Next-Generation Antimicrobial Susceptibility Testing

Abstract

Antimicrobial resistance has emerged as one of the most-significant health care problems of the new millennium, and the clinical microbiology laboratory plays a central role in optimizing the therapeutic management of patients with infection. This minireview explores the potential value of innovative methods for antimicrobial susceptibility testing of microorganisms that could provide valuable alternatives to existing methodologies in the very near future.

TODAY'S GLOBAL AST LANDSCAPE

A limited number of methods for antimicrobial susceptibility testing (AST) of medically important microorganisms have survived the maturation of modern diagnostic clinical microbiology. Surprisingly, one of these is the disk diffusion method first published in 1966 (1) and the various iterations thereof. Another is broth microdilution (BMD) testing, which has attained reference standard status to which all other AST methods are currently compared during development, verification, validation, and clinical trials. As such, BMD displaced agar dilution testing, the past gold standard methodology.

The most important outcome of any AST is the rapid and reliable prediction of antimicrobial success in the treatment of infection. Currently, AST is typically accomplished using either classical manual methods or growth-dependent automated systems, such as the Becton, Dickinson Phoenix, the Siemens Micoscan WalkAway, or the bioMérieux Vitek 2, all of which are based on BMD testing. The major limitations of these methods include the requirement for relatively large numbers of viable organisms, complicated preanalytical processing, limited organism spectrum, analytical variability, time to results, and cost. Table 1 provides a cursory review of current and future technologies and their strengths and weaknesses.

Table 1.

Partial inventory of contemporary, near-future, and long-term alternative methodologies for antimicrobial susceptibility testing^a

Antimicrobial testing technology	Test principle	Needs more than 105 cells	POP or CA	Cost	Automatic or manual	Heteroresistance detection	Real MIC	Test time (h)
Currently in use
Agar dilution testing	Growth inhibition on solid medium with antibiotics	Y	CA	L	M	+	Y	>10
Automated testing (Vitek, Phoenix, MicroScan)	Monitoring of growth or substrate conversion in a dedicated machine using optics	Y	CA	L	A	±	Y/N	<10
Broth dilution testing	Growth inhibition in liquid medium with antibiotics	Y	CA	L	M/A	±	Y	>10
Chromogenic agars	Metabolic conversion of chromogenic compounds in agar medium	Y	CA	I	M	+	N	>10
Disk diffusion	Measurement of growth inhibition around an antibiotic-containing disk	Y	CA	L	M/A	+	Y/N	>10
Etest	Measurement of growth inhibition around a strip containing an antibiotic gradient.	Y	CA	L	M/A	+	Y	>10
Fluorescent live/dead staining	Microscopy of (non)permeable cells in the presence of fluorescent stains	N	CA	L	M	−	N	<1
PCR gene detection	DNA amplification	N	CA	I	A/M	−	N	<1
Real-time microscopy	Filming bacterial division at the single-cell level	N	CA	L	M	−	N	<1
Near-future alternatives
Calorimetrics	Detection of heat produced by stressed bacteria	N	POP	NK	A	−	N	<5
Cantilever technology	Weighing bacterial cells by changes in cantilever vibrations	N	POP	NK	A	±	Y	<5
FACS	Sizing and measuring differential fluorescence between living and dead cells	N	CA/POP	I	A	−	Y/N	<5
Magnetic bead spin	Changes in spin of beads in a magnetic field as a function of the no. of attached bacteria	Y	POP	NK	A	−	N	<5
MALDI-TOF MS	Detection of antibiotic degradation products	Y	POP	L	A	−	N	<5
Microdroplets	Monitoring of growth or substrate conversion in nanoliter droplets	N	POP	NK	A	±	Y/N	<5
Next-generation sequencing	Sequencing of all cellular DNA and RNA	N	CA/POP	H	A	−	N	>10
Long-term alternatives
Apoptosis markers	Detection of compounds produced upon programmed cell death	Y	POP	I	M/A	−	N	<1
Bacteriophage amplification	Detection of phage reproduction in living cells only	N	CA	I	M	−	N	<10
Colorimetric detection of cell respiration	Optical detection of substrate or indicator color change at active cell respiration	Y	POP	L	A	−	N	<1
Electronic noses	Direct detection of volatile organic compounds	Y	POP	L	M/A	−	Y	<1
Impedance measurements	Changes in electrical characteristics of suspension with living or dead cells	Y	POP	NK	A	−	N	<5
Infrared spectroscopy	Absorption characteristics of bacteria exposed to IR	N	CA/POP	I	A	±	N	<5
LC-ESI MS	Proteomics of living/dead cells and resistance proteins	Y	POP	H	A	−	N	<5
Metabolomics (including ROS)	Detection of changes in intracellular composition focused on small molecules	Y	POP	NK	A	−	N	<1
Microsound measurements	Measuring vibrational differences between living and dead cells	N	POP	NK	A	−	N	<5
NMR	Assessment of molecular composition of complex mixtures	Y	POP	NK	A	−	N	>10
Raman spectroscopy	Absorption characteristics of bacteria exposed to laser light	N	CA/POP	L	A	±	N	<5
RNA sequencing	Definition of gene expression differences by sequencing	N	POP	H	A	−	N	>10

