Metalloid Reductase Activity Modified by a Fused Se⁰ Binding Peptide.

Abstract

A selenium nanoparticle binding peptide was isolated from a phage display library and genetically fused to a metalloid reductase that reduces selenite (SeO₃²⁻) to Se⁰ nanoparticle (SeNP) form. The fusion of the Se binding peptide to the metalloid reductase regulates size of the resulting SeNP to ~35 nm average diameter, where without the peptide, SeNPs grow to micron sized polydisperse precipitates. The SeNP product remains associated with the enzyme/peptide fusion. The Se binding peptide fusion to the enzyme increases the enzyme’s SeO₃²⁻ reductase activity. Size control of particles was diminished if the Se binding peptide was only added exogenously to the reaction mixture. The enzyme-peptide construct shows preference for binding smaller SeNPs. The peptide-SeNP interaction is attributed to His based ligation that results in a peptide conformational change on the basis of Raman spectroscopy.

Graphical Abstract

Introduction

Metal and metalloid reducing and/or oxidizing enzymes are a key component in cellular metal homeostasis. They can be involved in metal storage, for instance the ferroxidase centers in ferritin.¹ They can also be involved in diminishing metal/metalloid toxicity, for instance by reducing species from soluble to an insoluble form, as exemplified by mercuric reductase.^2–4 Nanoparticles comprised of cadmium sulfide, palladium, selenium, chromium, cobalt, tellurium and others have been formed by environmental bacterial isolates, suggesting enzymatic redox processes.^5–12 Various fungal species also reduce metal ions through quinones and/or intra- and extracellular enzymes.^13,14 From many of the species that form inorganic nanoparticles, enzymes are identified that reduce inorganic ions including as silica, chromate, cerium, selenite, selenate, tellurite, and tellurate.^15–20

For most of the enzymes identified, reduction occurs through an electron transfer utilizing the cofactors NADH or NADPH. In the case of bioremediase identified by Chowdury et al. the formation of SiO₂ from tetraethyl orthosilicate was catalyzed through a Zn²⁺ ion bound by the enzyme resulting in 25 ± 5 nm diameter nanoparticles.²¹ An extracellular cystathione γ-lyase from S. maltophilia, capable of removing a sulfide from cysteine was isolated and used to make CdS quantum dots. This same enzyme was then used to produce numerous core-shell quantum dots by utilizing cysteine and/or selenocysteine.^22–25 A glutathione reductase like-metalloid reductase (GRLMR) was identified capable of reducing seleno-diglutathione, selenite, and selenate to amorphous Se nanoparticles (SeNP).^26,27 GRLMR, which will be used in this present work is also capable of forming CdSe quantum dots of varying size when the enzyme is incubated with seleno-diglutathione, Cd⁺², and NADPH.

Whereas natural evolutionary processes can result in inorganic ion redox enzymes, in vitro selections have identified peptides that can bind to and sometimes exert control over the synthesis of inorganic materials such as CdS, FePt, ZnS, GeO₂, TiO₂, CoPt, and CaCo₃.^28–33 The pioneering works of Belcher and colleagues demonstrated the use of M13 phage for the ordering of quantum dots through the identification and fusion of inorganic interacting peptides to M13’s major coat protein, p8.³⁴ Subsequently, this approach produced highly structured hybrid materials.^35–37 Ahmad et al. isolated 2 peptides exhibiting a catalytic ability to form crystalline BaTiO₃ precipitates from barium acetate and potassium bis(oxalate) oxotitanate at near neutral pH, although without any major size controlling effects.³⁸ Two other sequence-unique peptides were identified by Bassindale et al. capable of interacting with AgNPs. Ag-22 formed triangular, quadrangular, and spherical shapes while Ag-28 formed uniformly spherical NPs³⁹. Feldheim et al. showed that a peptide originally isolated for binding to germanium could exert profound effects on the formation of Ag nanostructures.⁴⁰

The current work represents progress toward genetically encoded nanoparticle tags, analogous to green fluorescent protein (GFP), but visible in the electron microscope. This addresses a ‘contrast problem’ in biological electron microscopy.²⁷ Presently, imparting contrast to specific proteins in electron microscope imaging relies on immuno-labeling or two step deposition of diaminobenzidine followed by osmium tetroxide.^41–43 The use of a metal reducing enzyme to produce insoluble particulate from soluble metal salt precursors in concurrence with a peptide opens the door to currently difficult molecular identification in electron microscope images.

