Visual in Solution Detection of Denatured Insulin Coupled to Gold Nanoparticles in the Presence of an Aminosilane

The pharmaceutical industry is comprised of a myriad of active pharmaceutical ingredients, all applied to increasing human health and quality of life. One of the requirements of reaping the health beneﬁts of these drug products is maintaining them in a native or desired form. Such non-native forms can include, aggregated or fragmented structures. Unfortunately, the most widely used methods to detect non-native or denatured proteins require trained technicians, bulky instrumentation and large amounts of reagents. Deviation from the native structures can occur at all stages; from manufacturing and processing to storage. With these limitations in mind, a simplistic and highly sensitive in solution detection method was evaluated to visually detect denatured insulin proteins, utilizing gold nanoparticle aggregation via 3-Aminopropyltreithoxysilane. The insulin in this study was heat stressed using an 80 ◦ C water bath to create an accelerated heat stressed environment. The insulin, gold nanoparticle and aminosilane solution was then characterized utilizing, UV-Vis spectroscopy, dynamic light scattering and scanning electron microscopy. Captured images and resulting absorbance spectra of the trials demonstrated visual color changes detectable with the human eye as a function of the denaturation time. This work serves as an extended proof of concept for fast in solution detection methods for proteins that have experienced heat stress.

The pharmaceutical industry produces a myriad of active pharmaceutical ingredients, all applied to increasing human health and quality of life. Within this industry there are several components that contribute to the cost of a pharmaceutical drug product. These costs can include the active pharmaceutical ingredient, manufacturing processes/packing, quality testing, as well as administrative and marketing costs. 1 One specific cost associated with the production of pharmaceutical drug products include the man hours, materials and instrumentation applied to the creation of drug formulations. Drug formulations are extremely important as they maintain the native structure of the drug product and prevent undesired biological interactions with the host upon administration to patients. Such non-native forms can include, aggregated or fragmented structures which can cause a series of adverse effects ranging from irritation at the site of injection to mass allergic reactions. Unfortunately, the most widely used methods to detect non-native or denatured proteins require trained technicians, bulky instrumentation and large amounts of reagents. Deviation from the native structures can happen at all stages from manufacturing and processing to storage. Therefore, this work seeks to explore a fast and simple visual detection method with a high sensitivity for structural changes within a pharmaceutical drug product. For the purpose of this work insulin protein from recombinant DNA was utilized as the model protein for the study. Insulin is a hormone produced and secreted by the pancreas. 2 Specifically, within the pancreas exists a number of clustered cells, containing beta cell which are responsible for producing insulin and releasing it into the blood stream. 3 Its structure is composed of 51 amino acids separated in two peptide chains (A and B) connected by two disulfide bonds. [4][5][6] Insulin is a small protein with a molecular weight of approximately 5800 g/mol and a molecular radius of 1.34 nm. [7][8][9] This protein plays a major role in the control of hyperglycaemia for patients with type 1 diabetes and selected patients with type 2 diabetes, by regulating blood-glucose levels. 10,11 According to the National Center for Chronic Disease Prevention and Health Promotion, 9.3% of the population in the United State has some form of diabetes. 12 This translates currently into a product market of about $24 billion dollars (2014 numbers) and is expected to be a $48 billion dollar market by 2020. 13 Worldwide, the population affected by diabetes is increasing as well and has reached staggering numbers of over 350 million individuals. 14 Pharmaceutically and biologically produced insulin can often exist as a hexamer, which has a circular 'doughnut' shape and is stabilized by disulphide bridges. 15,16 Although insulin may exist as a hexamer or a six monomer structure, further aggregation within the system can occur based on changes in optimal environment conditions. 17 For example studies have shown evidence that insulin can aggregate at very low pH values, high temperatures and excessive agitation. 18 Introduction of abnormally aggregated insulin to diabetic patients will further decrease quality of life and overall patient care. The creation of a facile visual detection system that can confirm viability of continued drug use, may provide a high impact solution to the detection of aggregation within drug products. Due to gold nanoparticles (AuNps) possessing high biocompatibility and optical properties, these nano-metals were exploited for applications within the visual detection component of this study. [19][20][21] From previous work, we understood that bovine serum albumin (BSA) could be detected visually using gold nanoparticles in the presence of 3-Aminopropyltreithoxysilane (APTES). 22 The presence of the APTES within the gold nanoparticle solution induces a dipole causing aggregation and subsequent visual changes in the solution from red to purple to blue depending on the concentrations of APTES. The addition of a protein to the solution before the addition of APTES dictates the extent of aggregation upon exposure of the protein/gold nanoparticles solution to APTES. Taking this work a step further, we apply this methodology to the detection of changes in protein structure as opposed to protein concentration. Utilizing the concentration dependent induced aggregation of gold nanoparticles as a function of the protein denaturation time and APTES exposure, this works seeks to visually detect insulin proteins that have been heat stressed.

