Introduction

Approximately 7 % of the global soil surface (Mashali 1999), including 23 % of all cultivated land (Tanji 1990), is salt-affected. In Australia, 16 % of agricultural land is currently effected by salinity (Rengasamy 2006), while 67 % is highly susceptible to transient salinity, the seasonal accumulation of salts in the upper soil profile during dry periods (Rengasamy 2002). Saline soils have an electrical conductivity of the soil-water extract (EC e ) > 4 dS m−1 (Richards 1954) which correlates to an osmotic pressure of approximately 0.2 MPa (Munns and Tester 2008). A wide variety of metal salts are associated with salinised soils, with Na+ and Cl− the most common (Charman and Wooldridge 2000).

Plant salt tolerance is described using the salt tolerance threshold (STT). STT is quantified using EC e (Shaw 1999) which accounts for the effects of soil texture on the conductivity of the soil-water solution (Richards 1954). When all other factors are equal, salinities above the STT result in yield reductions (Maas 1986). Both salt tolerance and the degree of yield reduction above STT are highly variable between plant species and can range from 1 dS m−1 for salt-sensitive crops, to 10 dS m−1 for salt-tolerant crops (Maas 1986; Munns et al. 2006). The degree of reduced yield is a linear measure ranging from 2 % to 33 % yield decline per dS m−1 increase above the STT and generally has an inverse relationship with salt sensitivity, i.e. salt tolerant plants normally have a lower yield decline per dS m−1 above the STT (Maas 1986). Wheat (Triticum aestivum L.) is moderately tolerant to salinity with an STT of 6 and a yield decline of 7.1 % for each dS m−1 above the STT (Maas 1986). Salt sensitivity varies considerably over the growth cycle of wheat, with seedlings and early emergent plants more salt-sensitive than mature plants (Maas et al. 1977). Screening of wheat genotypes for salt tolerance is conducted at salinities above the STT to final salinities of between 12 and 25 dS m−1, or the equivalent mM NaCl (El-Hendawy et al. 2005a, b; Sayed 1985; Sharma et al. 1992).

Increasing levels of salt in soils effects plant growth in two main ways. The primary effect is an increase in osmotic potential, making water removal from the soil increasingly difficult (Rengasamy 2006). Secondly, increasing tissue salt concentrations can have toxic effects (Munns and Tester 2008). Plant salt tolerance is regulated by osmotic stress tolerance, ion exclusion and ion tolerance in plant tissues (Munns 2005; Munns and Tester 2008). Rising osmotic potential induces turgor loss (Munns and Tester 2008) and reduces photosynthetic efficiency via induced stomatal closure (Parida and Das 2005). The ability to exclude sodium from wheat leaves is correlated to enhanced plant salt tolerance as is the maintenance of K+ homeostasis (Tester and Davenport 2003; Zhu 2003). Ion tolerance in plant tissues is enhanced by antioxidant enzyme activity which protects the plant from salinity-induced reactive oxygen species (ROS) production (Benavides et al. 2000; Tester and Davenport 2003).

The observation of cyanobacteria–plant symbioses and their successful use as plant growth promoters (PGP) for rice cultivation (Kennedy et al. 2004; Kennedy and Islam 2001; Vaishamayan et al. 2001) has led to broader research into the use of cyanobacteria, and principally Nostoc, in cereal agronomy, including wheat cultivation. Cyanobacteria have been successfully isolated from wheat rhizospheres (Karthikeyan et al. 2009; Nain et al. 2010) and respond to hormogonia-inducing factors (HIFs) from wheat plants (Gantar et al. 1993). Cyanobacteria associate with wheat roots when exposed to light in liquid culture (Gantar et al. 1991b; Svircev et al. 1997) and penetrate through the epidermis to the root cortex (Gantar 2000; Gantar and Elhai 1999; Gantar et al. 1991a; Sood et al. 2011; Svircev et al. 1997). Diverse cyanobacteria have been assessed in laboratory and pot trial studies for plant growth promotion as well as specifically for ameliorating salt stress. This list includes Nostoc, Anabaena, Anabaenopsis, Calothrix, Fischerella, Westiellopsis, Hapalosiphon, Aulosira and Phormidium spp., as well as unclassified strains. Under laboratory and pot trial conditions cyanobacterial inoculants have colonised wheat roots, contributed to wheat N requirements (Gantar et al. 1995b; Karthikeyan et al. 2007; Nain et al. 2010; Obreht et al. 1993; Svircev et al. 1997) and produced phytostimulatory effects on root and shoot biomass (Mazhar and Hasnain 2011; Sood et al. 2011; Svircev et al. 1997). With regard to salt stress amelioration, cyanobacterial inoculation improved grain yield, straw yield and N uptake of rice, the dry weight of barley and antioxidant content of wheat grain when plants were exposed to salt (Abd El-Baky et al. 2008; Aziz and Hashem 2004; Hashem 2001; Issa et al. 1994; Rodriguez et al. 2006).

