Toxicity mechanism of Cu²⁺ ion individually and in combination with Zn²⁺ ion in characterizing the molecular changes of Staphylococcus aureus studied using FTIR coupled with chemometric analysis

Abstract

Copper and zinc have a high binding affinity with a Staphylococcus aureus bacterial community. This causes a change in the biomolecular composition of S. aureus. Our study aims at understanding the resistance mechanism of Cu and Zn either or in various combinations using FTIR and chemometric techniques. Zn toxicity resulted in a significant change in lipid content (3100-2800 cm⁻¹) compared to Cu. A significant decrease in protein content is observed for Cu treatment in the amide region. The bio-concentration factor shows a higher value for Cu compared to Zn. The increase in band area of carbohydrates moieties 1059 cm⁻¹ shows the secretion of EPS due to Cu toxicity. A significant change in nucleic acid compositions was noted in the region1200-900 cm⁻¹ due to Zn treatment. Secondary structural change in protein shows β sheet formation. The result of the finding shows Cu has greater toxicity than Zn. Further toxicity effects were greatly enhanced for metal mixtures ratio (Cu:2Zn). This shows Zn exhibits synergism effect with Cu. The obtained ROC (receiver operating characteristic) curve area gives good reliability of the experiments. The study attempts to understand the mechanism of toxicity removal of Cu and Zn metal mixtures by bacterial population using FTIR coupled with chemometric techniques.

Graphical abstract

Introduction

Water pollution due to the presence of toxic metals remains a serious threat to the environment. Microorganisms are widely affected by the release of toxic metals existing in the environment. They play an effective role in resisting the toxicity and detoxify the metals from the environment by removing them. Bacterial populations have evolved specific resistance mechanisms to exist in a polluted environment. General bacterial resistance mechanisms are efflux pumping of harmful material out from the cell, bioaccumulation, and converting more toxic ion species into a less toxic one. Thus they act as bioremediation of toxic wastes. Efflux pumping is important in the resistance mechanism. Hence understanding the same necessitates developing a process suitable for cleaning the environment from pollutants [1]. Metal toxicity resistance varies widely in a bacterial population. These differences in mechanisms shown by the bacteria are due to varying cell wall features that exist among them. This can be understood from metal accumulation at the cellular level. The metals are stopped from entering into the cell by forming metal binding agents. These agents protect microbes and their control on metal uptake. Thus, it plays a major role in waste treatment processes. This metal transport mechanism is significant in the bacterial population. In the environment, copper serves at lower levels as a nutrient, but is toxic at a higher level. Copper and zinc are the highest affinity binders among the divalent metals present widely in nature. Various investigations of Cu and Zn in the bacterial community show the significance of the storage and development of enzyme activities. The bacterial systems separate and expel these metals to maintain balance for various cellular actions. The basic nature of copper is to cause multiple damages at the cellular level in bacterial systems. Zn (II) is an important micronutrient for bacteria. It has a considerable effect on toxicity at higher concentrations [2]. Mostly the microorganisms can resist heavy metals ions like Zn²⁺, Cu²⁺, Cd²⁺, Ag²⁺, etc. by limiting the penetration levels to prevent toxicity. Bacteria are very sensitive to surrounding environmental conditions even with small changes. They can sustain sharp changes in the surrounding environment. This proves their ecological supremacy under external stress conditions. They secrete a protective coating called extra polymeric substances (EPS). It comprises polysaccharides, proteins nucleic acids, and lipids. It plays a vital role in forming the biofilm and an external covering from toxic substances [3].

Staphylococcus aureus (S. aureus) is a gram-positive bacterium consisting of cytoplasm and rigid cell walls with a complex molecular network, made of peptidoglycan, which forms a strong mesh-like structure that covers the cell. It acts as a barrier and prevents drugs and toxicants from entering into it. The Gram-positive bacteria have an additional polymer, covalently bound to the peptidoglycan, the teichoic acid built of phosphate backbone with a side chain of variable composition [4, 5]. Hence understanding the activities of bacterial cells, at the interfaces due to toxicants plays a critical role in studying the structural and biochemical changes at their surfaces. The metal ion has a strong binding with various functional groups on the cell surfaces even at low concentrations. This leads to enzyme activity changes at the intracellular level [6]. Molecular composition changes occurring in the bacterial populations are studied using infrared energy. The interaction of energy provides fundamental information regarding the molecules that arise at a cellular level. With the help of this interaction, FTIR provides precise and rapid detection of biochemical changes arising in the microorganisms. The spectral analysis of whole organisms in fingerprint gives the biochemical characteristics present in the bacteria. It gives information regarding cellular compositions such as proteins, carbohydrates, nucleic acids, and phospholipids [7].