^aFACS, fluorescence-activated cell sorting; MALDI-TOF MS, matrix-assisted laser desorption ionization–time of flight mass spectrometry; LC-ESI MS, liquid chromatography-electron spray ionization mass spectrometry; ROS, reactive oxygen species; NMR, nuclear magnetic resonance; Y, yes; N, no; POP, proof of principle; CA, commercially available; H, high; L, low; I, intermediate; NK, not known; +, detects heteroresistance; ±, may detect heteroresistance; −, fails to detect heteroresistance.

At present, nonphenotypic, mostly nucleic acid-based AST methods cannot detect all resistance markers, are expensive, and have not been widely adopted. Multiplex PCR detection of resistance determinants directly from positive blood cultures, however, has been shown to substantially reduce the time to clinically actionable results (2). Furthermore, digital PCR may allow for better quantification of target molecules present in starting material (3), and the refinement of aptamer technology (single-stranded short RNA or DNA molecules with antibody-like properties) may further facilitate nucleic acid diagnostics (4). Clearly, newer-generation, transcriptome, and whole-genome sequencing will provide near-future options to resistance prediction as databases mature. The per-strain assessment of detailed MICs for all relevant antimicrobials may be confounded by elevated degrees of genetic heterogeneity, as is, for instance, obvious among many Gram-negative bacterial species. This may still frustrate the genomic approach, but this would be the subject of an entire review by itself and will not be discussed here.

Novel options are available to supplant the existing toolbox, but the timing of such events is hard to foresee. As a disclaimer, we cannot provide a complete survey of all potential AST configurations but will try to highlight a number of phenotypic methods that provide insight into the wealth of the possibilities to come. Below, we will briefly describe six technologies that could represent competition for current reference standard methods—nucleic acid amplification and sequencing excluded.

NEAR-FUTURE ALTERNATIVES FOR ROUTINE AST

MS.

Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS), a powerful tool for the rapid identification of organisms with medical importance, may also prove to be of value as an AST method (5). Several approaches have been explored, including (i) documenting the activity of antibiotic-inactivating enzymes (e.g., β-lactamases), (ii) confirming the presence of a PCR product indicative of antimicrobial resistance (e.g., vanA, mecA, or NDM-1), and (iii) observing changes in the protein spectrum of an organism in the presence or absence of an antimicrobial agent that correlate with susceptibility changes.

As an example of the first, carbapenemase activity was detected in a variety of Gram-negative organisms using ertapenem as a substrate (6). A bacterial suspension was incubated with ertapenem, and serial samples were examined by MALDI-TOF MS. Organisms producing either NDM-1 or IMP-1 completely hydrolyzed ertapenem within 1 h. It should be noted that four distinct peaks were initially observed in the mass spectrum of the parent drug, which already included the inactive hydrolyzed form. Other carbapenemases (IMP-2, VIM-1, VIM-2, and KPC-2) hydrolyzed the drug more slowly. Similar data were generated using meropenem as a substrate for enzymes including NDM-1, VIM-1, KPC-2, KPC-3, OXA-48, and OXA-162 (7). The applicability of this method using other substrates (penicillin G, ampicillin, cefoxitin, and imipenem) was demonstrated with and without clavulanic acid as a means of distinguishing β-lactamase classes such as AmpC or TEM-1 (8). The possibility exists for multiplexing the assay using combinations of β-lactams and inhibitors, which would allow classification of extended-spectrum β-lactamases as well.

