Sponsored by HORIBAJun 12 2018
Fluorescence spectra are often subject to significant distortions. These are largely predictable and are caused by concentration-dependent Inner Filter Effects (IFE). It is therefore the responsibility of the user to evaluate returned absorbance values in relation to the concentration of the sample. This then gives the Beer-Lambert linearity requirement.
The Primary Inner Filter effect (PIF) can be seen when excitation light is absorbed by components in the sample. This happens before it reaches the area where the excitation and emission beam paths intersect, and it is at this point that fluorescence can be excited and therefore detected.
The Secondary Inner Filter effect (SIF) happens when the fluorescence emitted in the aforementioned intersected regions of the excitation and emission beam paths, is then absorbed by sample components before it is able to leave the cell and reach the emission detector.
Both PIF and SIF can be caused by sample components that absorb light, and this can occur regardless of whether or not these components fluoresce.
The effects are constantly present at any OD value in the overlapping absorbance and fluorescence regions.
There are a whole range of studies around IFE corrections and all of these relate to formulating, then testing equations which deal with the cell path lengths, cell positions, optical coefficients and beam geometry; the latter being defined by the focal properties (f-numbers) of excitation and emission beam paths.
It is common practice to measure the absorbance spectrum of the sample in question at the point where this overlaps the fluorescence excitation and/or emission spectral regions of interest. In an ideal situation this would be done at the same time as fluorescence measurements within the same cell, using the same instrument to help minimize differences due to the state or geometry of the sample.
A new method for simultaneously generating the individual excitation and emission spectra is presented here, and this can be used for all fluorescent sample components whilst providing useful information on components that have absorbed light but did not fluoresce. It is also able to provide the information needed in order to correct the fluorescence spectra where sample concentration-dependent IFE may be an issue.
Instrumentation
Within this study, the A-TEEM™ method is explored. This method offers simultaneous absorbance- transmission and fluorescence excitation emission matrix spectroscopy using the dedicated and patented Aqualog® from Horiba Instruments Inc.
Aqualog® is an optical bench that features a special aberration-corrected, double-grating excitation monochromator. It also includes an absorbance detector, a reference detector (both of which are Si photodiodes) and a specialized emission detector made up of a spectrograph and a thermoelectrically cooled, back-illuminated CCD.
This system can incrementally scan excitation at a variety of energy levels, ranging from high to low energy as required. It can then collect and collate the complete emission spectrum at each excitation increment, in order to quickly and simultaneously generate absorbance spectra as well as fully corrected excitation-emission matrices or spectral maps.
The Aqualog® includes bundled software which features a built in tool for normalization to water Raman scattering, or the use of quinine sulfate for the defined emission conditions. Additionally, correction for the influence of Rayleigh masking or IFE is included.
This Aqualog® Datastream package allows for the analysis of fully corrected EEM data and is based on the multivariate routine commonly known as PARAFAC.
Current Applications
Water Quality
Where drinking water treatment plants mainly use surface water, there is a need for regulation. This is because certain components within the water can be precursors to toxic Disinfection By-Products (DBPs) that can react with halogenated disinfectants over time, within the distribution system.
Total Dissolved Organic Carbon (TOC) is a common component that can be subject to considerable variations in unpredictable patterns. These patterns are associated with events that influence sporadic drainage of organic materials into the sourcewater, for example, snow-melt or rainfall.
As can be seen here, the Aqualog® has been utilized to monitor dissolved organic matter related to EPA Stage 2 Disinfection By-Product Rule (DBPR2) compliance in a highly typical water treatment plant that treats drinking water sourced from surface water.
Identical samples of raw sources and finished water have been filtered to 0.45 μm before analysis. Additionally, these samples were matched to room temperature of 25 °C before the analysis took place.
In this example, the Aqualog® EEM and absorbance measuring conditions included an excitation/absorbance range between 250 and 600 nm, using 3 nm for excitation intervals and 3.28 nm for emission. It also used a medium gain and 2 s integration for emission detection.
The blank sample, used for emission and absorbance, was a sealed and TOC-free water sample - Starna 3Q-10 - from Starna Scientific Inc. All of the 3.5 ml samples were analyzed using 1 cm path length suprasil 4-way clear fluorescence quartz cuvettes.
