Abstract

Hydrogen sulfide (H2S), independently of being a toxic gas with a characteristic smell of rotten eggs, is a crucial signaling molecule with significant physiological functions. Given the rapid diffusivity of the gas, it is a challenge to develop robust sensors and biomarkers to quantify free or bound H2S. In addition, there is the need to further develop a robust biosystem to efficiently trap or scavenge H2S from different producing environments. The work presented here uses recombinant met-aquo rHbI (rHbI-H2O) immobilization techniques on collagen to determine its ability to bind H2S due to its high affinity ( M-1). The hemeprotein will function as a scavenger on this scaffold system. UV-Vis absorption and UV-Vis diffuse reflectance (%R) spectroscopy of rHbI-H2O and rHbI-sulfide (rHbI-H2S) complex in solution and collagen scaffold demonstrated that the heme chromophore retains its reactivity and properties. UV-Vis diffuse reflectance measurements, transformed using the Kubelka-Munk function (K-M function), show a linear correlation ( and 0.9916) of rHbI-H2O and rHbI-H2S within concentrations from 1 μM to 35 μM for derivatives. The extraordinary affinity of rHbI-H2O for H2S suggests recombinant met-aquo HbI in a collagen scaffold is an excellent scavenger moiety for hydrogen sulfide. These findings give insight into H2S trapping using the rHbI-H2O-collagen scaffold, where the rHbI-H2S concentration can be determined. Future pathways are to work toward the development of a met-aquo rHbI collagen solution capable of being printed as single drops on polymer, cotton or chromatographic paper. Upon exposure of these matrixes to H2S, the rHbI-H2S complex is formed and its concentration determined using UV-Vis diffuse reflectance technique.

1. Introduction

Human physiology is influenced by exogenous hydrogen sulfide (H2S) at higher concentrations than 300 μM leading to death at 15 mM [1, 2]. However, significant discoveries suggest the endogenous synthesis of H2S in mammalians. Desulfuration of cysteine, homocysteine, and cystathionine by the enzymes cystathionine β-szynthase (CBS), cystathionine ɣ-lyase (CSE), D-amino acid oxidase (DAO), and indirectly by 3-mercaptopyruvate sulfurtransferase (3-MST) produces H2S in the body [3]. The amount generated can range from 10 to 300 μM or lower depending on enzyme distribution in the particular tissue [46]. H2S plasma values that vary between 50 and 150 μM are associated with several conditions [4]. For instance, SARS-CoV-2 survivors show serum H2S levels higher (≥150 μM) than those nonsurvivors [7]. Overexpressed CSE in the pancreas increases H2S levels, inducing the pathogenesis of diabetes in rats [8]. In contrast, reduction in the enzymatic activity of CSE promoted hypertension in rats and women with preeclampsia [9, 10]. Reduced H2S levels are also related to Parkinson’s, Alzheimer’s, and atherosclerosis disease [11, 12]. Despite this knowledge, fundamental questions and concerns remain. For example, H2S concentration values are subject of intense dispute because there is no suitable robust sensor capable of directly measuring H2S in tissues and plasma [13]. Therefore, an attractive alternative is the development of an H2S acceptor that can efficiently trap hydrogen sulfide gas from biological samples for detection and quantification.

Hydrogen sulfide is a lipophilic molecule that can diffuse through membranes without facilitating membrane channels [10]. At pH 7.4, there is approximately 80% HS- and 20% H2S. Hydrogen sulfide solubility is (~113 mM, 86 mM, and 68 mM at 20°C, 30°C, and 40°C, respectively) a function of temperature; hence, determining this gas concentration is complicated [14]. Therefore, the hydrogen sulfide generation rate, turnover, concentration, identification, products, transport, and concentration remain a significant challenge in tissues, organs, and plasma. The methylene blue method is the standard procedure for the quantitative determination of hydrogen sulfide, but its application is restricted to solutions using hazardous reagents [13]. Also, amperometric sensors have been reported for H2S detection [1518], being the polarographic hydrogen sulfide sensor (PHSS) one of the most cited [19].

