Molecular Investigation of the Achnanthidium minutissimum Complex (Bacillariophyceae) from the Transbaikal Region with the Description of Three New Species

Abstract

Representatives of the Achnanthidium minutissimum complex have been studied from the Transbaikal region, Russia. The Transbaikal region is known for its unique algal flora and high levels of endemism. Despite the group of monoraphid diatoms being studied quite extensively in this region in recent years, only four species of Achnanthidium have been reported so far. Three new species are described based on LM and SEM microphotographs and molecular data: Achnanthidium baicalonanum sp. nov., A. obscurum sp. nov. and A. angustum sp. nov. Comparison with similar species is given. The new species differ from other species of the genus by valve shape, striae and areolae structure, shape of external distal raphe ends and central area. According to molecular data, the new species form separate branches within the A. minutissimum complex. Ecological preferences of the new species are examined. All of the new species are only known from the Transbaikal region. This study contributes to the understanding of Achnanthidium taxonomy and biogeography and will be an asset to future biomonitoring studies.

1. Introduction

Lake Baikal is the largest freshwater lake in the world, well-known for its unique environmental conditions and high level of endemism. The diverse and distinctive diatom flora of the lake has been a focus of many studies since the early 20th century [1]. The non-monophyletic group of monoraphid diatoms has been studied quite extensively in Lake Baikal during the last two decades, with descriptions of many new species and several new genera, such as Trifonovia Kulikovskiy et Lange-Bertalot [2], Gliwiczia Kulikovskiy, Lange-Bertalot et Witkowski [3], Skabitschewskia Kulikovskiy et Lange-Bertalot [4], Platebaikalia Kulikovskiy, Glushchenko, Genkal et Kociolek [5], Gololobovia Kulikovskiy, Glushchenko, Genkal et Kociolek [6] and Platesiberia Kulikovskiy, Glushchenko, Genkal et Kociolek [7]. A comprehensive review and revision of diatoms from Lake Baikal was published in the 2010s [2,4], in which many monoraphid taxa were reported and illustrated, both previously known and new. The genus Achnanthidium Kützing, however, has been quite absent from these studies, despite being one of the most widespread freshwater diatom genera in the world [8]; only a handful of species have been reported in literature, namely A. minutissimum (Kützing) Czarnecki [9,10], A. affine (Grunow) Czarnecki [10], A. exile (Kützing) Heiberg [11], and A. sibiricum Kulikovskiy, Lange-Bertalot, Witkowski et Khursevich [1]. It is very likely that the diversity of Achnanthidium in Lake Baikal is still underestimated.

Achnanthidium is one of the most abundant and widespread freshwater diatom genera in the world [8]. Its definitive features are small, linear-lanceolate to lanceolate-elliptic valves, uniseriate striae that are often denser at the valve apices, external distal raphe ends that are straight or curved to one side, and a row of elongated areolae on the mantle [12]. Representatives of this genus are generally divided into two groups by morphology: the A. minutissimum complex that is characterized by mostly straight external distal raphe ends, and the A. pyrenaicum (Hustedt) Kobayasi complex in which the distal raphe ends are strongly curved to one side.

The A. minutissimum species complex has been drawing the attention of taxonomists for some time already. It is considered to be one of the most widespread freshwater diatoms in the world [13]. It is also frequently found in epilithic biofilms and studies have been performed that show the formation of extracellular insoluble carbohydrate capsules in A. minutissimum being induced by bacteria [14]. The taxonomy of this group is fairly problematic due to the small size of the cells, hard-to-resolve ultrastructure, unclear species boundaries, and the general scarcity of molecular data for representatives of this genus. It has been repeatedly suggested that the A. minutissimum complex in fact comprises many separate species with their own ecological preferences and biogeography. In recent years, many new species from this group have been described from various parts of the world [1,8,15,16,17,18,19,20,21,22] and the status of some taxa has been reassessed [8]. The type material of several established species has also been reexamined, including several examinations of A. minutissimum s.str. [8,16,23,24,25,26], and there have been studies that used morphometric [23] and molecular [13] methods to gain a better understanding of the taxonomy of this complex. Still, this issue is far from being completely resolved. Detailed investigations of Achnanthidium in less-studied regions such as the Transbaikalia are important for our understanding of the taxonomic composition of this genus; since this location is known for high levels of endemism in diatoms, the likelihood of finding new species is quite high as well. Discoveries of new species and descriptions of their morphology, phylogenetic relationships with other species, and environmental preferences are also relevant for biomonitoring studies of the Transbaikal region’s unique ecosystem.

