Introduction

The genus Euhalothece comprises unicellular, marine cyanobacteria that are morphologically diverse and ecologically versatile (Bandyopadhyay et al. 2011). This oxygenic photosynthetic group of microorganisms is highly adaptable to varying light conditions due to the presence of large antenna complexes called phycobilisomes (PBSs) in photosystem I which can absorb and transfer light of different wavelengths (500–650 nm range) to photosynthetic reaction centres (Prasanna et al. 2004). These PBSs are composed of pigmented phycobiliproteins (PBPs) and colourless linker polypeptides (Glazer 1994; Sun et al. 2009a). The PBS has a core made up of allophycocyanin (C-APC) with rods of C-phycocyanin (C-PC), and sometimes phycoerythrin (C-PE) or phycoerythrocyanin (C-PEC) (Singh et al. 2012). Each subunit of PBP contains one or more chromophores, phycourobilin (PUB), phycoviolobilin (PVB), phycoerythrobilin (PEB) and phycocyanobilin (PCB) linked to specific cysteines in apoprotein by thioether bonds (Eriksen 2008). C-phycocyanin exists as a trimer from the aggregation of three alpha and beta (αβ) monomers, with a molecular weight of ± 232 kDa (Patel et al. 2005), absorption of 620 nm and emission maxima of 640 nm (Kuddus et al. 2013).

C-PC and related PBPs have shown great commercial potential with applications in the pharmaceutical, nutraceutical, food and cosmetic industries, as well as in clinical diagnostics and biomedical research (Manirafasha et al. 2016). Although the market size for C-PC applications has not being accurately established, a market analysis indicated that in 2016, the food colourant market reached US$13 billion with approximately 6.5% growth rate per year (Kannaujiya et al. 2017). The market for the nutraceutical segment of food industry for C-PC is estimated at between US$6 billion and US$60 billion (Rodríguez-Sánchez et al. 2012; Leu et al. 2013; Johnson et al. 2014). The commercial value of C-PC is directly related to its purity ratio (A620/A280), varying from US$3 to US$25 mg−1 for food and cosmetic grade pigments, whereas highly purified reactive grade molecules can be sold for up to US$1500 mg−1. However, the successful application and production of C-PC depend on the nature of the organism, its growth characteristics, ability to scale up and, most importantly, accumulation and yield of C-PC (Eriksen 2008; Sekar and Chandramohan 2008; Kaushal et al. 2017).

The first steps in biosynthesis of PBP in cyanobacteria are the transcription and translation of amino acids, proteins and phycobilins, following the ligation of phycobilins to apoproteins with the aid of enzyme, lyases to form the C-PC (Wiethaus et al. 2010). Nutrients stimulate the production of PBPs by inducing the biosynthetic enzymes or by increasing the amount of heme and biliverdin (BV) precursors (Manirafasha et al. 2016). Nitrogen (N) is the major nutrient that is primarily used for cell growth, biomass production and protein and tetrapyrrole synthesis (Chaneva et al. 2007). Whereas, secondary nutrients, such as magnesium (Mg), and minor nutrients, which include iron (Fe), mainly serve as cofactors or prosthetic groups by forming a bridge between the biosynthetic enzymes and a substrate. There are some studies that report nutrient optimisation for C-PC production. For example, Johnson et al. (2014) found that when Nostoc sp. was grown in 1.5 g L−1 of NaNO3 and 0.8 g L−1 of Na2CO3, a maximum C-PC yield of 42 mg g−1 was achieved. Pandey and Pandey (2008) reported that C-PC production in Nostochopsis lobatus was increased up to 58.72 mg g−1 when the BG11 growth medium was supplemented with 10 mg L−1 of K2HPO4 and 3 mg L−1 of FeNH4 citrate. In another study, Geitlerinema sulphureum grown in Zarrouk’s medium supplemented with 3.5 g L−1 of NaNO3 and 6.24 g L−1 of Na2CO3 achieved 71 mg mL−1 of C-PC (Kenekar and Deodhar 2013). Very few studies have used response surface methodology (RSM) to enhance the C-PC production in cyanobacteria. Hong and Lee (2008) used a central composite design to optimise the culture media and reported that Synechocystis sp. produced 25.9 mg mL−1 when grown in 0.88 g L−1 NaNO3 and 0.32 g L−1 K2HPO4. Similarly, using RSM, Deshmukh and Puranik (2012) reported 23 mg L−1 C-PC yield when Synechocystis sp. was cultivated in 0.052 g L−1 CaCl2·2H2O and 0.10 g L−1 Na2CO3. To date, there has been limited research outlining the effect of these nutrients on major biosynthetic enzymes involved in the regulation and accumulation of C-PC in cyanobacteria. The focus of this study was to (1) enhance C-PC yield in Euhalothece sp. KZN, (2) evaluate the role of essential nutrients in C-PC production and (3) gain insights on interactions amongst the nutrients. This was performed by screening and optimisation of BG11 medium components via fractional factorial design (FFD) and RSM, respectively. To our knowledge, this is the first study that explained the role of these optimised nutrients in the C-PC biosynthesis while enhancing the C-PC yield in the Euhalothece sp. KZN.

