Intrapopulation gene flow, the Leptoxis of the Cahaba, and the striking of matches

Editor’s Note – This essay was subsequently published as: Dillon, R.T., Jr. (2023b)  Intrapopulation gene flow, the Leptoxis of the Cahaba, and the striking of matches.  Pp 133 – 146 in The Freshwater Gastropods of North America Volume 6, Yankees at The Gap, and Other Essays.  FWGNA Project, Charleston, SC.

Last month we reviewed a research project I conducted in the late 1970s on isolation-by-distance and barriers to dispersal in a population of Pleurocera proxima inhabiting a small stream in NW North Carolina [1].  The markers I used were allozyme variants – bands of protein migrating at different speeds in starch gels.  Even back in the 1970s, there was already a lot of hand-wringing about the genetic variation being missed by such a gross and clumsy technique.

I was missing, of course, all of the silent variation – variation due to the redundancy of the genetic code.  And of the non-redundant variation, mutations that yield variation in the amino-acid sequence, I was only catching a tiny fraction – that subset changing the charge on the protein.  And the technique I was using worked only on loci that encode enzymes, the evolution of which must certainly be constrained by selection, right?  What about junk DNA?  What about everything else? 

A lot of my colleagues worried about these problems in the 1970s.  And I can see why, if one invested in all that cumbersome equipment for protein electrophoresis, and all those expensive reagents to demonstrate allozyme bands, and conducted one’s field work, and ran one’s gels, and found no variation, and scraped one’s gels into one’s trash can, and washed one’s giant sink-full of dishes, one might grumble.

Allozyme variation in Naked Creek

But the grossness and clumsiness of allozyme electrophoresis never bothered me, through my entire career, even into the 21st century [5] because I did find allozyme variation.  With hard work, patience, and good technique, I found a lot of very useful genetic markers.  I realized that I was “binning” a bazillion silent variants together when I scored a snail homozygous for Odh106 back in 1979.  But I could distinguish those bazillion variants together from the bazillion silent variants I had binned together in snails I scored as homozygous Odh109F.  And those two big bins were inherited in Mendelian fashion.  Fine.

Well, technology marches on.  The first fully automated sequencing DNA machine came onto the market in 1987, even as Kerry Mullis’ patent for PCR amplification was approved, and by the 1990s everybody was sequencing DNA.  Massively-parallel (“Next-Generation”) DNA sequencing machines were introduced for commercial use around 2005, enabling rapid and cheap sequencing of gigabases of DNA.  And in 2008 a method was first proposed to identify single-nucleotide polymorphisms in random lengths of DNA amplified from population samples, “Restriction Associated DNA sequencing,” or RADseq for short.

The DNA from each individual (let’s say, for example, each snail) is isolated and cut into a bazillion little pieces with a carefully-chosen restriction enzyme or set of enzymes.  The millions of pieces in this mess that are (usually) a couple hundred base pairs long [6] are electrophoretically separated from the zillions of littler pieces and bigger hunks and ligated to adapters that facilitate their amplification and uniquely identify each individual snail.  These millions of pieces are called “reads.”  Feed all those reads through the front door of your local Next-Generation sequencing factory.

As mind-boggling as the previous paragraph most certainly is, I will now ask you to imagine repeating that process for (let’s say) 20 snails from (let’s say) 8 populations.  So 160 times.  And as mind-boggling as the preparation and the sequencing of all those 160 million reads most certainly is, I will now ask you to imagine that all of those reads can be “quality-filtered” to remove the crappy ones, screened by their utility as markers across the entire 160-individual population, and analytically matched to each other using some gargantuan gear-grinding smoke-belching computer program.  If you’re interested in the technical details, see the references at footnote [7] below.

What comes back out of the sequencing factory is, in our example, a comparison of 8 sets of 20 snails amplified at some huge number of random, anonymous (“restriction-site associated”) reads of DNA.  If there is any polymorphism, even for one single synonymous nucleotide, the researcher will know it.  The data are typically reported in tightly-packed bar graph form, standardized by the number of matching reads, coded by some color, let’s say orange to start.  If snail #1 and snail #2 are genetically identical, as far as can be determined by this mind-boggling technique, the bar graphs depicting their genomes will match in orange along 100% of their height.  If snail #1 and snail #2 differ by (let’s say) 40%, a second color is selected, let’s say blue, and the bar graph depicting snail #2 is 60/40 orange/blue.

