Profile cover photo
Profile photo
Gail Rosen
Gail's posts

Post has attachment
We usually think of antibiotic resistance as something that occurs on the level of the individual bacterial cell – if that bacterium has a trick (like a pump that pushes antibiotics out or a molecular product that disables them), then that cell will survive treatment to reproduce, generating a population of resistant bacteria.
Some bacteria, though, display a trait known as heteroresistance – they look susceptible in traditional tests but resist killing in “the wild”. Now we know why: some bacteria are good neighbors, and not just to their own kind.
The authors looked at Burkholderia cenocepacia, a soil-dwelling bacterium that sometimes causes opportunistic infections in vulnerable, immunocompromised people. When they exposed B. cenocepacia to different levels of several antibiotics, they made two important observations: a) some bacteria can survive higher levels of antibiotic exposure than others and b) the less resistant bacteria become more resistant when they are grown alongside their tougher siblings. They then refined the test by growing two B. cenocepacia strains known to be missing different genes that are important in antibiotic resistance. Usually, these bacteria are killed by exposure to antibiotics. In culture with the resistant strains, though, they survived. 
Other species were also protected by the resistant bacteria. Strains of Pseudomonas aeruginosa and Escherichia coli that normally die on exposure to the antibiotic they tested grew just fine in the mixed culture, even when there were 100 susceptible P. aeruginosa to every one resistant B. cenocepacia. This finding is illustrated in Figure 2, where you can see survival of P. aeruginosa without antibiotics alone and in culture with (first and second bars) and then with antibiotics alone (third bar; no survival) and in culture with B. cenocepacia (not as good as growth without antibiotics, but definitely better than nothing). 
From these results, the authors thought the mechanism of resistance was probably related to something the bacteria were releasing into their environment, so they filtered and compared the products in the culture media of treated and untreated bacteria.  Predictably, they found a ton of molecules related to antibiotic resistance, like BCAM2827, which stabilizes bacterial membranes against molecular assaults; flagellin, which is needed to rebuild the bacterial tails (used for mobility) that the antibiotic attacks; proteins used to import specific amino acids (protein building blocks); and YCEL, the function of which was previously unknown. 
The amino acids targeted by the import proteins were probably used to make a delightfully named (and horrific-smelling) molecule called putrescine, which increased survival of antibiotic treatment when it was added to the culture medium. Indeed, when the authors tested the resistance of bacteria missing a gene needed to make putrescine from those components, they found that these bacteria were susceptible to antibiotic treatment and unable to protect their neighbors. Putrescine works by binding to the outside of the bacteria in the same spot targeted by the antibiotic, blocking access and bumping out antibiotic that had already bound. It is probably the primary mechanism of resistance. 
YCEL, on the other hand, binds the antibiotic directly to keep it from reaching the bacteria. Deletion of the gene encoding YCEL led to deceased resistance, and when the authors attached fluorescent tags to the antibiotic, they were able to watch the binding happen. 
The clinical implications here are clear – just one strain of antibiotic-resistant bacteria can interfere with antibiotic treatment, even if the intended target is a different species. By understanding the underlying mechanisms, though, we can see opportunities to undermine them. For instance, the addition of drugs that inhibit putrescine synthesis to an antibiotic treatment regime could render the resistant bacteria susceptible again. Heteroresistance does, however, suggest potential problems with the increasing interest in testing bacteria for resistance before choosing an appropriate antibiotic treatment, as the target bacterium may not tell the whole story on its own.

