Transformation and regeneration of the holoparasitic plant Phelipanche aegyptiaca
© Fernández-Aparicio et al; licensee BioMed Central Ltd. 2011
Received: 31 July 2011
Accepted: 8 November 2011
Published: 8 November 2011
Transformation and subsequent regeneration of holoparasitic plants has never been reported, in part due to challenges in developing transformation protocols, but also because regeneration of obligate parasites is difficult since their survival depends completely on successful haustorium penetration of a host and the formation of vascular connections. The recent completion of a massive transcriptome sequencing project (the Parasitic Plant Genome Project) will fuel the use of genomic tools for studies on parasitic plants. A reliable system for holoparasite transformation is needed to realize the full value of this resource for reverse genetics and functional genomics studies.
Here we demonstrate that transformation of Phelipanche aegyptiaca is achieved by infection of 3 month-old in vitro grown P. aegyptiaca calli with Agrobacterium rhizogenes harboring the yellow fluorescent protein (YFP). Four months later, YFP-positive regenerated calli were inoculated onto tomato plants growing in a minirhizotron system. Eight days after inoculation, transgenic parasite tissue formed lateral haustoria that penetrated the host and could be visualized under UV illumination through intact host root tissue. YFP-positive shoot buds were observed one month after inoculation.
This work constitutes a breakthrough in holoparasitic plant research methods. The method described here is a robust system for transformation and regeneration of a holoparasitic plant and will facilitate research on unique parasitic plant capabilities such as host plant recognition, haustorial formation, penetration and vascular connection.
Parasitic weeds belonging to the genera Orobanche and Phelipanche have lost through evolution their autotrophic way of life, switching from photosynthesis to obtaining their resources by parasitizing other plants. Parasites capture host water and nutrients through a specialized organ, the haustorium, which invades the host root and connects with the host vascular system [1–3]. Phelipanche aegyptiaca (syn. Orobanche aegyptiaca) is an important parasitic weed attacking many crops in Asia and the Middle East. Conventional control based on cultural methods, herbicides, or host breeding for parasitic plant resistance, have not attained complete success due to several factors associated with the parasite life cycle, including high fecundity (hundreds of thousands of seeds per parasite), seed longevity in the soil, the subterranean location of the young parasite that effectively hides it from the farmer, the tightly coordinated parasitic and host life cycles, and the scarcity of sources of resistance in most affected crop species [4–6]. Biotechnological control approaches based on crops harboring transgenic resistance to the parasite, resistance to herbicides, or expressing silencing constructs targeting genes essential to an attached parasite remain rare, but have shown promise for protecting crops [7, 8].
Progress in understanding the unique biology of Orobanche and Phelipanche spp. has been hampered by a lack of genomic resources and transformation protocols for these parasites. A solution to the former problem has recently emerged through the Parasitic Plant Genome Project, which has produced more than 1.3 billion ESTs expressed at various life stages of the parasitic plants Triphysaria versicolor, Striga hermonthica and P. aegyptiaca. This is a major advance, but reverse genetics will be essential to discover the roles of these sequences in the parasitic phenotype and to be able to assign functions to parasite genes . To our knowledge there are no previous reports describing the development of a transformation system for any holoparasitic species. There is one report in which P. aegyptiaca gene expression was manipulated via hairpin RNAi targeted to the mannose 6-phosphate reductase gene. However, this construct acted indirectly through a transformed tomato host and relied on the systemic movement of the silencing signal through the haustorial connection, so no Orobanche transformation ocurred . Transformation has been accomplished in the facultative parasites T. versicolor and Phtheirospermum japonicum, and the former has proved useful in partially dissecting the haustorium signal transduction pathway .
The great challenge in generating transformed holoparasitic plants is in recovering a high percentage of transformants in a species for which the key step - successful establishement on a host - is a low frequency event. For Orobanche and Phelipanche seedlings, once germination is induced by host root-exuded stimulants [15, 16], survival depends on the ability to connect to the host or else perish after several days without nutritional supply from a host . In the field this short independence phase lasts for just a few days after germination, but under laboratory conditions it is possible to prolong independence for months by cultivating parasite callus in nutritive media [18, 19]. However studies based on in vitro growth of Orobanche and Phelipanche species have been considered difficult because the parasite requires a host for normal development . Here we report the development of a simple and efficient method for Agrobacterium rhizogenes-mediated transformation and subsequent regeneration by using cultured callus as starting tissue for both transformation and regeneration.
