BACKGROUND OF THE INVENTION

[0002] The invention relates to the detection, enumeration, and identification of replicating cells, especially microbial cells (e.g., bacteria, yeasts, and molds), in medical, industrial, and environmental samples. Microbial culture is the predominant methodology in these markets, because of its many attractive features. The invention addresses the chief drawback of microbial culture—the length of time needed to achieve results—while retaining the beneficial attributes of the method.

[0003] Microbial Culture for Detecting and Enumerating Microbes

[0004] During the 19th and 20th centuries an understanding emerged concerning the role of bacteria, yeast, and molds in causing infectious diseases and determining the quality of foods and beverages. Early on, a powerful method, microbial culture, was developed for detecting small numbers of microbes. Microbial culture allows simple visual detection of microbes by exploiting their propensity to reproduce in large numbers rapidly. For example, a single bacterial cell, which is much too small to see by eye (about one millionth of a meter), when placed in nutrient broth, can cause the broth to become visibly cloudy in less than 24 hours.

[0005] A related microbial culture technique, called microbial enumeration or colony counting, quantifies the number of microbial cells in a sample. The microbial enumeration method, which is based on in situ microbial replication, generally yields one visually detectable “colony” for each microbial cell in the sample. Thus, counting the visible colonies allows microbiologists to determine the number of microbial cells in a sample accurately. To perform microbial enumeration, bacterial cells can be dispersed on the surface of nutrient agar in petri dishes (“agar plates”) and incubated under conditions that permit in situ bacterial replication. The individual, visually undetectable, microbe replicates repeatedly to create a large number of identical daughter microbes at the physical site where the progenitor microbial cell was deposited. The daughter cells remain co-localized (essentially contiguous) with the original cell, so that the cohort of daughter cells (which may grow to tens or hundreds of millions of cells) eventually form a visible colony on the plate.

[0006] Electronic methods have been developed for enumerating microbial colonies. Most such methods automate colony counting but do not substantially increase the sensitivity or decrease the time to results compared to traditional enumeration by eye. Colony counters use a variety of optical methods for detecting colonies including detection of intrinsic optical properties of microcolonies (e.g., U.S. Pat. No: 3,493,772; U.S. Pat. No: 3,811,036; U.S. Pat. No: 5,290,701; Arkin, A. P., et al. (1990); Biotechnology (N Y) 8: 746-9) and color changes of pH indicator molecules in the matrix surrounding the colonies (U.S. Pat. No. 5,510,246). Methods that use stains or probes to label the colonies have also been developed and will be discussed below.

[0007] Microbial culture is a remarkably successful method, as evidenced by the fact that even after more than a century, the method still dominates medical microbiology and quality control testing in industrial microbiology (e.g., pharmaceutical, food, and beverage manufacturing). The method is inexpensive, relatively simple, and ultra-sensitive. The sensitivity of microbial culture can be seen in the common test for foodborne pathogens in ground beef. A single microscopic bacterial pathogen cell can be detected in 25 grams of ground beef using microbial culture. Another advantage of microbial culture is its ability to detect a large range of microbes of medical and industrial significance.

[0008] An advantage of in situ bacterial replication is the ability to generate a pure, or clonal, population of cells (called pure cultures, clones, or colonies). A pure culture is a large collection of identical living cells which descend from the same progenitor cell. Pure cultures are required for methods that identify microbes and for determining antibiotic resistance. Medical microbiology relies heavily on pure cultures, since bacterial pathogens are frequently isolated from non-sterile clinical samples (e.g., feces or wounds) along with non-pathogenic bacteria that are likely to be even more numerous than the pathogenic cell. Isolating pure microbial cultures is also important in industrial microbiology. For example, pharmaceutical and cosmetics manufacturers must test their products for the presence of microbial contaminants. Pure cultures of the contaminating microbes are used for microbial identification, which determines whether a production batch must be discarded and aids in investigating the source of the contamination in the industrial process. 1

[0009] The ability to culture microbes selectively is an essential tool for microbial identification and for determining resistance and susceptibility to antimicrobial agents such as antibiotics. Selective culture exploits the fact that different microbes require different growth conditions. These differences arise from the fact that strains of microbes differ in their biochemical makeup because of inherent genetic differences. For example, one type of microbe might be able to grow on nutrient medium containing the sugar sorbitol as the sole source of carbon atoms to fuel its growth, while another type of microbe cannot. Selective growth is important in the food industry. For example, a food sample can be scanned for a particular food pathogen, Salmonella, by plating the sample on media that allows Salmonella to grow but not other food microbes.

[0010] Similarly, selective culture is used to determine which antibiotic is most effective for killing a bacterial strain isolated from the spinal fluid of a child with bacterial meningitis. A pure bacterial culture (derived from a clonal colony from a nutrient agar plate) is used to inoculate growth medium containing various antibiotics at various concentrations. The optimal antibiotic therapy is determined by monitoring the ability of the microbe to grow in the presence of the various antibiotics. Determining antibiotic resistance and susceptibility by selective growth on the surface of solid nutrient agar medium is another common approach. For example, in the Kirby-Bauer method, small filter disks impregnated with different antibiotics are placed on the surface of nutrient agar plates coated with a pure culture of bacteria from a clinical sample. A gradient of antibiotic diffuses radially outward from the filter. Bacteria that are resistant to high levels of the antibiotic grow up to the edge of the filter. However, bacteria that are very sensitive to the antibiotic can not grow unless they are far from the edge of the filter. After incubating the plates (usually for one or two days) a microbiologist determines the level of resistance to an antibiotic by measuring the thickness of the growth-free ring or zone around the filter. A related method, the “E” test (Hardy diagnostics), uses a rectangular strip that is impregnated with a gradient of antibiotic. The level of bacterial resistance is determined by measuring the point on the strip with the highest antibiotic concentration next to which the bacteria continue to replicate.

[0011] The most serious drawback of microbial culture is that it is slow—it takes time to generate the number of cells required for visual detection. The long growth period required for microbial culture is a significant problem in both healthcare and industry. For example, because it requires days to culture and identify the microbe causing a patient's blood infection, a patient with a fungal blood infection could die before anti-fungal therapy is even begun. Some infectious agents, such as the bacterium that causes tuberculosis, generally require weeks to grow in culture. The long time required for detecting M. tuberculosis can result in a patient with tuberculosis infecting many others with the highly contagious disease or the costly quarantine of patients who do not have tuberculosis.