So-far, the investigations of metal binding peptides and metal reducing enzymes represent entirely separate fields of inquiry. For the first time, we investigate the effect of a selected peptide on enzymatic nanomaterial production for the eventual application as a GFP analogue for biological electron microscopy. In this case, we study a SeNP binding peptide effecting enzymatic production of SeNPs. We find that the peptide changes the SeNP sizes and distributions – an expected result. However, this is only observed robustly when the peptide is genetically concatenated to the enzyme, not when the peptide is freely available in solution. More unexpectedly, we observe a change in the enzyme’s kinetic properties. Based on Raman spectroscopy, we characterize the peptide binding to the SeNP through histidine ligation.

Results and Discussion

A SeNP binding peptide (SeBP) was isolated from a New England Biolabs PhD phage display library following the manufacturer’s protocol (supporting information). DNA encoding SeBP was concatenated to the C-terminus of GRLMR by outward PCR (GRLMR-SeBP). Presence of the SeBP was confirmed through sequencing and native-PAGE (Figure S2.). GRLMR-SeBP was expressed as previously described using T7 Express BL21(DE3) E. coli (NEB Catalog# C3013I).^26,27 After induction with 1 mM IPTG, the 1 L growth culture was supplemented with 1 μM of HNaSeO₃ and grown overnight at 37°C and 225 RPM. Supplementing with selenite increased the yield of the enzyme considerably by mitigating selenium deficiency in the culture due to scavenging of selenium by the expressed GRLMR-SeBP. GRLMR-SeBP was then purified from lysed cells with an Ni-NTA column.

Prior work revealed that in excess SeO₃²⁻ and NADPH, GRLMR activity produces precipitates of red-selenium.²⁷ Figure 1 shows this result (black trace, panel A; black cartoon, panel B). Size characterization of this precipitate shows polydisperse micron sized SeNPs.

Figure 1.

Panel A shows DLS traces (normalized to scattering intensity) for SeNPs produced by GRLMR (black), GRLMR in the presence of exongenously added SeBP (red), and the GRLMR-SeBP fusion (blue). Panel B. shows cartoon-based interpretations and photographs of the DLS cuvettes for GRLMR (top), GRLMR with exogenous SeBP (middle) and GRLMR-SeBP concatamers (bottom).

When SeBP is concatenated to GRLMR, the size and distribution of enzymatically produced SeNPs changes dramatically. The blue trace in Figure 1, panel A shows that the Se⁰ product size is restricted to a size of ~85 nm. The DLS data shown in Figure 1, panel A is intensity based, emphasizing larger particles (larger particles scatter light more intensely compared to smaller particles). When normalizing this data to account for net particle numbers, a ~35 nm average particle size is observed (Figure S4). Consistent with SeNP growth-arrest at a sub-100 nm diameter, we observe a red colloidal suspension, without observable preciptates.

Interestingly, for this size-control effect, it is compulsory that the peptide is concatenated to the enzyme. If the peptide is simply added exogenously to the solution, we observe no obvious limitation to GRLMR generaged Se⁰ product size, as shown in the red trace of Figure 1, panel A and the middle cartoon of panel B. The dramatically smaller hydrodynamic radius of SeNPs produced by GRLMR-SeBP implies a stable non-covalent binding interaction between the SeNP reaction product and the peptide-modified enzyme. To determine if GRLMR-SeBP had increased affinity to the SeNP products, we executed the pull-down assay shown in Figure 2, in which SeNPs are centrifugally removed from suspension, accompanied by any associated enzyme. This assay shows 83.1±1.0% and 14.4±2.6% of GRLMR-SeBP and GRLMR, respectively, as associated with the SeNP fraction. Data is shown in Figure 2, panel A. It is possible that this assay under-estimates the amount of enzyme associated with particles, as sufficiently small SeNPs may not be centrifugally removed at the speeds used here, yet remain enzyme-associated, as depicted in Figure 2, panel B (vide infra).

Figure 2.

