- Email: [email protected]
Holobionts, hybrids, and cladistic classification (reply to David P. Mindell)
BioSystems ELSEVIER S( IENTIFI( PUBLISHERS IRELAND
BloSystems 31 (1993) 127-130
Discussion
Holobionts, hybrids, and cladistic classification (reply to David P. Mindell Jan Zrzavy
,,a
, ZdenEk
Skfila b
aInstltute of Entomology and Faculty' of Bwlogtcal Sctences of the Umverstty of South Boherma, CeskOBudkjovt~e, Czech Repubhc bInstltute of Botany, Prthontce. Czech Repubhc
Abstract
David P Mlndell assumes that incorporation of the holoblonts (evoluhonardy stable symbiotic complexes) and hybrids into phylogenetlc trees necessardy distorts the hierarchical structure of cladlstxc classification, and must lead to reformulation of some basic cladlstac concepts (BtoSystems, 27, 53-62, 1992) He does not regard the Eukarya and Eubacterla as monophyletlc taxa, the former being 'polyphyletlc', the latter 'paraphyletlc' We attempt to show that the existence of holoblonts/hybrlds is not a cladlstlc problem Cladlstacs is a systematic methodology, not a system of evolutionary hypotheses, cladograms are statements about distribution of shared characters (synapomorphles) and, consequently, the cladograms are always branching and hierarchically structured A taxon is monophyletlc if it has a single root in the cladogram, not a single ancestor 'in reahty'. Cladlstlc methodology does not prowde a clue for distinguishing whether a confhctlng distribution of the potential synapomorphles is due to retlculatmn (e g , symbiogenesis) or convergent evolution Consequently, both Eubacterla and Eukarya either are or are not monophyletlc taxi if they are or are not determined to be so during cladlstlc analysis
Key words Cladlstlcs, Phylogenetlc reticulation, Symbiosis
The systematic status o f e v o l u t l o n a r l l y stable symbiotic complexes is an i m p o r t a n t b u t still scarcely discussed topic So D a v i d P M l n d e l l ' s (BwSystems, 27, 5 3 - 6 2 , 1992) a t t e m p t to redress this gap in the c u r r e n t t a x o n o m y should be welcomed. He assumes, in short, t h a t 'as a consequence o f considering symbioses in phylogenetlc analyses, the p r o p o s e d taxa E u b a c t e r l a a n d
* Corresponding author, Department of Morphology, Institute of Entomology, Czech Academy of Sciences, Bramgovska 31, 370 05 ~.eske Bud~jovlce. Czech Repubhc tBtoSystems, 27, 53-62. 1992
E u k a r y a are seen to be n o n - m o n o p h y l e t l c , and thus, p o o r indicators o f e v o l u t i o n a r y h i s t o r y ' (p 53) U n f o r t u n a t e l y , M l n d e l l ' s paper, in o u r view, shows some theoretical m i s c o n c e p t i o n s a n d contradictions We m u s t agree with Mlndell that ' both forms o f reticulate evolution represent taxa (hybrids or h o l o b l o n t s ) originating from m o r e than one species, thereby c o m p l i c a t i n g hierarchical representation o f genealogy a n d classification ' (p 54), and that is why they are fully equivalent from the v i e w p o i n t o f cladzsttcs (to which Mlndell a p p a r e n t l y adheres). Thus all conclusions concern-
0303-2647/93/$06 00 © 1993 Elsevier SoentlfiC Pubhshers Ireland Ltd All rights reserved SSDI 0303-2647(93)01412-M
128 lng inclusion of the holobionts in hierarchical classification must be apphcable to the hybrids as well, and vice versa (see also McDade, 1992) Mlndell presents both of the principal posslbihtles for deahng with the holobionts In the first, the members of a holoblont are each regarded as an independent species and, consequently, class1fled within the corresponding parental clades as its ecologically derived members (1) The second approach classifies the holoblont as a whole, each symbiotic event producing a new monophyletlc taxon (2) There is no non-arbitrary 'biological' way of deciding between the application of concept (1) or (2) in individual cases, in any case, they are entirely exclusive from the cladistlc viewpoint Either there is no classifiable taxon but, instead, a tightly coherent ecosystem (1), or there