We have studied these changes in the molecular profile of S. aures populations due to Cu-Zn metal mixtures. IR spectra of S. aureus show several bands describing the molecular compositions of the cells. They are sensitive to structural changes like intermolecular H bonding and membrane constitution. Hence our study utilizes the IR spectra to visualise the molecular changes due to this metal resistance. Our earlier studies [8, 9] on Ni-Cr metal interaction on E. coli and in S. aureus populations proved effective ways of molecular changes using infrared spectral analysis. There has not been any analysis done in the case of Cu-Zn interaction for S. aureus using FTIR. Considering the above discussion our study aims to know how S. aureus reacts to extracellular metal ions Cu and Zn. We have also studied the individual and combined toxicity of Cu and Zn stress on S. aureus to find any significant role played in their resistance to intoxication. The outcome of the study explains how the metal stress impacts phenotypic/genotypic changes occurring in them using FTIR combined with chemometric analysis.

Materials and methods

Chemicals used

Chemicals used Copper sulphate (CuSO₄) and Zinc sulphate (ZnSO₄) were bought from Sigma Aldrich of analytical grade. All the standard solutions were made ready in deionized water.

Test micro-organisms: Pure culture of Staphylococcus aureus was brought from B. K. Birla College, Kalyan. These cultures were maintained with Sterile Nutrient broth. They were regularly sub cultured for proper maintenance.

Minimum inhibitory concentration (MIC)

The MIC is defined as the lowest concentration in which metal inhibits the visible growth in the test. Staphylococcus aureus was checked for heavy metal tolerance. The sterile Nutrient broth was used as a test medium. Varying concentration of ZnCl₂ and CuCl₂ were used (50–500 ppb for Zn and Cu). The tubes were incubated at 37 °C for 24 h. Growth in the form of turbidity was observed. The MIC values were interpreted when the culture reached an OD600 of 0.55–0.59 [CL157 (ELICO)]. All the tests were carried out in triplicates. The MIC results show that S. aureus could tolerate Zn at 350 ppb and Cu at 200 ppb. For metal interaction studies equal concentration of Cu and Zn salt solution were used in the range of 10–300 ppb. The MIC value for metal mixtures was found to be 250 ppb for S. aureus. Three flasks were maintained. One flask, containing Nutrient broth without any heavy metal and other flasks containing nutrient broth along with the 1/4th MIC value of the respective heavy metals Cu and Zn. For metal interaction study ¼ th MIC (62.5 ppb) at three different proportions of concentration of Cu and Zn (1:1, 1:2, 2:1) was taken and the sample was harvested at the end of the period.

Experimental study

The experiments were performed in aerobic conditions. Microorganisms were grown in proposed metal concentrations having 100 ml of Sterile Nutrient broth (Himedia) in 150 ml Erlenmeyer flask at 30 °C at shaker conditions (90 rpm) for 18 h. The exponential phase culture was centrifuged at 5000Xg for 10 min and washed twice. The final sample was measured at OD600. Bacterial cells were collected from the liquid cultures by centrifugation (Remi, R-8C) at 10,000 rpm for 15 min. After removing the supernatants, the bacterial pellets were cleaned twice with phosphate buffer (pH 7.4; Na₂HPO₄-7H₂O 20.214 g; NaH₂PO₄H₂O 3.394 g; 1 L). They were used for further analysis in the FTIR spectrometer.

FITR analysis

The cells were separated by centrifugation at 10,000-g for 8 min. The pellet was resuspended in sterile normal saline (0.8% NaCl). All the samples were recorded in IIT SAIF, Bombay for FTIR analysis in the range 4000–400 cm⁻¹ (Bruker, Germany; 3000 Hyperion Microscope with Vertex 80 FTIR System). The sample of 2 ml of S. aureus solution is used in the ATR cell for spectral measurement. A total of 20 scans were taken for each spectrum at 4 cm⁻¹ resolution. All spectra were recorded for three identical samples. The collected spectra were vector normalized second derivatives with 15 points smoothing using the Savitzky–Golay algorithm using Origin 8.0 software. For curve fitting analysis Gaussian function was used to resolve peak pattern with least chi square value as described in our recent work [10].

Bioaccumulation of Cu, Zn, and its mixtures

Bioaccumulation of metals of Cu, Zn, and various combination ratios of the bacterial biomass was analysed using ICP-AES analysis (M/s. Spectro, Germany). Both the cells and the supernatant were analysed for residual heavy metal. The experiments were performed in triplicate. The bioconcentration factor (BCF) and metal removal efficiency (RF) were calculated as described by Phetsombat et al. [11] and Zhou et al. [12].

Principal component analysis (PCA)

Principal Component Analysis (PCA) was performed with SPSS 16.0. It is used commonly for multivariate data reduction. All spectral data are pre-processed before subjecting to PCA operations. The output of the PCA results in a small number of principal components. It explains the maximum variation of the data present in the samples. Each PC is composed of scores and loadings. It is utilized for quantitative approaches of discriminating samples. The scores of the component were plotted to gather data that accounts for variability in the FTIR information. The scores were plotted to obtain the data that describes the variation form the FTIR spectra.

