Fusarium circinatum causes pitch canker disease in a wide range of pine trees, including Pinus radiata, with devastating economic consequences.
To assess the spatial and temporal dynamics of growth of this pathogen in radiata pine, we examined the process of infection using both real-time PCR to quantify fungal biomass inside the plant host, and confocal microscopy using a green fluorescent protein (GFP)-tagged strain of F. circinatum.
Pathogen growth exhibited three distinct phases: an initial exponential increase in fungal biomass, concomitant with pathogen colonization of the cortex and phloem; a slowdown in fungal growth coincident with sporulating hyphae deep within the host; and stabilization of the fungal biomass when the first wilting symptoms appeared. The number of resin ducts in the xylem was found to increase in response to infection and the fungus grew inside both constitutive and traumatic resin ducts.
These results indicate that conidiation may contribute to the spatial or temporal dissemination of the pathogen. Moreover, the present findings raise the intriguing possibility that the generation of traumatic resin ducts may be of more benefit to the fungus than to the plant.
Fusarium circinatum Nirenberg & O'Donnell (teleomorph=Gibberella circinata Nirenberg & O'Donnell 1998) is the causal agent of pitch canker, one of the most serious pine diseases world-wide. The lifestyle of this filamentous fungus is typical of that of necrotrophic pathogens, which do not have specialized infection structures to derive nutrients from sacrificed cells (Lewis, 1973). The pathogen causes cankers which girdle the shoots, branches and even trunks of susceptible pine species. The characteristic feature of the cankers is excessive pitch flow resulting in complete impregnation of the wood inside the canker (Hepting & Roth, 1946). Multiple branch infections may cause severe crown dieback and lead to the death of the tree. However, the most significant loss from an economic perspective is caused by reduced growth of adult trees in forest plantations (Dwinell & Phelps, 1977; Arvanitis et al., 1984), as well as the detrimental effects of the pathogen in seedling nurseries (Dwinell & Barrows-Broaddus, 1978; Viljoen et al., 1994; Wingfield et al., 2002).
Pitch canker is known to affect over 50 Pinus species which vary widely in susceptibility (Dwinell, 1978; Gordon et al., 1998a; Mitchell et al., 2012). Pinus radiata D. Don is one of the most economically important pine species in the world, as a consequence of its fast growth rate, and is also one of the most susceptible species (Hodge & Dvorak, 2000). However, even for species considered to be susceptible, the degree to which individual trees develop pitch canker varies widely (Storer et al., 1999a). In addition, the degree of tolerance in an individual may change over time as a result of induced resistance, associated with the activation of natural defense mechanisms of the host plant, resulting in changes which diminish the effects of subsequent biotic attack (Eyles et al., 2010). This phenomenon, first described in gymnosperms by Christiansen et al. (1999), has been demonstrated to occur in radiata pine trees as a response to infection by F. circinatum (Bonello et al., 2001; Gordon et al., 2011).
Conifers are organisms that live for many years and part of their success is attributable to their capacity to integrate multiple constitutive and inducible defense mechanisms into an effective coordinated response. Sometimes the same defense traits can be both constitutively expressed and inducible (Franceschi et al., 2005). This is the case for oleoresin, the most familiar of the conifer defenses. Oleoresin is a complex mixture of terpenoids which is known to deter insect pests and their symbiotic fungal pathogens (Phillips & Croteau, 1999). Terpenoid resin is constitutively synthesized and stored in specialized secretory structures. Pinus spp. have developed a complex network of highly branched radial and axial resin ducts which are interconnected by anastomoses. In this genus, resin is synthesized by the thin-walled, long-lived secretory epithelial cells lining the lumen of the resin ducts. Resin ducts are formed as a normal feature of development, linked to secondary growth, or their formation may be induced by external factors, in which case they are referred to as traumatic resin ducts (TRDs). In contrast to constitutive axial ducts, which generally occur alone and have a scattered distribution, TRDs are normally found in tangential series of one or two rows within an annual ring of xylem (Bannan, 1936). The resin formed by TRDs can be more fungistatic than primary resin, because of altered terpene content and additional phenolics (Nagy et al., 2000; Krokene et al., 2003). Several studies have analyzed this defense mechanism in the pine response to fungal pathogen invasions (Reid et al., 1967; Hudgins et al., 2005; Luchi et al., 2005; Nagy et al., 2005). However, little is known about the role of the constitutive resin ducts or TRDs during pitch canker development, despite the fact that copious pitch flow is one of its principal symptoms, as evidenced by the disease denomination.
