Plant Cell Reports

, Volume 26, Issue 7, pp 977–987

Transgenic American elm shows reduced Dutch elm disease symptoms and normal mycorrhizal colonization

Authors

  • Andrew E. Newhouse
    • Faculty of Environmental and Forest BiologySUNY College of Environmental Science and Forestry
  • Franziska Schrodt
    • Faculty of Environmental and Forest BiologySUNY College of Environmental Science and Forestry
  • Haiying Liang
    • Department of Genetics and BiochemistryClemson University
  • Charles A. Maynard
    • Faculty of Forest and Natural Resources ManagementSUNY College of Environmental Science and Forestry
    • Faculty of Environmental and Forest BiologySUNY College of Environmental Science and Forestry
Genetic Transformation and Hybridization

DOI: 10.1007/s00299-007-0313-z

Cite this article as:
Newhouse, A.E., Schrodt, F., Liang, H. et al. Plant Cell Rep (2007) 26: 977. doi:10.1007/s00299-007-0313-z

Abstract

The American elm (Ulmus americana L.) was once one of the most common urban trees in eastern North America until Dutch-elm disease (DED), caused by the fungus Ophiostoma novo-ulmi, eliminated most of the mature trees. To enhance DED resistance, Agrobacterium was used to transform American elm with a transgene encoding the synthetic antimicrobial peptide ESF39A, driven by a vascular promoter from American chestnut. Four unique, single-copy transgenic lines were produced and regenerated into whole plants. These lines showed less wilting and significantly less sapwood staining than non-transformed controls after O. novo-ulmi inoculation. Preliminary observations indicated that mycorrhizal colonization was not significantly different between transgenic and wild-type trees. Although the trees tested were too young to ensure stable resistance was achieved, these results indicate that transgenes encoding antimicrobial peptides reduce DED symptoms and therefore hold promise for enhancing pathogen resistance in American elm.

Keywords

Ulmus americanaOphiostoma novo-ulmiDutch elm disease resistanceTransgenicMycorrhizae

Abbreviations

DED

Dutch-elm disease

GUS

Beta-glucuronidase

SSC

Saline sodium citrate

SDS

Sodium dodecyl sulfate (also known as sodium lauryl sulfate)

PDB

Potato dextrose broth

PDA

Potato dextrose agar

MIC

Minimum inhibitory concentration

AMP

Anti-microbial peptide

Introduction

In North America, the American elm (Ulmus americana L.) was originally found throughout the eastern United States and southeastern Canada. It has since been cultivated throughout temperate areas of the continent and has been characterized as one of American’s heritage trees. It can grow to be more than 500-years-old and 140 feet tall, typically with a spreading vase shape and a broad crown that can be wider than it is tall (Line 1997). Its aesthetically pleasing shape and tolerance of pollution, compacted soil, and drought make it a very desirable tree for lining city streets. When a street is lined on both sides with rows of large American elms, the canopies can intertwine to make a picturesque cathedral-like passageway. In fact, the American elm was for many years considered the most popular landscape tree in the United States (Sticklen et al. 1994).

Dutch-elm disease (DED) was first introduced to Europe in the early 1900s and North America in 1930, and has since swept through both continents twice in separate pandemics. Researchers identified the fungus Ophiostoma ulmi (Buism.) Nannf., (originally Graphium ulmi Buismann) as the causative organism of DED (Holmes and Heybrook 1990). The second pandemic was caused by a more virulent form of the fungus, named O. novo-ulmi (Brasier 1991). This new variant on DED was more virulent, killing trees more quickly, and infecting elm clones and species that were tolerant of the less aggressive strains of O. ulmi (Hubbes 1999). DED is spread when elm bark beetles (either native North American Hylurgopinus rufipes or the European Scolytusmultistriatus) hatch and emerge from dead elm logs harboring O. ulmi or novo-ulmi and proceed to feed on healthy trees (Anderson and Holliday 2003). Left untreated, nearly all American elm trees that are exposed to Dutch-elm disease will die within a few weeks to a few years.

