J. Eukaryot. Microbiol., 57(1), 2010 pp. 19–32 2009 The Author(s) … · 2010. 1. 4. · shapes observed in Phacus affect the whorled patterns of pellicle strips that have been

Evolution of Distorted Pellicle Patterns in Rigid Photosynthetic Euglenids(Phacus Dujardin)

HEATHER J. ESSONa,1 and BRIAN S. LEANDERa,b

aDepartment of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, Canada V6T 1Z4, andbDepartment of Zoology, University of British Columbia, 6270 University Boulevard, Vancouver, Canada V6T 1Z4

ABSTRACT. Members of the euglenid genus Phacus are morphologically differentiated from other photosynthetic species by the pres-ence of a rigid cytoskeleton (pellicle) and predominantly dorsoventrally flattened, leaf-shaped cells. In order to better understand theevolutionary history of this lineage, we used scanning electron microscopy to examine patterns of pellicle strips in Phacus acuminatus,Phacus longicauda var. tortus, Phacus triqueter, Phacus segretii, Phacus pleuronectes, Phacus similis, Phacus pusillus, Phacus orbic-ularis, Phacus warszewiczii, and Discoplastis spathirhyncha, a putative close relative of Phacus and Lepocinclis. Our observationsshowed that while the earliest diverging species in our analyses, namely P. warszewiczii, has three whorls of exponential reduction, mostmembers of Phacus have clustered patterns of posterior strip reduction that are bilaterally symmetrical distortions of the radially sym-metrical ‘‘whorled’’ patterns found in other photosynthetic euglenids. Comparative morphology, interpreted within the context of mo-lecular phylogenetic analyses of combined nuclear small subunit rDNA and partial nuclear large subunit rDNA sequences, demonstratesthat clustered patterns of posterior strip reduction arose after the divergence of Phacus from other photosynthetic euglenids and are theresult of developmental processes that govern individual strip length. Clustered patterns of pellicle strips in Phacus do not appear to beadaptively significant themselves; they evolved in association with the origin of cell flattening and cell rigidity, which may be adaptationsto a planktonic lifestyle.

Key Words. Character evolution, cytoskeleton, Discoplastis, morphology, phylogeny.

PHACUS (Dujardin, 1841) is a morphologically distinctiveclade of photosynthetic euglenids that includes rigid, dorso-

ventrally flattened cells. Most species have an elongated caudalprocess and longitudinally arranged pellicle strips (Fig. 1–13).Several species of Phacus consist of three lobes and are deltoid intransverse section, while other species have become twistedaround their longitudinal axis in a corkscrew fashion (e.g. Phacusinflexus and Phacus similis [ 5 P. smulkowskianus Zakrys], Fig. 5;Huber-Pestalozzi 1955). Molecular phylogenetic analyses havedemonstrated that the genus was polyphyletic, and several speciesformerly grouped within Phacus based on light microscopical ob-servations have subsequently been moved to other rigid photo-synthetic genera, namely Monomorphina and Cryptoglena (Marinet al. 2003). The phylogenetic relationships within Phacus sensustricto, however, remain poorly understood (Brosnan et al. 2003;Kosmala et al. 2007; Linton et al. 2000; Marin et al. 2003; Mullneret al. 2001; Nudelman et al. 2003; Triemer et al. 2006).

Comparative analyses of morphological data, particularly pell-icle characters, are expected to help build a phylogenetic frame-work for understanding the overall diversity of Phacus. Kosmalaet al. (2007) found that characters visible using light microscopy,such as the presence or absence of transverse struts, were goodtaxonomical characters in delimiting Phacus pleuronectes andPhacus orbicularis. Leander and Farmer (2000a, b, 2001a, b) andLeander, Witek and Farmer (2001) used scanning and transmis-sion electron microscopy (SEM and TEM, respectively) to de-scribe pellicle characters which, when incorporated into cladisticanalyses and compared with molecular data, provided robust in-ferences about euglenid phylogeny. Their sampling of Phacus,however, turned out to include only three members of Phacussensu stricto: Phacus oscillans, Phacus triqueter, and Phacusbrachykentron (reidentified as Phacus acuminatus; other taxa be-longed to Lepocinclis and Monomorphina; Leander and Farmer2001b; Marin et al. 2003; Triemer et al. 2006).

While the leaf-like morphology described by Dujardin (1841) ispredominant in Phacus sensu stricto, a number of Phacus taxa thatare not yet placed in other genera based on phylogenetic analysesdeviate from it in one or more characters. For example,P. triqueter and Phacus warszewiczii are conspicuously tri-lobedrather than being dorsoventrally flattened per se (Fig. 8–10; Hub-er-Pestalozzi 1955; Leander and Farmer 2001b), and P. wars-zewiczii is illustrated with helically arranged pellicle strips(Huber-Pestalozzi 1955). Moreover, taxa such as Phacus segretii(Fig. 7) and Phacus stokesii, lack a caudal process and insteadhave rounded posterior ends (Huber-Pestalozzi 1955). Other taxa,such as Phacus parvulus and Phacus pusillus, are described ashaving extremely blunt caudal processes (Huber-Pestalozzi 1955;Fig. 11). To date, very few of these atypical taxa (P. triqueter,P. parvulus, and P. pusillus) have been included in molecular ormorphological phylogenetic analyses (Leander and Farmer2001b; Marin et al. 2003; Triemer et al. 2006).

One pellicle character that has been informative in previousstudies of euglenid evolution and taxonomy is posterior strip re-duction: patterns formed on the posterior cell surface by pelliclestrips of different lengths (e.g. Leander and Farmer 2000a). Thepresence of uniquely modified patterns of posterior reduction insome species of Phacus indicate that it may be particularly usefulin resolving relationships within the genus and forming inferencesregarding pellicle character evolution (Leander and Farmer2001b). We were interested in exploring how the unusual cellshapes observed in Phacus affect the whorled patterns of pelliclestrips that have been characterized in other euglenid lineages. Ourknowledge of Phacus surface morphology is extremely spotty andin the vast majority of cases, non-existent, and this research takesthe initial steps needed to help illuminate this area of uncertainty.Because of the complex evolutionary history and developmentalprocesses underlying the formation of these patterns (Esson andLeander 2006, 2008), a brief review of their diversity and struc-ture is included below.

Evolutionary significance of posterior strip reduction. Inaddition to a corset of microtubules and a network of endoplasmicreticulum, the peripheral cytoskeleton of euglenids is reinforcedby 4–120 proteinaceous strips that lie beneath the plasma mem-brane and extend longitudinally or helically from the anterior

Corresponding Author: H. J. Esson, Department of Botany, Univer-sity of British Columbia, 6270 University Boulevard, Vancouver, Can-ada V6T 1Z4—Telephone number: 143 0664 60277 24032; FAXnumber: 143 1 4277 9240; e-mail: [email protected]

1Present Address: Max F. Perutz Laboratories, University of Viennaand Medical University of Vienna. Dr. Bohr-Gasse 9, A-1030 Vienna,Austria.

19

J. Eukaryot. Microbiol., 57(1), 2010 pp. 19–32r 2009 The Author(s)Journal compilation r 2009 by the International Society of ProtistologistsDOI: 10.1111/j.1550-7408.2009.00447.x

i:/BWUS/JEU/3779-447/[email protected]

Fig. 1–13. Scanning electron micrographs showing the diversity of Phacus. 1. Discoplastis spathirhyncha, a closely related lineage to Phacus with32 pellicle strips. 2. Phacus pleuronectes. 3. Phacus longicauda var. tortus. 4. Phacus oscillans. 5. Phacus similis. 6. Phacus orbicularis. 7. Phacussegretii, showing the rounded posterior end of the cell. 8. Phacus triqueter. 9. Phacus warszewiczii. 10. Posterior view of Phacus warszewiczii showingthree lobes of the deltoid shaped cell. 11. Phacus pusillus. 12. Phacus acuminatus, UBC isolate. 13. P. acuminatus (brachykentron), UTEX LB 1317. Allimages at same scale (bar 5 20 mm).