In this sense, a more integral analysis of the spatiotemporal events that occur during the process of infection by F. circinatum is needed. Conventional light microscopy (Barrows-Broaddus & Dwinell, 1983, 1984) and, more recently, advanced microscopy technologies (Thoungchaleun et al., 2008; Kim et al., 2009) have provided valuable insights into this pathogenic interaction. Confocal laser scanning microscopy (CLSM) and the use of green fluorescent protein (GFP)-tagged fungal transformants have been shown to be new and powerful tools with which to discover novel aspects of the in vivo pathogenic interaction between organisms (see e.g. Lagopodi et al., 2002; Jansen et al., 2005; Zvirin et al., 2010). Moreover, quantitative real-time polymerase chain reaction (q-RT-PCR) is currently the most accurate method for quantifying fungal colonization of host tissues (Stephens et al., 2008; Maciá-Vicente et al., 2009). Although rapid and sensitive protocols have been developed for the specific detection of F. circinatum DNA (Schweigkofler et al., 2004; Ioos et al., 2009), to date these have not been optimized for and applied to the monitoring of colonization of the natural substrate of this pathogen.
In the present work, we have applied the above-mentioned methodologies to systematically investigate the spatial and temporal dynamics of the infection process in planta during compatible interactions between P. radiata and F. circinatum which lead to pitch canker development. q-RT-PCR was used to estimate fungal DNA as an indirect measure of fungal biomass during infection, followed by histological CLSM analysis of stem colonization of P. radiata by a GFP-labeled F. circinatum strain. The role of the constitutive resin ducts and the formation of TRDs associated with pathogen invasion were also characterized, as well as the efficacy of this induced reaction to counteract fungal pathogen challenge.
Materials and Methods
Biological materials and growth conditions
The Fusarium circinatum Nirenberg & O'Donnell strain CECT20759 has previously been described in detail (Iturritxa et al., 2011) and was selected for GFP labeling. Fungal cultures were established from colonies resulting from the subculture of a single conidium (stored in 25% glycerol v/v at −80°C) and were routinely grown at 22–28°C on potato dextrose agar (PDA; Difco Laboratories, Detroit, MI, USA) for up to 2 wk.
Commercial seeds from selected Pinus radiata D. Don stands in the Basque Country were sown in 2 l plastic pots using sphagnum peat moss as a substrate. One hundred and ninety 2-yr-old seedlings were placed in a high-security glasshouse which was specially prepared for work with P2 quarantine organisms. The pines had a mean ± SD height of 58.69 ± 5.03 cm and a mean stem diameter of 4.72 ± 0.87 mm (measured at mid-height of the plant). Throughout the trial, water was supplied as needed by irrigation and by overhead misting, and glasshouse temperature and relative humidity were maintained at 25 ± 3°C and 70–80%, respectively. Seedlings were allowed to acclimatize for 1 month before F. circinatum inoculation.
The pNMk2 vector was constructed for GFP labeling of Fusarium oxysporum (García-Sánchez et al., 2010). Briefly, the GFP cassette from the pGPDAsGFP plasmid (Fernández-Ábalos et al., 1998; Lagopodi et al.,2002) was cloned into the polylinker region of the pBHt2 binary vector (Mullins et al., 2001). The transferable DNA region of pNMk2 contains the hygromycin B resistance gene (hph) under the control of the promoter of the tryptophan C gene (trpC ) from Aspergillus nidulans for selection of fungi and the coding region of the sgfp gene (termed SGFPS65T by Chiu et al. (1996)) behind the promoter of the glyceraldehyde-3-phosphate dehydrogenase gene (gpd ) from A. nidulans. pNMk2 was introduced into the AGL-1 Agrobacterium tumefaciens strain (Lazo et al., 1991) by electroporation. These bacteria were utilized for the transformation of F. circinatum spores as previously described (Mullins et al., 2001). Transformants were selected on PDA containing hygromycin B (100 μg ml−1). Mitotic stability of the transferable DNA in monoconidial cultures of the Pgpd::sGFP strains was tested on liquid and solid cultures (Covert et al., 2001) and the strain Fc-GFP1 was chosen for further work.