Treatment, control, and remediation efforts have included screening programs (Smucker 1944; Townsend and Douglass 2001), pesticides to control the beetle host (Heybroek et al. 1982), fungicidal treatments (Haugen and Stennes 1999), and biological controls (Hubbes 1999; Solla and Gil 2003). In general, these efforts have been impractical for widespread use due to cost, labor, or lack of effectiveness. Selection and breeding programs have produced American elm trees that are tolerant of DED to various degrees, but to date even the highly tolerant cultivars are still susceptible to elm yellows (Sinclair 2001), another serious disease of elms. Additionally, differences in ploidy further complicate breeding efforts between American elm (2n = 56) with other elm species (2n = 28) that are DED-tolerant. Finally, resistance produced by breeding or selection programs is likely quantitative, therefore clonal propagation would be necessary, leading to a lack of genetic diversity which may make large-scale plantings susceptible to environmental problems, pests, or new diseases.

In contrast, the process of producing a disease-resistant tree through genetic engineering can potentially be faster, more controllable, more predictable, and cheaper than traditional breeding (Giri et al. 2004). Through the use of varied explant material, more genetic diversity can be maintained, which is important for any plant population, and has been specifically recommended for new elm releases (Karnosky 1979). Potential transgenes and promoters can be synthesized de novo (i.e., Powell et al. 1995), isolated from related plants (i.e., wheat to maize, Sangtong et al. 2002) or even taken from completely unrelated organisms (i.e., moth to pear tree, Reynoird et al. 1999). Powell et al. (2005) gives a more thorough review of enhancing fungal and bacterial resistance in transgenic trees.

The first elm (and indeed one of the first tree species) cultured in vitro was U. campestris by R.J. Gautheret in 1940, though regeneration was not achieved at this time (Karnosky and Mickler 1986; Moore 2003). Regeneration of whole elm plants was accomplished about 35 years later with both U. americana (Durzan and Lopushanski 1975) and U. campestris (Chalupa 1975). American elm leaf pieces were first used as explants in 1991 (Bolyard et al. 1991a), and George and Tripepi (1994) studied American elm leaf explant tissue culture conditions in more detail to improve regeneration efficiency. Other species of Ulmus have been more commonly regenerated in tissue culture, including, U. carpinifollia (Mezzetti et al. 1988), U. procera (Fenning et al. 1993), U. pumila (Kapaun and Cheng 1997), and U. hollandica (Ben Jouira et al. 1998).

Transformation of American elm for resistance to Dutch-elm disease has been a goal almost as long as genetic engineering has been in existence (Sticklen et al. 1991), but transformation of U. americana with putative pathogen resistance enhancing genes has not been published prior to the present study (More detailed methods are reported in Newhouse et al. 2006). Foreign gene expression in elm tissue was first achieved on the ‘Pioneer’ hybrid elm, for which internode stem sections were shown to express the GUS gene after being transformed separately with Agrobacterium and biolistics (Bolyard et al. 1991b). Transgenic English elms have also been regenerated after shoots were transformed with wild-type Agrobacterium (Fenning et al. 1996) and Agrobacterium tumefaciens carrying GUS and nptII genes (Gartland et al. 2001; Gartland et al. 2000). More recently, English elm has been transformed with a putative pathogen resistance enhancing gene (Gartland et al. 2005).

The plasmid construct used to transform American elm in this project was designated pSE39 (Fig. 1). The three genes in this construct are a constitutively-expressed GUS marker gene, a synthetic antimicrobial peptide ESF39A, and a constitutively-expressed npt2 selectable marker. Newhouse (2005) gives a more complete description of the entire construct.
https://static-content.springer.com/image/art%3A10.1007%2Fs00299-007-0313-z/MediaObjects/299_2007_313_Fig1_HTML.gif
Fig. 1

pSE39 plasmid construct encoding the GUS marker gene, the ESF39A antimicrobial peptide (Powell et al. 2000), and the nptII selectable marker. The gene for ESF39A is preceded by a vascular-specific promoter (ACS2) cloned from American Chestnut (Connors et al. 2002), and followed by the Win6.39 terminator from the poplar chitinase gene (Clarke et al. 1994). Also note the restriction sites used for digestion before Southern hybridization

One aspect of plant ecology that is often neglected in transformation research is the role of mycorrhizal fungi and their relationship with the host plant. American elms typically form relationships with Paris-type arbuscular mycorrhizal (AM) fungi (Brundrett et al. 1990), though American elms have also been inoculated with ectomycorrhizae in greenhouse studies (Mushin and Zwiazek 2002).