20 J. EUKARYOT. MICROBIOL., 57, NO. 1, JANUARY– FEBRUARY 2010

canal region to the posterior end of the cell (Leander, Esson andBreglia 2007). The number of pellicle strips around the cellperiphery is more or less consistent within species and is referredto using the variable ‘‘P’’ (Leander and Farmer 2000a). In pho-tosynthetic euglenids, however, some strips are too short to reachthe posterior end of the cell and instead terminate at a certain pointalong the length of the cell. The length of any particular strip de-pends on its relative age: pellicle strips are duplicated and inher-ited semi-conservatively, where existing strips resume andterminate growth with each subsequent round of cytokinesis (Es-son and Leander 2006). Just before cytokinesis, a new strip formsbetween every pair of existing strips, and these are the youngeststrips on the pellicle of any cell. The youngest strips are shorterthan all other pellicle strips, while the oldest strips reach the pos-terior tip of the cell (Bouck and Ngo 1996; Esson and Leander2006; Hofmann and Bouck 1976; Mignot, Brugerolle, and Bri-cheux 1987). Younger strips terminate before reaching the poste-rior tip of the cell and form a radial pattern or ‘‘whorl’’ on the cellsurface. The older strips that lie between the strips forming awhorl extend either to the posterior tip or to a more posteriorwhorl, depending on the relative age of the strips and the degree oflength differentiation in the species (Fig. 14; Esson and Leander2006, 2008; Leander and Farmer 2000a).

Age-related length differentiation varies between species sothat some species will have strips of two different lengths, whileother species can have up to five different strip lengths (Leanderand Farmer 2000a). The number of posterior whorls, denoted as‘‘Wp,’’ can therefore be described in terms of the degree of striplength differentiation in a given species or culture strain. HigherWp values reflect increased differentiation in strip lengths and arethought to be the result of changes in developmental timing (i.e.heterochrony) associated with the termination and resumption ofstrip growth during pellicle duplication (Esson and Leander2006). Some Euglena and Lepocinclis species exhibit lengthdifferentiation within a single whorl, indicating that factors otherthan age contribute to strip length (Esson and Leander 2008;Leander and Farmer 2000a, b).

We studied the pellicle surface patterns of eight additionalPhacus taxa—Phacus acuminatus, Phacus longicauda var. tortus,P. pleuronectes, P. pusillus, P. orbicularis, P. segretii, P. similis,and P. warszewiczii—and a close relative of the clade consistingof Phacus and Lepocinclis, namely Discoplastis spathirhyncha.By generating several new sequences, we ensured that small sub-unit (SSU) and partial large subunit (LSU) rDNA sequences wereavailable from each of these species. This approach enabled us to(1) improve our understanding of Phacus diversity, (2) interpretpellicle characters in a molecular phylogenetic context, (3) ex-amine the significance of pellicle diversity in reconstructing evo-lutionary trends along the Phacus lineage, and (4) establish abroader framework for understanding developmental processesassociated with the diversification of the euglenid pellicle.

MATERIALS AND METHODS

Culture sources and conditions. Cultures were either pur-chased from culture collections or grown from single cells isolatedfrom freshwater sources located in or near Vancouver, Canada(Table 1). Cultures were maintained in LM7 (P. segretii, P. long-icauda var. tortus, P. acuminatus, P. inflexus, P. pleuronectes,P. warszewiczii, P. pusillus; ACOI, http://www1.ci.uc.pt/botanica/ACOI_M � 1.htm) or a modified soil water medium supple-mented with either 1/8 of a pea (P. orbicularis) or vitamin B12 (D.spathirhyncha) (modified from Pringsheim 1946) at 17–18 1Cwith a 12 h light:dark cycle.

Scanning electron microscopy and replicate observations.Cells in culture were placed in a Petri dish whose lid was fitted

with filter paper and fixed using osmium tetroxide vapor as de-scribed previously (Esson and Leander 2006). Cells were placedon filters and critical point dried with CO2. Once the filters weremounted on stubs, cells were coated with either gold or a combi-nation of gold and palladium. Stubs were viewed on a HitachiS4700 SEM (Pleasanton, CA).

While previous surface pattern descriptions (e.g. Leander andFarmer 2001b) were based on multiple cells with the same char-acter clearly visible, the flattening and twisting of many Phacusspecies results in cells lying in positions where a given character,especially posterior strip reduction, cannot be clearly viewed in itsentirety. Nevertheless, important information can be collectedfrom a number of cells and synthesized to provide an accuratedescription of a given character. Between 10 and 50 cells wereobserved for each taxon, and composite descriptions of relevantcharacters were created from these data.

DNA extraction, polymerase chain reaction (PCR) amplifi-cation, and cloning. Genomic DNA was extracted fromP. acuminatus, P. inflexus, P. longicauda var. tortus, P. orbicu-laris, P. pleuronectes, P. pusillus, P. segretii, and P. warszewicziiusing either a standard CTAB protocol (Breglia, Slamovits, and

Table 1. Taxon names, strain identification, and accession numbers ofsequences used for molecular phylogenetic analyses in this study.

Taxon Strainidentification

GenBank accessionnumbers

SSU LSU

Euglena viridis SAG 1224-17c AY523037 DQ140125Discoplastis spathirhynchaa SAG 1224-42 AJ532454 DQ140100Colacium mucronatumb UTEX 2524 AF326232 AY130224Monomorphina pyrumb UTEX 2354 AF112874 AY130238Trachelomonas lefevrei SAG 1283-10 DQ140136 AY359949Lepocinclis ovum SAG 1244-8 AF110419 AY130235Lepocinclis steinii(L. buetschlii in Leanderand Farmer 2000a)b

UTEX 523 AF096993 AY130815

Lepocinclis tripterisb UTEX LB 1311 AF445459 AY130230Phacus acuminatusa UBC culturec GQ422787 GQ422794Phacus acuminatus(P. brachykentron inLeander and Farmer2001b)b

UTEX LB 1317 AJ532481 AY130820

Phacus inflexus ACOI 1336 GQ422788 GQ422795Phacus longicauda var.tortusa

ACOI 1139 GQ422789 GQ422796

Phacus orbicularis ASW 08054 AF283315 DQ140126Phacus orbicularisa UBC cultured GQ422790 GQ422797Phacus oscillansb UTEX LB 1285 AF181968 AY1308238Phacus cf. parvulus ASW 08060 AF283314 DQ140127Phacus pleuronectes SAG 1261-3b AJ532475 AY130824Phacus pleuronectesa UTEX LB54 via

CCCM 7053GQ422791 GQ422798

Phacus pusillusa ACOI 1093 AJ532472 GQ422799Phacus pusillus UTEX 1282 AF190815 AY130237Phacus segretiia ACOI 1337 GQ422792 GQ422800Phacus similisa SAG 58.81 AJ532467 AY130239Phacus triqueterb SAG 1261-8

(5 UTEXLB1286)

AJ532485 Triemer et al.(In press)

Phacus warszewicziia ASW 08064 GQ422793 GQ422801

aTaxa for which pellicle surface morphology is described in this study.bTaxa for which pellicle surface morphology was described by Leander

and colleagues (Leander and Farmer 2001a, b; Leander et al. 2001).cIsolated from freshwater pond at the University of British Columbia.dIsolated from freshwater pond near Boundary Bay, British Columbia.Accession numbers for new sequences are shown in bold type.

21ESSON & LEANDER—PELLICLE PATTERNS IN PHACUS

http://www1.ci.uc.pt/botanica/ACOI_M&sim;1.htm







Leander 2007) or the MasterPureTM Complete DNA and RNAPurification Kit (Epicentre Biotechnologies, Madison, WI). ThePCR was performed using a total volume of 25 ml and the PuReTaq Ready-To-Go PCR beads kit (GE Healthcare, Buckingham-shire, UK). Nuclear SSU and LSU rDNA sequences were ampli-fied on either a PTC-100 Peltier Thermal Cycler or a MJ MiniPersonal Thermal Cycler (Bio-Rad, Hercules, CA) using the fol-lowing PCR protocol: an initial denaturation stage at 95 1C for2 min; 35 cycles involving 94 1C for 45 s (denaturation), 50–55 1Cfor 45 s (annealing), and 72 1C for 1.5 min (extension); and finalextension at 72 1C for 5 min. Small subunit rDNA sequences wereamplified as one or two fragments using combinations of theprimers listed in Table 2; partial LSU rDNA sequences were am-plified as one fragment using the primers in Table 2. Bands of theexpected size were excised from agarose gel and cleaned using theUltraCleanTM 15 DNA Purification kit (MO Bio, Carlsbad, CA)according to the manufacturer’s instructions. Purified sequenceswere cloned using the TOPO TA kit (Invitrogen, Carlsbad, CA).Plasmids were digested using EcoR1 and inserts were sequencedusing ABI BigDye 3.1 and forward and reverse vector primers.Sequencing reactions were processed using an Applied Bi-osystems 3730S 48-capillary sequencer.

