Water samples were obtained from the El-Mahmoudia Canal, the surface layer of water (upper 20 cm), Alexandria, Egypt (31°12′30.07″ N, 92°85′66.44″ E). The water samples were collected using sterilized bottles and promptly taken to the laboratory. A portable pH/temperature meter (Milwaukee MW102) was used to measure the temperature and pH in the field. The water turbidity and depth were also assessed using a Secchi disc with a diameter of 35 cm. Salinity was determined using a portable conductivity meter (Oakton, Eutech Instruments, Vernon Hills, IL, USA). Under control temperature conditions (25 ± 1°C) and continuous illumination (120 μmol photons m⁻² s⁻¹), selected water samples were inoculated and purified with the agar medium method using BG11 culture medium, as previously described by Robert (2005). After 2 weeks of maintenance, healthy colonies were transferred and inoculated into sterilized test tubes using the serial dilution method (Robert, 2005), with a total culture volume of 10 ml. Thereafter, the culture volume was scaled to 100 ml in 250-ml conical flasks for further subculture. Subsequently, the culture was upscaled in a volume of 20 L. Morphological examination of the isolate was first performed using a light microscope (Olympus BX51 Light Microscope, Tokyo, Japan) following standard identification references (Zaki et al., 2021). Finally, the results obtained from the morphological examination of the isolates were validated using molecular techniques.

The standard CTAB protocol was used to extract the entire genomic DNA from a dried cyanobacterial pellet according to Grube et al. (1995), with minor modifications as described by Elshobary et al. (2015). The universal cyanobacterial primers CYA106F (5′-CGGACGGGTGAGTAACGCGTGA-3′) and CYA781R (5′-GACTACAGGGGTATCTAATCCCWTT-3′) were used to amplify the 16S rRNA gene with the following PCR conditions: initial denaturation at 94°C for 5 min, 35 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min, followed by a final extension at 72°C for 7 min. The PCR product was then purified and sequenced using Sanger sequencing. Raw sequence data were edited and assembled using BioEdit software v7.2.5. The edited sequence was analyzed using BLAST against the NCBI nucleotide database for species identification. Multiple sequence alignment with retrieved reference sequences was performed using MEGA 11 software with the ClustalW algorithm. The sequence was then deposited in GenBank under the accession number OQ147028. For phylogenetic analysis, the obtained 16S rRNA sequences were aligned with 18 reference cyanobacterial strains obtained from the NCBI Ribosomal DNA database using Aulosira sp. as an outgroup to root the tree and provide evolutionary context. The neighbor-joining method was used to construct the phylogenetic tree.

Culture volumes of 10 ml were subjected to centrifugation for 10 min at 4,000 × g, with the resulting pellet kept at −20°C for biochemical analysis after discarding the supernatant. Biochemical analysis of the microalgal biomass measurement was according to AOAC (2003).

The PLs (1.68 ± 0.05 g) of L. vannamei were purchased from a private hatchery and transferred to NIOF laboratories, where they were adapted to laboratory conditions during a period of 15 days. During this acclimatization period, PLs were fed a commercial shrimp basal diet four times a day.

A total of 600 PLs were used, which were divided into four treatment groups (with three replicates per group). After a 15-day acclimation period, each treatment group received 50 PLs. The PLs were placed in net hapas measuring 0.7 m × 0.7 m × 1 m. These hapas were secured in concrete ponds measuring 4 m × 2 m × 1 m. Throughout the experimental period, the PLs were maintained under the recommended conditions for whiteleg shrimp PLs as reported by APHA (2005). The hapas are cleaned regularly, and the water daily exchange was 10%.

