Central Composite Design for Optimizing the Biosynthesis of Silver Nanoparticles using Plantago major Extract and Investigating Antibacterial, Antifungal and Antioxidant Activity

Central composite design (CCD) was applied to optimize the synthesis condition of silver nanoparticles (AgNPs) using the extract of Plantago major (P. major) seeds via a low cost and single-step process. The aqueous seed extract was applied as both reducing element and capping reagent for green production of AgNPs. Five empirical factors of synthesis including temperature (Temp), pH, volume of P. major extract (Vex), volume of AgNO3 solution (VAg) and synthesis time were used as independent variables of model and peak intensity of Surface Plasmon Resonance (SPR) originated from NPs as the dependent variable. The predicted optimal conditions was determined to be: Temp = 55 °C, pH = 9.9,Vex = 1.5 mL, VAg = 30 mL, time = 60 min. The characterization of the prepared AgNPs at these optimum conditions was conducted by Fourier transform infrared spectroscopy (FTIR), dynamic light scattering (DLS), transmission electron microscopy (TEM) and X-ray diffraction (XRD) to determine the surface bio-functionalities. Bio-activity of these AgNPs against bacteria and fungi were evaluated based on its assay against Micrococcus luteus, Escherichia coli and Penicillium digitatum. Furthermore, antioxidant capacity of these NPs was checked using the ferric reducing antioxidant power (FRAP) assay.

instruments. The UV-Vis absorbance measurements were done using a Carry 60 UV-vis spectrophotometer (Agilent, USA) by a quartz cuvette (10 mm). The X-ray diffraction (XRD) were performed by the XRD Bruker D8 Advance powder diffractometer (Bruker, Germany) by applying the reflection mode with Cu-K α radiation (λ = 1.5406 Å). The Fourier Transform-Infra Red (FT-IR) spectra were collected at room temperature using a "Spectrum RXI" Perkin-Elmer FT-IR spectrophotometer (PerkinElmer, USA) after tableting the AgNPs powder. The distribution of size and volume of NPs were also determined using Dynamic Light Scattering (DLS) using HORIBA SZ-100 nanoparticle analyzer (HORIBA, Japan).
The morphology of the particles was investigated using the Transmission Electron Microscopic examination (TEM). For this study, colloidal AgNPs were sonicated and a thin film was formed on the carbon coated grid Cu Mesh 300. TEM measurements were done using a Zeiss-EM10C microscope (Zeiss, Germany) operated at an accelerating voltage of 80 kV. The pH adjustment during synthesis was performed using a Metrohm 827 pH-meter equipped with a combined glass electrode (Metrohm, Switzer-land).
preparation of Plantago major extract. The seeds of P. major were prepared from local market (Shiraz, Iran) and were fast-washed with distilled water to eliminate the dusts and impurities. The clean and dried seeds were powdered mechanically. 20.0 g powder of P. major seeds was mixed in 100 mL of deionized water in 500 mL flask; and heated at 80 °C for 30 min by using a heater-magnetic stirrer. After cooling, the resulted solution was initially centrifuged and filtered through normal textile filter to eliminate the big particles; then the obtained extract was filtered with filtration paper (Whatman, No. #1). The filtered P. major extract was stored in the refrigerator at 4 °C before use. The extract solution was not utilized more than one week. It is clear that the P. major extract was applied as the stabilizing and reducing agent during the production of silver NPs.
www.nature.com/scientificreports www.nature.com/scientificreports/ preparation of Agnps using Plantago major extract. In the current research, the solution of silver nitrate salt in deionized water was applied as the metal source (Ag). The reaction vessel contained the determined volume (V e , mL) of the extract in desired volume of AgNO 3 solution 1 mM (V Ag , ml) under stirring at reaction temperature (Temp, °C) and desired reaction time (t, min) for the reduction of Ag + to Ag°. The reaction was continued until the color of the solution became reddish-brown that showed the construction of AgNPs. The effect of pH on the synthesis of AgNPs was first studied by fixing the pH of the reaction media to 3.0, 5.0, 7.0, 9.0, 10 and 11.0 to have an estimate regarding the useful range of pH. Then the effect of temperature, pH, extract content, concentration of AgNO 3 salt and mixing time were evaluated simultaneously using response surface methodology with CCD. The exact amounts of reagents and compounds, finally utilized for synthesis of AgNPs, were obtained during optimization with CCD and are reported and explained in the "Results and Discussion" section but is briefly denoted hare: Temp = 55 °C, pH = 9.9,V ex = 1.5 mL, V Ag = 30 mL, time = 60 min. optimization of synthesis conditions using ccD. As it was noted previously, in order to obtain optimum synthesis conditions for preparation of AgNPs, a CCD has been applied. This method has been used to achieve the best possible SPR intensity as a confirmation of NPs production. For all UV-Vis spectrophotometric studies, equivalent volume of the suspension (0.5 ml) were diluted with a constant volume (=2 ml) of deionized water and subsequently all measurements were done at room temperature. The design was done based on five levels for each noted factor in single-block mode and the obtained 32 runs was performed in a random order to minimize the effect of uncontrolled variables on the quality of NPs (followed by SPR).