Cyanobacteria, including Nostoc, Leptolyngbya, and Microcoleus, are prolific producers of extracellular polymeric substances (EPS) (Mazor et al. 1996). EPS is correlated to cyanobacterial salt tolerance (Ozturk and Aslim 2010) and is an important factor in root colonisation studies (Gantar et al. 1995a). Cyanobacterial EPS binds soil particles (Mager and Thomas 2011; Malam Issa et al. 2007) preventing their loss from wind erosion, thereby protecting the soil nutrients bound to soil particles (Belnap and Gillette 1998). Cyanobacterial EPS applied to the soil surface was also shown to enhance aggregate stability (Maqubela et al. 2009), organic C and the lipid content (Pardo et al. 2010) of soils. In saline soils cyanobacterial inoculation lowered soil EC e (Hashem 2001; Subhashini and Kaushik 1981) and improved organic matter and N content (Aziz and Hashem 2003).

The aim of the current study was to explore the effects of cyanobacteria isolated from wheat fields on the chlorophyll fluorescence and nutrition of wheat grown under salt stress. Based upon previous studies that demonstrated salt stress amelioration for barley and rice (Aziz and Hashem 2004; Hashem 2001; Issa et al. 1994) we hypothesised that: (1) cyanobacterial inoculation would ameliorate salt stress of wheat, (2) Nostoc inoculants would have better growth outcomes for wheat than inoculants of Microcoleus or Leptolyngbya. We also proposed that chlorophyll fluorescence parameters obtained via Pulse Amplitude Modulation (PAM) fluorometry would provide valuable indicators of plant health. The information generated in this study provides important insights into the application of cyanobacterial inoculants as growth promoters of wheat and expands the role of chlorophyll fluorescence in plant nutrition studies.

Methods

Pot trial set up and growth conditions

Fine Sydney sand was oven dried at 70 ○C overnight before 700 mL was added to 125 mm pots (RICO) along with 10 cm of 0.5” low density poly-pipe. The poly-pipe was placed in the centre of the pot and was used to apply all water and nutrient solutions. Pots were autoclaved at 121 ○C for 15 min. Autoclaving was repeated after 24 h. Pots were thoroughly flushed with RO water to remove seed germination inhibitors produced by autoclaving, then 350 mL modified 1/2 strength Hoaglands solution was added to remove nutrient limitations on wheat growth. Hoaglands recipe 1 (Hoagland and Arnon 1938) was used with the following modifications: The molar equivalent of 0.5 % Ferric tartrate was added as 0.22 % Ferric citrate (Sigma) to provide a more stable Fe source; BG11 micronutrient solution (Rippka et al. 1979) was used to ensure sufficient Mo supply for N2 fixation in heterocystous cyanobacteria. Pots were left to drain to field capacity and weighed.

