Bacterial endophytes: recent developments and applications


  • Editor: Richard Staples

Correspondence: David N. Dowling, Department Science & Health, Institute of Technology Carlow, Kilkenny Road, Carlow, Ireland. Tel.: +353 59 9170479; fax: +353 59 9170517; e-mail:


Endophytic bacteria have been found in virtually every plant studied, where they colonize the internal tissues of their host plant and can form a range of different relationships including symbiotic, mutualistic, commensalistic and trophobiotic. Most endophytes appear to originate from the rhizosphere or phyllosphere; however, some may be transmitted through the seed. Endophytic bacteria can promote plant growth and yield and can act as biocontrol agents. Endophytes can also be beneficial to their host by producing a range of natural products that could be harnessed for potential use in medicine, agriculture or industry. In addition, it has been shown that they have the potential to remove soil contaminants by enhancing phytoremediation and may play a role in soil fertility through phosphate solubilization and nitrogen fixation. There is increasing interest in developing the potential biotechnological applications of endophytes for improving phytoremediation and the sustainable production of nonfood crops for biomass and biofuel production.


Beneficial plant–microbe interactions that promote plant health and development have been the subject of considerable study. Recent work has also investigated their potential for the enhanced biodegradation of pollutants in soil. Most of these studies have focused on bacteria from the rhizosphere of plants (Lindow & Brandl, 2003; Kuiper et al., 2004; Berg et al., 2005). Endophytic bacteria can be defined as those bacteria that colonize the internal tissue of the plant showing no external sign of infection or negative effect on their host (Holliday, 1989; Schulz & Boyle, 2006), and of the nearly 300 000 plant species that exist on the earth, each individual plant is host to one or more endophytes (Strobel et al., 2004). Only a few of these plants have ever been completely studied relative to their endophytic biology. Consequently, the opportunity to find new and beneficial endophytic microorganisms among the diversity of plants in different ecosystems is considerable.

Bacterial endophytes colonize an ecological niche similar to that of phytopathogens, which makes them suitable as biocontrol agents (Berg et al., 2005). Indeed, numerous reports have shown that endophytic microorganisms can have the capacity to control plant pathogens (Sturz & Matheson, 1996; Duijff et al., 1997; Krishnamurthy & Gnanamanickam, 1997), insects (Azevedo et al., 2000) and nematodes (Hallmann et al., 1997, 1998). In some cases, they can also accelerate seedling emergence, promote plant establishment under adverse conditions (Chanway, 1997) and enhance plant growth (Bent & Chanway, 1998). Bacterial endophytes have been shown to prevent disease development through endophyte-mediated de novo synthesis of novel compounds and antifungal metabolites. Investigation of the biodiversity of endophytic strains for novel metabolites may identity new drugs for effective treatment of diseases in humans, plants and animals (Strobel et al., 2004).

Along with the production of novel chemicals, many endophytes have shown a natural capacity for xenobiotic degradation or may act as vectors to introduce degradative traits. The ability of some endophytes to show resistance to heavy metals/antimicrobials and degrade organic compounds probably stems from their exposure to diverse compounds in the plant/soil niche. This natural ability to degrade these xenobiotics is being investigated with regard to improving phytoremediation (Siciliano et al., 2001; Barac et al., 2004; Germaine et al., 2004, 2006; Porteous-Moore et al., 2006; Ryan et al., 2007a). This review aims to provide an overview of the potential applications for bacterial endophytes particularly in the area of phytoremediation and sustainable agriculture.

Isolation and biodiversity of bacterial endophytes

The endophytic niche offers protection from the environment for those bacteria that can colonize and establish in planta. These bacteria generally colonize the intercellular spaces, and they have been isolated from all plant compartments including seeds (Posada & Vega, 2005). Endophytic bacteria have been isolated from both monocotyledonous and dicotyledonous plants, ranging from woody tree species, such as oak and pear, to herbaceous crop plants such as sugar beet and maize. Classical studies on the diversity of bacterial endophytes have focused on characterization of isolates obtained from internal tissues following disinfection of plant surfaces with sodium hypochlorite or similar agents (Miche & Balandreau, 2001). A review by Lodewyckx et al. (2002) highlights the methods used to isolate and characterize endophytic bacteria from different plant species. A very comprehensive list of bacterial endophytes isolated from a broad range of plants is provided by Rosenblueth & Martinez-Romero (2006) and Berg & Hallmann (2006), which updates the groundwork laid by Hallmann et al. (1997) and Lodewyckx et al. (2002).

