Miniaturization of Anthracene-Containing Nonapeptides for Selective Precipitation/Recovery of Metallic Gold from Aqueous Solutions Containing Gold and Platinum Ions

: The separation and recovery of noble metals is increasingly of interest, in particular the recovery of gold nanocrystals, which have applications in medicine and industry. Typically, metal recovery is performed using liquid–liquid extraction or electrowinning. However, it is necessary to develop noble metal recovery systems providing high selectivity in conjunction with a one-pot setup, ready product recovery, and the use of dilute aqueous solutions. In prior work, our group developed a selective gold recovery process using peptides. This previous research showed that RU065, a nonapeptide containing an anthracene moiety (at a concentration of 2.0 × 10 − 4 M), is capable of selective reduction of HAuCl 4 to recover gold from a solution of HAuCl 4 and H 2 PtCl 6 , each at 5.0 × 10 − 5 M. However, peptide molecules are generally costly to synthesize, and therefore it is important to determine the minimum required structural features to design non-peptide anthracene derivatives that could reduce operational costs. In this study, we used RU065 together with 23 of its fragment peptides and investigated the selective precipitation/recovery of metallic gold. RU065 4–8 , a fragment peptide comprising ﬁve amino acid residues (having two lysine, one L-isoleusine, and one L-alanine residue (representing six amide groups) along with an L-2-anthrylalanine residue) provided an Au/Pt atomic ratio of approximately 8, which was comparable to that for the full-length original RU065. The structural features identiﬁed in this study are expected to contribute to the design of non-peptide anthracene derivatives for low-cost, one-pot selective gold recovery. the data from EDS-FESEM, we determined the most active miniaturized molecular structure derived from the original RU065 ( a ) and characterized its secondary structure using attenuated total reﬂection Fourier transform infrared (ATR-FTIR) spectroscopy. Inductively coupled plasma-optical emission spectroscopy (ICP-OES) was also employed to compare the residual metal ion concentrations of supernatants after the removal of metallic precipitates by centrifugation, in trials using the original RU065 ( a ) and the miniaturized peptides. We also examined the mechanism by which this peptide-based gold recovery system functioned, employing matrix-assisted laser desorption/ionization time-of-ﬂight mass spectrometry (MALDI-TOFMS). The ﬁndings of this study are expected to assist in the future design of non-peptidyl anthracene derivatives for low-cost, environmentally friendly, and selective recovery of gold.


Introduction
The separation and recovery of noble metals from industrial wastewater has recently received significant attention as a component of sustainable chemistry. Especially, recovery of metals from electronic waste and other sources may be preferable to standard mining practices on a per mass basis, because discarded electronics contain greater relative proportions of metals compared with ores [1]. Meanwhile, it is also important to note that electronic waste tends to contain noble metals at relatively low levels mixed with other metals [2]. In order to separate metal ions of interest, liquid-liquid extraction is a common technique where metals such as gold and palladium can be extracted from an aqueous phase by employing specific organic solvents [3][4][5][6][7]. This technique is commonly used but is labor intensive and requires organic solvents that can have environmental impacts. In recent years, new approaches to liquid-based separation of noble metals have been demonstrated. These include the use of ionic liquids supported on silica to extract platinum group However, peptide molecules are generally costly to synthesize and so it is important to determine the minimum requirements for the structural features of RU065 (a) for lowcost, selective gold recovery. This would contribute to the design of low-cost non-peptidyl anthracene derivatives. Therefore, the present study designed and synthesized 24 different anthracene-containing peptides, comprising the original RU065 (a) [27] and 23 of its fragment peptides (b-x, Figure 1C). These fragments were obtained by deleting amino acid residues one-by-one from both N-and C-termini of the original RU065 (a). Each of these compounds was applied to the selective precipitation/recovery of metallic gold from a mixture of dilute HAuCl4 and H2PtCl6 (5.0 × 10 −5 M each), using a peptide (a) or peptide fragment (b-x) at a concentration of 2.0 × 10 −4 M. The selectivity obtained from peptides (a-x) was examined by analysis of the precipitates using energy dispersive X-ray spectroscopy-field emission scanning electron microscopy (EDS-FESEM) after samples were collected by centrifugation to determine elemental contents and distributions. Based on the data from EDS-FESEM, we determined the most active miniaturized molecular structure derived from the original RU065 (a) and characterized its secondary structure using attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy. Inductively coupled plasma-optical emission spectroscopy (ICP-OES) was also employed to compare the residual metal ion concentrations of supernatants after the removal of metallic precipitates by centrifugation, in trials using the original RU065 (a) and the miniaturized peptides. We also examined the mechanism by which this peptide-based gold recovery system functioned, employing matrix-assisted laser desorption/ionization time-of- However, peptide molecules are generally costly to synthesize and so it is important to determine the minimum requirements for the structural features of RU065 (a) for low-cost, selective gold recovery. This would contribute to the design of low-cost non-peptidyl anthracene derivatives. Therefore, the present study designed and synthesized 24 different anthracene-containing peptides, comprising the original RU065 (a) [27] and 23 of its fragment peptides (b-x, Figure 1C). These fragments were obtained by deleting amino acid residues one-by-one from both N-and C-termini of the original RU065 (a). Each of these compounds was applied to the selective precipitation/recovery of metallic gold from a mixture of dilute HAuCl 4 and H 2 PtCl 6 (5.0 × 10 −5 M each), using a peptide (a) or peptide fragment (b-x) at a concentration of 2.0 × 10 −4 M. The selectivity obtained from peptides (ax) was examined by analysis of the precipitates using energy dispersive X-ray spectroscopyfield emission scanning electron microscopy (EDS-FESEM) after samples were collected by centrifugation to determine elemental contents and distributions. Based on the data from EDS-FESEM, we determined the most active miniaturized molecular structure derived from the original RU065 (a) and characterized its secondary structure using attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy. Inductively coupled plasmaoptical emission spectroscopy (ICP-OES) was also employed to compare the residual metal ion concentrations of supernatants after the removal of metallic precipitates by centrifugation, in trials using the original RU065 (a) and the miniaturized peptides. We also examined the mechanism by which this peptide-based gold recovery system functioned, employing matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS). The findings of this study are expected to assist in the future design of non-peptidyl anthracene derivatives for low-cost, environmentally friendly, and selective recovery of gold.

