Flower color polymorphism and speciation: Lysimachia monelli and Lysimachia arvensis as study systems

Flower color polymorphism and speciation: Lysimachia monelli and Lysimachia arvensis as study systems PhD Thesis María de las Mercedes Sánchez Cabrera Flower color polymorphism and speciation: Lysimachia monelli and Lysimachia arvensis as study systems Polimorfismo del color floral y especiación: Lysimachia monelli y Lysimachia arvensis como sistemas de estudio Tesis Doctoral María de las Mercedes Sánchez Cabrera Sevilla, 2022 Dpto. Biología Vegetal y Ecología Flower color polymorphism and speciation: Lysimachia monelli and Lysimachia arvensis as study systems Memoria presentada por la graduada en Biología María de las Mercedes Sánchez Cabrera para optar al título de Doctora en Biología por la Universidad de Sevilla. Directores: Dr. Montserrat Arista Palmero Dr. Pedro L. Ortiz Ballesteros Dr. Eduardo Narbona Fernández Este trabajo fue financiado por los proyectos del Plan nacional de investigación concedidos por el Ministerio de Economía y Competitividad del Gobierno de España (MINECO) a Montserrat Arista y Pedro L. Ortiz (proyectos CGL2012-33270 y CGL2015-63827), y Ministerio de Ciencia e Innovación (MICIN) a Montserrat Arista y Eduardo Narbona (proyecto PID2020-116222GB-100). La autora, María de las Mercedes Sánchez Cabrera, fue beneficiaria de una beca para la formación de Personal Investigador del MINECO (BES-C-2016-0023), que incluía financiación para estancias en el extranjero, realizada en University of Santa Clara, y una beca de movilidad internacional Erasmus de la Universidad de Sevilla para una estancia en la Università degli Studi di Sassari. A mi padre A mi madre A mi hermana ÍNDICE RESUMEN / ABSTRACT……………………………………………………………..11 INTRODUCCIÓN GENERAL………………………………………………………...15 El color de las flores…………………………………………………………....17 La bioquímica y la genética del color floral…………………………………....18 Variaciones del color floral y sus causas moleculares……………………….…22 El papel del color en la especiación………………………………………….…25 Los sujetos de estudio…………………………………………………………..26 El objetivo y la organización de esta memoria de tesis doctoral……………….29 Referencias……………………………………………………….......................31 CAPÍTULO 1 - Reproductive isolation between the blue and orange lineages of Lysimachia monelli. ………………………………………………………....................41 Abstract……………………………………………………………………...….42 Introduction………………………………………………………......................43 Material and Methods………………………………………………………......46 Results………………………………………………………..............................52 Discussion………………………………………………………........................65 Conclusions………………………………………………………......................67 References………………………………………………………........................68 CAPÍTULO 2 - Understanding biogeographical patterns and genetic diversity of Lysimachia monelli (Primulaceae), a flower colour polymorphic species endemic to the Mediterranean Basin………………………………………………………....................73 Abstract…………………………………………………………………………74 Introduction………………………………………………………......................75 Material and Methods………………………………………………………......77 Results………………………………………………………..............................83 Discussion………………………………………………………........................95 Conclusions………………………………………………………......................99 References………………………………………………………......................100 CAPÍTULO 3 - Changes at a critical branchpoint in the anthocyanin biosynthetic pathway underlie the blue to orange flower color transition in Lysimachia arvensis...109 Abstract………………………………………………………………………..110 Introduction………………………………………………………....................111 Material and Methods………………………………………………………....115 Results………………………………………………………............................119 Discussion………………………………………………………......................133 Conclusions………………………………………………………....................141 References………………………………………………………......................142 CAPÍTULO 4 - Flower color ancestral polymorphism transcends speciation………..153 Abstract………………………………………………………………………..154 Introduction………………………………………………………....................155 Material and Methods……………………………………………………........158 Results………………………………………………………............................163 Discussion………………………………………………………......................169 References………………………………………………………......................173 DISCUSIÓN GENERAL……………………………………………………..…........181 Implicaciones biológicas y ecológicas del polimorfismo del color floral…….182 Implicaciones bioquímicas y genéticas en el polimorfismo del color floral….184 Origen del polimorfismo del color floral y especiación en Lysimachia arvensis y L. monelli……………………………………………………….......................185 Perspectivas futuras para el estudio de los procesos de especiación en Lysimachia……………………………………………………….....................187 16 Introducción General 17 El color de las flores Las flores de las angiospermas presentan una gran variabilidad de tamaño, color, olor y forma. Se acepta que la enorme variedad de estos rasgos florales es, al menos en parte, el resultado de selección mediada por sus polinizadores (Stebbins, 1970; Feinsinger, 1983). Los polinizadores seleccionan las flores basándose en una variedad de señales visuales y olfativas y, por lo tanto, tienen el potencial de dar forma a la diversidad de rasgos florales (Faegri and van der Pijl, 1979). El color floral varía enormemente a diversas escalas taxonómicas, tanto dentro como entre especies, géneros y familias (Beardsley et al., 2003; Rausher, 2008; McEwen and Vamosi, 2010; Trunschke et al., 2021). Es frecuente encontrar especies hermanas que muestran patrones de color muy diferentes (e.g. Clegg and Durbin, 2000; Roguz et al., 2020; Skeels et al., 2021). Incluso a nivel intraespecífico son frecuentes las variaciones de color tanto de forma continua, como de forma discreta, dando origen a polimorfismos del color floral (Narbona et al., 2018; Sapir et al., 2021). Todos estos cambios se ven favorecidos por el hecho de que el color floral es un carácter que tiene una señal filogenética débil (Ortiz et al., 2021; Skeels et al., 2021), lo que lo hace susceptible a respuestas rápidas frente a presiones selectivas. El color de las flores está sujeto a presiones selectivas debidas tanto a factores bióticos como abióticos (Dalrymple et al., 2020). Entre los factores bióticos responsables de la variación del color floral, los polinizadores desempeñan un papel dominante (Stebbins, 1974; Meléndez-Ackerman and Campbell, 1998; Weiss, 2001), ya que el color es uno de los rasgos visuales más importantes que utilizan para localizar las flores y asociarlas con la oferta de recompensas (Waser and Price, 1985; Meléndez-Ackerman et al., 1997; Schoonhoven et al., 2007; Kelber and Osorio, 2010). Algunos polinizadores están condicionados por preferencias innatas por determinados colores florales (Giurfa et al., 1995; Lunau and Maier, 1995; Chittka et al., 2004; Raine and Chittka, 2007), pero muchos de ellos son capaces de superar ese condicionamiento si asocian el color con una determinada recompensa floral (Cate and Rowe, 2007; Russell et al., 2016). La importancia del color floral como diana de la selección mediada por polinizadores queda reflejada en el hecho de que este carácter es uno de los componentes de los llamados "síndromes de polinización" (Stebbins, 1974; Faegri and van der Pijl, 1979; Reverté et al., 2016). En el marco de los síndromes de polinización, los cambios en el color de las Introducción General 18 flores representan adaptaciones a diferentes grupos de polinizadores (Fenster et al., 2004; Schiestl and Johnson, 2013). Diversos trabajos han informado de selección direccional sobre el color floral mediada por polinizadores (e.g. Jones and Reithel, 2001; Ortiz et al., 2015). Sin embargo, el conjunto de polinizadores que visitan una determinada especie vegetal es a menudo diverso y varía tanto espacial como temporalmente (Herrera, 1996; Waser et al., 1996); incluso en una misma población muchas plantas son visitadas simultáneamente por polinizadores diversos. Puesto que distintos grupos de polinizadores presentan preferencias y sistemas visuales distintos, el color floral puede estar sometido a selección balanceada (Kellenberger et al., 2019). El color de las flores se debe fundamentalmente a la presencia de pigmentos, siendo las antocianinas y los carotenoides los más frecuentes (véase más abajo). Estos pigmentos, sobre todo las antocianinas, se producen en una ruta metabólica en las que se originan también compuestos secundarios que protegen a las plantas y a las flores frente a distintos tipos de estrés, tanto bióticos – herbívoros o patógenoscomo abióticos – temperaturas extremas, radiación ultravioleta, sequía o salinidad- (Strauss and Whittall, 2006; Landi et al., 2015; Narbona et al., 2021). Es por eso que, junto con los polinizadores, cualquiera de estos factores puede estar implicado en la evolución del color floral. Así por ejemplo, se ha documentado que el color floral puede ser usado como señal por los herbívoros y estar sometido a selección ejercida por éstos (Caruso et al., 2010). Igualmente, se han documentado distintos casos de selección sobre el color floral mediada por alguno de los citados factores abióticos (Arista et al., 2013). Es frecuente que varias de estas presiones selectivas actúen simultáneamente sobre el color floral, a veces con efectos aditivos y otras con efectos contrapuestos (Strauss and Irwin, 2004; Strauss and Whittall, 2006; Caruso et al., 2010). En algunos casos, incluso, son estos otros factores ajenos a la polinización los principales responsables de la selección sobre el color floral (Souto‐ Vilarós et al., 2018). La bioquímica y la genética del color floral El color floral es la percepción, mediada por el sistema visual del observador, de la radiación reflejada por los pétalos, la cual depende de las propiedades fisicoquímicas de éstos (Endler, 1992; van der Kooi et al., 2019). Por una parte, existen colores “estructurales” producidos por formas o disposiciones peculiares de las células de los pétalos que modifican las longitudes de onda de la luz reflejada y, por tanto, el color Introducción General 19 percibido (Moyround and Glover, 2017; Garcia et al., 2019). Si bien en los últimos años se han documentado varios casos muy llamativos en especies filogenéticamente muy alejadas (Moyround et al., 2017), el color estructural parece ser un proceso relativamente raro en plantas si lo comparamos con otros grupos taxonómicos como insectos o aves (Moyround and Glover, 2017). Así, el color de las flores se considera principalmente pigmentario, es decir, color producido por pigmentos que se acumulan en los tejidos florales. Los pigmentos son compuestos químicos que, cuando se exponen a la luz, pueden absorber ciertas longitudes de onda y reflejar las restantes. En última instancia, la longitud de onda específica que absorben los pigmentos depende de su estructura química (Harborne, 2014). Existen cuatro grandes grupos de pigmentos florales: clorofilas, carotenoides, flavonoides y betalaínas (Tanaka et al., 2008; Narbona et al., 2021a). Recientemente se ha demostrado que estos grupos de pigmentos son diferenciados por los polinizadores ya que producen colores y contrastes cromáticos distintos entre sí (Narbona et al., 2021b). Las clorofilas producen coloración verde y betalaínas generan coloraciones rosas, rojas o amarillas, pero son pigmentos relativamente poco frecuentes en las flores (Timoneda et al., 2019). Los carotenoides, comúnmente conocidos por producir color naranja en la zanahoria, pueden producir además coloraciones amarillas y rojas. No obstante, los carotenoides amarillos son con diferencia los más frecuentes en tejidos florales, formando parte de miles de especies con flores amarillas en las familias Asteraceae, Fabaceae o Amaryllidaceae (Ohmiya, 2001; Tanaka et al., 2008). Los flavonoides son sin duda el grupo de pigmentos más diverso y frecuente en las flores (Schwinn and Davies, 2004). Desde el punto de vista bioquímico existen a su vez varios subgrupos que generan coloraciones muy distintas (Tanaka et al., 2008; Iwashina, 2015). Por una parte, flavonoles, flavonas e isoflavonas producen color blanco al ojo humano, pero absorben la radiación ultravioleta por lo que pueden generar patrones invisibles para nosotros, pero percibidos por muchos grupos de polinizadores como abejas, moscas, mariposas y aves (Narbona et al., 2021a); de hecho, estos compuestos se encuentran frecuentemente en las guías de néctar (Harbone and Grayer, 1994). Las auronas y chalconas, así como algunos flavonoles glicosilados producen colores amarillos y crema. Por último, las antocianinas son el grupo de pigmentos que produce una mayor gama de colores florales de rosa a rojo pasando por el azul (Iwashina, 2015). Las antocianinas son formas glicosiladas de las antocianidinas, las estructuras centrales del pigmento. Existen más de 20 tipos de antocianidinas, pero son solo seis las más frecuentes en las flores: cianidina (tonalidades rosas a rojas), delfinidina (tonalidades azules), pelargonidina Introducción General 20 (tonalidades rojas a naranjas), peonidina (tonalidades rosas a púrpuras), petunidina (tonalidades azules a púrpuras) y malvidina (tonalidades azules a violetas) (Castañeda- Ovando et al., 2009; Davies, 2009). Sin embargo, las tonalidades de las antocianinas son muy variables ya que dependen del nivel de glicosilación y acilación, presencia de copigmentos y pH vacuolar, entre otros factores (Castañeda-Osvaldo et al., 2009; van der Kooi et al., 2019). Las antocianinas se producen en una rama biosintética derivada de la ruta de los flavonoides, que a su vez proviene de la de los fenilpropanoides (Tanaka et al., 2008; Davies, 2009; Okitsu et al., 2018). Generalmente se considera que en la ruta biosintética de las antocianinas (ABP en inglés) intervienen al menos seis enzimas principales, que están altamente conservados en todas las angiospermas (Ho and Smith, 2016). Algunas de ellas sirven también como puntos de ramificación para la producción de otros flavonoides importantes, como chalconas o flavonoles (Tanaka et al., 2008; Figura 1). La ABP se puede a su vez dividirse en tres ramas biosintéticas: la directa que daría lugar a las pelargonidinas, la rama de las cianidinas (que añade un grupo hidroxilo mediante la enzima flavonoide 3′-hidroxilasa; F3’H) y la rama de las delfinidinas (que añade dos grupos hidroxilo mediante la enzima flavonoide 3′,5′-hidroxilasa; F3’5’H). Las otras antocianidinas se producen por metilaciones sucesivas de la cianidina y la delfinidina (Castañeda-Ovando et al., 2009; Figura 1). La ABP es una de las rutas metabólicas mejor estudiadas en plantas (Kopp, 2009; Wheeler and Smith, 2019). Esto ha permitido que se conozca con bastante detalle su regulación y por tanto las bases genéticas de los cambios de color producidos por pigmentos antociánicos (Fattorini and Glover, 2020). La regulación de la ruta está controlada principalmente por el complejo regulador de la transcripción MYB-bHLH-WDR, que contiene los factores R2R3-MYB (MYB), bHLH y WD40 (Xu et al., 2015). Estos factores de transcripción pueden regular la actividad de uno o varios genes estructurales, pero los pasos exactos de la ABP que controla cada proteína pueden ser específicos de cada célula o tejido (Albert et al., 2014). Esta regulación tan especifica espacial y temporal permite que la biosíntesis de antocianinas se active o desactive en los órganos florales de forma independiente a la de los órganos vegetativos, lo cual puede ser ventajoso frente a posibles efectos pleiotrópicos (Streisfeld and Rausher, 2011). Además, permitiría la producción de patrones de pigmentación floral tales como máculas, guías de néctar o patrones de color complejos como los producidos en las orquídeas (Martins et al., 2017; Li et al., Introducción General 21 2021). Por otra parte, las mutaciones que afectan a la producción de antocianinas a nivel floral pueden ocurrir directamente sobre los genes estructurales de la ABP o bien en regiones reguladoras en cis de estos genes o de los factores de transcripción reguladores MYB-bHLH-WDR (Albert et al., 2014; Ho and Smith, 2016). Sin embargo, solo las dos últimas podrían eludir los efectos pleiotrópicos negativos de una falta de antocianinas en el resto de los tejidos de la planta, por lo que tienen mayor probabilidad de fijación evolutiva (Sobel and Streisfeld, 2013; Chesmore et al., 2016). Figura 1. Esquema general de la ruta biosintética de las antocianinas. Introducción General 22 Variaciones de color floral y sus causas moleculares Como ya se ha mencionado, las variaciones de color floral son frecuentes a distintas escalas taxonómicas. Estas variaciones de color pueden ser simplemente de intensidad, produciéndose por cambios de concentración de pigmentos (ej. intensidades de rosa; Rodríguez-Castañeda et al., 2020). También pueden producirse cambios del tipo de color derivados de variaciones en la composición de tipos pigmentos (ej. cambio de color amarillo a rojo o de azul a rojo; Ogutcen et al., 2020). El caso de las flores blancas es especial ya que esta colación se produce por la pérdida de pigmentos (Narbona et al. 2021b). Las variaciones de color floral producidas por cambios en la composición o pérdidas de pigmentos son mucho más importantes a nivel evolutivo que las producidas por cambio de concentración ya que están generalmente relacionados con procesos evolutivos (Sobel and Streisfeld, 2013; Narbona et al., 2018; Trunschke et al., 2021). La pérdida de pigmentación en las flores blancas es mucho más frecuente que los cambios de composición, tanto en especies polimórficas (Warren and Mackenzie 2001; Narbona et al. 2018), como entre especies filogenéticamente relacionadas (Ellis and Field, 2016, Landis et al., 2018). Esta pérdida de pigmentación se ha producido recurrentemente en la evolución de las angiospermas (Smith and Goldberg, 2015) y a nivel molecular se produce generalmente por cambios en la expresión génica, en lugar de por mutaciones en genes estructurales (Casimiro-Soriguer et al., 2016; Ho and Smith, 2016; Gates et al., 2017). Los cambios de color producidos por diferente composición de pigmentos pueden también involucrar pérdidas y ganancias de pigmentos, como por ejemplo el caso de Mimulus aurantiacus, en el cual existen morfotipos amarillos por acumulación de carotenoides y morfotipos rojos que, además de carotenoides, acumulan antocianinas (Streisfeld et al., 2013). También puede ocurrir cambios en la producción del tipo de pigmento predominante, como en el género Iocroma (Solanaceae), en el que las especies rojas acumulan pelargonidinas mientras que las azules acumulan delfinidinas (Larter et al., 2018). Existen además casos en el que se producen cambios de pigmentos y de su concentración, lo que genera morfotipos bien definidos de color floral, como en Phlox drumondii (Hopkins and Rausher, 2011) o especies con colores muy diferentes entre sí, como en las especies de Gesnerioideae (Ogutcen et al., 2020). Introducción General 23 La variación de color floral azul - rojo ha generado un relativo interés entre la comunidad de ecólogos evolutivos pues son cambios de color inducidos generalmente por presiones selectivas debidas a los polinizadores (Rausher, 2008; Ortiz-Barrientos, 2013; pero ver Arista et al., 2013). Así, las especies de Penstemon (Plantaginaceae) de flores rojas contienen pelargonidinas y son polinizadas por colibríes, mientras que las de flores azules contienen delfinidinas y son polinizadas por himenópteros (Wessinger and Rausher, 2014; Wessinger et al., 2019; Figura 2). Las especies de Ipomoea (Convolvulaceae), de flores azules-purpuras producen cianidinas y son generalmente polinizadas por abejas, mientras que las flores rojas producen pelargonidinas y son polinizadas por colibríes (Streisfeld and Rausher, 2009; Figura 2). Otro ejemplo sería el que ocurre en especies hermanas de Petunia (Solanaceae), así P. exserta acumula delfinidinas y cianidinas y es polinizada por colibríes mientras que P. secreta acumula malvidinas y petunidinas y es polinizada por abejas (Esfeld et al., 2018; Berardi et al., 2021; Figura 2). Las bases genéticas de los cambios de producción de antocianinas que provocan el cambio floral azul a rojo están relativamente bien estudiadas. En una revisión reciente, Wheeler and Smith (2019) observan que en los 15 casos de especies silvestres, los cambios entre tipos de antocianinas implicaban el bloqueo de los genes de las ramas de producción de cianidinas y delfinidinas y sus antocianidinas derivadas por metilación, es decir F3’H, F3’5’H o ambos, mientras que en otros dos casos implicaban también mutaciones en el gen dihidroflavonol-4-reductasa (DFR). En este último caso, la enzima DFR experimentaría un cambio en la especificidad del sustrato de forma que se produciría una pérdida de la capacidad de metabolizar sustratos de una determinada rama de la ABP (Forkman and Ruhnau, 1987; Zufall and Rausher, 2004). Recientemente se ha encontrado que las bases genéticas del color rojo de las flores de Petunia exserta son complejas e implican un cambio a nivel espacial de la acción de un regulador MYB, un reequilibrio de la expresión de los genes hidroxiladores y la regulación a la baja de una aciltransferasa, lo que produce matices rojizos en la delfinidinas, que son típicamente azules o púrpuras. Introducción General 24 Figura 2. Ejemplos de variaciones de color floral azul y rojo a nivel interespecífico o intraespecífico. (A) Penstemon grandiflorus (Nebraska, USA). (B) Penstemon barbatus (Arizona, USA). (C) Ipomoea nil (Taiwan). (D) Ipomoea quamoclit (Brasil). (E) Phlox drummondii (morfotipo rosa, Texas, USA). (F) Phlox drummondii (morfotipo rojo, Texas, USA). Fuente fotografías: iNaturalist (www.inaturalist.org). Introducción General 25 El papel del color en especiación En visión de Darwin (1859), la especiación es simplemente el resultado de una adaptación divergente mediada por la selección natural. Nuestra comprensión del proceso de especiación mejoró con las aportaciones de Mayr (1942, 1963), que definió la especiación como un proceso que genera grupos de poblaciones naturales que pueden cruzarse entre sí, estando aisladas reproductivamente de otras similares. Dobzhansky (1937) introdujo la idea de que el aislamiento reproductivo evoluciona entre las poblaciones alopátricas a medida que éstas acumulan mutaciones diferentes que resultan incompatibles en caso de nuevo contacto entre ellas. Posteriormente, se encontraron evidencias de que la especiación también puede ocurrir en simpatría (Schliewen et al., 1994; Gislason et al., 1999; Bolnick and Fitzpatrick, 2007). A diferencia de la alopátrica, la especiación simpátrica se produce en presencia de flujo génico entre los grupos divergentes, y se ha argumentado que para que esta especiación ocurra se requiere la concurrencia de selección natural divergente y de mecanismos que promuevan el apareamiento no aleatorio (Coyne, 1992; Dieckmann and Doebeli, 1999; Wu, 2001; Gavrilets, 2003; Papadopulos et al., 2011; Servedio, 2016). En estos casos, la especiación se ve facilitada si los caracteres sujetos a selección divergente, u otros correlacionados genéticamente, afectan a la compatibilidad reproductiva o la probabilidad de apareamiento (Servedio et al., 2011). El polimorfismo de color floral, definido como la presencia de dos o más morfotipos de color en una especie, genéticamente determinados e interfértiles, es uno de los caracteres que promueven la selección divergente y apareamiento no-azaroso (Gray and McKinnon, 2007; Forsman et al., 2008; McLean and Stuart‐Fox, 2014). Así, el polimorfismo del color se considera un “carácter mágico”, es decir, un carácter codificado por genes que están sujetos a selección divergente y que afectan pleiotrópicamente al aislamiento reproductivo (Servedio et al., 2011; Servedio and Brger, 2020). Aunque poco frecuente, el polimorfismo del color está representado en grupos taxonómicamente muy diversos (Hugall and Stuart-Fox, 2012). En plantas, el polimorfismo del color afecta principalmente a la flor, y se ha documentado en al menos 21 familias y más de 50 géneros (Narbona et al., 2014, 2018). Como se ha indicado antes, los polinizadores son importantes agentes de selección que pueden promover la diversificación en plantas. En relación con el polimorfismo del color Introducción General 32 Caruso, C. M., Scott, S. L., Wray, J. C., and Walsh, C. A. (2010). Pollinators, herbivores, and the maintenance of flower color variation: a case study with Lobelia siphilitica. Int. J. Plant Sci. 171: 1020-1028. Casimiro-Soriguer, I., Narbona, E., Buide, M. L., del Valle, J. C., and Whittall, J. B. (2016). Transcriptome and Biochemical Analysis of a Flower Color Polymorphism in Silene littorea (Caryophyllaceae). Front. Plant Sci. 7, 204. Castaeda-Ovando, A., Pacheco-Hernández, M. L., Páez-Hernández, M. E., Rodrguez, J. A. and Galán-Vidal, C. A. (2009). Chemical studies of anthocyanins: a review. Food Chem. 113: 859-871. Cate, C. T., and Rowe, C. (2007). Biases in signal evolution: learning makes a difference. Trends Ecol. Evol. 22, 380-387. Chesmore, K. N., Bartlett, J., Cheng, C., and Williams, S. M. (2016). Complex patterns of association between pleiotropy and transcription factor evolution. Genome Biol. Evol. 8, 3159-3170. Chittka, L., Ings, T. C., and Raine, N. E. (2004). Chance and adaptation in the evolution of island bumblebee behaviour. Popul. Ecol. 46, 243-251. Clegg, M. T., and Durbin, M. L. (2000). Flower color variation: a model for the experimental study of evolution. Proc. Natl. Acad. Sci. U.S.A. 97, 7016-7023. Coyne, J. A. (1992). Genetics and speciation. Nature 355, 511-515. Dalrymple, R. L., Kemp, D. J., Flores‐Moreno, H., Laffan, S. W., White, T. E., Hemmings, F. A., et al. (2020). Macroecological patterns in flower colour are shaped by both biotic and abiotic factors. New Phytol. 