Post has attachment
Bacteria have immune systems, too (bet you didn’t know that)! And they’re Lamarkian!
These systems protect against invading phages (the alien-invader-looking viruses that infect bacteria), and plasmids (circular, externally-acquired strands of DNA) carrying fitness-decreasing code. And just like us, bacteria and their parasites are engaged in an “arms race” of infection and immunity.
Nearly half of all bacteria (and 90% of archaea) whose genomes have been sequenced to date have regions called Clustered, Regularly Interspaced Short Palindromic Repeats (CRISPR). The name says it all, really – they’re just very short pieces of genetic code (24-47 letters) that form palindromes, repeated over and over again (2-250 times!), like so:
The CRISPRs don’t code for anything; however, next to these sequences are CRISPR-associated (Cas) genes, which do encode proteins. More than 40 different Cas genes have been found in various bacteria, but they all seem to share a basic set of functions:
1) One set of Cas proteins recognizes foreign DNA and, through an as-yet-unknown mechanism, grabs about 30 bases (letters) of it. The new sequence is than inserted between the CRISPRs: ATAGCCT…TCCGATA CCGAGT…ATTCG ATAGCCT…TCCGATA ATAGCCT…TCCGATA ATAGCCT…TCCGATA. This new sequence is now known as a “spacer”. We think, but we're not sure, that the CRISPR and its spacer are constitutively transcribed (that is, many RNA copies are being made all the time, rather than waiting for a signal to go). We don't know for sure whether they're all transcribed at once, or in sets, or one at a time, or whether it varies between bacteria or environments.
2) If that virus or plasmid ever tries to invade again, another Cas cuts out the piece containing the spacer from the RNA copies.
3) RNA is single-stranded, meaning in part that it is free to pair up with any exposed matching sequence (A to T and G to C). The spacer RNA pairs up with the matching sequence on the invading DNA (analogous to an antibody’s action) to mark it for degradation.*
4) When a new invasion occurs, the new spacer gets inserted between the next set of CRISPRs. We don’t know how many the bacteria can hold in total, but it’s a lot.
5) When the bacterium “reproduces” by copying its DNA and dividing, the CRISPRs are copied, too, so this immune system is inherited.
The bacteria avoid targeting their own spacers – they recognize the invading DNA by the flanking sequences (which won’t be CRISPRs).
Some crazy stuff going on here. First, this is basically Lamarkian evolution. Remember the guy who hypothesized that giraffes stretched their necks out to reach leaves and then had babies with long necks? Obviously (now), he was wrong; that's not how giraffe evolution worked. But these bacteria are acquiring traits in response to their environment and then passing them on, just like those hypothetical giraffes.
My favorite thing about CRISPRs is that the spacers are basically a written record of the host/pathogen arms race. For this reason, CRISPR bacteria and their viruses would be a great system to study for answers to basic questions about co-evolution between hosts and pathogens. Because they are highly variable and heritable, they could also provide a sensitive way of tracing bacterial "bloodlines" for classification.
CRISPRs are, however, a relatively new discovery, and much remains unknown.
How much do they cost? First, longer sequences take longer to transcribe, so CRISPR bacteria may grow slower than non-immune bacteria. Also, long repeats tend to cause copying errors, so longer CRISPRs could also result in more mutations that could be harmful to the cell.
How fast do they develop? We also don't know how fast bacteria can acquire CRISPRs -- clearly, in an invaded population of bacteria, some die before acquiring immunity.
How long do they keep working? Because we don't know whether the oldest CRISPRs are transcribed, we don't know how long immunity lasts. Also, just like the flu virus changes its coat to sneak by the human immune system, phages are under evolutionary pressure to change up their CRISPR-targeted sequences so that they can invade.

Coming up: math! A model system describes when CRISPRs are worth the trouble.

*Note for biologist readers – this mechanism is analogous to RNAi, but the two are not in fact homologous/related.

Post has attachment
How soon would antibiotic resistance go away if we stopped using antibiotics?
In about five years.
The usual argument goes something like this: antibiotic resistance is energetically expensive. Resistant bacteria put a lot of resources into building and maintaining the machinery that allows them to survive exposure to chemicals that would otherwise kill them. If we took away the antibiotics, these bacteria would be outcompeted because they’d be devoting energy and resources to useless projects. Like a moon base.
This phenomenon has been seen in the lab; logistically and ethically, however, it is much more difficult to observe in "the wild".
These authors recognized an opportunity to watch antibiotic resistance come and go in a real-world setting. Twice a year, for three years, almost everyone in 40 villages in Ethiopia was offered a dose of azithromycin, an antibiotic used to treat trachoma (a bacterial eye infection that is a major cause of blindness). The treatment was part of a study on eliminating trachoma, and the authors of this paper got permission from the people in the villages to piggyback. The villagers are a good group to track resistance in because their access to antibiotics is (unfortunately) very limited, so the researchers knew when they were treated last.

