A revision of Hypotrachyna subgenus Parmelinopsis (Parmeliaceae) in eastern North America

A taxonomic revision of the Hypotrachyna subgenus Parmelinopsis in eastern North America is presented based on molecular phylogenetic analyses of ITS and mtSSU data, extensive field observation and analyses of chemical and morphological data. Each species is described, illustrated with photographs, and the distribution in the region is mapped. An identification key is also presented. Eleven species are recognized: H. afrorevoluta, H. appalachensis, H. britannica, H. cryptochlora, H. horrescens, H. kauffmaniana, H. minarum, H. mcmulliniana, H. revoluta, H. showmanii and H. spumosa. Extensive discussion of prior studies is provided, particularly with respect to the delimitation of H. afrorevoluta and H. revoluta. Hypotrachyna kauffmaniana is described from the central and southern Appalachian Mountains and separated from H. afrorevoluta and H. revoluta by its ascending secondary lobes and pustulose soralia that are primarily confined to the secondary lobes. Hypotrachyna horrescens is shown to correspond to a taxon with narrow lobed, small thalli with ciliate isidia. Hypotrachyna mcmulliniana is described from material collected throughout southeastern North America that is chemically identical to H. horrescens but differs in having larger thalli and sparsely ciliate isidia. Hypotrachyna appalachensis is described to accommodate material previously referred to H. minarum but that differs in the production of 4,5-di-O-methylhiascic acid in high concentrations (vs. absent or present as a trace in H. minarum). Hypotrachyna britannica is reported for the first time from North America. Methods Field and herbarium study. – The genesis of this study derived from extensive fieldwork carried out by the authors in the southern Appalachian Mountains of eastern North America, a region long recognized as a center of diversity for the genus in North America north of Mexico (Dey 1978, Hale & Kurokawa 1964, Lendemer et al. 2017). Between 2010 and 2020 we visited numerous sites in the southern Appalachians in connection with multiple projects (e.g., Allen & Lendemer 2016, Lendemer & Allen 2015, Lendemer et al. 2017, Tripp et al. 2019). During this fieldwork, we collected many Hypotrachyna specimens and studied the species in the field. Additional species were studied and collected during field work in the Coastal Plain during this time (Lendemer et al. 2016). We made hundreds of Hypotrachyna subgenus Parmelinopsis collections, which we examined for this study together with the those already deposited in the herbarium of the New York Botanical Garden (NY) for a total of 2162 specimens. In addition to this material, we examined type specimens, sequenced vouchers, and selected general specimens from the following herbaria: KRAM, MAF, OSU, SP, US, WIS. Morphological and chemical study. – Morphological studies were