During the 8-week feeding experiment, shrimp PLs were fed the following four dietary groups: T₀, a commercial shrimp diet as the control diet (containing 45% protein, 7.9% lipid, 34% carbohydrate, 3.65% fiber, and 9.1% ash; Aller-Aqua, Giza Governorate, Egypt), and three other diets (T₁, T₂, and T₃) that were supplemented with 1, 2.5, and 5 g, respectively, of the dried powder of the microalgal isolate. The dried microalgal isolate levels were added to a specific diet according to the method previously described by Sharawy et al. (2022). In brief, four equal sections of the commercial diet, which served as the control, were crushed into a fine powder. The appropriate amount of cyanobacterial isolate was then added to each section, and the mixture was carefully mixed until homogeneous. In accordance with Sharawy et al. (2022), after dissolving the appropriate dose of cyanobacterial isolate in distilled water, it was sprayed onto the surface of the diet of the related section. An equivalent volume of distilled water without the cyanobacterial isolate was administered to the control diet (T₀). To cover the mixture of diet and cyanobacterial isolate, a 5-ml/kg diet of oil (sunflower oil) was sprayed over the basal diet after drying all diets for 48 h at 40°C to maintain a moisture content of approximately 10% (Zeraatpisheh et al., 2018). Lastly, 4°C was used to maintain the diet pellets until use.

The initial weight (IW, in grams) and the final weight (FW, in grams) of PLs were measured to determine the weight gain (WG), the specific growth rate (SGR), and the feed conversion ratio (FCR). The survival rate (SR, in percent) and the protein efficiency ratio (PER) were calculated using the following equations:

From each replicate, seven PLs were collected to analyze the shrimp’s whole body chemical composition. Random shrimp samples were mixed, dried, crushed into a powder, and stored at a low temperature (−20°C) until analysis. The chemical body composition was analyzed following AOAC (2003).

From each replicate, seven PL samples were randomly chosen after a 24-h fasting period. The samples were briefly washed with sterile seawater and then cut, weighed, and frozen until analysis at −80°C. The shrimp samples were homogenized [at a pH of 7.4 in phosphate-buffered saline (PBS)] and then centrifuged (20 min at 3,000 × g). Measurement of the homogenate lysozyme activity was carried out using Lysozyme (LZM) ELISA kits following the instructions given by the manufacturer (cat no. SL0050FI; SunLong Biotech Co., Ltd., Hangzhou, China). The antioxidant enzyme activities, including malondialdehyde (MDA), catalase (CAT), and superoxide dismutase (SOD), were measured using specific kits at wavelengths of 510, 534, and 560 nm, respectively, based on the instructions given by the manufacturer (cat nos. MD2529, SD2521, and CA2517, respectively; Biodiagnostic Company, Cairo, Egypt). For digestive enzyme (amylase and lipase) activity determination, GIT-homogenized tissues were separated by careful centrifugation. The digestive enzyme activities were measured using spectrophotometric assays at wavelengths of 580 and 660 nm.

Three shrimp samples from each group were taken at the end of the experiments. Total RNA was isolated from muscle tissue using Easy-RED (Easy-RED, iNtRON, Seongnam, South Korea) according to the manufacturer’s instructions. The total RNA concentration and purity were measured using NanoDrop (Nanophotometer, NP80 touch, Implen, München, Germany). The RNA products were used for the synthesis of cDNA using a commercial kit (Thermo Scientific™ RevertAid First Strand cDNA Synthesis Kit; Thermo Fisher Scientific, Waltham, MA, USA). The cDNAs were amplified through Rotor-Gene Q thermal cycling (Hassan et al., 2023; Wang et al., 2008; Jian et al., 2013) to quantify the expression of some growth- and immunity-related genes, including growth hormone (GH), insulin-like growth factor I (IGF-I), insulin-like growth factor II (IGF-II), prophenoloxidase (proPO), superoxide dismutase (SOD), and lysozyme (Lys). The quantitative PCR (qPCR) reactions were performed in 20 μl including 10 μl of ABT 2× qPCR Mix Kit (SYBR Green/low ROX), 0.5 μl of each primer (10 μM), 4 μl (50 ng) of cDNA, and 5 μl of RNAse-free water. The reaction followed the thermal profile: initial denaturation at 95°C for 10 min, followed by 40 cycles, each of which included denaturation for 10 s at 95°C, annealing at 58–60°C for 10 s, and extension for 30 s at 72°C. Subsequently, the temperature was increased by 0.5°C from 60°C to 95°C to establish a melting curve, which was then utilized to analyze the target gene products. In these reactions, the β-actin gene was included as a housekeeping gene (Hassan et al., 2023; Wang et al., 2008; Jian et al., 2013). The primer sequences and amplicon sizes are provided in Table 1 . According to the 2^{−ΔΔ C t} method, the β-actin gene was used to normalize the C _t values of the target genes (Rao et al., 2013). Statistical analysis was performed using GraphPad Prism software ver. 8.0.1 (GraphPad Prism, La Jolla, CA, USA). The results are presented as mean ± SD, with the level of significance considered at a probability value less than 0.05 (p < 0.05).

