The Biological Species Concept, or “BSC,” was devised and refined by evolutionist Ernst Mayr in the 1930s and 1940s as a way to conceptualize the distinct groups—”species”—that we see in most plants and animals. It runs as follows (in my words), which also includes how we conceptualize different biological species.
A biological species consists of a group of populations that, where they coexist in nature, exchange genes through reproduction. Two populations that coexist in one area but do not exchange genes are considered members of different biological species.
One of the advantages of the BSC is that it enables us to immediately solve the species problem that eluded Darwin: why is nature divided up into distinct clusters rather than existing as a continuum, clusters most visible where they coexist? Under the BSC, the problem of the “origin of species” simply becomes the problem of “the origin of those barriers that prevent interbreeding”—and that is a tractable problem. Again, see Coyne and Orr for our best take on how these clusters form.
Of course there are problems with this concept (it’s not an a priori definition, but an attempt to conceptualize in words what we see in nature). These problems include judging populations that live in different areas like islands of an archipelago, how we deal with groups that hybridize just a little where they coexist, and, most important for this article, what we do with species that are asexual, lacking the possibility of exchanging genes. We discuss all these issues in the first chapter of my book Speciation (2009) written with Allen Orr, but one issue we didn’t resolve properly was that of asexual organisms.
So what about those pesky “asexual” organisms? How can we conceptualize species in groups like bacteria? Well, the first thing we need to determine is whether they form distinguishable clusters like birds or turtles. If they don’t, then there’s no need to conceptualize nonexistent clumps. In our 2009 book, we reviewed the literature, which was scant at that time, and decided that the evidence was mixed about whether bacteria (considered asexual) formed species, but there are surely some clumps among them. So we restricted the rest of the book to sexually-reproducing organisms. Still, bacterial “species” are given names, like E. coli, but do all bacteria considered E. coli really comprise members of a distinct cluster? If so, how?
The literature has expanded since then, and the paper below, which I’d missed and which is now seven years old, makes a pretty good case that in bacteria, at least, there are species, and, more important, they are conceptualized in a way similar to that of the BSC. In other words, there are bacterial clusters, and each cluster is characterized by its ability to exchange genes among individuals. Members of different clusters, however, don’t exchange genes. In other words, bacteria do consist largely of genetically isolated clusters. The authors, though examining only bacteria (there are other asexually reproducing organisms, like bdelloid rotifers), conclude that life in general conforms to the BSC. That’s a bit too expansive a conclusion (see the title!), but their results for bacteria seem good.
Click to read, or see the pdf here.
The key to this paper is recognizing that bacteria are not in fact completely asexual, though they often reproduce that way. But they also have a form of sex in which genomes of two different individuals can sidle up to each other and recombine to produce new genes. This process, called homologous recombination, occurs via cell-to-cell contact or transfer of DNA through tubes (“pili”) connecting different individuals. This process is called conjugation.
Here’s a photo from Wikipedia showing two bacterial cells moving DNA through pili:
This movement is one-way: the DNA (a single chromosome with double-stranded DNA) from one individual moves to the other. After that, there can occur a form of “sexual” reproduction in which different copies of the same gene can line up and recombine, producing a new gene. A similar process happens during meiosis (gamete formation) in sexually-reproducing organisms.
In bacteria this mixing-up between similar genes is called homologous recombination because it changes the composition of a gene by recombining its DNA with the DNA from a similar gene in another bacterium. There are other forms of DNA exchange in bacteria in which a bit of DNA or a “plasmid” from one individual simply inserts itself somewhere else in the genome of another individual, but this is not recombination in the traditional sense, for it doesn’t involve two different copies of the same gene recombining to form a new gene. The paper by Bobay and Ochman deals with homologous recombination,
Their method of determining whether two individuals in a named bacterial species can recombine their DNA in this way is complicated, and I’d best leave it for the experts here. But I will say that it involves showing that individuals in a group share the same variants in a given gene segment (10,000 bases were sequenced) as do other individuals in a group. For example, in one ten-base stretch of DNA, an individual may have GTTACTCTAA, another would have GTTAGTCTAA, and another GTTACTCTAC, and still another GTTACTAC, representing combinations of DNA bases that could occur by recombination.
