Phylogeny and biogeography of wild roses with specific atten

Rose Genetics Ploidy Level
, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
1

“Conclusions The ancestral area reconstruction suggests that despite an early presence on the American continent, most extant American species are the results of a later re-colonization from Asia, probably through the Bering Land Bridge. The results suggest more recent exchanges between Asia and western North America than with eastern North America. The current distribution of roses from the Synstylae lineage in Europe is probably the result of a migration from Asia approx. 30 million years ago, after the closure of the Turgai strait. Directions for a new sectional classification of the genus Rosa are proposed, and the analyses provide an evolutionary framework for future studies on this notoriously difficult genus.”

The figures of the full paper are available free:

Correction, only low resolution versions of the figures are free.

2

These reports are always interesting. In particular, I note that the authors concluded, “The ancestral area reconstruction suggests that despite an early presence on the American continent, most extant American species are the results of a later re-colonization from Asia, probably through the Bering Land Bridge. The results suggest more recent exchanges between Asia and western North America than with eastern North America.”

This is also true for a wide range of plants of the eastern U.S. whose nearest relatives are found in eastern Asia.

The Rural Carolinian, 5(4): 281-282 (Feb 1873)
The Floras of America and Japan.
“Professor Asa Gray in his late address before the American Association for the advancement of science*, points out some singular coincidences between the plants of the United States and those of Japan. Our Rhus toxicodendron, or poison-ivy, is exactly repeated in Japan, but is found in no other part of the world, although a species like it abounds in California. Our other species of Rhus (R. venenata,) commonly called poison-dogwood, is in no way represented in Western America, but has so close an alliance in Japan that the two were taken for the same by Thunberg and Linnaeus, who called them both R. vernix. Our Northern fox-grape, Vitis labrusca, is wholly confined to the Atlantic States, except that it reappears in Japan and that region. Wistaria was named for a woody leguminous climber, with showy blossoms; native of the Middle Atlantic States. The other species which we prize so highly in cultivation, W. sinensis, is from China, as its name indicates, or perhaps only from Japan, where it is certainly indigenous. There is a shrub, ellittia, which is so rare and local that it is known only at two stations on the Savannah River, in Georgia. It is of a peculiar structure, and was without near relative until one was lately discovered in Japan. A specimen of a peculiar plant was detected in the herbarium of the elder Michaux, who collected it somewhere in the high Alleghany mountains, more than eighty years ago, since which no one has seen a living plant, or knows where to find one. At length the same plant is found in Japan. This close relation between the floras of the two countries may explain the mystery of our Japan Clover, (Lespedeza Striata) which, it is not unlikely, belongs to both.”

I may add that the “Rare Franklinia” (Franklinia alatamaha) has been hybridized with Asian camellias, and the Osage Orange (Maclura pomifera) with the Chinese Mulberry (Cudrania tricuspidata).

Prof. Gray’s paper can be found here:

3

Lysichitum and Campsis (larger area) nearest relatives are also on the Pacific opposite side and nowhere else.

4

What I find interesting is that the Pimpinellifoliae/Cinnamomeae and Synstylae clads split 30 MYA. It’s no wonder they are so difficult to cross with each other.

“The ancestral area reconstruction suggests that despite an early presence on the American continent, most extant American species are the results of a later re-colonization from Asia.”

According to the graph, R.minutifolia is from an entirely different line then all of the other NA roses and could be a survivor of the earlier colonization from Asia. Similar to how the Basque language is the lone survivor of an earlier language group and is not related to any other Indo-European language in Europe.

The rest of their graph has some surprises in it. R.foliolosa is in a separate group more closely related to Asian roses than to any other NA roses, where usually it’s been grouped in the Carolinae section closely related to R.nitida and R.palustris.

And I do question how they have R.woodsii, R.blanda and R.nitida related. It makes me wonder if they got R.nitida and R.woodsii mixed up. First they have R.nitida recently splitting with R.blanda where R.woodsii should be and not closely related to the other Carolinae roses. Then they have R.woodsii on the branch with R.virginiana, R.carolina and R.palutris where R.nitida should be. And finally R.woodsii var. woodsii and R.woodsii var. ultramontana split from each other 2.5 MYA which probably is not true since they are likely regional variations of the same species.