Measurement of differentiation index

The reproducibility of the samples was found from Pearson’s product-moment correlation coefficient (between replicates 1–2, 2–3, 3–1) which is expressed as the differentiation index (D) [4, 13]. It was calculated using the following equation:

$$r_{y1y2}\frac{{\sum }_{i=1}^n{y}_{1i}{y}_{2i}-n\overline{y_1{y}_2}}{\sqrt{{\sum }_{i=1}^n{y}_{1i}^2-n\overline{y_1^2}}\sqrt{{\sum }_i^n{y}_{2i}^2-n\overline{y_2^2}}}$$ 1

where y₁ and y₂ values mean observances of two kinds of comparative absorption spectra at a given wavelength; n represents the number of data points. From the correlation coefficient r_y1y2, the differentiation index Dy₁y₂ may be defined as

$$D_{y1y2}=\left(1-r_{y1y2}\right)\times 1000$$ 2

The D value was measured for the selected region. This includes W₁ (3050-2800cm⁻¹), W₂ (1700–1600 cm⁻¹), W₃ (1200–900 cm⁻¹) and whole spectrum W₄ (4000–450 cm⁻¹).

Statistical analysis

The results were expressed as ± standard error of the mean (SEM). S.aureus heavy metal resistant groups of zinc and copper and their metal mixtures vs. control group were analysed using the one-way ANOVA test using SPSS 16.0. A probability level (p value) of less than 0.05, 0.01, and 0.001 were considered statistically significant. Hierarchical cluster analysis and Receiver operating characteristic analysis was performed using the same software.

Hierarchical cluster analysis

Hierarchical cluster analysis (HCA) was utilized to know the differences among the samples tested. HCA analysis was carried out using SPSS 16 software using Pearson’s product-moment correlation coefficient with Euclidean separation ae utilized using Ward algorithm [14]. The clustering maximizes their similarity concerning measured characteristics to minimize errors in describing a large number of samples. Hierarchical groups are formed in the manner described for classification purposes of plants and animals in predicting the efficiency of the results [15].

Receiver operating characteristic analysis

Receiver operating characteristic (ROC) analysis allows the diagnostic performance of different tests by comparing the sensitivity and specificity of the samples tested. ROC curves were constructed by using SPSS 16 software. The ROC plot shows sensitivity against 1-specificity for each point with the threshold value between zero and one. The control was taken as reference point fixing the potential threshold limit of values as 95% of cut off points. The upper left-hand portion of the graph will have a region on the curve where potential threshold values occur that provides both high sensitivity and low specificity (1-specificity) of the curves [15]. Comparing the area under the ROC curves allows us to select the best test carried out in the experiment.

Results and discussion

Figures 1, 2 and 3 show the representative of the second derivative vector normalized spectra of control, Cu, Zn, and their metal mixtures. The peak assignment and the band area are measured for various treatments are shown in Table 1. The response to metal toxicity showed changes in the spectra of –CH groups of lipids at 3100–2800 cm^-1. Binding of metals ions with amino acids/proteins showed changes occurring at 1500–1800 cm⁻¹ regions. The phosphate groups of nucleic acids (DNA/RNA) structure occur in 600–1200 cm⁻¹ regions [16]. The lipid membranes of S. aureus are characterized in the region 2900–2800 cm⁻¹. The band ~2957 cm⁻¹ assigns to CH₃ asymmetric stretching of lipids. The band ~2930 cm⁻¹ corresponds to CH₂ asymmetric stretching of lipids groups. The symmetric CH₃ & CH₂ stretching assigns to 2875 cm⁻¹ and 2850 cm⁻¹ mainly contributing to lipids. CO stretching of polysaccharides corresponds to 1745 cm⁻¹. The band ~1635 cm⁻¹ was assigned to C=O stretching and N-H bending of amide I protein. Amide II rises due to N-H bending CN stretching of proteins at 1541 cm⁻¹. Amide III of proteins corresponds to 1316 cm⁻¹. The various bending vibration of lipids arises at 1462 cm⁻¹, 1458 cm⁻¹, and 1396 cm⁻¹. The phosphodiester groups (PO²⁻) arise from phospholipids bands at 1339 cm⁻¹, 1233 cm⁻¹,1174 cm^-1, and 1079 cm⁻¹ which contributes to nucleic acids of S. aureus. Carbohydrates moieties of C-O stretching present at 1059 cm⁻¹. DNA/RNA of the nucleus corresponds to 994 cm⁻¹ and 964 cm⁻¹. Also, the band area ratio of nucleic acid/protein, lipid/phosphorylated protein, and lipid/nucleic acid were calculated as suggested by Obinaju et al. [16]. It provides information on the compositional changes of lipids, proteins, nucleic acids, etc., in a bacterial population.

Fig. 1.

Characteristic second derivative vector normalized FTIR spectra of S. aureus in the 2908–2820 Cm⁻¹ region (A) Control, Cu and Zn toxicity (B) Metal mixtures at varying proportion

Fig. 2.

