Introduction

Occurrence of antimicrobial resistance is a worldwide menace posing threat not only to human health, but also hampering agricultural production system by developing resistant pathogenic bacterial strains which are difficult to manage by using conventional antibiotics like tetracyclines (Tc). Tc are used widely for treatment of farm animal diseases. For example, respiratory disease (BRD) [1] and digital dermatitis [2] in bovine, enteric red mouth [3], vibriosis [4] and bacterial hemorrhagic septicemia [5] in fishes, etc. In the bacteria, Tc binds to the 30S ribosomal subunit and inhibits the protein synthesis by blocking the binding of aminoacylated tRNA to the A site [6]. However, recently there are many reports from the agricultural farms on development of bacterial resistance against Tc. For example, occurrence of Tc resistant Pasteurella multocida (83%) causing BRD [7], Anaplasma species causing acute anaplamosis in cattle and sheep [8], fish pathogenic bacteria in mariculture [9,10,11], etc.

Tc resistance in pathogenic bacteria is caused by introduction of Tc resistance genes producing proteins responsible for binding-site alteration/ribosomal protection, efflux pump, enzymatic degradation, etc. [12, 13]. However, in gram negative bacteria it is primarily due to active transport of drug out of the bacterial cell by intrinsic membrane transport protein TetA [14]. TetA protein acts as an antiporter across the membrane and facilitates the efflux of Tc divalent cation complex [MgTc]+ out of the cell. However, the expression of TetA is highly controlled by the Tet repressor (TetR) protein which is having very high affinity (KA ~109 M−1) toward Tc [15]. TetR also regulates transcription of its own gene tetR and upon binding to [MgTc]+ it changes its conformation that cannot regulates transcription of both tetR and tetA. As a result, TetR and the resistance protein TetA are expressed which exports the antibiotics out of cell before inhibition of protein synthesis. TetR is a homodimer α-helix-turn- α-helix sequence motif and folds in to 10 α helices (α1 to α10) with connecting turns and loops [15, 16]. Tc binding pocket of TetR is composed of COOH-termini of α4 and α6 and the helices α5, α7, α8, α8’, α9’ [16]. Hence it was hypothesized that a change in the conformation of Tc can alter the binding affinity toward TetR and also improves the inhibition of protein synthesis by binding the ribosome.

In the present study molecular conformation of a model Tc antibiotics i.e. oxytetracycline (OTc) was changed by introducing Zn2+ as divalent cation and further zinc oxide nanoparticles (nZnO) was used as carrier of the complex [ZnOTc]+ so that the it has lower binding affinity toward TetR and bypass the efflux mechanism. Previously, a few studies reported improvement of antibacterial properties of drug molecule in association with Zn2+ and ZnO. For example, a combination of OTc and ZnO loaded on to polycaprolactone nanofibers showed excellent activity against a mixed bacterial culture with 97.5% growth inhibition [17]. Similarly, a recent clinical study reported that a combination of Zn and Tc antibiotics showed better bacteriacidal effect toward multidrug-resistant strains of Acinetobacter baumannii isolated from pneumonia affected COVID-19 patients [18]. Use of nanoparticle as carrier for antibiotics have proven as better option to enhance the efficacy of antibiotics and reinvigorate obsolete antibiotics against resistant pathogen [19]. For example, Tc loaded on para-amino benzoic acid functionalized grapheme oxide nanosheets performed better against resistant Escherichia coli. [20]. Similarly, loading of Tc into calcium phosphate nanoparticles (8 ± 5 nm) had shown to reduce MIC from 150–180 µg ml−1 to 20–40 µg ml−1 against resistant strains of E. coli, Salmonella kentuckey, Shigella flexneri [21]. Covalently attached ciprofloxacin on Single-walled carbon nanotubes (SWNT) was also found to be 16 times more effective against Staphylococcus aureus and Pseudomonas aeruginosa, and eight times more effective in case of E. coli [22]. SWNT and nanographene oxide was used to deliver Tc to kill E. coli with an efflux pump resistance mechanism and found to reduce MIC value from 200 to 38–75 µg ml−1 [23].

In the present study the efficacy of [ZnOTc]+ complex loaded on nZnO (OTc@ZnO) was tested against a model gram negative bacterium i.e., Aeromonas hydrophila which is a highly prevalent fish pathogenic bacteria and recently, it showed widespread occurrence of resistance [24, 25]. For example, OTc resistance incidences observed in Aromonas spp. isolated from Sparus aurata reared in Italian mariculture farms [26]. Similarly, a study in Czech carp fisheries found that 41% of isolated Aeromonas sp. (n = 159) were resistant to OTc [27]. This is first kind of any study where interaction of OTc and Zn2+ as a complex [ZnOTc]+ was reported to show breaking of Tc resistance.

Materials and methods

Materials

Oxytetracycline hydrochloride (OTc. HCl) or 5-hydroxytetracycline hydrochloride (C22H24N2O9. HCl) (95%) was purchased from Sigma-Aldrich. Zinc sulfate heptahydrate (ZnSO4.7 H2O) and sodium hydroxide pellet (NaOH) was purchased from Himedia Laboratories Pvt. Ltd. India. For routine work, analytical grade chemicals and solvents were employed.

Synthesis of nZnO and OTC@nZnO

The nZnO was prepared as per previous method with slight modification [28]. Briefly, a solution of ZnSO4 (20 mM) was prepared with MiliQ water (18 Ω) followed by stirring (400 rpm) and drop wise addition of equal volume of NaOH solution (12 mM) to get white suspension (pH recorded at 10.0). The suspension was vigorously stirred for 10 min followed by centrifugation (4000 rpm) to get precipitates. The white precipitates were dried first at 100 °C for 6 h and then calcinations at 300 °C for 1 h followed by grinding to get fine powder.