Experimental
The procurement locations and subsequent purities of all reagents used in this study are as follows. The insulin detemir utilized in this study was manufactured from Levemir (Rdna origin) liquid injection. The gold chloride metallic salt precursor was purchased from Sigma-Aldrich containing a 49% purity. American Chemical Society grade purity trisodium citrate reagent at a 99% purity used to reduce the gold chloride was obtained from Alfa Aesar. The aminosilane, 3-Aminopropyltreithoxysilane was procured from Sigma-Aldrich containing a 98% purity. All chemicals were dissolved in deionized water (DI H2O) without employment of any further purification methods.
Synthesis of the gold nanoparticles used within this study were prepared based on a one pot synthesis. 23 Precursor solutions were prepared by dissolving gold chloride and trisodium citrate in deionized water. The reaction was carried out at approximately 100 • C producing gold nanoparticles that were about 18 nm in size. A simple serial dilution was employed to obtain an insulin concentration of 22.06 ug/mL.
The accelerated heat stress trials consisted of 0.1 mL of the 22.06 ug/mL insulin solution and 0.8 mL of the gold nanoparticle (AuNp) solution. After thorough mixing, 0.1 mL of the respective APTES concentration is then added and subsequently characterized via ultraviolet visible spectroscopy. Protein denaturation was achieved by placing the diluted insulin in a glass vial submerged in a hot water bath set to 80 • C. As the protein was denatured, 0.1 mL aliquots of the protein solution was removed, mixed according to the above mentioned procedure and characterized. To properly capture the instantaneous color of the protein, AuNps, insulin and APTES solutions, a picture is taken directly after the addition of APTES and then loaded onto the UV-Vis.
Scanning electron microscopy images were taken using the HITACHI FE-SEM SU8010 instrument at 8 kV accelerating voltage and 5 milliamps. Dynamic light scattering measurements were taken to determine the aggregate sizes produced as a function of the insulin denaturation and subsequent aggregation of the gold nanoparticles in solution. The size and size distribution of the samples were measured using the Nanotrac Wave by Microtrac. The 60 second trials runs on the Nanotrac Wave by Microtac used a 780 nm laser diode. In order to quantify the extent of visual color changes in the solutions, the LAMBDA 950 UV/Vis/NIR Spectrophotometer by PerkinElmer was employed in absorbance mode to quantify the max peak wavelength shifts. The run parameters for this instrument included a 2 nm data interval and 2 nm slit width run from 800 to 400 nm.