A recent study by Porteous-Moore et al. (2006) describes the diversity of endophytes found in poplar trees, growing at a phytoremediation field site contaminated with toluene, with the aim of identifying potential candidates for enhancing phytoremediation of toluene, ethylbenzene, and xylene (BTEX) compounds. Endophytic bacteria were isolated from two varieties of Poplar. The isolated endophytic bacteria were characterized by comparative sequence analysis of partial 16S RNA genes, BOX-PCR profiling of genomic DNA and physiological characterization, with respect to substrate utilization, antibiotics and heavy-metal sensitivities. This study and those of Germaine et al. (2004, 2006) demonstrated that within the diverse bacterial communities found in Poplar trees, several endophytic strains were present that had the potential to enhance phytoremediation of volatile organics and herbicides.

Molecular approaches for the isolation and characterization of bacterial endophytes and plant-associated bacteria and communities have been reviewed recently by Franks et al. (2006). Microbial communities inhabiting stems, roots and tubers of various varieties of plants were analysed by 16S rRNA gene-based techniques such as terminal restriction fragment length polymorphism analysis, denaturing gradient gel electrophoresis as well as 16S rRNA gene cloning and sequencing. Five taxa exhibiting the most promising levels of colonization and an ability to persist were identified as Cellulomonas, Clavibacter, Curtobacterium, Pseudomonas and Microbacterium by 16S rRNA gene sequence, fatty acid and carbon source utilization analyses (Elvira-Recuenco & van Vuurde, 2000; Zinniel et al., 2002).

Studies that make use of both culture-based and culture-independent techniques can be particularly useful. The presence and taxonomy of endophytic bacteria of the entire aerial parts of Crocus (Crocus albiflorus) was investigated by Reiter & Sessitsch (2006). Their results suggest that Crocus supports a diverse bacterial microbial communities resembling the microbial communities that have been described for other plants, but also containing species that have not been described in association with plants before. A combination of plating of plant macerates, isolation and 16S rRNA gene sequence identification of isolates and whole-community fingerprinting was used. The results clearly indicated that a wide range of bacteria from diverse phylogenetic affiliation, mainly Gammaproteobacteria and Firmicutes, live in association with plants of Crocus albiflorus. The community composition of the culturable component of the microbial communities was different from that of the 16S rRNA gene clone library. Only three bacterial divisions were found in the culture collection, which represented 17 phylotypes, whereas six divisions were identified in the library analysis comprising 38 phylotypes. The predominant group in the culture collection was the low G+C Gram-positive group, whereas in the clone library, the Gammaproteobacteria predominated. This study confirms that the culturable endophytes are a subset of total endophyte biodiversity.

Plant colonization and inoculation with endophytes

Autofluorescent protein (AFP) methods are now a key tool for studying processes such as microbe–plant interactions and biofilm formation (for a recent review, see Larrainzar et al., 2005). These techniques have been utilized to detect and enumerate microorganisms in situ on plant surfaces and in planta (Gage et al., 1996; Tombolini et al., 1997; Tombolini & Jansson, 1998). One of these AFP strategies uses a marker system, which encodes the green fluorescent protein (GFP). This technique has shown promise in monitoring pseudomonads in root tissues (Tombolini et al., 1997; Tombolini & Jansson, 1998). GFP is a useful AFP biomarker because it does not require any substrate or cofactor in order to fluoresce. Workers have developed GFP cassettes for chromosomal integration and expression of gfp in a variety of bacteria (Tombolini et al., 1997; Tombolini & Jansson, 1998; Xi et al., 1999). Bacterial cells with chromosomal integration of gfp can be identified by epifluorescence microscopy or confocal laser scanning microscopy (Villacieros et al., 2003; Germaine et al., 2004).

Germaine et al. (2004) investigated a number of GFP-labelled Poplar endophytes for their colonization ability and also explored different methods of inoculation; a simple ‘stick dipping’ method was found to be very efficient, leading to extensive colonization of the specific tissues of Poplar plants (see Fig. 1) at levels of 102 to 104 CFU g−1 tissue depending on the strain, whereas a seed ‘imbibing’ method was efficient for inoculation of pea plants (Germaine et al., 2006; Germaine, 2007) (Fig. 2).