Figure 1 shows the morphological observations of the cyanobacterial isolate D. tharense NIOF17/006 using a light microscope (×400). The optical microscopic observations found that D. tharense NIOF17/006 is filamentous with a trichome width ranging from 2.2 to 3.6 μm.

Based on the recent taxonomic literature by Komárek (2005), the phenotypic characteristics of D. tharense NIOF17 classified it within the order Oscillatoriales. The common phenotypic features include a thin, pale to bright blue-green thallus. The filaments are either solitary or densely entangled, varying in length. This strain exhibited motility through gliding and oscillation movements, with tapered ends. The sheath surrounding the trichome is thin, colorless, and attached (agglutinated). The cylindrical cells have a uniform cell content, while the apical cells are long-conical with a rounded apex. The shape of the strains shows variability. Similar observations have been previously reported by Dadheech et al. (2012a), who isolated four strains of the cyanobacteria D. tharense: PD2001/TDC7, PD2001/TDC4, PD2001/TDC17T, and PD2001/TDC14.

A phylogenetic tree ( Figure 2 ) of the 16S rRNA gene was constructed using PCR-based nucleotide sequences that spanned more than 1,363 bp to provide a substantial amount of genetic information, allowing for a high-resolution analysis. This approach aligns with best practices in molecular phylogenetics, where longer sequences generally yield more reliable results (Patwardhan et al., 2014).

The neighbor-joining method was used for the construction of the phylogenetic tree, which is known for its efficiency and accuracy in depicting evolutionary relationships (Trees, 1987). The obtained 16S gene sequences were aligned with the 16S sequences of 18 cyanobacterial strains from the GenBank Ribosomal DNA database to ensure a comprehensive comparison, while the use of Aulosira sp. as an outgroup helps to root the tree and provide evolutionary context (Huelsenbeck et al., 2002). This methodological approach strengthens the reliability of the phylogenetic reconstruction. The resulting phylogenetic tree ( Figure 2 ) showed that each species formed a distinct clade, and the isolated strain was identified as D. tharense, which strongly matches D. tharense FJ158994, MK424816, and MW411006 with high similarity values of 98.53%, 98.46%, and 98.01%, respectively. It is worth noting that while 16S rDNA analysis is highly informative, complementary approaches such as whole-genome sequencing or multilocus sequence typing could provide even more comprehensive insights into the genetic makeup and evolutionary history of a strain (Huelsenbeck et al., 2002).

In the current study, the biochemical composition of the isolate D. tharense from cyanobacteria NIOF17/006 was examined, focusing on its protein, lipid, and carbohydrate contents. Figure 3 shows the biochemical composition of the cyanobacterial isolate D. tharense NIOF17/006 at the end of the late exponential growth phase. The average values for total protein, total lipid, and total carbohydrate were 37.74% ± 1.2%, 5.52% ± 0.75%, and 21.25% ± 0.87% (on a dry weight basis), respectively. Comparing these findings to the previous work by Hernández–Martínez et al. (2023), it is evident that D. tharense exhibits notable variations in its biochemical composition.

Table 2 illustrates the growth, survival, and nutrient utilization parameters of shrimp fed diets supplemented with the cyanobacterial strain. In this feeding trial, significant improvements (p < 0.05) were observed in the FW, WG, FCR, and SR between the control group (T₀) and the groups supplemented with cyanobacteria (T₁, T₂, and T₃). This improvement increased with increasing levels of cyanobacterial supplementation. However, no significant improvements (p < 0.05) were observed in other parameters of growth and nutrient utilization, i.e., SGR and PER.

