INTEGRATING PHYLOGENETICS, ECOLOGY AND EVO-DEVO TO

UNDERSTAND A KEY INNOVATION: FLORAL NECTAR SPURS IN LINARIA

AND RELATED GENERA
Mario FERNÁNDEZ-MAZUECOS*, Beverley J. GLOVER**
Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EA, UK
*mf546@cam.ac.uk **bjg26@cam.ac.uk

BACKGROUND
● Understanding the origin of species is one of the major challenges of modern biology, and requires the
integration of ecological, evolutionary and developmental approaches. However, the number of studies
using this integrative approach is still small.
A. MACROEVOLUTIONARY PATTERNS
● Nectar spurs are considered a key innovation that promotes speciation in angiosperms (Kay et al., 2006). 180
They may provide a mechanism of prezygotic reproductive isolation through differential pollinator visitation 160
(pollinator shift hypothesis). Under uniform diversification,
140 Linaria species diversity is expected to

Diversity (No. species)

● Toadflaxes (genus Linaria) belong to the snapdragon lineage (tribe Antirrhineae), a model group for plant be an exponential function of
120
clade age. The dating analysis of
development and evolution.
100 Antirrhineae revealed that
Linaria Linaria is one of the oldest
● They provide an ideal system to investigate pollinator-driven speciation because of their specialized flower 80
genera of the tribe. However, its
traits, including a nectar spur of variable length. The spur collects nectar, which is offered to pollinators. 4 No spur species diversity is much higher
60
Spur 40
Antirrhinum than that expected based on its

MAIN OBJECTIVE 20
age alone.

To understand the role of a potential key innovation (nectar spurs) in speciation processes in 1 0
plants, by using toadflaxes and relatives as a model system, and applying a multidisciplinary 0 5 10 15 20 25 30 35

approach integrating phylogenetics, ecology and developmental biology. Age (Ma)

2
A preliminary Binary State Speciation
λ lineages without nectar spur and Extinction analysis (BiSSE;
3 4
FitzJohn et al., 2009) supported higher
Anthers
speciation rates for those lineages of

Probability density
Anthers 3
λ lineages with nectar spur Antirrhineae displaying nectar spurs
than for spurless lineages. This
supports the role of nectar spurs as a
2 key innovation promoting
diversification.

We conducted a phylogenetic analysis and divergence time estimation 1 A more complete phylogenetic analysis
(BEAST) of 210 species (64% of c. 330 extant species) of Antirrhineae including c.90% of extant species and
using nuclear (ITS) and plastid (ndhF) DNA sequences. Bayesian additional analyses are in preparation to
0
ancestral state reconstructions supported multiple (3-4) acquisitions of confirm these results.
Nectar nectar spurs in the course of Antirrhineae diversification. One of these
spur 0.2 0.4 0.6 0.8 1.0 1.2
acquisitions resulted in the most diverse genus within the Antirrhineae
Pollination of Antirrhinum Pollination of Linaria (Linaria, >160 spp.). Speciation rate (λ)
(non spurred) by a bumblebee (spurred) by a honeybee

B. MODEL SYSTEM: THE IBERIAN CLADE OF LINARIA SECT. VERSICOLORES

● As a model system to study the role of nectar spurs in Spur lengths of the eight species of the Iberian clade of Linaria sect. Versicolores. Pollination of Linaria clementei
speciation processes, we have selected the Iberian clade of by Andrena flavipes
Linaria sect. Versicolores, which includes eight species 16

distributed in the Iberian Peninsula (Fernández-Mazuecos et al.,

Spur length (mm)

14
2013; Fernández-Mazuecos & Vargas, 2015). This group has
12
been chosen because:
(i) It is monophyletic. 10
(ii) It displays a wide range of spur lengths (see right).
8
(iii) It is a small group amenable to detailed analyses.
6
● Previous phylogenetic analyses based on nuclear and plastid
4 Mean
DNA sequences revealed the recent (Quaternary) radiation of
the clade (Fernández-Mazuecos & Vargas, 2015). 2 Mean±SD

0
● Preliminary pollinator observations revealed that all eight
L. clementei

L. salzmannii
L. spartea

L. incarnata
L. spicata

L. viscosa
L. onubensis

L. algarviana

species are mainly pollinated by bees (Fernández-Mazuecos et

al., 2013).