Panel A shows SeNP association of GRLMR (diagonal stripe) and GRLMR-SeBP1 (blue) as determined by Bradford assay. Panel B shows a schematic of the assay and interpreted results for GRLMR-SeBP (top) and GRLMR (bottom). Panel C shows a nondenaturing polyacrylamide gel of GRLMR-SeBP (lane 1), GRLMR (lane 3), the soluble fraction of GRLMR-SeBP after SeNP synthesis (lane 2) and the soluble fraction of GRLMR after SeNP synthesis (lane 4).

A native/nondenaturing polyacrylamide gel electrophoresis (native-PAGE) experiment (Figure 2, C) suggests qualitatively that indeed a larger fraction than 83.1±1.0% of GRLMR-SeBP is bound to SeNPs. The gel mobility in this pH 8.3 matrix of GRLMR-SeBP and GRLMR are shown in lanes 1 and 3, respectively. The addition of 5 His and 2 Lys residues in the SeBP concatenation that produces GRLMR-SeBP provokes a notable gel-shift relative to native GRLMR, which we attribute to a substantial change in net-charge to a more positive value. The presence of the SeBP on GRLMR increases the calculated pI from 6.06 to 6.37. To further accentuate the charge difference between GRLMR and GRLMR-SeBP, a pH 6.6 native-PAGE gel was run following the buffer solution described by McLellan.⁴⁴ GRLMR-SeBP only slightly migrates into the gel indicating a pI closer to 6.6 (Figure S5).

Lanes 2 and 4 show GRLMR-SeBP and GRLMR, respectively, present in the supernatant post-SeNP synthesis and centrifugation and dialyzed into milliQ water to remove excess salts from the reaction. For each reaction, we observe a gel-shift relative to unreacted enzyme, which we attribute to enzyme-SeNP complexation. In other words, SeNPs are bound to the enzyme, which alters its electrophoretic mobility. In the case of reacted GRLMR-SeBP/SeNP complexes, we observe an electrophoretic mobility shift toward a lower mass-to-charge product. Here we attribute the shift to the SeNP binding specifically to the SeBP component, thereby neutralizing the charge added to the GRLMR-SeBP by the SeBP concatenation. In the case of reacted GRLMR, the electrophoretic mobility shift is toward a higher mass-to-charge product, which we attribute to the increased mass of a GRLMR/SeNP complex relative to GRLMR. This provides strong evidence that the SeNPs bind to the SeBP. Charge neutralization of the SeBP fragment coupled with the mass change of SeNP binding to GRLMR gives both enzyme-SeNP products a similar gel-mobility.

The DLS data (vide supra) implies a substantial difference in SeNP size that depends upon the presence or absence of concatenated SeBP. We examined the enzymatic Se⁰ products by Scanning electron microscopy (SEM). As we had previously noted that SeO₃²⁻ precursor concentrations had a strong influence on resultant SeNP size, we examined the enzymatic products while varying SeO₃²⁻ concentrations from 40 μM to 10 mM in the presence of a fixed concentration of enzyme. NADPH was present in excess except for samples containing 10 mM SeO₃²⁻. GRLMR-SeBP produced SeNPs of diameter 37.38 ± 5.75 nm (n = 533), 32.60 ± 6.29 nm (n = 546), 37.34 ± 7.12 nm (n = 1288), and 45.89 ± 6.98 nm (n = 694) when the reactions contained 40 μM, 1 mM, 5 mM, and 10 mM HNaSeO₃, respectively (Figure 3, left column). Thus, at the lower three ‘physiological-like’ concentrations of SeO₃²⁻, the SeNP diameters were of identical size within measurement error. In contrast, GRLMR sans SeBP produces SeNPs that are larger, more polydisperse, and more prone to aggregation. GRLMR produced SeNPs of diameter 48.48 ± 17.63 nm (n = 1078), 61.59 ± 24.99 nm (n = 846), and 59.92 ± 21.19 nm (n = 1172) at HNaSeO₃ concentrations of 1 mM, 5 mM, and 10 mM, respectively (Figure 3, right column). We could not satisfactorily measure the diameters of particles produced by GRLMR at 40μM, so we do not report this value, although we do show a representative micrograph, implying that the resultant particles are much smaller than at higher concentrations.

