Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
1
Title: Eocene origin, Miocene diversification and intercontinental dispersal of
1
the genus Drosera (Droseraceae)
2
3
4
Sandeep Sen*1, Neha Tiwari1 and R Ganesan1
5
6
7
1Suri Sehgal Center for Biodiversity and Conservation, Ashoka Trust for Research in Ecology
8
and the Environment (ATREE), Bangalore, India
9
10
*Author for correspondence:
11
12
Sandeep Sen. Email: sensand@gmail.com, ORCiD- https://orcid.org/0000-0001-8804-8786
13
14
15
16
Acknowledgements: Research grants from the Department of Biotechnology, India for the
17
research Project Bioresource and Sustainable Development in Northeast India
18
(BT/01/17/NE/TAX Dt 29 March 2018), and the Biodiversity exploration grant from Christopher
19
Davidson (Flora of the World) is acknowledged. We thank Divya, Siddarthan Surveswaran, and
20
Priyadarsanan for their comments. We thank Abu Hang Samuel for sharing his field photographs
21
with us.
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
2
Abstract
46
47
Resolving the evolutionary history of plant carnivory is of great interest to biologists throughout
48
the world. Among the carnivorous plants, Genus Drosera (Droseraceae) is highly diverse with a
49
wide pantropical distribution. Despite being a group of interest for evolutionary biology studies
50
since the time of Charles Darwin, the historical biogeography of this group remains poorly
51
understood. In this study, with an improved species sampling from Genbank, we present a
52
reanalyzed phylogenetic hypothesis of the genus Drosera. We developed a dated molecular
53
phylogeny of Drosera from DNA sequences of nuclear ITS and chloroplast rbcL genes.
54
Divergence times were estimated on the combined dataset using an uncorrelated lognormal
55
relaxed clock model and a known fossil calibration implemented in BEAST. The maximum
56
clade credibility tree was then used for ancestral range estimations using DEC+J model
57
implemented in BioGeoBEARS. Our analysis suggests that Drosera evolved during the Mid
58
Eocene 36 Ma [95% HPD: 49.5-26] and have diversified and dispersed from the late Miocene
59
onwards. Ancestral areas estimated using the DEC+J models suggest an African origin followed
60
major radiation within Australia. Diversification in Drosera is temporally congruent with the
61
prevailing drier conditions during the Miocene. From Miocene, grasslands and open habitats
62
dominated across continents and might have provided ecological opportunities for their dispersal
63
and diversification. Several long-distance dispersals and range extensions and in situ radiations
64
coinciding with the evolution of drier conditions can explain their extant distribution across
65
continents. Overall our data set provides fresh insights into the biogeographic factors that shaped
66
the origin and evolution of the genus Drosera.
67
68
69
70
71
72
73
74
75
76
77
78
79
Key words: Molecular dating, carnivorous plants, biogeography, aridification, dispersals
80
81
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
3
1. Introduction
82
83
The evolution of carnivory in the plant kingdom poses an exciting biological phenomenon and
84
has fascinated biologists ever since Charles Darwin (Briggs 2009, Thorogood 2017). Carnivory
85
has evolved ten times independently in five different orders among the flowering plants
86
(Fleischmann et al., 2018). Carnivorous plants are distributed in open, moist sites, where they
87
acquire the nutrients from trapping insects (Givnish 2014). The remarkable morphological
88
adaptation to trap and digest animal/insect prey poses a selective advantage over other plant
89
lineages to establish them in infertile soils. Understanding of their ancient dispersal pathways
90
and the rate of interchange between different continents in the past is of utmost importance in
91
giving vital clues to the evolution of the species and shed light on the development of its habitat
92
conditions across time and space.
93
94
Consisting of ~ 250 species, Drosera L. (Droseraceae) is among the speciose carnivorous plant
95
genera with a cosmopolitan distribution and is highly diverse in Australia (Gonella et al., 2016;
96
Rivadavia et al., 2003). Drosera, commonly known as “sundews,” exhibits unmatched
97
adaptations to attract, trap, and digest insects for their nutritional requirements (Bothe and Drake
98
2007). A recent study by Palfalvi et al., (2020) suggests the carnivory within the family
99
Droseraceae evolved as a result of an early genome duplication event. The presence of
100
specialized colorful tentacles in its leaves tipped with nectar glands (flypaper traps), adhesives,
101
and digestive enzymes are used to attract, trap and digest insect prey (Hatcher et al., 2020). The
102
evolution of these flypaper traps is intriguing and might have evolved from a leaf only with
103
adhesive glands after the evolution of nastic and trophic glands (Rivadavia et al., 2012). The
104
sister genera Aldrovanda L. and Dionaea Sol. ex J. Ellis share exclusively another type of
105
trapping mechanism known as snap traps (Givnish 2014). Aldrovanda is a floating aquatic
106
species distributed in the old world and Australia, and Dionaea is an endemic genus and adapted
107
to USA’s marshy lands.
108
109
The systematics of Drosera by Rivadavia et al., (2003) described eleven sections and three
110
subgenera within the genus. Their phylogenetic hypothesis further suggests that genus Drosera
111
originated in Africa or Australia, followed by several long-distance dispersals from Australia to
112
Neotropics and the Nearctic (Rivadavia et al., 2003). Besides, the Australian species might have
113
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
4
expanded their distribution to Neotropics and Africa. The recent phylogenetic hypothesis of
114
Drosera developed by Biswal et al. (2017) discussed various biogeographic scenarios. Their
115
study aimed at providing a better understanding of the existing biogeographic hypotheses for
116
Drosera as well as phylogenetic positions of the three species found in India. However, their
117
divergence date estimation and tree topologies lacked resolution, which potentially makes the
118
biogeographic inferences hard to interpret. Thus, a comprehensive biogeographic analysis based
119
on an improved dated phylogenetic hypothesis can provide vital clues in the origin and evolution
120
of this interesting group across major biomes.
