Molecular phylogenetic data and seed coat anatomy resolve the generic position of some critical Chenopodioideae (Chenopodiaceae - Amaranthaceae) with reduced perianth segments.

The former Chenopodiumsubgen.Blitum and the genus Monolepis (Chenopodioideae) are characterised in part by a reduced (0-4) number of perianth segments. According to recent molecular phylogenetic studies, these groups belong to the reinstated genera Blitum incl. Monolepis (tribe Anserineae) and Oxybasis (tribe Chenopodieae). However, key taxa such as C.antarcticum, C.exsuccum, C.litwinowii, C.foliosumsubsp.montanum and Monolepisspathulata were not included and so their phylogenetic position within the Chenopodioideae remained equivocal. These species and additional samples of Blitumasiaticum and B.nuttallianum were incorporated into an expanded phylogenetic study based on nrDNA (ITS region) and cpDNA (trnL-trnF and atpB-rbcL intergenic spacers and rbcL gene). Our analyses confirm the placement of C.exsuccum, C.litwinowii and C.foliosumsubsp.montanum within Blitum (currently recognised as Blitumpetiolare, B.litwinowii and B.virgatumsubsp.montanum, respectively); additionally, C.antarcticum, currently known as Oxybasisantarctica, is also placed within Blitum (reinstated here as B.antarcticum). Congruent with previous studies, two of the three accepted species of Monolepis - the type species M.trifida (= M.nuttalliana) as well as M.asiatica - are included in Blitum. The monotypic genus Carocarpidium described recently with the type C.californicum is not accepted as it is placed within Blitum (reinstated here as B.californicum). To date, few reliable morphological characters have been proposed that consistently distinguish Blitum (incl. two Monolepis species) from morphologically similar Oxybasis; however, two key differences are evident: (1) the presence of long-petiolate rosulate leaves in Blitum vs. their absence in Oxybasis and (2) a seed coat structure with the outer wall of the testa cells lacking stalactites ('non-stalactite seed coat') but with an obvious protoplast in Blitum vs. seed coat with the outer walls of the testa cells having stalactites ('stalactite seed coat') and a reduced protoplast in Oxybasis. Surprisingly, the newly sequenced North American Monolepisspathulata nested within the tribe Dysphanieae (based on ITS and trnL-trnF + rbcL + atpB-rbcL analyses).The phylogenetic results, as well as presence of the stalactites in the outer cell walls of the testa and lack of the rosulate leaves, confirm the distinctive nature of Monolepisspathulata from all Blitum and, therefore, the recent combination Blitumspathulatum cannot be accepted. Indeed, the morphological and molecular distinctive nature of this species from all Dysphanieae supports its recognition as a new monotypic genus, named herein as Neomonolepis (type species: N.spathulata). The basionym name Monolepisspathulata is also lectotypified on a specimen currently lodged at GH. Finally, while Micromonolepispusilla is confirmed as belonging to the tribe Chenopodieae, its position is not fully resolved. As this monotypic genus is morphologically divergent from Chenopodium, it is retained as distinct but it is acknowledged that further work is required to confirm its status.

Introduction

The family Vent. comprises ~1500 species distributed worldwide (Sukhorukov 2014). It is divided into several subfamilies and at least one third of them belong to the core subfamily in the tribes G.Kadereit & Sukhor. ( L., L., Gueldenst.), incl. Atripliceae Duby ( G.L.Chu, L., S.Fuentes, Uotila & Borsch, L. s.str., Fenzl ex Moq., E.H.Zacharias, Hook. & Arn., Aellen, Hicken, Moq., Aellen, Hook.f., Ulbrich, Kar. & Kir., Stutz & G.L.Chu and E.H.Zacharias), ( L. incl. Ulbr., L.) and ( Moq., R.Br., A.Gray and Moq.) (Kadereit et al. 2003, 2010; Zacharias and Baldwin 2010; Fuentes-Bazan et al. 2012a, 2012b). While tribal boundaries are becoming well established, the status of a number of genera is far from stabilised, as ongoing molecular phylogenetic analyses continue to highlight new and sometimes unexpected relationships.

Some of the most recent and drastic taxonomic changes have been proposed by Fuentes-Bazan et al. (2012b) following their phylogenetic study of the large genus (~200–250 species) (Fuentes-Bazan et al. 2012a) and this classification is currently accepted by many authors (e.g. Iamonico 2011, 2014; Mosyakin 2013; Uotila 2017; Sukhorukov et al. 2013; Sukhorukov and Kushunina 2014, Hernández-Ledesma et al. 2015; Mosyakin and Iamonico 2017). According to the findings by Fuentes-Bazan et al. (2012b), sensu lato was shown to be polyphyletic and members previously included in the genus are now placed in tribes incl. Atripliceae ( s.str. 100–150 spp., ~12 spp., 8–9 spp., Moq., 1 sp.), ( >50 spp., 1 sp.) and ( ~12 spp.). To accomplish this, they reinstated the genera (type species Kar. & Kir. = (L.) S.Fuentes, Uotila & Borsch) and (type species (L.) Moq. ≡ L.) and recognised the new genus S.Fuentes, Uotila & Borsch. Finally, two of three known species of the genus Schrad. included in the study (the type species (Trev.) Schrad. = (Schult.) Greene) as well as Fisch. & C.A.Mey.) were shown to be nested within based on ITS (nrDNA) and trnF intergenic spacer with moderate statistical support (Fuentes-Bazan et al. (2012a). As is the oldest available name (Linnaeus 1753), was transferred and was re-instated as (Fisch. & C.A.Mey.) Fuentes et al. and Schult., respectively (Fuentes-Bazan et al. 2012b). The third species, A.Gray, was not sequenced, but also transferred into [as (A.Gray) Fuentes et al.] due to its morphological similarity to both and .