The recombinant hemoglobin I (rHbI) from the clam Lucina pectinata (Figure 1) with the heme in the FeIII oxidation state has an extraordinary affinity for H2S (M-1 s-1/ s-1) of M-1, which makes met-aquo rHbI exceptionally suitable for the detection of H2S [20]. Therefore, the underlying hypothesis is that recombinant met-aquo HbI (rHbI-H2O) collagen scaffold moiety could be a robust biosystem to efficiently trap or scavenge H2S from solution. Although previous literature has supported that rHbI can trap the H2S in different systems, its immobilization on collagen has not been yet studied. Previously, hemoglobin I had been immobilized as hexa-histidine tag rHbI in two gold electrodes detecting 25 to 800 nM H2S in solution [21, 22]. Additionally, immobilization of hexa-lysine tag rHbI on carbon nanotubes via covalent conjugation showed a solution electrochemical response toward 10 to 300 μM H2S [23]. Simultaneously, efforts have measured H2S as gas evolving from a biological system. For example, a silver/Nafion/polyvinylpyrrolidone (PVP) membrane on a polystyrene microplate array was used to quantify H2S from live glioma cells [24]. Furthermore, PVP membranes containing silver/Nafion were printed on chromatography paper to detect H2S in live cell cultures [16, 25]. Gold/Silver-iodide dimeric nanoparticles were immobilized in agarose gel to create test strips capable of measuring H2S from HepG2 cells with good sensitivity (500 nM), selectivity, and stability [26]. met-aquo rHbI encapsulated in tetramethyl orthosilicate (TMOS) sol-gel accepts hydrogen sulfide liberated from the rHbI-H2S complex in solution, -4 s-1, revealing that met-aquo rHbI in TMOS is an effective trap for H2S [27].

However, despite their advantages, TMOS sol-gels are difficult to work with and maintain in an overturned position [27]. This is not the case for collagen-based matrices, which have been used as assemblies to develop applications, such as tissue engineering, drug delivery, and biosensors [2834]. For example, a collagen-based electroconductive hydrogel was used to prove the concept of an injectable sensor of glucose in vivo and tissues [28]. Also, colorimetric data of collagen and gold nanoparticles moieties were used to detect glucose in a linear range of 3 to 25 mM [35]. Moreover, in vivo oxygenation for skin wounds in diabetic mice was tracked using a collagen-dextran conjugated with a phosphorescence oxygen sensor [36]. Furthermore, a peroxide biosensor was constructed upon the encapsulation of hemoglobin in a ZrO2-grafted collagen scaffold [37]. In contrast, a collagen hydrogel containing JK1, an H2S releaser, was developed to deliver H2S as a treatment for intervertebral disc degeneration [38]. Similarly, research has also been focused on developing matrixes to immobilize other H2S-donors, e.g., JK2 and GYY4137, to release hydrogen sulfide from silk fibroin scaffold [39, 40], fibrous membranes [41], polymers [42], and sodium alginate sponges [43]. These systems have been created to mimic physiological release and proposed as alternatives to H2S therapies.

The work presented here takes advantage of collagen versatility. A met-aquo rHbI collagen scaffold was constructed from a collagen sponge as an immobilization matrix for met-aquo Mb (Mb-H2O), met-aquo rHbI (rHbI-H2O), and rHbI-sulfide (rHbI-H2S) species. The scaffold hemeprotein complexes were analyzed with UV-Vis absorption and UV-Vis diffuse reflectance measurements (R%) represented by the Kubelka-Munk function.