228, 1972-1985. Darwin C. (1859). On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London, UK: John Murray. Davies, K. M. (2009). “Modifying anthocyanin production in flowers,” in Anthocyanins, eds Gould, K. S., Davies, K. M., and Winefield, C. New York, NY: Springer, pp49-83. Dieckmann, U., and Doebeli, M. (1999). On the origin of species by sympatric speciation. Nature 400, 354-357. Dobzhansky, T. (1937). Genetics and the origin of species. Columbia Univ. Press, New York. Ellis, T. J., and Field, D. L. (2016). Repeated gains in yellow and anthocyanin pigmentation in flower colour transitions in the Antirrhineae. Ann. Bot. 117, 1133-1140. Endler, J. A. (1992). Signals, signal conditions, and the direction of evolution. Am. Nat. 139, 125-153 Esfeld, K., Berardi, A. E., Moser, M., Bossolini, E., Freitas, L., and Kuhlemeier, C., (2018). Pseudogenization and resurrection of a speciation gene. Curr. Biol. 28, 3776- 3786. Introducción General 33 Faegri, K., and van der Pijl, K. (1979). The principles of pollination ecology. Pergamon Press. Fattorini, R., and Glover, B. J. (2020). Molecular mechanisms of pollination biology. Annu. Rev. Plant Biol. 71, 487-515. Feinsinger, P. (1983). Coevolution and pollination. In Futuyma, D. J. and Slatkin, M. (Eds.), Coevolution. Sunderland: Massachusetts: Sinauer. pp. 285-319 Fenster, C. B., Armbruster, W. S., Wilson, P., Dudash, M. R., and Thomson, J. D. (2004). Pollination syndromes and floral specialization. Annu. Rev. Ecol. Evol. Syst. 35, 375-403. Forkmann, G., and Ruhnau, B. (1987). Distinct substrate specificity of dihydroflavonol 4-reductase from flowers of Petunia hybrida. Z. Naturforsch. 42, 1146-1148. Forsman, A., Ahnesjö, J., Caesar, S., and Karlsson, M. (2008). A model of ecological and evolutionary consequences of color polymorphism. Ecology 89, 34-40. Freyre, R., and Griesbach, R. J. (2004). Inheritance of flower color in Anagallis monelli L. HortScience 39, 1220-1223. Garcia, J. E., Shrestha, M., Howard, S. R., Petersen, P., and Dyer, A. G. (2019). Signal or cue: the role of structural colors in flower pollination. Curr. Zool. 65, 467-481. Gates, D. J., Olson, B. J., Clemente, T. E., and Smith, S. D. (2018). A novel R3 MYB transcriptional repressor associated with the loss of floral pigmentation in Iochroma. New Phytol. 217, 1346-1356. Gavrilets, S. (2003). Perspective: models of speciation: what have we learned in 40 years?. Evolution 57, 2197-2215. Gibbs, P. E., and Talavera, (S. 2001). Breeding system studies with three species of Anagallis (Primulaceae): self-incompatibility and reduced female fertility in A. monelli L. Ann. Bot. 88, 139-144. Gislason, D., Ferguson, M. M., Skulason, S., and Snorrason, S. S. (1999). Rapid and coupled phenotypic and genetic divergence in Icelandic arctic char (Salvelinus alpinus). Can. J. Fish. Aquat. Sci. 56, 2229-2234. Giurfa, M., Nunez, J., Chittka, L., and Menzel, R. (1995). Colour preferences of flowernaive honeybees. J. Comp. Physiol. 177, 247-259. Grant, V. 1981). Plant speciation. Columbia university press. Gray, S. M., and McKinnon, J. S. (2007). Linking color polymorphism maintenance and speciation. Trends Ecol. Evol. 22, 71-79. Harborne, J. B. (2014). Introduction to ecological biochemistry. Cambridge, UK: Academic Press. Harbone, J. B., and Grayer, R. J. (1994). Flavonoids and insects. In Harborne, J. B. (ed.). The flavonoids. Advances in research since 1986. London, UK: Chapman and Hall, Introducción General 34 pp.589-618. Herrera, C. M. (1996). Floral traits and plant adaptation to insect pollinators: a devil’s advocate approach. In Floral biology. Springer, Boston, MA. pp. 65-87 Ho, W. W., and Smith, S. D. (2016). Molecular evolution of anthocyanin pigmentation genes following losses of flower color. BMC Evol. Biol. 16, 1-10. Hopkins, R., and Rausher, M. D. (2011). Identification of two genes causing reinforcement in the Texas wildflower Phlox drummondii. Nature 469, 411-414. Hugall, A. F., and Stuart-Fox, D. (2012). Accelerated speciation in colour-polymorphic birds. Nature 485, 631-634. Iwashina, T. (2015). Contribution to flower colors of flavonoids including anthocyanins: A review. Nat. Prod. Commun. 10, 529-544. Jiménez-López F. J. (2019). Polimorfismo en el color floral y especiación en Lysimachia arvensis. Ph.D. dissertation, University of Seville, Spain. Jiménez-López, F. J., Matas, L., Arista, M., and Ortiz, P. L. (2020a). Flower colour segregation and flower discrimination under the bee vision model in the polymorphic Lysimachia arvensis. Plant Biosyst. 154, 535-543. Jiménez-López, F. J., Ortiz, P. L., Talavera, M., Pannell, J.R., and Arista, M. (2020b). The role of lateral and vertical herkogamy in the divergence of the blue-and red-flowered lineages of Lysimachia arvensis. Ann. Bot. 125, 1127-1135. Jiménez-López, F. J., Viruel, J., Arista, M., Ortiz, P. L., and Talavera, M. (2022). Molecular approaches reveal speciation between red-and blue-flowered plants in the Mediterranean Lysimachia arvensis and L. monelli (Primulaceae). Bot. J. Linn. Jones, N. K., and Reithel, J. S. (2001). Pollinator‐mediated selection on a flower color polymorphism in experimental populations of Antirrhinum (Scrophulariaceae). Am. J. Bot. 88, 447-454. Kay, K. M., and Sargent, R. D. (2009). The role of animal pollination in plant speciation: integrating ecology, geography, and genetics. Annu. Rev. Ecol. Evol. Syst. 40, 637-656. Kelber A. and Osorio D. (2010). From spectral information to animal colour vision: experiments and concepts. Proc. R. Soc. B: Biol. Sci. 277, 1617-1625. Kellenberger, R. T., Byers, K. J., De Brito Francisco, R. M., Staedler, Y.M., LaFountain, A. M., Schönenberger, J., et al. (2019). Emergence of a floral colour polymorphism by pollinator-mediated overdominance. Nat. Commun. 10, 1-11. Kopp, A. (2009). Metamodels and phylogenetic replication: a systematic approach to the evolution of developmental pathways. Evol.; int. j. org. evol. 63, 2771-2789. Landi, M., Tattini, M., and Gould, K. S. (2015). Multiple functional roles of anthocyanins in plant-environment interactions. Environ. Exp. Bot. 119, 4-17. Introducción General 35 Landis, J. B., Bell, C. D., Hernandez, M., Zenil-Ferguson, R., McCarthy, E. W., Soltis, D. E., et al. (2018). Evolution of floral traits and impact of reproductive mode on diversification in the phlox family (Polemoniaceae). Mol. Phylogenet. Evol. 127, 878- 890. Larter, M., Dunbar-Wallis, A., Berardi, A. E., and Smith, S. D. (2018). Convergent evolution at the pathway level: predictable regulatory changes during flower color transitions. Mol. Biol. Evol. 35, 2159-2169. Li, X., Fan, J., Luo, S., Yin, L., Liao, H., Cui, X., et al. (2021). Comparative transcriptome analysis identified important genes and regulatory pathways for flower color variation in Paphiopedilum hirsutissimum. BMC Plant Biol. 21, 1-17. Lowry, D. B., Modliszewski, J. L., Wright, K. M., Wu, C. A., and Willis, J. H. (2008). The strength and genetic basis of reproductive isolating barriers in flowering plants. hilos. Trans. R. Soc. B: Biol. Sci. 363, 3009-3021. Lunau, K., and Maier, E. J. (1995). Innate colour preferences of flower visitors. J. Comp. Physiol. A 177, 1-19. Martins, T. R., Jiang, P., and Rausher, M. D. (2017). How petals change their spots: cisregulatory re-wiring in Clarkia (Onagraceae). New Phytol. 216, 510-518. Mayr, E. (1942). Systematics and the origin of species. Columbia Univ. Press, New York Mayr, E. (1963). Animal Species and Evolution, Harvard Univer-sity Press McEwen, J. R., and Vamosi, J. C. (2010). Floral colour versus phylogeny in structuring subalpine flowering communities. R. Soc. B: Biol. Sci. 277, 2957-2965. McLean, C. A., and Stuart‐Fox, D. (2014). Geographic variation in animal colour polymorphisms and its role in speciation. Biol. Rev. 89, 860-873. Meléndez‐Ackerman, E. and Campbell, D. R. (1998). Adaptive significance of flower color and inter‐trait correlations in an Ipomopsis hybrid zone. Evolution 52, 1293-1303. Meléndez-Ackerman, E., Campbell, D. R., and Waser, N. M. (1997). Hummingbird behavior and mechanisms of selection on flower color in Ipomopsis. Ecology 78, 2532- 2541. Moyroud, E., and Glover, B. J. (2017). The physics of pollinator attraction. New Phytol. 216, 350-354. Moyroud, E., Wenzel, T., Middleton, R., Rudall, P. J., Banks, H., Reed, A., et al. (2017). Disorder in convergent floral nanostructures enhances signalling to bees. Nature 550, 469-474. Narbona, E., Buide, M. L., Casimiro-Soriguer, I., and Del Valle, J. C. (2014). Polimorfismos de color floral: causas e implicaciones evolutivas. Ecosistemas 23, 36-47. Narbona, E., del Valle, J. C., and Whittall, J. B. (2021a). Painting the green canvas: How pigments produce flower colours. Biochemist 43, 6-12. Introducción General 36 Narbona, E., Del Valle, J. C., Arista, M., Buide, M. L., and Ortiz, P. L. (2021b). Major Flower Pigments Originate Different Colour Signals to Pollinators. Front. Ecol. Evol. 9, 743850. Narbona, E., Wang, H., Ortiz, P. L., Arista, M., and Imbert, E. (2018). Flower colour polymorphism in the Mediterranean Basin: occurrence, maintenance and implications for speciation. Plant Biol. 20, 8-20. Ohmiya, A. (2011). Diversity of carotenoid composition in flower petals. Jpn. Agric. Res. Q. 45, 163-171. Ogutcen, E., Durand, K., Wolowski, M., Clavijo, L., Graham, C., Glauser, G., et al. (2020). Chemical Basis of Floral Color Signals in Gesneriaceae: The Effect of Alternative Anthocyanin Pathways. Front. Plant Sci. 11, 604389. Okitsu, N., Noda, N., Chandler, S., and Tanaka, Y. (2018). Flower color and its engineering by genetic modification. In Ornamental Crops. Springer, Cham. pp. 29-62 Ortiz-Barrientos, D. (2013). The color genes of speciation in plants. Genetics 194, 39-42. Ortiz, P. L., Berjano, R., Talavera, M., Rodríguez-Zayas, L., and Arista, M. (2015). Flower colour polymorphism in Lysimachia arvensis: How is the red morph maintained in Mediterranean environments?. Perspect. Plant Ecol. Evol. Syst. 17, 142-150. Ortiz, P. L., Fernández‐Díaz, P., Pareja, D., Escudero, M., and Arista, M. (2021). Do visual traits honestly signal floral rewards at community level?. Funct. Ecol. 35, 369-383. Papadopulos, A. S., Baker, W. J., Crayn, D., Butlin, R. K., Kynast, R. G., Hutton, I., et al. (2007). Anatomical and biochemical studies of anthocyanidins in flowers of Anagallis monelli L.(Primulaceae) hybrids. Sci. Hortic. 112, 413-421. Pujadas, A. (1997). Anagallis L. In: Castroviejo, S., Aedo C, Laínz, M., Morales, R., Muñoz Garmendia, F., Nieto Feliner, G., Paiva, J. eds. Flora iberica, V. Ebenaceae- Saxifragaceae. Madrid: Real Jardín Botánico, CSIC, pp. 57–62. Raine, N. E., and Chittka, L. (2007). The adaptive significance of sensory bias in a foraging context: floral colour preferences in the bumblebee Bombus terrestris. PLoS One 2, e556. Rausher, M. D. (2008). Evolutionary transitions of flower color. Int. J. Plant Sci. 169, 7- 21. Reverté, S., Retana, J., Gómez, J. M., and Bosch, J. (2016). Pollinators show flower colour preferences but flowers with similar colours do not attract similar pollinators. Ann. Bot. 118, 249-257. Rieseberg, L. H., and Willis, J. H. (2007). Plant speciation. Science 317, 910-914. Rodrguez-Castaeda, N. L., Ortiz, P. L., Arista, M., Narbona, E., and Buide, M. L. (2020). Indirect Selection on Flower Color in Silene littorea. Front. Plant Sci. 11, 588383. Roguz, K., Gallagher, M. K., Senden, E., Bar-Lev, Y., Lebel, M., Heliczer, R., et al. Introducción General 37 (2020). All the colors of the rainbow: Diversification of flower color and intraspecific color variation in the genus Iris. Front. Plant Sci. 11, 569811. Russell, A. L., Golden, R. E., Leonard, A. S., and Papaj, D. R. (2016). Bees learn preferences for plant species that offer only pollen as a reward. Behav. Ecol. 27, 731-740. Sapir, Y., Gallagher, M. K., and Senden, E. (2021). What maintains flower colour variation within populations?. Trends Ecol. Evol. 36, 507-519. Schliewen, U. K., Tautz, D., and Paabo, S. (1994). Sympatric speciation suggested by monophyly of crater lake cichlids. Nature 368, 629–632 Schoonhoven, L. M., van Loon, J. J. A., and Dicke, M. (2007). Insect–plant biology. Oxford University Press, Oxford, UK. Schwinn, K. E., and Davies, K. M. (2009). “Flavonoids,” in Anthocyanins, Gould, K. S., Davies, K. M. and Winefield, C. (eds.) New York, NY: Springer, pp. 92–149. Servedio, M. R. (2016). Geography, assortative mating, and the effects of sexual selection on speciation with gene flow. Evol. Appl. 9, 91-102. Servedio, M. R., and Brger, R. (2020). The effectiveness of pseudomagic traits in promoting divergence and enhancing local adaptation. Evolution 74, 2438-2450. Servedio, M. R., Van Doorn, G. S., Kopp, M., Frame, A. M., and Nosil, P. (2011). Magic traits in speciation: ‘magic’ but not rare?. Trends Ecol. Evol. 26, 389-397. Skeels, A., Dinnage, R., Medina, I., and Cardillo, M. (2021). Ecological interactions shape the evolution of flower color in communities across a temperate biodiversity hotspot. Evolution letters, 5: 277-289. Smith, S. D., and Goldberg, E. E. (2015). Tempo and mode of flower color evolution. Am. J. Bot. 102, 1014-1025. Sobel, J. M., Chen, G. F., Watt, L. R., and Schemske, D. W. (2010). The biology of speciation. Evol.; int. j. org. evol. 64, 295-315. Sobel, J. M., and Streisfeld, M. A. (2013). Flower color as a model system for studies of plant evo-devo. Front. Plant Sci. 4, e321. Souto‐Vilarós, D., Vuleta, A., Jovanović, S. M., Budečević, S., Wang, H., Sapir, Y. et al. (2018). Are pollinators the agents of selection on flower colour and size in irises?. Oikos 127, 834-846. Stebbins, G. L. (1970). Adaptive radiation of reproductive characteristics in angiosperms, I: Pollination mechanisms. Annu. Rev. Ecol. Evol. Syst 1, 307-326. Stebbins, G. L. (1974). Adaptive shifts and evolutionary novelty: A compositionist approach. In Studies in the Philosophy of Biology. Palgrave, London. pp. 285-306 Strauss, S. Y., and Irwin, R. E. (2004). Ecological and evolutionary consequences of multispecies plant-animal interactions. Annu. Rev. Ecol. Evol. Syst. 35, 435-466. Introducción General 38 Strauss, S. Y., and Whittall, J. B. (2006). Non-pollinator agents of selection on floral traits. Ecology and evolution of flowers. In Harder, L.D., Barrett, S.C.H. (eds.) Ecology and evolution of flowers. Oxford University Press, Oxford. pp. 120-138 Streisfeld, M. A., and Rausher, M. D., (2009). Genetic changes contributing to the parallel evolution of red floral pigmentation among Ipomoea species. New Phytol. 183, 751-763. Streisfeld, M. A., and Rausher, M. D. (2011). Population genetics, pleiotropy and the preferential fixation of mutations during adaptive evolution. Evolution 65, 629-64 Streisfeld, M. A., Young, W. N., and Sobel, J. M. (2013). Divergent selection drives genetic differentiation in an R2R3-MYB transcription factor that contributes to incipient speciation in Mimulus aurantiacus. PLoS Genet. 9, e1003385. Savolainen, V. (2011). Speciation with gene flow on Lord Howe Island. Proc. Natl. Acad. Sci. U.S.A. 108, 13188-13193. Talavera, S., Gibbs, P. E., Fernández-Piedra, M. P., and Ortiz-Herrera, M. A. (2001). Genetic control of self-incompatibility in Anagallis monelli (Primulaceae: Myrsinaceae). Heredity 87, 589-597. Tanaka, Y., Sasaki, N., and Ohmiya, A. (2008). Biosynthesis of plant pigments: anthocyanins, betalains and carotenoids. Plant J. 54, 733-749. Timoneda, A., Feng, T., Sheehan, H., Walker‐Hale, N., Pucker, B., Lopez‐Nieves, S., et al. (2019). The evolution of betalain biosynthesis in Caryophyllales. New Phytol. 224, 71-85. Trunschke, J., Lunau, K., Pyke, G. H., Ren, Z. X., and Wang, H. (2021). Flower Color Evolution and the Evidence of Pollinator-Mediated Selection. Front. Plant Sci. 12, 617851. van der Kooi, C. J., Dyer, A. G., Kevan, P. G., and Lunau, K. (2019). Functional significance of the optical properties of flowers for visual signalling. Ann. Bot. 123, 263- 276. Warren, J., and Mackenzie, S. (2001). Why are all colour combinations not equally represented as flower‐colour polymorphisms?. New Phytol. 151, 237-241. Waser, N. M. (1986). Flower constancy: definition, cause, and measurement. Am. Nat. 127, 593-603. Waser, N. M. (2001). Pollinator behavior and plant speciation: looking beyond the “ethological isolation” paradigm. J. Pollinat. Ecol. 318-335. Waser, N. M., Chittka, L., Price, M. V., Williams, N. M., and Ollerton, J. (1996). Generalization in pollination systems, and why it matters. Ecology 77, 1043-1060. Waser, N. M., and Price, M. V. (1985). The effect of nectar guides on pollinator preference: experimental studies with a montane herb. Oecologia, 67, 121-126. Introducción General 39 Weiss, M. R. (2001). Vision and learning in some neglected pollinators: beetles, flies, moths, and butterflies. In: Chittka L. and Thomson T.D. (eds), Cognitive ecology of pollination: animal behavior and floral evolution. Cambridge University Press, Cambridge, UK, pp. 171– 190. Wessinger, C. A., Rausher, M. D., and Hileman, L. C. (2019). Adaptation to hummingbird pollination is associated with reduced diversification in Penstemon. Evol. Lett. 3, 521- 533. Wessinger, C. A., and Rausher, M. D. (2014). Predictability and irreversibility of genetic changes associated with flower color evolution in Penstemon barbatus. Evolution 68, 1058-1070. Wheeler, L. C., and Smith, S. D. (2019). Computational modeling of anthocyanin pathway evolution: biases, hotspots, and trade-offs. Integr. Comp. Biol. 59, 585-598. Widmer, A., Lexer, C., and Cozzolino, S. (2009). Evolution of reproductive isolation in plants. Heredity 102, 31-38. Wu, C. I. (2001). The genic view of the process of speciation. J. Evol. Biol. 14, 851-865 Xu, W., Dubos, C., and Lepiniec L. (2015). Transcriptional control of flavonoid biosynthesis by MYB–bHLH–WDR complexes. Trends Plant Sci. 20, 176-185. Zufall, R. A., and Rausher, M. D., (2004). Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428, 847-850. 40 41 CAPÍTULO 1 Reproductive isolation between blue and orange lineages of Lysimachia monelli Mercedes Sánchez-Cabrera, Montserrat Arista, Eduardo Narbona and Pedro L. Ortiz (Manuscript in preparation) Capítulo 1 48 of plants reaching flowering and their flower color. Additionally, we conducted three types of crosses between F1 plants: between blue-flowered plants from pure crosses, between orange-flowered plants from pure crosses, and between hybrid plants irrespective of flower color; we then recorded both fruit and seed production resulting from those crosses. Figure 1. Images showing the process to root the plants. (A) Lysimachia monelli cuttings obtained from natural populations. (B) Aeroponic cloner used to root the cuttings. (C) Cutting showing the new roots. (D) and (E), clones of the blue and orange lineages used as parents. 49 Table 1. Information about the plant material used for the reproductive barriers experiment. 