Further changes were subsequently proposed by Theodorova (2014), provided without a detailed explanation, suggesting that should be expanded to include , and , resulting in the proposed new combinations (L.) T.A.Theodorova (≡ (L.) S.Fuentes, Uotila & Borsch), (L.) T.A.Theodorova (≡ (L.) S.Fuentes, Uotila & Borsch) and (L.) T.A.Theodorova (≡ (L.) S.Fuentes, Uotila & Borsch). Recently, Zhu and Sanderson (2017) described a new monotypic genus S.C.Sanderson et C.L.Chu with the type species (S.Wats.) S.C.Sanderson & G.L.Chu (≡ S.Wats. ≡ (S.Wats.) S.Wats.), based on the fruits having a fleshy pericarp.

The recent split of sensu lato into genera belonging to different tribes as suggested by Fuentes-Bazan et al. (2012b) is supported in part by morphological characters. First, all species of with obvious glandular hairs, ovoid or roundish, yellow or orange subsessile glands and simple hairs now belong to the tribe (placed in either R.Br. or Moq.), while the remaining former (now included in and ) have an indumentum of white bladder (“mealy”) hairs, sometimes with scattered simple hairs (Reimann and Breckle 1988; Simón 1997; Sukhorukov et al. 2015b). The number of perianth segments was also traditionally thought to be a good diagnostic character, which usually corresponds to the number of stamens. s.str., and are characterised by the presence of five perianth segments and five stamens, while various genera across the subfamily are characterised by a lower number (1–4) of perianth segments and stamens, as observed in some and (), incl. () and many (), especially amongst Australian species (e.g. Ulbrich 1934; Wilson 1984; Judd and Ferguson 1999; Holmgren 2003). However, this character may not be consistently informative as species such as usually has 5 perianth segments and 5 stamens.

It has become apparent in recent years that fruit and seed characters are also useful in distinguishing members of the former , particularly amongst groups that are quite morphologically similar (Sukhorukov 2006, 2014; Sukhorukov and Zhang 2013; Sukhorukov et al. 2015a). A good example is Sukhor. Originally this species was described as a member of the former (Sukhorukov 1999). Its generic status was discussed by Fuentes-Bazan et al. (2012b) and finally resolved by Sukhorukov et al. (2013) as being a part of [ (Sukhor.) Sukhor. et Uotila] based on molecular phylogenetic data supported by morphological and seed characters. Almost all (, , , , , , , , and all with red or black seeds) possess a seed-coat testa with thickened outer cell walls impregnated with vertical or oblique stalactites and a reduced protoplast (hereafter ‘stalactite seed coat’) (Sukhorukov 2006; Kadereit et al. 2010; Sukhorukov and Zhang 2013; Sukhorukov 2014). There are a few exceptions, however, for example the seed coat in and three species endemic to Juan Fernández Archipelago (Chile) ( Skottsb., Johow, Skottsb.), does not contain the stalactites in the outer cell walls and possesses a visible protoplast (hereafter ‘non-stalactite seed coat’) (Sukhorukov 2014). These three geographically isolated Chilean species are closely allied and highly unusual, as they not only possess a non-stalactite seed coat but have a tree-like habit and fruits with an apically swollen pericarp. Of these, only has been included in molecular analyses (Kadereit et al. 2010), which confirmed its phylogenetic position within this genus. The non-stalactite seed coat morphology is also evident in the , Hook.f [≡ (Hook.f.) Mosyakin], almost all sensu Fuentes-Bazan et al. (2012b) with the exception of (A.Gray) S.Fuentes, Uotila & Borsch, or (Sukhorukov 2014).

Amongst the species of the former or investigated carpologically but not included in recent molecular phylogenetic studies, two taxa are of special interest. The first, , is endemic to western states of USA and North Mexico and was transferred to (as ) due to morphological affinities with other species of the genus. The second taxon, , is another poorly known taxon endemic to Tierra del Fuego (southernmost parts of Argentina and Chile) that still occupies a pending position within . Previously, it was described as Hook.f. (Hooker 1847) and later transferred by the same author to as (Hook.f.) Hook.f. (Bentham and Hooker 1880). The latter name was widely accepted in subsequent taxonomic treatments (Reiche 1911; Aellen 1929, 1931; Aellen and Just 1943; Moore 1983; Giusti 1984; Zuloaga and Morrone 1999). Recently, was transferred into by Mosyakin [2013, as (Hook.f.) Mosyakin] based on its morphological similarity to other . However, the stalactite seed coat morphology of and non-stalactite seed coat of contrast with those of other members of and , respectively (Sukhorukov 2014), which raises the question of their true phylogenetic position.

