Click for more "Microbes After Hours" videos
Here is a transcript of TWiM episode #49, "Grape-like clusters". Thanks to Frank Shinneman for transcription.
The transcript is also available as a pdf file - click here to download.
Hosts: Vincent Racaniello, Elio Schaechter, and Michael Schmidt
Aired 16 January 2013
Vincent Racaniello: This week in Microbiology is brought to you by the American Society for Microbiology, at microbeworld.org/twim. This Week in Microbiology, episode 49, recorded January 10th, 2013. Hello, everyone. I’m Vincent Racaniello and you’re listening to TWIM, the podcast that explores unseen life on earth. This is our first episode of 2013. Happy New Year, everybody.
Joining me today from Small Things Considered, Elio Schaechter.
Elio Schaechter: Howdy.
Vincent: Happy New Year.
Elio: Happy New Year to you and everybody who’s listening.
Vincent: That’s right. Another one. We started another year of TWIM among other things, of course. Also joining us today from the Medical University of South Carolina, Michael Schmidt.
Michael Schmidt: Happy New Year to everyone.
Vincent: Everyone have a good Holiday? Elio, you were in the desert right?
Vincent: I have this vision of you standing in the desert with nothing around. But there must be houses or something there, right?
Elio: There’s a town called Borrego Springs. But if you go a ways, you’re right. Actually it’s not like the Sahara, big dunes desert. There are some dunes around there but really there’s quite a bit of vegetation in between. So if you look from an airplane you would probably see half-sand and half-vegetation. There’s really little bushes. The creosote bush is the main one in the western desert. It’s abundant, as well, but there’s really, it is a desert by any definition.
Vincent: So do you walk around there? Nice to walk around, yes.
Vincent: Michael, you were in Chicago, right?
Michael: And I was fortunate that the snow that was snowing didn’t stick and I didn’t have to shovel.
Vincent: You didn’t want to shovel, huh?
Michael: No. Shoveling—my 91-year-old father still threatens to shovel and my brothers and I scold him for wanting to go out and shovel.
Vincent: Michael, You’re a young guy. You can go out and shovel.
Michael: No. I would have gone out and shoveled.
Vincent: What are you 40 years old?
Michael: No, I’m older than that.
Vincent: I just became old.
Michael: Yes. Yes. I wish you a Happy Birthday.
Elio: I saw a T-shirt recently that said, “This is the oldest I’ve ever been”. I can’t argue with that.
Vincent: That’s great. I like that. Elio wished me “Happy Birthday” by email. I said, “Well, I usually complain to most people but I can’t complain to you.” You would laugh at me, right?
Elio: Absolutely. Youngsters!
Vincent: We all look young, right. You look pretty young, actually.
Elio: Thank you.
Vincent: We have a couple follow-ups to last time. Actually, I have an email which I think I’ll bring up as a way of introducing a follow-up to Jo’s paper on subtle gender bias. We received an email from Robin, who writes:
The matriarch of a hunter-gatherer band might prefer males to be assigned to specialized training in hunting skills so the sub-conscious bias probably goes back a couple of million years to the early days of homo before sapiens. Who knows what the origin of the sub-conscious bias is.
Elio: Let’s just make sure that people who listening who may not have heard the last TWiM know what you’re talking about. I wasn’t here but I heard it afterwards and I was sorry I missed it in person because it was a fabulous TWiM. I mean both papers were incredible. And Jo presented a paper of her own, published in PNAS, which she did a remarkable little study that showed that in fact that there is an unconscious bias in people, both males and females. She did this—do you want to say how she did it, Vincent?
Vincent: This was specifically in science faculty. They identified 120 or so science faculty, different fields of science, at major universities, and sent them a CV or resume of a person applying for a lab position.
Elio: Lab manager.
Vincent: Lab manager position. They all received the same resume. Half had a male name and half had a female name on it: John or Jennifer. They were asked to rate this individual—would they hire them, how much would they pay them and would they mentor them. The results clearly show that the male applicant was vastly favored in all three categories over the female by both male and female science faculty.
Elio: One thing is, although the results are certainly statistically significant and cause for alarm, I don’t doubt that, still numerically the difference was about 20% on average. Wasn’t it?
Elio: Which means it’s important. It’s terribly important. But it not insurmountable. It’s not like a factor of two or something like that.
Vincent: I think it can be surmounted and they suggested some ways of doing that.
Elio: Right. So it’s a problem, and it’s an important problem, but it’s not a gigantic difference that they found. I was really taken by the experimental design. Experimental scientists, current or past, would have a hard time designing such an experiment and this was brilliant. Because it really had every factor conceivable that would ensure validity. I thought this was, in itself, quite dramatic. The ability that she showed, she and her colleagues. Jo worked with a bunch of social scientists on this, yes.
Michael: The leaders in the field, the ones that look at unconscious gender bias.
Elio: So they really put together a tricky but very, very successful instrument. They did, as Jo mentioned, have to be tricky. They didn’t tell the recipients what the purpose of this study was about. The recipients were just told straight—“Rate them because we’re trying to study how to best place our students” or something like that. It was something which was not quite true and they have to live with that. Jo mentioned this that she was uncomfortable with it. On the other hand, I think that most people of goodwill, at least, would not resent it at all and would say, “Well, this is what you have to do”.
Michael: And they found out at the end. It wasn’t like they were hood-winked. At the end of the survey they found out what actually was going on. So it’s not like they found out well after the fact. They found out literally when they completed the survey.
Elio: It’s probably an ethical question there. I wonder if this is any different than giving people a placebo or not giving a placebo and not telling them until afterwards. There’s some vague analogy here that I perceive.