1 Number of plants used for the crosses. Populatio n Id Flower color Location Latitud (Y) Longitud (X) Altitude (m) N plants1 1/17 Blue Spain. Huelva. Almonte 37°03'19.7"N 6°32'30.2"W 52 11 2/17 Blue Spain. Huelva. Hinojos 37°13'04.5"N 6°25'04.4"W 26 10 3/17 Blue Spain. Sevilla. Bormujos 37°22'18.7"N 6°04'55.7"W 98 11 4/17 Blue Spain. Sevilla. Bellavista 37°19'50.5"N 5°57'52.3"W 16 13 6/17 Blue Spain. Sevilla. Dos Hermanas 37°16'30.2"N 5°55'58.4"W 41 9 7/17 Blue Spain. Cádiz. Bornos 36°50'09.7"N 5°45'07.8"W 268 11 8/17 Blue Spain. Sevilla. Aznalcázar 37º16'40.8''N 6º14'09.8''W 32 8 10/17 Blue Spain. Sevilla. Utrera 37°09'44.5"N 5°45'36.9"W 62 12 LM_24 Blue Portugal 37°11'36.0"N 8°54'19.0"W 26 8 LM_28 Blue Portugal 38°00'55.2"N 8°49'11.1"W 61 8 MAR_2 Orange Morocco. Azrou 33°31'40.0"N 5°18'25.0"W 1455 5 MAR_4 Orange Morocco. Chefchaouen 35°11'01.6"N 5°13'28.4"W 1571 10 MAR_10 Orange Morocco. Ouezzane 34°52'48.0"N 5°31'48.0"W 200 13 MAR_12 Orange Morocco. Kenifra 33°01'40.6"N 5°19'28.7"W 1480 6 MAR_30 Orange Morocco. Zerhoun 34°04'08.7"N 5°27'20.0"W 765 7 MAR_32 Orange Morocco. Chefchaouene 35°05'00.0"N 5°10'00.0"W 1100 2 SAR_1 Orange Italy. Sardinia. Isola di San Pietro 39°11'06.8"N 8°18'21.5"E 12 7 SAR_2 Orange Italy. Sardinia. Isola di San Pietro 39°09'17.9"N 8°13'43.6"E 10 5 SAR_3 Orange Italy. Sardinia. Isola di San Pietro 39°08'38.8"N 8°13'26.6"E 58 13 SAR_6 Orange Italy. Sardinia. Sa Duchessa 39°22'20.7"N 8°35'51.0"E 322 17 50 To check whether fitness components depend on cross type by color, we performed a GLM for each evaluated component as dependent variable and cross type by color as independent one. We considered intra-lineage crosses (either BB or OO) irrespective of cross type by population, since no significant differences according to that factor were observed in most of the fitness components (Supplemental Table 1). For fruit-set, seedset, germination rate and fecundity we ran GLMs with a binomial error structure, while we ran a GLM with a Poisson error structure for number of ovules and number of seeds. We then performed a Tukey post-hoc test to check for differences in means among levels of fixed factors. All models were implemented in R v4.1.2 (R core team 2013) with package lme4 (Bates et al. 2015). Postzygotic reproductive barriers To assess postzygotic barriers between blue- and orange-flowered lineages, we compared fitness of intraversus inter-lineage offspring obtained from the above-mentioned controlled crosses. We considered fitness components throughout the life cycle: fruit-set of P plants, seed-set of P plants, number of seeds per fruit of P plants, germination rate of F1 seeds, fecundity of F1 seedlings, fruit-set of F1 plants, seed-set of F1 plants. We calculated the strength of isolation due to each of these fitness components by using the equation RI4A of Sobel and Chen (2014): RI = 1 − 2 (H/ (H + C)), where H is the fitness of the offspring from inter-lineage crosses and C is that of the offspring from intralineage crosses. This RI index ranges from –1 to 1, with –1 representing complete disassortative mating (gene flow between lineages being mandatory), 0 representing random mating (no reproductive isolation), and 1 representing complete assortative mating (total isolation). As asymmetries in RI can occur between both lineages and genders (Sobel and Chen, 2014; Sobel and Streisfeld, 2015), the RI index for each fitness component was calculated separately for each color lineage when acting as paternal (pollen donor) or maternal (pollen receptor) parent. We used a modified R script provided by Suni and Hopkins (2018) to calculate the 95% confidence intervals around individual RI estimates applying bootstrap resampling implemented in R. Lastly, we used the Excel spreadsheet provided by Sobel and Chen (2014) to calculate cumulative strength of all postzygotic barriers acting sequentially. RI barrier due to seed-set was not considered in calculating cumulative reproductive isolation because it would be redundant with RI barrier due to the number of seeds per fruit. We calculated total reproductive isolation, as Capítulo 1 51 well as the absolute contribution and the relative contribution of each partial postzygotic barrier to the total isolation, separately for each lineage acting either as mother or father. In any case, we did not consider those partial RI indexes with a 95% confidence interval spanning the zero to calculate cumulative reproductive isolation (Karrenberg et al. 2019). Characterization of flower color under pollinator vision To make a first approach to the flower color of L. monelli we obtained the reflectance spectral function of the adaxial surface of 75 blue and 50 orange petals of different plants from different populations. Additionally, we measured reflectance of the color categories recorded among F1 progenies to know if the most frequent pollinators could differentiate the color of the hybrid plants from that of the pure lineages. In all cases, we discarded the basal part of the petals, corresponding to the center of the flower, which is commonly known as the bullseye. We used a JAZ A1465 double-beam spectrophotometer from Ocean Optics, equipped with a UV–visible light source, to measure reflectance between 300 and 700 nm. We then explored how the most common groups of pollinators of L. monelli perceive each flower color by analyzing its reflectance spectral function under their vision models, which enabled us to categorize colors and measure contrasts among them under each pollinator perception (Chittka, 1992; Chittka et al., 1994). We first used procspec function in the R package pavo (Maia et al., 2019) to adjust to zero any negative value in the spectral functions obtained and to smooth each spectral function. To assess petal colors as perceived by the different pollinators, we then plotted the smoothed spectral function of each flower color category in the vision models of bees and diptera. For the bee vision model, we used photoreceptor spectral sensitivities of the honeybee Apis mellifera (Chittka et al., 1992), and for the model of diptera we used those of the hoverfly Eristalis tenax (An et al., 2018). We finally calculated the chromatic contrast against the background (i.e., the Euclidean distance between the color loci of the flower and the achromatic center of the color vision space; Rohde et al., 2013) according to the photoreceptor spectral sensitivities of either bees or hoverflies. For bees, we also calculated the achromatic contrast (i.e., green contrast) as it is known that bees use this contrast to detect distant flowers under small visual angles (Spaethe et al., 2001; Narbona et al., 2021). Since this behavioral information is not available for diptera (Ohashi et al., 2015), we did not calculate achromatic contrast for them. Capítulo 1 52 Results Distribution along the Mediterranean Basin A total of 855 blue and 96 orange points were obtained from different data collections. The blue lineage of L. monelli is distributed by Central and Southwestern Iberian Peninsula, Eastern Algeria and Tunisia while the orange lineage occurs through Morocco, Western Algeria, Northeastern Iberian Peninsula and Sardinia (Figure 2). Populations of the orange lineage that had been described as abundant on the west and east coasts of the Iberian Peninsula were exclusively found in the south of Tarragona. We did not find evidence of polymorphic populations in the Mediterranean Basin. Blue and orange lineages were separated by the sea at a minimum distance of 106.14 km (south Iberian Peninsula and north Morocco populations) and by land at a minimum distance of 212.62 km (center vs east Iberian Peninsula populations). Populations were found in a high diversity of habitats such as coastal areas, disturbed sites, well-preserved forests, and rocky areas (Figure 3). Figure 2. Blue and red lineages of Lysimachia monelli in the Mediterranean Basin. Contributions from herbarium vouchers, GBIF, iNaturalist and in situ observations from field trips (see Material and Methods section). Capítulo 1 53 Figure 3. Diversity of habitats where the blue (A-D) and the orange (E-H) lineages of Lysimachia monelli inhabit. Capítulo 1 54 Flower color and pollination attendance Each flower color presents a characteristic spectra pattern, and no differences were observed between the same flower colors from different populations. Blue petals mainly reflect in the UV, violet and blue spectra, between 330 and 450 nm (Figure 4A), while the orange petals reflect in the yellow, orange, and red spectra, between 600 and 700 nm, with a reflectance peak in the UV at 350 nm (Figure 4A). Figure 4. Reflectance spectra of parental and F1 hybrid flowers of L. monelli. (A) Spectra of blue and orange parental flowers. (B) Spectra of blue, orange, red and light pink F1 hybrid flowers. Dark lines indicate the mean values of each flower color and the lighter shadow represents the standard error. The basal part of the petal, the bullseye, was excluded for the measure. Censuses in natural populations revealed bees and dipterans as the main pollinators of blue and orange L. monelli (Table 2). A total of 40 morphospecies were identified as floral visitors of any of both lineages: 12 bees from four different families (Halictidae, Andrenidae, Apidae and Megachilidae), one non-identified wasp, 17 diptera from at least three known families (Syrphidae (10 of them), Bombiliidae and Muscidae), 9 coleoptera Capítulo 1 55 from at least five families (Mordelidae, Oedemeridae, Dermestidae, Merylidae and Bupestridae) and one lepidoptera from the Micropterigidae family (Table 2; Figure 5). The proportion of visits varied between populations (Figure 6), but short-tongued bees (Halictidae and Andrenidae) and diptera from Syrphidae family, were the pollinators that exhibit a greater number of visits per flower in both flower colors of L. monelli. Figure 5. Blue and orange L. monelli pollinators. Bees: Halictidae (a,b), Megachilidae (c,d); Diptera: Syrphidae (e,f), Bombyliidae (g); and Coleoptera: Buprestidae (h), Oedemeridae (i). Authors: M. Sánchez- Cabrera (a-c, e, g-i) and P. Fernández (d,f). Capítulo 1 56 Table 2. Flower visitors registered during the censuses in natural populations of blue and orange L. monelli. Morphospecies are indicated as the total number observed in each lineage and the number shared between the both lineages. No. morphospecies Pollinator Family Blue Orange Shared Short-tongued bees Halictidae 3 4 3 Andrenidae 2 1 1 Long-tongued bees Apidae (Apis, Bombus, Ceratina, Eucera) 2 3 1 Megachilidae (Anthidium, Rhodantidium) 1 1 - Syrphid Diptera Episyrphus sp. 1 1 1 Eupeodes sp. 1 1 1 Sphaeroforia sp. 1 1 1 Non identified - 7 - Non syrphid Diptera Bombyliidae (Anthrax) 1 1 - Muscidae 1 1 1 No identified 1 3 - Vespids Vespidae - 1 - Coleoptera Mordelidae 1 - - Oedemeridae - 1 - Dermestidae 1 1 - Melyridae (Psilothrix sp.) 1 - - Bupestridae 1 - - No identified - 3 - Lepidoptera Micropterix sp. - 1 - Capítulo 1 57 Figure 6. Visitation rate of pollinators for different populations of blue and orange L. monelli. Each bar plot represents one population and bar colors indicate the flower color. Pollinators were grouped in five groups: short-tongued bees (st.bee), long-tongued bees (lt.bee), syrphid diptera (syrphid), no syrphid diptera (no.syrphid) and other, which includes lepidoptera, coleopteran and wasps. Capítulo 1 64 Table 4. Cumulative strength (Cs) and absolute (Ac) and relative (Rc) contributions of postzygotic barriers between blue- and orange-flowered lineages of Lysimachia monelli calculated after Sobel and Chen (2014). Values are shown for each lineage acting as pollen donor (father) or pollen receptor (mother), as well as the mean between genders for each lineage and the mean between lineages. Bluefather Bluemother Bluemean Orangefather Orangemother Orangemean L. monelli mean Postzygotic barrier Cs Ac Rc Cs Ac Rc Cs Ac Rc Cs Ac Rc Cs Ac Rc Cs Ac Rc Cs Ac Rc Fructification of P 0.129 0.129 0.195 ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns Seed number of P 0.256 0.128 0.194 0.129 0.129 0.349 0.130 0.130 0.219 ns ns ns ns ns ns ns ns ns ns ns ns Germination of F1 0.492 0.236 0.358 ns ns ns ns ns ns ns ns ns 0.274 0.274 0.697 ns ns ns ns ns ns Fecundity of F1 0.660 0.168 0.254 0.368 0.240 0.651 0.368 0.238 0.400 ns ns ns ns ns ns ns ns ns 0.183 0.183 0.481 Fructification of F1 ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns Seed number of F1 0.797 0.138 0.173 0.595 0.227 0.381 0.595 0.227 0.381 0.133 0.133 1.000 0.392 0.119 0.303 0.133 0.133 1.000 0.380 0.197 0.519 “ns” indicates non-significant and thus not considered in calculating cumulative isolation. Capítulo 1 65 Discussion The results of our extensive survey indicate that blue and orange Lysimachia monelli lineages are currently allopatric, which implies a complete geographic isolation between them. The nearest populations of blue and orange lineages are separated by >100 km and given that (1) main flower visitors (bees and hoverflies) have short distance movements and (2) seeds disperse by gravity, gene flow between these populations is expected to be null. In fact, progenies obtained in the glasshouse from intra-lineage crosses of parental plants from natural populations always showed the same parental color; since other flower colors appeared after inter-lineage crosses, this result clearly indicates that plants collected from natural populations were pure lineages and corroborates no recent contact between lineages. When lineages were intercrossed in the glasshouse, more than 80% of the F1 hybrid progeny was orange, however up to three other color phenotypes were segregated. Quintana et al. (2008) performed crosses between blue and orange L. monelli lineages, obtaining a 100% orange F1 progeny that was segregated to four color phenotypes in the F2. As they have a reduced number of F1 progeny (seven plants), it is not surprising that those individuals had orange flowers. Quintana et al. (2008) suggested a four-gene model for the inheritance of flower color in L. monelli according to their F2 color segregation. However, our results suggest that this model needs to be reconsidered given that the four colors segregated in the F1. Despite differences in distribution and flower color, in nature both lineages were visited by the same groups of pollinators. Short-tongued bees and hoverflies were the main visitors of blue and orange L. monelli in all the habitat types where they occur, and in some cases, the same morphospecies visited the two lineages. Chromatic contrast values showed that in the bee-vision model, blue flowers have a higher contrast with the background than orange flowers. Given that this contrast is critical for detectability, this difference suggests that bees could detect blue flowers more quickly than orange ones. This situation had been found in the two lineages of L. arvensis where bees showed a clear preference for the blue lineage (Ortiz et al., 2015). Interestingly, the chromatic contrast values obtained in the hoverfly-vision model showed just the opposite, as they were higher for the orange lineage. Thus, although both lineages share pollinators, differences in detectability between bees and hoverflies open the door to a segregation of Capítulo 1 66 pollinator groups in case of secondary contact between lineages. Moreover, chromatic contrast values found between lineages in both vision systems were high enough to allow both pollinator groups to discriminate both colors, enabling them to develop flower constancy and thus to produce assortative mating, as reported for co-occurring lineages of L. arvensis (Jiménez-López et al., 2020a). It is tempting to forecast that in case of secondary contact between L. monelli lineages, pollinators would Loci of hybrid red flowers overlapped with those of orange flowers, partially in the bee visual space and almost completely in that of flies, which would suggest some difficulties for both pollinator groups in differentiating those colors. Nevertheless, it has been reported that bees are able to discriminate among very close color loci (García et al., 2017), and in fact, in L. arvensis, bees clearly preferred orange flowers over hybrid salmon flowers even although their loci partially overlapped (Jiménez-López et al., 2020a). Light pink flowers showed very low chromatic contrasts in both visual systems, which suggests these flowers are hardly perceived by both bees and flies. However, interestingly, this color produced a very high achromatic contrast for bees, nearly the highest value found among 98 species in a community study (Ortiz et al., 2021); this fact is remarkable because bees use such achromatic contrast to detect flowers at long distance (Giurfa et al., 1997; Spaethe et al., 2001). Geographical isolation is the main reproductive barrier between color lineages of L. monelli, and our results clearly indicate that, at least currently, reproductive isolation between these lineages is complete because they are allopatric. Geographical distribution of species depends on genetic, evolutionary, historical and/or ecological traits (Whittaker, 1972; Comes and Kadereit, 1998; Elith and Leathwick, 2009; Sexton et al., 2009). The two lineages seem to occupy similar habitats, so we cannot know if the isolation barrier is due to a difference in the ecological niche, which would imply an eco-geographical barrier. Under a possible secondary contact scenario, L. monelli lineages would present reproductive isolation, being fecundity and number of seeds of the F1 the main barriers contributing to the total isolation. However, the studied postzygotic barriers are slightly asymmetric, indicating higher likelihood for the occurrence of hybrid seedlings originating from L. monelli blue mothers. High reproductive isolation is common between closely related species expressing different flower colors. Mimulus cardinalis (red flowers) and M. lewisii (pink flowers) have an altitudinal distribution, coexisting at midelevation sites. Although they produce fertile F1 hybrids (Hiesey et al., 1971; Ramsey et Capítulo 1 67 al., 2003), they present an almost total reproductive isolation, based on cumulative pre- and postzygotic barriers (Ramsey et al., 2003). Blue and orange lineages of Lysimachia arvensis also present a high reproductive isolation, but also producing fertile F1 hybrids (Jiménez-López et al., in prep.). Lysimachia monelli produced fertile F1 hybrids, but only under experimental conditions. Total values of reproductive isolation found in this study are similar from those found in other studies of closely related species (Brys et al., 2014; Cahenzli et al., 2018; Jiménez- López et al., in prep.). The impact of multiple pre- and postzygotic barriers between blue and orange Lysimachia arvensis suggests that speciation is likely to be complete (Jiménez-López et al., in prep.). Although L. monelli showed a complete geographic isolation, meaning null gene flow between both lineages, analyses of barriers simulating a secondary contact are needed to test their speciation status (Sobel et al., 2010). In fact, L. monelli shows an elevated value of total isolation only considering postzygotic barriers (Table 3). Because reproductive isolating barriers act sequentially, early-acting barriers will contribute more to total isolation than late-acting barriers with the same individual strength (Ramsey et al. 2003; Nosil et al. 2005). Thus, as we only considered postzygotic barriers for a possible secondary contact scenario, our calculations probably underestimate the magnitude by which prezygotic barriers outweigh postzygotic barriers in reducing contemporary interspecific gene flow, meaning that both lineages of L. monelli present an even more isolated. Conclusions In this study we have shown that two geographically isolated flower color lineages of L. monelli present postzygotic barriers contributing to their reproductive isolation that, in case of a secondary contact, would reduce their hybridization, promoting the maintenance of the flower color lineages. The same group of pollinators visit both lineages. However, the main groups show differences in the perception of colors, suggesting a possible segregation of pollinator groups in case of secondary contact between lineages. Finally, four flower colors were obtained in the first generation of hybrids, being orange the most common. Crossed transplant experiments in the field are required to understand the pollinator behavior between flower colors and the development of hybrids in nature. Capítulo 1 68 Acknowledgements This work was financed by FEDER funds and MINECO grants (CGL2015-63827) to M.A. and P.L.O. and M.S.-C. (BES-C-2016-0023), and by MICINN (PID2020- 116222GB-10) to M.A. We want to specially thank Mohamed Abdelaziz for his valuable help in the organization and the survey during the field trip to Morocco to find L. monelli populations. We also thank the General Herbarium and Greenhouse Services of the Universidad de Sevilla (CITIUS). References An, L., Neimann, A., Eberling, E., Algora, H., Brings, S., and Lunau, K. (2018). The yellow specialist: Dronefly Eristalis tenax prefers different yellow colours for landing and proboscis extension. Journal of Experimental Biology, 221(22), jeb184788. Arista, M., Talavera, M., Berjano, R., and Ortiz, P. L. (2013). Abiotic factors may explain the geographical distribution of flower colour morphs and the maintenance of colour polymorphism in the scarlet pimpernel. J. Ecol. 101(6), 1613-1622. Arnold, S. E., Faruq, S., Savolainen, V., McOwan, P. W., and Chittka, L. (2010). FReD: the floral reflectance database—a web portal for analyses of flower colour. PloS one 5(12), e14287. Baack, E., Melo, M. C., Rieseberg, L. H., and Ortiz‐Barrientos, D. (2015). The origins of reproductive isolation in plants. New Phytol. 207(4), 968-984. Bates, D., Mächler, M., Bolker, B., Walker, S., Christensen, R. H., and Singmann, H. (2015). lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1-7. 2014. Bradshaw, H. D., and Schemske, D. W. (2003). Allele substitution at a flower colour locus produces a pollinator shift in monkeyflowers. Nature 426(6963), 176-178. Brys, R., Broeck, A. V., Mergeay, J., and Jacquemyn, H. (2014). The contribution of mating system variation to reproductive isolation in two closely related Centaurium species (Gentianaceae) with a generalized flower morphology. Evolution 68(5), 1281- 1293. Cahenzli, F., Bonetti, C., and Erhardt, A. (2018). Divergent strategies in pre‐and postzygotic reproductive isolation between two closely related Dianthus species. Evolution 72(9), 1851-1862. Chittka, L. (1992). The colour hexagon: a chromaticity diagram based on photoreceptor excitations as a generalized representation of colour opponency. J. Comp. Physiol. A 170(5), 533-543. Capítulo 1 69 Chittka, L. (1997) Bee color vision is optimal for coding flower color, but flower colors are not optimal for being coded - why? Isr. J. Plant Sci. 45, 115-127. Chittka, L., Shmida, A. V. I., Troje, N., and Menzel, R. (1994). Ultraviolet as a component of flower reflections, and the colour perception of Hymenoptera. Vision Res. 34(11), 1489-1508. Coberly, L. C., and Rausher, M. D. (2003). Analysis of a chalcone synthase mutant in Ipomoea purpurea reveals a novel function for flavonoids: amelioration of heat stress. Mol. Ecol. 12(5), 1113-1124. Comes, H. P., and Kadereit, J. W. (1998). The effect of Quaternary climatic changes on plant distribution and evolution. Trends Plant Sci. 3(11), 432-438. Coyne, J. A. (2007). Sympatric speciation. Curr. Biol. 17(18), R787-R788. Coyne, J. A., and Orr, H. A. (2004). Speciation (Vol. 37). Sunderland, MA: Sinauer Associates. Christie, K., and Strauss, S. Y. (2019). Reproductive isolation and the maintenance of species boundaries in two serpentine endemic Jewelflowers. Evolution 73(7), 1375-1391. Davies, K. M., Albert, N. W., and Schwinn, K. E. (2012). From landing lights to mimicry: the molecular regulation of flower colouration and mechanisms for pigmentation patterning. Funct. Plant Biol. 39(8), 619. Dieckmann, U., and Doebeli, M. (1999). On the origin of species by sympatric speciation. Nature 400(6742), 354-357. Dobzhansky, T. (1937). Genetics and the origin of species. Columbia Univ. Press, New York. Elith, J., and Leathwick, J. R. (2009). Species distribution models: ecological explanation and prediction across space and time. Annu. Rev. Ecol. Evol. Syst. 40, 677-697. Ferguson, L. F. (1972). Anagallis. In Tutin et al. (eds.) Flora Europaea. Vol. 3. Cambridge University Press, Cambridge, UK. Freyre, R., and Griesbach, R. J. (2004). Inheritance of flower color in Anagallis monelli L. HortScience 39(6), 1220-1223. Garcia, J.E., Spaethe, J., and Dyer, A.G. 2017. The path to colour discrimination is S- shaped: behaviour determines the interpretation of colour models. J Comp Physiol A. 203(12), 983-997. GBIF.org (2019), GBIF Home Page. Available from: https://www.gbif.org Gibbs, P. E., and Talavera, S. (2001). Breeding system studies with three species of Anagallis (Primulaceae): self-incompatibility and reduced female fertility in A. monelli L. Ann. Bot. 88(1), 139-144. Capítulo 1 70 Giurfa, M., Vorobyev, M., Brandt, R., Posner, B., and Menzel, R. (1997). Discrimination of coloured stimuli by honeybees: Alternative use of achromatic and chromatic signals. J. Comp. Physiol. A 180, 235-243. Grotewold, E. (2006). The Science of Flavonoids. New York, NY: Springer. Hiesey, W. M. (1971). Experimental studies in the nature of species (No. 575.1/C616). iNaturalist. (2022). Available from https://www.inaturalist.org Accessed via https://www.inaturalist.org/taxa/707889-Lysimachia-monelli [03 February 2022] Jiménez-López, F. J. (2019). Polimorfismo en el Color Floral y Especiación en Lysimachia arvensis. Ph.D. dissertation, University of Seville, Seville. Jiménez-López, F. J., Arista, M., Talavera, M., Morellato, L. P. C., Pannell, J.R., Viruel, J., et al. (in prep.). Preversus post-zygotic reproductive isolation between two neospecies of Lysimachia that differ in flower colour. Jiménez-López, F. J., Matas, L., Arista, M., and Ortiz, P. L. (2020a). Flower colour segregation and flower discrimination under the bee vision model in the polymorphic Lysimachia arvensis. Plant Biosyst. 