The complex [ZnOTc]+ loaded on to nZnO (OTc@nZnO) was prepared by mixing ZnSO4 solution (20 mM) with OTc solution (10 mM) while stirring (400 rpm) followed by drop wise addition of NaOH (12 mM) to yield yellowish suspension (pH recorded at 8.0) at dark condition. The suspension was vigorously stirred for 10 min followed by centrifugation (4000 rpm) to get precipitates which was washed three times with MiliQ water. The final suspension was dried in room temperature (29 °C) followed by storing in desiccator at dark.

Characterization of nZnO and OTc@nZnO

The morphology of nZnO and OTC@nZnO were measured by high resolution transmission electron microscope (HRTEM) (JOEL, JEM-2010). The known quantity of dried nZnO and OTC@nZnO were dispersed in MiliQ water following sonication (at 80% amplitude, 25 kHz). The dispersed materials were drop casted on carbon coated copper grid and analysis was performed in the HRTEM with at an acceleration voltage of 200 kV.

To confirm the loading of OTc on to nZnO, Fourier transformed infra-red (FTIR) spectroscopic analysis was performed. The dried powders of nZnO, OTc and OTc@nZnO were analyzed in ATR FT-IR (Spectrum 100 FTIR Spectrometer, PerkinElmer) and percent absorbance was recorded at range 650–4000 cm−1 with a resolution of 1 cm−1 [29].

The attachment position of Zn2+ on to OTc molecule was identified by comparing absorbance (λmax) values of OTc and [ZnOTc]+ using UV-Vis spectrophotometer. The λmax values of OTC was elucidated using Woodward-fieser rule for aromatic carbonyl compounds (D ring) and 6-member cyclic α, β unsaturated ketones (A ring) [30, 31].

Loading and release kinetics

The percent loading of OTc on to nZnO was determined by adding 1 ml Conc. HCl (37%) to 0.01 g OTc@nZnO followed by volume make to 100 ml and measuring absorption at 270 nm using UV-Vis spectrophotometer. The rate of release of [ZnOTc]+ from the OTc@nZnO was determined using dialysis method [32]. A known amount of OTc@nZnO was dispersed in MiliQ water using sonication (at 80% amplitude, 25 kHz) for 2 min and was taken in a dialysis bag with 10 kDa membrane. The dialysis bag was dipped in a beaker containing 100 ml phosphate buffer. The set up was maintained for 2 days and at regular interval 2 ml of aliquot was withdrawn from the beaker and the amount of OTc released was estimated using UV-Vis spectrophotometer at 270 nm.

Bacterial strains and antibacterial activity

Bacterial species used in the study was A. hydrophila which are abundantly found in the tropical aquatic ecosystems. A. hydrophila ATCC 7966 strain was used as sensitive strain to OTC whereas resistant A. hydrophila strain CIAHRT_11 (NCBI, MG754418) was isolated from blood and tail of infected Labeo rohita fish collected from aquaculture ponds of West Bengal, India.

All the isolates were revived in Tryptic Soy Broth (TSB) (Himedia Laboratories, India) for 18–20 h and incubated at 37 °C. After incubation the bacterial isolates were diluted up to 0.5 McFarland Standards which contains ~1.5 × 108 cfu ml−1 of bacterial population. Using the L-shaped spreader, the diluted bacterial suspension was then spread over the Tryptic Soy Agar (TSA) (Himedia Laboratories, India) plate. The experimental samples were loaded in various concentrations on a blank sterile disc (Himedia Laboratories, India) and placed on the agar plate. A 30 µg OTc disc (Himedia Laboratories, India) was also placed in the middle of the plate. Those plates were incubated for 24 h at 37 °C. The inhibition zone of the samples which were loaded on the sterile disc and the OTc disc were measured and recorded after 24 h.

Minimal inhibitory concentration (MIC) and minimum bactericidal concentration (MBC)

MIC is referred as the antimicrobial drug concentration which completely inhibit the growth of bacterial population (~106 cells ml−1) occurs during 18–20 h of incubation at 35 ± 2 °C [33]. In the present study overnight grown bacterium culture was inoculated into TSB at final concentration >106 cells ml−1 which was divided in different parts (2 ml each) to treat with OTc and OTc@nZnO at different concentrations. The treated bacterial cultures were incubated at 35 °C with shaking for 24 h followed by assessment of number of viable bacterial cell through serial dilution and spreading over TSA.

Determination of death rate

The death rate of test bacteria in the presence of OTc and OTc@nZnO was determined to understand the efficacy of developed drug nanomaterial composite. At first overnight grown bacterial cell cultures were ten times diluted in fresh TSB medium followed by incubation to grow (at 35 °C with shaking) up to log phase (~5 × 1010 cells ml−1). The test antimicrobials were then added at MIC and the bacterial cultures were further allowed to incubate at same condition. Culture aliquots were withdrawn at different time intervals (0, 2, 6, 11, 14, 16 and 24 h), serially diluted and spread on TSA plates to assay number of viable cells. The death rate of OTc and OTc@nZnO treated cells was determined from the slope of the plot “Log Nt/N0 vs. incubation time”. Survivor equation was used to describe the change in bacterial population as a function of time [34]:

$${{{{{\rm{Log}}}}}}\left[ {N/N_o} \right] = - t/D$$
(1)

where N denotes bacterial population at time t; No indicates initial bacterial population and D (D-value) is the time required for a 1-log cycle reduction in the bacterial population or decimal reduction time. Corresponding equation for reduction of microbial population is first-order kinetic model [21]:

$${{{{{\rm{Log}}}}}}\left[ {N/N_o} \right] = - k.t$$
(2)

where k denotes first-order rate constant. By comparing Eqs. (1) and (2) we can get the expression:

$$k = 2.303/D$$
(3)

The rate constant k indicates the date rate of the bacterial cell and D-value signifies the time required to kill 1-log cycle bacterial population.