If you see this pattern among individuals of a named bacterial species, that’s indicative that homologous recombination—bacterial “sex”—is going on. This form of recombination is called “homoplasic” recombination because the variants all come via mutation from a single original genome present in the individual that founded the species.
One alternative is that we are dealing with two related species in which similar DNA sequences only look as if they’ve undergone homologous recombination because two groups shared a common ancestor and then the descendants had similar (“convergent”) mutations. This, called “nonhomoplasic recombination”, is not caused by genetic exchange.
The authors have ways to distinguish these two types of recombination, and devise a ratio they call “h/m”, showing the ratio of the degree of homoplasic recombination (true sex) from nonhomoplasic recombination (independent mutations in different groups that superficially mimic sex). The higher the h/m ratio, the more sex individuals in that group are having.
The authors calculated h/m ratios for 91 named bacterial “species”, using, of course, a large number of genomes sequences for each species, because one needs to survey the variation among individuals in that 10,000-base segment. (They also did simulations to verify that they could tell “h” from “m”.) It turns out that over half of the 91 named bacterial species they examined conformed to biological species in which there was evidence of “h” recombination among individuals. Here’s one below, in which the h/m ratio increases. reaching an asymptote, as they looked at more strains. (This increases your power of detecting shared variants). 54 of the 91 named bacterial species looked like this, so the BSC holds for at least half of named bacterial species, and the authors sampled widely in bacteria.
A biological species in bacteria:
Here’s what was thought to be one species but, when they added more strains, they saw two clusters, one that behaved as like the one above, but the other, relative to the other group, showed very low h/m ratios, indicating that the two groups didn’t have homologous recombination between them. That is, they were different “biological species”. When they took out the low h/m group, B. pseudomallei behaved nicely. Here, then, we have two species that were given the same name, perhaps because they had similar morphologies or culturing requirements, or because the genetic distance between them (indicating the time of separation) was pretty low, suggesting a recent origin. These “cryptic species” were seen in 21 of the 91 named bacterial species.
Two biological species in bacteria that went under one name:
And the third group by itself had low h/m ratios no matter how many strains they included, so that there was no ability to assess gene flow at all—perhaps because these species simply don’t undergo any homologous recombination. Here’s one:
Thus 73/91 groups tested showed patterns consistent with a reproductive-isolation based species concept.
To test that their method did indeed detect groups analogous to biological species in more familiar animals, the authors did the same kind of h/m test for two pairs of related but clearly distinct biological species; one was the related species Drosophila melanogaster and D. simulans, and the other Homo sapiens and the chimp Pan troglodytes. As you see below, they were able to detect reproductive isolation between the group using a similar 10,000 base-pair fragment. (In all cases they looked at many replicates of the species on the left and a single sequence for the species on the right, which is why the “other” species forms a straight line: we have one sequence compared to many sequences in the other species, and all comparisons show a low h/m ratio.)
Humans vs. chimps, also good biological species:
The conclusion, then, is that the BSC is pretty good in conceptualizing species in bacteria: there are groups that exchange gene segments, and other groups (different “species”) that do not exchange DNA via homologous recombination. Remember: all of this was judged from looking at DNA sequences, not seeing gene exchange directly.
The big conclusion (from the paper):
That species can be universally defined based on gene flow implies that many of the same factors are operating in the process of speciation across all lifeforms. Differences in genomic properties (such as ploidy, recombination frequencies, and reproduction, and rates of gene acquisition) and demographic parameters (such as population sizes, geographic distribution, and rates of migration) will impact the pace at which microbes speciate relative to sexual organisms. However, the application of a single genomic-based BSC criterion to delineate species makes it possible to define species and study speciation under a similar framework across the tree of life.
