Germ Warfare

Germ Warfare

Bacteria and viruses adapt for battle

by Sam Sternberg

Colorized transmission electron micrograph of T2 virus invading bacterial host, Escherichia coli. Copyright: Lee D. Simon

There is an oft-repeated statistic that the human body is made up of ten times more bacterial cells than actual human cells. Indeed, we live in a world dominated by invisible microbes. Perhaps less well known is the fact that viruses are even more abundant than bacteria in almost every ecological setting on this planet, posing a constant threat to survival of cellular life. As a result, organisms have evolved countless strategies to protect themselves from infection as part of an effective immune response. Just seeing an image of a cell being attacked by dozens of virus particles is enough to appreciate the powerful role that these ancient life forms have played in shaping evolutionary processes.

Viruses are biological entities comprising two components possessed by all forms of life: genetic material made from either deoxyribonucleic acid (DNA) or ribonucleic acid (RNA), and proteins that form a surrounding shell to encase and protect the genetic material. By most definitions, however, viruses are not alive because they cannot replicate on their own. Rather, viruses have been fine-tuned to hijack the cellular environment of other living organisms in order to access the biomolecules necessary for their survival and propagation. Influenza virus (cause of the common flu), for example, adheres to specific attachment points on the cell surface and, after becoming internalized, transports its genetic material into the cell’s nucleus. Host enzymes are then co-opted to generate all of the proteins that ultimately lead to viral replication and propagation. Luckily, humans have evolved multiple immune responses to ward off infection. Among these, the innate immune system is the first line of defense, acting non-specifically to recognize and respond to interfering pathogens. The adaptive immune system is far more complex: specific pathogens are remembered so that subsequent infections can be combated with even stronger defenses.

Bacteria, nature’s simplest living organisms, have also evolved a robust innate immune system, an impressive feat given that bacterial viruses, or “bacteriophages” (meaning “eaters of bacteria”), represent one of the most common biological entities on Earth. One liter of seawater, for example, contains an average of almost one billion distinct bacteriophage particles. Most bacterial immune defense strategies are fairly well understood and have more sophisticated equivalents in the innate immune systems of higher organisms. Five years ago, however, the scientific community was shocked by reports suggesting bacteria might also possess a highly sophisticated adaptive immune system, one that looked nothing like the immune responses of other organisms. Since then, ambitious research efforts at UC Berkeley and elsewhere have revealed how bacteria maintain rapidly evolving molecular “vaccination cards” by preserving small chunks of viral DNA in their own chromosomes.

The birth of a field

Click to enlarge. Credit: Asako Miyakawa and Sam Sternberg

The central dogma of molecular biology describes the flow of information inside the cell and can be summarized in its simplest form as “DNA makes RNA makes protein.” DNA is the genetic blueprint of the cell, but because cells usually contain only a single precious copy of DNA, important regions that encode proteins (genes) are first copied into short-lived RNA molecules that are chemically similar to DNA but disposable. Using the genetic code, the sequence information in these RNA molecules is then translated into proteins that ultimately carry out most of the cell’s functions. Because DNA contains all the instructions that govern cellular physiology, it has been commonly assumed that knowing the entire sequence of a cell’s DNA—its genome—would readily reveal all its inner workings.

In reality, genomes are far more enigmatic than originally thought. Many genes encode proteins whose functions cannot be straightforwardly predicted or experimentally revealed, and certain regions of the genome do not encode proteins at all and have completely unknown functions (See “Reading Between the Genes,” BSR Spring 2011). The discovery of bacterial adaptive immunity finds its origin in precisely such a region. As early as 1987, scientists first described highly repetitive sequence elements in the genome of the model bacterium Escherichia coli. After a number of independent studies documented similar repetitive sequence elements in diverse species of bacteria, researchers eventually recognized a common theme to these genomic regions and grouped them under an all-encompassing term: clustered regularly interspaced short palindromic repeats, or CRISPRs. As the name suggests, these regions contain long arrays of identical DNA sequences, or “repeats,” interrupted at regular intervals by intervening sequences, known as “spacers”. Yet the functional role of CRISPRs remained a source of confusion and speculation for more than a decade. Finally, in 2005, three labs independently discovered that many spacers were in fact identical in sequence to the genomic regions of known phages. This finding suggested the novel possibility that CRISPRs might be the molecular memory of a previously unknown bacterial immune system.