Figure 1.

 Schematic diagram of the different plant–bacterial endophyte interactions that have been studied and their applications

Figure 2.

 Application of AFPs for studying bacterial endophyte–plant interactions. (a) Xylem tracheid pits of inoculated poplar trees showing colonization by Pseudomonas putida VM1453 cells (× 1000) (Germaine et al., 2004). Scale bar=10 μM. (b) Pseudomonas putida VM1450 micro-colony within the root cortex of inoculated pea plants exposed to 54 mg 2,4-D (× 1000) (Germaine et al., 2006).

Endophyte colonization has also been visualized with the use of the β-glucuronidase (GUS) reporter system. A GUS-marked strain of Herbaspirillum seropedicae Z67 was inoculated onto rice seedlings. GUS staining was most intense on coleoptiles, lateral roots, and also at some of the junctions of the main and lateral roots (James et al., 2002). This study by James et al. (2002) showed that endophytes entered the roots through cracks at the point of lateral root emergence. Herbaspirillum seropedicae subsequently colonized the root intercellular spaces, aerenchyma and cortical cells, with a few penetrating the stele to enter the vascular tissue. The xylem vessels in leaves and stems were also colonized.

Successful endophyte colonization also involves a compatible host plant. Recently, work by Miche et al. (2006) investigated the obligate nitrogen-fixing endophyte Azoarcus sp. strain BH72, which expresses nitrogenase (nif) genes inside rice roots. They used a proteomic approach to dissect responses of rice roots toward bacterial colonization and jasmonic acid treatment (which induces plant defence proteins). Data suggest that induced plant defence responses may contribute to restricting endophytic colonization in grasses.

Few studies have been published describing the molecular basis of the interactions between endophytic bacteria and plants. Adapting strategies that have been used to study bacterial gene expression in the rhizosphere and phyllosphere such as in vivo expression technology (IVET) and recombination in vivo expression technology (Leveau & Lindow, 2001; Preston et al., 2001; Zhang et al., 2006) may provide an insight into genes that are required by bacteria to enter, compete, colonize the plant, suppress pathogens and generally survive within the plant.

Genomics of endophyte bacteria

To date, few endophytic bacterial genome sequences have been published; however, genome sequencing of a number of endophytes including Enterobacter sp.638, Stenotrophomonas maltophilia R551-3, Pseudomonas putida W619, Serratia proteamaculans 568 and Methylobacterium populi BJ001 is underway at the United States Department of Energy Joint Genome Institute ( Recently, the complete genome sequence of the nitrogen-fixing endophyte, Azoarcus sp. strain BH72 (Hurek & Reinhold-Hurek, 2003; Krause et al., 2006) has been compared with that of the related soil bacterium strain Azoarcus sp. strain EbN1 and other plant-associated bacteria. The BH72 genome lacks genes encoding type III and type IV secretion systems, toxins, nodulation factors, common enzymes that hydrolyses plant cell walls and the N-acyl homoserine lactone-based quorum-sensing system, which is found in many plant-associated bacteria and plant pathogens (Rainey, 1999; Preston et al., 2001; Büttner & Bonas, 2006). However, Krause et al. (2006) identified other factors encoded by the BH72 genome that may be involved in the host interaction. These include type IV pili, surface polysaccharides, type I and II protein secretion systems, flagella and chemotaxis proteins and a large number of ferric-siderophore uptake systems. The BH72 genome provides valuable insights into the biology of bacterial endophytes, and as more endophyte genome sequences become available, this will provide a rational basis to design experiments to investigate the mechanisms involved in successful endophyte colonization.

Plant growth-promoting endophytes

Research has been conducted on the plant growth-promoting abilities of various rhizobacteria. They differ from biocontrol strains in that they do not necessarily inhibit pathogens but increase plant growth through the improved cycling of nutrients and minerals such as nitrogen, phosphate and other nutrients.

Endophytes also promote plant growth by a number of similar mechanisms. These include phosphate solubilization activity (Verma et al., 2001; Wakelin et al., 2004), indole acetic acid production (Lee et al., 2004) and the production of a siderophore (Costa & Loper, 1994). Endophytic organisms can also supply essential vitamins to plants (Pirttila et al., 2004). Moreover, a number of other beneficial effects on plant growth have been attributed to endophytes and include osmotic adjustment, stomatal regulation, modification of root morphology, enhanced uptake of minerals and alteration of nitrogen accumulation and metabolism (Compant et al., 2005a, b). The recent areas where these plant growth-promoting bacterial endophytes are being used are in the developing areas of forest regeneration and phytoremediation of contaminated soils.