2. Materials and Methods

2.1. Sample Preparation

The recombinant His-tagged hemoglobin I was expressed in Escherichia coli Bli5 cells and purified as previously detailed in the literature [20, 44, 45]. Briefly, the expression was conducted using Terrific Broth supplemented with 30 μg/ml hemin chloride and 1% glucose. Protein expression was induced with 0.001 M isopropyl β-D-1-thiogalacto-pyranoside (IPTG) at 30°C and yielded red cell pellets, which were eventually lysed and centrifuged. The soluble protein fraction was purified using Co2+ affinity columns (Talon, Invitrogen) followed by fast performance liquid chromatography (FPLC) in a Hi Load 26/60 Superdex 200 gel filtration column with an automated AKTA FPLC System (Amersham Biosciences). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) stained with Coomassie blue was performed on the purified protein. One band around 17 kDa confirmed the purity and integrity of the hexa-histidine tag rHbI [44]. The met-aquo rHbI and met-aquo Mb from equine skeletal muscle were prepared by heme oxidation, adding a 10% molar excess of potassium ferricyanide. After the reaction, the oxidant was removed with a centrifugal filter (Pall Corporation). Solutions containing from 1 to 35 μM of met-aquo derivatives have been prepared by transferring different aliquots (1-70 μL) of 0.3 mM stock solution to 1 cm square quartz cuvettes (Starna Scientific) filled with buffer. The H2S stock solution was prepared by dissolving Na2S∙9 H2O salt in an amber vial and completed until 1.5 mL with buffer. Aqueous solutions were prepared with buffer containing 100 mM succinic acid, 100 mM potassium dihydrogen phosphate, and 1 mM ethylenediaminetetraacetic acid (EDTA) and adjusted to pH 6.5. The Na2S∙9 H2O salt and buffer were purged and degassed before mixing to prevent oxygen contamination. Integra Life Sciences donated ultrapure lyophilized collagen sponges derived from the bovine tendon. Terrific Broth, hemin chloride, glucose (C6H12O6, ≥99.5%), isopropyl β-D-1-thiogalacto-pyranoside (IPTG, ≥99%), myoglobin from equine skeletal muscle (Mb, ≥95-100%), potassium ferricyanide (K3[Fe(CN)6], ACS reagent, ≥99.0%), sodium sulfide nonahydrate (Na2S∙9 H2O, ≥99.99%), succinic acid (C4H6O4, ≥99.0%), potassium dihydrogen phosphate (KH2PO4, ACS Reagent, ≥98.0%), and ethylenediaminetetraacetic acid (EDTA, ≥99.0%) were supplied by Sigma-Aldrich (USA) and used as received.

2.2. Hemeprotein Immobilization on Collagen Sponge

Immobilization of met-aquo Mb and met-aquo rHbI solutions was achieved individually by depositing 40 μL of protein solution (1-35 μM) onto the center of a collagen sponge with  cm in length, width, and height (thickness), respectively. All protein immobilization was done aerobically at 25°C. The collagen significantly reduced thickness where the hemeproteins drop solution was added. Thus, thickness size measurements were done on the spot for the met-aquo Mb and met-aquo rHbI collagen scaffold moieties. Ten samples were analyzed for the spot thickness (height) of the proteins collagen scaffold composite using a Leica MZ16 stereomicroscope coupled with a Nikon Digital Sight DS-5Mc-U1 camera. Each measure was repeated ten times for each sample, and a standard deviation was calculated for accuracy. The results showed that collagen with immobilized heme proteins had an average spot thickness of 0.041 cm. The rHbI-H2S complex was prepared by adding 6 μL of 3.7 mM H2S stock solution, equivalent to 482 μM, to the met-aquo rHbI collagen scaffold. Thus, the hydrogen sulfide ratio to met-aquo rHbI concentrations (1-35 μM) was 482 to 14.

2.3. Absorbance and Reflectance Measurement of Hemeproteins on Collagen

The concentration of met-aquo Mb and rHbI stock solution was determined by the rearrangement of the Beer-Lambert law equation, , where is the protein concentration (mM), is the sample absorbance (unitless), is the sample light path length (1 cm), and is the extinction coefficient for the particular proteins. For met-aquo Mb, met-aquo rHbI, and the rHbI-H2S complex, the extinction coefficient was 188 cm−1 M−1 at 407 nm, 178 cm−1 M−1 at 407 nm, and 102 cm−1 M−1 at 426 nm, respectively [46, 47]. UV-Vis absorbance measurements were performed using a Shimadzu UV 2700 spectrophotometer.