conducted in the field and the laboratory. Thalli were observed in the field to determine growth direction, lobe arrangement, and overall shape (e.g., rosette vs. lobes draping down, adnation to substrate), along with habitat and ecology. Herbarium specimens were then used to measure vegetative and reproductive structures with microscopy. Specimens were studied dry with an Olympus SZ-STB dissecting microscope. Microscopic morphology and anatomy were studied with an Olympus BX53 compound microscope and sections or preparations made by hand with a razor blade that were mounted in water. Chemistry was initially studied with standard spot tests (K, C, KC, P, UV) following Brodo et al. (2001). All specimens were also subjected to thin layer chromatography (TLC) using acetone extracts of thallus fragments run in solvent C following Culberson and Kristinsson (1970) as modified by Lendemer (2011) for use with a peanut butter jar and with the solvent ratio of 200:30 toluene:glacial acetic acid. The arrangement of substances in the species studied here, as detected with the above methods, is illustrated in Figure 1. DNA extractions, PCR, and sequencing. – Molecular data generation followed Muscavitch et al. (2017) for sequencing of mtSSU and ITS. Thallus fragments were taken from selected specimens and secondary metabolites were first extracted with an acetone wash that was used for TLC as outlined above. DNA was extracted with methods and reagents modified from the Qiagen DNeasy Plant Mini Kit. ITS4/5 (White et al. 1990) and mtSSU1/3R (Zoller et al. 1999) primers were used for PCR amplification of target regions. Sanger sequencing of PCR products was conducted at Macrogen USA. Sequencher 5.2.4 (GeneCodes, Ann Arbor, Michigan) was used to assemble and edit the resulting sequences. Taxon sampling and molecular dataset assembly (mtSSU). – To assess the relationships between clades within Hypotrachyna, and the placement of North American taxa in those clades, we assembled an alignment of mtSSU sequences guided by the results of Crespo et al. (2010) and Divakar et al. (2010). All sequences tagged with Cetrariastrum, Everniastrum, Hypotrachyna and Parmelinopsis were downloaded from GenBank (06 November 2016). Taxa belonging to the Remototrachyna clade as defined by Divakar et al. (2010) were pruned, as were taxa assigned to other genera (e.g., Bulbothrix). A selection of Myelochroa mtSSU sequences were also downloaded for use as an outgroup following Crespo et al. (2010). One-hundred and eleven newly generated sequences were also included (see Table 1). The sequences were assembled into a NEXUS file using Mesquite 3.04 (Maddison & Maddison 2015) which was then exported as a FASTA file and aligned using the MAFFT online interface. The resulting alignment from MAFFT was then downloaded and checked manually with adjustments made in Mesquite. During this initial alignment check, DQ287834 (H. rockii) was