Figure 3.

SEM images of SeNPs produced by either GRLMR-SeBP (left column) or GRLMR (right column). Rows correspond to the concentrations of SeO₃²⁻ used in the reactions.

Overall, the SEM images show that GRLMR-SeBP produces spherical, well-dispersed products of a narrow size range independent of SeO₃²⁻ concentration. This contrasts with GRLMR, wherein the SeNPs are of a much wider size range, not as apparently spherical, and prone to aggregation (consistent with the DLS measurements). Notably, when GRLMR was used to make SeNP products for these SEM experiments, precipitates were visible in the reaction vessels. Precipitation of larger particles and/or aggregates may have excluded larger materials from the SEM analysis, biasing our results toward measuring smaller Se⁰ products.

The narrow size distribution of particles produced by GRLMR-SeBP, implies that the SeBP or SeBP-GRLMR concatamer can discriminate among SeNP particle sizes. To probe this, amorphous red SeNPs with average sizes of of 30 nm, 50 nm, 125 nm, and 900 nm were abiotically synthesized by borohydride reduction. Figure S6 plots the DLS results showing these average sizes. We added GRLMR-SeBP to each of these SeNPs size preparations, at a final concentration of 100 μg/mL. We allowed 30 minutes to pass before assaying for GRLMR-SeBP / SeNP binding.

A centrifugation assay, similar to the pull-down assay of figure 2, was performed by spinning the solutions at 25,000 ×g for 20 minutes at 4 °C. Protein concentrations of the NP-bound (precipitate) and NP-unbound (soluble) fractions were determined and shown in Table 1.

Table 1.

Results of the centrifugation assay showing the percent of GRLMR-SeBP bound to various SeNP sizes.

Sample	Absorbance @ 595 nm	% Unbound GRLMR-SeBP	% Bound GRLMR-SeBP
Control	0.237	100%	N/A
30 nm SeNP	0.132	56%	44%
50 nm SeNP	0.130	55%	45%
125 nm SeNP	0.236	100%	0%
900 nm SeNP	0.188	79%	21%

The data reveals a preference for the association of GRLMR-SeBP with smaller SeNPs when compared to larger SeNPs. While nearly 50% of the enzyme is particle associated when the particles are 50nm diameter or smaller, when particles are larger, the largest bound fraction observed is 21%. The SeNP sizes that are apparently binding-preferred (50 nm and smaller) are in good agreement with the average size of SeNPs of ~35nm diameter when the SeNPs are enzymatically produced. Whereas nearly 100% of GRLMR-SeBP may be bound to SeNPs when the SeNPs are enzyme-synthesized, the smaller bound fractions here are attributed to the abiotic synthesis of the SeNPs, which may result in somewhat different particle surfaces and interfere with SeNP/SeBP interaction.

Because of the dramatic difference in SeBP size, morphology and aggregation state that depends on SeBP concatenation to GRLMR, we were curious if this concatenation also alters the kinetic properties of the enzyme. To explore this possibility, we monitored NADPH concentrations spectroscopically during GRLMR enzymatic reactions to establish fundamental enzymatic kinetics in the presence and absence of the peptide. Figure 4 depicts the plots of V₀s attained by the GRLMR with and without SeBP over a range of substrate concentrations. Each point was performed in triplicate to calculate standard deviation.

Figure 4.

V₀ plotted against substrate concentration of GRLMR-SeBP (▪) and GRLMR (•) comparing (A.) HNaSeO₃ and (B.) GSSG.

The activities of GRLMR and GRLMR-SeBP, as judged by K_M and k_cat values, are markedly different when HNaSeO₃ is used as the substrate (Figure 4, A). K_M and k_cat for both GRLMR-SeBP and GRLMR are shown in Table 2 against the substrates HNaSeO₃ and GSSG. The K_M of GRLMR-SeBP (0.22±0.06 mM) decreased compared to GRLMR (1.92±1.28 mM), indicating a more favorable enzyme-substrate complex. Measurements for k_cat were 40±2 min⁻¹ and 23±9 min⁻¹ for GRLMR-SeBP and GRLMR, respectively, indicating the peptide modification results in a faster enzyme for the reduction of SeO₃²⁻.