Well, they need to look at other putatively asexual groups to see if this method also shows the existence of interbreeding groups reproductively isolated from other such groups, but at least for bacteria we see that many of them form clusters. Two questions remain:
1.) What is “speciation” in bacteria, then? One of the paper’s most intriguing results is that if you take pairs of bacterial “species”, the degree of reproductive isolation between them isn’t positively correlated with the time separating them, as judged by the “genetic distance”, or whole-genome divergence, between them. This is hard to understand because it implies that, unlike sexually reproducing organisms like fruit flies and mammals, reproductive barriers don’t form as a simple byproduct of the time of divergence from their common ancestor. This is the case because in those groups reproductive barriers are usually byproduct of divergence between populations by natural selection and genetic drift, which drive species apart genetically as time passes.
Why isn’t this the case in bacteria? I have no idea! My only suggestion is that “species formation” might be so quick in bacteria that you simply don’t get a correlation of time with reproductive isolation. That would imply that it’s virtually instantaneous.
2.) Why do bacteria form clusters? In more familiar animals, clusters arise because after reproductive barriers arise, an interbreeding group is free to adapt to its environment without “pollution” from other species that would efface the clusters. The genetic divergence is reflected in not just reproductive isolation, but in the way organisms look or behave. This may also be true in bacteria: each cluster might represent a group adapted to a particular ecological niche. This would be hard to test for naturally-occurring bacteria, but might be tested in pathogenic bacteria, whose habitat (us) is more easily studied. As I recall, each bacterial species does its own thing in its own way, but that’s not really an answer to the question.
A final note: this paper was difficult, and I may have made some errors in summarizing its results. (I could read it only twice before I had to write about it here.) Perhaps the authors will read my summary and correct any mistakes.
_______________
Reference: Bobay LM, Ochman H. Biological species are universal across Life’s domains. Genome Biol Evol. 2017 Feb 10;9(3):491–501. doi: 10.1093/gbe/evx026.
👍 Very interesting!
“Asexual” shouldn’t be in quotation marks. Bacteria are just flat-out asexual.
If bacteria can have ‘sex’ without having their own gametes, then why are gametes fundamental to the definition of ‘sex’?
(If sex in bacteria doesn’t require gametes, then sexes in humans don’t need to correspond to gametes either.)
It isn’t meant to be contradictory. Sexual reproduction here is being used a bit liberally, but to expand on it: Sex in bacteria is exchange of DNA via conjugation. Sex in single-celled eukaryotes involves fusion between cells of different ‘mating types”. There are no males or females in these ways of sexual reproduction, so there is sexual reproduction without gametes in many species.
In multicellular eukaryotes, beginning with simple multicellular algae for example, we see distinct egg and sperm gametes. In many such species there are biological males and females, with separate kinds of gametes, while other species are hermaphrodites.
I do prefer to not use ‘sex’ for bacterial conjugation, except as an informality.
I don’t even think it should be used as an informality.
What grounds do we have to criticize people who define the “sexes” based on secondary sex characteristics, if some biologists also loosely define “sex” such that even asexual species like bacteria can engage in it.
Technical definitions shouldn’t be used metaphorically. As far as I can tell, the only reasons people refer to bacterial “sex” are: (1) for bacteriologists to make their work sound “sexier” and more relatable, (2) confusion about the meaning of sex.
In your last sentence in parentheses you are confusing having the act of sex (behaviour that in the natural scheme of things results in genetic recombination to form a new individual) with having the property of sex (body plans that make one (or both in hermaphroditic species) of two possible gametes.) Bacteria have sex (of a sort) in the first sense but not in the second sense. Higher eukaryotes have sex in both senses.
“Sex”-competent bacteria are the ones that can make sex pili. This is mediated by an F gene (which causes undergraduates to snicker). As DNA replicates it is fed through the sex pilus to the recipient cell which, statistically, is more likely to be F-negative. The transferred DNA incorporates into the recipient chromosome. I confess I don’t remember if an F+ cell can conjugate with another F+ cell. I don’t see why it couldn’t but here’s the important thing: if the conjugation event lasts long enough before being interrupted by too much jiggling, the DNA sequence copy transferred can include a copy of the F+ gene itself. Then the recipient cell becomes F+, too.