is such a taxon, possessing a unique position within the cladogram and classification (2) Some of Mindelt's statements are, therefore, hardly intelligible" 'Holoblont taxa will have constituent symblonts Included m dtfferent monophylenc groups' (p. 53, our italics). Classification of both/all participants of a holobiont as independent species (1) could be awkward in practice (e.g, if the human mltochondria were to be classified as a species of the Eubacterla), but no methodological problems arise if this is accepted, The problem is how to classify those complexes which will be considered coherent and, as such, classifiable species (2, e g , a eukaryotlc cell) Mlndell assumes that incorporation of the holobionts/hybrlds into the phylogenetlc tree necessarily distorts the hierarchical structure of cladistxc classification, and must lead to reformulatmn of some basic cladlstic concepts, especially that of monophyly. In fact, in the conceptual framework of cladlstics one must distinguish sharply between evolutzonary and systemanc concepts Monophyly (p 56 'Including only and all the descendants of a single common ancestral species') is not an evolutionary term, it is not a feature of an evolutionary lineage but of a taxon, and, consequently, monophyletlc/nonmonophyletic taxa cannot be depicted in the phylogenettc tree but only in the cladogram They are not the same cladograms are statements about the distribution of shared characters - - the synapo-
J Zrzavy, Z Skfla/BtoSystems 31 (1993) 127-130 morphles - - and they are represented as branching diagrams, which recall phylogenetlc trees merely for traditional and historical reasons; cladograms are about characters and taxa, phylogenetlc trees are about ancestors and descendants. Consequently, the cladogram, as a result of cladlstlc analysis, is always branching and hierarchically structured, it is a diagram in which a character pattern is displayed, monophyly being the manner in which this pattern is structured (Fig 1) (Nelson and Platnick, 1981; Forey et al., 1992) The holoblonts/hybrlds should, of course, be classified 'regardless of size, weight, status as host, or other measure of relative dominance' (p 57), since all these measures are out of the scope of cladlstlc methodology The only parameter Important for cladlstics is the number of synapomorphles shared by individual couples of taxa within a three-taxon system (e g., by the holoblont/hybrld and its two parents) There is no reason to exclude complex taxa, if found classifiable, from this rule all species, regardless of their supposed mode of speclatlon, must therefore be treated cladlstlcally in the same way (Fig 1). Mindell says that 'to discount one symblont within a holoblont when reconstructing the holoblont's overall phylogenetlc history and proposing a classification . . is equivalent to disposal of data' He is right, but this is a typical hazard of systematlsts. The distributional pattern of characters is usually reticulate some of potential synapomorphles are usually in conflict These conflicts must be resolved by elimination of those 'synapomorphies' which are incongruous with the others and, therefore, considered superficial, nonsynapomorphic, 'convergent'. Mlndell asserts that 'the history of no one symblont takes precedence over another in assessing genealogy (monophyly) and classification ' (p. 57) In fact, this precedence of a certain symblont happens because the number of synapomorphles binding each constituent symblont to its parental clade is usually unequal As the complex species is usually formed asymmetrically, with more characters descending from one parent than from the other, the characters one symblont shared with its parental clade are regarded as 'synapomorphies' of the complex species and the pertinent parental clade (Fig I taxa D + E,
J Zrzavy, Z Sk6la/BtoSystems 31 (1993) 127-130 A
B
C
0
129
E
D
A
0
I]
E
a
b C
d e
f g
A
B
0
D
E