Few studies have been published regarding the effect of transgenes on mycorrhizae. Two reports of transgenic tobacco, expressing pathogenesis-related proteins or chitinases, concluded that in general, these antifungal genes did not affect mycorrhizae (Vierheilig et al. 1995; Vierheilig et al. 1993). Turrini et al. (2004) showed that aubergine (Solanum melongena) constitutively expressing a natural AMP isolated from Dahlia was resistant to pathogenic Botrytis and Verticillium, while symbioses with mycorrhizal Glomus mosseae were not significantly affected. A single publication on mycorrhizae in field-planted transgenic aspen found that mycorrhizal fungi associated with most transgenic trees showed the same colonization levels and diversity as those on nontransgenic trees (Kaldorf et al. 2002). However, the same study reports that one of the common EM fungal species on one transgenic clone showed significantly different morphology and lower abundance compared to other transgenic trees transformed with the same construct. Therefore, positional effects of transgenes may influence mycorrhizal associations as well as transgene expression levels (Peach and Velten 1991). In a separate ongoing study, preliminary data has indicated that reduced growth in hybrid poplar transformed with constitutively expressed antifungal genes (Liang et al. 2001) may be correlated with reduced mycorrhizal colonization (Horton and Powell unpublished). Therefore, mycorrhizal observations should certainly be considered when performing future field trials of transgenic plants.

Materials and methods

Transformation

All American elm seed used in this study was from a single batch purchased from F.W. Schmacher Co. Inc. (Sandwich, MA, USA). American elm seedlings were grown on a light bench with a 16-h day, 8-h night cycle. Leaves were collected from seedlings, sterilized, sliced into approximately 1 cm squares, and co-cultivated with Agrobacterium tumefaciens strain EHA105, containing the binary vector pSE39. Transformed shoots were regenerated under 100 μg/ml kanamycin selection in tissue culture according to the method of Newhouse et al. (2006). Potential transformants were screened with a GUS assay (Newhouse 2005) to confirm the presence of T-DNA in the plant genome.

MIC assay

Sap was extracted from petioles of large, fresh green leaves using a portable Pressure Bomb (PMS Instrument Co., Corvallis, OR, USA). Approximately 20–50 μL of sap was extracted from each leaf. ESF22A, which is the processed central portion of the ESF39A antimicrobial peptide (Powell et al. 2000), was tested against O. novo-ulmi using a minimum inhibitory concentration (MIC) assay (Catranis 1999). Dilutions of pure synthetic peptide ESF22A (Sigma-Genosys) were suspended in Potato Dextrose Broth (PDB, Difco) and low-melting temperature agarose in a Corning 96-well plate, and approximately 100 spores of an O. novo-ulmi conidia solution were added to each well.

Molecular tests

DNA was extracted from young elm leaves (transgenic and wild-type) with a CTAB method based on that of Lodhi et al. (1994). PCR was performed with custom primers (Table 1) to confirm the presence of the GUS and ESF39A genes in all transformants. Newhouse et al. (2006) gives a more detailed description of DNA extraction and PCR techniques.
Table 1

PCR primer data

Name

Sense sequence

Ta (°C)

Product length (bp)

ESF39_F

GTGATGGTGATGGTGATG

55

157

ESF39_R

TCGTCTTCTTCCTCCTTG

SE39long_F

AACGGATCACTGTCAGCATACG

61

926

SE39long_R

GTCTTCTTCCTCCTTGCCTTGG

GUS_F

TTACAGAACCGACGACTC

55

327

GUS_R

ATGCTCCATCACTTCCTG

GUSlong_F

TGTGTCTATGATGATGATG

48

910

GUSlong_R

CCAAAGCCAGTAAAGTAG

**_F indicates Forward/Sense primer, **_R indicates Reverse/antisense, Ta indicates annealing temperature used with each primer pair