The authors took nasal swabs from 15 children in each of these villages. They also took samples from kids in nearby villages where no antibiotic was offered to get a sense of what the bacteria were like at baseline (0.9% of bacteria resistant). They sampled again just after each treatment, and they continued sampling for two years after the treatment ended. They then tested Streptococcus pneumoniae, which causes disease but is not related to trachoma, from each sample for the two main mechanisms of resistance to azithromycin.

Two years into the study, after only four treatments, the proportion of bacteria that were resistant to azithromycin skyrocketed to 28.2%. After six treatments, resistance reached an incredible 76.8%. More than ¾ of the bacteria they tested were resistant to an antibiotic that had been used only six times in three years. Clearly, the resistant bacteria had a substantial advantage when the antibiotic was present, even when it was being used very infrequently (though universally).

After the trachoma study ended, however, the resistant strains dropped quickly in prevalence. Two years after the last treatment, resistance was 20.6%, lower than it had been two years in. These data were used to calculate the “cost” of resistance -- resistant bacteria reproduced at only about 86% of the rate of susceptible bacteria – and included in a mathematical model that accounted for variables like the number of individuals who were uninfected, infected with drug-susceptible strains and infected with resistant strains. Five years after treatment, there was less than a 5% chance of any resistant bacteria sticking around.

This is, of course, just one bacterial species and one antibiotic. Quite reasonably, the authors remind us that five years is a long time and that prevention of resistance is the best policy. They don’t make any particular suggestions, however. Clearly, trachoma treatment is not a case of frivolous or inappropriate use of antibiotics, but some possibilities include antibiotic cycling (described in an earlier post) and trachoma prevention by improved access to hygienic facilities and clean water.

Post has attachment
Zheng Lab - Bad Project (Lady Gaga parody)
Amazeballs Lady Gaga parody: "Caught in a Bad Project".
We just watched this in our lab. Alas, I think many of us know how this feels.

Post has attachment
US readers: If you haven't signed yet, please consider doing so. Show the Senate that we're paying attention.

Post has attachment
A single mutation could make a Clostridium difficile infection much more unpleasant.
C. diff., as it is called, is a hospital-acquired bacterium that causes severe diarrhea. It is resistant to many antibiotics, which means that it can completely take over the gut in people who have been taking antibiotics that destroy the normal gut flora. It’s so nasty and so stubborn that people are willing to take extraordinary, gross measures to treat it ( Serious C. diff. infections are getting progressively more common, and these authors tried to identify the mutation responsible. They didn’t succeed, but their results are still interesting and instructive.
These new strains reproduce faster and produce more of the bacterial toxins (toxin A and toxin B) than the old-time strains. Production of these toxins is controlled by proteins called TcdR and TcdC; TcdR is needed to make toxin, while TcdC limits the activity of TcdR by grabbing onto it and stopping it from acting.
The authors of this paper hypothesized that in the new, hypervirulent strains, a mutation in the gene encoding TcdC might make the protein non-functional, so production of the toxins would no longer be limited. Toxins A and B would be made continuously, rather than depending on a “go” signal.
The authors tested their hypothesis by genetically manipulating C. diff. to produce two strains that were completely identical, except in their ability to make functional TcdC. They started with a strain that lacked a functioning TcdC gene. They then constructed a plasmid – a circular piece of mobile, transferable DNA – containing a whole TcdC gene and inserted it into some of their bacteria.
The strain that could not make working copies of this protein made significantly more toxin (16-32 times more) than the strain in which this protein was fully functional. When they were incubated with monkey cells in a petri dish, the toxin-producing strain also killed more cells faster. Hamsters infected with the strain lacking TcdC also got sick faster. When the authors transferred a working copy of TcdC into the highly virulent strain, toxin production plunged.
All of this evidence sounds pretty solid – but it doesn’t appear to hold up in the real world, where many factors can influence the virulence of a given strain. When the authors looked at a wide array of C. diff. strains taken from human patients, they found that whether they made good TcdC didn’t predict how much toxin they made or how fast they killed the monkey cells. For one thing, C. diff can also make a toxin called binary toxin, which is not controlled by TcdR and TcdC. Strains that make binary toxin are better at sticking to host cells, making them better at invading a new person and possibly more damaging once they’re in. Some strains are also better than others at forming the tough, highly infectious spores through which C. diff. can spread or cause relapse in people who were apparently cured.
These researchers have demonstrated a potentially important mechanism of virulence in an increasingly important pathogen. Even though this one mutation doesn’t consistently match up with the virulence seen “in the wild”, this work will still be useful as a reference point for future studies on why the new C. diff. strains are as nasty as they are.

