Fast and reliable identification of infectious disease agents is among the most important challenges for the healthcare system. The discrimination of individual components of mixed infections represents a particularly difficult task. In the current study we further expand the functionality of a ratiometric sensor array technology based on small-molecule environmentally-sensitive organic dyes, which can be successfully applied for the analysis of mixed bacterial samples. Using pattern recognition methods and data from pure bacterial species, we demonstrate that this approach can be used to quantify the composition of mixtures, as well as to predict their components with the accuracy of ∼80% without the need to acquire additional reference data. The described approach significantly expands the functionality of sensor arrays and provides important insights into data processing for the analysis of other complex samples.

No abstract

Timely and efficient identification of pathogenic microorganisms is of high importance for public health and safety. Current bacterial pathogen analysis relies on phenotypic identification of the bacterial species by Gram staining and culturing in combination with other biochemical methods and PCR. These methods, however, are sometimes inconclusive, time-consuming and equipment-demanding, which delays treatment and limits their application to laboratory settings. The development and use of molecular sensing techniques has a potential for overcoming the limitations of the traditional phenotypic approaches. Most of currently developed sensing methods are based on optical and/or electrochemical techniques for analysis of biomarkers or whole bacteria [1], where selectivity is usually achieved by using antibodies, their fragments, natural or engineered peptides, or aptamers. Important limitations for using these elements as a part of the sensor are cost, stability and adaptability: they often require special handling and are specific to only a narrow range of analytes. Multivariative analysis using sensor arrays has been actively developed for various applications, including detection and analysis of small molecules and ions, proteins and even whole microorganisms [2]. Ratiometric dyes have been long explored as sources of a self-referenced analytical signal that is independent from the concentration of the probe. However, relatively few sensor arrays have been explored leveraging this principle. Derivatives of 3-hydroxyflavone have been widely explored as powerful and sensitive tools to study different parameters of their microenvironment [3]. Owing to the excited-state intramolecular proton transfer (ESIPT) taking place in their molecules, they exhibit two distinct emission bands in their fluorescence spectra. In addition to that, these molecules can participate in other protolytic equilibria. For example, in basic media or under the influence of strong external hydrogen bond acceptors (solvent or other molecules), 3-hydroxyflavones form anions whose fluorescence is located between the emission bands of the normal and tautomer forms of the neutral molecule undergoing ESIPT. Leveraging both ground- and excited-state equilibria, two independent ratiometric signals can be obtained from a single probe dye. Sensor array components were designed by introducing various substituents into the 3-hydroxyflavone core, thus leading to differences in the specific interactions of the dyes with their microenvironment, and potentially in their localization in bacteria [4]. The response patterns were investigated upon interaction with eight different bacterial species equally representing Gram positive and Gram negative strains. Linear discriminant analysis is usually used for data processing, however, other methods such as support vector machines can also be applied. Analysis of the canonical score plots using the first two discriminants results in excellent separation of all groups of sensor responses originating from different bacterial strains. An interesting detail can be further observed: signals from the Gram-positive bacteria tend to group on the left side of the canonical score plot, whereas those coming from the Gram-negative bacterial species cluster on the right side. This evidences for some distinctions in the spectral features of the dyes fluorescence upon interaction with bacteria of different Gram status. The discriminant analysis of these data revealed the ability of the sensor array to correctly identify bacterial species and perfectly recognize the Gram status of the analyzed bacteria. This illustrates the potential versatility of applications of the proposed sensor. From the general perspective, being “trained” to recognize the responses of the eight bacterial species, the system is intrinsically incapable of recognizing other strains beyond this list. However, this limitation is not necessarily valid for the broader application of the obtained dataset, which can be exemplified by Gram status recognition and “one against the rest” analysis. In these cases, samples representing bacterial species excluded from the training dataset will still belong to one of the “known” classes. To test this hypothesis with Gram status determination, a truncated training dataset was used. The results show that the system was able to correctly recognize the Gram status in 100% of the cases, even when the samples represented bacterial species excluded from the training dataset. A similar test for “one against the rest” discrimination showed correct attribution in 89% of the cases. We further expand the functionality of a ratiometric sensor array by demonstrating its successful application for the analysis of mixed bacterial samples. Using pattern recognition methods and data from pure bacterial species, we demonstrate that this approach can be used to quantify the composition of mixtures, as well as to predict their components with the accuracy of approximately 80% without the need to acquire additional reference data. Thus, the proposed sensor array not only can be used for classification and multiparametric analysis of the samples, but is also capable of characterizing the unknown samples beyond the species used in training datasets. The sensor’s functionality can be expended to allow analysis and quantification of mixed samples, and even prediction of the mixture components. The described approach also provides important insights into data processing for the analysis of other complex samples. References [1] J. Kubicek-Sutherland, D. Vu, et al. Detection of Lipid and Amphiphilic Biomarkers for Disease Diagnostics, Biosensors 7, 25 (2017). [2] J. R. Carey, K. S. Suslick, et al. Rapid Identification of Bacteria with a Disposable Colorimetric Sensing Array, J. Am. Chem. Soc. 133, 7571-7576 (2011). [3] G. Duportail, A. Klymchenko, et al. Neutral Fluorescence Probe with Strong Ratiometric Response to Surface Charge of Phospholipid Membranes, FEBS Lett. 508, 196-200 (2001). [4] D. Svechkarev, M. R. Sadykov, et al. Ratiometric fluorescent sensor array as a versatile tool for bacterial pathogen identification and analysis, ACS Sensors 3, 700-708 (2018). Figure 1

No abstract