Previous findings of Achnanthidium in the Transbaikal region were mostly registered during floristic studies. This study is one of the first taxonomic analyses of this genus in the region of Lake Baikal; moreover, there had not previously been any molecular investigations of Achnanthidium representatives in this region, which is especially important for small-celled species and difficult species complexes. The goal of this study is to provide a morphological and molecular analysis of several strains of Achnanthidium isolated from the Transbaikal region with the description of three new species; to gain information on the taxonomy of the A. minutissimum species complex and to contribute to the study of the unique flora of Lake Baikal and its vicinity.

2. Materials and Methods

Sampling. The samples for this study were collected in 2012 and 2014 in the Transbaikal region, Russia. The full list of strains with the relevant information is given in

Table 1. List of strains examined in this study.

Strain	Sampling Locality	Collection Date	Latitude	Longitude	pH	Water Temperature, °C	Sample Type	Slide No.
Achnanthidium baicalonanum sp. nov. B387	Lake Frolikha	26 July 2012	55°26′15.2″ N	110°01′16.1″ E	N/A	18.2	benthos	B387
Achnanthidium baicalonanum sp. nov. B397	puddle by the road on the way from Lake Ilchir	7 July 2014	51°56′49.3″ N	100°49′34.8″ E	7.66	11.6	benthos	B397
Achnanthidium baicalonanum sp. nov. B398	puddle by the road on the way from Lake Ilchir	7 July 2014	51°56′49.3″ N	100°49′34.8″ E	7.66	11.6	benthos	B398
Achnanthidium cf. minutissimum B443	Lake Nizhnee Vesy	19 July 2014	55°27′29.9″ N	110°16′08.2″ E	5.02	13.9	plankton	B443
Achnanthidium angustum sp. nov. B522	Ehe-Ugui river	5 June 2014	51°41′66″ N	101°40′547″ E	7.90	9.5	epilithic periphyton	B522 (01313)
Achnanthidium angustum sp. nov. B543	Ehe-Ugui river	5 June 2014	51°41′66″ N	101°40′547″ E	7.90	9.5	epilithic periphyton	B543 (01334)
Achnanthidium obscurum sp. nov. B722	Levaya Frolikha river	21 July 2014	55°22′57.2″ N	110°07′43.7″ E	N/A	12.0	sand near the river bank with detritus	B722 (01512)
Achnanthidium obscurum sp. nov. B723	Levaya Frolikha river	21 July 2014	55°22′57.2″ N	110°07′43.7″ E	N/A	12.0	sand near the river bank with detritus	B723 (01513)

.

Culturing. A part of each sample was transferred to the WC liquid culturing medium [27]. Monoclonal strains were established by micropipetting a single cell using an inverted microscope Axio Vert. A1 (Zeiss, Oberkochen, Germany). Non-axenic unialgal cultures were maintained in the WC culturing medium at 22–25 °C in a growth box with a 12:12 h light:dark photoperiod. Microscopical investigation and DNA extraction was performed after one month of culturing.

Preparation of slides and microscope investigation. Cells for LM and SEM investigations were processed with a standard procedure that involves treatment with concentrated hydrogen peroxide. Distilled water was used to wash the material. Permanent diatom preparations were mounted in Naphrax^® mounting medium (Naphrax^®, Brunel Microscopes Ltd., Chippenham, UK). Light microscopic (LM) observations were made using the microscope AxioScope A1 (Zeiss, Oberkochen, Germany) with an oil immersion objective (×100/n.a.1.4, DIC). Ultrastructure of the valves was examined with the scanning electron microscope JSM-6510LV (Jeol, Tokyo, Japan).

Molecular study. Genomic DNA of the studied strains B387, B397, B398, B443, B522, B543, B722, and B723 was extracted from fresh cultures by Chelex 100 Chelating Resin (Bio-Rad Laboratories, Hercules, CA, USA) using protocol 2.2. One nuclear gene and one plastid gene were amplified (18S rDNA and rbcL, respectively). For the highly variable V4 region of 18S rDNA (432–434 bp), D512for and D978rev primers were used [28]. The plastid rbcL (1260–1431 bp) was amplified using rbcL66+ [29] and dp7- [30] primers.

PCR amplifications were performed using premade mastermixes (ScreenMix by Evrogen, Moscow, Russia). The amplification of the 18S rDNA was performed using the following program: 5 min of denaturation at 95 °C; next, 35 cycles of denaturation at 94 °C (30 s), annealing at 52 °C (30 s), and elongation at 72 °C (50 s); and a final extension at 72 °C (10 min), subsequently held at 12 °C. The amplification of the rbcL gene was performed using the following program: 4 min of denaturation at 94 °C; followed by 44 cycles of denaturation at 94 °C (50 s), annealing at 53 °C (50 s), and elongation at 72 °C (80 s); with a final extension at 72 °C (7 min), subsequently held at 12 °C.