Fig 1 Two hypothetical phylogenetlc trees (one with convergent evolution, the other with reticulation) with five terminal species (A-E, D represents holobxont/hybrxd m the latter tree) and seven potentml synapomorphxes (a-g), and their translauon into a smgle cladogram with five terminal taxi (A-E), six synapomorphaes (a-f) and one homoplasy (g) The species analyzed should be classified m four monophylette taxi ABCDE (synapornorphy a), BCDE (synapomorphy b), CDE (synapomorphy c) and DE (synapomorphles
d-f)
characters d-f), while the other symblont's characters are regarded as 'homoplaslous' and discounted (taxi C + D, character g) Such conflicts between characters, caused either by reticulation or convergent evolution, cannot be distinguished by cladlstlcs, since tt has no methodological means of doing so. In conclusion, we must emphasize that the extstence of holobaonts/hybrads is not a cladlstlc problem and, therefore, not a problem of elass~catton Cladastxcs is a systematic methodology
(and, therefore, either 'useful' or 'useless'), not a system of evolutionary hypotheses (either 'true' or 'false'), and it mvolves cladogram construction, not phylogeny reconstruction A taxon is monophyletlc if it has a single 'root' m the cladogram, not a single ancestor in the phylogenetlc tree (or 'in reality'). We all tend to interpret monophyly, synapomorphy, homoplasy, and related terms from the viewpoint of evolutionary biology, but this stems from our own theoretical background, and not from the character matrix observed or
130
J Zrzavy, Z Skfla/BtoSystems 31 (1993) 127-130
from the methodology used So, both Eubacterla and Eukarya either are or are not monophylettc taxa if they are or are not determined to be so during cladlstic analysis (McDade, 1992). Their symbiotic relationships should be included in analysis, but they will not necessarily change the overall cladogram topology We thank Nelson Platnlck (American Museum of Natural History, New York) and Pavel Stys (Charles University, Praha) for their comments on the manuscript drafts
References Forey, F L, Humphrles, C J , Kltchlng, I J , Scotland, R W, Slebert, D J and Wdhams, D M , 1992, Cla&stlcs A Practical Course m Systematlcs (Oxford Umv Press, Oxford, etc ) McDade, L A , 1992, Hybrids and phylogenetlc systernatlcs II The impact of hybrids on cladlstlc analysis Evolution 46, 1329-1346 Mmdell, D P, 1992, Phylogenetlc consequnces of symbmses Eukarya and Eubacterm are not monophyletxc taxa BloSysterns 27, 53-62 Nelson, G and Platmck, N I, 1981, Systematlcs and Blogeography Cla&stlcs and Vlcanance (Columbm Unlv Press, New York)
BioSystems ELSEVIFR SCIFNq If'l( PUBLISHLRS IRFLAND
BloSystems 31 (1993) 130-133
Merger of taxa and the definition of monophyly (reply to Jan Zrzav3 and Zden6k Skfila) D a v i d P. M i n d e l l Department of Btologzcal Sctences, Untverslty of Cmclnnatt, Cmcmnatt, OH 45221, USA
In a recent paper I discussed the consequences for phylogeny reconstruction of merger among symbiotic taxa (Mxndell, 1992). I make the point that prior independent histories for symblont taxa that have merged within a composite taxon, or holoblont, can be included in phylogenetic studies, thereby Improving the accuracy of phylogenetlc reconstruction. I also point out that recognition of the prior independent htstones for symbIonts in phylogeny reconstruction requires their consideratlon in diagnosis of monophyletlc groups Recognition of symbioses leading to merger of taxa (as in the endosymblotlc origin of some cellular organelles) is relatively recent and is only now SSDI 0303-2647(93)01413-N
being incorporated into the body of phylogenetlc systematlcs theory The justification for incorporating merger of taxa, and corresponding reticulate branches, in phylogenetic trees is to provide a more complete estimate of orgamsmal evolutionary history Zrza@ and Skf.la's (1993) objections stem from their embrace of what has been called pattern cladlstlcs. This is apparent from their primary focus on &strlbutlon of characters rather than the evolution of organisms, as seen in their redefinition of the term 'monophyly' (relative to the definition which I provided) They substitute their definition of monophyly for mine, and then, not surprisingly, they find my statements