The second approach was presented in a study that compared two methods of single-nucleotide polymorphism (SNP) analysis for epidemiological typing of 147 strains of methicillin-resistant Staphylococcus aureus (MRSA) at 16 distinct loci (9). The predicate method of analysis was a real-time SYBR green PCR assay. The comparator method employed the Sequenom MassARRAY iPLEX SNP typing platform (Sequenom, Brisbane, Australia), which combines multiplexed single-base extension PCR with MALDI-TOF MS of amplicons to determine the location of SNPs. Both methods proved comparable, and the mecA PCR amplicon was successfully identified by MALDI-TOF MS for all 147 strains. A combination of a primer extension (PEX) reaction with MALDI-TOF MS also led to the detection of ganciclovir resistance mutations in cytomegalovirus (CMV) among viremic heart transplant patients (10). Compared to a combination of real-time PCR and Sanger sequencing, the PEX/MALDI-TOF MS method disclosed resistance mutations earlier without loss of specificity. Similar analyses of PCR-generated amplicons are the basis for resistance detection using the more-advanced electrospray ionization mass spectrometry (PCR/ESI-MS). In one study, the quinolone resistance-determining regions of parC and gyrA of multidrug-resistant strains of Acinetobacter spp. were identified with adequate correlation to BMD testing (11).

Finally, there are a number of examples of MALDI-TOF MS being used to highlight the effects of antimicrobial agents on the protein spectral profile of susceptible organisms. Comparisons of the profiles of Candida albicans grown in the presence of increasing concentrations of fluconazole led to the formulation of a minimal profile change concentration (MPCC) that was defined as the lowest concentration of the drug at which a change in the profile could be documented (12). The authors found a very high concordance between the MPCC and the MIC values obtained by the CLSI broth-based reference method. Similarly, MALDI-TOF MS was used to assess caspofungin resistance secondary to fks mutations in 34 Candida species and 10 Aspergillus isolates (13). Strains were exposed to increasing concentrations of caspofungin in a BMD format, along with a drug-free control well, and incubated for 15 h prior to MALDI-TOF MS. For each drug concentration, an MPCC was calculated for each strain. This group found 100% essential agreement for all of the isolates using CLSI breakpoints for MIC or minimal effective concentration (MEC). Only two Candida isolates were incorrectly interpreted as nonsusceptible, generating a categorical agreement of 94.1%.

FC.

Flow cytometry (FC) permits changes in the morphology, physiological and metabolic activity, and viability of microorganisms to be followed after exposure to antibiotics. Through a process of staining with nucleic acid dyes that do not permeate the cell walls of healthy organisms, the proportion of cells in a dying or dead state (and everything in between) can be rapidly assessed by examining emission spectra after the cells pass individually through a flow channel and when the dye is excited by a laser (14). In early studies, a primitive flow cytometer constructed from a fluorescence microscope was used to assess cellular morphology or DNA after bacteria were exposed to antimicrobial agents. It was concluded that the effects of antimicrobial treatment could be detected within a few hours, suggesting a promising application for AST. The earlier studies were also useful for elucidating various dye/fixation combinations that would better differentiate the state of cell viability by FC (15). In 1997, propidium iodide was used to differentiate live/dead C. albicans cells treated with amphotericin B or fluconazole, which could then be rapidly and sensitively quantified by FC (16). Similar assays were developed for the echinocandins, caspofungin, and additional azoles (17). The results of these AST strategies were 96 to 99% concordant with other forms of AST (18). The overall duration of AST could be reduced from overnight incubations to 1 to 2 h. Also in 1997, bis(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4[3]) was used to visualize anionic membrane potential changes using fluorescence, which proved to be very adequate for AST of Escherichia coli (19). Tests of organisms causing urinary tract infection showed 94% agreement between classical disk diffusion testing and DiBAC4[3]-FC testing (20). FC AST was also described for Mycobacterium tuberculosis. Pyrizinamide susceptibility testing by FC was 93% concordant with the Bactec MGIT assay, and the former was conclusive within 24 h (21). For Yersinia spp. also, significantly faster testing was achieved using FC (22). FC AST was at least 20% faster than classical methods for E. coli, Pseudomonas aeruginosa, and S. aureus, and extended-spectrum β-lactamases could be reliably detected by FC in 1 to 2 h (23). Although significant advances in the design and performance of FC instrumentation have occurred, the technology has not yet emerged as a major player in the AST market, although commercial assays have been launched. Some of the perceived hurdles for the methodology are, among others, the ability to differentiate cellular damage caused by cidal versus static antibiotics, autofluorescence of certain bacterial species, and the tremendous amount of work required for verification/validation of the clinical database and the method itself.