Materials and methods

Cyanobacterium isolation

The cyanobacterium was isolated from a hypersaline environment (St. Lucia Estuary, KwaZulu-Natal, South Africa; 27.976 S, 32.362 E). The physiochemical parameters such as temperature (30 °C), pH (8.0), dissolved oxygen (9.3 mg L−1) and salinity (120 g L−1) of the water samples were measured during the time of sample collection using a multiparameter YSI 556 MPS system (Yellow Spring Systems, USA). Isolation and purification of the cyanobacterium was carried out by continuous streak plating on 1% agar BG11 medium (pH 7.0) (Allen and Stanier 1968; Rippka et al. 1979; Waterbury and Stanier 1981). Obtained single colonies were inoculated into BG11 broth medium and incubated under the following cultivation conditions: Sylvania Grolux T8 illumination of 80 μmol photons m−2 s−1, for 16-h light/8-h dark cycles at 29 ± 2 °C for 3 weeks. Wet mounts were performed using light microscopy (Nikon Eclipse 80i phase contrast, Japan) to ensure purity of the culture and observe morphological features. The isolate was identified to genus level on the basis on the proposed classification by Komárek (2016).

Partial 16S ribosomal RNA gene sequencing and analysis

Genomic DNA was extracted from exponentially grown cyanobacterial culture in BG11 medium using the FastDNA SPIN Kit for Soil (Mo Bio, USA) according to the manufacturer’s instruction. The partial 16S ribosomal ribonucleic acid (rRNA) gene amplification was done using cyanobacteria-specific primers CYA106F (5′CGG ACG GGT GAG TAA CGC GTG A3′) and CYA781R (5′GAC TAC TGG GGT ATC TAA TCC CAT T3′) according to Nübel et al. (1997) in a Verity 96 well thermal cycler (Applied Biosystems, USA). The PCR conditions optimised include the following: initial denaturation at 94 °C for 3 min, followed by 35 cycles of denaturation at 94 °C for 1 min; primer annealing at 59 °C for 1.5 min; and extension at 72 °C for 2 min, followed by a final extension at 72 °C for 5 min. The amplified products were confirmed using gel electrophoresis and thereafter sent for Sanger sequencing at a commercial lab (Inqaba Biotechnical Industries (Pty), Ltd., South Africa). The partial 16S rRNA nucleotide sequences obtained were compared to the available nucleotide sequences in the National Centre for Biotechnology Information (NCBI) GenBank database using Basic Local Alignment Search Tool (BLAST). The sequence showing > 97% similarities was selected and used to construct the phylogenetic tree (maximum likelihood) using MEGA 7 software.

Transmission electron microscopy

The thylakoid membranes and PBS arrangement in Euhalothece sp. KZN were analysed using transmission electron microscopy (TEM). Preparation for electron microscopy was done according to the method of Loza et al. (2013) with some modification. In brief, samples were transferred to 2% bacteriological agar in 0.1 M phosphate buffer (pH 7.2), following fixation in 3% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2), and incubated for 24 h at 4 °C. Subsequently, the samples were washed using 0.1 M phosphate buffer (pH 7.2) and post-fixed in 1% osmium tetroxide made up in the same buffer for 2 h at 4 °C. Samples were then dehydrated in increasing concentrations of ethanol (30–100% v/v). Thereafter, were progressively infiltrated the samples then embedded in Spurr’s resin and allowed to polymerise at 65 °C for 72 h. Ultrathin sections were cut with an ultramicrotome using glass knives and stained with uranyl acetate and lead citrate. The sections were observed using a JEOL 1010 transmission electron microscope (JEOL, USA) operated at 100 kV.

Experimental setup

The growth medium used in this study was BG11 medium (Stanier et al. 1971). The medium was prepared using filtered (100-μm mesh stainless sieve) seawater (Sigma-Aldrich) and supplemented with ± 90 g L−1 of natural sea salt (total salinity 120 g L−1). The salinity of the seawater used to make up the medium was measured using a multiparameter YSI 556 MPS system (Yellow Spring Systems, USA). Screening and optimisation experiments were performed in 250 mL Erlenmeyer flasks with 100 mL medium. The flasks were inoculated with an exponentially growing culture [10% (v/v)] under previously mentioned cultivation conditions.

Experimental designs

The components of BG11 medium were used in two statistically designed experiments in order to optimise C-PC production. Design-Expert 9.0 software (Stat-Ease, Inc., USA) was used for the experimental design, quadratic model building and analysis of the experimental data (Montogomery 2012).