[0012] In food manufacture, long testing cycles can increase food spoilage or result in moving inadequately tested material through subsequent processing steps. Slow microbial culture also adversely impacts the production of biopharmaceuticals and vaccines. In these applications, the manufacturing process often requires pooling of batches. Because of long microbial culture testing cycles and the need to move material through the manufacturing process, contaminated batches are sometimes not detected until after a batch pooling step. If it is subsequently found that a contaminated batch was combined with uncontaminated batches, the whole pool of combined batches must be discarded.

[0013] Other disadvantages of microbial culture, such as tedious manual procedures and inability to culture some microbes, are considered less problematic than the long time required. For example, manual methods for microbial enumeration predominate, even though instruments for automated plating and analysis have been introduced. Most types of microbes found in the environment cannot be grown in the laboratory. However, such microbes are often not harmful to humans or are destroyed in industrial manufacturing processes and are therefore ignored for most applications. However, several important exceptions of critical medical importance include hard or impossible to culture bacteria such as Chlamydia, strains of which can cause sexually transmitted disease and pneumonia. Fortunately, alternative culture-independent methods are available in these cases (see below).

[0014] Rapid Microbial Culture Enumeration Methods

[0015] A number of microbial culture methods for more rapid microbial enumeration have been developed. One rapid microbial culture method deposits bacterial cells on microscope slides coated with nutrient medium. Using microscopic examination, microbial growth can be detected much earlier than with the naked eye, since microscopes can detect microcolonies resulting from a small number of cell divisions. However, this method is not effective for testing large samples containing low numbers of microbial cells, because only a very small volume of sample can be observed in a microscopic field of view. The low sensitivity of microscopic methods generally limits their usefulness to samples containing more than ten thousand bacterial cells per milliliter—these methods are much less sensitive than traditional microbial culture.

[0016] The advent of electronic imaging systems has led to the development of numerous automatic “colony counters.” Although, most of these counters are designed to aid the user by automating the colony counting process and do not decrease the time to result, some systems have demonstrated the ability to detect colonies before they are large enough to be seen easily by eye. For example, the Colifast Rapid Microcolony Counter (Colifast) can detect small fluorescently labeled colonies of coliform bacteria hours before they can be seen by eye. The Colifast system achieves enhanced detection by using a fluorogenic compound (a substance that is not fluorescent until metabolized by coliform bacteria) included in the nutrient agar media.

[0017] A system for rapid enumeration of microbial colonies using bioluminescent labeling has recently been commercialized. The MicroStar system (Millipore) uses the cellular ATP in microcolonies to generate light via the action of applied luciferase enzyme and substrates. The method reduces time to detection substantially. The MicroStar imaging system has also been used in conjunction with labeled probes to identify specific bacteria (Stender, H., et al. J Microbiol Methods 46: 69-75 (2001)). A drawback of the system is that the detection method kills the microbes, precluding isolation of pure cultures from the colonies. The system also requires an expensive image intensifier module.

[0018] An instant film-based method for detecting microcolonies containing specific bacteria has been developed by Boston Probes (Perry-O-Keefe, H., et al. Journal of Applied Microbiology 90: 180-9 (2001)). Microbial microcolonies on membranes are labeled using microbe-specific PNA probes tagged with an enzyme capable of generating a chemiluminescent signal. The membranes are then placed on X-ray or instant-film for imaging. The method is limited to scanning for a particular microbe in one experiment. A similar method uses fluorescently labeled PNA probes and an array scanner (Stender, H., et al. Journal of Microbiological Methods 45: 31-9 (2001)). These approaches require substantially more expertise than traditional culture methods.

[0019] Rapid Microbial Enumeration without Microbial Culture

[0020] The fastest methods for microbial enumeration forgo microbial culture. Medical and industrial microbiologists are generally interested only in enumerating viable microbes—only living microbes are capable of replicating during microbial culture. Therefore, to be most effective, methods that detect individual cells without reliance on cellular replication must distinguish living from dead microbes by using physiological surrogates for cellular replication (e.g., Nebe-von-Caron, G., et al., J Microbiol Methods 42: 97-114., 2000; Mignon-Godefroy, K., et al., Cytometry 27: 336-44, 1997). Cells are stained with dyes that measure a biochemical property that is generally correlated with the ability to replicate (e.g., esterase activity or biochemical respiration). Validating and instituting surrogate methods have been problematic since samples that are known to meet regulatory standards and that are scored as sterile using traditional plate culturing methods often have thousands of cells that score positive for the surrogate biochemical activity.

[0021] An example of a system that directly detects viable cells is the ScanRDI system (Chemunex). ScanRDI enumerates microbial cells that are stained with a fluorogenic esterase substrate using laser scanning technology (U.S. Pat. No: 5,663,057; Mignon-Godefroy, K., et al., Cytometry 27: 336-44, 1997). A laser-scanning system (including an optical collection system using photomultiplier tubes (PMTs)) captures an image of the filter and can detect individual labeled cells. The system illuminates and queries a microscopic area (generally 4-14 μm) but scans the beam progressively so as to cover a macroscopic area (e.g., a 25 mm diameter circle). The system is designed to detect cells with intact membranes and active esterase enzyme. There is a correlation between the numbers of such cells and the number of cells that can form colonies on growth medium. However, this approach often results in substantial “overcounting”—i.e., higher numbers of cells than are detected by traditional culture (Costanzo, S., et al. (2002). PDA Journal of Pharmaceutical Science and Technology 56: 206-219). Another disadvantage of the ScanRDI system is that it kills the microbes during the staining process precluding generation of pure cultures from the detected microbes. Finally, laser scanning systems for cellular enumeration are complex and expensive (hundreds of thousands of dollars) making them difficult to justify for routine microbiological applications. Other laser scanning systems have also been commercialized (Miraglia, S., et al., J Biomol Screen 4: 193-204,1999; Tibbe, A. G., et al., Nat Biotechnol 17: 1210-3,1999; Kamentsky, L., 2001, Laser Scanning Cytometry . In Cytometry, Z. Darzynkiewicz, H. Crissman and J. Robinsnon, eds. Methods in Cell Biology Vol. 63, Part A, 3rd ed, Series Eds. L. Wilson and P. Matsudaira. (San Diego: Academic Press)).