154(4), 535-543. Jiménez-López, F. J., Ortiz, P. L., Talavera, M., Pannell, J. R., and Arista, M. (2020b). The role of lateral and vertical herkogamy in the divergence of the blue-and red-flowered lineages of Lysimachia arvensis. Ann. Bot. 125(7), 1127-1135. Jiménez-López F. J., Viruel J., Arista M., Ortiz P. L., and Talavera M. (2022). Molecular approaches reveal speciation between red and blue flowered plants in the Mediterranean Lysimachia arvensis and Lysimachia monelli (Primulaceae). Bot. J. Linn. Karrenberg, S., Liu, X., Hallander, E., Favre, A., Herforth‐Rahmé, J., and Widmer, A. (2019). Ecological divergence plays an important role in strong but complex reproductive isolation in campions (Silene). Evolution 73(2), 245-261. Kulmuni, J., Butlin, R. K., Lucek, K., Savolainen, V., and Westram, A. M. (2020). Towards the completion of speciation: the evolution of reproductive isolation beyond the first barriers. Philos. Trans. R. Soc. B 375(1806), 20190528. Lowry, D. B., Rockwood, R. C., and Willis, J. H. (2008). Ecological reproductive isolation of coast and inland races of Mimulus guttatus. Evolution: International Journal of Organic Evolution, 62(9), 2196-2214. Maia, R., Gruson, H., Endler, J. A., and White, T. E. (2019). pavo 2: new tools for the spectral and spatial analysis of colour in R. Methods Ecol. Evol. 10(7), 1097-1107. Narbona, E., Del Valle, J. C., Arista, M., Buide, M. L., and Ortiz, P. L. (2021). Major Flower Pigments Originate Different Colour Signals to Pollinators. Front. Ecol. Evol. 9. Nosil, P., Harmon, L. J., and Seehausen, O. (2009). Ecological explanations for (incomplete) speciation. Trends Ecol. Evol. 24(3), 145-156. Capítulo 1 71 Nosil, P., Vines, T. H., and Funk, D. J. (2005). Reproductive isolation caused by natural selection against immigrants from divergent habitats. Evolution 59(4), 705-719. Orr, H. A., and Turelli, M. (2001). The evolution of postzygotic isolation: accumulating Dobzhansky‐Muller incompatibilities. Evolution 55(6), 1085-1094. Ortiz, P. L., Berjano, R., Talavera, M., Rodríguez-Zayas, L., and Arista, M. (2015). Flower colour polymorphism in Lysimachia arvensis: How is the red morph maintained in Mediterranean environments?. Perspect. Plant Ecol. Evol. Syst. 17(2), 142-150. Ortiz, P. L., Fernández‐Díaz, P., Pareja, D., Escudero, M., and Arista, M. (2021). Do visual traits honestly signal floral rewards at community level?. Funct. Ecol. 35(2), 369- 383. Ortiz-Barrientos, D. (2013). The color genes of speciation in plants. Genetics 194(1), 39- 42. Ohashi, K., Makino, T. T., and Arikawa, K. (2015). Floral colour change in the eyes of pollinators: Testing possible constraints and correlated evolution. Funct. Ecol. 29, 1144- 1155. Pujadas A. (1997). Anagallis L. In: Castroviejo S, Aedo C, Laínz M, Morales R, Muñoz Garmendia F, Nieto Feliner G, Paiva J. eds. Flora iberica, V. Ebenaceae-Saxifragaceae. Madrid: Real Jardín Botánico, CSIC, 57–62. QGIS.org. (2019). QGIS Geographic Information System. QGIS Association. URL http://www.qgis.org Quintana, A., Freyre, R., Davis, T. M., and Griesbach, R. J. (2008). Genetic studies of flower color in Anagallis monelli L. HortScience 43(6), 1680-1685. R Core Team. (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Ramsey, J., Bradshaw, H. D., and Schemske, D. W. (2003). Components of reproductive isolation between the monkeyflowers Mimulus lewisii and M. cardinalis (Phrymaceae). Evolution 57, 1520-1534. Rausher, M. D. (2008). Evolutionary transitions in floral color. Int. J. Plant Sci. 169(1), 7-21. Rohde, K., Papiorek, S., and Lunau, K. (2013). Bumblebees (Bombus terrestris) and honeybees (Apis mellifera) prefer similar colours of higher spectral purity over trained colours. J. Comp. Physiol. A 199(3), 197-210. Seehausen, O., Butlin, R. K., Keller, I., Wagner, C. E., Boughman, J. W., Hohenlohe, P. A., et al. (2014). Genomics and the origin of species. Nat. Rev. Genet. 15(3), 176-192. Sexton, J. P., McIntyre, P. J., Angert, A. L., and Rice, K. J. (2009). Evolution and ecology of species range limits. Annu. Rev. Ecol. Evol. Syst. 40, 415-436. Capítulo 1 72 Sobel, J. M., Chen, G. F., Watt, L. R., and Schemske, D. W. (2010). The biology of speciation. Evol.; int. j. org. evol. 64(2), 295-315. Sobel, J. M., and Chen, G. F. (2014). Unification of methods for estimating the strength of reproductive isolation. Evolution 68, 1511-1522. Sobel, J. M., and Streisfeld, M. A. (2015). Strong premating reproductive isolation drives incipient speciation in Mimulus aurantiacus. Evolution, 69(2), 447-461. Spaethe, J., Tautz, J., and Chittka, L. (2001). Visual constraints in foraging bumblebees: flower size and color affect search time and flight behavior. Proc. Natl. Acad. Sci. U.S.A. 98(7), 3898-3903. Strauss, S., and Whittall, J.B. (2006) "Non-pollinator agents of selection on floral traits" in Ecology and Evolution of Flowers. Eds. L. D. Harder, and S. C. H. Barrett (Oxford, UK: Oxford University Press), 120-138. Suni, S. S., and Hopkins, R. (2018). The relationship between postmating reproductive isolation and reinforcement in Phlox. Evolution 72(7), 1387-1398. Turelli, M., Barton, N. H., and Coyne, J. A. (2001). Theory and speciation. Trends Ecol. Evol. 16(7), 330-343. Van der Niet, T., Peakall, R., and Johnson, S. D. (2014). Pollinator-driven ecological speciation in plants: new evidence and future perspectives. Ann. Bot. 113(2), 199-212. Via, S. (2009). Natural selection in action during speciation. Proc. Natl. Acad. Sci. U.S.A. 106(Supplement 1), 9939-9946. Warren, J., and Mackenzie, S. (2001). Why are all colour combinations not equally represented as flower-colour polymorphisms? New Phytol. 151, 237–241. Whittaker, R. H. (1972). Evolution and measurement of species diversity. Taxon 21(2-3), 213-251. Wilson, P., Castellanos, M. C., Hogue, J. N., Thomson, J. D., and Armbruster, W. S. (2004). A multivariate search for pollination syndromes among penstemons. Oikos 104(2), 345-361. Wilson, P., Wolfe, A. D., Armbruster, W. S., and Thomson, J. D. (2007). Constrained lability in floral evolution: counting convergent origins of hummingbird pollination in Penstemon and Keckiella. New Phytol. 176(4), 883-890. 73 CAPÍTULO 2 Understanding biogeographical patterns and genetic diversity of Lysimachia monelli (Primulaceae), a flower color polymorphic species endemic to the Mediterranean Basin Mercedes Sánchez-Cabrera, Francisco Javier Jiménez-López, Pedro L. Ortiz, María Talavera and Montserrat Arista (Manuscript in preparation) Capítulo 2 80 AFLP and nSSR genotyping Two different nuclear markers were applied to define the current spatial genetic structure of L. monelli: Amplified Fragment Length Polymorphisms (AFLPs), a dominant marker, and nuclear microsatellites (nSSR), a highly polymorphic co-dominant marker. Considering Vos et al. (1995) protocols for AFLP, an initial screening of selective primers was performed on eight individuals (four blue and four orange) from four populations of L. monelli, using 36 primer combinations with three selective nucleotides. The six primer combinations with higher variability were selected for the selective PCR (EcoRI (Ned)- AAC/MseI-CTG, EcoRI (Vic)-ACG/MseI-CTG, EcoRI (Fam)-ACT/MseI CAG, EcoRI (Ned)-AGC/MseI CTC, EcoRI (Vic)-ACG/MseI-CAT, and EcoRI (Fam)-ACC/MseI- CTC). We analysed 324 leave samples of L. monelli and, in order to exclude nonreproducible bands and to calculate the error rate according to Bonin et al. (2004), a 10% of the samples were replicated. Regarding nSSR primers, seven primers (Lys12, Lys24, Lys28, Lys29, Lys30, Lys31) were selected and applied to all the samples of L. monelli. These primers were specifically developed for Lysimachia arvensis (Jiménez-López et al., 2015), a close related species. Following the procedure applied in Jimenez-López et al. (2020b), the fluorescence ranges emitted by each sample were analysed. The fluorescence labelled amplification products, both AFLP and nSSR, were separated by capillary gel electrophoresis on an automated sequencer (3730 DNA Analyser, PE Applied Biosystems, Foster City, CA, USA) with an internal size standard (GeneScan 500 LIZ, Applied Biosystems) at STABVIDA Lda. (Oeiras, Portugal). Raw data were exported to GeneMarker 1.9 (SoftGenetics, State College, PA, USA) for manual scoring of the fragments, after normalisation of the profiles. Peaks were validated when their fluorescence intensity exceeded the fluorescence threshold (set at 100 relative fluorescent units). Amplified fragments from 100 to 500 base pairs (bp) were scored. The scoring results from AFLP data were exported as a presence/absence matrix and imported into Famd v 1.3 (Schlüter and Harris, 2006). The nSSR screening were exported to a matrix with the fragment size of each locus. Moreover, a simulated analysis of multilocus genotype accumulation curves were conducted with the R package “poppr” (Kamvar et al., 2014), to evaluate if seven loci were an adequate sample size, using 1,000 randomly permutations to create the distribution. Genetic variability Capítulo 2 81 Different standard parameters of genetic diversity were estimated for each marker. For AFLP data, within-population genetic diversity was assessed by the total number of AFLP fragments present (Ftot), the percentage of polymorphic fragments (Fpoly), the private fragments (Fpriv) and the average gene diversity (HD); genetic distances between individuals and populations were estimated from the standard similarity by the pairwise FST (Jacard). For nSSR data, we estimated the observed and expected heterozygosity (Ho and He), the inbreeding coefficient (Fis) and the null allele frequency (No) with GENEPOP v4.0.10 (Raymond and Rousset, 1995; Rousset, 2008); two neutrality tests, Tajima’s D-test (T’s D) and Fu’s Fs-test (Fu’S), were conducted with ARLEQUIN v. 3.01 (Excoffier et al., 2005). Isolation-by-distance between populations of each lineage was assessed by computing the correlation between the matrix of pair-wise population genetic distance obtained for both markers and the matrix of geographical distances, and then applying a Mantel test (10,000 permutations) for nSSR data with GENODIVE v3.0 (Meirmans and Van Tienderen, 2004). Population structure To determine the number of independent genetic and evolutionary entities, the overall population genetic structure was inferred by the Bayesian clustering method with STRUCTURE 2.3.4 (Pritchard et al., 2000; Pritchard, 2010), which allow to identify the optimal number of clusters (K) in AFLP and nSSR datasets. Clustering of individuals was conducted without using the geographic origin of the samples as a prior information. Analyses were based on an admixture ancestry model with correlated allele frequencies, for a range of K genetic clusters from one to 28. The analyses were performed with a burn-in period of 100,000 and a run length of the Monte Carlo Markov Chain (MCMC) of 1,000,000 iterations, respectively. The most likely number of genetic clusters (K) was determined according to the ∆K and Mean LnP(K) (Evanno et al., 2005; Pritchard, 2010), using STRUCTURE Harvester (Earl and vonHoldt, 2012). Additionally, the K-mean clustering function implemented in GENODIVE were also used with nSSR dataset. Analyses were based on Euclidean distances calculated from within-populations allele frequencies with simulated annealing using 50,000 steps and 20 random starts for a range of K genetic clusters from 2 to 25. The best clustering was defined according to Calinski and Harabasz' pseudo-F and Bayesian Information Criterion (BIC). Capítulo 2 82 We developed patterns and levels of population genetic differentiation for four different groups using analyses of molecular variance (AMOVA) in AFLP data, conducted with ARLEQUIN: (1) population level, (2) flower color level, and (3) first optimum from Bayesian clustering analysis in AFLP dataset (K = 3; see Results). We also calculated the AMOVA-derived fixation index FST with ARLEQUIN, which is an indirect approach to estimate gene flow (Wright, 1951), describing the heterozygosity reduction within populations relative to the total population. Furthermore, permutation tests were used to determine statistical significance for nSSR dataset. Statistics for the significance (OSxstatistic; Goudet, 1995) across all groups were obtained by 9,999 randomizations using GENODIVE. The possible influence of population size on the genetic diversity of lineages, measured as observed heterozygosity, was tested by an ANOVA with population size and flower color as factors. In addition, we used principal coordinate analysis (PCoA) among individuals of the different populations and lineages to assess relationships between them, plotted in IBM SPSS Statistic 24 (Armonk, NY: IBM Corp.). We used BOTTLENECK 1.2.02 (Piry et al., 1999) to detect recent bottlenecks in the studied populations. Wilcoxon’s tests with three mutational models of microsatellite variation were performed: the infinite alleles model (IAM), the stepwise mutation model (SMM) and the two-phase mutational model (TPM). Environmental Niche Modelling Presence location data of L. monelli were obtained from sampling campaings, verified herbarium vouchers from University of Seville and from the Global Biodiversity Information Facility data portal (http://www.gbif.org/). A total of 292 presence points (234 from blue-flowered and 58 from orange-flowered plants) were used, covering the reported geographic distribution of the species throughout the Mediterranean Basin. To evaluate the current potential distribution of both flower color lineages of L. monelli in the Mediterranean basin, we selected variables relevant to the ecological conditions which likely determine the niche of terrestrial plant species. We used four datasets from two different sources: 19 bioclimatic variables, 12 solar radiation variables and the elevation data were extracted from WorldClim (Fick and Hijmans, 2017) at a resolution of 5 km2 spatial resolution; 12 edaphic variables were extracted from SoilGrids (Poggio et al., 2021) at a resolution of 250 m2 spatial resolution. All the layers were rescaled to 5 Capítulo 2 83 km2 with Resample function in R v.4.1.0 software (R Core Team 2020). Layers were cut to the Mediterranean region extent and data from L. monelli presence coordinates were extracted with QGIS 3.10 (QGIS.org, 2021). We used the variance inflation factor with a threshold of 10, indicating collinearity, to remove the highly correlated variables using R software (López-Jurado et al., 2018). Finally, a dataset of 19 variables was obtained to construct the models: six bioclimatic, two solar and 11 edaphic (Supplemental Table 1). Furthermore, we explored past projections, using available paleoclimatic information. The 19 bioclimatic variables from three different time periods were extracted from WorldClim (Community Climate System Model, CCSM4): Mid Holocene (MH; ~6,000 years ago), Last Glacial Maximum (LGM; ~20,000 years ago) and Last Interglacial (LIG; ~125,000 years ago). For past projections, we used all the bioclimatic variables, which model presents the highest area under the curve (AUC) value (> 0.9). Maximum entropy algorithm was used to evaluate the potential geographic distribution of blue and orange L. monelli using MAXENT v.3.4.4 software (Phillips and Dudík, 2008). Environmental niche modeling (ENM) were performed under present and three past scenarios (MH, LGM and LIG) for each flower color lineage (blue and orange). For model accuracy, we employed the cross-validation method with 15 replicates, 10,000 background points and a maximum number of iterations of 5,000. We included AUC to evaluate the model and the permutation importance of each variable in the final model. Results Genetic variability We scored 503 AFLP fragments ranging from 100 to 500 bp. The genetic diversity (HD) fluctuated from 0.110 to 0.161 (0.112-0.161 in blue-flowered plant populations and 0.110-0.137 in orange-flowered plant populations). Mean HD value was 0.134 ± 0.07 (0.135 ± 0.071 in blue and 0.128 ± 0.067 in orange; Table 2). There was a trend towards lower HD in orange populations, but not statistically significant (F1,25=2.211, p=0.150). Likewise, lower Fpoly and Fpriv mean values were observed in orange populations (191.67 and 8.83 respectively) than in blue ones (208.38 and 21.8 respectively) (Fpoly: F1,25=3.608, p=0.069, mean value range from 163 to 251; Fpriv: F1,25=5.266, p=0.030, mean value ranged from 3 to 48; Table 2). Capítulo 2 84 The seven nSSR analysed in all the samples showed a total of 108 alleles (10-25 per locus), with a mean number of alleles per locus and population ranging from 2.57 to 5.14. The genotype accumulation curves analysis indicated that five-six loci describe the genetic diversity of the studied populations adequately (Supplemental Figure 1). Ho and He ranged from 0.512-0.821 (mean 0.641) and 0.457-0.681 (mean 0.574) respectively (Table 2). Fis ranged from -0.477 to 0.197 (mean -0.117). Blue-flowered plants showed similar Ho and He mean values and slightly higher Fis than the orange-flowered plants (Ho: 0.650 blue vs. 0.610 orange; He: 0.580 blue vs. 0.552 orange; Fis: -0.127 blue vs - 0.112 orange). Twelve studied populations showed deviation from Hardy-Weinberg Equilibrium (HWE; P < 0.05) after Bonferroni correction; the most significant cases were related to high levels of Fis, indicating heterozygote excess (negative Fis) or deficiency (positive Fis). Low values of No were estimated (0.000 to 0.122; mean 0.039; Table 2). Hence, no evidence of linkage disequilibrium has been observed for each locus pair across all populations. Finally, comparing the genetic parameters between flower color lineages and geographic groups (see Bayesian analysis below), no significant differences were observed in Ho, He or Fis. However, Ho (OSx=0.291, p=0.024) and Fis (OSx=0.556 p=0.011), showed significant differences across geographic clusters (K=5; Supplemental Table 2): Portugal populations (LM1-5) presented the highest Ho mean (0.733) and lowest Fis mean (-0.299) and South Spain populations (LM6-10 and LM14) showed lowest Ho mean (0.560) and highest Fis mean (0.020). Capítulo 2 85 Table 2.- Genetic diversity parameters estimated for AFLP and nSSR datasets of studied L. monelli populations. N, number of individuals analysed. Parameters estimated with AFLP dataset: Ftot, total number of fragments per populations; Fmean, mean number fragment per individual in each population; Fpoly, polymorphic allele number; Fpriv, private fragments; HD, average gene diversity. Parameters estimated with nSSR: A, average number of alleles; Ho, observed heterozygosity; He, expected heterozygosity; Fis, inbreeding coefficient (+ p<0.05; ++ p<0.01 of Hardy-Weinberg equilibrium deviation); No, null allele frequency mean of 7 nSSR; T’s D, Tajima’s D-test; Fu’S, Fu’s Fs-test; and SSD, sum of square deviations (ns: not significant, ** p < 0.01, * p < 0.05). Pop N AFLP SSR Ftot Fmean Fpoly Fpriv HD A Ho He Fis No T's D Fu'S SSD LM1 12 215 66.250 ± 12.226 194 21 0.123 ± 0.064 3.430 0.571 0.544 -0.051 0.032 0.797ns -3.756* 0.029ns LM2 12 240 73.083 ± 9.765 214 26 0.132 ± 0.069 4.140 0.821 0.610 -0.346++ 0.033 0.468ns -4.152** 0.025ns LM3 12 226 75.083 ± 5.632 201 25 0.135 ± 0.071 3.430 0.762 0.516 -0.477++ 0.000 0.731ns -5.085** 0.017ns LM4 12 226 72.750 ± 5.675 203 23 0.134 ± 0.071 3.710 0.786 0.596 -0.318++ 0.011 -0.012ns -3.643ns 0.015ns LM5 12 264 76.333 ± 4.397 244 20 0.154 ± 0.080 3.570 0.726 0.556 -0.307++ 0.002 -0.425ns -4.547* 0.016ns LM6 12 234 80.167 ± 7.322 204 30 0.126 ± 0.066 5.140 0.619 0.655 0.055 0.048 -0.344ns -3.672* 0.032* LM7 12 238 101.917 ± 5.680 190 48 0.133 ± 0.069 3.570 0.536 0.582 0.080 0.084 -0.048ns -3.973** 0.010ns LM8 12 261 96.333 ± 7.152 218 43 0.137 ± 0.072 4.290 0.655 0.615 -0.064 0.044 0.100ns -3.712* 0.014ns LM9 12 253 93.583 ± 7.317 215 38 0.144 ± 0.075 4.570 0.524 0.652 0.197++ 0.122 -0.371ns -3.692ns 0.014ns LM10 12 225 95.083 ± 5.178 182 43 0.130 ± 0.068 3.570 0.512 0.457 -0.120 0.017 0.417ns -4.398** 0.013ns LM11 12 258 67.917 ± 6.841 251 7 0.161 ± 0.084 4.000 0.667 0.597 -0.116 0.049 0.092ns -4.131* 0.009ns LM12 12 225 65.750 ± 6.851 214 11 0.139 ± 0.072 3.140 0.667 0.497 -0.341++ 0.000 0.035ns -3.996** 0.016ns LM13 12 216 63.833 ± 4.239 200 16 0.129 ± 0.067 5.000 0.738 0.678 -0.089 0.021 -0.149ns -3.470** 0.008ns LM14 12 233 80.417 ± 8.382 202 31 0.133 ± 0.070 2.570 0.512 0.462 -0.108 0.069 -0.300ns -4.083* 0.02ns LM15 12 251 74.833 ± 6.043 241 10 0.158 ± 0.082 5.000 0.679 0.681 0.004 0.102 -0.105ns -3.263* 0.014ns LM16 12 217 62.667 ± 8.711 204 13 0.135 ± 0.070 3.430 0.595 0.570 -0.044 0.054 -0.372ns -4.059** 0.013ns LM17 12 213 55.667 ± 17.619 206 7 0.129 ± 0.068 3.860 0.679 0.622 -0.091 0.052 -0.013ns -3.716** 0.017ns LM18 12 227 58.917 ± 7.879 215 12 0.130 ± 0.068 4.000 0.667 0.605 -0.102 0.014 0.210ns -4.552* 0.007ns LM19 12 215 67.000 ± 6.836 199 16 0.135 ± 0.071 3.860 0.643 0.626 -0.027 0.037 0.591ns -4.276** 0.068* Capítulo 2 86 LM20 11 205 62.800 ± 8.703 193 12 0.137 ± 0.073 2.860 0.584 0.497 -0.175+ 0.028 -0.233ns -3.527** 0.023ns LM21 12 179 63.750 ± 6.628 163 16 0.123 ± 0.068 3.000 0.512 0.482 -0.062 0.056 0.320ns -4.469* 0.012ns LM22 12 202 60.333 ± 8.606 195 7 0.131 ± 0.068 3.860 0.619 0.548 -0.129+ 0.012 0.206ns -3.876** 0.009ns LM23 12 219 59.500 ± 6.908 212 7 0.134 ± 0.070 3.710 0.583 0.667 0.126 0.110 0.275ns -3.749* 0.019ns LM24 12 202 61.833 ± 5.686 194 8 0.133 ± 0.070 4.140 0.738 0.608 -0.214++ 0.027 -0.019ns -4.512* 0.011ns LM25 13 196 48.923 ± 6.897 193 3 0.111 ± 0.058 3.710 0.626 0.513 -0.221++ 0.002 0.473ns -5.732** 0.019ns LM26 12 172 53.000 ± 4.023 166 6 0.112 ± 0.059 4.140 0.571 0.491 -0.164+ 0.012 0.438ns -5.160* 0.013ns LM27 12 219 61.083 ± 5.664 213 6 0.141 ± 0.074 3.710 0.714 0.569 -0.256++ 0.028 -0.140ns -7.140** 0.019ns Blue mean 229 73.413 208 22 0.135 ± 0.071 3.911 0.650 0.580 -0.127 0.040 0.076 -4.213 0.019 Orange mean 200 59.523 192 9 0.128 ± 0.067 3.547 0.610 0.552 -0.112 0.039 0.170 -4.311 0.016 Total mean 223 70.326 205 19 0.134 ± 0.070 3.830 0.641 0.574 -0.117 0.039 0.097 -4.235 0.018 Capítulo 2 87 Pairwise genetic distance between markers was highly correlated (R2 = 0.588, p<0.05; Supplemental Figure 2), although it did not show a clear pattern with any marker between populations. Nevertheless, orange-flowered populations presented less distance between them than with blue-flowered populations with AFLP data (Supplemental Figure 3). Portugal populations (LM1-LM5) were more similar between them than the remainder populations, South Spain populations (LM6-LM10 and LM14) presented a high dissimilarity (>0.4) with all the populations. With some exception, nSSR data showed higher distance between Portugal populations and the other blue-flowered populations (Pairwise PhiST> 0.3); also, orange populations (excluding LM25) showed lower genetic distance (PhiST<0.3) between them, compared to other populations (Supplemental Figure 2). Only five populations (LM6, LM10, LM11, LM12 and LM20) showed evidence of genetic impoverishment by a recent bottleneck for some of the applied models (Supplemental Table 3). Population structure Geographic vs. genetic distances Mantel tests showed association between genetic and geographical distances in all pairs of L. monelli populations with AFLP (r=0.313, p=0.010) and nSSR data (r=0.435, p= 0.000). The same pattern was observed for each color lineage: between pairs of blueflowered populations (AFLP: r= 0.345, p=0.010; nSSR: r=0.480, p= 0.000) and between pairs of orange-flowered populations (AFLP: r=0.747, p=0.010; nSSR: r=0.592, p= 0.023). Hence, the isolation-by-distance hypothesis was accepted for both molecular markers and color lineages. Pairs of populations within geographic groupsalso showed an association between genetic and geographic distances, except for the south-western IP, Tunisia, and Sardinia populations from nSSR data (Supplemental Table 4). In addition, no significant differences were observed in genetic diversity among different sized populations, either considering flower color (blue-flowered plants F2,18=0.724, p=0.498; orange-flowered plants F2,3=0.491, p=0.654) or not (F2,24=0.851, p=0.440). Capítulo 2 88 Figure 1.