Characteristic second derivative vector normalized FTIR spectra of S. aureus in the 1750–1250 Cm⁻¹ region (A) Control, Cu and Zn toxicity (B) Metal mixtures at varying proportions

Fig. 3.

Characteristic second derivative vector normalized FTIR spectra of S. aureus in the 1250–950 Cm⁻¹ region (A) Control, Cu and Zn toxicity (B) Metal mixtures at varying proportions

Table 1.

Frequency assignment and the band area of S. aureus treated with Cu, Zn and its mixtures

Peak Number	Wavenumber Cm−1	Functional group	Control	Cu	Zn	Cu:Zn	2Cu:Zn	Cu:2Zn
1	2957	CH3 antisymmetric stretching: lipids	5.26 ± 0.84	1.68 ± 0.07 a	1.80 ± 0.04 a	4.85 ± 0.96 a	3.82 ± 0.58 a	0.24 ± 0.03 a
2	2930	CH2 antisymmetric stretching	7.92 ± 1.02	10.62 ± 1.13 a	10.77 ± 0.97 b	1.46 ± 0.09 a	0.38 ± 0.06 a	0.24 ± 0.02 a
3	2875	CH3 symmetric stretching	2.99 ± 0.83	0.94 ± 0.08 a	0.56 ± 0.04 a	9.67 ± 0.86 a	5.47 ± 0.79 c	0.18 ± 0.04 a
4	2850	CH2 symmetric stretching: mainly lipids	2.62 ± 0.63	1.28 ± 0.29 a	0.87 ± 0.07 a	0.14 ± 0.02 c	9.59 ± 1.03 b	0.15 ± 0.01 a
5	1745	CO stretch: polyester storage compounds, polysaccharides	66.78 ± 4.94	14.05 ± 1.58 b	5.98 ± 0.89 b	80.10 ± 7.04 a	9.46 ± 1.32 a	11.42 ± 1.75 b
6	1635	Amide I: protein C=O stretching, N––H bending	271.64 ± 18.9 4	171.21 ± 11.47 a	98.61 ± 6.84 a	322.68 ± 12.36 a	325.66 ± 13.06b	307.02 ± 9.69 a
7	1541	Amide II: protein, N-H bending, C––N stretching	61.71 ± 9.46	17.38 ± 1.37 a	1.42 ± 0.73 b	145.35 ± 10.65 a	79.72 ± 4.87 b	94.89 ± 16.37 a
8	1462	CH2 scissoring: lipids	62.31 ± 4.82	23.56 ± 2.08 a	1.78 ± 0.77 b	34.92 ± 2.04 a	18.31 ± 1.26	14.32 ± 1.07 a
9	1458	CH2 bending: lipids	79.82 ± 6.92	23.56 ± 3.05 c	1.78 ± 0.45 b	8.05 ± 0.79 b	21.87 ± 2.52c	11.77 ± 1.83 b
10	1396	COO− symmetric stretching: fatty acids	156.38 ± 9.49	32.19 ± 2.04 b	4.65 ± 0.92 a	34.32 ± 1.57 a	43.74 ± 2.35 b	31.26 ± 2.02 a
11	1339	PO2− antisymmetric stretching	29.89 ± 1.47	3.82 ± 0.68 b	1.94 ± 0.06 a	19.22 ± 1.62 a	3.61 ± 0.58 b	2.14 ± 0.08 b
12	1316	Amide III	14.98 ± 1.30	23.61 ± 3.31 a	9.96 ± 0.72 a	41.98 ± 3.72 b	5.23 ± 0.25 a	0.78 ± 0.06 a
13	1233	PO2− antisymmetric stretching: mainly nucleic acids with minor contribution from phospholipids	21.17 ± 3.52	22.71 ± 2.79 b	35.08 ± 1.03 a	69.08 ± 5.07 b	92.79 ± 4.99 a	58.34 ± 2.04 a
14	1174	CO––O––C antisymmetric stretching: phospholipids	46.74 ± 2.47	48.62 ± 1.96 c	8.52 ± 0.75 a	54.02 ± 2.68 a	105.42 ± 8.42 a	63.85 ± 3.08 a
15	1112		34.72 ± 1.37	32.14 ± 1.57 a	0.63 ± .04 a	42.69 ± 1.73a	39.47 ± 1.38a	24.74 ± 1.02a
16	1079	PO2− symmetric stretching: nucleic acids and phospholipids	27.45 ± 2.53	8.41 ± 0.69 b	1.40 ± 0.06 a	21.15 ± 3.71 c	32.29 ± 4.02 a	13.58 ± 1.38 a
17	1059	C–O stretching: polysaccharides	59.57 ± 2.41	91.94 ± 4.07 a	1.27 ± 0.04 a	18.96 ± 1.12 b	19.56 ± 1.40 b	5.43 ± 0.06 a
18	994	Ribose skeleton	4.85 ± 0.04	7.52 ± 0.85 a	5.25 ± 0.34 a	17.99 ± 0.04 a	5.63 ± 0.67 a	3.45 ± 0.64 a