[0022] Flow cytometry is another powerful method that can rapidly enumerate microbes without relying on cellular replication (Alvarez-Barrientos, A., et al., Clin Microbiol Rev 13: 167-195, 2000). Individual organisms or particles are forced to flow through a narrow channel, one at a time, past a laser beam. Besides enumeration, information about size/shape and composition is gathered by analyzing the fluorescence emission and light scattering caused by the organisms. Thousands of individual cells or particles can be analyzed per minute. Pathogens can by identified using flow cytometry by binding fluorescently labeled species-specific antibodies or nucleic acid probes to fixed organisms (Alvarez-Barrientos, 2000, supra).

[0023] Pathogens can by identified using flow cytometry by binding fluorescently labeled species-specific antibodies or nucleic acid probes to fixed organisms (Alvarez-Barrientos, 2000, supra). Individual cells of one particular type are usually the targets. Flow cytometric methods have been used more extensively for quantitatively detecting particular cell types on the basis of the ability to bind labeled probes, usually either antibodies or nucleic acids. For example, flow cytometry is used to quantify the population sizes of classes of lymphocytes in patients with AIDS. Flow cytometry is a more complex and expensive method than traditional culture. Although faster than traditional culture, flow cytometry does not have a comparable limit of detection to the traditional method. Traditional microbial culture can detect one bacterial cell in 0.1 liter of water, while flow cytometry is most effective when there at levels that are many thousands of times higher than that. Furthermore, microbial targets are often killed by the staining methods used for detection, eliminating the ability to produce pure cultures.

[0024] Using microscopic imaging to visualize and enumerate microorganisms directly can be rapid and relatively simple to perform (Amann, R. I., et al., Microbiological Reviews 59: 143-69, 1995). Direct fluorescent assays (DFA) in which a fluorescently labeled antibody reacts with a fixed sample is a common method in clinical diagnostics laboratories. For example, specimens suspected of containing bacterial agents are routinely stained with Gram stain. Similarly, to test for M. tuberculosis , samples are subjected to acid fast staining. The drawback of this technique is that it is many thousands of times less sensitive than microbial culture. The low sensitivity is due to the small fields visualized at high magnification. Only at high target cell concentrations are small fields likely to contain a target cell. Thus, for example, reliable identification of bacterial pathogens in sputum using fluorescent in situ hybridization requires titers of about 4×10 ⁵ cells/ml or more. Clinical samples obtained in common medically significant infections may contain fewer than 100 cells/ml—a concentration that is not nearly high enough to expect to find a cell in a high power microscopic field.

[0025] A system that does have the sensitivity to detect single bacterial cells using large area non-magnified imaging has been developed by researchers at Hamamatsu Corporation (Masuko, M., et al., FEMS Microbiol Lett 67: 231-8,1991; Masuko, M., et al., FEMS Microbiol Lett 65: 287-90,1991; Yasui, T., et al., Appl Environ Microbiol 63: 4528-33,1997). Large area imaging of individual microscopic target cells is accomplished using an ultrasensitive photon-counting CCD camera coupled to a fiber optic system, image intensifier, and Image-Processor. A disadvantage of this system is the great expense incurred because of the incorporation of the image intensifier and associated optics. Furthermore, unlike microbial culture methods, the system can not detect any microbe, distinguish between living and dead microbes, or generate pure cultures.

[0026] Rapid Microbial Enumeration by Quantifying Molecular Constituents of Cells

[0027] Numerous methods for detecting and identifying microbes based on their molecular constituents have been developed in the last half-century. Although some of these methods are substantially faster than microbial culture, none offers all of the features of culture that are critical to microbiologists. For example, although numerous immunoassays for microbes have been commercialized, this technique is not inherently quantitative, is much less sensitive than microbial culture, and is not as powerful as culture for detecting many types of microbes in a single test. Or, as another example, nucleic acid amplification methods can be as sensitive as microbial culture, but they do not distinguish between living and non-living cells and can not deliver pure cultures for antibiotic susceptibility testing. Methods for biochemical analysis (e.g., of fatty acids, nucleic acids, or proteins) using electrophoresis, mass spectroscopy, and chromatography can be powerful for microbial identification, but such methods are usually inappropriate for microbial enumeration and are generally too expensive and complex for routine microbial diagnostics.

[0028] Unmet Needs for Microbial Enumeration

[0029] In summary, current microbial enumeration testing is dominated by microbial culture. Microbial culture has the important advantages of being simple, ultra-sensitive, inexpensive, and quantitative but has the significant drawback of being slow. The long time required for results has major costs in healthcare and in manufacturing. More rapid methods have been developed, but while improving the time to results, they have sacrificed one or more of the critical advantages of microbial culture.

[0030] Thus, there is need for a test that is faster than traditional microbial culture but that retains the key benefits of the traditional method.

SUMMARY OF THE INVENTION

[0031] The invention enable efficient, rapid, and sensitive enumeration of living cells by detecting microscopic colonies derived from in situ cell division using large area imaging. Microbial enumeration tests based on the invention address an important problem in clinical and industrial microbiology—the long time needed for detection of traditional tests—while retaining key advantages of the traditional methods based on microbial culture. Embodiments of the invention include non-destructive aseptic methods for detecting cellular microcolonies without labeling reagents. These methods allow for the generation of pure cultures which can be used for microbial identification and determination of antimicrobial resistance.

[0032] The invention features a method for detecting living target cells in a sample including the steps of providing living target cells present in the sample in a detection zone including a detection area at a density of less than 100 target cells per mm ² of the detection area, allowing the formation of one or more microcolonies of the target cells by in situ replication; and detecting one or more microcolonies, wherein the replication produces one or more microcolonies; wherein the longest linear dimension of the detection area is greater than 1 mm; within the detection area, the cells are randomly dispersed and immobilized; the detecting detects one or more microcolonies that have a mean measurement of less than 50 microns in at least two orthogonal dimensions; and the cells in the one or more microcolonies remain competent to replicate following the detection step.

[0033] The invention further features a method for detecting microcolonies of target cells including the steps of providing target cells in a detection zone, wherein within the detection area, the cells are randomly dispersed and immobilized; allowing the formation of one or more microcolonies of the target cells by in situ replication, wherein at least one of the microcolonies includes fewer than 100 target cells; and detecting one or more naturally occurring optical properties of the one or more microcolonies using less than 5 fold magnification.