Table 2.

Kinetics values K_M and k_cat for GRLMR-SeBP1 and GRLMR for the substrates HNaSeO₃ and GSSG.

	HNaSeO3		GSSG
	GRLMR-SeBP	GRLMR	GRLMR-SeBP	GRLMR
KM (mM)	0.217±0.057	1.921±1.279	0.365±0.043	0.253±0.033
kcat (min−1)	40.316±2.486	22.823±8.839	14784±720	14498±624

The differences in activity indicated by these kinetic experiments, notably, are also reflected in the SEM examination of enzymatic product (Figure 3, vide supra). Specifically, at the lowest concentrations of SeO₃²⁻ (40uM) we observe no enzymatic product for GRLMR but do observe it for GRLMR-SeBP. At higher concentrations of HNaSeO₃ (10 – 40 mM) both GRLMR-SeBP and GRLMR had the same overall activity within error (data not shown), and both enzymes produce abundant nanoparticle product.

In the reduction of GSSG, the k_cat for the two enzymes was unaffected by the presence of the SeBP as both values were within error of each other. These results indicate that the presence of SeBP on GRLMR does not hinder native function of the enzyme but does increase the enzymes ability to reduce a secondary substrate, SeO₃²⁻. We hypothesize that some of the increased activity of GRLMR-SeBP arises from the peptide introducing 5 positively charged residues proximal (34 Å) to the enzyme active site. The introduction of these residues results in a favorable charge-charge interaction with the negatively charged SeO₃²⁻ anions and may allow for a higher effective concentration of the substrate near the active site.

Raman Spectroscopy allows structural insight into the nature of the interaction between SeBP and the SeNP surface. Figure 5 shows an overlay of the Raman spectra of the peptide in the presence and absence of SeNPs, where the large differences in the spectra suggest a substantial interaction between peptide and nanoparticle.

Figure 5.

Raman spectra of bound (dotted) and unbound (solid) SeBP. The peak at ~1230 cm⁻¹ represents an approximate assignment depicting the shift of the amide III mode to a β-sheet like structure when bound to the surface of the particle. The broad peak seen at 1680 cm⁻¹ indicates a random or α-helical structure of the SeBP backbone when SeNP is omitted from the sample.

We interpret these spectral changes in the presence of SeNPs as arising from two changes in the peptide: First, a change in the peptide backbone conformation; Second, multiple changes to the ligation environment of the imidazole rings of the histidine residues in the peptide.

Changes in peptide backbone conformation are ascertained in Raman spectra from the so-called Amide I, Amide II and Amide III vibrational modes. This corresponds to complex amide related vibrational modes in the peptide backbone. The amide I backbone mode is located in the 1600 – 1700 cm⁻¹ region.⁴⁵ This mode is considerably influenced by transition dipole coupling, which describes the conformational dependence of the dipole interaction energy on spatial separation and orientation.^45,46 The amide III mode is located in the 1200 – 1320 cm⁻¹ region and is also sensitive to structural rearrangement.^45–47

The amide I mode is in the 1650 – 1710 cm⁻¹ region as shown in Figure 6. The broadness and location of peak corresponding to the amide I mode in the absence of SeNPs is suggestive of a disordered backbone structure (Figure 6, right panel, solid trace). In the presence of SeNPs, the peak corresponding to the amide I mode narrows and downshifts to 1667 cm⁻¹, suggestive of a β-sheet type backbone structure (Figure 5, dashed trace). This assignment of a change in backbone structure from disordered to β-sheet-like after SeNP binding interaction is also suggested in the amide III mode.

Figure 6.

Traces of SeBP (solid) and SeBP/SeNP (dashed) from Figure 5, expanded to the 1540 – 1660 cm⁻¹ region to show the relevant His modes described. The marks are color coded to the structure that represents each shift shown. Histidine is represented in the lower right corner.