Antibacterial activity. The antimicrobial activity of AgNPs against Gram-positive (Micrococcus luteus ATCC 9341) and Gram-negative (Escherichia coli ATCC 25922) indicator bacteria was first evaluated using the disk diffusion method described by Kumar et al. 42 . In brief, the overnight culture of bacterial strains was adjusted to a 0.5 McFarland turbidity standard (1.5 × 10 8 cfu/mL). The inoculum was then added to the molten brain heart infusion (BHI) agar (Merck, Darmstadt, Germany) to achieve a final concentration of about 10 6 cfu/mL. Afterwards, 6 mm-diameter blank disks were placed onto the seeded agar plates and 10 µL aliquots of different concentrations of AgNPs (full strength, 50%, 25%, 12.5%, 6.25%, and 3.125% v/v) as well as P. major extract were added to each disk. After 24 h incubation at 37 °C, the growth inhibition zone diameters were measured. Standard gentamicin disks (10 µg) were used as positive control. All tests were conducted in triplicate.
The minimal inhibitory concentrations (MIC) of AgNPs were also determined for test organisms using a broth microdilution method. The full-strength concentration of AgNPs (10 mM or 1078 µg/mL; which is shown as 100% in this research) was two-fold serially diluted in the wells of a microtiter plate containing 70 µL of BHI broth to achieve reduced concentrations (50%, 25%, 12.5%, 6.25%, and 3.125% v/v). Following addition of resazurin solution (final concentration of 1 mg/mL) and bacterial suspensions (final concentration of 10 6 cfu/mL) to each well, the plate was incubated at 37 °C for 24 h. A set of controls were used in this experiment: a) the culture medium; b) the culture medium inoculated with test strains; c) the culture medium mixed with P. major extract and inoculated with test strains; and d) the culture medium mixed with AgNPs or P. major extract without bacteria. Bacterial growth was verified by color change from purple-blue to pink or colorless. The lowest concentration without any discoloration was considered as MIC value 43 . The experiments were carried out in 2 replicates. Antifungal activity. A disc diffusion assay was also applied to evaluate the antifungal activity of biosynthesized AgNPs against Penicillium digitatum IR1037c (P. digitatum). The conidia suspension was prepared by washing the surface of potato dextrose agar (PDA) containing a 7-day old culture of P. digitatum with 2 mL of normal saline (9 g/L NaCl). The suspension was then adjusted to 10 6 conidia/mL using a Neubauer's chamber and seeded in PDA to achieve a final concentration of 10 4 conidia/mL. Blank disks were placed on the agar surface and after adding AgNPs or P. major extract (10 µL), the plates were incubated at 25 °C for 48 h 44 . ferric reducing antioxidant assay for antioxidant activity. The antioxidant potential of Plantago major extract and AgNPs were checked using an in vitro assay, using ferric-reducing antioxidant power (FRAP) 45 . In FRAP test, the principal is on the potential of antioxidant compounds for reduction of the intense blue ferric tripyridyltriazine complex and formation of its ferrous complex as the product, which can lead to change the absorbance of ferric complex at 593 nm. The preparation of the required reagent for FRAP was done daily by adding acetate buffer (pH 3.6, 300 mM) to 10 mM solution of tripyridyl-s-triazine (TPTZ) in 40 mM HCl and 20 mM FeCl 3 .6H 2 O. The ratio of TPTZ: HCl: FeCl 3 .6H 2 O was 10:1:1 and was kept at 37 °C. Stock concentration of each target sample (P. major extract and prepared AgNPs) was firstly prepared in deionized water (1078 μg/mL) and different concentrations of NPs or P. major extract were prepared by dilution (808.5, 539, 269.5, 134.75 μg/mL). Then, 1.5 ml of the FRAP reagent was mixed with 50 μl of target sample and the mixture was then retained at 37 °C for 10 min in the dark. Here, excess amount of Fe(III) was utilized, and the color formation due to the construction of Fe(II)-TPTZ was the rate-limiting criteria to show the reducing power of the extract or investigated AgNPs. In the FRAP assay for each target sample, the blank solution was prepared with just FRAP reagent. In all runs of assay, the final absorbance of sample (i.e.: FRAP reagent + target solution) was corrected by subtraction of the blank absorbance. The calibration or standard curve of this assay was prepared using a set of solution with known concentrations of FeSO 4 (100 to 2000 µM). All the obtained values in this assay were reported as mM of ferrous ion (Fe (II)) equivalent.