Inoculation of radiata seedling stems was carried out after gently removing one leaf fascicle from the middle of the stem. A 1-μl droplet of a spore suspension of the Fc-GFP1 strain, previously adjusted to 2.5 × 105 spores ml−1, was placed at the wound site. Mock inoculation control seedlings received a 1-μl droplet of sterile distilled water (Davis et al., 2002). In order to evaluate disease incidence and the amount of exuded resin in time-course experiments, three independent inoculation assays were performed in which 20 replica plants were fungal-inoculated and 10 replica plants were mock-inoculated. Aerial symptoms were scored weekly up to 2 months on a severity scale from 0 to 4, where a score of 0 indicated a healthy plant with no necrosis; 1 indicated healthy foliage, with necrosis only at the point of inoculation; 2 indicated healthy foliage, with necrosis > 2 cm beyond the point of inoculation; 3 indicated wilting, with crown dieback; and 4 indicated dead and reddening foliage above the point of inoculation as described by Correll et al. (1991). The amount of exuded resin was also scored during the infection process as 0, no resin exudation; 1, low resin exudation, ≤ 5 mm from the wound; 2, moderate exudation, ≤ 4 cm from the wound; and 3, high exudation, > 4 cm from the wound (Bonello & Blodgett, 2003).
Five replica plants inoculated with Fc-GFP1 and two mock-inoculated plants were randomly sampled for CLSM analysis at 0, 7, 14, 21, 28, 42 and 56 d post-inoculation (dpi). The radial progression of pine stem colonization was analyzed using transverse sections sampled at the inoculation zone, whereas vertical colonization along the longitudinal axis of the pine stem was analyzed with transverse sections sampled at seven locations, each 1 cm from the next, with the inoculation point (IP) as the center.
Plant material was hand-sectioned into 0.5–1-mm-thick slices with a razor blade and fixed overnight in a 2.5% paraformaldehyde (PFA) solution, pH 7.4, at 4°C in the dark. Then slices were washed three times with phosphate-buffered saline (PBS) and maintained in the dark at 4°C, until histological examination. Fixed tissue slices were placed directly on a glass slide in a water drop with a cover slip. Confocal images were acquired on a Leica LCS-SP2 AOBS spectral confocal microscope (Leica Microsystems, Wetzlar, Germany). Excitation was provided by an argon laser (488 nm). GFP fluorescence was recorded from 500 to 520 nm, whereas endogenous plant autofluorescence was recorded from 530 to 690 nm. Confocal optical section stacks during three-dimensional scanning of sampled slices were processed using the Leica Confocal Software (Leica Microsystems). Image merging to obtain microphotographs of the whole pine cross-section was performed with Adobe Photoshop CS (Adobe Systems Inc., San Jose, CA, USA). Histological descriptions as well as estimations of xylem resin duct densities were based on observations of microphotographs. The number of xylem constitutive resin ducts and TRDs in a quarter of the stem cross-section was counted and normalized to the area of xylem tissue in each quarter. This sample area of xylem was measured with the aid of ImageJ software (Rasband, WS; ImageJ, US National Institutes of Health, Bethesda, MD, USA).
Fungal biomass quantification
Five replica infected plants were randomly selected and harvested at 0, 1, 3, 7, 14, 21, 28 and 42 dpi for q-RT-PCR analysis. Each sample consisted of a 5-mm segment of the stem tissue around the point of inoculation. This was immediately frozen in liquid nitrogen and ground to a fine powder with a mortar and pestle. Genomic DNA was extracted using a Qiagen DNeasy Plant mini kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol.
For detection of F. circinatum DNA, the specific and highly sensitive primers FCIR-F and FCIR-R (Ioos et al., 2009) were employed, as they recognize repeated target sequences in the intergenic spacer (IGS) region of the nuclear ribosomal DNA which are present in high copy numbers. The P. radiata NEEDLY (NLY ) gene (GenBank accession number U76757) was used as an endogenous gene to normalize differences in DNA template amounts. The primers NLY-F (5′-GCAGTGAGAGCGATGAAAGGAA-3′) and NLY-R (5′-TCTCCTGGTTCCGTGACAATGA-3′), which generate a 121-bp amplicon, were designed with the aid of Primer Express software 3.0 (Applied Biosystems, Foster City, CA, USA). The fungal target sequence and the plant endogenous gene were amplified in singleplex reactions, in a final volume of 20 μl. The reaction mix (15 μl per reaction) which was prepared independently was made up of 10 μl of 2X SYBR Green PCR master mix (Applied Biosystems) and 150 nM of each specific primer. This was added to each well in the reaction plate followed by 5 μl of DNA template previously adjusted to 20 ng μl−1. All PCR reactions were run in triplicate on a StepOnePlus Real-Time PCR System (Applied Biosystems), with the following cycling conditions: a first step at 95°C for 10 min to activate the AmpliTaq Gold DNA polymerase (Roche Diagnostics, Basel, Switzerland), followed by 40 amplification cycles consisting of 15 s denaturation at 95°C and 1 min annealing at 60°C.