So in 2019 our colleague Nathan Whelan, together with six coworkers, published the first RADseq study of a population of pleurocerid snails [8].  His results nicely augment and compliment the results I myself obtained using allozyme variants in the 1970s, and as such make an important contribution to our understanding freshwater gastropod evolution generally.

Whelan et al. Figure 2 [8]

Nathan collected samples of 20 Leptoxis from each of 8 sites in the Cahaba River drainage of central Alabama, stretching approximately 45 km (30 miles) from the Helena suburbs at County Road 52 (“CR52”) downstream to Centreville.  Four of his eight samples came from the main river, and four from the tributaries, as shown in his Figure 2 reproduced above.  This is a marvelously data-rich figure, from which those of us interested in pleurocerid evolution can learn at least four important lessons.  Let’s unpack each lesson, one at a time.

First, all eight of Nathan’s samples (let’s call them “subpopulations”) were genetically unique – some entirely unique, others just mostly unique.  I am not a fan of Nathan’s palate, but what he is trying to convey with his four shades of blue and four shades of orange is eight unique genomes associated with eight distinct Leptoxis subpopulations.  And the shade of orange 100% covering the Shades Creek box is completely different from the shade of orange 100% covering the Schultz Creek box.  And both of those oranges are completely different from the shade of blue 100% covering the CR52 box and the shade of blue 100% covering the Bulldog Bend box.  Those four subpopulations seem to be entirely unique, as far as Nathan’s data extend.  The other four subpopulations – Schultz Creek, Marvel Slab, Sixmile, and Centreville, are mostly unique, with other shades of color more or less impinging around the edges of their mostly-uniquely-colored box interiors.

Apparently, gene flow among most of Nathan’s subpopulations of Leptoxis is zero, at least on the time scale of single-nucleotide mutations.  That is the most surprising result of the 2019 study conducted by Whelan and colleagues.  All eight of their subpopulations are connected by water, with stepping-stone distances generally in the 10 – 30 km range.  The research we reviewed in September would have led us to expect slow but measurable active migration upstream, with rapid and episodic transport downstream [9].  My previous research on Pleurocera [1] suggested an average gene flow of 6.5 migrants among sites sampled 1 kilometer apart which, one would assume, ought to extend 10 km to some value greater than 0.0.  No?

The second important lesson conveyed by Nathan’s Figure 2 is that what little dribble of gene flow does occur among a few of his subpopulations seems to occur 100% downstream.  The four completely-unique subpopulations (Shades Creek, Schultz Creek, CR52, and Bulldog Bend) are from the four most upstream sites.  All four of Nathan’s downstream boxes show at least a little bit of upstream color.  One’s eye is especially drawn to the box at Nathan’s downstream-most subpopulation, Centreville, which is primarily painted dark blue, but demonstrates significant bars of NWR yellow, Schultz Creek dark orange, Shades Creek orange, Marvel Slab blue, and Bulldog blue.

More than any other pleurocerid, Leptoxis populations are strongly associated with rock substrate and rapid water flow.  Individuals of the genus Pleurocera, by contrast, are at least occasionally observed grazing across softer substrates.  This includes P. proxima, which although apparently adapted for small, trout-stream-sized creeks tumbling through the southern Appalachians, is not uncommonly spotted crawling on sand and firmer mud.  See the last photo I published in last month’s post [12Oct21].

But I can never, in my 60 years of field experience, ever remember collecting an individual Leptoxis on anything but rock.  So, Nathan’s eight Leptoxis subpopulations were collected from eight riffle areas that must (inevitably) have been separated by pool areas, with extensive bottoms of mud substrate.  Leptoxis can wash downstream through such pools but (apparently) cannot effectively crawl upstream through them.

Then how did Leptoxis get upstream in the first place?  The two answers to that question are great age + dirty birds.  While the lower regions of the Mobile Basin were yet covered by the Cretaceous embayment, the upper Mobile Basin had long been flowing free from the mountains of what is now North Georgia.  In 2009 I offered several lines of evidence suggesting that the pleurocerid populations of this region are “The Snails The Dinosaurs Saw,” living fossils of great antiquity [10].  I subsequently penned a series of essays showing that aerial dispersal among such populations is not as unlikely as one might think [11].

Whelan et al. Fig 1, modified [8]

And in 2016, Nathan together with our colleague Ellen Strong published a paper documenting extensive mitochondrial superheterogeneity among these same Cahaba River populations of Leptoxis for which he and his co-workers now report RADseq data [12].  Nathan’s 2016 results strongly imply very long periods of isolation, punctuated by very rare introductions of genomes from very great distances away.  Now in light of Nathan’s 2019 research findings, his 2016 paper makes more sense. 