Results and discussion
Response surface methodology. In this work, CCD strategy was used to find optimum experimental condition for green synthesis of AgNPs. CCD has been employed to achieve the best SPR band of synthesized AgNPs with considering of the peak intensity. The goal was the investigation of the five empirical factors of synthesis including temperature (Temp), pH, volume of extract used for synthesis (V ex ), volume of 1 mM AgNO 3 solution (V Ag ) and synthesis time (time).
It is worthy to note that based on our previous studies, pH is one of the most effective variables in synthesis of AgNPs. Thus, to determine the operational value range of pH, a serie of synthesis was done by one-at-a-time approach with varying the pH and constant amounts of other factors (Temp = 40 °C, V ex = 2 ml, V Ag (0.1mM) =25 ml and time=60 min). As it is represented in Fig. 1, by increasing pH the quality of synthesis was enhanced and acidic pH (3.0 and 4.0) had no good results and doesn't show SPR which can be related to the suppression of the AgNPs construction. At pH 5.0, as an acidic media, a broad SPR peak was observed with λ max at 490 nm which can be due to increase in NPs size because of their agglomeration in acidic pH 46 . As it is observed in Fig. 1 in alkaline and neutral pH, the construction of particles was instantaneous which led to higher intensity in SPR band. This can be related to the ionization of the phenolic functional groups in the utilized seed extract and enhancement its reduction ability 47 . By increasing the pH from 7.0 to alkaline values, the target SPR was become narrower with blue shift (to lower wavelengths).
Thus the pH in neutral to alkaline range (7, 8.5 and 10) was the main levels which were used in CCD. Also two syntheses was done in lower and higher pH as their boundaries points (−α and +α), i.e in 5.5 and 11.5 respectively. Five experimental factors used for biosynthesis of AgNPs and their corresponding five levels (3 levels and 2 boundary levels) are represented in Table 1. Table 2 represents 32 runs which has been designed by CCD and was made of a 2 4 full factorial design. 16 experiments of these 32 were the factorial runs, which were included as the cubic points, and 6 replication runs were suggested in center points of levels, which were used for clarification of pure error. In addition, 8 axial runs were included as star points of the suggested CCD. To have a rotatable design for experiments, the α star points have been adjusted in a value of 2.0.
"Peak intensity" of the SPR was the utilized as responses of the CCD design and modeling which resulted from recording the absorbance of 32 synthesized beaches of AgNPs at 490 nm.