To resolve this issue, we have included these two species, in addition to several accessions of taxa sampled for the first time [, (C.Loscos) Uotila, (Paulsen) Uotila, Uotila and ], as well as an additional sample of (Fisch. & C.A.Mey.) S.Fuentes, Uotila & Borsch. in expanded molecular analyses based on nrDNA (ITS region) and cpDNA (atpB-rbcL intergenic spacers + rbcL and trnL-trnF intergenic spacer + rbcL, hereafter as atpB-rbcL and trnL-trnF, respectively) to determine their phylogenetic position within the . Furthermore, we discuss the role of fruit and seed characters for delimitating morphologically similar but phylogenetically distant taxa and conclude with proposed taxonomic changes that reflect our findings.

Methods

Taxon sampling

Several new taxa were included in the phylogenetic analysis for the first time: (Hook.f.) Hook.f. [≡ (Hook.f.) Mosyakin: Chile, Tierra del Fuego, December 1971, Moore & Goodall s.n. (LE)]; (C.Loscos) Uotila: Algeria, Zenina, July 1968, V.P. Boczantsev 681 (LE); Uotila: Iran, prov. Tehran, Elburz, June 1977, K.-H. Rechinger 57243 (B); (Paulsen) Uotila: Afghanistan, Parwan prov., Salang, 8 August 1969, J.E. Carter 602 (LE); A.Gray: USA, California, Susanville, August 1983, I.Yu. Koropachinsky & al. 404 as (MHA). Additionally, we have included a new accession of (Fisch. et C.A.Mey.) S.Fuentes, Uotila et Borsch (Russia, Yakutiya, Ust-Yansky distr., August 1976, E.V. Ter-Grigoryan 1009, MHA). The taxa included in the molecular analyses and their GenBank accession numbers are given in the Table 1.

Table 1.

Voucher information and GenBank accession numbers for the species of and outgroups included in the phylogenetic analysis (arranged in alphabetical order). The newly sequenced samples are highlighted in bold. Some vouchers in GenBank may be stored under old names.