The PCR products were sized on a 1.0% agarose gel stained with SYBR^TM Safe (Life Technologies, Waltham, MA, USA) and then cleaned using FastAP, 10× FastAP Buffer and Exonuclease I (Thermo Fisher Scientific, Waltham, MA, USA), mixed with water. The purified PCR products were sequenced by Sanger Sequencing method using a Genetic Analyzer 3500 instrument (Applied Biosystems, St. Louis, MO, USA).

Newly obtained sequences were manually edited in Ridom TraceEdit ver. 1.1.0 (Ridom GmbH, Münster, Germany) and Mega ver. 7 software [31]. The reads were supplemented with GenBank-extracted sequences of 44 diatom species and two outgroup Placoneis Mereschkowsky species (taxa names and Accession Numbers are given in Figure 1). The 18S rDNA and rbcL nucleotide sequences were aligned separately using the G-INS-i algorithm in the Mafft ver. 7 software (RIMD, Osaka, Japan) [32]. The resulting data set comprised 437 and 1431 nucleotide sites for nuclear 18S rDNA and plastid rbcL regions, respectively. After removal of the unpaired regions, the aligned 18S rRNA gene sequences were combined with the rbcL gene sequences into a single matrix for the concatenated 18S rDNA and rbcL tree.

The Bayesian inference (BI) method was performed to infer the phylogenetic position of B387, B397, B398, B443, B522, B543, B722, and B723 strains using BEAST ver. 1.10.1 software (BEAST Developers, Auckland, New Zealand) [33]. The most appropriate partition-specific substitution models, a proportion of invariable sites (pinvar) and shape parameter α were recognized by the Bayesian information criterion (BIC) in jModelTest ver. 2.1.10 software (Vigo, Spain) [34]. The following models, shape parameter α and a proportion of invariable sites (pinvar) were selected by the BIC-based model selection procedure: TrN+I+G, α = 0.3570 and pinvar = 0.7170 for 18S rDNA; TPM1uf+I+G, α = 0.6410 and pinvar = 0.8480 for the first codon position of the rbcL gene; JC+I, pinvar = 0.9290 for the second codon position of the rbcL gene; and TVM+G α = 0.5430 for the third codon position of the rbcL gene. However, the HKY model was applied instead of TrN and TPM1uf, the F81 applied instead of JC, and the GTR applied instead of TVM as the most similar applicable options for BI. A speciation model was performed by a Yule process tree prior. Five MCMC analyses were run for 5 million generations (burn-in 1000 million generations). The convergence diagnostics was performed in the Tracer ver. 1.7.1 software (MCMC Trace Analysis Tool, Edinburgh, UK) [33]. The initial 15% trees were removed and the rest retained to construct a final chronogram with 90% posterior probabilities. The robustness of tree topologies was assessed by bootstrapping the data set with Maximum Likelihood (ML) analysis using RAxML software [35]. The ML bootstrapping was performed with 1000 replicas. Trees were viewed and edited using FigTree ver. 1.4.4 (University of Edinburgh, Edinburgh, UK) and Adobe Photoshop CC ver. 19.0.

3. Results

On the phylogenetic tree based on concatenated 18S rRNA and rbcL genes sequence data (Figure 1) and in the individual trees for 18S rDNA (Figure S1) and the rbcL separately (Figure S2), all of our strains are included in various subclades of the Achnanthidium minutissimum complex. Strains B522 and B543 form a separate lineage, which with high support (LB 90, PP 0.99) is a sister to clade lineage, which includes strains of A. minutissimum complex isolated from Spitsbergen. Strain B443 also forms a separate lineage (LB 95, PP 1) with strain A. minutissimum AT-196Gel02 from Germany. In the phylogenetic tree strains B398, B387 and B397 occupy a separate position and form single clade with maximum statistical support (LB 99, PP 1). With high support (LB 97, PP 1) strains B722 and B723 form a common clade with strains A. minutissimum complex MIC10-53 and MIC10-61 from Marion Island (sub-Antarctica).

The morphological analysis carried out with the use of light and scanning electron microscopy, together with the molecular data, allows us to describe three new species of the genus Achnanthidium.

Achnanthidium baicalonanum Tseplik, Genkal, Maltsev et Kulikovskiy sp. nov. (Figure 2).

Holotype. Slide no. B387, in collection of Maxim Kulikovskiy, Institute of Plant Physiology, Russian Academy of Sciences, Russia, represented here by Figure 2C.

Reference strain. B387, isolated from sample 75/2.

Type locality. Russia, Lake Frolikha, sample 75/2, benthos (55°26′15.2″ N 110°01′16.1″ E), 26.07.2012.

