Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Author for correspondence: K. K. Newsham Tel: +44 (0) 1223 221400 Email: firstname.lastname@example.org
•Dark septate endophytes (DSE) frequently colonize roots in the natural environment, but the effects of these fungi on plants are obscure, with previous studies indicating negative, neutral or positive effects on plant performance.
•In order to reach a consensus for how DSE influence plant performance, meta-analyses were performed on data from 18 research articles, in which plants had been inoculated with DSE in sterile substrates.
•Negative effects of DSE on plant performance were not recorded. Positive effects were identified on total, shoot and root biomass, and on shoot nitrogen (N) and phosphorus contents, with increases of 26–103% in these parameters for plants inoculated with DSE, relative to uninoculated controls. Inoculation increased total, shoot and root biomass by 52–138% when plants had not been supplied with additional inorganic N, or when all, or the majority, of N was supplied in organic form. Inoculation with the DSE Phialocephala fortinii was found to increase shoot and root biomass, shoot P concentration and shoot N content by 44–116%, relative to uninoculated controls.
•The analyses here suggest that DSE enhance plant performance under controlled conditions, particularly when all, or the majority, of N is available in organic form.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Fungi-forming dematiaceous, septate hyphae are commonplace in roots in the natural environment. These fungi, usually referred to as dark septate endophytes (DSE), are often ascomycetes belonging to the order Helotiales. They have been recorded in the roots of almost 600 plant species (Jumpponen & Trappe, 1998a) and are present in a wide range of terrestrial ecosystems, but are especially common in polar and alpine habitats (Read & Haselwandter, 1981; Newsham et al., 2009). Over a decade ago, Gardes & Dahlberg (1996) stated that ‘because of the abundance and widespread distribution (of DSE) in all Arctic and alpine environments, the elucidation of their ecological and physiological role is of primary interest’, and yet little progress has been made since then in understanding the influence of DSE on their plant hosts. In these cold- and water-stressed environments, arbuscular mycorrhizal (AM) fungi, the typical mutualists of forb and grass roots at lower altitudes and latitudes, are either absent or occur only sporadically (Newsham et al., 2009), and, given that plants with which DSE are associated typically do not display disease symptoms, it has been suggested that DSE might act as surrogate mycorrhizas in these habitats (Bledsoe et al., 1990), and that the association between these fungi and their hosts is mutualistic (Haselwandter & Read, 1980). These are rather controversial views: although there is some evidence that DSE can improve the growth of plants in vitro, other evidence indicates that these fungi can also have neutral or negative effects on the growth of their hosts (Jumpponen, 2001). A consensus for how DSE influence plant growth and nutrient status has thus yet to be reached.
Meta-analysis, in which data from different experimental studies are collectively analysed using weighted statistical methods, is used to determine a mean response to a treatment across a range of studies (Rosenberg et al., 2000). Typically, a meta-analysis measures the magnitude of a response to a treatment by calculating effect sizes for each study, which are then weighted with the inverse of their sampling variances, so that the results from more precise studies are given more weight in the analysis. This form of analysis has, for example, been used to determine the effects of AM associations and ectomycorrhizas on plant performance and nutrient balance, and the responses of these mutualisms to nutrient application and CO2 treatment (Treseder, 2004; Alberton et al., 2005; Lekberg & Koide, 2005). Recently, as a corollary to a study examining the effects of elevated concentrations of CO2 on the growth of plants colonized by DSE, Alberton et al. (2010) reported a meta-analysis of data from 11 studies in the literature on the effects of DSE on shoot and root biomass. They concluded that DSE have little or no effect on shoot biomass, and that these fungi may increase root biomass (Alberton et al., 2010). However, it is apparent that not all of the data available in the literature on plant responses to DSE were included in the study reported by Alberton et al. (2010), with those from several studies having been omitted from the analysis. Furthermore, several issues associated with the data reported in five of the 11 studies used by Alberton et al. (2010) preclude their use in an accurate meta-analysis (Supporting Information Table S1). There is hence some doubt over whether the meta-analysis reported by Alberton et al. (2010) gives an accurate view of how plants respond to inoculation with DSE.