There are no gametes involved here, so we have to say that bacteria have sex without gametes. The “maleness” (if I may deliberately misuse the term for illustrative purposes) of an F+ cell can be transferred to an F- cell making it newly “male”. It doesn’t work at all that way in higher eukaryotes. Where gamete-forming organs evolve, the two sexes become thus absolutely distinguished.
I am not confusing anything. “Sex” a reproductive act that results in the union of gametes. The sexes are defined by their gametes.
The first part of the sentence is about the act; the second part is based on the definition of the sexes based on that act.
If you define sex differently (such that it does not require the union of gametes), then defining the sexes based on gametes is contradictory and nonsensical.
F+ and F- types are not gametes. Bacteria are asexual, full stop.
Polysemy. The word ‘sex’ has various meanings in different contexts. If you’re not confusing anything then you’re insisting on your preferred single ‘correct’ meaning. The authors of the bacteria paper are simply using a different meaning than you would like. To them, ‘sex’ means a process that recombines genes. Just as reproduction may be sexual or asexual, to them sex can be reproductive or nonreproductive. They are not “wrong”.
That’s one reason this silly argument about sex as a binary or a spectrum never stops; different meanings of ‘sex’ get swapped and mixed with different meanings of ‘gender’ at will for rhetorical purposes.
Because even sexual reproduction does not require different sexes (i.e. isogamy), it seems to me that we should be insisting that ‘sexes’, not ‘sex’, are binary. That would obviate at least some of the intentional confusion.
It’s not polysemy in this case.
“Sex” is a technical term in biology. It has a precise meaning: reproduction based on union of gametes. And sex is binary when used with that meaning, because there are only two gametes.
Biologists have to clearly be able to say “sex is binary” without being pilloried by non-biologists who don’t know any better. (Or by people intentionally trying to obfuscate for political reasons.)
Biologists should use the word “sex” consistently and unapologetically in its technical sense, and avoid using it in any looser, technically incorrect sense that invites confusion.
That sex is binary should have no political ramifications at all. (Who cares if someone produces big or small gametes!) No sane person denies that the two sexes overlap substantially in almost every phenotype. But they don’t overlap in their “sex”.
You’re incorrect. Biologists do in fact use the term ‘sex’ in different ways.
Fungi and many algae engage in unambiguous sexual reproduction but do not have morphologically different gametes. Hence the concept of ‘sexes’ (binary) must be different than the concept of ‘sex’, however the latter is construed.
“If sex in bacteria doesn’t require gametes, then sexes in humans don’t need to correspond to gametes either.”
Your sentence is illogical. There are two logical constructions. One is “If sex in bacteria doesn’t require gametes, then sex in humans doesn’t require gametes.” The other is “If the sexes in bacteria don’t require gametes, then the sexes in humans don’t require gametes.” Both constructions are false.
Clarity requires a three part logical construction. One is “If sex in bacteria doesn’t require gametes, and sex in humans requires gametes, then bacteria aren’t human.” The other is “If the sexes in bacteria don’t require gametes, and the sexes in humans require gametes, then bacteria aren’t human.”
Really nice post! I think this is still a major open question.
I really enjoyed this one – made a difficult idea simple enough for me to follow
Hear hear!
It sounds interesting and I’ll need to read it properly. As far as I know, conjugation is quite rare and transfer of chromosomal DNA through conjugation is even rarer. I’d need to check those things. I’m also a bit concerned that it’s on a very small number of bacterial species and the ones for which we have a large number of genomes are not necessarily representative.
I’d also note that this is not how bacterial species are commonly defined. The current standard is still average nucleotide identity, see the GTDB FAQ (https://gtdb.ecogenomic.org/faq#how-are-gtdb-species-clusters-formed). I’m not sure if Bobay and Ochman’s idea is being ignored because it’s not well-known or if it has major flaws.
“But they also have a form of sex in which genomes of two different individuals can sidle up to each other and recombine to produce new genes. This process, called homologous recombination, occurs via cell-to-cell contact or transfer of DNA through tubes (“pili”) connecting different individuals. This process is called conjugation.”
