Nano-anatase TiO₂ was prepared from a commercial TiO₂ nanopowder (Sigma-Aldrich, USA). Titanium (IV) oxide have a purity of 99.7%, their specific surface area is 45–55 m²/g and their formula weight is 79.87 g/mol. X-ray diffraction (XRD) was performed using a Philips X-ray diffractometer with Cu Kα radiation (λ = 0.15406 nm). The crystalline structure of the samples was analyzed using a PAN analytical X’PERT PRO model X-ray diffractometer, on the instrument operating at a voltage of 50 kV and a current of 30 mA. Fourier Transform Infrared spectra were recorded under identical conditions in the 400–4000 cm⁻¹ region using a Fourier Transform Infrared Spectrometer (Shimadzhu). The morphology distribution was characterized using SEM (JEOL JSM 500-F).

Seeds of fenugreek (Trigonella foenum-graecum L.) were purchased from the Space green company, Tunisia. They were disinfected with 2% of sodium hypochlorite for 10 min and rinsed thoroughly to remove the disinfectant and soaked in distilled water at 4 °C for 24 h to obtain an initial stage [29]. The same number of seeds (20 seeds per 9 cm Petri dishes) was germinated in the dark at 24–26°C over two sheets of filter paper imbibed with 10 mL of distilled water for three days. Fenugreek seedlings were then transferred in pots containing Hoagland's nutrient solution [30] containing or not two concentrations of TiO₂NPs (50 or 100 mg·L⁻¹). TiO₂NP preparations were dispersed by ultrasonic treatment for 60 min and were maintained in the dark. Growth conditions were performed at 25 °C, with relative humidity of 70% and a light–dark cycle of 16:8 h. Nutrient solutions were changed every four days during the experiment. In the greenhouse stage, an experiment was conducted with five replicas, and each pot contained forty seeds.

The TiO₂NP and control seedlings were thoroughly washed with distilled water to remove nanoparticles and nutrients on the surface every four days up to 16 days. The seedlings were then used for growth analyses. Following the harvest of plants, fresh weight (FW) was immediately measured using Electronic Balance and was dried in a hot air oven at 65 °C for 72 h for dry weight (DW) determination. The remaining seedlings were then divided into leaves and stem individually and kept at –80 °C in an ultra-deep freezer for measurements of total soluble protein, pigment contents, antioxidative enzyme activities (GPOX, POX and CAT), and lipid peroxidation (MDA content).

For the determination of photosynthetic pigments (chlorophyll a, b, and carotenoid levels), 100 mg of fresh leaves from control and treated seedlings were ground with mortar in 5 mL of 80% (v/v) ice-cooled acetone. The pigment extract was measured against a blank of 80% (V/V) acetone at wavelengths of 647 and 663 nm for chlorophyll assays and at wavelengths of 470 nm for carotenoids. The determination of the chlorophyll a, b and carotenoid contents level was carried out using the equations proposed by Lichtenthaler [31] and expressed as mg·g⁻¹ FW.

Chla(μg·mL⁻¹) = 12.21 A₆₆₃–2.81 A₆₄₆

Chlb (μg·mL⁻¹) = 20.13 A₆₄₆  −  5.03 A₆₆₃

Car (μg·mL⁻¹) = (1000 A₄₇₀–3.27 Chla–104 Chlb)/229

Anthocyanin samples were analyzed according to Gould et al.’s method [32] Frozen tissues were soaked immediately in acidified methanol (methanol:water:HCl: 16:3:1), then crushed using a glass pestle and kept at 25 °C for 72 hours in the dark. The relative amount of anthocyanin was estimated spectrophotometrically at 530 nm and 653 nm. Determination of anthocyanins level was expressed as μg g⁻¹ FW. Finally, amounts of traits were calculated using the following formula: Anthocyanin (μg·mL⁻¹) = A₅₃₀–0.24 A₆₅₃.

After drying at 60 °C in an oven for 72 h, 5 g of powder were introduced into 100 mL of boiling water in a 250-mL Erlenmeyer flask. After 15 min, the sample was filtered, and the filtrate was adjusted to 100 mL with distilled water [33].