In silico [MgOTc]+ and [ZnOTc]+ compound preparation

The molecular structures of [MgOTc]+ and [ZnOTc]+ were prepared in the ChemDraw Ultra (CambridgeSoft) and subjected to energy minimization at the Avogadro platform by using universal force field (UFF) with 5000 steps of steepest descent algorithm [35]. UFF was reported to be does well with organometallic compounds [36]. These steps were performed to generate best conformation of these two molecules. After energy minimization and optimization of geometry the [MgOTc]+ and [ZnOTc]+ showed total energy of 676.9 kJ mol−1 and 826.17 kJ mol−1, respectively. The distances between coordinating atoms and metal ion in case of [MgOTc]+ ranged between 1.88–2.00 Å and for [ZnOTc]+ it was between 1.89–1.91 Å (Fig. S1).

TetR and TetA sequence retrieve process and protein structure prediction

Amino acid sequence of Tc resistance transcriptional repressor TetR(E) of A. hydrophila was collected from NCBI database with accession id: WP_139390343 [14]. By using NCBI BLASTp server we identified crystal structure (PDB ID: 5MRU_A) with respect to the query sequence of test TetR(E) [37]. Based on sequence alignment between the target protein and the template sequence, homology modeling was used for predicting 3D model of the target protein (PDB ID: 5MRU) through MODELER10.1 package, which was further used in molecular docking purpose [38]. DoGSiteScorer in the webserver https://proteins.plus/ was used to detect potential binding site which revealed that TetR(E) has one prominent binding site (P_0) (Fig. S2Ai, ii). The binding pocket P_0 showed volume of 1380.61 ų and surface area of 1693.21 Å2 with high druggability score (0.81) (Table S1). The amino acid residues constructing the P_0 binding site are Ala(5), Asn(2), Asp(1), Gln(4), Glu(1), Gly(2), His(3), Ile(5), Leu(10), Lys(1), Met(1), Phe(1), Pro(2), Ser(6), Thr(4) and Val(1) (Table S1). The same binding pocket in TetR(E) was also indicated by CASTp (http://sts.bioe.uic.edu/castp/) (Fig. S2Aiii). Similarly, amino acid sequence of tetracycline resistance protein TetA(E) of A. hydrophila was collected from NCBI database with accession id: PKD25314.1. By using NCBI BLASTp server we identified three crystal structures (PDB ID: 6S4M_A, 7BC6_A and 6T1Z_A) with respect to the query sequence of test TetA(E) [37]. The alignment of query protein sequence was done with the identified crystal structures using Phyre2 web platform (397 residues (100%) modeled at >90% confidence) and 3D model for the target protein was obtained. Binding-site analysis using DoGSiteScorer in the webserver https://proteins.plus/ revealed that TetA(E) has four prominent binding sites (P_0, P_1, P_2 and P_3) with druggability score more than 0.73 (Fig. S2Bi, ii) (Table S2). The details of these binding pockets are given in Table S2 which showed that binding pockets, P_0, P_1 and P_2 were having volume and surface area more than the Connolly solvent excluded volume and accessible area of ligand molecules. However, in CASTp, a single binding pocket was observed covering sites of P_0 and P_1 (Fig. S2Biii). SAVeS server (http://nihserver.mbi.ucla.edu/SAVES/) was used to check the stereo chemical quality of TetR(E) and Tet(A) on PROCHECKweb software which generated Ramachandran plot (RP) [39] (Fig. S3).

Ab initio 16S RNA structures prediction of A. hydrophilla

Partial sequences of 16S ribosomal RNA gene of A. hydrophila CCM 7232 strain (sensitive) (NCBI, NR_043638.1) and A. hydrophila strain CIAHRT_11 (NCBI, MG754418) were collected from NCBI database. At first, the secondary structures of 16S ribosomal RNA genes are prepared using mFold platform (http://www.unafold.org/) deltaG in plot file 12.kcl/mole [40]. The secondary structure of the partial sequences of 16S ribosomal RNA genes are presented in the Supplementary File (Figs. S4 and S5). From secondary structure, the tertiary structures were prepared in 3dRNA (Xiao Lab) using Ab initio technique [41, 42]. For CIAHRT_11, a model with 3dRNAscore of 26.54 was selected and for CCM 7232 it was 25.85.