Vincent: And in both cases the only way you can get an answer, you have to have a placebo, right? And in this case…
Michael: You know in the new work for the discovery and testing of new antibiotics we can no longer give placebos because it’s not ethically appropriate. It’s non-inferiority trials that are being conducted today. So Elio’s got a point in terms of whether or not we had─ she needed to look at something like a non-inferiority trial. I’ll share with you one of the things that I did after that particular episode. I was so intrigued with the paper, when I got home I wrote an email to the dean of the College of Medicine recommending the paper that she would then recommend it to the chairs of the college and recommend it as a departmental journal club for the faculty to consider in terms of looking at this whole issue of gender equity. I also forwarded it on to the folks at my institution who are concerned with this use of gender equity and really encourage folks to take a look at the paper and to think about it. That’s, after all, what journal clubs are for is to get people to begin to think. I think that was one of my take-homes from that really intriguing paper is to go out and become, if you will, an evangelist to say, “Hey, this is out there and we need to think seriously about it and develop a strategy to help our female colleagues excel”.
Elio: Yes. Good for you. Did you get a response from the dean?
Michael: I got a response from the dean. She liked the idea and she forwarded it on to the committee that worries about those issues. They are currently noodling on that idea trying to figure out how to roll it out and see what happens. They have seminars that they often offer on this particular topic, all college seminars. So it may be something that gets presented or at least gets discussed.
Elio: I think it’s great and I think we all should probably try to do something like that because it seems to me that awareness on the part of everybody, males and females, awareness of this being a problem is the first step. I hardly knew this. I kind of vaguely would have thought that there is an unconscious bias but I thought, “If we don’t have a conscious bias, it should be able to take care of it,” but in reality it’s more pervasive than that. People should know about it. So let’s all do that. Let’s all send it out to various deans and people. Okay.
Vincent: Well, I blogged about it and I got a lot of quite interesting comments. I felt the same way, I wanted to do something in addition and my blogging platform reaches a lot of people, and Twitter and Google+, so I think it’s important. I think you’re right, Elio, that awareness will probably will go a good way although not all the way.
Elio: Right. Now I actually─ I liked what you wrote in your blog so much that I asked your permission to re-print it in our blog so it’s going to have another shot in February, to keep the issue alive.
Vincent: I think that Jo, unfortunately, should go around talking about this.
Elio: She probably is.
Vincent: She is but she could do it an awful lot. She could spend her whole life doing that. But that really helps too, I would love to have her come here and give a talk on this. I think you have to let people know this. They’re not just going to read just the paper. You have to emphasize it. Okay. Was there something else we wanted to a…?
Michael: I wanted to ask you, Vincent, since you’re our resident influenza expert. I wanted to gauge your opinion. The press has been making a lot out of this year’s influenza attack rate. There have been some intriguing hypothesis that have been promulgated in the scientific and non-scientific literature that one of the reasons the attack rate may be so high is because the mid-west was so dry this past year and that’s why it’s not attacking certain states. That hypothesis has appeared to have died over the last few weeks because the attack rate now looks pretty much countrywide. Would you care to comment on what you—because Boston is really having a tough time of things—in our local community we’re restricting visitation and hospitals to only immediate family and so any thoughts or comments?
Vincent: Yeah. Interesting. Actually a reporter from USA Today called me this morning and asked me the same question. He said, “You’re good at debunking myths. What’s going on here?” First of all, the strain circulating this year, the predominant strain, is H3N2 and that’s different from the last few seasons where H1N1 strains were circulating. H3N2 tends to be a little bit more lethal than H1N1. So we may see more deaths this year. It’s too early to tell what the mortality will be. So that’s one issue. There’s no doubt about thatH3N2 is predominating and who knows why. We can’t really tell why.
The onset of the first peak of illness or infectious—CDC does a wonderful job of collecting samples from all over the country and then doing diagnostic tests looking for whether they have influenza in them and, if so, what type it is. They have a wonderful website cdc.gov/flu/weekly. There, you can see every week they put of a graph and you can compare it to previous years. This year the onset of the first peak of flu is earlier than usual. I wouldn’t say there are that many more cases than previously. For example, I’m looking the graph for 2010/11 right now and the peak number of positive specimens is about 6,000. That was in week six of the—which would be January, of course, so the peak there that year was a bit later. I don’t know if we have peaked so far this year. We may have. If you look at the CDC numbers for this week the numbers are going down compared to the last few weeks but that is subject to revision. We could have another peak in January or so. So, we had a quicker onset. I don’t’ think the numbers are any higher.
If you look at New York City ─ I’m looking at a graph from the New York City Department of Public Health here. Now, this isn’t as good as the CDC because this is just influenza-like illness surveillance data. These are people who have certain words like: “cough and fever” or “sore throat and fever” or “flu” and they say, “Okay, that’s influenza-like illness but it may not all be flu.” There is quite a spike in New York which is a little steeper than in previous years. Although in 2010/11 again there was quite a substantial spike of illness at about the same time as now. So my feeling is this is a little bit more serious of a flu year. It’s nothing off the charts. There have been some recent years that have probably had as many or more cases. Maybe the onset is not as early but I don’t think there’s anything unusual. I think there’s still time for people to get their influenza vaccines which they should do.
Now, unfortunately, as an example, I read an article on Forbes today, by a professor of epidemiology at Yale University. It’s about Tamiflu, this antiviral for flu. He introduces this article: “Influenza-like illness is sweeping the country with CDC reporting that most areas of the country experience high rates”. Remember that’s still like influenza-like illness. It’s not diagnosed again. So some of it may not be flu. “I should know”, he writes, “my family is in the midst of it despite having been vaccinated”. Now that makes people think: “Huh, this guy got vaccinated and he still got flu”. But he might not have flu because there are other viruses that can cause an influenza-like illness or make you think you have flu. So just because you have the vaccine and got some kind of respiratory illness is no reason for others not to get the vaccine. Remember, the vaccine is not 100% effective so you might get it and still get influenza anyway. It’s not the best vaccine that we have.