To clarify, homologous recombination only refers to the exchange between similar DNA strands, not the whole process with conjugation. Conjugation is only the exchange of DNA through the pili but that is often just a plasmid with no homologous recombination.
Very good. I also thought that chromosomal exchange in bacteria during conjugation was more of an accident, while its main role was to exchange plasmid DNA. In particular, the F plasmid which carries genes for building the pili. But that just shows that what I know on this – and maybe what is in the textbooks? – is miniscule.
That’s my understanding, too. Conjugation mostly involves the transmission of small circular extra-chromosomal plasmid DNA. The F gene itself is often on a plasmid. Certainly this is how many antimicrobial resistance genes disseminate among bacteria that are not so closely related as to be considered the same species but still rhyme with each other.
To define species by average nucleotide identity (a form of “genetic distance”) seems to me a useless exercise. What does it mean, unless genetic distance is correlated with some biological feature of interest? Otherwise you’re just using an arbitrary and subjective criterion to create meaningless groups.
It’s not entirely arbitrary. The problem is, of course, that bacteria are asexual, so the definitions developed in animals seldom apply (current paper under discussion notwithstanding). We’ve long known about specific bacterial species even before we could compare their genomes; this has been through phenotypic characteristics such as morphology, growth conditions and metabolism. But those criteria can take a lot of effort to get, especially since many bacteria are not easy to culture. But when you find that bacterial species defined by those traits are all more similar to each other than to other species, then it gives you an easier way to try and define bacterial species. While it’s not perfect it does also fit with phylogenetic history.
But note what I said above: “UNLESS GENETIC DISTANCE IS CORRELATED WITH SOME BIOLOGICAL FEATURE OF INTEREST.” You’ve just listed biological features of interest that you think might be correlated with genetic distance. Therefore my argument stands.
Save those correlations, yes, the genetic-distance measure is completely arbitrary.
Interestingly, sexual animals also possess within their cells an asexually transmitted bacteria-like genome: mitochondrial DNA (mtDNA). Do mtDNA “clusters” agree with the organismal clusters recognized as species under the traditional biological species concept? The general answer appears to be “yes”, as we described in a paper published in 1999 entitled “Species realities and numbers in sexually reproducing vertebrates: perspectives from an asexually transmitted genome.”
Here’s the Abstract summarizing that paper:
“A literature review is conducted on the phylogenetic discontinuities in mtDNA sequences of 252 taxonomic species of vertebrate animals. About 140 of these species (56%) were subdivided clearly into two or more highly distinctive matrilineal (mtDNA) phylogroups, the vast majority of which were localized geographically. However, only a small number (two to six) of salient phylogeographic subdivisions (those that stand out against mean within-group divergences) characterized individual species. A previous literature summary showed that vertebrate sister species and other congeners also usually have pronounced phylogenetic distinctions in mtDNA sequence. These observations, taken together, suggest that current taxonomic species often agree reasonably well in number (certainly within an order-of-magnitude) and composition with biotic entities registered in mtDNA genealogies alone. In other words, mtDNA data and traditional taxonomic assignments tend to converge on what therefore may be ‘‘real’’ biotic units in nature. All branches in mtDNA phylogenies are nonanastomose, connected strictly via historical genealogy. Thus, patterns of historical phylogenetic connection may be at least as important as contemporary reproductive relationships per se in accounting for microevolutionary unities and discontinuities in sexually reproducing vertebrates. Findings are discussed in the context of the biological and phylogenetic species concepts.”
Very nice–clever to ask that question!
One of my graduate students did something like this as well, but applying a specific algorithm for mtDNA clustering that usually generates clusters that covary with other biological traits. Can be used to discover species or assign individuals to a known species. People use the method often.
Edit to add: I’m embarrassed to find that we didn’t cite John’s 1999 paper 🙁
https://doi.org/10.1098%2Frsbl.2007.0307
Agreed – this is a very clever question.
Jerry – Thanks for summarizing this article for us. I found it to be one of the most interesting summaries you have written. It will be good discussion material for my Intro Bio class when we get to speciation. Google Scholar says it has been cited 122 times. I would have thought more, but perhaps that it is a difficult paper to read has held back its impact. Yet another example of why writing skills are so important for the dissemination and understanding of science.