Statistical analysis of ccD. To start the statistical analysis using a multivariate optimization approach the first step was construction of a regression model between the SPR intensity (target response) and the experimental factors as the independent variables of model. For this goal, various linear and polynomial regression models were evaluated and different statistical parameters such as Fisher F-value using F-test (for both regression and lack of fit), squared correlation coefficient (R 2 ), adjusted R 2 (R 2 adj ) and the R 2 of prediction (R 2 pred ) were utilized to select the best regression model. Finally based on these statistics, a quadratic model was proposed to correlate the relationship between the experimental factors and the response (peak intensity of SPR). Since the goal of this part  www.nature.com/scientificreports www.nature.com/scientificreports/ was finding conditions for obtaining maximum SPR intensity, the analysis of variance (ANOVA) was constructed for this goal which is represented in more details in Table 3. According to Table 3, the performance parameters obtained for the quadratic full model (FM) including all experimental variables (Temp, V ex , V Ag , and time), their binary interactions (Temp. × pH, Temp. . Table 3 by ANOVA that the full model (FM) is valid according to the p-value of Fisher F-test to explain the SPR intensity (model responce) properly. In this regard, the low p-value of the model (0.0131) indicated the significance of the mean square of model in comparison to the mean square of residual. Also, the "lack of fit" (LOF) of Eq. (1) is not significant as shown by the p-value of the lack of fit. In spite of these results, it can be observed in Table 3 that the p-values of some of the variables in FM are not significant. According to which is common in CCD modeling, terms with P-values greater than 0.1 are not significant and must be eliminated from www.nature.com/scientificreports www.nature.com/scientificreports/ the final selected model at the desired confidence limit. Thus a variable selection strategy (backward elimination) was applied to eliminate the redundant variables or interactions in Eq. (1) and development of a refined linear model (RM), as shown in the following: As it can be seen by comparison of Eqs. (1) and (2), among 20 terms in FM (Eq. 1), five interaction terms such as Temp × pH, Temp × V ex , Temp × V Ag , pH × time, V ex × V Ag and three self-interaction variables such as pH 2 , V Ag 2 and time 2 was excluded from full model by backward elimination and Eq. (2) was suggested as the simpler model named as refined model (RM) contained 12 variables (original ones, interactions and self-interactions). However it is very worthy to note that the constraint which was applied during variable selection was obtaining a "hierarchy" model i.e. not removing a variable in presence of each kind of interactions of the removed variable. In Analysis of variance (ANOVA)  www.nature.com/scientificreports www.nature.com/scientificreports/ other words, the variable "time" was retained in model because of presentation of its interactions. Thus As it can be seen in Table 3, the p-value of coefficients of each remaining variables in Eq. (2) as the RM (except "time") are lower than 0.1 which confirms the significance of coefficients. On the other hand the p-value of the total model in RM is also lower than 0.0001 and indicates the significance of obtained linear model as well. Here again, as another good criteria, the p-value of the lack of fit of model suggests that the LOF of Eq. (2) is not significant.

Source of variation
By comparison of FM and RM for the optimization of SPR intensity in Table 3, it can be seen that F-values of the model improved significantly from 3.78 to 7.44. Also F-value of lack of fit in both FM and RM remained below 1 (0.84 and 0.63 respectively) and P-values of LOF increased from 0.5891 to 0.7710 which shows the decreasing of error in reduced model after variable selection. Thus, the Eq. (2) seems acceptable. In addition to the equations obtained based on codded factors shown in Eqs. (1) and (2), the equations obtained from real values of experimental factors (without coding) before and after variable selection are also represented in Table 3.
For more evaluation of the models, the correlation coefficient of the model (R 2 ) and adjusted R 2 (R 2 adj ) were calculated which were equal to 0.83 and 0.72 respectively for Eq. (2) which have improvement in compare with Eq. (1) and show enough goodness of fit in Eq. (2) 48 . It is noteworthy that for limited number of runs in models with high number of variables, R 2 adj is a better parameter than R 2 because R 2 adj is adjusted to the number of variables and the size of runs (here 32) in model. The R 2 pred for Eq. (2) was also equal to 0.78 which indicates an acceptable prediction ability 48 . "Adequate precision" is also another statistical indicator for validation of our model and can be an estimate for signal to noise ratio. The "adequate precision" in our suggested refined model (Eq. (2)) was equal to 13.8 which was significantly greater than 4, as the cut-off value for acceptance of precision 49 2)).
The direct effect of temperature on the SPR may be due to enhancing in the rate of NPs formation 50 . As it was expected increasing the pH has positive effect on the SPR intensity which can be because of increasing the ionization of active functional groups especially phenolic groups in P. major extract and enhancing the reduction ability of extract and production of AgNPs. Here, it can be observed that V Ag have a direct relationship on SPR intensity but but it has been reported in literature that has reverse effect on the size of AgNPs 26,46,51 . Increasing SPR intensity shows enhancing NPs concentration. However it should be emphasized that the proposed model has a multi-parameter nature and the effect of other factors and interactions should be considered simultaneously. The positive sign of time in the proposed model also shows that as the mixing time increased, the formation of AgNPs enhanced as well. The negative sign of V ex indicates that decreasing in volume of P. major extract causes increasing in the SPR intensity of the obtained AgNPs. In high volume of P. major extract, the SPR intensity decreased and the AgNPs were not stable due to an excess amount of reducing compounds which may cause instantaneous precipitation of particle 52 .
