Whole fungal elicitors boost paclitaxel biosynthesis induction in Corylus avellana cell culture

Paclitaxel is an effective natural-source chemotherapeutic agent commonly applied to treat a vast range of cancers. In vitro Corylus avellana culture has been reported as a promising and inexpensive system for paclitaxel production. Fungal elicitors have been made known as the most efficient strategy for the biosynthesis induction of secondary metabolites in plant in vitro culture. In this research, C. avellana cell suspension culture (CSC) was exposed to cell extract (CE) and culture filtrate (CF) derived from Camarosporomyces flavigenus, either individually or combined treatment, in mid and late log phase. There is no report on the use of whole fungal elicitors (the combined treatment of CE and CF) for the elicitation of secondary metabolite biosynthesis in plant in vitro culture. The combined treatment of CE and CF significantly led to more paclitaxel biosynthesis and secretion than the individual use of them. Also, multivariate statistical approaches including stepwise regression (SR), ordinary least squares regression (OLSR), principal component regression (PCR) and partial least squares regression (PLSR) were used to model and predict paclitaxel biosynthesis and secretion. Based on value account for (VAF), root mean square error (RMSE), coefficient of determination (R2), mean absolute percentage error (MAPE) and relative percent difference (RPD) can be concluded that mentioned regression models effectively worked only for modeling and predicting extracellular paclitaxel portion in C. avellana cell culture.


Introduction
Paclitaxel is a well-known chemotherapeutic agent widely applied as a therapy for various types of cancers [1], and this is also used as a treatment for non-cancerous human diseases [2]. Paclitaxel as a cytoskeletal drug arrests the proliferation of tumor cells following stabilizing the microtubules. Indeed, this valuable metabolite blocks cell cycle in G0/G1 and G2/M phases by interaction with tubulin [3].
This fantastic diterpene alkaloid, paclitaxel, was originally found in the bark of Taxus brevifolia tree [4]. Since Taxus trees are slow-growing and harvesting the bark is destructive [5], a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 thus the continuous harvest of native plant bark for commercial production of paclitaxel was untenable. Nowadays, plant cell factories offer a promising and environment-friendly approach for large-scale paclitaxel production [6][7][8][9][10]. The rising demand for paclitaxel, and Taxus recalcitrant behavior under in vitro conditions have caused extensive efforts toward finding alternatives for producing this invaluable secondary metabolite.
In vitro culture of Corylus avellana (European filbert) has been made known as a promising and inexpensive strategy for producing paclitaxel [7][8][9][10][11][12]. The major advantage of producing paclitaxel through hazel cell culture is that in vitro culture of C. avellana is easier as compared to that of Taxus, and given that C. avellana is a dicotyledonous plant, possibly C. avellana response to genetic manipulation by Agrobacterium is more promising than that of Taxus [7][8][9][10]. Obtaining high-producing in vitro cultures is a key step for producing secondary metabolites through plant cell culture (PCC) [13]. The various factors affect the biosynthesis of bioactive compounds in plants [7-10, 12, 14-17]. Amongst the available techniques for enhancing secondary metabolites biosynthesis in PCC, the elicitation is likely the most effective one for dramatic increment in yield [18,19]. The mass biosynthesis of secondary metabolites even in plant cells overexpressed for key genes of biosynthetic pathway yet requires elicitation [20]. Among the various elicitors, fungal elicitors are widely applied for eliciting the biosynthesis of secondary metabolites in plant in vitro culture, as a result of their high effectuality and little toxicity on plant cells [21].
Endophytic fungi synthesize the conserved molecules known as microbe-associated molecular patterns (MAMPs). The first plant defense line is recognizing MAMPs by receptors on plant cell surface. Indeed, plant pattern recognition receptors localized on cell surface recognize MAMPs, and thus induce plant defense system [22][23][24][25]. The concentration levels and types of MAMPs are different in fungal cell extract (CE) and culture filtrate (CF). Accordingly, it seems that the potential of these fungal elicitors (CE and CF) is likely different in paclitaxel biosynthesis induction. The previous studies [8][9][10] demonstrated the positive influences of CE and CF of endophytic fungi on paclitaxel biosynthesis in cell suspension culture (CSC) of C. avellana. Nevertheless, no information is available regarding the simultaneous use of fungal CE and CF on secondary metabolites production in PCC including paclitaxel biosynthesis in C. avellana CSC. Fungal elicitor type, concentration and adding-time, and also exposure time of cell culture with it should be optimized to achieve the maximum biosynthesis of paclitaxel in C. avellana CSC [8][9][10]. However, the optimization of these factors is not only time-consuming but also costly.
Analyzing the relationship amongst input variables "CE and CF concentration levels, elicitor adding day and CSC harvesting time" and paclitaxel biosynthesis and secretion could help to optimize the conditions for the biosynthesis and secretion of this valuable secondary metabolite. Multivariate statistical approaches including stepwise regression (SR), ordinary least squares regression (OLSR), principal component regression (PCR) and partial least squares regression (PLSR) have been used in biological studies [26][27][28][29]. There are no studies to evaluate regression methods including SR, OLSR, PCR and PLSR to model and predict paclitaxel biosynthesis. SR is a well-known data-mining method selecting the explanatory variables for regression model from a group of input variables [30]. OLS is a statistical method estimating the relationship amongst independent variable(s) and dependent variable by minimizing sum of square differences among the predicted and observed values of dependent variable [31]. PCR is a regression method established on principal component analysis (PCA) [32]. PLSR, combining PCA and multiple regression, is a powerful modeling technique especially when the factors (input variables) are highly collinear [33]. Indeed, PLSR is an alternative for PCR which selects principal components that are related to independent variable [32].
The objectives of this research were (i) to evaluate the efficiency of CE and CF derived from endophytic fungus "Camarosporomyces flavigenus" isolated from C. avellana, either individually or as a combined treatment, on paclitaxel biosynthesis and secretion in C. avellana CSC, (ii) to estimate growth and paclitaxel biosynthesis (intracellular, extracellular and total) (model parameters) in C. avellana cell culture treated with fungal elicitors derived from C. flavigenus using regression methods (SR, OLS, PCR and PLSR) and (iii) to suggest the best regression model for prediction of growth, paclitaxel biosynthesis and its secretion in C. avellana cell culture.