Description. LM. Valves small, linear-elliptic to elliptic, with broadly rounded ends. Length 5.8–8.7 μm, width 2.1–3.4 μm. Raphe straight, filiform. Axial area on the raphe valve narrow, linear, central area indistinct, often asymmetrical, formed by one or two central striae being shortened or absent on one or both sides of the valve. Axial area on the rapheless valve also narrow, linear, central area absent. Striae on both valves radiate, may be hard to resolve in LM.

SEM. Raphe valve. External central raphe ends straight, external distal raphe ends straight or slightly bent to one side; sometimes only one distal end is slightly bent while the other is straight. Internally, central raphe ends are bent to opposite sides, distal raphe ends terminate in weakly developed helictoglossae. Two or three striae in the center of the valve are usually spaced wider than on the rest of the valve, in addition to central striae being shortened. Striae uniseriate, radiate, 27–39 in 10 μm. Areolae rounded or transapically elongated, 42–60 in 10 μm. Rapheless valve. Striae uniseriate, weakly radiate, 31–35 in 10 μm. Areolae rounded or transapically elongated, 47–56 in 10 μm. There are 4–5 areolae per stria on both valves, sometimes three on the valve ends.

Sequence data. Nuclear-encoded 18S rDNA partial sequence (GenBank accession OR543355 for strain B387, OR543353 for strain B397, OR543356 for strain B398), plastid gene rbcL partial sequence (GenBank accession OR545508 for strain B387, KR709273 for strain B397, KR709274 for strain B398).

Etymology. The epithet refers to the type locality—the Transbaikal region, and the similarity of the new species to Achnanthidium nanum (F. Meister) Novais et Jüttner.

Distribution. As yet known only from type locality.

Achnanthidium obscurum Tseplik, Genkal, Maltsev et Kulikovskiy sp. nov. (Figure 3).

Holotype. Slide no. B723, in collection of Maxim Kulikovskiy, Institute of Plant Physiology, Russian Academy of Sciences, Russia, represented here by Figure 3B.

Reference strain. B723 (01513), isolated from sample 172.

Type locality. Russia, Levaya Frolikha river, sample 172, sand near the river bank with detritus (55°22′57.2″ N 110°07′43.7″ E), 21.07.2014.

Description. LM. Valves small, linear to linear-lanceolate, with slightly protracted and rounded ends. Length 10.1–12.0 μm, width 2.5–3.3 μm. Raphe straight, filiform. Axial area on raphe valve narrow, linear, central area asymmetrical. Axial area on rapheless valve also narrow, linear, central area absent. Striae not resolvable in LM.

SEM. Raphe valve. Central area approximately rectangular, sometimes shortened striae are present on one side of the central area. External central raphe ends straight, teardrop-shaped, external distal raphe ends slightly bent to one side. Internal central raphe ends bent to opposite sides, internal distal raphe ends terminate in helictoglossae. Striae uniseriate, radiate, 31–33 in 10 μm. Areolae close to the axial area rounded, close to valve margin transapically elongated, 44–53 in 10 μm. Rapheless valve. Striae uniseriate, radiate, 31–36 in 10 μm. Areolae rounded or transapically elongated, 45–51 in 10 μm. There are usually 3–5 areolae per stria on both valves.

Sequence data. Nuclear-encoded 18S rDNA partial sequence (GenBank accession OR543359 for strain B722, OR543354 for strain B723), plastid gene rbcL partial sequence (GenBank accession OR545511 for strain B722, OR545512 for strain B723).

Etymology. The epithet was given because the new species is quite similar to other species of the genus and can be easily misidentified or overlooked.

Distribution. As yet known only for type locality.

Achnanthidium angustum Tseplik, Genkal, Maltsev et Kulikovskiy sp. nov. (Figure 4).

Holotype. Slide no. B543, in collection of Maxim Kulikovskiy, Institute of Plant Physiology, Russian Academy of Sciences, Russia, represented here by Figure 4D.

Reference strain. B543 (01334), isolated from sample 14-3.

Type locality. Russia, Ehe-Ugui river, sample 14-3, epilithic periphyton (51°41′66″N 101°40′547″E), 05.06.2014.

Description. LM. Valves narrow linear-lanceolate, with protracted and rounded ends. Length 12.3–14.8 μm, width 2.4–3.1 μm. Raphe straight, filiform. Axial area on the raphe valve narrow linear, very slightly widened towards the central area, central area formed by wider spaced or shortened striae in the center of the valve. Axial area on the rapheless valve narrow lanceolate, central area absent. Striae are hard to resolve in LM.