Fusarium circinatum biomass in unknown samples was estimated, using the standard curve method, as a relative quantity of F. circinatum DNA normalized to an endogenous control of P. radiata DNA. Standard curves for F. circinatum and P. radiata were constructed based on the relationship of cycle threshold (Ct) values and known host and pathogen DNA concentrations. Genomic DNA isolated from fungal pure cultures and from a stem surface-sterilized segment of noninoculated radiata pine seedlings was used to create the standard curves. All calculations were conducted as described in the ABI PRISM 7700 Sequence Detection System User Bulletin # 2 (Applied Biosystems).
Data were analyzed using spss v.15.0 for Windows. ANOVA was used to determine the significance of differences between treatments in the amount of exuded resin and in the normalized density of TRDs. Post-ANOVA mean separation was conducted using the HSD Tukey test. The chi-square test and the Fisher exact test were used to estimate the association between the amount of exuded resin and survival of the seedlings.
Development of pitch canker visual symptoms
Pathogenicity assays performed with the Fc-GFP1 F. circinatum strain revealed that the integration and expression of the transforming construct did not alter the pathogenic and virulence abilities of the original strain (data not shown). During the first week following inoculation, almost all seedlings appeared to be healthy without any visible disease symptoms (Fig. 1a). By 14 dpi, c. 60% of seedlings displayed necrotic lesions around the IP (Fig. 1b). Subsequently, the characteristic brown to purple and depressed necrotic area extended both upward and downward from the IP (Fig. 1c) and chlorosis was observed in the more proximal needles. The first symptoms of dieback appeared by 21 dpi (Fig. 1d). All seedlings that reached this stage finally died (Fig. 1e), whereas not all seedlings that displayed initial symptoms of necrosis at the inoculation zone finally died. By the end of the pathogenicity assay (56 dpi), c. 10% of seedlings survived; in these cases, the injured tissue had formed a scar and the seedling remained healthy (not shown). The amount of resin exuded by the inoculated seedlings increased over time, whereas resin production in mock-inoculated seedlings decreased rapidly after the first week post-inoculation (see Fig. 9a). The largest significant increase in resin flow was scored in inoculated seedlings with respect to control seedlings at 21 dpi. The proportion of seedlings in each of the resin exuding ranges (low, moderate or high) scored at 21 dpi that had died or survived by the end of the experiment was analyzed with a contingency table. The chi-square test for independency ((χ2 = 10.363) > (χ20.01(3-1)(2-1) = 9.210)) and the exact method of Fisher (P =0.006 < 0.01) indicated that some association did indeed exist between resin exudation at 21 dpi and seedling survival. Indeed, all seedlings that exuded high or moderate amounts of resin at 21 dpi died by the end of the experiment. By contrast, all the surviving seedlings were found to have exuded low amounts of resin by 21 dpi.
Fungal biomass during the infection process
We next examined the growth curve of F. circinatum around the inoculation zone during colonization of pine stems by means of q-RT-PCR analysis. Three distinct phases of growth could be identified (Fig. 2). During the first week following inoculation (phase 1), pathogen biomass exhibited an exponential increase, reaching 6000-fold the level of the initial inoculum. Phase 2 (from 7 to 21 dpi) was a transition period in which fungal biomass continued to increase, but more slowly. The slope of the growth curve decreased continuously and by 21 dpi, the increase in the relative level of fungal DNA was 8-fold that at 7 dpi. Phase 3, which began at 21 dpi, was a stationary phase in which the growth rate stabilized. Thus, the increment in fungal biomass by 42 dpi was a mere 1.2-fold with respect to that at 21 dpi.
Radial progression of pine stem colonization by F. circinatum
CLSM of asymptomatic inoculated stem sections at the inoculation level revealed that, by 7 dpi, the fungus had already reached the cambial zone and showed a tendency to migrate toward the pith (Fig. 3a). The cambial zone frequently exhibited tissue breakdown between the secondary phloem and xylem. Later, in seedlings that exhibited a necrotic lesion at the IP, invasion was seen to progress tangentially through the cortex and phloem in the region of the inoculation zone (Fig. 3b). By this stage (c. 14 dpi), the F. circinatum pathogen had reached the pith of all the analyzed seedlings and, in some of them, invasion of the xylem had also commenced. A few medullary rays were invaded and most xylem resin ducts close to the invaded areas were also found to be invaded (‘XRD’ arrows in Fig. 3b). At more advanced stages of infection (e.g. 21 dpi), in which the necrotic lesion had progressed up and down along the stem of the seedlings, the pathogen was found to have extensively invaded the xylem tissue and a large increase in the number of colonized medullary rays was evident (Fig. 3c). In all of the seedlings analyzed, the proportion of the cross-section area invaded by the fungus increased considerably, although the degree of colonization varied. Xylem invasion by the pathogen followed a centripetal-like spiral course that progressively led to the massive occupation of the entire stem circumference (Fig. 3d). As invasion of inner layers of the seedling stem progressed, the presence of live F. circinatum in the cortex and the phloem appeared to be reduced (Fig. 3b vs Fig. 3d).
