New and advanced methods of molecular diagnostics are changing the way we practice clinical microbiology, which affects the practice of medicine. Signal amplification and real-time nucleic acid amplification technologies offer a sensitive and specific result with a more rapid turnaround time than has ever before been possible. Numerous methods of postamplification analysis afford the simultaneous detection and differentiation of numerous microbial pathogens, their mechanisms of resistance, and the construction of disease-specific assays. The technical feasibility of these assays has already been demonstrated. How these new, often more expensive tests will be incorporated into routine practice and the impact they will have on patient care remain to be determined. One of the most attractive uses for such techniques is to achieve a more rapid characterization of the infectious agent so that a narrower-spectrum antimicrobial agent may be used, which should have an impact on resistance patterns.
It has been >2 decades since the inception of the PCR [1, 2]. Although this technique was almost immediately implemented in research laboratories to study a variety of infectious diseases, among other processes, it has been within only the past 5–7 years that these assays have been made so user-friendly that they may be implemented in any microbiology laboratory . In addition to PCR, there are numerous competitive technologies that have been developed and shown to be useful for the detection and characterization of microorganisms. Many of these more recently developed assays may be performed by medical technologists who, although highly skilled, do not have advanced training in molecular biology. The simplification of this type of testing means that it may be performed after hours or in adverse conditions (i.e., in the field). Several of these assays are US Food and Drug Administration approved or cleared, and more are in this process. Today, in the modern microbiology laboratory, there truly is a “molecular revolution” under way. Although this is exciting, the responsibility of the laboratorian and the ordering physician is to ensure that these tests are used appropriately. The abuse of this technology has serious implications for cost as well as patient care, because, as with any test, false-positive and false-negative reactions may occur.
The clinical microbiologist is occasionally asked to exclude the presence of a particular microorganism in a clinical specimen. For example, CSF may be submitted for the exclusion of herpes simplex virus, or a blood smear may be submitted for the exclusion of Plasmodium species. However, in many instances, several tests are ordered per clinical specimen, and the questions asked are of a broader nature. For example, when a blood or wound culture is ordered, the question really is “Are there any bacteria or fungi present in this clinical specimen, and to what antimicrobial agents are they susceptible?” The clinician rarely draws a blood sample for culture for the detection or exclusion of a single type of microorganism (i.e., blood is not drawn for culture to “rule out [r/o] Pseudomonas” but, rather, to “r/o bacteremia and fungemia”). The task of the clinical microbiologist is complicated, because he or she must be aware of all of the potential pathogens that may be present in a clinical specimen and must devise or utilize means by which to detect and characterize these organisms.
The molecular diagnostic techniques used for the detection and characterization of microorganisms may be separated into several broad categories. These are direct hybridization, nucleic acid amplification, and a variety of methods for postamplification analysis. These methods may be used to either detect or exclude particular pathogens or, more usefully, to assess a specimen or positive culture sample for a wide variety of microorganisms. The value of such applications must be rigorously assessed, because these tests are usually added to the battery of routinely performed assays. The benefits of such assays are less likely to be found in the laboratory and more likely to be seen by the clinician or the health care administrator. For example, the more rapid detection and characterization of an infecting pathogen afford the clinician the opportunity to tailor antimicrobial therapy. In addition to benefiting the patients, this may save health care dollars through the pharmacy by identifying situations wherein less expensive antimicrobial agents may be used [4, 5]. For example, Forrest et al.  demonstrated that $1729 in pharmacy costs per patient could be saved by switching the patient's treatment from a more expensive echinocandin to a less expensive azole. The use of narrower-spectrum agents, in contrast to the use of broad-spectrum agents, should also slow the selection for and spread of antimicrobial-resistant microorganisms. Other benefits that have been noted include better use of ancillary diagnostic tests (e.g., radiology services) and of hospital beds (i.e., decreasing lengths of stay or the number of unnecessary admissions) .