Cyanobacterial inoculum preparation and watering regime

Three cyanobacterial isolates originally isolated from wheat trials in agricultural research stations at Yanco, New South Wales and Kerang, Victoria were selected due to their salt physiology, strong ability to associate with wheat roots in agar media and their robust growth. Nostoc PCC73102 was selected for use as a reference strain as it has a fully sequenced genome (Meeks et al. 2001) and was originally isolated as an endosymbiont from Macrozamia in Australia (Rippka et al. 1979). Nostoc sp. ‘K1.3’ and Nostoc PCC73102 cultures were maintained in liquid BG110 (Rippka et al. 1979) while Leptolyngbya sp. ‘Y4.2’ and Microcoleus vaginatus ‘K2.2’ were maintained in liquid BG11. Triplicate pots were inoculated at each salinity with 250 μg of wet weight stationary phase culture. Nostoc cultures were applied by diluting with 1/2 strength modified Hoaglands to a total volume of 1 mL which was applied across the soil surface. Leptolyngbya and Microcoleus were shredded and applied via forceps across the soil surface. The cyanobacteria were applied in these two manners to try to mimic natural cyanobacterial crust formation as opposed to the previous methods used by Gantar et al. (1995b) and Svircev et al. (1997) which used liquid inoculum in flooded pots to induce thick crust formation. A fifth set of pots were used as controls. Controls were not supplied with any inoculum but were otherwise treated the same as other pots. Pots were completely randomised and watered to field capacity every two days for 14 days by weighing plants on a balance. Soil electrical conductivity (EC e ) was adjusted over the 14 day growing period, commencing the day after inoculation. The target salinities were 6 and 13 dS m−1 which correspond, respectively, to the salt tolerance threshold (STT) and 50 % yield reduction of wheat (Maas 1986). Hereafter the non-salinised, 6 and 13 dS m−1 salinities are referred to as low, medium and high salinity, respectively. To raise the soil EC e , volumes of 5 M NaCl were added via the irrigation pipe at watering. A separate series of pots were used to monitor the EC e of the medium and high salinity treatments. Salinity levels were determined using the soil-water extract method (Richards 1954). The experiment was conducted on three replicates in a Conviron PGR14 growth chamber with a 16-h light/8-h dark cycle at 25 ○C (± 3 ○C)/20 ○C (± 2 ○C) and 70 μmol photons m−2 s−1.

Wheat source, germination, planting and culture maintenance

Wheat (Triticum aestivum L. ‘EGA Gregory’) seeds were surface sterilised by soaking in 70 % ethanol for one min, rinsed three times with sterile milliQ water and left to germinate for two days in sterile petri dishes. Five germinated wheat seeds were transferred to each pot 14 days after cyanobacterial inoculation and planted 1 cm beneath the cyanobacterial crust. To supplement nutrient leaching prior to planting, 50 mL 1/2 strength modified Hoaglands solution was added to each pot. Pots were weighed and watered to field capacity daily. Soil EC e was monitored intermittently and adjusted by adding 5 M NaCl at watering. On day 10, leaves on the uninoculated low salt wheat plants started turning yellow, indicating nutrient deficiency. Nutrient deficiency was relieved by adding 50 mL 1/2 strength modified Hoaglands solution to all pots. The trial was terminated on day 14 for the low and medium salinity treatments. For sufficient biomass production the high salinity pots were grown until day 21.

Pulse amplitude modulation (PAM) fluorometry

Pulse Amplitude Modulation (PAM) fluorometry was conducted using a PAM2500 (Walz) just prior to harvesting plant material. Dark leaf clips were applied to the midpoint of the newest fully-emergent leaves for 30 min to allow all photocentres to open. The F o and F m values were measured in the dark with a red saturation pulse followed by 5 s far-red illumination. Red actinic light was applied at 1/2 the growth level of photosynthetically active radiation (PAR), growth PAR and 4× growth PAR, 35, 70 and 240 μmol photons m−2 s−1, respectively, for five min intervals with red light saturation pulses applied every min to calculate F t and F m ’. Each saturation pulse was followed by 5 s of far-red illumination for the calculation of F o ’. Chlorophyll fluorescence parameters follow the nomenclature and calculations of van Kooten and Snel (1990) and Schreiber et al. (1986). Values of F o , F m , F t , F m ’ and F o ’ were used to calculate the fluorescence parameters: maximum photochemical efficiency of photosystem II [F v /F m  = (F m – F o )/F m ], non-photochemical quenching [qN = 1– [(F m ’ – F o ’)/(F m – F o )]] and photochemical quenching [qP = (F m ’ − F t )/(F m ’ − F o )]. All qP and qN values are annotated relative to the proportion of growth PAR applied at measurement, i.e.: qP0.5, qP1 and qP4 and qN0.5, qN1 and qN4. Dark leaf clips were reapplied 3/4 of the way along the same wheat leaf and again left for 30 min. The F o and F m values were measured as above and red actinic light was applied at growth PAR, 70 μmol photons m−2 s−1, for five min with red light saturation pulses applied every min to calculate F t and F m ’. Each saturation pulse was followed by five s far-red illumination for the calculation of F o ’. The F o and F m values were used to calculate the maximum photochemical efficiency of photosystem II (F v /F m ) to determine that there was no lag effect from the previous treatment to the leaf photosynthetic response. The F m ’ and F o ’ values were used to calculate the effective photochemical efficiency of photosystem II (Φ II ).