The amide III mode is in the 1200 – 1320 cm⁻¹ region as shown in Figure 5, left panel. and when unbound the mode may be hidden behind the His modes centered at 1268 cm⁻¹ (Figure 5, solid trace). An amide III mode centered about 1260 cm⁻¹ suggests an unordered structure, and one that is centered at 1265 cm⁻¹ is suggestive of either an α-helix or a polyproline II (PPII) type structure. The presence of PPII structure is likely, since the peptide contains a Pro residue, which is the cause of this structure type.³¹ Furthermore, the presence of a proline makes helical structure unlikely in a dodecapeptide. Upon binding to a NP, we observe the amide III mode downshifts to 1230 cm⁻¹ which also corresponds with a β-sheet structure.^47,48

In addition to the evidence for an ordering in backbone structure that is induced by the presence of SeNPs, we also observe evidence that binding of the SeNP to the peptide is driven by interactions with the imidazole rings of the 5 His residues in the peptide. Imidazoles are well-known to coordinate with metal ions. In the context of Raman spectra, tautomer markers in the 900 – 1630 cm⁻¹ region are used to identify His-metal binding.⁴⁹

By convention, each atom within the imidazole ring of His is labeled as shown in the Lewis diagram of histidine shown as an inset in Figure 6. Vibrations arising from bonds between nitrogen and carbon atoms (and bonded hydrogen atoms) that are labeled C², C⁴, C⁵, N^τ and N^π are of special interest in examining His-metal binding.^49–52 Tautomer markers used to assign His – SeNP binding are listed in Table S2.

The Figure 6 peak spanning 1631 – 1636 cm⁻¹ correlates with the histidine C4-C5 stretch mode, consistent with a specific Nτ-H/Nπ-H protonation state (HisH²⁺). This assignment is supported by the presence of the peak at 1268 cm⁻¹ (Figure S7), which also corresponds to a HisH²⁺ imidazolium ion. In the SeNP bound SeBP trace, the C4-C5 stretching mode downshifts to a broad peak centered at 1571 cm⁻¹. The features of this peak are shown in Figure 6 and are color-coded with the respective vibration depicted. The region of the peak from 1565 – 1573 cm⁻¹ likely corresponds to a neutral imidazole form (HisH) that is unbound with the Nτ tautomer protonated and the Nπ unprotonated. The portion of the peak above 1573 cm⁻¹ likely corresponds with a metal bound His in the form Nτ-H/Nπ-M.

This assignment is supported by the shift in the band at 1268 cm⁻¹ (Figure S7).⁴⁹ Upon metal binding to SeNP with the tautomer form Nτ-H/Nπ-M, the band splits into a doublet with peaks at 1267 cm⁻¹ and 1271 cm⁻¹. The 1267 cm⁻¹ band is attributed to HisH in the form Nτ-H and unbound Nπ whereas the 1271 cm⁻¹ band correlates with a Nτ-H/Nπ-M form, which agrees with the findings by Takeuchi.¹⁵ The peak at 1556 cm⁻¹ is likely due to a metal bridging form of His (Nτ-M/Nπ-M) with the corresponding peak arising at 1292 cm⁻¹ (Figure S7).

Overall, the Raman spectra of the peptide in the presence and absence of SeNPs support a specific interaction between SeBP and SeNPs. Here, SeNP binding is mediated primarily by previously described His-metal ligation interactions, and the aggregate of these interactions appears to drive a change in peptide backbone conformation that is consistent with a change from random coil to beta-sheet-like.

Conclusions

In this work we have demonstrated the added functionality of a metal reducing enzyme concatenated to a dodecapeptide selected for its affinity to SeNPs. When fused to the enzyme, SeBP arrests SeNP growth to a uniform size and shape. GRLMR-SeBP also showed a much greater affinity for the SeNP products increasing the colloidal stability of resulting SeNP. Surprisingly, the fusion of SeBP to GRLMR was also able to increase the activity of GRLMR at lower SeO₃²⁻ allowing for SeNP production even at concentrations in the μM range. The results presented here GRLMR-SeBP has potential applications in bioimaging and cancer therapy and will need to be explored further as such.

Methods

Materials

The Ph.D.™ Phage Display Kit, BSA, BspQI, T4 Ligase, Q5 High Fidelity polymerase, dNTPs, and BL21(DE3) E. coli were purchased from New England Biolabs. Antibiotics were purchased from GoldBio. Na₂SeO₃ and HNaSeO₃ were purchased from Alfa Aesar. NADPH was purchased from BioVision and Coomassie Plus Bradford Reagent from Thermo Scientific. GeneJet Plasmid Miniprep Kit (Cat# K0503) and PCR Cleanup Kit (Cat# K0702) were purchased from ThermoFisher Scientific.