In silico molecular docking studies

Molecular docking is widely used to predict binding affinity of a ligand to a protein chain. In the current work, to investigate the binding of [ZnOTc]+ on to resistance proteins i.e. TetR(E) and TetA(E), and 16S rRNAs, molecular docking was performed by using the Graphical User Interface program of “Autodock 4.2” (Ver. 4.2.6). The software is considered as best suited automatic docking tool for modeling flexible small drug molecules binding to receptor macromolecules [43]. According to a previous study 50% among the 1300 protein–ligand complexes have their top score conformations below 2 Å threshold using Autodock and the correlation between Autodock score and experimental binding affinity data was about 0.27 [44]. Lamarckian Genetic Algorithm (LGA) was used for conformational search to explore binding conformational landscape of [ZnOTc]+ and [MgOTc]+ docked against TetR€ and T€(E) efflux protein of A. hydrophila. Before docking, proteins structures and drug chemical structures were prepared by removing water molecule, assigned Kollman charges each atom of the residues, defining the rotatable bonds, adding partial charges, adding hydrogen atoms to target protein for better interaction. TetR(E) is evaluated with already reported binding pocket [16] with grid size of 32 × 28 × 30 xyz points with center coordinates of −1.099, 8.795, 10.963 (X, Y, Z). Before docking to check the correct binding pose, self-docking was executed between 5A,6-anhydrotetracycline with TetR(E) to see our docking protocol is optimum. During self-docking experiment it was found that the target ligand binds exactly to the site of identified crystal structure (PDB ID: 5MRU_A) with LBE value of −11.13 kcal/mol (Fig. S6 and Table S3). The docking of 5A,6-anhydrotetracycline in the TetR(E) of A. hydrophila generated similar binding efficiency (LBE, −11.79 kcal mol−1) with 3 H-bonds (Fig. S6 and Table S3). The RMSD from the reference structure are below 2.0 Å in both the cases (Table S3) indicating adequacy of the binding protocol. LGA was used for docking simulation by setting the initial position, orientation, and torsions of [ZnOTc]+ and [MgOTc]+ ligand molecules in random position. Ten runs of LGA were set to terminate after a maximum of 250,000 energy evaluations for each docking with population size of 150. Since TetA(E) showed a larger binding pocket (Fig. S2Bi–iii) a blind docking was performed with a larger grid size of 56 × 36 × 40 xyz points with center coordinates of −23.393, 12.054, 30.453 (X, Y, Z:). In total, 100 runs of LGA were set to identify the exact binding site from the binding pockets of the macromolecular structure of TetA(E). Similarly, electrostatic grid map of 16S rRNA of A. hydrophila CCM 7232 strain was calculated to identify the best binding sites through blind docking with grid size of 126 × 126 × 126 xyz points with center coordinates of 23.83, 114.99, −78.49 (X, Y, Z) and for 16S rRNA of A. hydrophila CIAHRT_11 strain it was with grid size of 126 × 126 × 126 points with center coordinates of 245.48, 117.79, −146.10 (X, Y, Z). In total, 100 runs of LGA were set to identify binding site from the most of 16S rRNAs structure. Finally, the best pose was extracted and aligned with the receptor structure with the lowest binding energy for final analysis. The molecular docking results were visualized in the Discovery Studio and PyMol, and the receptor-ligand interactions like H-bond and other non-bonded energies are analyzed in the Ligplot [45].

Data presentation and statistical analysis

The antibiogram data of OTc and OTc@nZnO against sensitive and resistant strains of A. hydrophila is presented as box plot (variables distribution with quartiles and outliers) and prepared using JMP Pro 10 software. The bacterial populations at different doses of antibiotics are presented as Log CFU ml−1 vs. OTc (µg ml−1) in both contour and linear regression line with good fit (RMSE: 1.51 to 1.92 for OTc and 0.38–0.54 for OTc@nZnO) and prepared using JMP Pro 10 software. The death kinetics data of both the strains of A. hydrophila are presented as Log (Nt/No) vs. time (hours) against the test antibiotics and kinetic parameters were analyzed in Microsoft Excel.

Results and discussion

Synthesis and characterization of OTc@nZnO

In the present study OTc loaded nZnO (OTc@nZnO) was synthesized with an objective to attach divalent cation Zn2+ to the OTc molecule and delivering the same using nZnO to bypass the cellular drug efflux barrier and exert the antibacterial effect to Tc resistant A. hydrophila by binding 30S ribosomal subunit. The OTc@nZnO was synthesized by a co-precipitation method in which OTc molecule was added at the stage of formation of ZnO nanoparticles (nZnO). The synthesis protocol is simple, quick and prepare stable dispersion of OTc@nZnO in water. Since Tcs are light and pH sensitive, the synthesis was done in dark condition and the final pH during co-precipitation was kept at pH 8 [46].

Detailed information on the crystal structure of the samples was analyzed through HR-TEM analysis. At lower magnification TEM images revealed that nanoparticles of ZnO were hexagonal in shape (Fig. 1ai) and at higher resolution TEM micrograph depicts the fringes of the spacing 0.29 nm (d spacing) corresponding to (100) ZnO plane (Fig. 1aii). The SAED pattern of nZnO (Fig. 1aiv) revealed the diffraction rings corresponding to (100), (002), (101), (102), (110) and (112) planes of ZnO. Fig. 1aiii shows the size distribution of nZnO analyzed with help of a histogram of 70 particles (range from 42 to 211 nm) and after Gaussian fit the estimated average particle size is 99.42 nm. However, crystal structure analysis using HR-TEM micrograph of OTC@nZnO revealed that the d-spacing value has increased to >0.3 nm in several regions (Fig. 1bii) and the average particle size has reduced to 17.33 nm (range 8.13–32.70 nm) (Fig. 1biii). This observation revealed that introduction of OTc hindered the crystal growth of ZnO making it smaller sized and also distorted the general crystal structure of ZnO as evident from the diffused diffraction rings of OTc@nZnO (Fig. 1biv).