General
All solvents and reagents, unless otherwise noted, were purchased from Wako Pure Chemical Industries (Osaka, Japan) and used as received. The 9-fluorenylmethoxycarbony (Fmoc)-protected amino acid derivatives and reagents for peptide synthesis were purchased from Watanabe Chemical Industries (Hiroshima, Japan). Acetonitrile (high performance liquid chromatography (HPLC) grade) was purchased from Nacalai Tesque Inc. (Kyoto, Japan) and used for HPLC analysis and peptide purification. Ultrapure water (hereafter water) was purchased from Wako Pure Chemical Industries (Japan).
Fmoc-protected Rink amide resin was treated with 20% piperidine (PPD) in Nmethylpyrrolidone (NMP) (v/v) at room temperature for 15 min and washed with NMP three times to remove the Fmoc protecting groups. A solution of Fmoc-protected amino acid (10 eq), HBTU (10 eq), HOBt·H 2 O (10 eq), and N,N-diisopropylethylamine (DIEA, 20 eq) in NMP was added to a reaction vessel containing the N-terminal free Rink amide resin (1 eq), and the mixture was reacted at room temperature for 90 min for peptide chain elongation. The reaction mixture was filtrated and the resulting Fmoc-amino acid-bound resin was washed with NMP three times. The N-terminal Fmoc protecting groups of the resin were removed with 20% PPD /NMP (v/v) at room temperature for 15 min, and the resulting resin was washed with NMP three times. These coupling and deprotection cycles were repeated as required for peptide elongation. Subsequently, the Fmoc protecting groups at the N-termini of the peptide-bound resin were removed with 20% PPD/NMP (v/v) at room temperature for 15 min, and the resulting resin was washed with NMP three times. A solution of acetic anhydride (10 eq) and DIEA (10 eq) was added to the reaction vessel containing the N-terminal free peptide-bound resin (1 eq) and the mixture was reacted at room temperature for 30 min for N-terminal acetylation. The reaction mixture was filtrated and the resulting N-acetylated peptide-bound resin was washed with NMP three times, washed with chloroform three times, and then dried in vacuo to prepare a fully protected peptide-bound resin.
Final deprotection and cleavage of the resin-bound peptides were performed by treatment with trifluoroacetic acid (TFA)/thioanisole/m-cresol (90/7.5/2.5, v/v/v) at room temperature for 60 min. The reaction mixtures were concentrated and then precipitated with cold diethyl ether to provide the crude peptides. The crude peptides were washed with cold diethyl ether three times, dried in vacuo, and purified on a Hitachi LaChrom Elite HPLC system (Hitachi, Tokyo, Japan) using Cosmosil 5C 18 -AR-II packed columns (4.6 × 150 mm and 10 × 250 mm, Nacalai Tesque, Japan) with a linear gradient of acetonitrile/0.1% TFA, at a flow rate of 3.0 mL min −1 . The purified peptides were lyophilized and characterized by MALDI-TOFMS (AXIMA-CFR Plus or MALDI-8020, Shimadzu, Kyoto, Japan).
Peptide stock solutions (ca. 1.0 × 10 −3 M) were prepared by dissolving each purified peptide powder in 2,2,2-trifluoroethanol (TFE) to prevent self-assembly during storage. The concentrations of the stock solutions of peptides (a-x) were determined by absorption spectroscopy (Shimadzu UV-3100 spectrophotometer, Japan) using an extinction coefficient of 5300 M −1 cm −1 at 377 nm for an Ant residue in methanol solutions containing 1% TFE (v/v) [27]. The peptide stock solutions were stored at −20 • C until use.

Preparation of Sample Solutions
According to a procedure previously described in the literature [27], a portion of each peptide stock solution in TFE was transferred to a microtube, dried with a N 2 gas stream, Processes 2021, 9, 2010 5 of 16 then further dried in vacuo for 1 h. Water was added to the microtube to give a peptide concentration of 4.0 × 10 −4 M and the solution was sonicated at 40 • C for 2 min to break up any peptide aggregates. Mixtures of equal volumes of the peptide solution and of an aqueous solution containing HAuCl 4 and H 2 PtCl 6 (together with Ni(NO 3 ) 2 if required; 1.0 × 10 −4 M each) were incubated in a dry block heater (As One EB-303, As One, Osaka, Japan) at 40 • C for 24 h in the dark to prepare "sample solutions".

Absorption Spectra of Reaction Mixtures
A portion of each sample solution was transferred from the microtube to a 1.0 cm path length quartz cell and its UV-Vis absorption spectrum was recorded on a spectrophotometer (UV-3100, Shimadzu, Japan).