Fractional factorial design

The purpose of the screening step was to identify the medium components that had a significant effect on C-PC synthesis. Six medium components, i.e. NaNO3, K2HPO4, MgSO4, CaCl2, trace metal and minor nutrients at varying concentrations, were evaluated. The selection of the low and high levels for each nutrient is important; therefore, the range of each of the nutrients selected was based on the previous literature (Burrows et al. 2008; Singh et al. 2009; Deshmukh and Puranik 2010; Moraes et al. 2013; Venkata Ramana Reddy et al. 2015). The concentration and levels of the variables utilised are listed in Table 1. C-PC (mg g−1) was selected as the response, which can be calculated using Eq. (1)

$$ y={\beta}_0+\sum \limits_i^k{\beta}_i{X}_i+\sum 0\sum \limits_{i<j}{\beta}_{ij}{x}_i\ {x}_j+\in $$
(1)

where y is the predicted response, β0 is the intercept of the plane, βi is the constant coefficient of the equation, xi and xj are coded independent factors and ϵ is the random error term. Analysis of variance (ANOVA) (type III) was conducted to test the significance of the model and selected factors as shown in Online Resource 1.

Table 1 Detailed experimental design for FF0621 with the coded and actual values of the variables tested indicating low, high and central points as well as actual, predicted and residual values for C-PC production from Euhalothece sp. KZN
Full size table

Central composite design

Central composite design is one of the most widely used RSM designs which allows one to evaluate the relationship between independent factors and to predict the response in an effective experimental design (Myers et al. 2009; Montogomery 2012). Based on the results from the FFD analysis, MgSO4, NaNO3 and minor nutrients were found to have a significant effect on C-PC yield and were subsequently selected to determine their individual and combined effects on C-PC yield. A 20-run CCD was used for three independent factors with five levels (− α, − 1, 0, + 1, + α), six replicates of the central points and six axial points, leading to a total of 20 sets of experiments. The design matrix indicating the concentration of the three media components with respect to their values in actual and coded form, as well as the actual and predicted values of the C-PC yield, is outlined in Table 2. The results from the previous experiment were used to establish the minimum and maximum ranges. The optimum levels of the selected factors were obtained by solving the regression equation and also by analysing the response surface contour plots. The responses were optimised by the second-order model shown in Eq. (2)

$$ y={\beta}_0+\sum \limits_{i=1}^k{\beta}_i{X}_i+\sum \limits_{i=1}^k{\beta}_{ii}{x_i}^2+\sum 0\sum \limits_{i<j}{\beta}_{ij}{x}_i\ {x}_j+\in $$
(2)

where y is the predicted response; β0, βi, βii and βij were the intercept, the linear coefficient, the squared coefficient and the interaction coefficients of the model, respectively; k is the number of factors; and xi and xj are the level of independent factors in the study. In order to confirm the reliability of the developed RSM model, the predicted results were compared with the experimental values (actual C-PC yield).

Table 2 Central composite design was used for three independent factors with five levels (− α, − 1, 0, + 1, + α) and six replicates of the central points and six axial points, leading to a total of 20 sets of experiments
Full size table

C-phycocyanin analysis

Cyanobacterium cells from each flask were harvested by centrifugation at 8000×g for 10 min at 4 °C. The pellet was washed with 10 mM sodium phosphate buffer (pH 7.0) and subjected to repeated freeze-thaw cycles of − 20 at 4 °C. The buffer was prepared using ultrapure water (18.2 MΩ-cm) supplemented with sodium azide 0.02% (w/v). The cell debris was removed by centrifugation at 17,000×g for 30 min, and the supernatant layer containing C-PC was collected for further analysis. The concentration of the total crude C-PC was determined at 620 and 650 nm according to Bennett and Bogorad (1973). The C-PC concentration in the sample was calculated using Eq. (3)

$$ \mathrm{C}\hbox{-} \mathrm{PC}\ \left(\mathrm{mg}\kern0.5em {\mathrm{mL}}^{-1}\right)=\frac{{\mathrm{OD}}_{620\kern0.5em \mathrm{nm}}-0.7\times {\mathrm{OD}}_{650\kern0.5em \mathrm{nm}}}{7.38} $$
(3)

The total C-PC (mg g−1) extracted from the biomass was determined using Eq. (4)

$$ \mathrm{C}\hbox{-} \mathrm{PC}\ \left(\mathrm{mg}\kern0.5em {\mathrm{mg}}^{-1}\right)=\frac{\mathrm{C}\hbox{-} \mathrm{PC}\ \left(\mathrm{mg}\kern0.5em {\mathrm{mL}}^{-1}\right)\times \mathrm{Volume}\ \left(\mathrm{mL}\right)}{\mathrm{Biomass}\ \left(\mathrm{g}\right)} $$
(4)

Results

Cyanobacterium identification

The cyanobacterium cells were unicellular and blue-green in colour with a wide oval-to-cylindrical shape approximately 8–10 μm in size (Fig. 1a). Cell division was perpendicular to the longer axis of oval cells. The actively diving cells appeared to be smooth, whereas the older cells had irregular net-like granule appearance and were buoyant. On the basis of morphological features observed using the light microscope, the cyanobacterium belonged to the order Chroococcales (subsection 1). The molecular identity of the isolate based on 16S rRNA gene sequencing and BLAST analysis confirmed that the isolate falls within a clade of halophilic cyanobacteria and was found to be closely associated with form genus Euhalothece strains, a group that receives strong support (RLS = 100%) (Fig. 1b). Partial 16S rRNA gene sequence of the cyanobacterium was deposited in the NCBI database (accession number KF017202).