Post has attachment
Bacteria can become resistant to antibiotics even without antibiotic exposure.
Triclosan is widely used in soap, mouthwash, cleaning products, etc. to kill bacteria. Although triclosan is not an antibiotic, its use might select for antibiotic resistance in bacteria that use the same mechanism to survive exposure to both chemicals. This study provides a closer look at the specific mechanism underlying resistance to triclosan in the opportunistic pathogen Stenotrophomonas maltophilia; this bacterium causes disease and death in people with compromised immune systems, like patients with cancer and cystic fibrosis.
One common form of antibiotic resistance uses a pump called a multidrug-resistance (MDR) efflux pump, which removes the drugs from the cell. Bacteria with more of these pumps are more resistant to antibiotics and biocides, and some bacteria start making more pumps in response to antibiotic exposure. In S. maltophilia, the most common pump type is called SmeDEF.
SmeDEF production is “opt-out”; it is controlled by a sequence of DNA upstream of the pump gene, a repressor called SmeT. The pumps don’t get made as long as the SmeT protein is around because SmeT binds to the promoter – the “go” signal for making SmeDEF – and halts production by blocking RNA polymerase. This system prevents the bacteria from making lots of energetically expensive pumps that they have no use for.
In bacteria exposed to triclosan, however, the structure of SmeT is altered. The protein can no longer prevent the production of SmeDEF pumps. SmeDEF are then constitutively produced, meaning that they are made at a constant and consistent rate. These authors used x-ray diffraction (a method that aims x-rays at molecules to determine the arrangement of the atoms) to observe SmeT being deformed by triclosan.
When it isn’t bound to anything, the assembly of helices that make up SmeT can wobble around relatively freely. Triclosan, which is tiny compared to the protein, grabs the ends of two of these loops, holding them together and stabilizing SmeT in a structure that is too wide at the binding site to fit the DNA. You can picture it as being like pinching a clothespin and then taping the ends together. SmeT floats off, and the bacterium starts producing SmeDEF pumps. The pumps remove the triclosan, as well as antibiotics like ciprofloxacin.
This study demonstrates definitively how antibiotic resistance can be selected for in the absence of antibiotics. It’s always good to have a reminder of something we’ve known for a long time – pathogens evolve, and the environment we have created is shaping their evolution (usually, to our disadvantage).

Post has attachment
Children with a common parasitic infection might be more vulnerable to malaria, but they might also be protected from one of the most dangerous complications of the disease.
Malaria and schistosomiasis are the two most significant parasites of humans in the developing world, affecting 500 million and 200 million people each year, respectively. In malaria-endemic regions, especially in Africa, parasitic co-infection is extremely common – indeed, polyparasitism is the rule, not the exception. School-age children are especially likely to have both worms and acute malaria infections.
Children are also most vulnerable to lethal complications of malaria. In malaria-endemic regions, most adults have been exposed their whole lives. The “memory” portion of the immune system is active, so they have low, controlled levels of parasites in their blood; children are more likely to have high parasite levels because their bodies have not yet “learned” to recognize the parasites and keep them under control.
One consequence is that children also have strong inflammatory reactions going all the time. This inflammation can result in a sometimes-fatal condition called cerebral malaria, in which the blood supply to the brain is blocked by parasitized red blood cells. Blockage occurs when inflamed membranes and sick cells become “sticky”. The cells pile up, cutting off the brain’s access to oxygenated blood. Coma often results, and without treatment, the child usually dies within a few days. In adults, the “memory” portion of the immune system has taken over malaria control and down-regulated that initial inflammatory response, so they don’t tend to get cerebral malaria.
Schistosomes also dampen that inflammatory response. They set up house permanently in the blood vessels next to your bladder or intestines, so they’ve evolved proteins that actively dial down attacks by the immune system.
The consequences of coinfection are potentially very complicated, and studies following infected humans and experimentally infected mice have yielded some mixed results. In general, though, there are two main outcomes of coinfection:
1) More severe malaria. Because the worms weaken the first-line immune response, children with schistosomes have more malaria parasites in their blood than children with malaria alone. They’re also more prone to new infections with variant malaria strains, which can make it harder for their immune systems to get the disease under control. Most studies on the subject have found this effect (in humans and mice both). Higher parasitemia probably also makes these patients more infectious, one way in which molecular-level events can affect a whole disease system.
2) Protection against cerebral malaria. This dangerous condition results from inflammatory action by the immune system. By cranking down inflammation, low levels of worms may reduce the risk the infected person will “self-destruct” in responding to the malaria parasites. This effect is suggested indirectly by studies focused on the levels of specific immune factors and was demonstrated in mouse models. It has not, however, been directly observed in human studies, though one study of adults in Thailand found that a different worm species gave some protection. It may be that protection only occurs within a narrow window, probably in pre-adolescent children with low levels of established schistosomes.
Our immune systems are complicated. Within the human ecosystem, as in the ecosystem outside, species that share space will have various effects on each other. In this case, the practical conclusion is that control efforts shouldn’t treat these two important diseases as isolated problems. Malaria treatment campaigns will probably be more effective if they are paired with deworming treatments, which are cheap and easy to administer. And those directing deworming campaigns should be alert for unwanted “side effects” of eliminating the worms.
Grau GE, Taylor TE, Molyneux ME, Wirima JJ, Vassalli P, Hommel M, Lambert PH. (1989) Tumor necrosis factor and disease severity in children with falciparum malaria. N Engl J Med. Jun 15;320(24):1586–1591.
Nacher M, Gay F, Singhasivanon P, Krudsood S, Treeprasertsuk S, Mazier D, Vouldoukis I and Looareesuwan S. (2000). Ascaris lumbricoides infection is associated with protection from cerebral malaria. Parasite Immunol. 22:107-114