Fungal and plant cell culture reagents
Culture medium components and paclitaxel standard applied in this research were purchased from Sigma (USA) and Merck (Germany) chemical companies.

Cell suspension culture
Callus of C. avellana (ecotype Gerd Ashkorat) was produced from seed cotyledons on MS medium supplemented with 2 mg l −1 2,4-Dichlorophenoxyacetic acid (2, 4-D) and 0.2 mg l −1 6-Benzylaminopurine (BAP), and 8 g l −1 agar agar [7]. C. avellana CSC was established with cultivating 5 g fresh callus into 250 ml flasks containing 100 ml of the same medium and the cultures were maintained at 25˚C in darkness on gyratory shakers at 110 rpm. CSCs were also subcultured until the cells reached homogeneity.

Preparation of elicitors and elicitation experiment
Endophytic fungus applied in this research was a strain of Camarosporomyces flavigenus, HEF 17 , isolated from the leaf of C. avellana grown in Iran. CE and CF elicitors were prepared as described previously [8]. For elicitation, 1.5 ± 0.1 g of C. avellana cells (fresh mass) was cultured in 100 ml flasks containing 30 ml MS medium supplemented with 2 mg l −1 2,4-D and 0.2 mg l −1 BAP.

Cell growth measurement
Cell growth was determined by the measurement of cell dry weight (DW). Cell biomass was separated from culture medium by the filtration (Whatman No. 1) and washed with distilled water to remove the residual medium, afterward freeze-dried to constant weight by a vacuumfreeze drier.