121
122
The fossil records for insectivorous plants are scarce, except for the genus Aldrovanda and
123
family Roridulaceae (Sadowski et al., 2014). For Droseraceae, the prime being an early Eocene
124
microfossil aged between 55-38 Ma and another pollen fossil record of Drosera appeared is
125
sediments of Miocene aged between 22-5 Ma. (Degereef et al., 1989). However, previous dated
126
phylogenetic hypotheses on Drosera using these fossil ages and other calibration schemes
127
produced contrasting age estimates. For example, fossil-based estimates by Magallón et al.,
128
(2015) suggested a late cretaceous origin 76.8 Ma [95% HPD: 93-53] of Drosera. Later, the
129
likelihood-based time tree developed by Biswal et al. (2017) using the two known fossil dates
130
estimated the age of Drosera to be 82 Ma. Similarly, the angiosperm phylogenetic tree
131
developed by Smith and Brown (2018), indicated that the age of Drosera to be 63 Ma. Together,
132
these studies suggest an age distribution of Drosera from Cretaceous to Oligocene. In addition to
133
the variations in divergence dates proposed by previous studies, an influential study on
134
angiosperm origin by Magallón and Sanderson (2001) suggested that several cosmopolitan
135
angiosperm lineages radiated during recent times but not in the Cretaceous.
136
137
In this context, with a lack of consensus among the previous studies, we re-investigated the
138
biogeographic history of genus Drosera with a more comprehensive sampling of the nuclear ITS
139
and chloroplast rbcL sequences available from the Genbank compared to previous studies. The
140
primary goals of the current study are i) to generate an improved dated phylogenetic hypothesis
141
using three previously defined calibrations and compare them after calculating their marginal
142
likelihoods (MLEs) following a path sampling or stepping stone method. And ii) infer the
143
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
5
biogeographical history of Drosera and the mechanisms that shaped the origin and spread of the
144
genus across continents. Owing to its association with habitats with drier and nutrient-poor
145
conditions (Brewer and Schlauer 2018), iii) we specifically aim to test whether global
146
aridification from Miocene and Plio-Pleistocene have played a significant role in its
147
diversification and establishment in different biomes?
148
149
2 Materials and methods
150
151
2.1 Phylogenetic analysis
152
The DNA sequences for nuclear ITS and chloroplast rbcL for 86 species of Drosera and two
153
species of Nepenthes were downloaded from the GenBank and then aligned with MAFFTv 7.213
154
(Katoh and Standley 2013) using L-INS-I (accurate) settings. Sequence gaps in the alignment
155
were treated as missing data. The best fit models and optimum partition schemes were evaluated
156
using the Bayesian information criterion (BIC) and a greedy algorithm employed in
157
PartitionFinder v1.1 (Lanfear et al., 2012). Phylogenetic trees were reconstructed using
158
maximum likelihood (ML) and Bayesian inference (BI). ML tree topologies were developed
159
after 1000 bootstrap replicates in RAXML-HPC V.8 (Stamatakis 2006). A Bayesian inference (BI)
160
was further performed in the aligned data set using the program Mr. Bayesv3.2.6 (Ronquist et al.,
161
2012) implemented on CIPRES Science gateway (Miller et al. 2010). Two independent runs of
162
MCMC generations were run for 50 million generations of Markov Chain Monte Carlo (MCMC)
163
by sampling every 1000 generations until the average standard deviation of split frequencies falls
164
less than 0.01. MCMC chain convergence was further assessed by checking the Estimated
165
sample size (ESS) values. The initial 20% trees from the Mr. Bayes analysis was discarded as
166
burn-in. All the phylogenetic trees were rooted using two outgroups from the genus Nepenthes.
167
2.2 Divergence time estimation
168
For divergence time estimations, we used the program BEAST V 1.8 (Drummond et al., 2012)
169
implemented in CIPRES science gateway V3.3 (Miller et al., 2010). Molecular dates were
170
estimated using the combined dataset of ITS and rbcL. The dataset was partitioned by gene, and
171
an uncorrelated lognormal relaxed clock model was employed after unlinking the substitution
172
and clock models and linking trees in BEAUTI.v.18.3 (part of the BEAST package). We tested
173
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
6
three different calibration schemes to infer the node ages of the Drosera phylogeny. In
174
calibration scheme 1 (C1) the crown age of Droseraceae was constrained between 53-92 Ma with
175
mean age of 76 Ma using a normally distributed age prior since these ranges were estimated
176
previously by Magallón et al., (2015). In calibration scheme 2 (C2), we used the known
177
Dorseraceae pollen records from Eocene to constrain the crown age of Droseraceae (Degreef
178
1997). A uniform prior was set between 34-56 Ma, which is mentioned in Smith et al. (2017).