[0034] The invention also features an instrument for detecting microcolonies of target cells that includes a photoelectric array detector having an optical resolution of less than 20 microns and encircled energy values of greater than 70% per pixel; and an illumination source, wherein the instrument is capable of illuminating and simultaneously imaging a detection area having at least one dimension that is 24 cm, and wherein the instrument does not optically magnify more than 5 fold.

[0035] Advantages of the Invention

[0036] Some advantages of various embodiments of the invention are listed in Table 2. 2

[0037] The invention's short time needed to achieve results derives from the invention's ability to detect microcolonies containing only a small fraction of the cells that are required by the traditional methods. Since cell replication requires time, detecting small microcolonies using the invention provides results faster than detecting the large visible colonies using traditional enumeration methods. To detect small microcolonies the invention uses a combination of efficient signal generation and signal detection methods.

[0038] The ultra-sensitivity—its ability to detect small numbers of microscopic cells in large samples—stems, in part, from the use of large area imaging. For example, the invention can detect microscopic colonies without magnification. This feature allows a large area to be surveyed for microcolonies in a single image. Imaging a large area is a key to the invention's ability to efficiently analyze large sample volumes. For example, the microbial contaminants in a large volume of a sample can be deposited on a membrane using membrane filtration. The invention using large area non-magnified imaging of microcolonies can analyze the entire membrane efficiently. In contrast, using a high magnification microscope to evaluate the microcolonies on the same filter might require thousands of images.

[0039] The power to enumerate small numbers of microcolonies in a large area efficiently also comes from the invention's ability to use imaging approaches that compare object signals to local backgrounds. This ability improves the signal to background ratio for samples containing few cells over methods that integrate the total signal and background in a large area.

[0040] Assay robustness for samples with few cells is provided by the invention's inherent ability to enumerate growing microcolonies. Thus, the invention can decrease false positives over methods which detect a single integrated signal, such as methods that quantify the presence of biomolecules (e.g., ATP, antigens, or nucleic acids). Any artifact that causes a signal can generate a false positive when using methods that rely solely on integrated signal. Consider a sample that contains 482 microbial cells each of which generate 100 fluorescent units. The result of an integrative method is a single number (48,200 fluorescent units). Artifacts that generate a similar number of fluorescent units, for example, a large fluorescent dust particle may be indistinguishable. The invention, however, can easily distinguish between a single large dust fluorescent dust particle and 482 individual growing microcolonies.

[0041] Detecting growing microcolonies is a powerful method for discriminating against false positive signals from inanimate objects and cells incapable of growth under the test conditions. For example, consider a test to detect microbial microcolonies on a membrane lying on solid growth media in a petri dish. In one embodiment of the invention, the detection area is imaged before allowing the microbes in the detection area to grow into microcolonies. If some fluorescent dust particles or autofluorescent mammalian cells are present in the detection area some positive signals will be apparent in this “zero time” image. After incubating the petri dish to allow for microbial replication another image is taken. When the two images are aligned in register, the positive signals that correspond to microcolonies can be distinguished from the false positives since the false positives are present (usually unchanged) in the “zero time” image and the post-incubation image. Only growing microcolonies should appear over time. To confirm the microcolony signals, images can be acquired and compared at multiple time points during the incubation. Only growing microcolonies should increase in signal strength and in size over time.

[0042] Tests constructed using the invention can have a large dynamic range compared to tests constructed using methods in the prior art. Thus, for example, a test based on the invention designed can detect from one to 10 ⁶ microcolonies in a single image. In contrast, traditional microbial enumeration methods work best when about 30 to 150 colonies are deposited on a filter (47 mm diameter). New enumeration methods (e.g., Chemunex's ScanRDI and Millipore's MicroStar) also have limited dynamic ranges.

[0043] To achieve efficient signal generation, the invention can exploit either the intrinsic optical properties of the microcolonies (e.g., autofluorescence, reflectance, or light scattering) or various externally applied labeling reagents. The ability to exploit a range of optical properties and labeling methods enables creation of important microbiological tests. For example, using a method that detects a ubiquitous property of microcolonies (e.g., autofluorescence or infrared absorption) is useful for tests that enumerate total microbial content of a sample. Such tests are critical in food processing for determining the likelihood of spoilage and for finished product release testing in pharmaceutical manufacture. One important embodiment of the invention uses a reagent-less system based on detecting cellular autofluorescence to detect small microbial microcolonies. This embodiment provides a simple, non-destructive, aseptic approach to microbial enumeration. To detect specific types of cells, category-specific labeling reagents can be used. For example, a fluorescently labeled antibody that specifically binds to Listeria monocytogenes can be used to detect microcolonies derived from cells of this important food pathogen.

[0044] Like traditional microbial culture, the invention can exploit the diagnostic power of measuring microbial growth under selective conditions. For example, to determine bacterial resistance to antibiotics bacteria can be grown on growth medium onto which antibiotic disks have been placed. The size of the no-growth zone near the disks determines antibiotic resistance. The invention can be used to detect the size of this zone more rapidly. Similarly, the invention can be used to detect the growth of specific microbes on selective medium rapidly.

[0045] Simplifying the obligatory test validation cycle in which a new method is shown to be equivalent to the “gold standard” method is another advantage of the invention that derives from non-destructive enumeration. The invention facilitates equivalence to the “gold standard” culture tests by allowing an internal comparison of the new and old methods. Briefly, after imaging the microcolonies derived from microbes in a sample at an early time point, the samples can be re-incubated incubated for the amount of time required when using traditional visual detection of colonies. In this way an internal comparison can be made between the invention's enumeration of the microcolonies and the enumeration of the same colonies at a later time by the traditional method.

[0046] Other features and advantages will be apparent from the following description and the claims.

[0047] By target cell is meant a cell that is potentially present in a sample and whose presence is assayed by the invention.

[0048] By category of target cells is meant multiple target cells that are considered identical for the purposes of a test constructed using the invention.

[0049] Consider a test designed to detect any strain of E. coli bacteria. For the purposes of the test, the category “ E. coli ” would thus include any bacterium in the species E. coli . Such a test would be designed to detect, without differentiation, any bacterium in the species E. coli . Bacteria, and other target cells, that are not E. coli would either not be detected in this test, or would be detected and identified as not being members of the group E. coli . In contrast, consider a test designed to detect the pathogen E. coli O157:H7, a subgroup of the E. coli species. In this case, the subgroup E. coli O0157:H7 is a category of target cells. Bacteria in the subgroup, i.e., in the category “ E. coli O0157:H7”, are detected without differentiation. E. coli that are not in the E. coli O0157:H7 subgroup are not detected by the test and are therefore not in the E. coli O157:H7 category.