The UV-Vis diffuse reflectance measurements of the hemeprotein collagen scaffold spot were monitored using the Shimadzu UV 2700 spectrophotometer equipped with an MPC 2600 unit, providing accurate reflectance measurements of hemeprotein on collagen. Figure 2 shows the standard plate setup on the reference (R) and sample (S) sides. Barium sulfate powder on both sides was used as a standard white plate to execute the diffuse reflectance baseline correction. The immobilized protein then replaced the S side to start data collection. Collagen sponges were mounted on the instrument, but only the center of the strip was exposed to radiation. The spectra were collected employing an incident angle of 0 degrees, a photomultiplier with 5.0 nm slit-width, 0.5 nm sampling pitch, 0.1 nm resolution, and six scans. A relation between the Kubelka-Munk function obtained from diffuse reflectance measurement has been described in the literature [4850]. Therefore, all acquired reflectance spectra were converted using a Kubelka-Munk function by the UVProbe 2.3 software (Shimadzu Corporation). The Kubelka-Munk function represents the UV-Vis diffuse reflectance measurements, , where is the experimental reflectance intensity [5154].

Diffuse reflectance data were obtained for met-aquo Mb and met-aquo rHbI collagen scaffold in two different scenarios. First, the spectra were acquired on the collagen sample 30 minutes after the added hemeprotein drop solution. Then, these samples were stored at 4°C for 24 hours, and diffuse reflectance data were collected. In both circumstances, the spectra were the same. Next, the rHbI-H2S complex on the collagen scaffold was prepared by reacting the immobilized met-aquo rHbI with hydrogen sulfide. Finally, the diffuse reflectance data was converted to the Kubelka-Munk function, facilitating monitoring of the transformation of met-aquo rHbI to rHbI-H2S. The hydrogen sulfide heme complex is evidenced by the decreased intensity of the met-aquo rHbI 407 nm transition and the increased 426 nm band of rHbI-H2S derivatives. Also, the electronic transitions of rHbI-H2S species at 542 nm and 574 nm support the complex formation [47, 55]. All the experiments were conducted in triplicate.

3. Results and Discussion

3.1. Absorbance and Reflectance of Hemeproteins on Collagen Sponge

Our study took advantage of the conventional UV-Vis spectroscopy used to measure absorbance in solutions with the UV-Vis coupled to reflection measurements of solid samples. Figure 3 shows UV-Vis absorption spectra of 6 μM met-aquo Mb in solution (blue line) and immobilized on collagen sponge (black line). The inset shows a view of the 450-700 nm region. The results indicate that with identical met-aquo Mb concentrations (6 μM), there is a significant intensity change between the UV-Vis absorption spectrum of the hemeprotein solution (blue line) and the collagen sponge (black line). Attempts were made to improve the absorbance intensity of the met-aquo Mb on the collagen sponge by subtracting a blank collagen sponge. Still, it did not improve the sample signal significantly. The variation in absorption can be attributed to the clearness between the hemeprotein in solution versus the collagen scaffold environment. Although the protein concentration on the collagen spot was not directly calculated, the initial hemeprotein content was identical to the solution concentration. Furthermore, the same immobilized met-aquo Mb sample was used to obtain the UV-Vis absorption (black line) and the reflection data, translated with the Kubelka-Munk function (Figure 3, red line). The data clearly shows that the UV-Vis diffuse reflectance spectroscopy significantly improves the band intensity in the sample. The observation suggests that processes and reactions of hemeproteins that adsorb collagen sponges can be monitored using diffuse reflectance techniques. Also, the identical electronic transitions between met-aquo Mb in solution and adsorbed on collagen at 409, 504, and 633 nm reveal that the heme chromophore pocket structure was preserved. The data also suggests that the surface did not interfere with the characteristic bands of the heme protein active center.