● Flower and spur development have been previously studied in L. clementei L. salzmannii
the model species of the same genus L. vulgaris. A potential The two species with the shortest (L. clementei) and
involvement of KNOTTED1-like homeobox (KNOX) genes in longest (L. salzmannii) nectar spurs are the main focus
Species
spur development has been suggested (Box et al., 2011). of our research.

(1) Phylogeny (2) Ecology (3) Development

Objectives: Objectives: Objectives:
● To reliably infer phylogenetic relationships and ● To determine the relative roles of pollinator specificity (potentially related to ● To understand the genetic and developmental basis of
divergence times within the study clade using a spur length variation) and other factors (habitat isolation, post-pollination spur length variation.
phylogenomic approach. breeding barriers) as components of reproductive isolation among species. ● To test whether variation in expression patterns of
● To reconstruct evolutionary changes in spur length and ● To investigate the selective pressures involved in spur length evolution, regulatory genes determines developmental changes that
their potential association to speciation events. including the interaction between floral rewards and pollinator behaviour. underlie spur length evolution in the course of speciation.

L. spicata ● We are focused on the two species (L. clementei and L. salzmannii) with extreme spur lengths, both of ● We are comparing nectar spur
● Previous analyses based on conventional 0.95 which are endemic to the same small region in Southern Spain. ontogeny in L. clementei and L.
phylogenetic markers (Fernández-
L. salzmannii salzmannii. A comparison of
Mazuecos et al., 2013; Fernández- 0.47 ● Pollinator censuses have been conducted in four wild populations of L. clementei and two of L. salzmannii
spur growth timing revealed that
Mazuecos & Vargas, 2015) were unable to for >500 min of observations per population. Both species are mainly pollinated by bees (>95% of flower 1 cm
L. clementei length variation is the result of
resolve fine-scale relationships between 1 visits). Different main pollinators are found in L. clementei and L. salzmannii (below), but there does not
differences in growth rate rather
species (right). seem to be a relationship between spur length and proboscis length of pollinators.
0.44
L. algarviana than growth duration (left).
14
Mean
L. clementei L. salzmannii Therefore, spur length evolution
● A sister relationship between the two 12 Mean±SD
L. incarnata in this case involves
species (L. clementei and L. salzmannii) 0.77
Spur length (mm)

10 L. salzmannii
heterochronic changes such as
with extreme spur lengths was suggested
neoteny and acceleration. This
8
L. viscosa
but not statistically supported. This 0.58 6
result contrasts with that found
relationship would be congruent with a role
L. spartea 4
in a different study system
0.56
of spur length evolution in recent L. clementei
2
(Aquilegia; Puzey et al., 2011).
speciation events. L. onubensis 0
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3
Amegilla Xylocopa Andrena Rhodanthidium Anthophora Time (days to flower opening)
● In the model species L. vulgaris, the
● In this scenario of recent speciation, a phylogenomic approach using
majority of spur growth is attributed to
next-generation sequencing data is required to resolve phylogenetic relationships.
2.0 cell expansion (Box et al., 2011). Thus
To that end, we are using genome-wide single-nucleotide polymorphisms (SNPs)
the contrasting spur lengths of L.
obtained through a genotyping-by-sequencing (GBS) approach. Mean