Species	Old names (if applicable)	GenBank accession number
		ITS	rbcL	trnL-trnF	atpB-rbcL
Atriplex hortensis	–	HM005854	KX678160	HE577500	–
Atriplex patula	–	HE577358	MG249776	HE577498	HM587650
Atriplex spongiosa	–	–	AY270060	–	HM587661
Atriplex undulata	–	–	AY270061	–	HM587665
Atriplex phyllostegia	–	HM005870	HM587590	–	HM587651
Atriplex peruviana	–	HM005867	–	–	–
Atriplex watsonii	–	HM005871	–	–	–
Atriplex rusbyi	–	HM005865	–	–	–
Atriplex patagonica	–	HM587541	–	–	–
Atriplex lentiformis	–	HM005872	–	–	HM587637
Atriplex cinerea	–	HM587491	–	–	–
Atriplex centralasiatica	–	DQ086481	HM587583	–	HM587621
Atriplex suberecta	–	HM005863	–	–	–
Axyris amaranthoides	–	AM849227	KX678411	HE577510	–
Axyris hybrida	–	HE577371	–	HE577511	–
Blitum antarcticum	Chenopodiumantarcticum (Oxybasisantarctica)	MH155315	MH632743	MH632745	MH152573
Blitum asiaticum	Monolepis asiatica	MH150882	MH731231	MH731229	–
Blitum bonus-henricus	Chenopodium bonus-henricus	HE577372	KF613023	HE577512	HM587670
Blitum californicum	Chenopodium californicum	HE577376	MF963177	HE577516	–
Blitum capitatum	Chenopodium capitatum	KJ629064	MG249277	HE577513	–
Blitum litwinowii	Chenopodium litwinowii	MH153781	MH632744	MH632746	MH632749
Blitum nuttallianum	Monolepis nuttalliana	HE577375	JX848452	HE577515	HM587702
Blitum petiolare	Chenopodium exsuccum	MH150883	–	MH632747	MH152574
Blitumvirgatum L.	Chenopodium foliosum	JF976147	AY270081	HE577518	HM587673
Blitum virgatum subsp. montanum	Chenopodium foliosum subsp. montanum	MH155242	–	–	–
Ceratocarpus arenarius	–	AY556430	HM587594	HE577505	–
Chenopodiastrum coronopus	Chenopodium coronopus	HE577403	HM587595	HE577543	HM587671
Chenopodiastrum hybridum	Chenopodium hybridum	HE577530	–	HE577530	–
Chenopodiastrum murale	Chenopodium murale	HE577392	HM849890	HE577531	HM587675
Chenopodium album	–	JF976146	JF941270	HE577609	MF073794
Chenopodium atrovirens	–	KP226648 /	KX679232	HE577587	–
Chenopodium auricomum	–	KP226671	–	–	–
Chenopodium bengalense	Chenopodium giganteum	HE577458	–	–	–
Chenopodium berlandieri var. boscianum	–	HE577426	MG249740	HE577564	–
Chenopodium berlandieri var. zschackei	–	HE577425	–	–	–
Chenopodium desertorum	–	HE577417	AY270042	HE577555	HM587672
Chenopodium desiccatum	–	HE577412	KX678128	HE577550	–
Chenopodium ficifolium	–	HE577466	KM360714	HE577606	–
Chenopodium fremontii	–	HE577408	KX679065	HE577572	–
Chenopodium hians	–	HE577470	MG248000	HE577610	–
Chenopodium iljinii	–	HE577468	–	–	–
Chenopodium incanum	–	HE577410	MG246401	HE577548	–
Chenopodium leptophyllum	–	HE577428	MG248863	HE577566	–
Chenopodium neomexicanum	–	KJ629054	–	–	–
Chenopodium nevadense	–	HE577411	–	–	–
Chenopodium opulifolium	–	HE577454	MG248036	HE577594	–
Chenopodium pallescens	–	HE577409	–	–	–
Chenopodium pallidicaule	–	KJ629055	–	–	–
Chenopodium nutans	Einadia nutans	–	KM896090	–	HM587686
Chenopodium parabolicum	Rhagodia parabolica	–	KU564859	–	HM587704
Chenopodium quinoa	–	HE577443	KY419706	–	KY419706
Chenopodium standleyanum	–	KJ629051	MG249838	HE577560	–
Chenopodium subglabrum	–	HE577465	MG249459	HE577605	–
Chenopodium vulvaria	–	HE577407	JN892907	HE577591	–
Chenopodium watsonii	–	HE577462	MG246238	HE577602	–
Cycloloma atriplicifolium	–	HQ218998	HM587598	–	HM587681
Dysphania ambrosioides	Chenopodium ambrosioides	DQ005963	MG249540	HE577493	HM587682
Dysphania botrys	Chenopodium botrys	KJ629068	MG247946	DQ499383	HM587683
Dysphania cristata	Chenopodium cristatum	KJ629066	AY270046	–	HM587684
Dysphania glomulifera	Chenopodium glomuliferum	–	AY270086		HM587685
Dysphania pumilio	Chenopodium pumilio	HE577343	MG248652	HE577485	–
Dysphania schraderiana	Chenopodium schraderianum	HE577349	–	–	–
Exomis microphylla	–	–	HM587601	–	HM587687
Grayia brandegeei	–	HM005845	HM587604	HE577497	HM587690
Grayia spinosa	–	HM005844	HM587605	HE577496	HM587691
Halimione verrucifera	Atriplex verrucifera	HM587575	HM587606	–	HM587695
Halimione pedunculata	Atriplex pedunculata	HM587573	AY270093	–	HM587694
Holmbergia tweedii	–	HM005842	AY270100	–	HM587696
Krascheninnikovia ceratoides	–	HE577367	AY270105	HE577507	HM587697
Krascheninnikovia ceratoides subsp. lanata	Krascheninnikovia lanata	HE577368	MG248963	HE577508	HM587698
Lipandra polysperma	Chenopodium polyspermum	KJ629061	KX677934	HE855686	–
Micromonolepis pusilla	–	–	HM587608	–	HM587701
Neomonolepis spathulata	Monolepisspathulata (Blitumspathulatum)	MH675518	MH731232	MH731230	MH152575
Oxybasis glauca	Chenopodium glaucum	KJ629060	MG249300	HE577527	MF073807
Oxybasis rubra	Chenopodium rubrum	HE577381	MG249329	HE577525	–
Oxybasis urbica	Chenopodium urbicum	KJ629057	MG246691	HE577524	HM587678
Oxybasis micrantha	–	KU359325	–	–	–
Spinacia oleracea	–	EU606218	–	AJ400848	–
Suckleya suckleyana	–	HE577347	–	–	–
Teloxys aristata	Chenopodiumaristatum; Dysphaniaaristata	KJ629070	AY270140	–	HM587708
Outgroups	–
Bassia laniflora	Kochia laniflora	KF785942	–	–	–
Bassia prostrata	Kochia prostrata	KF785963	AY270104	HE577478	KF785926
Beta vulgaris	–	AY858597	–	–	DQ074969
Hablitzia tamnoides	–	AY858590	AY270092	HE577475	JQ407841
Polygonum aviculare	–	–	MF158792	HQ843161	JN234937
Polygonum aviculare subsp. buxiforme	–	GQ339988	–	–	–

DNA extraction

Total genomic DNA was extracted from herbarium samples according to Krinitsina et al. (2015). Following the homogenisation of plant fragments (MiniLys, Bertin Technologies, France), total DNA was extracted using the CTAB-method (Doyle and Doyle 1987) and further purified using AMPure Beads (Beckman Coulter, USA).

PCRs for two chloroplast markers (atpB-rbcL and trnL-trnF) and nrDNA (ITS region) were carried out in a Thermal Cycler T100 (Bio-Rad, USA) using primers and cycler programmes listed in Table 2. A 10 ng aliquot of DNA was used to make a 25 μl total volume reaction, containing 1 μM of each primer, 200 μM of each dNTP and 0.5 U Encyclo polymerases (Evrogen, Russia). PCR products were checked on 1.2% agarose gels and purified using AMPure Beads (Beckman Coulter, USA) according to the owner’s manual. AMPure Beads suspension was mixed with a solution containing PCR-product ratio 1 vol. PCR-mix: 1.2 vol. AMPure Beads for atpB-rbcL and ITS primer pairs and 1 vol. PCR-mix: 1.4 vol. AMPure Beads for rbcL, Tab C/Tab D and Tab E/Tab F primer pairs.

Table 2.

Primers and cycler programmes used for the molecular analysis.