The results show that collagen with immobilized heme proteins had an average spot thickness of 0.041 cm. Therefore, the appropriate thickness should be determined to prevent loss of reflected radiation and avoid misinterpretation of results. This finding supports previous research where the diffuse reflectance revealed a significant depth dependency [56, 57]. Although reflectance is a limited measurement of the superficial analysis, with a penetration depth from 0.5 to 7 mm [52, 58, 59], it still depends on the optical absorption and scattering of the selected surface [57, 58]. The role of immobilized sample wetness versus 24 hours dried environments was evaluated by monitoring variations in the intensity of the UV-Vis reflectance measurements. The spectra were the same in both circumstances, and no significant changes in the linear regression were observed. These results agree with the diffuse reflectance measurements to quantify the met-aquo Mb of a wet meat sample [52] and share similarities with analysis in dry conditions of human skin in vivo [60]. Nevertheless, a slight increase in the slope for the freshly immobilized sample revealed that the water content in the protein spot reflected the light, increasing the reflectance signal [61]. Consequently, it demonstrated that the water content had a small influence in the 350-700 nm region measures because it altered the light path length [56].

3.2. Hemeprotein Distribution on a Collagen Scaffold

Figure 4(a) illustrates the four (4) sites of an expanded drop solution containing 12 μM met-aquo Mb added to a collagen surface. The hemeprotein distribution from the center was evaluated by diffuse reflectance. Figure 4(b) shows reflectance (%) plotted against wavelength, while Figure 4(c) represents the conversion to the Kubelka-Munk function for every colored site. Characteristic bands are typical of those expected for the met-aquo myoglobin derivative, evidencing that the chromophore retained its optical properties. The spectrum with less percent reflectance (blue line) and the intense peaks can be associated with higher protein concentration [58]. The spectrum with a lower transition intensity (black line) was associated with measures obtained at the edge of the spot, with less chromophore. The hemeprotein single drop spread on the collagen revealed irregularities in the sample path length. In reflectance measurements, it is challenging to control path length, which could depend on the concentration of the sample [62]. A highly concentrated sample could result in short average path lengths, while the average path length will increase if the sample has a low concentration. As a consequence, quantitative analysis must be interpreted with caution. Thus, the evidence points out that the measurements should be performed close to the protein’s deposition site to decrease systematic errors and misunderstandings in quantitative analysis.

3.3. Diffuse Reflectance of Recombinant met-aquo rHbI on Collagen Sponge

Spectral analysis of rHbI on a collagen sponge was performed by monitoring the intensity of the UV-Vis diffuse reflectance as a function of the hemeprotein concentration. Figure 5 shows reflectance converted to absorbance using the Kubelka-Munk function for various met-aquo rHbI solutions immobilized in collagen sponges. There is an intense band at 407 nm and small intensity bands at 502 and 633 nm [55]. The higher energy transition is approximately ten times more intense than the lower energy bands. This behavior is similar to that of the protein in the solution. These bands are attributed to ππ electronic transitions of the heme porphyrin being sensitive to the heme oxidation, spin, and coordination states [63]. Thus, the evidence denotes that the HbI preserved its optical properties upon immobilization, suggesting that the heme porphyrin remains stable. These results are consistent with the work done on a slightly different assembly of immobilized myoglobin on collagen studied by UV spectroscopy and electrochemistry [64]. The met-aquo rHbI was analyzed in absorbance from 0.1 to 1.2. An in concentrations ranging from 1 μM to 35 μM established a linear correlation between K-M function and met-aquo rHbI supported on collagen (Figure 5, inset). These results coincided with earlier findings where UV-Vis diffuse reflectance was implemented to quantify mercury in water [65] and myoglobin derivatives in pork and beef [52]. The standard deviation of 0.003 to 0.2 indicates that the data points are close to the mean and exhibit low spreading.