179
And in calibration scheme 3 (C3), we place a log-normal prior offset of 63 Ma and a standard
180
deviation equal to 1. This age constraint was derived from the larger seed plant phylogeny
181
developed by Smith and Brown (2018). A starting tree satisfying the age priors was set for all the
182
analyses and were generated using ‘chronopl’ function in R using the package ape (Paradis et al.,
183
2004). We removed the subtree slide, Wilson balding, narrow, and wide exchange operators
184
from the XML file to prevent BEAST from exploring the topology space and only to estimate the
185
branch lengths. These three models were then compared using the marginal likelihood estimates
186
following a path sampling/stepping stone method implemented in BEAST. Two independent
187
MCMC runs were performed for 100 million generations by sampling every 1000 steps. Outputs
188
of independent runs for the best model were combined using the software Logcombiner 1.8.2
189
(Rambaut and Drummond 2010a). The software program TRACER (Rambaut and Drummond
190
2010b) was used to inspect the estimated sample size (ESS) values and proper mixing. The
191
maximum clade credibility (MCC) tree was generated using tree annotator after discarding 20%
192
of the initial runs are burn in.
193
2.3 Biogeographical Inference
194
195
The biogeographic analysis was performed in the R package BioGeoBEARS (Matzke, 2013) in
196
the BEAST MCC tree. To estimate the ancestral ranges at each node, we used the dispersal
197
extinction and cladogenesis (DEC) model and its variant DEC+ J model (Matzke, 2014), which
198
accounts for founder event speciation. Furthermore, the other models implemented in
199
BioGeoBEARS viz. DIVALIKE, BAYAREA, and its J variants were also used to estimate the
200
ancestral ranges. However, this approach received several conceptual criticisms by Ree and
201
Sanmartin (2018), especially on the direct model comparisons between these models based on
202
their likelihood scores. For the biogeographic analysis, we designated seven areas for the
203
ancestral range estimations based on the distribution of Drosera i.e Afro-Madagascar (A),
204
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
7
Neotropics (B), Nearctic (C), Western Palearctic (D), Eastern Palearctic (E), Southeast Asia (F)
205
and Australasia (G) following the extant distribution of Drosera taken from published records
206
and other online resources (Table S1).
207
208
3. Results
209
210
3.1 Phylogenetic analysis
211
Our improved phylogenetic hypothesis based on the combined analysis of chloroplast rbcL and
212
nuclear ITS represents the genetic relationships of 85 species of Drosera. The combined data
213
sets consist of 2004 bp, of which 1227bp was chloroplast rbcL, and 777 bp belong to the nuclear
214
ITS. The models of sequence evolution identified by PartitionFinder varied between all three
215
analytical schemes (Table 1). Genus Drosera forms a monophyletic clade sister to Aldrovanda
216
and Dionaea in BI and ML analysis, as in previous studies (Rivadavia et al. 2003, Magallón et
217
al. 2015). The analysis also recovered two major clades within Drosera. Clade 1 consisted of
218
species from Africa, Europe, and Asia (ML bootstrap = 97, posterior probability support=1,
219
(supplementary file Figure S1). Clade 2 consists majorly of members from South Pacific
220
(Australia and New Zealand) and the ancestral species (D. regia Stephans) from Africa. We
221
found uncertainties in the placement of the previously defined taxonomic sections, and some
222
taxonomic relationships contradict the previous relationships. This difference can be attributed to
223
larger sequence sampling used in this study compared to the previous works by Rivadavia et al.
224
(2003; 2012). Hence, we refrain from making inferences of the taxonomic sections in this study
225
due to our lack of access to several voucher specimens and leave this section open for
226
interpretation.
227
3.2 Divergence time estimation and biogeographic analysis
228
The highest likelihood corresponded to the C2 (-19578.03), but this was not significantly
229
different than the marginal likelihood scored for C1 (Table 2). The ESS values ad convergence
230
was better in C2; hence we used these settings for a two replicate MCMC search (see
231
supplementary file Figure S2 for alternative tree topologies). Molecular dating analysis inferred
232
that Droseraceae diverged from its sister group during Eocene between 42 Ma [95% HPD: 34-54
233
Ma] (Figure 1). The crown age estimate of genus Drosera was 36Ma [95% HPD: 26-49 Ma].
234
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
8
Split between the two main clades within Drosera occurred during the Oligocene 30Ma [95%
235
HPD: 21-42 Ma]. DEC+J model was favored amongst all other models (-lnl=130.069, AICc=
236
266.4; Table 3) and the null model rejected at p< 0.0001. All the ancestral range estimations
237
identified an African origin of Drosera and a further dispersal to Australia during the Eocene
238
(Node 1; Figures 1, 2; Table 4). Further, the genus Drosera had undergone massive radiation
239
within Australia, starting from Oligocene. These radiations were followed by several dispersals
240
to Neotropics and Africa starting from the Mid Miocene (Node numbers 2, 3, 4, 5, 6; Figures 1,
241
2). A single back dispersal to Africa and dispersal to Southeast Asia from Australia was observed
242
during this period (Node numbers 7, 8, 9; Figures 1, 2). Furthermore, we observed several in situ
243
radiations within the Neotropics, and the Afro-Madagascar in the late Miocene, followed by
244
dispersal from Australia (Node numbers 10,11,12; Figures 1, 2).
245
246
4. Discussion
247
248
Our results support a complex biogeographic history of genus Drosera and explain the historical
249
factors that influenced its dispersal and diversification patterns. It consists of an African origin
250
followed by Australian radiations, long-distance dispersals, range extensions, and several in situ
251
radiations in other continents. In situ radiations in multiple lineages were followed by the
252
intensification of grassland spread and reductions in global temperatures. The implications of our
253
new, improved analysis and its potential shortcomings are discussed herein.