Protein isolation and characterization

10mL cultures of BL21(DE3) cells containing the GRLMR-SeBP or GRLMR were started and grown O/N in a shaker at 37°C and 225 RPM. The dense cultures were diluted into 1 L of LB Kan/Cam and allowed to grow until an OD₆₀₀ ~0.5–0.6. Induction was started using a 1 mM final concentration of IPTG and was supplemented with 1 μM of HNaSeO₃. Growth was O/N at 37°C. Cells were then spun down at 14000 RPM for 20 minutes and resuspended into B-PER and sonicated to lyse the cells. The insoluble cell debris was removed by centrifugation and the soluble cell lysate was collected for Ni-NTA purification. Nickel columns were prepared using Ni-NTA agarose beads. Beads were washed as follows: 3 × 3 column volumes of H2O, 3 × 3 column volumes of binding buffer (50 mM Tris-HCl (pH 8), 5 mM imidazole, 100 mM NaCl). The lysate was then run through the column 3 × before 4 cycles of washing the column using washing buffer (50 mM Tris-HCl (pH 8), 20 mM imidazole, 300 mM NaCl). Finally, the column was incubated with column volume of elution buffer (50 mM Tris-HCl (pH 8), 300 mM imidazole, 50 mM NaCl) for at least 5 minutes before the elution buffer was collected. The isolated protein solution was dialyzed into PBS before the concentration was collected using UV-Vis and stored at −80°C in aliquots for further study. A native PAGE gel was then run to ensure the positively charged SeBP was present on the isolated GRLMR.

Enzymatic SeNP formation

100 μg of enzyme and aliquots of a 100 mM HNaSeO₃ solution were added to PBS, pH 7.4. The reaction was then started by the introduction of NADPH and allowed to react for several hours. After which time, the SeNPs were spun down and separated from the supernatant for further study.

Protein Bradford Assay

Stocks of enzyme in PBS, pH 7.4 at various concentrations were prepared of which 100μL were diluted by 900μL of Bradford reagent. Standard curves were collected for both the GRLMR and GRLMR-SeBP by monitoring the absorbance at 595nm, each point being performed in triplicate. Samples were then prepared for measurement in the same fashion by taking 100μL of the target solution and diluting it up with 900μL Bradford reagent. The concentration was then calculated by monitoring the absorbance at 595nm and performed in triplicate.

PAGE Electrophoresis

Native gels were prepared as follows: 1.25 mL of 40% polyacrylamide, 1.25mL of Tris buffer (pH 8.8), and 50 μL of a 10% ammonium persulfate (w/v) were diluted into 2.45mL of milli-Q water with or without 50 μL of a 10% SDS solution for denaturing or native gels. Polymerization was initiated by adding 7μL of TEMED before pouring the PAGE solution into a cast and allowing to solidify. PAGE gels were run in tris/glycine, pH 8.3 buffer with or without SDS at 150 V for 2.5 h at 4°C. For pH 6.6 PAGE gels, the gel buffer and running buffer was replaced with buffer containing 25 mM histidine and 30 mM MOPS giving a buffer of similar ionic strength and pH of 6.6.³⁹ Gels were then submerged in Coomassie blue and microwaved for 30 sec and incubated for an addition 5 min at RT. Coomassie stain was then replaced by milliQ water and microwaved for 5 min, then washed again in the same way before imaging. Native gels would also be run and later soaked in 5 mM HNaSeO₃ and 1 mM NADPH inside of a plastic bag under nitrogen. These gels would result in red bands of reduced selenium that could then be visualized the next day.

Dynamic Light Scattering

Reactions for DLS monitoring were prepared in disposable plastic cuvettes as described above. SeNP formation was monitored using a refractive index of 2.6 and an absorbance of 0.5 for α-Se and a refractive index and viscosity of PBS of 1.332 and 0.8898 cP, respectively. Reactions ran for at least 4 hours at RT. Each point in DLS was an aggregate of 1 – 14 reads depending on the quality of data collected, determined by the Zetasizer Nano ZS software.