As can be observed in Eq. (2), among the interaction terms, the sign of Temp × time, pH × V Ag and V Ag × time are negative which shows their reverse effect on AgNPs synthesis and the sign of others (V ex × time, Temp 2 and V ex 2 ) are positive which emplies direct effect on the synthesis quality and subsequently on their SPR intensity. Two additional approaches which are useful to evaluate the accuracy of models are following the normal probability of residual and also the internally studentized residual versus predicted response (SPR intensity). The normal probability plot of the residuals is represented in Fig. 2(a) and shows that all the residual values lie along the straight line with no big deviations and confirms the normality of errors-distribution. It is also shown in Fig. 2(b) that not only the residual are randomly distributed in both sides of zero line (without any systematic pattern in one side), but also lie in the range of ±2σ (even lower than permitted range of ±3σ). Therefore, Fig. 2 declares that the proposed linear model is acceptable and has not any systematic error and can be applied for optimization of the synthesis process.

Visualization and comparison of interactions. To investigate the effect of various operational factors
(independent variables) on SPR intensity of AgNPs (dependent variable), three-dimensional (3D) surface representations of the factors are illustrated in Fig. 3. In this regard, graphs have been prepared by drawing the response surface (i.e. SPR intensity) versus the two effective factors that shows binary interactions whereas other factors have been retained constant at the central level of the applied CCD. Figure 3(a) shows that Temp has higher impact on the SPR intensity in compare with time, because slope of the obtained surface in the direction of Temp axis is steeper than time axis. On the other hand, by comparison of Fig. 3(a) with Fig. 3(d,e) it can be observed that the steepest slope of the surface plot can be observed for time versus V Ag and the lowest slope was obtained versus V ex . Interestingly only interaction of pH with V ex and V Ag were entered in model which are shown in Fig. 3(b,c) respectively. As it was expected and according to Fig. 3(b.c), the synthesis of AgNPs was better done in alkaline pH. It is also clear from Fig. 3(c) that the interaction of pH and V Ag is the biggest effect on the synthesis of AgNPs based on the slope of curve.
According to the Fig. 3(a,c and e), there are no abnormal interaction between factors because all of the obtained surfaces are smooth in each plot and both factors have similar effect on the SPR intensity. It is also clear that, when the factors in Fig. 3(a,c and e) are in their higher possible amount, the SPR intensity and the quality of synthesis increased. But as can be seen in Fig. 3(b,d) the interaction of pH × V ex and V ex × time are more complex because there is more curvature in their 3D surface. On the other hand the increasing of factors during enhancement of SPR intensity is not observed in Fig. 3(b,d).
optimal operational condition of synthesis. After development and validation of the multivariate linear models, the optimal synthesis conditions have been discovered with a multi-response approach. Here, it was assumed simultaneously to maximize the SPR intensity to have maximum amounts of AgNPs with lower sizes. The predicted optimal conditions was determined to be: Temp = 55 °C, pH = 9.9,V ex = 1.5 mL, V Ag = 30 mL and (2020) 10:9642 | https://doi.org/10.1038/s41598-020-66357-3 www.nature.com/scientificreports www.nature.com/scientificreports/ time = 60 min. To check the accuracy of the predicted optimal conditions, some synthesis batches of AgNPs were prepared under the noted condition and one of the obtained UV-Vis spectra from the synthesized particles is demonstrated in Fig. 4(a). Good SPR spectra were obtained from the synthesized NPs. This result illustrates good agreement between predicted values by model and experimental ones for the optimal conditions. For example, in the optimal conditions, the relative errors for prediction of PSR intensity in different synthesis runs were 2-5.5%. These results confirm the ability of P. major extract for biosynthesis of AgNPs with good repeatability and also show the possibility of controlling the conditions with a mathematical approach.
Reduction ability of Plantago major extract. In general, the phytochemicals and phenolic compounds in the herbal extracts could be act as the reducing components to convert the Ag + ions into Ag°, to stabilize the synthesized AgNPs and to inhibit agglomeration of NPs via chelate effect base on binding to the metals 53 . The P. major extract also acted as the stabilizing agent that obviates the addition of any other chemical reducing agent or external stabilizing compound and surfactant.
The P. major extract including alkaloids, flavonoids, phenolic acid derivatives, terpenoids, fatty acids, iridoid glycosides, polysaccharides and vitamins 54 . The main flavonoids present are flavones, including apigenin and luteolin 55,56 . Plantamajoside and acteoside, also named as verbacoside are the determined caffeic acid derivatives in P. major extract 54 . Thus, the utilized plant extract is a good option for reduction of silver ion and formation of AgNPs.
