Zahora ait-atta Lemmel & M.Koch, a new species from the Moroccan Sahara, is described and documented here and constitutes a monotypic new genus. The new taxon belongs to the tribe Brassiceae (Brassicaceae), and cytogenetic and phylogenetic analyses reveal that this diploid species has a remote status of Miocene origin in the northwestern Sahara Desert. We examined the morphological differences between morphologically related genera and provide photographs of the new species. The new genus may play a key role in future Brassica-Raphanus crop research since it is placed phylogenetically at the base of a generically highly diverse clade including Raphanus sativus, and it shows affinities to various Brassica species.

Brassiceae, Brassicaceae, flora of the Sahara, Morocco, new genus, Zahora ait-atta

The tribe Brassiceae is among the most complex monophyletic lineages within Brassicaceae. The tribe underwent an early genome triplication (Lysak et al. 2005) affecting subsequent diversification (Arias et al. 2014) and gave rise to approximately 50 genera comprising 250 species (Huang et al. 2019; Koch et al. 2017). The entire tribe started to evolve about 23 million years ago, and centers of origin and diversity are the entire circum-Mediterranean region (Arias et al. 2014). Phylogenic analyses have identified eight clades within the tribe Brassiceae: Vella L., Zilla Forssk., Henophyton Coss. & Durieu, Crambe L., Cakile Mill., Savignya DC., “Nigra”, and “Oleracea” (Arias and Pires 2012), but several genera remain poly- and paraphyletic such as Brassica L. and Diplotaxis DC. (Arias and Pires 2012), Raphanus L. (Ziffer-Berger et al. 2014), or Sinapis L. (Arias and Pires 2012). The tribe Brassiceae is not only characterized by an ancient triploidization, but extensive hybridization and reticulate evolution may have been involved while forming numerous polyploids. The actual amount of polyploids in the entire tribe is of about 28%; and approximately 43% of the species are monocarpic (Hohmann et al. 2015), which may coincide with arid and high-temperature environments preferred by numerous species of the tribe.

The new taxon was (re)discovered in 2015 by Claude Lemmel at isolated stands near the national road between Merzouga and Taous (Morocco) close to the border with Algeria. Since then the species has been continuously monitored by the second author and has been found at various places in that region. It is likely that in February 1951 Ph. Guinet and Ch. Sauvage might have noticed the same plant species near Tafilalet, but the plants were in bloom only and fruits were missing, therefore the botanists listed the species as Brassica spec. and putatively unknown (Guinet and Sauvage 1954). Originally, we thought that this plant could be of recent hybrid origin, because in this area there are many wild or cultivated cabbage-related species belonging to the genera Brassica, Eremophyton Bég., Eruca Mill., Moricandia DC., Erucaria Mill. and Diplotaxis, and hybrids are often observed between genera of Brassiceae (reviewed in Warwick et al. 2009). However, the species also occurs at sites with no other Brassicaeae nearby; and Aït-atta nomadic herders, who roam with their sheep and goats in this area, told us that they knew this plant as «Zizaou n’oudad» meaning «Barbary-sheep’s cabbage» and that it grows more or less frequently in the region in some oueds (a stream-bed that remains dry except during the rainy season) depending on local and seasonal rainfall. This field evidence encouraged us to analyze in greater detail this taxon unknown to science before.

Morphological characters of the new species do not match any known generic circumscription within Brassicaceae; although the new taxon combines characters, which are typically found in members of tribe Brassiceae. Therefore, we also analyzed chromosome number and genome size to compare results with known karyotypes, and we obtained DNA sequence information for phylogenetic placement analysis and phylogenetic reconstructions.

Morphological observations and measurements of the new species were carried out based on living plant material, either from the wild or cultivated at Heidelberg Botanical Garden (HEID), as well as prepared voucher specimens. Characters were measured using a dissecting microscope. Seeds were collected in the wild from the type locality (Meknés-Tafilalet/Drâa-Tafilalet: Border region with Algeria. Near Errachidia. Oued Bou-Ibourine), and grown for subsequent analysis of chromosome number (root tips) and genome size (leaf material) following protocols provided in detail with earlier studies (Hohmann et al. 2015).

Molecular analysis following the procedure of (i) DNA extraction, (ii) PCR amplification of nuclear encoded ribosomal DNA (ITS1-ITS2 region), and (iii) direct sequencing of the PCR product as it has been described earlier in detail (Karl and Koch 2013). For DNA extraction we used leaf material from the herein presented holotype.

ITS sequence information was added to a tribal-wide alignment of Brassiceae (Huang et al. 2019) and analyzed using maximum-likelihood inferences (Stamatakis 2014) with the same settings as described earlier (Huang et al. 2019). In total 193 taxa from tribe Brassiceae plus two additional outgroups have been combined with the new ITS sequence for phylogenetic analysis. The sampling, therefore, covers approximately 77% of all known species. The entire alignment is presented with Suppl. material 2. Initial phylogenetic tests have been conducted using the phylogenetic placement tool for Brassicaceae as implemented in BrassiBase (https://brassibase.cos.uni-heidelberg.de/; Koch et al. 2012).

