Introduction

Critical (rare) metals (Chakhmouradian et al. 2015), including antimony (Sb), beryllium (Be), chromium (Cr), cobalt (Co), gallium (Ga), germanium (Ge), indium (In), lithium (Li), niobium (Nb), tantalum (Ta), the platinoids (such as Ru, Pd, and Pt), the rare-earth elements (REEs), and tungsten (W), are becoming increasingly expensive. The increasing cost of these materials has been attributed to their growing use in many modern industrial technologies, such as the production of catalytic filter neutralizers for exhaust gases, fiber optics, lasers, oxygen sensors, phosphors, and superconductors. Moreover, numerous biological (trace) metals are essential or beneficial for life (i.e., microorganisms, plants, and animals), including arsenic (As), boron (B), calcium (Ca), Cr, Co, copper (Cu), iron (Fe), magnesium (Mg), manganese (Mn), molybdenum (Mo), nickel (Ni), potassium (K), selenium (Se), silicon (Si), sodium (Na), tin (Sn), vanadium (V) and zinc (Zn). Critical metals are generally regarded as being unconnected to biological processes in living systems. However, in extreme environments, such as the deep-sea sediments from the critical metal-rich crust/nodule deposit regions of the Pacific ocean (Huo et al. 2015; Liao et al. 2011), several interesting interactions have been observed between critical metals and microorganisms. The microbial communities found in these environments are well adapted to the extreme conditions and contain many specific microorganisms (bacteria, archaea) bearing uncommon structural surface polymers and cell metabolism processes.

Some naturally occurring polymers bearing negatively charged residues can function as ion-exchange (chelating) resins in a similar manner to synthetic polymers (e.g., polyacrylate) (Beauvais and Alexandratos 1998). These naturally occurring systems could therefore be used in environmentally friendly biochemical processes and applications. Natrialba aegyptiaca, which is an extremely halophilic archaeon, produces extracellular poly-γ-glutamate (PGA) (Hezayen et al. 2001). PGA (Fig. 1a) possesses a nylon-like backbone bearing numerous polyacrylate-like carboxylate groups, making it highly soluble in water. PGA can form multiple (e.g., non-covalent crosslinking) interactions with certain cationic compounds, thereby suppressing its extreme hydrophilicity (Ashiuchi et al. 2013). This archaeal polymer is composed entirely of l-glutamate residues and is also called stereoregular L-PGA, making it different in stereochemistry to the DL-PGA of Bacillus subtilis (Ashiuchi et al. 2015). Yamasaki et al. (2010) revealed the extremolyte-like (versatile) function of archaeal L-PGA. However, it remains unclear whether water-soluble PGA could be used to adsorb heavy metals, including the critical elements. Adsorption is generally defined as the transfer of a solute (e.g., metal ion) from a liquid phase to a solid phase via the formation of specific interactions with an absorbent (e.g., functional polymer) (Sampranpiboon et al. 2014). However, there is very little information available pertaining to the mode of adsorption of different metal ions by free (water-soluble) polymers (Fig. 2), compared with immobilized (water-insoluble) polymers derived from biomasses and cell debris (Das et al. 2008; Ogi et al. 2012; Vijayaraghavan and Yun 2008).

Fig. 1
figure 1

Chemical structures of a poly-γ-glutamate (PGA), b polyacrylate (PAC), and c polyvinyl alcohol (PVA). Given that the molecular weight of archaeal L-PGA is on average over 800,000, its number-average degree of polymerization of (a), i.e. its n value, was estimated to be >6200

Full size image
Fig. 2
figure 2

Schematic diagrams of chemical adsorption using water-insoluble (a) and water-soluble (b) polymer materials for metal ions. In this case, the polymer moieties of a and b are negatively charged; metal ions are expressed by positive particles

Full size image

Herein, we present the adsorption capacities of L-PGA and two water-soluble synthetic polymers, i.e. polyacrylate (PAC) and polyvinyl alcohol (PVA) (Fig. 1b and c), towards six industrially valuable metal ions (Co2+, Ni2+, Mn2+, Ga3+, In3+, and Dy3+). The adsorption profiles of these polymers were also analyzed using the non-ideal competitive adsorption (NICA) model (Foo and Hameed 2010; Ringot et al. 2007).