Molecular phylogenetic analyses. In addition to the se-quences we obtained for this study, previously published nuclearSSU rDNA and LSU rDNA sequences for Phacus strains andother taxa were acquired from GenBank (Table 1). Sequencesfrom strains identified as belonging to the same species as our owncultures (i.e. P. orbicularis, P. pleuronectes, and P. acuminatus)were included to ensure accurate taxonomic identification (usinglight microscopy) of these cultures. New SSU rDNA sequencesranged in length from 2,110 to 2,208 bp; despite many repeatedattempts, only the second half of the SSU rRNA gene (977 bp)was obtained from P. segretii. New LSU sequences ranged inlength from 1,106 to 1,654 bp. We performed molecular phylo-genetic analyses on three alignments that combined SSU rDNAand partial LSU rDNA sequences: (1) a 24-taxon alignment withnine outgroup sequences (2,017 sites), (2) a 23-taxon alignmentwith nine outgroup sequences and excluding the shorter SSUrDNA sequence from P. segretii (2,017 sites), and (3) an align-ment of 15 Phacus spp. with outgroup sequences removed (2,017sites). The euglenid sequences were pairwise aligned in MacClade4 (Maddison and Maddison 2000) using an alignment from a pre-viously published study as our guide (Triemer et al. 2006). Se-quences were further aligned by eye. Gaps and ambiguouslyaligned bases were excluded from all three alignments.

PhyML (Guindon and Gascuel 2003; Guindon et al. 2005) wasused to analyze all three datasets with one heuristic search per

dataset and with maximum-likelihood (ML) using a general-timereversible (GTR) model of base substitutions (Rodrıguez et al.1990) that incorporated invariable sites and a discrete g distributionwith eight rate categories (GTR1I1G model). The model param-eters were estimated by PhyML from each of the three originaldatasets: 24-taxon alignment (a5 0.768; i 5 0.348; A 5 0.22914,C 5 0.24677, G 5 0.30259, T 5 0.22150; Ao-4C 5 1.15579,Ao-4G 5 2.40124, Ao-4T 5 1.36811, Co-4G 5 0.44077,Co-4T 5 5.38119, Go-4T 5 1.0), 23-taxon alignment(a5 0.788; i 5 0.353; A 5 0.22959, C 5 0.24616, G 5 0.30201,T 5 0.22224; Ao-4C 5 1.17369, Ao-4G 5 2.46732, Ao-4T 5 1.40475, Co-4G 5 0.43655, Co-4T 5 5.50249, Go-4T 5 1.0), and 15-taxon alignment (a5 0.495; i 5 0.249;A 5 0.22822, C 5 0.24627, G 5 0.30132, T 5 0.22418; Ao-4C 5 0.86600, Ao-4G 5 1.92150, Ao-4T 5 1.23231, Co-4G 5 0.28591, Co-4T 5 4.92935, Go-4T 5 1.0). The ML boot-strap analyses were conducted on each alignment with the samesettings described above (i.e. 100 pseudoreplicates; one heuristicsearch per pseudoreplicate).

The three alignments were also analyzed with Bayesian meth-ods using the MrBayes program 3.1.2 (Huelsenbeck and Ronquist2001; Ronquist and Huelsenbeck 2003). The program was set tooperate with a g distribution, invariable sites and four Monte Car-lo Markov chains starting from a random tree. A total of 2,000,000generations was calculated with trees sampled every 50 genera-tions and with a prior burn-in of 100,000 generations (2,000 sam-pled trees were discarded; burn-in was checked manually). Amajority rule consensus tree was constructed from 38,000 post-burn-in trees. Posterior probabilities correspond to the frequencyat which a given node was found in the post-burn-in trees.

In order to rule out alternative topologies that were not recov-ered in the ML analysis of the 24-taxon alignment, but mightnevertheless be supported by our data, eight additional topologiesto the one generated by the ML analysis were created throughbranch swapping in MacClade 4 (Maddison and Maddison 2000);this yielded nine topologies for comparison. RaxML 7.0.3(Stamatakis, Ludwig and Meier 2005) was used to determinesite-by-site likelihoods for all topologies using our 24-taxon align-ment and the settings described above. The approximately unbi-ased (AU) test was performed using CONSEL 0.1 (Shimodairaand Hasegawa 2001) to compare the alternate topologies.

RESULTS

Description of clustered reduction. Clustered strip reductionin various forms appeared in all taxa whose surface morphologywas examined, except for D. spathirhyncha and P. warszewiczii(Fig. 20, 21). In order to describe the differences in pellicle pat-terns between taxa, it is helpful to identify the main componentsof clustered reduction and compare these patterns to radially sym-metrical, whorled patterns (Fig. 14–19). The main features of two-whorled exponential reduction are summarized in Fig. 14, 17.Clustered reduction (Fig. 16, 18, 19) is a distortion of whorledreduction that is often associated with dorsoventral cell flatteningin Phacus. Length differentiation between different generations ofstrips still results in whorls of reduction, but the ventral and dorsalstrips forming a whorl terminate closer to the posterior tip of thecell than the lateral strips do, resulting in whorls that are ovoid orotherwise misshapen rather than circular in outline. Furthermore,some mature strips that would reach the posterior tip in cells withregular whorled strip reduction terminate before reaching the pos-terior tip, forming clusters on either side of the cell (Fig. 15).

The relationship between whorled and clustered strip reductionis best described in terms of the relative distance of strip termi-nations from the posterior tip of the cell: the actual differentiallength of each strip (Fig. 16). Clustered patterns are derived from

Table 2. Primers used in this study for amplification of ribosomal DNA.

SSU primer name Sequence

475 EUGF (forward) 50-AAGTCTGGTGCCAGCAGCYGC30

M917FD (forward) 50-GGTGAAATTCTTAGAYCG-30

PF1 (forward)a 50-GCGCTACCTGGTTGATCCTGCC-30

PHACF (forward) 50-CTGTGAATGGCTCCTTACATCA-30

EUGR (reverse) 50-TCACCTACARCWACCTTGTTA-30

FAD4 (reverse)b 50-TGATCCTTCTGCAGGTTCACCTAC-30

Inf 870R (reverse) 50-CAAGAGGCTGCTTTGAGCACA-30

PhR4 (reverse) 50-CAGGTTCACCTACAACAACC-30

R4 (reverse) 50-GATCCTTCTGCAGGTTCACCTA-30

LSU primer name1F (forward)c 50-TTAAGCATATCACTCAGTGGAGG-30

CIR (reverse)c 50-GCTATCCTGAGGGAAACTTCG-30

aPreviously published by Keeling (2002).bPreviously published by Deane et al. (1998) and Keeling (2002).cPreviously published by Brosnan et al. 2003.


developmental processes whereby strips in a given generation un-dergo unequal length differentiation, so that some strips terminatecloser to the posterior end of the cell than their co-generationalstrips. For example, if dorsal and ventral strips in a whorl of re-duction extend while the lateral strips belonging to the same whorldo not, that whorl acquires an ovoid shape. Similarly, if lateralstrips that would normally reach the posterior tip of the cellshorten while ventral and dorsal tip strips retain their length, clus-ters are formed from the lateral strips (Fig. 15, 16, 18, 19).

Descriptions of pellicle surface patterns in Discoplastis andPhacus. Surface patterns and sample sizes for the taxa examinedin this study are summarized in Table 3. Discoplastis spat-hirhyncha had P 5 32 pellicle strips arranged in a clockwise he-lix (when viewed from the posterior end). Strips reduced over twowhorls of exponential reduction (Wp 5 2) of 16 and 8 terminatingstrips, respectively (Fig. 20). Most cells observed had tips withfour–seven strips instead of the predicted eight, but additionalwhorls were never detected—the few additional terminating stripsobserved near the posterior tip in some cells did not conform toany recognizable pattern. Cells gradually tapered over the poste-rior half to form a sharp caudal process.

Phacus warszewiczii had P 5 32 pellicle strips that were ar-ranged in an anti-clockwise helix when viewed from the posterior

end (Fig. 10). The helical pitch of the strips was reduced at thecaudal process so that strips were arranged almost longitudinally(Fig. 10, 21). Strips reduced over three exponential whorls of re-duction (Wp 5 3), leaving four strips at the posterior tip. Strutswere present on strips until they reached the caudal process (Fig.21).

Phacus segretii (P 5 32) had more or less longitudinal stripsthat began to twist in an anti-clockwise direction near the posteriorend of the cell, which completely lacked a caudal process (Fig. 7).Two distorted whorls of exponential reduction (Wp 5 2) were ob-served. The eight strips that comprised the second whorl usuallyformed a ‘‘figure eight’’ shape when straight lines were used toconnect the posterior ends of consecutive terminating strips (Fig.22). In addition to the exponential whorls, there were two addi-tional terminating strips that were located on opposite sides of thecell from one another; toward the anterior of the cell and beforestrip reduction, these two strips were separated from one anotheron both sides by 15 strips.