There are many direct-hybridization techniques used in clinical microbiology. Some of these have been used for the rapid identification and characterization of bacteria and fungi in blood culture samples, whereas others have been used for different applications (e.g., human papillomavirus [HPV] detection) but could also potentially be used in this manner. In situ hybridization will predominate this discussion, because it is the newest signal amplification technique to be used in the clinical microbiology laboratory. The detection process used in in situ hybridization may be fluorescent (FISH) or chromogenic (CISH). These tools, which were once used solely in the domains of researchers and anatomic pathologists, have been demonstrated to be useful for the detection of pathogens in blood culture samples as well as in other clinical specimens [7–16]. A chemiluminescence-generating chemical reaction used after the direct hybridization of oligonucleotide probes to microorganism ribosomal RNA (rRNA) is another technology that may be used for the rapid characterization of microorganisms. This technology has been used for more than a decade for the identification of microorganisms, such as commonly occurring Mycobacterium species and the dimorphic fungi, in positive culture samples [17–21]. Another technology, target-capture technology, used predominantly for the detection of viruses (HPV, cytomegalovirus, and hepatitis B virus), could also be used for the detection of bacteria and fungi in positive blood culture samples [22–25].
In situ hybridization. In situ hybridization has been introduced into the clinical microbiology laboratory and should prove to be a useful technology for the rapid characterization of bacteria and fungi in positive blood culture samples, which is an area of particular interest to many [5, 7, 9–11, 13, 15, 26–30]. In situ hybridization has also been used successfully in direct clinical specimens and in histologic sections for the detection of microorganisms [8, 12, 14]. Regardless of the type of direct hybridization technology used, identifications using signal hybridizations are most accurately obtained after microorganism growth or biological amplification of the organism. This technology is often used to differentiate microorganisms of similar morphotypes in positive culture samples (e.g., gram-positive cocci in clusters or acid-fast bacilli). One of the advantages of this technology is that it may be used in conjunction with traditional methods, such as Gram staining. The information derived from Gram staining may be used to select the appropriate probes from a battery to more specifically address select questions. However, the use of the full battery of probes on all blood culture bottles that indicate positivity without the Gram stain-derived information would be cost-prohibitive and an inefficient use of resources. The information derived from Gram staining is used to select the one or few probes necessary for bacterial or fungal characterization. For example, if gram-positive cocci in clusters are demonstrated in the Gram stain, then one would select a probe that would detect Staphylococcus aureus and differentiate it from coagulase-negative staphylococci. Similarly, a battery of probes for the most commonly occurring yeast species would be used when yeasts are seen in the Gram stain, whereas probes for Pseudomonas and members of the Enterobacteriaceae would be used when gram-negative bacilli are seen in the Gram stain smear. This is an excellent example of optimizing the use of new technology through the use of established methods. A variety of oligonucleotide probe sequences have already been described that generate family-, genus-, and species-specific information [13, 15].
Two types of oligonucleotide probes are most frequently used for in situ hybridization. These are traditional nucleic acid oligonucleotide probes and peptide nucleic acid (PNA) probes. Traditional oligonucleotides used for in situ hybridization need no further explanation for most in the field of molecular diagnostics, but a limited description of PNA probes is warranted [31, 32]. Peptide nucleic acid probes are synthetic molecules that do not naturally occur. The most important difference between DNA and PNA resides in the “backbone” of the molecule. The backbone of the DNA polymer consists of repeating deoxyribose sugar and phosphate molecules, the latter of which have a net negative charge. In contrast, the backbone of the PNA molecule, used in the studies described here, consists of repeating glycine molecules. Glycine, importantly, is a sterically small and neutrally charged amino acid. The similarity between DNA and PNA oligonucleotide probes is the 4 common nucleotide bases, which participate in Watson and Crick base pairing. These unique properties afford the construction of a probe molecule that can penetrate through the cell wall and cell membrane (i.e., a hydrophobic lipid bilayer) of microorganisms while retaining the specificity achieved with Watson and Crick base pairing . The charge of the in situ probe is particularly important in microbiology, wherein the microorganism undergoing FISH analysis is intact, and in distinct contrast to many applications in molecular pathology (e.g., FISH for Her-2 amplification), wherein the nucleus, which contains the chromosomal targets for the FISH assay, has been sectioned, and the probes are applied directly onto their nucleic acid targets (i.e., penetration through a lipid bilayer is not necessary).