To have an estimate on the functional groups of the P. major extract FT-IR and UV-Vis spectroscopy were carried out and these spectra are shown in Fig. 4. The UV-Vis spectrum of P. major extract in aqueous media is shown in Fig. 4(a). The bonds ranging 262 to 356 nm with λ max at 315 nm can be related to the transition within the cinnamoyl ring system and the double bonds (i.e. π → π* transitions) 57 . It is noteworthy that the extract of P. major was prepared in aqueous media and thus the highly polar phytochemical compounds such as polyphenolics www.nature.com/scientificreports www.nature.com/scientificreports/ and some alkaloids are extracted from seeds via water thus it is contained many conjugated aromatic species and double bonds. Therefore, the bond which is appeared with a maximum around 206 nm is originated from the benzoyl system and π → π* transitions due to the existence of polyphenolics [57][58][59] .
FT-IR spectroscopy for the extract was also carried out to appear the possible bio-compounds with reducing ability to reduce the Ag + ions and to stabilize the constructed AgNPs by utilized herbal extract. As can be observed in Fig. 4(b), there are some transmission peaks at 3440, 2926, 2338, 1600, 1360, 1050, 790, 588 cm −1 . The band in 3600 to 3400 cm −1 is an indicator of stretching in OH group within free hydroxyl or involved in hydrogen bonds and the broad bonds z in range of 3500 to 3200 cm −1 is also an indicator of Ar-OH or -OH group bond in aromatic structures which confirm the existence of phenolic compounds in P. major extract. On the other hand the band in 3440 cm −1 can be a representative of stretching in N-H group. The band in 2926 cm −1 is appeared in the indicator range of carboxylic acids. The band in 1600 cm −1 is another representative range of carboxylic acids (1610 to 1550 cm −1 ) and the band in 1600 cm −1 can be also a typical band of conjugated dienes. The band in 1360 cm −1 can show the S=O band or N-O group. The observed bond in 1050 cm −1 shows the alcohols in the www.nature.com/scientificreports www.nature.com/scientificreports/ utilized herbal extract and the band in 790 cm −1 illustrates "out-of-plane stretching" in C=C or out-of-plane stretching of C-H in aromatics or S-OR within ester compounds. The band in 588 cm −1 are appeared around 600 cm −1 which representative of bending in C=C bond of alkenes or stretching in alkyl halides. Based on results, the flavonoid and other phenolic molecules in the P. major extract can act as reducing agent of Ag + ions and stabilization of the constructed NPs.
Stability and reproducibility of nanoparticles. The stability of the AgNPs (synthesized in optimum conditions) was checked by storing the colloidal dispersion of NPs in dark and at 4 °C and measuring its absorption spectra during time. After 30 days, the dispersion was stable and no considerable spectral changes and precipitation were detected which showed its stability in the investigated time period (data not shown). The synthesis of NPs was repeated more than 10 times in optimum conditions and the characteristics were reproducible. characterization of Agnps. XRD analysis. The XRD of biosynthesized AgNPs was recorded at 2θ values ranging from 20°-80° to clarify the crystal structure of NPs which is represented in Fig. 5(a). XRD analysis showed distinct peaks at 2θ equal to 38.1°, 44.2° and 64.5° and 77.4°, which indicated the planes (1 1 1), (2 0 0), (2 2 0) and (3 1 1) of the faced center cubic (fcc) silver respectively which determines the crystalline nature of the prepared AgNPs using P. major extract 60 . Thus, the obtained diffraction pattern showed that the most abundant crystals of the prepared NPs are in (111) plane. The average crystalline size was also calculated based on Scherrer's equation which was equal to 11 nm 61 .
TEM and SEM. The size and morphology of the prepared AgNPs was investigated utilizing TEM image which can be observed in Fig. 5(b). TEM images show relatively uniform particles in diameter with almost spherical shape and well-distributed particles. An estimation was also done on the particle size of NPs based on different recorded images and the range of size was found between 10 nm to 39 nm with an average size of 23.5 nm. A SEM micrograph of the dry mass of synthesized AgNPs is shown in Fig. 5(c), which represents NPs with an average size of ~30 nm.