Fig. 1
figure 1

HR-TEM (i), d spacing (ii), SAED (iii) and particle size distribution (iv) of nZnO (a) and OTc@nZnO (b)

Further, the loading of OTc in the nZnO was confirmed using FT-IR spectroscopy (Fig. 2a). nZnO showed absorption bands at ~3400 cm−1 and 1669 cm−1 for O-H stretching and bending vibrations, respectively of free H2O that are absorbed on the surface of nZnO [47]. Distinct bands at ~600 cm−1 and 1111 cm−1 are attributed to Z-O stretching [47, 48] and Zn-O-H bending [49] vibrations, respectively. The FT-IR spectra of OTc showed characteristics bands at 3377 cm−1 and 2924 cm−1 assigned to O-H stretching and C-H stretching vibration [50]. The band at 1695 cm−1 was assigned to C=O stretching vibration of –CONH2 or Amide I and the spectral band for Amide II (N-H bending and C-N stretching) appeared at 1525 cm−1 [51]. The C=O stretching vibration of conjugated ketone of ring A appeared at 1619 cm−1 whereas C=O of ring C appeared at 1583 cm−1 [51]. The spectral band at 1455 cm−1 was assigned to aromatic C=C vibration [50]. The FT-IR spectra of OTc@nZnO showed distinct major bands of nZnO and OTc. However, the characteristic spectra at 1695 cm−1 because of Amide I and at 1583 cm−1 due to >C=O of ring C was absent in OTc@nZnO. Thus, it can be postulated that co-precipitation of nZnO in the presence of OTC has led to complexation of Zn2+ in the following two moieties of OTC, (1) the Amide moiety of ring A and (2) >C=O moiety of ring C. Similar observation was also reported in a previous study with interaction of Ca2+ in the designated positions of OTc leading to elimination intermolecular H-bond [21].

Fig. 2
figure 2

FT-IR spectra of nZnO, OTc and OTc@nZnO (a), UV spectra of OTc (b) and OTc@nZnO (c); in vitro release kinetics of [ZnOTc]+ from OTc@nZnO (d)

To further confirm the complexation nature between Zn2+ and OTc, λmax for OTc and OTc@nZnO were observed in the UV-Vis spectrophotometer. OTc showed two major absorptions peak at 256 and 354 nm (Fig. 2b). Following the Woodward Fieser’ rule for correlating λmax with molecular structure, the λmax at 256 nm is assigned to the carbonyl substituted aromatic moiety at ring D and the λmax at 354 nm is assigned to α, β unsaturated carbonyl compound at ring A. The detailed calculation is provided in Table S4. Interestingly, the UV absorbance of OTc@nZnO showed bathochromic or red shift of λmax to 276 nm for aromatic moiety and to 376 nm for α, β unsaturated carbonyl compound (Fig. 2c). Hence, besides the FT-IR spectroscopic results, this bathochromic shift of UV λmax also postulates that Zn2+ was complexed with the OTc at ring A and ring D with respective amide and aromatic carbonyl moieties. Previous literature also reports bathochromic shift of OTc from 353.35 to 374 nm due to formation of ligand-to-metal charge transfer coordination compound (LMCT) by OTc molecule in coordination with the d orbitals of the Zn2+ [52]. Further, it was extensively reported that Mg2+ ion forms coordination compound with Tc molecule near the aromatic moiety at ring D get binds with 16S RNA to block binding of aminoacylated tRNA to the A site. There are extensive reports that the amide group at ring A could bind with various metal ion like Cd, Mo, etc. to form metal coordination compound [53].

Loading and kinetics of [ZnOTc]+ release from OTc@nZnO

Addition of strong acid cause dissolution of metal oxide framework thus releasing the encapsulated drug molecule [54]. In the present study, addition of HCl acid to OTc@nZnO cause complete release of OTc in the water and the same was estimated by measuring absorbance at λmax 270 nm. The calculated OTc loading on to OTc@nZnO was 9.35%. Since the drug molecule must desorb from the zinc oxide crystal framework to exert its antibiotic properties the release of [ZnOTc]+ from the OTc@nZnO suspension was studied and the release data were fitted in to first-order kinetics equation:

$$\frac{{M_t}}{{M_ \propto }} = \left( {\frac{{M_t}}{{M_\infty }}} \right)_{{{{{{\rm{max}}}}}}}[1 - \exp ( - k_f \times t)]$$
(4)

where \(\frac{{M_t}}{{M_ \propto }}\) is the fraction of [ZnOTc]+ released at a time t, \(\left( {\frac{{M_t}}{{M_\infty }}} \right)_{{{{{{\rm{max}}}}}}}\) stands for the maximum fraction of [ZnOTc]+ released during the total period and kf denotes first-order rate constant. The in vitro release profile depicted a slow and sustained release of [ZnOTc]+ from OTc@nZnO and 10–28% of the loaded OTc was released with 2 h and more than 0.74% release occurred within 40 h (Fig. 2d). Hence it may be postulated that during in vivo condition, the initial bulk release of drug is sufficient to kill the bacterial cell and subsequent sustained release maintain the killing environment for long duration. It was also found that the rate of release of [ZnOTc]+ was pH dependent and a faster rate was observed at pH 4 and 9 as compared to 7. The higher release rate at pH 4 can be attributed to the dissolution of metal oxide framework of nZnO by presence of HCl. It was reported that ZnO dissolves in water as Zn2+ at pH 4 through hydrolysis releasing hydroxyl ions (OH) to increase the pH [55]. Though ZnO is fairly stable at alkaline solution, a faster release rate of [ZnOTC]+ at pH 9.0 may be explained by the formation of fairly water-soluble sodium zincate in the presence of NaOH [56]. The release data fitted well to the first-order kinetic equation with high (>0.98) coefficient of determination (R2) and the rate constant followed the order pH 4 (Kf 0.21) > pH 9 (Kf 0.13) > pH 7 (Kf 0.07) (Table S5).