Here, a meta-analysis of data from a wider selection of studies than that used by Alberton et al. (2010) is reported. Eight responses other than shoot and root biomass are included in the analysis, including the concentrations and contents of nitrogen (N) and phosphorus (P) in plant tissues. As previous studies have shown that N applied in either inorganic or organic forms elicits different plant responses to DSE (Usuki & Narisawa, 2007; Upson et al., 2009a), whether or not plants had been supplied with inorganic N, and the form in which N had been applied to plants, were included as categorical explanatory variables in the analysis.
Materials and Methods
A total of 56 research articles on plant response to DSE inoculation, published in English language peer-reviewed research journals over the past 40 yr, were located in the literature during spring 2010. The articles were found by online database searches using the keyword phrases ‘DSE’ and ‘dark septate’, by examining the reference lists of relevant articles and by contacting researchers who had published studies investigating the effects of DSE on plant growth. Of these articles, 38, including the five referred to earlier that were used by Alberton et al. (2010), were rejected from the analysis, primarily on the grounds that they did not report values of SE or SD (Table S1), which are necessary for weighted meta-analyses. The articles which report data that were suitable for analysis, and which were hence included in the meta-analysis here, are listed in the Appendix.
For the purposes of the analysis here, DSE were defined as being chiefly members of the phylum Ascomycota that form dematiaceous, septate hyphae in the roots of living plants that do not display visible symptoms of disease. Taxonomic data, primarily based on analyses of internal transcribed spacer region sequences of ribosomal DNA, confirm that all but one of the DSE represented in the analyses could be assigned to the Ascomycota. Taxonomic information for the one remaining isolate, termed C1 by Haselwandter & Read (1982), is not available. Of the 15 different genera of DSE represented in the meta-analysis, eight are members of the order Helotiales. The remainder are members of the Chaetothyriales, Pleosporales, Capnodiales, Chaetosphaeriales and Sordariales, or cannot be assigned to orders (Appendix). Taxa represented in the analysis included Cadophora finlandica, Cryptosporiopsis rhizophila, Heteroconium chaetospira, Leptodontidium orchidicola, Phialophora graminicola and Phialocephala fortinii sensu lato (Appendix), each of which are reported by Addy et al. (2005) to be DSE. Data for fungi that form dark septate hyphae in roots but which have mutualistic roles, such as Meliniomyces vraolstadiae, Meliniomyces bicolor, Oidiodendron spp., Rhizoscyphus ericae and its anamorph Scytalidium vaccinii (Smith & Read, 2008), were excluded from the analysis.
All but two of the studies included in the analysis report that hyphae of DSE were observed either in root cells or on the surfaces of roots. Ruotsalainen & Kytöviita (2004) indicate that hyphae were not observed in roots, whilst Haselwandter & Read (1982) do not provide details of root colonization by DSE. Although the majority of DSE do not form specialized plant–fungal interfaces for nutrient transfer, three studies included in the analysis report that H. chaetospira, P. fortinii and Leptodontidium sp. form pelotons or hyphal coils in the root cells of orchids and ericaceous plant species (Vohník et al., 2003; Usuki & Narisawa, 2005; Hou & Guo, 2009). Nineteen plant species were represented in the analysis, with data included for members of the Cyperaceae, Poaceae, Asteraceae, Brassicaceae, Ericaceae, Orchidaceae, Lauraceae and Pinaceae.
Data were only included from experiments that examined the effects of a single DSE taxon on the growth of an individual plant species in a sterilized substrate in which N was available. Two experiments (Newsham, 1999; Violi et al., 2007) took place in glasshouses and the remainder were conducted in growth chambers. Data on total (shoot + root) biomass, shoot and root biomass, and root : shoot ratio were extracted from the literature, as were those for shoot N and P concentration, root N concentration, shoot N and P content and root N content (Appendix). Only two or three research articles reported height, root length, and the contents and concentrations of P in roots and of N and P in total biomass, and so these responses were deleted from the analysis. In order to meet the assumptions of meta-analysis, only one observation was included for the influence of each DSE taxon on each plant species per study (Hedges et al., 1999). As in previous studies (e.g. Treseder, 2004; Alberton et al., 2005), data from experiments in which plants had been grown under different conditions, e.g. the different N sources used by Usuki & Narisawa (2007) and Upson et al. (2009a), were considered to be independent observations.