Although there is limited commercial availability of FISH probes in microbiology, some are available, and more are likely on the horizon. The PNA FISH assays that are currently available in North America detect and differentiate S. aureus from other gram-positive cocci in clusters and Candida albicans from other yeasts. The assay may be performed as soon as the blood culture bottle indicates positivity and Gram staining has been performed, or the slides to be tested may be tested in batches. Although batch testing clearly increases the time to result that is possible with immediate testing, it is a viable alternative for laboratories with limited personnel and still has a better time to result than traditional methods. It is now possible, with this technology, to provide a definitive species-level identification of microorganisms, such as S. aureus and C. albicans, on the same day that the blood culture bottle indicates positivity and demonstrates gram-positive cocci in clusters or yeast, respectively [7, 9–11, 26, 30]. In addition to the commercially available applications, several studies have demonstrated the feasibility of definitively identifying other medically important bacteria and yeast species [13, 15, 28, 29]. Importantly, these assays are simple to perform and represent a molecular diagnostic tool that may be introduced into laboratories without previous experience in molecular microbiology. They have already been adopted and are performed routinely in the several clinical microbiology laboratories in the United States. From an economic standpoint, the use of the C. albicans PNA FISH probe has been shown to be cost-effective, particularly when linked to limiting the use of more expensive antifungal drugs, such as an echinocandin, when a less expensive but equally efficacious drug like fluconazole is a viable alternative [4, 5]. In addition to its use in the differentiation of the causes of bacteremia and fungemia, in situ hybridization has been used to identify the most common causes of infection in patients with cystic fibrosis, to differentiate mycobacteria in culture and in histologic sections, and to detect trypanosomes in patients with African sleeping sickness [8, 14, 27].
The same hybridization probes used in the clinical microbiology laboratory could be used in in situ hybridization reactions in histologic sections. Hayden et al. have demonstrated, in a series of articles [12, 33–35], that chromogenic in situ hybridization may be used successfully to differentiate morphologically similar microorganisms. They showed the feasibility of this technology to detect Legionella pneumophila and differentiate filamentous bacteria (i.e., aerobic actinomyces), filamentous fungi, and yeast and yeastlike fungi. Importantly, they showed the ability to separate the systemic dimorphic fungi, which appear as yeast (e.g., Histoplasma capsulatum and Blastomyces dermatitidis) or spherules (e.g., Coccidioides immitis), from Candida and Cryptococcus species. Although it may seem a morphologically simple task to differentiate a well-developed spherule of C. immitis from budding yeast, it is considerably more complicated when intact spherules are not seen and, rather, only immature spherules and endospores are present. In situ hybridization has also been used to identify Pneumocystis jiroveci . Finally, one can imagine an economic benefit when the same probe sets are used for both microbiology and infectious disease pathology applications. For example, PNA FISH probes have also been used successfully to identify mycobacteria in histologic sections and from positive culture samples .
Another exciting application that may have practical uses in the clinical microbiology laboratory is the combined use of in situ hybridization and flow cytometry. Small-scale flow cytometers that can detect microorganisms have been developed and are routinely used in the food and pharmaceutical industries. These applications often use nonspecific dyes that bind to the external aspect of the bacteria. Although such assays are useful for determining overall bacterial counts in a product, the results generated do not yield genus- or species-level information. However, it has been demonstrated that this technology can be used to detect microorganisms to which in situ probes are hybridized [37, 38]. This combination of technologies affords the replacement of human interpretation, which is subjective and requires advanced training and experience to perfect, with objective, quantitative measures that are performed by an instrument.
The precise role that in situ hybridization will play in the clinical microbiology laboratory and its impact on patient care remain to be seen. Several other competitive technologies, some of which are described below, offer similar advantages with regard to rapid identification. Regardless, it has been demonstrated that the identification of the most common microorganisms that cause bacteremia and fungemia may be rapidly determined by use of this technology.
Other methods of signal amplification. There are several other signal-amplification methods that are currently used in many clinical microbiology laboratories, the applications of which could be expanded. The chemiluminescent rRNA probe hybridization technology that has been available from GenProbe for years could be expanded/promoted to detect many of these pathogens. This technology uses probes that target rRNA, which is present in numerous copies in each cell. The hybridized probe is then detected using a proprietary chemical reaction that results in the generation of light. Products that utilize this technology have been used for years in many laboratories for the rapid identification of the most commonly occurring mycobacteria and for isolates suspected to represent systemic dimorphic fungi [18–21, 26, 39–41]. Although products have been developed for rapid detection of many of the more commonly occurring bacteria, such as S. aureus, these have not gained widespread adoption for the rapid identification of the common causes of bacteremia and fungemia [26, 42, 43]. There have, however, been some hints that this company might be interested in developing panels useful for the identification of the commonly encountered bacteria and yeast species, although a commercial product has not become available.