In addition, plastid DNA markers trnL intron and trnL-trnF intergenic spacer were amplified and sequenced (Koch et al. 2017), and results have been subsequently used for megaBLAST searches (high similarity) in GenBank to identify taxa with similar plastid (maternal) genotypes.

Temporal inferences about divergence time and age of the new species have been analyzed using BEAST (Drummond et al. 2012). Here we also used the same setting as presented earlier (Huang et al. 2019) analyzing the ITS alignment (Suppl. material 2), and details can be found with Huang et al. (2019).

In Brassicaceae there are hardly any apomorphic characters (or character states) defining genera sufficiently. This resulted in numerous poly- and paraphyletic taxa (Al-Shehbaz 2012). Therefore, we used our results from phylogenetic analysis (see below) to identify those genera «widely» associated phylogenetically with the new taxon. In total, there are 13 genera: Brassica, Cordylocarpus Desf., Crambella Maire, Diplotaxis, Erucastrum Gaertn., Guiraoa Coss., Hirschfeldia Moench, Raffenaldia Godr., Rapistrum Crantz, Sinapis, Trachystoma O.E.Schulz, Morisia J.Gray and Otocarpus Durieu. For entire Brassicaceae we developed a morphomatrix scoring 37 characters to deliminate genera (« morphology tool » in BrassiBase; https://brassibase.cos.uni-heidelberg.de/; Koch et al. 2018). This matrix builds upon an interactive key presented earlier by Ihsan Al-Shehbaz and now integrated into BrassiBase. We used the character matrix to identify corresponding genus-level discriminative characters (Suppl. material 1). The combination of 37 characters allow to separate the genus from 12 out of the 14 genera with the following character states found in Zahora: (1) annual, (2) herb, (3) glandular hairs are absent, (4) trichomes are simple, (5) stem thorns are absent, (6) basal leaves are rosette-forming, (7) leaf margin is pinnately lobed, (8) stem leaves are present, and (9) petiolate, (10) stem leaves are entire to sinuate, (11) leaf thorns are absent, (12) raceme is ebracteate, (13) petals are distinctly longer than sepals (14) petals are (pale) yellow, (15) petal are wide-shaped in upper part, (16) petal apex is obtuse, (17) petal margin is entire, (18) sepals are erect, and (19) free, (20) stamen number is 6, (21) lower part of filaments and petal claws are without any structure, and (22) filaments are free, (23) flower symmetry is actinomorphic, (24) fruit type is a silique, (25) fruits are terete, (26) fruit wall is thin and leathery, (27) fruit is dehiscent (at least in proximal part), (28) gynophore in fruit is absent, (29) septum in mature fruits is complete, (30) stigma is entire, (31) fruit appendices are missing, (32) there are more than 20 seeds per fruit, (33) seeds are arranged biseriate in middle of fruit, (34) seed wing is missing, (35) cotyledons are conduplicate, (36) seed mucilage is present, and (37) fruit orientation is spreading. The combination of 37 character states does not distinguish between Brassica and Diplotaxis, two polyphyletic genera. And the new genus shows the same general generic morphotype, too.

However, Zahora is different from both, Brassica and Diplotaxis, because of its peculiar fruit type. Occurrence of heteroarthrocarpic fruits with seeded beak have been described for Brassica (Gómez-Campo and Prakash 1999), but Diplotaxis shows some trend only towards this feature.

But in none of these cases functional seeds are constantly developed in the distal part. Neither Brassica nor Diplotaxis have been shown to contain species with heteroarthrocarpic fruits with disarticulation of the joint (Hall et al. 2011). In various heteroarthrocarpic species, a joint forms a novel separation layer such that the distal segment may separate and is dispersed independently of seeds from the proximal segment, a phenomenon referred to as disarticulation. Zahora ait-atta variant of heteroarthrocarpy can be defined as “proximal segment dehiscent with disarticulation of the indehiscent distal part”.

Zahora ait-atta is described as a new species of a new monotypic genus. Zahora shows a peculiar fruit feature, namely heteroarthrocarpic fruits, and the species might mediate evolutionary between Core Oleracea clade (e.g. Brassica oleracea, Brassica napus) and Raphanus sativus and related genera. Both represent important crop plant groups with seeds playing an enormous agronomical role. The diploid new species might, therefore, serve as important germplasm reservoir to study traits and characters in a number of Brassiceae crop plants.

We thank Peter Sack for curation of germplasm, support in the herbarium and cytogenetic analyses. Bärbel Schwarz is curating our living collections and we are grateful for her excellent work. We thank the local ait-atta herders for their hospitality, sharing with us their bread, tea and their knowledge of desert flora, Zahora Attioui for her participation in field work and Dr Abdelmonaim Homrani-Bakali, rangeland ecologist at INRA Morocco for sharing a field trip and his data. Valuable comments from Ihsan Al-Shehbaz and an anonymous reviewer are greatly acknowledged.