Myriophyllum rubricaule sp. nov., a M. aquaticum look-alike only known in cultivation

. A confusingly labeled water-milfoil of obscure status, known only in cultivation, is here formally described as a new species, Myriophyllum rubricaule Valk. & Duist. sp. nov. This species has fully replaced M. aquaticum in the horticultural trade in Europe since the addition of M. aquaticum to the list of invasive alien species of Union concern (EU regulation no. 1143/2014) in 2016. This manuscript provides a morphological description of M. rubricaule sp. nov., and its distinction from M. aquaticum is further supported by molecular data (chloroplast and nuclear loci).


Introduction
Incorrect labeling of plants in trade and misidentification are widespread and may be caused by negligence or willful disrespect of regulations (Brunel 2009;Thum et al. 2012;Verbrugge et al. 2014;Hulme et al. 2018). Mislabeling may consist of simple misspelling of names, or treating a variety as a true species, using synonyms or just preferring a name that sounds nice or a name that customers are familiar with. The latter case actually equals misidentification of the plant in trade.
In addition to the above mentioned specimens used for the present study, all Myriophyllum collections from South America at Naturalis (AMS, L, U, WAG) were consulted as well as most relevant revisions and new taxonomical publications on the genus Myriophyllum to ascertain the novelty of the new species (Meijden & Caspers 1971;Orchard 1980Orchard , 1981Orchard , 1986Orchard & Kasselmann 1992;Wang et al. 2002;Yu et al. 2002;Wang & Yu 2007).
Morphological description is based on both living and herbarium material and follows Meijden & Caspers (1971) and Crow & Ritter (1999) for characters considered important in species recognition. No plant material was consulted of the Myriophyllum sp. 'red 1' and 'red 2' that were used in the molecular analysis as they were reported to be sterile (Moody & Les 2010).
An annotated collection of M. aquaticum (Pedersen 3977 in Orchard 1981) from Argentina, as well as a diminutive collection of M. aquaticum from Brazil (Lindeman 8405) were included to ascertain the molecular novelty of the new species. A Laurembergia tetrandra (Schott) Kanitz collection (van Proosdij 33) was used as an outgroup (Appendix 1).

DNA extractions
Genomic DNA was isolated from approximately 100 mg of plant material with the DNeasy plant mini kit (Qiagen, Venlo, the Netherlands) using the TissueLyser procedure and eluted with 50 μl of prewarmed (65ºC) AE buffer. DNA was stored at -20°C until use.
The cycle conditions for rbcL and trnH-psbA loci were as follows: 5 min at 95ºC, followed by 5 cycles of 30 s at 94ºC, 30 s at 45ºC, 30 s at 72ºC and 35 cycles of 30 s at 94ºC, 30 s at 50ºC, 30 s at 72ºC and a final extension for 10 min at 72ºC. For the ITS locus we used the following conditions: 5 min at 95ºC, followed by 40 cycles of 30 s at 94ºC, 30 s at 52ºC, 100 s at 72ºC and a final extension for 10 min at 72ºC.
The obtained PCR products were purified using the QIAquick PCR Purification Kit (Qiagen, Venlo, the Netherlands) preceding bidirectional cycle sequencing with the BigDye Terminator ver. 1.1 Cycle Sequencing Kit (Thermo Fisher Scientific, Bleiswijk, the Netherlands) using the aforementioned amplification primers as sequencing primers in separate reactions according to the manufacturer's instructions. Cycle sequence products were purified with the DyeEx 2.0 Spin Kit (Qiagen,Venlo, the Netherlands) and subsequently sequenced using a 3500 Genetic Analyzer (Thermo Fisher Scientific). Consensus sequences were generated from an assembly with trace files from both Sanger sequencing runs in Geneious Prime® 2021.1.1 (Biomatters Auckland, New Zealand). Amplification primer sequences were trimmed in the assembly, and when needed, additional trimming was performed to obtain high-quality (PHRED > 30) consensus sequences.
Samples that were sequenced with the NovaSeq 6000 were processed using the NEBNext ® Ultra II FS DNA module and the NEBNext ® Ultra II Ligation module. Fragmentation, A-tailing and ligation of sequencing adapters and PCR using NEBNext ® Ultra II Q5 master mix of the resulting product was performed according to the procedure described in the NEBNext Ultra II FS DNA module and NEBNext Ultra II Ligation module instruction manual. HiSeq 2500 samples were processed using the NEBNext® Ultra DNA Library Prep kit for Illumina (New England Biolabs, Ipswich, USA). Fragmentation of the DNA using the Covaris, ligation of the sequencing adapters and PCR amplification of the resulting product were performed according to the procedure described in the NEBNext Ultra DNA Library Prep kit for Illumina instruction manual.
After preparation, the quality and yield for all samples was measured with the Fragment Analyzer (Agilent Technologies, USA). The size of the resulting products was consistent with the expected size of approximately 500-700 bp.

