Stanley Falkow is a Professor of Microbiology and Immunology and Lucy S. Tompkins is Professor of Medicine (Infectious Diseases and Geographic Medicine) and of Microbiology and Immunology at Stanford University School of Medicine, Stanford, Calif. Stanley Falkow is a Past President of ASM.

Links to Other ASM Pages:

Table of Contents
ASM News Issues

The Application of Molecular Biology to Medical Bacteriology Yesterday, Today, and Tomorrow

Over the years, clinical microbiologists have striven to find better and faster ways to identify infectious microbes

Stanley Falkow and Lucy S. Tompkins 

In the summer of 1952, one of us (S.F.) began his professional life as a medical technologist in a hospital bacteriology laboratory, which back then was a colorful, somewhat aromatic place to work. Specimens were plated on blood agar and eosin methylene blue plates (to be replaced by MacConkey agar a few years later). The first act of examining the culture plates was to quickly flip the plate lid open and sniff the growth, hoping for the telltale scent that heralded the identification of the offending microbe. Colony morphology and Gram staining brought one to the next level of identification, which was based on quick tube tests (like the coagulase test for staphylococcus) or disc tests (like the bacitracin test for group A streptococci or the optochin test for the pneumococcus). Many bacteriologists would flood cervical cultures with oxidase reagent looking for the purple-pink colonies of the oxidase-positive gonococcus; others dropped small amounts of bile salts on suspected pneumococci, looking to see if the colonies disappeared as they autolysed.

Early tests for antibiotic sensitivity were set up using discs impregnated with the relatively few antibiotics available at that time. Sensitivity was not measured in the current quantitative terms of minimal inhibitory concentration (MICs). Instead it was measured qualitatively as either sensitive, intermediate, or resistant depending on the relative size, or the complete absence, of a zone of inhibition. For many of us in that era, a chief joy was the enteric culture yielding a non-lactose-fermenting colony that required setting up a battery of fermentation tubes and media to detect special bacterial end products. The next morning's array of data brought us to books of tables where we compared the results to see if they were consistent with our secret hope that they were Salmonella or Shigella. Several afternoons a week were spent painstakingly examining acid-fast stains of sputum, spinal fluid, and other clinical material. The antibiotic revolution had not yet emptied the sanitariums, and there was much tuberculosis in small communities.

This era has long since passed, of course, and has been supplanted by miniaturized identification schemes read off in a binary-like identification code. A plethora of tests now exists to identify cultured bacteria, including the anaerobes, which we had ignored in 1955-65. The great revolution in medical bacteriology has been the technology of direct antigen detection in clinical material. Now the use of molecular biology methods to identify the nucleic acid of microbes directly in clinical material produces a more rapid result that may lead to a specific therapy. The molecular techniques currently in use include PCR amplification and molecular fingerprinting. These tests are wonderful and easily achieved in the test tube, although technical problems and economic concerns remain significant impediments to widespread usage.

Looking for Bacterial DNA

Until about 1960, the nature of bacterial DNA was not well understood. Watson and Crick may have solved its structure, but the biology of the molecule at that time was only in the realm of experimental laboratories and not in textbooks. During the period Salvador Luria called the "Golden Age of Molecular Biology," a number of fundamental discoveries were made, including the description of messenger RNA, the solving of the genetic code, the theoretical concepts of the operon and the episome, and the significance of the variation seen in bacterial DNA molecules.

Even after Julius Marmur presented the world with a simple reproducible way to extract high-molecular-weight DNA by winding alcohol-precipitated DNA on a stirring rod from deproteinized bacterial cell extracts, none of us appreciated the potential information present in the base composition of bacterial DNA. Melting curves of bacterial DNA and DNA banding properties derived by ultracentrifugation provided us with gross measurements at the level of the guanine-cytosine content of nucleic acid. We took comfort from the fact that the differences in bacterial base composition followed (more or less) the taxonomy we used in the clinical laboratory, but it was far from perfect. Many DNA samples from different organisms had no obvious difference in their overall DNA base composition.

The subsequent explosion in the methodology to study DNA using the technique of DNA-DNA and RNA-DNA hybridization extended the Watson-Crick model and permitted physical measurements of DNA composition. Reduced to practice by Ellis Bolton and Brian McCarthy, agar-based DNA hybridization studies between different enteric bacteria revealed that though some microbes shared a similar DNA base composition, the actual linear arrangement of bases along the strands was measurably different.