Results and Discussion
P. aegyptiaca in vitro growth
Development of Agrobacterium-mediated P. aegyptiaca transformation
Clonal propagation and regeneration of transgenic Phelipanche
The Phelipanche transformation method takes advantage of the ability to asexualy propagate Phelipanche such that multiple calli can be generated for each transformation event (Figures 3E and 3F). This clonal propagation step is a novel feature of our protocol and serves the dual function of producing sufficient material for direct assays of parasite infection via lateral haustoria and greatly enhancing chances of successfully inoculationg a host and completing the full life cycle of a transformed plant. Phelipanche in vitro calli that developed from individual seeds were previously reported to be infective on host roots growing in aseptic conditions, but the percentage of successful infections was low . Subsequently this method was improved by Zhou et al.  using different media and hormonal supplementation. A remarkable feature of our protocol is that it avoids the use of any hormonal supplementation, which simplifies both the procedure itself and subsequent analyses of tubercle gene expression by eliminating potentially confounding hormonal interactions with parasite gene expression and development.
The methodology described here presents a robust system for generating and evaluating Phelipanche transformants. First, the system is efficient in producing large numbers of parasite calli for both the Agrobacterium-mediated transformation step and the multiplication of transformed calli for experimental evaluation (e.g., Figures 3G and 3H). A principle advantage is that a single cultured tubercle can produce enough explants for a transformation experiment and such genetic uniformity can be important when studying a genetically variable organism such as P. aegyptiaca. Clonal propagation is also advantageous for multiplying transformed tissue because it ensures that even a single transformation event can be subdivided to produce sufficient numbers of transformed calli to satisfy experimental requirements. For example replication can be achieved at all levels by using more than one transformation event created from the same A. rhizogenes inoculation and then clonally propagating each event to reach any level of replication need for statistical analysis.
With respect to evaluating transgenic plants, the infective calli produced by this method are suitable for addressing many of the key questions in parasitic plant biology. Chief among these are questions that revolve around the haustorium and its initiation, development, and function. Other questions relating to post attachment development such as how tubercle growth is regulated and how roots and shoots develop are equally tractable. In contrast, questions relating to aspects of parasite seed development and germination will require completing the life cycle of the parasite. Although we have not described it here, production of transgenic P. aegyptiaca seed should be possible as this species is partially autogamous and can set seed under greenhouse or growth chamber conditions. During pilot studies we have successfully transplanted hosts and attached parasites to soil and observed development of a normal floral shoot (data not shown). While we plan to produce transgenic seed for future experiments, the currently reported method is analogous to established methods of studying transgenic facultative parasites [12–14] or using chimeric plants with transformed roots [25, 26]. The transgenic P. aegyptiaca we describe are not chimeric, but are similar to products of other A. rhizogenes transformation protocols in being useful research materials that do not require the extra time for sexual reproduction. For P. aegyptiaca the time required from A. rhizogenes inoculation to having transgenic calli ready for evaluation on a host is approximately four months.
Obligate parasitic plants are interesting examples of plant evolution and can be important agricultural pests . There is increasing interest in understanding how these plants have evolved and how crops can be made resistant to parasitism. However, research on obligate parasites has been hampered by the inherent challenges in genetically manipulating an organism that is completely dependent on a host plant. Here we describe a protocol for growing and clonally propagating P. aegyptiaca in culture such that tissue can be generated in sufficient quantity for transforming the parasite and studying the resultant transformants. The roots of the resulting transgenic calli are parasitically competent and can thus be used to study processes of haustorial initiation and function. The development of this method is timely in that it comes shortly after the first large scale release of expressed gene sequences from P. aegyptiaca by the Parasitic Plant Genome Project . Thus, it is now possible to envision research that systematically silences P. aegyptiaca gene candidates for essential parasitism genes. Many questions remain and much more work is needed, but the advent of this relatively simple and robust transformation system should greatly accelerate progress in understanding obligate parasitic plants.