Yogurt-producing bacteria held the key. Researchers at Danisco, a prominent food ingredients company, were investigating viral defense strategies employed by the bacterium Streptococcus thermophilus, in the hopes of developing phage-resistant strains that would be less susceptible to infection in large-scale industrial food processes. Working in the laboratory, they discovered something remarkable: after infecting S. thermophilus cultures with one kind of phage, some cell strains developed robust immunity after integrating and thereby preserving a piece of the viral DNA in the CRISPR region of their own genome. A second distinct virus could still kill most of these cells, but again some cells persisted, and invariably, these phage-resistant strains had also inserted a new piece of viral DNA into the CRISPR. Somehow, sequence information provided by these repetitive genomic regions allowed the bacteria to both recognize and evade foreign transgressors that they had previously encountered.

But the CRISPR alone was not sufficient for this immune response; numerous proximal CRISPR-associated (Cas) genes were also required, indicating that proteins encoded by these genes might act together with the CRISPR to defend the bacteria against infection. Indeed, a second pioneering study published a year later revealed that CRISPRs are actually copied into RNA inside the cell, and that Cas proteins use these RNA molecules to target the destruction of any matching phage DNA. The implication of RNA in this process immediately suggested a plausible mechanism for CRISPR/Cas function. Since RNA can associate with complementary DNA sequences just as in the two strands of a DNA double helix, CRISPR-derived RNA molecules could be used to identify DNA sequences indicative of an infecting phage and then trigger its destruction. Remarkably, bacteria had achieved an elegant solution to the problem of pathogen detection that relied on nucleic acids instead of proteins, like the antibodies employed by the human adaptive immune system.

As with most breakthroughs, these seminal findings provoked more questions than answers and spawned an exciting new area of research in which UC Berkeley has figured prominently. In addition to hosting annual, international CRISPR conferences over the last five years, UC Berkeley Professors Jillian Banfield and Jennifer Doudna and their teams of researchers have been at the forefront of the blossoming CRISPR/Cas field. Working in such a nascent field is not an opportunity afforded many scientists, and the potential for discoveries has been vast.

CRISPRs highlight virus-host evolutionary dynamics

Professor Jillian Banfield on-site collecting samples in the Richmond mine at Iron Mountain in northern California. Sequencing the genetic material extracted from this natural environment enabled her and her colleagues to learn how CRISPRs and phages co-evolve. Credit: Eric Ramahatra

Much of what we know about bacteria-phage interactions has been learned from carefully controlled studies in laboratory settings. While invaluable, these experiments do not accurately reflect natural ecological habitats and so have limited relevance to real ecosystems, where a single strain of bacteria would never encounter a single type of phage as in the S. thermophilus experiment. In nature, innumerable microbial species are constantly battling diverse mixtures of infectious viruses. As CRISPRs were identified in laboratory strains of bacteria, researchers began wondering how CRISPRs and phages function and evolve in the wild.

Banfield (see also, “Manipulative Microbes,” 30), a faculty member in the Departments of Earth & Planetary Science and Environmental Science, Policy, & Management, has made substantial contributions to the field of geomicrobiology, the study of how natural environments and microorganisms interact. In the early 2000s, Banfield began pioneering a new approach in the study of microbial diversity in ecological niches. Traditional methods were limited by the requirement for laboratory isolation of genetically identical strains from natural environments, but Banfield and her team of researchers recognized that the advent of inexpensive and robust DNA sequencing technologies offered a powerful alternative. By recovering all genetic material contained in environmental samples and sequencing hundreds of thousands of random DNA fragments from this heterogeneous mixture, the genomes of numerous microorganisms and phages from the sample could be reconstructed in parallel with the help of powerful computer algorithms, an approach called metagenomics. In addition to shedding light on the species diversity in these samples, this technique could also reveal highly mutable regions within a single species since the sequenced pieces of DNA originate from non-identical strains. “Examining microbial populations rather than single isolates is particularly critical since CRISPR regions can be extremely diverse within a population,” explains Christine Sun, a PhD student in the Banfield laboratory. “In some cases, different cells from the same bacterial species have acquired completely unique sets of spacers to mediate CRISPR immunity.”