PS -- Graduate school applications are done, so I am returning to semi-regular posting status. Please direct any good wishes or gifts to the PhD admissions committees (!).

Post has attachment
You may have heard of New Delhi Metallo-β-lactamase (NDM-1), the gene behind the new ohcrapwe'reallgonnadie superbug from India. Bacteria carrying this gene were resistant to all but two antibiotics tested. This is how it works (fair warning: organic chemistry ahead, but I kept it simple).
The β-Lactam antibiotics include penicillin derivatives. They are the most widely-used antibiotics. They get their name from the β-Lactam ring, a square arrangement of a nitrogen atom and three carbons, with an oxygen atom sticking out of one carbon and a sulfur attached to another. They work by stopping the bacteria from building the tough, flexible bacterial overcoats called cell walls.They can do that because they resemble a protein that the bacteria use in assembling the cell wall closely enough that they can substitute for it; when the antibiotics bind irreversibly to unassembled cell wall subunits, the bacteria can't link the pieces together, and the whole process grinds to a halt. Meanwhile, the bacteria see the precursors building up, think they have plenty of material around to make new cell wall pieces, and start breaking down the existing wall to get ready for the replacement parts that will never come.
So far, so good. But, as we know, bacteria have developed defenses against this trickery. Most bacteria resistant to these antibiotics break down the antibiotic itself by breaking open the β-Lactam ring through hydrolysis: they break the bond between the carbon with the oxygen on it and the nitrogen by deploying a protein that binds temporarily to the antibiotic and introduces new bonds to an -OH group (on the carbon) and a hydrogen atom (on the nitrogen), effectively introducing a water molecule between the two. With this ring broken, the antibiotic's whole structure changes; it no longer resembles the bacterial protein, so it can't interfere in cell wall synthesis (see an illustration here:
Other resistant bugs (like MRSA) have new binding sites on their cell wall subunits that the antibiotics can't recognize, but bacteria with NDM-1 are members of the former group (hence, β-lactamase, something that breaks down β-lactams). For most bacteria in this category, we have a class of backup drugs called carbapenems. These drugs still have the β-Lactam ring, but the sulfur atom has been replaced with a carbon. Sulfur is more dynamic, reaction-wise, than carbon, so this change stabilizes the molecule against hydrolysis.
Somehow, though, NDM-1-positive bacteria still hydrolyze the ring. These authors created virtual 3D models of the NDM-1 structure, (using the gene sequence and the known structures of other β-Lactamases as a guide) to figure out how. As it turns out, the key is zinc. NDM-1 comes attached to two zinc ions, but it is otherwise very similar to the other β-lactamases, and it appears to use the same basic mechanism. These zinc ions carry a positive charge, meaning that they can attract electrons; because electrons are the basis of molecular bonds, the "tug" of these ions creates a very polar (and thus chemically active) water molecule. A zinc ion also forms a bond with the dangling nitrogen, stabilizing the opened β-Lactam ring so that it doesn't snap closed again. Figure 3 in this paper offers a very helpful illustration of these interactions.
To test their model, the researchers transferred the NDM-1 gene into harmless bacteria and introduced mutations. When mutations were introduced into the area of the gene encoding regions that interacted with the antibiotic, the bacteria became susceptible again. Other mutations didn't affect the enzyme's ability to deactivate antibiotics. From looking at their model, the researchers also suggest that colistin, the last remaining antibiotic defense against these bacteria, is probably just too big for NDM-1 to grab hold of.
This study may seem dry and technical, but it's important to know how antibiotic resistance works if we're interested in stopping it. For example, we might add compounds to the antibiotic that specifically inhibit the resistance mechanism (NB: we already do this, and bacteria are already resistant to at least one of these additives). Or we can use our understanding of the chemistry to track genes that could potentially come to underlie resistance mechanisms, which may alter the way we choose to treat particular infections or alert us to problems before they evolve. More obviously, if we know how resistance mechanisms work, we are better equipped to develop new antibiotics that existing bacteria are not resistant to.