Phacus acuminatus (P 5 32) had one flattened whorl of expo-nential reduction. Whorl I strips on the dorsal and ventral sides ofthe cell contributed to the short, blunt caudal process and termi-nated closer to the posterior tip than the lateral whorl I strips.Additional terminating strips formed a cluster on each side of the

Fig. 14–19. Illustrations comparing whorled (ancestral state) and clustered (derived state) posterior strip reduction. 14. Whorled strip reduction is theresult of length differentiation between pellicle strips of different generations: every alternate strip terminates before reaching the posterior of the cell,forming a radial pattern (i.e. a whorl) on the posterior cell surface. Because half of the strips terminate on each whorl, this pattern is also referred to as‘‘exponential’’ strip reduction. A cell with two whorls of strip reduction, for example, has pellicle strips of three lengths: the younger, shortest strips(black) form the first, anterior whorl of posterior reduction. Slightly longer and older strips (dark gray) form the second whorl of reduction, and the longestand oldest strips (white) extend to the posterior of the cell. 15, 16. Clustered strip reduction is a modification of whorled reduction (Wp 5 2 or 1) that isassociated with dorsoventral compression of cells. Dorsal and ventral strips belonging to whorl I (black) and whorl II (dark gray) now terminate closer tothe posterior cell tip than co-generational lateral strips do. Furthermore, mature lateral strips (light gray) no longer extend to the posterior cell tip as theywould in whorled reduction, but form clusters of adjacent terminating strips on either side of the posterior tip. 17–19. The relationship between whorledreduction and different forms of clustered reduction is represented schematically. Dorsoventral compression of a cell with two whorls of exponentialreduction (Wp 5 2; Fig. 17) results in a cell with two whorls of reduction supplemented by lateral clusters (Wp 5 2; Fig. 18). Loss of one of these whorlsyields cells with one whorl of reduction with lateral clusters (Wp 5 1; Fig. 19). Dotted lines indicate the schematic outlines of clustered strips.


caudal process, leaving five–six strips at the posterior tip of thecell (Fig. 23, 24). Pellicle strips maintained a longitudinal orien-tation over the entire cell surface. However, six cells showed avery slight clockwise twist at the posterior tip.

Phacus similis (P 5 20) had longitudinally oriented strips andanti-clockwise twisted cells when viewed from the posterior endof the cell (Fig. 5, 25, 26). Strips reduced over one whorl of ex-ponential reduction that followed the deformation of the cell.Some whorl I strips extended down the caudal process near theposterior tip of the cell, and some terminated further up the cell onridges formed by cell flattening and twisting (Fig. 25, 26). Addi-tional terminating strips formed clusters of one–three strips on oneside of the caudal process and two–four strips on the other side.The total number of cluster strips on a cell ranged from five toseven, leaving three–five strips at the posterior tip of the cell. Aswith P. acuminatus, some cluster strips terminated so close to theposterior tip of the cell that the exact number of strips in a clusterwas difficult to determine (Fig. 25).

Phacus pusillus cells had P values ranging from 20 to 26 strips(Table 3). Pellicle strips were longitudinal to slightly helical untilthey reached the posterior end of the cell, where their helical pitchincreased to produce a pronounced anti-clockwise pattern (Fig.27, 28). Strips reduced over one whorl of reduction that was de-formed along with cell flattening and twisting. In some cells, con-secutive terminating strips formed a ‘‘figure eight’’ pattern whenconnected by straight lines, similar to whorl II in P. segretii (cf.Fig. 22, 28). In most cells, five or six strips reached the posteriortip, leaving four or five clustered strips in cells with P 5 20 (i.e. 20strips less 10 whorl I strips, less five or six tip strips, leaves four–five clustered strips) and seven or eight clustered strips in cellswith P 5 26 (Fig. 27, 28).

Phacus pleuronectes (P 5 32) (Fig. 29, 30) and P. orbicularis(P 5 32) (Fig. 31, 32) each had two flattened whorls of exponen-tial reduction. Strips were longitudinally oriented along the cellbody and twisted very slightly at the caudal process. Most of thecaudal processes in P. pleuronectes were twisted anti-clockwise(19 of 21 cells); of these, 12 were twisted anti-clockwise alongpart of the process and then clockwise at the posterior tip. In P.pleuronectes, three–four strips reached the posterior tip of the cell,leaving two lateral clusters of two–three strips on one side andthree strips on the other (Fig. 29). In P. orbicularis, clusters wereeach composed of one or two strips, leaving five–six strips at theposterior tip of the cell (Fig. 31, 32). Strips were oriented longi-tudinally down the entire length of the cell; however, the caudalprocess often displayed a slight clockwise twist, as viewed fromthe posterior end. In P. orbicularis, the strips exhibited robusttransverse struts.

Phacus longicauda var. tortus had P 5 32 strips and two flat-tened whorls of strip reduction (Fig. 33, 34), much likeP. pleuronectes and P. orbicularis. Strips were oriented longitu-dinally along the cell and anti-clockwise at the cell posterior end;the cell body itself was twisted in a slight anti-clockwise helix.Whorls I and II were distorted due to twisting of the cell (Fig. 34).Whorl I strips terminated before reaching the long, thin caudalprocess, while whorl II strips extended slightly past the base of thecaudal process (Fig. 33). Clusters of one–two strips on one sideand two–three strips on the other side of the cell were present (Fig.34). Additional terminating strips along the extremely narrowcaudal process resulted in only two–three strips reaching the sharpposterior tip of the cell (Fig. 33). Transverse struts were present onpellicle strips over most of the cell but not the caudal process (Fig.34); the struts were less well defined than those in P. warszewicziiand P. orbicularis.

Phylogeny of Phacus as inferred from SSU and LSUrDNA. The ML and Bayesian analyses of all three alignmentsproduced very consistent results, and several phylogenetic rela-

Tab

le3

.S

um

mary

of

novel

pell

icle

surf

ace

chara

cte

rsdesc

ribed

inth

isst

udy.

Taxon

Pell

icle

surf

ace

chara

cte

rsN

um

ber

of

cell

so

bse

rved

Nu

mb

er

of

stri

ps

aro

un

dcell

peri

ph

ery

(P)

Str

ipo

rien

tati

on

Nu

mb

er

of

po

steri

or

wh

orl

s(W

p)

Nu

mb

er

of

tip

stri

ps

(T)

Nu

mb

er

of

stri

ps

inla

tera

lclu

sters

(N)

Dis

co

pla

stis

spa

thir

hyn

ch

a3

2(3

1,

33

)(8

/n5

10

)C

lock

wis

eh

eli

cal

(18

/n5

18)

2(1

2/n

51

2)

4–

7,

n0 5

9—

18

Ph

acu

sa

cu

min

atu

s3

2(3

0)

(8/n

51

0)

Lo

ng

itu

din

al

(20

/n5

20

)1

(16

/n5

16

)T

55

–6

,n0 5

15

4–5

an

d5

–6

,n0 5

14

20

Ph

acu

slo

ng

ica

ud

av

ar.

tort

us

32

(28

,2

9)

(3/n

55

)L

on

git

ud

inal

foll

ow

ed

by

an

ti-c

lock

wis

ep

ost

eri

or

twis

t(1

8/n

51

8)

2(6

/n5

6)

2–

3,

n0 5

91

–2

an

d2

–3

,n0 5

51

7

Ph

acu

so

rbic

ula

ris

P5

32

(3/n

53

)L

on

git

ud

inal

foll

ow

ed

by

slig

ht

clo

ck

wis

ep

ost

eri

or

twis

t(2

9/n

52

9)

2(2

8/n

52

8)

5–

6,

n0 5

27

1an

d1

–2

,n0 5

12

30

Ph

acu

sp

leu

ron

ecte

s3

2(3

0,

35

,3

6,

39

,4

0,

42

)(3

/n5

9)

Lo

ng

itu

din

al

foll

ow

ed

by

asl

igh

tan

ti-

clo

ck

wis

ep

ost

eri

or

twis

t(3

1/n

53

1)

2(2

2/n

52

2)

3–

4,

n0 5

28

2–3

an

d3

,n0 5

23

31

Ph

acu

sp

usi

llu

s2

0,

22

,2

3,

24

,2

6(n

57

)L

on

git

ud

inal

foll

ow

ed

by

an

ti-c

lock

wis

ep

ost

eri

or

twis

t(3

9/n

53

9)

1(2

9/n

52

9)

5–

6,

n0 5

28

1–4

an

d2

–5

,n0 5

26

37

Ph

acu

sse

gre

tii

32

(29

,3

0)