P. aegyptiaca in vitro culture
P. aegyptiaca seeds were collected in 2000 from multiple parasitic plants growing on tomato in the location of May Ami, Israel (seeds provided by Dr. D.M. Joel). Seeds were surface-sterilized by sequential immersions in 75% ethanol and 1% sodium hypochlorite followed by three rinses in sterile distilled water as described previously . Sterilized seeds were spread on sterilized glass fiber filter paper (GFFP) moistened with sterile distilled water and placed inside 10 cm Petri dishes and incubated in the dark at 25/20°C for 7 days to allow the conditioning period needed to germinate. Conditioned P. aegyptiaca seeds were induced to germinate by applying the synthetic germination stimulant GR24 at 2 mg L-1. One day after GR24 application seeds were removed from the GFFP using a 1 ml pipette and plated on full strength Murashige & Skoog medium (MS) supplemented with 3% (w/v) of sucrose and 0.8% (w/v) of agar (slightly modified from Ben-Hod et al ) with the pH adjusted to 5.75. P. aegyptiaca germination percentage was observed by counting the percentage of emerged radicles seven days after GR24 (Figure 1A). Thirty days after GR24 application, the percentage of P. aegyptiaca radicles that developed tubercle-like swelling was evaluated (Figure 1B) and these calli were transferred to 150 ml Erlenmeyer flasks containing MS medium as described before except without agar. Flasks were placed in the dark at 20°C in a shaker at 100 rpm for two months. The calli were transferred to fresh medium approximately every 20 days. No hormones were added.
Plasmid and bacteria
The vector used in this work was pHG8-YFP modified to contain the mas2' promoter controlling the YFP reporter gene [14, 28]. The vector pHG8-YFP was transformed into A. rhizogenes MSU440 by electroporation .
Prior to transformation A. rhizogenes was grown for two days at 27°C in solid medium composed of peptone 5 g L-1, yeast extract 3 g L-1, and agar 8 g L-1, supplemented with spectinomycin (100 μg/μl) and acetosyringone (400 μM). A single colony was transferred to 3 ml of liquid medium containing the same components except for agar, and was allowed to grow for 16 hours at 27°C. After this incubation the A. rhizogenes was collected by centrifugation at 8,000 rpm for one minute at 4°C and the pellet resuspended in half strength MS (1/2X MS) medium without sucrose, at optical density of 0.2 at 600 nm. In a laminar flow hood, a three month-old P. aegyptiaca callus (Figure 1C) was immersed in a 10 cm Petri dish containing 15 ml of the A. rhizogenes suspension and chopped into approximately 100 small pieces, with each piece containing a few developing roots. The dish was incubated at room temperature for 20 minutes. Subsequently, the P. aegyptiaca explants were blotted against sterile filter paper and transferred to solid1/2X MS sucrose-free medium supplemented with acetosyringone 400 μM, where they were co-cultivated with A. rhizogenes in the dark at 20°C for 5 days (Figure 1D). P. aegyptiaca explants were then washed with sterile distilled water and transferred to 150 ml Erlenmeyer flasks (Figure 1E) containing full strength MS liquid medium with 3% of sucrose and supplemented with 300 mg L-1 of the antibiotic timentin (SmithKline Beecham Pharmaceuticals) and 100 μg/ml kanamycin. Thirty-five days after transformation, putative transgenic tissue was identified by the presence of YFP sectors (Figures 3A and 3B) and transformation frequency was calculated as the number of P. aegyptiaca explants with at least one area of YFP expressing tissue, using Olympus SZX12 stereomicroscope with a GFP filter set (excitation 461 to 500 nm, detection > 510 nm) and YFP-positive explants returned to fresh full strength MS medium for another 55 days.
The YFP-positive sectors developed on the calli at 90 days after transformation (Figures 3B and 3C). To select for the transgenic tissue, YFP-negative tissue was manually removed by observing and photographing calli placed inside sterile Petri dishes under UV light and then cutting away YFP-negative tissue under sterile conditions in a laminar flow hood. Then the YFP-positive tissue was returned to fresh MS liquid media and the process of observing and cutting repeated several times over the next 20 days, until all negative tissue was removed. In addition, the growing YFP-positive calli were divided into pieces to create a collection of clones for each transformation event (Figures 3E and 3F). After 110 days from the time of A. rhizogenes transformation, the P. aegyptiaca calli were transferred to liquid medium in which the strength of MS salts and sucrose was reduced by half for 10 aditional days in order to promote root and haustorium organogenesis. As negative controls, in vitro grown P. aegyptiaca calli were used following the same procedure from A. rhizogenes inoculation to callus division, but using A. rhizogenes that did not carry pHG8-YFP.