Effect sizes were calculated using Hedges’d, rather than the more commonly used response ratio. Hedges’d was selected because it is not biased by small sample sizes (Rosenberg et al., 2000). The mean value of each parameter for DSE-inoculated (treatment, ) and uninoculated (control, ) plants was determined. Values of n and SE or SD were extracted from each article, and SE was converted to SD where necessary. The 95% CIs reported by Vohník et al. (2005) were similarly converted to SD. Hedges’d was then calculated using the following formula:
where S is the pooled SD and J is a correction factor accounting for small sample sizes (Rosenberg et al., 2000). The effect size and variance for each observation were calculated and the cumulative effect size for each response parameter was determined, along with 95% bootstrap CIs, using fixed-effects models and resampling tests generated from 4999 iterations. A cumulative effect size was considered to be significant at P < 0.05 when its 95% bootstrap CIs did not bracket zero (Rosenberg et al., 2000). In cases where a cumulative effect size was significant, the data were backtransformed in order to determine the percentage change in a parameter in response to DSE inoculation, relative to uninoculated control plants. Publication bias, i.e., the selective publication of significant over nonsignificant data, was assessed for each response parameter by rank correlating effect size against sample size (Spearman’s rho; Rosenberg et al., 2000).
Categorical analyses were also made on the data in order to determine the influence of the form in which N was supplied to plants, and of different DSE and plant taxa, on cumulative effect sizes. All studies were coded for whether or not supplementary inorganic N was made available to plants, either in nutrient solutions or in artificial growth media (Appendix). The materials and methods sections of articles were also examined in order to determine whether all, or the majority, of N had been applied in organic or inorganic forms, and the studies were coded accordingly. Those of Usuki & Narisawa (2007) and Upson et al. (2009a) supplied N either solely in organic form or in inorganic form. There was sufficient detail in another six articles to determine whether all or the majority of N had been applied to plants in organic or inorganic forms. These determinations assumed that acid-washed quartz sand (Haselwandter & Read, 1982; Violi et al., 2007) contains negligible organic N, that the total N content of wheat grain (Newsham, 1994, 1999) is 2% and that the majority of this N is in organic form (Debaeke et al., 1996), and that 70% of N in peat is in dissolved organic form (Bragazza & Limpens, 2004; Cundill et al., 2007). Based on these assumptions, the experiments reported by Haselwandter & Read (1982), Violi et al. (2007) and Alberton et al. (2010) were found to have applied the majority of N in inorganic form, and those reported by Newsham (1994, 1999) and Jumpponen et al. (1998) were found to have applied the majority of N in organic form.
Categorical analyses on the effects of different DSE taxa on effect sizes were restricted to the effects of P. fortinii sensu lato or members of the Helotiales on plant growth, as there were too few data for other taxa to include in the analyses. Analyses on the influence of different plant groups on effect sizes were similarly restricted to a comparison of dicotyledonous, monocotyledonous and gymnosperm plant species. For each categorical analysis, the model heterogeneity (QM) for the response parameter was calculated and tested against a χ2-distribution with n-1 degrees of freedom. Significant QM values indicated differences between groups (Rosenberg et al., 2000).
In order to reduce bias towards individual studies, cumulative effect sizes are only reported when data were derived from three or more research articles. Statistical analyses were made in MetaWin 2.1 (Rosenberg et al., 2000).
Effects of DSE inoculation on plant biomass
Noncategorical analyses indicated significant positive effects of DSE inoculation, relative to uninoculated controls, on total, shoot and root biomass (Fig. 1a). Analyses of backtransformed data indicated that DSE inoculation increased total, shoot and root biomass by 79%, 45% and 71%, respectively. Inoculation with DSE did not influence the root : shoot ratio (Fig. 1a). Significant negative effects of DSE inoculation were not found on biomass parameters (Fig. 1a). Publication bias was not detected for any plant biomass parameter.
Effects of DSE inoculation on plant nutrient concentration and content
Noncategorical analyses indicated no significant effects of DSE inoculation on the concentrations of N or P in shoots, or of N in roots (Fig. 1b). Root N content was similarly unaffected, but shoot N and P contents were significantly increased by DSE inoculation (Fig. 1b), with analyses of backtransformed data indicating 103% and 26% increases in these parameters, relative to uninoculated controls, respectively. Significant negative effects of DSE inoculation were not recorded on N or P concentrations or contents in any plant parts (Fig. 1b). Publication bias was detected for root N concentration (rs = −0.548; P = 0.019), but not for any other plant nutrient concentration or content parameter.