Agreed: thank you very much, Jerry. As a non-biologist with a considerable interest in evolution I have often wondered about this.
Super interesting! I feel like I need to post a long comment so that you don’t think I don’t appreciate how much time and effort went into studying and summarizing the article. (But I will resist.)
I can’t tell you how many (friendly) battles we paleontology graduate students had regarding whether bacteria were biological species back in graduate school. Mayr himself, IIRC, limited his BSC to sexually reproducing organisms, perhaps just to avoid having to deal with them. (I suspect he wrote about why somewhere in his voluminous works, but I don’t remember for sure.)
It’s way amazing indeed that reproductive isolation exists in bacteria, and the experimenters were super clever in the way they figured it out. Maybe they were a bit exuberant in their title, but I can understand their excitement.
+1
This post was very interesting to me! Until now, I regarded prokaryotic species as a concept useful for scientists but completely irrelevant to prokaryotes themselves.
Excellent stuff. This is one of the main things that the internet is for – explanations of “difficult” recent work by experts in the field.
Oh dear, so many errors.
Bacteria have several processes that sometimes move pieces of DNA from one cell to another (called ‘parasexual processes’). None of these processes exists because of selection for recombination of chromosomal genes. In each case the transfer of chromosomal DNA (always unidirectional, never ‘exchange’) is a side effect of machinery selected to transfer other DNAs between cells (self-transfer machinery encoded by conjugative plasmids, viruses and other genetic parasites) or to bring free DNA into the cell as a source of nucleotides (by ‘natural competence’).
Whether particular bacteria form species-like clusters is likely to mainly depend on the host specificities of their parasites and whether or not they are naturally competent.
Want more? You could read my 2002 ‘Do bacteria have sex?’ paper.
Fascinating article. I’m a relative novice in the field of biology, so the guidance of your summary and analysis will be very helpful when trying to read the actual article!
Thanks PCC!
Basic question (i.e. I’m not disputing anything) :
What is the distinction between “strain” and “species”?
I see the discussion above about adding strains to one species which found more clusters. I’m just wondering if I might get a clear idea of what that means –
e.g. I know E. coli has lots of strains like O157:H7 (?) that appears in the news sometimes – some bad hamburgers – but how do we know O157:H7 is still E. coli?
A “strain” is more arbitrary even than species. In clinical work it usually describes a serial clone that varies enough from typical in some interesting way that we kept it for further study instead of just tossing it into the Chlorox at the end of the day. It might be a particularly nasty strain that is resistant to all available antibiotics and we want to find out why, how it spreads, and maybe find an antibiotic that will work against it. Not all E. coli O157:H7 isolates are necessarily the same “strain” in this sense. If they came from places widely separated in space and time, you could probably find enough base-pair differences on genetic testing to call them different “strains”. If they were identical, you might still call them the “Manila strain” or the “Cincinnati strain” just to keep them straight in your collection in case you want to go back and look at them further.
In routine lab work, bacteria are identified phenotypically to species* level first, and then only the ones that check out as E. coli are further tested to see if they also have the O and H antigens of interest in diarrhea.
————-
* Here we recognize that “species” is a colloquial term intended so that clinical doctors who have not studied bacteriology since medical school will understand what the lab is trying to tell them. Making busy clinicians learn the new terms for common bacteria refined by more sophisticated DNA relatedness techniques is not fruitful.
Great post. Thank you!!
Excellent post and discussion. Perhaps because I come from a long line of bacteriologists (both parents, grandfather, great grandfather), my issue with the BSC is “but what about the bacteria”? Back in 2011, I had the pleasure of participating in the dedication of a memorial to my great grandfather (Herbert Conn, cofounder of the American Society for Bacteriology). One of the others was Fred Cohan from Wesleyan University who has done seminal research on the question of bacterial species/speciation. As I remember it, his basic model incorporated both occurrence of recombination and ecological differentiation, but I struggled to grasp the basic concepts. Back then this mattered for my teaching. Now, I may reinvestigate it just for fun.