Biocontrol and endophytes

Endophytic bacteria are able to lessen or prevent the deleterious effects of certain pathogenic organisms. The beneficial effects of bacterial endophytes on their host plant appear to occur through similar mechanisms as described for rhizosphere-associated bacteria. These mechanisms have been reviewed in great detail by Kloepper et al. (1999) or, more recently, by Gray & Smith (2005) and Compant et al. (2005a). Diseases of fungal, bacterial, viral origin and in some instances even damage caused by insects and nematodes can be reduced following prior inoculation with endophytes (Kerry, 2000; Sturz et al., 2000; Ping & Boland, 2004; Berg & Hallmann, 2006).

It is believed that certain endophyte bacteria trigger a phenomenon known as induced systemic resistance (ISR), which is phenotypically similar to systemic-acquired resistance (SAR). SAR develops when plants successfully activate their defence mechanism in response to primary infection by a pathogen, notably when the latter induces a hypersensitive reaction through which it becomes limited in a local necrotic lesion of brown desiccated tissue (van Loon et al., 1998). ISR is effective against different types of pathogens but differs from SAR in that the inducing bacterium does not cause visible symptoms on the host plant (van Loon et al., 1998). Bacterial endophytes and their role in ISR have been reviewed recently by Kloepper & Ryu (2006).

Natural products from endophytic bacteria

Many endophytes are members of common soil bacterial genera, such as Pseudomonas, Burkholderia and Bacillus (Lodewyckx et al., 2002). These genera are well known for their diverse range of secondary metabolic products including antibiotics, anticancer compounds, volatile organic compounds, antifungal, antiviral, insecticidal and immunosuppressant agents. While a wide range of biologically active compounds have been isolated from endophytic organisms, they still remain a relatively untapped source of novel natural products.

While most research has focused on fungal-based production of antimicrobial products, a number of low-molecular-weight compounds, active at low concentrations against a range of human, animal and plant pathogenic bacteria, have been isolated from bacterial endophytes (Table 1). One member of the plant-associated fluorescent pseudomonads, Pseudomonas viridiflava, which has been isolated on and within the tissues of many grass species (Miller et al., 1998), was found to produce two novel antimicrobial compounds called ecomycins. Ecomycins represent a family of novel lipopeptides and are made up of some unusual amino acids including homoserine and β-hydroxy aspartic acid. It was found that these compounds were able to inhibit the human pathogens Cryptococcus neoformans and Candida albicans (Miller et al., 1998). While viral inhibitors have been isolated from Cytomaema sp. of fungi, such as cytonic acids A and D, these were found to inhibit the human cytomegalovirus (Guo et al., 2000). However, comprehensive screens for antiviral compounds from bacterial endophytes have yet to be reported.

Table 1.   Natural products that have been derived or produced from various endophytic bacteria
OrganismPlant associationActive agentActivityReference
Taxomyces andreanaeTaxus brevifoliaTaxolAnticancerStrobel et al. (1993)
Pseudomonas viridiflavaGrassEcomycins B and CAntimicrobialMiller et al. (1998)
Streptomyces griseusKandelia candelp-Aminoacetophenonic acidsAntimicrobialGuan et al. (2005)
Streptomyces NRRL 30562Kennedia nigriscansMunumbicins
Munumbicin D
Castillo et al. (2002)
Streptomyces NRRL 30566Grevillea pteridifoliaKakadumycinsAntibioticCastillo et al. (2003)
Serratia marcescensRhyncholacis penicillataOocydin AAntifungalStrobel et al. (2004)
Paenibacillus polymyxaWheatFusaricidin A–DAntifungalBeck et al. (2003)
Lodge pine  Li et al. (2007)
Green beans   
Arabidopsis thaliana   
Canola  Beatty & Jensen (2002)
Cytonaema sp.Quercus sp. 103Cytonic acids A and DAntiviralGuo et al. (2000)
Streptomyces sp.Monstera sp.CoronamycinAntimalarial antifungalEzra et al. (2004)

Bioplastics are biomaterials that are receiving increasing commercial interest. Lemoigne (1926) first described a bioplastic, poly-3-hydroxybutyrate (PHB) produced by Bacillus megaterium. They are polyesters, produced by a range of microorganisms cultured under different nutrient and environmental conditions. The most widely produced microbial bioplastics are poly-3-hydroxyalkanoate (PHA) and PHB. Genomic analysis indicates that many species of bacteria have the potential to produce bioplastics (Kalia et al., 2003).

