Protist, Vol. 165, 177–195, March 2014 http://www.elsevier.de/protis Published online date 17 February 2014
ORIGINAL PAPER
Molecular Phylogeny of Amphora sensu lato (Bacillariophyta): An Investigation into the Monophyly and Classification of the Amphoroid Diatoms Joshua G. Stepanek 1 , and J. Patrick Kociolek Department of Ecology and Evolutionary Biology and, Museum of Natural History, University of Colorado, Boulder, CO 80309, USA Submitted September 27, 2013; Accepted February 8, 2014 Monitoring Editor: Marina Montresor
Amphora sensu lato encompasses a large group of raphid diatoms, diverse in both form and ecology. The defining feature of this group has been an extreme asymmetry of the valve mantle and girdle bands, bringing both faces of the cell onto a single plane. Although this ‘amphoroid’ structure has long been the diagnostic feature and thus considered ‘conservative’ for the group, many have argued that the diversity of forms presently assigned to Amphora likely does not represent a monophyletic group. With the exception of several taxonomic transfers and the recent elevation of Halamphora to the level of genus, much of Amphora classification has remained unchanged for over 100 years. This study presents a phylogenetic analysis of Amphora s.l. based on a concatenated molecular alignment including the nuclear marker SSU rDNA and the chloroplast markers rbcL and psbC. These results are discussed within the framework of the current classification system of Amphora and Halamphora and lay the groundwork for a taxonomic revision of the group based on monophyly. The results of this analysis demonstrate that the genus Amphora is polyphyletic and that lineages assigned to the genus are distributed widely across the raphid diatom tree of life. The feature of amphoroid symmetry appears to have evolved independently several times. We discuss the nature of conservative characters in the raphid diatoms and their usefulness as a guide to phylogenetic relationships. © 2014 Elsevier GmbH. All rights reserved. Key words: Amphora; Halamphora; phylogeny; SSU; rbcL; psbC.
Introduction Amphora Ehrenberg ex Kützing sensu lato is a large (>800 species; Fourtanier and Kociolek 2009; Levkov 2009) and morphologically diverse group of raphid diatoms found in fresh, brackish and salt water habitats from the tropics to the poles (BérardTherriault et al. 1986; Hohn and Hellerman 1966; 1
Corresponding Author; fax +1 303-735-0128 e-mail [email protected] (J.G. Stepanek).
© 2014 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.protis.2014.02.002
Patrick and Freese 1961; Stepanek and Kociolek 2013; Stoermer and Yang 1971; Wachnicka and Gaiser 2007). Partially due to this morphological diversity, the genus is primarily defined by a distinct asymmetry in the valve mantle and girdle bands. This asymmetry, which places both valve faces on the same plane, is distinct in the light microscope and has been the single feature unifying the genus since its description by Kützing in 1844 (Cleve 1895; Kützing 1844; Smith 1873). For nearly as long there have been concerns as to whether
178 J.G. Stepanek and J.P. Kociolek
the genus Amphora, as defined by this asymmetry, represented a natural group (Cleve 1895; Krammer 1980; Mann 1994) and early attempts at a comprehensive Amphora classification proved difficult (Cleve 1895; Mereschkowsky 1903). Now, over 100 years after these concerns were first raised, questions regarding the monophyly of the genus Amphora have yet to be fully resolved. In a classification system that largely persists today, Cleve (1895) separated the genus into nine subgenera (Amphora, Diplamphora, Halamphora, Oxyamphora, Amblyamphora, Psammamphora, Cymbamphora, Calamphora, and Archiamphora) based on a number of morphological characters such as a ‘simple’ or ‘complex’ connecting zone (girdle area), the presence or absence of ‘longitudinal lines’, as well as several raphe and striae features. In creating his classification, Cleve struggled with the relationships between the subgenera and their position within the larger diatom classification system. Cleve recognized frustular asymmetry as a feature shared by members of the genus, but unlike previous authors (Smith 1873) he did not believe this characteristic was appropriate for the formation of a natural group (Cleve 1895, p. 99). A notable