[0050] Categories need not be taxonomically related as in the previous paragraph. For example, a test might be designed to detect the category of bacteria that makes a protein that is required to confer resistance to the antibiotic vancomycin. This protein could be made by bacterial strains that are not closely related, i.e., that are members of disparate species. A vancomycin resistant strain in one species, however, is likely to be very closely related to vancomycin sensitive strains in the same species. The category of bacteria that make the vanA protein (important for achieving vancomycin resistance), for instance, includes vancomycin-resistant bacteria in the genus Enterococcus and in the genus Staphylococcus, while the majority of enterococci and staphylococci are not included in the category. Thus, in this case, it can be seen that the category encompasses target cells that are considered, for the purposes of the test, to be identical because of a common feature, in this case a molecular component (a category-specific binding site) rather than to a common phylogenetic (genealogical) relationship.

[0051] By non-overlapping categories of target cells is meant sets of target cells whose union is the null set. That is, the category of all E. coli bacteria, the category of all bacteria in the genus Pseudomonas, and the category of all fungi are non-overlapping categories. That is, no member of any of the categories is a member of any of the other sets.

[0052] By the categorical complexity of a test is meant the number of non-overlapping categories that are detected in the test.

[0053] By a category-specific binding site is meant a site on a target cell that specifically binds to a category-binding molecule under specific-binding conditions and that distinguishes target cells that are members of a particular category to be identified in a test from target cells that are not members of that category but might also be present in the test sample. That is, the site is present typically on all members of one category, and typically not on any members of non-overlapping categories. Category-specific binding sites specifically bind to category-specific binding molecules.

[0054] If a test scans a sample for a category of target cells that constitutes a taxonomic group, a category-specific binding site is one that is present in essentially all members of that taxonomic group, but is not present in essentially all members of other taxonomic groups that might be present in the test sample.

[0055] Alternatively, a test might scan a sample for category-specific binding sites that are shared by members of different taxonomic groups. Examples of this type of category-specific binding site include various macromolecules (e.g., DNA) and genes, mRNAs, and proteins that confer antibiotic resistance, confer virulence, or indicate viability. A category-specific binding site is often a part of a larger molecule or complex. For example, a category-specific genomic sequence can be used as a category-specific binding site in a test. Such a category-specific binding site is part of a much larger genome that contains (1) sections that are not category-specific; (2) sections that are category-specific binding sites but for which the test does not scan; and (3) other sections that are distinct category-specific sequences for which the test does scan.

[0056] Binding sites that are present, e.g., in 80%, 90%, 95%, or more than 99% of the target cells that are members of a category but that are absent, e.g., in 80%, 90%, 95%, or more than 99% of the target cells that are members of all other categories of the same class, are considered category-specific binding sites. Note that a category-specific binding site can be trivially or exceptionally absent from a target cell that is a member of the category. Similarly, a category-specific binding site can be trivially or exceptionally present in a target cell that is not a member of a category. For example, consider a protein site that occurs in essentially all E. coli bacteria but in no other bacterial species. If, as might be the case in less than one cell out of millions of bacteria, a mutation causes the protein not to be produced, the marker will not be present in that strain of E. coli . However, this protein site is still considered a category-specific binding site. Alternatively, the gene for the same protein is transferred to a strain of a different species of bacteria by recombinant DNA technology or by natural means (e.g., by viral transduction). In this case, a bacterial strain that is not a member of the category E. coli would express what would still be considered an E. coli -specific binding site.

[0057] By category-binding molecule is meant a molecule or molecular complex that specifically binds to a category-specific binding site. Examples of category-binding molecules are nucleic acid probes that hybridize to genomic DNA; nucleic acid aptamers that have been selected or “evolved” in vitro to bind specifically to sites on proteins; antibodies that bind to cellular antigens or serum proteins; and ligands such as epidermal growth factor or biotin that bind specifically to hormone receptors or to binding molecules, such as avidin. Two category-binding molecules are said to be distinct if they bind to distinct and non-overlapping category-specific binding sites. Category-binding molecules may be referred to according to their molecular composition, e.g., a category binding oligonucleotide, probe, antibody, ligand, etc.

[0058] By a category-binding molecule that specifically binds to a category of target cells is meant a category-binding molecule that binds under defined binding conditions to essentially all target cells that are members of a category scanned for by a test, but to essentially no target cells that are not members of the category but that are likely to be present in the sample. The number of category-binding molecules that are bound by target cells in a category scanned for as compared to the number bound by target cells(not in such a category, are typically two-fold, five-fold, ten-fold, or greater than fifty-fold greater.

[0059] By binding conditions is meant the conditions used in a test to achieve specific binding of category-binding molecules to category-specific binding sites. For example, when the category-binding molecules are category-specific DNA probes, the binding conditions for a particular test might be stringent DNA hybridization conditions. The appropriate stringent DNA hybridization conditions depend on the nature of the probes, as is well known by those familiar with the art. For example, for typical DNA probes of length greater than 500 bases, an appropriate binding condition (usually referred to as a “washing condition” in the hybridization vernacular) is 65° C. at 0.2×SSC. For binding an antibody to an antigen, typical binding conditions are room temperature in PBS-TB.

[0060] By a family of category-binding molecules is meant a set of category-binding molecules that specifically bind to a particular category of target cells.

[0061] Polyclonal antibodies generally constitute families of category-binding molecules since they generally comprise multiple distinct category-binding molecules that bind to the same category of target cell. Note that, unless affinity purification is used, polyclonal antibody preparations typically also contain antibodies that do not bind to the chosen category of target cell and may contain antibodies that bind to other categories. Additional antibodies are present because the antibody repertoire of an animal is determined by the animal's infection history. Therefore, polyclonal antibodies are preferably purified by affinity methods. Category-binding molecules in a family might bind to some target cells in the category but not to others.

[0062] Another example of a family of category-binding molecules is a set of 80 category-specific genomic DNA sequences that occur in all E. coli O157:H7 strains but that do not occur in members of other groups of bacteria. This family of category-binding molecules can hybridize as a group to suitably prepared E. coli O157:H7 cells, but does not hybridize to other categories of cells. Families can include different types of category-binding molecules. For example, a monoclonal antibody that specifically binds to the O157 antigen and one that binds to the intimin protein (a virulence factor) could also be included in the above family of category-binding molecules. A family of category-binding molecules can comprise any number of category-binding molecules (i.e., one or more).