Southern hybridization was performed to determine copy number and further confirm transformation. Twenty micrograms of genomic DNA was digested with 10 units/μg EcoRI (New England Biolabs), and an additional 20 μg was digested with 10 units/μg HindIII (Promega). Each of these enzymes has only one restriction site inside the T-DNA, but outside of the GUS probe sequence (Fig. 1). Each digestion was allowed to proceed for 3–4 h and then stopped with a 10-min incubation at 65°C. The digested DNA was separated by electrophoresis in a 0.7% agarose/TBE gel. The pSE39 plasmid (100 picograms) was treated the same way and used as a positive control. The DNA was then transferred from the agarose gel to a positively charged nylon membrane (Tropilon-Plus, Tropix, Bedford, MA, USA) according to the ‘alkaline transfer’ method of Sambrook and Russell (2001), modified by using a large cellulose sponge as a support. A hybridization probe was produced from a 910 base pair segment of the GUS gene, amplified via PCR from the pSE39 plasmid, and purified with the High-Pure PCR Purification Kit (Roche, Mannheim, Germany). This segment was labeled with 32P Redivue dCTP using the Redi-Prime II Random Prime Labeling System (Amersham Biosciences). The probe was hybridized to the membrane for 2.5 h at 65°C in Rapid-Hyb buffer (Amersham). The blot was washed for 20 min at room temperature with 2× SSC, 0.1% (w/v) SDS, and four times for 15–20 min at 68°C in 0.1× SSC, 0.1%SDS. Blots were wrapped in cellophane and exposed to X-ray film (Kodak Biomax MS) between intensifying screens for 18–48 h at −80°C.

RNA for RT-PCR was extracted from young stem and leaf midvein tissue using the ‘Hot-Borate Method’ of Wilkins and Smart (1996). Reverse transcription and PCR were carried out with the Qiagen One-Step RT-PCR Kit according to the manufacturer’s instructions. The template consisted of diluted RNA (1:10 dilutions for approximately 400 ng total RNA per reaction) from each of four transgenic lines (FS5, AN1C, and two AN1X), and a nontransgenic wild-type (Wt) sample. ESF39 and GUS primers (Table 1) were used (separately) with all samples, as both of these produce products inside the transcribed region of their respective genes. Identical reactions were assembled using ESF39 primers without reverse transcriptase as DNA-contamination controls, and these were added to the cycler as soon as the 15-min 95°C activation step finished. Reactions with reverse transcriptase and RNA-free water instead of template RNA were included as RNA-contamination controls.

Inoculation

Inoculation tests were performed on potted wild-type and transgenic elms (1–3 growing seasons old, 0.6–1.2 m tall) to test for enhanced resistance to Dutch-elm disease. Ophiostoma novo-ulmi (ATCC #34359) was grown on Potato Dextrose Agar (PDA). Conidia were harvested by washing the surface of the colony with sterile water. The O. novo-ulmi conidia injection solution was adjusted to a concentration of approximately 1 × 108 cells/ml (counted using a Neubauer Hemacytometer). A #11 scalpel blade was then used to make a vertical slice into the stem of each tree at 1/3 to 1/2 its total height, and 10–20 μL of the conidia injection solution was pipetted slowly into the wound, allowing transpiration to pull the liquid into the vessels. Sapwood staining was observed 14 weeks after inoculation and quantified by measuring the stained distance above and below the inoculation site.

Mycorrhizae

Mycorrhizal colonization was quantified in roots collected from healthy plants (transgenic and wild-type) with new growth approximately 3 months after these plants were transplanted from the greenhouse to the field. Staining was performed with Chlorazol Black E according to Brundrett et al. (1996) and stored in 50% glycerol at room temperature.

Mycorrhizal colonization was quantified on 30 root samples from each plant using the ‘magnified intersections’ technique (McGonigle et al. 1990) with minor modifications. Stained root sections were arranged in parallel lines on standard microscope slides in 50% glycerol, and examined under 200× magnification on a Nikon Eclipse E600 DIC microscope. Each intersection between a horizontal root fragment and the vertical ocular micrometer was scored for presence or absence of mycorrhizal structures (hyphae, spores, vesicles, or arbuscules, Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00299-007-0313-z/MediaObjects/299_2007_313_Fig2_HTML.jpg
Fig. 2

Mycorrhizal fungal structures [vesicles (a and b) and hyphae (c)] in field-grown transgenic American elm roots. Similar structures were seen in roots from wild-type controls (not shown)