5

I liked Bruneau and Joly’s in depth work on the carolinae, which doesn’t seem entirely consistent with this new graph. Before they said that “In the east, two groups of diploids were found: one consists of R. blanda and R. woodsii and the other of R. foliolosa, R. nitida, and R. palustris. Only eastern diploids are involved in the origins of the polyploids. Rosa arkansana is derived from the blanda–woodsii group, R. virginiana originated from the foliolosa–nitida–palustris group, and R. carolina is derived from a hybrid between the two diploid groups.” If you look at this new chart, those groups don’t seem to exist, and at various points reconciling the papers is hard. Like figuring out which the two diploid parents of arkansana were, when they were from the blanda-woodsii group, but blanda (we’re now told) hadn’t evolved yet and woodsii isn’t very genetically close. Do I need to assume the involvement of extinct diploid species, or is there a more sensible way to harmonize these papers?

6

Yes, the graphs do have some surprises.

I don’t know how important the ploidy of a species is to the conclusions reached about its relation to other species in this study, but both specimens of Rosa laxa, Retzius in Fig. 2 are labeled as tetraploid. Some years ago R. laxa, Retz was routinely listed as tetraploid, and it is not impossible that tetraploid R. laxa, Retz. might exist. However, all the samples of R. laxa, Retz. we’ve checked are diploid–including those descended from stock collected by Hansen, those descended from plants grown from seeds sent from Kew Gardens to Canada’s Central Experimental Farm in 1897, and those grown in 2009 from seeds collected in Mongolia in more recent times and obtained from USDA’s GRIN. Flora of China says that R, laxa, Retz is diploid (2n=14), and various other studies also have found that their specimens of R. laxa, Retz. were diploid.

Only two specimens are included in the current study, and they come from the same garden, Maowen Rose Garden in Sichuan. I wonder whether other species in the study were representatively sampled.

The connections a machine makes depend on the data given to the machine.

Peter

7

It’s one thing to have a firm handle on phylogeny. It is entirely another to try to use that information wisely in a breeding program.

Until now I have relied on the Bayesian dendrogram given in Figure 3 in Koopman et al (cited on p.15 of Fouge`re-Danezan et al.) as the best-guestimate phylogeny for roses. This new paper finally shows positioning for R. chinensis spontanea which has been missing from Koopman and most other reliable studies so it really fills a big void for planing breeding strategies. A surprise in that regard is that multiflora appears to be not-so-close to chinensis as I had assumed (and that the American multiflora seems to actually be a ‘horticultural hybrid’ rather than the pure Asian species).

The genetic isolation of Hesperhodos (minutifolia is given but they seem to have omitted stellata?) is no surprise but looks to be s-o-o far out on a limb as to be in another forest.

My biggest take-away from a first-pass reading is the assignment of praelucens as a hybrid of roxburghii (a Platyrhodon) and an unknown Cinnamomeae which will inform future choices of partners for my roxie and moyesii F1’s.

8

Papers like these are wonderful and help add new information to try to clarify phylogeny. They should be valued in light of their context (what they sequenced and very importantly how representative what they chose is for their rose genotypes) and placed into the greater framework of other understanding (other molecular work, morphological data, crossability, etc.). All that information together gives us a more complete picture. DNA sequence data is very very valuable of course and tells us things the other kinds of information can’t, but there are limitations with it. With studies like this they look at very limited and strategic areas of the genome, which is reasonable at this point in time and it being resource intensive to sequence large regions of the genome. One needs to aim for genes that tend to be conserved and not change too quickly or be too variable hopefully in a species (studies like this often have very limited representatives of a species and what they specifically choose may or may not be representative, which can introduce problems too), but on the flip side hopefully these regions/genes vary enough that over time that at a species level we can pick up divergence. It is nice they looked at and compared multiple genes to add extra information to their proposed interrelationships of the species.

My rambling here is my thinking out loud trying to put into perspective the questions we are having with some of the proposed interrelationships.