Plant growth and nutrient concentrations

The number of fully emergent wheat leaves was recorded. Plants were harvested and plant material from each pot was pooled. Sand was removed from the roots by gentle washing. Roots were visually examined for the presence of root colonisation by cyanobacteria and any persistent seeds were removed before root and shoot material was separated and dried overnight at 70 ○C. Dry weights were measured before dried material was ground with a mortar and pestle. Plant material was digested using distilled 70 % nitric acid (AJAX Chemicals) for 24 h followed by hot digestion at 120 ○C for four h. Once cool, 2 mL 30 % hydrogen peroxide (AJAX Chemicals) was added and gently heated at 120 ○C for 30 min. After cooling milliQ water was added to give a total volume of 25 mL. A Perkin-Elmer Optima 7300 DV optical emission spectrometer was used with WinLab32 software to measure shoot and root concentrations of Ca, Cu, Fe, K, Mg, Mn, Na, P, S and Zn.

Statistical analysis

Spearman’s rank correlations were calculated between all fluorescence parameters and all plant nutrient concentrations (Table 1) using the cor and cor.test functions in R version 12.0.2 (R Development Core Team 2010). To determine the effects of salinity the results were pooled by salinity treatment, and the effect of salinity was assessed using one-way ANOVA. Pairwise comparisons were made using Tukey’s Honest Significant Difference (HSD) when ANOVA demonstrated significant differences. Since we were most interested in the effects of the individual inoculants rather than cyanobacterial inoculation per se, one-way ANOVA was performed between each inoculant and control. One-way ANOVA and Tukey’s HSD were performed using Minitab version 16.1.1. To clarify whether inoculation was confounding the effects of salinity on plant biomass and nutrient concentrations a two-way ANOVA was performed using R version 12.0.2 (R Development Core Team 2010). To reduce the heterogeneity of variance observed between inoculants, shoot dry weight and many shoot nutrient concentrations were log10 transformed.

Table 1 Spearman’s rank correlation coefficients between photosynthetic parameters and wheat seedling characteristics. Nutrient concentrations are in mg kg-1. Significant correlations are shown at P = 0.05 (*), 0.01 (**), 0.001 (***)
Full size table

Results

Electrical conductivity fluctuations and cyanobacterial crust formation in sand pots

During the course of seedling growth EC e values varied considerably for both the medium and high salinity treatments. In the case of the medium salinity treatment the EC e of the soil at planting was 5.17 dS m−1 and varied from 2.62 to 5.37 during the growth period with a final EC e value of 5.63 dS m−1 at completion. The EC e values for the high salinity treatment were more variable. At planting, the EC e was 15 dS m−1 and varied from 5.57 to 12.54 during the growth period with a final EC e of 18.1 dS m−1 at the conclusion of the experiment.

Cyanobacterial crusts formed in all inoculated pots, covering between 50 and 100 % of the soil surface. Before washing, some persistent crustal material from Nostoc sp.’K1.3′ and Microcoleus vaginatus’K2.2′ was observed on plant shoots at the soil surface. No root colonisation was observed on wheat seedlings growing in any of the inoculated pots.

Chlorophyll fluorescence parameters correlate well with shoot dry weight, shoot Na concentrations and all root nutrient concentrations

Significant Spearman’s rank correlations between fluorescence parameters and biomass and nutrient concentrations were limited to shoot dry weight, shoot Na concentrations and all root nutrient concentrations measured (Table 1). Of the fluorescence parameters measured only qP4 and F v /F m significantly correlated to all these variables. Overall, Φ II was the best indicator of shoot biomass and plant nutrient concentrations. Φ II was the strongest indicator of shoot biomass (rs = 0.63) and the root nutrient concentrations of Ca (rs = 0.7), Fe (rs = 0.49), Na (rs = 0.59), P (rs = 0.72) and S (rs = 0.44). Φ II was significantly correlated with shoot Na concentrations and, with the exception of Zn (rs = 0.027, P = 0.07), all root nutrient concentrations. However, Φ II had a much poorer correlation with root concentrations of Cu (rs = 0.36) and K (rs = 0.57) relative to F v /F m (rs = 0.48) and qN4 (rs =−0.72) respectively.