Results

Transformation and molecular tests

One leaf piece cultured with Agrobacterium produced a cluster of at least 30 shoots (designated AN1_). Two additional shoots from a single leaf piece in a previous transformation experiment were also evaluated (designated FS5). All of these transformants grew on selection media (100 μg/ml Kanamycin), showed strong GUS expression, and PCR amplified distinct bands in both the GUS and ESF39A genes (Newhouse 2005). Since all three genes in the construct are thus shown to be present in each transgenic line, it is unlikely that mere DNA fragments are responsible for transformation. Southern hybridization confirmed presence of the transgene construct, revealed that the many shoots in the current experiment originated from at least three unique transformation events (designated AN1C, AN1F, AN1X), and that the two shoots from the previous experiment originated from a single transformation event (FS5) (Fig. 3). All transgenic lines tested were shown to be single-copy. None of the wild-type trees produced bands on Southern blots.
https://static-content.springer.com/image/art%3A10.1007%2Fs00299-007-0313-z/MediaObjects/299_2007_313_Fig3_HTML.gif
Fig. 3

Southern hybridization of 32P-labelled GUS probes to transgenic elm DNA digested by HindIII

Transcription of the ESF39A antimicrobial peptide transgene in the target tissues (stem and midvein) was confirmed by RT-PCR performed on RNA isolated from these tissues. RNA templates from transgenic trees produced bands of the appropriate length (∼160 bp), while wild-type RNA and transgenic lines without reverse transcriptase did not produce this band (not shown, Newhouse 2005).

MIC assay

In five separate assays, the ESF22B antimicrobial peptide inhibited growth of O. novo-ulmi at concentrations of 2.5–5.0 μM in PDB+agarose. (ESF22B is the processed central portion of ESF39A, without the amino end AP24 signal peptide sequence used to traffic the peptide out of the cell and the carboxyl end 6× histidine tail. (Powell et al. 2000).) However, when wild-type American elm sap was mixed with ESF22B and incubated at room temperature for 30 min prior to the assay, peptide concentrations of at least 42 μM were required to inhibit O. novo-ulmi. Two separate assays showed that the MIC of ESF22B, when exposed to wild-type American elm sap, was about ten times greater than the MIC without sap (Fig. 4). Sap from FS5, AN1F, AN1X, AN1H (other AN1) or wild-type control American elm leaves, when mixed 1:5 or 1:1 with PDB+agarose, did not inhibit growth of O. novo-ulmi in these in vitro assays. Therefore, the peptide concentration as expressed in vivo must be less than 84 μM in sap from all transgenic lines tested.
https://static-content.springer.com/image/art%3A10.1007%2Fs00299-007-0313-z/MediaObjects/299_2007_313_Fig4_HTML.gif
Fig. 4

In vitro MIC assay showing growth inhibition of O. novo-ulmi by ESF22B antimicrobial peptide. Top three rows contain ESF22B (the central portion of ESF39A) mixed with wild-type elm sap; bottom three rows contain only ESF22B. Substrate is Potato Dextrose Broth in low-melting-point Agarose. X-ed wells are empty

Inoculation

Both wild-type and transgenic American elm trees inoculated with the O. novo-ulmi conidia solution showed typical DED symptoms (leaf wilt and sapwood staining) to varying extents. O. novo-ulmi could be reisolated from stem sections of inoculated trees 14–18 weeks after inoculation indicating that the inoculations were successful. Some symptomatic differences were observed after an O. novo-ulmi inoculation experiment in which six non-transformed tissue culture-derived trees were compared to six FS5 and a mix of six AN1 transgenic trees. Eleven weeks after inoculation, non-transformed trees retained an average of 9% of their leaves, FS5 trees retained 41%, and AN1 trees retained 67%. These numbers decreased to 7, 18, and 52%, respectively, after 13 weeks (Fig. 5a), and were proportionally similar after 14 weeks. Despite the consistent trends, none of these differences in leaf loss between clones were statistically significant (P > 0.05). Staining was present to some extent in all trees, but the maximum distance at which the staining was observed below the inoculation point (Scala et al. 1997) varied significantly 14 weeks after inoculation (Fig. 5b). ANOVA and Tukey’s HSD test revealed that in AN1 trees, staining spread a significantly shorter distance below the inoculation point than in non-transgenic control trees (P < 0.05), but spread of staining in FS5 trees was not significantly different from the other groups. Additionally, stem tissue above the inoculation point was dried and dead in four of the six non-transgenic controls and four of the six FS5 trees, but the entire stem was alive in all six AN1 trees.
https://static-content.springer.com/image/art%3A10.1007%2Fs00299-007-0313-z/MediaObjects/299_2007_313_Fig5_HTML.gif
Fig. 5

O. novo-ulmi inoculation experiment results, n = 18 (6 individuals/group). AN1 events are pooled due to small sample sizes for individual events. a Mean percentage of live leaves remaining on trees 13 weeks after inoculation. b Mean distance staining was observed below inoculation point

Mycorrhizae

Both transgenic and wild-type seedling American elm trees formed extensive mycorrhizal associations within 3 months of field planting (Fig. 2 shows mycorrhizal fungal structures found in both transgenic and wild-type roots). Mycorrhizal fungi colonized 76% (±3%) of transgenic elm root length, and 75% (±15%) of wild-type seedling root length. This difference was not significant (P = 0.959) in an unpaired t test.