Elio: By the way, this year’s vaccine that’s going around against H3N2?
Vincent: Yes, in fact that’s a good point. The match between the vaccine and the circulating strains is very good this year. They made a good choice which they don’t always do. It’s a hard choice to make. So that’s my take Michael.
Michael: I think it’s good. The watchword that I always tell folks about Tamiflu is that you have to take it within two days of onset of flu symptoms because afterwards the horse has left the barn and there’s no point in closing the door because you just literally have too much viral load. The symptoms that you’re going to feel or how you feel from the flu is just too far gone. It’s no longer going to be effective because the way Tamiflu works is it prevents the virus from being released from the replicating cells. You sort of lock the virus in so it can’t get out.
Vincent: The only thing Tamiflu will do is shorten the duration of symptoms. It’s important to remember that when they did the clinical trial for that, that is what they measured, that’s where they got an effect and that’s what it’s licensed for. So don’t think Tamiflu is going to make you miraculously healed. It will just shorten the number of days that you feel lousy.
Michael: It’s not penicillin, which is the generic concept of how antimicrobials should work. Because when penicillin was first released to treat bacterial illness it was truly a wonder drug that really stopped everything cold in its tracks.
Vincent: Right. Yes. Tamiflu’s not a wonder drug.
Elio: By the way, Vincent, you did discuss this in TWiV, did you?
Vincent: This flu story? Not yet, no.
Elio: So when you do, our listeners of TWiM should go over to TWIV to hear the thing in better detail.
Vincent: Yeah, I don’t know when. Maybe in the next few weeks. If we do it I’ll just announce it on here again. I was thinking of doing an episode on this whole issue at some point. Because it’s relevant now, obviously, people are worried. It’s on the front page of the New York Times today.
Michael: It was on the front page of our paper.
Vincent: Okay. Let’s move on to our two papers today. The first one has to do with Staphylococcus. Those of you who are faithful listeners should remember that not too long ago, episode number 42, we talked about this three-star pathogen, the meeting that Michael attended in Lyon, right?
Vincent: So I remind everyone to go there and bone up on your Staph biology.
This is a paper from Nature. It is called: CCR5 is a Receptor for Staphylococcus aureus leukotoxin ED. It’s from right here in New York City: New York University School of Medicine. This again has to do with Staph aureus. You all should know from listening to that episode this is a significant pathogen that causes all sorts of diseases: skin infections, soft tissue infections, bacteremias, pneumonia, sepsis.
The important aspects for this paper are that these bacteria make a variety of protein products that are secreted and allow the bacterium to do its nasty business: it makes super-antigens, it makes antibody-binding proteins, which we talked about last week, those proteins that bind—or two weeks ago—FC receptors of antibodies, FC portions of antibodies. They also make pore-forming cytotoxins. These are the leukotoxins that we’re going to talk about today.
These leukotoxins are soluble proteins that are secreted and they recognize the membrane of a cell. Typically for these leukotoxins they’re leukocytes or white blood cells like lymphocytes or macrophages. They make a pore in the membrane, basically make a pore, poke a hole in the cell membrane, this is the eukaryotic cell. That, of course, is bad for the eukaryotic cell and eventually the cell will die. Staph aureus makes four different of these pore-forming toxins. What we don’t really understand about them is what they target on the host cell, what do they recognize. This is important information to know so you can understand the mechanism by which they work and maybe to design ways to subvert their activity.
So what they did in this paper, was to express these four toxins. They did three of them and they simply added them to cells in culture as a start. They say, “Do these toxins kill the cells?” It’s very easy to tell if a cell in a culture is killed or not. There are a variety of assays. They actually used flow cytometry here. You can add the toxin, you wait a bit, and then you say, “Are the cells living or dead?” You can express that as the percentage of dead cells.
What they found, interestingly, with one of these leukotoxins, LukED, now this is a toxin composed of two protein subunits, the E-subunit and the D-subunit. It would kill cells, a T-cell line actually from humans, which expresses a chemokine receptor called CCR5. They had another T-cell line which did not have this chemokine receptor, CCR5, which was not killed by the leukotoxin. Alright, so two cell lines, both T-Cell lines from humans, one is killed by the leukotoxin that expresses CCR5. The other does not. It doesn’t express CCR5. So that suggests, of course, that CCR5 is the target.
So what they did was to prove that they knocked the expression of CCR5 in a different cell line with small interfering RNA and they showed that, in fact, that confers resistance to this leukotoxin. On the other hand, if they start with the cell line that does not have CCR5 and then they express it that confers killing on those cells by the leukotoxin ED. And they also show that both subunits are needed and that none of the other chemokine receptors are able to make a cell susceptible to this toxin. So it looks like, at least in cultured cells, CCR5 is recognized by LukED. Now those of you who dabble in virology may recognize CCR5. I bet you do Michael.
Michael: I was wondering whether you were going to bring up our friend HIV to introduce…
Elio: You’re friend! You’re friend!
Michael: HIV is this really remarkable virus that I think has done more for immunology than any other virus because it’s really helped us tease apart the immune system. Remember, during the pandemic phase of HIV, there was a lot of talk about CCR5 and delta 32. They go back to plague. They go back to smallpox in looking at CCR5. So I’m going to be very much interested in listening to what Vincent has to say about that.