Marker	Primer sequences and combination	Reference	Cycler programmer
ITS	ITS5 5'-GGA AGT AAA AGT CGT AAC AAG G-3'	White et al. (1990)	95 °C for 5 min, 33 cycles of amplification (95 °C for 15 s, 55 °C for 30 s, 72 °C for 40 s), 72 °C for 5 min
	ITS4 5'-TCC TCC GCT TAT TGA TAT GC-3'
rbcL (partial)	rbcLaF 5'- ATG TCA CCA CAA ACA GAG ACT AAA GC-3'	Levin et al. (2003)	95 °C for 5 min, 35 cycles of amplification (95 °C for 10 s, 55 °C for 30 s, 72 °C for 40 s), 72 °C for 5 min
	rbcLaR 5'-GTA AAA TCA AGT CCA CCR CG-3'	Kress et al. (2009)
atpB-rbcL spacer	atpB-rbcL F 5'-GAA GTA GTA GGA TTG ATT CTC-3'	Golenberg et al. (1993)	95 °C for 5 min, 35 cycles of amplification (95 °C for 20 s, 56 °C for 30 s, 72 °C for 60 s), 95 °C for 20 s, 56 °C for 80 s, 72 °C for 8 min
	atpB-rbcL R 5'-CAA CAC TTG CTT TAG TCT CTG-3'
trnL-F	Tab C 5'-CGA AAT CGG TAG ACG CTA CG-3'	Taberlet et al. (1991)	95 °C for 5 min, 35 cycles of amplification (95 C for 1 min, 50 °C – 65 °C (increasing in 0.3 C per cycle) for 1 min, 72 °C for 4 min), 72 °C for 5 min
	Tab D 5'-GGG GAT AGA GGG ACT TGA AC-3'
	Tab E 5'- GGT TCA AGT CCC TCT ATC CCC-3'
	Tab F 5'ATI' TGA ACT GGT GAC ACG AG 3'

Sequencing and alignment

Sequencing was performed following Sanger methods on an Applied Biosystems 3730 DNA Analyser using ABI PRISM BigDye Terminator v. 3.1 (Center of Collective Use “Genome”, Institute of Molecular Biology, Moscow, Russia). The sequencing primers were the same as the amplification primers.

The raw forward and reverse sequences were checked and combined in BioEdit sequence alignment editor v. 7.0.5.3 (Hall 1999). Sequences were edited and aligned using Muscle 3.6 (Edgar 2004). The obtained alignments were manually edited using PhyDe (version 0.9971: Müller et al. 2010) following the rules outlined in Löhne and Borsch (2005). Mutational hotspots (regions of uncertain homology) were excluded from the analysis (Borsch et al. 2003). Gaps were treated as missing data during the phylogenetic inference.

Phylogenetic inference

To show the relationships between taxa, we reconstructed various phylogenies using Bayesian analysis, maximum likelihood (ML) and maximum parsimony (MP) methods for the ITS and combined trnL-trnF + rbcL + atpB-rbcL datasets. Models of nucleotide substitution were selected using the MrModeltest 2.1.7 (Nylander 2004) via the Akaike information criterion (AIC: Akaike 1974). The substitution model was set to GTR + G + I. For the ML analyses, we employed RAxML Version 8 (Stamatakis 2014). Bootstrap analyses were conducted with 2500 replicates for ML. Parsimony analyses were conducted in PAUP* 4.0a162 (Swofford 2002) with the following settings: all characters have equal weight, MaxTrees set to 1000 (auto increased by 1000), TBR branch swapping and with 20000 jackknife (JK) replicates to calculate node support. Bayesian analyses were conducted in BEAST 2.5.0 (Bouckaert et al. 2014). Four Markov Chain Monte Carlo analyses with four chains were run for 20 million generations for every dataset, sampling every 1000 generations. Burn-in was set to remove 5% of the total trees sampled after assessing likelihood convergence by inspection of the trace plots in the programme Tracer v.1.6 (Rambaut et al. 2014). A birth and death prior was chosen for branch lengths (Gernhard 2008). The maximum clade credibility tree was calculated in the programme TreeAnnotator v1.4.8 (Drummond and Rambaut 2007) with a posterior probability limit of 0.7. Final trees were edited in the programme TreeGraph ver. 2.14.0 (Stöver and Müller 2010).

Morphology and anatomy

The carpology of the tribe was described in detail in a previous study by Sukhorukov (2014). In this study, we pay particular attention to the fruit and seed of and to the general structure of the reproductive shoot of that were not illustrated in Sukhorukov (2014). The samples were observed using a scanning electron microscope (SEM) JSM–6380 (JEOL Ltd., Japan) at 15 kV after sputter coating with gold-palladium in the laboratory of Electron Microscopy at Lomonosov Moscow State University. Prior to SEM, the fruits were dehydrated in aqueous ethyl alcohol solutions of increasing concentration, followed by alcohol-acetone solutions and pure acetone. No dehydration of the seeds is required prior to SEM observation due to the absence of soft tissues (e.g. papillae or trichomes) on their surface.

The cross-sections of the seeds were prepared using a rotary microtome Microm HM 355S (Thermo Fisher Scientific, USA) and then examined using a Nikon Eclipse Ci (Nikon Corporation, Japan) light microscope and photographed using a Nikon DS-Vi1 camera (Nikon Corporation, Japan) at the Department of Higher Plants, Lomonosov Moscow State University. Before sectioning, the seeds were soaked in water:alcohol:glycerine (1:1:1) solution, dehydrated in ethanol dilution series and embedded in the Technovit 7100 resin (Heraeus Kulzer, Germany).