PCR analysis and Southern blot hybridization
To confirm further the integration of the YFP gene in the YFP-fluorescing tissue, genomic DNA was isolated from YFP-positive and YFP-negative calli using a DNeasy Plant Mini Kit (Qiagen). Two polymerase chain reaction (PCR) primers were designed to amplify a 1000 bp section of the mas-YFP promoter-reporter gene construct (, GenBank accession number: AY995145). A second set of PCR primers were designed to amplify the P. aegyptiaca 18S gene as a positive control. Sequences from the related parasitic species Orobanche fasciculata, Orobanche ludoviciana, Orobanche multiflora and Conopholis americana (respective GenBank accession numbers: U59960.1, U59953.1, U59952 and U59954.1) were aligned using Muscle in the Geneius Pro 5.3.4 software. The more conserved regions were used to design a set of forward and reverse 18S primers. Each PCR reaction contained 5 ng of genomic DNA or plasmid DNA, 0.5 μM of each forward and reverse primers, 12.5 of 2× iProof Master Mix (BIO-RAD) and conditions used as described in the manufacturer's protocol. PCR products were separated by electrophoresis through a 1% agarose gel (Figure 4).
Approximately 10 μg of genomic DNA from each of three different P. aegyptiaca YFP-positive calli, and YFP-negative control calli were digested with 100 u of XbaI, EcoRV or EcoRI respectively at 37°C for 16 hours. 300 ng of the plasmid DNA pHG8-YFP was similarly digested with 40 u XbaI. DNA was separated on a 0.65% TBE agarose gel at 60 V for about 24 hours. The gel was stained and photographed and then treated with 0.25 M HCl for 25 minutes. The gel was treated with 0.5 M NaOH+1.5 M NaCl two times 25 minutes each followed by treatment with 0.5 M tris, pH8.0+1.5 M NaCl two times 25 minutes each. DNA was transferred to a nylon membrane Nytran Supercharge (Whatman) overnight using a 'TurboBlotter' and 10× SSC. The membrane was UV linked, trimmed and air-dried. The filter was prehybridized using 6× SSC, 5× Denhardt's solution, and 0.5% SDS at 68°C for 4 hours. The filter was hybridized using a32P-labeled 1000 bp probe specific to the mas-YFP sequence. The probe concentration was ~3.5 × 106 dpm/ml in the HYBE buffer (same as the prehybe buffer). The hybridization was carried out at 68°C for 19 hours. The filter was then washed in 2× SSC+0.1%SDS at 68°C with three buffer changes over a period of 60 minutes. The filter was autoradiographed for 19 hours with an intensifier screen at -80°C. Due to excess loading of the control lane 1 (plasmid pHG8-YFP), the filter was re-exposed for 1 hour and then the control lane was trimmed off and the remaining filter was re-exposed for 9 days.
Demonstration of transgenic parasitic competence
In order to demonstrate parasitic competence of the transgenic calli, three YFP-positive clones from a single transformation event and three YFP-negative control clones were removed from the1/2X MS medium in which they were growing, washed with sterile distilled water and placed next to roots of six tomato plants that were 20 days old and growing in a mini-rhizotron system . To increase the chance of tomato root infection, each P. aegyptiaca callus was again chopped in small pieces containing from 3 to 6 P. aegyptiaca roots of approximately 5 mm length, just before inoculation on tomato. Ten P. aegyptiaca explants per clone were placed in close contact with the tomato roots (Figures 3G and 3H, Figures 5A and 5B). Twenty days after inoculation of transgenic P. aegyptiaca explants on tomato roots, the percentage of P. aegyptiaca attachment to the host was scored as the percentage of inoculated explants that attached to the host and produced a tubercle out of the total number of explants inoculated per tomato plant (Figures 5E and 5F). Thirty days after P. aegyptiaca explants were inoculated onto tomato roots, the percentage of shoot formation was evaluated out of the total number of regenerated P. aegyptiaca tubercles (Figures 5I and 5J).