^ap < 0.05; ^bp < 0.01; ^c p < 0.001 with respect to control (n = 3)

Toxicity impact of heavy metals Cu and Zn and their combinations in the lipid region

FTIR spectra in the 3100–2800 cm⁻¹ region are generally dominated by the spectral characteristics of various membranes of phospholipids. We studied the effect of extracellular Zn and/ or Cu metal ions that can exert significant and potentially deleterious effects on bacterial cell membranes composition and integrity. This happens through the variation of excess unsaturated fatty acids, which affects membranes fluidity [17]. Hence, we examine these metal ions stress on the fatty acid content of S. aureus. Our analyses show that S. aureus has a change in the band area of fatty acid compositions (CH₃/ CH₂ symmetric vibrations) as shown in Fig. 1a, b. The band area in the selected region 2900–2800 cm⁻¹ shows changes in lipids due to toxicity compared with control. The reduction in CH₃ asymmetric band area of lipids arises for Cu and Zn metal treatment of S. aureus compared with control. For metal mixtures treatment (Cu: Zn) lipids bands decrease showing combined toxicity is greater than individual toxicity. The computed band area of Lipid/Phosphorylation (Fig. 4) shows a substantial reduction in lipid contents due to Zn treatment (p < 0.001). Goswami et al. [18] studied using FTIR and showed that Cu and Zn toxicity resulted in lipid contents reduction in bacterial strain Rhodococcus opacus. Rafig Gurbanov et al. [19] studied similar studies of Pb & Cd intoxication on E.Coli and S. aureus bacterial populations. Our reports (Fig.1a) in the band area of 2930 cm⁻¹, 2875 cm⁻¹, and 2850 cm⁻¹ are well in good agreement with the literature discussed above. This shows a decrease in lipid content because of Cu and Zn intoxication. Our study shows that Zn toxicity has a higher effect on lipid contents compared to Cu. Further, the Cu and Zn interaction at various ratios 1:1, 1:2, 2:1 shows band area (Fig.1b) reduction resulting in, increase in toxicity.

Fig. 4.

Mean computed band area ratio of (a) Lipid/Phosphorylation (b) Nucleic acid/Protein (c). Lipid/Nucleic acid (d). DNA/RNA/ Nucleic acid (*p < 0.05, ** p < 0.01, ***p < 0.001) with respect to control

Hassan et al. [17] studied Zn + Cu metal ion results in a change in the composition of fatty acid in A.baumannii. The results of our experiments show that Cu with the addition of Zn (Cu:2Zn) causes a decrease in lipid content. This shows Zn exhibits synergism effect with Cu. Further, the band area of lipid/phosphorylation ratio confirms the decreases in lipid significantly (p < 0.001) indicating that Zn exhibits greater toxicity than Cu. Thus, the toxicity influence of Cu depletion in Zn stressed cells is because of failure to acquire enough amounts of Cu. This predicts cytoplasmic Cu accumulation of bacterial populations that also recognize other cations Zn²⁺ [17]. This further established changes in the phosphate band discussed in section 3.4. This complimentary information obtained in region 1470–1400 cm⁻¹ shows the lipid bands at 1462 cm⁻¹, 1458 cm^-1, and 1396 cm⁻¹ corresponds to CH₂ scissoring, CH₂ bending, and COO⁻ of fatty acids. The reduction in this band area is because of metal intoxication. This agrees well with other lipid regions (3100–2800 cm⁻¹) discussed in Section 3.2. Hassan et al. [17] obtained similar results due to Zn + Cu treatments showing a synergistic impact on A. baumannii. Similarly, the band emerging from the COO⁻ symmetric of fatty acids (~1396 cm⁻¹) decreases in Cu and Zn treatment of S. aureus. The Presence of Zn/ and Cu or in combination causes a fatty acid reduction. These changes in the fatty acid will influence membrane fluidity. It is proposed to be one of the mechanisms of bacterial adaption to extracellular stress [20].

Toxicity impact of heavy metals Cu and Zn and their combinations in the protein region