The ecological importance of CRISPR/Cas immune systems in shaping virus-host dynamics was immediately revealed by Banfield’s work. On the host side, bacterial CRISPR regions evolved on extremely short time scales, even faster than the natural accumulation of mutations that invariably occurs during DNA replication. By rapidly integrating new fragments of viral DNA into the CRISPR, bacteria were able to continuously mount an effective immune response against phages in their environment. Just as surprising was the finding that viruses similarly evolved an effective strategy to evade the CRISPR immune system. Through a process called homologous recombination, phages had shuffled large chunks of their genomes with other phages in the community to effectively rid themselves of the very sequences matching spacers in the CRISPR regions, enabling them to still trigger a successful infection. Collecting samples from different time points further enabled Banfield and her colleagues to observe—in real time—how both virus and host evolved over the course of months and years.

When asked about the relative advantages of her work as compared to more controlled laboratory experiments, Banfield stresses that “the phage-host interactions that rapidly generate diverse populations make sense in the context of complex natural systems. This diversity is ultimately required for the long-term survival of both populations.” More recently, Banfield has teamed up with Professor Wayne Getz of the Department of Environmental Science, Policy, & Management, to build a more complete picture of the evolutionary dynamics shaping CRISPRs using a combination of metagenomics and population-scale mathematical modeling. Taken together with other work, this study helps to explain the long-term advantages of maintaining such an elaborate immunological memory in microbial CRISPR regions.

Inner workings of CRISPR/Cas immune systems

The CRISPR/Cas immune response proceeds in three stages: fragments of viral DNA are inserted into CRISPR regions during adaptation, CRISPR RNAs are produced and processed during biogenesis, and complementary viral DNA sequences are targeted and destroyed during interference. Credit: Asako Miyakawa and Sam Sternberg

The work of Banfield and others established that CRISPRs evolve rapidly in natural ecosystems, and that these genomic regions also require CRISPR-associated genes to confer viral immunity. But Doudna, a joint appointee in the Departments of Chemistry and Molecular and Cell Biology, wanted to know what molecular mechanisms achieve this objective inside the cell. What biomolecules are involved, what do they do, and how do they work together to efficiently recognize and mount an immune response against infecting phages? Such questions are left to the domain of biochemists and molecular biologists, who rely on purified systems studied in vitro (literally “in glass,” i.e. test tube experiments) to tease out the biochemical steps that underlie a biological outcome.

Doudna has spent her career investigating the molecular structures and functions of various kinds of cellular RNA molecules. With early reports documenting a critical role for RNA in CRISPR/Cas immune systems, her lab was an ideal setting for studying the molecular details of this pathway, though Doudna credits Banfield for initially setting her on this path. “Jill brought the system to my attention six years ago, before there was any evidence for the molecular mechanisms involved. She suggested that my lab would be able to conduct biochemical and structural biology experiments to dissect the function of CRISPRs, and we’ve been having fun doing this ever since.”

To appreciate the contributions made by Doudna and her team of researchers (including this author), one must first be familiar with the three stages of CRISPR/Cas immune system function (see illustration below). In stage one—adaptation—an infected bacterium, faced with imminent death, pilfers a small piece of the phage DNA and integrates it into one end of the CRISPR region in order to quickly mount an immune response before it’s too late. During stage two—CRISPR RNA biogenesis—the entire CRISPR region is copied into long RNA molecules, which are precisely chopped into smaller pieces by an enzyme called a ribonuclease; each mature, processed CRISPR RNA molecule contains one unique spacer sequence encoded by the CRISPR. Finally, during stage three—interference—additional Cas proteins interact with each CRISPR RNA molecule to form large surveillance complexes that search for DNA sequences in the cell that are complementary to the CRISPR RNA. If such sequences are detected in foreign DNA such as that originating from a phage (the CRISPR DNA is excluded during this step), Cas enzymes degrade the targeted DNA and the phage is disabled.