Quantification of paclitaxel
C. avellana cells were separated from culture medium by a filter paper (Whatman No. 1). Intracellular and extracellular paclitaxel were extracted from the cells and culture broth using a procedure described by Salehi et al. [7,8,12]. Filtering all samples was performed by 0.22 μm cellulose acetate syringe filters before HPLC analysis. Paclitaxel in the samples was analyzed by HPLC (Waters, USA) with a C18 analysis column (Machereye-Nagel EC 250/4.6 Nucleodur). Each sample (20 μl) was injected and detected at 230 nm using a UV detector. The mobile phase was methanol: water (80:20 v/v) at a flow rate of 1.0 ml/min. The quantification of paclitaxel was based on an external standard of genuine paclitaxel (Sigma).

Regression methods
The data were randomly divided into a training subset (70%) and testing one (30%), respectively. Training subset was applied to develop regression models, and testing subset was applied to test the predictability of developed models [34].
Stepwise regression (SR), ordinary least squares regression (OLSR), principal component regression (PCR) and partial least squares regression (PLSR) were used to predict DW, intracellular, extracellular and total yield of paclitaxel, and also extracellular paclitaxel portion. XLSTAT software [35] and Minitab [36] software were used for the development and evaluation of regression models. Also, the graphs were made by GraphPad Prism 5 [37] software.
RMSE ¼ ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffiffi X n i¼1 Where "y act " are the measured values, "y est " are the predicted values, and "n" is the number of data. Evaluation of regression models was performed according to RPD values (Table 1) [38].

Effects of elicitors derived from C. flavigenus on C. avellana cell growth
DW of the cells in C. avellana CSCs exposed with E 100/0 , E 75/25 , E 50/50 , E 25/75 and E 0/100 derived from C. flavigenus in mid (day 13) and late (day 17) log phase were measured as cell growth.
Analysis of variance (ANOVA) demonstrated that the main effects of factors "elicitor type, concentration level and elicitor adding day" and also their reciprocal interactions on DW were not significant ( Table 2), suggesting that fungal elicitors, regardless of its adding day and concentration levels, did not have an impact on cell growth in C. avellana CSC (Fig 1). It displays adaptability of C. avellana cells to elicitors derived from C. flavigenus, in view of the fact that strain HEF 17 is a symbiont of this plant. It is reported that endophytic fungi display no negative effect on plant cell growth, and even a few of them are capable of enhancing the growth [8,9,39,40].

Effect of exposure time of C. avellana cells with fungal elicitor on paclitaxel biosynthesis
To study the relations between paclitaxel biosynthesis and exposure time of C. avellana cells with E 100/0 , E 75/25 , E 50/50 , E 25/75 and E 0/100 derived from C. flavigenus, paclitaxel content in C. avellana CSCs treated with various concentration levels (2.5, 5 and 10% (v/v)) of five fungal elicitors at days 13 and 17 (mid and late log phase) were measured in 2-day periods after elicitation (Fig 2). Taken together, paclitaxel biosynthesis increment was recorded in the course of cell growth, and its highest statistically significant level was biosynthesized on day 21. Paclitaxel content in C. avellana CSCs reduced on 23 rd day, likely because of paclitaxel molecular structure changes or its degradation beginning. The degradation of bioactive molecules could happen either extra-or intracellularly [41]. Paclitaxel productivity (Pr) and paclitaxel biosynthesis elicitation index (ratio of elicited culture paclitaxel productivity to control one, Pr e /Pr c )  (Fig 2).

Effects of elicitors derived from C. flavigenus on paclitaxel biosynthesis
The effects of elicitors "E 100/0 , E 75/25 , E 50/50 , E 25/75 and E 0/100 " derived from C. flavigenus on paclitaxel biosynthesis were surveyed in an elicitor adding time-, exposure time-and concentration level-dependent way. Maximum significant paclitaxel level was recorded on day 21 ( Fig  2). Thus, this time point, day 21, was selected as the benchmark of paclitaxel biosynthesis. The results of paclitaxel biosynthesis induction in C. avellana cell cultures using elicitors derived from C. flavigenus displayed that the biosynthesis of this metabolite was significantly affected by above mentioned fungal elicitors ( Table 2). The results of ANOVA showed that the main effects of factors (fungal elicitor type, concentration level and adding day) and their interactions (reciprocal and trilateral effects) on paclitaxel biosynthesis (intracellular, extracellular and total yield) were highly significant (p < 0.01) ( Table 2). According to these significant interactions (Table 2), it can be concluded that fungal elicitor adding day and concentration levels affected paclitaxel biosynthesis variously at each elicitor type. Therefore, elicitor concentration level and its adding day were further scrutinized on each elicitor to carefully analyze these significant interactions.