[0063] By non-overlapping families of category-binding molecules is meant families of category-binding molecules in which each family binds specifically to one, and only one, category in a set of non-overlapping categories. That is, a set of non-overlapping families of category-binding molecules map to a congruent set of non-overlapping categories. For example, in a test that scans the 4 USP objectionable organisms E. coli , Salmonella, Pseudomonas spp., and Staphylococcus aureus , there are four non-overlapping categories. Such a test might incorporate four different non-cross-reacting polyclonal antibodies, each specific for one of the test categories. Thus, the test comprises four non-overlapping families of category-binding molecules. The non-overlapping families of category-binding molecules in a test are called an ensemble of category-binding molecules.

[0064] By an ensemble of category-binding molecules is meant a set of one or more non-overlapping families of category-binding molecules that are combined in a mixture for a particular test. Tests that scan for multiple non-overlapping categories of target cells comprise one family of category-binding molecules per category. The entire set of category-binding molecules, that comprise these families, is referred to as an ensemble.

[0065] By the category-binding molecule complexity of an ensemble is meant the number of distinct category-binding molecules or moieties in an ensemble. For example, if an ensemble of category-binding molecules consisted of 234 oligonucleotide probes, the category-binding molecule complexity of the ensemble would be 234.

[0066] By the family complexity of an ensemble is meant the number of non-overlapping families of category-binding molecules in an ensemble. The family complexity is the same as the minimum number of target cells required to bind a category-binding molecule from each of the families in an ensemble. The family complexity of a test corresponds to the categorical complexity of a test—i.e., the number of distinct categories for which the sample is scanned. In general, the family complexity also corresponds to the number of distinct signal signatures used in a test.

[0067] By signal element is meant a molecule or particle that directly generates a detectable signal. The phrase “directly generates” refers to the fact that signal elements are the immediate source or critical modulator of the detectable signal. Thus, if the signal is photons that arise from a fluorophore, the fluorophore is the immediate source of the photons and, therefore, is a signal element. If the signal is photons scattered by an RLS particle, the RLS particle is a signal element. Alternatively, if the signal is the light transmitted or scattered from a chromogenic precipitated product of the enzyme horseradish peroxidase, the chromogenic product is the signal element.

[0068] A characteristic of a signal element is that such an element cannot be divided into parts such that each part generates a signal that is comparable (in character, not necessarily in intensity) to the whole. Thus, a 2 nM diameter quantum dot is a signal element, as dividing it changes the character (emission spectrum) of the resulting nanocrystals. A 5 μm particle impregnated with a fluorescent dye such as fluorescein, is not a signaling element, since it could be divided into parts such that each part has signaling characteristics comparable to the intact particle. The molecule fluorescein, in contrast, is a signaling element. The detectable products of signal generating enzymes (e.g., luciferase, alkaline phosphatase, horseradish peroxidase) are also considered signal elements. Such signal elements (or their precursors when there is a chemical conversion of a precursor to a signal element) may be diffusible substances, insoluble products, and/or unstable intermediates. For example, the enzyme alkaline phosphatase converts the chemiluminescent substrate CDP-Star (NEN; catalog number NEL-601) to an activated product, which is a photon-emitting signal element.

[0069] By signaling moiety is meant a molecule, particle, or substance comprising or producing (in the case of enzymes) one or more signal elements and that is or can be conjugated to a category-binding molecule. The signaling moiety can be attached to the category-binding molecule either covalently or non-covalently and either directly or indirectly (e.g., via one or more adaptor or “chemical linker” moieties). Examples of signaling moieties include carboxylated quantum dots; a fluorophore such as Texas Red that is modified for binding to a nucleic acid probe or an antibody probe; streptavidin-coated fluorescent polystyrene particles (which can be conjugated to biotinylated category-specific binding proteins); a rolling-circle replication product containing repeated nucleic acid sequences each of which can hybridized to several oligonucleotides tailed with fluorescently modified nucleotides and which contains a category-specific binding oligonucleotide at the 5′ end. A signaling moiety can comprise physically distinct elements. For example, in some cases the signaling moiety is an enzyme (e.g., alkaline phosphatase) that is conjugated to a category-binding molecule (an antibody, for example). Signal is generated when a substrate of alkaline phosphatase (e.g., CDP-Star, or BM purple from NEN and Roche, respectively) is converted to products that are signal elements (e.g., an unstable intermediate that emits a photon, or a precipitable chromogenic product). It is not unusual for the category-binding molecules, enzymatic signaling moieties, and substrate to be applied to the reaction at distinct times.

[0070] By signaling moiety complex is meant a physical cell that comprises more than one signaling moiety and more than one category-binding molecule. The physical association of the signaling moieties and category-binding molecules in a signaling moiety complex must be stable (e.g., the signaling moieties and category-binding molecules should have mean half-lives of association with the complex of at least one day in PBS at 4° C.). As an example of a signaling moiety complex, consider a polystyrene microparticle that is coated with thousands of molecules of two types: a target cell-specific antibody and alkaline phosphatase. Such a signaling moiety complex binds to the target cell via the conjugated antibody category-binding molecule. When incubated with a chromogenic alkaline phosphatase substrate (the signal element; e.g., BM purple, Roche), a colored spot can be generated that can be detected by eye. Alternatively, the same signaling moiety complex, when incubated with either a chemiluminescent or a fluorescent alkaline phosphatase substrate, generates either a chemiluminescent or fluorescent signal. Further examples of signaling moiety complexes include: nanogold particles conjugated to fluorescein-labeled antibodies, and latex particles conjugated to both oligonucleotide category-binding molecules and acridinium esters that chemiluminesce upon addition of hydrogen peroxide.

[0071] By signal character of a signal element or signal moiety is meant the aspect or aspects of a signal generated by the signal element signaling moiety that is useful for distinguishing it from other signal elements or signaling moieties. For example, the signal character of a signaling moiety labeled with fluorescein and rhodamine is fluorescence. The character of a radio transponder is radio frequency. Examples of photonic signaling character are fluorescence, light scattering, phosphorescence, reflectance, absorbance, chemiluminescence, and bioluminescence. All but the latter two examples of photonic signaling character depend on external illumination (e.g., a white light source, a laser light source, or daylight). In contrast, chemiluminescence and bioluminescence are signaling characters that are independent of external light sources.