- Geographical distribution and Bayesian analyses of the genetic structure of the 27 sampled populations of Lysimachia monelli (LM) in the Mediterranean region (see Table 1). (A) First maximum clustering on AFLP dataset (K = 3). (B) Second maximum clustering on AFLP dataset (K = 7). (C) First maximum clustering on nSSR dataset (K = 5). Colour coding indicates different genetic clusters. The proportion of the inferred genetic clusters is shown for each population. Phylogeographical analyses Bayesian clustering analysis differed between both markers. AFLP dataset showed a first optimum number of K = 3 to capture the most variation in genetic structure (DK = 306.783, mean LnP(K) = -58,178.90; Supplemental Figure 4), thus generating three distinct clusters (Figure 1A): (1) blue-flowered populations from southwestern IP (LM1– LM10 and LM14); (2) blue-flowered populations from the rest of IP (LM11-13 and LM15-19) and Tunisia (LM26-LM27); and (3) all the orange-flowered populations. Bayesian mean FST values varied between clusters: (1) 0.207, (2) 0.186 and (3) 0.228. Capítulo 2 89 In contrast, nSSR dataset showed an optimum number of K = 5 to capture the most variation in genetic structure (DK = 66.535, Mean LnP(K) = -6715.1; Supplemental Figure 4), generating five distinct clusters (Figure 1B): (1) five blue-flowered populations from Portugal (LM1-5); (2) six blue-flowered populations from SW Spain (LM6-10 and LM14); (3) eight blue-flowered populations from C Spain (LM11-13 and LM15-19); (4) orange-flowered populations from NE Spain and Morocco (LM20-24); and (5) one orange-flowered population from Sardinia (LM25) and two blue-flowered from Tunisia (LM26-27). Bayesian mean FST values for each group were 0.299, 0.2587, 0.192, 0.301 and 0.219 respectability. Additionally, according to both, the Calinski and Harabasz' pseudo-F and the Bayesian Information Criterion, the best clustering was K = 25, thus indicating a high degree of independence between populations. PCoA analysis showed a clear separation between flower color lineages, both at individual and population level (Figure 2). At individual level, a total of 28.96% and 22.20% of the total variation was explained for AFLP and nSSR datasets respectively (Figure 2A, B), while at population level, 52.98% and 39.42% of variation were explained for these markers (Figure 2C, D). These ordination results indicated two patterns: a separation of the south-western IP populations from the rest and a separation between blue- and orange-flowered plants. Furthermore, the main genetic/geographical pattern observed in the Bayesian clustering analysis was congruent with the data dispersion in the PCoA analysis at individual level, mainly with AFLP data (K=3, Figure 2A). Capítulo 2 96 nearly mutually exclusive with those of orange lineage. Climate conditions and soil characteristics play a major role in lineage distribution. Blue L. monelli is distributed preferentially by dry habitats while orange L. monelli prefers wet habitats with low micronutrient availability in soils (alkaline soils; Gentili et al., 2018) and poor water holding capacity. Thus, our results suggest different environmental properties that enable the presence of each color lineage and explain their allopatric distribution. Interestingly, climatic requirements for L. monelli lineages are similar to those of the two-color lineages of the sister species L. arvensis, where blue-flowered plants are distributed in drier areas than orange-flowered plants (Arista et al., 2013). Past Niche models only showed with a very high probability the overlapping or the occurrence of very nearby populations of blue and orange lineages in N Marocco and Algeria during LIG and in Algeria and Sardinia during LGM. Genetic diversity and structure through Lysimachia monelli distribution L. monelli has a high genetic diversity, similar or even higher than those from other Mediterranean species of Lysimachia (Calero et al., 1999; Jiménez-López et al., 2020b; Kwiecińska-Poppe et al., 2020). The low Fis values observed in most populations indicate a heterozygote excess that may be a consequence of the species' self-incompatibility system. This excess of heterozygotes may contribute to the maintenance of allele frequencies, increasing genetic diversity, mating opportunities and fertility (Wright et al., 2008), as reported in others self-incompatible species as Leavenworthia alabamic (Koelling et al., 2011), Linaria vulgaris (Bartlewicz et al., 2015) or Bromelia hieronymi (Godoy et al., 2018). He values may also influence population structure (Beaumont and Nichols, 1996) and our results suggest two different structuring levels. On the one hand, a set of parameters indicated a high independence among populations: (1) high Fpriv values, (2) a higher percentage of explained variability within populations than within groups in all cases, (3) a Calinski and Harabasz' pseudo-F k-clustering optimum number of K = 25, and (4) isolation-by-distance evidence with Mantel test. In fact, gene flow was very low between most populations. On the other hand, both AFLP and nSSR showed a geographical structure among L. monelli populations, revealing a particular disjunctive distribution with K = 3 and K = 5 groups, respectively, highly supported by the PCoA. By using AFLP, blue and orange L. monelli lineages were separated in two clusters, which indicates Capítulo 2 97 their genetic distance. In species with blue/red or orange polymorphism, the most frequent color transition is from blue to red (Rausher, 2008), so the red L. monelli lineage is more likely to be derived from the blue one. The genetic diversity data are in agreement with this possibility, since all the genetic parameter indicated that blue linage is much more diverse, suggesting that is older than the orange lineage. Interestingly, an orange population from Sicily and a blue population from Tunisia shared fragments, indicating the existence of gene flow between them in the past. Additionally, blue individuals were separated by AFLP in two different clusters related to their location: a cluster from the Atlantic coast of SW IP much more diverse, and other from the Mediterranean Basin including populations from CE IP and Tunisia. This kind of divergence has been reported in other coastal perennial plant species (Juan et al., 2012). By using nSSR, a clear differentiation was also found between color lineages although populations were grouped on five cluster. The most significant difference with AFLP was the separation of the blue Mediterranean populations in two cluster (instead of one), and more significant, the grouping of blue populations from Tunisia and the orange population from Sicily in a single cluster. Differences between markers could be explained by changes in the paleoclimatic dynamics, which allowed different patterns of gene flow over time, or because the two sets of markers segregate independently (Gaudeul et al., 2004). The grouping between C. Mediterranean populations differing in color is congruent with the fact that these populations shared AFLP fragments indicating the existence of past gene flow between them. Niche modelling and phylogeography All molecular results showed that L. monelli populations have diversified along the longitudinal gradient with western populations showing higher genetic diversity than eastern ones. The same pattern has been reported in other groups of plants, such as Atripex halimus, Taxus baccata or Rosmarinus officinalis, with western Mediterranean populations differing from central or eastern Mediterranean populations (Ortiz-Dorda et al., 2005; González-Martínez et al., 2010; Mateu-Andrés et al., 2013). It has been suggested that populations from the SW IP may have acted as a nucleus of diversification and genetic reservoir in relation to the Mediterranean populations (Kadereit and Westberg, 2007). This differentiation is usually explained by both current orographic elements (e. g. Kadereit and Westberg, 2007; Piñeiro et al., 2007) and phylogeographic Capítulo 2 98 history of the region (Comes and Kadereit, 1998; Hewitt, 1999; Médail and Diadema, 2009), mainly related to the bioclimatic fluctuation of the Strait of Gibraltar, that has played an important role in genetic differentiation of Mediterranean plant species (Kadereit et al., 2005; Arroyo et al., 2008; Escudero et al., 2008; Terrab et al., 2008). However, genetic differentiation of L. monelli has been estimated around 2.1 Myr, just after the definitive opening of the Strait of Gibraltar, so it does not seem that this biogeographical event could have had any influence on the divergence of their populations. Bioclimatic changes across the region due to Quaternary fluctuations (Comes and Kadereit, 1998) and geographical distance seems to influence the divergence between L. monelli populations. The putative Quaternary long-term refugia correspond to the southern PI and Morocco, and coastal areas from the central Mediterranean basin (Medail and Diadema, 2009). Genetic diversity tends to be higher in areas of refuge and contact zones (Bonin et al., 2007). Niche modelling for the blue lineage showed that during the LGM its most favourable habitat was restricted to the SW of the Iberian Peninsula, that may have acted as a climatic refuge. Thus, after LGM, a eastwards colonization of genotypes of the blue lineage originated in the East could have taken place. This situation is consistent with the decreasing pattern of genetic diversity from the east of the center of the Mediterranean of the blue lineage and is a relatively common pattern in Mediterranean lineages (Rodríguez-Sánchez et al., 2009; Migliore et al., 2012). Niche modelling prediction for the red lineage showed that the favorable area did not decrease during the LMG, although it showed a high likelihood of occurrence in N African habitats and in the center west Iberian Peninsula. Among the orange lineage, Moroccan populations were the most diverse, while the Sardinian population was the less diverse, strongly suggesting that Moroccan populations are the oldest. It is likely that, during the Quaternary, a fluctuating island ‘stepping stone’ situation permitted several expansionary pathways from Morocco to other central Mediterranean areas (Ortiz et al., 2008; Del Valle et al., 2019). Sardinian population may have fixed genetic differences due to founder effects following the colonization of this Mediterranean island by a small number of individuals and/or subsequent population bottlenecks, explaining its low current genetic diversity. Alternatively, another possible explanation would involve a phylogeographic split, which might have resulted from contraction of formerly continuous ranges, with populations remaining in three areas: Morocco, NE Iberian Capítulo 2 99 Peninsula and Sardinia. Although this pattern has been found in other Mediterranean species (REF), it would be much less probable because the low genetic diversity of the Sardinian population. Interestingly, niche modelling also predicted the existence of very nearby populations of both lineages in Tunisia and Sardinia during LGM. In fact, a narrow hybrid zone between orange and blue lineage needs to be postulated to explain the intermediate genetic composition of individuals. Gene flow between Sardinian and Tunisian populations has been found with AFLP and molecular nSSR grouped these populations in a sole cluster that can be explained by the mixture of two gene pools. However, gene flow between lineages is currently very low as they do not co-occur in populations. Conclusions The two color lineages of L. monelli show clear differences in distribution that seems to be related to environmental conditions. The two lineages also showed a strong genetic separation with very low gene flow between them. All populations of the orange lineage form a single gene cluster, while those of the blue lineage split into two or three. Blue and orange populations from the C. Mediterranean seems to have much more genetic similarity than other blue and orange populations between them, suggesting a history of past gene flow between them. Overall, our results indicate a complete geographical and genetic separation of the L. monelli color lineages suggesting they are near the end of a speciation process. Author contributions MA and PO conceived the ideas; All the authors collected the data; FJ and MS analyzed the data; FJ and MS led the writing and all the authors revised the manuscript. Acknowledgments This work was supported by the European Regional Development Fund (ERDF) and grants from the Spanish Ministerio de Ciencia e Innovación MICINN (PID2020- 116222GB-I00) and Ministerio de Economía y Competitividad MINECO (CGL2015- 63827, CGL2012-33270) to MA and PLO, (BES-2013-062859) to FJL and (BES-C- Capítulo 2 100 2016-0023) to MSC. We thank the Herbarium of the University of Sevilla (CITIUS) for their collaboration with specimen conservation and the use of the molecular lab. References Alvarado‐Serrano D. F., and Knowles, L. L. (2014). Ecological niche models in phylogeographic studies: applications, advances and precautions. Mol. Ecol. Resour. 14, 233-248. Arista, M., Talavera, M., Berjano, R., and Ortiz, P. L. (2013). Abiotic factors may explain the geographical distribution of flower colour morphs and the maintenance of colour polymorphism in the scarlet pimpernel. J. Ecol. 101: 1613-1622. Arroyo, J., Aparicio, A., Albaladejo, R. G., Muñoz, J., and Braza, R. (2008). Genetic structure and population differentiation of the Mediterranean pioneer spiny broom Calicotome villosa across the Strait of Gibraltar. Biol. J. Linn. Soc. 93: 39-51. Avise, J. C., Bowen, B. W., and Ayala, F. J. (2016). In the light of evolution X: Comparative phylogeography. Proc. Natl. Acad. Sci. U.S.A. 113, 7957-7961. Bartlewicz, J., Vandepitte, K., Jacquemyn, H., and Honnay O. (2015). Population genetic diversity of the clonal self-incompatible herbaceous plant Linaria vulgaris along an urbanization gradient. Biol. J. Linn. Soc. 116, 603-613. Barton, N. H., and Hewitt, G. M. 1985. Analysis of hybrid zones. Annu. Rev. Ecol. Evol. Syst. 16, 113-148. Beaumont, M. A., and Nichols, R. A. (1996). Evaluating loci for use in the genetic analysis of population structure. Proc. R. Soc. B: Biol. Sci. 263, 1619-1626. Benítez‐Benítez, C., Escudero, M., Rodríguez‐Sánchez, F., Martín‐Bravo, S., and Jiménez‐Mejías, P. (2018). Pliocene–Pleistocene ecological niche evolution shapes the phylogeography of a Mediterranean plant group. Mol. Ecol. 27,1696-1713. Bonin, A., Bellemain, E., Bronken, E. P., Pompanon, F., Brochmann, C., and Taberlet, P. (2004). How to track and assess genotyping errors in population genetic studies. Mol. Ecol. 13, 3261-3273. Calero, C., Ibáñez, O., Mayol, M., and Rosselló, J. A. (1999). Random amplified polymorphic DNA (RAPD) markers detect a single phenotype in Lysimachia minoricensis JJ Rodr. (Primulaceae), a wild extinct plant. Mol. Ecol. 8, 2133-2136. Casimiro-Soriguer, R., Talavera, M., Balao, F., Terrab, A., Herrera, J., and Talavera, S. (2010). Phylogeny and genetic structure of Erophaca (Leguminosae), a East-West Mediterranean disjunct genus from the Tertiary. Mol. Phylogenet. Evol. 56, 441-450. Comes H. P., and Kadereit J. W. 1998. The effect of Quaternary climatic changes on plant distribution and evolution. Trends Plant Sci. 3, 432-438. Capítulo 2 101 Del Valle, J. C., Casimiro-Soriguer, I., Buide, M., Narbona, E., and Whittall, J. B. (2019). Whole plastome sequencing within Silene section Psammophilae reveals mainland hybridization and divergence with the Balearic Island populations. Front. Plant Sci. 1466. Duggen, S., Hoernle, K., van den Bogaard, P., Rüpke, L., and Morgan, J. P. (2003). Deep roots of the Messinian salinity crisis. Nature 422, 602-606. Dyer, A. G., Jentsch, A., Burd, M., Garcia, J. E., Giejsztowt, J., Camargo, M. G., et al. (2021). Fragmentary blue: resolving the rarity paradox in flower colors. Front. Plant Sci. 11, 2212. Earl, D. A., and VonHoldt, B. M. (2012). STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv. Genet. Resour. 4(2), 359-361. El Bakkali, S., Gourgaud, A., Bourdier, J. L., Bellon, H., and Gundogdu, N. (1998). Postcollision neogene volcanism of the Eastern Rif (Morocco): magmatic evolution through time. Lithos 45(1-4), 523-543. Ellegren, H., and Galtier, N. (2016). Determinants of genetic diversity. Nat. Rev. Genet. 17, 422-433. Escudero, M., Vargas, P., Valcárcel, V., and Luceño, M. (2008). Strait of Gibraltar: an effective gene-flow barrier for wind-pollinated Carex helodes (Cyperaceae) as revealed by DNA sequences, AFLP, and cytogenetic variation. Am. J. Bot. 95, 745-755. Estoup, A., and Angers, B. (1998). Microsatellites and minisatellites for molecular ecology: theoretical and experimental considerations. In: Carvalho G (ed.) Advances in Molecular Ecology. NATO Press: Amsterdam. Evanno, G., Regnaut, S., and Goudet, J. (2005). Detecting the number of clusters of individuals using the software structure: a simulation study. Mol. Ecol. 14, 2611-2620. Excoffier L. (2004). Patterns of DNA sequence diversity and genetic structure after a range expansion: lessons from the infinite-island model. Mol. Ecol.13, 853-864. Excoffier, L., Laval, G., and Schneider, S. (2005). Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evol. Bioinform. Online 1, 47-50. Feliner, G. N. (2014). Patterns and processes in plant phylogeography in the Mediterranean Basin. A review. Perspect. Plant Ecol. Evol. Syst. 16, 265-278. Ferguson, L. F. (1972). Lysimachia L. in Tutin TG, et al. eds. Flora Europaea vol 3. Cambridge: Cambridge University Press. 26-27. Fick, S. E., and Hijmans, R. J. (2017). WorldClim 2: new 1km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37, 4302-4315. Forsman, A., Ahnesjö, J., Caesar, S., and Karlsson, M. (2008). A model of ecological and evolutionary consequences of colour polymorphism. Ecology 89, 34-40. Capítulo 2 102 García-Castaño, J. L., Terrab, A., Ortiz, M. A., Stuessy, T. F., and Talavera, S. (2014). Patterns of phylogeography and vicariance of Chamaerops humilis L. (Palmae). Turk. J. Bot. 38, 1132-1146. García, P. L, Talavera, S., Ortiz, P. L., and Arista, M. (1997). Números cromosomáticos de plantas occidentales 734-737. An. Jard. Bot. Madr. 55, 344-345. Gaudeul, M., Till-Bottraud, I., Barjon, F., and Manel, S. (2004). Genetic diversity and differentiation in Eryngium alpinum L. (Apiaceae): comparison of AFLP and microsatellite markers. Heredity 92, 508-518. Gentili, R., Ambrosini, R., Montagnani, C., Caronni, S., and Citterio, S. (2018). Effect of Soil pH on the Growth, Reproductive Investment and Pollen Allergenicity of Ambrosia artemisiifolia L. Front. Plant Sci. 9, 1335. Gibbs, P. E., and Talavera, S. (2001). Breeding system studies with three species of Anagallis (Primulaceae): self-incompatibility and reduced female fertility in A. monelli L. Ann. Bot. 88, 139-144. Gray, S. M, and McKinnon, J. S. (2007). Linking colour polymorphism maintenance and speciation. Trends Ecol. Evol. 22, 71-79. Godoy, F. M. D. R., Lenzi, M., Ferreira, B. H. D. S., Silva, L. V. D., Zanella, C. M., and Paggi, G. M. (2018). High genetic diversity and moderate genetic structure in the selfincompatible, clonal Bromelia hieronymi (Bromeliaceae). Biol. J. Linn. Soc. 187(4), 672- 688. González-Martínez, S. C., Dubreuil, M., Riba, M., Vendramin, G. G., Sebastiani, F., and Mayol, M. (2010) Spatial genetic structure of Taxus baccata L. in the western Mediterranean Basin: past and present limits to gene movement over a broad geographic scale. Mol. Phylogenet. Evol. 55(3), 805-815. Goudet J. (1995). FSTAT (version 1.2): a computer program to calculate F-statistics. J. Hered. 86, 485-486. Gueguen, E., Doglioni, C., and Fernandez, M. (1998). On the post-25 Ma geodynamic evolution of the western Mediterranean. Tectonophysics 298, 259-269. Hamrick, J. L., Godt, M. J. W., and Sherman-Broyles, S. L. (1992). Factors influencing levels of genetic diversity in woody plant species. In Adams, W. T., et al. eds. Population genetics of forest trees. Springer: Dordrecht. 95-124. Hewitt, G. M. (1999). Post-glacial re-colonization of European biota. Biol. J. Linn. Soc. 68, 87-112. Imbert, E., Wang, H., Anderson, B., Hervouet, B., Talavera, M., and Schatz, B. (2014). Reproductive biology and colour polymorphism in the food-deceptive Iris lutescens (Iridaceae). Acta Bot. Gall. 161, 117-127. Irwin, R. E., Strauss, S. Y., Storz, S., Emerson, A., and Guibert, G. (2003). The role of herbivores in the maintenance of a flower colour polymorphism in wild radish. Ecology 84, 1733-1743. Capítulo 2 103 Jacquemyn, H., Roldán-Ruiz, I., and Honnay, O. (2010). Evidence for demographic bottlenecks and limited gene flow leading to low genetic diversity in a rare thistle. Conserv. Genet. 11, 1979-1987. Jakob, S. S., Ihlow, A., and Blattner, F. R. (2007). Combined ecological niche modelling and molecular phylogeography revealed the evolutionary history of Hordeum marinum (Poaceae)—niche differentiation, loss of genetic diversity, and speciation in Mediterranean Quaternary refugia. Mol. Ecol. 16(8), 1713-1727. Jiménez-López, F. J., Matas, L., Arista, M., and Ortiz, P. L. (2020a). Flower colour segregation and flower discrimination under the bee vision model in the polymorphic Lysimachia arvensis. Plant Biosyst. 154, 535-543. Jiménez-López, F. J., Ortiz, P. L., Talavera, M., and Arista, M. (2020b). Reproductive Assurance Maintains Red-Flowered Plants of Lysimachia arvensis in Mediterranean Populations Despite Inbreeding Depression. Front. Plant Sci. 11, 563110. Jiménez-López, F. J., Talavera, M., Ortiz, P. L., and Arista, M. (2015). Nuclear microsatellite primers in the annual herb Lysimachia arvensis (Myrsinaceae) and closely related taxa. Biochem. Syst. Ecol. 58, 242-246. Jiménez-López, F. J., Viruel, J., Arista, M., Ortiz, P. L. and Talavera, M. (2022). Molecular approaches reveal speciation between red- and blue-flowered plants in the Mediterranean Lysimachia arvensis and L. monelli (Primulaceae). Bot. J. Linn. Soc. Juan, A., Fay, M. F., Pastor, J., Juan, R., Fernández, I., and Crespo, M. B. (2012). Genetic structure and phylogeography in Juniperus oxycedrus subsp. macrocarpa around the Mediterranean and Atlantic coasts of the Iberian Peninsula, based on AFLP and plastid markers. Eur. J. For. Res. 131, 845-856. Kadereit, J. W., Arafeh, R., Somogyi, G., and Westberg, E. (2005). Terrestrial growth and marine dispersal? Comparative phylogeography of five coastal plant species at a European scale. Taxon 54, 861-876. Kadereit, G., and Yaprak, A. E. (2008). Microcnemum coralloides (Chenopodiaceae- Salicornioideae): an example of intraspecific East-West disjunctions in the Mediterranean region. An. Jard. Bot. Madr. 65, 415-426. Kadereit, J. W., and Westberg, E. (2007). Determinants of phylogeographic structure: a comparative study of seven coastal flowering plant species across their European range. Watsonia 26, 229-238. Kalia, R. K., Rai, M. K., Kalia, S., Singh, R., and Dhawan, A. K. (2011). Microsatellite markers: an overview of the recent progress in plants. Euphytica 177, 309-334. Kamvar, Z. N., Tabima, J. F., and Grünwald, N. J. (2014). Poppr: an R package for genetic analysis of populations with clonal, partially clonal, and/or sexual reproduction. PeerJ 2, e281. Koelling, V. A., Hamrick, J. L., and Mauricio, R. (2011). Genetic diversity and structure in two species of Leavenworthia with self-incompatible and self-compatible populations. Heredity 106, 310-318. Capítulo 2 104 Kwiecińska-Poppe, E., Haliniarz, M., Sowa, S., and Paczos-Grzęda, E. (2020). Genetic diversity and population structure of endangered plant species Anagallis foemina Mill. [Lysimachia foemina (Mill.) U. Manns & Anderb.]