The three-dimensional structure of a Cas ribonuclease bound to a piece of CRISPR RNA was determined using X-ray crystallography and revealed the atomic-level interactions that confer specificity. Hydrogen bonding interactions between the protein and RNA are indicated by dashed lines. Credit: Sam Sternberg

Having previously made significant advances in our understanding of ribonucleases involved in a process known as RNA interference, Doudna was naturally drawn to analogous Cas ribonucleases involved in CRISPR RNA biogenesis. In particular, she wanted to better understand how a single enzyme could selectively ‘cut’ the long CRISPR-derived RNA into smaller pieces, at a defined location, while leaving other cellular RNA molecules untouched. To answer this question, her team needed to visualize the biomolecules at high resolution, and because proteins and RNA are too small to see by light microscopy, they turned to a technique known as X-ray crystallography. Doudna and her team were the first to visualize detailed interactions between a Cas ribonuclease and CRISPR RNA by determining atomic-resolution, three-dimensional images of the protein-RNA complex. These images revealed that each repeat sequence of the long CRISPR RNA molecule folds into a characteristic “hairpin” shape that is perfectly complemented by the shape of the Cas ribonuclease, allowing the two biomolecules to form a tight binary complex. Furthermore, their data revealed that specificity for the correct RNA sequence is provided by a set of energetically favorable hydrogen bonds between the two molecules, ensuring that only this RNA sequence is cut by the ribonuclease. Subsequent work from Doudna and her colleagues revealed that similar strategies of RNA recognition are conserved in numerous other bacterial CRISPR/Cas immune systems, demonstrating the evolutionary importance of this protein–RNA interaction.

The work leading up to a high-resolution structure determined using X-ray crystallography is challenging and time-consuming, so a successful result often leads to considerable excitement. “It has been fantastic to see the three-dimensional structures of these biomolecules for the first time,” says Doudna. “I still get chills down my spine each time a discovery like this is made.” The first objective of the approach is to grow well-ordered, microscopic crystals of the biomolecule(s) of interest, something that occurs serendipitously and only through trial and error. After countless rounds of optimization, the crystals are exposed to high-energy X-ray beams such as those produced at the Advanced Light Source synchrotron at Lawrence Berkeley Laboratory, and interactions between the X-rays and electrons from the biomolecule lead to diffraction patterns that are collected by a camera. Finally, the three-dimensional position of each atom in the biomolecule is calculated from the raw data using advanced mathematics, producing the final model that describes the precise shape of the molecule.

To the frustration of many researchers, however, some biomolecules are simply intransigent to X-ray crystallography. Blake Wiedenheft, a former post-doctoral fellow in the Doudna laboratory who is now an assistant professor at Montana State University, found himself with this problem. “When I first started in the Doudna lab, one of my primary objectives was to get a high-resolution structure of the protein-RNA complex that mediates the interference stage of the CRISPR/Cas pathway. I threw everything but the kitchen sink into this project, but five years later we still have not found crystallization conditions suitable for generating well-behaved crystals.” So Wiedenheft took advantage of two other powerful techniques at his disposal in the Berkeley community that do not require crystal growth: small-angle X-ray scattering (SAXS) and electron microscopy (EM). Like X-ray crystallography, these approaches can provide three-dimensional images of biomolecules, though the lower resolution renders atomic-level detail invisible. “I initially determined the first three-dimensional structures of the protein-RNA complex using SAXS,” explained Wiedenheft. “Later, I teamed up with Gabe Lander, a post-doctoral fellow in the laboratory of Eva Nogales [also at UC Berkeley], and we froze these complexes in ice and analyzed them using cryogenic EM. The first images of the complex with this approach were stunning.” In addition to revealing how each protein component of the complex interacts with the CRISPR RNA, these structures, in conjunction with other experiments, showed how matching DNA sequences are recognized and marked for destruction by other Cas enzymes.

Collaboration is a must

The protein-RNA complex involved in the interference stage of CRISPR/Cas immunity was imaged with this electron microscope (center), which enables 100,000x magnification of cryogenically frozen samples. Involved in the research collaboration were, from left to right, Blake Wiedenheft, Eva Nogales, Gabe Lander, and Jennifer Doudna. Credit: Roy Kaltschmidt

Speaking of his research on CRISPR/Cas immune systems while at UC Berkeley, Wiedenheft is quick to stress the importance and influence of his colleagues. “The atmosphere of the Doudna lab was enthusiastic, intense, and stimulating, and my collaborations both within and outside of the lab were critical to my success.” Wiedenheft added that “the rapid progress in the CRISPR field is largely due to intellectual contributions from research groups with diverse backgrounds.” In fact, if you’ve spoken with anyone studying CRISPR/Cas immune systems, then you’ve undoubtedly heard about the importance of collaborative research. Evidently the stereotype about scientists working in isolation does not apply here.