The third important lesson to be taken from Nathan’s RADseq study is that divergence among Leptoxis subpopulations of the Cahaba is phenotypic, as well as genotypic.  The shells born by most of Nathan’s eight samples were almost entirely smooth, as shown in (B) of his Figure 1, modified above.  But the subpopulation inhabiting the Little Cahaba River below Sixmile Creek bears lightly tuberculate shells, typically with carination, as shown in (C).

In the figure below I have reproduced my diagram of Naked Creek from last month’s post and inset a slice of topographic map showing the Little Cahaba River between Bulldog Bend and Sixmile Creek.  These two maps are depicted at the same scale, see the Naked Creek scale bar at upper left.  The water distance from Bulldog to Sixmile is a serpentine 7.85 km (flowing from right to left), comparable to the distance between Naked Creek Site 7 and Naked Creek Site 8.

The Bulldog Bend box is 100% blue, demonstrating no gene flow from any other subpopulation sampled.  The Sixmile box is almost entirely orange but shows a bunch of little Bulldog-blue nibbles at the bottom, plus one big dramatic blue cut.  That singleton snail, the one individual whose genotype seems [13] to match Bulldog more than Sixmile, also bore a smooth shell (B) like the Bulldog subpopulation, not a tuberculate/carinate shell (C) like the other 19 in Nathan’s Sixmile sample.

Clearly this phenomenon is attributable to washdown gene flow from Bulldog to Sixmile.  The Smooth-Shelled Singleton in Nathan’s Sixmile Sample (Let’s call him “5S.”) did not demonstrate a 100% Bulldog genome, however, but rather only about 50% [13] matching Bulldog.  The implication is that Snail 5S is not a first-generation washdown, but a second generation washdown, born at Sixmile but retaining the shell morphology along with half the genome of a Bulldog parent.  This is indirect but nevertheless compelling evidence for the heritability of tuberculate/carinate shell morphology in pleurocerid snails.  But wait, there’s more.

The final lesson from Nathan’s RADseq study, and the most important lesson, is this.  Although these eight subpopulations have diverged both genetically and morphologically, they have not speciated.  The isolation between them is physical, not reproductive.  When snails wash down, albeit rarely, they are apparently able interbreed freely with the snails in the riffles downstream.  The Cahaba River at Centreville is not populated by an admixture of five different Leptoxis species.  All 20 of the snails Nathan collected at Centreville belonged to the same biological species as the seven subpopulations Nathan sampled upstream.

Nathan identified all eight of his subpopulations as “Leptoxis ampla.”  OK, that’s a good start.  We all agree on the conspecific status all the Leptoxis subpopulations of the Cahaba.  Now let us see if we can apply the lessons we have learned in the Cahaba to the greater Mobile Basin beyond.

I have reviewed the tangled taxonomic history of the Mobile Basin Leptoxis fauna on several occasions in the tangled epistemological history of the FWGNA blog [15].  But not recently.  So to refresh the collective memory.

Timothy Abbot Conrad got the ball rolling back in 1834, describing four species, two from the Alabama/Coosa and two from the Black Warrior.  Isaac Lea [16] added seven, J. G. Anthony added three, H. H. Smith added eleven, and Calvin Goodrich [17] one, so that by 1922, Goodrich tallied 26 nominal species of Leptoxis in drainages of the Mobile Basin [18].  I reproduced Goodrich’s figure of all 26 in my essay of [15Sept09] and have re-reproduced it below.

Most of these nominal species were nominally-extincted by extensive damming and impoundment conducted throughout the Mobile Basin, starting in 1912, accelerating in the 1920s and 1930s, and continuing into the 1960s.  By the 1990s, Goodrich’s Leptoxis list had been reduced to four nominal species, each inhabiting small fragments of its former range: L. picta (Conrad 1834) in the main Alabama River, L. ampla (Anthony 1855) in the Cahaba, L. taeniata (Conrad 1834) in the lower reaches of three creeks in the Coosa drainage, and L. plicata (Conrad 1834) in the Black Warrior.

Goodrich [18]

How many of these might be biologically valid?  In 1998 I published a paper with Chuck Lydeard [19], reporting that the allozyme divergence among L. picta, L. ampla, and L. taeniata was no greater than the allozyme divergence among a set of conspecific Leptoxis praerosa controls sampled from equally-distant quarters of Tennessee [20].  We suggested that ampla and taeniata be synonymized under picta (Conrad 1834).