Naturally, when the clinical microbiologists heard that the molecular biologists were excited because E. coli and Salmonella DNA were clearly distinct, they thought that the molecular biologists were pretty naive. The bacteriologists already knew that this was the case. Nevertheless, with the refined hybridization methods and the exhaustive studies by people like Don Brenner at Walter Reed and then the Center for Disease Control and those of DeLey and Crosa in the early 1970s, the era of molecular bacteriology began producing the first precise molecular descriptions of bacterial species. The subsequent consequences to clinical microbiology are difficult to overemphasize. In a way, this is the first example of using genomics to understand the biology of microbes.

Cloning Genes and Using the Methods in the Clinical Laboratory-the Birth of Molecular Epidemiology

H. influenzae ampicillin-resistant plasmid

In the 1970s, many of us in the medical microbiology and clinical bacteriology fields became enchanted with plasmids. It was not just that antibiotic resistance had raised its ugly head, but we were beginning to truly understand the biological aspects of R-factors and their role in the biochemical basis of transmissible antibiotic resistance. We also discovered that many of the important features of medically important bacteria actually were carried on bacterial plasmids and not on the bacterial chromosome. Together with the appearance of antibiotic resistance in community-acquired infections like Haemophilus influenzae meningitis, these discoveries took on a new and ominous significance.

Plasmid analysis, not to mention fine-structure DNA characterization, was technologically difficult then. To characterize a plasmid, it was necessary to carefully extract high-molecular-weight DNA from a culture and then purify it through some physical separation, either classically by alkaline-sucrose gradient ultracentrifugation or by banding in a preparative CsCl density gradient. The estimate of the sedimentation coefficient by alkaline centrifugation or the electron microscopic measurement of the circular DNA spreads from the plasmid band of the CsCl gradient was used to estimate the molecular mass of the plasmid. In the early 1970s, the adoption of the cleared lysate methodology to partially purify plasmids away from the bacterial chromosome had made life somewhat easier, but it was still a labor-intensive procedure.

The discovery of DNA cloning in 1972-1973 brought forth a variety of new methods, particularly the use of restriction endonucleases and argarose-gel electrophoresis to study the cleavage products of the DNA. Historically, it is interesting to note that agarose gel electrophoresis initially was not considered for examining circular DNA; conventional wisdom held that circular plasmid molecules would not enter argarose gels because they were too large. However, in our Seattle lab in 1974, an accidental observation allowed us for the first time to identify, isolate, and even "type" plasmids using agarose gel electrophoresis. The simplicity, ease, and economy of the method when applied to the analysis of bacterial plasmids from clinical isolates led to the first phase of the field now called molecular epidemiology. Bacterial strains could be typed on the basis of the number, size, and restriction site polymorphisms of their plasmids. Now, the separation of plasmids using pulsed-field gel electrophoresis is a standard technique used in clinical epidemiology and is performed in many public health and routine clinical laboratories. The same is true of chromosomal DNA, as shown by Jim Kaper in 1978 when he typed different clinical isolates of Vibrio cholerae by restriction endonuclease polymorphism. We anticipate that these molecular methods will be supplanted by the more comprehensive application of the methods of bacterial genomics in the near future. In fact, some diagnostic laboratories are already performing DNA sequence analysis of viral and bacterial DNA targets.

If restriction analysis of plasmids and genomic nucleic acid had important implications for identifying the nature of strains causing global and isolated outbreaks of nosocomial disease, cloning of so-called virulence genes also had a major impact on our understanding of medically important bacteria. The comparison of commensal strains and virulent bacteria of the same or closely related species and genera gave rise to the understanding that the pathogens often possessed a constellation of genes that were distinct from nonpathogenic species. We had the hope then of using DNA-DNA hybridization to directly detect specific bacteria in clinical material by their genes. The approach had some success in the early 1980s when applied to in situ hybridization of viral agents in tissue, but initial application of DNA hybridization studies to identify clinically important bacteria did not succeed. Then the observation that enterotoxi- genic E. coli bacteria differed from ordinary coliforms by possessing plasmid genes encoding one or more enterotoxins led Steve Mosely to apply the methodology of DNA dot blots to directly identify enterotoxigenic strains from diarrheal stool samples from patients in Bangladesh. Although this approach has been considerably refined since those first attempts, the idea remains the same-use the specificity of DNA to distinguish between different species and even strains of bacteria.