9

Remember to take all phylogenies with a grain of salt, or two. Finally enough birds species have had near complete DNA sequences done so that a fairly decent association can be made to identify a likely tree. Just published in Science a couple weeks ago.This despite centuries of study and striking morphological differences to work with. With roses, as suggested, the purported species may be so similar that they are not correctly assigned names by traditional methods. Hence we may have a GIGO situation. I would want to have full sequencing of whole chromosomes at minimum, and these days we could do it from voucher specimens- even original type specimens collected by prominent authorities. At very least I’d want a reliable taxonomist to verify the correct ID of every specimen analyzed. If I’m not mistaken, Lewis has done a pretty thorough revision of the North American species. There must be lots of specimens that he authenticated. Same would be true for EIleen Erlanson’s work at a couple universities where she was for many years. I could go on, but the point is, more, better work is needed using deep sequencing and next-gen sequencing methods that are getting cheaper by the day. We must also apply unbiased fitting methods.

10

Ploidy can be relevant if the tetraploids are differential (composed of chromosomes from two diploid species). ‘Basye’s Amphidiploid’ (4x) vs. R. rugosa, for example. Autotetraploids would be a different matter: R. macrophylla (2x) vs. R. macrophylla ‘Korolkowii’ (4x).

Furthermore, there is the possibility that a differential tetraploid may give the occasional differential diploid offspring. This was reported by Crane and Thomas (1949). The Veitchberry (4x) was derived from a diploid blackberry and a tetraploid raspberry. Its offspring did not revert to one or the other parent. When pollinated by a diploid raspberry, it produced some diploid offspring that were intermediate between blackberry and raspberry. If this occurred in the wild, one could conclude that the Veitchberry “species” exists in both diploid and tetraploid forms … if one also ignored the “pure” blackberries and raspberries in the vicinity.
http://bulbnrose.x10.mx/Roses/Hurst/CraneVersatility1949/CraneVersatility1949.html

But on the subject of R. laxa:

Iowa State Journal of Science: 35: 255-260 (1960)
Some notes on the use of Rosa laxa as a source of hardiness in rose breeding
Griffith Buck
Pollen from the ‘Crimson Glory’-> R. laxa > seedling was effective in producing viable seed on a wide range of cultivars of the Hybrid Tea, Floribunda, and Grandiflora garden classes. The seedlings segregate into two sharply defined groups of approximately equal numbers. One group bears a pronounced resemblance to > R. laxa > in growth habit, foliage and prickle characteristics. The flowers, which are borne only in June, are single, two to three inches in diameter, and are in the lighter tones of pink, salmon, and yellow. All the plants in this group are as hardy as the species parent. The plants of the second group resemble the garden rose parent in floral and foliage characteristics. The plants are June-blooming; the everblooming habit of the Hybrid Tea and the remontance of > R. laxa > being absent. All plants of this group winter-killed during the winter of 1958-59, even though they had been given winter protection. All the seedlings retained the freedom from powdery mildew and the susceptibility to blackspot of the ‘Crimson Glory’-> R. laxa > parent.

http://bulbnrose.x10.mx/Roses/breeding/Buck/BuckLaxa1960.html

The fact that “Laxa-ness” was inherited as a unit strongly suggests (a) Buck’s clone a diploid, and (2) “Laxa-ness” is likely associated with a single chromosome.

As Buck wrote, other traits from R. laxa (PM and BS resistance) were passed to grandchildren that did not look Laxa-ish.

David Austin (1993) had a similar experience with Rugosa-ness:

The third line we pursued was by way of the Rugosa hybrid ‘Conrad Ferdinand Meyer.’ At first we harbored no great hopes of success, for we feared that the resulting seedlings from a cross with this excessively vigorous hybrid would be altogether too gross in character. ‘Conrad Ferdinand Meyer’ was itself a cross between the very popular and beautiful Climbing Noisette Rose ‘Gloire de Dijon,’ and an unknown Rugosa hybrid. It also had one of the most powerful and delicious fragrances. As before, we crossed with some of our better English Rose in particular ‘Chaucer,’ and had one of those pieces of luck that sometimes turn up in rose breeding. Some of the seedlings from this cross were of typical rugosa appearance, while others bore absolutely no resemblance to a Rugosa Rose. It seemed that some of our hybrids had taken the genes only from the ‘Gloire de Dijon’ half of ‘Conrad Ferdinand Meyer,’ while others had inherited those from the Rugosa side. What we had in many instances were in effect hybrids of ‘Gloire de Dijon.’