Antibacterial efficacy, MIC and death kinetics

The antibiogram study (Fig. 3a) revealed that ATCC 7966 strain of A. hydrophila is sensitive to both free OTc and OTc@nZnO with similar trend in bacterial zone of inhibition, 17–20 mm and 18.5–22 mm, respectively. However, the A. hydrophila strain CIAHRT_11 did not show any zone of inhibition for free OTc at the given dose (30–60 µg). Interestingly, the resistant strain did show a considerable bacterial zone of inhibition (9–16 mm) for OTc@nZnO. Antibiogram studies for synthesized nZnO alone were also done against both the strains, however, no zone of inhibitions was observed (Fig. S7). To better understand the bacteriotoxicity of developed formulation, bacterial cells were exposed to different concentration of free OTc and OTc@nZnO in TSB medium. In case of A. hydrophila ATCC 7966 strain, it was observed that free OTc at 50 µg ml−1 completely inhibit bacterial cell growth. However, OTc@nZnO exert the same phenomenon at very low concentration of 28.125 µg ml−1 which is equivalent to 2.6 µg ml−1 OTc (OTc loading 9.35%) (Fig. 3b). Similarly, bacterial population of A. hydrophila strain CIAHRT_11 as an effect of various concentration of OTc and OTc@nZnO was also observed (Fig. 3c). Free OTc at 150 µg ml−1 concentration could be able to completely inhibit the bacterial population, whereas OTc@nZnO @ 75 µg ml−1 (~7.02 µg ml−1 OTc) could be able to inhibit bacterial growth. This drastic reduction of MIC was also reported previously where conjugation of silver nanoparticles with methicillin lead to reduction of MIC from 250 to 7.8 µg ml−1 against Staphylococcus epidermidis [57]. Similarly, conjugation of gentamycin gold nanoparticles led to reduction of MIC from 125 to 31.25 μg ml−1 against S. epidermidis [58]. In case Tc, nanotization of with calcium phosphate nanoparticles was reported to reduce MIC from 180–150 to 13–35 µg ml−1 for multi-drug resistant diarrhea-causing bacteria [21].

Fig. 3
figure 3

Antibiogram of OTc and OTc@nZnO against resistant and sensitive strains of A. hydrophila (a); minimum inhibitory concentration of OTc@nZnO against OTc sensitive (b) and resistant strain (c) of A. hydrophilla; death kinetics in presence 15.5 μg ml−1 OTc@ZnO and equivalent concentration of free OTc and free nZnO against sensitive A. hydrophila (d); and 77.78 μg ml−1 OTc@ZnO and equivalent concentration of free OTc and free nZnO against OTc resistant A. hydrophila (e)

Cell death kinetics by free OTc (Fig. 3d) and OTc@nZnO (Fig. 3e) at their MIC concentration revealed that free OTc caused gradual decline CFU number for A. hydrophila ATCC 7966 strain during 24 h study. However, it could not be able to do the same for the strain CIAHRT_11. Similarl to antibiogram study, free nZnO could not be able to exert any antibacterial efficacy during 24 h study against both the strain. In contrary, OTc@nZnO performed excellent by gradual decline of CFU number for both the strains i.e., ATCC 7966 and CIAHRT_11. According to the first-order kinetic model, Otc@ZnO showed a higher (−1.14) death rate constant for A. hydrophila (ATCC 7966) as compared to A. hydrophila (CIAHRT_11) (−0.93) (Table 1). This signifies that Otc@nZnO was faster in killing the sensitive strains of A. hydrophila as compared to the resistant one. This might due the fact the sustained or delayed release of [ZnOTc]+ from the Otc@nZnO. However, interestingly the D-value, which signifies the time required to kill 1-log cycle bacterial population, of Otc@nZnO against both the strains are nearly same (2.49 and 2.43, respectively) (Table 1). This fact signifies that the efficacy rate of Otc@nZnO against both the sensitive and resistant strains of A. hydrophila were similar. Tc incorporated in calcium phosphate nanoparticles was also reported to kill faster multidrug-resistant bacteria as compared to free Tc [21].

Table 1 Death kinetic parameters of OTc@nZnO against sensitive and resistant strain of A. hydrophila