Energy Dispersive X-ray Spectroscopy-Field Emission Scanning Electron Microscopy (EDS-FESEM)
According to a procedure previously described in the literature [27], a sample solution was centrifuged at 15,000 rpm (himac CT15E, Hitachi, Japan) for 3 min and the supernatant was removed. The resulting solid was redispersed in a small volume of water, and droplets of the suspension were applied to a transmission electron microscopy (TEM) grid (Cu 200 mesh covered with a collodion membrane, Nisshin EM, Japan), allowed to stand for 1 min, and dried with filter paper. The resulting sample was further dried in vacuo before acquiring images and performing an EDS analysis (JIB-4601F, JEOL, Akishima, Japan). The Au/Pt atomic ratio in the precipitate was calculated by comparing the intensities of the characteristic X-ray signals for gold and platinum obtained from 12 to 27 randomly selected areas per sample.

Hydropathy Assessment of Peptides by Reversed-Phase High-Performance Liquid Chromatography
According to a procedure described in the literature [31,32], 10 µL aliquots of each peptide stock solution in TFE were analyzed by HPLC (LaChrom Elite HPLC system, Hitachi, Japan) using Cosmosil 5C 18 -AR-II packed columns (4.6 × 150 mm, Nacalai Tesque Inc., Japan) with a linear gradient from H 2 O/0.1% TFA to CH 3 CN/0.1% TFA over a span of 30 min, using a flow rate of 1.0 mL min −1 and UV detection at 220 nm. The retention times of the analytes were converted to relative hydrophobic index values (H) using the equation: where t R is the retention time for the peptide of interest and t R (RU065 4-7 , p) and t R (RU065 6 , x) are the retention times for the most hydrophilic and hydrophobic peptides, respectively.

Attenuated Total Reflection Fourier Transform Infrared (ATR-FTIR) Spectroscopy
According to a procedure previously described in the literature [25], an ATR-FTIR spectrum of peptide film prepared from a peptide solution obtained in the manner described above (peptide = 1.0 × 10 −3 M in water, incubated at 40 • C for 24 h) was acquired on a Shimadzu Fourier transform infrared spectrophotometer IRTracer-100 (Japan) equipped with an ATR-8200HA unit (Shimadzu, Japan).

Transmission Electron Microscopy
TEM samples were prepared using the procedure described above for the EDS-FESEM analyses and TEM images were acquired using a JEM-2100 instrument (JEOL, Japan) at an accelerating voltage of 200 kV.

X-ray Diffraction (XRD)
The precipitates obtained after centrifugation of each sample solution were dried and mounted on a quartz sample cell. The XRD pattern was recorded on a RINT2000 Processes 2021, 9, 2010 6 of 16 spectrometer (Rigaku, Tokyo, Japan) with Cu-Kα radiation with a scan area from 2.0 to 80.0 • , a scan step of 0.02 • , and a scan speed of 2.00 • min −1 .

Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES)
The supernatant (15 mL) obtained after centrifugation of each sample solution for 1 h at 15,000 rpm (himac CT15E, Hitachi, Japan) was analyzed by ICP-OES (Optima 5300DV, Perkin Elmer, Yokohama, Japan). The residual metal ion concentrations were calculated based on the working curves prepared by three-point-calibration using standard solutions of gold, platinum, and nickel.

Results and Discussion
Design, Synthesis, and Characterization of the Peptides. As described in the introduction, in prior work, our group synthesized [Ant 6 ]-RU006, a nonapeptide forming amphiphilic β-sheets (referred to herein as RU065 (a)) and having the structure Ac-Ala-Ile-Ala-Lys-Ala-Ant-Lys-Ile-Ala-NH 2 ( Figure 1B) [25,26]. The L-isoleucine (Ile) residues at the second and eighth positions in this peptide along with the Ant at the sixth position were all situated on the same side and so generated a hydrophobic face. A hydrophilic side was also produced based on the L-alanine (Ala) residues at the first, third, fifth, and ninth positions and the Lys at the seventh position, all of which were either less hydrophobic or, in the case of the Lys, were hydrophilic (see Figure 1B and Figure S1). A Lys residue was also present at the fourth position in proximity to the Ant on the hydrophobic side to allow interactions with AuCl 4 − and PtCl 6 2− ions, so as to both concentrate and reduce the metal ions in the self-assembled peptides to generate metallic precipitates. As noted, 23 different peptides having smaller structures than the original RU065 (that is, specimens b-x) were obtained by sequentially removing amino acid residues from both N-and C-termini to the Ant group at the sixth position of the original molecule. These fragment peptides are referred to herein as RU065 x-y , where x and y are the positions on the original RU065 (a) at which these fragments begin and end, respectively ( Figure 1C and Figure S1). As an example, RU065 2-9 (b) had the structure Ac-Ile-Ala-Lys-Ala-Ant-Lys-Ile-Ala-NH 2 , and so spanned the second to ninth positions of the original peptide, based on removing the Ala at the first position.
A typical solid-state peptide synthesis technique based on Fmoc chemistry was used to synthesize these peptides [30]. This synthesis was followed by purification using HPLC ( Figure S2) and characterization by MALDI-TOFMS (Table S1 and Figure S2).