Microbial cell weighing by vibrating cantilevers.

Cantilevers containing small canals which facilitate microbial passage can be made to vibrate continuously. When bacteria pass through, their weight (in the femtogram range) will cause a change in the frequency of cantilever movement (24). Less-dense cells will cause a different change than more-dense cells. When cells are treated with antimicrobial agents, their buoyant mass density changes, and this is measurable (25). The principle has been proven using ampicillin-resistant and -susceptible variants of Citrobacter rodentium. It was also shown that resuscitation of both phenotypes after osmotic shock in the presence or absence of ampicillin allows rapid differentiation in a reduced time span. Cantilevers can be multiplexed using nanotechnology such that multiple antibiotics in various concentrations could be tested for a single growing culture simultaneously.

A rapid biosensor for the detection of bacterial growth was developed using vibrating cantilevers containing a certain number of fixed but still viable bacteria (26). The change in resonance frequency as a function of the increasing mass on the cantilever forms the basis of the detection scheme. The calculated mass sensitivity according to the mechanical properties of the cantilever sensor is approximately 50 pg/Hz; this mass corresponds to about 100 E. coli cells. The sensor was able to detect active growth of E. coli cells within 1 h. The number of E. coli cells initially attached to the cantilever was on average 1,000 cells. Furthermore, the noninhibited growth of resistant cells could be documented within 2 h after the addition of antibiotics (27). Cantilever technology has also been used to assess vancomycin binding to cell wall precursors (28) and to measure the effects of colistin on P. aeruginosa (29).

Using suspended nanochannel resonators (SNRs), it was demonstrated that the measurement of bacterial mass in solution was even more precise. The SNR consisted of a cantilever with an embedded nanochannel. In addition, a new method was introduced that uses centrifugal force caused by vibration of the cantilever to trap particles at the free end of the SNR (30). This approach eliminates the intrinsic position-dependent error of the SNR and also improves the mass resolution by increasing the average “time of presence” for each particle. In addition, it would facilitate the continuous mass monitoring of a limited number of bacteria during (changing) exposure to antibiotics. Clearly, cantilever systems and precise weight measurements provide an interesting option for the development of multiplexed AST.

IMC.

Isothermal microcalorimetry (IMC) is a dynamic technique that allows the measurement of heat production either as a flow rate (μW/unit time) or as total accumulation over time (Joules/unit time) stemming from the metabolism of actively growing cells. Cumulative heat production generally parallels conventional growth curves in that the slope and shape of the accumulating heat production correspond with classical lag, log, and stationary phases. Maximum heat values represent the total number of cells produced over time (31). The method has been successfully adapted to small culture volumes (e.g., 1 to 3 ml) and can be used in conjunction with either solid or liquid culture medium (32). Using IMC, bacterial species identification from urine specimens was performed even at low bacterial counts within 3 h on the basis of dynamic heat flow patterns (33). When adapted to AST, measurements of heat production are made passively from sealed vials containing the organism, growth medium, and antimicrobial agents in doubling-dilution concentrations. The minimal heat inhibition concentration (MHIC) can be defined as the lowest drug concentration to either inhibit 50% of total heat production or result in a 50% reduction in heat flow rate, depending on the drug being tested. The process requires specific IMC instrumentation (e.g., TAM III; TA Instruments, New Castle, DE) for real-time measurement of heat generation with a detection limit in the 0.2-μW range.