Histological and cellular features
During invasion of the outermost layers of the pine stem, fungal hyphae were observed in the intercellular spaces surrounding parenchyma cells of the cortex and around cortical resin ducts (arrows in Fig. 4a). The phloem collapsed in the presence of abundant fungal hyphae (Fig. 4b) and infection progressed toward inner tissues of the pine stem by fungal growth through the medullary rays and the closely associated resin ducts in the proximity of the cambial zone (asterisk in Fig. 4a). The invasion of xylem tracheids occurred by fungal penetration from adjacent ray parenchyma cells across the pits (Fig. 4c), leading to the formation of little groups of invaded axial tracheids around the medullary rays (Fig. 4d). These groups of invaded axial tracheids were initially found more frequently in the xylem tissue close to the cambial zone (arrowhead in Fig. 3b) and then progressively, in inner locations within the xylem (arrowheads in Fig. 3c), normally coinciding with zones of more lignified and densely compressed wood (“CW” in Fig. 3c). Once the pathogen had reached the pith, fungal hyphae were first observed in the intercellular spaces between neighboring thin-walled parenchyma cells (Figs 3b, 4e). Then, the parenchymatous tissue began to disintegrate, leading to the formation of big gaps in which various polyphialidic conidiophora were readily observed (Fig. 4f). Abundant sporulating hyphae orientated toward these hollow cavities were seen in slices analyzed at the inoculation level by 14 and 21 dpi (Fig. 5a–c) and some were also observed at 28 dpi (Fig. 5d).
Vertical progression of stem colonization by F. circinatum
Fusarium circinatum was found to migrate both upward and downward from the IP, with a similar strategy in all the seedlings visualized, although with some differences in the degree of colonization. Vertical invasion occurred first through the outermost cortex and phloem tissues, although colonization of the pith was also very intense proximal to the IP in some seedlings at early stages (Fig. 6). The expansion of infection through axial tracheids in the xylem occurred later than the expansion through the interconnected system of living cells, that is, the ray parenchyma and resin canal epithelium. The latter acquired high relevance in vertical colonization from the very beginning of the infection process. It is noteworthy that, in asymptomatic tissue sampled distal to the IP, fungal hyphae were frequently found only in the proximity of some xylem resin ducts (Fig. 7a).
By 14 dpi, fungal presence was observed in four of the five slices analyzed 1 cm above and below the IP. Infection had progressed 2 cm around the inoculation zone, but vertical colonization was more intense in slices above the IP, mainly in the parenchymatous cells within the xylem and the pith, than in slices below the inoculation zone. By 21 dpi, fungal presence was observed in 60% of slices visualized 2 cm above the IP, and in 40% of the equivalent lower sections. A similar proportion was observed at 28 dpi. However, by that time-point, no sections sampled 2 cm above the IP showed fungal presence in the pith. By contrast, the pathogen had colonized the pith in all the lower sections in which fungus was present. Thus, at stages later than 28 dpi, fungal presence in the pith occurred predominantly below the IP (compare Fig. 7a,b vs Fig. 7f,g). Moreover, in those slices in which the pith was pathogen-free, we typically observed pathogen in the cortex and the xylem, indicating that invasion through the pith had slowed down compared with invasion through other tissues in the stem.
Vertical colonization through the xylem also exhibited distinct patterns depending on whether invasion was upward or downward from the IP. This was particularly noticeable at advanced stages of infection in which upper slices normally showed a scattered distribution of pathogen through the tracheids and inside the lumen of several xylem resin ducts (Fig. 7b,c), whereas in lower slices it was more frequent to observe small groups of invaded axial tracheids arranged as a discontinuous tangential ring (coincident with the xylem growth ring boundary or with zones of compression wood) than fungal hyphae in the medullary rays or resin ducts (Fig. 7e,f). Vertical colonization of the pine stem by F. circinatum continued actively until the final stages of the infection process. We observed conidiogenous cells in the pith of cross-sections sampled both above and below the IP at 42 and 56 dpi (Fig. 5e,f). By 42 dpi, the fungus had already advanced 3 cm above and below the IP in 60% of analyzed sections. Two weeks later, at 56 dpi, all analyzed sections showed this feature (data not shown).