The hybrid capture (Digene) technology, which has been used most successfully for the detection of high-risk HPV subtypes, is also a signal amplification technology . This technology has also been used for the detection of cytomegalovirus, hepatitis B virus, Neisseria gonorrhoeae, and Chlamydia trachomatis [45–47]. Although signal amplification assays that do not employ direct microscopy are often not as sensitive as nucleic acid amplification, some have shown the contrary [44, 48]. These assays have adequate sensitivity for many clinical applications and afford high specificity, depending on the specificity of the probes used. In addition to the types of applications described, this technology could also be used to characterize bacteria or fungi that grow in culture and require further identification.
Nucleic Acid Amplification
Nucleic acid amplification includes not only PCR but also alternate technologies, such as strand-displacement amplification and transcription-mediated amplification (i.e., nucleic acid sequence-based amplification), which have also proven useful in clinically important assays. In addition to the various chemical reactions that may be used for amplification, the design of these assays varies widely. Monoplex assays are amplification reactions that detect a single target and provide a “present or absent”-type result. For example, a Legionella pneumophila-specific PCR could be used to rapidly detect this pathogen. If used in conjunction with quantitative standards, these could be used to provide quantitative information (e.g., viral loads). The combination of ⩾2 nucleic acid amplification reactions into 1 assay is a multiplex reaction. These assays have been used to detect multiple pathogens simultaneously but require advanced design and critical assessments, because intramolecular interactions between the various primers and probes are potential limitations of such assays. Several multiplex (and monoplex) assays are commercially available as “analyte-specific reagents.” These are commercial products that contain all the necessary components to conduct a particular amplification reaction, which are manufactured under Good Manufacturing Practice conditions, but the responsibility of validating such assays resides with the testing laboratories. One of the more successful of the multiplex analyte-specific reagents that has been released is a real-time RT-PCR that simultaneously detects and differentiations the most important respiratory viral pathogens—influenza A, influenza B, and respiratory syncytial virus. The real-time version currently available is a truncated version of another multiplex assay produced by the same company that detected and differentiated 6 respiratory viruses [49, 50].
Broad-range nucleic acid amplification is another approach that could be used to detect any of a wide variety of microorganisms are in a taxonomically related group [51–54]. The results of broad-range PCR simply indicate that one of the members of a particular group is present. Quantitative data (i.e., organism load) may also be obtained if quantitative standards are included in the amplification run and a standard curve has been produced. Although quantitative data may be of use, much of the strength of broad-range nucleic acid amplification assays comes from the information that may be derived from the amplified product (i.e., the amplicon) through postamplification analysis. There are many methods of postamplification analysis, but only a few will be briefly discussed here, given the scope of the present article. These are postamplification melt curve analysis, reverse hybridization, DNA sequencing (traditional and pyrosequencing), and solid- and liquid-phase microarray analysis.
It seems that advances in laboratory diagnostics more commonly occur quite rapidly in conjunction with technological advances (i.e., as in Stephen J. Gould's theory of punctuated equilibrium), rather than occurring though the accumulation of small advances over a long period (i.e., a more Darwinian-type change) [55, 56]. The history of the use of PCR in clinical laboratories supports this notion. Although PCR and other nucleic acid amplification technologies (note: hereafter, “PCR” will be used generically to encompass all methods of nucleic acid amplification) have penetrated into the larger diagnostic and university laboratories for the detection of microorganisms that are usually difficult to culture, they have not until recently become readily available in small, perhaps less molecularly savvy laboratories or for the detection of more common pathogens. This is because traditional PCR is cumbersome and slow, especially in light of the time and process necessary to specifically determine the nature of the amplified product by Southern blot analysis. However, all this changed with the advent of real-time PCR, also known as “homogeneous, rapid cycle nucleic acid amplification.”
The advent of real-time PCR awaited complementary advances in engineering and chemistry. Engineers devised ways of decreasing the PCR cycling time. This has been accomplished through the use of heated and cooled air to achieve more rapid heating and cooling, respectively. This is in contrast to traditional block cycle heating, wherein, traditionally, a metal or solid block was heated and cooled to change the temperatures of the reaction vessels contained in the block to achieve those temperatures necessary for the PCR. This latter method, although effective, is naturally slower and reflects the physical differences between conduction and convection heating. Next, the engineers needed to develop sufficient optical systems for these instruments, to both introduce a beam of light of a particular wavelength and detect minute changes in light emitted from the solution within the reaction vessel. Fortunately, many of these technical issues had been addressed by scientists who used laser and fluorescent dye technology for various applications, such as flow cytometry. This afforded the adaptation of such technology to real-time PCR platforms.