2.6.1 Determination of the total flavonoid content

2.6.2 Determination of the total phenolic content

The protein concentration was determined using the BioRad reagent and BSA as a standard [36]. The total soluble protein contents were expressed in mg/FW.

Fresh tissues were ground with a mortar in a homogenization medium (w/v = 1/3) (pH 7.5) consisting of Tris-HCl (50 mmol·L⁻¹), saccharose (0.4 mol·L⁻¹), and EDTA-Na₂ (5 mmol·L⁻¹). The homogenate was centrifuged at 800 g for 5 min. The resulting supernatant was centrifuged again at 1500 g for 10 min. The chloroplast pellet was suspended in the homogenization medium and kept at 4 °C.

2.8.1 Guaiacol peroxidase (EC 1.11.1.7) enzyme assay

2.8.2 Catalase (CAT, EC 1.11.1.6) enzyme assay

2.8.3 Ascorbate peroxidase (EC 1.11.1.11) enzyme assay

Lipid peroxidation was estimated by determining the malonyldialdehyde (MDA) [40]. One hundred milligrams of fresh samples were homogenized in 1 mL of 0.1% trichloroacetic acid (TCA). The homogenate was centrifuged at 15,000 g for 10 min at 4 °C. A 0.5-mL aliquot of the supernatant was mixed with 1.5 mL of 0.5% of thiobarbituric acid (TBA) prepared in TCA (20%), and incubated at 90 °C for 20 min. After stopping the reaction in an ice bath, samples were centrifuged at 10 000 g for 5 min. The absorbance of supernatant was then measured at 532 nm. After subtracting the non-specific absorbance at 600 nm, MDA concentration was determined using an extinction coefficient of 155 mM⁻¹·cm⁻¹.

The experimental data were analyzed using two-way ANOVA and differences were considered significant when they reached the 0.05 level of probability.

Fig. 1A presents the XRD patterns of the TiO₂ powder. The observed XRD peaks are in good agreement with the standard JCPDS file (84–1285), which confirmed an anatase structure at all peaks. Particle size was estimated according to Debye–Scherer's equation [41] as follows (Eq. (1)):

The FT–IR spectrum of anatase TiO₂ nanoparticles is shown in Fig. 1B. A broad absorption peak in the range of 3000–3500 cm⁻¹ is present in both figures, which can be attributed to the characteristic absorption of hydroxyl groups (O–H) [42]. The FT–IR spectra of TiO₂NPs exhibiting a strong absorption band between 800 and 450 cm⁻¹ remained, which was attributed to the formation of TiO₂ nanoparticles.

The surface morphology of the TiO₂NPs was studied using SEM. The SEM graph was shown in Fig. 1C for TiO₂NPs. It is clear from the SEM of the powder that the size of the particles is less than 20 nm. The SEM results revealed the presence of agglomerates of nanoparticles and showed that the morphology is replete with spheres.

The most important phonological stages are the vegetative growth period, due to its impact on the yields [43]. Various findings on biological impacts of TiO₂NPs especially on plant kingdom, have contradictory results, ranging from positive and negative or inconsequential effects [44]. Nanoscale TiO₂ proved no significant effects on stem length, stem dry weight (Table 1, Fig. 2). Our results agree with those of Haghighin and Teixeira da Silva [45], who stated that shoot length and shoot dry weight of greenhouse grown seedlings of onion were not significantly affected by TiO₂NPs for sizes from 8 to 15 nm. Also, the result corroborates the work, reported earlier, by Seeger et al. [46], who found no differences in the growth of willow trees in the range of 1–1000 mg·L⁻¹TiO₂NPs. The same work by Haghighin and Teixeira da Silva [45] proved thatTiO₂NPs decreased shoot length in tomato and increased it at most for 100 mg·L⁻¹ of radish, but the dry weight of radish increased in greenhouse. Indeed, Antisari et al. [9] reported that concentrations of 10 and 100 ppm of TiO₂NPs with a size of 20 nm enhanced the shoot of wheat. The fresh weight of the stem decreased approximately by 40% over 16 days in 100 mg·L⁻¹ of TiO₂NPs. The fresh weight of the stem was reduced by 40% for 100 mg·L⁻¹of TiO₂NPs compared to that of controls after 16 days of treatment (Table 1) [9,45]. There was no significant effect on fresh, dry weight of fenugreek leaves of greenhouse grown seedlings between treatments. Internodes increased by 89–100% over 16 days of TiO₂NPs treatment, although internodes were not significantly affected by NP-TiO₂ treatments compared to control after 16 days of treatment (Table 1).