Screening of Peptides by UV-Vis Absorption Spectroscopic Measurements.
We initially examined the abilities of peptides a-x to reduce the metal ions by preparing equivolume mixtures of aqueous peptide solutions (4.0 × 10 −4 M) and aqueous solutions containing HAuCl 4 and H 2 PtCl 6 (1.0 × 10 −4 M each) that were assessed using UV-Vis spectroscopy. Figure S3 shows the spectra of these sample solutions. It is evident that the majority of the peptide solutions (a-u) generated a surface plasmonic absorption band at approximately 540 nm, while the short fragments (v-x) produced a band at 580 nm. These results suggest the formation of spherical gold nanocrystals via the reduction of Au ions by the anthracene moieties in the former peptides (a-u), and also indicate that peptides (v-x) caused aggregation of the nanocrystals and/or provided anisotropic gold nanocrystals [25,33,34]. It should be noted that the formation of platinum nanocrystals could not be ascertained from the UV-Vis absorption spectra because this material did not generate characteristic peaks in the visible region of the spectrum [35].
Screening of Peptides by EDS-FESEM Measurements. We next separated the precipitates recovered from the sample solutions containing peptides (a-x), HAuCl 4 , and H 2 PtCl 6 following centrifugation and analyzed their morphologies and elemental distributions by EDS-FESEM. Figure S4 presents representative elemental mapping images of the precipitates for all peptides. In these images, the Au and Pt distributions are shown in yellow and purple, respectively, along with the corresponding EDS spectra. The results of replicate trials are summarized in Figure 2A. The Au/Pt atomic ratio for the precipitates generated using the original RU065 (a) was 8.4 ± 0.4, while the fragment peptides lacking one, two, and three amino acids from the N-terminus of RU065 (a) (RU065 x-9 (x = 2-4, b-d)) all provided similar ratios of 8.6 ± 0.2, 8.6 ± 0.6, and 8.8 ± 0.5, respectively. Surprisingly, the deletion of four and five amino acids from the N-terminus of the RU065 (peptides e and f ) gave much lower ratios of 4.9 ± 0.3 and 5.1 ± 0.6, respectively. characteristic peaks in the visible region of the spectrum [35].
Screening of Peptides by EDS-FESEM Measurements. We next separated the precipitates recovered from the sample solutions containing peptides (a-x), HAuCl4, and H2PtCl6 following centrifugation and analyzed their morphologies and elemental distributions by EDS-FESEM. Figure S4 presents representative elemental mapping images of the precipitates for all peptides. In these images, the Au and Pt distributions are shown in yellow and purple, respectively, along with the corresponding EDS spectra. The results of replicate trials are summarized in Figure 2A. The Au/Pt atomic ratio for the precipitates generated using the original RU065 (a) was 8.4 ± 0.4, while the fragment peptides lacking one, two, and three amino acids from the N-terminus of RU065 (a) (RU065x-9 (x = 2-4, bd)) all provided similar ratios of 8.6 ± 0.2, 8.6 ± 0.6, and 8.8 ± 0.5, respectively. Surprisingly, the deletion of four and five amino acids from the N-terminus of the RU065 (peptides e and f) gave much lower ratios of 4.9 ± 0.3 and 5.1 ± 0.6, respectively. We also examined the series of fragment peptides lacking an Ala residue at the ninth position of the RU065 (a) (RU065x-8 (x = 1-6, g-l)) and found that the deletion of one, two, and three amino acids from the N-terminus of RU0651-8 (g) provided excellent gold selectivity, with ratios of 7.8 ± 0.2 (RU0651-8 (g)), 7.9 ± 0.2 (RU0652-8 (h)), 7.6 ± 0.5 (RU0653-8 (i)) We also examined the series of fragment peptides lacking an Ala residue at the ninth position of the RU065 (a) (RU065 x-8 (x = 1-6, g-l)) and found that the deletion of one, two, and three amino acids from the N-terminus of RU065 1-8 (g) provided excellent gold selectivity, with ratios of 7.8 ± 0.2 (RU065 1-8 (g)), 7.9 ± 0.2 (RU065 2-8 (h)), 7.6 ± 0.5 (RU065 3-8 (i)) and 7.8 ± 0.4 (RU065 4-8 (j)), respectively. These values were comparable to those for peptides a-d. In contrast, greatly decreased Au/Pt atomic ratios of 5.0 ± 0.7 and 3.0 ± 0.8 were determined for RU065 5-8 (k) and RU065 6-8 (l), respectively. These observations strongly indicate that the Lys residue at the fourth position of the original RU065 (a) was important with regard to obtaining significant gold selectivity.
The fragment series RU065 x-7 (x = 1-6, m-r), lacking two amino acid residues at the eighth and ninth position of the original RU065 (a), were also assessed by EDS-FESEM. RU065 1-7 (m) and RU065 2-7 (n) provided suitable levels of selectivity, with ratios of 7.9 ± 0.3 and 7.8 ± 0.4, respectively. However, the deletion of the Ile residue at the second position of RU065 1-7 (m) afforded significantly decreased Au/Pt ratios of 3.8 ± 0.3 and 4.3 ± 0.5 in the case of the RU065 3-7 (o) and RU065 4-7 (p), respectively. The RU065 x-7 (x = 5-6, q-r) specimens did not provide large variations in the ratio even though the Lys residue at the fourth position was deleted in both, with RU065 5-7 (q) and RU065 6-7 (r) giving ratios of 3.2 ± 0.4 and 2.8 ± 0.7, respectively. These data suggest that the presence of a hydrophobic Ile residue at the second position of RU065 1-7 (m) was vital to realizing high selectivity for gold. This effect is attributed to the lack of hydrophobic Ile and Ala residues at the eighth and ninth positions of the original RU065 (a) in the RU065 x-7 (x = 1-6, m-r) fragments.
Lastly, we examined the selectivity of the RU065 x-6 (x = 1-5, s-w) specimens, which lacked three amino acid residues at the seventh, eighth, and ninth positions of the original RU065 (a), and of Ac-Ant-NH 2 (RU065 6 (x)), which comprised only an Ant residue. The selectivity data trend for these peptides was similar to that for the RU065 x-7 (x = 1-6, m-r) series. That is, the ratios for RU065 1-6 (s) and RU065 2-6 (t) showed slightly reduced values of 6.2 ± 1.2 and 6.4 ± 1.0, respectively, compared with those for peptides a-d, g-j, m, and n. The removal of the Ile residue at the second position of RU065 1-6 (s) evidently lowered the ratio, giving values of 2.8 ± 0.3, 4.0 ± 1.2, 2.5 ± 0.4, and 2.6 ± 0.3 for the RU065 x-6 (x = 3-5, u-w) and RU065 6 (x) fragments, respectively. From these data, it appears that the incorporation of a hydrophobic Ile residue and a positively charged Lys residue is necessary for good selectivity. This is likely because the Ile and Lys residues facilitate the initial self-assembly of the peptide through hydrophobic interactions and also promote the accommodation/enrichment of AuCl 4 − ions within the self-assembled peptide nanostructure, based on the proposed mechanism [25,26].