254
255
4.1. Origin and historical spread of Drosera
256
257
The results from our time-calibrated phylogeny suggest that the ancestors of Drosera have
258
evolved in Africa during the Eocene and dispersed to Australia as hypothesized by Rivadavia et
259
al., (2003). The molecular clock dated this event to Eocene, and the time corresponds well to the
260
Eocene-Oligocene glacial maximum. This period exhibited arid conditions in the southern
261
regions of Australia and Africa (Buerki et al., 2013), which might have provided a suitable
262
environment for Drosera to disperse from Africa to Australia via Antarctica which was once a
263
common route for several lineages to disperse into Australia (Benda et al., 2019). For example,
264
Sapindaceous lineages were shown to have using this dispersal route from Australia to Africa
265
during the Middle Paleocene-Late Eocene (Buerki et al., 2013). These age estimates are too
266
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
9
young to invoke a Gondwanan vicariance hypothesis. And another possibility is dispersal
267
through the now-submerged Crozet plateaus and Kerguelen islands to Australia (see Kayaalp et
268
al., 2017, Renner et al., 2010).
269
270
The dispersals from Africa were followed by tremendous in situ radiations in Drosera within the
271
Australian continent. Starting from the Oligocene, a monsoonal drier condition has evolved in
272
northern Australia. Due to the drop in global temperature, there was a reduction in the extent of
273
broad-leaved forests. Previous studies have shown that these changes are known to affect the
274
native vegetation of Australia. For example, in situ radiations in Eucalyptus subgenus
275
Symphyomyrtus (Ladiges et al., 2003) coincides well with this time frame. The early to mid-
276
Miocene weather was mostly warm, followed by cool and dry conditions in the late Miocene.
277
Especially the fossil pollen from freshwater swamps suggests the existence of a relatively open
278
heath scrubland, which has highly acidic infertile soils with the presence of repeated wildfire.
279
The late Miocene witnessed a reduction in rainforests and an increase in Eucalyptus forests,
280
indicating a colder and drier condition. And the existing rainforests continued to reduce with an
281
increase in grassland and arid climate in Pliocene (Martin 2006). Our phylogenetic hypotheses
282
suggesting several species radiations in Australia is highly congruent with this period. As shown
283
elsewhere (Givnish 2014), these unfavorable conditions are advantageous for a carnivorous plant
284
genus like Drosera. Moreover, during this time frame, a change in climate to drier state and the
285
incidence of fire and the spread of open grasslands might have provided several new niches for
286
Drosera to diversify within the Australian continent (see Lamont and Downes 2011).
287
288
The diversification in Drosera also corresponds well to the variation in global climatic trends
289
derived from ∆18O data to the Cenozoic to present (Zachos et al., 2008), which was evident from
290
Oligocene -Miocene radiations in Australia (Figure 1). These radiations in Australia were
291
followed by several long-distance dispersals from Australia to South America and range
292
extensions to Southeast Asia (Node numbers 2, 3, 4, and 5). For example, D. meristocaulis
293
Maguire & Wurdack, is shown to have undergone dispersal from Australia to the Neotropics in
294
the Mid-Miocene. Our biogeographic analysis and molecular dates are also in agreement with a
295
possible long-distance dispersal from Australia to Neotropics (see Thornhill et al., 2015). These
296
dispersals are intriguing, but, can take place via Antarctica (35-10Ma) and possibly mediated
297
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
10
through an establishment of the West Wind Drift (WWD) and equatorial currents during this
298
period (see Buerki et al., 2013). And this long-distance dispersal can also be through the islands
299
that have existed in the Pacific basin since the Oligocene, such as Hawaiian archipelago and Fiji
300
Islands (Chen et al., 2014). However, such long-distance dispersal from Australia to Neotropics
301
via Antarctica is not entirely rare. Similar patterns are found in the Ampelopsis clade of Vitaceae
302
(Nie et al. 2012) and the cosmopolitan genus Ranunculus (Emadzade et al., 2011). To better
303
understand the major dispersal events, we marked the results in Figure 2 (insets A and B).
304
305
A lineage dispersed to Neotropics from Australia (Node 6), has undergone in situ radiation and
306
then dispersed to Nearctic and Western Palearctic (Node 7, 9) during the Miocene. A back
307
dispersal then followed this dispersal to Neotropics (Node 8). Interestingly, our age estimates of
308
these nodes align well will the expansion of grassland biomes within the Neotropics and the
309
Nearctic (see Fig 6c in Meseguer et al., 2015). Further, there were long-distance dispersals from
310
Australia to Africa during the Miocene (node 10) and might be a result of via wind or ocean
311
currents. (Dick et al., 2007). During the late Miocene, Africa also several ecological transitions,
312
mainly the spread of C4 type grasslands; overall, these transitions might have been favorable for
313
Drosera to diversify (Meseguer et al., 2015).
314
315
We observe a single dispersal from Africa to South America during the Miocene, followed by
316
species radiation in that lineage (Node 12). This pattern of in situ radiation again corresponds
317
well with the timing of grassland expansion in South America, which could be a favorable time
318
for Drosera to diversify within. Drosera paleacea (Node 12) originate via dispersal from Africa
319
to Australia during the late Miocene. A long-distance dispersal from Africa to Australia remains
320
the most likely explanation supported by well-studied taxa such as Simaroubaceae (Clayton et
321
al., 2009). Overall findings from this study suggest the role of then prevailing drier conditions
322
from Miocene and long-distance dispersals (possibly by birds) explains the inter-continental
323
dispersals and radiations in the genus Drosera.