SEM. Raphe valve. External distal and central raphe ends are straight, distal ends are quite long, extending past the last striae almost to the junction between the valve face and mantle. Internally, central raphe ends are slightly bent to opposite sides and distal ends terminate in helictoglossae. Striae uniseriate, radiate, 31–37 in 10 μm. Areolae small, rounded or transapically elongated, 48–49 in 10 μm. Rapheless valve. Striae uniseriate, radiate, 31–35 in 10 μm. Areolae also small, rounded or transapically elongated, 43–48 in 10 μm. There are typically 3–4 areolae in a single stria on both valves.

Sequence data. Nuclear-encoded 18S rDNA partial sequence (GenBank accession OR543357 for strain B522, OR543358 for strain B543), plastid gene rbcL partial sequence (GenBank accession OR545509 for strain B522, OR545510 for strain B543).

Etymology. The epithet refers to the slender valve shape and narrow tapered ends of the new species.

Distribution. As yet known only from type locality.

4. Discussion

All our new species clearly belong to the genus Achnanthidium, as they possess the definitive traits of the genus [12]: small linear-lanceolate valves, radiate uniseriate striae and straight external distal raphe ends. The shape of the external distal raphe ends which are straight or only slightly bent in one direction places the new species in the A. minutissimum species complex.

Achnanthidium baicalonanum sp. nov. is a small-celled species that can be hard to identify in LM. There are several Achnanthidium species with small elliptical valves that the new species may be confused with (see

Table 2. Comparison of Achnanthidium baicalonanum sp. nov. and similar species.

	Achnanthidium baicalonanum sp. nov.	A. straubianum	A. nanum	A. atomoides	A. duriense
Valve shape	linear-elliptic to elliptic	elliptic	linear to linear-elliptic	linear-elliptic	elliptic to linear-elliptic
Valve apices	broadly rounded	broadly rounded	broadly rounded to slightly subrostrate	broadly rounded	broadly rounded to slightly protracted
Valve length, μm	5.8–8.7	6.5–8.5	6.4–10.9	4.3–11.6	5.0–9.7
Valve width, μm	2.1–3.4	2.6–3.7	1.9–2.8	2.2–3.2	2.0–2.7
Raphe valve
Distal raphe ends	straight or slightly bent to one side	straight	straight	more or less straight	straight or slightly bent to one side
Axial area	narrow linear	narrow lanceolate	narrow linear	narrow linear	narrow linear
Central area	indistinct and asymmetrical	almost absent	indistinct	bowtie-shaped	indistinct
Striae pattern	radiate	weakly radiate to parallel	radiate	radiate and curved	almost parallel to weakly radiate
Striae density in 10 μm	27–39	27–30	28–36	25–32	35
Areolae	rounded to transapically elongated	rounded to elongated to rectangular	rounded to elongated	rounded	rounded or square
Rapheless valve
Axial area	narrow linear	narrow lanceolate	narrow linear	narrow linear	narrow linear
Central area	absent	almost absent	indistinct	absent	indistinct
Striae pattern	weakly radiate	weakly radiate to parallel	parallel to weakly radiate	parallel	almost parallel to weakly radiate
Striae density in 10 μm	31–35	27–30	28–36	23–28	35
Areolae	rounded to transapically elongated	rounded to elongated to rectangular	rounded to elongated	rounded	rounded or square, sometimes slit-like
Source	this study	[24]	[8]	[15]	[8]

). The first similar species is A. nanum [8], it can be distinguished from A. baicalonanum sp. nov. by a more lanceolate shape of the raphe valve and more radiate striae on the rapheless valve. Achnanthidium nanum also has a higher maximum length than A. baicalonanum sp. nov. (10.9 μm in the type material of A. nanum, 15.7 μm in samples from Portugal [8] vs. 8.4 μm in A. baicalonanum sp. nov.); however, all the quantitative features overlap in these two species and cannot be reliably used to differentiate them. External distal raphe ends in both species are straight, but in A. baicalonanum sp. nov. they appear longer, reaching the junction between valve face and mantle, while in A. nanum they do not reach the edge of the valve face. Achnanthidium nanum also does not exhibit slit-like areolae on the valve face, while in A. baicalonanum sp. nov. they can be seen.

Another similar species is A. straubianum (Lange-Bertalot) Lange-Bertalot [24]. The most obvious distinguishing feature is the valve shape: the valves of A. straubianum are generally wider elliptical, while A. baicalonanum sp. nov. has a more linear-elliptical shape. The striae on the raphe valve of A. baicalonanum sp. nov. are spaced denser than in A. straubianum (32–39 vs. 27–30 in 10 μm, respectively); the striae on both valves of A. straubianum are usually well resolvable in LM, which is not always the case in A. baicalonanum sp. nov. The striae on the raphe valve of A. straubianum are weakly radiate or almost parallel, while in A. baicalonanum sp. nov. they are distinctly radiate. Achnanthidium straubianum also may have two types of areolae on the mantle—elliptical and slit-like; however, this feature is variable [24] and thus cannot be used to reliably distinguish this species.