Modern fashion has trended in the other direction, however.  Even as our 1998 paper was in review, a pleurocerid population identified as “Leptoxis downei” was discovered in the Oostanaula River of Georgia, a nomen subsequently dropped in favor of L. foremani.  And in 2011 a population identified as Leptoxis compacta was discovered in the Cahaba River at the Shades Creek confluence, sympatric with snails Nathan identifies as L. ampla [21].  Today the list of nominal Leptoxis species inhabiting the Mobil Basin has rebounded to six [22].

Now research results have crossed our desk demonstrating that the Leptoxis population of the Cahaba River is strikingly fragmented into subpopulations, that these subpopulations have diverged both genetically and morphologically, and that they have not speciated.  Another 50 km downstream will bring us to the main Alabama River, and 100 km back up the Alabama/Coosa will bring us to the mouth of Buxahatchee Creek.  Does this new evidence support the assignment of three different specific nomina to “Leptoxis ampla” in the Cahaba, “L. picta” in the main river, and “L. taeniata” in Coosa tributaries such as Buxahatchee Ck?

Forty years ago, while I was yet a doe-eyed graduate student, it was clear to me that the key to understanding speciation was to understand population divergence, and the key to understanding population divergence was to understand intrapopulation gene flow.  I did not have any big grants, and I did not have any fancy tools, and I did not have legions of collaborators.  But I did have, even at that tender age, quite a few years of field observation on the biology of pleurocerid populations in rivers of the American southeast, and an openness to learn more, and that took me a long way.

Now I am delighted to discover colleagues in Alabama bringing sophisticated tools to bear on questions I myself pondered in my youth.  Can my colleagues extend their newfound understanding of intrapopulation gene flow forward through population divergence and generalize to the species level?  Can they bring dawn to the darkness that has enveloped the pleurocerid fauna of the Mobile Basin for 200 years?  That remains to be seen.  But Nathan Whelan and his colleagues have struck the first match.

Notes 

[1] My research project on gene flow in the Naked Creek population of Pleurocera proxima was ultimately published in three different ways.  The 1979 results, in their entirety, were published as Chapter 2 of my 1982 dissertation [2].  The barrier-to-dispersal portion of that 1979 study was combined with data from 1980 and 1985 and published in 1988 [3].  The isolation-by-distance portion of my 1979 study was published just last year in Ellipsaria [4].  For an overview of the entire research program, see last month’s post:

• Intrapopulation gene flow: The polymorphic Pleurocera of Naked Creek [12Oct21]

[2] Dillon, R.T. Jr (1982)  The correlates of divergence in isolated populations of the freshwater snail, Goniobasis proxima.  Ph.D. Dissertation, University of Pennsylvania. 182 pp.  Dissertation Abstracts 43: 615B

[3] Dillon, R.T., Jr. (1988) The influence of minor human disturbance on biochemical variation in a population of freshwater snails. Biological Conservation 43: 137-144.  [PDF]

[4] Dillon, R. T. (2020) Fine scale genetic variation in a population of freshwater snails. Ellipsaria 22(1): 24 - 25.  [PDF]

[5] I was still frantically running allozyme gels when I was kicked out of my lab at the College of Charleston in the spring of 2016.  And still getting interesting results, too!  See:

• The best estimate of the effective size of a gastropod population, of any sort, ever [14Jan19]

[6] Actually, Nathan and his colleagues used a clever modification called 2b-RADseq, involving a special restriction enzyme called ALF1, that cuts DNA into fragments exactly 36 bp long.

[7] A few of the better references on RADseq:

• Baird N, Etter P, Atwood T, et al. (2008) Rapid SNP Discovery and Genetic Mapping Using Sequenced RAD Markers. PLoS ONE 3:e3376.
• Davey, JW & M.L Blaxter (2010) RADSeq: next-generation population genetics.  Briefings in Functional Genomics 9: 416-423. doi: 10.1093/bfgp/elq031
• Rubin, B.E.R., R.H. Ree, and C.S. Moreau (2012)  Inferring phylogenies from RAD sequence data.  Plos One 7(4): e33394.
• Wang, S, E. Meyer, J.K. McKay, and M. Matz (2012)  2b-RAD: A simple and flexible method for genome-wide genotyping.  Nature Methods 9: 808 – 810.