Fig. 1
figure 1

a Micrograph showing blue-green Euhalothece sp. KZN cells. b Reconstructed phylogenetic tree of Euhalothece sp. KZN and related cyanobacteria that were aligned based on 16S rRNA sequences. The Kimura two-parameter model along with 1000 bootstrap replications was selected to construct the maximum likelihood tree. Bootstrap values (distance) of ≥ 50% are indicated on the branches. The bar shows genetic distance. All sequences were extracted from NCBI database (www.ncbi.nlm.nih.gov). Multiple sequence alignments were performed using the BioEdit sequence alignment editor. Sequence alignment as well as tree building were performed using MEGA7

Full size image

Phycobiliproteins and phycobilisomes in Euhalothece sp. KZN

The blue-green colour of the cells is due to the major PBP, i.e. C-PC, occurring in the PBS rods. The blue colour and peak of the absorption spectrum at 620 nm confirm the production of C-PC (Fig. 2). In the case of C-PC, denaturation leads to a loss of absorbance at 620 nm (first excited state) and an increase in absorbance at 360 nm (second excited state). The spectra of the extract did not reveal the characteristic PE peak in the range of 565 nm. In Fig. 3, the PBSs are the dense structures lying on the thylakoid membrane of the Euhalothece sp. KZN.

Fig. 2
figure 2

UV-vis absorbance spectra of crude PBP extract. Inset vividly coloured blue PBPs

Full size image
Fig. 3
figure 3

Transmission electron micrographs of Euhalothece sp. KZN cell showing parallel thylakoid membranes (th), and dense structures representing PBS, a schematic drawing of PBS, are also shown. Scales bars are 50 nm

Full size image

Effect of nutrients on C-PC production using fractional factorial design

The initial aim was to determine the effect of individual BG11 nutrient concentrations on C-PC production in Euhalothece sp. KZN. Pareto chart illustrates the magnitude of significance of each nutrient affecting C-PC production (Fig. 4). This was estimated as the difference between averages of measurements made at the high level (+ 1) and the low level (− 1) of that factor. Based on the Pareto chart and small p value, minor nutrients (citric acid, ferric ammonium citrate, EDTA and Na2CO3), NaNO3 and MgSO4 were identified to have positive impact factors on C-PC production.

Fig. 4
figure 4

Pareto chart indicating the magnitude of significance of media components on C-PC production. The individual components minor nutrients (X6), and NaNO (X1) and MgSO (X3) showed significant positive effect, while the interaction between NaNO3 (X1) and trace metals (X5) showed negative effect towards CPC production. The other nutrients, K HPO (X) and CaCl (X) had a positive but insignificant effect. Solid black bars indicate positive effects, and white bars indicate negative effects. Bars above the Bonferroni limit (3.8) are certainly significant while those above the t-value limit (2.2) are possibly significant

Full size image

The yield of C-PC may be best predicted using Eq. (4)

$$ {\displaystyle \begin{array}{l}Y\ \left(\mathrm{mg}\kern0.5em {\mathrm{g}}^{-1}\right)=2.57+0.50\times {X}_1+0.06\times {X}_2+0.25\times {X}_3+0.06\times {X}_4+-0.06\times {X}_5+\\ {}0.52\times {X}_6+0.09\times {X}_1{X}_2+-0.1\times {X}_1{X}_5+0.08\times {X}_1{X}_6+-0.07\times {X}_2{X}_6\end{array}} $$
(4)

Minor nutrients (citric acid, ferric ammonium citrate, EDTA and Na2CO3) showed the highest positive effect on C-PC yield with an F value of 383,630 and a significant p value < 0.0001. The positive effect for minor nutrients suggests that the medium requires a higher level of minor nutrients than optimum to enhance C-PC production. Thus, the minor nutrient solution was selected as the first essential medium component to be optimised. N also showed a positive effect and contributed 19.13% to the total effects of the F and p values of 365.95 and < 0.0001, respectively, showing the significance of N in C-PC production. The significance of MgSO4 in C-PC production is evident from the low p value of 0.0101. Potassium dihydrogen phosphate and CaCl2 were found have a positive but insignificant effect towards C-PC production in terms of the percentage contribution of 1.74 and 2.67, respectively. Increasing the concentration of the inorganic salt KH2PO4 did not show any effect on C-PC synthesis in this study. This indicated that 0.06 g L−1 in the medium is sufficient for C-PC synthesis in Euhalothece sp. KZN. On the contrary, trace metals and the interaction between trace metals and NaNO3 showed negative effects of − 0.12 and − 0.28. Trace metals showed a 2.55% contribution, and the interaction between trace metals and NaNO3 showed 2.88% contribution with an F value of 7.99. This demonstrates that − 1 level (i.e. at low concentration) of trace metals in the medium favoured C-PC production in this cyanobacterium. The above results indicate that the trace metals did not play an active role in the C-PC production.