It can be concluded that each plant is affected differently by nanoparticle titanium dioxide. Also, research proved that TiO₂NPs impacts on the crops depended on nanoparticle concentrations, size, timing, method of application. However, the morphology of nanoparticles also plays an important role in the phytotoxicity. One study reported a significant difference in toxicity to macrophages between two carbon nanotubes with distinct morphologies (MWCNTs and SWCNTs) [47]. Xiang et al. [48] reported that the phytotoxicity of the small ZnO-50 spheres and the large ZnO-90 columns were comparable, which does not support the concept of size-dependent toxicity, instead suggesting a morphology-dependent tendency. The differences in the morphology of nanoparticles may alter the sites and strength of the interaction between the particles and their biological targets, leading to a difference in toxicity.

Toxicity appears to be aggravated by the fibrous or filamentous form of the nanoparticles. Long particles such as nanotubes or nanofilaments would be more toxic than spherical particles of identical chemical composition. This effect is notable for particles with a length greater than 8 μM and a diameter less than 0.25 μM, irrespective of their chemical composition.

Notably, after 16 days of treatment, there were no differences between the water contents of stem and leaves of fenugreek exposed to TiO₂NPs compared to controls (Table 1). These results are consistent with the study of Seeger et al. [46], who have shown that TiO₂NPs have limited effects on willow tree cutting growth in terms of water usage and transpiration.

Interestingly, at 16 days of stress, higher concentrations of nanoscale TiO₂ decreased leaf area by 20% compared to controls (Table 1) and proved a chlorosis (Fig. 3). These results agree with Asli and Neumann [10], who observed that 30 and 1000 mg·L⁻¹ of TiO₂NPs with a size of 30 nm inhibited the growth of the leaf of Zea mays.

Photosynthetic pigments are very important metabolites and play a fundamental role in plant metabolism. The amount of Chl a increased about 16% and threefold due to treatment with 50 and 100 mg·L⁻¹ TiO₂NPs respectively after 16 days of treatment. However, the results show that Chl a did not change following the addition of 50 mg·L⁻¹ ofTiO₂NPs, but the application of 100 mg·L⁻¹ of TiO₂NPs made the Chl a level decrease by 50% compared to controls after 16 days of stress treatment (Fig. 4a), while the amount of Chl b increased by about 145 and 126% under 50 and 100 mg·L⁻¹ TiO₂NPs, respectively, for 16 days. The exposure of fenugreek to 50 mg·L⁻¹of NP-TiO₂ increased the amount of Chl b by about 36% compared to that of controls after 16 days of treatment. Moreover, 100 mg·L⁻¹ NP-TiO₂ significantly diminished by about 51% the amount of Chl b compared to control (Fig. 4b).

In greenhouse experiments, the levels of carotenoids increased by 45% after 16 days of developpement in control plants. The level of carotenoids in TiO₂-treatment plant is higherthan that of control. Nanoparticles of TiO₂ at concentrations of 50 mg·L⁻¹ increased by 38% the amount of carotenoids compared to the control, although this character declines by 36% after the application of 100 mg·L⁻¹ of TiO₂NPs (Fig. 4c).

These results are in accordance with those reported by Lei et al. [24] and Zheng et al. [50]; therefore, Gao et al. [19], in a study on the effects of different concentrations of nanoparticles titanium dioxide on the spinach trait, concluded that the chlorophyll amounts and the photosynthetic rate amount in treatments with NP-TiO₂ were significantly increased. The results concur with other studies that have shown that 10 and 30 ppm of TiO₂ NPs with size < 25 nm altered chlorophyll contents in Phaseolus vulgaris [13]. Also, Lei et al. [15] observed that 0.25% of NP-TiO₂ (colloidal) with a size of 5 nm damaged the chloroplast of the spinach plant, resulting in a reduction of the photosynthetic rate.