Results

Phylogenetic analysis

The phylogenetic analysis based on nrDNA (ITS) and combined cpDNA analyses (trnL-trnF + rbcL + atpB-rbcL) revealed that the tribes , s.str., and are well-supported within and congruent with previous molecular analyses by Fuentes-Bazan et al. (2012b) (Figures 1–2). The results outlined below focus on the phylogenetic position of the newly included taxa [≡ ], , , and .

Figure 1.

Best tree from the BEAST analysis of the ITS dataset. Bayesian posterior probabilities are given above the branches, jackknife values (left) and bootstrap percentages of the maximum likelihood analyses (right) are given below branches.

Figure 2.

Best tree from the BEAST analysis of the combined trnL-trnF + rbcL + atpB-rbcL dataset. Bayesian posterior probabilities are given above the branches, jackknife values (left) and bootstrap percentages of the maximum likelihood analyses (right) are given below branches.

In the ITS analysis (Figure 1), the tribe is placed sister to the remaining . The next diverging lineage is a well-supported , with + forming a sister lineage to the remaining representatives of the tribe. , , and fall well within , which is sister to a well-supported . and (L.) C.A.Mey. form part of the polytomy with the rest of the genus.

Like the ITS phylogenetic analysis, the combined trnL-trnF + rbcL + atpB-rbcL tree (Figure 2) shows the as an early branching lineage in , sister to a polytomy of , and . Within the , and form a polytomy with the remaining representatives of the tribe, which includes nested within . , and are nested within ( is not included in the combined tree). is sister to + – .

Carpological studies

This study highlighted the fact that these species, with the exception of , possess the same fruit and seed anatomy as other species such as a mamillate pericarp (Figure 3) and non-stalactite seed-coat with obvious (visible) protoplast (Table 3; Figure 4). In contrast, the carpology of somewhat resembles the morphology observed in species of and many other in having a papillate pericarp and a stalactite seed coat with a highly reduced protoplast (Figure 5). Other important characters such as life history, the degree of fusion of reduced perianth segments, pericarp structure and adherence, the colour, shape and morphology of seeds and an embryo position, are recorded for representative species of each genus, as summarised in Table 3.

Figure 3.

Pericarp of . Scale bar: 200 μm.

Table 3.

Additional noteworthy characters evolved in and . This table summarises life history and carpological data from Sukhorukov and Zhang (2013), Sukhorukov et al. (2013), Sukhorukov (2014), with additional information included for and .

Taxon/Character	Life history	Perianth segments	Cells of the outer pericarp layer	Pericarp adherence to the seed coat	Seed shape and colour	Seed surface	Seed keel	Thickness of seed-coat testa (µm)	Acicular outgrowths of the testa cells	Presence of spatial heterospermy	Seed embryo position
Blitum antarcticum	short-lived perennial herb	basally connate	spongy	scraped off the seed	roundish, red	alveolate	–	12–20	–	–	vertical
B. asiaticum	annual	free	not spongy	easily ruptured	roundish, red	undulate	+	7–10	–	–	vertical
B. atriplicinum	annual or short-lived perennial herb	basally connate	not spongy	hardly removed	roundish, red	alveolate, with hairy-like outgrowths	–	17–25	+	–	vertical
B. bonus-henricus	perennial herb	basally connate	spongy	scraped off the seed	roundish, red	smooth	–	37–45	–	+	vertical, rarely horizontal
B. californicum	perennial herb	basally connate	spongy	scraped off the seed	roundish, red	alveolate	–	25–30 and 37–45 (heterospermous)	–	+	vertical
B. capitatum	annual or short-lived perennial herb	basally connate	not spongy	hardly removed	ovate, red	undulate	+ (two keels and a groove between them)	12–15	–	+	vertical
B. hastatum	annual or short-lived perennial herb	connate to 1/3	not spongy	hardly removed	ovate, red	undulate	+ (two keels and a groove between them	15–18	–	+	vertical
B. korshinskyi	annual or short-lived perennial herb	almost free	not spongy	hardly removed	ovate, red	undulate	+ (two keels and a groove between them)	10–12	–	–	vertical
B. litwinowii	annual or short-lived perennial herb	basally connate	not spongy	hardly removed	ovate, red	alveolate	+ (two keels and a groove between them)	10–12	–	–	vertical
B. nuttallianum	annual	free, or perianth absent	not spongy	hardly removed	roundish, red	alveolate, with hairy-like outgrowths	–	8–10	+	–	vertical
B. petiolare	annual or short-lived perennial herb	basally connate	not spongy	hardly removed	ovate, red	alveolate	+ (two keels and a groove between them)	15–17	–	–	vertical
B. virgatum	annual or short-lived perennial herb	basally connate	not spongy	hardly removed	ovate, red	undulate	+ (two keels and a groove between them)	10–12	–	+	vertical
Oxybasis chenopodioides	annual	fused in almost all flowers, free only in some flowers	not spongy	easily ruptured	roundish, red	minutely pitted	–	10–15	+		vertical and horizontal
O. glauca	annual	basally connate	not spongy	easily ruptured	roundish, red	minutely pitted	–	10–15 and 17–25 (heterospermous)	+	+	vertical and horizontal
O. gubanovii	annual	basally connate	not spongy	hardly removed	roundish, red	smooth (minutely pitted)	+ (one keel)	12–15	+	–	vertical
O. macrosperma	annual	connate to the middle or almost to the top	spongy	scraped off the seed	roundish, red	reticulate with minutely pitted dots	–	12–20	+	–	vertical and horizontal
O. mexicana	annual	basally connate	not spongy	easily ruptured	roundish, red	reticulate with minutely pitted dots	–	20–25	+	+	vertical and horizontal
O. micrantha	annual	basally connate	not spongy	scraped off the seed	roundish, red	minutely pitted	+ (one keel)	12–15	+	–	horizontal, rarely vertical
O. rubra	annual	basally connate	not spongy	easily ruptured	roundish, red	reticulate with minutely pitted dots	–	10–15	+		vertical and horizontal
O. urbica	annual	basally connate	papillate	scraped off the seed	roundish, black	minutely pitted	–	42–50	+	–	horizontal

