CRISPR/Cas（Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR associated）系统是一种高效的基因编辑系统，也是现阶段应用最广泛的基因组编辑工具，但是由于在基因编辑过程中存在DNA双链断裂导致的基因编辑产物异质性以及染色体不稳定性等问题，研究人员将产生DNA单链断裂的Cas9切口酶与脱氨酶或者逆转录酶融合，开发出碱基编辑器（Base editor，BE）和先导编辑器（Prime editor，PE），其中碱基编辑器主要包括胞嘧啶碱基编辑器（Cytosine base editor，CBE）和腺嘌呤碱基编辑器（Adenine base editor，ABE）。ABE能在编辑窗口内高效地引起A•T到G•C的碱基转换，理论上可以修复近一半的人类致病性点突变，在基因治疗中展现出极佳的应用潜力。但ABE的应用依然受到靶向编辑效率不理想、编辑范围有限以及脱靶编辑等因素的影响。为了进一步拓展腺嘌呤碱基编辑工具的应用范围，本研究在ABEmax-SpRY的基础上通过基因工程改造，获得一种几乎不受PAM限制、编辑活性高但脱靶活性低的新型腺嘌呤碱基编辑器，并对其在疾病相关位点中的编辑能力进行了初步探索。

本研究首先在ABEmax-SpRY的基础上以腺嘌呤脱氨酶替换的方式得到靶向编辑活性更高的8e-SpRY，接着对8e-SpRY通过引入突变或者Cas蛋白嵌入方式得到四种8e-SpRY突变体，其中将TadA-8e插入到SpRY-Cas9中间的CE-8e-SpRY的编辑活性略高于8e-SpRY，将V106W引入脱氨酶结构域的V106W-SpRY的编辑活性略低于8e-SpRY，其余两种突变体的编辑活性均显著低于8e-SpRY。进一步的编辑活性分析结果显示8e-SpRY和其两种编辑活性相当的突变体，在NRN PAM（R代表A或G）位点上的编辑活性高于NYN（Y代表C或T），与SpRY核酸酶一致；并且8e-SpRY和其突变体的编辑窗口均是3-10位，比传统腺嘌呤碱基编辑器4-7位的编辑窗口更宽，因此能使基因组上更大范围的碱基得到编辑。

本研究接下来评估SpRY腺嘌呤碱基编辑器的脱靶编辑情况。结果显示，在RNA脱靶活性方面，CE-8e-SpRY的RNA脱靶编辑水平最低；在sgRNA依赖性DNA脱靶编辑方面，CE-8e-SpRY也产生低于8e-SpRY的脱靶活性；在sgRNA非依赖性DNA脱靶方面，CE-8e-SpRY仅产生略高于对照组的随机脱靶活性。因此，CE-8e-SpRY是一种兼顾靶向编辑效率高和脱靶编辑活性低的新型腺嘌呤碱基编辑器。

本研究进一步检测CE-8e-SpRY在疾病相关位点中的编辑能力。结果显示，CE-8e-SpRY可以编辑传统ABE编辑窗口之外的疾病位点，还可以通过sgRNA优化，实现疾病位点的精确编辑、最高效率编辑或者特异性编辑等需求。

最后为了进行有效的体内编辑，本研究对双AAV递送CE-8e-SpRY（Split-CE）的系统进行优化，得到编辑活性较高的Split573-CE。本研究还针对肝脏细胞的特异性编辑进行了启动子优化，细胞中的检测结果显示P3启动子可能具有更好的肝脏特异性。

综上所述，本研究通过基因工程改造得到一种靶向编辑活性高但脱靶活性低的新型腺嘌呤碱基编辑工具CE-8e-SpRY，CE-8e-SpRY不仅可以拓展腺嘌呤碱基编辑工具的编辑范围，还能通过sgRNA优化满足基因组位点编辑的不同需求。优化得到的AAV-CE-8e-SpRY为实现CE-8e-SpRY在体内肝脏细胞的高效编辑奠定了基础。

The CRISPR/Cas (Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR associated) system could induce efficient DNA editing at target sites and has become the most widely used genome editing tool in various organisms. However, undesired byproducts and chromosomal instability resulting from DNA double strand breaks (DSBs) generated by CRISPR-Cas, also raise many concerns. To overcome these issues, base editors (BEs) and prime editors (PEs) were developed by fusing deaminase enzymes or reverse transcriptase with Cas9 nickase. Two classes of base editors were mainly included: cytosine base editors (CBEs) and adenine base editors (ABEs). ABEs mediate efficient A•T-to-G•C conversions without creating DSBs and requiring the DNA donor template, and could theoretically correct nearly half of the pathogenic point mutations, showing great potential in the treatment of genetic diseases. However, the applications of ABEs are still hindered by undesired editing efficiency, limited editing scopes, and off-targeting effects. To further expand the application scopes of adenine base editing tools, we developed a new near PAMless ABE with high editing activity but low off-target activity based on ABEmax-SpRY, and preliminarily explored its potentials in editing disease-associated mutations.

We firstly engineered 8e-SpRY with high editing activities by replacing the adenine deaminase domain of ABEmax-SpRY with TadA-8e, then four 8e-SpRY variants were developed by introducing mutations or Cas-embedding strategy. Among them, CE-8e-SpRY, generated by inserting TadA-8e into the tolerant site of SpRY-Cas9, induced a little higher editing activities than 8e-SpRY; V106W-SpRY, engineered by installing V106W mutation in deaminase domain, mediated a little lower editing activities relative to 8e-SpRY; while the other two 8e-SpRY variants exhibited significantly reduced editing activities in comparison to 8e-SpRY. Further analysis revealed that three 8e-SpRY variants with comparable activities displayed higher editing activities at sites with NRN (R refers to A or G) than with NYN (Y refers to C or T) PAMs, in line with SpRY nuclease; moreover, 8e-SpRY variants performed efficient editing in a wider activity window (positions 3-10) than conventional ABEs (positions 4-7), enabling a wider range of nucleotides editable.

We next compared the off-target effects among developed SpRY-ABEs. In terms of RNA off-target effects, CE-8e-SpRY induced the lowest off-target editing at transcriptome level; when it concerns to sgRNA-dependent DNA off-targets, CE-8e-SpRY exhibited lower editing activities than 8e-SpRY; as to the sgRNA-independent DNA off-targets, CE-8e-SpRY only generated a little higher unguided DNA editing than SpRY-Cas9 nuclease. Therefore, CE-8e-SpRY was considered as the best SpRY-derived ABE with high on-target but low off-target editing activities.

Then, CE-8e-SpRY was adopted to edit the disease-relevant mutations in cells. Results suggested CE-8e-SpRY could induce efficient editing at sites which were outside the activity windows of conventional ABEs, moreover, specific editing requirements, like precise editing, or editing with the highest efficiency or specificity, could be achieved through sgRNA optimization.

Finally, to enable efficient editing in vivo, we optimized the dual AAV delivery system of CE-8e-SpRY (Split-CE) and found Split573-CE displayed higher editing activities. Besides, the promotor of Split573-CE was also optimized to improve the tissue specificity and P3 promotor might have a better liver specificity for Split-CE.

In summary, we developed a new SpRY-ABE with high on-target but low off-target editing activities, named CE-8e-SpRY. CE-8e-SpRY could expand the editing scope of adenine base editing tools and meet different requirements raised by target genome editing. Moreover, our optimized AAV-P3-CE-8e-SpRY system laid the foundation for efficient somatic editing in liver.