The advances in chemistry that made real-time PCR possible were also significant, and modifications of these chemical reactions continue today. The challenge was to devise a way that amplification and detection reactions could occur within the same reaction vessel without necessitating the opening of the vessel and the manipulation of the amplified product. Maintaining a closed system is important in a clinical microbiology laboratory, because chemical contamination of the environment with amplicon could lead to false-positive reactions . This was achieved through the discovery and construction of molecules that were nonfluorescent in a nonhybridized state but were able to generate a fluorescent signal when both bound to their DNA target and excited by the appropriate electromagnetic energy. The signal generated may be nonspecific, demonstrating only the presence of double-stranded DNA (i.e., evidence that the PCR has occurred), or specific, which is achieved through the hybridization of specific, fluorescently labeled oligonucleotide probes. An example of a nonspecific probe is SYBR Green dye I, which binds to the minor groove of double-stranded DNA. The signal derived from this molecule, while PCR is under way (thus, in real time), demonstrates to the user that amplification is occurring. Whether this amplification is specific cannot be determined at this point in the reaction (see the melt curve analysis below). Conversely, if one uses a specific, fluorescently labeled oligonucleotide probe, then the signal that is detected during the PCR is evidence that the target DNA molecule is present, because the oligonucleotide probe hybridizes to its target through specific Watson and Crick base-pairing rules. There are many different types of oligonucleotide probes available. The 3 most commonly used types are the hydrolysis or “TaqMan” probes, molecular beacons, and fluorescence resonance energy transfer (FRET) probes. A positive real-time PCR result for any of these probes is demonstrated by an exponential increase in fluorescence with respect to amplification time or cycle number (figure 1). Specimens that do not contain the target molecule (i.e., that are negative) will demonstrate minimal or no change in fluorescence. The qualitative information derived from such reactions is useful for demonstrating the presence or absence of a microorganism of interest. For example, assays have been devised to address the presence or absence of Legionella pneumophila by use of real-time PCR. Such an assay may offer times to results that are similar to those of direct immunofluorescence antigen detection, but the real-time PCR has been shown to be more sensitive than direct immunofluorescence antigen detection; additionally, the same assay was described as being more rapid and possibly more sensitive than culture while retaining a specificity equivalent to that of culture .
The specificity of a PCR assay is determined by the target DNA sequence under evaluation, the sequence of the oligonucleotide probe, similar sequences that may exist elsewhere in nature, and the intentions of the assay designer. Many assays are species specific, but genus-generic assays (i.e., a more broad-range PCR) may also be designed. These would be useful when it is important to detect all members of a group that may be divergent in nature. For example, one may use a species-specific PCR for Salmonella serotype Typhi, to detect that particular serotype, or a broad-range Salmonella assay, to detect all of the members of that genus .
In addition to the detection of microbial pathogens, these applications may be used to detect genetic elements that encode for mechanisms of antimicrobial resistance (e.g., the mecA gene), those that determine subtypes of such genes for epidemiologic purposes (e.g., the mecA cassette IV associated with community-acquired Staphylococcus aureus infection), and the detection of virulence factors (e.g., Panton-Valentine leukocidin). Although useful for the detection of certain mechanisms of resistance, nucleic acid amplification assays for more complicated and genetically diverse mechanisms of antimicrobial resistance (e.g., the extended-spectrum β-lactamases) have remained elusive and are not likely to replace traditional phenotypic susceptibility testing in the near future.
Melt curve analysis. Another feature of some of the chemical reactions used with real-time PCR is the ability to perform a postamplification melt curve analysis. This provides different information, depending on the type of probe used. If SYBR Green dye I or another nonspecific DNA binding dye is used, then the information derived is similar to that derived from gel electrophoresis. Larger amplicons are held together by more hydrogen bonds, whereas smaller amplicons, such as primer dimers, are held together by fewer hydrogen bonds. The postamplification melt curve derived using these types of DNA binding molecules demonstrates high melting temperatures (95°C–98°C) for large amplicons and lower temperatures for smaller amplicons and primer dimers. Some researchers have demonstrated the detection and differentiation of single nucleotide polymorphisms (SNPs) by use of high-resolution melt curves and nonspecific DNA binding dyes, but, for the most part, these uses remain experimental .