3.4. rHbI-H2S Complex Formation on a Collagen Scaffold

The hydrogen sulfide solution (482 μM) reacts with the met-aquo rHbI collagen scaffold concentrations from 1 to 35 μM to generate a hemeprotein to H2S ratio of 482 to 14. The range was chosen to ensure the reactivity of H2S with met-aquo rHbI to create the rHbI-H2S complex. If necessary, individual complex concentrations could be determined using a molar absorptivity coefficient of 102 cm-1 mM-1 [47]. Figure 6 shows the 426 nm, 542 nm, and 574 nm electronic transition characteristic of the six coordinated low spin ferric rHbI-H2S complexes. Therefore, the data shows that the protein chromophore conformation was preserved after immobilization [47]. A linear correlation () shown in the inset reflected an association between diffuse reflectance measurements and protein concentrations. The standard deviation range was ±0.004 to 0.2, having the most significant deviation at a higher concentration of the rHbI-H2S complex. Various factors can influence the diffuse reflectance measurements, including beam penetration depth and high hemoglobin I concentration [66].

The data shows that 482 fold of H2S to hemeprotein ratio did not produce deoxy heme FeII species, characterized by the Soret transition at 436 nm and dependent on H2S concentration [45, 67, 68]. The reduction could not be observed because the transition was weak and may have remained inside the 426 nm envelope. The formation of the deoxy heme FeII species could have been prevented because the reaction was done aerobically [68, 69], causing a smaller than expected H2S concentration. Also, the collagen matrix may have been stabilizing the rHbI-H2S complex independently of the hydrogen sulfide concentration. However, the rHbI-H2S formation clearly shows a linear dependence of the diffuse reflectance intensities and the rHbI-H2S complex formation. Therefore, recombinant hemoglobin I on a collagen scaffold could act as a hydrogen sulfide scavenger, whose complex concentration can be determined using UV-Vis diffuse reflectance spectroscopy.

4. Conclusions

In our approach, met-aquo Mb and rHbI-H2O were absorbed on a collagen porous scaffold and analyzed using UV-Vis absorption and diffuse reflectance spectroscopy. The techniques also evaluated the reaction between met-aquo rHbI and H2S to generate a stable rHbI-H2S complex on the scaffold moiety. These measures were then transformed to absorbance by the Kubelka-Munk function, which conserved the same peaks as in hemeprotein solution. These results suggest that the hemeproteins active center was not affected upon insertion into the collagen scaffold. Furthermore, the hydrogen sulfide concentration ratio to rHbI-H2O (1-35 μM) was 482 to 14 shows a linear behavior between the K-M function and heme protein concentration. Overall, the results demonstrate the ability of the rHbI-H2O collagen composite to be an excellent trap for H2S by forming the rHbI-H2S complex due to its extraordinary affinity for H2S ( M-1 s-1/ s-1) of ( M-1). Hence, met-aquo rHbI is exceptionally suitable for the detection of H2S [20]. The results also suggest that the process is a direct function of the rHbI-H2O concentration present in the scaffold collagen system. Future work in this direction includes producing a met-aquo rHbI collagen gel capable of being printed as single drops in polymer, cotton, or chromatographic paper, as has been done for other chemical systems [25, 70, 71]. The process determines H2S gas concentrations from different biological sources by measuring the rHbI-H2S concentration in the single drop moiety, using UV-Vis diffuse reflectance techniques, which are simple, nondestructive, and low-cost tools. Although further research is needed, the design of rHbI-H2S-collagen as an H2S donor can extend their use to new therapeutic prospects.

Data Availability

The data used to support the findings of this study are included within the article. Additional information is available upon request by Jennifer Vargas Santiago.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This research was supported by PR-INBRE NIH/NIGMS under Grant No. P20GM103475. In addition, we recognize the support of the Alfred Sloan Foundation. Finally, we are grateful for the guidance provided by the Graduate Writing Facilitators of the Graduate Research and Innovation Center (GRIC) at the University of Puerto Rico-Mayagüez.

Supplementary Materials

Scheme of the immobilization of the hemoglobin I on the surface of a collagen sponge and the measurement of reflectance. (Supplementary Materials)