These results are in good agreement with others, which have shown that the physiological effects are probably attributed to the small size of the particles, which allows their penetration into the plant during exposure [50,51]. An important implication of these finding is that higher chlorophyll accumulation at a concentration of 50 mg·L⁻¹ TiO₂NPs may be due to complementary effects of other inherent nutrients like magnesium (Mg), iron (Fe) and sulfur (S) [52]. TiO₂NPs can improve the structure of chlorophyll, increase light absorbance, facilitate the formation of pigments, allow better capture of sunlight and transfer of light energy to active electrons, chemical activities, and have an effect on photosynthesis [52–55]. Leaf chlorosis became pronounced in plants treated with 100 mg·L⁻¹ of TiO₂NPs. In agreement with these symptoms and with the reduction in leaf area, significant decreases in Chl a, Chl b and carotenoids were observed under 100 mg·L⁻¹ TiO₂ NP stress. The observed decreasing chlorophyll and carotenoids contents may be due to an oxidative process, which is an indicator of tissue ageing, as a result of the stress factors in the environment [56]. At the same time, pigment loss is due to rising levels of toxic metals and metalloids [57]. It can be concluded from this experimental study that the observed chlorosis in leaves and reduction in leaf area were not due to a direct interaction of nanoparticles with the chlorophyll biosynthesis pathway, and that it was caused most probably by a decrease in chloroplast density [33].

Other authors invoked that the decrease in the chlorophyll contents explains that anatase enhances the activity of chlorophyllase [58]. On the other hand, the decreases in Chl a and Chl b at 100 mg·L⁻¹ were due to the damage exerted on the chloroplasts by TiO₂. Our results disagreed with those of Cao et al. [59] who stated that no damage was exerted on the chloroplasts by either type of CeO₂ NPs at the 100 and 500 mg kg⁻¹ concentrations when increasing chlorophyll a and decreasing chlorophyll b. In our results carotenoids content was less affected by TiO₂NPs than chlorophyll. These compounds have antioxidant activity as triplet chlorophyll and singlet oxygen quenchers [60].

Under environmental stresses, ROS overproduction can damage important biological molecules (lipids, protein, and DNA) in plants [61]. Plant species have many enzymatic and non-enzymatic antioxidant pathways that can be simultaneously activated to defend against oxidative damage [61,62]. Anthocyanin, a type of flavonoid located in vacuole system [63], is one of most common non-enzymatic antioxidants and operates as a superoxide radical scavenger, hydrogen donor, and metal chelator.

Exposure of fenugreek to TiO₂NPs during greenhouse experiments results in an increase in the anthocyanin content: about 2.9- and 1.29-fold for 50 and 100 mg·L⁻¹ of TiO₂NPs, respectively, after 16 days of treatment compared to controls (Fig. 5). These results are consistent with those of Kumaret al. [64]; they demonstrated that Pb at concentrations up to 1 mM significantly increased anthocyanin in Ceylon spinach (Talinum triangulare L.). The biosynthesis of anthocyanin in plants is catalyzed by phenylalanine ammonium-lyase (PAL) through a phenylpropanoid pathway [65]. It is possible that highly stressful conditions caused by 100 mg·L⁻¹ NPs may induce the production of high levels of H₂O₂, which could inhibit the PAL activity, and thus anthocyanin biosynthesis becomes somewhat disrupted.

Even after 16 days of treatment, there was a significant decrease in the concentration of flavonoids by about 60 and 78% for 50 and 100 mg·L⁻¹ NP-TiO₂, respectively, compared to the control stems. The flavonoid concentration in leaves of fenugreek did not change significantly following the addition of 50 mg·L⁻¹ TiO₂NPs, but the flavonoid content was reduced with 100 mg·L⁻¹of TiO₂NPs (Fig. 6a). These results disagree with those reported by Rabia et al. [66] in a study on the effect of zinc oxide (ZnO) nanoparticles on physiology and steviol glycosides production in the micropropagation of Stevia rebaudiana. Bertoni concluded that total flavonoid content was significantly increased by the treatment with NP-TiO₂.