aspect of Cleve’s treatment was his assertion that his subgenera likely represented a number of distinct and potentially unrelated genera, though he was hesitant to elevate his new subgenera because of nomenclatural concerns as well as affording an “opportunity of testing my views, which are entirely new, before admitting the proposed new genera” (Cleve 1895, p. 100). Mereschkowsky (1903) put forth the first challenge to Cleve’s classification system, in which he proposed a new classification based on plastid number and morphology. Mereschkowsky separated Amphora into four large groups based on plastid number, and within these groups the taxa were further divided into nine subgroups based on the position and shape of the chloroplast. This new classification system was not widely adopted and the genus Amphora continued to be recognized for this entire group of species by most taxonomists. With the increased use of electron microscopy came new investigations and revisions to groups and taxa within Amphora s.l. Paddock and Sims (1980) where the first to transfer taxa from the genus Amphora to other genera, transferring taxa from the subgenus Oxyamphora, to the newly created genus Undatella Paddock and Sims. Several other transfers based on scanning electron microscope (SEM) observations followed, including several taxa to the genus Biremis Mann and Cox (Mann and Cox in Round et al. 1990) and
Colliculoamphora Williams and Reid (Williams and Reid 2006), as well as the elevation of the Amphora subgenus Halamphora to generic status (Levkov 2009) and the transfer of the subgenus Cymbamphora to the newly created genus Seminavis Mann (Mann in Round et al. 1990). In addition to the morphological investigations, several molecular phylogenetic studies, while not specifically focusing on Amphora, have included Amphora s.l. in their analyses (Bruder and Medlin 2008; Medlin and Kaczmarska 2004; Ruck and Theriot 2011; Sorhannus 2004). The molecular markers used varied between studies from single marker alignments (Medlin and Kaczmarska 2004; Sorhannus 2004) to three marker concatenated alignments (Bruder and Medlin 2008; Ruck and Theriot 2011). Although these studies included few Amphora s.l. taxa (2–10) and produced somewhat variable results, trends that emerged included the non-monophyly of Amphora + Halamphora and a close association of some of these taxa and the canal raphid orders Rhopalodiales Mann and Surirellales Mann (Bruder and Medlin 2008; Medlin and Kaczmarska 2004; Ruck and Theriot 2011; Sorhannus 2004). Although provocative, the small taxon sampling and continued low support has done little to advance our understanding of the systematics of this group. In the most extensively sampled phylogeny to date, and the first to include taxa outside Amphora s.s. and Halamphora, Sato et al. (2013) recently published a phylogeny based exclusively on SSU rDNA data that includes 22 Amphora taxa: nine from Amphora s.s., nine from Halamphora, three from the subgenus Diplamphora, and one from the subgenus Oxyamphora. This analysis supported the non-monophyly of Amphora s.l., with Halamphora and the lone representative from the subgenus Oxyamphora most closely related to the Rhopalodiales and Surirellales, and Amphora s.s. most closely related to taxa from the subgenus Diplamphora. Although returning a result supporting the hypothesis of a paraphyletic Amphora s.l., hypothesis testing performed on the single marker alignment was unable to reject the hypothesis of monophyly of Amphora s.l. at a 0.05 significance level (Sato et al. 2013, p. 233). This investigation presents a molecular phylogeny of Amphora s.l. inferred from a three marker dataset that includes the largest sampling of Amphora s.l. taxa and subgenera. Taxa presented here include representatives from Amphora s.s. and Halamphora as well as taxa from the subgenera Oxyamphora, Diplamphora and Amblyamphora. The aims of this study are to 1)
Molecular Phylogeny of Amphora s.l. 179
investigate hypotheses concerning the monophyly of Amphora s.l. and of the represented subgenera, and 2) discuss the systematic implications of the phylogenetic analysis within the framework of the hypotheses preceding it, most notably Cleve’s (1895) and Mereschkowsky’s (1903) classification, and offer insights towards a revised classification of the group.