The example draws comparisons between common EEM and absorbance profiles for raw and finished water samples that have been measured in the same daily sample set.
EEM data for the raw water showed an approximately 6.25 fold higher peak intensity than the finished water, as well as broader and considerably red-shifted primary emission band.
The raw water’s absorbance also showed higher extinction at every wavelength when contrasted against its finished water counterpart.
In order to properly evaluate quantitative changes in EEMS associated with the treatment, PARAFAC analysis has been applied to every sample in order to decompose the excitation spectra, concentration loadings and emission spectra for the core fluorescent components.
Three key components have been identified within this example:-
- Component 1 (C1) is identified as a humic/fulvic component with comparatively lower molecular weight and aromaticity.
- Component 2 (C2) is identified as a humic/fulvic component with a higher molecular weight and aromaticity.
- Component 3 (C3) is identified as a protein-like component.
The image above contrasts the normalized PARAFAC concentration loadings for the three components. This is done for each of the daily raw and finished water samples. As can be seen, Component 2 was consistently more prominent in the raw water; while in the finished water, the most prominent component was Component 1.
The relative concentration of Component 3 was largely the same between the treatments, while the improved removal of Component 2 was in line with the expected effects of coagulation to remove organics with a higher molecular weight more effectively than lower weighted species.
As can be seen, the pattern showed an obvious correlation with the broader, red-shifted spectra for the EEM bands in the raw water when compared to the finished water shown in an earlier image.
Wine Analysis
Common red wines contain a number of colored and fluorescent components. These are primarily polyphenolic in nature and are central to the determination of several quality parameters like flavor and color.
Fluorescence and UV-vis spectroscopy offer the potential to detect and resolve these components, meaning that these can be effectively characterized to reveal the unique compositional properties of red wines.
Absorbance, transmission and fluorescence EEM data can be combined in order to evaluate lot-to-lot, regional and varietal characteristics. This combination can also be used to sense the effects of aging as well as oxidation and sulfite treatment.
The image above shows EEMs that have been recorded simultaneously (A) and (B), and absorbance and percent transmission spectra (C) and (D) for a typical Italian red wine from a freshly opened bottle. It also highlights before (A) and (C) and after a week-long exposure to air (B and D).
The Aqualog® EEM and absorbance measuring conditions have included an excitation/absorbance range that extends from 200 to 800 nm with a 3 nm increment. They also include an emission range of 250 to 800 nm with a 4.65 CCD bin increment at medium gain and 0.1 second integration time.
These complicated EEMs highlight the major contours in the UV excitation-emission range with the major excitation/emission peak occurring at around 275/309 nm.
Samples taken both before and after show that both absorbance and transmission spectra exhibited a major extinction peak at around 275 nm, a smaller shoulder peak around 320 nm and a second, minor peak at around 520 nm.
This 275 nm peak is at least partly associated with phenolic compounds while the 520 nm peak region is often associated with anthocyanin compounds. When compared to the before (C) and after (D) samples, these showed increased extinction across the whole absorbance spectrum. This is generally associated with the phenomenon of oxidation.
EEM spectral contours are incredibly complex, and these are made up of multiple overlapping excitation and emission components. Their complexity can severely limit quantitative and qualitative visual interpretation of major contour elements.
Because of this, multivariate analysis is often applied in order to decompose the EEM components within the contexts of quality and quantity.
This study evaluated the PARAFAC model, and in this instance, this was constrained to yield non-negative values for all score and loading parameters. The table below highlights the five spectral loading components that were resolved in the PARAFAC model developed using all the replicated fresh and oxidized samples of wine. For additional tentative identification, these components were compared to existing literature.
The image below shows a comparison between the effects of oxidation of each of the PARAFAC components for the wine sample. Here, Component 1 was the most prominent component both before and after oxidation, with the deeper UV emitting components (1, 2 and 3) increasing comparatively more than the longer emission wavelength components (4 and 5).
Insulin Structure and Analysis
As a protein hormone, insulin is produced by the pancreas and is essential for basic metabolic processes. Commercial insulin therapeutics will usually fall into two primary categories, these being fast acting and long acting insulin.
Short and long acting insulin are both very similar, with the difference between them being as small between one and three residues in the protein sequence. These tiny sequence changes alongside controlled pH of storage and delivery are used to either trigger or prevent insulin dimers and hexamers forming in the blood stream.