Effects of adding day and concentration level of fungal elicitors on paclitaxel biosynthesis
According to means comparison, adding 5 and 10% (v/v) E 100/0 on 13 th day of cell culture cycle led to significantly higher paclitaxel biosynthesis (1.4-fold) as compared to control, but adding 2.5% (v/v) E 100/0 on 13 th day did not significantly improve paclitaxel biosynthesis (Fig  3). In the same way, C. avellana cell cultures treated with 2.5% (v/v) E 100/0 on 17 th day of cell culture cycle displayed no significant paclitaxel biosynthesis as compared to control (Fig 3). However, adding 5 and 10% (v/v) of this elicitor at day 17 significantly improved paclitaxel biosynthesis (137.04 μg l -1 ; 1.9-folds). Paclitaxel biosynthesis in C. avellana cell cultures treated with 5 and 10% (v/v) E 100/0 in late log phase, day 17, were significantly higher (1.4-fold) than that in mid log phase. The optimized concentration level for E 100/0 on 17 th day of cell culture cycle was 5% (v/v) (Fig 3). As shown in Fig 3, adding 2.5% (v/v) E 75/25 (1.875% (v/v) CE and 0.625% (v/v) CF) on 13 th and 17 th days did not affect paclitaxel biosynthesis in C. avellana CSC, whereas cell cultures exposed with 5 and 10% (v/v) of mentioned elicitor in mid and late log phase significantly enhanced paclitaxel biosynthesis. Adding 5 and 10% (v/v) E 75/25 at day 17 resulted in the significantly higher contents of paclitaxel (1.7-fold) as compared to that at day 13. The most significant total yield of paclitaxel (166.6 μg l −1 ) in cell cultures subjected to E 75/25 was biosynthesized using 5% (v/v) of this elicitor at day 17, about 2.3-fold that biosynthesized in control culture (Fig 3). It is noteworthy that no significant difference was observed between cell cultures treated with 5 and 10% (v/v) E 75/25 at day 17.
As illustrated in Fig 3, adding different concentration levels of E 25/75 derived from C. flavigenus in mid and late log phase (except 2.5% (v/v) of it at day 13) resulted in the higher significantly biosynthesis of paclitaxel as compared to control culture. Paclitaxel biosynthesis in cell cultures treated with 5 and 10% (v/v) E 25/75 derived from C. flavigenus in late log phase (day 17) were significantly higher (1.8-and 1.7-fold, respectively) than that in mid log phase (day 13) (Fig 3).
According to mean comparison, C. avellana cell cultures treated with 5 and 10% (v/v) E 0/ 100 in mid log phase (day 13), and also 2.5, 5 and 10% (v/v) of it in late log phase (day 17) displayed a significant increment in paclitaxel biosynthesis (Fig 3). As Fig 3 shows adding 2.5, 5 and 10% (v/v) E 0/100 at day 17 resulted in the significantly higher contents of paclitaxel (1.4-, 2.0-and 1.8-fold, respectively) than that on day 13. The maximum significant total yield of paclitaxel in cell cultures exposed to E 0/100 (203.5 μg l −1 ) was biosynthesized using 5% (v/v) of it on day 17, about 2.8-fold that biosynthesized in control culture (Fig 3).
In this research, whole fungal elicitors (the combined treatment of fungal CE and CF) have been used for the elicitation of secondary metabolite biosynthesis in PCC for the first time.
The results clearly showed the high potential of whole fungal elicitors for paclitaxel