Dynamic Light Scattering. The size distribution of the obtained AgNPs in the optimum conditions was also checked with DLS as a standard particle size analyzer. As it is represented in Fig. 5(d) FTIR spectra of NPs. FTIR has been used as an important method to find the responsible functional groups in stabilizing and surface modifying of NPs 62 . In addition to the FTIR spectrum of P. major extract which was discussed in previous parts, the FTIR spectrum of biosynthesized AgNPs was recorded which is represented in www.nature.com/scientificreports www.nature.com/scientificreports/ Fig. 4(b). The FT-IR of AgNPs indicates three representative peaks at 3475, 1652 and 714 which has some changes in shape and position in compare with to the observed peaks in the P. major extract. This can be due to the interaction between silver(I) ions and active sites of phytochemical compounds during stabilizing and producing of AgNPs. Appearance and shift in the position of peaks at 3600-3200 cm −1 , 1610-1550 cm −1 and 714 cm −1 , indicated the presence of OH functional groups, carboxylic acid, C=C aliphatic and probably amine groups in the capping agents which act as the stabilizers of the synthesized AgNPs using P. major. The main possible adsorbed compounds on the AgNPs surface can be polyphenolic molecules probably by π-electrons interaction in the www.nature.com/scientificreports www.nature.com/scientificreports/ absence of common strong ligating agents. These results confirmed the reduction of Ag + ions to Ag° and capping the surface of produced NPs by the natural polyphenolic materials in P. major extract 61 ferric-Reducing/Antioxidant power. Because of bioactive compounds in the P. Major, the reducing ability of the P. major extract and biosynthesized AgNPs capped by extract was checked and the results are illustrated in Fig. 5(e). It is observed from this figure that, the reducing activity enhanced by increasing the amount of P. major extract or the produced biosynthesized AgNPs and a dose-dependent manner was observed in this study. The similar result also reported by other researchers and also by our research group in synthesized AgNPs using other herbal extracts 26,63 . The phytochemical constituent in the P. major extract, which is effective in the bioreduction of Ag + to Ag 0 , can be appeared by the FRAP assay. The FRAP values were obtained in range of 303 to 1540 mM (as Fe (II)) dependent on different diluted sample of P. major extract (concentrations=134.75 ̶1078 μg/ml), which can be due to the high concentration of phenolic compounds and flavonoids in the extract. This can be also an evidence to show that the compounds in P. major extract were responsible in the reduction of Ag + and construction of AgNPs 64 . As it was observed in results, the synthesized AgNPs has also acceptable reducing ability which can be related to the polyphenolic compounds in the P. major extract encapsulated on the surface of NPs 65 .
CCD was done for optimizing the effective parameters in the synthesis of AgNPs. The hypothesis was obtaining highest quality of synthesis based on monitoring the SPR of the NPs and the response vector of CCD was not any one of biological activities. But, to evaluate the relationship between the SPR and anti-oxidant activity in FRAP assay, the FRAP assay was performed in some runs in Table 2 (i.e runs #4, #5, #15, #16, #18 and #20), with various SPR intensity from low to high. It was observed that the antioxidant power in the optimum run was bigger than all of these studied runs (data not shown). On the other hand among these six runs, run #20 had higher activity and run#5 has the lowest value which denote the effect of SPR on the anti-oxidant power in FRAP assay. Antimicrobial activity. The antibacterial effect of different strengths of AgNPs evaluated by disk diffusion method is represented in Table 4. Compared with E. coli, M. luteus was sensitive to lower concentrations of AgNPs which was confirmed by the MIC levels. The concentrations of 12.5% and 25% (v/v) were recorded as the MIC for M. luteus and E. coli, respectively. Although several studies have shown a remarkable activity of nano-silver against E. coli compared to the Gram-positive strain Staphylococcus aureus [66][67][68] , its effect is highly dependent on the species/strain variation in uptake, general susceptibility and tolerance mechanism 42 . In the study conducted by Kumar et al. 42 , M. luteus and Bacillus subtilis were found to be the most sensitive strains to biosynthesized AgNPs using Couroupita guianensis flower buds extract, while S. aureus exhibited much more resistance to AgNPs than E. coli. As shown in Fig. 6, AgNPs also revealed antagonistic activity against the indicator mold P. digitatum at all the concentrations tested with inhibition zone diameters of 15.5-16.5 mm. The aqueous extract of P. major was ineffective on all test organisms. The inefficacy of aqueous extract of P. major against E. coli has been reported by other researchers as well 69,70 . Therefore, the inhibitory effects found in this study can most likely be attributed to AgNPs. There are several reports on the antimicrobial activity of AgNPs synthesized using plant extracts against pathogenic bacteria and fungi [66][67][68][71][72][73] . The antimicrobial effects of AgNPs can be at least partly due to their ability to bind to bacterial cell membrane which may in turn lead to the increased membrane permeability 66 . Moreover, they may disrupt enzymatic activity of bacteria by interacting with the sulfhydryl group (-SH) of bacterial enzymes 71 .