Figure 4.

Cross-section of the seed of . Abbreviations: T – testa, TE- tegmen, PE – perisperm.

Figure 5.

Cross-section of the seed of . Abbreviations: T – testa, TE – tegmen, PE – perisperm, ST – stalactites in the outer walls of the testa cells.

Discussion

The phylogenetic position of [≡ (Uotila) S.Fuentes, Uotila et Borsch], [= Link] and [≡ S.Fuentes, Uotila et Borsch] within as proposed by Fuentes-Bazan et al. (2012b) was supported by the findings of this study. Indeed, the results were predictable due to the shared morphological and carpological affinities of these species to , such as the presence of a leaf rosette, tight adherence of the pericarp to the seed coat and the ovoid and keeled seeds having the same anatomical structure (e.g. Uotila 1993, 1997; Sukhorukov 2014). For this reason, while (Litv.) Minkw. has not been included in any molecular phylogenies to date, it should be treated as Litv. (Fuentes-Bazan et al. 2012b) due to the shared presence of these diagnostic traits. It is also evident, based on phylogenetic and carpological data from this study, that (formerly ) must be treated as as proposed by Hooker (1847). Moreover, as is the type of (Aellen) Mosyakin [including and (R.Br.) Mosyakin: Mosyakin 2013], this section may be recognised within but this requires further exploration as the phylogenetic position of remains equivocal.

Diagnostic characters for and

The importance of morphological characters used to delineate species within the genus that are now considered to belong to either or have been discussed by various authors (e.g. Moquin-Tandon 1840, 1849; Aellen and Just 1943; Scott 1978; Fuentes-Bazan et al. 2012b). However, the morphological similarity of some species has led to taxonomic confusion. For example, many macromorphological characters overlap in and , including previous diagnostic traits such as: reduced (1–4) number of perianth segments, presence of the vertical seed embryo position and emergence of spatial heterospermy. Such characters are clearly homoplastic in , and some other groups of the (Sukhorukov and Zhang 2013). Only one trait visible to the naked eye, the presence of leaf rosette in (Figure 6) and its absence in , can be used for the delimitation of both genera (see diagnostic key and generic descriptions in Fuentes-Bazan et al. 2012b). However, it should be noted that the leaf rosette in some , especially in species previously included in (, ), is reduced to 1–2 leaves that may wither away completely by anthesis. From this study and from previous work (Sukhorukov and Zhang 2013; Sukhorukov et al. 2013; Sukhorukov 2014), it is evident that another character, the structure of the testa cells of the seed coat, is also diagnostic. In , as well as almost all other , the seed testa cells have a reduced protoplast and “stalactites” hanging vertically in the outer wall (stalactite seed coat). In contrast, the presence of non-stalactite seed coat with a highly visible protoplast, unambiguously distinguishes . Other characters, such as reduced perianth segments, mamillate pericarp, red seeds, seed keel, vertical embryo position of note for representative species of each genus, are summarised in Table 3 and they play a role for the diagnostics at the species level or species group (see Sukhorukov 2014 for further detail).

Figure 6.

Habit of showing the senescing leaf rosette. Photographer: Igor Pospelov (Russia, Krasnoyarsk prov., Taymyr, Khatanga, August 2014).

In the absence of molecular phylogenetic data, it is clear that carpological characters must be taken into consideration when determining the generic placement of taxa in either or . Molecular data from this study and previous investigations (Kadereit et al. 2010; Fuentes-Bazan et al. 2012a, 2012b), when examined in conjunction with carpological evidence (Sukhorukov 2014), show that two taxonomic changes recently proposed: (1) the merger of , and () into an extended () as suggested by Theodorova (2014) and (2) the description of a new monotypic genus S.C.Sanderson et G.L.Chu with the type (≡) by Zhu and Sanderson (2017), cannot be accepted. Additionally, it should be noted that the pericarp of is not fleshy as previously described (Zhu and Sanderson 2017), but its outer layer consists of spongy (mamillate) cells that imitate a “fleshy” pericarp. This type of mamillate pericarp is present in some and (Figure 3, see also Table 3) and so this character is clearly not unique to .