Hydropathy Assessment. The EDS-FESEM experiments described above highlighted the importance of characterizing the hydropathy of the original RU065 (a) and its fragment peptides (b-x) by HPLC analysis [31,32]. Figure 2B summarizes the H values for these specimens. Note that more positive values are associated with more hydrophobic compounds while negative values indicate more hydrophilic species. These data demonstrate that the original RU065 (a) was relatively hydrophobic but a significant transition from hydrophobic to hydrophilic was induced by removing the hydrophobic Ile residue at the second position between RU065 2-9 (b) and RU065 3-9 (c). The deletion of the hydrophilic Lys residue at the fourth position led to another significant transition from hydrophilic back to hydrophobic between RU065 4-9 (d) and RU065 [5][6][7][8][9] . Similar transitions also occurred between RU065 2-8 (h) and RU065 3-8 (i) and between RU065 4-8 (j) and RU065 5-8 (k). In the case of the RU065 x-7 (x = 1-6, m-r) series, there was a change from hydrophobic to hydrophilic following the removal of the hydrophobic Ile residue at the second position on going from RU065 2-7 (n) to RU065 3-7 (o), resulting in the most hydrophilic peptide lacking two hydrophobic Ile residues at the second and eighth positions. Although the deletion of the Lys at the fourth position to generate RU065 5-7 (q) decreased the relative hydrophilicity of this peptide (or more accurately increased the relative hydrophobicity), its H value was still within the hydrophilic range because it did not have two of the Ile resides.
The H values for the RU065 x-6 (x = 1-5, s-w) series and RU065 6 (x) also demonstrated a significant transition from hydrophobic to hydrophilic between RU065 2-6 (t) and RU065 3-6 (u) and from hydrophilic to hydrophobic between RU065 4-6 (v) and RU065 5-6 (w). These transitions were similar to those that occurred in the other series. The most hydrophobic fragment was RU065 6 (x), which contained only one Ant residue. The smaller peptide structures afforded higher H values, likely due to the greater contributions of individual amino acid residues.
Determination of Miniaturized Active Structures. The relationship between the Au/Pt atomic ratios determined from Figure 2A and the H values from Figure 2B for all peptides, along with the amino acid residues comprising each specimen, are presented in Figure 3. Those fragments having H values between −2 and +2 provided the best ratios (in the vicinity of 8). Specifically, these specimens were a, b, c, d, g, h, i, j, m, and n, among which the smallest peptide was RU065 4-8 (j), having five amino acid residues including two Lys and one Ile ( Figure 4A). On this basis, we determined that RU065 4-8 (j) Processes 2021, 9, 2010 9 of 16 had the minimum structure required to precipitate/recover gold from a mixture of gold and platinum ions. amino acid residues.
Determination of Miniaturized Active Structures. The relationship between the Au/Pt atomic ratios determined from Figure 2A and the H values from Figure 2B for all peptides, along with the amino acid residues comprising each specimen, are presented in Figure 3. Those fragments having H values between −2 and +2 provided the best ratios (in the vicinity of 8). Specifically, these specimens were a, b, c, d, g, h, i, j, m, and n, among which the smallest peptide was RU0654-8 (j), having five amino acid residues including two Lys and one Ile ( Figure 4A). On this basis, we determined that RU0654-8 (j) had the minimum structure required to precipitate/recover gold from a mixture of gold and platinum ions. Next, we characterized the secondary structure of the miniaturized active peptide RU0654-8 (j) by measuring the ATR-FTIR spectrum of a peptide film prepared from aqueous solution ( Figure 4B). The amide I region, originating from the amide carbonyl stretching frequencies between 1600 and 1700 cm −1 , is commonly used to determine the amide mode. The ATR-FTIR spectrum showed a strong amide I band around 1630 cm −1 and a weak band around 1675 cm −1 , indicating that an antiparallel β-sheet conformation was predominant [25], suggesting that even such a small fragment peptide (j) preferably formed β-sheet conformation as seen for RU006 and RU065 (a, [Ant 6 ]-RU006) in the literature [25]. This presumably contributed to the enrichment of gold ions during self-assembly and the densification of the obtained metallic gold-peptide composites resulting in easy separation [27].  Interestingly, peptides RU0654-8 (j) and RU0653-7 (o), each of which comprised five amino acids, showed very different Au/Pt atomic ratios even though the only difference was that the Ile residue in RU0654-8 (j) was replaced with an Ala residue in RU0653-7 (o). Further, peptides RU0654-8 (j) and RU0655-7 (q) had similar hydropathy characteristics but different numbers of amino acid residues, and also showed a significant difference in ratios. These variations can presumably be ascribed to the important role of amide moieties in forming the hydrogen bond networks that facilitated peptide self-assembly and metal precipitation. These results strongly suggest that adequate hydrophobicity and the presence of several amide moieties are important to the self-assembly process that, in turn, accommodates and concentrates gold ions with the subsequent reduction of these ions within the assemblies to provide high Au/Pt atomic ratios.