To evaluate the relationship between the SPR and antibacterial activity, the activity against M. luteus were evaluated for some of runs in Table 2, having various SPR intensities (form low to high). It was observed that higher SPR can cause higher inhibition zone. For example the inhibition zone for run#4 (I.Z. = 10 mm), run#5 (I.Z. = 0 mm), run#15 (I.Z. = 0 mm), run#16 (I.Z. = 9 mm), run # mm), run #20 (I.Z. = 14 mm) confirmed this point. It was also observed that the antibacterial activity in the optimum run was bigger than all of these studied runs (see Table 4 for comparison).

conclusion
CCD as a known RSM can be used to optimize the experimental conditions of the biosynthesis process of AgNPs using herbal extracts. In this approach the interaction of factors and also the original factors (temperature, pH, volume of extract used for synthesis, volume of AgNO 3 solution and synthesis time) can be covered during NPs preparation/optimization simultaneously and the intensity of SPR spectra of the AgNPs was used as the  www.nature.com/scientificreports www.nature.com/scientificreports/ response (dependent variable). A linear model with R 2 = 0.825 with statistical significance in ANOVA as well as not-significant lack of fit showed the efficiency of CCD in biosynthesis of AgNPs. In optimum conditions, (Temp = 55 °C, pH = 9.9, volume of P. major extract=1.5 mL, volume of silver salt=30 mL, time of synthesis= 60 min), stable and high quality AgNPs were prepared. It is worthy to note that involving the interaction between experimental factors in the obtained model indicates the importance of multi-parameter optimization and its potential in compare with one-at-a-time optimization in nano synthesis systems as complex media.
The biosynthesis of AgNPs using extract of P. major seeds can be an effective method in pH ~10. FTIR showed that the main possible adsorbed compounds on the AgNPs' surface can be polyphenolic moleculesprobably by π-electrons interaction without common strong ligating agents. Because of bioactive compounds in the P. major, the reducing power of the P. major extract and biosynthesized AgNPs was evaluated using FRAP assay. The FRAP values indicate a dose dependent response and the reducing activity was increased in higher concentration of P. major seed extract or final AgNPs which shows its possible anti-oxidant ability.
DLS confirmed that 10.0% of the prepared AgNPs using this method was of the sized of 22.4 nm, 50% of the size of 26.1 nm and 90% of the size of 32.3 nm, which was in agreement with the SEM and TEM studies. XRD also confirmed the potential of the P. major seeds extract in the reduction of Ag + and preparation of Ag° NPs with average crystalline size around ~11 nm.
According to the best of our knowledge this is also the first work on the evaluation of antibacterial and antifungal activity of AgNPs synthesized by P. major seeds extract which showed an acceptable activity against Micrococcus luteus, Escherichia coli, and Penicillium digitatum. However the antibacterial activity was not tested against other kinds of bacteria and fungi which is a limitation of this work and should be done in future studies.
Since green synthesis of AgNPs is cost-effective, environmental friendly and compatible with human health due to less waste and safer products, this method can be considered as an alternative method of synthesis. Because of safe process of synthesis and the confirmed and antioxidant, antibacterial and antifungal activities of AgNPs synthesized by Plantago seeds,, the product can be used in medical applications such as surgical gowns, face masks and arms/ knees braces in the future studies. However various papers have reported the antibacterial, antifungal and antioxidant activity of the AgNPs, but the difference of the herbal extract which is used in the current work is popularity and availability of Plantago major in society and for people who are familiar with traditional herbals. It can make it a power point for general applications in society especially as antimicrobial agents. Thus, not only the elimination of chemical agents with toxic properties but also, covering both reducing and stabilizing agents by Plantago major, as a popular plant with known biological activity, can introduce it as a reliable method. On the other hand the production of AgNPs using the suggested method does not need high temperature and can be completed within one hour.