Two other species that can be confused with A. baicalonanum sp. nov. are A. atomoides Monnier, Lange-Bertalot et Ector [15] and A. duriense Novais et Ector [8]. The first can be distinguished from A. baicalonanum sp. nov. by the striation patterns: A. atomoides has distinctly parallel striae on the rapheless valve and radiate and curved striae on the raphe valve. The striae density on both valves of A. atomoides is lower than in A. baicalonanum sp. nov. (25–32 in 10 μm on the raphe valve, 23–28 in 10 μm on the rapheless valve). Achnanthidium atomoides also has quite characteristic small, rounded areolae on both valves, while A. baicalonanum sp. nov. has rounded or transapically elongated areolae typical for many Achnanthidium species. Lastly, A. baicalonanum sp. nov. does not have such a clear bowtie-shaped central area on the raphe valve as A. atomoides. Achnanthidium duriense has a more distinctly lanceolate valve shape than A. baicalonanum sp. nov., with often slightly protracted ends which is never the case in A. baicalonanum sp. nov. External distal raphe ends in A. duriense are expanded and teardrop-shaped, while in A. baicalonanum sp. nov. they are simple and not expanded.

The most similar species to Achnanthidium obscurum sp. nov. is A. lineare W. Smith [25] (

Table 3. Comparison of Achnanthidium obscurum sp. nov. and similar species.

	Achnanthidium obscurum sp. nov.	A. lineare	A. minutissimum	A. maritimo-antarcticum	A. indistinctum
Valve shape	linear to linear-lanceolate	linear to narrow lanceolate	linear-lanceolate	linear-lanceolate	narrow lanceolate
Valve apices	slightly protracted and rounded	broadly rounded	subrostrate to subcapitate	rostrate to subcapitate	rostrate
Valve length, μm	10.1–12.0	9.0–13.5	8.9–18.5	12–15	8.5–13.0
Valve width, μm	2.5–3.3	2.2–2.8	2.5–3.0	2.3–2.7	1.8–2.2
Raphe valve
Distal raphe ends	straight, teardrop-shaped	long, almost straight	straight or slightly bent	straight, short	straight, short
Axial area	narrow linear	narrow linear	narrow linear	linear	narrow linear
Central area	asymmetrical, more or less rectangular	broad rectangular fascia	variable	irregular	small, indistinct
Striae pattern	radiate	radiate	radiate	weakly radiate	radiate
Striae density in 10 μm	31–33	28–32	28–32	30–33	36
Areolae	rounded to transapically elongated	rounded to slit-like	rounded to transapically elongated	rounded to square to rectangular	rounded to elongated to rectangular
Rapheless valve
Axial area	narrow linear	narrow lanceolate	narrow linear	narrow linear	narrow linear
Central area	absent	almost absent to small elliptical	absent	weakly elliptical	almost absent
Striae pattern	radiate	radiate	radiate	weakly radiate	radiate
Striae density in 10 μm	31–36	28–32	28–32	30–32	30–35
Areolae	rounded to transapically elongated	rounded to slit-like	rounded to transapically elongated	rounded to square to rectangular	rounded to elongated to rectangular
Source	this study	[25]	[23]	[20]	[20]

). It has a similar size and shape to the new species and may be confused with it in LM. It can be differentiated from A. obscurum sp. nov. by a broad rectangular fascia on the raphe valve, which the new species does not have. Achnanthidium lineare also has only 1–3 areolae per stria on both valves, while A. obscurum sp. nov. tends to have 3–5. The external distal raphe ends in A. lineare are long and mostly straight, in A. obscurum sp. nov. they are slightly bent to one side.

Achnanthidium obscurum sp. nov. is also similar to the type of A. minutissimum [23]. These species can be distinguished from each other by valve shape: A. minutissimum has linear-lanceolate valves with rostrate ends, A. obscurum sp. nov. has elliptic-lanceolate valves with only slightly protracted ends. The maximum length is higher in A. minutissimum (18.5 μm vs. 12.0 μm in A. obscurum sp. nov.); however, the quantitative features overlap and thus cannot be a reliable differentiating feature. The striae in A. minutissimum are visibly denser spaced on the valve ends than at the center, which does not appear to be the case in the new species.