Resin duct alteration
The most dramatic anatomical change in inoculated seedlings was the morphology and quantity of xylem resin ducts following pathogen invasion. Xylem resin ducts in infected tissues had larger surrounding epithelial cells in comparison to those in uninfected stems or in uninfected areas of the stem (Fig. 8a vs Fig. 8b). The intense yellow autofluorescence (530–600 nm) of the epithelium of resin canals is probably attributable to newly synthesized resin. Note also the bright yellow autofluorescence close to the resin ducts in the reaction zone shown in Fig. 8(d) (arrows). The distribution of activated xylem resin ducts was found to be polarized in all the inoculated sections analyzed (Figs 3b, 6a–c), and these always coincided with the presence of pathogen. Fungal hyphae were frequently observed close to the epithelial cells and even in the lumen of resin ducts (Fig. 8b) during the early (Figs 3b, 4f) and final stages of infection (Fig. 7c). By contrast, the pathogen did not show the same high affinity for cortical resin ducts. Thus, we observed fungal hyphae around cortical resin ducts, but rarely inside the lumen or internal to the surrounding epithelial cells (Figs 4a,b, 8c). Occasionally, the outgrowths that had developed from hypertrophied epithelial cells ended up by forming tylosoids that filled the lumen of the resin ducts (asterisk in Fig. 7g).
Mock-inoculated control seedlings had only a few constitutive axial resin ducts by 42 dpi (Fig. 6d), whereas F. circinatum-infected stems at this stage presented a tangential row of axial TRDs in the cambial zone, closely corresponding to sites in which the pathogen was present (Fig. 7b,c). The formation of TRDs occurred mainly in the cambial zone (Figs 7b,c,g, 8e), but also in association with the closed parenchyma ray cells of the wood (Fig. 8f). Different stages in the formation of the TRDs in the cambial zone were observed as aggregations of mother xylem cells with a swollen or rounded shape (Fig. 8e). Such immature resin ducts were attacked by fungal hyphae even during their earlier stages of formation, leading to the generation of big gaps in the cambial zone (Figs 7c,d,e,g, 8g). Despite pathogen attack and destruction of TRDs, image analysis revealed that the total number of xylem resin ducts (constitutive resin ducts plus TRDs) in the inoculated seedlings increased during the course of the experiment, with this increment being statistically significant at 14 and 42 dpi. By contrast, total resin duct density in wounded controls remained stable throughout the assay (Fig. 9b).
Using advanced confocal microscopy and molecular q-RT-PCR analyses, novel features of infection have been revealed which extend our understanding of infection dynamics that occur during compatible interactions between F. circinatum and susceptible P. radiata seedlings. Fusarium circinatum exhibited a homogeneous spatial pattern in colonizing the pine stem. Initially there appeared to be two predominant pathways for fungal advance which were taken simultaneously: a radial advance toward the pith via medullary rays; and a tangential invasion of the outermost layers of the stem through the phloem and the cortex. These tendencies during the initial stages of colonization may reflect fungal foraging for host nutrients, as food reserves in the form of starch and fat are stored principally in the pith and cortex, but also in the ray parenchyma and resin canal epithelium in the wood (Sinnot, 1918). We observed what appeared to be starch grains inside these parenchymatous tissues (Fig. 6e). Our findings indicate that initial colonization of the cortex and pith by F. circinatum is through growth in intercellular spaces (Fig. 4a,e). It is likely that the fungus releases cell wall-degrading enzymes into the apoplast which are required to liberate large amounts of nutrients from these plant tissues. Indeed, Chimwamurombe et al. (2001) have sequenced and characterized a gene encoding an F. circinatum endopolygalacturonase. This enzyme is involved in the digestion of polysaccharides present in primary cell walls. Subsequent stages of pitch canker disease involve vertical colonization of the pine stem by three principal pathways: through the cortex and the phloem in the stem periphery (consistent with the external visual necrotic lesion); through the xylem via axial tracheids and resin ducts, and through the inner parenchymatous pith tissue.
We also analyzed the temporal progression of pathogenic invasion of the host. Three phases of colonization from the inoculation zone can be distinguished. The exponential growth phase, during the first week following inoculation, correlates with germination of spores and hyphal development across the parenchymatous cortex from the inoculation zone to the phloem. Rapid proliferation of the pathogen may be attributable to its initial invasion of host tissues which present more available and nutrient-rich resources, that is, the phloem. The lack of a lag phase, typically characteristic of an adaptation period before exponential growth, suggests prompt adaptation of the pathogen to the new growth milieu. Then, F. circinatum took 2 wk, from 7 to 21 dpi, to reach the stationary phase. This period was coincident with the colonization of the xylem and the pith, and also with the production of conidiophora in the pith. Stephens et al. (2008) showed a 1-wk transition phase in the growth curve of Fusarium graminearum during crown rot disease of wheat (Triticum aestivum) using a similar fungal biomass assay. In our study, the slowdown in the fungal growth during the large transition period is probably a result of the large energy investment required for the morphogenic switch from vegetative mycelium to conidia. This requires an alteration in the pattern of gene expression (Iida et al., 2006; Garzia et al., 2013) and is likely to occur intensively once the fungus appropriates the required starch reserved in the parenchyma cells of the pith.