There is disagreement over the parentage of ‘CFM’, but the inheritance of Rugosa-ness as a "unit character’ is the interesting bit (to me).
Karl

11

I suppose I should reduce my salt intake. Still, I take your point. One thing that all these phylogenies seem to share is that they are all different. And generally the authors draw conclusions that involve implicit assumptions. For example, do species evolve in isolation? Probably not. Are chloroplasts immune to the effects of natural selection? I don’t know.

I do think it would be interesting to see a map of chloroplast genotypes, ignoring (for the moment) the species in which they are found. Would there be a pattern? Would some genotypes be found more commonly in damp woodlands, others in the prairies, and still others in deserts?

The same principle might apply to DNA snippets. The bits studied might not have selective value, but they could be linked to other bits that are useful to the plants in some conditions.

Nevertheless, the various studies raise some interesting possibilities. I recently learned a couple of things that surprised me. For one, Fragaria (strawberries) has been merged with Potentilla. I must have been asleep that year. And according to one DNA-phylogenetic study, Rosa is closer to Potentilla/Fragaria than it is to Rubus. This is striking because Burbank raised hybrids from Raspberry x Strawberry, which are (supposedly) more remote than Rosa and Fragaria.
http://bulbnrose.x10.mx/Heredity/Burbank/Burbank_raspXstraw.html

Does this mean we could breed yellow roses from yellow cinquefoils? I wouldn’t bet, but you never know. The flowers of Potentilla reptans contain β-Carotene, 5,6-monoepoxyde of α-carotene, 5,6:5′,6′-diepoxide of α-carotene, 5,6:5′,6′-diepoxide β-carotene, cryptoxanthin, epoxide of cryptoxanthin, mutatochrome, lutein, trans-epoxide of lutein, cis-epoxide of lutein, flavoxanthin, auroxanthin.

Oh, and one more thing. I had never heard of Dalibarda, an herbaceous plant with heart-shaped leaves (like a violet), and both chasmogamous and cleistogamous flowers (like a violet). Now I learn that it nests within Rubus (DNA-wise), and has been taxonomically reunited with that genus.

12

That’s really interesting that potentilla/strawberry seems to be more closely related to Rosa than Rubus (raspberries/blackberries, etc.). There are commercialized potentilla/fragaria hybrids that bring pink flower color from I think herbaceous potentilla into ornamental strawberries. Potentilla, strawberry, and rose all have aggregate fruits consisting of achenes unlike Rubus which is an aggregate fruit of drupes (basically like little cherries all fused together with a hard “pit” or endocarp). Basically a strawberry is an inside out rose hip. Hopefully with the strawberry genome sequenced (rose is on the way it sounds like with the French group taking the lead) maybe it can serve as a better model to probe for common genes contained in rose and other structural commonalities that Rubus.

13

Does this mean we could breed yellow roses from yellow cinquefoils?

I have not yet tried it although several years ago I did acquire a potentilla specifically for that purpose. Maybe this is the year to give it a go.

14

I tried crossing roses with potentilla. No success, but that does not mean that it could not be done as there are probably an infinite number of crosses and growing conditions to vary.

15

I tried to find the report, but came up with a different one - with different conclusions. Potter et al. (2007) rejected the merger of Fragaria and Potentilla. “Fragaria is not combined with Potentilla, as suggested by Mabberley (2002), because it would make Potentilla polyphyletic.”

They also disagree with the previous report I read. “In Rosoideae, … Rosa and Rubus are both included in Rosodae but not in any tribe, and Potentilla is included in Potentilleae but not in any subtribe, although the remaining genera are placed in Fragariinae.”

In addition, they concluded that Gillenia (x=9) is sister to all the pomes (x=17) and other species with x=15. Which is to say that all the pears, apples, medlars, loquats and such are (presumably) descended from American ancestors.