The use of IMC for susceptibility testing is not new but has been successfully adapted for AST of bacteria (31), mycobacteria, including M. tuberculosis (32), and fungi (34, 35). The advantages of this technique include the following: (i) testing is conducted in sealed ampules, alleviating safety concerns when evaluating high-risk organisms, such as M. tuberculosis or fungal species; (ii) all monitoring during testing is passive and requires no manual manipulation of the test vials; and (iii) the completed analysis provides information about the maximum growth rate of the organism, static versus cidal activity of an agent, and delays to log-phase growth (extended lag phase) caused by the agent. The latter provides useful information concerning antimicrobial activity at subinhibitory concentrations of the drug and can help predict the actual MIC when inhibitory concentrations have not been developed (31). This is pronounced with fungal testing, where delays or an abbreviation of maximum heat flow can be readily appreciated with subinhibitory concentrations of antifungal agents (34, 35). Furthermore, IMC is not prone to subjective interpretations, such as trailing MBD wells or the determination of MECs based on morphological changes at a microscopic level. IMC correlates well with BMD testing and CLSI and/or EUCAST breakpoints when net heat production over time is used as a surrogate for growth. As a bonus, IMC can be used to evaluate the synergistic activity of antimicrobial combinations. Chip calorimetry is a monitoring tool for determining the physiological state of biofilms. Its potential use for the study of the effects of antibiotics was tested using an established model. The real-time monitoring potential of chip calorimetry was successfully demonstrated: a dosage of antibiotics initially increased the heat production rate, probably due to activity of energy-dependent resistance mechanisms (36). The subsequent reduction in heat production was probably due to the loss of activity and the death of the biofilm. This new analytical tool provided fast, quantitative, and mechanistic insights into the effects of antibiotics on biofilm activity. In short, the maximum bacterial growth rate and the start of the lag phase can be quantified by microcalorimetric technology in an affordable and sensitive manner. Hence, antibiotic-associated changes in these parameters can be efficiently measured as well.

Magnetic bead rotation.

When magnetic beads are brought into a revolving magnetic field, they self-assemble and assume a specific rotational spin. The frequency of rotation can be influenced by the binding of molecules, viruses, or bacteria. So, if the beads are equipped with a ligand that specifically captures bacterial cells, the rotation of the beads changes at the moment of capture. This change can be measured. If all beads in a broth culture are paired with one or two cells, something that can be accomplished by incubating ligand-modified beads with a diluted bacterial suspension followed by washing, they will resume a constant rotational frequency. As bacteria start to divide, the rotation frequency changes. If cell division is inhibited or blocked by antimicrobials due to susceptibility, the change is arrested. If the bacteria are resistant to the antimicrobial applied, again, a change in rotational frequency occurs. In this way, antimicrobial resistance can be detected and precisely quantified.

A growth-based antimicrobial susceptibility assay based on asynchronous magnetic bead rotation (AMBR) biosensors has been described (37). In this system, the effects of bacterial growth on the rotation and shape of a cluster of self-assembling magnetic microbeads in a rotating magnetic field can be observed over time. The rotational period (RP) is indirectly proportional to the drag coefficient of the surrounding medium containing bacteria, broth medium, and antimicrobial agents at various dilutions. The RP increases as organisms multiply and attach to the organism-specific antibody-coated beads or if the viscosity of the growth medium is altered. The addition of antimicrobial agents in increasing concentrations prevents an increase in RP over time. This process can be observed directly by illuminating the culture broth (in the form of a hanging drop) with a light-emitting diode (LED) or laser. The hanging-drop format also acts as a lens to create magnification of up to 100× such that the structure and rotational rate of the bead aggregates can be observed microscopically or when projected onto a detector. Serial 2-fold dilutions of streptomycin and gentamicin were added to Mueller-Hinton broth containing antibody-sensitized magnetic beads prebound to a standardized inoculum of E. coli. A hanging drop was formed with the mixture that was subjected to an oscillating magnetic field. Changes in the RP were observed microscopically over time and recorded. As expected, the RP increased relative to bacterial growth, with solutions containing higher concentrations of antibiotic demonstrating the lowest increases. This method can be miniaturized to nanoliter-volume water-in-oil droplets containing 50 or fewer bacterial cells per droplet (38). These changes substantially reduced the duration of a test.

Testing in microdroplets.