[0072] By the class of a signal element or signaling moiety is meant the distinct quality of the signal that is useful for distinguishing it from other signal elements or signaling moieties. For example, a liposome that is labeled with red dye is distinguished from differently colored liposomes. The color red is its class. For a micro-transmitter that broadcasts a particular radio-frequency signal, the quality of the radio-frequency signal that differentiates the micro-transmitter from other micro-transmitters constitutes the signal element class.

[0073] By signal signature is meant the distinctive signaling quality of the combination of signaling moieties that bind to a category of target cells in a test. A target cell that is bound to four types of antibodies, one of which is conjugated to a fluorescein molecule, and three of which are conjugated with rhodamine molecules has a signal signature that is described by the combined weighted absorbance and emission spectra of fluorescein and rhodamine.

[0074] By signal complexity of a test or an ensemble of labeled category-binding molecules is meant the number of categories of target cells that can be distinctly labeled in the test or by binding to the ensemble. Alternatively, the signal complexity is defined as the number of distinct signal signatures that would be expected to occur if a member of each category of target cell were present. For some tests, the signal complexity of an ensemble of category-binding molecules is the same as the number of categories for which the test scans. Other tests, which scan for many categories, may only have a signal complexity of one.

[0075] By selection force is meant a force that is used to capture, isolate, move, or sequester target cells. Examples of selection forces include gravity, magnetism, electrical potential, centrifugal force, centripetal force, buoyant density, and pressure. Target cells can be mobilized by a selection force acting on the target cell alone. Alternatively, selection forces can act specifically on target cells that are associated with selection moieties (see definition below).

[0076] Examples of the application of selection forces to mobilize target cells include: centrifugation of target cells; magnetic selection of target cells bound to magnetic particles; gravitational sedimentation of target cells labeled with metallic particles; and deposition of target cells on a porous membrane by vacuum filtration.

[0077] By selection moiety is meant an atom, molecule, particle, or cell that can be conjugated to a category-binding molecule and that confers on the category-binding molecule the ability to be selectively captured, isolated, moved, or sequestered by a selection force. When a category-binding molecule:selective moiety complex is specifically bound to a target cell, the target cell can also generally be selectively captured, isolated, moved, or sequestered by the selection force. Selective refers to the preferential conferring of susceptibility to mobilization by the selection force on selection moieties and associated cells over cells not associated with selection moieties.

[0078] Paramagnetic particles and ferritin are examples of selection moieties. A dense silica particle that sinks in solution is another type of selection moiety. Such particles, when coated with category-binding molecules and bound to a microbial target cell will cause the target cell to sink in aqueous solution, thus enabling separation of the bound target cell from other sample unbound constituents.

[0079] By selective character is meant the aspect or aspects of a selection moiety that is useful for capturing, selecting, or moving the selection moiety. For example, the selective character of a paramagnetic particle is magnetism. The selective character of a silica particle that rapidly sinks in aqueous solution is mass.

[0080] By a roughly planar surface or substrate is meant a surface that can be aligned in parallel to an imaginary plane such that when the distance is measured from points in any 1 mm×1 mm square on the surface to the closest points on the imaginary plane, the absolute value of the mean distance is less than 50 micrometers.

[0081] By detection surface is meant the surface of a roughly planar substrate onto which target cells are deposited. In embodiments using photonic signaling character, if the detection surface is optically transparent, detection can be effected via either face of the detection surface. If the detection surface is opaque, detection is effected via the face of the detection surface on which the target cells are deposited.

[0082] By detection area is meant the area of the detection surface that is simultaneously sampled by a detection device. For example, the section of a glass slide that is simultaneously imaged by an optical device that includes a collection lens and a CCD chip might measure 0.8 cm×0.5 cm. The detection area is then 0.4 cm ² .

[0083] By detection zone is meant the volume in which replicating target cells can be detected by the detection device. The detection zone has the same dimensions as the detection area but has a depth corresponding to the depth in which the signal from replicating target cells can be detected and identified. The depth of the detection zone is therefore dependent on the threshold criteria used to score for positive signal. When optical detection is used, the depth of the detection zone is dependent on the optical depth of field.

[0084] By the longest dimension of a detection area is meant the line of maximum length that can be drawn between two points on the perimeter of the detection area. For example, if the detection area is a rectangle measuring 0.3 cm×0.4 cm, the longest dimension of the detection area is the diagonal, 0.5 cm. If the detection area is an ellipse with semi-major axis of length 7 mm and semi-minor axis of length 2.5 mm, the longest dimension of the detection area is 14 mm.

[0085] By large area detection or large area imaging is meant a method for detecting microscopic target cells in which the detection area (the area that is simultaneously analyzed by the detection device) is much larger than the dimensions of the target cells or microcolonies. The detection area for large area detection has at least one linear dimension that is ≧1 mm. In contrast, the microscopic colonies are substantially smaller, typically measuring less than 50 μm in at least two orthogonal dimensions. Examples of large area detection include imaging a 9 mm diameter detection area with a CCD camera; imaging a 2 cm×1 cm rectangle by scanning with a linear array detector that has a long dimension of 1 cm; and imaging a 4 cm×4 cm filter using direct exposure on photographic film.

[0086] Some technologies scan samples for microcolonies but do not exploit large area detection. Examples include solid phase laser microbeam scanning cytometry and microscopic examination of multiple high power microscopic fields on a slide.

[0087] By conjugated or stably associated is meant a physical association between two entities in which the mean half-life of association is least one day in PBS at 4° C. Consider, for example, the complex case of passive protein adsorption to polystyrene particles. There are several different classes of adsorbed proteins. Some proteins are stably associated to the surface with half-lives of many months. Other proteins, such as those that are loosely bound on the outer layer of adsorbed protein, may not be stably associated with the particles and can leach out within hours.

[0088] By particle is meant a rigid matrix (i.e., with at least some characteristics of a solid), which measures less than one millimeter along any axis. Particles can be doped with or conjugated to signal elements. Particles are often referred to as particles or with terms that reflect their dimensions or geometries. For example, the terms nanosphere, nanoparticle, or nanobead are used to refer to particles that measures less than 1 micron along any given axis. Similarly, the terms microsphere, microparticle, or microbead are used to refer to particles that measure less than one millimeter along any given axis. Examples of particles include latex particles, polyacrylamide particles, magnetite microparticles, ferrofluids (magnetic nanoparticles), quantum dots, etc.