Vincent: So CCR5, besides being a chemokine receptor on certain immune cells—and chemokines are proteins that are elaborated during an immune response. Their function is to recruit cells to the infected areas, say, so they bind receptors and that’s how they exert their activity. CCR5 is one of these receptors. So it has a function in the cell. It also happens to be one of the two receptors that HIV needs to get into cells. So the most famous receptor for HIV is CD4 which is a molecule on T cells. But it requires a second protein, and there are two possible second proteins and one of them is CCR5. So there are some people, as Michael alluded to, that have a mutation in the gene for CCR5, it’s called “delta 32”. They don’t produce this protein and they are naturally resistant to infection with HIV. There are about, I think, 10% of the people in the world have this allele this delta CCR5.
Michael: I thought it was only Europe. I thought it was only people of European extraction but I may be wrong.
Vincent: It is. So that 10% is basically all pretty much in Europe. We don’t know what’s selected for but as you suggest that some people think that it may have to do with the plague. There was once a hypothesis that smallpox selected for it but that’s been debunked because the mutation’s been around for a long time and in places where smallpox was not. We don’t actually know what’s selected for but it’s in the human population. That will come into this story in a moment. Anyway, CCR5, because it’s a co-receptor for HIV there has been a drug developed. There’s several drugs developed that target CCR5. One of them is called “Maraviroc”. It binds to CCR5 and it prevents HIV infection. They show in this paper that this drug blocks killing of cells by LukED, by the purified leukotoxin. So I think this is amazing that the drug that blocks HIV also blocks LukED killing. It’s pretty neat. So far we’ve been working with purified proteins. Then, of course, they ask, “Does Staph aureus bacteria kill CCR5 expressing cells?” They have to jack up the expression of LukED in staph aureus because normally it’s not made at very high levels. So they do that with a genetic trick. They find, in fact, that if you add those Staph aureus to cells expressing CCR5, those cells are killed. That killing can be blocked again by Maraviroc, that drug that binds to CCR5. If there’s no CCR5 present, the bacteria will not kill the cells.
Michael: So this leads me to bring up the question of—you know, the epidemiologists were the ones who tumbled into this CCR5 and HIV and the correlation to Black Death or smallpox. So the immediate question that I have is, during those times, was it a co-infection when you had plague or when you had smallpox where you actually had frank staphylococcal aureus pneumonia, and it was CCR 5 and the Staph aureus was the true selective pressure rather than the Yersinia pestis or the Variola. Here we have these authors who, unequivocally, mechanistically demonstrate via the knockdowns with the silencing RNA as well as the delta 32 mutation which tumbled out of the work in the ‘90s with HIV. It really is a mechanistic argument driving home that Staph aureus does indeed use CCR5. If you’re expecting natural selection to work on individuals before they’re able to deposit their genes to the next generation, you can well imagine how getting deleted for CCR5 would be a selective advantage if you actually got a co-infection that compromised your lung and the staph sort of snuck in. But if you were deleted with CCR5, you may be able to survive plague.
Vincent: Yeah, I think that’s a reasonable hypothesis. Sure. In fact, they say later that this adds now another pathogen that could have selected for the delta 32 mutation.
Elio: It’s more common. Being it’s more common that makes some sense, doesn’t it?
Vincent: It could, yeah.
Michael: In the most recent New England Journal the Supreme Court has now said that it’s possible to talk about off-label use of drugs because they’re now—off-label discussions are subject to free speech. It’s funny but true. The Supreme Court has decided that off-label discussions of drugs is subject to the free speech clause of our constitution. Maraviroc, which is used for HIV, could we use Maraviroc and some of those horrific cases where no drugs are working against the staph pneumonia.
Vincent: Well, I suspect someone’s going to do a clinical trial, right, Michael?
Vincent: Yeah, because it makes perfect sense. As we’ll see in a moment it does have some effect in a model for disease. Now they showed very nice experiments that this leukotoxin, in fact, the LukE part of the leukotoxin binds directly to CCR5, so LukE but not LukD. Again, there are two polypeptides that make up this toxin. LukE binds directly with CCR5 at the end terminus and so this begins to give you some idea of how the toxins are targeted to the membrane of leukocytes. They bind directly to CCR5 and then, of course, that’s followed by making a pore in the membrane. We also would like to know what cells are killed by LukED. So they do some very nice studies to identify the particular subsets of T-cells that are depleted when you add this leukotoxin. Again, it’s important for us to understand the pathogenic effects of infections.
The last experiment is really striking to me. They ask in a mouse model for Staph aureus infection, what is the effect of removing CCR5. They have mice that lack the gene for CCR5. This is, as you might have guessed, from the delta 32 mutation in people, this is not an essential gene. They infect wild type mice and CCR5 null mice with staph aureus. The CCR null mice, the bacteria can get in to the infection site and they elicit an immune response but there is less bacteria in the kidneys of these animals and there’s less mortality associated with infection in mice lacking CCR5.
Elio: Actually, this struck me so much. I was wondering if the title of the paper couldn’t be a parallel title that says: Under some conditions leukotoxin ED is sufficient to cause death. It still is a difficult thing. Here’s a multifactorial pathogen par excellence. Staph has half-a-dozen to a dozen major classes of what I call “virulence factors.” In this case, under these conditions using mice, et cetera, it looks like that’s all it takes. In other words, if you were to clone this toxin into an otherwise nonpathogenic organism, it could cause disease and kill mice.
Vincent: Yeah, I think that’s a good experiment to do. Yes.
Elio: It would be. Yes, the kind of thing that Victor Nizet lab’s likes in doing. Things like put the Lactococcus and puts the strep genes in it and stuff like that. It may not be sufficient but it comes close to defining a situation where you can make—tease out from this multifactorial, complex situation, you can tease out something relatively straightforward. It’s not as simple as I say it but it would be very interesting thing to do. So the title could be: “Leukocidin ED is a major or, if not, sole virulence factor in staphylococcal infection.” I don’t blame them for not doing that, by the way.