[8] Whelan, N.V., M.P. Galaska, B.N. Sipley, J.M. Weber, P.D. Johnson, K.M. Halanych, and B.S. Helms (2019)  Riverscape genetic variation, migration patterns, and morphological variation of the threatened Round Rocksnail, Leptoxis ampla.  Molecular Ecology 28: 1593 – 1610.

[9] For a review of previous research on this important topic, see:

• Intrapopulation gene flow: King Arthur’s lesson [7Sept21]

[10] Dillon, R.T., Jr. and J.D. Robinson (2009)  The snails the dinosaurs saw: are the pleurocerid populations of the Older Appalachians a relict of the Paleozoic Era?  Journal of the North American Benthological Society 28: 1 – 11 [PDF].  For more, see:

• The snails the dinosaurs saw [16Mar09]

[11]  My four-part series on aerial dispersal:

• Freshwater gastropods take to the air, 1991 [15Dec16]
• A previously missed symbiosis? [11Jan17]
• Accelerating the snail’s pace 2012 [24Apr17]
• Freshwater snails and Passerine Birds [26May17]

[12] Whelan, N.V. & E. E. Strong (2016)  Morphology, molecules and taxonomy: extreme incongruence in pleurocerids (Gastropoda, Cerithiodea, Pleuroceridae). Zoologica Scripta 45: 62 – 87. For an independent analysis of these fascinating results, see:

• Mitochondrial superheterogeneity: What we know [15Mar16]
• Mitochondrial superheterogeneity: What it means [6Apr16]
• Mitochondrial superheterogeneity and speciation [3May16]

[13] The singleton blue streak in the orange Sixmile box seems to slice much more than halfway through.  Possibly 75 – 80%?  I feel sure that’s just slop [14].  If a snail collected at Sixmile doesn’t bear 100% of the Bulldog genome, it must bear 50% or less.

[14] The only other explanation would be that Snail 5S has one parent and one grandparent washed down from Bulldog.  In other words, the mother of 5S washed down from Bulldog and was inseminated by a father with one parent born at Bulldog, yielding Snail 5S, with 75% Bulldog genome. 

That scenario seems quite unlikely.  The proportion of first-generation Bulldog-washdowns at Sixmile seems to be less than 1/20 = 0.05, and the proportion of second-generation washdowns approximately 1/20 = 0.05, so the likelihood of a first-generation x second-generation mating must be less than 0.05^2 = 0.0025. 

The implication of the highly-unlikely scenario outlined above would be that the father of Snail 5S actively sought the mother of 5S.  In other words, there is some sort of reproductive isolation between the Bulldog and Sixmile Leptoxis populations.  The two populations have speciated.

Might a unique species of Leptoxis have evolved on the Sixmile rapids in the middle of the Cahaba? Nah.  Just look downstream at Centreville.  I count six snails bearing hunks of Sixmile genome together with native Centreville genome, upstream Schultz Creek genome, and everything else.  So I agree with Nathan on this one.  The apparent excess in the length of that skinny blue cut in the orange Sixmile box is almost certainly just slop.

[15] For additional background on the taxonomy of Leptoxis populations in the Mobile Basin:

• Mobile Basin I: Two pleurocerids proposed for listing [24Aug09]
• Mobile Basin II: Leptoxis lessons [15Sept09]

[16] For the record:

• Isaac Lea Drives Me Nuts [5Nov19]

[17] Actually, if you’re digging around in these footnotes for more homework, I would recommend reading my 2007 biographical sketch of Calvin Goodrich before reading the two Mobile Basin essays I posted in 2009.  Start here:

• The Legacy of Calvin Goodrich [23Jan07]

[18] Goodrich, C. (1922) The Anculosae of the Alabama River Drainage.  University of Michigan Museum of Zoology Miscellaneous Publication 7: 1 – 57.

[19] Dillon, R.T., and C. Lydeard (1998) Divergence among Mobile Basin populations of the pleurocerid snail genus, Leptoxis, estimated by allozyme electrophoresis.  Malacologia. 39: 111-119. [PDF]

[20] Leptoxis plicata populations of the Black Warrior appear to be genetically distinct.

[21] Whelan, N.V. P.D. Johnson and P.M. Harris (2012)  Rediscovery of Leptoxis compacta (Anthony 1854) (Gastropoda: Cerithioidea: Pleuroceridae)  PlosOne 7(8) e42499 [html]

[22] Shelton-Nix, E. (2017)  Alabama Wildlife, Volume 5.  University of Alabama Press, Tuscaloosa. 355 pp.