Herbaspirillum seropedicae, a diazotrophic endophyte, can colonize a variety of higher plants and utilize a diverse range carbon sources. Catalán et al. (2007) have shown that H. seropedicae accumulates significant levels of PHB, when grown on a range of individual carbon sources. The design and development of bacteria and higher plants able to accumulate PHAs may also help to streamline cost-effective production and to produce novel heteropolymers for a range of applications (Aldor & Keasling, 2003).

Endophytic microorganisms with the potential to improve phytoremediation

Siciliano et al. (2001) showed that plants grown in soil contaminated with xenobiotics naturally recruited endophytes with the necessary contaminant-degrading genes. Indeed, in field sites contaminated with nitro-aromatics, genes encoding for nitro-aromatic compound degradation were more prevalent in endophytic strains than within rhizospheric or soil microbial communities. Van Aken et al. (2004) also showed that a phyto-symbiotic strain of Methylobacterium, which was isolated from hybrid Poplar trees (Populus deltoids x nigra), was capable of biodegrading numerous nitro-aromatic compounds such as 2,4,6-trinitrotoluene. In the absence of a natural biodegradation ability, genetically engineered strains can be constructed and tailor-made for the desired application. Lodewyckx et al. (2001), demonstrated that endophytes of yellow lupin, genetically constructed for nickel resistance, were able to increase the nickel accumulation and tolerance of inoculated plants.

An application of bacterial endophytes with considerable biotechnological potential was described by Barac et al. (2004), who showed that engineered Burkholderia cepacia G4 could increase plant tolerance to toluene, and decrease the transpiration of toluene to the atmosphere. Because toluene is one of the four components of BTEX contamination, this has the potential to improve phyto-remediation by decreasing toxicity and increasing degradation of the xenobiotic (Barac et al., 2004). Table 2 outlines a number of studies that demonstrate the potential role of endophytes in phytoremediation.

Table 2.   A non-exhaustive list of pollutants that have been associated with bacterial endophyte phyto-remediation strategies
CompoundPlant associationOrganismReference
  1. TNT, 2,4,6-trinitrotoluene; 2,4-D, 2,4-dichlorophenoxyacetic acid; TNT, 2,4,6-trinitrotoluene; RDX, hexahydro-1,3,5-trinitro-1,3,5-triazene; HMX, octahydro-1,3,5,7-tetranitro-1,3,5-tetrazocine; NDAB, aliphatic nitramine 4-nitro-2,4-diazabutanal; BTEX, benzene, toluene, ethylbenzene, and xylene; TCP, 2,3,4,6-tetrachlorophenol; PCB, polychlorinated biphenyl.

Mono- and dichlorinated benzoic acidsWild rye (Elymus dauricus)Pseudomonas aeruginosa strain R75 and Pseudomonas savastanoi strain CB35Siciliano et al. (1998)
2,4-DPoplar (Populus) and willow (Salix)P. putida VM1450Germaine et al. (2006)
MethanePoplar tissues (Populus deltoidesnigra DN34)Methylobacterium populi BJ001Van Aken et al. (2004)
TNT, RDX, HMXPoplar tissues (Populus deltoidesnigra DN34)Methylobacterium populi BJ001Van Aken et al. (2004)
MTBE, BTEX, TCEPopulus cv. Hazendans and cv. HoogvorstPseudomonas spGermaine et al. (2004), Porteous-Moore et al. (2006)
ToluenePoplar (Populus)B. cepacia Bu61(pTOM-Bu61)Taghavi et al. (2005)
TCP and PCBWheatHerbaspirillum sp. K1Mannisto et al. (2001)
Volatile organic compounds and tolueneYellow lupine (Lupinus luteus L.)Burkholderia cepacia G4Barac et al. (2004)

Germaine et al. (2006) inoculated pea plants with a Pseudomonas endophyte capable of degrading the organo-chlorine herbicide, 2,4-dichlorophenoxyacetic acid (2,4-D). When inoculated plants were exposed to 2,4-D, they showed no accumulation of the herbicide into their tissues and experienced little or no signs of phytotoxicity, whereas uninoculated plants showed significant accumulation of 2,4-D and displayed signs of toxicity including a reduction in biomass, leaf abscission and callus development on their roots. Large rhizosphere populations were also observed during these experiments, which were responsible for enhanced degradation of 2,4-D in soil.