Interestingly, after 16 days of stress, the phenol content in stems was improved by factors of approximately 2.8 and 1.45 by a treatment with 50 and 100 mg·L⁻¹TiO₂NPs, respectively, compared with the control. The phenol level in leaves enhanced significantly by factors of 2.8 and 1.2 after the application of 50 and 100 mg·L⁻¹TiO₂NPs respectively after 16 days of treatment (Fig. 6b).

Fig. 5(c and d) shows that translocation ratio of flavonoids declined significantly by 62 and 71% after treatment with 50 and 100 mg·L⁻¹ TiO₂NPs, respectively, after 16 days of treatment (Fig. 6c). Interestingly, in the present experiment, NP-TiO₂ stress had no apparent effects on the translocation ratio of polyphenols (Fig. 6d). The decrease in the translocation ratio of flavonoids can be explained by a chelating process. Flavonoids contain an enediol group, by which they act as bidentate ligands for anatase TiO₂ nanoparticles. If the chemical bond between flavonoids and TiO₂ nanoparticles is strong, desorption of the flavonoids from the anatase surface should be minimal, and thus the replacement of flavonoids with other bidentate ligands present in the vicinity of the nanoparticles (e.g. in a cellular environment) is expected to be negligible. Therefore, anatase TiO₂ nanoparticles are predicted to be an efficient platform for the isolation of flavonoids [67].

Transformation or complexation is a critical factor that affects the fate and toxicity of NPs in living organisms. After absorption by roots, plants can transform NPs into other forms [68]. For example, nano-anatase at size 2.8 nm conjugated with alizarin red S accumulate in plantlets of Arabidopsis thaliana grown in hydroponics [69]. These nanoconjugates can pass through cell walls and distribute in distinct subcellular compartments. Peng et al. [70] reported that copper oxide nanoparticles (CuO NPs) entered the stele through lateral roots in rice and translocated to leaves; Cu (II) was mainly combined with cysteine, citrate, and phosphate ligands and was even reduced to Cu (I). For maize, Wang et al. [71] reported that 100 mg·L⁻¹ CuO NPs at size 20–40 nm was translocated from roots to shoots via xylem and re-translocated from shoots to roots via phloem; CuO NPs, during this translocation, could be reduced from Cu (II) to Cu (I). Excess Cu is toxic to plants and caused reduction in growth and yield of rapeseed (Brassica napus L.) and wheat. Parsons et al. [72] observed NiNPs in roots and shoots of mesquite (Prosopis sp.) treated with uncoated nickel hydroxide nanoparticles (Ni(OH)₂NPs), whereas leaves had a Ni(II)-organic acid type complex.

The data presented in Fig. 6 show that the protein content increased approximately by 24 and 103% in control leaves and stem of fenugreek over 16 days. In leaves, the protein content increased by factors of 1.26 and 4 over 16 days with 50 and 100 mg·L⁻¹ of TiO₂NPs. The soluble protein in TiO₂NP-treated chloroplast leaves was ∼ 22–64% lower than in controls after 16 days of exposure to 50 and 100 mg·L⁻¹ TiO₂NPs, respectively (Fig. 7a). However, in stems of fenugreek plants, the protein levels increased by factors of approximately 1.2 and 3.3 over 16 days with 50 and 100 mg·L⁻¹ TiO₂NPs. In addition, there was no impact on the soluble protein content in chloroplast following the exposure to TiO₂NPs compared with controls (Fig. 7b). This finding agrees with previous studies by Salama [73], who reported the effects of silver nanoparticles on the protein contents of common bean and corn. The improvement of the protein concentration depends on the optimum dose of nanoparticles of silver for the growth of common bean and corn plants. However, a decrease in the protein contents was observed; it was due to the toxic effect of AgNPs. The soluble protein contents depend also on the content of chlorophyll in the leaves of the plant [74].