. Physiol. Mol. Biol. Plants 26, 1675- 1683. Lavergne, S., Hampe, A., and Arroyo, J. (2012). In and out of Africa: how did the Strait of Gibraltar affect plant species migration and local diversification?. J. Biogeogr. 40, 24- 36. Lonergan, L., and White, N. (1997). Origin of the Betic-Rif mountain belt. Tectonics 16, 504-522. Lopes, I. F., Miño, C. I., and Del Lama, S. N. (2007). Genetic diversity and evidence of recent demographic expansion in waterbird populations from the Brazilian Pantanal. Braz. J. Biol. 67, 849-857. López-Jurado, J., Mateos-Naranjo, E., and Balao, F. (2018). Niche divergence and limits to expansion in the high polyploid Dianthus broteri complex. New Phytol. 222, 1076- 1087. Macaya‐Sanz, D., Heuertz, M., López‐de‐Heredia, U., De‐Lucas, A. I., Hidalgo, E., Maestro, C., et al. (2012). The Atlantic–Mediterranean watershed, river basins and glacial history shape the genetic structure of Iberian poplars. Mol. Ecol. 21(14), 3593-3609. Martín‐Hernanz, S., Martínez‐Sánchez, S., Albaladejo, R. G., Lorite, J., Arroyo, J., and Aparicio, A. (2019). Genetic diversity and differentiation in narrow versus widespread taxa of Helianthemum (Cistaceae) in a hotspot: The role of geographic range, habitat, and reproductive traits. Ecol. Evol. 9, 3016-3029. Maruyama, T., and Fuerst, P. A. (1985). Population bottlenecks and nonequilibrium models in population genetics. II. Number of alleles in a small population that was formed by a recent bottleneck. Genetics 111, 675-689. Mateu-Andrés, I., Aguilella, A., Boisset, F., Currás, R., Guara, M., Laguna, E., et al. (2013). Geographical patterns of genetic variation in rosemary (Rosmarinus officinalis) in the Mediterranean basin. Bot. J. Linn. 171(4), 700-712. McLean, C. A., and Stuart‐Fox, D. 2014. Geographic variation in animal colour polymorphisms and its role in speciation. Biol. Rev. 89, 860-873. Migliore, J., Baumel, A., Juin, M., Mdail, F., (2012). From Mediterranean shores to central Saharan mountains: key phylogeographical insights from the genus Myrtus. J. Biogeogr. 39, 942-956. Médail, F., and Diadema, K. (2009). Glacial refugia influence plant diversity patterns in the Mediterranean Basin. J. Biogeogr. 36, 1333-1345. Meirmans, P. G., and Van Tienderen, P. H. (2004). GENOTYPE and GENODIVE: two programs for the analysis of genetic diversity of asexual organisms. Mol. Ecol. Notes 4, 792-794. Capítulo 2 105 Muñoz‐Pajares, A. J., Perfectti, F., Loureiro, J., Abdelaziz, M., Biella, P., Castro, M., et al. (2018). Niche differences may explain the geographic distribution of cytotypes in Erysimum mediohispanicum. Plant Biol. 20, 139-147. Myers, N., Mittermeier, R. A., Mittermeier, C. G., Da Fonseca, G. A., and Kent, J. (2000). 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853-858. Narbona, E., Wang, H., Ortiz, P. L., Arista, M., Imbert, E. (2018). Flower colour polymorphism in the Mediterranean Basin: Occurrence, maintenance and implications for speciation. Plant Biol. 20, 8-20. Nosil, P. (2012). Ecological speciation. Oxford University Press. Ortiz, M. Á., Tremetsberger, K., Terrab, A., Stuessy, T. F., García‐Castaño, J. L., Urtubey, E., et al. (2008). Phylogeography of the invasive weed Hypochaeris radicata (Asteraceae): from Moroccan origin to worldwide introduced populations. Mol. Ecol. 17, 3654-3667. Ortiz, P. L., Berjano, R., Talavera, M., Rodríguez-Zayas, L., and Arista, M. (2015). Flower colour polymorphism in Lysimachia arvensis: How is the red morph maintained in Mediterranean environments?. Perspect. Plant Ecol. Evol. Syst. 17, 142-150. Ortiz-Dorda, J., Martínez-Mora, C., Correal, E., Simón, B., and Cenis, J. L. 2005. Genetic structure of Atriplex halimus populations in the Mediterranean Basin. Ann. Bot. 95, 827- 834. Phillips, S. J., and Dudík, M. (2008). Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31, 161-175. Pineiro, R., Aguilar, J. F., Munt, D. D., and Feliner, G. N. (2007). Ecology matters: Atlantic–Mediterranean disjunction in the sand-dune shrub Armeria pungens (Plumbaginaceae). Mol. Ecol. 16, 2155-2171. Piry, S., Luikart, G., and Cornuet, J. M. (1999). BOTTLENECK: a program for detecting recent effective population size reductions from allele data frequencies. J. Hered. 90, 502- 503. Poggio, L., De Sousa, L. M., Batjes, N. H., Heuvelink, G., Kempen, B., Ribeiro, E., et al. (2021). SoilGrids 2.0: producing soil information for the globe with quantified spatial uncertainty. Soil 7(1), 217-240. Pritchard, J. K. (2010). Documentation for STRUCTURE software: version 2.3. Available at: http://pritchardlab.stanford.edu/structure_software/release_versions/v2.3.4/html/structur e.htm Pritchard, J. K., Stephens, M., and Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics 155, 945-959. Pujadas, A. (1997). Anagallis L. In Castroviejo S, et al. eds. Flora Iberica V. Madrid: Real Jardín Botánico, CSIC. 57-62. Capítulo 3 112 commonly, flower color polymorphisms are caused by changes in anthocyanin composition as in Ipomoea nil (blue, magenta, purple and red; Morita et al., 2005), Phlox drummondii (blue and red; Hopkins and Rausher, 2011) and Lysimachia arvensis (blue and orange; Arista et al., 2013). Streisfeld and Rausher (2011) predicted that evolutionary transitions in floral color are more likely caused by regulatory mutations in ABP genes instead of mutations in the coding sequence (CDS), because regulatory changes can be tissue-specific and are thus less likely to have deleterious pleiotropic effects. In this study, we focus on the flavonoid genes and compounds underlying the blue and orange flower color variation of Lysimachia arvensis. Anthocyanins are synthesized by the anthocyanin biosynthetic pathway (ABP) — a highly conserved pathway among angiosperms and one of the best studied pathways for the production of secondary metabolites in plants (Holton and Cornish, 1995; Davies and Schwinn, 2006; Ho and Smith, 2016; Wheeler and Smith, 2019). The ABP is part of the flavonoid biosynthetic pathway, a broader metabolic pathway emerging from phenylpropanoid metabolism that includes flavone, flavonol, and anthocyanin synthesis plus the post-translational modifying steps like acylation and glycosylation (Clegg and Durbin, 2000; Sobel and Streisfeld, 2013; Chen et al., 2018; Gao et al., 2019). Approximately 100 structural genes (KEGG pathway database1) are involved in this broad flavonoid biosynthetic pathway. In addition, tissue-specific transcription factors play an important role in this pathway (Quattrocchio et al., 2006; Zhao et al., 2013), regulating those structural genes through the MYB-bHLH-WDR complex (MBW; Xu et al., 2015; Gao et al., 2019). This regulatory complex is conserved among vascular plants; however, the structural genes that they regulate varies among angiosperm lineages (Hichri et al., 2011; Lai et al., 2013; Albert et al., 2014; Xu et al., 2015). Regulatory complexity (e.g., tissue-specific regulation) is conferred by the large gene families for these proteins (e.g., R2R3 MYBs; Stracke et al., 2001; WD40; Xu et al. 2015). In particular, MYB regulators tend to be disproportionately involved in floral anthocyanin variation both within and between species. The number of studies describing the causes of flower color variation in the wild is rapidly growing (e.g. Dick et al., 2011; Yuan et al., 2013; Casimiro-Soriguer et al., 2016; Kellenberger et al., 2019). Wheeler and Smith (2019) reviewed 15 studies involving natural evolutionary transitions between anthocyanin pigment types in which all 15 Capítulo 3 113 examples involved decreased expression in either F3'H or F3'5'H, or both genes (Wessinger and Rausher, 2012; Wheeler and Smith, 2019). Furthermore, two of these cases involved DFR mutations conferring substrate specificity, a secondary modification to redirect the flux directly down the pelargonidin branch of the ABP (Johnson et al., 2001, Zufall and Rausher, 2004; Smith and Rausher, 2011; Wheeler and Smith, 2019). According to Streisfeld and Rausher (2011), flower color variation associated with anthocyanin loss is more likely to be caused by regulatory genes mutations to avoid deleterious pleiotropic effects, while flower color variation associated with changes in anthocyanin type are more likely to be caused by mutations in structural genes, however Wheeler and Smith's findings (2019) suggest both regulatory and structural changes are often correlated with the blue to red transition in natural flower color changes. Traditional approaches to finding the genes underlying flower color variation involve sequencing and determining expression in a relatively narrow group of candidate genes in the ABP (Maitre et al. 2019) which we will refer to as the narrow ABP. Alternatively, transcriptomic approaches using RNA-Seq allow a much broader investigation of the underlying causes of flower color variation (e.g., across the broader flavonoid biosynthetic pathway). For non-model species, de novo transcriptome assembly can be used to estimate genes with significant differential expression (DEGs; Wang et al., 2009; Martin and Wang, 2011; Grabherr et al., 2011; Anders et al., 2013; Haas et al., 2013; Butler et al., 2014; Casimiro-Soriguer et al., 2016; Chen et al., 2018; Duan et al., 2020) Even using a transcriptome approach, most studies of anthocyanin loss have focused on a narrow group of ABP candidate genes (Lulin et al., 2012; Lou et al., 2014; Casimiro- Soriguer et al., 2016; Gao et al., 2016; Lang et al., 2019; Li et al., 2019a; Duan et al., 2020; Zhang et al., 2020). Herein, we describe the de novo assembled transcriptome for blue- and orange-flowered Lysimachia arvensis to determine differentially expressed genes among the broader flavonoid biosynthetic pathway. Lysimachia arvensis (L.) Manns & Anderb. (Primulaceae), formerly Anagallis arvensis, is a self-compatible (Gibbs and Talavera, 2001), tetraploid (2n=40; Pujadas, 1997; Moneim et al., 2003) annual species displaying a petal color polymorphism. Blue- and orange-flowered plants, traditionally described as blue- and red-flowered plants (Ferguson, 1972), occur in monomorphic and polymorphic populations across its native range in Europe (Arista et al., 2013; Ortiz et al., 2015). The blue petals of L. arvensis contain malvidin derivatives, whereas orange petals contain pelargonidin derivatives Capítulo 3 114 (Lawrence et al., 1939; Harborne, 1968; Ishikura, 1981). Although both L. arvensis flower colors have long been described as different morphs, recent molecular phylogenetic studies indicate that these flower color phenotypes reflect very closely related, yet distinct, evolutionary lineages (Jiménez-López, 2019; Jiménez-López et al., in review) with some degree of reproductive isolation (Jiménez-López et al., 2020). Pollinators (mainly Halictus spp. bees; Ortiz et al., 2015; Jiménez-López et al., 2019) exhibit a preference for the blue individuals when given a choice. Across Europe, there is a latitudinal cline in flower color from north to south. Completely orange populations are common in the more mesic temperate latitudes of northern and central Europe whereas completely blue populations are most common in the more xeric Mediterranean environments of southern Europe with polymorphic populations in between (Arista et al., 2013). In combination with the ecological divergence of these two flower color variants across the species range, this case of incipient speciation provides an opportunity to investigate the role of the flavonoid biosynthetic pathway during speciation (Coyne and Orr, 1998; Widmer et al., 2009; Reiseberg et al., 2010). In this study, we aim to identify the underlying genetic and biochemical causes of flower color variation in L. arvensis. Previous studies have found that gene expression and/or gene structure at a critical branch-point in the anthocyanin biosynthetic pathway involving two genes, F3'5'H and DFR, are involved in most blue to red/orange flower color shifts (malvidin/cyanidin to pelargonidin; Davies, 2009; Streisfeld and Rausher, 2009; 2011; Smith et al., 2012; Ho and Smith, 2016; Wheeler and Smith, 2019). In order to identify the underlying genetic causes of petal color polymorphism in L. arvensis, we performed de novo transcriptome assembly from RNA-Seq data. We then tested for differential expression across a broad panel of 94 structural and 46 regulatory genes involved in the flavonoid biosynthetic pathway. We also examined these two candidate genes for non-synonymous single nucleotide polymorphisms (SNPs) fixed between blue and orange flowers in reference to the known functional domains as described for other species (Johnson et al., 2001; Seitz et al., 2007 Smith et al., 2012; Guo et al., 2019). Finally, we interpret our transcriptome results in light of the flavonoid biochemical profiles of both flower colors using Ultra-High Performance Liquid Chromatography (UHPLC-MS). Capítulo 3 115 Material and Methods Plant material For transcriptomic analysis, we collected seeds from blue-flowered plants that belong to a primarily blue-flowered population (blue:orange ratio = 10:1; Spain, Andalusia, Hinojos; 37º17'37''N, 6º25'15''W) and orange-flowered plants that belong to a monomorphic orange population (France, Corsica, Solenzara; 41º51'19''N, 9º21'16''E). Voucher specimens were deposited in the University of Seville Herbarium (Herbarium Numbers: SEV286467 and SEV279123). Although there are benefits from studying a mixed population in order to control for differences in genetic background that can arise due to geographic separation and/or local adaptation in allopatry, we used these two populations because we previously obtained four generations of inbred lines that exhibit reduced heterozygosity, which facilitates transcriptome assembly and analysis. These seeds were sown in a glasshouse at Santa Clara University (California, USA), where the RNA extractions of blue- and orange-flowered samples were performed on the same days, alternating the preparation of samples between flower colors, in relatively homogeneous environmental conditions. We sampled all five petals from ~30 flowers per plant, from eight blue- and eight orange-flowered plants. We avoided the "bullseye" at the base of each petal because this study is focused on the blue and orange flower color polymorphism, and we have no evidence that the bullseye is different between morphs based on visible and UV digital images (Supplemental Figure 1). All samples were taken from first-day anthesis flowers. Immediately after excision, petals were placed in cryotubes that were immediately immersed in liquid nitrogen and stored at -80ºC to avoid subsequent changes in gene expression and RNA degradation. For flavonoid profiling, we sampled one blue and one orange plant from Spain and confirmed with 13 blue and 10 orange glasshouse-grown introduced plants from California. Each sample was composed of between 10 and 15 petals without the central bullseye. Petal samples were stored in microtubes containing 200µl of MeOH with 1% HCl. RNA extraction and de novo assembly Petals were homogenized using a mortar and pestle. Total RNA was then extracted following the Qiagen RNeasy®Plant Mini Kit protocol (Qiagen, Germany) with the addition of PEG 20,000 mol. wt. (550µl, 2%; Gehrig et al. 2000) before the first filtering Capítulo 3 116 step. The addition of PEG was essential to achieve reasonable RNA concentrations for library preparation and sequencing. RNA samples were stored at -80ºC until further analysis. RNA concentration and purity was initially assessed with a Nanodrop Nd-1000 (ThermoFisher) and agarose gel, and then confirmed with a Bioanalyzer (Agilent, Santa Clara, CA, USA) before sequencing. Individual libraries were barcoded, multiplexed and sequenced as 150 bp paired-end reads using two lanes on an Illumina Hi-Seq 2000 (Illumina, San Diego, CA, USA) through Novogene (Beijing, China). Raw paired-end Illumina reads were assessed for quality using FastQC (Andrews, 2010) and were processed using Rcorrector v1.0.4 (Song and Florea, 2015) to correct random sequencing errors. Then, reads were trimmed with Trimmomatic v0.39 (Bolger et al., 2014) to remove any read containing bases with Phred scores lower than 20, low quality reads less than 50 bp long, and any adapter or other Illumina-specific sequences that were still present. The remaining reads were filtered with Kraken 2 (Wood et al., 2019) to remove Small and Large Subunit ribosomal RNA (SILVA database; Quast et al., 2013) and contaminating reads (minikraken2_v2 database). Additionally, we used custom-built databases, derived from RefSeq libraries: UniVec_Core, viral, mitochondrion, plastid, plasmid, archaea, bacteria, protozoa, human and fungi. Filtered reads were combined across all samples into a single RNA-Seq data set. We conducted a de novo transcriptome assembly using Trinity v2.8.5 (Grabherr et al., 2011) to generate a single reference transcriptome assembly for L. arvensis. Statistical information on the assembly quality was obtained with the "TrinityStats.pl" commands in the Trinity package and Bowtie2 software (Langmead and Salzberg, 2012). Trinity produces assembled genes and isotigs (which may represent alleles, isoforms, and/or splice variants). We used the assembled genes for the DEGs in order to have a single locus per expression measure and isotigs for the SNP and phylogenetic analysis allowing a balance of certainty in the assembly quality while maximizing ortholog coverage for phylogenetic inference. Identification of differentially expressed genes (DEGs) Since the de novo transcriptome assembly was done using the entire dataset as a single sample (blue- and orange-flowered plants combined), we then mapped the reads from each sample back onto the reference transcriptome. Trimmed mean of M-values (TMM, mean of log-expression ratios; Robinson and Oshlack, 2010), were calculated for each Capítulo 3 117 gene with RSEM software (Li and Dewey, 2011) using a Pearl script (align_and_estimate_abundance.pl) provided with the Trinity distribution. Then, we used EdgeR package (Robinson et al., 2009) in R v4.0.0 (R Core Team, 2020) to determine statistically significant differentially expressed genes (DEGs) between different colored samples. In order to be conservative in DEG identification, we set two thresholds for determining significant differential expression: the false discovery rate (FDR) was less than 10-5 and the expression difference threshold was greater than one log2 fold-change (log2FC). Wilcoxon rank-sum test for ABP and non-ABP gene comparisons were conducted in R v4.0.0 (R Core Team, 2020). In order to focus on a broadly defined set of genes associated with the broader flavonoid biosynthetic pathway, we explored 94 structural loci (Supplemental Tables 1 and 2). These genes start at the phenylpropanoid biosynthetic pathway and continue through the entire flavonoid biosynthetic pathway to include the ABP and three non-anthocyanin flavonoid side-branches: flavone, flavonol and isoflavonoid biosynthetic pathways (KEGG pathway database1). In addition, we selected 46 regulatory genes involved in flavonoid pathway regulation, including the gene families for R2R3-Myb, bHLH and WD40 loci (Li, 2014; Xu et al., 2015; Gates et al., 2017; Supplemental Table 2). To identify and annotate the flavonoid biosynthetic pathway genes in our de novo transcriptome assembly, we used Kakapo2. As input, we specified a list of 140 Genbank accessions for the genes involved in flavonoid biosynthetic pathway, restricted to the Order Ericales (if the gene was not found within Ericales, we removed the Order restriction). Phylogenetic analysis of F3'5'H and DFR Based on previous studies and our expression results, we carefully examined the RNA sequences of F3'5'H and DFR in relation to a set of Genbank reference sequences. In order to generate sequences of the coding region for these two enzymes, we mapped the Illumina reads of every sample to each F3'5'H or DFR isotig that arose from our Trinity de novo assembly. We used the "Map to Reference" function in Geneious v9.0.4 (Kearse et al., 2012), applying "highest sensitivity/slow" and "trim sequences'' parameters. A subset of these isotigs, across orange- and blue-flowered samples for each gene, were used as queries in BLASTn tool from NCBI3 to locate homologous loci from a group of Capítulo 3 118 reference sequences (See results section; Supplemental Figure 4). We selected a set of Genbank references that represented the top hits and some of the best-studied model systems in flower color research for our phylogenetic analysis. For each gene, the isotigs and Genbank reference sequences were aligned using the MUSCLE plug-in from within Geneious v9.0.4. Isotigs missing large portions of the coding sequence or containing >30% ambiguities were removed from the alignment, thus excluding some putative splice variants to improve accuracy. For DFR, we removed the 5' and 3' ends (24bp and 52bp, respectively) of the coding sequence (CDS) due to alignment ambiguities before conducting phylogenetic analysis. We reconstructed the relationships using the RAxML 7.2.8 plug-in from within Geneious v9.0.4, searching for the maximum likelihood tree with 100 bootstrap replicates. We then interpreted the phylogenetic tree in light of normalized expression values (TMM) in R v4.0.0 (R Core Team, 2020). SNP frequencies for F3'5'H and DFR In addition, we examined the frequency of non-synonymous SNPs that showed some tendency to differentiate flower color types for F3'5'H and DFR-1 (DFR-2 was only present in orange-flowered individuals so no sequences were available from blue-samples for comparison). For each sample, every isotig was determined to be fixed or variable. Although L. arvensis is tetraploid, we treated these as either homozygotes or heterozygotes (respectively) for our frequency calculations. We calculated the frequency of each non-synonymous SNP that showed differences between the flower color types providing BLOSUM62 scores (Henikoff and Henikoff, 1992), an empirical measure that estimates the biological probability of an amino acid substitution based on an alignment, and noting the biochemical properties of the amino acids involved. Furthermore, we extended this survey to include synonymous color-differentiating SNPs in the CDS for F3'5'H and DFR-1, including SNPs in the 5' and 3' UTRs for F3'5'H. In addition, since there were two