The possible advantages of using endophytic microorganisms to improve xenobiotic remediation were summarized by Newman & Reynolds (2005), a major advantage being where genetic engineering of a xenobiotic degradation pathway is required, bacteria are easier to manipulate than plants. In addition, quantitative gene expression of pollutant catabolic genes within the endophytic populations could be a useful monitoring tool for assessing the efficiency of the remediation process. The unique niche of the interior plant environment provides the xenobiotic degrader strain with an ability to reach larger population sizes due to reduced competition. Another important advantage of using endophytic pollutant degraders is that any toxic xenobiotics taken up by the plant may be degraded in planta, thereby reducing phytotoxic effects and eliminating any toxic effects on herbivorous fauna residing on or near contaminated sites.

The endophyte niche is a hot spot for horizontal gene transfer (HGT) as demonstrated by Taghavi et al. (2005). In this study, the degradative plasmid, pTOM-Bu61, was found to have transferred naturally to a number of different endophytes in planta. This HGT activity promoted more efficient degradation of toluene in poplar plants. HGT in planta is likely to be widespread, as studies in pea with Pseudomonas endophytes harbouring the plasmids pWWO and pNAH7 also show high rates of transfer into a range of autochthonous endophytes (Ryan et al., 2007b; E. Keogh & D.N. Dowling, unpublished data). This approach may have practical applications in equipping the natural endophyte populations with the capacity to degrade a pollutant and does not require long-term establishment of the inoculant strain.

It is has been demonstrated that endophytic bacteria efficiently expressing the necessary catabolic genes can promote the degradation of xenobiotic compounds or their metabolites as they are accumulated or while being translocated in the vascular tissues of the host plant. With phytoremediation playing an ever-increasing role in the clean-up of contaminated land and water, it is envisaged that endophytes will play a major role in enhancing both the range of contaminants that can be remediated and the rate of their degradation.

The endophyte niche and emerging pathogens

Various opportunistic bacterial pathogens including Burkholderia, Enterobacter, Herbaspirillum, Ochrobactrum, Pseudomonas, Ralstonia, Staphylococcus and Stenotrophomonas have been identified as colonizers of the plant rhizosphere (Berg et al., 2005). Many facultative endophytes are recruited from the large pool of soil and rhizospheric species and bacteria adapted for living in planta may include opportunistic human/animal pathogens. Some bacteria associated with human infections have been isolated from the interior of alfalfa (Ponka et al., 1995). This research area needs further investigation to establish the risks, if any, associated with the development of the endophytic niche for biotechnological applications.

Concluding remarks

Exploitation of endophyte–plant interactions can result in the promotion of plant health and can play a significant role in low-input sustainable agriculture applications for both food and nonfood crops. With the availability of complete genome sequences of key endophytic bacteria, the genes governing colonization and establishment of endophytic bacteria in planta can be identified. This information will form the foundation for transcriptome and proteome analysis currently optimized in studying other plant–microbe interactions. Incorporating this information with well-established techniques such as IVET and recently advanced ‘omic’ technologies offers the ability to search for genes on a global scale that are found to be induced or repressed during colonization of plant tissues. An understanding of the mechanisms enabling these endophytic bacteria to interact with plants will be essential to fully achieve the biotechnological potential of efficient plant–bacterial partnerships for a range of applications. One promising area of research for future studies is developing endophytes (and rhizobacteria) to promote the sustainable production of biomass and bioenergy crops in conjunction with phytoremediation of soil contamination.


The authors thank J. Maxwell Dow for useful advice and for important suggestions and the other members of the BIOMERIT Research Centre for comments on this work. The authors acknowledge the funding by the Technological Sector Research Strand I and III programmes Higher Education Authority of Ireland (PRTLI Cycle 3) and Science Foundation of Ireland (SFI) BRG programme, and the DAF Stimulus programme. A.F. is supported by a fellowship from the Irish Research Council for Science, Engineering and Technology.