Results and Discussion
The following experiment was executed to quantify the optical changes in the solutions containing denatured vs. non-denatured insulin. As previously described in the Experimental sections, 0.8 mL of the gold nanoparticle solution was mixed with 0.1 mL of the heat stressed insulin. To complete the trials, 0.1 mL of either 0.3% or 0.2% APTES was added to the solution and immediately characterized via ultraviolet visible spectroscopy. In Figure 1 we observe systematic blue shifts in the max peak wavelength and a decrease in the peak broadening as the heat stress time was increased for all trials. Specifically, within Figures 1a and 1c, the insulin was heat stressed at 80 • C using 0.3% and 0.2% APTES respectively. In Figure 1c we observe a significant decrease in the peak broadening after only 1 minute of heat stress. The resulting absorbance spectra of the 1-minute trial in Figure 1c is almost indistinguishable from the control of gold nanoparticles and native insulin without the addition of APTES. In contrast, it takes 6 minutes for the trial containing 0.3% APTES to produce an appreciable decrease in the peak width. In Figure 1b the heat stress temperature was decreased to 70 • C and we observe the temperature dependence of absorbance spectra as well. In Figure 1b it took 3 minutes of the accelerated heat stress to induce regression of the max peak wavelength back to approximately 500 nm. In Figure 1d a higher APTES concentration of 0.4% was utilized and the subsequent blue shifts were also monitored as a function of the accelerated heat stress.
The changes in the absorbance spectra of the samples correspond to visual in solution color detection of the denatured insulin utilizing the gold nanoparticles and APTES. Figure 2 displays the solution hues of the insulin, gold nanoparticles and APTES trials. In Figure 2a, the resulting solution hues of the 80 • C heat stress and 0.3% APTES trials were captured. Within Figure 2a, as the insulin is heat stressed, the solution transitions from a purple hue back to a pink/red hue indicating visual detection at 6 minutes of heat stress. Figure 2b displays the 70 • C heat stress and 0.2% APTES trial which produced a 3-minute detection with a similar trend of a purple to red color transition. Figure  1c    occurred after 12 minutes to a lavender purple. If the heat stress of the insulin was continued, it is projected that there would be a defined level of insulin denaturation that would produce a red hue and give a second detection marker for the 0.4% APTES concentration trial.
In order to further characterize the in solution detection of denatured insulin, the trials were imaged using scanning electron microscopy and the size and size distributions of the samples were quantified via dynamic light scattering. Figure 3 displays the scanning electron microscopy images of the gold nanoparticles, insulin and 0.3% APTES solutions. In Figure 3a we observe the large scale aggregation of the non-denatured insulin, gold nanoparticles and 0.3% APTES solution which was drop casted and air dried on a p-type silicon. This aggregation corresponds to the visual change in the solution color of the insulin and gold nanoparticle solution from red to purple upon the addition of APTES. In contrast to Figure 3a, we see in the denatured insulin trial (Figure 3b) a less densely packed morphology with minimal aggregation visible.
To quantify the extent of aggregation within the denatured and nondenatured insulin trials, the size and size distribution of the samples were measured via dynamic light scattering. In Figure 4 we observe the size and size distribution for the trials containing, gold nanoparticles and native insulin. Here we observe no appreciable difference between the two curves corresponding to both solutions maintaining a red hue as observed in the captured solution images of Figure 2a. The majority of the gold nanoparticles in the solution ranged from 11-20 nm. In Figure 4b we observe the size and size distributions for the denatured insulin and non-denatured insulin containing 0.3% APTES. The nondenatured trial within Figure 4b exhibited an increase in the size and size distribution of the gold nanoparticle aggregates in contrast with the denatured curve which was very similar with respect to size and size distribution of the controls. In Figure 4b we observe a size range from 18-122 nanometers for the non-denatured insulin trial. Within Figure 3, we observe micron scale aggregation, however within the dynamic light scattering technique only submicron aggregation is observed. We believe this difference is due to differences in sample measurement times. Immediately after mixing the gold nanoparticle, insulin and APTES solution, the trials were characterized using the  dynamic light scattering technique. However, in order to image the samples via scanning electron microscopy, the samples were air dried overnight. During that time, the samples can continue to aggregate until they are immobilize on the surface of the silicon as the water evaporates.