324
325
4.2 Molecular dating and previous hypotheses
326
Maximum likelihood RelTime time estimates of Drosera by Biswal et al., (2017) computed
327
based on branch lengths optimized by ML (Tamura et al., 2012) lack resolution and was also
328
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
11
characterized by several polytomies. Usage of the second fossil calibration (25-5Ma) was also
329
not well justified. and we refrained from adding it in our divergence dates due to the confusion in
330
its placement. The RelTime method is computationally efficient for estimating divergence times
331
(Mellow et al., 2019). Their dated tree topology resulted in several polytomies and raised
332
practical difficulties in interpreting the results (See Fig. 7 Biswal et al., 2017). An incomplete
333
taxon sampling from the GenBank might be another reason that prevented the authors from
334
developing a resolved tree topology.
335
Further, inadequate taxon sampling is also known to underestimate the divergence times (Linder
336
et al., 2005). A densely sampled phylogeny and genome sequences can improve the taxonomic
337
relationships and biogeographic inferences within the genus Drosera, which is a way forward.
338
Interestingly, our molecular dating approaches suggest an Eocene origin of the family Drosera,
339
opposing the previous date estimates. Besides, our newly inferred dates are more congruent with
340
the global decrease in temperature, which might have set the stage for diversification of Drosera
341
in different biomes. It is also noteworthy that the recent large scale dated phylogeny developed
342
by Ramrez-Barahona et al., (2020) also identified a Cretaceous origin of Droseraceae 65.8 M
343
instead of an Eocene origin as per our dating analysis, but with a larger credible interval
344
spanning from 30-94 Ma. We did not include these in our calibration schemes, as these ages
345
ranges were covered in C1 and C3.
346
5. Conclusions
347
348
Aridifications can exert intense selective pressure, influencing diversification in plant lineages
349
(Gutiérrez-Ortega et al., 2017). Our analysis confirms an Eocene origin and Miocene
350
diversification of Drosera followed by a complex dispersal pathway, including long-distance
351
dispersals, to establish its presence across continents. Drosera's diversification is temporally
352
congruent with the aridification and presence of open grassland and fire in Australia, Africa, and
353
Neotropics. Our study also confirms that Drosera exhibits an early history of range expansion
354
between major continental areas. Overall, our finding agrees with previously discussed dispersal
355
pathways for major angiosperm lineages (Buerki et al., 2013; Chen et al., 2014). Our results will
356
have implications for understanding the biogeographic history of Drosera and also expected to
357
shed light on the evolution of modern-day biomes. Further studies with a more extensive taxon
358
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
12
sampling and large sequence datasets are required to deepen the knowledge in phylogeny,
359
biogeographic history and diversification dynamics of Drosera.
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
13
References
403
404
Benda, D., Nakase, Y., & Straka, J. (2019). Frozen Antarctic path for dispersal initiated parallel
405
host-parasite evolution on different continents. Molecular Phylogenetics and Evolution, 135,
406
67–77. https://doi.org/10.1016/j.ympev.2019.02.023
407
Biswal, D. K., Yanthan, S., Konhar, R., Debnath, M., Kumaria, S., & Tandon, P. (2017).
408
Phylogeny and biogeography of the carnivorous plant family Droseraceae with
409
representative Drosera species from Northeast India. F1000 Research, 6, 1454.
410
https://doi.org/10.12688/f1000research.12049.1
411
Bothe, H., & Drake, H. (2007). Interactions among organisms that result in enhanced activities of
412
N-cycle reactions. In Biology of the Nitrogen Cycle (pp. 397-405). Elsevier.
413
Brewer, J. S., & Schlauer, J. (2018). Biogeography and habitats of carnivorous plants. In Oxford
414
Scholarship Online. Oxford University Press.
415
https://doi.org/10.1093/oso/9780198779841.003.0002
416
Briggs, J. C. (2009). Darwin’s biogeography. Journal of Biogeography, 36(6), 1011–1017.
417
https://doi.org/10.1111/j.1365-2699.2008.02076.x
418
Buerki, S., Forest, F., Stadler, T., & Alvarez, N. (2013). The abrupt climate change at the
419
Eocene–Oligocene boundary and the emergence of South-East Asia triggered the spread of
420
sapindaceous lineages. Annals of Botany, 112(1), 151–160.
421
https://doi.org/10.1093/aob/mct106
422
Chen, L.-Y., Zhao, S.-Y., Mao, K.-S., Les, D. H., Wang, Q.-F., & Moody, M. L. (2014).
423
Historical biogeography of Haloragaceae: An out-of-Australia hypothesis with multiple
424
intercontinental dispersals. Molecular Phylogenetics and Evolution, 78, 87–95.
425
https://doi.org/10.1016/j.ympev.2014.04.030
426
Clayton, J.W, Soltis, P.S, Soltis, D.E.(2009). Recent long-distance dispersal overshadows ancient
427
biogeographical patterns in a pantropical angiosperm family (Simaroubaceae,
428
Sapindales) Systematic Biology, 58(4), 395–410. doi: 10.1093/sysbio/syp041.
429
Degreef, J.D. 1997. Fossil Aldrovanda. Carnivorous Plant Newsletter 26: 93–97.