Molecular docking to resistant proteins

Tc antibiotics like OTc preferably bind to highly conserved 16S rRNA in 30S ribosomal subunit arresting translation by sterically hindering docking of aminoacyl t-RNA [12]. However, the process gets interrupted by the presence of resistance gene encoding TetA efflux protein which transports intracellular Tc out of cell cytoplasm. Further, the expression of TetA protein is governed by TetR protein which upon binding with Tc lost its regulation over expression of TetA [14]. It was reported that the Tc binding pocket of TetR protein is composed of the COOH-termini of α4 and α6 and the helices α5, α7, α8, α8’, and α9’ [16]. Tc binds in this pocket in coordination with Mg2+ and H-binding with the amino acid residues. The Mg2+ forms octahedral coordination with the chelating ketoenolate moiety at O-11/O-12 of Tc, His100Nϵ and three H2O molecules which further forms H-bond to the carboxylate of Glu147. The residues Gln116NϵH and His64NϵH also forms H bonds to the amide oxygen in position 2 and negatively charged enolate O-3. The O-3 and the positively charged ammonium N-4 forms H-bond to amide chain of Asn82. The hydroxyl at O-121 form aromatic H-bond with Phe86. The aromatic ring D form hydrophobic interaction with Pro105, Leu170, leu174, Met177 and Arg104 [16]. Hence binding of drug molecule to TetR is very vital in resistance development in bacteria and in the present study the binding characteristics of [ZnOTc]+ toward TetR(E) is evaluated by molecular docking studies with the already reported binding pocket using AutoDock. Though it was reported that binding poses and affinities information obtained from Autodock may not be exact with the experimental finding [44], in the present study it was performed to get an idea on comparative binding properties between two antibacterial compounds. The predicted binding energy of [MgOTc]+ and [ZnOTc]+ for TetR is summarized in Table 2. It was found that the calculated lowest binging energy (LBE) of [MgOTc]+ is low for TetR(E) as compared to [ZnOTc]+. This signifies that the binding affinity of [ZnOTc]+ toward the resistance protein TetR is less as compared to [MgOTC]+. Further, the inhibition constant (Ki) of [ZnOTc]+ for TetR(E) was also calculated to be in micromolar range (µM) as compared to nanomolar (nM) for [MgOTc]+. The docking score or LBE indicates a direct relationship between the energy of the binding affinity, referring to the lowest docking scores and stability [59].

Table 2 Molecular docking simulation results and interaction analysis between resistance proteins and test ligands ([MgOTc]+ and [ZnOTc]+)

The interaction analysis between 5A,6-anhydrotetracycline TetR in PDB ID: 5MRU_A using LigPlot+ revealed that the ligand binds to the specific pocket of TetR with 8 H-bonds with Thr103, His100, Ser67, His64 and Asn82; and Hydrophobic interaction with Pro105, Leu60, Leu134, Phe86, Gln116, Ser138, Met135 and Arg104 (Fig. S6aiii and Table S6). However, in the present study despite the same Tc binding pocket in TetR protein was used for docking studies [16], [MgOTc]+ formed different orientation which might be due to introduction of extra hydroxyl moiety at the 5 position in OTc and change in the amino acid sequence of TetR of A. hydrophila. In the present study, Mg2+ did formed octahedral coordination but did not participate in binding with any amino acid residue. However, the His100 formed H-bond with the hydroxyl group present at O-9 position (Fig. 4aii). Like reported earlier the residues Gln116 did participated actively, however made H-bonds with the O-8 and O-12. The aromatic ring D formed hydrophobic interaction with Pro105, Thr103 and Gln135 (Fig. 4aii and Table 2). Further, the [ZnOTc]+ also showed different orientation in the same pocket with higher binding energy (LBE, −5.49 Kcal mol−1) (Table 2). The103, His100 and Ser138 formed 6 H-bonds with Zn-OH, O-9 and N of amide moiety (Fig. 4bii). Due to presence of Zn2+ at the amide moiety, Asn82 and Gln116 formed H-bonds with Zn-OH. The hydrophobic interaction existed between aromatic ring D with Pro105, Ser104 and Gln135. Previously, molecular docking with OTc with parental TetR in a different platform (YASARA 16.2.18) reported good binding energy of 9.03 kCal/mol with hydrophobic interaction with five amino acid residues (His64, Phe86, Pro105, Gly138 and His139) and pi-pi bond with Phe86 [60, 61]. However, in that study OTc intercalation with Mg2+ in octahedral complex was not considered for docking purpose. In another study caffeic acid and gallic acid were evaluated for assessing their binding affinity toward TetR of Streptococcus spp. using AutoDock 4.2 [43]. Both the acids were reported to have promising inhibitory effect against TetR with binding energy of −5.93 Kcal mol−1 and –3.92 Kcal mol−1 with 7 and 4 H-bond interactions.

Fig. 4
figure 4

3D representation (i) and binding mode (ii) of TetR(E) with [MgOTc]+ (a) and [ZnOTc]+ (b)

Recently, there are many reports on in silico evaluation of chemical compounds for their inhibitory effects toward antibiotics efflux pumps. For example, MepA efflux pump inhibitory effect of 1,8-naphthyridines sulfonamides was studied using molecular docking techniques against Staphylococcus aureus [62]. Further, the same compounds are also evaluated in silico for their inhibitory action against NorA efflux pump [63]. Similarly, anti-fungal azoles compounds were also evaluated for their inhibitory effect toward Tet(K) efflux pump of Staphylococcus aureus through simulation studies [64]. Similarly, in silico docking studies (AutoDock Vina) of TetA efflux protein of E. coli with berberine and Tc resulted higher free energies of −7.3 and −6.7 kcal/mol, respectively [65]. However, in the present study, the binding efficacy of drug molecules toward TetA(E) efflux protein was done to understand the better antibacterial properties of [ZnOTc]+. The binding affinity of [ZnOTc]+ with TetA(E) efflux protein was calculated using molecular docking and compared with [MgOTc]+. [MgOTc]+ showed higher binding affinity toward TetA(E) with low LBE (−8.26 Kcal/mol) and low Ki (nanomolar range). However, [ZnOTc]+ performed poorly in binding the pockets of TetA(E). It was having high LBE (−4.4 Kcal mol−1) and Ki constant (millimolar range). Hence it could be postulated that [ZnOTc]+ can effectively bypass the efflux protein as compared to [MgOTc]+ to render the antibacterial activity by binding with 16S rRNA of 30S ribosomal subunit. It was reported that motif A of TetA consisted with G64ALSD68RFG71R72RP is responsible for binding with Tc [65]. In the present study, the [MgOTc]+ was found to bind of TetA(E) with hydrophobic interactions with Phe138, Gly135, Ile221, Gly312, Gly225, Phe254, Leu316 and 5 H-bonding with His258 and Gln222 (Fig. 5aii). However, [ZnOTc]+ only formed 7 H-bond (Gln222, His344, Gln54, Gly108, Gln320) and five hydrophobic interactions (Leu316, His258, Gly135, Thr343, Tyr50) with the amino acid residues of TetA(E) (Fig. 5bii).