removed because it contained a large section of missing data. AY611131 (H. immaculata), AY611168 (P. minarum), and DQ287793 (C. ecuadorense) were removed because nearly one third of the sequences were lacking compared to the rest of the alignment. After pruning the above sequences the alignment was again subjected to auto-alignment in MAFFT with subsequent manual adjustment in Mesquite. Ambiguously aligned regions were then defined as part of an exclusion set in Mesquite. Taxon sampling and dataset assembly (ITS). – To examine the relationships between members of Parmelinopsis as recovered by Divakar et al. (2010) we assembled a dataset of ITS sequences. All sequences of Hypotrachyna, Parmelinopsis and Remototrachyna were downloaded from GenBank (06 November 2016). The sequences were assembled into a NEXUS file, together with all newly generated sequences (see Table 1) and aligned in the same manner as was outlined for the mtSSU dataset above. After the initial stage of alignment and checking, KM250245 (H. minarum) was removed because it was highly divergent from the other sequences and did not BLAST to other Hypotrachyna. Following a subsequent cycle of alignment and checking, ambiguously aligned regions were excluded and the dataset was subjected to a rapid maximum likelihood (ML) analysis in RAxML 8.2.10 (Stamatakis 2006) with 500 bootstrap replicates and implementing the most complex nucleotide substation model available in the program (GTRGAMMA). The results were visualized in FigTree (Rambaut 2017). Using the results of the initial ML analysis as a guide, we pruned the large ITS dataset into a smaller dataset corresponding to Hypotrachyna subgenus Parmelinopsis, selected members of Hypotrachyna subgenus Hypotrachyna, and selected members of Hypotrachyna subgenus Everniastrum for use as an outgroup following Divakar et al. (2013). The alignment was then subjected to another round of auto-alignment and manual adjustment, and the ambiguously aligned regions were defined as part of an exclusion set. Molecular phylogenetic analyses. – The mtSSU dataset and the final, smaller Hypotrachyna subg. Parmelinopsis ITS dataset were subjected to a rapid ML analysis in RAxML with 500 bootstrap replicates and implementing GTRGAMMA as the nucleotide substitution model. That model was selected because it is the most complex available in RAxML. The results were visualized in FigTree (Rambaut 2017).本研究基于内转录间隔区（Internal Transcribed Spacer, ITS）与线粒体小亚基核糖体DNA（mitochondrial small subunit, mtSSU）数据的分子系统发育分析、大量野外观察以及化学与形态学数据分析，对北美东部的树花属（Hypotrachyna）Parmelinopsis亚属开展了分类学修订。本研究对每个物种进行了详细描述，辅以照片图示，并绘制了该类群在研究区域内的分布地图，同时提供了物种鉴定检索表。本研究共确认11个物种：非洲卷叶树花（H. afrorevoluta）、阿巴拉契亚树花（H. appalachensis）、不列颠树花（H. britannica）、隐绿树花（H. cryptochlora）、糙枝树花（H. horrescens）、考夫曼树花（H. kauffmaniana）、米纳姆树花（H. minarum）、麦克马林树花（H. mcmulliniana）、卷叶树花（H. revoluta）、肖曼树花（H. showmanii）以及泡沫树花（H. spumosa）。本研究对既往相关研究进行了详尽讨论，尤其聚焦于非洲卷叶树花与卷叶树花的分类界定问题。 