430
Dick, C. W., Bermingham, E., Lemes, M. R., & Gribel, L. (2007). Extreme long-distance
431
dispersal of the lowland tropical rainforest tree Ceiba pentandra L. (Malvaceae) in Africa
432
and the Neotropics. Molecular Ecology, 16(14), 3039–3049. https://doi.org/10.1111/j.1365-
433
294x.2007.03341.x
434
Drummond, A. J., Suchard, M. A., Xie, D., & Rambaut, A. (2012). Bayesian Phylogenetics with
435
BEAUti and the BEAST 1.7. Molecular Biology and Evolution, 29(8), 1969–1973.
436
https://doi.org/10.1093/molbev/mss075
437
Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high
438
throughput. Nucleic Acids Research, 32(5), 1792–1797. https://doi.org/10.1093/nar/gkh340
439
Emadzade, K., Gehrke, B., Peter Linder, H., Hörandl. E. (2011). The biogeographical history of
440
the cosmopolitan genus Ranunculus L. (Ranunculaceae) in the temperate to meridional
441
zones. Molecular Phylogenetics and Evolution, 58(1), 4–21. doi:
442
10.1016/j.ympev.2010.11.002.
443
Fleischmann, A., Schlauer, J., Smith, S. A., & Givnish, T. J. (2018). Evolution of carnivory in
444
angiosperms. In Oxford Scholarship Online.
445
https://doi.org/10.1093/oso/9780198779841.003.0003
446
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
14
Givnish, T. J. (2014). New evidence on the origin of carnivorous plants. Proceedings of the
447
National Academy of Sciences, 112(1), 10–11. https://doi.org/10.1073/pnas.1422278112
448
Gonella, P. M., Fleischmann, A., Rivadavia, F., Neill, D. A., & Sano, P. T. (2016). A revision of
449
Drosera (Droseraceae) from the central and northern Andes, including a new species from
450
the Cordillera del Cóndor (Peru and Ecuador). Plant Systematics and Evolution, 302(10),
451
1419–1432. https://doi.org/10.1007/s00606-016-1341-3
452
Gutiérrez-Ortega, J.S., Yamamoto, T., Vovides, A. P., Angel Pérez-Farrera, M., Martínez, J. F.,
453
Molina-Freaner, F., Watano, Y., & Kajita, T. (2017). Aridification as a driver of
454
biodiversity: a case study for the cycad genus Dioon (Zamiaceae). Annals of Botany,
455
121(1), 47–60. https://doi.org/10.1093/aob/mcx123
456
Hatcher, C. R., Ryves, D. B., & Millett, J. (2019). The function of secondary metabolites in plant
457
carnivory. Annals of Botany, 125(3), 399–411. https://doi.org/10.1093/aob/mcz191
458
Katoh, K., & Standley, D. M. (2013). MAFFT Multiple Sequence Alignment Software Version
459
7: Improvements in Performance and Usability. Molecular Biology and Evolution, 30(4),
460
772–780. https://doi.org/10.1093/molbev/mst010
461
Kayaalp, P., Stevens, M. I., & Schwarz, M. P. (2017). “Back to Africa”: increased taxon
462
sampling confirms a problematic Australia-to-Africa bee dispersal event in the Eocene.
463
Systematic Entomology, 42(4), https://doi.org/10.1111/syen.12241
464
Ladiges, P. Y., Udovicic, F., & Nelson, G. (2003). Australian biogeographical connections and
465
the phylogeny of large genera in the plant family Myrtaceae. Journal of Biogeography,
466
30(7), 989–998. https://doi.org/10.1046/j.1365-2699.2003.00881.x
467
Lamont, B. B., & Downes, K. S. (2011). Fire-stimulated flowering among resprouters and
468
geophytes in Australia and South Africa. Plant Ecology, 212(12), 2111–2125.
469
https://doi.org/10.1007/s11258-011-9987-y
470
Lanfear, R., Calcott, B., Ho, S. Y. W., & Guindon, S. (2012). PartitionFinder: Combined
471
Selection of Partitioning Schemes and Substitution Models for Phylogenetic Analyses.
472
Molecular Biology and Evolution, 29(6), 1695–1701.
473
https://doi.org/10.1093/molbev/mss020
474
Linder, H. P., Hardy, C. R., & Rutschmann, F. (2005). Taxon sampling effects in molecular
475
clock dating: An example from the African Restionaceae. Molecular Phylogenetics and
476
Evolution, 35(3), 569–582. https://doi.org/10.1016/j.ympev.2004.12.006
477
Magallón, S., & Sanderson, M. J. (2001). Absolute diversification rates in angiosperm clades.
478
Evolution, 55, 1762–1780. https://doi.org/10.1111/j.0014-3820.2001.tb00826.x
479
Magallón, S., Gómez‐Acevedo, S., Sánchez‐Reyes, L. L., & Hernández‐Hernández, T. (2015). A
480
metacalibrated time‐tree documents the early rise of flowering plant phylogenetic diversity.
481
New Phytologist, 207(2), 437–453. https://doi.org/10.1111/nph.13264
482
Martin, H. A. (2006). Cenozoic climatic change and the development of the arid vegetation in
483
Australia. Journal of Arid Environments, 66(3), 533–563.
484
https://doi.org/10.1016/j.jaridenv.2006.01.009
485
Matzke, N. J. (2013). Probabilistic historical biogeography: new models for founder-event
486
speciation, imperfect detection, and fossils allow improved accuracy and model-testing.