Transmission Electron Microscopy and X-ray Diffraction. The precipitates generated from mixtures of RU0654-8 (j) with HAuCl4 and H2PtCl6 were characterized using TEM ( Figure 5A-C). The image in Figure 5A demonstrates a spherical particle with a diameter of approximately 30 nm along with numerous other spherical particles with diameters on the order of 2 nm. The image in Figure 5B presents a magnified view of the nan- Next, we characterized the secondary structure of the miniaturized active peptide RU065 4-8 (j) by measuring the ATR-FTIR spectrum of a peptide film prepared from aqueous solution ( Figure 4B). The amide I region, originating from the amide carbonyl stretching frequencies between 1600 and 1700 cm −1 , is commonly used to determine the amide mode. The ATR-FTIR spectrum showed a strong amide I band around 1630 cm −1 and a weak band around 1675 cm −1 , indicating that an antiparallel β-sheet conformation was predominant [25], suggesting that even such a small fragment peptide (j) preferably formed β-sheet conformation as seen for RU006 and RU065 (a, [Ant 6 ]-RU006) in the literature [25]. This presumably contributed to the enrichment of gold ions during selfassembly and the densification of the obtained metallic gold-peptide composites resulting in easy separation [27].
Interestingly, peptides RU065 4-8 (j) and RU065 [3][4][5][6][7] (o), each of which comprised five amino acids, showed very different Au/Pt atomic ratios even though the only difference was that the Ile residue in RU065 4-8 (j) was replaced with an Ala residue in RU065 3-7 (o). Further, peptides RU065 4-8 (j) and RU065 5-7 (q) had similar hydropathy characteristics but different numbers of amino acid residues, and also showed a significant difference in ratios. These variations can presumably be ascribed to the important role of amide moieties in forming the hydrogen bond networks that facilitated peptide self-assembly and metal precipitation. These results strongly suggest that adequate hydrophobicity and the presence of several amide moieties are important to the self-assembly process that, in turn, accommodates and concentrates gold ions with the subsequent reduction of these ions within the assemblies to provide high Au/Pt atomic ratios.
Transmission Electron Microscopy and X-ray Diffraction. The precipitates generated from mixtures of RU065 4-8 (j) with HAuCl 4 and H 2 PtCl 6 were characterized using TEM ( Figure 5A-C). The image in Figure 5A demonstrates a spherical particle with a diameter of approximately 30 nm along with numerous other spherical particles with diameters on the order of 2 nm. The image in Figure 5B presents a magnified view of the nanosphere surrounded by the yellow square in Figure 5A. This image indicates a d-spacing of 2.35 Å that corresponds to the Au<111> direction in crystals having a face-centered cubic (fcc) structure [27,36]. A magnified view of the region within the blue square is provided in Figure 5C, with a d-spacing of 2.21 Å that equates to the Pt<111> direction in an fcc crystal [27,36]. These results were verified by acquiring X-ray diffraction (XRD) data. The powder XRD pattern in Figure 5D contains four broad peaks at 2θ values of 38.2 • , 44.4 • , 64.6 • , and 77.6 • that are attributed to (111), (200), (220), and (311) crystal planes, respectively, in metallic gold having an fcc crystal structure. XRD peaks corresponding to elemental platinum (2θ = 39.8 • , 46.3 • , and 67.5 • , assigned to (111), (200) and (220) crystal planes, respectively) were not observed, likely because the platinum nanospheres had crystallite sizes smaller than those detectable by powder XRD. These results suggest that the precipitates from the reaction of RU065 4-8 (j), HAuCl 4 , and H 2 PtCl 6 were in a metallic state and that metallic platinum formed much smaller particles than metallic gold. Gold Selectivity Assessment by ICP-OES. ICP-OES was used to better assess the precise selectivity of the peptide-based recovery system for gold by quantifying residual ion concentrations in the supernatants obtained after centrifugation of mixtures comprising the peptides, HAuCl4 and H2PtCl6 ( Figure 6) [27,37]. As shown in Figure 6A, ICP-OES Gold Selectivity Assessment by ICP-OES. ICP-OES was used to better assess the precise selectivity of the peptide-based recovery system for gold by quantifying residual ion concentrations in the supernatants obtained after centrifugation of mixtures comprising the peptides, HAuCl 4 and H 2 PtCl 6 ( Figure 6) [27,37]. As shown in Figure 6A, ICP-OES analyses of the supernatants of the reaction mixtures containing the original RU065 (a) revealed a residual Au(III)/Au(I) concentration of 2.7 × 10 −6 M (equivalent to 95% recovery) and a residual Pt(IV)/Pt(II) concentration of 3.7 × 10 −5 M (26% recovery). These results are comparable to those in our previous work [27].
imately 1% recovery) for Ni(II). These results suggest that positively charged transition metal ions such as nickel were more difficult to reduce than noble metals and so primarily remained in solution as a consequence of charge repulsion between the Ni(II) ions and RU0654-8 (j). Therefore, the present peptide-based system could be robust for selective gold recovery from a mixture of base metal ions with redox potentials lower than those for noble metal ions.
As reported by Kim et al., a tannin-TiO2 heterostructure selectively adsorbed gold ions from a solution containing nine different metal ions ([metal ion] = 1.0 × 10 −3 M each) with an Au/Pt ratio of ca. 50 under light irradiation [38]. Dogan et al. showed that nitrile and amidoxime-containing nanoporous polymer microspheres adsorbed gold ions from a solution containing eleven different metal ions ([metal ion] = 100 ppb each) with an Au/Pt ratio of ca. 3.8 [39]. These results suggest that for high metal ion concentrations, an excellent Au/Pt ratio can be achieved, whereas for low metal ion concentrations, it is difficult to improve the gold selectivity against platinum. Thus, our results (metal ion = 5.0 × 10 −5 M each) are competitive and it is possible to design low-cost gold-selective recovery agents based on the active short peptide RU0654-8 (j).