David Páez, a post-doctoral fellow in the Banfield lab, couldn’t agree more. “It is very important to collaborate in our field. Combining information from bacterial isolates, metagenomics, mechanistic studies, and computational models is critical to future discovery.” Having arrived at UC Berkeley with a background in molecular biology, Páez challenged himself to switch gears and develop the skills necessary for metagenomic research. But instead of working only with natural ecosystems, Páez teamed up with some of the lead authors on the groundbreaking S. thermophilus study to apply a metagenomic approach to isolated cultures challenged with phage. Rather than focusing on species diversity, millions of DNA sequencing reads from individual S. thermophilus cells were instead used to learn about what regions from the phage genome are chosen for insertion into the bacterial CRISPR during an infection. “We are able to recover the CRISPR content of hundreds of thousands of cells with different immune system potentials on a daily basis using time-series experiments, to reveal underlying trends in spacer acquisition patterns,” explains Páez. By blending experimental approaches used by the Banfield lab and Danisco (now DuPont), Páez and his colleagues are deciphering the mechanistic details of the adaptation stage in living bacteria.

Although collaborations can and do happen between scientists separated by large distances, it certainly helps to have an international meeting to bring everyone together once a year. Rachel Haurwitz, a recently graduated PhD student from the Doudna laboratory, has a unique perspective on the topic; she is one of the few individuals who has attended all of the five annual CRISPR conferences at UC Berkeley to date. “It has been quite exciting to watch the CRISPR field grow rapidly over the past five years,” says Haurwitz. “Despite its growth, though, the field has remained cordial and tight-knit. It seems that nearly everyone is collaborating with someone else, and the CRISPR conferences have been an integral part of making those personal connections. I hope they will continue for many years to come.”

What the future holds

Far from peaking in interest, CRISPR/Cas immune systems continue to attract new scientists and students. “It’s easy to get excited about a field when it’s so young,” says Megan Hochstrasser, a new graduate student in the Doudna lab. “There are always fresh questions to ask, novel mechanisms to elucidate, and new techniques to try. Our understanding of CRISPR systems advances almost daily, and it’s great to feel like you’ve gotten in on the ground floor.” A brief survey of the literature also attests to the steady increase in academic research efforts, as the number of CRISPR-related publications has risen exponentially ever since its initial discovery.

The coming years will undoubtedly bring a deeper understanding of the CRISPR/Cas bacterial immune system, both in how it plays out in natural environments and in the molecular details of its function. But slowly, new initiatives are surfacing that aim to develop the knowledge from basic research for completely distinct purposes. “I think the field has naturally split in two directions,” says Haurwitz,  “focusing in parallel on basic aspects of the biology underlying CRISPRs and on potential applications of the CRISPR/Cas system.” As to the kinds of applications we can expect, Páez has some ideas. “Information from model systems will be used towards improving industrial processes [such as dairy cultures or microbial biofuel production], bacterial strain identification, and most probably in synthetic biology,” he says.

An illuminating example of a promising synthetic biology application can be found in a recent study published in the journal Science by Doudna’s group and colleagues at Umeå University in Sweden. While characterizing the CRISPR interference stage in a subset of bacteria, they stumbled upon a Cas enzyme that could be engineered to generate precise cuts in both strands of a DNA double helix at any desired position, as specified by the CRISPR RNA sequence. Scientists hoping to cure genetic diseases through site-specific genome editing have long sought designer enzymes with this functionality, but until now, every new gene target has required a complete re-engineering of the enzyme, an arduous task. In the case of the Cas enzyme, though, only the RNA sequence needs to be modified, an adjustment that can be made practically overnight. Time will tell if the full potential of this enzyme in genome editing is realized, but one thing is certain: no amount of foresight could have predicted that a bacterial immune system might offer so much biomedical potential.

Perhaps here—at the intersection of pure science unadulterated by utility, and engineering pursuits motivated by everyday problems—do we find research at its most exciting. As Haurwitz puts it, “neither of these research directions happens in a vacuum. Basic research feeds into the applications, and knowledge learned from the applications feeds back into the basic research.” Both ventures will surely keep UC Berkeley scientists engaged in this microbial battle for years to come.