Results Maximum likelihood and bayesian analyses were performed on the concatenated three-marker alignment (SSU + rbcL + psbC) as well as on individual markers. Presented are the ML phylograms. Node labels correspond to maximum likelihood bootstrap values (BS)/bayesian posterior probability (BPP) reported as a percentage. Incongruence between the ML and bayesian trees is denoted as (-) in the bayesian node support. The results from the analysis of the three marker concatenated matrix is shown in Figure 1. There was congruence between the ML and Bayesian analysis in both the relationships within Amphora clades, as well as the positions of these clades within the larger tree. The analysis of the three-marker alignment indicates that Amphora s.l. is polyphyletic, both including and excluding the genus Halamphora, and has members distributed widely across raphid diatom lineages. Trees inferred from single marker alignments are largely congruent with the major patterns shown in the concatenated analysis. The results of the ML analyses are presented in Supplementary Material Figures S1, S2 and S3 for the markers SSU, rbcL and psbC, respectively. All three markers indicate Amphora s.l. is polyphyletic. As in the three-marker analysis, Amphora and Diplamphora were consistently monophyletic, Halamphora was monophyletic, and Oxyamphora was polyphyletic. Support for the deeper nodes was low in the single markers analyses, making the position of these groups within the larger raphid diatom lineage somewhat ambiguous. Only the sister relationship between Amphora s.s. and the Diplamphora + Oxyamphora clade was consistently well supported (95/100, 86/100 and 48/97 for SSU, rbcL and psbC, respectively). Only the psbC analysis returned a tree in which Amphora s.l. was largely monophyletic with a large clade including Amphora s.s., the Diplamphora + Oxyamphora clade, and the Halamphora + Oxyamphora clade, although support for this relationship was low (16/56).
Table 1. Results from the testing of monophyletic hypotheses using the approximately unbiased (AU) test statistic. OBS = the unconstrained ML tree score subtracted from the constrained tree score. PAU = pvalues from the AU test. H0: unconstrained topology from the maximum likelihood analysis. H1: Large Amphora and Halamphora clades form a monophyletic group. H2: Large Amphora, Halamphora and Oxyamphora lineolata clades form a monophyletic group, H3: Amphora sensu lato is monophyletic. Hypothesis
OBS
H0 H1 H2 H3
-19.3 19.3 70.2 162.1
PAU 0.991 0.010
ORIGINAL PAPER
Molecular Phylogeny of Amphora sensu lato (Bacillariophyta): An Investigation into the Monophyly and Classification of the Amphoroid Diatoms Joshua G. Stepanek 1 , and J. Patrick Kociolek Department of Ecology and Evolutionary Biology and, Museum of Natural History, University of Colorado, Boulder, CO 80309, USA Submitted September 27, 2013; Accepted February 8, 2014 Monitoring Editor: Marina Montresor
Amphora sensu lato encompasses a large group of raphid diatoms, diverse in both form and ecology. The defining feature of this group has been an extreme asymmetry of the valve mantle and girdle bands, bringing both faces of the cell onto a single plane. Although this ‘amphoroid’ structure has long been the diagnostic feature and thus considered ‘conservative’ for the group, many have argued that the diversity of forms presently assigned to Amphora likely does not represent a monophyletic group. With the exception of several taxonomic transfers and the recent elevation of Halamphora to the level of genus, much of Amphora classification has remained unchanged for over 100 years. This study presents a phylogenetic analysis of Amphora s.l. based on a concatenated molecular alignment including the nuclear marker SSU rDNA and the chloroplast markers rbcL and psbC. These results are discussed within the framework of the current classification system of Amphora and Halamphora and lay the groundwork for a taxonomic revision of the group based on monophyly. The results of this analysis demonstrate that the genus Amphora is polyphyletic and that lineages assigned to the genus are distributed widely across the raphid diatom tree of life. The feature of amphoroid symmetry appears to have evolved independently several times. We discuss the nature of conservative characters in the raphid diatoms and their usefulness as a guide to phylogenetic relationships. © 2014 Elsevier GmbH. All rights reserved. Key words: Amphora; Halamphora; phylogeny; SSU; rbcL; psbC.