The deleterious effects of ROS and lipid peroxidation products are counteracted by the antioxidant defense system [75]. Phenols, flavonoid and stress enzymes are part of the plants defense mechanisms.

GPOX activities increased approximately by a factor of 3 over 16 days in control leaves. They increased by factors of 1.5 and 1.9 compared to controls, due to the treatment with 50 and 100 mg·L⁻¹ of TiO₂NPs, respectively, after 16 days (Fig. 8a). CAT activities decreased by approximately 72% during 16 days in control leaves. The activity in TiO₂NP-treated leaves was 40–400% higher than in controls after 16 days of exposure for 50 and 100 mg·L⁻¹ of TiO₂NPs, respectively (Fig. 8b). APX activity in controls leaves of fenugreek decreased by 21% after 16 days and it was higher than 6.2 and 2.6-fold in TiO₂NPs treated leaves compared with normal conditions (Fig. 8c).

These results agree with other reports that have shown that nano-anatase treatment could activate SOD, CAT, APX and GPX in spinach chloroplasts. SOD converts O₂- into H₂O₂ and O₂; moreover, CAT, APX, and GPX can reduce H₂O₂ into H₂O and O₂. Therefore, SOD, CAT, APX, and GPX maintain a low level of ROS, prevent ROS toxicity, and protect cells [15]. GPOX, CAT and APX activities increased approx. 24 and 55% over 16 days in control stems of fenugreek, whereas GPOX and CAT activities did not change significantly following the addition of 50 mg·L⁻¹ of TiO₂NPs compared to controls after 16 days of treatment (Fig. 8d–f). Compared to the 100 mg·L⁻¹ TiO₂NPs condition, GPOX activity decreased approximately by 11% after 16 days of exposure under control conditions (Fig. 8d). Similar finding noted that CAT and GPOX activities were enhanced in the presence of 10–30 μg·mL⁻¹ of TiO₂NPs, but that their activities decreased with higher concentrations of TiO₂NPs [76]. Interestingly, in the present experiment, NP-TiO₂ stress had no apparent effects on CAT activity (Fig. 8e). No changes in CAT activity in Spirodela polyrrhiza under TiO₂ nanoparticles has been previously reported [77]. In contrast, when incubated with NP-TiO₂, APX activity was by 95–252% higher than that of controls after 16 days of treatment for 50 and 100 mg·L⁻¹TiO₂NPs, respectively (Fig. 8f). Other works reported an increase in CAT activity in response to 250–750 mg kg⁻¹of TiO₂NPs, but a decrease in APX activity with 500 mg·kg⁻¹ in cucumber plants [14].

Lipid peroxidation measured by the accumulation of the MDA-TBA complex was also assayed in plant tissues (stems and leaves). As shown in Fig. 9, MDA concentration in stems increased significantly by factors of about 1.7 and 2 compared to controls due to the application of 50 and 100 mg·L⁻¹ of TiO₂NPs, respectively, after 16 days (Fig. 9a). MDA concentration in NP-TiO₂-treated leaves was higher by factors of 1.6 and 1.4 than that of controls after 16 days of exposure for 50 and 100 mg·L⁻¹ TiO₂NPs respectively (Fig. 9b). MDA activities significantly increased in nano-stressed leaves, suggesting higher membrane damage. Our results concurred with earlier studies that reported that CuONPs induced an increase in H₂O₂ and MDA concentrations in the leaves of barley [78], rice, and chickpea seedlings [79,80]. We suggest that fenugreek cells accumulated MDA after decomposition of polyunsaturated fatty acid. Stress induced a progressive enhance in H₂O₂ content and MDA accumulation indirectly due to oxidative stress. Hydroperoxides were used as markers for oxidative membrane damage [33]. The latter generally acts as an indicator of free radical production [81]. The accumulation of MDA was found to be higher in stems than in leaves. This may be related to the macroscopic signs of toxicity we observed, as membrane damage can induce apoptosis and premature senescence [33].