The FRET probes and another oligonucleotide probe called an eclipse probe are the types of specific oligonucleotide probes that provide the best postamplification melt curves. Foremost, a melt curve analysis after amplification provides information regarding the specificity of the reaction. A particular melting temperature is expected for the hybridization between the oligonucleotide probe and its target. The presence of the expected melt curve after amplification helps to assure the user that the results are correct. If, for example, the primers were degenerate and also amplified a related microorganism, the melt curve generated would have a lower temperature than expected, if nucleotide polymorphisms existed at the probe hybridization site. This molecular misidentification would not be detected in the instrument's “quantification” mode, and, if such a scenario occurred, the specimen would erroneously be deemed positive. However, if a postamplification melt curve analysis was performed, then the nucleotide mismatches between the probe and target sequences would be detected, because the probe could not hybridize to its target sequence with the optimal affinity and would “melt off” at a lower temperature. This technology is so sensitive that it is commonly used to determine SNPs in human genetic pathology (e.g., the SNP responsible for factor V Leiden) . Thus, postamplification melt curve analysis is a method whereby one may obtain additional data concerning the specific nature of the amplified DNA product after real-time PCR.
This difference in melt curves created by imperfect complementarity between probes and their target hybridization sites may be exploited to differentiate taxonomically related microorganisms to the species level. Assays have been designed to differentiate HSV type 1 and 2, the BK and JC polyoma viruses, human herpesvirus 6 types A and B, the commonly occurring Bartonella species, and M. tuberculosis from nontuberculous mycobacteria (figure 2), among others [53, 62]. This application is quite useful, but there is some evidence that there may be diminished sensitivity for the target that has the lower melting temperature (i.e., the imperfect complementarity) . Therefore, it would be prudent to not use this type of application if the target that had this profile caused serious disease (i.e., if there would be a significant sequela associated with a false-negative reaction) or if such a target organism is present in small quantities and may be missed because of the limited sensitivity. The approach we have taken with the mycobacteria has been to utilize this technology to differentiate mycobacteria that are detected either in the acid-fast smear or in histologic sections stained by the Ziehl-Neelsen method. In such an instance, it is not a question of “if” mycobacteria are present but, rather, a question of “which type of” mycobacteria is present, with the primary goal being the separation of M. tuberculosis from the nontuberculous mycobacteria (figure 2). The hybridization probes have 100% homology to M. tuberculosis, with the nontuberculous mycobacteria having a ⩾1-nt mismatch at the probe hybridization site. This design makes the assay most sensitive for the detection of the most important mycobacteria that we encounter in North American microbiology, M. tuberculosis. If acid-fast bacilli are seen in the smear or stained tissue section, but amplified product is not detected, then one may categorize the result as “quantity of mycobacteria below the limit of detection of this assay,” which is a more useful result than a false-negative result.
There are a variety of other methods of postamplification amplicon analysis that require the opening and manipulation of amplified PCR product. These are acceptable methods, but care must be taken to avoid the introduction of amplified product into the areas of master mix preparation or specimen preparation, to avoid false-positive reactions. The types of methods are too numerous to describe in an exhaustive manner in this article, so I have selected a few to highlight that either have or will likely have an impact in clinical medicine; these are reverse hybridization, DNA sequencing—including pyrosequencing—and microarray analysis, with a focus on limited microarrays of both the solid and liquid formats.
There is often a need to assess for or identify >1 microbial pathogen. It is more time- and cost-effective if these assessments can be performed simultaneously or in a single reaction than if they are performed in numerous monoplex amplification reactions. For example, it is necessary to assay for numerous high-risk HPVs to appropriately diagnose or exclude a high-risk HPV infection in a woman. It is preferable to perform a single assay to detect and/or differentiate these viruses rather than to perform >10 individual PCRs for the various high-risk HPV subtypes. Postamplification analyses are often employed when the number of results obtained exceeds the technical differentiation capabilities of multiplex reactions wherein each reaction contains a different fluorescent label, or when the number of results achieved may exceed the differentiation capabilities of a melt curve analysis.
Reverse hybridization. One of the most simple-to-perform methods of detecting a variety of pathogens, after either a multiplex or a broad-range amplification reaction, is through reverse hybridization. This technique is the opposite of Southern blot analysis in many aspects and is therefore termed “reverse.” Herein, a variety of probes are placed on a nitrocellulose membrane strip, and the amplicon rather than the probe is labeled. The amplicon is applied to the strip under appropriate hybridization conditions, and then wash steps are performed. If the amplicon hybridizes to one of the immobilized probes, it may be visualized through the application of a color development reagent. The position of the product on the strip reveals the nature of the pathogen. This technology has been used to detect and differentiate the commonly occurring mycobacteria (figure 3), clinically important fungi, HPV subtypes, hepatitis C virus genotypes, and certain resistance-associated mutations in HIV [64–67]. This technology is simple to use, but some applications have been limited because of the expense of the product. Occasionally, light bands of nonspecific hybridization are seen (i.e., “ghost bands”) that may confuse interpretation or suggest dual infections.