PLOS ONE
biosynthesis increment in C. avellana cell culture. Various fungal elicitors applied in this research led to the different responses regarding paclitaxel biosynthesis elicitation. E 50/50 consisting 5% (v/v) CE and 5% (v/v) CF derived from C. flavigenus strain HEF 17 , isolated from C. avellana leaf, was an efficient elicitor for biosynthesizing paclitaxel in C. avellana cell culture. Endophytic fungi synthesize conserved molecules known as microbe-associated molecular patterns (MAMPs) that can activate plant defense response [22,42]. Indeed, the recognition of MAMPs via plant cell surface receptors is the initial phase of defense response induction. The particular structure of receptors leads to particularly recognize specific MAMPs [43]. The specific and diverse responses of plant cells to different fungal elicitors in the elicitation of secondary metabolite biosynthesis as seen in our research can be due to unique interactions of plant cell-surface receptors with fungal MAMPs [44].
Our results illustrated that paclitaxel secretion was also increased by E 50/50 , 24.1% more than that in control (Fig 4). Paclitaxel secretion from the cells to culture medium decreases toxicity and feedback inhibition of paclitaxel [10,45]. Besides, the secretion of paclitaxel to culture medium undoubtedly makes easy extraction and the purification of it which is required for the steady production of paclitaxel at the commercial level. The influence of fungal metabolites on plant cell membrane transporters likely leads to secrete paclitaxel from the cells into culture medium that is a key strategy for a continuous production system.
The various fungal elicitors have been used in C. avellana cell culture to enhance paclitaxel productivity. C. avellana cell culture treated with CF (2.5% (v/v) on 17 th day) of Paraconiothyrium brasiliense strain HEF 114 isolated from C. avellana displayed a 3.0-folds increment in paclitaxel biosynthesis [8]. Also, C. avellana cell culture subjected to CE (10% (v/v) on 17 th day) of Chaetomium globosum strain YEF 20 led to a 4.1-fold increment in paclitaxel biosynthesis [8]. In another report, the addition of 2.5% (v/v) CE of Epicoccum nigrum strain YEF 2 to C. avellana CSC resulted in a 3.6-fold in paclitaxel biosynthesis [9]. In another attempt to find the efficient elicitors, out of different fungal elicitors, 5% (v/v) CE and also 2.5 and 5% (v/v) of cell wall (CW) derived from Coniothyrium palmarum added at day 17 resulted in maximum paclitaxel biosynthesis (3.6-folds), while a combined treatment of CW (2.5% (v/v) on 17 th day) and 50 mM of methyl-β-cyclodextrin (MBCD) highly enhanced paclitaxel biosynthesis (5.8-fold) in C. avellana cell culture with displaying a synergistic effect [10]. In another report, a slight increase in paclitaxel biosynthesis displayed in C. avellana cell culture affected by silver nanoparticles [46]. Also, the joint effects of phenylalanine (3 μM) and vanadyl sulfate (0.05 and 0.1 mM) in culture medium supplemented with fructose (3% (v/v)) led to a 2.3-fold increment in paclitaxel biosynthesis [47].
In the light of the prominent positive effect of MBCD on paclitaxel biosynthesis, it can be suggested to assess the effects of whole fungal elicitors (E 50/50 consisting 5% (v/v) CE and 5% (v/v) CF) and MBCD, in a combined treatment on paclitaxel productivity in C. avellana cell culture.
Regardless of previous studies on the effects of fungal CE and CF on paclitaxel biosynthesis and secretion in C. avellana CSC, there was still the question that needed to be answered; whether whole fungal elicitors (the combined treatment of CE and CF) would lead to a higher enhancement of paclitaxel biosynthesis as compared to individual fungal elicitor? This research introduced "whole fungal elicitors" as an efficient treatment for boosting the biosynthesis of secondary metabolites in plant in vitro culture, paclitaxel biosynthesis in C. avellana CSC as a case study.
Finally, it is important to note that paclitaxel biosynthesis in C. avellana cell culture was not as high as that reported for Taxus. However, fast-growing in vitro culture of C. avellana and a series of genetic manipulation may compensate for the lower yield of paclitaxel in C. avellana.