Effects of DSE inoculation in the presence and absence of supplementary inorganic N
Categorical analyses indicated that there were no significant differences between the effect sizes for the total, shoot and root biomass of plants that had, and had not, been supplied with additional inorganic N either in nutrient solutions or in artificial growth media (QM= 0.448–7.31, all P > 0.08). Inoculation with DSE did not influence total, shoot or root biomass when supplementary inorganic N had been made available to plants (Fig. 2a). By contrast, significant positive effects of DSE inoculation on each of these parameters were recorded when supplementary inorganic N had not been applied to plants (Fig. 2a), with analyses of backtransformed data indicating 92%, 52% and 70% increases in these parameters, relative to uninoculated plants, when inorganic N had not been applied.
Categorical analyses showed no significant difference between the effect sizes for shoot N concentration of plants that had, and had not, been supplied with supplementary inorganic N (P > 0.05). Shoot N concentration was unaffected by DSE inoculation, either when supplementary inorganic N had been applied to plants or not (Fig. 2b). The effect sizes for shoot P concentration and shoot N content did, however, differ between plants to which supplementary inorganic N had, and had not, been applied (QM = 26.1 and 51.7, respectively, both P < 0.01). Shoot P concentration was increased by DSE inoculation, but only when supplementary inorganic N had been applied to plants (Fig. 2b). Analyses on backtransformed data indicated a 114% increase in the shoot P concentration of DSE-inoculated plants to which supplementary inorganic N had been applied, relative to uninoculated controls. By contrast, shoot N content was increased by DSE inoculation, but only when supplementary inorganic N had not been applied to plants (Fig. 2b). Analyses of backtransformed data indicated that DSE inoculation increased shoot N content by 170%, relative to uninoculated controls, in the absence of supplementary inorganic N sources.
Effects of DSE inoculation in the presence of inorganic and organic N
Categorical analyses indicated that there were significant differences between the effect sizes for shoot and root biomass of plants to which all, or the majority, of N had been applied in inorganic or organic forms (QM= 42.0 and 33.6, respectively, P < 0.001). The effect sizes for the total biomass of plants in these two groups did not differ (P > 0.05). No effects of DSE inoculation were found on plant biomass when all, or the majority, of N had been applied to plants in inorganic form (Fig. 3). By contrast, significant positive effects of DSE inoculation were recorded on total, shoot and root biomass when N had been applied to plants either solely, or chiefly, in organic form (Fig. 3). Analyses of backtransformed data indicated that DSE inoculation increased total, shoot and root biomass by 138%, 79% and 109%, respectively, relative to uninoculated control plants, when all, or the majority, of N had been applied to plants in organic form.
Effects of specific DSE taxa on plant biomass and nutrient status
Inoculation with P. fortinii was found to have significant positive effects on several parameters. Analyses of effect sizes indicated that shoot and root biomass were significantly increased by inoculation with this DSE, as were shoot P concentration and shoot N content (Fig. 4). Analyses of backtransformed data indicated that inoculation with P. fortinii increased shoot and root biomass by 44% and 88%, and shoot P concentration and shoot N content by 116% and 56%, respectively, relative to uninoculated controls. Inoculation with members of the Helotiales was also found to increase the N content of aboveground plant parts. Shoot N content of plants inoculated with members of this order was increased by 113%, relative to uninoculated controls (n = 27; effect size = 0.862; bootstrap CIs = 0.427 and 1.415).
Effects of different plant groups on biomass
Categorical analyses indicated that there were no differences between the effect sizes for total, shoot and root biomass of monocotyledonous, dicotyledonous and gymnosperm plant species (all P > 0.05), with wide overlap of the 95% bootstrap CIs for each of these groups (Fig. 5). However, significant positive effects of DSE inoculation were recorded on the total biomass of monocotyledonous and gymnosperm species (Fig. 5), with analyses of backtransformed data indicating 161% and 23% increases in the total biomass of these species, in response to DSE inoculation, relative to uninoculated controls, respectively. Positive effects of DSE inoculation were also recorded on the shoot biomass of gymnosperm species (Fig. 5), which increased by 73% relative to uninoculated controls.