Mechanistic Study.
Our prior work demonstrated that the original RU065 (a) was able to generate metal nanocrystals after a span of 4 h. This was a more rapid process than was evident using another peptide incorporating a naphthalene moiety in place of an anthracene ring. Specifically, EDS-FESEM indicated that a combination of RU065 (a), HAuCl4, and H2PtCl6 decreased the Au/Pt ratio from 12.3 to 7.3 after 4 h. On this basis, we suggest that gold nanocrystals were initially generated as the HAuCl4 was reduced by the anthracene rings, while nanocrystals of platinum were formed more gradually. Consequently, Au(III) was rapidly changed to metallic gold with limited formation of Au(I) as an intermediate, while Pt(IV)/Pt(II) ions received electrons less slowly to produce a platinum precipitate [27][28][29].
The present work also attempted to elucidate the mechanism of the selective reduction/recovery of metallic gold from the mixture of HAuCl4 and H2PtCl6 by RU0654-8 (j). This was done by analyzing sample solutions using MALDI-TOFMS. Figure 7A shows the MALDI-TOFMS profile for the solution containing RU0654-8 (j) prepared using the standard conditions described in the Experimental Section but without HAuCl4 and H2PtCl6. A single peak was obtained at m/z = 748.5, corresponding to the molecular ion peak (that is, [M + H] + ) along with a less intense peak at m/z = 770 likely corresponding to [M + Na] + . These data indicate that RU0654-8 (j) remained intact after a 24 h incubation in water at 40  Figure 6B. The results for RU065 (a) and RU065 4-8 (j) were similar, suggesting that the latter has the same potential to selectively reduce/recover gold ions as RU065 (a), even though it has four fewer amino acids. Consequently, the minimum requirement for the structure would be Ac-Lys-Ala-Ant-Lys-Ile-NH 2 . The high recovery yields for gold with RU065 (a) and RU065 4-8 (j) suggest that RU065 4-8 (j) also concentrated gold ions on the basis of a mechanism involving self-assembly, reduction of gold ions to the metallic state and concentration, enabling easy separation of the metallic solids by simple centrifugation, as RU065 (a) did in our previous work [27]. However, the selectivity (that is, the Au recovery %/Pt recovery %) determined by ICP-OES was somewhat lower than the Au/Pt atomic ratio obtained from the EDS-FESEM analyses, as shown in Figure 2A. These small differences between the selectivity determined by ICP-OES and EDS-FESEM could be attributed to the formation of platinum nanoclusters that were difficult to separate by centrifugation (as shown in Figure 5A,C), such that different amounts of gold were determined using the two methods.
We next demonstrated selective gold reduction/recovery from a mixture of gold, platinum, and nickel ions using RU065 4-8 (j) to examine the effects of a transition metal. Figure 6C shows the residual ion concentrations in the supernatant obtained from a mixture of RU065 4-8 (j, 2.0 × 10 −4 M), HAuCl 4 , H 2 PtCl 6 , and Ni(NO 3 ) 2 (metal ion = 5.0 × 10 −5 M each) in water. The data show residual concentrations of 1.8 × 10 −6 M (96% recovery) for Au(III)/Au(I), 4.1 × 10 −5 M (18% recovery) for Pt(IV)/Pt(II) and 5.0 (4.96) × 10 −5 M (approximately 1% recovery) for Ni(II). These results suggest that positively charged transition metal ions such as nickel were more difficult to reduce than noble metals and so primarily remained in solution as a consequence of charge repulsion between the Ni(II) ions and RU065 4-8 (j). Therefore, the present peptide-based system could be robust for selective gold recovery from a mixture of base metal ions with redox potentials lower than those for noble metal ions.
As reported by Kim et al., a tannin-TiO 2 heterostructure selectively adsorbed gold ions from a solution containing nine different metal ions ([metal ion] = 1.0 × 10 −3 M each) with an Au/Pt ratio of ca. 50 under light irradiation [38]. Dogan et al. showed that nitrile and amidoxime-containing nanoporous polymer microspheres adsorbed gold ions from a solution containing eleven different metal ions ([metal ion] = 100 ppb each) with an Au/Pt ratio of ca. 3.8 [39]. These results suggest that for high metal ion concentrations, an excellent Au/Pt ratio can be achieved, whereas for low metal ion concentrations, it is difficult to improve the gold selectivity against platinum. Thus, our results (metal ion = 5.0 × 10 −5 M each) are competitive and it is possible to design low-cost gold-selective recovery agents based on the active short peptide RU065 4-8 (j).
Mechanistic Study. Our prior work demonstrated that the original RU065 (a) was able to generate metal nanocrystals after a span of 4 h. This was a more rapid process than was evident using another peptide incorporating a naphthalene moiety in place of an anthracene ring. Specifically, EDS-FESEM indicated that a combination of RU065 (a), HAuCl 4 , and H 2 PtCl 6 decreased the Au/Pt ratio from 12.3 to 7.3 after 4 h. On this basis, we suggest that gold nanocrystals were initially generated as the HAuCl 4 was reduced by the anthracene rings, while nanocrystals of platinum were formed more gradually. Consequently, Au(III) was rapidly changed to metallic gold with limited formation of Au(I) as an intermediate, while Pt(IV)/Pt(II) ions received electrons less slowly to produce a platinum precipitate [27][28][29].
The present work also attempted to elucidate the mechanism of the selective reduction/recovery of metallic gold from the mixture of HAuCl 4 and H 2 PtCl 6 by RU065 4-8 (j). This was done by analyzing sample solutions using MALDI-TOFMS. Figure 7A shows the MALDI-TOFMS profile for the solution containing RU065 4-8 (j) prepared using the standard conditions described in the Experimental Section but without HAuCl 4 and H 2 PtCl 6 . A single peak was obtained at m/z = 748.5, corresponding to the molecular ion peak (that is, [M + H] + ) along with a less intense peak at m/z = 770 likely corresponding to [M + Na] + . These data indicate that  Figure  7B,C). These results strongly suggest that the anthracene rings were oxidized by gold and/or platinum ions to generate ketone, hydroxide, quinone, and/or hydroquinone-like functional groups during the self-assembly of the peptides and the subsequent reduction of the metal ions to form gold-rich metallic precipitates [40]. However, since only limited reaction kinetics data were obtained in our previous study [27], further experiments are required to clarify the reaction mechanisms that lead to good gold selectivity.
around m/z = 767 likely attributable to [M + K] + and/or [M + O + Na] + ( Figure 7B,C). These results strongly suggest that the anthracene rings were oxidized by gold and/or platinum ions to generate ketone, hydroxide, quinone, and/or hydroquinone-like functional groups during the self-assembly of the peptides and the subsequent reduction of the metal ions to form gold-rich metallic precipitates [40]. However, since only limited reaction kinetics data were obtained in our previous study [27], further experiments are required to clarify the reaction mechanisms that lead to good gold selectivity.