copies of DFR, we also examined all non-synonymous amino acid substitutions between DFR-1 and DFR-2 in light of the location of variation in known functional domains responsible for substrate specificity and binding (Johnson et al., 2001; Petit et al., 2007; Smith et al., 2012). For both genes the amino acid residue numbers are based on the L. arvensis alignment, which translates to Vitis vinifera plus two amino acids (Petit et al., 2007). Identification of flavonoid compounds by UHPLC-MS Capítulo 3 119 Flavonoids were extracted by placing 10-15 petals of blue and orange L. arvensis flowers in microtubes containing 200µl of MeOH with 1% HCl. Homogenization was performed with a BeadBeater using 3 x 3.2 mm steel balls for 30 seconds. Samples were stored at - 80°C until they could be loaded on a Dionex UltiMate3000 ultra-high pressure liquid chromatography system equipped with a diode array detector and connected to a Thermo Fleet LCQ mass spectrometer (purchased from Thermo-Fisher, Waltham, MA USA). Data were analyzed using Xcalibur software (Thermo Fisher Scientific). Crude extracts were eluted through a Phenomenex Gemini Column (50 x 2 mm, 3 micron; acquired from Phenomenex, Torrance, CA, USA) with a multi-step gradient of two solvents: 0.5% formic acid in water (solvent A) and 0.5% formic acid in methanol (solvent B) at a flow rate of 0.5 mL/min. The multi-step gradient involved an initial 2 min elution at 15% B, followed by a linear gradient over 5 min to 50% B, then isocratic elution at 50% B for 3 min, then flushing the column with 100% B for 3 min. The diode array detector collected UV-Vis spectra of eluted compounds from 190-700 nm. Mass spectra of eluted compounds were ionized using electrospray ionization and the mass detector was set to positive ion detection mode with a mass range of 150-800 Da. Flavonoid identification was based on the compound's retention time, UV-Vis spectra and whenever possible, chromatographic comparisons with authentic standards. We used flavonoid standards including quercetin, luteolin, kaempferol, isorhamnetin, malvidin, pelargonidin, cyanidin and delphinidin (purchased from Extrasynthese, Genay, France) that were previously reported for L. arvensis and other Lysimachia species (Supplemental Table 3). Putative identifications were also compared with published flavonoid data4. Results RNA extraction and de novo assembly Bioanalyzer results confirmed we extracted high-quality RNA without substantial degradation. All RNA integrity numbers (RIN) were above the recommended RIN=6.3 (Novogene, unpublished). The mean RIN value was 8.0 (6.7-9.0). The total number of filtered reads in the transcriptome analysis ranged from 21 million to 41 million in blue samples (mean of 26 million) and from 21 million to 26 million in orange samples (mean of 24 million). Approximately 98% of filtered reads of each sample mapped to the de novo assembled L. arvensis reference transcriptome. Isotig number per sample varied Capítulo 3 120 from 146 thousand to 193 thousand and the number of genes ranged from 37 thousand to 42 thousand (Supplemental Table 4). The GC percentage of the assembly was 38.27% and it had an N50 value of 1,635 bp. The average contig length was 973.5 bp and the total number of assembled bases equaled 389,888,932 bp. A total of 189,189 trinity genes and 400,523 trinity isotigs were obtained. Identification of differentially expressed genes (DEGs) Initially, we identified 48,306 differentially expressed genes and 102,750 differentially expressed isotigs in the L. arvensis transcriptome with EdgeR. After applying the cut-off values for FDR and log2FC, we reduced the number of DEGs to 10,624 genes. There were 6,966 DEGs with higher expression in blue samples and 3,658 DEGs with higher expression in orange samples (Supplemental Figure 2). The average log2FC of DEGs for blue>orange (B>O) was 2.39 and for O>B was 3.16. Restricting our analysis to the flavonoid biosynthetic pathway, our Kakapo results identified 60% (58/94) of the structural genes and 20% (10/46) of the regulatory genes in our 140 gene panel. Since some loci were found in multiple copies, a total of 153 structural gene copies and 13 regulatory gene copies were examined for differential expression (Supplemental Tables 5 and 6; these tables contain the number of copies for each gene identified). Gene expression analysis showed 38 structural genes and one regulatory gene with significant differential expression between flower colors (Figure 1; Table 1). Among the DEGs, 22 showed higher expression in blue-flowered plants, while 17 showed higher expression in orange-flowered plants (Figure 1). Structural DEGs from the early steps of the flavonoid biosynthetic pathway have lower expression levels (TMM) compared to DEGs from the narrow ABP (Wilcoxon rank-sum test, W= 46606, p= 2.2x10-16; Figure 2). These trend is also present within flower colors (orange ABP gene expression is 4.8x higher than orange non-ABP gene expression; Wilcoxon ranksum test, W = 12111, p = 1.3x10-12; blue ABP gene expression is 6x higher than blue non-ABP gene expression; Wilcoxon rank-sum test, W = 11278, p = 8.5x 10-9). In general, structural DEGs with higher expression in blue samples tended to have ~2x higher overall expression than those genes with higher expression in orange samples (average TMMblue = 55 vs. average TMMorange = 29; Figure 2). Focusing on the narrow ABP, we obtained eight structural DEGs with higher expression in blue L. arvensis (Table 1; Figure 2). On the contrary, we only found two with higher expression in orange L. Capítulo 3 121 arvensis (Table 1; Figure 2). Two candidate genes involved in the blue to orange transition had significant differential expression (F3'5'H and two copies of DFR). F3'5'H and DFR-1 show 2.5x and 2x B>O, respectively, while DFR-2 has >600x O>B (this locus is nearly undetectable in blue flowers). In regard to the regulatory genes, we found 13 trinity genes from four different gene families (Supplemental Figure 3; Supplemental Table 6). Only MYB61 (MYB4-2) showed significant differential expression with 26x higher expression in orangethan in blue-flowered plants (Figure 1; Table 1). This MYB was initially identified as a MYB4 from Hibiscus syriacus by Kakapo, but a BLASTn of this L. arvensis sequence showed that it is more similar to Camelia sinensis MYB61, and will be referred to as LaMYB61 from here onwards. Figure 1. Clustering and heatmap of differentially expressed genes (DEGs) from the flavonoid biosynthesis pathway in blue- and orange-flowered plants of L. arvensis. The color scale represents the scaled TMM expression values (trimmed mean of M- values). Green represents high expression and purple represents low expression. Capítulo 3 128 Table 2. Comparison of amino acid (AA) variation in the F3'5'H non-synonymous SNP sites of the coding sequences between blue and orange L. arvensis isotigs. Site numbers refer to base pairs (bp). In bold are indicated the AA changes between blue and orange isotigs (blue AAs are used as reference). *Not including blue sample L19, which is closely related to orange samples. Since there are well characterized functional domains in DFR that provide substrate specificity conferring blue-to-red shifts in other species (Johnson et al., 2001; Petit et al., 2007; Des Marais and Rausher, 2008; Smith et al., 2012), we compared DFR-1 and DFR- 2 for fixed amino acid differences at these sites. For the 26 residue substrate specificity region identified by Johnson et al. (2001), (amino acids 133 to 158 in our alignment), the L. arvensis DFR-2 has fixed differences in 13 of them compared to DFR-1 (Figure 5B). For the additional 12 substrate binding sites identified by Petit et al. (2007), we found only one site (amino acid 210) where DFR-2 was fixed with a serine (S) compared to a threonine (T) in DFR-1 (BLOSUM62 score = +1; both polar uncharged amino acids). Two more color-differentiating sites identified by Smith et al. (2012) were fixed in both DFRs (Figure 5B). In addition to these five amino acid substitutions in known functional domains, we identified 63 more amino acid changes outside known functional domains that were fixed between DFR-1 and DFR-2 and another six that were variable in one copy of DFR, yet fixed in the other. Isotig Site 6 Site 222 Site 472 Site 1426 Blue* i1 E K N/H K/E i2 E K N/H K/E i3 E K N/H K/E i4 E K N/H K/E i9 E K N/H K/E i10 E K N/H K/E Orange i1 E K N E i2 E K N E i3 E K N E i4 D N N E i9 E K N E i10 E K N E BLOSUM 62 score +2 0 +1 +1 Biochemistry both negatively charged positive (K) to polar uncharged (N) positive (H) to polar uncharged (N) positive (K) to negative (E) Capítulo 3 129 Figure 5. Non-synonymous SNP analysis of F3’5’H and DFR for L. arvensis. SNP frequency graphs are aligned with the gene diagram indicating known functional domains (grey). For F3’5’H (A) there are five binding regions (SRS; Seitz et al., 2007) and other known functional domains (I-IV; Guo et al., 2019). The DFR analysis (B) shows non-synonymous SNP frequencies in DFR-1 in the graph. Gene diagrams show the substrate specificity region (Johnson et al., 2001) and other previously identified substrate binding sites (Petit et al., 2007; Smith et al., 2012). Amino acid changes between blue and orange DFR-1 located in a substrate binding region are marked with *. Amino acid changes between DFR-1 and DFR-2 that may correlate with flower color variation in L. arvensis (Petit et al., 2007; Smith et al., 2012) are indicated in the diagram (pink) and their location and single letter abbreviation are shown between the DFR diagrams. An alignment of DFR-1 and DFR-2 consensus sequences of the 26 amino acids from the substrate specificity region is provided, highlighting all the amino acid changes in a black box (sites 133-158; relevant amino acid changes shown in the gene diagram are highlighted in pink). Since DFR-2 is only expressed at considerable levels in orange flowers, we focus our color-differentiating SNP analysis on DFR-1 where we have sequences of both blue- and orange-flowered individuals. Over all 54 DFR-1 SNPs (synonymous and nonsynonymous), the frequency differences between blue- and orange-flowered samples (across isotigs) ranged from 0.12 to 1 (mean 0.46). There were ten non-synonymous SNPs with considerable differences in frequency between flower colors (Table 3). The nonsynonymous SNPs located in the binding specificity region (bp434 and bp474; Figure 5B) showed clear patterns of color frequency differences between the two flower color types (bp434 = 0.50 SNP frequency difference; bp474 = 0.52 SNP frequency difference; Capítulo 3 130 Supplemental Table 8). Both have a BLOSUM62 score of +1 and are amino acid substitutions within the same general biochemical category (bp434 = both polar uncharged amino acids; bp474 = both hydrophobic amino acids). Three non-synonymous SNPs in the 3' half of the DFR-1 coding sequence that (only present in i2 and i10) were completely differentiated between blue- and orange-flowered samples (bp859, bp863 and bp959). Their BLOSUM62 scores range from 0 to +2 and two of them represent biochemically distinct amino acid replacements (bp859 and bp913 = negatively charged to polar uncharged amino acids; Table 3). There were 44 additional synonymous SNPs in the DFR-1 CDS (Supplemental Table 8). Identification of flavonoid compounds by UHPLC-MS UHPLC-MS analysis confirmed that malvidin-3-rhamnoside is the main flavonoid that accumulates in the blue petals of L. arvensis (relative content = 98.4%; Table 4; Figure 6). In addition, blue petals also contain traces of the aglycone malvidin and delphinidin derivatives. In contrast, orange petals contain four derivatives of pelargonidin with different glycosylation patterns: pelargonidin-3-glucoside (60.1% relative content), pelargonidin-3-glucoside-glucuronide (26.5%), pelargonidin-3-rhamnoside (11.6%), and pelargonidin-3-diglucoside (trace amounts; Table 4; Figure 6). We also detected traces of the flavone luteolin-7-glucoside in the blue samples, and small amounts of two flavonols present in some orange samples - syringetin (trace) and kaempferol-3-glucoside (1.7%). The putative locations of these compounds in the flavonoid biosynthetic pathway are shown in Figure 7. Figure 6. UHPLC-MS chromatograms of the main anthocyanins present in blue (above) and orange (below) petals of L. arvensis detected at 520nm. 131 Table 3. Comparison of amino acid (AA) variation in the DFR_1 non-synonymous SNP sites of the coding sequences between blue and orange L. arvensis isotigs. Site numbers refer to base pairs (bp). In bold are indicated the AA changes between blue and orange isotigs (blue AAs are used as reference). *Not including blue sample L19, which is closely related to orange samples. Isotig Site 14 Site 97 Site 358 Site 434 Site 474 Site 743 Site 859 Site 863 Site 913 Site 959 Blue* i10 I/T T K/G N/S I/M E E S E N i11 T T E N I - - - - i2 I T K/G N/S I/M E E S E N i7 I T T N M - - - - Orange i10 T T K S I A Q N Q T i11 T T K N I - - - i2 I S/T K S I E Q N Q/E T i7 I S/T T/I N I - - - - - i8 I S/T K S I - - - - - BLOSUM 62 score -1 +1 NA +1 +1 -1 +2 +1 +2 0 Biochemistry hydroph obic (I) to polar uncharge d (T) both polar uncharge d NA both polar uncharge d both hydrophoa bic hydroph obic (A) to negative charged (E) negatively charged (E) to polar uncharged (Q) both polar uncharge d negatively charged (E) to polar uncharged (Q) both polar uncharge d Capítulo 3 132 Table 4. Putative flavonoid identification in blue and orange petals of L. arvensis from the UHPLC–MS biochemical analysis (MS analysis was acquired in positive mode). Petal colour Flavonoid type Putative flavonoid identification N samples* tR (min)  max (nm) Mass (m/z) Relative content (%)** Blue Anthocycanin Delphinidin 3-glucoside-rhamnoside 1/14 2.3 233, 302, 522 609.90 trace (14 plants) Anthocycanin Delphinidin 3-rhamnoside-rhamnoside 1/14 3.3 235, 310, 522 592.45 trace Anthocycanin Malvidin 3-rhamnoside 14/14 5.8 227, 280, 527 476.71 98.4 Anthocycanin Malvidin 1/14 7.0 231, 268, 350, 533 331.28 1.6 Flavone Luteolin 7-glucoside 2/14 7.9 230, 290, 340, 377 448.48 trace Orange Anthocycanin Pelargonidin 3-glucoside-glucuronide 4/11 2.2 279, 427, 502 606.66 26.5 (11 plants) Anthocycanin Pelargonidin 3-diglucoside 2/11 2.8 239, 280, 428, 500 594.72 trace Anthocycanin Pelargonidin 3-glucoside 11/11 3.0 283, 427, 502 433.09 60.1 Anthocycanin Pelargonidin 3-rhamnoside 6/11 5.2 232, 268, 483, 506 416.72 11.6 Flavonol Syringetin 3/11 6.9 232, 264, 294, 346 346.64 trace Flavonol Kaempferol 3-glucoside 5/11 7.9 232, 264, 294, 346 448.75 1.7 Petals do not include the darker basal area (i.e. bullseye). Identification was based on retention time (tR), maximum absorbance wavelength (λmax) observed in the UV-Vis spectra of the compound that eluted at the tR indicated, and molecular mass comparisons with standards and values of previously reported flavonoids for Lysimachia spp. (see Supplemental Table 3 and the database http://metabolomics.jp). *The number of samples where each flavonoid was observed. **Relative content of each putative compound in each flower color, calculated from the areas of the UV-Vis peaks. 133 Discussion We found 38 genes with significant differential expression between flower colors of L. arvensis across 94 genes in the flavonoid biosynthetic pathway. Our two sampled populations are geographically separated and may represent incipient species based on recent phylogenetic results (Jiménez-López, 2019; Jiménez-López et al., in review), which may have introduced some differences between the populations not correlated with color. Nevertheless, we still detected patterns correlated with color that are consistent with known functions of the ABP loci and previous studies of comparable flower color transitions. First, we focus on the ABP genes including discussion about the expression differences, SNP detection and biochemical changes associated with two enzymes at a critical branchpoint in the ABP in order to shed light on the molecular underpinnings of the blue to orange shift in L. arvensis. Then, we expand the conversation to include additional genes in the flavonoid biosynthetic pathway not previously associated with such a color shift and discuss potential pleiotropic consequences not directly related to pollinator attraction. Finally, we mount evidence for a directional transition from blue to orange flowers in L. arvensis. Within the narrow ABP, there were two DEGs in L. arvensis concentrated in the latter half of the pathway that are likely directly involved in the blue to orange shift in flower color (F3'5'H and DFR). These genes had the highest overall expression of all the narrowly defined ABP genes. This flower color transition could be caused by either of these or some combination of the two since these genes are responsible for the synthesis of two different types of anthocyanins: delphinidin-derivatives (in the case of L. arvensis this is malvidin that produces the blue color) and pelargonidin-derivatives (orange; Figure 7; Smith and Rausher, 2011; Smith et al., 2012; Wheeler and Smith, 2019). These biochemical results are consistent with previous findings in L. arvensis (Lawrence et al., 1939; Harbone, 1968; Ishikura, 1981; Kawashty et al., 1998). Several additional ABP genes with differential expression in blueversus orange-flowered plants were consistent with the post-translational modifications of these principle pigments. Here we discuss three of these genes (Figure 2). BZ1-2 showed high levels of expression in orange petals (TMM >290) compared to blue (expression 950x O>B). This enzyme introduces a glucose group to the 3-position of the anthocyanin precursor's C-ring to stabilize the color (Koes et al., 2005). The high expression of BZ1-2 in orange-flowered L. arvensis plants Capítulo 3 134 correlates with the biochemical results showing that orange petals are composed of pelargonidin 3-glucoside derivatives whereas blue petals are malvidin glycosylated with rhamnoside. Similarly, enzymes OMT and 3GT showed higher levels of expression in blue than orange samples (5.3x and 2.9x, respectively). OMT methylates delphinidin to form malvidin, and 3GT introduces the rhamnoside (Saito et al., 2013) both of which are unique to blue flowered samples, likely involved in stabilization and protection of the pigment, but unlikely solely responsible for the color change. Figure 7. Tentative scheme of the anthocyanin biosynthetic pathway for L. arvensis petals. Structural genes involved in the broader flavonoid biosynthetic pathway are shown in green: Phenylalanine ammonia-lyase (PAL), Cinnamic 4-hydroxylase (C4H), 4-coumarate-CoA ligase (4CL), Chalcone synthase (CHS), Chalcone isomerase (CHI), Flavanone 3-hydroxylase (F3H), Flavonoid 3',5'-hydroxylase (F3'5'H), Dihydroflavonol 4-reductase (DFR), Anthocyanin synthase (ANS), Anthocyanidin-3-o-glucosyltransferase (BZ1), Anthocyanin-3'-o-beta-glucosyltransferase (3GT), O-methyltransferase (OMT), Flavonol synthase (FLS), Flavonol-3-o-glucosyltransferase (Fl. 3-o-glucosyl.), Flavone synthase I (FNSI) and Flavonol 3-O- Capítulo 3 135 methyltransferase (Fl. 3-O-methyl.). Narrow ABP genes differentially expressed in orange and blue L. arvensis are marked with red and blue circles, respectively. Blue and orange boxes mark the anthocyanin compounds found by UHPLC-MS in each flower color. Blue and orange dashed lines mark the nonanthocyanin compounds found by UHPLC-MS in each flower color. The grey portion indicates the location of the inactivated branch leading to cyanidin-derivatives that would exist if F3’H outcompeted DFR-2 for dihydrokaempferol. We hypothesize that DFR-2 outcompetes F3’H at this step suggested by the black “<<<” symbols. Role of F3'5'H expression and amino acid substitutions F3'5'H catalyzes the hydroxylation of dihydrokaempferol to produce a delphinidin precursor (Halbwirth, 2010). Our expression results indicate a 2.5x decrease in expression in orange petals consistent with several previous studies investigating the blue- to orangeflowered shift among a diversity of eudicots (Rausher, 2008; Wessinger and Rausher, 2014; Larter et al., 2018). For example, the loss of function of F3'5'H in Iochroma spp. (Smith and Rausher, 2011), Antirrhinum spp. (Ishiguro et al., 2012) and Penstemon barbatus (Wessinger and Rausher, 2013) or the decreased expression of F3'5'H in Phlox drummondii (Hopkins and Rausher, 2011), confers blue to red color shifts. Rausher (2008) suggests that unidirectional changes at this key step in the ABP play an important role in the repeated transitions from blue to red/orange flowers observed in many angiosperm lineages. In L. arvensis, F3'5'H activity is not completely lost since F3'5'H isotigs were detected (at low levels) in orange petals and we detected traces of a flavonol requiring F3'5'H activity (syringetin) in almost 30% of the orange petal samples. In addition to the decreased expression of F3'5'H, there are four flower color differentiating non-synonymous SNPs that may contribute to the blue to orange shift, but would require future enzyme activity assays to be determined. Role of DFR in the blue to orange transition The transition from delphinidin-like precursors (malvidin in this case) to pelargonidin precursors caused by changes in F3'5'H, creates a new substrate which requires subsequent modifications in downstream enzymes like DFR. Several lines of evidence suggest that DFR in L. arvensis, which catalyzes the reduction of dihydroflavonols to leucoanthocyanins, may also be involved in the evolutionary transition from blue to orange flowers. Our transcriptome assembly identified two distinct copies of DFR. One copy, DFR-2, is expressed primarily in orange flowered individuals (nearly undetectable in blue samples). This copy presents an N residue in the third position of the substrate specificity region (residue 135 in L. arvensis), which is known to confer substrate affinity Capítulo 3 136 to dihydrokaempferol, the pelargonidin-like precursor, in a wide group of angiosperms (Johnson et al. 2001; Shimada et al., 2005). Alternatively, DFR-1 is primarily expressed in blue-flowered individuals and exhibits a D residue at this site, which confers less or no substrate affinity for dihydrokaempferol in other species (Johnson et al., 2001; Shimada et al., 2005). We propose that DFR-1 in L. arvensis is likely more efficient in processing dihydromyricetin, the delphinidin-like precursors to malvidin. Like F3'5'H, a combination of expression and changes in the amino acid composition of DFR likely contributes to the completion of the blue to orange color transition. In Petunia, alteration of the amino acids in the substrate specificity domain of DFR confers a flower color shift by increasing catalytic efficiency in metabolizing dihydrokampferol, a pelargonidin precursor (Johnson et al., 2001). In addition, three more amino acid changes between DFR-1 and DFR-2 have been suggested to be involved in substrate specificity differentiation in other species. In Iochroma and Vitis (Petit et al., 2007; Smith et al., 2012), the number of sites responsible for substrate specificity is larger and more physically spread out along the DFR coding region than the 26 amino acid stretch suggested by Johnson et al. (2001). All these amino acid changes between the two copies of DFR present in L. arvensis are correlated with blue to orange flower color transition, yet causality would require further genetic manipulation. Within the DFR-1 coding sequence of L. arvensis, there