Amide bands dominating in the region 1600–1500 cm⁻¹ are used for protein changes confirmation (Fig. 2a, b). Any decrease/increase in the band area is related to protein changes. The band area measured shows a high reduction in protein content due to Zn toxicity compared to Cu (P < 0.01). This may well support an increase in the binding of zinc with protein metallothionein. This regulates zinc level distribution in the microbes [21]. Further the combined toxicity effects resulted in the reduction of proteins when compared with individual toxicants. In bacteria, Zn, or Cu binds with cysteine-rich polypeptides in bacterial resistance [22]. The Cu and Zn combinations (Cu:2ZN), surprisingly increases toxicity showing a band area of protein. Hassan et al. 2017 showed accumulation of Zn resulted in depletion of Cu in A baumannii. Bart A. Eijkelkamp et al. [23] studied that in Streptococcus pneumonia Zn (II) inhibits Mn (II) by binding with protein leading to increased susceptibility to oxidative stress. Our results show a change in Amide bands (I, II, III) of proteins playing an effective role in the defence mechanism. This decrease in the Amide band of proteins is due to inhibition of Cu (II) ions by the peptide. This forms a complex formation of Cu bonding to side peptide as described by Tsuneo Ishida [24]. The bioaccumulation study shows the accumulation of more Cu. The computed BCF values show a higher value as against Zn (Fig. 5). This shows the complex binding of the peptide chain by Cu in S. aureus. Besides, a copper transport mechanism (copA and copB) determines uptake of these metal ions which favours bioaccumulation [25]. The order of preference for cation binding to metallothionein is high for Zn²⁺ and low for Cu²⁺. This probably explains a higher reduction in amide bands due to Zn toxicity. The presence of Zn abundance with Cu (Cu:2Zn) increases in the BCF value of Cu showing higher toxicity (Fig. 5). This high BCF probably explains the role of Cu in the periplasmic space that is considered as a covering membrane. In S. pneumoniae excess, Zn (II) antagonizes Mn (II) by protein binding mechanism reported by McDevitt et al. [26].

Fig. 5.

Computed Bioconcentration factor (BCF) and Metal removal efficiency of S. aureus populations exposed to Cu, Zn and metal mixtures

Toxicity impacts of heavy metals of Cu and Zn and their combinations in the carbohydrate region

The Fig. 3a, b shows the FTIR spectra of secondary derivative vector normalized in the regions 1200–950 cm⁻¹. The carbohydrate regions are dominating in 1200–950 cm⁻¹ region. The measured band areas of the selected region are shown in Table 1. It is observed that the band area of S. aureus treated to metal toxicity increases for all treatment except for Cu:2Zn. The toxicity impact of Cu depletion in Zn stressed cells could be due to a failure of the cells to acquire enough amounts of Cu [17]. This may be because of cytoplasmic Cu accumulation of the bacterial systems that recognize other cations Zn. This confirms a decrease in phosphate band 1450 cm⁻¹ observed at (2Zn: Cu). Further, the changes in the carbohydrate band area are influenced by Cu (II) ions on S. aureus owing to the damages of PGN as suggested by Tsuneo Ishida [22]. Our result confirms the changes in carbohydrates as studied from FTIR spectra. Similar results were reported by Joana Campos et al. [27] using FTIR in Salmonella enterica bacteria showing changes of phospholipids/DNA/ RNA with the sharp changes in the polysaccharide region. Similarly, Tom Grunert et al. [5] studied distinct spectral variation in 1200–800 cm⁻¹ of S. aureus. Studies reported by Whelan et al. [28, 29] and Rafig Gurbanov et al. [30] shows how FTIR is used to monitor changes in nucleic acid bands by changes in DNA confirmation. The antisymmetric PO²⁻ gives information regarding nucleic acid which appears at 1233 cm⁻¹. This is also confirmed in our study with Cu-Zn metal interaction where the changes in band area is predominant when compared to control. This confirms interaction of Cu-Zn resulted in greater toxicity. Hence our results are in agreement with the other reports using FTIR to monitor DNA changes due to metal toxicity in E. coli and S. aureus bacterial populations [31, 32].

The complex formation of metal is important in the removal of toxic metals from the environment. It is well known that EPS are polysaccharides that are secreted by bacterial to overcome the metal toxicity. The measured band 1059 cm⁻¹ corresponds to C-O stretching of polysaccharides. The decrease in band area resulted due to toxicity of these metals. This may be due to metal binding cation with sugar polymers resulting in changes in the band area [30–32]. Increase in band area due to Cu toxicity resulted in the production of extracellular polymers growth. Ferreira et al. [33] found similar results of extracellular polymers due to Cu, Zn, and Cd toxicity. The measured band area ~ 1059 cm⁻¹ agrees with the results for Cu treatment against Zn treatment.

Toxicity impact of heavy metal and its mixture in the phosphate region

The band that arises in the 1250–950 cm⁻¹ (Fig. 3a, b) region corresponds to the nucleic acid composition of S. aureus treated to various metals. Our FTIR study supports the changes in 994 cm⁻¹ where the band area of Zn treatment has a significant reduction compared to Cu treatment. The band area measurement shows a 26% reduction in band area for DNA/RNA/ nucleic acid due to Cu:2Zn treatment. This shows Cu:2Zn treatment has greater toxicity affecting the DNA backbone of C-C stretching resulting in decreased band area (Fig. 3b). This result was supported by Guerra et al. [34] and Sanson et al. [35] in the case of B. subtilis where Zn (II) binds with high affinity to the DNA. Binding of zinc triggers an impairment of DNA but allows metallothionein production thereby resulting in cell survival at higher zinc levels. Similarly, Obinaju et al. [16] studied using FTIR and observed changes in DNA/RNA, carbohydrates bands due to toxicity bebzo[a]pyren on the bacterial cell population. The interaction of bacterium with metals is enveloped by S-layer of proteins in which phosphate groups on the cell wall bind closely. The main interaction mechanism of metals with a bacterial population is binding to organic phosphate and COO⁻ group. Merroun et al. [36] showed that carboxyl groups of S layer participate in binding of palladium which was shown by EXAFS and FTIR spectra.

