Galeidinium parvulum, Kutzing 1849 (Kutz.)

Novis, Phil M., Braidwood, Jasmine & Kilroy, Cathy, 2012, Small diatoms (Bacillariophyta) in cultures from the Styx River, New Zealand, including descriptions of three new species, Phytotaxa 64 (1), pp. 11-45 : 26-28

Gomphonema parvulum Kützing 1849 ( Figs 65–75 , 163 )

Frustules lanceolate-clavate in valve view, 5.5–5.8 µm wide, 19.5–21.8 µm long, asymmetrical to the transapical axis, biraphid, with broadly subrostrate headpole and footpole. Apical pore field present at footpole on both sides of raphe terminus ( Fig. 74 ). Central area elliptical, bounded by one short stria at each valve margin, with single stigma to one side. Stigma opening round externally, elongate internally ( Fig. 73 ). Striae radiate, uniseriate, 12–14 in 10 µm, composed of reniform areolae. Raphes straight to slightly undulate, with slightly widened central termini and apical termini weakly hooked towards the secondary valve side.

Reference: — Krammer & Lange-Bertalot (1986), p.358, fig. 154:1–25.

Specimens examined: —CHR618413! (cleaned frustules from cultured strain LCR-S:22:1).

Distribution: — G. parvulum and G. affine Kützing 1844 are regarded as widespread and cosmopolitan, and appeared in the Styx in cultures from site 5. Both names are recorded previously from New Zealand.

Molecular data:—The closest matches to the 18S sequence of the Styx strain were G. affine strains AT-196Gel03 and AT-219Gel06, and G. parvulum strain UTEX FD241 (p-distance = 0.001). The first two strains were isolated from a lake and river in Germany ( Bruder & Medlin 2008) and the last from South Carolina. However, four strains of Cymbella Agardh 1830 were almost as close (pdistance = 0.002), and the Styx strain formed a robust clade with all these strains. Relationships within the clade could not be resolved with the 18S data alone ( Fig. 163 ). The length of the 18S fragment of the Styx strain was 849 bp, in a total dataset 1765 bp long, with 66 variable sites (55 parsimony informative, with 22 occurring within the Styx fragment). The variation in the 18S gene was thus lower in Gomphonema Ehrenberg 1832 than in the other taxa studied. The model chosen and implemented in the Bayesian analysis was TN93+I. The closest match to the rbc L sequence of the Styx strain was the rbc L pseudogene of G. affine strain AT-196Gel03; the two fragments were identical over their 295 bp region of overlap (except for the stop codon in AT-196Gel03, which was removed before analysis). The analyses of rbc L placed the Styx strain in a robust polytomy with G. affine strains AT-196Gel03 and AT-219Gel06 (not shown). The length of the Styx rbc L fragment was 575 bp, in a total dataset 1473 bp long, with 133 variable sites (69 parsimony informative, 47 occurring within the Styx fragment). The model chosen and implemented in the Bayesian analysis was T92+G.

When the 18S and rbc L data were combined into a single dataset, the closest match to the Styx strain was G. affine strain AT-196Gel03 (p-distance = 0.001). The Styx strain formed a robust clade with this strain and G. affine strain AT-219Gel06 ( Fig. 163 ). G. parvulum UTEX FD 241 emerged as sister to the G. affine clade. However, this G. affine / G. parvulum clade was nested within a further clade containing G. affine UTEX FD 173. A further strain, G. cf. parvulum AT-161.15, was sister to this clade. Consequently, we were not able to resolve the strains of G. affine and G. parvulum currently accessioned in Genbank as distinct species. The combined dataset failed the partition homogeneity test (P=0.021). However, a minimum conclusion is that the individual genes both robustly supported the grouping of the Styx strain (which has characteristic features of G. parvulum ) with two strains of G. affine to the exclusion of other G. parvulum strains. Examination of the combined dataset represented as a neighbournet graph in Splitstree 4 ( Huson & Bryant 2006) showed the combined data to be strongly treelike.

Comments: —This strain most closely resembles G. parvulum in size ( G. parvulum : 10–36 µm long, 4–8 µ m wide; G. affine : 30–100 µ m long, 7–12 µm wide; Krammer & Lange-Bertalot 1986) and shape ( G. parvulum has a subrostrate headpole, whereas G. affine has a rounded headpole; Krammer & Lange-Bertalot 1986). Taken at face value it appears that these features do not describe monophyletic groups of Gomphonema. Our analyses of rbc L sequences, plus a combined dataset with 18S, demonstrated an affinity of our strain with others identified as G. affine . Vouchers of Theriot et al. (2010), accessible at http:// www.protistcentral.org/index.php/Taxa/get/taxa_id/97227, indicate that strains G. affine UTEX FD 173 and G. parvulum UTEX FD 241 correspond to their classical morphologies as shown by Krammer & Lange-Bertalot (1986), although we note that the latter authors illustrate considerable variation in morphology of the valve poles within G. parvulum . Potapova & Charles (2007) regarded G. parvulum as indicative of high total N and total P, and G. affine as indicative of low total P in the USA; N and P in the Styx River are much more likely to be high than low. Thus, classical morphology and ecology place our specimens in G. parvulum , whereas molecular data suggest G. affine . Unfortunately, we have been unable to confirm the identification of strains AT-196Gel03 and AT-219Gel06, listed as G. affine by Bruder & Medlin (2008).

Rather than suggest that these two species will need to be merged (as G. affine , which has nomenclatural priority) as part of a genus-wide revision, we suggest that further data are needed to properly examine this question. Great variation in valve shape is well known in Gomphonema, with some species having significant differences in morphology between vegetative cells and auxospores, although these have been interpreted as reflecting vulnerability of morphology to environmental influence during the development of initial valves ( Kociolek & Stoermer 1988; Passy-Tolar & Lowe 1995; Kociolek & Stoermer 2010). Even if the strains AT- 196Gel03 and AT-219Gel06 are referable to G. parvulum based on morphology, there is considerable genetic variation within the species ( Fig. 163 ).

The rbc L fragment from G. affine strain AT-196Gel03, which the Styx strain most closely matches, is described as a pseudogene; however, it is not possible to determine whether the fragment from the Styx strain is also a pseudogene, since there are no stop codons over the region sequenced (including the position in the AT-196Gel03 pseudogene where the stop codon occurs; this was removed prior to analysis). In the absence of evidence to the contrary, we assume our fragment is a functional gene.

Order Naviculales

Family Naviculaceae

Figures

FIGURE 163. Bayesian phylogenies of the Gomphonema strain from the Styx River, for 18S rDNA (left) and concatenated rbcL and 18S rDNA (right). Support numbers above branches are Bayesian posterior probabilities / Maximum Parsimony bootstrap values. Scale bar = 0.1 subsitutions/site. See text for substitution models and description of data sets.

FIGURE 65–75. Gomphonema parvulum, strain LCR-S:22:1, CHR618413. Scales: Fig. 65, 10 µm; Fig. 66, 5 µm (use for Figs 66–71), all others 1 µm. Fig. 65. Live cells in valve and girdle views by LM. Figs 66–71. Cleaned frustules viewed by LM, with single stigma visible in some valves (girdle view in Fig. 71). Figs 72–75. SEM of cleaned valves. Fig. 72, valve exterior showing stigma; Fig. 73, valve interior showing helictoglossae; Fig. 74, external view of footpole showing apical porefield; Fig. 75, internal view of footpole showing apical porefield and helictoglossa.