In the next century, microbiologists likely will use DNA-based identification schemes based on genetic sequences of the hundreds of genes from virtually all known pathogens arrayed on disposable "chips" of some supporting matrix. Sample analysis will be automated and the readout provided within a very short period of time. Even today, hand-held, battery-operated devices are capable of using the magic of fluidics and microminiaturization to perform PCR analysis of food and environmental samples to detect specific bacteria. Applications to the clinical laboratory will not be very far behind. Within the next five years, we will have at least one completely sequenced genome of virtually every known bacterial pathogen. This information explosion, coupled with the ability to characterize nucleic acid "in silico," will have extraordinary impact on the clinical laboratory. Identification of the bacteria could move to the bedside.

The Years To Come-Does the Bell Toll for Medical Microbiology?

With the arrival of genomics we are not Cassandras foretelling the end of research in medical microbiology. Although there is clearly a dramatic need for new anti-infective agents and vaccines to cure and protect us from disease, the study of pathogenesis of medically important microbes must still continue. Genomics per se will not establish the relevance of genetic sequences nor the mechanisms at play in the host-pathogen interaction. There hardly can be any doubt that the clinical microbiology lab will be as different 40 years from now as it was 40 years ago. Indeed, it may not exist at all as a distinct entity. Perhaps future clinical microbiologists will suffer from the same nostalgia that we feel for the sights and smells of yesteryear (although we would not want to return to the same level of knowledge and sophistication). Nevertheless, the fact remains that for many (most) infectious diseases seen in the world, a definitive diagnosis is never established. Moreover, new pathogens emerge continuously. Although infectious diseases were declared conquered in the late 1960s, Louis Pasteur was correct: bacteria will always have the last word.

Now there is an increasing awareness and burgeoning experimental proof for an infectious origin of many diseases currently thought to be noninfectious in nature. We are not very advanced in understanding the precise nature of the interaction of the pathogen with its host during their initial encounter. For all of our immunologic sophistication, there is a considerable void in our knowledge of the steps between the initial encounter and the development of antibodies (protection not necessarily withstanding). We stand on the threshold of examining new facets of the host-parasite relationship. Already, we have seen the identification of the microbial factors that subvert existing host cell biochemical cascades so that the pathogen can breach the host's epithelial barrier. None of us thought a decade ago that pathogens actually could come to the surface of the host cell and efficiently inject bacterial proteins that interact with the host cell cytoskeleton, signal transduction pathways, and that even induce programmed host cell death. This new discipline, cellular microbiology, is the harbinger of a new era of using pathogenic microorganisms as biological probes of their animal hosts as much as using host cell infection to learn more about pathogenic bacteria.

In the future, we will be able to deliberately manipulate our immune system to respond in ways we wish, rather than following the dictate of the offending microbe. We have just begun to understand that not only does the invading organism overcome our defenses, it often directs the host response to the pathogen in a way designed to best suit the survival of the microbe rather than the survival of the host. This is not restricted to just those bacteria that infect mammals. The same principles apply to the study of plant-bacterial interaction, protozoan-bacterial interaction, and the myriad of other interactions that occur between microbial communities and the rest of the biosphere

We remain truly ignorant about the microbes that inhabit our bodies. It is estimated that the vast majority of the bacteria to which we are heirs remain uncultivated and unidentified. Our flora is not simply an accident of which species colonized first and fastest but rather reflects ongoing evolutionary adaptation. Even the final stages of our development, which take place ex utero, seem dependent upon the nature of the interaction between our microbial flora and our cells. For example, the microbial flora actually define our immune system and are critically involved in maturation of our gut epithelium. Just as the world of environmental bacteriology is in the throes of a revolution of understanding about microbial societies in every habitat, so too will the medical bacteriologist of the next century need to explore the community of microorganisms that inhabit our bodies.

In retrospect, it is always instructive in science to see how small increments in technology have such profound implications for theoretical and practical aspects of an entire field of endeavor for decades to come. We make fundamental observations about biology; we then seek to answer a rather simple question experimentally, but simple answers require a simple way to ask the question and to understand the answer. In some ways those interested in clinical and medical microbiology have been fortunate, because the developing methods have always been applied so quickly to the practical problem of disease. For the clinical microbiology laboratory, the future will become one of technology and informatics (in one form or another). For the medical microbiologist, it may be that we actually are beginning our excursion into the "Golden Age." In our judgment, not since the time of Pasteur and Koch have we been at the threshold of understanding so many new fundamental principles about medically important bacteria and host-parasite relationships. We embark now from a history of research largely restricted to the laboratory flask to a new frontier of research that probes the molecular biology, as well as the cellular and immunobiology of microbe-host interactions in real time in infected tissue.