Achnanthidium obscurum sp. nov. can also be compared to two species that were described from the Antarctic region, A. maritimo-antarcticum Van de Vijver et Kopalová and A. indistinctum Van de Vijver et Kopalová [20]. Achnanthidium maritimo-antarcticum has longer and slenderer valves than A. obscurum sp. nov. (length 12–15 μm vs. 10.5–11.7 μm in A. obscurum sp. nov.) and weakly radiate to almost parallel striae on both valves while in A. obscurum sp. nov. the striae are distinctly radiate. The valves of A. indistinctum are narrower than in A. obscurum sp. nov. (width 1.8–2.2 μm vs. 2.5–2.9 μm, respectively). Both of the Antarctic species also exhibit quite large areolae that are almost square or rectangular, while in A. obscurum sp. nov. the areolae are smaller, rounded to elongated.

Achnanthidium angustum sp. nov. is quite similar to the type of A. minutissimum [23] (

Table 4. Comparison of Achnanthidium angustum sp. nov. and similar species.

	Achnanthidium angustum sp. nov.	A. lineare	A. minutissimum	A. jackii	A. indistinctum
Valve shape	narrow linear-lanceolate	linear to narrow lanceolate	linear-lanceolate	linear-lanceolate to weakly lanceolate	narrow lanceolate
Valve apices	protracted and rounded	broadly rounded	subrostrate to subcapitate	rostrate	rostrate
Valve length, μm	12.3–14.8	9.0–13.5	8.9–18.5	8.0–17.0	8.5–13.0
Valve width, μm	2.4–3.1	2.2–2.8	2.5–3.0	3.0–3.9	1.8–2.2
Raphe valve
Distal raphe ends	straight, long	long, almost straight	straight or slightly bent	straight, short	straight, short
Axial area	narrow linear	narrow linear	narrow linear	narrow linear	narrow linear
Central area	indistinct	broad rectangular fascia	variable	asymmetrical fascia	small, indistinct
Striae pattern	radiate	radiate	radiate	weakly radiate to almost parallel	radiate
Striae density in 10 μm	31–37	28–32	28–32	25–30	36
Areolae	rounded to transapically elongated	rounded to slit-like	rounded to transapically elongated	rounded to slit-like	rounded to elongated to rectangular
Rapheless valve
Axial area	narrow lanceolate	narrow lanceolate	narrow linear	narrow linear-lanceolate	narrow linear
Central area	absent	almost absent to small elliptical	absent	absent to small fascia to asymmetrical	almost absent
Striae pattern	radiate	radiate	radiate	weakly radiate to almost parallel	radiate
Striae density in 10 μm	31–35	28–32	28–32	28–30	30–35
Areolae	rounded to transapically elongated	rounded to slit-like	rounded to transapically elongated	rounded to slit-like	rounded to elongated to rectangular
Source	this study	[25]	[23]	[36]	[20]

). These species can be differentiated by valve shape: A. angustum sp. nov. has slenderer valves with ends that are more tapered and narrower than in A. minutissimum; the ends in the new species are protracted and rounded and never rostrate. The axial area on the rapheless valve of A. angustum sp. nov. widens more towards the centre, having a narrowly lanceolate shape rather than narrow linear as in A. minutissimum. The striae in A. angustum sp. nov. appear to be evenly distributed along the whole valve and do not become denser at the ends.

Another similar species is A. jackii Rabenhorst [36]. This species also differs from the new species by the outline: the ends of the valves are less tapered, more bluntly rounded than in A. angustum sp. nov. The striae in A. jackii are almost parallel vs. the clearly radiate striae of A. angustum sp. nov. The central area on the rapheless valve of A. jackii is variable, from absent to a well-defined, small rectangular fascia; in A. angustum sp. nov. the central area seems to be always absent. The external distal raphe ends of A. jackii are fairly short, ending just beyond the last striae, while in A. angustum sp. nov. they extend almost to the edge of the valve face.

Achnanthidium angustum sp. nov. can also be compared to A. lineare [25]. The latter has more linear valves with broadly rounded ends and a broad rectangular fascia on the raphe valve that is absent in the new species. Achnanthidium lineare also tends to have only 1–3 areolae per stria, while A. angustum sp. nov. has 3–4. Finally, A. indistinctum [20] has a similar shape to the new species, however, its valves are generally narrower (width 1.8–2.2 μm vs. 2.4–3.1 μm in A. angustum sp. nov.) and it has rather large, square to rectangular areolae while A. angustum sp. nov. has small rounded or elongated ones. The axial area on the rapheless valve of A. angustum sp. nov. also has a wider, more lanceolate shape while in A. indistinctum it is narrow linear.