Michael: It also suggests probably why folks making vaccines against Staph aureus have had such a terrifically hard time. If you think about it, if Staph aureus is able to infect T cells. T cells is a very important arm of the immune system, you can begin to understand how Staph aureus can literally take out that arm of the immune system with its toxin. As Elio alluded to, it may not be sufficient by itself, which is why Nizet has to do that fancy experiment, but it may suggest how staph is able to escape the vaccines where they’ve pulled out antigens and actually used them to vaccinate animals. In the animal, it works, but the animals don’t have CCR5 unless it happens to be a humanized mouse. So I wonder—I’ll have to go back and look at some of those vaccine papers and see what they used. Because, I think, this really will help the vaccinologist understand why the staphylococcal vaccines may have failed.
Vincent: Right. Basically to summarize this, this effect in mice, it’s quite a substantial effect, about 80% survival as opposed to 20% survival when you have CCR5 present. So this toxin, LukED targets, remember, T-cells; it targets macrophages and dendritic cells, I didn’t mention that. So basically if you take away the targeting of these cells the mice do a lot better so that tells us we have to look to see the role of these cells in protection against infection. The other interesting point they make is that the gene encoding this leukotoxin, LukED gene, is present in over 70% of strains that cause infections in the US and Germany. That’s good. On the other hand, it is not present in some strains that cause hospital-acquired infections in the UK, for example. So it’s not going to be the answer to everything, right, Michael?
Vincent: It’s going to be a good subset though perhaps. So you might consider treating some of these infections—I guess you’d have to type the infections for the presence of LukED before you would treat with Maraviroc, right, Michael?
Michael: Right. It would be the prudent thing to do. With the advent of PCR being so quick you could ask the simple question, does it have the LukED gene? You could just do a quick PCR and answer the question based on the strain. That’s routinely done for a large number of viruses in the hospital today because they no longer culture viruses. But in the case of bacteria, a prudent use of an antimicrobial is always the key to preventing the emergence of resistance.
Vincent: Right. So that’s what I think is quite a nice piece of work.
Elio: I thought so too. I was well quite taken. My bias about the title reveals that I’m on the bug’s side and not the host’s side. But it’s really quite a paper.
Vincent: They may have tried to broaden the title but the journal may have discouraged them, for the reviewers, it’s difficult, reviewers. Okay, let’s move on to a topic close to all of us, wine.
Michael: Well since it’s after the first of the New Year, I’ll tell you how I discovered this paper. When I was first reading the CCR5 paper on Staph I did a quick search and I stumbled into another paper on community-acquired Staphylococci in PLOS ONE and this paper happened to be in the same issue. So I started to read this paper—and the paper I’m going to discuss is: The Vineyard Yeast Microbiome, A Mixed Model Microbial Map.
Vincent: You have to tell us where it’s from Michael.
Michael: I am. It’s from, and this is the dream job for most microbiologists. It’s from the Institute for Wine Biotechnology at Stellenbosch University in Stellenbosch, Western Cape, South Africa. The paper is by Setati, Jacobson, Andong and Bauer.
Elio: I’m going to dare you, Michael, to make a connection between this paper and the previous one. Let’s see what you can do.
Michael: Oh, god. The challenge flag has been thrown down. The easy one is, this is about grapes and if you take the name of “staphylococci” apart it means “grape-like cluster”. That’s the easy one. But, in reality, this is about vineyards. Anyone who has ever talked to an enologist, someone who is enamored with wine, they will use this fancy French word that I’m going to probably butcher in my pronunciation, so I apologize to my French colleagues, called “terroir.”
Terroir is this remarkable concept. It has to do with the geography of where the vines are. It takes into account everything: the microbes that are in the soil, the microbes that are on the grapes, the microbes that are in the wine. It really takes into account everything because a lot of vintners use a natural crush where they don’t add any yeast starter culture to control the fermentation of the grapes that then result in wine. This is a substantial economic issue for people who grow grapes for a living because you get paid on how well your grapes perform. That is, the wine that results from the crushing of the juice and the flavor, the bouquet, the complex odors that come out of wine that then are all a component of the flavor profile that we so enjoy if you happen to enjoy wines. You can do this experiment at home. You can take the famous bottle of Two-Buck Chuck that you can buy at one of those grocery stores out in California. In South Carolina we don’t have any two dollar bottles of wine. That’s Elio’s advantage living in California. He can get a decent $2.00 bottle of wine.
Elio: Actually, they are selling in quite a few other places, as well, in the east.
Vincent: Yeah, we have it in the east but it’s three bucks now.
Michael: Yeah, unfortunately, South Carolina we don’t have anything as inexpensive.
Vincent: But I could send you a case Michael.
Michael: I think that’s illegal in South Carolina too—sending, mailing wine.
Vincent: My name is Racaniello.
Michael: The bootlegger! So I was really intrigued by—and again this is the proper crafting of a good abstract. As I was reading the abstract, the authors used this term “Theory of Sampling”. We know that terroir is important for wine but the other thing folks should appreciate is terroir or...
Elio: Try terroir.
Michael: Terroir, there’s no L in it. Terroir is important is consider coffee and consider chocolate. Something that most folks cannot live without. Chocolate is also similarly a natural fermentation that takes place in the rain forest and we don’t even have a starter culture for making chocolate. It’s a natural fermentation and it’s one of the principal reasons for preserving the rain forest is if we lose the rain forest we lose chocolate because there is no starter culture for chocolate while there is a starter culture for wine. It has all sorts of other applications. In thinking about this, the authors, because the principal point of this paper is they’re asking a very straight-forward question: Does the quality of grape that they will then use in wine, is it influenced by the farming practice? They evaluated three farming practices. The first, of course, is conventional farming where pesticides may be involved. The second is biodynamic, which is their phrase, which I think in the United States we can probably appreciate as being organic where they don’t use any commercial or chemical-type company insecticides. They use compost and they use additional minerals to augment the soil. Then they used a third method of farming called the “Integrated Production of Wine,” which is a mixture of organic farming as well as the conventional farming. There is no set of codified rules that say, “You must do this. This and that.” Basically their Integrated Production of Wine results that they use mulches and oak cover crops to improve moisture and soil fertility. They have bait traps to accumulate insects and they used ducks of all things…
Elio: Used Ducks. Yeah I like the duck part.