Materials and methods

Materials

Sodium polyacrylate (PAC; polymerization degree (Mn), 2700–7500), polyvinyl alcohol (PVA; Mn, 3100–3900), CoCl2·6H2O, NiCl2·6H2O, and MnCl2·4H2O were obtained from Wako (Osaka, Japan). GaCl3 and InCl3 were purchased from Tokyo Chemical Industry (Tokyo, Japan). Safranin O and DyCl3·6H2O were purchased from Sigma (St. Louis, MO, USA). A QMA anion-exchange cartridge and 4-(2-pyridylazo) resorcinol (PAR) were purchased from Millipore (Bedford, MA, USA) and Dojindo (Kumamoto, Japan), respectively. All of the other chemicals used in this study were purchased as the analytical grade.

Preparation of archaeal L-PGA

N. aegyptiaca was inoculated into 200 ml of S medium (pH 7.2) comprising 25 % NaCl, 0.2 % KCl, 1 % trisodium citrate, 1 % yeast extract, and 0.75 % casamino acids, cultured at 37 °C for 4 days, before being harvested by centrifugation (8000×g, 30 min) at room temperature (~25 °C). The harvested cells (0.4 g) were cultured on a 2 % agar plate (145 mm diameter) of S medium (50 ml) at 37 °C for 2 weeks. Extracellular PGAs were collected along with growing cells from the solid-culture system using 10 ml of 25 % NaCl solution, and centrifuged at 8000×g for 30 min. The resulting supernatant was injected onto QMA anion-exchange cartridge (Park et al. 2005), which was washed with 10 ml of distilled water, followed by 5 ml of 0.5 M NaCl solution. L-PGA (Mw, ~800,000; Fig. 1a) was subsequently eluted with 5 ml of 0.7 M NaCl solution.

Examination of PGA

PGA was examined as follows. A small samples of the polymer (100 µl) was diluted with 200 µl of 60 mM citrate buffer (pH 6.0), and the resulting mixtures was treated with 200 µl of 0.1 % safranin O solution, followed by 500 µl of saline and incubated for 10 min at room temperature. The mixture was then centrifuged at 12,000×g for 10 min to remove the water-insoluble safranin-PGA complex and the supernatant was collected and diluted 50-fold with saline. The residual concentration of safranin was determined using a Shimadzu UV-1600 spectrophotometer (Kyoto, Japan). For data standardization, the reagent blanks (lacking only safranin, normally zero) and the reactant blanks (lacking only PGA) were also analyzed. The calibration curve used to determine the PGA concentration was as follows: y = 0.348x, where y represents a decrease in absorbance at 520 nm relative to the reactant blank and x corresponds to the polymer concentration (mg ml−1). Notably, the “x” value was linear in the range of 0.07–0.7 mg ml−1 (R2 = 0.9946).

Examination of metal ions

Sample solutions of the different metal ions (100 µl) were mixed with 200 µl of 0.5 M citrate buffer (pH 6.8) and diluted with 500 µl of distilled water. The diluted solutions were added to 0.1 % solution of PAR in 1 M aqueous NaOH, which can form a colored chelating complex with various metal ions (Ghasemi et al. 2007), including Co2+, Ni2+, Mn2+, Ga3+, In3+, and Dy3+. The solutions were then incubated for 10 min at room temperature, and centrifuged at 12,000×g for 10 min. The concentrations of the different metal ions in the resulting supernatants were estimated spectrophotometrically. The calibration curves for the different metal ions were tested as follows:

  • Co2+ y = 0.589x (0.13 ≤ y ≤ 1.2; R2 = 0.9989);