As these aggregates are formed they allow the body to absorb insulin more slowly. In contrast, the absence of these aggregates causes the body to absorb insulin faster. The changes in protein stability and structure like those central to the pharmacokinetics of insulin are generally measured using fluorescence emission spectra, UV–vis absorbance spectra or occasionally even both, using inherently fluorescent amino acids.
Commercial insulin formations are highly concentrated (4 mg ml−1), meaning that the IFE correction is crucial for the accurate measurement of the A-TEEM™ fingerprint. This is shown in the image below.
Here, three different insulin proteins have been analyzed at varying concentrations. These were at pH levels of 4.5 to 8.5 and at temperatures of 5 °C to 37 °C. The insulin types were:
- Recombinant human insulin [10]
- Insulin-aspart (commercially available, fast acting)
- Insulin-glargine (commercially available, long-acting)
The use of PARAFAC analysis for measuring intrinsic fluorescence A-TEEMs™ allowed for the characterization of the aggregate state of each type of insulin. Four components were identified in the commercial formations:-
- Components 1, 2 and 3 (Comp1, Comp2 and Comp3) identified as Tyrosine [11]
- Component 4 (Comp4) identified as m-cresol [12] - a preservative in insulin formulations
The red data points in the image above illustrate the scores from a four component PARAFAC model as fit to recombinant human insulin solutions at pH 4.5 (the first five data points), pH 7.4 (the second five data points) and pH 8.5 (the last five data points).
Each of the five repeats were at temperatures 5 °C, 20 °C, 25 °C, 30 °C, and 37 °C in that order. The models were repeated and plotted for insulin glargine (green data) and insulin aspart (blue data) under identical sequential conditions.
As the example shows, insulin glargine – which is generally sequenced to produce aggregates at biological pH and temperature – exhibits a far larger score for Component 1 (stars). This could be attributed to the blue-shifted spectrum of tyrosine when insulin is in an aggregate form.
In contrast, solutions of human insulin and insulin aspart (which are less likely to aggregate under the same conditions) demonstrate far lower scores for the same component.
Given the results presented, it is clear that insulin sequences that vary by just 1 to 3 amino acids can be differentiated using intrinsic fluorescence A-TEEMs™.
Because commercial insulin formations have high concentrations and absorbance values, the A-TEEM™ method is an ideal solution for the characterization of such commercial protein therapeutic formulations, especially where more traditional fluorescence methods may fail to achieve the desired results.
It is important to note that this study does not provide new information on insulin or how these specific proteins do or do not aggregate according to the formulation condition. It does however, provide clear insight into how the A-TEEM™ method can be utilized for the characterization of protein therapeutic formations for aggregation behavior, even in conditions where there are only miniscule differences in protein sequences.
EEM has already proven itself to be a highly relevant method within the field of biology. Additionally, the A-TEEM™ fingerprinting technique could theoretically be used for other protein therapeutics such as vaccines, inhibitors, enzymes and antibodies.
Conclusion
Fluorescence is already highly regarded as an incredibly sensitive tool for the characterization of a multitude of sample types. However, when combined with other measurement techniques, fluorescence’s potential becomes far greater.
The A-TEEM™ method that is implemented within the patented Aqualog® is the only instrument of its kind that is able to perform real-time, simultaneous inner filter correction of fluorescence spectra in order to provide true, NIST-traceable molecular fingerprints.
This level of accuracy and quality is of vital importance where there is a need for absolute and reproducible libraries of molecular fingerprints, which must be relied upon for credible quantitative analysis of a wide and varied selection of essential substances.
As this piece has explored, the A-TEEM™ method – used in combination with the Aqualog® spectrofluorometer - has been used to monitor raw water treatment, affording its users a quick, accurate and easy to use tool.
This application has been extended into other areas such as wine study, where the technique has offered the opportunity to characterize the main compounds of the wine as well as the level of oxidation present.
This multimodal methodology has huge scope for use in a myriad of fields. It has the potential for practical application in the food industry, the biomedical sector and the petrochemicals industry. Additionally, the A-TEEM™ has been shown to be an enormously powerful method for the discrimination of the different types of existing insulins.
This information has been sourced, reviewed and adapted from materials provided by HORIBA.
For more information on this source, please visit HORIBA.