(21

/n5

24)

Lo

ng

itu

din

al

foll

ow

ed

by

po

steri

or

an

ti-

clo

ck

wis

etw

ist

(31

/n5

31

)2

(27

/n5

27

)6

(5,

7),

(20

/n5

26

)1

an

d1

(11

0,

11

11

1)

(22

/n5

27

)3

1

Ph

acu

ssi

mil

is2

0(2

4)

(15

/n5

19)

Longit

udin

al

wit

hanti

-clo

ckw

ise-t

wis

ted

cell

(30

/n5

30

)1

(23

/n5

23

)3

–5

,n0 5

30

1–3

an

d2

–4

,n0 5

39

57

Ph

acu

sw

ars

zew

iczi

i3

2(4

/n5

4)

An

ti-c

lock

wis

efo

llo

wed

by

lon

git

ud

inal

po

steri

or

(31

/n5

31

)3

(28

/n5

28

)3

–4

(2/n

52

)—

31

Alt

ern

ate

chara

cte

rst

ate

sare

indic

ate

din

bra

ckets

aft

er

the

most

frequentl

yobse

rved

chara

cte

rst

ate

.W

here

chara

cte

rst

ate

scould

be

obse

rved

dir

ectl

y,th

en

um

ber

of

cell

sd

isp

lay

ing

thatst

ate

are

sho

wn

as

afr

acti

on

of

the

tota

lsa

mp

lesi

ze

for

that

ch

ara

cte

r,n.

Wh

ere

ch

ara

cte

rst

ate

sare

co

mp

osi

tere

co

nst

ructi

on

s,th

en

um

ber

of

cell

so

bse

rved

toarr

ive

at

that

syn

thesi

sis

giv

en

as

n0 .


tionships were repeatedly recovered with robust statistical sup-port, such as the monophyly of the genus Phacus (Fig. 35; resultsfrom the analyses of the 15-taxon-unrooted alignment not shown).The nearest sister group to the Phacus clade was a well-supportedclade consisting of Lepocinclis species. The early divergence of P.warszewiczii from the other Phacus species, and a clade consist-ing of P. longicauda var. tortus and P. triqueter, were alsostrongly supported in the analyses of all three datasets (Fig. 35).A subclade of Phacus consisting of P. oscillans, P. similis, P. in-flexus, P. parvulus, and both strains of P. pusillus—the so-called‘‘oscillans clade’’ (after Marin et al. 2003)—was also recoveredwith high support in all analyses. The oscillans clade, in turn,consisted of two strongly supported subclades: (1) P. oscillans,P. similis, and P. inflexus, and (2) P. parvulus and both P. pusillusstrains. The position of the oscillans clade within the genus wasessentially unresolved with our data, but did form a strongly sup-ported clade (posterior probability 5 0.95) with P. acuminatus inthe Bayesian analysis of the 23-taxon dataset (data not shown).

Of the nine alternative phylogenies subjected to the AU test,four could not be rejected. One of these was the topology shown inFig. 35, which was ranked as the second most probable topology(P 5 0.514) The topology with the greatest statistical support

(P 5 0.558) differed from that shown in Fig. 35 by placingP. acuminatus and P. pleuronectes together as a sister clade tothe oscillans clade. The third best-supported topology (P 5 0.491)included sister relationships between P. triqueter and P. wars-zewiczii and P. pleuronectes and P. segretii, respectively. Thefourth topology (P 5 0.489) placed P. acuminatus together withthe oscillans clade in a larger clade; this was the sister clade to aclade comprised of P. pleuronectes and P. orbicularis in a sisterrelationship with P. segretii.

DISCUSSION

A molecular phylogenetic framework for Phacus. Althoughsome of the deepest branches within Phacus were not consistentlyrecovered or highly supported in our molecular phylogenetic an-alyses, several conclusions can be drawn from these trees. Phacusas it is currently defined, is monophyletic and all taxa used in thisstudy that have been classified as Phacus should remain in thegenus. Phacus warszewiczii is among the earliest divergingPhacus species, a conclusion reinforced by the morphologicaldata (discussed below). Both ML and Bayesian analyses of allthree alignments indicate that P. longicauda var. tortus and

Fig. 20–22. Posterior strip reduction in Discoplastis spathirhyncha, Phacus warszewiczii, and Phacus segretii. 20. Discoplastis spathirhyncha hastwo whorls of exponential strip reduction (Wp 5 2) and 32 strips around the cell periphery (P 5 32) (bar 5 2 mm). Inset: high-magnification scanningelectron micrograph showing transverse strut-like striations on the pellicle strips (bar 5 0.5 mm). 21. Phacus warszewiczii has three whorls of exponentialreduction on the caudal process (Wp 5 3). Transverse struts (arrowheads) are present on the pellicle strips (bar 5 2 mm). 22. Phacus segretii has twowhorls of exponential strip reduction (Wp 5 2); whorl I is slightly distorted, and whorl II forms an asymmetrical ‘‘figure eight’’ shape when adjacentterminating strip are connected by straight lines. Two additional strips (arrows) terminate on opposite sides of the cell (bar 5 2 mm).


Fig. 23–28. Phacus species with clustered strips associated with one whorl of exponential strip reduction (Wp 5 1). 23. Posterior view of Phacusacuminatus (P 5 32) showing a flattened whorl of strip reduction (asterisks) and symmetrical lateral clusters of four terminating strips (arrows)(bar 5 5mm). 24. Lateral view of P. acuminatus, showing whorl I strips (asterisks) extending up the dorsal/ventral surfaces of the caudal process and onelateral cluster of four terminating strips (arrows) (bar 5 3mm). 25. A lateral view of Phacus similis (P 5 20) showing a distorted whorl of strip reduction(asterisks, white line) and a lateral cluster of three or four terminating strips (arrows). The posterior-most terminating strip (left arrow) is a tip stripeffectively shortened by cell twisting (bar 5 4mm). 26. A lateral view of P. similis showing a distorted whorl of strip reduction and two clustered strips(right arrows) (bar 5 3mm). 27. A view of Phacus pusillus (P 5 20–26) showing a distorted whorl extending down the ventral or dorsal surface of the celland a cluster of four terminating strips (lower arrows). The upper arrow indicates a terminating strip belonging to the cluster on the other side of the cell(bar 5 2mm). 28. Posterior view of P. pusillus showing a figure eight-shaped whorl of reduction and two clusters (arrows) consisting of one and threeterminating strips, respectively (bar 5 2mm).


P. triqueter are closely related to one another despite their some-what divergent morphology (Fig. 3, 8). However, the longbranches associated with these taxa might be prone to long-branchattraction artifact, so we interpret these results with caution. Theprecise phylogenetic position of P. segretii within Phacus, an-other morphologically derived taxon with long branches in ourmolecular analyses, also proved elusive, but there was high sup-port for a sister relationship between P. segretii and P. pleur-onectes in Bayesian analyses of the 24-taxon and 15-taxonalignments (data not shown).

A close relationship between P. pleuronectes and P. orbiculariswas never recovered in ML or Bayesian analyses. Support for thepositions of P. pleuronectes and P. orbicularis within Phacus,however, was low in all of our phylogenetic analyses, and one ofthe four topologies that could not be rejected based on the AU testincluded a clade comprised of P. pleuronectes and P. orbicularis.

The close phylogenetic position of P. parvulus (ASW 08060)with P. pusillus sequences suggests that this strain, or perhaps oneor both of the P. pusillus strains, might be misidentified. None-theless, the ‘‘oscillans clade,’’ consisting of P. oscillans, P. simi-lis, P. inflexus, both P. pusillus strains, and P. parvulus, wasrecovered with high support in all analyses. This is consistent withpreviously published SSU rDNA and SSU/LSU rDNA phyloge-nies where any combination of these taxa invariably branched to-gether to the exclusion of other Phacus taxa (Brosnan et al. 2003;Kosmala et al. 2007; Marin et al. 2003; Triemer et al. 2006).Moreover, the sister relationship between P. acuminatus and theoscillans clade in some analyses reinforces the comparative mor-phological data discussed below.