487
Frontiers of Biogeography, 5(4), 242-248. http://escholarship.org/uc/item/44j7n141
488
Matzke, N. J. (2014). Model Selection in Historical Biogeography Reveals that Founder-Event
489
Speciation Is a Crucial Process in Island Clades. Systematic Biology, 63(6), 951–970.
490
https://doi.org/10.1093/sysbio/syu056
491
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
15
Meseguer, A. S., Lobo, J. M., Ree, R., Beerling, D. J., & Sanmartín, I. (2015). Integrating
492
Fossils, Phylogenies, and Niche Models into Biogeography to Reveal Ancient Evolutionary
493
History: The Case of Hypericum (Hypericaceae). Systematic Biology, 64(2), 215–232.
494
https://doi.org/10.1093/sysbio/syu088
495
Miller, M.A., Pfeiffer, W., and Schwartz, T. (2010) "Creating the CIPRES Science Gateway for
496
inference of large phylogenetic trees" in Proceedings of the Gateway Computing
497
Environments Workshop (GCE), 14 Nov. 2010, New Orleans, LA pp 1 - 8.
498
Nie, Z.-L., Sun, H., Manchester, S. R., Meng, Y., Luke, Q., & Wen, J. (2012). Evolution of the
499
intercontinental disjunctions in six continents in the Ampelopsis clade of the grape family
500
(Vitaceae). BMC Evolutionary Biology, 12(1), 17. https://doi.org/10.1186/1471-2148-12-17
501
Palfalvi, G., Hackl, T., Terhoeven, N., Shibata, T. F., Nishiyama, T., Ankenbrand, M., …
502
Hedrich, R. (2020). Genomes of the Venus Flytrap and Close Relatives Unveil the Roots of
503
Plant Carnivory. Current Biology. https://doi.org/10.1016/j.cub.2020.04.051
504
Paradis, E., Claude, J., & Strimmer, K. (2004). APE: Analyses of Phylogenetics and Evolution in
505
R language. Bioinformatics, 20(2), 289–290. https://doi.org/10.1093/bioinformatics/btg412
506
Rambaut A, Drummond AJ. (2010a). LogCombiner version 1.6.1 [computer
507
program] http://beast.bio.ed.ac.uk .
508
Rambaut A, Drummond AJ. (2010b). TreeAnnotator version 1.6.1 [computer
509
program] http://beast.bio.ed.ac.uk .
510
Rambaut, A., Drummond, A. J., Xie, D., Baele, G., & Suchard, M. A. (2018). Posterior
511
Summarization in Bayesian Phylogenetics Using Tracer 1.7. Systematic Biology, 67(5),
512
901–904. https://doi.org/10.1093/sysbio/syy03
513
Ramírez-Barahona, S., Sauquet, H., & Magallón, S. (2020). The delayed and geographically
514
heterogeneous diversification of flowering plant families. Nature Ecology & Evolution.
515
https://doi.org/10.1038/s41559-020-1241-3
516
Ree, R. H., & Sanmartín, I. (2018). Conceptual and statistical problems with the DEC+J model
517
of founder-event speciation and its comparison with DEC via model selection. Journal of
518
Biogeography, 45(4), 741–749. https://doi.org/10.1111/jbi.13173
519
Renner, S. S. (2010). Biogeographic insights from a short-lived Palaeocene island in the
520
Ninetyeast Ridge. Journal of Biogeography, 37(7), 1177–1178.
521
https://doi.org/10.1111/j.1365-2699.2010.02355.x
522
Rivadavia, F., Kondo, K., Kato, M. & Hasebe, M. (2003). Phylogeny of the
523
sundews, Drosera (Droseraceae), based on chloroplast rbcL and nuclear 18S ribosomal
524
DNA sequences. American Journal of Botany 90: 123–130.
525
http://dx.doi.org/10.3732/ajb.90.1.123
526
Rivadavia, F., Miranda, V.F.O., Hoogenstrijd, G., Pinheiro, F., Heubl, G. & Fleischmann, A.
527
(2012). Is Drosera meristocaulis a pygmy sundew? Evidence of a long-distance dispersal
528
between Western Australia and northern South America. Annals of Botany 110: 11–21.
529
http://dx.doi.org/10.1093/aob/mcs096
530
Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D. L., Darling, A., Höhna, S., Larget, B.,
531
Liu, L., Suchard, M. A., & Huelsenbeck, J. P. (2012). MrBayes 3.2: Efficient Bayesian
532
Phylogenetic Inference and Model Choice Across a Large Model Space. Systematic
533
Biology, 61(3), 539–542. https://doi.org/10.1093/sysbio/sys029
534
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
16
Sadowski, E.-M., Seyfullah, L. J., Sadowski, F., Fleischmann, A., Behling, H., & Schmidt, A. R.
535
(2014). Carnivorous leaves from Baltic amber. Proceedings of the National Academy of
536
Sciences, 112(1), 190–195. https://doi.org/10.1073/pnas.1414777111
537
Sauquet, H. (2013). A practical guide to molecular dating. Comptes Rendus Palevol, 12(6), 355–
538
367. https://doi.org/10.1016/j.crpv.2013.07.003
539
Smith, S. A., & Brown, J. W. (2018). Constructing a broadly inclusive seed plant phylogeny.