The gold nanoparticles utilized in this study were synthesized via a one pot hot injection method using trisodium citrate as the reducing agent. This method produced gold nanoparticles that are capped with negatively charged citrate molecules, creating a stabilized particle through repulsive forces. [24][25][26] Upon the addition of APTES to the gold nanoparticle solution, dipoles are formed and initiate the observed aggregation of the AuNps. 27 The dipoles are induced due to the positively charged amine group on APTES molecule; which is electrostatically attracted to the negatively charged citrate groups on the surface of the AuNps. 28,29 The newly formed dipoles are higher in energy and so to align with the laws of entropy, the AuNps aggregate to minimize effects of the dipole causing aggregates which then stimulates a visual color change.
In Figure 5. the size and size distribution of the non-denatured and denatured insulin as a function of the accelerated heat stress exposure time was quantified. The size majority of the non-denatured insulin samples exhibited a diameter of about 4.5 nm, which is indicative of hexamer formation. Within the denatured trials ranging from 1-20 minutes of heat stress, a systematic increase in the measured diameter size was observed. It is important to note that the insulin is aggregating and this is the cause of the larger size and size distribution as opposed to individual particle growth. In the 1-minute heat stress trial we observe a size majority of 102 nm for the aggregated insulin. As the heat stress is further applied, we observe 172 nm and 223 nm size diameter aggregates for the 5-minute and 10-minute accelerated heat stress trials ( Figure 5). Within the 20-minute heat stress trial we observed two size diameter maxima at 204 nm and 530 nm.
Without the addition of APTES, the gold nanoparticle solution is relatively mono disperse and maintains an optical absorbance of about 500 nm. These values were comparable to previously reported sizes and corresponding wavelengths. 30,31 The amine group present on APTES is electrostatically attracted to the negatively charged citrate surface of the gold nanoparticles. 32,33 This coulombic attraction causes a dipole in the gold nanoparticles, increasing the overall energy state of the particle. To decrease this energy state, the gold nanoparticles aggregate which induces a visual color change in the solution. Therefore, aggregation is induced as a result of dipoles that form on the gold nanoparticles due to the presence of APTES. 34 Maintaining the concentration of APTES, the extent to which these dipoles are formed are also dependent on the concentration of insulin present in solution and the state of the insulin (denatured or non-denatured).
Based on the mode of aggregation due to the presence of APTES, we propose the following mechanism for the visual detection method based on the observed insulin aggregation (Figure 4). Without insulin present in solution, the addition of 100 uL of 0.2% and 0.3% APTES will produce a blue hue within the gold nanoparticle solution. With the addition of 100 uL of the insulin solution in the presence of gold nanoparticles and APTES, a purple color is observed instead. However, as the protein begins to denature, the insulin agglomerates forming barriers that prevent the interaction of the APTES with the gold nanoparticles as a function of the extent of insulin denaturation ( Figure 6). We propose that as the insulin is heat stressed, it more effectively blocks the creation of the dipoles thus preventing the gold nanoparticles from closing the inter-particulate space and limiting the degree of aggregation. This causes a blueshift in the absorbance spectra of the solutions as the heat stress is applied to the protein, thus producing visual means of detection.

Conclusions
Through this work we demonstrate the ability to visually detect denatured insulin utilizing concentration dependent induced aggregation of gold nanoparticles. The need to visually detect denatured or non-native proteins are crucial as the adverse effects of non-native structures can cause sever patient discomfort and a decrease in drug potency. This phenomenon poses a continuous problem within the pharmaceutical industries and can have detrimental and long lasting effects on patient quality of life. Therefore, this work provides a fast, facile and visual detection method for the presence of denatured proteins. Continued proof of concepts with other biological analytes such as antibodies and antibody conjugates using this system may prove Figure 6. Displays the proposed mechanism for the blueshift observed in the ultraviolet visible absorbance spectra of the insulin, gold nanoparticle and APTES solutions as the insulin is denatured.
) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 207.241.231.82 Downloaded on 2018-07-20 to IP to be a highly effective visual detection method during all stages of a drug product lifetime including; procession, formulation, filling, packing, shipping and storage.