These analyses indicated that effect sizes for shoot N content were negatively associated with duration of experiment (slope = −1.75 × 10−2; r2 adjusted = 12.8%; P = 0.022). Dark septate endophytes did not have positive effects on shoot N content after 150 d. The effect sizes for root N concentration and root N content were both positively associated with the length of root colonized by DSE (slopes = 2.8 × 10−2 and 2.3 × 10−2; r2 adjusted = 26.4% and 16.8%; P = 0.017 and P = 0.050, respectively).
In accordance with a previous meta-analysis of plant responses to DSE (Alberton et al., 2010), the current study found that inoculation with these fungi increases root biomass. However, in contrast to the findings of Alberton et al. (2010), the analysis here indicates that shoot biomass also responds positively to DSE inoculation. It also indicates that inoculation with these fungi significantly increases total plant biomass and shoot N and P contents, with the effects on biomass not differing between monocotyledonous, dicotyledonous and gymnosperm plant species. Given that data from a wider range of research articles were included in the analysis here than in that of Alberton et al. (2010), and that more stringent criteria for inclusion were applied to the data used here (see Table S1), it is likely that the present analysis gives a more accurate view of how plants respond to inoculation with DSE.
Other than hyphae, the only structures that are usually formed by DSE in root cells are microsclerotia, which are thought to be resting structures or propagules (Peterson et al., 2008; Upson et al., 2009a). In contrast with, for example, the arbuscules formed by AM fungi, the hyphae and microsclerotia of DSE in root cells lack a host-derived perifungal membrane and interfacial matrix material, and hence cannot be regarded as specialized interfaces for the transfer of nutrients between plant and fungus (Peterson et al., 2008). The length of root colonized by DSE is thus unlikely to be associated with plant performance. The analyses here support this view and indicate that DSE probably do not influence plant growth through direct contact with roots, because the effect sizes for only two parameters – root N content and root N concentration – were associated with the length of root colonized by DSE structures. Although these positive associations suggest increased N transfer to roots via hyphae of DSE, the noncategorical analyses indicated no significant effects of DSE inoculation on these parameters, and the associations may hence be an artefact, possibly explained by the presence of N-rich hyphae on root surfaces.
In the absence of specialized interfaces for the transfer of nutrients between the majority of DSE and their hosts, then what mechanisms might explain the positive effects of these fungi on plant growth? Several theories have been put forward to explain such positive effects, including enhanced protection from soil pathogens, the synthesis of phytohormones or the mineralization by DSE of organic N-containing compounds in the rhizosphere (Newsham, 1999; Addy et al., 2005; Mandyam & Jumpponen, 2005). The former theory is unlikely to explain the responses reported here because all of the data included in the meta-analysis were derived from experiments that took place in sterilized substrates. Whilst it is not possible to discount fully the theory that some DSE may synthesize phytohormones, the categorical analyses reported here support the idea that these fungi may enhance plant growth by mineralizing organic N compounds in the rhizosphere. These analyses showed no effects of DSE on plant biomass when inorganic N had been applied in supplementary nutrient solutions or in artificial growth media, and positive effects when it had not. They indicate that when inorganic N is in short supply to roots, DSE may enhance plant growth, possibly by mineralizing organic compounds such as proteins, peptides and amino acids in the rhizosphere, making inorganic N more freely available to roots. By contrast, plants apparently do not benefit from DSE when roots can readily access inorganic N.
The same categorical analyses also showed that the concentration of P increased in the shoots of DSE-inoculated plants to which inorganic nutrients had been applied, suggesting that DSE also enhance the assimilation of P when it is co-applied with N in nutrient solutions or in growth media. The analyses were hence refined by coding for studies in which all, or the majority, of N had been applied in either inorganic or organic forms. These analyses indicated that DSE inoculation enhances plant biomass when all, or the majority, of N is made available to plants in organic form. Thus, although DSE might also facilitate the assimilation from soil of other nutrients, such as iron (Haselwandter, 2009), these analyses further suggest that these fungi, which are known to synthesize proteolytic enzymes (Caldwell et al., 2000), benefit plants by mineralizing organic N-containing compounds in the rhizosphere. They are thus in accordance with studies that report positive effects of DSE on plant biomass when hyphae are not observed in roots (Ruotsalainen & Kytöviita, 2004). The prevalence of organic N in cold-stressed soils (Roberts et al., 2009), caused by the slow decomposition of organic matter at low temperature, increases the likelihood that DSE will have positive effects on plant performance in ecosystems at high latitudes or altitudes, and may explain the abundance of these fungi in roots in these habitats (Read & Haselwandter, 1981; Gardes & Dahlberg, 1996; Newsham et al., 2009).