  • Ni2+ y = 0.589x (0.13 ≤ y ≤ 1.2; R2 = 0.9985);

  • Mn2+ y = 0.122x (0.024 ≤ y ≤ 0.25; R2 = 0.9952);

  • Ga3+ y = 1.04x (0.021 ≤ y ≤ 0.22; R2 = 0.9989);

  • In3+ y = 2.91x (0.060 ≤ y ≤ 0.64; R2 = 0.9932);

  • Dy3+ y = 0.340x (0.089 ≤ y ≤ 0.69; R2 = 0.9925),

where y represents an increase in absorbance at 510 nm relative to the reactant blank (lacking only the indicated metal ions) and x corresponds to the metal concentration (mM).

Metal adsorption tests

The reaction mixture used for the metal adsorption test consisted of a metal ion (as a target) and an adsorbent polymer. The final pH values of the different mixtures are shown in Supplementary Fig. S1. Based on the detectability of the PAR assay (see above), the highly sensitive metal ions (Co2+, Ni2+, and In3+) and several other metal ions (Mn2+ and Dy3+) were examined at concentrations of 0.035–1.8 and 0.14–2.1 mM, respectively. For gallium, it is noteworthy that a high concentration of Ga3+ (>1 mM) can lead to the spontaneous formation of undesirable aggregates in the solution, making it impossible to examine the adsorption capacities of L-PGA, PAC, and PVA at Ga3+ concentrations >1 mM. During these examinations, L-PGA, PAC, and PVA were tested as metal adsorbent candidates, and 0.05 % solutions of these polymers (pH ~7) were added to solutions of the different metals. The final concentration of L-PGA, PAC, and PVA were maintained at final concentrations of 0.09, 0.05, and 0.03 mg ml−1 (for Co2+, Ni2+, Ga3+, and In3+) and 0.34, 0.19, and 0.11 mg ml−1 (for Mn2+ and Dy3+), respectively, so that the concentrations of the carboxyl groups from L-PGA and PAC, and the hydroxyl groups from PVA were the same (i.e., 0.7 mM for Co2+, Ni2+, Ga3+, and In3+; 2.8 mM for Mn2+ and Dy3+). After being incubated for 10 min at room temperature, the mixtures were filtrated using a 0.2-µm membrane filter, and subjected to the PAR assay. The critical (rare) metal adsorption capacities of the polymers were determined according to the following equation: q e = (m i − m r)/p, where q e is the adsorption capacity (mmol g−1), m i and m r correspond to the initial and the residual amounts of metal in the reaction mixture being tested (V = volume of the mixture), and p represents the initial loading polymers. Alternatively, the q e value can be determined using the initial and equilibrium metal concentrations (µM), which are also known as C i (=m i/V) and C e (=m r/V) (Sampranpiboon et al. 2014).

Equilibrium studies

Adsorption equilibrium is established when the amount of solute being adsorbed (q e) is equal to the amount being desorbed; where the equilibrium concentration (C e) of the solute is becomes constant. In this study, we used the NICA model (Koopal et al. 1994; 2001) to evaluate the adsorption equilibria of the different polymers. This model uses Eq. (1), which is similar to the Hill equation (Hill 1910). Notably, this equation assumes that adsorption is a cooperative phenomenon, which is dependent on the ability of solute binding at one site on a polymer to influence solute binding at a different site on the same polymer.

$$q_{e} = \frac{{Q_{0} C_{e}^{n} }}{{K_{\text{D}} + C_{\text{e}}^{n} }}$$
(1)

This equation can be linearized using logarithmic terms (Eq. 2):

$$\log \left( {\frac{{q_{e} }}{{Q_{0} - q_{e} }}} \right) = n\log C_{e} - \log K_{\text{D}}$$
(2)

The K D corresponds to the Hill constant, which can be related to the affinity of adsorption sites: K D = \(K_{\text{d}}^{n}\), where K d is the dissociation constant per site (or residual solute concentration at half saturation). The term “n” represents the Hill coefficient of the cooperative interaction between the polymer molecules during the adsorption. Theoretically,

  • n > 1, positive cooperativity;

  • n = 1, non-cooperativity (or hyperbolic adsorption);

  • n < 1, negative cooperativity.