Micro- or nanodroplets can be used as small individual reaction wells. The droplets can be individually manipulated, and when they contain bacteria in sufficient numbers, the metabolic activity and viability of cells can be monitored. The development of this system became feasible once the emulsification process was refined and the long-term stability of the droplets could be guaranteed (39). A system consisting of 100-nl droplets containing 10³ bacteria per droplet and differing concentrations of antibiotics was developed (40). By following the droplets over time using epifluorescence, growth curves can be monitored at each drug concentration. More recently, droplets were prepared that contain a single bacterial cell (41). This technology can be miniaturized and easily multiplexed with respect to the number of antibiotics tested per organism The duration of testing can be as short as a single or a few bacterial replication cycles. Obviously, the assessment of technical reproducibility and development of adequate reference MIC databases will be necessary before this approach can be evaluated as a routine AST tool.

TECHNOLOGIES FOR APPLICATION IN THE MORE-DISTANT FUTURE

Several innovative AST methods have been designed and presented over the past decade (Table 1, Long-term alternatives), but most are far from introduction into routine clinical microbiology. Some, including bacteriophage-based AST, may be closer than the others. The bacteriophage-based strategy has been successfully adapted for testing of M. tuberculosis (42), with optimization achieved by including recombinant phages containing luciferase genes (43). It was shown that the luciferase assay could be performed at low cost in 2 days (44) and could be applied successfully in laboratories in developing countries (45). Unfortunately, this innovative approach has not been broadly accepted in diagnostic microbiology laboratories, probably due to contamination (46) or phage resistance issues. As stated above, as promising as some innovative methods may seem to be, broad acceptance has not yet been achieved.

Real-time microscopy is one of the innovative technologies that may be applicable to AST in the not-too-distant future. High-resolution camera-based systems have been commercialized, and these show a high degree of efficacy (47). This technology has been extended for use with microcolonies and serial photography in the presence or absence of antibiotics. Such technologies can generate a “1-day AST” for bacteria from positive blood culture bottles (48). This demonstrates that simple, seemingly old-fashioned test formats can still be adapted into systems that better serve the patient's needs.

Several modern technologies have been proposed as being possible long-term future alternatives to today's technologies in the clinical microbiology laboratory; these are briefly described in Table 1. Proof of principle has been demonstrated, and in some cases, the first trials of feasibility have been published. However, for all of these strategies, database development has yet to be completed, and until then, the prospects of these technologies remain nebulous.

CONCLUSIONS

Conventional AST has its diagnostic limitations: it is generally time-consuming, and actionable results have a tendency for late arrival. The current methods are very solid and well-respected and do generally have CE and FDA certification. Any new technology has to compete with current reference standards, and the method that shows significant improvements has yet to be published. As we described here, several technologies may be knocking on the door shortly.

Biography

Alex van Belkum graduated as a biologist at the University of Leiden, The Netherlands, in 1983. In 1988, Professor van Belkum did his Ph.D. examination in Biochemistry at the same university. In 1996, he received a second Ph.D. in Molecular Microbiology at the Erasmus University, Rotterdam, The Netherlands. Between 1988 and 1990, he was involved in malaria vaccine research as a research scientist at the Biomedical Primate Research Centre (BPRC-TNO), Department of Infectious Diseases, Rijswijk, The Netherlands. Between 1990 and 1991, he was the Head of the Department of Infectious Diseases, MedScand Ingeny B.V., Leiden, The Netherlands, after which he joined the Department of Molecular Biology, Diagnostic Centre (SSDZ), Delft, The Netherlands (1991 to 1994), as a staff member. In both positions, his focus was on the development of molecular tests for the detection and characterization of infectious pathogens. From 1994 until 2010, he was a staff member at the Erasmus University Medical Center Rotterdam (EMCR), Department of Medical Microbiology & Infectious Diseases, Rotterdam, The Netherlands. Between 2002 and 2010, van Belkum was the head of the Unit for Research and Development. Since 2003, he has been a Professor of Molecular Microbiology at Erasmus MC. From 2010 to 2011, he worked for bioMérieux as R&D Director in the La Balme Microbiology Unit. In 2011, Dr. van Belkum became the Corporate Vice President for Microbiology R&D at bioMérieux (La Balme les Grottes, France). In his current position at bioMérieux, he heads an international team of microbiology researchers in the field of in vitro diagnostics of bacterial diseases. Professor van Belkum has authored or coauthored more than 440 peer-reviewed publications, 100 chapters in books, and a variety of editorials, letters, etc. Dr. van Belkum is Editor in Chief of the European Journal of Clinical Microbiology and Infectious Diseases.