Comparative genomics reveals the molecular basis of the synthesis of triterpenoid saponin and distinct flower morphology in Sapindaceae species

Sapindaceae, also named as soapberry family, is a large family of plants of great morphological diversities. The majority of them are monoecious, producing separate male and female flowers on the same plant. Some have petal-less flowers, some bear fruits with fleshy arils, and some are great source of certain secondary metabolites, like saponins. With the rapid development of next generation sequencing technologies, a number of Sapindaceae species have been sequenced, providing great resources for research from the perspective of comparative genomics analysis. Here, we sequenced and assembled a high-quality genome of balloon-vine (Cardiospermum halicacabum L.), the first liana plant with a full-genome sequenced in Sapindaceae, with the smallest genome size reported so far, low heterozygosity and repeatability, conferring it an ideal model plant for functional genomics research in Sapindaceae. In combination with other six publicly available Sapindaceae genomes, we reconstructed the Ancestral Sapindaceae karyotype (ASK) consisting of 18 protochromosomes and main chromosomal diversification fusion events shared or unique to these species. We characterized a Sapindale-specific biosynthetic gene cluster likely responsible for the biosynthesis of triterpenoid saponins in yellowhorn, soapberry and litchi. Intriguingly, we found that the absence of petals in litchi flowers is caused by the repressed expression of AP1 and AP3-1, two MADS-box genes essential for petal identity specification. Species-specific insertion of transposon elements in the promoter region of these two genes contributed to their expression repression via RNA-directed DNA methylation. Our work provides new insight and a great demonstration of using comparative genomics to dissect the molecular basis of characteristic trait development in Sapindaceae.无患子科（Sapindaceae），又称肥皂草科，是一类形态多样性极为丰富的大型植物科。该科绝大多数物种为雌雄同株，同一植株可产生单性雄花与雌花。部分物种的花朵无花瓣，部分物种的果实带有肉质假种皮，还有部分物种是皂苷等次生代谢产物的优质来源。随着下一代测序技术的快速发展，已有多种无患子科物种完成全基因组测序，为比较基因组学研究提供了宝贵资源。本研究对倒地铃（Cardiospermum halicacabum L.）进行了全基因组测序与组装，该物种是无患子科中首个完成全基因组测序的藤本植物，其基因组为目前已报道的无患子科物种中最小，且具有低杂合度与低重复序列占比的特征，是开展无患子科功能基因组学研究的理想模式植物。结合另外6个已公开的无患子科物种基因组数据，本研究重构了由18条祖先染色体构成的无患子科祖先核型（Ancestral Sapindaceae karyotype, ASK），并解析了这些物种共有的或物种特有的主要染色体进化融合事件。本研究鉴定到一个无患子目特异性的生物合成基因簇，该基因簇可能参与文冠果、无患子与荔枝的三萜皂苷生物合成。有趣的是，本研究发现荔枝花朵无花瓣的表型，是由AP1与AP3-1两个调控花瓣特征发育的关键MADS-box基因（MADS-box）的表达受抑制所导致的。这两个基因的启动子区域发生了物种特异性的转座子插入，通过RNA介导的DNA甲基化途径抑制了其表达。本研究为利用比较基因组学解析无患子科物种特征性状发育的分子机制提供了全新视角与优质范例。