Introduction Amphora Ehrenberg ex Kützing sensu lato is a large (>800 species; Fourtanier and Kociolek 2009; Levkov 2009) and morphologically diverse group of raphid diatoms found in fresh, brackish and salt water habitats from the tropics to the poles (BérardTherriault et al. 1986; Hohn and Hellerman 1966; 1
Corresponding Author; fax +1 303-735-0128 e-mail [email protected] (J.G. Stepanek).
© 2014 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.protis.2014.02.002
Patrick and Freese 1961; Stepanek and Kociolek 2013; Stoermer and Yang 1971; Wachnicka and Gaiser 2007). Partially due to this morphological diversity, the genus is primarily defined by a distinct asymmetry in the valve mantle and girdle bands. This asymmetry, which places both valve faces on the same plane, is distinct in the light microscope and has been the single feature unifying the genus since its description by Kützing in 1844 (Cleve 1895; Kützing 1844; Smith 1873). For nearly as long there have been concerns as to whether
178 J.G. Stepanek and J.P. Kociolek
the genus Amphora, as defined by this asymmetry, represented a natural group (Cleve 1895; Krammer 1980; Mann 1994) and early attempts at a comprehensive Amphora classification proved difficult (Cleve 1895; Mereschkowsky 1903). Now, over 100 years after these concerns were first raised, questions regarding the monophyly of the genus Amphora have yet to be fully resolved. In a classification system that largely persists today, Cleve (1895) separated the genus into nine subgenera (Amphora, Diplamphora, Halamphora, Oxyamphora, Amblyamphora, Psammamphora, Cymbamphora, Calamphora, and Archiamphora) based on a number of morphological characters such as a ‘simple’ or ‘complex’ connecting zone (girdle area), the presence or absence of ‘longitudinal lines’, as well as several raphe and striae features. In creating his classification, Cleve struggled with the relationships between the subgenera and their position within the larger diatom classification system. Cleve recognized frustular asymmetry as a feature shared by members of the genus, but unlike previous authors (Smith 1873) he did not believe this characteristic was appropriate for the formation of a natural group (Cleve 1895, p. 99). A notable aspect of Cleve’s treatment was his assertion that his subgenera likely represented a number of distinct and potentially unrelated genera, though he was hesitant to elevate his new subgenera because of nomenclatural concerns as well as affording an “opportunity of testing my views, which are entirely new, before admitting the proposed new genera” (Cleve 1895, p. 100). Mereschkowsky (1903) put forth the first challenge to Cleve’s classification system, in which he proposed a new classification based on plastid number and morphology. Mereschkowsky separated Amphora into four large groups based on plastid number, and within these groups the taxa were further divided into nine subgroups based on the position and shape of the chloroplast. This new classification system was not widely adopted and the genus Amphora continued to be recognized for this entire group of species by most taxonomists. With the increased use of electron microscopy came new investigations and revisions to groups and taxa within Amphora s.l. Paddock and Sims (1980) where the first to transfer taxa from the genus Amphora to other genera, transferring taxa from the subgenus Oxyamphora, to the newly created genus Undatella Paddock and Sims. Several other transfers based on scanning electron microscope (SEM) observations followed, including several taxa to the genus Biremis Mann and Cox (Mann and Cox in Round et al. 1990) and
Colliculoamphora Williams and Reid (Williams and Reid 2006), as well as the elevation of the Amphora subgenus Halamphora to generic status (Levkov 2009) and the transfer of the subgenus Cymbamphora to the newly created genus Seminavis Mann (Mann in Round et al. 1990). In addition to the morphological investigations, several molecular phylogenetic studies, while not specifically focusing on Amphora, have included Amphora s.l. in their analyses (Bruder and Medlin 2008; Medlin and Kaczmarska 2004; Ruck and Theriot 2011; Sorhannus 2004). The molecular markers used varied between studies from single marker alignments (Medlin and Kaczmarska 2004; Sorhannus 2004) to three marker concatenated alignments (Bruder and Medlin 2008; Ruck and Theriot 2011). Although these studies included few Amphora s.l. taxa (2–10) and produced somewhat variable results, trends that emerged included the non-monophyly