DNA sequencing: traditional (Sanger) and pyrosequencing. The sequencing of the amplicon, through traditional Sanger sequencing, is the ultimate in postamplification analysis. The entire composition of the amplified product is thereby determined. Although initially difficult to perform, this technology is now readily available to any molecular diagnostic laboratory. Advances in capillary and gel electrophoresis, as well as computer-assisted sequence analysis, afford the expanded use of this technology to many more laboratories. Traditionally, sequencing after RT-PCR is the method of choice for determining resistance-associated mutations in HIV, and 2 Food and Drug Administration-approved tests have been developed for this application. This technology is also commonly used to identify bacteria, fungi, and mycobacteria on the basis of sequences within the ribosomal gene complex [68–71]. Once the sequence of an unknown microorganism is determined, it may be queried against a genetic database that contains thousands of sequences. The results of a query are returned as a percentage match, which assists the molecular microbiologist with the identification of the microorganisms. There is clearly great promise for this approach, and commercial products (e.g., MicroSeq [Applied Biosystems]) and nonpublic databases (e.g., SmartGene) have been established. Full-length sequencing is particularly powerful when several nucleotide polymorphisms are needed for the characterization of a microbe and these are distant from one another within the amplicon (e.g., some of the mutations associated with HIV resistance). The drawbacks associated with this technology would include the cost of the equipment, reagents, and access to private databases; the need to analyze long lengths of sequenced DNA, which is time and labor consuming; and the inability to generate high-quality sequence just downstream of the sequencing primer hybridization site.
Another type of DNA sequencing that has been more recently described is sequencing by synthesis, or pyrosequencing [72, 73]. This technology uses proprietary chemical reactions including several enzymes that generate light when a nucleotide is incorporated into the strand of DNA that is being synthesized. The amount of this light is in proportion to the pyrophosphate released during nucleotide incorporation. For example, if 2 of the same nucleotides (e.g., ...GG) are incorporated, then the light generated will be twice that generated by the incorporation of a single nucleotide (e.g., ...G). The instrument records the dispensation of the nucleotides and the light generated. Nucleotides that are dispensed but not incorporated are hydrolyzed by apyrase, one of the enzymes in the mixture. Hereby, the DNA sequence is determined as it is being synthesized. The strengths of this technology are that it is user friendly, the reactions occur in the commonly used 96-well plate, and the sequences immediately downstream of the sequencing primer are readily determined, which is ideal for SNP analysis. The limitations of this approach include the cost of the instrumentation and reagents, its limited ability to appropriately determine the sequence of homopolymers (i.e., numerous nucleotides of the same type occurring one after another), and the length of high-quality DNA sequence that may be generated, which is <100 nt and frequently is <50 nt. As noted, this technology is particularly useful for SNP analysis, because the assessment of only a limited number of nucleotides is necessary. In addition, this technology, although perhaps not as discriminating as Sanger sequencing, may be used in the taxonomic categorization of clinically relevant microorganisms. For this application, it is necessary to determine the sequence of variable regions that contain unique or “signature” sequences for microorganisms within a group. This technology has been used to appropriately categorize mycobacteria and nocardiae into clinically important groups (e.g., M. tuberculosis complex) . In our experience at the Cleveland Clinic, this technique of identifying mycobacteria was less expensive than our traditional approach that utilized genetic probes and biochemical reactions (data not shown). It has been used to identify yeasts (figure 4), filamentous fungi, and a variety of bacteria [75, 76]. It has also been used to differentiate the BK and JC polyoma viruses, as well as HPV subtypes [77–79].
Regardless of the type of sequencing technology used, both the laboratorian performing the sequencing and the ordering clinician must be aware of the strengths and limitations of these assays. In addition, the identifications achieved using these methods are only as sound as the entries in the genetic databases that are being queried. Genetic database entries should be based on isolates that have been soundly identified using traditional methods (if feasible) and should be of good sequence quality. Unfortunately, this is not always the case. Finally, an adequate number of well-identified isolates from each species should be examined, so that normal genetic variation may be understood and incorporated into identification schema (i.e., some variation may reflect only strain variation, whereas other variation is taxonomically meaningful for the differentiation of species). The bottom line is that the sequence-based identification of microorganisms is only as good as the sequences in the genetic library used and the characterization of the strains from which these sequences were derived.