On the phylogenetic tree, all of our strains are distributed into separate branches or subclades within the A. minutissimum complex. This large clade also includes strains of A. digitatum Pinseel, Vanormelingen, Hamilton et Van de Vijver and A. saprophilum (Kobayashi et Mayama) Round et Bukhtiyarova and is sister to another clade that comprises strains of other monoraphid genera, such as Lemnicola Round et Basson, Planothidium Round et Bukhtiyarova, Pauliella Round et Basson and Psammothidium Bukhtiyarova et Round. Together with a thorough morphological analysis, these data confirm our conclusion that our taxa are indeed separate species previously unknown to science.

In our material we were able to study another strain, B443, in LM and SEM (Figure 5 and Figure 6). On the phylogenetic tree, this strain forms a separate branch that is sister to strain AT-196Gel02 identified as A. minutissimum. Morphologically, this strain is very similar to the type of A. minutissimum and cannot be confidently delineated from it. However, since there are no molecular data on the type of A. minutissimum, the true identity of this taxon remains a question; we suspect that it may be a cryptic species that is not A. minutissimum s. str.

Previous studies have already shown that according to molecular data, the A. minutissimum complex comprises many different lineages [13]. Our study confirms these findings. Most likely, these lineages represent separate species, as is the case with A. digitatum [13] and the new species described in the current study. Our study shows that often there are subtle morphological differences that can support molecular evidence, but that is not always the case; in these cases, more data are needed to draw a definitive conclusion regarding the identity of a concrete taxon. Morphological investigations alone are evidently not enough to resolve the taxonomy of A. minutissimum s.l. Quantitative features overlap very often between different species, and in those cases cannot be used for species delineation. The main morphological features that separate our new species from others are the valve shape, the striation patterns and areolae shape, and the shape of the axial and central areas, and these features are also used in other studies; however, the differences between species are often quite subtle and this may add a degree of subjectivity to identification of various natural populations. Moreover, little is known about the intraspecific variability in Achnanthidium species, which may also affect the process of taxonomic attribution. Combined application of morphological and molecular methods seems to be the best course of action in regards to this issue. However, the currently available molecular data on Achnanthidium taxa are quite scarce, which hinders molecular identification. This applies to both historically established and recently described species. A much larger reference database is needed; in particular, strains that are confidently assigned to A. minutissimum s. str. would be extremely useful in resolving the taxonomy of the A. minutissimum species complex.

Representatives of A. minutissimum s.l. have been reported from a wide variety of environmental conditions, from acidic to alkaline and from oligotrophic to eutrophic [8]. It has been shown that A. minutissimum s.str. demonstrates a wider tolerance range to various environmental factors than other species from this complex [37]; ecological preferences of other species have also been assessed [8,13,15,37]. Studies show that while ecological preferences of different species may be similar, the ranges of species distribution across nutrient and pH gradients may vary considerably [37]. Our new species have mostly been found in fairly cold water (9.5–13.9 °C), apart from strain B387 of A. baicalonanum sp. nov. that was found at 18.2 °C (Table 1). Achnanthidium baicalonanum sp. nov. and A. angustum sp. nov. were recorded from circumneutral to slightly alkaline waters, while strain B443 (A. cf. minutissimum) was found in a slightly acidic habitat. More records of the new species are needed to draw definitive conclusions about their ecological preferences and to compare them with other Achnanthidium species.

Currently, there are two main hypotheses regarding freshwater diatom distribution [38]. The first one suggests that most diatom species have cosmopolitan distributions, while the other argues that limited distribution and endemism are more characteristic. For a long time, A. minutissimum has been regarded as a widespread species with cosmopolitan distribution; however, recent studies show that some lineages of this complex represent separate species with limited distribution and specific ecological preferences [13], and our study confirms this as well, which supports the idea of endemism being more common for diatoms than previously thought. As indicated in Figure 1, the strains identified as A. minutissimum originate from various parts of the world, such as Europe, the USA, and sub-Antarctic islands. Strains from the same geographic locations tend to form different lineages that do not necessarily appear as closely related to each other on the phylogenetic tree. This further supports the idea that A. minutissimum is most likely a complex of cryptic or semi-cryptic species. A detailed investigation of worldwide diversity of A. minutissimum might reveal more species that are characteristic for only a specific region. Such an investigation would lend a greater understanding of the biogeography of this complex and would be an important tool for bioindication and biomonitoring; thus, it might be a direction worth exploring in further studies.

5. Conclusions

Our study contributes to the knowledge of the global diversity and biogeography of monoraphid diatoms, providing morphological and molecular data on several new species exclusive for the Transbaikal region. Our results support the idea of A. minutissimum being a complex of cryptic species and the hypothesis of endemism being common in diatoms. Research into different Achnanthidium populations, especially from unique and less-studied locations as Lake Baikal and its surrounding area, will not only shed light onto the taxonomy of the group, but also help our understanding of biogeographical patterns and increase the accuracy of biomonitoring methods; thus, it is likely an interesting direction for further studies.