[0089] By image intensifier or image tube is meant a device that amplifies a photonic signal, as defined in the glossary of Inoué, Shinya, et al., Video microscopy: the fundamentals (Plenum Press, New York, 1997; p. 665): “A device coupled (by fiber optics or lenses) to a video camera tube to increase sensitivity. The intensifier is a vacuum tube with a photocathode on the front end that emits electrons according to the image focused upon it, an electron lens and/or microchannel plate(s) that focuses the electrons onto a phosphor at the back end, and a high voltage accelerator that increases the energy of the electrons. Can be single or multiple stage.” A variety of such image intensifiers is described in detail in Chapter 8 of the same reference.

[0090] By simultaneous detection in a section of the detection area is meant detection of the signal from a section of a roughly planar detection surface in one step. Large area imaging of targets in a detection area using a CCD chip, visual detection, or photodiode-based signal integration are examples of simultaneous detection.

[0091] By identification is meant determining the category or categories of which a target cell is a member.

[0092] By sample is meant material that is scanned by the invention for the presence of target cells.

[0093] By direct visual detection is meant visual detection without the aid of instrumentation other than wearable corrective lenses.

[0094] By photoelectric detector is meant a man-made device or instrument that transduces photonic signals into electric signals. Examples of photoelectric detectors include CCD detectors, photomultiplier tube detectors, and photodiode detectors, e.g., avalanche photodiodes.

[0095] By encircled energy or ensquared energy is meant the percentage of photons from an infinitely small light source that are captured on a pixel of a photodector array.

[0096] By thermal radiation is meant black body radiation.

[0097] By cellular autofluorescence or autofluorescence is meant the fluorescence exhibited by cells due to the fluorescence of natural intrinsic cellular constituents, such as NADH and oxidized flavoproteins. Cells expressing fluorescence due to recombinant fluorescent proteins such as green fluorescent protein are not considered to be autofluorescent.

[0098] By in situ replication is meant the replication of a target cell in place, so that the daughter cells remain essentially co-localized with the progenitor target cell. For example, in in vitro biological culturing of bacteria on nutrient agar plates, single dispersed bacteria are deposited on a plate and incubated under conditions that permit bacterial replication. A bacterium in a certain location replicates giving rise to progeny cells that also replicate. All of the cells remain co-localized (essentially contiguous) with the original cell, eventually giving rise to a visible colony on the plate. Where there was formerly a single cell, there is now a colony of more than 10 ⁷ cells.

[0099] By a microcolony of target cells is meant a set of target cells that lie in close physical proximity to each other, that lie on (or are anchored to) a surface, and that are the clonal descendants via in situ in vitro replication-based amplification of a single ancestral target cell. A microcolony is generally too small to be visible by the naked eye (e.g., less than 50 microns in diameter).

[0100] Any type of dividing target cell can give rise to microcolonies in situations that lead to physical co-localization of the clonal descendents of the target cells. For example, microcolonies could contain animal or plant cells, fungi, or bacteria.

[0101] By illuminating is meant irradiating with electromagnetic radiation. Electromagnetic radiation of various wavelengths can be used to illuminate. It includes, for example, radiation with wavelengths in the X-ray, UV, visible, or infrared regions of the spectrum. Note that illuminating radiation is not necessarily in the visible range.

[0102] By signal elements or signaling moieties with photonic signaling character is meant signal elements or signaling moieties that are detectable through the emission, reflection, scattering, refraction, absorption, capture, or redirection of photons, or any other modulation or combination of photon behavior. Some examples of signal elements or signaling moieties that have photonic signaling character include: the fluorophore Texas Red (fluorescent signaling character); CDP-Star (chemiluminescent signaling character); luciferase (bioluminescent signaling character); resonance light scattering particles (light scattering signaling character); BM purple (light absorption or chromogenic signaling character); and up-converting phosphors (absorption of two long wavelength photons and emission of one shorter wavelength photon).

[0103] By ‘number’ X '‘solution name’ is meant an aqueous solution comprising the constituents of solution name at number times the concentration of the solution (except for water). For example, 10×EE contains 10 mM EDTA/100 mM EPPS (EE, or 1×EE, contains 1 mM EDTA/10 mM EPPS).

[0104] EE is a solution that is 1 mM EDTA/10 mM EPPS. Before mixing them together, the conjugate acids of both components are brought to pH 8.0 with NaOH

[0105] PB is 0.1 M sodium phosphate buffer pH 7.4.

[0106] PBS is a phosphate-buffered saline solution containing: 120 mM NaCl, 2.7 mM KCl and 10 mM phosphate buffer (sodium salt) pH 7.4.

[0107] PBS-B is 0.1% BSA (IgG Free; Sigma Cat. No. A-7638) in PBS.

[0108] PBS-T is 0.05% Triton ×-100 (Sigma Cat. No. ×-100) in PBS

[0109] PBS-TB is PBS/0.1%BSA/0.05% Triton ×-100

[0110] PBT is PBS/0.1% BSA (IgG Free; Sigma Cat. No. A-7638)/.05% Tween-20 (Sigma Cat. No ×-100)

[0111] LB is Luria Broth for growing bacteria and is made as described previously (Ausubel 1987, supra).

[0112] SSC is 150 mM NaCl/15 mM Na ₃ citrate adjusted to pH 7.0 with HCl.

[0113] EDAC is (1-Ethyl-3-(3-dimethylaminopropyl)) carbodiimide.

[0114] TSA is Tryptic Soy Agar (Becton Dickinson/Difco; cat. num. 236950).

[0115] TSB is Bacto™ Tryptic Soy Broth (Becton Dickinson cat. num. 211822).

[0116] AP is alkaline phosphatase.

[0117] BSA is Bovine Serum Albumin.

[0118] CCD is charged coupled device.

[0119] Cfu is Colony forming unit (a measure of bacterial concentration that corresponds to the number of viable bacterial cells).

[0120] FITC is fluorescein isothiocyanate.

[0121] PNA is peptide nucleic acid.

[0122] Unless otherwise noted, microbiological strains described in the specifications are obtained from the American Type Culture Collection (ATCC), Manassas, Va.