We report here for the first time the capacity of F. circinatum to complete its asexual cycle inside the host by 14 dpi, producing conidiophora orientated toward the hollow cavities of the pith at the moment when the first pitch canker symptoms appeared. Up to now, F. circinatum sporodochia (1–3 mm in diameter) and microscopic sporodochia (0.06–0.2 mm in diameter) have been known to be produced on the surface of the dead shoots or needles of infected plants, serving as an inoculum source to spread the fungus to neighboring host plants (Blakeslee et al., 1978; Barrows-Broaddus & Dwinell, 1984). To the best of our knowledge, sporulating hyphae deep within the host have never been reported in gymnosperm plants. In angiosperm crops, the production of conidia in the inner tissues inside the host has been reported to be related to the spread of the pathogen to different parts of the same plant, as occurs with vascular wilt parasites (Di Pietro et al., 2003), or to perpetuation of the fungus over time, as a reservoir of inoculum, as has been reported for some endoparasites, such as the rice blast fungus Magnaporthe grisea (Terui, 1940). In this regard, we did not observe conidia within the vascular tissues which could be carried upward in the xylem or downward in the sap stream. Neither did we observe translocation of conidia to other tissues. However, the development of F. circinatum conidia in the cavities of the pith may lead to dispersion through these gaps and the generation of new infection foci distal to the IP within the pith. As spore dissemination is passive, spores will be subject to the force of gravity. Thus, the germination of secondary mycelium from new infecting foci may partially explain the preponderance of fungus in the pith at locations below the IP as seen at stages later than 21 dpi, and also the slowdown of fungal progression through this tissue. Indeed, the colonization of the pith in positions below the IP appears to be intrinsic, as fungus was absent in the immediate vicinity of this tissue (Fig. 7e–g). By contrast, the colonization of the pith in upper locations appears to be associated with invasion of the pathogen from the xylem (Fig. 7a–c).
Symptoms of dieback began to appear at 21 dpi and were apparent in the majority of seedlings by 28 dpi. Dieback consists mainly of chlorosis, wilting of needles and desiccation of the upper part of the seedling tip, symptoms that were consistent with those described in pine seedlings by other authors (Solel & Bruck, 1990; Correll et al., 1991; Viljoen et al., 1994). Solel & Bruck (1990) attribute wilting of young shoots to the pathogen's ability to obstruct water flow through the vascular bundle. In our case, q-RT-PCR analysis showed that fungal biomass around the inoculation zone had reached carrying capacity at 21 dpi (Fig. 2) and confocal microscopy revealed that xylem colonization at the IP was almost complete by 21–28 dpi (Fig. 3c,d). However, by this time, extensive invasion of the three-dimensional network of living cells in the xylem as well as the highest amounts of exuded resin (Fig. 9a) were both observed. These data together indicate that F. circinatum may play an active role not only in the physical obstruction of the tracheid pits, but also in the generalized collapse of the network of living cells in the xylem and in the stimulation of resin production, all of which combined may ultimately produce plant death.
CLSM analysis showed that, in general, fungal colonization was predominantly upward when dieback symptoms were still not apparent and then, after wilting of the seedlings, fungal colonization occurred predominantly in a downward direction. On this basis, we propose that upper or lower colonization along the stem probably relies on the direction of water flow and on the hydration state of the host. Thus, analysis of the colonization pattern during the consummated stage of infection represented in Fig. 7 suggests an initial upward trend due to the transpiration stream, which leads to intense colonization of upper locations close to the inoculation zone (Fig. 7c) and which may also explain the scattered distribution of the pathogen seen in xylem tissue above the IP (Fig. 7b,c). After disruption of water flow, this ascending invasion loses protagonism, in favor of a descending invasion via phloem and pith through lower positions in the stem which are still alive and have more available succulent tissues (Fig. 7e–g).