Regression models
All regression models (SR, OLSR, PCR and PLSR) displayed statistically significant relationships between each output variables (DW, intracellular paclitaxel, extracellular paclitaxel, total yield of paclitaxel and extracellular paclitaxel portion) and input variables (CE and CF concentration levels, adding day and harvesting day) (Table 3). SR, OLSR, PCR and PLSR models developed for output variables "DW, intracellular paclitaxel, extracellular paclitaxel, total yield of paclitaxel and extracellular paclitaxel portion" regarding CE and CF concentration levels, elicitor adding day and CSC harvesting day were shown in Table 3. Goodness-of-fit test of SR, OLSR, PCR and PLSR models was performed to detect the best model for predicting each output variables (DW, intracellular paclitaxel, extracellular paclitaxel, total yield of paclitaxel and extracellular paclitaxel portion). High VAF, R 2 and RPD values and low RMSE and MAPE values displayed the model capability.
As shown in Table 3, SR had slightly higher VAF, R 2 and RPD and lower RMSE and MAPE values as compared to OLSR, PCR and PLSR for predicting paclitaxel biosynthesis (intracellular, extracellular and total) and secretion. However, OLSR, PCR and PLSR displayed slightly higher VAF, R 2 and RPD and lower RMSE and MAPE values as compared to SR. Goodnessof-fit displayed no difference regarding the accuracy of OLSR, PCR and PLSR for all output variables, 0.67, 0.57, 0.62, 0.60 and 0.86 for DW, intracellular paclitaxel, extracellular paclitaxel, total yield of paclitaxel and extracellular paclitaxel portion, respectively for training subset (Table 3). R 2 values for predicting DW, intracellular paclitaxel, extracellular paclitaxel, total yield of paclitaxel and extracellular paclitaxel portion using SR models were estimated 0.65, 0.62, 0.68, 0.65 and 0.85, respectively for training subset ( Table 3).
The fit of regression models was presented by R 2 (Fig 5) for testing subset, suggesting the best-mentioned models can explain 67, 62, 68, 65 and 86% of the variability in intracellular paclitaxel, extracellular paclitaxel, total yield of paclitaxel and paclitaxel extracellular portion, respectively, when they faced unseen data.
In this study, multiple regression models were applied to determine the relationships among four factors "CE and CF concentration levels, elicitor adding day and CSC harvesting time" and each of parameters "DW, intracellular, extracellular and total yield of paclitaxel and extracellular paclitaxel portion", and also the possibility of predicting of paclitaxel biosynthesis by the determining factors. Such mathematical predictions have been applied to predict cell growth and paclitaxel biosynthesis in C. avellana cell culture for the first time.
According to RPD, all developed regression (SR, OLSR, PCR and PLSR) displayed superior ability for predicting extracellular paclitaxel portion (Tables 1 and 3). Based on VAF, RMSE, R 2 , MAPE and RPD (Table 3), it can be concluded that mentioned regression models effectively worked only for modeling and predicting extracellular paclitaxel portion in C. avellana cell culture.
Since mentioned various regression methods showed fair predictive and fitting ability for growth and paclitaxel biosynthesis (Tables 2 and 3). It is recommended that artificial intelligence is used for accurate modeling and predicting growth and paclitaxel biosynthesis.

Conclusion
This research presents the high potential of whole fungal elicitors (the combined treatment of CE and CF) derived from C. flaveginus strain HEF 17 isolated from C. avellana for paclitaxel biosynthesis induction in C. avellana cell culture. The combined treatment of fungal CE and CF led to more paclitaxel biosynthesis than the individual use of them, suggesting the presence of different MAMPs with different concentrations in fungal CE and CF. High RPD shows the superior ability of regression models for predicting extracellular paclitaxel portion. Great accordance between the predicted and observed values of extracellular paclitaxel portion ( Fig  5) supports the excellent performance of developed SR, OLSR, PCR and PLSR models for predicting extracellular paclitaxel portion.