考夫曼树花的模式标本采自阿巴拉契亚山脉中南部，其与非洲卷叶树花及卷叶树花的区别在于：具有向上伸展的次级裂片，且主要生于次级裂片上的疱状粉芽堆（soralia）。糙枝树花被证实对应一类具有狭窄裂片、小型地衣体（thallus）且具缘毛裂芽（isidia）的类群。麦克马林树花的研究材料采自北美东南部全域，其化学特征与糙枝树花一致，但地衣体更大，且裂芽的缘毛较为稀疏。阿巴拉契亚树花的建立是为了容纳此前被归为米纳姆树花的材料，二者的区别在于前者可高浓度产生4,5-二-O-甲基海桑酸，而米纳姆树花中该物质要么不存在，要么仅以痕量存在。不列颠树花为北美地区首次报道的新记录物种。 ## 研究方法 ### 野外与标本馆研究 本研究的初衷源于作者在北美东部阿巴拉契亚山脉南部开展的大量野外考察工作——该区域长期以来被认为是墨西哥以北北美地区该属的多样性分布中心（Dey 1978；Hale & Kurokawa 1964；Lendemer et al. 2017）。2010年至2020年间，作者依托多项研究项目对阿巴拉契亚山脉南部的众多样点进行了考察（如Allen & Lendemer 2016；Lendemer & Allen 2015；Lendemer et al. 2017；Tripp et al. 2019）。本次野外考察期间，我们采集了大量树花属标本，并在野外对该属物种进行了研究。同期，我们还在沿海平原开展野外工作时对其他物种进行了采集与研究（Lendemer et al. 2016）。本研究共采集了数百份树花属Parmelinopsis亚属的标本，结合纽约植物园标本馆（NY）已馆藏的标本，总计2162份标本供本研究检视。除此之外，我们还检视了以下标本馆的模式标本、测序凭证标本及代表性普通标本：KRAM、MAF、OSU、SP、US、WIS。 ### 形态学与化学研究 形态学研究分别在野外与实验室开展。野外观察地衣体的生长方向、裂片排列方式、整体形态（如莲座状 vs. 裂片下垂、贴生基质情况）及其生境与生态特征。随后利用显微镜对馆藏标本的营养与繁殖结构进行测量。标本检视采用Olympus SZ-STB体视显微镜对干燥标本进行观察。显微形态与解剖结构则通过Olympus BX53正置显微镜进行观察，切片或临时装片由手工用刀片制作，以水封片。化学分析首先参照Brodo等（2001）的方法，采用标准斑点试验（K、C、KC、P、UV试剂）进行初步检测。所有标本还采用薄层色谱法（TLC）进行分析：取地衣体碎片的丙酮提取物，参照Culberson与Kristinsson（1970）的方法，并经Lendemer（2011）针对以花生酱罐作为展开容器的体系优化，采用甲苯:冰乙酸=200:30的溶剂系统展开，溶剂体系记为C。本研究中通过上述方法检测到的物种内物质分布情况如图1所示。 ### DNA提取、PCR扩增与测序 分子数据生成参照Muscavitch等（2017）的方法进行mtSSU与ITS片段的测序。从选定的标本上取地衣体碎片，先用丙酮萃取次级代谢产物，该萃取液可用于前述的薄层色谱分析。DNA提取采用改良自Qiagen DNeasy植物微量提取试剂盒的方法与试剂。采用ITS4/5（White et al. 1990）与mtSSU1/3R（Zoller et al. 1999）引物对目标区域进行PCR扩增。PCR产物的桑格测序由Macrogen美国分部完成。使用Sequencher 5.2.4（GeneCodes，密歇根州安阿伯）对测序结果进行序列拼接与编辑。 ### 类群取样与分子数据集组装（mtSSU） 为评估树花属内各演化支的亲缘关系，以及北美类群在这些演化支中的系统位置，本研究参照Crespo等（2010）与Divakar等（2010）的研究结果，组装了mtSSU序列比对矩阵。2016年11月6日，我们从GenBank下载了所有标注为Cetrariastrum、Everniastrum、Hypotrachyna及Parmelinopsis的序列。参照Divakar等（2010）的定义，我们移除了Remototrachyna演化支类群以及其他归入其他属的类群（如Bulbothrix）。参照Crespo等（2010）的方法，我们还下载了部分Myelochroa的mtSSU序列作为外类群。本研究还纳入了111条新生成的序列（详见表1）。 序列使用Mesquite 3.04（Maddison & Maddison 2015）组装为NEXUS格式文件，随后导出为FASTA格式文件，通过MAFFT在线界面进行序列比对。将MAFFT生成的比对结果下载后，使用Mesquite进行手动检查与调整。在初始比对检查阶段，我们移除了DQ287834（H. rockii），因其存在大量缺失数据；移除了AY611131（H. immaculata）、AY611168（P. minarum）及DQ287793（C. ecuadorense），因其与其余比对序列相比，近三分之一的位点存在缺失。移除上述序列后，再次使用MAFFT进行自动比对，随后在Mesquite中进行手动调整。最后在Mesquite中将比对模糊的区域定义为排除集。 ### 类群取样与数据集组装（ITS） 为验证Divakar等（2010）所恢复的Parmelinopsis亚属类群间的亲缘关系，本研究组装了ITS序列数据集。2016年11月6日，我们从GenBank下载了所有Hypotrachyna、Parmelinopsis及Remototrachyna的序列。将所有新生成序列（详见表1）与下载序列一同组装为NEXUS格式文件，并参照前述mtSSU数据集的比对流程进行序列比对。初始比对与检查阶段结束后，移除了KM250245（H. minarum），因其与其他序列差异极大，且BLAST比对未匹配到其他树花属物种。经过新一轮的比对与检查后，排除比对模糊的区域，随后使用RAxML 8.2.10（Stamatakis 2006）对数据集进行快速最大似然（ML）分析，设置500次自展重复，并采用程序中最复杂的核苷酸替换模型GTRGAMMA。分析结果使用FigTree（Rambaut 2017）进行可视化。以初始ML分析结果为指导，参照Divakar等（2013）的方法，我们将大型ITS数据集修剪为小型数据集，该数据集包含树花属Parmelinopsis亚属类群、树花属树花亚属的选定类群，以及树花属Everniastrum亚属的选定类群作为外类群。随后对该比对矩阵进行新一轮自动比对与手动调整，并将比对模糊的区域定义为排除集。 ### 分子系统发育分析 对mtSSU数据集以及最终修剪后的树花属Parmelinopsis亚属ITS小型数据集，使用RAxML进行快速最大似然分析，设置500次自展重复，采用GTRGAMMA作为核苷酸替换模型。选择该模型是因为其为RAxML中最复杂的核苷酸替换模型。分析结果使用FigTree（Rambaut 2017）进行可视化。