Regardless of cyanobacterial inoculation, altering EC e strongly effects fluorometry parameters and nutrient concentrations

The response of fluorescence parameters and nutrient concentrations to salt stress was highly variable (Table 2). Salt stress primarily effected fluorescence parameters and root nutrient concentrations though shoot dry weight and some shoot nutrient concentrations were also affected. A significant difference was found between the high salinity treatment and both the low and medium treatments with the high salinity treatment having significantly higher qN and shoot Na concentrations and lower F v /F m , Φ II , qP and shoot dry weight, root Mg and root P concentrations. A significant difference was found between medium and high salinity nutrient concentrations with higher shoot P and S and lower root Mn at high salinity, however there was no difference compared to the controls. Root Ca, K and Na concentrations were lowest in the high salinity treatment and increased from lowest to highest concentration in the order high < low < medium salinity. Both root Cu and Zn concentrations were significantly higher under the low salinity treatment than in salinised treatments. No significant difference was found between low and high salt for root S concentrations, however, medium salinity samples had significantly higher concentrations.

Table 2 Pooled ANOVA results of all pots at EC e of 0, 6 and 13 dS m−1. Mean of 15 pots ± standard error. Different letters indicate significant differences between columns (P = 0.05) using Tukey’s HSD. Dry weights are in g, nutrient concentrations are in mg kg−1
Full size table

Cyanobacterial inoculation detrimentally effects wheat nutrition but not biomass or chlorophyll fluorescence

Cyanobacterial inoculation did not significantly affect leaf number, root dry weight or shoot dry weight at any salinity (Tables 3, 4, 5, 6). Only one chlorophyll fluorescence parameter showed a significant response to cyanobacterial inoculation, and in only one instance, where wheat plants in the non-saline treatment had significantly lower qP4 values (P = 0.049) when inoculated with Leptolyngbya sp. ‘Y4.2’. Cyanobacterial inoculation had few significant effects on root nutrient concentrations (Tables 3, 4, 5) with the exception of root Fe concentrations, while salinity significantly effected all root nutrient concentrations (Table 6). Plants inoculated with Nostoc sp. ‘K1.3’ had lower root Na concentrations (8,726 ± 153 mg kg−1, P = 0.002) relative to controls (10,317 ± 177 mg kg−1) in the high salinity treatment.

Table 3 One-way ANOVA between individual isolates and the control at low salinity (0 dS m−1). Mean of 3 pots ± standard error. Significant differences between treatment and control are shown by P = 0.05 (*), 0.01 (**), 0.001 (***). Nutrient concentrations are in mg kg−1
Full size table
Table 4 One-way ANOVA between individual isolates and the control at medium salinity (6 dS m−1). Mean of 3 pots ± standard error. Significant differences between treatment and control are shown by P = 0.05 (*), 0.01 (**), 0.001 (***). Nutrient concentrations are in mg kg−1
Full size table
Table 5 One-way ANOVA between individual isolates and the control at high salinity (13 dS m-1). Mean of 3 pots ± standard error. Significant differences between treatment and control are shown by P = 0.05 (*), 0.01 (**), 0.001 (***). Nutrient concentrations are in mg kg−1
Full size table
Table 6 Two-way ANOVA results on the effects of inoculation verses salinity on plant biomass and nutrient concentrations. Unless otherwise indicated, all shoot variables were log10 transformed
Full size table

In all the other cases, plants inoculated with cyanobacteria had lower nutrient concentrations. Lower root Mg concentrations were observed in non-salinised treatments when inoculated with Nostoc sp. ‘K1.3’ (P = 0.012), Microcoleus vaginatus ‘K2.2’ (P = 0.005) and Leptolyngbya sp. ‘Y4.2’ (P = 0.003) and also at medium salinity when inoculated with Microcoleus vaginatus ‘K2.2’ (P = 0.036). Lower root Mn concentrations were observed at medium salinity when inoculated with Nostoc PCC73102 (P = 0.042). All other significant root effects occurred in the non-salinised treatments. These included lower concentrations of K (P = 0.05) when inoculated with Nostoc sp. ‘K1.3’, Ca (P = 0.042) and Zn (P = 0.027) when inoculated with Microcoleus vaginatus ‘K2.2’, and S (P = 0.047) when inoculated with Leptolyngbya sp. ‘Y4.2’.