Microarray technology. Microarray technology has proven an important tool for discovery [80–82]. These platforms may be used to assay for numerous (i.e., hundreds) of targets simultaneously and have been a critical means of studying the expression of multiple genes simultaneously. This technology, regardless of its technical format, consists of numerous individual probe-target hybridization reactions that are assayed for simultaneously. Not surprisingly, this technology has been used to differentiate numerous microbial pathogens, study their expression profiles, and detect the genetic determinants of drug resistance [83–86]. For example, soon after its introduction, it was shown to be capable of differentiating medically important mycobacteria and detecting the most common causes of resistance to antimycobacterial drugs [87, 88]. Unfortunately, the limitations of this technology, such as cost, inconsistent reactions in some systems, and data management and interpretation, have surpassed its practical uses. From a practical standpoint, simply more data are generated from a single, large array than are necessary in clinical medicine, and the cost of the array is prohibitive for routine use. Fortunately, this is changing. One of the most promising areas in molecular diagnostics is the use of limited microarrays for the assessment of numerous genetic targets after either a multiplex reaction or a broad-range PCR.
A number of different microarray platforms have been described, including bioelectric arrays and liquid microarrays. As with any technology, there are strengths and limitations with each system. These systems used today have improved hybridization and reporting for the individual probe/target combinations (i.e., they have been made more reliable), and, perhaps most importantly, some have been scaled down from large platforms to smaller-sized versions that address clinically important issues that require multiple results. In clinical microbiology, a microarray with 20–30 hybridization sites would be sufficient to address the most common causes of bacteremia and fungemia (i.e., the most common causes of positive blood culture results), and similarly sized arrays could be used to identify the typical causes of invasive fungal and mycobacterial infections (figure 5). This approach is optimized with the inclusion of broad family or kingdom probes. For example, a Raoutella species is an uncommon cause of bacteremia, and genus- or species-level probe sites are not likely to be included in a “bacteremia” microarray. The presence of an uncommon bacterium such as this in the blood culture sample could be detected through the inclusion of a broad-range bacterial probe site on the microarray. Hybridization at this site, with the absence of hybridization at a species-determining site, would demonstrate the presence of a bacterium other than any of the more common causes of bacteremia that are included in the panel.
The possibility of producing disease-specific microassays also exists, because these formats may be used after broad-range PCR, multiplex PCR, or a combination of such assays. For example, one could envision a microarray that could distinguish between the various causes of meningoencephalitis. Such an assay may include a complex multiplex reaction that consisted of a broad-range bacterial PCR, a species-specific PCR for Cryptococcus neoformans, a broad-range RT-PCR for the enteroviruses, and a specific PCR for HSV. Although not exhaustive, such an array would, in a single postamplification assay, address all the most common etiologic agents causing meningoencephalitis. Furthermore, additional probe sites could be added, as needed. For example, a West Nile virus site could be added, if desired. Limited microarrays, from a business perspective, represent a potentially disruptive technology, because the widespread use of these could change the way that we practice microbiology. However, their use will depend on numerous factors, such as the development of useful products, cost, Food and Drug Administration approval, reimbursement, and, most importantly, their performance in the clinical laboratory and their impact on patient care.
In summary, molecular methods are changing the way that we practice laboratory medicine and thereby are affecting the practice of clinical medicine. New methods of direct hybridization, rapid-cycle homogeneous PCR, and a variety of postamplification methods of analysis are making the genetic-based identification and characterization of pathogenic microorganisms more rapid and, in many instances, more accurate than ever before. However, for each of these technologies, the additions to health care costs must be weighed against the potential advantages of more rapid diagnostics. The techniques and the technologies are here. Well-controlled outcome studies are now needed to demonstrate the efficacy of these technologies.
Supplement sponsorship. This article was published as part of a supplement entitled “Annual Conference on Antimicrobial Resistance,” sponsored by the National Foundation for Infectious Diseases.
Potential conflicts of interest. G.W.P. is a scientific advisor for Luminex, AdvanDX, Roche, Cepheid, bioMerieux, and BD GeneOhm and receives royalties for licensing from Biotage and BD GeneOhm.