Evolution of clustered strip reduction patterns in Phacus.Clustered reduction was originally described in a strain ofP. acuminatus (brachykentron; UTEX LB 1317) with one whorlof exponential reduction (Wp 5 1) (Leander and Farmer 2001b).Before the present study, clustered reduction had not been ob-served in any other species of euglenid, even in the other twoPhacus species previously examined: P. oscillans has a singleadditional terminating strip, while P. triqueter was reported aspossessing three distorted whorls of exponential reduction (Lean-der and Farmer 2001b). Our SEM data show that clustered reduc-tion is in fact widespread within Phacus (Fig. 36). Its absence inP. warszewiczii indicates that clustered reduction was derived af-ter the divergence of the genus from a photosynthetic ancestorwith two or three whorls of strip reduction (Wp 5 2–3) (Fig. 36,37). This character state is shared with D. spathirhyncha and allmembers of Lepocinclis for which posterior reduction has beendescribed, with the exception of L. salina (Conforti and Tell 1983;Leander and Farmer 2000a; Leander et al. 2001). We have shownthat clustered reduction can be associated with two whorls ofstrip reduction (Wp 5 2) as well as with one whorl of strip reduc-tion (Wp 5 1) (Fig. 37). The phylogenetic distribution of taxawith two-whorled clustered reduction suggests that this stateevolved before Phacus taxa with a single whorl and clustered re-duction, which is limited to P. acuminatus and members of theoscillans clade (Fig. 36, 37). Because of the low statistical supportfor some branches within the Phacus clade, this inference remainsessentially untested using the current molecular phylogeneticdata.

The clusters of strips described in P. acuminatus by Leanderand Farmer (2001b) were symmetrical—each cluster was com-prised of four terminating strips positioned laterally on the cell.None of the species with clustered reduction described in thisstudy had consistently symmetrical clusters, and most had con-sistently asymmetrical clusters. Members of the oscillans cladehad particularly exaggerated asymmetry: the smaller cluster inP. pusillus was sometimes comprised of one strip; the additionalterminating strip observed in P. oscillans (Leander and Farmer

2001b) appears to be the only remnant of the clusters that werepresent in its ancestors (Fig. 36, 37).

Clustered reduction has also been minimized in P. segretii andP. triqueter. We interpret the two additional terminating strips inP. segretii as vestigial clusters (Fig. 37); they do not belong toeither whorl I or whorl II and are therefore relatively mature strips.Furthermore, their location opposite one another is reminiscent ofthe location of lateral clusters of strip reduction observed in othertaxa. Phacus triqueter, on the other hand, has distorted whorls butlacks strip clusters per se. Reexamination of electron micrographsused in a previous study (Leander and Farmer 2001b), combinedwith the distribution of character states inferred from the presentstudy (Fig. 36), suggest that this taxon actually has two whorls ofexponential reduction and several other terminating strips alongthe narrow caudal process, rather than three whorls of reduction.Although P. triqueter and P. warszewiczii share a deltoid cellshape, the three whorls on P. warszewiczii are positioned on ornear the caudal process and thereby avoid distortion by the pro-nounced deltoid shape of the cell. The whorls in P. triqueter arecomprised of some strips that terminate on the caudal process andother strips that terminate further up the cell body, resulting indistorted whorls (see Fig. 4b in Leander and Farmer 2001b). Thisdistortion is considered to be a vestige of the clustered reductionpresent in P. triqueter’s inferred ancestor (Fig. 36, 37).

The vestiges of clustered strip reduction in P. oscillans,P. segretii, and P. triqueter raise some questions regarding thedevelopmental origins of clustered reduction. While clusters andwhorl distortion are related to dorsal–ventral flattening in Phacus,there appear to be other factors leading to the development andevolution of these patterns. Phacus segretii and P. oscillans haverounded cells that do not pose any spatial restrictions on strips attheir posterior ends that would necessitate clustered reduction (i.e.a modification of the ‘‘optimal packing hypothesis’’ as proposed,and somewhat refuted, by Leander et al. 2001). Similarly, thethree-lobed cells of P. triqueter do not, in and of themselves, re-quire distorted whorls, as P. warszewiczii clearly demonstrates(Fig. 37). Phacus pusillus, moreover, has clusters located on itsventral and dorsal surfaces, rather than its lateral margins, wherespace should be more restricted. It is possible that further taxonsampling will demonstrate Phacus species that have lost all tracesof clustered reduction, indicating that as the degree of dorsoven-tral flattening is decreased, clusters of reduction are lost. On theother hand, flattened and twisted heterotrophic euglenids, such asHeteronema spirale, a bacteriovore that completely lacks poste-rior strip reduction (Breglia, pers. commun.), show that there mustbe other factors underlying the complex length differentiationpatterns observed in Phacus and other photosynthetic euglenids.

The presence of clustered strip reduction in Phacus, but not inother photosynthetic taxa, suggests that there are developmentalprocesses governing differential strip length that are peculiar tothis genus of rigid cells. Scanning electron microscopy studiessimilar to those previously undertaken (e.g. Esson and Leander2006) should be integrated with previous observations of cell di-vision in Phacus (Pochmann 1942) to better understand the inter-action between bilateral symmetry and other developmentalstages in pellicle duplication and cytokinesis, such as the place-ment of cleavage furrow strips (Esson and Leander 2006).

Two culture strains identified as the same species (i.e.P. acuminatus/brachykentron UTEX LB 1317 and P. acuminatusUBC) have similar, but not identical, patterns of strip reduction.Both strains have one whorl of exponential reduction supple-mented by lateral clusters, but the clusters in the UTEX strain aresymmetrical (Leander and Farmer 2001b), while the clusters inthe UBC strain appear to be asymmetrical in most cells. Further-more, the strips are arranged in a conspicuous clockwise helix atthe tip of the caudal process in the UTEX culture (see Fig. 5g in


Leander and Farmer 2001b), while in the UBC culture a longitu-dinal orientation is maintained in most cells. These differences,and indeed all differences in posterior strip reduction described inthis study, particularly cluster strip distribution, may eventually

prove to be useful characters for taxonomic delimitation. Theyrequire SEM to observe, however, which may be impractical forsome ecological/biogeographical studies where the use of suchtaxonomic characters would be desirable.


Evolution of other pellicle surface characters in Phacus.The evolution of total strip number (P) in euglenids has previouslybeen described in terms of behavioral ecology (i.e. a large P value

facilitates metaboly via sliding between pellicle strips and, there-fore, allows ingestion of large prey) and developmental processes(i.e. pellicle duplication combined with failure to divide has re-

Fig. 35. Rooted Maximum Likelihood tree of Phacus species and related photosynthetic euglenids (24-taxa total) inferred from combined smallsubunit and partial large subunit rDNA sequences. Maximum Likelihood bootstrap values above 55 are shown above the branches; Bayesian posteriorprobabilities above 0.80 are shown below the branches. Values shown in italics indicate where exclusion of Phacus segretii from the phylogeneticanalyses resulted in improved statistical support. The dash (-) indicates a branch that was not recovered in the 24-taxon Bayesian analysis.

Fig. 29–34. Phacus species with clustered strips associated with two whorls of exponential strip reduction (Wp 5 2). 29. Posterior view of Phacuspleuronectes (P 5 32) showing two dorsoventrally flattened whorls of strip reduction. Strips belonging to whorl I (outlined in white) do not extend downthe caudal process, while strips belonging to whorl II (outlined in black) do extend down the caudal process. Lateral clusters are formed by threeterminating strips on either side of the caudal process (arrows) (bar 5 5mm). 30. Lateral view of P. pleuronectes showing the positions and lengths of twocluster strips (arrows) relative to whorl I strips (asterisks) and whorl II strips (} ) (bar 5 5mm). 31. Lateral view of Phacus orbicularis (P 5 32) showingtwo flattened whorls of strip reduction and one extra terminating strip (arrows) on both sides of the caudal process. Transverse struts (arrowheads) arepresent on pellicle strips (bar 5 10 mm). 32. View of the posterior end of P. orbicularis showing the relative positions and lengths of whorl I strips(asterisks), whorl II strips (} ), and one extra terminating strip (arrow). Note that whorl II strips extend along the caudal process (bar 5 5mm). 33. Lateralview of Phacus longicauda var. tortus (P 5 32) showing a long, twisted caudal process. Whorl I strips (asterisks) terminate anterior to the caudal process,while whorl II strips ( � ) occupy the base of the caudal process. Extra terminating strips (arrows) are also shown (bar 5 5mm). 34. Posterior view of P.longicauda var. tortus showing an extremely flattened and twisted whorl I (white lines); whorl II (black lines) is distorted. Cluster strips (arrows) arearranged asymmetrically with two strips on one side of the caudal process and one strip on the other. Transverse struts (arrowheads) are visible on moststrips but are absent on the caudal process and the most posterior region of the cell body (bar 5 5mm).


sulted in several ‘‘strip doubling’’ events throughout euglenidevolution; conversely, division combined with failure to duplicatethe pellicle results in ‘‘strip halving’’ events) (Leander 2004,Leander et al. 2001, 2007). Photosynthesis originated in euglenidsvia a secondary endosymbiotic event involving eukaryovorouseuglenids and green algal prey cells (Gibbs 1978; Leander 2004;Leander et al. 2007). In rigid photosynthetic euglenids likePhacus, behavioral and other locomotive requirements associ-ated with predatory modes of feeding (e.g. gliding motility andmetaboly) are no longer selected for. This could be one reasonwhy many members of the Phacus and Lepocinclis clades have arelatively low number of strips (P � 32).