K_M and k_cat Studies

Initial velocities (V₀) of the GRLMR-SeBP were performed in various concentrations of HNaSeO₃ or GSSG while monitoring 340 nm correlating to NADPH absorbance. The reactions were run in PBS, pH 7.4 and with 13 μg of enzyme with substrate SeO₃²⁻ (1.5 μg of enzyme with substrate GSSG) and 200 μM of NADPH with 0.05 mM – 2 mM of substrate. Enzymatic V₀s were plotted against substrate concentration in OriginPro from which K_M and k_cat could then be calculated. Each data point was run in triplicate to calculate standard deviation.

Synthesis of SeBP capped Selenium Nanoparticles (SeBP/SeNPs)

Peptide-capped SeNPs were synthesized based on the method from Nath et. al. For the SeNPs, 50 – 100 μL of acidic 10 mM SeO₃²⁻ and 0.5 – 1.0 mL of 10 mM NaBH₄ (aq) were added to a 15 mL conical tube and diluted to a final volume of 1.5 – 3.0 mL with milliQ water. The solution was mixed and placed on a rocker for 90 – 300 seconds, after which 50 – 100 μL of 10 mM of the SeBP was added and thoroughly mixed into the solution and placed back on the rocker. Within 10 minutes the tube was placed in an ice bag and allowed to conjugate for 4 hours on the rocker. After this, the solution was dialyzed using a 3,500 MWCO cassette on ice in 2.0 L milliQ water for at least 2 hours. The resulting mixture was lyophilized and stored in a refrigerator until future analysis. The sample was diluted into 50 μL PBS, pH 7.4.

Raman Spectroscopy

Raman spectra were collected using an inverted Raman microscope with an Olympus IX73 frame and objectives with a Horiba iHR 550 Spectrometer with a neural synapse thermoelectrically cooled charge-coupled device (CCD) detector attached to a Horiba ONDAX T-Hz Raman 532 nm laser provided by Justin Sambur. This setup was accompanied with a LabSpec software package. The specimen was prepared by drop-casting 3–7 μL of sample onto a glass cover slip and allowed to dry in air at room temperature. We used double-sided tape to seal the sample and to adhere the coverslip to a glass slide. Spectra were collected using an incident laser power of 83 mW. We used a 60x water objective with a 1200 blazes/mm grating, which has a resolution of approximately 2 cm⁻¹ per pixel. We manually focused the laser on the sample using the optical setup. If signal was insufficient, we refocused the laser until signal was obtained. We used the software’s denoiser program, which is essentially a smoothing algorithm, to obtain a smoother curve. Backscatter collection ranged from 30 – 300 seconds per acquisition, and a total of 1 – 15 spectra were accumulated and averaged, depending on the level of noise. Any spikes caused by cosmic rays were removed using the software’s spike removal function.

Centrifugation Assay

A centrifugation assay was performed using inorganically synthesized SeNPs to test whether GRLMR-SeBP exhibited size preference. Bare, amorphous SeNPs of varying size were produced by varying the concentration of HNaSeO₃ precursor by adding 100 μL of 1 mM, 10 mM, 100 mM, and 1 M stocks diluted into a final volume of 1.1 mL 1x PBS pH 7.4 and 10 mM NaBH₄ at 4 °C. The reactions were allowed to proceed for 30 minutes, after which DLS size data were recorded using a Malvern Zetasizer Nano ZS. After the SeNP reaction was allowed to proceed for 30 minutes, GRLMR-SeBP was added at a final concentration of 100 μg/mL and allowed to incubate for an additional 30 minutes before centrifugation. The samples were all spun at 15,000 rpm (~25,000 ×g) for 20 minutes at 4 °C. Unbound enzyme concentration was found using a modified Bradford Assay to measure absorbance at 595 nm of 100 μL aliquots of the supernatant mixed with 900 μL Bradford Reagent. The concentration of unbound enzyme was normalized to a control composed of 100 μL of 100 μg/mL GRLMR-SeBP in 1x PBS pH 7.4 mixed with 900 μL Bradford Reagent.

ABBREVIATIONS

GRLMR

glutathione reductase-like metalloid reductase

SeBP

Se(0) binding peptide

EM

electron microscopy