Secondary structural study of protein in amide I band analysis

The Fourier self-deconvolution applied in the amide I regions reveals the existence of four bands as shown in Fig. 6a-f. Table 2 corresponds to the band area measurement of the secondary structure of proteins for control, Cu, Zn, and metal combination treatments. The band ~1677 cm⁻¹ is assigned to ß turns of proteins. The α helix structure is assigned at 1656 cm⁻¹. Random coil and ß sheet secondary structure of proteins corresponds to 1635 cm⁻¹ and 1616 cm⁻¹. The Fig. 6a-f shows the secondary structure variation due to various treatments of metals. Among the various treatments examined studied Cu toxicity results in higher changes in ß sheet formation. She et al. [37] studied the secondary structural variation of protein due to the magnetic field on the bacterial population. The metal interaction Cu: Zn also shows significant changes in ß sheet formation. It is observed from our study that an increase in ß sheet and decrease in an α helix occurs for all treatment except for 2Cu:Zn treatments. Also, the random coil increases for this treatment. These results confirm the structural changes occurring in protein. This result confirms 2Cu:Zn treatment shows that Cu exhibits an antagonism effect.

Fig. 6.

Fourier self deconvolution deduced by curve fitting in Amide region (a). Control (b) Cu treatment (c) Zn treatment (d) Cu:Zn treatment (e) 2Cu:Zn treatment (f) Cu:2Zn treatment

Table 2.

Secondary structural variation of protein in the Amide I regions of S. aureus population due to various metal treatments

Band Assignment	Spectral Range Cm−1	% Band area
		Control	Cu	Zn	Cu:Zn	2Cu:Zn	Cu:2Zn
1680–1690	β turns	8.36 ± 0.42	5.46 ± 0.36 a	5.93 ± 0.43b	9.84 ± 0.68 a	5.64 ± 0.45 a	5.94 ± 0.61 a
1660–1650	α helix	43.49 ± 1.05	15.18 ± 0.27 b	45.38 ± 1.12 a	22.46 ± 0.86 a	31.41 ± 0.94 a	59.22 ± 1.07 a
1645–1630	Random Coil	41.42 ± 0.78	57.43 ± 1.04 a	41.26 ± 0.86 a	53.95 ± 1.32 a	53.86 ± 1.04 b	33.25 ± 0.86 a
1610–1635	β sheet	7.32 ± 0.41	21.27 ± 1.08 a	7.72 ± 0.66 a	13.93 ± 0.79 b	9.28 ± 0.53 a	1.86 ± 0.06 a

^ap < 0.05; ^bp < 0.01 with respect to control (n = 3)

Principal component analysis of the S. aureus sample subjected to heavy metal treatments

Figure 7 shows the PCA results obtained for various metals treatment of S. aureus. The plots indicate the control and treated samples are well separated according to eigenvalues. It shows that all the components are plotted in the positive region. The highest absolute eigenvalue relates to components 1 of 52% variation. Component 2 has 26% variation. The 3rd component is neglected because of the least variation and low eigenvalues. This PCA was obtained from absorption values in the entire region. The plot shows that control has the least value along PC1 whereas Cu and Zn treatment has high value along PC2. For metal interaction (Cu:Zn) of various ratios shows the highest positive value along PC1 and they are distinctly separated from control samples. FTIR spectra support this distinct variation due to various metal treatment studied in our case. The Cu:2Zn treated sample is predominant in influencing the changes studied followed by other combination ratios. This variation is because of changes in lipids, proteins, polysaccharides, and nucleic acid. Loading plots (Fig. 8) facilitate prominent frequency change that enables most of the variables studied. From the loading plot, it is seen that the amide region of proteins has a higher value (Fig. 8). The loading plots show a prominent peak responsible for discrimination of the samples treated to various Cu, Zn metal treatments. The loading plot shows PC1 has a higher value occurring in protein which accounts for maximum variation in the samples studied. Carbohydrates, fatty acids show the next important factors influencing the variation as observed from PC2 loading values. This account for the distinct grouping of S aureus treated due to metal treatments. This alteration and adaptation in the S. aureus are supported by bimolecular changes as revealed from the variation in the score plot. Similar studies were made by Kepenek et al. [38], Partouche et al. [39], Gupta et al. [9], and Kochan et al. [40] where PCA was employed on FTIR spectral data to determine the variability of a bacterial population.

Fig. 7.

PCA scatter plots of S. aureus populations of control, Cu, Zn, and metal mixture treated samples

Fig. 8.