of Amphora + Halamphora and a close association of some of these taxa and the canal raphid orders Rhopalodiales Mann and Surirellales Mann (Bruder and Medlin 2008; Medlin and Kaczmarska 2004; Ruck and Theriot 2011; Sorhannus 2004). Although provocative, the small taxon sampling and continued low support has done little to advance our understanding of the systematics of this group. In the most extensively sampled phylogeny to date, and the first to include taxa outside Amphora s.s. and Halamphora, Sato et al. (2013) recently published a phylogeny based exclusively on SSU rDNA data that includes 22 Amphora taxa: nine from Amphora s.s., nine from Halamphora, three from the subgenus Diplamphora, and one from the subgenus Oxyamphora. This analysis supported the non-monophyly of Amphora s.l., with Halamphora and the lone representative from the subgenus Oxyamphora most closely related to the Rhopalodiales and Surirellales, and Amphora s.s. most closely related to taxa from the subgenus Diplamphora. Although returning a result supporting the hypothesis of a paraphyletic Amphora s.l., hypothesis testing performed on the single marker alignment was unable to reject the hypothesis of monophyly of Amphora s.l. at a 0.05 significance level (Sato et al. 2013, p. 233). This investigation presents a molecular phylogeny of Amphora s.l. inferred from a three marker dataset that includes the largest sampling of Amphora s.l. taxa and subgenera. Taxa presented here include representatives from Amphora s.s. and Halamphora as well as taxa from the subgenera Oxyamphora, Diplamphora and Amblyamphora. The aims of this study are to 1)
Molecular Phylogeny of Amphora s.l. 179
investigate hypotheses concerning the monophyly of Amphora s.l. and of the represented subgenera, and 2) discuss the systematic implications of the phylogenetic analysis within the framework of the hypotheses preceding it, most notably Cleve’s (1895) and Mereschkowsky’s (1903) classification, and offer insights towards a revised classification of the group.
Results Maximum likelihood and bayesian analyses were performed on the concatenated three-marker alignment (SSU + rbcL + psbC) as well as on individual markers. Presented are the ML phylograms. Node labels correspond to maximum likelihood bootstrap values (BS)/bayesian posterior probability (BPP) reported as a percentage. Incongruence between the ML and bayesian trees is denoted as (-) in the bayesian node support. The results from the analysis of the three marker concatenated matrix is shown in Figure 1. There was congruence between the ML and Bayesian analysis in both the relationships within Amphora clades, as well as the positions of these clades within the larger tree. The analysis of the three-marker alignment indicates that Amphora s.l. is polyphyletic, both including and excluding the genus Halamphora, and has members distributed widely across raphid diatom lineages. Trees inferred from single marker alignments are largely congruent with the major patterns shown in the concatenated analysis. The results of the ML analyses are presented in Supplementary Material Figures S1, S2 and S3 for the markers SSU, rbcL and psbC, respectively. All three markers indicate Amphora s.l. is polyphyletic. As in the three-marker analysis, Amphora and Diplamphora were consistently monophyletic, Halamphora was monophyletic, and Oxyamphora was polyphyletic. Support for the deeper nodes was low in the single markers analyses, making the position of these groups within the larger raphid diatom lineage somewhat ambiguous. Only the sister relationship between Amphora s.s. and the Diplamphora + Oxyamphora clade was consistently well supported (95/100, 86/100 and 48/97 for SSU, rbcL and psbC, respectively). Only the psbC analysis returned a tree in which Amphora s.l. was largely monophyletic with a large clade including Amphora s.s., the Diplamphora + Oxyamphora clade, and the Halamphora + Oxyamphora clade, although support for this relationship was low (16/56).
Table 1. Results from the testing of monophyletic hypotheses using the approximately unbiased (AU) test statistic. OBS = the unconstrained ML tree score subtracted from the constrained tree score. PAU = pvalues from the AU test. H0: unconstrained topology from the maximum likelihood analysis. H1: Large Amphora and Halamphora clades form a monophyletic group. H2: Large Amphora, Halamphora and Oxyamphora lineolata clades form a monophyletic group, H3: Amphora sensu lato is monophyletic. Hypothesis
OBS
H0 H1 H2 H3
-19.3 19.3 70.2 162.1
PAU 0.991 0.010