Specifically, during hyperbolic adsorption (n = 1), this expression becomes almost identical to that of Langmuir model (Langmuir 1916) (Eq. 3):

$$q_{e} = \frac{{Q_{0} C_{e} }}{{K_{\text{d}} + C_{e} }}$$
(3)

The linearized form of the Langmuir equation is shown below (Eq. 4).

$$\frac{1}{{q_{e} }} = \frac{{K_{\text{d}}}}{{Q_{0} }}\frac{1}{{C_{e} }} + \frac{1}{{Q_{0} }}$$
(4)

The K d value is also the reciprocal of the association constant K a. Under the analytic conditions, the K a value was generally referred to as Langmuir constant. Furthermore, the Q 0 value indicates that maximum specific uptake corresponds to the saturation of the adsorption sites. Recent studies pertaining to the interactions formed between metal ions and anionic polyelectrolytes (e.g., PAC) have indicated the presence of different carboxylate–metal cation stoichiometries, including 2:1 and 3:1 for divalent and trivalent ions, respectively (Kogej et al. 2013). Accordingly, the theoretical Q 0 values (mmol g−1) of L-PGA, PAC, and PVA were calculated to be 3.88, 6.98, and 0 (for divalent ions); 2.58, 4.65, and 0 (for trivalent ions), respectively.

Results and discussion

Several studies have been published pertaining to the chemical adsorption of divalent cations by water-insoluble (solid) materials bearing carboxylate moieties (Ewecharoena et al. 2009; Kwiatkowska-Marksa and Wójcika 2014; Wu and Li 2013). However, the water-soluble polymers tested in the current study showed apparently little or no adsorption of Co2+, Ni2+ and Mn2+ (Fig. 3), indicating difficulty in the spontaneous transfer of the divalent metal ion-carboxylate polymer complexes into the solid phase. This difficulty could be attributed to negligible conformational changes to their separable (water-insoluble) complexes (Fig. 2b). In contrast, L-PGA and PAC showed good potential for the adsorption of Ga3+, In3+ and Dy3+ (Fig. 3), suggesting that their complexes to trivalent cations could be readily separated from aqueous solutions without using ultra- and nanofiltration procedures (Rivas et al. 2009). These results therefore imply that anionic polyelectrolytes exhibit beneficial valence-dependent selectivity (viz., water-soluble polymers bearing multiple carboxylates) for metal ions. We have also confirmed that PVA does not adsorb metal ions under the conditions tested in the current study (see Fig. S1), because its hydroxyl groups are weakly acidic and would be deprotonated only at pH values much higher than 7 (Yokoi et al. 1986). It is well known that some critical metal ions are solubilized and stabilized in acidic solutions. For example, leached In3+ is converted to undesirable hydroxide precipitates at pH ~5 (Sawai et al. 2015). The crucial role of carboxyl groups in the adsorption of crucial (rare) metal ions therefore cannot be replaced by hydroxyl groups (Fig. 1). Although PVA is essentially a non-ionic polymer, it was used as a negative control in this study.

Fig. 3
figure 3

Adsorption plots of a Co2+, b Ni2+, c Mn2+, d Ga3+, e In3+, and f Dy3+ for PGA (circles), PAC (triangles), and PVA (squares). Data have been given as the mean values (symbols) ± the standard deviations (bars) of 12 independent tests