So, whatever theory or model one prefers, there is likely to be some highly technical support.

16

For example, do species evolve in isolation? Probably not.

Reticulation has already been proved in roses, if that’s what you are getting at. I’d go a step further and anticipate that phylogenetic trees will be proven obsolete by DNA and that we’ll eventually have a proven network model rather than an hierarchical tree model for the evolution of roses.

17

Right. And in this, there is the possibility that organelles (mitochondria, chloroplasts) might be selectively favored in some environments. I mean that organelles might be introgressed. A “hybrid swarm” descended from a rare hybrid could be swallowed by the more common species. If one of the organelles provided a selective advantage, it might displace the original in the predominant species. Thus, the similarity of chloroplast DNA would tell more about a relatively recent contact than about the truly ancient phylogeny of the species.

The only example I can think of is what happened in corn (maize) after cytoplasmic male sterility became fashionable among corn breeders. The various inbred lines used to raise “Hybrid Corn” were bred into the CMS line. When the Southern Corn Blight abruptly headed north in 1970, breeders were alarmed. Researchers soon figured out that the mitochondrial mutation responsible for male sterility also gave a high susceptibility to the blight. And so, that CMS line was abandoned (mostly). Which is to say that in 1970, most of the corn raised commercially in the U.S. shared the Texas Cytoplasm. A few years later that cyto-type was mostly gone, though the strains of cultivated corn did not change very much.

Whether this sort of thing can occur with different types of chloroplasts is another question.

And as I have noted previously, in some cases that have been studied, many of the distinguishing characteristics of a species may sometimes involve a single chromosome. Other chromosomes contribute to other aspects of the plants, of course. But diagnostic characters that are linked in one chromosome may “introgress” other species with different histories. That is, the various chromosomes in one supposed species may have come from different sources, with only one pair dominating the “species-ness”.

So, I do agree that full sequencing will very likely tell us a different story than we’re getting from the various “snapshots” that are currently in print. And when the sequencing gets down to individual chromosomes, I anticipate even more surprises.

18

For example, do species evolve in isolation? Probably not.

Why not?

At least there is a derivation as it is obvious considering some island species with a different species for each island.
Same for mountains.
Same for relictual populations of a once widespread (invasive?) species such as white Oaks around Mediterrannea.

19

Pierre,
I was not as clear as I should have been. By “isolation” I meant not only isolation from other species, but isolation from the environment. In the back of my mind I was remembering a brief email conversation I had 20-odd years ago with a biochemist. He expressed his contempt for the “phenetic” aspects of heredity. To him, “evolution” meant “molecular evolution”. Period. End of discussion.

It is now pretty well established that the phenotype can change dramatically with little change in the genotype. For examples, marsupials have much the same gene-set as placental mammals, but regulate the genes differently. And in the early 19th century, Van Mons made a side-career of breeding fruit trees without deliberate crossing. Wild pears needed 10-12 years from seed to first fruits. Van Mons bred them, by a variety of techniques, to bear in as few as 3 or 4 years from seed. He was just as successful with roses, apples, peaches, and so on. It is hard to imagine that “random gene mutations” were happening in every breeding line he developed.

Del Pozo, et al. (2000) studied the Burr Medic (Medicago polymorpha) in Chile and its native Sardinia.
“Flowering time appears to be a common adaptive trait in annual legumes along aridity gradients, provided that they have sufficient seed dormancy (e.g. Ehrman and Cocks, 1996; Piano et al., 1996). Here it has been shown that, along the 1000-km-long aridity gradient of the Mediterranean-climate region of central Chile, clear ecotypic differentiation in the reproductive phenology of burr medic has occurred, allowing adaptation to the various bioclimatic zones present. This has occurred despite the fact that very little genetic diversity has been detected in populations from Chile (Paredes et al., pers. comm.) or from Sardinia (Bullita et al., 1994).”

It is apparent that this adaptive shift in flowering time took place with little or no genetic modification. Genomic shock is the most plausible explanation, so far as I can see.

20

Genomic shock

Do tell.