are many color differentiating SNPs. There are two amino acid changes nearly fixed between colors that also differ from DFR-2, and are located in the 26 amino acid substrate specificity region, suggesting interesting residue positions for future studies of differential substrate binding affinity. Overall, we suggest that DFR-1 may have had an original affinity for dihydromyricetin, which is maintained in blue flowers, while in orange-flowered individuals, a transregulatory change to DFR-2 expression (e.g., via LaMYB61 O>B) may confer dihydrokaempferol specificity to accommodate the new substrate being produced by the loss of F3'5'H activity. If DFR-2 outcompetes F3'H for dihydrokaempferol as described in other angiosperms producing pelargonidin (Des Marais and Rausher 2008), then the flux would be directed to the pelargonidin pathway (even if F3'H enzymes are present as we have relatively equal expression in blue and orange-flowered samples for F3'H). In order to produce orange flowers, there must be no (or very low) expression of DFR-2 in blue L. arvensis to allow F3'5'H to win the competition for dihydrokaempferol redirecting the flux down the delphinidin branch of the ABP. Capítulo 3 137 The maintenance and detectable expression of both DFR-1 and DFR-2 in L. arvensis is somewhat enigmatic. A similar situation exists in the Convolvulaceae with three DFR copies of relatively recent origin in that family (Des Marias and Rausher, 2008). However, the duplication of DFR-1 and DFR-2 we report is likely independent of the gene duplications reported by Des Marias and Rausher (2008), because it appears to be much older, perhaps predating the split of Asterales and Ericales (see annotation in Figure 4). Next, we plan to investigate the petal transcriptomes of the closely related L. monelli, which exhibits phenotypically similar petal color polymorphism. We expect to find two copies of DFR in that case, as well, and will compare its color-specific expression differences to those we report for L. arvensis herein. Broader insights from the flavonoid biosynthetic pathway & the roles of nonpollinator agents of selection We hypothesized that there would be some flower-color specific differences in the transcriptome for genes outside the narrow ABP that might open the door for agents of selection not associated with pollinator attraction (Strauss and Whittall, 2006; Rausher, 2008; Ortiz-Barrientos et al., 2013). In this regard, we found 30 DEGs in the broader flavonoid pathway, outside the narrow ABP (15 DEGs with B>O with an average foldchange of 9.1 and 15 DEGs with O>B with an average fold-change of 18.8). Among these non-ABP DEGs, we looked for loci with potential functionality that may correlate with the ecological differences of the two color types (Arista et al. 2013). For example, Caffeoyl CoA O-methyl (only present in blue samples with TMM=1.27) and Caffeoyl 3- O-methyl (ATTSM1-1 and ATTSM1-2, with B>O and an average fold-change of 4.8 and 3.7, respectively) both promote anthocyanin methylation following drought stress in Vitis vinifera (Giordano et al., 2016) and have been implicated in lignin biosynthesis, an important barrier against pathogens (Liu et al., 2018). Additionally, we found differential expression for three CYP84A genes with O>B fold-changes ranging from 3.3 to 28.1. These loci are involved in sinapate ester biosynthesis, a key flavonoid compound providing UV protection (Ruegger, 1999). Finally, in defense against herbivores and pathogens, we found differential expression of two IEMT (with O>B average fold-change of 3.3 and 186.8) involved in the synthesis of isomethyleugenol and methyleugenol for deterring herbivores (and/or attracting pollinators; Barckman, 2003) and two PKR (also known as CHR; one with B>O fold-change of 30.4, and the other with O>B fold-change Capítulo 3 144 Fan, X., Fan, B., Wang, Y., and Yang, W. (2016). Anthocyanin accumulation enhanced in Lc-transgenic cotton under light and increased resistance to bollworm. Plant Biotechnol. Rep. 10(1), 1-11. Feng, C., Ding, D., Feng, C., and Kang, M. (2020). The identification of an R2R3-MYB transcription factor involved in regulating anthocyanin biosynthesis in Primulina swinglei flowers. Gene 144788. Fenster, C. B., Armbruster, W. S., Wilson, P., Dudash, M. R., and Thomson, J. D. (2004). Pollination syndromes and floral specialization. Annu. Rev. Ecol. Evol. Syst. 35(1), 375- 403. Ferguson, L. F. (1972) Anagallis L. in Flora Europaea, Vol. 3 (eds T.G. Tutin, V.H. Heywood, N.A. Burges, D.H. Valentine, S.M. Walters & D.A. Webb), pp. 28-29. Cambridge University Press, Cambridge, UK. Freyre, R., and Griesbach R. J. (2004). Inheritance of flower color in Anagallis monelli L. HortScience 39(6), 1220-1223. Freyre, R., Wilson, S. B., Knox, G. W., Uzdevenes, C., Gu, L., and Quesenberry, K. H. (2015). Breeding and genetic studies of Ruellia simplex (Mexican petunia). Acta Hortic. 1087, 113-120. Gao, J., Peng, H., Chen, F., Luo, M., and Li, W. (2019). Genome-wide analysis of transcription factors related to anthocyanin biosynthesis in carmine radish (Raphanus sativus L.) fleshy roots. PeerJ 7, e8041. Gao, L., Yang, H., Liu, H., Yang, J., and Hu, Y. (2016). Extensive transcriptome changes underlying the flower color intensity variation in Paeonia ostii. Front. Plant Sci. 6, 1205. Gates, D. J., Olson, B. J. S. C., Clemente, T. E., and Smith, S. D. (2017). A novel R3 MYB transcriptional repressor associated with the loss of floral pigmentation in Iochroma. New Phytol. 217(3), 1346-1356. Gehrig, H. H., Winter, K., Cushman, J., Borland, A., and Taybi, T. (2000). An improved RNA isolation method for succulent plant species rich in polyphenols and polysaccharides. Plant Mol. Biol. Rep. 18(4), 369-376. Gibbs, P. E., and Talavera S. (2001). Breeding system studies with three species of Anagallis (Primulaceae): Self-incompatibility and reduced female fertility in A. monelli L. Ann. Bot. 88(1), 139-144. Gigord, L. D. B., Macnair, M. R., and Smithson, A. (2001). Negative frequencydependent selection maintains a dramatic flower color polymorphism in the rewardless orchid Dactylorhiza sambucina (L.) Soo. Proc. Natl. Acad. Sci. U.S.A. 98(11), 6253-6255. Giordano, D., Provenzano, S., Ferrandino, A., Vitali, M., Pagliarani, C., Roman, F., et al. (2016). Characterization of a multifunctional caffeoyl-CoA O -methyltransferase activated in grape berries upon drought stress. Plant. Physiol. Biochem. 101, 23-32. Capítulo 3 145 Gómez, J. M., Perfectti, F., Armas, C., Narbona, E., González-Megías, A., Navarro, L., et al. (2020). Within-individual phenotypic plasticity in flowers fosters pollination niche shift. Nat. Commun. 11, 4019. Grabherr, M. G., Haas, B. J., Yassour, M., Levin, J. Z., Thompson, D. A., Amit, I., et al. (2011). Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat. biotechnol. 29, 644-652. Guo, L., Gao, L., Ma, X., Guo, F., Ruan, H., Bao, Y., et al. (2019). Functional analysis of flavonoid 3′-hydroxylase and flavonoid 3′,5′-hydroxylases from tea plant (Camellia sinensis), involved in the B-ring hydroxylation of flavonoids. Gene 717, 144046. Haas, B., Papanicolaou, A., Yassour, M., Grabherr, M., Blood, P. D., Bowden, J., et al. (2013). De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat. Protoc. 8, 1494-1512. Halbwirth, H. (2010). The creation and physiological relevance of divergent hydroxylation patterns in the flavonoid pathway. Int. J. Mol. Sci. 11, 595-621. Harborne, J. B. (1968). Comparative biochemistry of the flavonoids. VII. Correlations between flavonoid pigmentation and systematics in the family Primulaceae. Phytochemistry 7, 1215-1230. Henikoff, S., and Henikoff, J. G. (1992). Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci. U.S.A. 89(22), 10915-10919. Hichri, I., Barrieu, F., Bogs, J., Kappel, C., Delrot, S., and Lauvergeat, V. (2011). Recent advances in the transcriptional regulation of the flavonoid biosynthetic pathway. J. Exp. Bot. 62(8), 2465-83. Ho, W. W., and Smith, S. D. (2016). Molecular evolution of anthocyanin pigmentation genes following losses of flower color. BMC Evol. Biol. 16, 98. Holton, T. A., and Cornish E. C. (1995). Genetics and biochemistry of anthocyanin biosynthesis. Plant Cell 7(7), 1071-1083. Hopkins, R., and Rausher, M. D. (2011). Identification of two genes causing reinforcement in the texas wildflower Phlox drummondii. Nature 469(7330), 411-414. Huxley, J. S. (1955). Morphism and evolution. Heredity 9, 1-51. Irwin, R. E., and Strauss, S. Y. (2005). Flower color microevolution in wild radish: evolutionary response to pollinator‐mediated selection. Am. Nat. 165(2), 225-237. Ishiguro, K., Taniguchi, M. and Tanaka, Y. (2012). Functional analysis of Antirrhinum kelloggii flavonoid 3'-hydroxylase and flavonoid 3',5'-hydroxylase genes; critical role in flower color and evolution in the genus Antirrhinum. J. Plant Res. 125(3), 451-456. Ishikura, N. (1981). Flavonoids in the petal cells of Anagallis arvensis f. coerulea containing a blue crystalline anthocyanin. Z. Pflanzenphysiol. 103(5), 469-473. Capítulo 3 146 Jiang, N., Doseff, A., and Grotewold, E. (2016). Flavones: From Biosynthesis to Health Benefits. Plants 5(2), 27. Jiménez-López F. J. (2019). Polimorfismo en el color floral y especiación en Lysimachia arvensis. Ph.D. dissertation, University of Seville, Spain. Jiménez-López, F. J., Matas, L., Arista, M., and Ortiz, P. L. (2019). Flower colour segregation and flower discrimination under the bee vision model in the polymorphic Lysimachia arvensis. Plant Biosyst. 154(4), 535-543. Jiménez-López, F. J., Ortiz, P. L., Talavera, M., Pannell, J. R., and Arista, M. (2020). The role of lateral and vertical herkogamy in the divergence of the blue- and red-flowered lineages of Lysimachia arvensis. Ann. Bot. 125(7), 1127-1135. Jiménez-López, F. J., Viruel, J., Arista, M., Ortiz, P. L., and Talavera, M. (2022). Molecular approaches reveal speciation between red and blue flowered plants in the Mediterranean species of Lysimachia arvensis and Lysimachia monelli (Primulaceae). Bot. J. Linn. Soc. Johnson, E. T., Ryu, S., Yi, H., Shin, B., Cheong, H., and Choi, G. (2001). Alteration of a single amino acid changes the substrate specificity of dihydroflavonol 4-reductase. Plant J. 25(3), 325–333. Kähkönen, M. P., and Heinonen, M. (2003). Antioxidant activity of anthocyanins and their aglycons. J. Agric. Food Chem. 51(3), 628-633. Kawashty, S., El-Garf, I., and El-Negoumy, S. (1998). Chemosystematics of Anagallis arvensis L. (Primulaceae). Biochem. Syst. Ecol. 26(6), 663-668. Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., et al. (2012). Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28(12), 1647-1649. Kellenberger, R. T., Byers, K. J. R. P., De Brito Francisco, R. M., Staedler, Y. M., LaFountain, A. M., Schönenberger, et al. (2019). Emergence of a floral colour polymorphism by pollinator-mediated overdominance. Nat. Commun. 10(1). Koes, R., Verweij, W., and Quattrocchio, F. (2005). Flavonoids: a colorful model for the regulation and evolution of biochemical pathways. Trends Plant Sci. 10(5), 236-242. Koski, M. H., MacQueen, D., and Ashman, T.-L. (2020). Floral pigmentation has responded rapidly to global change in ozone and temperature. Cur. Biol. 30, 4425-4431. Lai, Y., Li, H., and Yamagishi, M. (2013). A review of target gene specificity of flavonoid R2R3-MYB transcription factors and a discussion of factors contributing to the target gene selectivity. Front. Biol. 8(6), 577-598. Lang, X., Li, N., Li, L., and Zhang, S. (2019). Integrated Metabolome and Transcriptome Analysis Uncovers the Role of Anthocyanin Metabolism in Michelia maudiae. Int. J. Genomics 2019, 1-14, 4393905. Capítulo 3 147 Langmead, B., and Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2. Nat. Methods 9(4), 357-359. Larter, M., Dunbar-Wallis, A., Berardi, A. E., and Smith, S. D. (2018). Convergent evolution at the pathway level: predictable regulatory changes during flower color transitions. Mol. Biol. Evol. 35(9), 2159-2169. Lawrence, W. J. C., Price, J. R., Robinson, G. M., and Robinson, R. (1939). The distribution of anthocyanins in flowers, fruits and leaves. Phil. Trans. Roy. Soc. Lond. B 230(567), 149-178. Le Maitre, N. C., Pirie, M. D., and Bellstedt, D. U. (2019). An approach to determining anthocyanin synthesis enzyme gene expression in an evolutionary context: an example from Erica plukenetii. Ann. Bot. 124(1), 121-130. Lee, W. J., Jeong, C. Y., Kwon, J., Van Kien, V., Lee, D., Hong, S. W., et al. (2016). Drastic anthocyanin increase in response to PAP1 overexpression in fls1 knockout mutant confers enhanced osmotic stress tolerance in Arabidopsis thaliana. Plant Cell Rep. 35(11), 2369-2379. Lepiniec, L., Debeaujon, I., Routaboul, J.-M., Baudry, A., Pourcel, L., Nesi, N., et al. (2006). Genetics and biochemistry of seed flavonoids. Annu. Rev. Plant Biol. 57(1), 405– 430. Li, B., and Dewey C.N. (2011). RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics, 12, 323. Li, B., Fan, R., Guo, S., Wang, P., Zhu, X., Fan, Y., et al. (2019b). The Arabidopsis MYB transcription factor, MYB111 modulates salt responses by regulating flavonoid biosynthesis. Environ. Exp. Bot. 166, 103807. Li, L., Zhai, Y., Luo, X., Zhang, Y., and Shi, Q.Q. (2019a). Comparative transcriptome analyses reveal genes related to pigmentation in the petals of red and white Primula vulgaris cultivars. Physiol. Mol. Biol. Plants 25, 1029-1041. Li, S. (2014). Transcriptional control of flavonoid biosynthesis. Plant Signal Behav. 9(1), e27522. Liang, G., He, H., Li, Y., Ai, Q., and Yu, D. (2014). MYB82 functions in regulation of trichome development in Arabidopsis. J. Exp. Bot. 65(12), 3215-3223. Liu, Q., Luo, L., and Zheng, L. (2018). Lignins: biosynthesis and biological functions in plants. Int. J. Mol. Sci. 19(2), 335. Lou, Q., Liu, Y., Qi, Y., Jiao, S., Tian, F., Jiang, L., et al. (2014). Transcriptome sequencing and metabolite analysis reveals the role of delphinidin metabolism in flower colour in grape hyacinth. J. Exp. Bot. 65(12), 3157-3164. Lulin, H., Xiao, Y., Pei, S., Wen, T., and Shangqin, H. (2012). The first Illumina-based de novo transcriptome sequencing and analysis of safflower flowers. PLoS One 7(6), e38653. Capítulo 3 148 Martin, J. A., and Wang, Z. (2011). Next-generation transcriptome assembly. Nat. Rev. Genet. 12(10), 671-682. McCarthy, E. W., Berardi, A. E., Smith, S. D., and Litt, A. (2017). Related allopolyploids display distinct floral pigment profiles and transgressive pigments. Am. J. Bot. 104(1), 92-101. Moneim, A., I. A. Atta, and A. A. Shehata. (2003). On the delimitation of Anagallis arvensis L. (Primulaceae) 1. Evidence based on micromorphological characters, palynological features and karyological studies. Pak. J. Biol. Sci. 6(1), 29-35. Morita, Y., Hoshino, A., Kikuchi, Y., Okuhara, H., Ono, E., Tanaka, Y., et al. (2005). Japanese morning glory dusky mutants displaying reddish-brown or purplish-gray flowers are deficient in a novel glycosylation enzyme for anthocyanin biosynthesis, UDP- glucose:anthocyanidin 3-O-glucoside-2''-O-glucosyltransferase, due to 4-bp insertions in. Plant J. 42(3), 353-363. Narbona, E., Wang, H., Ortiz, P. L., Arista, M., and Imbert, E. (2018). Flower colour polymorphism in the Mediterranean Basin: occurrence, maintenance and implications for speciation. Plant Biol. 20, 8-20. Ng, J., Freitas, L. B., and Smith, S. D. (2018). Stepwise evolution of floral pigmentation predicted by biochemical pathway structure. Evolution 72(12). Nguyen, N. H., Kim, J. H., Kwon, J., Jeong, C. Y., Lee, W., Lee, D., et al. (2016). Characterization of Arabidopsis thaliana FLAVONOL SYNTHASE 1 (FLS1) - overexpression plants in response to abiotic stress. Plant Physiol. Biochem. 103, 133-142. Noda, K., Glover, B. J., Linstead, P., and Martin, C. (1994). Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor. Nature 369(6482), 661-664. Ortiz, P. L., Berjano, R., Talavera, M., Rodríguez-Zayas, L., and Arista, M. (2015). Flower colour polymorphism in Lysimachia arvensis: How is the red morph maintained in Mediterranean environments? Perspect. Plant Ecol. Evol. Syst. 17(2), 142-150. Ortiz, P. L., Berjano, R., Talavera, M., Rodríguez-Zayas, L., and Arista, M. (2015). Flower colour polymorphism in Lysimachia arvensis: How is the red morph maintained in Mediterranean environments? Perspect. Plant Ecol. Evol. Syst. 17(2), 142-150. Ortiz-Barrientos, D. (2013). The color genes of speciation in plants. Genetics 194(1), 39- 42. Petit, P., Granier, T., d' Estaintot, B. L., Manigand, C., Bathany, K., Schmitter, J.M., et al. (2007). Crystal structure of grape dihydroflavonol 4-reductase, a key enzyme in flavonoid biosynthesis. J. Mol. Biol. 368(5), 1345-1357. Price, J. R., and Sturgess, V. C. (1938). A survey of anthocyanins. VI. Biochem. J. 32(9), 1658-1660. Pujadas A. (1997). "Anagallis" in Flora Ibérica Vol. V.Eds. S. Castroviejo, S. Aedo, M. Laínz, R. Morales, F. M. Garmendia, G. N. Feliner, et al. (Madrid, Spain: C.S.I.C). Capítulo 3 149 Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al. (2013). The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41(D1), D590–D596. Quattrocchio, F., Baudry, A., Lepiniec, L., and Grotewold, E. (2006). "The regulation of flavonoid biosynthesis" in The Science of Flavonoids, ed. E. Grotewold (New York: Springer-Verlag), 97-122. R Core Team (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: https://www.R- project.org/ (Accessed October 20, 2020) Rausher, M. D., and Simms, E. L. (1989). The evolution of resistance to herbivory in Ipomoea purpurea. I. Attempts to detect selection. Evolution 43(3), 563-572. Rausher, M. D. (2008). Evolutionary transitions in floral color. Int. J. Plant Sci. 169(1), 7-21. Reiseberg, L. H., and Blackman, B. K. (2010). Speciation genes in plants. Ann. Bot. 106(3), 439-455. Roberts, W. R., and Roalson, E. H. (2017). Comparative transcriptome analyses of flower development in four species of Achimenes (Gesneriaceae). BMC Genomics 18, 240. Robinson, M. D., McCarthy, D. J., and Smyth, G. K. (2009). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26(1), 139-140. Ruegger, M., Meyer, K., Cusumano, J. C., and Chapple, C. (1999). Regulation of Ferulate-5-hydroxylase expression in Arabidopsis in the context of sinapate ester biosynthesis. Plant Physiol. 119(1), 101-110. Robinson, M. D., and Oshlack, A. (2010). A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol. 11, R25. Schemske, D. W., and Bierzychudek, P. (2007). Spatial differentiation for flower color in the desert annual Linanthus parryae: was wright right? Evolution 61(11), 2528-2543. Seitz, C., Ameres, S., and Forkmann, G. (2007). Identification of the molecular basis for the functional difference between flavonoid 3′-hydroxylase and flavonoid 3′,5′- hydroxylase. FEBS Lett. 581(18), 3429-3434. Shimada, N., Sasaki, R., Sato, S., Kaneko, T., Tabata, S., Aoki, T., et al. (2005). A comprehensive analysis of six dihydroflavonol 4-reductases encoded by a gene cluster of the Lotus japonicus genome. J. Exp. Bot. 56(419), 2573-2585. Smith, S. D., and Rausher, M. D. (2011). Gene loss and parallel evolution contribute to species difference in flower color. Mol. Biol. Evol. 28(10), 2799-2810. Smith, S. D., Wang, S., and Rausher, M. D. (2012). Functional evolution of an anthocyanin pathway enzyme during a flower color transition. Mol. Biol. Evol. 30(3), 602-612. Capítulo 3 150 Sobel, J. M., and Streisfeld, M. A. (2013). Flower color as a model system for studies of plant evo-devo. Front. Plant Sci. 4, 321. Song, L., and Florea, L. (2015). Rcorrector: efficient and accurate error correction for Illumina RNA-seq reads. GigaScience 4(1), 48. Stracke, R., Werber, M., and Weisshaar, B. (2001). The R2R3-MYB gene family in Arabidopsis thaliana. Curr. Opin. Plant Biol. 4(5), 447–456. Strauss, S., and Whittall, J.B. (2006) "Non-pollinator agents of selection on floral traits" in Ecology and Evolution of Flowers. Eds. L. D. Harder, and S. C. H. Barrett (Oxford, UK: Oxford University Press), 120-138. Streisfeld, M. A., and Rausher, M. D. (2009). Genetic changes contributing to the parallel evolution of red floral pigmentation among ipomoea species. New Phytol. 183(3), 751- 763. Streisfeld, M. A., and Rausher, M. D. (2011). Population genetics, pleiotropy, and the preferential fixation of mutations during adaptive evolution. Evolution 65(3), 629-642. Tanaka, Y., Sasaki, N., and Ohmiya, A. (2008). Biosynthesis of plant pigments: anthocyanins, betalains and carotenoids. Plant J. 54(4), 733-749. Tripp, E. A., and Manos, P. S. (2008). Is floral specialization an evolutionary dead-end pollination system transitions in Ruellia (Acanthaceae). Evolution 62(7), 1712-1737. Tsuchiya, Y., Nambara, E., Naito, S., and McCourt, P. (2004). The FUS3 transcription factor functions through the epidermal regulator TTG1 during embryogenesis in Arabidopsis. Plant J. 37(1), 73-81. Van der Kooi, C. J., Kevan, P. G., and Koski, M. H. (2019). The thermal ecology of flowers. Ann. Bot. 124(3), 343-353. Van der Niet, T., Peakall, R. and Johnson, S.D. (2014). Pollinator-driven ecological speciation in plants: new evidence and future perspectives. Ann. Bot. 113(2), 199-212. Wang, Z., Gerstein, M., and Snyder, M. (2009). RNA-Seq: a revolutionary tool for transcriptomics. Nat. Rev. Genet. 10(1), 57-63. Wessinger, C. A., and Rausher, M. D. (2012). Lessons from flower colour evolution on targets of selection. J. Exp. Bot. 63(16), 5741-5749. Wessinger, C. A., and Rausher, M. D. (2013). Predictability and irreversibility of genetic changes associated with flower color evolution in Penstemon barbatus. Evolution 68(4), 1058-1070. Wessinger, C. A., and Rausher, M. D. (2014). Ecological transition predictably associated with gene degeneration. Mol. Biol. Evol. 32(2), 347-354. Wheeler, L. C., and Smith, S. D. (2019). Computational modeling of anthocyanin pathway evolution: Biases, hotspots, and trade-offs. Integr. Comp. Biol. 59(3), 585-598. Capítulo 3 151 Widmer, A., Lexer, C. and Cozzolino, S. (2009). Evolution of reproductive isolation in plants. Heredity 102, 31-38. Winkel-Shirley, B. (2002). Biosynthesis of flavonoids and effects of stress. Curr. Opin. Plant Biol. 5(3), 218-223. Wood, D. E., Lu, J., and Langmead, B. (2019). Improved metagenomic analysis with Kraken 2. Genome Biol. 20, 257. Xu, W., Dubos, C., and Lepiniec, L. (2015). Transcriptional control of flavonoid biosynthesis by MYB–bHLH–WDR complexes. Trends Plant Sci. 20(3), 176-185. Yuan, Y. W., Byers, K. J., and Bradshaw, H. (2013). The genetic control of flower– pollinator specificity. Curr. Opin. Plant Biol. 16(4), 422-428. Zhang, Y., Zhou, T., Dai, Z., Dai, X., Li, W., Cao, M., et al. (2020). Comparative transcriptomics provides insight into floral color polymorphism in a Pleione limprichtii orchid population. Int. J. Mol. Sci. 21(1), 247. Zhao, L., Gao, L., Wang, H., Chen, X., Wang, Y., Yang, H., et al. (2013). The R2R3- MYB, bHLH, WD40, and related transcription factors in flavonoid biosynthesis. Funct. Integr. Genomics 13(1), 75-98. Zheng, X., Om, K., Stanton, K.A., Thomas, D., Cheng, P.A., Eggert, A., et al. (2020). MYB5a/NEGAN activates petal anthocyanin pigmentation and shapes the MBW regulatory network in Mimulus luteus var. variegatus. bioRxiv [Preprint]. Zufall, R. A., and Rausher, M. D. (2004). Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428(6985), 847-850. 152 153 CAPÍTULO 4 Flower color ancestral polymorphism transcends speciation Mercedes Sánchez-Cabrera, Eduardo Narbona, Montserrat Arista, Pedro L. Ortiz, Amelia A. Fuller and Justen B. Whittall (Manuscript in preparation)