The Oscillatorean isolates, Microcoleus vaginatus ‘K2.2’ and Leptolyngbya ‘Y4.2’, had a stronger effect on shoot nutrient concentrations than the Nostocalean inoculants (Tables 3, 4, 5). When inoculated with the Nostocalean cyanobacteria, shoot concentrations of Cu were significantly lower in the high salinity treatments. Shoot Zn concentrations were significantly lower when plants were treated with Nostoc PCC73102 at high salinity and when treated with Nostoc sp ‘K1.3’ at all salinities. Wheat plants inoculated with Nostoc sp. ‘K1.3’ also had significantly lower Mn and significantly higher Na concentrations in plant shoots at the low and medium salt treatments. When analysed in a two-way ANOVA, inoculation significantly effected most shoot nutrient concentrations (Table 6). In the case of shoot Mg, P, S and Zn salinity was also significant. An interaction was found between inoculation and salinity for shoot Ca, K and Na. The interactions resulted from differences in the effects of low, medium and high salinity depending on the inoculum used.

Inoculation effects from Oscillatorean cyanobacteria were more frequent than effects from inoculation with Nostoc. Microcoleus vaginatus ‘K2.2’ detrimentally affected most nutrient levels, with higher shoot Na and lower shoot concentrations of other nutrients relative to controls at all salinities. For the high salinity treatments all P values were below 0.1 for shoot Ca, Mn, P and S. Leptolyngbya sp. ‘Y4.2’ had fewer significant effects, especially at high salinity. Treatment with Leptolynbgya sp. ‘Y4.2’ followed the same pattern as Microcoleus vaginatus ‘K2.2’ with significantly higher shoot Na concentrations and significant reductions in other shoot nutrient concentrations. An exception to this trend was the higher shoot concentrations of K in Leptolyngbya sp. ‘Y4.2’ plants without added NaCl.

Discussion

Root colonisation absent

The successful colonisation of wheat roots by heterocystous and non-heterocystous cyanobacteria in liquid culture is well established (Ahmed et al. 2010; Gantar et al. 1991a, b, 1995a; Sood et al. 2011; Svircev et al. 1997) whereas colonisation in sand pots has been variable (Gantar et al. 1995b; Svircev et al. 1997). The effectiveness of hormogonia-inducing factors (HIFs) for both hormogonia induction and chemotaxis varies between plants (Nilsson et al. 2006; Svircev et al. 1997) as does the response by cyanobacteria (Gorelova 2006; Nilsson et al. 2005; Rasmussen et al. 1994; Svircev et al. 1997). From the available studies the differences between plants may also apply at the cultivar level, with some wheat cultivars producing more HIFs (Svircev et al. 1997). Most crustal biomass is concentrated in the top 3 mm of soil (Garcia-Pichel and Belnap 1996; Veluci et al. 2006) with cyanobacterial populations falling by approximately 90 % to approximately 0.5 × 105 cells g soil−1 at 6–9 mm below the soil surface (Veluci et al. 2006). This suggests that the movement of cyanobacteria into the rhizosphere or bulk soil is driven primarily by the flushing of cyanobacterial filaments into soil pores. Once they are in darkness cyanobacteria may then become responsive to plant chemoattractants.