Polynomial regression is a basic model in statistical machine learning (ML). It enables researchers to model how the predictor parameters and the outcome variables are related. Polynomial regression is an extension of a standard linear regression model. The nonlinear relationship between a predictor and an outcome variable is modeled using polynomial regression. Nonlinear relationships can be accurately modeled using polynomial regression (Maulud and Abdulazeez, 2020). Figure 4 shows the polynomial regression model of the dietary supplementation levels (in grams) of the cyanobacterial isolate D. tharense NIOF17/006 and the WG and FCR of the whiteleg shrimp L. vannamei. Figure 4 illustrates that, with the increase of D. tharense NIOF17/006 supplementation in the diet, the WG polynomial regression increased (r ² = 0.814) while the FCR polynomial regression decreased (r ² = 0.994). In conclusion, based on Figure 4 , the explanation lines of the highest and the lowest peaks of WG and FCR, respectively, presented in the polynomial regression model of ML predict that the ideal supplementation level of the probiotic cyanobacteria D. tharense NIOF17/006 ranges from 3.4 to 4.2 g/kg diet. This ideal range exists between groups T₂ and T₃ (2.5 and 5 g/kg, respectively).

In the aquaculture domain, El-Khodary et al. (2021) conducted a study on the growth, survival, and pigmentation of the larvae of Solea aegyptiaca using different microalgal species, underscoring the importance of microalgae in fish nutrition and development. The same findings were observed by Sharawy et al. (2022), who utilized Tetraselmis suecica as an aquafeed additive for L. vannamei. To the best of our knowledge, this is the first work to report on the potential application of the cyanobacterial isolate D. tharense NIOF17/006 as an aquafeed additive for L. vannamei. In the current study, the improvements in shrimp WG, FW, FCR, and SR may be due to the nutrient composition of the probiotic cyanobacterial isolate D. tharense NIOF17/006. However, it is commonly known that several cyanobacterial strains, particularly A. platensis, have high concentrations of multiple biologically active components that support the growth performance, immunity, and antioxidant capabilities of a variety of aquatic animals, such as the shrimp L. vannamei (Sharawy et al., 2022), gilthead sea bream (Galafat et al., 2022), Nile tilapia (Mabrouk et al., 2022), hybrid red tilapia (El-Sheekh et al., 2014), and common carp (Suantika et al., 2016).

Table 3 illustrates the biochemical composition analysis of shrimp fed the diet supplemented with different levels of the cyanobacteria-isolated strain D. tharense NIOF17/006. As presented in the table, significant differences (p < 0.05) were found in the total composition (protein, fat, dry matter, and ash) of L. vannamei in the supplementation groups (T₁, T₂, and T₃) compared with the control group (T₀). The percentages of carcasses increased with increasing supplementation doses. These improvements may be due to the high contents of protein (37.74%), lipid (5.52%), and carbohydrate (21.25%) of the strain of D. tharense isolated from cyanobacteria NIOF17/006. This is the first time this strain was used as an aquafeed additive for the whiteleg shrimp L. vannamei. The data shown in Table 3 are consistent with the findings by several authors who concluded that high supplementation levels of the cyanobacterial strain A. platensis significantly increased the shrimp carcass composition even in the whiteleg shrimp L. vannamei (Ashour et al., 2024) and the freshwater prawn Macrobrachium rosenbergii (Radhakrishnan et al., 2016).

The results showed that, compared with those in T₀, shrimp in the cyanobacteria supplementation groups (T₁, T₂, and T₃) showed the highest significant (p < 0.05) Lys, SOD, and CAT values. On the other hand, these groups exhibited the lowest significant (p < 0.05) MDA value compared with T₀. The shrimp immune system primarily relies on nonspecific processes due to its lack of adaptive immunity (Farzanfar, 2006). The enzymes Lys and SOD break down the cell walls of harmful bacteria and destroy free radicals in cells (Naiel et al., 2021). MDA is commonly used as an indicator of oxidative stress as it also shows an increase in the generation of free radicals (Chen et al., 2016). These shrimp humoral substances, e.g., Lys, SOD, MDA, and CAT, play essential roles in both specific and nonspecific immunity (Ashour et al., 2024). The findings of the current work aligned with those of the study by Ashour et al. (2024), who concluded that the nonspecific immunity and antioxidant activities of L. vannamei significantly improved when fed diets containing different levels of the cyanobacteria-isolated strain A. platensis. Furthermore, several works have reported that the use of different algal strains significantly improved the nonspecific immunity and antioxidant activities of aquatic animals (Mansour et al., 2022c). According to Jerez-Cepa and Ruiz-Jarabo (2021), immune-related factors have been studied as predicted indications for the examination of shrimp health. Figure 5 illustrates the immunological responses and antioxidant activities of L. vannamei fed diets containing different levels of the cyanobacterial strain D. tharense NIOF17/006.