Phylogenetic analyses
Sequences were aligned using MUSCLE ver. 3.8.425 (Edgar 2004) within Geneious Prime® 2021.1.1 (Biomatters Auckland, New Zealand) with standard settings. We performed model testing on the alignments to select the most appropriate model for constructing a maximum likelihood tree using the model selection tool in MEGA X with standard settings (Kumar et al. 2018). Selected models were used to generate the phylogenetic trees in MEGA X.

Data accessibility
The annotated DNA sequences are available from the NCBI GenBank database under the following accession numbers ( Sequences are also available at https://qbank.eppo.int/plants/.

Etymology
The species epithet is based on the purplish red color of the stem.
Amphibic herb with pinnate leaves in whorls of 4 or 5, only known from female plants and (escapes from) horticulture. Differs from M. aquaticum in its generally more modest dimensions, the stems being purplish red, the emergent leaves being green to bluish green and not glaucous (i.e., leaves without waxy coating) and the female flowers being pinkish (Fig. 1).

Distribution
Origin unknown. Known from cultivation (Netherlands, Belgium) and as escapes from cultivation since 2018 in the Netherlands (Hoogeveen, Steenwijk, Hattem, Klazienaveen) and Belgium (Maasmechelen, Houthalen-Helchteren, Brugge, Gent, Beauraing, Waimes; all confirmed from photographs at waarnemingen.be), and at least since 2019 in Hungary (Kács). It is unclear whether the observations in Belgium before 2018 (in retrospect the first observation dates from 2012; see waarnemingen.be records 70176185 and 95286205) refer to escapes or planted material.

Note
If grown in particularly nutrient rich and/or high light conditions plants develop much bigger submerged leaves (e.g., Valkenburg 3298 and 4116 WAGPD).

Similar species
Because both the submerged and emergent leaves are whorled, pectinate and not much different in length, and the inflorescence is emergent, the species nov. in having green stems only turning red when grown as a potted plant or in unfavorable conditions, but never purplish red, with leaves in whorls of 4-6, submerged leaves 25-45 mm long and green or red brown, emerged leaves 25-35 mm long and bluish green glaucous (i.e., with a waxy coating that can be rubbed of), and white flowers. Myriophyllum mattogrossense has green stems, small, globular, sessile glands on leaves and stems, submerged leaves 10-35 mm wide, and bisexual flowers (Crow & Ritter 1999).
While performing the DNA barcoding study of M. aquaticum (Ghahramanzadeh et al. 2013), we initially thought that the plants described here as a new species belonged to M. robustum. This is a species resembling M. aquaticum with similar emergent pectinate leaves and solitary axillary flowers. However, M. robustum is described as having hermaphrodite flowers whereas M. aquaticum is dioecious (Orchard 1980). Shortly after our paper was published, plants of M. robustum were received from New Zealand and were grown in a greenhouse. The cultivated plants had more robust erect emergent stems that were pink in the upper part, and subglaucous leaves that were oblong in outline with an acute tip. Surprisingly, the plants produced only female flowers in the greenhouse (Valkenburg 3739 WAGPD). Grown outdoors in a mesocosm in later years, the plants first produced female flowers, a row of hermaphrodite flowers and then, distally, male flowers (Valkenburg 3853 WAGPD).
The trnK3 and matK loci were extracted from the sequence data and subsequently concatenated, before tree generation. Similarly to ITS, the trnK3-matK sequences were identical for all samples of M. rubricaule sp. nov. (Fig. 3) Finally, we compared two other well-known chloroplast loci, rbcL and trnH-psbA, to identify more distinguishing molecular features between M. aquaticum and M. rubricaule sp. nov. We incorporated previously described sanger sequences (Ghahramanzadeh et al. 2013)