540
American Journal of Botany, 105(3), 302–314. https://doi.org/10.1002/ajb2.1019
541
Smith, S. A., Brown, J. W., Yang, Y., Bruenn, R., Drummond, C. P., Brockington, S. F., Walker,
542
J. F., Last, N., Douglas, N. A., & Moore, M. J. (2017). Disparity, diversity, and duplications
543
in the Caryophyllales. New Phytologist, 217(2), 836–854. https://doi.org/10.1111/nph.14772
544
Stamatakis, A. (2006). RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with
545
thousands of taxa and mixed models. Bioinformatics, 22(21), 2688–2690.
546
https://doi.org/10.1093/bioinformatics/btl446
547
Tamura, K., Battistuzzi, F. U., Billing-Ross, P., Murillo, O., Filipski, A., & Kumar, S. (2012).
548
Estimating divergence times in large molecular phylogenies. Proceedings of the National
549
Academy of Sciences, 109(47), 19333–19338. https://doi.org/10.1073/pnas.1213199109
550
Thornhill, A. H., Ho, S. Y. W., Külheim, C., & Crisp, M. D. (2015). Interpreting the modern
551
distribution of Myrtaceae using a dated molecular phylogeny. Molecular Phylogenetics and
552
Evolution, 93, 29–43. https://doi.org/10.1016/j.ympev.2015.07.007
553
Thorogood, C. J., Bauer, U., & Hiscock, S. J. (2017). Convergent and divergent evolution in
554
carnivorous pitcher plant traps. New Phytologist, 217(3), 1035–1041.
555
https://doi.org/10.1111/nph.14879
556
Zachos, J. C., Dickens, G. R., & Zeebe, R. E. (2008). An early Cenozoic perspective on
557
greenhouse warming and carbon-cycle dynamics. Nature, 451(7176), 279–283.
558
https://doi.org/10.1038/nature06588
559
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
17
560
Figure 1: Maximum clade credibility (MCC) chronogram from molecular dating analysis of
561
Drosera. The temperature curve represents the major trends in climate from Cenozoic (-65Ma) to
562
present (0Ma) which is derived from oxygen isotope (∆18O) data from Zachos et al., (2008) were
563
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
18
converted to absolute temperatures. The divergence time and HPD intervals of the numbered
564
nodes are represented in Table 4. Inset: D. indica.
565
566
567
568
569
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
19
570
Figure 2: Historical biogeography of Drosera. Colored boxes to the left of species names shows
571
current geographical distributions indicated in the distribution map. Pie charts on the nodes
572
represents probabilities of the most likely ancestral ranges estimated from the DEC+J model. The
573
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
20
relative continental positions are represented in the globes and the arrows and the numbers
574
represents important dispersal events.
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
21
Table 1: Partition schemes and substitution models used for the Maximum likelihood (ML),
619
Bayesian Inference (BI) and Molecular dating (BEAST)
620
621
Sl.no
Gene
RAXML
Mr. Bayes
BEAST
1
rbcl codon position1
GTR+ G
JC+I+G
TRNef+I+G
2
rbcl codon position2
GTR+ G
JC+I+G
TRNef+I+G
3
rbcl codon position3
GTR+ G
GTR+G+I
GTR+I+G
4
Nuclear ITS
GTR+ G
GTR+G
GTR+G
622
623
624
625
626
627
628
Table 2: BEAST model comparisons based on three different calibration schemes used in this
629
study
630
631
C1
C2
C3
Path sampling
-19676.385
-19578.03
-19594.457
Stepping stone
-19671.036
-19486.71
-19542.725
632
633
634
635
636
637
638
Table 3: Model comparisons in BioGeoBEARS.
639
640
Model
-LNL
No. of
parameters
d
e
j
AICc
p
DEC
-141.26
2
0.016
0.002
0
286.7
DEC+j
-130.069
3
0.010
1X10-12
0.032
266.4
>0.0001
DIVALIKE
-138.348
2
0.018
1X10-12
0
280.8
DIVALIKE+j
-132.404
3
0.011
1X10-12
0.029
271.1
0.0006
BAYAREA
-182.473
2
0.016
0.043
0
369.1
BAYAREA+j
-136.858
3
0.008
0.001
0.046
280
>0.0001
d=rate of dispersal; e= rate of extinction; j= relative probability of founder event speciation
641
642
643
644
645
646
647
648
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
22
Table 4: Divergence time and the distribution of its highest posterior density (HPD) intervals of
649
biogeographically important nodes. Node numbers correspond to the ones in Figures 1, 2, 2A,
650
and 2B.
651
652
Node
Description
Divergence time (Ma)
Mean
95%HPD
1
Drosera crown
36.4
49.5 - 26
2
Split: D. peltata and D. auriculata
6.5
11.5 – 2.8
3
Stem of D. meristocaulis
11.8
18.4 – 6.7
4
Stem of D. allantostigma
2.3
4.3 - 0.9
5
Split: D. stenopetala and
D. uniflora
5.3
10.6 – 1.4
6
Stem D. brevifolia
6.6
10.4 -3.6
7
Crown of the Nearctic clade
4.1
6.8 – 2.7
8
Stem of D. biflora
0.5
1.6 - 0.2
9
Split: D. tokaiensis – D. rotundifolia
1.1
2.6 -0.2
10
Crown of the clade dispersed to Africa
from Australia
8.7
13.2 – 5.1
11
Stem of the clade dispersed to
Neotropics from Africa
5.2
8.4 – 2.7
12
Stem of D. paleacea
3.5
6.6 – 1.3
653
.CC-BY-NC-ND 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 6, 2020. . https://doi.org/10.1101/2020.08.06.240234doi: bioRxiv preprint