Inoculation confounds the effects of salinity on wheat seedling nutrition

Most studies agree that biomass of shoots and roots and shoot nutrient concentrations of wheat decrease incrementally with increasing salinity, whereas root nutrient concentrations increase under the same conditions. This includes shoot and root dry weights, leaf number, shoot Na, Ca, Mg, K, P and S, root Na, K, Ca and S (Atak et al. 2006; El-Hendawy et al. 2005a, b; Genc et al. 2007; Hu and Schmidhalter 1997; Husain et al. 2004; Islam et al. 2007; Rascio et al. 2001; Sharma et al. 1992; Yang et al. 2007; Zheng et al. 2008). Very few of the pooled data results, that is, when data was pooled by salinity rather than inoculant or control, matched the established trends of salt effects on wheat biomass and nutrient concentrations. Instead, most nutrient concentrations fluctuated in response to increasing salinity with maximal negative effects occurring with medium salt treatment (STT of 6 dS m−1). However, shoot dry weights, root P and shoot Na concentrations only changed under high salinity. Two important characteristics of inoculation are worth noting. Firstly, inoculation with all four cyanobacterial inoculants was generally detrimental to wheat seedling health, with higher shoot Na and lower concentrations of other nutrients (Tables 3, 4, 5). Overall, inoculation rather than salinity was the main effector of shoot nutrient concentrations (Table 6). Secondly, inoculation had greater negative effects on wheat seedling nutrition at lower salinities (Tables 3, 4, 5). As a result it seems that cyanobacterial inoculation is confounding the trends that should be observed by imposing salt stress on wheat seedlings. This would also explain why correlations between chlorophyll fluorescence parameters and root nutrient concentrations were so strong (Table 1), since inoculation had fewer effects on root nutrient concentrations while salinity significantly affected them all (Table 6). In the absence of root colonisation, cyanobacterial exopolymeric substances in the cyanobacterial soil crust are the most likely mechanism that could explain how cyanobacterial inoculation is affecting wheat nutrition.

The role of exopolymeric substances (EPS)

In the absence of physical contact between cyanobacteria and wheat roots, the mechanism responsible for cyanobacterial effects on shoot nutrient concentrations is most likely cyanobacterial extracellular polymeric (exopolymeric) substances (EPS). Cyanobacterial EPS production is correlated with cyanobacterial salt tolerance (Ozturk and Aslim 2010) as well as being an effective sponge of metal ions (Blanco et al. 1998; Decho 2000) due to the presence of charged polysaccharides and proteins (Flemming 2011). Filamentous cyanobacteria are effective sorptive agents of Cu, Fe, Ni, Zn, Cd and Pb (Blanco et al. 1998, 1999; de Philippis et al. 2007; 2003; Singh et al. 1989; Yee et al. 2004). The sorption of Zn to cyanobacterial EPS is of particular concern for global nutrition since Zn-deficiency is the largest of the micronutrient deficiencies facing global agriculture (Alloway 2009). While all inoculants tested reduced shoot Zn concentrations, plants inoculated with either Nostoc sp. ‘K1.3’ or Microcoleus vaginatus ‘K2.2’ had lower shoot Zn at all salinities.

Both synergistic and competitive effects of metal ions on sorption in multimetal solutions have been reported (de Philippis et al. 2003, 2007; Pradhan and Rai 2001). In this regard, Na is less competitive for sorption sites than Ca (Khattak et al. 1989), however, increasing salinity by adding NaCl decreases the sorption of Mn (Khattak et al. 1989) and Cu (Bhaskar and Bhosle 2006), presumably by Na increasing competition for sorptive sites. The role of Na in competitive sorption in cyanobacterial EPS in the soil crust may explain why fewer detrimental effects on shoot nutrient concentrations occurred with cyanobacterial inoculation at high salinity (Tables 3, 4, 5).

In sand, cyanobacterial EPS encapsulates cyanobacteria within capsule, slime and sheath morphologies that bind sand particles (Mager and Thomas 2011; Malam Issa et al. 2007). Despite producing similar amounts of EPS in non-salinised conditions (Mazor et al. 1996), crusts of Microcoleus have stronger sand binding properties, lower water penetration and higher runoff rates than Nostoc when established on sand (Hu et al. 2002, 2003). The different nature of EPS between the Oscillatorean cyanobacteria, such as Microcoleus, and Nostoc is the most likely explanation for the differences observed between Nostocalean and Oscillatorean inoculants. Leptolyngbya sp. ‘Y4.2’ and Microcoleus vaginatus ‘K2.2’ are also expected to have greater effects since they grew faster and covered almost 100 % of the sand surface before the end of the experiment (data not shown).