Barrows-Broaddus & Dwinell (1983) suggested that reddening and wilting of needles was the result of invasion by F. circinatum hyphae from the infected cortex in the shoot to the adjacent needle. We also observed invasion of the vascular tissue of needle traces arising from the infected pith, as well as an extensive reaction zone in the cortex adjacent to the needle fascicle (Figs 7b, 8d). This reaction zone exhibited a different fluorescence pattern which was probably a result of secreted plant substances such as phenolic compounds with different autofluorescence properties (Hutzler et al., 1998). Resin synthesized by TRDs is considered to be enriched with phenolics (Nagy et al., 2000) and we indeed observed an intense yellow fluorescence emission (530–600 nm) when blue excited (488 nm) from the secretory epithelial cells lining resin ducts (Figs 3b, 5d, 6a–c, 8b).
Despite the general effectiveness of the enhanced biosynthesis and accumulation of oleoresin as an induced chemical defense mechanism in many conifer species (Eyles et al., 2010), our results suggest that the opposite may also be the case, that is, that resin production following F. circinatum infection may be proportional to the degree of susceptibility to the pathogen. Evidence for this possibility can also be found in the data provided by Enebak & Stanosz (2003) and Kim et al. (2008, 2010), who evaluated tolerance to infection by F. circinatum in different pine species. Indeed, Barrows-Broaddus & Dwinell (1983) affirmed that resin production did not appear to be a protective advantage against F. circinatum. Our study indicates that F. circinatum tolerates resin and even stimulates its production, initially in the epithelial cells lining preformed resin ducts (Fig. 3b) and then indirectly by inducing the reprogramming of the cambial zone to produce TRDs (Fig. 7b,c,g). We visualized new TRDs as early as 14 dpi (Fig. 8e), in keeping with observations by others in Picea abies (Nagy et al., 2000; Krekling et al., 2004). Xylem resin duct density increased during pitch canker development in radiata pine seedlings (Fig. 9b) with the subsequent significant increase in resin exudation 1 wk later (Fig. 9a).
Fusarium circinatum appears to exploit to its benefit one of the classical conifer defense mechanisms, that is, enhanced formation of TRDs and associated increase in resin production. On the one hand, the pathogen uses resin ducts for vertical colonization of new zones in the pine stem. Indeed, Barrows-Broaddus & Dwinell (1984) had suggested that resin ducts may be being used as portals for the vertical spread of F. circinatum. On the other hand, high amounts of resin accumulation are thought to restrict the supply of water (Gordon, 2011), thereby contributing to plant death. Moreover, epithelial cells lining the resin ducts and surrounding parenchyma cells contain starch grains (Hudgins et al., 2005; Nagy et al., 2005) which provide the pathogen with additional food. Supporting this idea, Krekling et al. (2004) found that susceptible clones of P. abies produced more TRDs in response to Heterobasidion annosum inoculations than resistant clones; Hudgins et al. (2005) reported that the accumulated resin produced by constitutive resin ducts and by TRDs in Pinus monticola did not directly affect Cronartium ribicola development; and Luchi et al. (2005) concluded that resin flow may not be a good predictor of host resistance to Sphaeropsis sapinea and Diplodia scrobiculata canker pathogens. The fact that F. circinatum induced a defense mechanism which appeared to damage pine health by favoring fungal establishment is quite intriguing. However, others have also suggested that some alleged defense-related molecules, such as stilbenes, function more as markers of disease or are autotoxic to the plant (Bonello et al., 1993).
To conclude, the present findings extend our knowledge of the mechanisms underlying the interaction between fungal pathogens and the most successful gymnosperm plants, the conifers. Today, coniferous forests are dominant in eastern North America, Europe and Siberia, so a more precise characterization and understanding of the mechanisms that underlie disease development is very important in order to improve management systems for minimizing the damage or loss from both an ecological and economic point of view. In this sense, our results support the idea proposed by Lombardero et al. (2006) that increased resin flow in conifers following fungal inoculation should not be interpreted exclusively as an effective inducible defense mechanism against some fungal pathogens. In addition, with respect to pitch canker disease, our findings of conidiophora inside the pine stem at early stages point to the importance of preventing primary infection from this source to other plants as a consequence of incorrect management procedures. Moreover, we have also identified signs of disease before the manifestation of external symptomatology, using confocal microscopy and q-RT-PCR techniques. This provides the groundwork for the detection of compatible and incompatible plant–pathogen interactions during the first week post-inoculation, although the optimization of specific protocols for this purpose will require further studies.
This study was supported by a PhD trainee grant from the National Institute of Agricultural Research (INIA) and by the Basque Government (grant IT526-10). The authors would like to express their thanks to Dr David J. Fogarty for critical reading of the manuscript and improvement of its English. Technical and human support provided by SGIker (UPV/EHU, MICINN, GV/EJ, FEDER and FSE) and in particular by Dr Ricardo Andrade is also gratefully acknowledged.