Optimisation and role of nutrients in C-PC production using response surface methodology

Response surface methodology was used to determine the optimal level of the nutrients as well as to study the interactions amongst the selected media components: NaNO3, MgSO4 and minor nutrients (citric acid, ferric ammonium citrate, EDTA and Na2CO3) for enhancing C-PC synthesis in Euhalothece sp. KZN. Equation (5) describes the mathematical model relating to the production of C-PC with independent variables and the second-order polynomial through multiple regression analysis using Design-Expert 9.0. The equation provides an estimate of the level of C-PC (Y) as a function of NaNO3 (X1), MgSO4 (X3) and minor nutrients (X6). The response equation from the above set of experiments can be written as

$$ {\displaystyle \begin{array}{l}Y\ \left(\mathrm{mg}\kern0.5em {\mathrm{g}}^{-1}\right)=42.18+5.62\times {X}_3--0.38\times {X}_1-1.14\times {X}_6+4.19\times {X}_3{X}_1-2.62\times \\ {}{X}_3{X}_6-3.37\times {X}_6-6.60\times {X}_3^2-4.55\times {X}_1^2-6.24\times {X}_6^2\end{array}} $$
(5)

An ANOVA of the quadratic regression model (Eq. (5)) showed that the model significantly fitted the data, as was evident from the low p value (< 0.0001) of Fisher’s test. The results showed that the linear effect of MgSO4 (X3) and minor nutrients (X6) as well as the interactive effects of NaNO3 and MgSO4 (X1X3), NaNO3 and minor nutrients (X1X6) and MgSO4 and minor nutrients (X3X6) were also significant model terms. The fit of the model variables was determined by the coefficient of determination (R2). Based on the ANOVA results (Online Resource 2), the model reported a high R2 value of 0.9847, indicating that 98.47% of the variability in the responses could be explained by the model. An adjusted R2 value of 0.9709 was also found, and since they are close to 1.0, this indicates a high correlation between the experimental values and the predicted values. The lack of fit was found to be insignificant (p value = 0.0642) which indicated that the developed model data fitted the variables. Based on above, the actual relationship between the response and the significant variables could be confidently described by the developed model.

The optimum levels of the selected medium components were obtained by analysing the response surface plots. The contour plots described by the regression model were drawn to illustrate the individual and interactive effects of MgSO4, NO3 and minor nutrient stock (citric acid, ferric ammonium citrate, EDTA and Na2CO3) on C-PC production by Euhalothece sp. KZN. The response surface plot (Fig. 5a) clearly shows the interaction between NO3 and minor nutrients. Increasing NaNO3 (1 to 1.8 g L−1) and minor nutrient (7.5 to 10 mL L−1) concentrations in the media led to an overall increase in C-PC yield (42.65 mg g−1). However, a further increase in NO3 concentration to 2 g L−1 led to a decrease in C-PC yield (27.01 mg g−1). The minor nutrients which composed of citric acid, ferric ammonium citrate, EDTA disodium salt and Na2CO3 were found to be crucial for stimulating C-PC synthesis.

Fig. 5
figure 5

Response surface contour plots for C-PC production by Euhalothece sp. KZN, showing (a) minor nutrients and MgSO while NaNO was kept at a constant level (1.5 g L-1), (b) minor nutrients and NaNO while MgSO was kept at 0.08g L-1, and (c) MgSO and NaNO while minor nutrients were kept at a constant level (11.25 ml L-1)

Full size image

Similarly, the C-PC concentration increased proportionally with an increase in MgSO4 and minor nutrients in the medium (Fig. 5b). Increasing MgSO4 concentration from 0.04 to 0.11 g L−1 and that of minor nutrients from 7.5 to 10 mL L−1 while keeping NaNO3 constant at 1.5 g L−1 greatly increased the C-PC yield from 3.82 to 5.62 mg g−1. It is important to note that when the MgSO4 concentration was increased beyond α level (0.15 g L−1), the C-PC yield decreased. The predicted optimum concentrations of minor nutrients and MgSO4 were observed near 10 mL L−1 and 0.11 g L−1, respectively. When both nutrient factors were kept at low levels, the C-PC yield was low.

Experimental validation

A confirmatory experiment at different levels of the factors, as suggested by the numerical modelling to support data under optimised conditions, was performed. This experiment was conducted to maximise C-PC content while keeping other responses within the range. On the basis of numerical optimisation, the quadratic model predicted that the maximum C-PC production was a mean of 95% CI of 43.97 ± 1.76 mg g−1 when the optimal values of test factors were 0.10 g L−1 of MgSO4, 1.67 g L−1 of NaNO3 and 10 mL L−1 of minor nutrients. The predicted optimum concentration of nutrients was validated at the suggested levels in triplicate. The maximum C-PC obtained under optimum conditions was 45 mg g−1, which was within the 95% CI of the prediction value (Fig. 6). This indicated that the model was precise when predicting the optimal nutrient concentrations for improved C-PC production. Moreover, the use of RSM significantly enhanced the C-PC yield by > 280% in comparison to the un-optimised BG11 medium. Therefore, the applied model was considered to be reliable for predicting the production of C-PC from Euhalothece sp. KZN.