BloSystems 31 (1993) 127-130
Discussion
Holobionts, hybrids, and cladistic classification (reply to David P. Mindell Jan Zrzavy
,,a
, ZdenEk
Skfila b
aInstltute of Entomology and Faculty' of Bwlogtcal Sctences of the Umverstty of South Boherma, CeskOBudkjovt~e, Czech Repubhc bInstltute of Botany, Prthontce. Czech Repubhc
Abstract
David P Mlndell assumes that incorporation of the holoblonts (evoluhonardy stable symbiotic complexes) and hybrids into phylogenetlc trees necessardy distorts the hierarchical structure of cladlstxc classification, and must lead to reformulation of some basic cladlstac concepts (BtoSystems, 27, 53-62, 1992) He does not regard the Eukarya and Eubacterla as monophyletlc taxa, the former being 'polyphyletlc', the latter 'paraphyletlc' We attempt to show that the existence of holoblonts/hybrlds is not a cladlstlc problem Cladlstacs is a systematic methodology, not a system of evolutionary hypotheses, cladograms are statements about distribution of shared characters (synapomorphles) and, consequently, the cladograms are always branching and hierarchically structured A taxon is monophyletlc if it has a single root in the cladogram, not a single ancestor 'in reahty'. Cladlstlc methodology does not prowde a clue for distinguishing whether a confhctlng distribution of the potential synapomorphles is due to retlculatmn (e g , symbiogenesis) or convergent evolution Consequently, both Eubacterla and Eukarya either are or are not monophyletlc taxi if they are or are not determined to be so during cladlstlc analysis
Key words Cladlstlcs, Phylogenetlc reticulation, Symbiosis
The systematic status o f e v o l u t l o n a r l l y stable symbiotic complexes is an i m p o r t a n t b u t still scarcely discussed topic So D a v i d P M l n d e l l ' s (BwSystems, 27, 5 3 - 6 2 , 1992) a t t e m p t to redress this gap in the c u r r e n t t a x o n o m y should be welcomed. He assumes, in short, t h a t 'as a consequence o f considering symbioses in phylogenetlc analyses, the p r o p o s e d taxa E u b a c t e r l a a n d
* Corresponding author, Department of Morphology, Institute of Entomology, Czech Academy of Sciences, Bramgovska 31, 370 05 ~.eske Bud~jovlce. Czech Repubhc tBtoSystems, 27, 53-62. 1992
E u k a r y a are seen to be n o n - m o n o p h y l e t l c , and thus, p o o r indicators o f e v o l u t i o n a r y h i s t o r y ' (p 53) U n f o r t u n a t e l y , M l n d e l l ' s paper, in o u r view, shows some theoretical m i s c o n c e p t i o n s a n d contradictions We m u s t agree with Mlndell that ' both forms o f reticulate evolution represent taxa (hybrids or h o l o b l o n t s ) originating from m o r e than one species, thereby c o m p l i c a t i n g hierarchical representation o f genealogy a n d classification ' (p 54), and that is why they are fully equivalent from the v i e w p o i n t o f cladzsttcs (to which Mlndell a p p a r e n t l y adheres). Thus all conclusions concern-
0303-2647/93/$06 00 © 1993 Elsevier SoentlfiC Pubhshers Ireland Ltd All rights reserved SSDI 0303-2647(93)01412-M
128 lng inclusion of the holobionts in hierarchical classification must be apphcable to the hybrids as well, and vice versa (see also McDade, 1992) Mlndell presents both of the principal posslbihtles for deahng with the holobionts In the first, the members of a holoblont are each regarded as an independent species and, consequently, class1fled within the corresponding parental clades as its ecologically derived members (1) The second approach classifies the holoblont as a whole, each symbiotic event producing a new monophyletlc taxon (2) There is no non-arbitrary 'biological' way of deciding between the application of concept (1) or (2) in individual cases, in any case, they are entirely exclusive from the cladistlc viewpoint Either there is no classifiable taxon but, instead, a tightly coherent ecosystem (1), or there is such a taxon, possessing a unique position within the cladogram and classification (2) Some of Mindelt's statements are, therefore, hardly intelligible" 'Holoblont taxa will have constituent symblonts Included m dtfferent monophylenc groups' (p. 53, our italics). Classification of both/all participants of a holobiont as independent species (1) could be awkward in practice (e.g, if the human mltochondria were to be classified as a species of the Eubacterla), but no methodological problems arise if this is accepted, The problem is how to classify those complexes which will be considered coherent and, as such, classifiable species (2, e g , a eukaryotlc cell) Mlndell assumes that incorporation of the holobionts/hybrlds into the phylogenetlc tree necessarily distorts the hierarchical structure of cladistxc classification, and must lead to reformulatmn of some basic cladlstic concepts, especially that of monophyly. In fact, in the conceptual framework of cladlstics one must distinguish sharply between evolutzonary and systemanc concepts Monophyly (p 56 'Including only and all the descendants of a single common ancestral species') is not an evolutionary term, it is not a feature of an evolutionary lineage but of a taxon, and, consequently, monophyletlc/nonmonophyletic taxa cannot be depicted in the phylogenettc tree but only in the cladogram They are not the same cladograms are statements about the distribution of shared characters - - the synapo-
J Zrzavy, Z Skfla/BtoSystems 31 (1993) 127-130 morphles - - and they are represented as branching diagrams, which recall phylogenetlc trees merely for traditional and historical reasons; cladograms are about characters and taxa, phylogenetlc trees are about ancestors and descendants. Consequently, the cladogram, as a result of cladlstlc analysis, is always branching and hierarchically structured, it is a diagram in which a character pattern is displayed, monophyly being the manner in which this pattern is structured (Fig 1) (Nelson and Platnick, 1981; Forey et al., 1992) The holoblonts/hybrlds should, of course, be classified 'regardless of size, weight, status as host, or other measure of relative dominance' (p 57), since all these measures are out of the scope of cladlstlc methodology The only parameter Important for cladlstics is the number of synapomorphles shared by individual couples of taxa within a three-taxon system (e g., by the holoblont/hybrld and its two parents) There is no reason to exclude complex taxa, if found classifiable, from this rule all species, regardless of their supposed mode of speclatlon, must therefore be treated cladlstlcally in the same way (Fig 1). Mindell says that 'to discount one symblont within a holoblont when reconstructing the holoblont's overall phylogenetlc history and proposing a classification . . is equivalent to disposal of data' He is right, but this is a typical hazard of systematlsts. The distributional pattern of characters is usually reticulate some of potential synapomorphles are usually in conflict These conflicts must be resolved by elimination of those 'synapomorphies' which are incongruous with the others and, therefore, considered superficial, nonsynapomorphic, 'convergent'. Mlndell asserts that 'the history of no one symblont takes precedence over another in assessing genealogy (monophyly) and classification ' (p. 57) In fact, this precedence of a certain symblont happens because the number of synapomorphles binding each constituent symblont to its parental clade is usually unequal As the complex species is usually formed asymmetrically, with more characters descending from one parent than from the other, the characters one symblont shared with its parental clade are regarded as 'synapomorphies' of the complex species and the pertinent parental clade (Fig I taxa D + E,
J Zrzavy, Z Sk6la/BtoSystems 31 (1993) 127-130 A
B
C
0
129
E
D
A
0
I]
E
a
b C
d e
f g
A
B
0
D
E