This research is supported by International Outgoing European Marie Curie postdoctoral fellowship (PIOF-GA-2009-252538) to M Fernández-Aparicio; and projects P07-AGR-02883 and AGL2008-01239/AGR co-financed by FEDER funds to D Rubiales, NSF Plant Genome award DBI-0701748 to JH Westwood and JI Yoder, and U.S. Department of Agriculture (Hatch Project no. 135798) to JH Westwood. We thank Dr. DM Joel for supplying P. aegyptiaca seeds and Norma Constanza Manrique Carpintero for technical assistance with DNA extraction.
- Press MC: Autotrophy and heterotrophy in root hemiparasites. Trends Ecol and Evol. 1989, 4: 258-263. 10.1016/0169-5347(89)90195-X.View ArticleGoogle Scholar
- Parker C, Riches CR: Parasitic Weeds of the World: Biology and Control. 1993, CAB Int, Wallingford, UKGoogle Scholar
- Nickrent DL, Musselman LJ: Introduction to parasitic flowering plants. The Plant health instructor. 2004Google Scholar
- Rubiales D, Fernández-Aparicio M, Wegmann K, Joel D: Revisiting strategies for reducing the seedbank of Orobanche and Phelipanche spp. Weed Res. 2009, 49: 23-33.View ArticleGoogle Scholar
- Yoder JI, Scholes JD: Host plant resistance to parasitic weeds; recent progress and bottlenecks. Curr Opin Plant Biol. 2010, 13: 478-484. 10.1016/j.pbi.2010.04.011.View ArticlePubMedGoogle Scholar
- Perez-De-Luque A, Fondevilla S, Perez-Vich B, Aly R, Thoiron S, Simier P, Castillejo MA, Fernandez-Martinez JM, Jorrin J, Rubiales D, Delavault P: Understanding Orobanche and Phelipanche-host plant interactions and developing resistance. Weed Res. 2009, 49 (Suppl.1): 8-22.View ArticleGoogle Scholar
- Aly R: Conventional and biotechnological approaches for control of parasitic weeds. In Vitro Cell Dev Biol Plant. 2007, 43: 304-317. 10.1007/s11627-007-9054-5.View ArticleGoogle Scholar
- Gressel J: Crops with target-site herbicide resistance for Orobanche and Striga control. Pest Manag Sci. 2009, 65: 560-565. 10.1002/ps.1738.View ArticlePubMedGoogle Scholar
- Westwood JH, dePamphilis CW, Das M, Fernández-Aparicio M, Honaas L, Timko MP, Wickett N, Yoder JI: The parasitic plant genome project: new tools for understanding the biology of Orobanche and Striga. Weed Sci. 2012,Google Scholar
- Westwood JH, Yoder JI, Timko MP, Depamphilis CW: The evolution of parasitism in plants. Trends Plant Sci. 2010, 15: 227-235. 10.1016/j.tplants.2010.01.004.View ArticlePubMedGoogle Scholar
- Aly R, Cholakh H, Joel DM, Leibman D, Steinitz B, Zelcer A, Naglis A, Yarden O, Gal-On A: Gene silencing of mannose 6-phosphate reductase in the parasitic weed Orobanche aegyptiaca through the production of homologous dsRNA sequences in the host plant. Plant Biotechnol J. 2009, 7: 487-498. 10.1111/j.1467-7652.2009.00418.x.View ArticlePubMedGoogle Scholar
- Tomilov AA, Tomilova NB, Yoder JI: Agrobacterium tumefaciens and Agrobacterium rhizogenes transformed roots of the parasitic plant Triphysaria versicolor retain parasitic competence. Planta. 2007, 225: 1059-1071. 10.1007/s00425-006-0415-9.View ArticlePubMedGoogle Scholar
- Ishida JK, Yoshida S, Ito M, Namba S, Shirasu K: Agrobacterium rhizogenes-Mediated Transformation of the Parasitic Plant Phtheirospermum japonicum. PLoS ONE. 2011, 6: e25802. 10.1371/journal.pone.0025802.PubMed CentralView ArticlePubMedGoogle Scholar