Fig. 6
figure 6

C-phycocyanin yield obtained in the original and optimised BG11 medium from Euhalothece sp. KZN. Error bars represent means ± SEM. n = 3

Full size image

Discussion

Cyanobacterium identification

Bergey’s manual of systematic bacteriology classifies the cyanobacteria in ‘form genera’ which, in turn, are divided into clusters or subclusters, but not an ‘official classification’ (Ramos et al. 2017). Waterbury and Rippka (1989) proposed that the coccoid to rod-shaped cyanobacteria with cells larger than 3 μm in diameter dividing by binary fission in one plane perpendicular to the long axis and lacking a sheath were placed in the provisional Cyanothece group. Garcia-Pichel et al. (1998) described the common characteristics of strains belonging to the Halothece cluster of the form genus Cyanothece are extremely halotolerant with optimal salinities between 5 and 20%, moderate thermophiles, i.e. growth at 45 °C, and only contain C-PC pigment. Marine and freshwater unicellular Cyanothece that cannot grow at high NaCl concentration do not cluster within Halothece. The results from our analysis (cell shape, size, pigment composition and salt tolerance) indicate the isolated strain-resembled members of the order Chroococcales, genus Cyanothece, cluster 3 Halothece/Euhalothece. Traditionally, cyanobacteria are identified on the basis of their morphological characteristics; however, recently, molecular characterisation is widely used to confirm the identity of cyanobacterial species. The isolated cyanobacterium is closely related cyanobacteria that are associated with extreme conditions. For example, the isolate shared identities (> 96%) to Euhalothece sp. Z-M001 isolated from Soda Lake, Magadi (Mikhodiuk et al. 2008); Dactylococcopsis salina PCC 8305 isolated from a Solar Lake in Sinai, Egypt (Walsby et al. 1983); Euhalothece sp. from the United Arab Emirates (Mckay et al. 2016); and Cyanothece sp. 104 isolated from seawater off China’s coast (Zhang et al. 2007). Garcia-Pichel et al. (1998) reported Halothece PCC 7418 isolated from Solar Lake, Sinai, Egypt, along with 11 other cyanobacterial isolates from solar evaporation and hypersaline ponds. The analysis showed that amongst all the matches for the 16S rRNA gene sequences from GenBank, the isolate showed 100% similarity to Euhalothece and Halothece genera.

The genus Halothece was first created by Garcia-Pichel et al. (1998) following phylogenetic, 16S rRNA gene-based analysis of 13 extremely halotolerant unicellular cyanobacterial strains previously isolated from various hypersaline environments. Two distinct subclusters were described: (1) ‘Euhalothece’, comprising 12 closely related strains of similar morphology and were originally identified as members of three traditional genera (Cyanothece, Aphanothece and ‘Dactylococcopsis’ (Myxobakton)), and (2) ‘true Halothece’, consisting only a single strain, MPI 96P605. Similarly, using 16S rRNA sequence analysis, Margheri et al. (1999) confirmed 11 new halotolerant strains that formed a monophyletic cluster of previously proposed ‘Halothece’ genus. Although the strains belonging to this cluster are supported by a molecular analysis, to date, the form genus Halothece has no standing under either the International Code of Nomenclature (ICN) for algae, fungi and plants or the International Code of Nomenclature of Prokaryotes (ICNP) as it has not been validly described (Komárek et al. 2014). Even though Halothece is not a validated genus, to date, there are approximately 45 isolates that have been assigned to Halothece, subcluster Euhalothece.

Phycobiliproteins and phyobilisomes in Euhalothece sp. KZN

C-phycocyanin is a water-soluble pigment-bound protein used to capture light energy for photosynthesis that constitutes up to 15% of its dry cell weight when harvested (Hazra and Saha Kesh 2017). This pigment is unique to cyanobacteria and is responsible for the distinctive deep blue colour. The blue colour and prominent peak at 620 nm of the absorption spectrum indicated the presence of mainly C-PC (Fig. 2). Due to low concentrations of C-APC in the cell, its major absorption band (650 nm) is masked by those of C-PC. Transmission electron microscopy was done to examine the thylakoid membrane and PBS. The cytoplasmic surface of the thylakoid membranes is separated by the PBS. Phycobilisomes are hemi-discoidal pigment-bound protein complexes that are attached to the cytoplasmic side of the thylakoid membrane (Chang et al. 2015). The PBSs are composed of a core and clusters of rods made up of short stacks of discs (Li et al. 2004). It is known that PBSs naturally have a rectangular shape and stand upright on the thylakoid membrane in the cell (Yi et al. 2005). The appearance of the PBS in thin-section electron micrographs is dependent upon the plane of sectioning of the samples (Cosner 1978). When the rows of PBS are cut in cross section in a plane perpendicular to the thylakoid surface, the PBSs display an irregular semi-circular outline with the flat side attached to the thylakoid membrane and, therefore, appear in face view (Bryant et al. 1979). The PBSs are seen in edge view standing on the thylakoid surface, when the rows of PBS are either cut in longitudinal section in a plane normal to the thylakoid surface or in tangential section, i.e. when the plane of sectioning is parallel to the thylakoid surface (Bryant et al. 1979). The PBSs appear as electron-dense cords running at an angle to the long axis of the cells. As seen in transverse sections (Fig. 6), PBS appears to be dense structures that are anchored on the thylakoid membrane. Similarly, Porta et al. (2000) found that in Cyanothece PCC 8303, the alternative distribution of the PBS between two thylakoids gave the PBS a zipper-like appearance. Distortion of PBS shapes within cells could be caused by close contact with the abundant α-granules which compete for the same stromal space (Gray et al. 1973).