Variation of the factor loading obtained from the PCA with the corresponding wavenumber of the S. aureus populations

Hierarchical cluster analysis (HCA)

An understanding of the S. aureus under different treatments are subjected to hierarchical clustering analysis using squared Euclidean distance method. They are displayed graphically (Fig. 9) as a dendrogram, by the Ward algorithm which relates to the dissimilarity scale. Similar cases are related by smaller distances which are visualized by closer separations. The distance present in the dendrogram shows how the composition of samples differs within an organism [14]. The first cluster corresponds to toxicity due to Cu-Zn metal interaction. The second cluster has two well-differentiated sub-clusters of Cu and Zn treatment. The third cluster results from control and Cu:2Zn treatment. This differentiation due to metal interaction results from the toxicity of Cu, playing a dominant role compared to Zn. The fourth cluster is well separated from individual metal treatments.

Fig. 9.

Dendrogram of S. aureus bacterial populations treated with various Cu, Zn, and metal mixtures treatments showing group linkage obtained using the Ward algorithm

Analysis of reproducibility

The reproducibility of the sample replicates is measured by the parameter differentiation index D [13, 41]. Using correlation coefficients, the differentiation index (D value) of the identical spectra can be computed given by the relation [41]. The larger the deviation between the spectra, the greater the index values, thereby reaching the completely non-correlated spectra. The spectral region was separated as w₁ fatty acid, W₂, protein, and W₃ carbohydrates and nucleic acid groups. The window W₁ reflects changes in fatty acids. The window W₂ displays variation in the amide groups of proteins. The mixed regions W₃ detects the influence of polysaccharides and nucleic acid groups. These spectral windows are used to obtain the best identification of inter replicates variability as much as possible. Mean D values below 10 are considered good for strain reproducibility. The computed D value for the samples subjected to different metal treatments is represented in Table 3. The full spectra region is also considered to measure the reproducibility variations among the samples studied. In our study w₂ shows the highest D values. The windows W₁ and W₃ have a lower value (Table 3). The W₂ comprises of proteins region used for differentiation [4, 9, 41–43]. To check the reliability of our work the receiver operating characteristic curve was obtained to measure the diagnostic of our findings.

Table 3.

Differentiation index of S. aureus strain exposed to metal toxicity and its mixtures calculated for the 3 spectral windows and for the whole spectrum

Sample Treatment	Spectral window			Whole spectra
	W1	W2	W3
Control	0.36	4.78	2.70	1.63
Cu	2.68	4.90	4.41	5.71
Zn	7.64	2.17	1.02	4.94
Cu:Zn	4.54	1.29	3.47	6.97
2Cu:Zn	8.67	2.68	7.42	4.70
Cu:2Zn	2.28	2.50	1.84	1.75

Receiver operating characteristic analysis

The ROC curve represents the sensitivity versus 1-specificity for all possible values of the experimental values. It gives information about the cut-off point of various samples studied. The diagnostic ability was measured by the area of the curve. The obtained ROC curve for various metal treatments lies in the higher range showing the reliability of our study (Fig.10). The highest ROC curve area was obtained for Cu:2Zn and 2Cu:Zn treatments. A similar conclusion was made by Sharaha et al. [44] using the ROC model from infrared spectroscopy to determine the accurate classification analysis of the S. aureus bacterial population. The measured Youden index ranges from 0.82 to 0.72 for all treatments. The highest Youden index is measured for Cu treatments. It gives the response and validation of the study.

Fig. 10.

Receiver operating characteristic curve for Cu, Zn and metal mixtures treated on the S. aureus populations

Conclusion

We examined the various molecular variations in the FTIR spectra due to Cu and Zn toxicity on S. aureus. Our study provides information that the toxicity mechanism of Cu is more pronounced than Zn. Further, the combined toxicity level Cu:2Zn shows increased toxicity resulting in changes at polysaccharides and protein levels. This shows the synergism exhibited among the metals studied. The spectral variation of S. aureus, resistant to metal treatment shows the changes occurring in the carbohydrates and protein regions. S. aureus bacteria produce cellulose-rich polysaccharide which was observed from FTIR spectra. This confirms binding occurs in this region. However, the secretion of EPS depends on the availability of nutrients and morphology of bacteria studied. This may be due to uptake or efflux of the metal ions from the cells. This results in the translocation of Zn and/or Cu from the cytoplasm in S. aureus and depends on the particular metal concentration present in the bacterial strain. Secondary structural changes of protein reveal an increase in ß sheet and a decrease in α helix due to metal treatments. PCA plots show the possible distinct variation of the S. aureus samples studied. The selected spectral range observed helps in evaluating the toxicity due to metal interaction on S. aureus. The ROC curve and the Youden index values validate our experimental study. This work shows how S.aureus can resist stress due to Cu-Zn metal mixtures in the environment. The study will help in understanding the capability of S. aureus in the defence mechanism due to Cu and/or Zn metals. The study gives an overview of the interaction mechanism of complex metal mixtures with regard to the molecular changes using FTIR coupled with chemometric techniques.

Funding

This research does receive any funding from government or private bodies.

Compliance with ethical standards

Conflict of interest

There is no conflict of interest in this research work as declared by authors.

Ethical approval

This study does not contain any animal model/experiment and hence does not require any ethical approval.