Fig. 5
figure 5

3D representation (i) and binding mode (ii) of TetA(E) with [MgOTc]+ (a) and [ZnOTc]+ (b)

Molecular docking to 16s rRNA

The complexation with divalent Mg2+ is also crucial for the binding of Tc to acceptor site for aminoacylated tRNA between head and body the 30S ribosomal subunit [66]. Tc binding pocket is formed by H34 (RNA residues 1196–1200:1053–1056) in combination with residues 964–967 from the H31 stem-loop in Thermus thermophilus [6]. It primarily forms H-bonds with the exposed backbone sugar phosphate oxygen atoms of H34. The hydrophobic interaction of Tc happens with the bases of 1054 and 1196 RNA residues. The flat fused-ring structure of Tc molecules with hydrophilic functional groups along one side can make both charged and hydrophobic or interactions. The 30S ribosomal subunit was reported to have three binding sites for binding tRNA viz. A (aminoacyl), P (peptidyl) and E(exit) sites [6]. A general consensus is that Tc particularly blocks binding of aminoacyl-tRNA to the A site of the protein synthesis machinery of mRNA-ribosome complex [67, 68]. However, the exact binding site and mechanism is not fully confirmed [6].

Since binding of drug molecule to the A site of 16S rRNA is necessary to exert its protein inhibition property, molecular docking of [ZnOTC]+ was also performed with 16S rRNAs of both the strains (ATCC 7966 and CIAHRT_11) of A. hydrophila. Since crystallographic data of 16s rRNA of A. hydrophila is not available we prepared 3-D fold for both the strains using Ab initio technique. It was reported that intercalation of divalent Mg2+ is very crucial for binding of Tc to the rRNA [69]. The Mg2+ forms salt bridges with phosphate oxygen atoms of G1197 and G1198. Further, presence of Mg2+ ion in absence of Tc was also reported to maintain local structure of 30S in that region [70]. Hence in the present study docking was performed Mg intercalated OTc i.e. [MgOTc]+. It binds with rRNA with very low binding energy (−19.10 to −16.27 kcal mol−1) and the Ki is also very low (femto to picomolar range) (Table 3). The very low binding energy between [MgOTc]+ and 16S rRNA may explained by formation of many H-bonds. The binding pocket of [MgOTc]+ in 16S rRNA of CCM 7232 consisted of 12 H-bonds with U1121, G1122, C1123, G1238, A1125 and U1236; whereas in case of CIAHRT_11 it formed 10 H-bond with U386, G387, A803, A377, C802 and U738 (Table 3 and Fig. S8). Similarly, [ZnOTc]+ also showed very low binding energy (−15.76 to −11.43 kcal mol−1) with both the 16S rRNAs and the Ki of [ZnOTc]+ for both 16S rRNA are found to nano to pico molar range (Table 3). [ZnOTc]+ also formed 8 H-bonds with U312, A348, G332 and C347 of 16S rRNA of CCM 7232, and 10 H-Bonds with U386, G387, A803, A377, C802 and U738 (Table 3 and Fig. S8). Hence it could be postulated that besides having low binding affinity toward the Tc resistance proteins, [ZnOTc]+ is also having high binding affinity toward the 16s rRNA thus exerting antibacterial property to resistant microorganism.

Table 3 Molecular docking simulation results and interaction analysis between 16S rRNA of A. hydrophila and test ligands ([MgOTc]+ and [ZnOTc]+)

Conclusion

The present study describes simple synthesis method of Zn2+ intercalated OTc molecules encapsulated in crystal structure of ZnO nanoparticles. The particles were hexagonal in shape with average size of 99 nm. Spectroscopic investigation revealed Zn2+ intercalation on the ring structure of OTc at ring A and ring D with respective amide and aromatic carbonyl moieties. In vitro release profile of Zn2+ intercalated OTc demonstrated sustained release of nearly 80% loaded OTc over 2 days. Due to nanotization and intercalation of Zn2+, modified OTc molecule showed superior antibacterial properties toward resistant strain of Aeromonas hydrophila as compared to free OTc. The minimum inhibitory concentration of free oxytetrcycline was 150 µg ml−1 and nanotized Zn2+ intercalated oxytetrcycline performed better with very low inhibitory concentration equivalent to 7.02 µg ml−1 of OTc. It was also observed that nanotization and Zn2+ intercalation also reduced the minimum inhibitory concentration against sensitive strains of A. hydrophila with faster cell death kinetics. To understand the mechanism of superior action of nanotized Zn2+ intercalated oxytetrcycline, molecular docking was performed that revealed less binding affinity toward Tc resistant proteins i.e. TetR(E) and TetA(E) as compared to free OTc. Molecular docking also revealed that Zn2+ intercalated oxytetrcycline had good binding affinity toward 16S rRNA of 30S ribosomal subunit of A. hydrophila. Finally, this study suggests that nanotization of OTc through Zn2+ intercalation and encapsulation in ZnO nanoparticles could break the Tc resistance pattern of A. hydrophila strains which are causing several fish mortality in aquaculture farm. However, further in vivo studies are required for its feasibility for use in animal health care.