Screening and optimisation of nutrients for C-PC production in Euhalothece sp. KZN

The selection of the optimal combination of nutrients is essential to increase C-PC production in cyanobacteria. Statistical design strategies for screening and optimisation allow for quick identification of essential nutrients and interactions between them. Amongst the BG11 nutrients screened, three components, viz. minor nutrients (i.e. citric acid, ferric ammonium citrate-EDTA, Na2CO3), NaNO3 and MgSO4, showed a significant effect (p > 0.005) on C-PC yield (Fig. 4). Based on the results of the CCD, 10 mL L−1 of minor nutrients, 1.67 g L−1 of NaNO3 and 0.10 g L−1 of MgSO4 were the optimum concentrations required for C-PC production. The maximum C-PC yield was demonstrated by a confirmatory experiment using the optimised medium that the predicted values (43.7 mg g−1) agreed well with the experimental values (45 mg g−1) (Fig. 1). The interaction effects of minor nutrients and NaNO3 on C-PC production are shown in Fig. 5a. As seen, increasing the minor nutrient and MgSO4 concentrations up to the optimum point increased the C-PC yield to a maximum level. However, the C-PC yield decreased following a further increase in the nutrient levels. Similar observations were also noticed for minor nutrients and MgSO4 (Fig. 5b) and for MgSO4 and NaNO3 (Fig. 5c) on C-PC production by Euhalothece sp. KZN. The C-PC pathway is shown in Online Resource 3 (Frankenberg and Lagarias 2003; Wiethaus et al. 2010; Alvey et al. 2011; Aoki et al. 2011; Czarnecki and Grimm 2012).

Although, in this study, minor nutrients are not supplied as separate components, ferric ammonium citrate-EDTA and Na2CO3 serve as the sources of the essential minerals required for C-PC synthesis. The key enzymes, ferrochelatase and phycocyanobilin/ferredoxin oxidoreductase (PcyA), are regulated by Fe. Ferrochelatase (which contains two Fe-S clusters) is responsible for the insertion of Fe2+ into PIX to form heme in step 5 (Alvey et al. 2011) (Online Resource 3). Thus, the increase in ferrochelatase activity is associated with accumulation of heme. The initial step in the PCB biosynthesis is catalysed by heme oxygenase (HO) where heme serves as both the substrate and cofactor for its degradation (Montellano 2000). Spiller et al. (1982) reported that the level of HO enzyme is a rate-controlling factor in PCB synthesis. Biliverdin IXα produced from heme is further reduced by ferredoxin-dependent PcyA (step 7). Thus, the requirement of high concentration of Fe ions is attributed to its inevitable role as a redox cofactor for PcyA. Moreover, Chakdar and Pabbi (2016) reported that Fe regulates the expression of the PC αβ subunit gene. It is well established that Fe deficiency results in oxidative stress; thus, a significant increase in reactive oxygen species triggers the synthesis antioxidants, which include PCB to act as free radical scavengers (Latifi et al. 2005). Tu et al. (2004) observed that Fe limitation (or removal of chelators) results in the accumulation of chlorophyll precursor, protoporphyrin IX monomethyl ester (Mg-ProtoMe). Even though Fe is not a constructional element in PCBs, it is significantly important, since the PCB synthesis has Fe-dependent steps, and Fe serves as a substrate and cofactor for these essential enzymes (Tooley et al. 2001; Wang et al. 2010).

Cyanobacteria also need C compounds in the form of keto acids (oxaloacetate and 2-oxoglutarate) to produce the macromolecules such as amino acids and apoproteins. A study by Stöckel et al. (2013) showed that the glutamate-utilising enzyme, i.e. GluRS that converts glutamate to glu-tRNA (step 2), revealed higher transcript abundance under elevated C conditions. Furthermore, the reduction in the inorganic carbon source (Na2CO3) promotes the arrangement of the PBP complex. Previously, a suitable concentration of Na2CO3 was found helpful in enhancing the PBP synthesis (Johnson et al. 2014). Deshmukh and Puranik (2012) reported that CaCl2 and Na2CO3 were the most influential nutrients for C-PC in Synechocystis sp., producing 23 mg L−1 of C-PC when cultivated in 0.058 and 0.115 g L−1 of CaCl2 and Na2CO3, respectively. Sun et al. (2009b) reported that C and Fe limitation resulted in partial loss of pigment since a lack of these essential compounds produces a protein factor (NblC) that initiates the degradation of C-PC.