Discussion

The transgenic elms in this study represent the first American elm trees transformed with a putative resistance gene. American elm thus joins an increasing number of woody plants transformed for food (Norelli et al. 1998), pulp (Tournier et al. 2003), timber (Charity et al. 2005), and disease resistance (Escobar et al. 2002; Liang et al. 2002; Pappinen et al. 2002; Polin et al. 2006). Antimicrobial peptides (AMPs) in particular are becoming more common as transgene products in plants, and a variety of transgenic plants expressing AMPs for disease resistance have been reported recently (Alan et al. 2004; Bi et al. 1999; Powell et al. 2005).

Transformation and molecular tests

Transformation experiments in the current work produced at least three separate transgenic lines, designated AN1C, AN1F, and AN1X, which all arose from a single leaf piece (Not all putatively transgenic shoots have been analyzed via Southern hybridization, so further testing may reveal additional lines). Earlier work with the same explant type and plasmid construct produced one additional transgenic line, designated FS5. The fact that all tested events are single-copy is unusual (Pappinen et al. 2002; Polin et al. 2006), but chances of false single-copy results were reduced by probing separate blots of DNA digested with different restriction endonucleases.

Reverse-transcriptase PCR (RT-PCR) with custom primers ESF39F and ESF39R confirmed that the ESF39a peptide was being expressed. RNA from both wild-type and transgenic elm vascular tissue was tested, and the presence of the RNA transcript in the transgenic samples, visible as a band of the expected length (approximately 160 bp), confirmed transcription.

Inoculation and peptide inhibition

Inoculation tests performed in this project should be considered preliminary, since young elm trees are not as susceptible to DED as mature trees (Gartland et al. 2000; Townsend et al. 1995). Therefore, most researchers inoculate field-grown trees at a minimum age of 3 years and report resistance as a function of percent crown dieback (Elgersma 1973; Scala et al. 1997; Sherald et al. 1994; Townsend and Douglass 2004). Also, elms are most susceptible to DED in the early spring soon after they produce new growth (Hubbes 1999; Santini et al. 2005), and though trees at this stage were selected whenever possible from the growth chamber or greenhouse, they may not represent true field conditions. Despite these limitations, AN1 transgenic trees showed a consistent trend of reduced spread of O. novo-ulmi compared to wild-type trees. (Multiple AN1 transformation events were pooled in these tests due to the small available sample size for each individual event). A statistically significant difference between clones was shown in the distance which sapwood staining was observed below the inoculation point. The same measurement was taken by Scala et al. (1997), who differentiated between O. ulmi and O. novo-ulmi pathogenicity based on distance of fungal spread in sections below the inoculation point. Measurements of percent crown dieback were not possible given that most of the inoculated trees consisted of a single shoot, but percentage of leaves remaining after inoculation is a relatively close approximation of this at a reduced scale, and is similar to the percent defoliation measurement used by Santini et al. (2005). The variations in remaining leaves and fungal reisolation between clones were not statistically significant, but they support the trend of reduced spread of O. novo-ulmi in AN1 transgenic trees.

The core of the ESF39A synthetic antimicrobial peptide (ESF22B) has been shown to inhibit growth of multiple bacterial and fungal pathogens in vitro, and Septoria musiva specifically in vivo (Liang et al. 2002; Powell et al. 2000). However, the concentration at which this peptide inhibits growth of Ophiostoma novo-ulmi had not previously been determined. Therefore, O. novo-ulmi was cultured with 0–100 μM dilutions of ESF22B (the central, processed portion of ESF39A) to determine the in vitro MIC (Minimum Inhibitory Concentration). Dilutions of ESF22B were also mixed with American elm sap to more closely simulate in vivo conditions. The resulting MIC was approximately an order of magnitude higher than the MIC without sap. Therefore, American elm sap reduces the activity of ESF22B.

