Fluorescence-based bioaerosol sensing project advances to third phase

Printer-friendly Version

Home > View Article

A flurry of recent activity to make fast-response bioaerosol detectors has been spawned by the need to protect against bioterrorism threats; however, available devices often false alarm on non-biological particles. Sandia researchers (Tom Kulp, Scott Bisson, Tom Reichards, Bob Crocker, and Scott Ferko) have been involved in a program to characterize one such class of sensors—those based on single-particle laser-induced fluorescence (LIF)—and to augment their operation to reduce false alarms. This program, called the Enhanced Bioaerosol Detection System (EBADS), is funded by the Department of Homeland Security (DHS). It is a multilab collaboration between Sandia, Oak Ridge, Lawrence Livermore, Pacific Northwest, and Los Alamos National Laboratories (SNL, ORNL, LLNL, PNNL, and LANL). Within this group, SNL and ORNL scientists have been collaborating to develop an improvement on LIF-based detection that uses fluorogenic reactions (those producing a fluorescent product upon reaction). Recently, the SNL-ORNL team passed a blind-sample testing exercise conducted by DHS and the Edgewood Chemical and Biological Center (ECBC), which qualifies their approach to proceed to a third phase of development. In the first phase of EBADS, false alarms generated by two commercial LIF-based sensors were correlated with the readings of single-particle analytical measurements to find out which classes of materials produce "biological-like" LIF spectra. That work occurred at the San Francisco International Airport (SFO), where the ambient particles in the air-handling system were monitored using single particle mass spectrometry and selective collection and analysis of particles causing alarms in the LIF systems. The results showed that most of the SFO particles that triggered LIF sensors were not innocuous bacteria or spores but were non-biological in composition.

Figure 1. Comparison of the EBADS hydrosol measurement (right) with that of a typical LIF-based aerosol sensor. The EBADS measurement is made in a liquid, rather than an air stream, and it adds a fluorimetric stain to the particles to label protein molecules within them.

After the tests, the SNL-ORNL team proposed and demonstrated an improved method to better select between biological and non-biological particles. To maintain low operating costs, they turned to fluorimetric chemical reactions to allow better recognition of biomolecules within aerosols. Specifically, the protein-sensitive dye fluorescamine was used because it creates a fluorescent product when it reacts with primary amine groups on proteins. They found it is best to measure the response of individual aerosols to this dye when the aerosols are entrained in the fluidic stream of a flow cytometer (see Figure 1), because this allows simultaneous measurement of the laser scattering by each particle, in addition to its LIF; and this method provides the needed solvent requirement for the dye to react. Scattering measurement occurs in two directions, forward (sensitive to particle size) and sideways (sensitive to particle shape or structure). When in the fluid, the particles are referred to as hydrosols.

LIF/scattering data from each hydrosol contributes to a three dimensional distribution (see Figure 2) depicting the response of the particle population to each of these three measurements. Populations of similar particles (e.g., spores) produce characteristic patterns in these plots that allow their discrimination from background mixtures. Because they are composed of a broad range of particle sizes and composition, the latter produce broad distributions of points. The response of samples to fluorescamine further distinguishes them from backgrounds because their uniform "stainability" causes their LIF to increase dramatically on staining, relative to the generally non-biological background materials (see Figure 2).

Figure 2. Plots of the pre-stained (left) and stained (right) data collected from bacillus subtilis spores (red) and house dust background (blue). Each dot represents the meaurement from a single particle. Note that the spores are affected greatly (the LIF scale is logarithmic) by staining, whereas the house dust is not.

A challenge to this process is the fact that some threat agents (i.e., viruses and toxic proteins) are too small to trigger the described particle measurement. Although they may have been imbedded in respirable-sized salt (or other co-precipitate) aerosols when collected, they likely will dissolve in their native form in the hydrosol state.

This problem has been addressed by measuring the stainability of the fluid as a measure of the presence of viruses and protein toxin surrogates. In this case, it is important to ensure that a triggerable sample particle is not in the fluid volume being probed, because that would bias the measurement. To ensure this, the sample was seeded with non-fluorescent, uniformly-sized silica beads. The fluorescence of those beads was measured by the cytometer in the stained and unstained state. Because they are both non-fluorescent and non-stainable, any LIF in those measurements is assumed to arise from the fluid around the beads. The concentration of large particles in the fluid is low enough to ensure none is in the volume with the seeded bead.

These methods were applied to the measurement of twenty unknown mixtures of cells, spores, viruses, and/or proteins in various background particles. Analysis of the particle "clouds" was enabled by decomposing observed data distributions into linear combinations of pre-measured distributions for various threat and background materials using three dimensional linear analysis. In all cases, the staining approach correctly identified the presence or absence of the threat agent simulants. The project is now proceeding to a phase in which it will address "real-world" samples collected at a remote site, such as SFO. This study will be accomplished by bringing samples from SFO to SNL and, ultimately, by deploying an instrument on site at SFO.

Article taken from the CRF News Volume 29 Number 2 (PDF - 1323K)


	© 2006 Sandia Corporation \| Questions and Comments \| Site Index \| Privacy and Security P.O. Box 969 * Livermore, California 94551