Conclusions
The anthracene-containing peptide RU065 (a) and its fragment peptides (b-x) were synthesized and investigated with regard to the selective metallic recovery of gold. These trials used peptides (2.0 × 10 −4 M each) in homogenous aqueous solutions containing dilute HAuCl4 and H2PtCl6 (5.0 × 10 −5 M each) to determine the minimum required structural composition. Ten peptides provided suitable Au/Pt atomic ratios on the order of 8 based on the data obtained from EDS-FESEM and HPLC analyses. The smallest useable peptide structure was determined to be RU0654-8 (j), a fragment peptide made of five amino acid residues (an Ant, two Lys, one Ile, and one Ala). Therefore, an anthracene ring accompanied by residues providing adequate hydrophobicity, two positive charges, and several amide groups are required to facilitate selective metallic gold precipitation/recovery. This process results in the subsequent oxygenation of the anthracene rings as determined by MALDI-TOFMS. We also demonstrated peptide-based selective gold recovery from a

Conclusions
The anthracene-containing peptide RU065 (a) and its fragment peptides (b-x) were synthesized and investigated with regard to the selective metallic recovery of gold. These trials used peptides (2.0 × 10 −4 M each) in homogenous aqueous solutions containing dilute HAuCl 4 and H 2 PtCl 6 (5.0 × 10 −5 M each) to determine the minimum required structural composition. Ten peptides provided suitable Au/Pt atomic ratios on the order of 8 based on the data obtained from EDS-FESEM and HPLC analyses. The smallest useable peptide structure was determined to be RU065 4-8 (j), a fragment peptide made of five amino acid residues (an Ant, two Lys, one Ile, and one Ala). Therefore, an anthracene ring accompanied by residues providing adequate hydrophobicity, two positive charges, and several amide groups are required to facilitate selective metallic gold precipitation/recovery. This process results in the subsequent oxygenation of the anthracene rings as determined by MALDI-TOFMS. We also demonstrated peptide-based selective gold recovery from a mixture containing gold, platinum, and nickel ions using the RU065 4-8 (j) fragment. This system successfully precipitated and recovered metallic gold by simple centrifugation with high selectivity and good recovery, as indicated by ICP-OES data. We believe that the minimum structural features for miniaturized active molecules as demonstrated herein will assist in the future design of small, non-peptidyl anthracene derivatives that selectively precipitate/recover metallic gold with a high gold recovery rate per unit mass of recovery agents. This could allow the development of a low-cost, one-pot gold recovery system applicable on a large scale.