Generally, the C-PC yield increases with nitrate concentration up to a critical level; however, at a very high nitrate concentration (2 g L−1), C-PC yield decreases. Therefore, in the present study, the optimal nitrate concentration (1.67 g L−1) predicted was justified. Wakte et al. (2011) reported that nutrients such as NO3 and Mg are required for growth, metabolism as well as the structural and functional components of nucleic acids, proteins and cell wall. Nitrogen accounts for approximately 10% of dry weight, and it is very essential for synthesis of the amino acids which constitute proteins and other cellular compounds including the C-PC subunits of the cyanobacteria (Herrero et al. 1990; Flores and Herrero 1994). Nitrate is actively taken up by cyanobacterial cells and is sequentially reduced by nitrate reductase to nitrite and then to ammonium by nitrite reductase (Liotenberg et al. 1996). Nitrogen assimilation operates through the glutamine synthetase (GS)/glutamate synthase (GOGAT) pathway, which allows incorporation of N to glutamate and glutamine (step 1). Glutamate is a vital precursor for aminolevulinic acid (ALA), the immediate precursor for PCB synthesis (Online Resource 3). Tandeau de Marsac and Houmard (1993) reported that PBP synthesis depends on the supply of assimilable N. The chemical structure of phycobilin consists of four pyrrole rings, where the N atoms form a strong bond between the Fe+ ions. At a low nitrate concentration (1 g L−1), C-PC biosynthesis is also low, since an insufficient amount of assimilable N is present in the medium. Lack of N represses the synthesis of new C-PC by suppressing the de novo synthesis of the αβ apoprotein subunits (Xie et al. 2015). It is known that PBPs play a role as N reservoirs; thus, during N limitations, cyanobacteria degrade the PBS and linker proteins to release essential amino acids to synthesise other proteins required to support the metabolic activity of the cells (Richaud et al. 2001). Researchers also have reported that during N and S limitation, the PBP content and the level of C-PC mRNA decrease (Singh et al. 2003; Eisenhut et al. 2007; Ludwig and Bryant 2012). It has also been previously reported that a higher level of nitrate results in an increase in the amount of Fe absorbed by cells (Schwarz and Forchhammer 2005; Leu et al. 2013; Esen and Ozturk Urek 2015). However, a further increase in NO3 concentration to 2 g L−1 led to a decrease in C-PC yield. Comparably, Johnson et al. (2014) reported the synthesis of PBP decreased at a very high concentration of NaNO3 due to an inhibition of PBP biosynthetic enzymes in the presence of excess N.

Similarly, Mg also plays a crucial role in C-PC synthesis, and it is primarily used as a cofactor for porphobilinogen (PBG) synthase enzyme. The formation of ð-ALA is the first committed and rate-limiting step in the synthesis of phycobilins requiring Mg2+. In cyanobacteria, ALA is normally synthesised from glutamate via the C5 pathway seen in step 3 (Online Resource 3) (Shoolingin-Jordan et al. 2002; Chow 2012; Esen and Ozturk Urek 2015). The PBG synthase enzyme catalyses the condensation of two molecules of ALA with the addition of Mg2+ to form the pyrrole ring of PBG (Frankenberg and Lagarias 2003). However, as described previously, excess Mg will be incorporated into the protoporphyrin molecule, to synthesise chlorophyll (Beale 1999), thus decreasing C-PC. Metal ions are often found in the active site of PBGS and are commonly required for enzymatic activity. Jaffee (2003) reported that cyanobacteria, Synechococcus sp. and Synechocystis sp. which are distantly related to Euhalothece sp., have aspartate-rich and allosteric Mg metal binding region, respectively, thus requiring Mg (Jaffee 2003). An increase in Mg concentration up to 0.11 g L−1 resulted in an increase in C-PC yield. This finding is in agreement with Wakte et al. (2011) where they have reported the essentiality and significant influence of MgSO4 on C-PC production in Spirulina sp.

Conclusions

This study successfully identified and optimised the most critical nutrients required for C-PC production of Euhalothece sp. KZN using a strategy of selection and modelling. The optimised strategy could reduce the production cost of C-PC, thus enhancing the feasibility of commercial application of this versatile pigment. The optimum media components for producing maximum C-PC (45 mg g−1) in Euhalothece sp. KZN were found to be MgSO4 (0.10 g L−1), NaNO3 (1.67 g L−1) and minor nutrients (10 mL L−1). The outcome of this study resulted in an in-depth understanding of the roles these nutrients play in increasing the C-PC yield, as well as an integrated strategy for enhancing the C-PC in Euhalothece sp. KZN.