Post has attachment
We may be able to enlist the mosquito immune system to fight malaria. When mosquitoes are infected with my favorite organism, the bacterium Wolbachia , they start producing proteins that are involved in the immune response. Between this effect and the other kickass things Wolbachia does, the bacterium may provide a novel mechanism for malaria control.
A strain of Wolbachia called wMelPop dramatically shortens mosquito lifespans, killing bugs that have fed on blood. In a mosquito species called Aedes aegypti, it induces the mosquito's immune system to attack the parasitic worms the insects carry. It also does wild things to mosquito reproduction, making some loving couples sterile with each other: uninfected females cannot produce fertile eggs with infected males, so infected females have more babies. Because the bacteria are maternally inherited (passing into the eggs from mom), Wolbachia can move into populations fast. Even though the bacterium is "bad" for the host, being uninfected is a lot worse if it means you can't find a mate. Wolbachia even manipulates the host immune system to make the mosquitoes leave it alone.
Consider what would happen if we could make this work for Anopheles gambiae, the mosquito that carries malaria. The parasites need about 9 days to incubate inside the mosquitoes; if few mosquitoes make it that long, we could see serious impacts on transmission of malaria. And because Wolbachia infection knocks down parasite numbers, surviving mosquitoes would be considerably less efficient vectors of disease. The authors of this study experimentally infected An. gambiae with two Wolbachia strains and found that the infection shortened the mosquitoes' lifespans, induced mating type incompatibility, and inhibited maturation of Plasmodium falciparum, the malaria parasite.
This Wolbachia strain is not naturally found in malaria mosquitoes, and it is not maternally inherited in this species, so it might take a while before infection becomes possible in the wild; however, this research exemplifies a new and creative approach to malaria control. If stable, self-perpetuating infections can be induced, a non-pathogenic Wolbachia species might even work better for malaria control, so that selection against the infection is less strong; the mosquitoes must live long enough to lay eggs if the parasite is to remain in the population.
Such approaches are badly needed. A child in Africa dies of malaria every 45 seconds. There is a vaccine in the last phase of trials, but it is not fully effective; even natural infection doesn't produce full immunity to malaria. Efforts to control these mosquitoes have been variably effective. They thrive in tiny, temporary puddles, so the first rain can bring the bugs back, and they don't lay eggs in ponds deep enough to hold larvae-eating fish. DEET is effective, but its impact on the surrounding ecology makes it an unsustainable choice, and, unlike Wolbachia, DEET is not heritable or self-perpetuating.
One of the coolest things about this study is that it implicitly recognizes the need for control on an ecological scale; malaria control projects that focus narrowly on the human side of the infection (e.g., vaccines or bednets) are not wrong, but they are missing half of the equation. Meanwhile, a species-specific infection that does not even cause host extinction could potentially control the vector population without dramatically changing the surrounding ecosystem, so that we are not forced to choose between children's lives now and their livelihoods later.
Wait while more posts are being loaded