylodon (Vogel), Vigna (Savi).
Soybeans may be planted year round in most arable locations in Hawai‘i and this increases the likelihood of contact between the cultivated crop and native plant species. Distribution analysis has shown that populations of each of these native species fall into the major agricultural areas. Although, no evidence of intergeneric hybridization occurring under natural conditions with G. max as one parent was found, this does not necessarily mean that it does not occur. Factors suggest that cross-hybridization is unlikely to occur between G. max and the native plant species are the divergent cytotypes found within this tribe, as well as the low frequency of cross pollination that have been found between different cultivars of G. max under field conditions (Ray et al., 2003; Caviness, 1966; Boerma et al., 1975).  

It seems unlikely that hybridization will occur naturally between the cultivated crop plants discussed in this report and native Hawaiian plant species (with the exception of G. tomentosum). We were surprised at the lack of information with regards to this issue. Insular species are at greater risk of extinction than mainland populations (Levin et al., 1996; Wolf et al., 2001), and hence hybridization is a potential threat. Various factors contribute to a higher extinction rate of insular species, including inbreeding depression, a loss of genetic variation through genetic assimilation, or genetic adaptation to island conditions.  Insular species are coming into contact with non-native species at higher frequencies due to human processes that increases the potential for gene flow between these groups (Levin et al., 1996). The USDA-APHIS, suggests isolation distances between transgenic and non-transgenic plant species, to minimize the potential for hybridization to occur. Gene flow occurs in both non-transgenic and transgenic species (Haygood et al., 2003). We recommend that the USDA-APHIS isolation distances be employed to separate any agricultural crop from native Hawaiian plant relatives. Further, small scale pollination studies should be performed for the crops and native genera that were considered in this report. In addition to the agricultural products discussed in this study, many native plant species have relatives that have become naturalized in the Hawaiian Islands. Often these established relatives share habitats that bring them into close proximity of each other. Also, the phenotypic and genotypic similarities between them may increase the potential for hybridization. It is recommended that hybridization and introgression between these groups be further investigated to evaluate the risk of extinction by this means.

Many of the native plant species in the Hawaiian Islands are threatened by extinction. We were surprised to see a lack of specific location information for most of these species. Especially for those species with low population numbers, this type of information could prove crucial to the management of these populations. It is therefore vital that all the endangered plant species populations be geographically pinpointed to help save them from extinction.