Full size image

Breakthrough has been required to date for developing high-performance adsorbent polymers that can maintain their functionality even under acidic conditions. Although PAC is generally considered to be a good chelating polymer, the pK a value of its carboxyl groups was calculated to be ~ 4.8 (Rivas et al. 2009), implying incomplete functionalization during the adsorption of some critical metal ions, such as In3+. The potential of L-PGA as an adsorbent for trivalent metal ions (Fig. 3) therefore appears to be far superior to that of PAC, which has been studied extensively and used in a wide range of applications (Beauvais and Alexandratos 1998; Kogej et al. 2013). Interestingly, several recent studies (Moriyama et al. 2015; Parka et al. 2014) have reported that the pK a of B. subtilis PGA is 4.8, which is virtually identical to that of PAC. In contrast to chemically synthesized polymers, the naturally occurring polymer PGA possesses an unusual molecular mechanism that enables it to overcome the pH problems associated with the chemical adsorption of critical metal ions. Then, the adsorption profiles of Ga3+, In3+, and Dy3+ for L-PGA and PCA (Fig. 3, panels d, e, and f) was analyzed using the NICA isotherm model. As shown in Fig. 4, PAC exhibited non- or somewhat negative cooperative adsorption towards these metal ions, whereas L-PGA showed considerable positive cooperative adsorption against all of the metal ions. We also determined the K d values (μM) of L-PGA and PAC which were 103 and 541 (for Ga3+); 117 and 1290 (for In3+); and 352 and 5140 (for Dy3+), respectively. The binding sites in L-PGA therefore exhibited apparently much higher affinities for trivalent cations than those in PAC (see also Fig. S2). Based on these results, we concluded that the success of high-performance metal adsorption using L-PGA is heavily dependent on the ability of this polymer to adopt cooperative polymer architecture. As shown in Fig. 1, PGA can be considered as a PAC analogue, with a backbone structure consisting of multiple amide linkages. The development of a deeper understanding of the features of the amide linkages in PGA could therefore provide further insights into it use as an environmentally friendly bio-sorption material for the recovery of various metal ions. This would be especially useful for the recovery of metal ion that are in great demand, including those associated with specific industrial and analytical applications.

Fig. 4
figure 4

a NICA isotherm plots of Ga3+, In3+, and Dy3+ for PGA (circles) and PAC (triangles) and b calculations of Hill parameters

Full size image

Critical metals are being used in an increasing number of applications in medical (e.g., anti-infectious, anti-diabetic and anti-tumor agents) (Chan et al. 2014; Kelson et al. 2013; Levina and Lay 2011; Ma et al. 2013) and environmental research, as well as several other metal-dependent industries. This increase in the use of these materials therefore highlights the dramatic rise in the level of interest in the use of critical (or rare) metals for biology and biotechnology. However, our observation (Figs. 3 and 4) that L-PGA strongly adsorbs trivalent metal ions according to a sigmoidal curve (as does hemoglobin), while PAC describes a hyperbolic curve (as does myoglobin), still requires a physiological explanation. We recently discovered that a PGA-producing microorganism exhibits increased PGA productivity and growth rate in the presence of Ga3+ or Dy3+ (unpublished results). It is envisaged that these results will prompt considerable interest within the field of extremophiles to elucidate the physiological role of water-soluble PGA as a potent bio-adsorbent of multivalent metal ions.

Conclusion

Although the extremolyte-like (versatile) function of archaeal L-PGA has been revealed, it remains unclear whether this archaeal polymer can bind to critical (rare) metals. In the study, L-PGA was found to prefer trivalent metal ions (Ga3+, In3+, Dy3+) and to discriminate against divalent metal ions (Co2+, Ni2+, Mn2+), and these properties could be used to develop a selective polymer for the collection of valuable (multivalent) metal ions. In this sense, the fact that L-PGA afforded a sigmoidal curve for its metal adsorption represents a considerable advantage, because a process for metal desorption could also be established using facile manipulation steps, which would have very little impact on the adsorption capability of the polymer. In contrast, PAC exhibited hyperbolic behavior, which would be more suitable for the removal of metal ions rather than their collection. However, further work is required to improve the metal ion-binding performances of these materials. Overall, our report highlights the potential of L-PGA as a useful metal adsorbent.