All but one of the DSE included in the meta-analysis here have been confirmed to be members of the phylum Ascomycota, and the majority of the DSE studied are members of the order Helotiales. It is widely reported that R. ericae, a member of this order and the typical mycorrhizal associate of ericaceous plant species, enhances the uptake of organic N from the acidic heathland soils that its hosts inhabit (Smith & Read, 2008). The analyses here indicate that other fungi in the Helotiales also enhance plant N uptake, with members of this order increasing shoot N content. The most widely researched DSE in the order, P. fortinii, the effects of which on plant growth are poorly defined (Addy et al., 2005), was similarly found to enhance shoot N content. This DSE also enhanced shoot and root biomass and the concentration of P in shoots, which is consistent with the suggestion that it may be able to transport P to host cells (Peterson et al., 2008). Given the widespread occurrence of members of the Helotiales, and particularly P. fortinii, in the roots of plants in alpine and polar habitats (Stoyke et al., 1992; Grünig et al., 2008; Upson et al., 2009b), it is plausible that these fungi could have a significant influence on plant biomass and nutrient balance in these regions.
Whilst the AM, ecto- and ericoid mycorrhizal symbioses are well researched, the DSE association remains relatively understudied, despite the fact that it is as common as mycorrhizas in many ecosystems, and is more frequent than mycorrhizal symbioses at high altitudes and latitudes (Gardes & Dahlberg, 1996; Newsham et al., 2009). Consequently, data from only 18 research articles could be included in the meta-analysis here. However, this compares favourably with meta-analyses examining the effects of the AM association on fungal and nematode pathogens in roots, and the effects of P fertilization and elevated CO2 on mycorrhizal symbioses (13–24 articles; Borowicz, 2001; Treseder, 2004; Alberton et al., 2005). Nevertheless, further studies on plant responses to DSE inoculation are clearly needed in the literature. These studies should report mean response values, SE or SD and n, and should also focus on how root length and plant height respond to DSE inoculation, because there are few data for these parameters, which have important consequences for plant fitness in the natural environment, that are suitable for meta-analysis in the literature.
Contrary to the results from previous studies (Wilcox & Wang, 1987; Stoyke & Currah, 1993), negative effects of DSE on plant performance were not recorded here. However, the current analysis indicates that positive effects of DSE on shoot N content are only apparent in the first 150 d after inoculation, suggesting that DSE might reduce shoot N content after several months. Further work is therefore required to determine the long-term effects of these fungi on plant performance. In addition, experiments using designs similar to those of Usuki & Narisawa (2007) and Upson et al. (2009a), in which plants inoculated with DSE are supplied with N solely in organic or inorganic forms, should also be made in order to resolve the issue of how plants respond to DSE in the presence of different forms of N. Finally, because all of the data that are available in the literature on plant responses to DSE are from experiments conducted either in glasshouses or in growth cabinets, studies should now focus on whether or not DSE exert positive effects on plant performance in the natural environment.
This work was funded by the Natural Environment Research Council. Kurt Haselwandter, Larry Peterson, Martin Vohník, Kan Narisawa, Yali Lv and Chu-Long Zhang kindly supplied unpublished materials and details of their studies. Dean Adams provided helpful advice. Three anonymous referees supplied helpful comments. All are gratefully acknowledged.
Table Appendix A1. Literature from which the data for the meta-analysis were taken
Dark septate endophyte (DSE)
Root : shoot ratio
‘x’ indicates that a parameter was included in the meta-analysis.
BAM, basal agar medium; LSM, liquid basal salts medium; MS, Murashige and Skoog medium; MMN, modified Melin-Norkrans medium; WA, water agar medium.