With a few exceptions (such as the larger, semi-rigidLepocinclis helicoideus [P 5 80] [Leander and Farmer 2000b]and the taxa with P 5 20 strips described here), members of thePhacus and Lepocinclis clades possess P 5 32 strips (Fig. 36, 37;Leander et al. 2001). The taxa examined in this study, with theexception of the members of the oscillans clade, all have P 5 32strips—including D. spathirhyncha, which forms the sister lin-eage of the clade comprising Lepocinclis and Phacus in somephylogenies (Marin et al. 2003) (Fig. 36). While D. spathirhynchashares other morphological features with members of this clade(i.e. multiple disc-shaped plastids lacking pyrenoids; Triemer etal. 2006), further molecular and morphological work are requiredto more robustly resolve its relationship to these taxa (Triemer etal. 2006). Because P 5 20 strips is shared by all members of theoscillans clade whose surface morphology has been described (i.e.P. oscillans, P. pusillus, and P. similis), it may be regarded as asynapomorphy for this clade. It is interesting to note, however,

that the P values recorded for P. pusillus in this study have a widerrange than those recorded for other taxa (Table 3).

Transverse struts were present on the pellicle strips in three ofthe taxa described in this study: P. warszewiczii, P. longicaudavar. tortus, and P. orbicularis. Leander and Farmer (2001b) alsoobserved struts in P. triqueter and to a much lesser degree inL. tripteris. Because the relationships between these Phacusspecies are unresolved and L. tripteris is the only Lepocinclisspecies in which struts have been observed, it is impossible tomake conclusive statements about the evolution of this charac-ter at this time. However, the presence of struts in both generasuggests that it was present in the most recent common ancestorof both Phacus and Lepocinclis (Fig. 36). Moreover, we ob-

Fig. 36. Hypotheses of character evolution in Phacus as inferred fromthe combined small subunit/large subunit rDNA phylogenetic analyses andcomparative morphology (see Fig. 37). Position 0: the number of stripsaround the cell periphery stabilizes at 32 (P 5 32); cells have two whorlsof posterior reduction (Wp 5 2) and are capable of metaboly. Position1: rigid cells with helically arranged strips, a caudal process, and two orthree undistorted, exponential whorls of strip reduction (Wp 5 2 or 3) de-marcate the origin of Phacus. Position 2: potential acquisition of an ad-ditional whorl of posterior reduction (Wp 5 3) and deltoid cell shape.Position 3: the origin of longitudinal strip orientation, dorsoventral cellcompression, and clustered reduction associated with two whorls of ex-ponential reduction. Position 4: whorls of posterior strip reduction twisted.Position 5: deltoid cell shape secondarily prominent and clusters reducedor lost. Position 6: loss of the caudal process and reduction of clusters.Position 7: secondary loss of one whorl of posterior reduction. Position8: secondary loss of one whorl of posterior reduction. Uneven strip re-duction event from P 5 32 strips, resulting in cells with P 5 20 strips.Cells within the oscillans clade acquired stronger asymmetry in clusterdistribution, and posterior reduction patterns became twisted.

Fig. 37. Illustration of the evolution of distorted posterior reductionpatterns in Phacus (see Fig. 36). Taxon names with a given pattern areshown at the bottom of each box. The most recent ancestor of Phacuslikely possessed a rigid pellicle and undistorted whorled reduction (Wp 5 2or 3) (position 1). These cells gave rise to both the deltoid cell shape andthree whorls as seen in Phacus warszewiczii (Wp 5 3; position 2) and flat-tened cells with two whorls and clustered strip reduction (Wp 5 2) similarto those described here for Phacus pleuronectes and Phacus orbicularis(center) (position 3). Modification of this latter type of posterior strip re-duction resulted in the remaining distortions described in this study. Forinstance, twisting of the cell and exaggerated elongation of the caudalprocess produced the misshapen whorls and extra terminating strips ob-served in Phacus longicauda var. tortus (top, center) (position 4). A sec-ondary modification of the strip clusters associated with a prominentdeltoid cell shape resulted in the pattern observed in Phacus triqueter(upper right) (position 5). Loss of the caudal process and reduction of stripclusters resulted in the pattern observed in Phacus segretii (right, secondfrom top) (position 6). Loss of one whorl resulted in the pattern observedin Phacus acuminatus (right, second from top) (position 7). Comparativemorphology indicates that this condition might be ancestral to the patternsof strip reduction observed in the oscillans clade (position 8). The absenceof one whorl of strip reduction (like that in P. acuminatus), a reduction inthe overall number of strips (P 5 20, rather than 32), and increased asym-metry of the strip clusters produced the strip reduction patterns found inthe oscillans clade (bottom) (position 8). Twisting of the flattened ances-tral cell resulted in the distorted whorls observed here in Phacus pusillusand Phacus similis (bottom right and center).


served fine, strut-like striations on the pellicle strips of D. spat-hirhyncha, which suggests that this feature evolved before themost recent common ancestor of Discoplastis, Lepocinclis, andPhacus.

With the exception of P. warszewiczii, most Phacus taxa havelongitudinally arranged strips over most of the cell surface and atwisted caudal process. As previously observed by Leander andFarmer (2001b), however, the handedness of the posterior twistvaries between taxa. Some taxa, such as P. acuminatus (brachy-kentron) and P. oscillans, exhibit a clockwise helix when viewedfrom the posterior end; other taxa, such as P. triqueter, P. long-icauda, and P. pusillus, have an anti-clockwise helix. Our obser-vations affirm the observations by Leander and Farmer (2001b)that handedness of pellicle strip orientation is not phylogeneticallyinformative. Members of the well-resolved oscillans clade haveboth clockwise and anti-clockwise helices, and there is no certainrelationship between the few taxa that exhibit a clockwise twist(i.e. P. orbicularis, P. oscillans, P. acuminatus, and D. spat-hirhyncha). Based on previous developmental research regardingthe semi-conservative nature of pellicle duplication (e.g. Bouckand Ngo 1996; Hofmann and Bouck 1976; Mignot et al. 1987),Leander and Farmer (2001b) hypothesized that there should bedevelopmental constraints on the handedness of pellicle strip ori-entation. Similar studies using strains of Phacus, given the com-bination of pellicle rigidity and the usual change in striporientation at the cell posterior in this genus, could be particu-larly informative regarding developmental mechanisms for theevolution of different strip orientations.

CONCLUSIONS

Based on the molecular and morphological data presented here,Phacus shares with its sister genus Lepocinclis the widespreadpossession of a semi-rigid or rigid pellicle with P 5 32 pelliclestrips. Furthermore, molecular phylogenetic support and the pres-ence of 32 pellicle strips in the metabolic taxon D. spathirhynchasuggest that Discoplastis is the sister taxon to the clade formed byPhacus and Lepocinclis. Patterns of clustered strip reduction arecommon in Phacus; however, they evolved after the divergence ofP. warszewiczii and are not, therefore, considered a synapomor-phy for the genus.

The posterior strip reduction patterns described in this studysuggest that other, unknown factors contribute to strip lengthdifferentiation, which has previously been explained in terms ofstrip maturity (Esson and Leander 2006) and position relative toparental strips (Esson and Leander 2008). Strip clusters and stripsthat terminate outside of any particular exponential whorl arecomprised of mature strips belonging to different generations thatnevertheless terminate sooner than their co-generational strips onthe ventral and dorsal cell surfaces. The presence of these clustersdoes not always appear to be directly correlated with cell flatten-ing and twisting and often reflect modifications of ancestral stateswithin the group. Therefore, distorted patterns of pellicle stripsoffer important insights into the development and evolutionaryhistory of the cytoskeleton in Phacus, and have the potential tomake contributions to our understanding of the overall diversifi-cation of the eukaryotic cell.