Michael: Yeah, to go around and eat the system. Now imagine their problem. The first rule of agriculture, that I learned, is that farmers are barely making ends meet. So consequently these folks don’t have a lot of money to do an exhaustive sampling of their territory. They’re evaluating plots of land in the same geographic region and they’re asking the question: Does the microbial diversity differ based on the farming. The way folks traditionally determine whether or not you have reached saturation is you sample until you see a plateau in the number of species that you’re sequencing your 16S, your 18S, whatever you happen to be using, reaches that plateau. Which, as you can well imagine, would be very expensive because how many sequences, how many bacteria do you have to isolate, how many yeast do you have to isolate in order to recognize whether or not you have a representative sample. So these authors did something brilliant. They employed the theory of sampling. Without getting into too much math you can go into their paper and they explain the math in gory detail but I’m going to try a metaphor. Since we just finished an election cycle in the United States, did you ever wonder how the polling people know who are going to win an election? What they do is they survey likely voters. They employ this theory of sampling. What they do is they take a particular geographic region and they convert it into a two-dimensional plot and they run it through this algorithm that effectively tells them where to sample. In the case of Vincent’s neighborhood, where people live at many different levels, they tell you which floor to sample. Now imagine them. They’re looking at grapes. Grapes don’t just grow in the ground like a potato. They grow at all different levels, they may or may not have leaves shading them and the yeast that…
Elio: The usually grow on trellises, don’t they?
Michael: They grow on trellises. But some of the trellises may be in full sun. Some of the trellises may have leaf cover where they’re covering the grapes. What these authors did is they used the grape as the perfect collection vehicle. They had to understand where to pick the grapes. So they had five trapezoidal plots of land that were evaluating these three different farming methods. What they did is they linearized them so that they could then determine in incremental sampling. What happens is the the computer spits out a 2D map telling them at what height and at what location to accumulate the grape. They then accumulated the grapes. They rinsed them in normal saline and there was a comment. The neat thing about PLOS ONE is people can write into the authors in the comment section and say: “By using saline aren’t you osmotically stressing the yeast and killing them?” These authors simply looked at yeast. They didn’t care about bacteria in this instance because what they were looking at are the epiphytic yeast. An epiphyte, of course, is just something that is growing on something. It’s not actually eating the grape, it’s just on the grape. The yeast are somehow making their living out of the air sort of like an orchid.
So they asked the first question of their paper is: Did the farming methodology change the density of the bacteria based on the farming practices?
Michael: Yeah, excuse me. I got bacteria on the brain today. The yeast. They did a really nice job of showing us the concentration of the bacteria.
Michael: Yes. Keep correcting me. The yeast and the concentration was not that high. It was between—the total yeast population ranged between 4 – 8 X 104 CFU per gram of grape. That was found not to be statistically significant between the different farming systems. The way they asked about the genomic differences or the differences between the yeasts is they did it in two methods. The first is the old-time religion. They cultured them then they identified them via traditional methodology for identifying yeast. I got it right that time. What they learned is that they didn’t find all that much of a difference between the cultivars. They report in their paper that in the conventional farming they had one predominant species, Aureobasidium pullulans, was principally there at 70% distribution. The Integrated Farming method, which if you are thinking intuitively about complexity, you would expect that it would be a little bit more complex than the conventional farming. Conventional farming methodologies because the thought is that harsh chemicals and insecticides will limit biodiversity. You would expect that predominance of one species intuitively would go away and they did indeed see that. The Organic methodology, or the bio-diverse methodology, or Biodynamic method of farming you would expect to be more heterogeneous. They then further asked the question whether or not this was indeed the case. They used a methodology called “ARISA”. ARISA is “Automated Ribosomal Intergenic Spacer Analysis”. Now, we’re all familiar with the bacterial ribosomal RNA operon in that there is a 16S gene and then there is a spacer gene and then the 23S gene. Since yeast are Eukaryotes, they did a variation of this for fungi and there is tucked in-between the 18S ribosomal RNA in yeast and the 28S ribosomal RNA in yeast, there’s the 5.8S ribosomal RNA and they have a fluorescently-tagged 5-prime primer. This 5.8S ribosomal RNA region is amplified and it gives you a different sized piece. Since one end is fluorescently tagged, they run it either on an agarose gel or they were fortunate in they had a capillary gel electrophoresis system that has a specific detector in it that fluoresces and it literally gives them a read out of how big the region between the 18 and 20S gene is. They can determine the complexity of the population. Of course, they can go on to sequence it. What they learned is that there were, as you can well imagine, as they ran this through the statistical packages that are now quite common in any laboratory that does DNA sequencing, they asked the question about species richness and species evenness and then species diversity. I hadn’t looked at the math on these subjects in a while and I was really intrigued about how clearly this paper helps me visualize whether or not their sampling technique really worked. You look at the Melnyk's Index of Species Richness and the conventional farming gave them a 0.96, the intermediate form of farming gave them a 1.06 and then the biodynamic gave them a 1.45.