- Bandaranayake PCG, Filappova T, Tomilov A, Tomilova NB, Jamison-McClung D, Ngo Q, Inoue K, Yoder JI: A single-electron reducing quinone oxidoreductase is necessary to induce haustorium development in the root parasitic plant Triphysaria. Plant Cell. 2010, 22: 1404-1419. 10.1105/tpc.110.074831.PubMed CentralView ArticlePubMedGoogle Scholar
- Fernández-Aparicio M, Flores F, Rubiales D: Recognition of root exudates by seeds of broomrape (Orobanche and Phelipanche) species. Ann Bot. 2009, 103: 423-431.PubMed CentralView ArticlePubMedGoogle Scholar
- Xie X, Yoneyama K, Yoneyama K: The strigolactone story. Annu Rev Phytopathol. 2010, 48: 93-117. 10.1146/annurev-phyto-073009-114453.View ArticlePubMedGoogle Scholar
- Joel DM, Hershenhorn Y, Eizenberg H, Aly R, Ejeta G, Rich PJ, Ransom JK, Sauerborn J, Rubiales D: Biology and Management of Weedy Root Parasites. Horticultural Reviews. Edited by: Janick J. 2007, John Wiley & Sons, Inc. Hoboken, NJ, 33: 267-350.View ArticleGoogle Scholar
- Ben-Hod G, Losner D, Joel DM, Mayer AM: In vitro culture of Orobanche aegyptiaca. Ann Bot. 1991, 68: 413-416.Google Scholar
- Batchvarova RB, Slavov SB, Bossolova SN: In vitro culture of Orobanche ramosa. Weed Res. 1999, 39: 191-197. 10.1046/j.1365-3180.1999.00137.x.View ArticleGoogle Scholar
- Zhou WJ, Yoneyama K, Takeuchi Y, Iso S, Rungmekarat S, Chae SH, Sato D, Joel DM: In vitro infection of host roots by differentiated calli of the parasitic plant Orobanche. J Exp Bot. 2004, 55: 899-907. 10.1093/jxb/erh098.View ArticlePubMedGoogle Scholar
- Johnson AW, Rosebery G, Parker C: A novel approach to Striga and Orobanche control using synthetic germination stimulants. Weed Res. 1976, 16: 223-227. 10.1111/j.1365-3180.1976.tb00406.x.View ArticleGoogle Scholar
- Sonti RV, Chiurazzi M, Wong D, Davies CS, Harlow GR, Mount DW, Signer ER: Arabidopsis mutant deficient in T-DNA Integration. Proc Natl Acad Sci. 1995, 92: 11786-11790. 10.1073/pnas.92.25.11786.PubMed CentralView ArticlePubMedGoogle Scholar
- Ben-Hod G, Losner D, Joel DM, Mayer AM: Infection of tomato roots using calli of broomrape. Physiol Plantarum. 1991, 83: 353-356. 10.1111/j.1399-3054.1991.tb00105.x.View ArticleGoogle Scholar
- Paran I, Gidoni D, Jacobsohn R: Variation between and within broomrape (Orobanche) species revealed by RAPD markers. Heredity. 1997, 78: 68-74. 10.1038/hdy.1997.8.View ArticlePubMedGoogle Scholar
- Veena V, Taylor CG: Agrobacterium rhizogenes: recent developments and promising applications. In Vitro Cellular & Developmental Biology-Plant. 2007, 43: 383-403. 10.1007/s11627-007-9096-8.View ArticleGoogle Scholar
- Collier R, Fuchs B, Walter N, Kevin Lutke W, Taylor CG: Ex vitro composite plants: an inexpensive, rapid method for root biology. Plant Journal. 43: 449-457.Google Scholar
- Westwood JH: Characterization of the Orobanche-Arabidopsis system for studying parasite-host interactions. Weed Sci. 2000, 48: 742-748. 10.1614/0043-1745(2000)048[0742:COTOAS]2.0.CO;2.View ArticleGoogle Scholar
- Subramanian C, Woo J, Cai X, Xu XD, Servick S, Johnson CH, Nebenfuhr A, von Arnim AG: A suite of tools and application notes for in vivo protein interaction assays using bioluminescence resonance energy transfer (BRET). Plant J. 2006, 48: 138-152. 10.1111/j.1365-313X.2006.02851.x.View ArticlePubMedGoogle Scholar
- Fernández-Aparicio M, Pérez-de-Luque A, Prats E, Rubiales D: Variability of interactions between barrel medic (Medicago truncatula) genotypes and Orobanche species. Ann App Biol. 2008, 153: 117-126. 10.1111/j.1744-7348.2008.00241.x.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.