Mean±SD L. clementei L. salzmannii

1 cm
clementei and L. salzmannii are likely
1.5
Nectar volume (µl)

attributable to changes in the rate of

● 80 individuals of the study clade representing the whole distribution of the eight
cell expansion. We will verify this
species were sampled. 16 individuals of additional species of Linaria sect.
1.0 using Scanning Electron Microscopy.
Versicolores were sampled to be used as the outgroup.
● We will then use RNA-Seq to compare transcriptome profiles of the two species
● A GBS library was prepared using the PstI-HF restriction enzyme and following the 0.5 during spur development. This will allow us to identify genes with a likely role in
protocols of Grabowski et al. (2014) and Escudero et al. (2014) with minor growth rate variation and thus in spur length evolution.
modifications.
0.0
ind.1 ind.2 ind.3 ind.1 ind.2 ind.3
● Our observations suggest that pollinators of L. clementei

PRELIMINARY CONCLUSIONS
● The Bioanalyzer profile L. clementei L. salzmannii
collect pollen deposited on the lower lip (palate) of the
(left) indicates successful
● Nectar quantification confirmed the corolla. Spur reduction in this species would be related to a
amplification of the library,
presence of nectar in L. salzmannii, but shift in main reward type from nectar to pollen. We are now ● Macroevolutionary analyses are consistent with a role of nectar spurs as a key innovation
which will be submitted to
revealed its absence in L. clementei investigating additional traits that may have evolved in promoting speciation in Antirrhineae.
Illumina single-end
(above). This indicates that rewards other correlation with this shift. For instance, the higher density of ● However, our study system does not fit a classical pollinator shift hypothesis in which
sequencing. SNP calling
than nectar (e.g. pollen) are involved in trichomes on the palate of L. clementei (above) may divergent spur lengths constitute adaptations to nectar-feeding pollinators with different
and phylogenetic
pollinator attraction in L. clementei. facilitate pollen presentation. tongue lengths. Instead, spur reduction in L. clementei, as compared to L. salzmannii,
analyses will be
conducted using seems to be the result of loss of function as a consequence of a shift from nectar to
● While post-pollination barriers between study species seem to be weak (cf. Viano, 1978), ecological niche (presumably) pollen as the main reward for pollinators.
recently-developed
modelling revealed that habitat isolation (based on divergent lithologies), rather than pollinators, may be ● Diverse ontogenetic mechanisms seem to be involved in spur length variation in distantly
software.
the main factor reproductively isolating L. clementei and L. salzmannii. related plant lineages.

REFERENCES
• Box MS, Dodsworth S, Rudall PJ, Bateman RM, Glover BJ (2011). The Plant Journal 68(4): 703-714.
• Escudero M, Eaton DAR, Hahn M, Hipp A (2014). Molecular Phylogenetics and Evolution 79: 359-367.
• Fernández-Mazuecos M, Blanco-Pastor JL, Gómez JM, Vargas P (2013). Annals of Botany 112(9): 1705-1722.
• Fernández-Mazuecos M, Vargas P (2015). Plant Systematics and Evolution 301(5): 1411-1423.
• FitzJohn RG, Maddison WP, Otto SP (2009). Systematic Biology 58(6): 595-611.
• Grabowski PP, Morris GP, Casler MD, Borevitz JO (2014). Molecular Ecology 23(16): 4059-4073.
ACKNOWLEDGEMENTS
This project is funded by the European Comission through a Marie Curie Intra-European
• Puzey JR, Gerbode SJ, Hodges SA, Kramer EM, Mahadevan M (2012). Proceedings of the Royal Society B 279(1733): 1640-1645.
Fellowship (LINARIA-SPECIATION). We thank Pablo Vargas, Concepción Ornosa, José Luis
• Kay KM, Voelckel C, Yang JY, Hufford KM, Kaska DD, Hodges SA (2006). In: Harder LD, Barrett SCH (Eds.) Ecology and Evolution
Blanco-Pastor, Marcial Escudero and all members of Beverley Glover’s lab for their support.
of Flowers. Oxford University Press. Pp. 311-325
Antirrhinum photo by Isabel Liberal.
• Viano J (1978). Caryologia 31, 383-425.