Figure 6 shows the digestive enzyme activity of L. vannamei fed diets supplemented with different doses of the cyanobacterial strain D. tharense NIOF17/006. The results revealed that shrimp in the cyanobacteria supplementation groups (T₁, T₂, and T₃) exhibited the highest significant (p < 0.05) amylase and lipase activities compared with those fed T₀. The highest significant (p < 0.05) amylase and lipase values were observed in shrimp reared in groups T₃ and T₂, respectively. These current findings may be attributed to the fact that the bioactive materials of D. tharense NIOF17/006 may improve the secretion of digestive enzymes and, subsequently, improve feed absorption and digestibility (Xu et al., 2018). Moreover, it could be explained by the fact that shrimp fed diets containing higher levels of D. tharense NIOF17/006 might stimulate recycling, which is believed to be the outcome of the movement of fluid in the midgut lumen and the compartmentalization caused by the peritrophic membrane (Alexandre et al., 2014). Furthermore, according to this theory, if aquafeed supplementation increases the amount of protein in the diet, the corresponding digestive enzymes will be displaced, which ultimately results in a greater recovery of these types of enzymes in the feces (Ozorio et al., 2015).

The results of the gene expression analysis showed that the expression of the growth-related genes (GH, IGF-I, and IGF-II) increased almost significantly with increasing levels of supplementation with the cyanobacterial strain D. tharense NIOF17/006 in the formulated diets, except for T₂, which had a lower expression compared with T₁ ( Figures 7A–C ). The highest expression in shrimp was observed in group T₃ (5 g/kg diet). Similarly, the expression of the immune-related genes (proPO, SOD, and Lys) showed the same pattern and was upregulated in the muscles of L. vannamei with increasing concentrations of the cyanobacterial strain D. tharense NIOF17/006 in the T₁ and T₃ formulated diets ( Figures 7D–F ). With regard to the growth-related genes, GH is synthesized in the shrimp’s pituitary gland and exerts its action on the target cell to stimulate shrimp growth (Chandhini et al., 2021). IGF-I is a peptide hormone that is essential for the control of growth and development in shrimp (Pang et al., 2021). IGF-II is another protein that belongs to the IGF family. It plays a role in fetal growth and development and participates in tissue repair and replacement in adult shrimp (Li et al., 2024). With regard to the immunity-related genes, ProPO is a protein and an enzyme that participates in the defense system of shrimp.

When shrimp encounter pathogens, proPO is converted into the active form and catalyzes the melanization process. Melanization assists in pulling into the body these foreign microbes and destroys them, hence acting as a barrier against diseases (Amparyup et al., 2013). SOD is one of the antioxidant enzymes, the role of which is to protect shrimp cells against damage. SOD sustains the proper work of cells and shields the shrimp body from free radicals by transforming superoxide radicals into less harmful molecules (Chirawithayaboon et al., 2020). Finally, Lys is an antimicrobial enzyme that comprise an essential part of the shrimp immune system defense due to its content in the hemolymph. It is a versatile enzyme that is capable of attacking a wide variety of bacteria, hence offering an indication of protection against bacterial infections (Ferraboschi et al., 2021). According to the literature, there are no previous studies on the potential application of the cyanobacterial strain D. tharense NIOF17/006 as a feed additive for the shrimp L. vannamei or as an aquafeed additive in general. However, the findings of the current study have been previously confirmed by several studies reporting that the dietary inclusion of micro/macroalgae positively improved the growth and immune gene expression of L. vannamei (Ashour et al., 2024), the rainbow trout Oncorhynchus mykiss (Ratti et al., 2023), and the Nile tilapia Oreochromis niloticus (Mabrouk et al., 2024).