ACKNOWLEDGMENTS

We wish to thank S. A. Breglia for performing SEM onD. spathirhyncha, M. Hoppenrath for collection of P. orbicular-is, C. Chantangsi for his help in primer design, amplification of theSSU rDNA of P. segretii, and execution of phylogenetic analyses,S. Rueckert for her assistance with the literature, and E. Gill andA. Horak for performing AU tests. The unpublished LSU rDNA

sequence of P. triqueter was generously provided by R.E. Trie-mer. E.W. Linton provided a comprehensive combined SSU andLSU rDNA alignment of photosynthetic euglenids from a previ-ous study, greatly facilitating our molecular analyses. Three anon-ymous reviewers significantly improved the clarity of thismanuscript. This research was supported by grants to B. S. L.from the Natural Sciences and Engineering Research Council ofCanada (NSERC 283091-04) and the Canadian Institute for Ad-vanced Research, Program in Integrated Microbial Biodiversity;H.J.E. was also supported by a UBC Graduate Fellowship. Thismanuscript was submitted by H.J.E. in partial fulfillment of therequirements for the Ph.D. degree, University of British Colum-bia, Vancouver, BC, Canada.

LITERATURE CITED

Bouck, G. B. & Ngo, H. 1996. Cortical structure and function in eugleno-ids with reference to trypanosomes, ciliates, and dinoflagellates.Int. Rev. Cytol., 169:267–318.

Breglia, S. A., Slamovits, C. H. & Leander, B. S. 2007. Phylogeny ofphagotrophic euglenids (Euglenophyceae) as inferred from hsp90 genesequences. J. Eukaryot. Microbiol., 52:86–94.

Brosnan, S., Shin, W., Kjer, K. M. & Triemer, R. E. 2003. Phylogeny ofthe photosynthetic euglenophytes inferred from the nuclear SSU andpartial LSU rDNA. Int. J. Syst. Evol. Microbiol., 53:1175–1186.

Conforti, V. & Tell, G. 1983. Disposicion de la Bandas y Estrias de laCuticula de Lepocinclis salina Fritsch, (Euglenophyta) observadas enM.E.B. Nova Hedwigia, 38:165–168.

Deane, J. A., Hill, D. R. A., Brett, S. J. & McFadden, G. I. 1998. Hanusiaphi gen. et. sp. nov. (Cryptophyceae): characterization of ‘Cryptomonassp.F’. Eur. J. Phycol., 33:149–154.

Dujardin, F. 1841. Histoire naturelle des Zoophytes. Infusoires. LibrairieEncyclopedique de Roret, Paris, 684 p.

Esson, H. J. & Leander, B. S. 2006. A model for the morphogenesis ofstrip reduction patterns in photosynthetic euglenids: evidence for he-terochrony in pellicle evolution. Evol. Develop., 8:378–388.

Esson, H. J. & Leander, B. S. 2008. Novel pellicle surface patterns onEuglena obtusa Schmitz (Euglenophyta), a euglenophyte from a benthicmarine environment: implications for pellicle development and evolu-tion. J. Phycol., 44:132–141.

Gibbs, S. 1978. The chloroplasts of Euglena may have evolved from sym-biotic green algae. Can. J. Bot., 56:2883–2889.

Guindon, S. & Gascuel, O. 2003. A simple, fast, and accurate algorithmto estimate large phylogenies by maximum likelihood. Syst. Biol., 52:696–704.

Guindon, S., Lethiec, F., Duroux, P. & Gascuel, O. 2005. PHYML Online:a web server for fast maximum likelihood-based phylogenetic infer-ence. Nucleic Acids Res., 33(Web Server Issue):W557–W559.

Hofmann, C. & Bouck, G. B. 1976. Immunological and structural evi-dence for patterned intussusceptive surface growth in a unicellular or-ganism. J. Cell Biol., 69:693–715.

Huber-Pestalozzi, G. 1955. Das Phytoplankton des Su�wassers; System-atik und Biologie: 4 Teil; Euglenophyceen. E. SchweizerbartscheVerlagsbuchhandlung, Stuttgart, Germany, 606 p.

Huelsenbeck, J. P. & Ronquist, F. 2001. MRBAYES: Bayesian inferenceof phylogenetic trees. Bioinformatics, 17:754–755.

Keeling, P. J. 2002. Molecular phylogenetic position of Trichomitopsistermopsidis (Parabasalia) and evidence for the Trichomitopsiinae. Eur.J. Protistol., 38:279–286.

Kosmala, S., Bereza, M., Milanowski, R., Kwiatowski, J. & Zakrys, B.2007. Morphological and molecular examination of relationships andepitype establishment of Phacus pleuronectes, Phacus orbicularis, andPhacus hamelii. J. Phycol., 43:1071–1082.

Leander, B. S. 2004. Did trypanosomatid parasites have photosyntheticancestors? Trends Microbiol., 12:251–258.

Leander, B. S. & Farmer, M. A. 2000a. Comparative morphology of theeuglenid pellicle. I. Patterns of strips and pores. J. Eukaryot. Microbiol.,47:469–479.

Leander, B. S. & Farmer, M. A. 2000b. Epibiotic bacteria and a novelpattern of strip reduction on the pellicle of Euglena helicoideus (Ber-nard) Lemmermann. Eur. J. Protistol., 36:405–413.


Leander, B. S. & Farmer, M. A. 2001a. Comparative morphology of theeuglenid pellicle. II. Diversity of strip substructure. J. Eukaryot. Mi-crobiol., 48:202–217.

Leander, B. S. & Farmer, M. A. 2001b. Evolution of Phacus (Eugleno-phyceae) as inferred from pellicle morphology and SSU rDNA. J. Phy-col., 37:143–159.

Leander, B. S., Esson, H. J. & Breglia, S. A. 2007. Macroevolution ofcomplex cytoskeletal systems in euglenids. BioEssays, 29:987–1000.

Leander, B. S., Witek, R. P. & Farmer, M. A. 2001. Trends in the evolutionof the euglenid pellicle. Evolution, 55:2215–2235.

Linton, E. W., Nudelman, M. A., Conforti, V. & Triemer, R. E. 2000. Amolecular analysis of the euglenophytes using SSU rDNA. J. Phycol.,36:740–746.

Maddison, D. R. & Maddison, W. P. 2000. MacClade. Sinauer AssociatesInc., Sunderland, MA.

Marin, B., Palm, A., Klingberg, M. & Melkonian, M. 2003. Phylogeny andtaxonomic revision of plastid-containing euglenophytes based on SSUrDNA sequence comparisons and synapomorphic signatures in the SSUrRNA secondary structure. Protist, 154:99–145.

Mignot, J. P., Brugerolle, G. & Bricheux, G. 1987. Intercalary strip de-velopment and dividing cell morphogenesis in the euglenid Cyclidiopsisacus. Protoplasma, 139:51–65.

Mullner, A. N., Angeler, D. G., Samuel, R., Linton, E. W. & Triemer, R. E.2001. Phylogenetic analysis of phagotrophic, photosynthetic, andosmotrophic euglenoids by using the nuclear 18S rDNA sequence.Int. J. Syst. Evol. Microbiol., 51:783–791.

Nudelman, M. A., Rossi, M. S., Conforti, V. & Triemer, R. E. 2003. Phy-logeny of Euglenophyceae based on small subunit rDNA sequences:taxonomic implications. J. Phycol., 39:226–235.

Pochmann, A. 1942. Synopsis der Gattung Phacus. Arch Protistenkd.,95:81–252.

Pringsheim, E. G. 1946. The biphasic or soil-water culture method forgrowing algae and flagellata. J. Ecol., 33:193–204.

Rodrıguez, F., Oliver, J. L., Marin, A. & Medina, J. R. 1990. The generalstochastic model of nucleotide substitution. J. Theor. Biol., 142:485–501.

Ronquist, F. & Huelsenbeck, J. P. 2003. MRBAYES 3: Bayesianphylogenetic inference under mixed models. Bioinformatics, 19:1572–1574.

Shimodaira, H. & Hasegawa, M. 2001. CONSEL: for assessing theconfidence of phylogenetic tree selection. Bioinformatics, 17:1246–1247.

Stamatakis, A., Ludwig, T. & Meier, H. 2005. RAxML-III: a fast programfor maximum likelihood-based inference of large phylogenetic trees.Bioinformatics, 21:456–463.

Triemer, R. E., Linton, E., Shin, W., Nudelman, A., Monfils, A., Bennet,M. & Brosnan, S. 2006. Phylogeny of the Euglenales based uponcombined SSU and LSU rDNA sequence comparisons and descriptionof Discoplastis gen. nov. (Euglenophyta). J. Phycol., 42:731–740.

Received: 03/26/09, 07/24/09; accepted: 07/26/09


Documents

J. Eukaryot. Microbiol., 57(1), 2010 pp. 19–32 2009 The Author(s) … · 2010. 1. 4. · shapes observed in Phacus affect the whorled patterns of pellicle strips that have been