Discussion
During our study of the confusingly labeled water-milfoil in trade, we consulted the most relevant revisions and new taxonomical publications on the genus Myriophyllum (Meijden & Caspers 1971;Orchard 1980Orchard , 1981Orchard , 1986Orchard & Kasselmann 1992;Wang et al. 2002;Yu et al. 2002;Wang & Yu 2007). After attempting to identify the plants using the revised key for New Zealand (Orchard 1980), the two choices where M. aquaticum and M. robustum. Firstly, our taxon cannot be M. aquaticum because of the color of the leaves in the field, and secondly it is not M. robustum because the flowers are not hermaphroditic. The revision for South America (Orchard 1981) likewise does not yield a result as M. aquaticum has glaucous leaves and M. mattogrossense is disqualified by the plants being weak and having hermaphrodite flowers. In the revision for Australia (Orchard 1985), M. aquaticum is also the only realistic option. None of the species covered by the Myriophylllum treatment for Malesia (Meijden & Caspers 1971), nor recent publications for new species in Asia Yu et al. 2002;Wang & Yu 2007) match our material as all of those species have dimorphic leaves.
The new species fits in section Pectinatum M.L. Moody & Les (2010: 136 Species within this section have in common the emergent inflorescences and all leaves whorled and pectinate; the emergent leaves are not highly reduced.
The definition as such would also fit for M. robustum, that is placed in M. subsection Myriophyllum but is the exception in the subsection for having all pectinate emergent leaves that are not highly reduced in relation to submerged leaves.

Fig. 2.
Maximum likelihood tree based on ITS (partial 18S, ITS1, 5.8S, ITS2, partial 28S) sequences. The Tamura 3 parameter model was used (Tamura 1992). The tree with the highest log likelihood (-1398.71) is shown. Bootstrap values after 1000 replicates are expressed as percentages. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.4031)). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Laurembergia tetrandra AP 33 was used as an outgroup. Fig. 3. Maximum likelihood tree based on concatenated trnK3-matK sequences. This tree was inferred by using the Maximum Likelihood method and Tamura 3 parameter model (Tamura 1992). The tree with the highest log likelihood (-2853.86) is shown. Bootstrap values after 1000 replicates are expressed as percentages. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Laurembergia tetrandra AP 33 was used as an outgroup.
ITS, and the chloroplast locus trnH-psbA demonstrate that M. aquaticum markedly differs from M. rubricaule sp. nov. (Fig. 2 and Supp. file 3). The rbcL locus is not suitable to distinguish between the species, little variation was observed even with the more distantly related L. tetrandra (Supp. file 4). Intriguingly, we observed that the concatenated loci trnK3-matK of M. aquaticum JL 8405 were identical to M. rubricaule. However, only two SNPs were observed in comparison with the other M. aquaticum specimens (Fig. 3). Furthermore, stark genetic differences between M. aquaticum JL 8405 and M. rubricaule were found in comparison with the other loci.
Moody & Les (2010)  In addition to species in trade of unknown origin as further mentioned by Thum et al. (2012) incorrect labeling of plants is widespread and may be caused by negligence or willful disrespect of regulations.
What we have noticed so far with respect to species of Myriophyllum in trade in the Netherlands is that, in general, negligence of proper labeling causes confusion in the proper identity of the species in trade (Van Valkenburg & Boer 2015;. Concerning the new species described, the only mislabeling observed was M. proserpinacoides Gillies ex Hook. & Arn. (Valkenburg 3314) a name that was also used for M. aquaticum in trade.