This species was initially described as Torr. ex Watson (Watson 1871) and it is noteworthy to consider its morphology and phylogenetic position in context with other species previously known as . It is a small annual herb covered with bladder hairs that has fleshy leaves (Figure 7), unisexual flowers with reduced (1–3) perianth segments and tiny papillate fruits. Due to its unusual habit, was transferred into a new monotypic genus (Ulbrich 1934). The species was included in a atpB-rbcL molecular analysis, where it was unexpectedly placed within the “ I” clade comprising , and a part of s.l. (Kadereit et al. 2010). The papillate pericarp and the stalactite seed coat provide a good support for its placement into , based on cpDNA being a part of s.str. (Kadereit et al. 2010, as I; Figure 2). However, the limited number of taxa used in the atpB-rbcL analysis, the lack of additional molecular data and the significant morphological differences evident between and the remaining species in this clade, such as the presence of fleshy leaves and reduced perianth segments, precludes the formal transfer of to . Further work is needed to evaluate the exact position of within .

Figure 7.

Shoot of showing the characteristic fleshy leaves. Photographer: Steve Matson (USA, California, Mono County, Long Valley, 2007).

is neither nor

Recently, was transferred to (as ) based on its resemblance to other species of the genus due to the presence of a reduced number of perianth segments (Fuentes-Bazan et al. 2012b). It is evident, however, that the reduced number of perianth segments independently evolved in (e.g. in and some ), and many (Sukhorukov and Zhang 2013). In light of carpological evidence (Sukhorukov 2014), it seemed doubtful that should be included in , as this species possesses a stalactite seed coat with a reduced protoplast. Our phylogenetic results show that is not closely related to the other species in (, and ) that are now included in () as and , respectively. This species falls within forming a polytomy with and + . is a glabrous annual and differs from all by the absence of simple hairs and subsessile glands that are diagnostic characters of this tribe. Additionally, is found to have the stalactite seed coat, a character missing in all (Sukhorukov 2014). The close relationship between and the , evidenced by molecular data, is unexpected given the obvious morphological and carpological differences. Indeed, is considered so distinct that it warrants recognition at the generic level. As the type for , (Trev.) Schrad. [= (Schult.) Greene], is synonymised within (as ), a new name is required for . As such, a new monotypic genus named Sukhor., gen. nov. is established here.

Taxonomy

Sukhor. gen. nov.

urn:lsid:ipni.org:names:77191294-1

Type species.

(A.Gray) Sukhor., comb. nov.

Description.

Annual, glabrous, branched or not; lateral branches if present ascending; leaves cauline (rosulate leaves absent), densely located, spatulate-oblong, with a short petiole up to 1 cm or sessile, entire; inflorescence leafy (bracts similar to stem leaves); flowers sessile or shortly pedicellate, unisexual intermixed in small glomerules (Figure 8); male flowers with 2-lobed hyaline perianth, stamens 1–2, anthers 0.10–0.15 mm long; female flowers without perianth, fruits 0.55–0.65 mm in diameter, almost round, with blackish papillate pericarp (when dry) that is easily raptured, styles 2(3); seeds 0.4 × 0.3 mm, reddish, with smooth surface, with small irregular pits (seen at a higher magnification), seed-coat testa with stalactites in the outer cell walls and reduced protoplast; embryo vertical.

Figure 8.

SEM detail of the inflorescence of . Abbreviations: B – bract (stained in green), FF – female flowers (orange), MF –male flower (perianth stained in blue, stamen in yellow), S – stem.

(A.Gray) Sukhor. comb. nov.

urn:lsid:ipni.org:names:77191295-1

≡

≡

Morphological notes.

As is a monotypic genus, the description of corresponds to the generic description above. is morphologically distant from all (, A.Gray, R.Br. and Moq.) in being glabrous in all parts (vs. glandular and/or simple hairs), having unisexual flowers (vs. bisexual or polygamous) and ‘stalactite’ seed-coat testa (vs. ‘non-stalactite’). For this reason, we prefer to refer to the clade with the above-mentioned genera as the ‘ + ’ clade.

Typification.

The type specimen lodged at GH contains several plants collected from different areas in California and almost all of them were collected after the description of (Gray 1868). The lectotype selected here (lower right-hand specimen on the GH00037208 sheet) is a part of original material cited in the protologue as “Sierra Nevada, at Mono Pass, in loose soil, Bolander” (Gray 1868) and it is chosen in accordance with Art. 9 of ICN (Turland et al. 2018). The description of the species is consistent with the image of the lectotype. Gray (1868) also noted that the seeds of are notably smaller than those of [= ]. The small seed dimensions of (0.4 × 0.3 mm) are similar to those observed in many Australian (Wilson 1984 sub ; Sukhorukov 2014).

Distribution.

South-western North America (USA, North Mexico).

Etymology.

The new generic name is composed by the prefix “neo” (new) and the core name .

Conclusion

In the , some phylogenetically distant taxa often look similar due to convergence of various morphological characters, some of which were previously thought to be diagnostic such as the number of perianth segments. A remarkable example is highlighted by the different phylogenetic positions occupied by members of the former genus , which are currently included in ( ≡ ; ≡ ), ( ≡ ) and ( ≡ ). This study shows that fruit and seed characters such as seed-coat structure are valuable traits for taxonomic study. These features are particularly useful in distinguishing the morphologically similar but phylogenetically distinct genera and .