Species richness is really an interesting concept because what they do is they simply ask the question: How close in numbers each species in an environment are? The math is really pretty complicated but if you think about—and the example I tumbled into when I was trying to figure out how to explain this so I would survive my good friend Elio’s telling me “Make the math simple”. So if there are 40 foxes and 1,000 dogs the community is not very even. But if there are 40 foxes and 42 dogs then the community is quite even. The evenness of the community is then represented by that Pielou's Index of Species Evenness. The working hypothesis is that conventional farming should have a less even coefficient at 0.5 while the organic, which is at the other extreme, should be less even. It is, indeed, at 0.76. The inverse of that is the diversity index or the species diversity. What they found is the conventional had a species diversity index of 1.2 telling you it’s not very much diverse. The intermediate was 1.45 and then the bio-dynamic version was 2.15. The reason I brought this paper to TWiM is principally because I think papers like this will help people begin to ask the question about human microbiomes. How do you know if something is remarkable? Recall back to that TWiM where we asked the question about the bacterial link between autism and traditional gastrointestinal upset. And we found the Suterilla species in the autistic population, but it was a very small sample size, only 27. What this sampling theory will allow us to do is teach us how to answer the question whether or not we are indeed seeing a representative sample. This is important for anyone doing community analysis. I was very much intrigued by sampling analysis because I’d been working in the area of hospital-associated infections and we’ve been sampling the built environment of the hospital. Trying to understand how bacteria move in the hospital, how antibiotic resistance moves. So how are you going to be able to afford to do large scale studies when molecular biology is not cheap? Think about sequencing. Even though it’s come down substantially in cost, you’re going to have to sequence quite a few if you use the traditional modality where you reach a plateau in finding something new. Whereas if you could use the power of statistics—you know, the polling folks have used that to really forecast elections ± 2% or ± 5% in order to predict the outcome of an election. You know the census folks have similarly been arguing that rather than doing a traditional head count of everyone in the United States, they can use statistical methodology in order to arrive at how many toilets are in a particular geographic location.
Vincent: So Michael, basically, when you’re looking at large complex populations you use these mathematical methods to allow you to compare, in this case, different farming methods, without having to enumerate every single member of the different populations. Right?
Michael: Yeah, because the way you would do it traditionally is you would literally go in there with a harvester and take every single grape and literally sequence the yeast epiphyte population from each grape.
Vincent: So this you do it by sampling specific ones.
Elio: It goes beyond it. Let me tell you why I’m so happy that we did this today. It illustrates what you’ve just been saying that statistical methods can, in fact, be used throughout. They can be used to answer questions. But the thing it really emphasizes is the analogy of the question. In other words, you can ask, in microbiology, questions clear across fields and they come out to be the same. These are population structures in certain environments and how they are influenced by modification. So this brings up what I think is the power of modern microbiology. We teach a course—in an hour I’m going to be going over there—Jo Pogliano and I at UCSD, teach a course, we call “Integrative Microbiology”. We think microbiology is an integrative science. It integrates prop questions that pop up from all kinds of different places. So I think this is delightful that we have been able to deal in—the same people, I mean, who are able to talk about something which sounds as disparate as Staphylococcus aureus and yeasts on grapes. I think this is the power of modern microbiology really.
Vincent: By the way Elio, we did you friend Jo’s paper on TWiV not too long ago. You should tell him that.
Elio: Did you?
Vincent: Yeah, one of his papers about phage making microtubule in host cells. Just tell him that we did it.
Elio: Okay. I’ll tell him.
Vincent: So Michael, this is the cool take-home message but in the end they wanted an answer to their question about the variation between the methods and what is that?
Michael: The take-home message is what you would expect. Is that the organic farming did indeed give them a little bit more complexity and the culturable yeast population, you know the god-fearing old-time religion culture, did not reproducibly predict the fully representative cohort that was present. It’s only through this molecular ARISA analysis that we’re able to discriminate amongst farming practices. The yeast distribution was significant and it was subject to significant intra-vineyard spatial fluctuations. The frequent heterogeneity of tank samples of grapes harvested from single vineyards at the same stage of ripeness might therefore be at least due to these different microbiota in different sections of the vineyard. So looking at this from an economic perspective in how the farming industry is going to take advantage of this paper, they’re going to literally do ARISA on grape lots where they will go out to a field and they will develop an index to understand how much to pay the farmer for his grape crop based on what type of wine that particular ARISA pattern turns out. The analogy that I thought when I was reading this, is this is very similar to what the dairy industry used to do for farmers. They would take a sample of raw milk from a dairy farmer and they would add methylene blue to it. It would immediately turn milk, this blue color, the stuff that you saw on the first Star Wars movie when Luke Skywalker was drinking blue fluids. Then they would start a stopwatch. Based on the bacterial load in milk, the milk would go from that Star Wars blue to milky-white. The faster it turned white, the less the farmer got paid per pound for his milk because it had…
Elio: No kidding? That’s funny. That’s great.
Michael: It had greater bacterial contamination. Moving this into the modern era, I can imagine this happening in California immediately because this technology, all you need is capillary electrophoresis system, two PCR primers, PCR up, you know, grab the grapes based on the farmer’s plot. You can then assess whether a Cabernet Sauvignon is going to make you a $100-bottle of wine versus a $2-bottle of wine. So there is real economics here and it’s up to the, again, enterprising microbiologist to go out and do this mapping but it’s all based on this sampling technique.
Elio: Michael, you’re making a jump here. You’re making a jump that you’re going to tell the quality of the wine from the composition of the yeast population. Is that, in fact, true? Can you make that jump?
Michael: Well a lot of commercial wineries are now using starter cultures to overwhelm the microbiome that comes with the grape.
Vincent: Yeah, then you can control it. Right?
Michael: Then you can control it. Our friends out at University of California at Davis have a whole department of microbiology and enology where they actually marry the two of these things together. So I’m sure we’ll get letters from them if I got it wrong.