The data presented here suggest that the cyanobacterial inoculants used in this study are not suitable for establishing wheat crops in saline soils. It must be highlighted that plants in previous studies that demonstrated a reduction of salt stress on rice and barley by cyanobacterial inoculants were grown until harvest (Aziz and Hashem 2004; Hashem 2001; Issa et al. 1994). As well as plant growth promotion, cyanobacterial inoculation has been shown to have a range of benefits to saline soils including lower soil EC e (Hashem 2001; Subhashini and Kaushik 1981), improved organic matter and N contents (Aziz and Hashem 2003). As a result, further exploration of cyanobacterial inoculants for use on salinised soils is justified, with particular focus on the use of cyanobacteria in stabilising and improving these degraded soils as has already been demonstrated on non-saline degraded soils (Maqubela et al. 2009; Pardo et al. 2010).

High salt stress, not inoculation, reduced the photochemical efficiency of wheat

Due to its non-invasive nature, chlorophyll fluorescence is a useful tool in assessing plant responses to environmental stress (Logan et al. 2007; Sayed 2003). Of the inoculants tested, the fluorescence parameters were only affected by inoculation with Leptolyngbya sp. ‘Y4.2’ that reduced qP4 at low salinity (Tables 3, 4, 5). When the chlorophyll fluorescence data was pooled by salt there was no difference between the low and medium salinities whereas high salinity had strong effects on chlorophyll fluorescence parameters (Table 2). The reductions in Φ II , F v /F m and qP and increases in qN measured at high salinity indicate that the shift above the STT to high salinity led to a reduction in photochemical efficiency (Φ II ) caused by a reduction in the number of open photocentres for conducting photosynthesis (qP) and a decrease in non-photochemical quenching efficiency (F v /F m , qN) as previously described by Maxwell and Johnson (2000). The observed reduction of qP4 values to below 0.6 and the reductions in F v /F m suggest reaction centre damage occurred while protective energy dissipation may or may not be occurring at the same time (Horton et al. 1996). Shoot Na was the only plant nutrient variable that followed the same pattern as the chlorophyll fluorescence parameters, suggesting that shoot Na concentrations are the primary mechanism of both reaction centre damage and the induction of protective energy dissipation.

When plant biomass and nutrient concentrations correlated with chlorophyll fluorescence parameters, they generally correlated with all chlorophyll fluorescence parameters (Table 1). It is noteworthy that, with the exception of shoot Na, cyanobacterial inoculation effects on shoot nutrient concentrations were not closely associated with chlorophyll fluorescence parameters (Tables 3, 4, 5). Significant correlations between pooled biomass and nutrient concentrations were often highest with Φ II , suggesting it is the most useful parameter for indicating the effects of salt stress on a broad range of physiological parameters for wheat. Both F v /F m and Φ II have previously been suggested as the easiest and most effective chlorophyll fluorescence parameters to measure (Maxwell and Johnson 2000; Woo et al. 2008) and both are useful indicators of crop stress (Sayed 2003). In agreement with this study, Φ II has been recommended as an appropriate plant physiology indicator since it can be performed in the field under light conditions (Maxwell and Johnson 2000). F v /F m has previously been recommended to measure drought stress in Arabidopsis thaliana because it is faster to measure (Woo et al. 2008), however it was a poor indicator of salt stress above STT in 25–50 day old rice (Lutts et al. 1996) and flag leaves of barley (Belkhodja et al. 1999).

Conclusion

This study investigated the effects of cyanobacteria on the chlorophyll fluorescence and nutrition of wheat seedlings grown under salt stress and evaluated the potential of chlorophyll fluorescence parameters to indicate plant nutrition. Rather than improving wheat salt tolerance as hypothesised, cyanobacterial inoculation produced neutral or negative effects on plant nutrition when compared to non-inoculated wheat seedlings. The hypothesis that Nostoc would have greater ameliorative effects than Leptolyngbya or Microcoleus was partially correct, in that Nostoc inoculants induced fewer negative growth effects on wheat seedlings. Furthermore, the effects of cyanobacterial inoculation on wheat shoot nutrition were so strong they confounded the effects of salinity at 6 and 13 dS m−1. EPS is the most likely cause of the observed effects of cyanobacterial inoculation on wheat seedling nutrition. The effects of cyanobacterial EPS and the absence of root colonisation would also explain the shift in focus of cyanobacterial inoculant research in recent years from plant growth promotion to using cyanobacteria to improve saline soil structure and as substrates for biosorption. Salinity rather than inoculation was principally responsible for the observed effects on the chlorophyll fluorescence of wheat, suggesting that chlorophyll fluorescence holds some promise for indicating shoot Na concentrations, although this requires further validation.