It is possible that plant proteases in the sap are responsible for these differences, but they do not have a completely inhibitory effect. Alternatively, several studies have shown that antimicrobial peptide activity is reduced in the presence of mono- and divalent cations including Na+, K+, Mg2+, and Ca2+ (Harrison et al. 1997; Jacobi et al. 2000; Shin et al. 2002). Further studies have shown some antimicrobial peptides to be effective in vitro, but to have no effect in vivo (De Bolle et al. 1996; Yeaman et al. 2002), or to have a reduced effect in the presence of salts similar to those in biological fluids (Mills and Hammerschlag 1993; Terras et al. 1993). American elm leaves are known to have relatively high levels of calcium (Bey 1990), which could interfere with the AMP activity and therefore account for the higher MIC when ESF22B was mixed with sap. This would explain the presence of some wilt symptoms and staining in transgenic trees known to express this AMP, if the expression was below the level necessary to completely inhibit O. novo-ulmi. Complete inhibition of the fungus, however, may not be necessary to provide useful levels of field resistance. Slowing down fungal growth could allow the plant’s natural defenses such as tyloses or mansonones (Rioux et al. 1995; Sticklen et al. 1991) to inhibit the fungus or slow its spread enough to manage the infection and keep the tree alive. Additionally, disease resistance in transgenic plants could be improved by the use of a different promoter, ‘pyramiding’ transgenes (Dixon 2001; Li et al. 2003), or simply creating and testing more transformation events, since expression can vary greatly between multiple events transformed with the same construct (Peach and Velten 1991).

Mycorrhizae

All American elm roots observed were colonized to some extent by arbuscular mycorrhizal (AM) fungi. Mycorrhizal effects on plant growth were not measured in this project, but mycorrhizal relationships are formed with almost all types of plants, and in many cases, the relationship is beneficial or essential for both organisms (Smith and Read 1997). One study specifically examining mycorrhizal effects on American elm concluded that mycorrhizal roots showed increased hydraulic conductance and conductivity, which would help plant growth, especially in cold conditions (Mushin and Zwiazek 2002). It is important that any new transgenic plants are able to form regular mycorrhizal relationships, especially those intended for use outside managed (regularly fertilized) agricultural areas. The results of one study linking AM colonization to phytoplasma-induced yellows disease resistance in tomato (Lingua et al. 2002) are especially encouraging for applications involving American elm, which is susceptible to elm yellows.

Mean colonization rates for transgenic and wild-type elms were essentially identical. Therefore, according to the data collected so far, mycorrhizal colonization of American elm trees is not affected by transformation with the ESF39A antimicrobial peptide.

Future work

One important consideration for any American elm disease-resistance research program is susceptibility to elm yellows. The causative organism of this disease, variably called a phytoplasma (Lingua et al. 2002) or a mycoplasmalike organism (Zhao et al. 2004), is part of a weakly defined group of microorganisms that cause yellows diseases by infecting phloem tissue in a variety of plant species. Symptoms of elm yellows, as the name suggests, start with yellowing of leaves and usually result in death within a few seasons of infection (Sinclair 2000). The vascular-specific promoter and antimicrobial peptide used in the current project should be tested to determine whether they are active in both phloem and xylem vessels and whether they convey resistance to elm yellows, which has been reported for homologous gene promoters in other plant systems (Zhang et al. 2000).

Conclusions

Transgenic American elm trees, expressing the synthetic antimicrobial peptide ESF39A, show some signs of resistance to Dutch-elm disease. Three new single-copy transgenic lines were produced and shown to express ESF39A transgene, whose processed peptide product was shown to inhibit O. novo-ulmi in vitro. This represents the first transformation protocol for American elm using putative resistance-enhancing transgenes. Mycorrhizal associations with transgenic trees in this study are not significantly different than those on wild-type trees. Therefore, transformation with an antimicrobial peptide gene is a potentially safe and effective method for restoring American elm trees.

Acknowledgments

ArborGen, LLC provided funding for this research. Thanks to Drs. Tom Horton and Larry Smart for shared expertise and use of lab equipment. Thanks and best of luck to Nick Kaczmar for carrying on this research. Many thanks to Megan Newhouse for valuable advice, support and encouragement.

Copyright information

© Springer-Verlag 2007