Vincent: Well then they can come on and set us straight.
Michael: That’s right.
Vincent: So, in France, where the best wines are made, do they used to…
Vincent: I know California’s pretty good.
Michael: South Africa too.
Vincent: But do they use starter cultures—is that considered okay to starter culture? Is that blasphemous?
Michael: It’s not blasphemous. The brewing industry, of course, no longer lets things to chance. After all most of the baking yeast that we buy in the grocery store to make bread comes from breweries.
Vincent: I guess you could say if you get a great bottle of wine it doesn’t matter how you do it. Right?
Michael: That’s right.
Vincent: That’s really cool. I think we ought to get a person from this industry to tell us all about it one day. That would be fun. Then they could send us a case.
Michael: That would be wonderful.
Vincent: You know, are you done with that so we can move on?
Michael: Yes. I’m done.
Vincent: We have actually an email which is totally relevant to this so let me read it. This is from Chip. Chip writes:
“Saw this paper in my inbox this morning and thought It might be a good fun discussion paper on an upcoming TWiM.”
So this paper is called: “Brewhouse-Resident Microbiota Are Responsible for Multi-Stage Fermentation of American Coolship Ale.” So this is beer. American Coolship Ale is a type of beer that we make in the US. It uses production methods developed in Belgium by the Lambic brewers. Traditional Lambic fermentation is done using all the microbes in the worts. They expose the worts to the atmosphere overnight in a vessel which is called a “Coolship.” That’s why it’s called Coolship Ale. It becomes inoculated with autochthonous yeasts and bacteria that then, over the next one to three years, carry out the fermentation in oak casks. This is a kind of beer, Michael, where you use the endogenous yeasts and bacteria, you don’t use a starter culture. I guess it’s a traditional kind of Ale. So they wanted to know how this culture develops over time. They used a bunch of methods. They used restriction fragment length polymorphism, barcoding sequencing, quantitative PCR and culture dependent analysis. They traced over the time of the fermentation, it’s a couple of years, the sequence of populations that arise. For example, in the first month, they get Enterobacteriacea, a bunch of oxidative yeasts, then they get Saccharomyces spp. and Lactobacillales. Then a year later they get Brettanomyces bruxellensis, and then Candida, Pichia. It’s just amazing.
Michael: Pichia, which is a biotechnology wonder, wonder yeast.
Vincent: Right. Use it for expression. So they define this whole sequence of events that happened in this culture. I guess the idea would be that maybe you could use this information to make sure things are going properly.
Michael: I think you could go and figure out how to sample the floor of the brewery where this is made so that when you make another brewery you can answer the question of whether or not—you know the whole issue of brewing is to turn out a consistent product.
Vincent: Right. So you’re depending on the microbes coming from wherever you have this Coolship sitting, right? Atmosphere, so figure out ─ maybe from this they can from this figure out a starter culture to take care of it more consistently. I never heard of Coolship Ale. Have you?
Vincent: Well it comes from the Belgian Lambic brewers. That’s pretty cool. Alright, let’s do one more email. This is from Meimei:
“I do not have the talent for a M.D-Ph.D. at Columbia, but I am definitely a huge fan. It's striking how and where the interaction of antibody and bacterium occurs (TWIM #48)! Learned a lot from listening to your interpretation of the paper. Many Thanks! Just one question about one antibody and one bacterium. Is it commonly considered more than one antibody molecule per bacterium?”
The answer is yes. If you looked at the images in that paper there are multiple antibodies bound to each fiber, actually, on each bacteria in one or the other orientation.
“P.S. Can this finding extend to the interaction between virus (e.g. HIV) and neutralizing antibody? It might be we just look at the wrong place for all these years!”
So he’s asking whether antibodies can bind to viruses via their FC receptors, right, just the way they bind to bacteria by FC-receptor-binding proteins. So it turns out that there is a virus, Herpes simplex virus, where the virions on their surfaces have FC receptors. Those are believed to bind antibodies as a way of evading immune responses, just like the bacterial FC-binding proteins are an immune evasion mechanism. So it’s believed that that’s what Herpes simplex does too. It’s the only example I know of so it seems to be less common amongst viruses, but it does happen.
Michael: And the problem, in the case of viruses, is they do all their replication inside the house, inside our cells. So you would have to have an antibody level at a high enough level at all times to effectively—because in HIV the whole conundrum is you need to prevent infection rather than disease, which most vaccines try to do this. They prevent the disease from happening rather than the infection. That’s why it’s been so challenging to develop a vaccine for HIV is that it must prevent the infection because of the cell that HIV infects. The immune system and it takes it out. So it’s a really good question.
Elio Schaechter is at Small Things Considered. Thanks for joining us Elio.
Elio: Sure. I missed the last one. I must say this was pretty unusual and I enjoyed it a lot. Thank you very much everybody.
Vincent: Michael Schmidt is at the Medical University of South Carolina.
Michael: Thank you Vincent. Elio, I’m glad I was able to survive the math.
Vincent: You did a great job.
Elio: You did nobly. You really did.
Vincent: You know I think the title for this could just be Grape-like Clusters.
Michael: Grape-like Clusters. That is true.
Elio: Hey, not bad.
Vincent: It will be fine. I’m Vincent Racaniello. You can find me at virology.ws. Many thanks to the American Society for Microbiology for their support of TWIM. Chris Condayan and Ray Ortega for all their help behind the scenes. In case you’re wondering, the music you hear at the beginning and the end of TWiM is composed and performed by Ronald Jenkees and we use it with his permission. You can check out his work at ronaldjenkees.com.
Thanks for listening everyone. We’ll see you next time on This Week in Microbiology.
- End –
Transcribed by: Frank Shinneman