Fig 1 Two hypothetical phylogenetlc trees (one with convergent evolution, the other with reticulation) with five terminal species (A-E, D represents holobxont/hybrxd m the latter tree) and seven potentml synapomorphxes (a-g), and their translauon into a smgle cladogram with five terminal taxi (A-E), six synapomorphaes (a-f) and one homoplasy (g) The species analyzed should be classified m four monophylette taxi ABCDE (synapornorphy a), BCDE (synapomorphy b), CDE (synapomorphy c) and DE (synapomorphles
d-f)
characters d-f), while the other symblont's characters are regarded as 'homoplaslous' and discounted (taxi C + D, character g) Such conflicts between characters, caused either by reticulation or convergent evolution, cannot be distinguished by cladlstlcs, since tt has no methodological means of doing so. In conclusion, we must emphasize that the extstence of holobaonts/hybrads is not a cladlstlc problem and, therefore, not a problem of elass~catton Cladastxcs is a systematic methodology
(and, therefore, either 'useful' or 'useless'), not a system of evolutionary hypotheses (either 'true' or 'false'), and it mvolves cladogram construction, not phylogeny reconstruction A taxon is monophyletlc if it has a single 'root' m the cladogram, not a single ancestor in the phylogenetlc tree (or 'in reality'). We all tend to interpret monophyly, synapomorphy, homoplasy, and related terms from the viewpoint of evolutionary biology, but this stems from our own theoretical background, and not from the character matrix observed or
130
J Zrzavy, Z Skfla/BtoSystems 31 (1993) 127-130
from the methodology used So, both Eubacterla and Eukarya either are or are not monophylettc taxa if they are or are not determined to be so during cladlstic analysis (McDade, 1992). Their symbiotic relationships should be included in analysis, but they will not necessarily change the overall cladogram topology We thank Nelson Platnlck (American Museum of Natural History, New York) and Pavel Stys (Charles University, Praha) for their comments on the manuscript drafts
References Forey, F L, Humphrles, C J , Kltchlng, I J , Scotland, R W, Slebert, D J and Wdhams, D M , 1992, Cla&stlcs A Practical Course m Systematlcs (Oxford Umv Press, Oxford, etc ) McDade, L A , 1992, Hybrids and phylogenetlc systernatlcs II The impact of hybrids on cladlstlc analysis Evolution 46, 1329-1346 Mmdell, D P, 1992, Phylogenetlc consequnces of symbmses Eukarya and Eubacterm are not monophyletxc taxa BloSysterns 27, 53-62 Nelson, G and Platmck, N I, 1981, Systematlcs and Blogeography Cla&stlcs and Vlcanance (Columbm Unlv Press, New York)
BioSystems ELSEVIFR SCIFNq If'l( PUBLISHLRS IRFLAND
BloSystems 31 (1993) 130-133
Merger of taxa and the definition of monophyly (reply to Jan Zrzav3 and Zden6k Skfila) D a v i d P. M i n d e l l Department of Btologzcal Sctences, Untverslty of Cmclnnatt, Cmcmnatt, OH 45221, USA
In a recent paper I discussed the consequences for phylogeny reconstruction of merger among symbiotic taxa (Mxndell, 1992). I make the point that prior independent histories for symblont taxa that have merged within a composite taxon, or holoblont, can be included in phylogenetic studies, thereby Improving the accuracy of phylogenetlc reconstruction. I also point out that recognition of the prior independent htstones for symbIonts in phylogeny reconstruction requires their consideratlon in diagnosis of monophyletlc groups Recognition of symbioses leading to merger of taxa (as in the endosymblotlc origin of some cellular organelles) is relatively recent and is only now SSDI 0303-2647(93)01413-N
being incorporated into the body of phylogenetlc systematlcs theory The justification for incorporating merger of taxa, and corresponding reticulate branches, in phylogenetic trees is to provide a more complete estimate of orgamsmal evolutionary history Zrza@ and Skf.la's (1993) objections stem from their embrace of what has been called pattern cladlstlcs. This is apparent from their primary focus on &strlbutlon of characters rather than the evolution of organisms, as seen in their redefinition of the term 'monophyly' (relative to the definition which I provided) They substitute their definition of monophyly for mine, and then, not surprisingly, they find my statements